This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project
to make the world's books discoverable online.
It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books
are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover.
Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the
publisher to a library and finally to you.
Usage guidelines
Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to
prevent abuse by commercial parties, including placing technical restrictions on automated querying.
We also ask that you:
+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for
personal, non-commercial purposes.
+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the
use of public domain materials for these purposes and may be able to help.
+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find
additional materials through Google Book Search. Please do not remove it.
+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner
anywhere in the world. Copyright infringement liability can be quite severe.
About Google Book Search
Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web
at http : //books . google . com/|
Journal of the Chemical Society
Digitized by
Goog
Digitized by
Google
Digitized by
Google
Xll CONTENTS.
PAGE
CXLY. — ^The Action of Nitric Acid on Bromophenolic Com-
pounds. By William Kobbrtson, A.RG.S. . 1475
CXLVI. — Derivatives of Normal- and iso-Butyrylpyruvic Acids.
By Arthur Lapworth and A. C. Osborn Hann . . 1485
CXLVIl.— Optically Active Esters of j8-Ketonic and jS-Aldehydic
Acids. Part I. Menthyl Formylphenylacetate. By
Arthur Lapworth and A. C. Osborn Hann . .1491
CXLVIII.— Optically Active Esters of j8-Ketonic and j8-Alde-
hydic Acids. Part II. Menthyl Acetoacetate. By A.
Lapworth and A. C. Osborn Hann 1499
CXLIX. — The Mutarotation of Camphorquinonehydrazone and
Mechanism of Simple Desmotropic Change. By Arthur
Lapworth and A. C. Osborn Hann 1508
CL. — l^he Action of Sodamide and Acyl-substituted Sodamides
on Organic Esters. By Arthur Walsh Titherlby, D.Sc.,
Ph.D 1520
CLI. — 3 : 6-Dichloro-o-xylene and 3 : 5-Dichloro-o-phthalic Acid.
By Arthur William Crosslby and Hbnry Rondel Le
Sueur 1533
CLII. — Non-existence of the Gaseous Sulphide of Carbon
described by Deninger. By Edward John Russell and
Norman Smith 1538
CLIII. — Note on the Localisation of Phosphates in the Sugar
Cane. By Chas. Henry Graham Spranklino . 1543
CLIV. — Isometric Anhydrous Sulphates of the Form
W&O^jK^O^. By Frbdebio R. Mallet . .1546
CLY. — Asymmetric Optically Active Selenium Compounds and
the Sezavalency of Selenium and Sulphur, d- and ^Phenyl-
methylselenetine Salts. By William Jackson Pope, F.RS.,
and Allen Neville, B.Sc. 1552
CLVI. — ^Hydroxyoxamides. Part II. By Robert HowsonPiokard,
Charles Allen, William Audley Bowdler and William
Carter 1563
CLVII. — ^The Constituents of Commercial Chrysarobin. By
Hooper Albert Dickinson Jowbtt and Charles Etty
Potter . . ' 1576
CLVIII.— The tJonstituents of an Essential Oil of Rue. By
Frederick B. Power and Frederic H. Lees . 1585
CLIX. — Methyl j3-Methylhexyl Ketone. By Frederic Herbert
1594
CLX. — ^The Constitution of the Metallic Cyanides as Deduced
from their Synthetic Interactions. The Constitution of
Hydrogen Cyanide. By John Wade, D.Sc. . . 1596
Digitized by
Google
JOURNAL
OF
THE CHEMICAL SOCIETY.
TRANSACTIONS.
H. E. Armstrong, Ph.D., F.R.S.
E. DiVBRS, M.D., F.R.S.
Wtndham K Dunstan, M.A., F.R.S
H. J. H. Fbnton, M.A., F.KS.
P. F. Frantcland, LL.D., F.R.S.
H. McLtod, F.R.S.
Commttttt 0f ^ttblicHtion :
Sir William Ramsay, K.C.B., LL.D.,
F.R.S.
J. Emrrson Reynolds, ScD., F.R.S.
A. SooTT, D.Sc, F.R.S.
T. E. Thorpe, C.B., LL.D., F.KS.
W. A. TiLDBN, D.Sc, F.R.S.
(Ebiior:
W. P. Wynne, D.Sc, F.R.S.
Sixth' fiiiax :
A. J. Greenawav.
1902. Vol. LXXXI. Part I.
LONDON:
GURNEY & JACKSON, 1, PATERNOSTER ROW.
1902.
Digitized by
Google
Richard Clay A Soys, I^imitrd,
London b Dohoay
Digitized by
Google
JOURNAL
OF
THE CHEMICAL SOCIETY.
TRANSACTIONS.
I. — The Oxidation of Sulphurons Acid to Dithionic
Acid hy Metallic Oxides.
By H. C. H. Cabpbntkb.
Thebb are, so far as I have been able to ascertain, only two metallic
oxides which have been found to react with sulphorous acid and produce
dithionic acid. These oxides are manganese dioxide and hjdrated
ferric oxide.
The object of the present research was the investigation of reactions
which lead to the formation of dithionic acid, with particular precautions
as to the purity of the sulphurous acid and the various oxides used. In
this study, the author has been helped by the discovery that dithionic
acid is obtained when sulphurous acid reacts with (a) manganic hydr-
oxide, MnjOj(0H)2, and {b) cobaltic hydroxide, Co2(OH)(j.
k It is a priori possible in all cases for the ' available oxygen ' in the
metallic oxides to form either sulphuric acid or dithionic or both, as
the following equations show : —
(i) SOj + HaO + O^HjSO^. (ii) 2SO2 + HjO + 0 = HjSjOe
In the case of ferric hydroxide, Gelis (Ann. Ckim» Phy8 , 1862, [iii],
65, 222) states that the whole of the available oxygen goes first to
dithionic acid. On the other hand, it has always been found in the
investigations with manganese dioxide that a mixture of dithionic and
sulphuric acids results.
G^lis's experiments with ferric hydroxide have been repeated by the
author of this paper and his conclusion confirmed that a nearly theo-
retical yield of dithionic acid is obtained. Although it has not been
VOL. LXXXI, B
Digitized by VjOOQIC
2 carpenter: the oxidation of sulphurous acid to
found possible to obtain theoretical yields of this acid from the corre-
sponding hydroxides of manganese, cobalt, or nickel, yet on account
of the greater development of energy in the reduction of these oxides,
there is very strong reason for believing that a partial or even complete
decomposition of dithionic acid into sulphuric and sulphurous acids
would take place. The facts which point to this conclusion are con-
tained in the two following tables. The first shows the percentages of
dithionic and sulphuric acids obtained from the four hydroxides
already mentioned; the second indicates the changes of energy ex-
pressed in thermal units, involved in the reduction of the oxides : —
I.
Hydroxide. Percentage of dithionate. Percentage of sulphate.
Ferric 9606; 96-23 Not estimated.
Manganic 7652; 74-53 25-42.
Cobaltic 36-97; 35-07 63-80 ; 63-33.
Nickelic Nil. 101-04.
11.
Reduction of the hydroxides. Heat of reaction.
Fe2(0H)g = 2Fe(OH)2 + O + H^O - 546 calories.*
Mn2(OH)fl = 2Mn(OH)3 + 0 + HjO -448 „
Co8(OH)e =20o(OH)2 +O+H3O -225 „
Nij(OH)g «2Ni(OH)2 +O + H2O +13
It will be seen, on comparing the two tables, that the greater the
energy required for the reduction of the hydroxides, the larger is the
* The value of this calorie is that of the unit K used by Ostwald. It la that
quantity of heat which is given up by 1 gram of water as it cools from the boiling
to the freezing point. The values quoted for the reductions of the hydroxides of
iron, cobalt, and nickel have been taken from Ostwald's Lehrbueh der aXlgmrmrun
Uhemie,
Thermal data relating to the oxide Mn^O, appear to be entirely lacking, and the
figure given is only an approximate one, for purposes of comparison, obtained in the
following way :
MnO+O = MnO, +844 K.
3MnO + 0 = Mn,04 + 562K.
The mean of these values is taken to represent the heat of formation of
manganic from manganous oxide.
2MnO + 0 a Mna0, + 448K.
This value refers to the oxide, not the hydroxide. Judging by the analogy of
the diflference between MnO and Mn(OH)j, the value for Mnj(OH)j would be larger
by 8 or 4 units in the middle figure, but either value serves quite well for the
relation which it is desired to emphasise. The seaquioxides of iron, manganese, and
cobalt thus show negative heats of reduction the values of which vary in the order
mentioned, beginning with iron as the largest. In strikmg contrast to these, nickeUc
oxide shows a decided positive heat of reduction.
Digitized by
Google
DITHIONIC AHD BT METALLIC OXIDES. 3
percentage yield of dithionic aoid. The maximum yield is obtained
with ferric hydroxide where the energy that needs to be supplied is
such that it is possible to stop the reaction almost wholly at the stage
represented by the equation
(iii) Fea(OH)e + SSOg * FeSgOg + FeSOg + SH^O.
Whether the slight deficit from the theoretical number is due to a
slight decomposition of the nature
(iv) FeSgOe * ^^SO^ + SO^,
or to a defect in the method of estimation used, is a question which it
has not been found possible to decide.
In the case of manganic hydroxide, about one-fourth, and in the case
of cobaltic hydroxide, about two^thirds, of the dithionic acid formed is
decomposed in the manner indicated in equation (iv).
The reduction of nickelic hydroxide differs essentially from that of
the other three in being an exothermic change, and the energy liberated
is such that the reaction cannot be stopped at the intermediate stage,
but proceeds wholly to the last stage, and the only oxidation product
is sulphuric acid. The data in Table II are thus sufficient to account
in a simple manner for the results obtained by the author which have
been given in Table I, although the thermochemical values of the
reactions between the four hydroxides and sulphurous acid are not
known.
ExPBBiMEirrAL Fabt.
The essential part of the apparatus used in the author's experiments
will be found in the diagram on p. 4.
il is a three-way cock in a T-tube bent at right angles in the manner
indicated. The two arms can be connected, that at B with an apparatus
evolving carbon dioxide, that at C with an apparatus giving off sulphur
dioxide. At the other end, the T-piece is connected by thick rubber
tubing with the flask, Jt, in which the reaction takes place. The flask
has a capacity of about 300 c.c. and its neck is fitted with a rubber
cork bored with four holes. Through these pass —
(i) The tube delivering the gas either from B or C.
(ii) The exit tube, F^ which is further connected with the stopcock, S,
(iii) A small mercury pressure-gauge, Q.
(iv) A dropping funnel, E^ which becomes necessary when it is desired
to estimate the sulphuric acid formed in the reaction.
The apparatus was tested in the following way to see whether a
solution of sulphurous acid, free from sulphuric acid, could be obtained
in the flask, £
B 2
Digitized by VjOOQIC
4 CARPENTER: THE OXIDATION OF aULPHUROUS ACID TO
It was fitted together >in the manner indicated, except that neither
the gauge nor the dropping funnel was fitted into the cock, the holes
being temporarily plugged with glass rods, and further that the stop-
cock, jGT, was not fitted on to the exit tube, F. About 200 c.c. of distilled
water were placed in the flask and a quick current of carbon dioxide
was sent through to displace the air from the apparatus, the water
being boiled steadily, but not too rapidly. The carbon dioxide was
evolved by heating pure sodium hydrogen carbonate carefully packed
in a glass tube placed in a gas furnace. It was found that after l^ — 2
hours the limit of air displacement possible by this method had been
reached, but a slight froth, unabsorbed by potassium hydroxide solution,
always remained. When this was the case, the water in the flask was
gradually cooled, first by water and then by ice, the current of gas being
maintained. From 75 — 100 c.c. of water remained. The cock, H^ was
now attached and when the air had been expelled it was closed. The
first of the plugs in the cork was next removed, the dropping funnel
inserted, its air displaced, the tap closed, and a stopper placed in the
neck. Lastly, the second plug was taken out, the empty gauge put in,
and after the expulsion of its air, mercury poured in. By this time
the water, which had a temperature of about 5^, was generally saturated
with carbon dioxide, so that when the cock, A, was turned, isolating
the apparatus from the arm, B^ and connecting the latter with (7, only
at most a slight diminution of pressure inside the apparatus was
noticed on standing; carbon dioxide was passed through the tube,
C, until it had displaced all the air. The pressure can, of course,
always be increased again by connecting the apparatus with B for a
few seconds.
Meantime, during the last 20 — 30 minutes, a solution of sodium
hydroxide previously saturated with sulphur dioxide had been slowly
warmed in a flask provided with a leading tube, in order to expel the
air by a gentle but steady current of gas. When all the air had been
Digitized by
Google
DITHIONIC ACID BT METALLIC OXIDES. 5
removedy the apparatus was attached to C, the cock, A^ quickly turned
so that the gas passed into the flask, Z, the cock, ZT, opened, and the arm,
By disconnected from the sodium hydrogen carbonate tube. Sulphur
dioxide now passed through tdie water contained in the flask, gradually
saturating it.
In the meantime, a solution of barium chloride in water acidified
with hydrochloric acid had been steadily boiling in an open flask for
1^ — 2 hours to expel dissolved air. The solution was then rapidly
transferred to the dropping funnel. There it was cooled to the ordin-
ary temperature in a stream of carbon dioxide and the current main-
tained up to the moment when the solution was run into the flask.
This was done after sulphur dioxide had been passing in for 1^ hours,
care being taken that the tap was turned ofiE before the last c.c. of
liquid could run through. No turbidity of the liquid in the flask
could be detected. When, after shutting the cocks at A and H and
isolating the apparatus, a pressure from within was registered by the
mercury gauge, the apparatus was permanently isolated between A
and H and at the same time the flask allowed gn^lually to regain the
laboratory temperature. This ensured a steady pressure of sulphur
dioxide from within, the gas forcing its way through the gauge. Even
after 18 hours no precipitate had formed, and it was accordingly con-
cluded that the method was successful in giving a solution of sul-
phurous add free from sulphuric acid. This test was sufficient for
the purpose of the experiments, although solutions of sulphurous acid
under the influence of diffused light, gradually undergo a chemical
change which is not yet understood.
It should be mentioned that both the carbon dioxide and sulphur
dioxide were passed through a small empty wash bottle, where the
greater part of the water given off with them was retained, before
entering the T-piece at B and C respectively. The method of ob-
taining sulphur dioxide described is very convenient; a continuous
stream for 10 hours may, by carefully regulating the heating, be
obtained from about 3 litres of saturated sodium hydrogen sulphite
solution.
Separate experiments were needed for the estimation of dithionic
and sulphuric acids formed in the reaction between a metallic oxide
and sulphurous acid.
JSatimalion of the Sulphurio Acid.
The method is very similar to that adopted when the solution of
sulphurous acid was tested for the presence of sulphuric acid. In
the experiments with the dried metallic oxides, which it was
desired to keep in a compact state, the latter were put into the
Digitized by
Google
6 CARPENTER: THE OXIDATION OF SULPHUROUS ACID TO
flasky K^ by momentarily removing the rubber-cork and then replacing
it at the stage where the water had already been boiled and was
cooled in ice, a steady current of carbon dioxide being all the while
maintained. On the other hand, in those with the moist hydrated
oxides in a fine state of division, known volumes of these, sus-
pended in water, were introduced by a pipette into the flask at the
beginning of the experiment instead of distilled water alone.
After the reaction with sulphurous acid was finished, the sul-
phuric acid was estimated in the flask in which the reaction had
taken place, to avoid the oxidation of the dissolved sulphurous acid.
The liquid was precipitated by an acidified barium chloride solution
in the manner already described and was allowed to stand until it
had clarified. The supernatant liquid was poured off through a
weighed filter and the precipitate repeatedly washed by decantation,
first with an air-free, dilute solution of hydrochloric acid and after-
wards with distilled water, and the estima ion of the barium sul-
phate carried out in the usual way.
Eatimaiion of the DUhioni Aoid,
In this case, except that the dropping funnel is not needed, the
method is the same as the foregoing one up to the point where the
action of sulphurous acid on the hydroxide is complete. A ther-
mometer dipping below the surface of the liquid was substituted
for the dropping funnel. A stream of carbon dioxide was then
passed through the liquid, which was gradually warmed until it had
acquired a temperature not exceeding 46^ and well shaken in order
to expel as much of the dissolved sulphur dioxide as possible.
After cooling in a current of the same gas, it was precipitated
with an excess of a warm saturated barium hydroxide solution. In
this way, sulphurous and sulphuric acids were removed whilst a
solution of barium dithionate and the excess of barium hydroxide
remained. The precipitate was allowed to settle, the liquid poured
off through a filter, and the residue washed six times with boiling
water by decantation, the washings being successively added to the
main filtrate. The barium hydroxide was then precipitated with
carbon dioxide at 100^, filtered, and the filtrate evaporated on a
steam bath to dryness on account of a small quantity of barium
carbonate which always remains dissolved, fhe residue was extracted
six times with hot water and the solution of barium dithionate thus
obtained filtered into a weighed platinum dish. It waa evaporated to
dryness on the steam-bath and ignited until the weight was constant.
* At temperatures above 50% Bolations of the dithionates of Iron, cobalt, and
nickel begin to deoompoae into the correaponding Bolphatesland solphunms acid.
Digitized by
Google
DITHIONIC ACID BY METALLIC OXIDES, 7
Iq the estimations of ditbionic acid resulting from cobaltic hydroxide,
whioh contained a small quantity of alkali, the barium dithionate
solution was at the last stage converted into barium sulphate by pro-
longed boiling in a reflux apparatus with a few drops of hydrochloric
acid solution. This modification is necessary, but not so convenient as
the other method.
The disadvantage attaching to this method of estimating ditbionic
acid is that it involves the washing out of a small quantity of barium
dithionate from a large quantity of barium sulphate, with consequently
a possible loss of the dissolved salt by adsorption. This applies to the acid
as obtained from the hydroxides of iron and cobalt where the reduction
is slow and the amount of sulphurous acid used is large, but not to the
other two cases where the reduction is rapid and the quantity of suN
phurous acid used is small.
Eatimaiicn of the ' Available ' Oxygen in the Metallic Oxides.
Ferric hydroxide was estimated gravimetrically as ferric oxide. The
other three oxides were made to react with an excess of standard oxalic
acid, which was afterwards titrated with standard potassium permangan-
ate solution, and in this way their available oxygen determined. Owing
to the influence of light on solutions of oxalic acid, the latter were
always freshly prepared, immediately before they were needed, from pure
ammonium oxalate, which can easily be obtained and keeps well. The
potassium permanganate solutions were standardised against this.
For the estimations of available oxygen^in the moist precipitates, the
method of withdrawing a known volume by a pipette was used here
just as in the estimations of ditbionic and sulphuric acids.
It was not found that the pink colour of the cobalt solutions or the
green colour of the nickel solutions interfered with the permanganate
titrations or masked the colour of the permanganate. Seeing that
these colours were complementary, the plan was tried of decolorising
whichever liquid was being titrated with a few drops of a solution of
the sulphate of the other metal, but in neither case was any effect on
the results noticeable.
Particulars of the methods adopted for the preparation of the metallic
hydroxides will be found in the sections dealing with the reduction of
each oxide. The specimens referred to as * moist ' were kept in closed
vessels and suspended in water or, in the case of nickelic hydroxide, in
dilute alkali. For the estimations, a measured volume was withdrawn
by a pipette after the liquid had been thoroughly shaken to ensure a uni-
form distribution of the solid. Any portion of the latter adhering to the
walls of the pipette was afterwards washed in. The ' dried ' specimens
were obtained from the 'moist' by allowing the latter to dry gradually
Digitized by
Google
8 carpenter: the oxidation of sulphurous acid to
on clean porous plates. On drying, the precipitate shrinks and as a
rule becomes completely detached from the plate, or needs at most only
a slight touch to loosen it. After removing it, the surface of the plate
is found to be traced with a delicate pattern resembling vegetable oell-
structures in a remarkable degree. Here and there, spirals are formed,
caused by the precipitate having adhered with varying degrees of in-
tensity to the plate. These facts were noticed with all the precipitates
dried in this way.
Reduction of Ferric Hydroxide, Formation of lerrous DUhionaie,
Gelis (loe, cit.) claims to have proved that the reaction between
ferric hydroxide and sulphurous acid takes place in the following
way : —
(v) Fe2(0H), + 3S0j = Fe2(S03)3 + mfi.
(vi) Fe,(SO,)s - FeS^O. + FeSO^.
He found that when a current of sulphur dioxide is passed into
water containing ferric hydroxide in suspension, the latter dissolves
and a liquid is obtained which has a red colour, an acid reaction, and
a strong odour of sulphur dioxide, and concluded from his analyses that
neutral ferric sulphite, Fe2(S03)3, is formed.
When the solution has stood for some hours in a closed vessel, the
red colour is found to have changed to pale green. The analytical re-
sults quoted show that ferrous dithionate and ferrous sulphite have
been formed in approximately equal and quantitative amounts as shown
in equation (vi). A small, but not negligible, quantity of ferrous
sulphate was shown to be also present. Gelis regarded this as an
accidental product, probably due to the oxidation of some of the
sulphite by the air.
It is curious that so remarkable an oxidation of sulphurous acid,
from which, according to the explanation given, no sulphuric acid, but
solely dithionic acid, results, should have apparently attracted no atten-
tion. Gelis's work, published 39 years ago, has never been repeated,
and yet certain criticisms obviously suggest themselves to anyone who
studies the paper referred to.
It is, in the first place, noteworthy that the specimens of iron used in
those experiments do not appear to have been tested for the presence
of manganese, and there can be hardly any doubt that small quantities
of this metal were present. It is a natural question to ask whether this
impurity may not really have been the cause of the formation of di-
thionic acid, possibly by acting catalytically. In the second place, the
question whether sulphuric acid is formed or not cannot be dismissed
in the manner indicated, but must, if possible, be tested by experiment.
Digitized by
Google
DITHIONIC ACID BY METALLIC OXIDES. 9
It is evident that no pains were taken to exclude air from the appar-
atus and liquids used, but this is one of the fundamental conditions
which must be realised if this point is to be settled.
In my own experiments, the ferric hydroxide was prepared in the
following way. The ferric chloride used was purified by separating
iron as the basic formate and repeating the precipitation until it was
quite free from the last traces of manganese. This was held to be the
case when fusion of the ferric formate with potassium nitrate and
sodium carbonate yielded a product from which the green tint of a
manganate was entirely absent, and this test was usually satisfied after
the third precipitation of the iron. Separation by the formate is much
quicker than by either the acetate or succinate. The precipitate of
basic ferric formate breaks ap, on warming, into a very fine powder,
which can be thoroughly well washed with boiling water by deoantation^
and settles quickly. When pure, it was dissolved in hydrochloric acid,
precipitated with ammonia, and the hydroxide washed until it was so
finely divided that it remained for several days suspended in the
liquid.
The reduction of such ferric hydroxide by sulphurous acid is very
slow and I can by no means confirm Gelis's statement that the sub-
stance dissolves in that quantity of the acid corresponding to the
formation of the neutral sulphite of iron. To take an example, 0*2859
gram of ferric hydroxide suspended in water was acted upon by a
current of sulphur dioxide for 7^ hours. The liquid was repeatedly
shaken, but even after 3| days 0*1041 gram ;of hydroxide still re-
mained— ^indeed the analytical results quoted for iron were, in both
cases, derived from experiments in which the residual ferric hydroxide
had to be estimated. This was done by filtering it ofiE just before the
precipitation with barium hydroxide. The precipitate was well washed,
the washings being added to the main filtrate, and ignited until the
weight was constant.
The reason why Gelis found ferric hydroxide dissolved readily in
sulphurous acid probably was that the substance was not thoroughly
washed. Freshly precipitated and unwashed ferric hydroxide is very
quickly dissolved by sulphurous acid. The more pains that are taken
to wash out adhering salts, the slower does the reaction become, until
a condition is reached similar to that of which an instance has already
been given. In such a case, the liquid begins to acquire a yellow
colour after 4 — 5 hours, and this increases somewhat in intensity if the
current of gas is maintained, but a red solution has not been obtained.
After standing 12 — 18 hours, the yellow colour changes to pale green
and the solution evidently contains iron in the ferrous state. The
dithionic acid which has been formed may then be estimated. The
results already quoted show that its formation is quite independent of
Digitized by
Google
10 CARPENTER: THE OXIDATION OF SULPHUROUS ACID TO
the presence of manganese. They also show that less than 4 per
cent, of the dithionic acid formed undergoes any decomposition.
Whether this slight decomposition takes place has not been actually
decided. The fact that some unchanged ferric hydroxide always
remained in the liquid rendered it impossible to test for the presence
of sulphuric acid. When an acidified solution of barium chloride was
introduced, with the needful precautions, into the flask, no precipitate
was noticeable for a few seconds, and in one case for nearly a minute,
but after this interval a turbidity could always be detected, and this
gradually increased in intensity. This can be accounted for by the
diffusion through the liquid of the hydrochloric acid introduced and its
reaction with the ferric hydroxide. The ferric chloride formed would
react with the sulphurous acid present, oxidising it, as is well known,
to sulphuric acid, and this would be precipitated by the barium chloride
present. There can, however, be little hesitation in concluding that if
any sulphuric acid is formed in the reaction, it can only be a minute
quantity.
Enough has been said to show that there can be no question of testing
Gelis's first equation with ferric hydroxide which has been thoroughly
washed. The author's experiments, however, agree with those of Crelis
in 8howing that there is a stage of the reaction preceding that of
ferrous dithionate and ferrous sulphite, corresponding probably to the
formation of a ferric sulphite. Evidence of a similar stage has been
obtained, although to a less^degree, in the experiments with cobaltic
hydroxide, but not with the other two oxides.
Reduction of Mcmganic Hydroxide, 'M.nfi^{0^)2' FarmcUion qf Man*
^notis Dithionate cmd Mangcmoua Sulphate.
The reduction of this hydroxide stands in sharp contrast with that
of ferric hydroxide. It takes place with the utmost ease and this is
the more remarkable because previous investigators have concluded
that exactly the opposite is the case. Heeren {Pogg, Ann., 1826, 7,
55), discussing the formation of sulphuric acid in the reaction between
manganese dioxide and sulphurous acid, quotes Berzelius to the effect
that if the dioxide contains no manganic hydroxide, the sole product
is manganous dithionate ; but that as this impurity is nearly always
present, a certain quantity of sulphate is also formed. After question-
ing whether this statement is founded on direct experiment, Heeren
gives analytical results to show that some of the sulphuric acid must
be formed from the manganese dioxide, and even calculates how much
has come from this on the assumption that the manganic hydroxide
cannot give rise to any dithionic acid. He investigated the action of
sulphurous apid on manganic oxides of various states of aggregation
Digitized by
Google
DITfflONIC ACID BY METALLIC OXIDES. 11
and found it to be very slow except with precipitated specimens. He
noticed that small quantities of dithionic acid were formed, but
snggosts that they were due to some of the dioxide being present.
So far from this being the case^ the yields of dithionic acid obtained
by the author from manganic hydroxide are higher by more than
5 per cent, than the highest obtained by Spring and Bourgeois (Btdl Soe,
Ghm,^ 1886| [ii], 46, 151) in their investigation of the yields obtained
from manganese dioxide in different states of aggregation. An even
greater percentage of dithionic acid was obtained by using manganic
hydroxide which had been dried at 100^ and thus assumed a more com-
pact state. This was evident from the fact that it took fully 30 times
as long to dissolve in sulphurous acid as the undried specimen.
In preparing the oxide for these experiments, a specimen of man-
ganous nitrate containing only a very small quantity of iron was used.
The latter was first removed as basic ferric formate. The manganese
was then precipitated with ammonia and bromine, and the hydrated
dioxide thoroughly washed with nitric acid and then with water. In
preparing manganic hydroxide from this, the directions given by
Garius {Anndlm^ 1856, 96, 63) were followed exactly.
The reduction of the ' moist ' oxide by sulphurous acid is very rapid,
0'25 gram, the amount usually treated, dissolving after about one
minute's passage of the gas through the liquid. The estimation of
the dithionic or sulphuric acid can be proceeded with at once. There
is no evidence of the intermediate formation of a sulphite.
Reduction of CohaUie Hydroxide. Formation of CobalUma Lithionate
and CobaUous Sulphate,
The cobalt nitrate used for the preparation of the oxide was first
freed from a small quantity of iron present by precipitating the latter
as basic ferric formate. It was then purified by converting it into
potassium cobaltinitrite. The latter, after being carefully washed,
was dissolved in hydrochloric acid^ and afterwards precipitated with
the necessary precautions by bromine and potassium hydroxide. It
was repeatedly washed, first with potassium hydroxide and afterwards
with hot water. It is well known that the hydroxide thus obtained
does not correspond to the formula Oo2(On)0. The amount of avail-
able oxygen it contains depends on the exact conditions of its prepara-
tion. Further, it contains alkali, which cannot be removed without
decomposing the hydroxide. On account of these facts, the percentage
of available oxygen was estimated by oxalic acid in the manner already
described.
For the complete reduction of about 0'25 gram of the ' moist' hydr-
p^e, approximately 5 hpurs are needed. The li<][uid acc^uires a brown
Digitized by
Google
12 CARPENTER: THE OXIDATION OF SULPHUROUS ACID TO
colour, which gives place to light yellow as the solution of the hydr-
oxide proceeds, and this in its turn is replaced by the pink colour
characteristic of cobaltous salts. The liquid is then ready for the esti-
mation of the dithionic and sulphuric acids formed.
In order to test whether the percentage of dithionic acid could be
increased by retarding the rate of the reaction, a number of experi-
ments were made with specimens of the hydroxide of different degrees
of compactness. The results are illustrated in the subjoined table :
Approximate time
of reduction Percentage of
Condi cion of the hydroxide. of 0*25 gram. dithionic acid.
(i) Moist 6 hours 36-97; 35-07
(ii) Dried at 140—150° and finely
po.wdered 4 days 28-14
(iii) Dried at 140—150° and coarsely
powdered 5 „ 23-41
(iv) Dried at 140—150°, but not
powdered 6 „ 10-87
In experiments ii, iii, and iv, a current of sulphurous acid was
passed through the liquids for about 7 hours. At the end of 2 days,
the current was renewed for a further period of 2 hours on account of
the gauge registering a gradual diminution of pressure inside the
apparatus. In these three cases, no colour, except' the pink of
cobaltous salts, was noticed during the reduction.
A comparison of the figures shows that the amount of decomposition
of dithionic acid is increased by rendering the hydroxide more com-
pact. This fact is accounted for, if it is borne in mind that in these
cases, where reaction is taking place at the surface of solid particles,
local heating is bound to take place, and the effect of such a cause
upon dithionic acid is easily comprehended.
In the reductions of cobaltic hydroxide and ferric hydroxide, where a
prolonged current of sulphur dioxide is necessary, care must be taken
that the temperature of the liquid is kept just above the point at
which the solid hydrate, SOjiicHgO, crystallises out (8°), otherwise the
tube in the flask where the gas enters becomes choked.
Reduction of Nichdic Hydroxide. Formation of Nickeloua Sulphate.
' Preliminary experiments with this hydroxide had shown that both
dithionic and sulphuric acids are produced by reduction with sulphurous
acid, but when the estimations with the hydroxide obtained in as pure
a state as possible came to be carried out, it was found that sulphuric
acid is the sole product of the oxidation of the sulphurous acid
used.
Digitized by
Google
DITHIONIC ACID BIT METALLIC OXIDES. 13
The nickelic hydroxide was prepared from a specimen of nickel
nitrate containing a minute quantity of cobalt. The latter metal was
removed as potassium cobaltinitrite. The filtrate was boiled with
excess of hydrochloric acid, precipitated with potassium hydroxide,
filtered, well washed, and then redissolved in hydrochloric acid. From
this solution, it was precipitated with bromine and potassium hydroxide
and in consequence of the readiness with which it loses oxygen in
presence of water, it was washed with cold alkali only. In order to
keep the moist hydroxide in an alkaline medium until the moment of
its reaction with sulphurous acid, nitrogen was used instead
of carbon dioxide to displace the air from the apparatus and liquids
used. Further, as it is not safe to expose the hydroxide to a temper-
ature of 100^, a modification of the methods of estimation of dithionio
and sulphuric acids was necessary. After the expulsion of the air
from the apparatus by nitrogen, and while the gas was still passing
through, one of the glass plugs was removed and the cold liquid con-
taining the hydroxide in suspension introduced by a pipette. The
plug was then re-inserted and the usual course of the determination
followed.
The reaction between this hydroxide and sulphurous acid is the
most rapid of the four investigated. The absorption of the gas is so
energetic that unless a very quick evolution is taking place at the
moment when it is admitted into the apparatus, the liquid ;and precipi-
tate in the flask are sucked back into the tube, G. The green colour
of the nickelous salt is seen almost immediately.
The rate can be considerably diminished by increasing the size of
the particles of the hydroxide and thus exposing relatively less
surface. This was done by drying the substance at the ordinary
temperature on a porous plate. Experiments were made with these
dried specimen?, both powdered and not powdered, in which the dura*
lion of the reaction was between ten and twenty minutes. But in
none of these cases has any dithionic acid been detected, and this
negative result has been confirmed for the moist hydroxide by the
finding of rather more than the theoretical percentage for sulphuric
acid.
Note an the Action of Sulphurous Acid on t?ie Dithionatea qf Lead and
Barium.
The only account of any experiments performed with the object of
testing whether dithonic acid is obtained by the action of sulphurous
add on the peroxides of lead and barium is to be found in the paper of
Gay Lussac and Welter {Ann. Chim. Pht/a., 1818, 10, 312), where
they describe the discovery of this acid. Their results were negative.
Digitized by
Google
14 TU6kBR and MOODY: THE PRODUCTION OF
The author has made a number of experiments with these oxides in
the apparatus described at an early stage of this paper, but has never
found that any dithiopic acid is produced. In view of the negative
character of these results, the action of sulphurous add on the
dithionates of these metals was studied. It was found that
solutions of barium dithionate are quite unaltered except at a temper*
ature at which the influence of heat alone begins to be seen, when a
gradual decomposition into barium sulphate (and sulphur dioxide sets
in. On the other hand, solutions of lead dithionate are decomposed
instantaneously, even at 5^, lead sulphite being precipitated and free
dithionlc acid remaining dissolved.
The experimenlti were carried out in a similar manner to those
already described, the solid dithionates being introduced into the
flask, Ki after the water had been cooled in a stream of carbon dioxide.
In the case of the decomposition of lead dithionate, the precipitate was
filtered o£E and washed with air-free water. It dissolved completely in
aqueous hydrochloric acid and liberated sulphur dioxide, which was
recognised by its odour and the reduction of chromate paper. The
filtrate was rendered alkaline with barium hydroxide and barium
dithionate obtained after the removal of sulphurous and carbonic
acids.
In conclusion, the author acknowledges with great pleasure the
assistance of Mr. F. H. Palmer in the .preliminary experiments with
nickelic hydroxide and the dithionates of lead and barium, and of Mr.
C. W. May in those with manganic hydroxide.
Addendum. — This paper had been placed in the hands of the
Secretaries of the Chemical Society before the publication of the
paper, by Julius Meyer, on the formation of dithionic acid (^«r., 1901,
d4, 3606). The ground covered in the two papers is very nearly the
same.
Thb Owxnb Gollsgb,
Hanchbstbb.
11. — The Production of hitherto unhnovm Metallic Borides,
By Samuel Auchmutt Tucker, Ph.B., and Hebbebt R.
Moody, B.S., M.A.
Until the electric furnace made their formation comparatively easy«
the borides were almost unknown and even now there have not as yet
been reports concerning many of them. Moissan has described a few of
these compounds, notably those of iron, cobalt, nickel, carbon, caloimo,
strontium, barium, and, lately, silicon.
Digitized by
Google
HITHERTO UNKNOWN METALLIC BORIDES. 15
So far as they have been investigated, the borides present evidence
of definite composition and crystallisation, they are stable and fuse at
comparatively high temperatures. As a consequence of their high
fusing point, hardness, and good crystallisation, it is quite possible that
some of them may prove to have industrial uses.
The available processes for the production of borides are two in
number. In the first, the two elements are heated together in the
electric furnace, and in the second, boron chloride is passed over the
metallic element. The former of these was selected as being the most
practicable for the preparation of the borides described in this paper.
The utmost care was taken to prevent the Addition of carbon, silicon,
&C., to the product in each case, and in our opinion the borides
described were entirely free from these elements. The boron in each
case was determined directly by Gooch's method (Amer, Cham. «/*., 1887,
0, 23).
Zirconium Boride, — The zirconium salt available for the preparation
of this boride happened to be the nitrate. In order to reduce this to
the elementary state, two processes were tried.
In the firsts the nitrate was ignited until it was wholly converted
to the oxide. This was then subjected to the regular Goldschmidt
process, which did not prove to be satisfactory.
In the second, the nitrate was dissolved in cold water and the
hydroxide precipitated from this solution by sodium hydroxide. Cold
water was used in washing the precipitate, inasmuch [as hot water
causes it to become insoluble. After being washed, the hydroxide was
dissolved in hydrofluoric acid and to this solution neutral potassium
fluoride solution was added, forming a precipitate of the double fluoride,
SKEyZrE^. When dry, this salt was reduced by means of powdered
aluminium and the cake thus formed was boiled for three days with
concentrated hydrochloric acid. It was found impossible to filter the
product rapidly, even with the aid of suction. After being washed
with hot water, the metal was ready for use.
The elementary boron was prepared by fusing boric acid and
reducing the oxide thus formed ^ith metallic magnesium* To remove
magnesium salts, the cooled mass was boiled with dilute hydrochloric
acid, filtered, washed, and then boiled for three days with hydrochloric
acid of sp. gr. 1*2. After filtering and washing, the residue was
boiled for several hours with hydrofluoric acid, and after a final
washing it was dried.
For preparing the sirconium boride, 15 grams of the eirconium
were mixed with 2-2 grams of boron and the whole heated for 5 minutes
in a carbon crucible with the aid of a current of 200 amperes and 65
volts. The product was a button, blackish on the outside, brittle, and
of a steel grey colour on fracture. Under the microscope^ it proved to
Digitized by
Google
16 PRODUCTION OF HITHERTO UNKNOWN METALLIC BORIDES.
be an agglomeration of brilliant, tabular, translacent to transparent
crystals, many of these being colourless. It had a sp. gr. 3*7 and a
hardness 8. It was slowly attacked by hot concentrated acids and
aqua regia. Boiling liquid bromine attacked it feebly.
Analyses of the compound were made and 86 per cent, of zirconium
was found to be present. This corresponds very closely with the
theoretical amount of zirconium in a boride in which zirconium is
quadrivalent ; therefore the formula of this compouod is undoubtedly
Zr,B,.
Chromium Boride. — ^This boride was made by heating a charge con-
sisting of 10 grams of metallic chromium and 2*1 grams of boron for 6
minutes by the aid of a current of 175 amperes and 60 volts. The
product was a well formed button, greenish on the outside and of a
greyish metallic lustre on fracture. It had a sp. gr. 5, a hardness 8,
was distinctly crystalline, and had a conchoidal fracture. It was
weakly attacked by hot acids and was not altered by exposure to the
air. Analyses of the product gave 82 per cent, of chromium, a
result which indicates CrB as the probable formula of the compound.
TungBten Boride. — As tungsten is closely related to chromium, it was
selected as a promising element and a trial was made of its affinity for
boron. The metal tungsten may be prepared from alkali tungstaies
by acidifying their solutions with hydrochloric acid. This causes the
precipitation of the trioxide. After being dried, the trioxide can be
reduced in the electric furnace, the charge used containing 10 parts of
tungsten trioxide to one part of carbon.
For the preparation of the boride, 4 grams of tungsten were mixed
with 0*2 gram of boron and then heated for 5 minutes by the aid of a
current of 175 amperes and 65 volts. This produced a good fusion
and the product was silvery and metallic on fracture. It was very
brittle and xinder the microscope was seen «to be crystallised in
perfect octahedra. Its hardness was 8 and its sp. gr. 9 '6. It was
slowly attacked by concentrated acids, and vigorously by aqua regia.
Analyses of the product Ehowed the presence of 89 per cent, of
metallic tungsten, a result which indicated the formula to be WB^.
Molybdenum Boride* — The final compound prepared was a boride of
molybdenum. This was selected, as the element molybdenum is closely
related to chromium and tungsten and the metal is rather easily
prepared.
The molybdenum was obtained by heating 300 grams of molybdenum
trioxide and 30 grams of coke for 25 minutes with a current of 200
amperes and 65 volts.
For making the boride, 6 grams of metallic molybdenum were mixed
with 1 gram of boron and heated for 20 minutes by the aid of a current
of 230 amperes and 70 volts. Th's gave a homogeneous button with a
Digitized by
Google
CONSTITUTION OF ACIDS OBTAINED FBOM a-DIBBOMOCAMPHOR. 17
hardness of 9. It was quite brittle and on fracture showed a brilliant
metallic lustre resembling that of pale brass. It was crystalline in
struct urcy and its sp. gr. was 7 '105. The substance was moderately
attacked by hot concentrated acids and vigorously by hot aqua regia.
The formula MogB^ was -given to this compound as the result of
several analyses which showed the presence of 86 per cent, of
molybdenum.
An attempt to make the borides of copper or bismuth failed
entirely ; in fact, there does not seem to be any affinity between boron
and the members of the copper group.
Columbia Univebsity,
New Tobk.
III. — The Constitution of the Acids obtained from
a-Dibromocamphor.
By Arthur Lapworth and Walter H. Lenton.
Whbn a-dibromoeamphor is warmed with moist silver salts, it is in
part transformed into the unsaturated monocyclic acid, bromocam-
phorenic acid (Trans., 1899, 75, 1134), in which, as has already been
shown, the complex
(2) :(f — 9: (I)
(3) :C CMej,
•5E 5c
CHg- c:
IMe-COgH
must be assumed to be present, the ethylenic union existing between
the atoms 1 and 2, or 2 and 3, to one of which also the bromine atom
mnst be attached. The facts on which these statements were based
are, briefly, as follows : — (1) the substance readily affords homocam-
phoronic acid on oxidation with mild oxidising agents (Trans., 1899,
75, 988), and (2) it is obtained from camphor, which contains the
group "CMej* 0Me*CO', by a process which involves no violent action.
In the first paper, in which the constitution of bromocamphorenic
add was discussed, it was shown that when the acid is converted into
a-monobromocampholid, the lactonic oxygen atom becomes attached to
the ring at the brominated carbon atom, the group *OBrIOn* becoming
•CBr-CHL'
converted into X This conclusion was confirmed later by the
observation that camphonic acid, the acid which is formed on hydro-
lydng the lactone, contains tHe carbonyl group *00' in the ring
indicating that the first stage in the hydrolysis is the formation of an
acid in which the group *G(OH)Br* is present.
VOL. LZXXI. C
Digitized by
Google
18 LAPWOBTH AND LENTON : THE CONSTITUTION OF THE
Arguing from the behaviour of simpler lactones, it was presumed
that a y-laotone would be formed in preference to a 8-lactone, the
bromine atom would then be attached to the carbon atom labelled
1 or 3, and it is here that the first dubious point in the argument
appears, for it is by no means always legitimate to apply generalisa-
tions based on the behaviour of open chain compounds to substances
containing closed rings. For this reason, we have again taken up the
investigation, with the object of ascertaining to which of the carbon
atoms, 1, 2, or 3, the bromine atom is attached, for, in so doing, the
constitution of the interesting series of compounds obtained from
a-dibromocamphor would be determined beyond question. ' *
It may be worth while to point out that the view to which one of
us had come respecting the position of the bromine atom in question
led to conclusions which did not appear to be altogether sati^actory^
Thus, assuming that the bromine atom was attached to the carbon
atom in position (1), the behaviour of camphonic acid towards substi-
tuting agents was not easily explained (compare Trans., 1900,77, 451),
whilst the supposition that it was associated with carbon atom (3) led
to the view that both camphonic and camphononic acids must be re-
presented by formulsB containing the grouping 'CHj'^^'^^s* C^^i^s-y
1899, 75, 1139), a conclusion which, although in excellent agreement
with the properties of camphonic acid, is altogether unsatisfactory in
regard to the other substance, which behaves exactly as would an acid
containing the complex C-^'CO, forming, for example, no additive
compound at all with hydrogen cyanide (compare Trans., 1901,
79, 379).
The first part of the investigation was therefore devoted to proving
that, in the formation of camphonic acid from bromocamphorenic acid,
no change of structure occurs, and that the ketonic oxygen atom
occupies the position of the lactonic oxygen atom in the campholids.
For this purpose, camphonic acid was reduced with sodium amalgam,
and the product shown to be identical in all respects with the hydroxy-
acid obtained by Forster on hydrolysing campholid itself (Trans., 1896,
69, 57). The view already advanced of the mode of formation of
camphonic acid seems, therefore^ to be correct.
For the second part of the investigation, the tribromolactone ob-
tained from camphonic acid (Trans., 1900, 77, 458) was used as a
starting point. This substance, being obtained by gentle treatment of
camphonic acid with bromine, must be supposed to be derived from the
tetrabromo-acid containing the group 'CBr^* CO'CBr^*. It was hydro-
lysed by careful treatment with alkali, and the product, which we did
not attempt to isolate, was oxidised by means of cold sodium hypo-
Digitized by
Google
ACIDS OBTAINED FBOM a-DIBROMOCAMPHOB. 19
bromite, which was found to be the most suitable agent for the purpose.
The oxidation product was isolated in the usual way and found to be a
mixture, which, on further investigation, proved to contain trimethyl-
suocinic add, a small quantity of camphoronic acid, and a relatively large
amount of an acid, C^H^jOg, which proved to be the substance hitherto
known as /3-hydroxycamphoronic acid (Bredt, AntuUen, 1898, 200, 168).
The formation of camphoronic acid from tribromocamphonolactone
shows that the latter contains the complex
. -9-
:0 CMoj .
CHj-CMe-C:
On inspecting the skeleton formula of camphorenic acid (p. 17) and
remembering that the bromination of camphonic acid probably results
in the first instance in the formation of a tetrabromo-add containing
the group 'OBr^'CO'CBrj*, it will be seen that the only possible formula
for this tetrabromo^acid is expressed by I, hence camphonic acid itself
must have the constitution represented by II :
CO — (j3Br2 9O-9H3
I. CBrg CMej 11. CHj CMe^
CHj--CMe-COaH CBj-CMe'COjH
Tetrabromo-acid. Camphonic acid.
It follows of necessity, therefore, that in bromocamphorenic acid the
bromine atom occupies the position of the ketonic oxygen in this
formula, or in other words it is attached to the carbon atom 2 in the
skeleton formula.
This, in itself, does not enable us to decide whether the double bond is
situated between the atoms 1 and 2, or 2 and B ; the first alternative,
however, is in all probability the correct one, for in this event homo-
camphoric acid and camphononic acid will have the formulie III
and IV.
(jJOjH (jJOjH OHj-CO
III. CHj CMoj, IV. I CMoj
CHj CMe-OOgH OBLj-CMe-COjH
Homocamphoronic acid. Camphononic acid.
This conclusion is in complete agreement with the feeble ketonic pro-
perties of camphononic acid, whilst the readiness with which cam-
phonic acid forms additive compounds is explained by the presence in
it of the complex •CHg-OO-CHg-.
The conclusion thus arrived at harmonises with the whole of the
known properties of the substances obtained by Forster (Trans., 1896,
60, 36) and by Lapworth and Chapman (Trans., 1899, 75, 986;
1900, 77, 446) and explains the apparent anomalies which have
C 2
Digitized by VjOOQIC
20 LAPWORTH AND LBNTON : THE CONSTITUTION OF THE
appeared from time to time. The formula of camphononic acid,
moreover, is what it should be on the basis of Bredt's formula for
camphor and the relationship which has been firmly established
between this acid and the simpler camphor derivatives (Trans., 1900,
77, 1066 ; 1901, 79, 1284).
The formation of bromocamphorenic acid from a-dibromocamphor
may be expressed by the scheme,
CHj-CH CBrj CHa'CBrlCH
j CMe„ I + HjO « I CMe, + HBr.
CHj-CMe^O CHj CMe-OOjH
It is not impossible that at an intermediate stage a trimethylene ring
is produced and that this afterwards breaks down, the carbon atom
originaUy exterior to the ring having become merged in it. Thus
intermediate compounds such as
OH
OH
CBr
...I —
CO
\
CMcj
or
CHj- ^^CMe
OH
OMeo
OMe-OOjjH
might afford bromocamphorenic acid by scission at the points indicated
by the dotted lines.
The ready formation of trimethylene rings in certain cases, notably
in the production of carone and of the caronic acids (Perkin and
Thorpe, Trans., 1899, 76, 522), makes it appear likely that the phe*
nomenon is not so infrequent as is generally supposed.
The assumption that an unstable trimethylene ring is formed in
many other changes would probably be of great value in explaining
their progress. Thus the curious transformations of campholytic
and Molauronolic acids one into the other (Walker, Trans., 1900, 77,
378), and into derivatives of tetrahydro-xylic acid (Perkin and Lees,
Trans., 1901, 79, 323), may be the result of reactions like the following :
CMe/ '
OHj«CH-CO,H
Campholytic
acid.
OHj-iH'OOgH
Intermediate
compoand.
CHs-cJ-OOjH
ifoLaaronolic acid.
OH,-(:]Me
I 9^
I OHMe
OHj-C-OOjH
Tetrahydro-zylic acid.
Digitized by
Google
ACIDS OBTAINED FROM a-DIBROMOGAMPHOR. 21
Similarly, the formation of camphene derivatives from bornyl chlor-
ide and the allied substances, may be the result of a similar parti-
cipation of a methyl group in the formation of a trimethylene ring, and
most of the apparently anomalous properties of these compounds point
to such an explanation as being the correct one (compare Marsh, Proa,
1899, 15, 64).
As already mentioned, *' /3-hydroxycamphoronic acid " is produced to
a far greater extent than camphoronic acid in the oxidation of the pro-
duct of hydrolysis of tribromocamphonolactone. It may appear at first
sight to be somewhat remarkable that this should be the case, as it is
impossible to suppose that the substance is obtained by the oxidation
of camphoronic acid. If it be remembered, however, that the product
of hydrolysis consists largely of ketonic substances, which are usually
easily attacked by hypobromites, there is no great difficulty in explain-
ing the formation of the " hydroxy "-compound.
Thus, for example, the product in the first instance may consist of a
mixture of substances such as
(jX)-90 90jH 9O2H
CO CMcj and CO CMe^ ,
CHj-CMe-COjH CHj— CMe-COsH
and either of these might be attacked by the hypobromite, the latter,
for example, affording successively the substances represented by the
formulsB
OOjH COjH OOgH yCO CO^H /CO
CO CMe« -> CO/O CMe- -^
II 1/ I
CHBr-CMe-COaH CH OMe-COjjH CH CMe-COgH
The substance known as /3-hydroxycamphoronio acid is, in reality,
a lactonic acid containing water of crystallisation and should more
correctly be termed ^-camphoranic acid, employing the word used by
Bredt for the isomeric substance. The hydrated acid, OgHj40g,2H20,
is dibasic and may be boiled with excess of iT/lO alkali for half an hour
without suffering any appreciable amount of hydrolysis into the
hydroxy-acid. This fact rendered its identification a matter of con-
siderable difficulty, for it is described as a tribasic acid both by Kach-
ler and Spitzer and by Bredt. It was necessary, therefore, to prepare
the acid directly from camphoronic acid, by the process which is
described later, and it was, then found that the conclusions which we had
arrived at with regard to the acid from tribromocamphonolactone held
good with regard to the other, the two substances being identical in
every respect.
Digitized by
Google
22 TJLPWOBTH AND LEKTON: THE OONSTrTUTION OF THE
EXPBBIHBKTAL.
Reduction of Gcvmphonic Acid.
Oamphonic acid, dissolved in 10 per cent, aqueous sodium hydroxidoi
was placed in an evaporating dish, carbon dioxide passed rapidly
through the solution, and sodium amalgam added in small quantities
during the course of several hours, until a small portion, after acidifi-
cation with acetic acid, gave only a slight precipitate with p-htomo-
phenylhydrazine acetate, indicating the absence of all but a trace of
the ketonic acid. The liquid was separated from the mercury, acidified
with dilute sulphuric acid and extracted repeatedly with ether ; the
ethereal solution was then washed with a very little water, dried, and
evaporated. The colourless, oily residue slowly solidified to a mass
of needles, which was crystallised from ethyl acetate. The hydroxy**
acid finally formed large prisms, which melted and evolved gas at
1 78—1 79°. On analysis :
0-1236 gave 0-2902 00, and 0-1073 H,0. 0 = 64-0 ; H- 9-6.
Cj^jHigOg requires 0 = 64*5 ; H - 9*7 per cent
The acid dissolved slowly in strong sulphuric acid, and on pouring
the colourless solution into water a flocculent, white mass separated.
This was collected, dried, and crystallised from light petroleum, from
which it was deposited in fern-like, camphoraceous crystals melting
at 177 — 178°; it had the properties of a lactone and was identical
with the campholid obtained by the action of strong sulphuric acid on
camphorenic acid (Forster, Trans., 1896, 69, 56). The hydroxy-add
was identical with that which Forster obtained on hydrolysing the
lactone (2oe. cU,).
Degradation qf Camphonic Aeid.
Oamphonic acid was first converted into tribromocamphonolactone
by the method described in a former paper (Trans., 1900, 77, 458),
and the lactone carefully purified by crystallisation from chloroform.
The pure substance, which was in the form of large crystals, was
finely powdered and covered with a 25 per cent, solution of potassium
hydroxide containing some alcohol. No considerable rise of tempera-
ture occurred. The whole was allowed to remain for a week, then
warmed on the water-bath for 15 minutes and poured into twice its
bulk of water, the alcohol being got rid of by repeated evaporation
with water nearly to dryness. The aqueous solution of the residue
was then acidified, and extracted repeatedly with ether in the usual
way. The ether, on evaporation, deposited an oily mass which slowly
soUdified. This was not closely examined, but was found to contain
Digitized by
Google
ACIDS OBtAlHEB I'BOlt a-DlBROMOCAHPHOB. 23
onlj a trace of bromine ; it exhibited marked ketonio properties and
its solution in alkalis had a distinct yellow colour.
The oil was dissolved in dilute sodium hydroxide, cooled to 0°, and
to this solution sodium hypobromite solution was added, in small quan-
tities at a time ; after each such addition^ a notable rise in tempera-
ture occurred, and the process was continued until, after the lapse of
15 minutes, hypobromite could be detected in the liquid. Sodium
sulphite was then added, the solution neutralised with hydrochloric
add and evaporated to a small bulk, a large excess of hydrochloric
add added, and the deposit of sodium chloride and bromide removed
by filtration and thoroughly washed with ether. The filtrate was ex-
tracted twenty times with ether, and the ethereal solution dried and
evaporated.
The oily residue thus obtained was dissolved in a little water, the
solution made alkaline with baryta water, and the very slight deposit
of insoluble matter removed. The filtrate was then heated to boiling,
when a second and much larger deposition of insoluble substance
occurred ; this was removed, washed with water, decomposed by means
of hydrochloric add, and the product examined.
The amount of add obtained from the precipitate was too small for
analysis as well as satisfactory examination. The substance was
found to melt at 137^ when very slowly heated, and at higher tem-
peratures when the capillary tube was plunged into sulphuric acid
already at that temperature. It formed an anilic add melting at 146^,
and a faintly alkaline solution of the ammonium salt gave no precipi-
tate with barium or calcium chloride in the cold, but a copious one on
boiling. In fact, the chemical and crystallographical properties of the
acid were identical in every respect with those of camphoronic
acid.
As it appeared, from the small quantity of camphoronic acid ob-
tained, that this substance did not constitute the principal product of
the oxidation, the acids in the filtrate from the barium camphoronate
were liberated, extracted with ether, and, after the usual process of
purification, were allowed to remain with a little water for several
months. During this time, the mixture became semi-solid, and was at
last spread on porous earthenware to drain. The solid portion was
crystallised repeatedly from boiling water, when it was finally obtained
in beautiful, lustrous prisms, which, after drying in the air for 2 days,
did not lose their brilliancy, but on exposure at 100° rapidly became
opaque and diminished in weight, owing to loss of water. On analysis:
0-1648 gave 0-3212 Hp and 0-0852 HjO. 0 = 498 ; H - 6-7.
O^H^O^ requires 0» 50*0; H-*6*6 per cent.
The equivalent was determined by titration against i\r/10 sodium
Digitized by VjOOQIC
24 LAPWORTH AND LENTON : THE CONSTITUTION OF THE
hydroxide in preaence of phenolphthalein. The number found
waa 107, whereas a dibasic acid of the formula O^Kifi^ requires
108.
The substance dissolved fairly readily in hot water, and in ethyl
acetate, alcohol, or acetone, but only very sparingly in benzene, and
was insoluble in light petroleum. It melted sharply at 246^.
The crystals from water were well-formed, rectangular plates or
stout prisms, belonging apparently to the rhombic system; in the
plates, the axial plane was parallel to the large face, and the direction
of the acute bisectrix was at right angles to the direction of greatest
length. The anhydrous substance, when molted on a glass slip be-
neath a cover-glass, solidified rapidly, forming radiate or fan-shaped
structures split up by linear air-spaces.
The function of the third pair of oxygen atoms was not easy to
determine, but as the substance gave no oxime, phenylhydrazone, or
acetyl derivative, it was surmised that a lactone ring was present in
the molecule. A small quantity of the acid was therefore heated to
boiling with a known excess of iT/^^ sodium hydroxide for half an hour,
and, after cooling, the excess of alkali remaining was determined. It
was found that no hydrolysis had occurred, the acid remaining dibasic,
' as before.
As the acid had a composition and a melting point identical with
those of *' /3-hydroxycamphoronic acid," obtained by Kachler and
Spitzer from camphoronic acid, it was thought possible that the two
substances might be identical, although ''^-hydroxycamphoronicacid"
is stated to be tribasic.
To obtain further evidence on the point, the acid was converted
into its ethyl ester by treatment with absolute alcohol and hydrogen
chloride. The substance thus obtained crystallised from a mixture of
ethyl acetate and light petroleum in thin, six-sided plates melting at
161^, which is exactly the melting point given by Kachler and Spitzer
for the ester of their acid. The identification of the acid was com-
pleted by preparing " /3-hydroxycamphoronic acid " by the method
described later.
In order to ascertain whether the oxidation of the hydrolytic pro-
duct of tribromocamphonolactone had proceeded further than to
^3-camphoranic acid, the syrupy mother liquors were extracted from
the porous plate by hot water, and subjected to distillation in a cur-
rent of steam for several hours in order to separate the volatile acids.
The aqueous distillate was then carefully neutralised with milk ot
lime, filtered, and evaporated nearly to dryness. A granular salt
separated towards the end of this operation, and was collected and
decomposed by hydrochloric acid, the acid being extracted with pure
ether in the usual way. The residue obtained on evaporating the
Digitized by
Google
ACIDS OBTAINED FROM a-DIBBOMOGAHPHOR. 25
ethereal solution was again converted into calcium salt, which was
collected and decomposed once more.
The acid which was thus obtained melted at 150 — 151°, formed an
anhydride which melted at 36 — 37°, and was identical with that pre-
pared from the trimethjlsuccinic]acid obtained by fusing a-camphoranic
acid with potassium hydroxide.
It appears, therefore, that the product obtained from tribromocam-
phonolactone by the above process consists mainly of )9-camphoranic
acid with small quantities of camphoronic acid and trimethylsuccinic
acid, and it is remarkable that no a-camphoranic acid could be de-
tected, although it is a substance which would probably be isolated
easily from such a mixture.
Braminatian of Camphoronic Acid,
The action of bromine on camphoronic acid takes place only under
pressure in closed tubes at 140° or thereabouts, and bromo-acids are
not obtained, as hydrogen bromide is at once eliminated, and a mix-
ture of the lactones of a- and )3-hydroxycamphoronic acids is formed.
Bredt {Annatenj 1898, 299, 158) did not succeed in brominating cam-
phoronic acid or any of its derivatives under the ordinary pressure,
but found it necessary to conduct the bromination in closed tubes and
to employ the purified anhydro-chloride.*
The authors have found that, as in so many other cases, the action of
bromine on the mixture obtained by treating the acid with phosphorus
pentabromide does not lead to satisfactory results, but that if phos-
phorus pentachloride is employed, an excellent yield of the mono-
brominated compounds can be obtained. The procedure was as follows.
Camphoronic acid was converted into the anhydro-acid by heating
it in a flask at about 130 — 140° until water vapour ceased to be
evolved. Thp cooled and powdered product was then carefully mixed
with phosphorus pentachloride (1 moL), heated on the water- bath for
half an hour, and allowed to cool. Bromine (1^ mols.) was then
added, the temperature gradually raised to 100° during about an hour,
maintained at that point for about 6 hours, and the product then
poured on to ice and allowed to stand overnight.
The granular product thus obtained consisted almost entirely of a
mixture of the anhydro-chlorides of a- and j3-bromocamphoronic acids,
and these may be converted into the bromo-acids by boiling with
nearly anhydrous formic acid.
To obtain the a- and ^-camphoranio acids, the mixture of bromo-
acids was boiled with water for several hours and the liquid then cooled,
and rendered faintly alkaline with baryta water. The barium salt of
a-campboranic acid separated almost at once as a fine, crystalline
Digitized by
Google
26 COAEN AND DAKlK: REDUOTIOK OF TBtHrTROBllNZEinE ARt}
powder, and the aoid obtained from this was used for the fweparation
of trimethjlsuccinic acid for purposes of identification.
The filtrate from the barium a-camphoranate was acidified with
hydrochloric acid, extracted with ether, and the /S-camphoranie add
examined. It was found to be identical in eyery respect with the
acid obtained by the former process.
A quantity of ^-camphoranio acid prepared in this way was crystal-
lised from water ; the clear crystals were then allowed to dry in the
air and at once analysed :
0-2921 gave 0-4612 00, and 01622 H^O. C-431 ; H-6-2.
CpHi20^,2HjO requires C = 42-8 ; H - 6-3 per cent.
The equivalent of the acid in the hydrated crystals was determined
by titration with i\^/10 sodium hydroxide, using phenolphthalein as
indicator. The number obtained was 130, whilst that required for a
dibasic aoid of the formula CoH^p^^SH^O is 126.
Chemical DiSparthent, Sohool of Pharkact,
17, Bloqhbbubt Square, W.O.
IV. — Note on the Reduction of Trinitrohenzene and
Trinitrotoluene with Hydrogen Sulphide.
By Julius B. Cohen and Henby D. Dakin.
Thb reduction of the 2:4: 6-trinitrotoluene was originally undertaken
with the object of producing an amino-group in the pararposition, and
by its removal of obtaining eventually 2 : 6-dinitrotoluene, a compound
which we required in the study of the chlorination products of toluene.
The redaction of trinitrotoluene to 2 : 6-dinitro-4-toluidine by means
of ammonium sulphide is described by Tiemann {Ber,^ 1870, 3^ 218)
and Beilstein {Ber., 1880, 18, 243), but the yield we obtained was
small, and we did not succeed in improving it or in suppressing a
quantity of tarry impurity which makes its appearance at the same
time. After many unsuccessful attempts to effect reduction with
ammonium sulphide and other agents, we tried a methyl alcoholic
solution of crystallised ammonium sulphide, passing in hydrogen
sulphide at the same time to displace air and keeping the whole well
cooled. We found that the reaction proceeded vigorously even when
the quantity of ammonium sulphide present was very far below the
theoretical amount. Finally, we simplified the method by adding a few
drops of concentrated ammonia to an alcoholic solution of the trinitro*
Digitized by
Google
TBIKITROTOLXTENB WITH HrDROGEN SULPHIDE. 27
oompoand and saturating with hydrogen sulphide. The product, filtered
from sulphur and poured into water, formed a bright yellow, crystalline
precipitate, which was not the anticipated dinitrotoluidine, but, as we
eventually discovered^ 2 : 4-dinitro-6-tolylhydrozylamine,
OH-NHi
Precisely the same reaction occurs with trinitrobenzene.
2 : i-Dinitro-6'tolythydroon/lamine*
Twenty grams of finely powdered trinitrotoluene were suspended in
about 100 c.c of absolute alcohol, about 0*6 c.o. of concentrated
ammonia was added, and the mixture cooled in ice. Hydrogen sulphide
was then passed in with frequent shaking. In a short time, the colour
of the solution deepened and the heavy crystals of trinitrotoluene,
which at first settled to the bottom, were soon replaced by a thick,
bulky, deep yellow precipitate, which filled the liquid. After about
an hour, no further increase in the'quantity of precipitate appeared,
and the mixture was warmed for a moment on the water-bath and
filtered quickly into a flask standing in ice. The precipitate was
washed with hot alcohol until the filtrate was colourless. A residue
of sulphur remained which weighed 6*3 grams. The alcoholic filtrate
deposited^ on standing, a mass of needle-shaped crystals, which were
separated and amounted to 4*7 grams. This fraction contained a small
proportion of dinitrotolylhydrozylamine, mixed with some compound
of high melting point, and melted indefinitely from 130 — 160°. The
product of high melting point is 2 : 6-dinitra-4-toluidine, for, on boiling
0*5 gram of this fraction for 2 hours with concentrated hydrochloric
add so as to convert the hydroxylamine compound into the insoluble
2 : 4-dinitro-6-toluidine, diluting and filtering, 0*3 gram of orange crys-
tals melting at 167 — 169°, which is the melting point of the 2:6-
dinitro-base, was deposited from the filtrate.
The filtrate was poured into water, which precipitated the bulk of
the hydroxylamine compound. It was filtered, washed with water,
and carefully dried. The weight was 11*5 grams. It was extracted
with successive quantities of benzene, in which it all eventually dis-
solved, each portion being kept separate. The last extracts yielded
crystals melting sharply at 143 — 145°, which did not change by
successive recrystallisations and were therefore regarded as pura The
Bubstance was analysed with the following results :
Digitized by
Google
28 REDUCTION OF TRIKITROBENZENE WITH HTDROQEN SULPHIDE.
«
0*2377 gave 39*5 c.c. moi&t nitrogen at 1 7"" and 764 mm. N « 19 -45.
0-1595 „ 27*25 „ „ 17° „ 760mm. N« 19-79.
C^H^OgNg requires N» 19*71 per cent.
A molecular weight determination by the boiling point method gave
the following result :
0*181 gram in 1 1 43 grams of benzene raised the boiling point by 0197°.
Mol. wt, found « 215 ; calculated « 213.
The compound reduces alcoholic silver nitrate, depositing a mirror ;
it also reduces Fehling's solution. It readily dissolves in alcohol, but
is less soluble in benzene and insoluble in light petroleum. From
benzene, it crystallises in rhombohedra. It dissolves in boiling dilute
hydrochloric acid unchanged and then crystallises in small, pale yellow
needles. On prolonged boiling, it becomes insoluble and changes to a
colourless, crystalline compound. The same result is much more rapidly
effected by concentrated hydrochloric acid. Half a gram of the
hydrozylamine compound, boiled with about ten times its weight of
strong hydrochloric acid for half an hour, yielded 0*3 gram of the
colourless substance.
2 : i-DinitroS-toluidine,
The colourless compound was crystallised from benzene, from which
it separated in needles melting at 212 — 213°. It was analysed with
the following results :
01857 gave 0*2884 CO, and 00575 H,0. 0 = 42*35 ; H = 3*44.
0*1475 „ 27*7 c.c. moist nitrogen at 26° and 759 mm, N= 20*8.
OyHyO^Nj requires 0-42*6 ; H = 3*5 ; N = 21*3 per cent.
0*220 gram in 8*1 grams of beozene raised the boiling point by 0*160°.
Mol. wt., found = 221 ; calculated - 197.
The substance is insoluble in sodium hydroxide solution or in dilute
hydrochloric acid. Neither stannous chloride nor sodium nitrite in
acid solution has any action on it. It dissolves unchanged in strong
sulphuric acid and is reprecipitated by water.
The conversion of p-phenylhydroxylamine into p-aminophenol by
mineral acids has been studied by Bamberger (Ber., 1894, 27, 1349),
and takes place by intramolecular rearrangement :
OgHg-NH-OH -> OH-OeH^-NHy
In the present case, the substance produced is not a phenol, but,
according to analysis, a dinitrotoluidine. The conversion must there-
fore be accompanied by the removal of oxygen.
OH8-OeHj(NO,)3-NH-OH - CH8-CeH,(N0,),-NH, + O.
Digitized by
Google
THE SYNTHESIS OF ALKTLTRICARBALLYLIC ACIDS. 29
Thiais precisely what happenfl, for if the hydrozylamine compound
be boiled with hydroohlorio a^id and indigo solution, the colour is
slowly discharged, or with hydrochloric acid and potassium iodide,
iodine is liberated.
As the amino-compound melts at 212 — 213°, it must be the 2 : 4-di-
nitro-6-toluidine, as the only other possible isomeride melts at 1 66 — 1 68°.
This substance has not been previously prepared. As we have seen,
the hydrozylamine compound acts both as a reducing and an oxidising
agent. It is also worthy of remark that, whereas ammonium sulphide
conyerts trinitrotoluene into 2 :6-dinitro-4-toluidine, in which the|>-nitro-
group is reduced, the action of hydrogen sulphide is to reduce the nitro-
group in the ortho-position.
1 : S-DinUro-d-phenylhydroocf/lamins.
This substance is prepared from 1:3: 5-trinitrobenzene in exactly
the same way as the tolyl derivative, but although the yield is
smaller, the product is more readily obtained in a pure state. It
forms dark orange crystals melting at 114 — 116°. Nine grams of
trinitrobenzene yielded 4*5 grams of pure hydroxylamine derivative.
Dinitrophenylhydroxylamine reduces alcoholic silver nitrate solution.
On analysis, the following result was obtained :
0*113 gave 208 c.c. moist nitrogen at 15° and 756 mm. N>» 21*53.
OeHjOjNg requires N = 21 -10 per cent.
On boiling with concentrated hydrochloric acid, the substance at first
passes into solution, but very soon a precipitate appears. The product
is then poured into water, boiled up, and allowed to crystalliBe. On
cooling, dark orange needles separate out, which melt sharply at
158 — 159°. This is the melting point of 3 : 5-dinitroaniline, with which
it is undoubtedly identical, a fact which serves to confirm the nature
of the reaction in the case of the tolyl derivative. We propose to con-
tinue this investigation.
The Tobkshire College.
V. — The Synthesis of Alkyltricarhallylic- Acids,
By William A. Bone and Chables H. G. Spsanklino.
In a previous communication (Trans., 1899, 75, 839), we described a
method for the preparation of ethyl esters of cyanosuccinic acid and
its alkyl derivatives ; for some time past, we have been investigating
a general method for the synthesis of alkyltricarballylic acids, based
Digitized by
Google
30 BONE AND SPBANKLINQ: THE SYNTHESIS OF
on the interaction of the sodium compounds of these ethyl cyano-
Buocinates with the ethyl esters of a-bromo-fatty adds, as indicated by
the general equation,
CR,.CO,Et i,^^^l
where B indicates hydrogen or any alkyl radicle.
A reference to the literature of the subject shows that tricarballylio
acid and its a-alkyl derivatives have been generally obtained by the
condensation of the sodium compound of ethyl malonate, or one of its
alkyl derivatives, with ethyl fumarate or maleate, and subsequent
hydrolysis of the resulting ester, as follows :
CH-COBt 9^<^0,Et),
CRNa{OO^t), + h**^"«^' = CH-0O,Bt .
' CH.CO,Et JUAEt
Auwers and others {Ber,, 1891, 24, 307, 2887) prepared a number
of a-alkyltricarballylic acids from ethyl fumaiate, but so far as we
know no attempt has been made to see whether dialkyltricarballylio
acids can be obtained by any similar method.
In 1896, Zelinsky {Ber.j 29, 333, 616) showed that three apparently
stereoisomeric ay-dimethyltricarballylic acids are obtained when the
highest fraction of the oil which results from the interaction of ethyl
cyanoacetate (1 mol.), sodium ethozide (2 mols.), and ethyl a-bromo-
propionate (2 mols.) in alcoholic solution is hydrolysed with sulphuric
acid. His investigation of the subject was incomplete and he has
since abandoned it.
Just as we were beginning our experiments, Haller and Blanc
(CoTnpt. rend,, 1900, 131, 19) synthesised aa-dimethyltricarballylic
acid from ethyl cyanosuccinate, but except in this one instance the
practicability of preparing tricarballylio acids from ethyl alkylcyano-
succinates has not so far been studied.*
As the result of a long and systematic investigation of the matter,
we have shown that anyalkyltricarballylic acid in which the alkyl radicle
or radicles occupy an a-position with respect to either' of the two
extreme carboxyl groups may readily be prepared by the method we
have indicated.
Besides the method just discussed, there is obviously another
possible way of passing over from an acid of the succinic to one of
the tricarballylic series, namely, by the interaction of an ethyl mono-
♦ Since this paper was written, however, Dr. H. A. D. Jowett has published an
account of the preparation of a-ethyltricarballylio acid from ethyl a-cyano-iS-ethyl-
succinate and ethyl bromoacetate (Trans. 1901, 79, 1846).
Digitized by
Google
ALKTLTBICABBALLYLIO ACIDS. 31
bromosuocinate with ethyl sodiooyanoacetate or malonate as represented
by the general equation,
OOjEt-CRa-ORBr-OOaEt + NaCH<^g^ =
COjEt-CIl3-CR(CO2Et)-0H(0]Sr)'CO2Et + NaBr,
where B represents hydrogen or an alkyl radicle (or radicles). Hitherto,
only tricarballylic acid itself has been prepared by this method (Emery,
Befr. 1890, 23, 3759), and we therefore extended our experiments in this
direction in order to ascertain whether this second method possesses
any advantages over the first, or vice verad. The results are very
decisive on this point, for they indicate that whereas the first {" cyano-
succinate ") method is a general one, the second can only be applied
in certain cases (owing partly to the circumstance that the bromination
of many substituted succinic acids does not proceed regularly, and
partly also to the tendency which some ethyl bromosuccinates exhibit
to lose hydrogen bromide and form unsaturated compounds). And,
further, even when the second method can be applied, the yields of
resulting tricarballylic acids are comparatively poor. Incidentally,
however, we have accumulated valuable information as to the bromina*
tion of alkylsuccinio acids, which will be briefly discussed later.
During the course of the experiments, we have added to the know-
ledge of the properties of various alkyltricarballylic acids, their
anhydro-acids and monomethyl salts, but have not been able con-
jointly to finish the scheme of work originally drawn up; the
results 80 far obtained are, however, sufficiently complete in themselves
to justify their publication. The investigation of this interesting and
important series of acids will be continued.
L Pbbparation of Ethtl Ctanotbioabballylatbs fbom Ethtl
Cyanosucoinates.
The method adopted may be briefly described as follows : To a solu-
tion of 5 '75 grams of sodium in alcohol is added one-fourth of a gram-
mol. of the ethyl cyanosuccinate j the sodium compound of the latter,
which is at once formed, remains in^solution. Bather more than the
calculated quantity of the a-bromo-fatty ester is then cautiously added
in small portions at a time. The interaction which follows is generally
very vigorous, much heat is developed, and sodium bromide separates.
The liquid usually becomes neutral after being heated for 30 to 60
minutes on the water-bath, after which it is poured into water and
the ethyl cyanotricarballylate extracted with eUier in the usual manner.
On fractionating the crude oil under diminished pressure (20 — 35 mm.),
a certain amount of it passes over at temperatures below 150^; the
Digitized by
Google
32 BONE AND SPBANKTJNG: THE SYNTHESIS OF
thermometer then rieee rapidly to ahout 200^, when the ethyl cyano-
tricarhallylate begins to distil.
The following are the particulars concerning the yields, Ac, obtained
in the various preparations, and the properties of the refractionated
ethyl cyanotricarballylates.
Fthyl Cyanotricarballf/late,
The yield of refractionated oil obtained from ethyl cyanosuccinate
and ethyl bromoacetate amounted to 75 per cent, of that theoretically
possible; under 28 mm. pressure, it boiled at 206 — 212^, On being
rapidly cooled, the distillate solidified. By dissolving the solid in the
minimum quantity of warm glacial acetic acid, then adding hot water
until a faint turbidity appeared, and setting the liquid aside to cool
slowly, the whole of the substance separated after some hours in
prismatic and transparent crystals which melted sharply at 40 — 4 P.
On analysis :
02213 gave 04421 COj and 01376 HjO. C « 5448 ; H - 690.
0*3116 „ 13*8 c.c. nitrogen at 18"" and 752 mm. N«5*07.
CijHjgOeN requires C = 54*73 ; H « 6*66 ; N = 4*91 per cent.
CH3-0H((X)2Et)-C-CH,-CO2Et
Ethyl a-Methj/lcyanatricarbcUlylale, y/\
CN COjEt
This substance may be prepared by the interaction of either the
sodium derivative of ethyl )9-methylcyanosuccinate and ethyl bromo-
acetate or of the sodium derivative of ethyl cyanosuccinate and ethyl
a-bromopropionate. The first named method is much the better of the
two, and the yield obtained by it amounted to 70 per cent, of the
theoretical. The refractionated oil boiled at 202 — 204^ under 23 mm.
pressure, had a density d 0^/4° » 1*1329, and a refractive index fiN»'»
1 '446 1 . On analysis :
0-2002 gave 04137 COj and 01316 H^O. 0 = 55*86 ; H = 7-31.
0-3102 „ 12-76 C.C. nitrogen at 15° and 771 mm. N = 4*87.
Cj^HjiOgN requires 0 = 56*18; H = 7*02 ; N = 4*68 per cent.
JEthyl ay-DimethylcyanolricarballylaUf
CH3*CH(C03Et)-C-CH(CH,)-C02Et
/\
ON OOjEt
The yield of refractionated oil obtained from ethyl ^methylcyano-
succinate and ethyl a-bromopropionate amounted to 65 per cent, of the
theoretical. It boiled at 208-— 210° under 30 mm. pressure, had a
Digitized by
Google
ALKYLTRICARBALLTLIC ACIDS. 33
density d 074''»M215, and a refractive index /Ay. =° 1*4484. On
analysis :
0-1976 gave 0-4146 CO^and 0-1353 HjO. C = 57-21 ; H = 7-61.
0-2863 „ 11-3 C.C. nitrogen at 16° and 758 mm. N = 459.
CjsHjjOgN requires C = 57*50 ; H = 7-34 ; N = 447 per cent.
(CH3)8C(C02Et)-C-CH,-CO,Et
Ethyl aa-Dimethi/leyanoiriccbrbdUlt/lcUe, /\
CN COgEt
This may be prepared either by the method adopted by Haller and
Blanc {loc. cii.) by the interaction of ethyl sodiocyanosuccinate and ethyl
a-bromoiaobntyrate, or by the interaction of ethyl )3/3-dimethyl8odio-
cyanosuccinate and ethyl bromoacetate. We have tried both methods
and find that the second is by far the better one ; the yield obtained
by it amounts to 55 per cent, of the theoretical, and if after fraction-
ating the crude product the portion of lower boiling point be again
heated with a small quantity of sodium ethoxide in alcohol, a further
quantity of the cyanotricarballylate is formed, bringing the total yield
up to nearly 70 per cent, of the theoretical.
The refractionated oil boiled at 202 — 204° under 17 mm. pressure,
had a density d 0°/4° « 1*1353, and a refractive index /ai^. =1*4503.
On analysis :
0*2006 gave 0*4206 COj and 01341 HjO. 0 = 5719; H = 7*43.^
CigHggOgN requires 0 = 57*50; H = 7-34 per cent.
£thyl ay-DiiBopropylcyanoiricarballylcUe*
This was prepared by the interaction of ethyl )9-t«opropylsodiocyano-
Buccinate and ethyl a-bromotsovalerate ; the experiment cannot be
properly carried out in an open vessel on the water-bath, as the reaction
only proceeds very slowly under these conditions. The mixture was
accordingly heated in soda-water bottles at 100° under pressure for
10 — 12 hours ; on fractionating the resulting crude oil under 15 mm.
pressure, we obtained from 61 grams of ethyl /3-tffopropylcyanosuccinate
originally taken the followiog fractions :
(a) Below 150°... 33 grams. (y) 205—215°... 10 grams.
{P) 150—205° ... 37 grams. (8) Above 215°, a few drops only.
The fraction (/}) contained large quantities of nitrogen and bromine,
and evidently consisted of a mixture of unchanged cyanosuccinate and
bromotffo valerate ; the fraction (a) contained no nitrogen to speak cf,
but a large quantity of bromine. They were accordingly mixed,
and after determining the amount of bromine in the mixture, the
VOL, LXXXI. A>
Digitized by VjOOQIC
34 BONK AND SPRANKLINQ : THE SYNTHESIS OF
corregponding quantity of sodium ethozide in alcoholic solution was
added to it. The whole was then heated in a soda-water bottle at 100^
for 10 hours, as beforoi and on fractionating the resulting oil a further
10 grams passed over at 205 — 215° under 15 mm. pressure. This was
mixed with the fraction y obtained in the first part of the experiment,
and the mixed oils were afterwards refractionated under 16 mm. presr
sure ; finally, 17 grams of a nearly colourless oil boiling at 208 — 212°
were obtained, which represent an 18 per cent, yield. On analysis :
0-2016 gave 04672 CO^ and 01596 H^O. 0 = 63-2 ; H = 88.
0-3259 „ IM C.C. nitrogen at 5° and 762 mm. N = 4-09.
^iftHsiOftN requires 0 = 63-51 ; H = 8-65 ; N = 3-90 per cent.
The oil was very thick and viscous, it had a density d 0°/4° =1-075
and refractive index fiff^ = 1 '4595.
Hydr6ly9%8 qfthe Ot&.— With the exception of ethyl ay-di«opropyl-
cyanotricarballylate, all the oils just described can be readily hydro-
lysed by boiling them in a reflux apparatus for 10 to 20 hours with
strong hydrochloric acid. This method we accordingly adopted. In
no case did any solid acid separate on cooling the liquid after all the
oil had dissolved, nor did we find it feasible to isolate the acids by
means of their calcium salts, a plan which answers very well in the
case of alkylsuccinic acids. We therefore resorted to the simple ex-
pedient of saturating the liquid in each case with ammonium sulphate
and then thoroughly extracting it with pure ether. After drying the
ethereal solution over anhydrous sodium sulphate and distilling off the
solvent, there remained an oily residue which usually solidified in the
course of a few hours. This was then either recrystallised from a
suitable solvent, or, in cases where it consisted of a mixture of isomeric
acids, was submitted to a suitable process for their separation. In one
case, namely, that of aa-dimethyltricarballylic acid, the oil which
remained after distilling off the ether did not solidify even after stand-
ing many days, and there was evidence that the hydrolysis had been
incomplete; on heating the oil with dilute (10 per cent.) hydrochloric
acid under pressure at 190° for a few hours, and afterwards evaporating
the liquid in a vacuum over strong sulphuric acid, the pure acid was
obtained.
II. Trioabballylig Acids, their Anhydbo-aoids and Monomethtl
Salts.
TriccvrhaUylic Add,
The acid, after being recrystallised from a mixture of glacial acetic
acid and chloroform, melted at 157 — 159°. On analysis :
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 35
01806 gave 0-2703 COj and 0-0762 H2O. U = 40-81 ; H = 4*69.
0-2164 silver salt gave 0-1407 Ag. Ag = 6602.
CeHaOg requires C = 40-90 ; H = 4-54 per cent.
OeHjOgAgj „ Ag = 66-16 per cent.
The dissociation constant of the acid is 0*022, a value practically
identical with that given by Walker (0*0224) for tricarballylic acid
(Trans., 1892,61, 707).
The calcium salts of this and the other acids of the tricarballylic
series described in this paper are readily soluble in cold water, but are
almost entirely precipitated when the solution is boiled. When a
25 per cent, solution of calcium chloride is added to a cold solution of
the neutral ammonium salt of tricarballylic acid, no separation of the
calcium salt occurs ; on boiling the solution, however, a dense, crystal-
line precipitate instantly appears which entirely redissolves when the
liquid is cooled again. The process of alternately precipitating and
' then redissolving the calcium salt may be repeated several times, but
the precipitate seems very gradually to become less soluble in cold
water. The behaviour of these calcium salts may be contrasted with
those of the succinic acids which, when once precipitated from a hot
solution of the ammonium salts, do not redissolve when the liquid is
cooled. Acids of the two series may be readily separated by means of
their calcium salts.
Anhydro-aeid, — ^The characteristic property which the tricarballylic
acids possess of yielding anhydro-acids (generally crystalline) when
they are boiled with acetyl chloride, or maintained at a temperature of
200° or upwards, was first noticed by Emery {Ber,, 1891, 24, 696) in
the case of tricarballylic acid itself. These anhydro-acids combine the
functions of a true anhydride and a monobasic acid, but it has nob yet
been shown whether in their formation from the tricarballylic acid,
the elements of water are eliminated from the afi- or the ay-carboxyl
groups, or, in other words, whether, say in the case of tricarballylic
acid, the anhydro-acid has the formula I or II.
C^Hj-COjH 9H2— COv
1. GK-QQhy^ 11. CH-COaH >0.
The best way of preparing these anhydro-acids is to dissolve the
tricarballylic acid in warm acetyl chloride, and, after boiling the solu-
tion for 2 — 3 hours in a reflux apparatus^ to distil off the solvent and
afterwards fractionate the residual liquid under diminished pressure.
In the case of tricarballylic acid, the anhydro-acid passed over between
216° and 226° under 46 mm. pressure; on cooling, it completely
solidified y and after recrystallisation from a mixture of chloroform and
glacial acetic acid, melted at 130 — 131°.
D 2
Digitized by
Google
36 BONE AND SPRANKLING: THE SYNTHESIS OF
01715 gave 02830 COg and 00615 HgO. C« 46-10; H = 3-98.
CeHgOg requires 0 = 45-57 ; H = 3-79 per cent.
The following investigation of the monomethyl salts of tricarballjlic
acid shows that the anhydro-acid has probably the constitution ex-
pressed by the formula I.
Monormthyl Salts, — There are two possible isomeric monomethyl
salts of tricarballylic acid, and three methods by which they may be
prepared, namely, (a) by the direct partial esterification of the acid ;
{b) by the partial hydrolysis of the trimethyl ester, and (c) by the
solution of the anhydro-acid in methyl alcohol. We have carefully
investigated these methods as follows.
(a) Direct Esterification of the Acid. — Five grams of the acid were
heated for 10 minutes with methyl alcohol containing just sufficient
dry hydrogen chloride to effect the esterification of one carboxyl
group. The excess of alcohol was then distilled off under reduced
pressure. A colourless oil "^ remained, which entirely dissolved in a
cold solution of sodium carbonate, and on being titrated with a
standard solution of barium hydroxide proved to have an acidity
corresponding to that of a methyl dihydrogen salt. The silver salt,
prepared by adding silver nitrate to a solution of the oil exactly
neutralised with dilute ammonia, was analysed as follows :
01726, onignition, gave 0-0920 Ag. Ag = 53-30.
OyHgOgAgg requires Ag = 53-47 per cent.t
There can be no doubt, therefore, that the oil had the composi-
tion of a methyl dihydrogen tricarballylate. The next question
to be decided was whether the oil was a single substance or a
mixture of the two isomeric monomethyl salts. We according deter-
mined its dissociation constant on the supposition that whereas a
single monomethyl salt would give a value for the constant K which
would remain practically the same for successive dilutions, a mixture
of two isomeric monomethyl salts would be indicated by well-marked
variations in the value of K on dilution. The results indicated that
the oil was a single substance.
* None of the methyl dihydrogen salts of tricarballylic acids investigated by us
are solids, so that it was impossible to purify them by crystallisation ; nor did
distillation under reduced pressure serve the purpose ; the evidence of their purity
is derived from a stHidy of their dissociation constants.
t Besides analysing the silver salts of the monomethyl dihydrogen tricarballylates
described in the paper, we always ascertained the acidity of each by titration with a
standard barium hydroxide solution. In each case, practically the calculated
amount of the alkali was required.
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 37
Diesoeit
i^t'on Constant
. K-
= 0-0075.
{Temj.
K 25°. )
V.
1^
m.
K.
7-62
8-27
00236
0-00748
15-24
11-65
00333
0-00753
30-48
16-42
0-0469
0-00756
60-96
2303
00658
0-00761
(b) Partial Hydrolysis of Trimethyl Triearballylate, — It was first of
all necessary to prepare the trimethjl ester from the acid by saturat-
iog a solution of it in methyl alcohol with dry hydrogen chloride in
the usual manner. The resulting oil was washed with a dilute sodium
carbonate solution, and distilled under 48 mm. pressure, when it passed
over at 205 — 208°. It was then quite colourless, having a density
d 0*^/4° = 11381, and a refractive index /Ana- 1'4398. On analysis :
0-2110 gave 0-3823 COg and 0-1246 H,0. C = 49-41 ; H = 6-56.
CgHj^Oe requires 0 = 49-50 ; H = 642 per cent.
Six grams of the oil were added to a quantity of potassium hydr-
oxide, dissolved in methyl alcohol, just sufficient to effect the hydrolysis
of two methoxy-groups. A drop of a methyl alcoholic solution of
phenolphthalein was added, and the liquid allowed to stand at the
ordinary temperature in an atmosphere free from carbon dioxide until
only the faintest pink tinge remained. Two drops of a methyl alcoholic
solution of methyl-orange were then added, and dry hydrogen chloride
passed into the well-cooled liquid until a pink colour first appeared.
The liquid was at once filtered from the potassium chloride which had
separated, and the filtrate evaporated in an exhausted desiccator over
sulphuric acid. The residual oil was dissolved in a slight excess of
sodium carbonate solution, and the liquid extracted with pure ether
in order to remove any trace of unchanged trimethyl ester. Finally,
the solution was acidified with hydrochloric acid, and again extracted
with pure ether. About 4*6 grams of a colourless oil were thus
obtained ; the silver salt was prepared and analysed as follows :
0-2610 gave on ignition 01389 Ag. Ag » 53*32.
C^HgOgAgj requires Ag = 63-47 per cent.
Its dissociation constant was then determined as follows :
Dissoetatitm constant. ^=0-00925. {Temp, 25°)
V. /iy. VI. K.
11*47 9*15 0*0320 000922
22-94 15*78 0*0457 000926
46-88 22-12 0*0632 000929
91-76 30-94 00884 000933
Digitized by VjOOQIC
38 BONE AND SPRANKLINO: THE SYNTHESIS OF
These numbers show that the oil was a single monomethyl di-
hydrogen tricarballylate and isomeric with that obtained by the direct
esterification of tricarballylic acid. Now it has been shown by Y.
Meyer, Sudborough, and other workers on the subject of esterifi-
cation that a carboxyl attached to a primary carbon atom is much
more easily esterified than one attached to a secondary carbon atom ;
consequently we must regard the monomethyl dihydrogen tricarballylate
obtained by the direct esterification of the acid as the a-compound,
C02Me*CH2*CH(C02H)*CH2'C02H, and therefore the isomeric ester
obtained by the partial hydrolysis of trimethyl tricarballylate must be
the ^-compound, CO2H-CH2-CH(CO2Me)-0H2-CO2H.
(c) By Solution of the Anhydro-ctcid in Methyl Alcohol — The anhydro-
acid was boiled for 45 minutes in a reflux apparatus on a sand-bath
with a quantity of pure dry methyl alcohol slightly in excess of that
required to effect its conversion into the monomethyl dihydrogen salt.
The liquid was then placed in a vacuum over sulphuric acid in order
to get rid of the slight ^excess of alcohol, and, after some days, the
residue was subjected to a further purification by means of sodium
carbonate as described under (6). The silver salt of the purified oil
was analysed as follows :
0-1167 gave on ignition 0-0611 Ag. Ag = 52-35.
C^HgOjAgj requires Ag = 52-47 per cent.
The dissociation constant of the monomethyl salt was determined as
follows :
Diasoc
iation constant.
K^
= 0-00945.
(Temp,
25°.)
V.
M'.
7?l.
K,
12-82
9-74
00342
0-00945
25-64
16-80
0-0480
0-00944
51-28
23-55
00673
0-00946
102-56
32-87
0-0939
0-00949
This shows, therefore, that the monomethyl dihydrogen tricarballyl-
ate obtained by dissolving the anhydro-acid in methyl alcohol is the
/S^jompound, C02H-CB[2-CH(C02Me)»CH2-C02H, and such as can
only result from an anhydro-acid of the constitution represented by
formula I (p. 35).
a-Methyliricarballylic Acids, CH3*CH(C02H)-CH(C02H)-CH2*C02H.
Since this acid contains two asymmetric carbon atoms, it exists in
two inactive forms, meso- and racemic. Auwers, von Meyenberg,
and Kobner (Ber., 1891, 24, 307, 2887) succeeded in isolating these
from the hydrolysed product of the condensation of ethyl fumarate
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 39
(1 mol.) with ethyl sodiomethylmalonate (2 mols). Their acids melted
at 134^ and 184^ respectively, and it was shown that the isomeride of
lower melting point is partially converted into the other on being
boiled with strong hydrochloric acid.
Oar experiments showed that when 26 grains of ethyl a-methyl-
cyanotrioarballylate were hydrolysed with strong hydrochloric acid in
the manner described, 1 6'5 grams of a miztore of isomeric acids were
obtained ; this only solidified after being kept for some days in ica
The substance, however, still contained a little nitrogen, and it was
therefore heated with dilute (10 per cent.) hydrochloric acid in sealed
tubes at 180—200° for 24 hours. The solid which finally remained
after evaporating the liquid to dryness melted between 160° and 170°.
On rapidly extracting this residue with small quantities of cold water,
one of the stereoisomeric acids dissolved, and the melting point of the
residue gradually rose to 179° and afterwards remained constant.
The washings, on evaporation, yielded a residue melting between 136°
and 145°, and when this was once again subjected to fractional extrac-
tion with cold water, an acid melting at 134—135° was obtained from
the first washings.
The acids were analysed and their dissociation constants determined
as follows :
tr&nB-AGid, m. p. 179°.
0-1624 gave 0-2633 COj and 00769 H^O. 0 = 4411 ; H = 5-35.
0-3007 silver salt gave 0'1897 Ag. Ag = 6310.
C^HioOg requires C -44-21 ; H = 5-26 per cent.
CyHYOgAgg „ Ag=. 63-39 per cent.
DissoGiatUm Cmatant. K^ 00322. (2W;>. 25° )
V,
m. K,
20-0 27-46 00767 00319
40-0 37-42 01069 0-0320
80-0 52-14 0-1489 0-0326
160-0 71-10 0-2032 0-0324
cia-Acidy m. p. 134—135°.
0-2113 gave 0-3425 00, and 01023 H^O. 0 = 44-09 ; H-6-38,
OyHioOg requires 0 = 44-21 ; H = 5-26 per cent.
Dissociation Conatcmt. Jr= 0-0480. {Temp. 25°)
V.
Mf
m.
K.
20-64
32-83
00938
0-0470
41-28
46-89
0-1311
0-0479
82-56
66'52
01900
0-0481
65-12
90-00
0-2583
0O486
Digitized by VjC
40 BONE AND SPRANKLINQ: THE SYNTHESIS OF
Anhydro-acid* — We have found that each acid on being dissolved
in acetyl chloride yields its oton liquid anhydra«cid, and that even after
being distilled under reduced pressure neither of the anhydro-acids
solidifies. Each anhydro-acid, however, with water yielded the acid
from which it was originally derived, and on heating the (ra9W-anhydro-
acid with acetyl chloride, or acetic anhydride, for several hours, it was
completely transformed into the ct^-isomeride.
Conversion qf cis- into trAnn-Acid, — We are able to confirm Auwers'
observation that the cis-eucid is partially converted into the trans-iso'
meride on being treated with hydrochloric acid under pressure at
190 — 200^ and find that equilibrium is established when 80 per cent,
is so transformed.
The behaviour of the anhydro-acids leaves no doubt as to the con-
stitution of the two isomeric acids from which they are derived ; the
ct^anhydride is the more easily formed from its acid, and is more
stable than the ^a^w-isomeride. The two acids, therefore, have the
following constitutions :
9H3 9H3
H-C-COaH H-C-COgH
•C-H
H-C-H H-
6O2H 6O2H
trans- Acid f m. p. 179°. cw-Acid, m. p. 134 — 135".
Monameihyl Salts, — So far we have only studied the monomethyl
salts of the cM-acid ; on determining the dissociation constants of
those prepared by the three methods described in the case of tri-
carballylic acid (pp. 36 — 38), we obtained practically identical numbers
as follows :
Mean values of
Monomethyl salt prepared by K at 25**.
Direct esterification of acid 0-00893
Partial hydrolysis of trimethyl ester 0 0085 7
Solution of anhydro-acid in methyl alcohol 0 '00888
At this stage of the inquiry we do not feel able to express any
decided opinion as to the interpretation of these results, and the matter
is receiving further investigation.
ay- DimethyUricaballylic AddSj
CH,-CH(C02H)-CH(COjH)-CH(CH3)-C02H.
By the hydrolysis of the oil of higher boiling obtained by the inter-
action of sodium ethoxide (2 mols.) ethyl cyanoacetate, (1 mol.), and
* Auwers did not study these substances.
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 41
ethyl a-bromopropionate (2 mols.), Zelinsky {loc, eit.) obtained three
isomeric acids, G^^^^q^ ^ foUows :
^for
M.p.
M. p. acid.
acid.
anhydro-acid.
(1)
203—204°
0-042
111—113°
(2)
175—176
0-054
129—130
(3)
148—149
0051
117—119
and although his experiments were not quite conclusive, he brought
forward evidence in favour of the view that the three acids are stereo-
isomeric ay-dimethyltricarballylic acids. If this be so, it is the only
instance of the synthetical formation of three inactive stereoisomeric
forms of a compound, C(abc)'C(ab)*C(abc), corresponding to the three
trihydroxyglutaric acids (the one laevorotatory, to which there is, of
course, a corresponding ' racemic ' acid, and the other two ' meso '-
inactive) obtained by Fischer (Ber., 1891, 24, 1842, 2686, 4222) by
the oxidation of ^-arabinose, xylose, and ribose respectively. The
point seemed to us sufficiently important to warrant further and
independent investigation. Briefly stated, our results are as follows.
When ethyl ay-dimethylcyanotricarballylate was hydrolysed by
boiling it with excess of strong hydrochloric acid for 12 hours, and
the resulting liquid extracted with ether, a solid mass was obtained
which melted gradually between 140° and 160°. By boiling it for
some time with successive small quantities of hydrochloric acid, part
dissolved, leaving finally an insoluble constituent which melted at
206—207°, and was not altered by further treatment with hydrochloric
acid. On concentrating the hydrochloric acid solution in a vacuum
over sulphuric acid, two sucessive crops of crystals were obtained which
melted at 1 70—1 88° and 1 7 1 —1 73° respectively. This second fraction
was twice recrystallised from strong hydrochloric acid and then melted
sharply at 174° We were unable to isolate any third acid either
from the first crop of crystals melting at 170 — 188°, or from the
hydrochloric acid mother liquors. The two acids melting at 206 — 207°
and 174° were analysed, and their dissociation constants determined,
as follows :
Acid, m. p. 206—207°
0-1706 gave 02951 00^ and 00922 HjO. 0 = 47-19 ; H = 601.
0-1064 silver salt gave 00654 Ag. Ag = 61*49.
OgHjgOg requires 0 = 47-58; H = 5-88 per cent.
OgHgO^Agg „ Ag = 61-70 per cent.
Digitized by
Google
42 BONE AND SPRANKLING: THE SYNTHESIS OF
Dissoeiation eonstcmt.
jr=
= 0-0445.
{T»mp.
25°)
V.
M^
m.
K.
33-71
4018
0-1148
0-0441
67-42
55-81
0-1594
00448
134-84
75-93
0-2169
0-0446
269-68
101-70
0-2911
00443
Add, m. p. 174°.
01 926 gave 0-3302 COj and 01038 H^O. C « 47-49 ; H = 599.
01099 silver salt gave 0-0676 Ag. Ag = 61-56.
CgHijOg requires C -47-58 ; H = 5-88 per cent.
CgHgO^jAgj^ „ Ag = 61-70 per cent.
Dissociation constant.
jr= 00545.
{Temp, 25°. )
20-7 35-03
41-4 48-67
82-8 66-72
m.
0-1002
0-1387
0-1906
K.
0-0559
0-0540
0-0642
166-6 90-60 0-2588 0-0646
Anhydro-aeids. — Each acid dissolved in acetyl chloride, yielding its
own solid anhydro-acid ; that obtained from the acid of higher melting
point (206—207°) fused at 110—112°, and that from the acid of lower
melting point (174°) fused at 130°.
Mtaual Conversion.— (!) The acid melting at 206—207° was heated
for 4 hours at 180° under pressure with acetic anhydride, and from
the dark-coloured liquid the acid melting at 174° was recovered by
means of its potassium salt. It is clear, therefore, that the anhydro-
acid of the former is at high temperatures converted into that of the
latter.
(2) The acid melting at 174° was partially converted into that melt-
ing at 206 — 207° by heating it with strong hydrochloric acid at 210°
for several hours.
There can be no doubt, therefore, that these two acids are identical
with two of the acids obtained by Zelinsky, and, further, that they
are stereoisomeric.
Zelinsky hydrolysed the oil from which he obtained his three acids
with sulphuric acid ; we therefore hydrolysed another portion of the
ethyl ay-dimethylcyanotricarballylate by boiling it with 50 per cent,
sulphuric acid. The operation was rather a slow one, and was
only complete after 2 or 3 days. On cooling the liquid a crop of
crystals, A, separated, melting at 190° or thereabouts ; on further
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 43
concentrating the mother liquor, two more crops of crystals were
obtained, namely, B, melting at 170 — 180^, and C, at temperatures be-
low 139^. From fractions A and B, by further purification, were obtained
two acids melting at 204 — 206° and 174 — 176°, identical in all respects
with those obtained in the earlier experiments. Fraction 0 was sub-
mitted to two or three recrystallisations from water ; its melting point
gradually rose to 143° and then remained constant ; analysis showed
that it had the empirical formula CgH^^O^. On being treated with
acetyl chloride, it yielded an anhydro-aoid, CgH^QOg, melting quite
sharply at 116 — 117°, which with water regenerated the original acid.
Disaoeiation constant
A"= 00572.
{Temp, 25° )
V. n^
m.
K.
21-78 36-97
0-1056
0-0673
43-56 5711
01460
00573
87-12 69-80
0-1994
00570
174-24 94-29 0-2693 0056J
The most curious point about this acid is that on being warmed
with strong hydrochloric acid it is very quickly and quantitatively
transformed into the acid melting at 174°; for example, on recrystal-
lising a portion of it from warm hydrochloric acid its melting point
rose to 160 — 164°, after a second recrystallisation to 171 — 173°, and
after a third to 174°.
The question therefore arises : Is this acid melting at 143° a third
inactive stereoisomeric form of ay-dimethyltricarballylic acid, or is it
merely a molecular mixture of the other two forms 9 Three facts are
in favour of the first view, namely (1) that it yields its own anhydro-
acid with acetyl chloride; (2) that its dissociation constant varies
very little with successive dilutions, and is higher than the correspond-
ing values for the other two acids ; and (3) that treatment with strong
hydrochloric acid converts it into the second (174°) acid, whereas the
acid melting at 206° remains absolutely unchanged when heated with
hydrochloric add under the ordinary pressure.
One of the three acids must be the racemic {trcmS') form, the other
two must be meso-modifications of ay-dimethyltricarballylic acid
which we may distinguish as the cm^- and ci^^-acids. Since the an-
hydro^acid of the acid melting at 174° is the most stable of the three
anhydro-acids at high temperatures, it is probably one of the da- (meso-)
forms ; the other ds-iovm is, therefore, the acid melting at 143°. The
someride having the highest melting point must therefore be the
iranS' or racemic form, as under :
Digitized by
Google
44 BONE AND SPRANKLING: THE SYNTHESIS OF
H,
s
9H3 9^8 9
H-Cp-COaH H-C-CO^H H-C-COaH
H-Q-002H H-A-COaH . C02H-6-H
COgH-^-H H-t-COjjH H-C-COjH
Trans- or racemic, cwj- and cisj- Acids (meso).
m. p. 206— 207^
At present we are unable to decide which of the two acids, melting
at 174° and 143° respectively, is the cis^- and which the cis^'f orm.
The further investigation of the subject is in hand, however.
aaDimetkyUHcarballylic Add, (CHg)2C(C02H)-CH(C02H)-OH2-002H.
This acid is a very interesting member of the series, inasmuch as it
is an oxidation product of pinonic acid (Tiemann and Semmler, JBer,,
1895, 28, 1349), also of fenchone (Gardner and Cockburn, Trans.,
1898, 63, 710) and camphoceenic acid (Jagelki, Ber., 1899, 32, 1498).
The acid we obtained by hydrolysing ethyl aa-dimethylcyanotricarb-
ally late melted at 143°. On analysis :
0-2136 gave 03722 CO2 and 0-1145 H2O. 0 = 47*51 ; H = 5-96.
OgHjgOg requires 0 = 47*58 ; H = 5*88 per cent.
Dissociation constant. Jr= 00318. (Temp. 25°)
V. fXp. m. A'.
23-67 2916 0-0833 00320
47-34 40-39 01154 0-0318
94-68 54-67 01562 00315
189-36 75-12 0*2146 00309
The aniif/dro-acid, recrystallised from chloroform, melted at
135 — 136°. On analysis :
0-2022 gave 0-1900 COj and 01003 H2O. 0 = 51 -25 ; H = 5-50.
OgHjoOg requires 0 = 51*61 ; H = 5-37 per cent.
The trimethyl ester was a thick oil boiling at 170 — 174° under
33 mm. pressure; it had a density d 0°/4° = 1-1403 and a refractive
index fjLjg^ = I'i^ll.
01829 gave 03588 OO2 and 0-1236 HjO. 0 = 535 ; H = 7-51.'
OiiHjgOg requires 0 = 53-7 ; H = 7-32 per cent.
Monomethyl Salts. — ^There are three possible isomeric monomethyl salts
of this acid, namely, (a) (CH3)20(002H)*CH(002H)-OH2-002Me, (6)
(CH3)20(0O2H)-0H(0O2Me)-CH2-0O2H, and
(c)(OH3)20(002Me)-OH(002H)-OH2-002H.
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 45
We prepared ikionomethyl salts from the acid, the trimethyl ester, and
the anhydro-add, by the methods already desoribed (pp. 36 — 38), with
the following results :
(i) By Direct Esierification of the Acid. — A colourless oil.
0-1526 of its silver salt gave 0076 Ag. Ag = 49-79.
CgHj^OgAgg requires Ag = 50*00 per cent.
Duaoeiation ecnatant.
K.
= 0-0180.
{Ttmp.
25°)
V. 11^
OT.
K.
31-2 25-34
0-0724
00181
62-4 3510
0-1003
00179
124-8 48-52
01386
00179
249-6 65-92
01883
00175
There can be no doubt, therefore, that the oil was a single sub-
stance, and from the fact that it was formed by the direct esterification
of the acid (which contains only one primary CO^H group), we may
conclude that it has the formula (a).
(ii) By Partial Hydrolysis of the Trimethyl Ester, — A colourless oil.
0-2038 of its silver salt gave 0-1022 Ag. Ag = 5012.
Dissociation constant.
Jr= 000865. (5
"emp. 25° )
V, fl^
m.
K.
8-95 9-63
00275
0-00870
17-90 13-51
0-0386
000866
35-80 18-94
00541
000863
71-60 26-28
0-0751.
0-00859
There can be no doubt that the oil was a single substance and
quite different from that obtained by direct esterification of the acid,
but we have no means of judging at present which of the two
formuliB, (b) and (c), represents its constitution.
(iii) Frovi the Anhydro-cund, — A colourless oil.
0-1286 of its silver salt gave 0-0642 Ag. Ag = 49-91 per cent.
£■=00186.
(Tmnp. 25°.)
r. /v
m.
K.
12-52 15-89
00454
0-0189
25-04 2314
0-0661
0-0186
5008 31-96
00913
0-0183
This monomethyl salt, therefore, is probably the same as that
obtained by the direct esterification of the acid. Comparing now the
Digitized by
Google
Acid. Trimethyl ester.
Anhydro-acid.
00075 0-00925
0-00945
O'OISO 0O0865
0-01860
46 BONE AND SPRANKLING: THE SYNTHESIS OF
values for JT, determined for tricarballylic and]aa-dimethyltricarballylic
acids and their monomethyl salts,
Monomethyl salt from
Acid.
Tricarballylic 0022
oa-Dimethyltricarballylic 0032
we see that in both cases the monomethyl salt obtained by the direct
esterification of the acid is quite different from that obtained by the par-
tial hydrolysis of the trimethyl ester ; but that the salt obtained from
the anhydro-acid is in the one case identical with that obtained from
the trimethyl ester, and, in the other case, with that yielded by direct
esterification of the acid.
ay- Diisopropi/Uricarballylie Acids.
These acids were prepared with the view of determining whether
the substitution of two wopropyl groups in ay-positions has an
influence upon the dissociation constant of tricarballylic acid at all com-
parable with that exerted upon the constant of succinic acid by the
symmetrical substitution of two hydrogen atoms by i^opropyl groups
(compare Trans., 1900, 77, 667).
Ethyl ay'dii8opropylci/anotrica7*ballylcUe is a difficult oil to hydrolyse ;
we found it best to perform the operation in two stages, namely, (1),
with alcoholic potassium hydroxide, and (2), with 50 per cent, sulphuric
acid. Finally, on extracting the acid liquid with ether we obtained
from 17 grams of oil 9'8 grams of a solid mixture of stereoisomerrc
acids. These were difficult to separate, but on dissolving the mixture
in water, saturating the solution with hydrogen chloride, and allowing
it to stand for some time, we were able to resolve it into fractions of
higher and lower melting point, by reason of the greater solubility of
the latter. Two pure stereoisomeric acids were finally obtained,
melting at 173^ and 156° respectively. Each yielded its own liquid
anhydro-acid, but we had not sufficient material to investigate these
properly, and it is possible that, had we been able to purify them
further they would have solidified.
The acid of higher melting point was transformed into the anhydro-
isomeride on being boiled for many hours with acetyl chloride*
nd was analysed, and its dissociation constant determined as
. p. 173°.
gave 0-4707 CO.^ and 01653 HgO. 0 = 55-18; H-7-90.
silver salt gave 0*0943 Ag. Ag = 56*40.
C^gHaoOg requires C = 55 38 ; H = 7*69 per cent.
CijHj^OgAgg „ Ag « 55-90 per cent.
Digitized by
Google
ALKYLTRICARBALLYLIC ACIDS. 47
Dissoci
ation eotutant.
K=
= 0-193. (Tmtp. 26°)
V.
ih-
OT. K.
171-5
151-6
0-4332 0-193
3430
191-5
0-5469 0-192
686-0
2331
0-6660 0-194
1372-0
270-7
0-7731 0192
Acid, m. p. 156°.
0-1971 gave 0-3982 COj and 0-1397 HjO. 0 = 5509 ; H = 7'88.
0-2018 silver salt gave 01 130 Ag. Ag = 55-99.
^12^20^6 requires 0 = 55-38 ; H = 7*69 per cent.
OijHiyO^Agg „ Ag = 55*90 per cent.
Diaioeieaion eonatant.
K=
=0-1625.
(Temp.
25°)
V.
lu.
m.
jr.
95-9
113-7
0-3241
0-1621
191-8
148-8
0-4250
0-1628
383-6
188-3
0-5380
0-1633
767-2
230-0
0-6570
0-1640
If we compare these values with those for tricarballjlic acid (0*022)
and a-i«opropyltricarballylic acid (0*0434 — Auwers, loc. dt.), we see at
once that, in both cases, the introduction of the two taopropyl radicles
has had a very marked ' raising ' effect on the dissociation constant, but
there is no such enormous difference between the constants of the two
isomerides as there is between those of da- and inma-a-diiso^ro^yl'
succinic acids.
The subject of the variation of dissociation constants with molecular
constitution in this series of acids presents many interesting features,
and will be discussed more fully in a future communication.
III. Tbicabballtlic Acids fbom Ethyl Bbomosuccinates.
Ab already stated, we have studied the preparation of ethyl cyano-
tricarballylates by the interaction of ethyl bromosuccinates with the
sodium compound of ethyl cyanoacetate, and have been able to carry
it out in the following instances.
TriccvrhaUylie Acid.
The best method for preparing tolerably pure ethyl bromosuccinate,
is to act on succinic anhydride with the calculated quantity of dry
amorphous phosphorus and bromine, to form the dibromide of mono-
bromosuccinic acid, and afterwards to pour the product into excess of
Digitized by
Google
48 BONE AND SPRANKLINO: THE STNTHESIS OF
alcohol. In this way we obtained an 80 per oent. yield of ethyl
bromosuccinate boiling at 140 — 143° under 29 mm. pressure. On
condensing this with the calculated quantity of ethyl sodiocyanoaoetate
suspended in alcohol, a 70 per cent, yield of ethyl cyanotricarballylate
resulted ; when hydrolysed, this yielded tricar bally lie acid, melting at
167—169°. On analysis :
0-2023 gave 03028 CO^ and 00864 H2O. C = 40-81 ; H = 4-69.
CgHgOg requires 0 = 40*90 ; H = 464 per cent.
a-MeihyUriaxrhallylic Acid.
On brominating 26 grams of monomethylsuccinic acid by the Hell-
Yolhard-Zelinsky method, pouring the product into alcohol, and
extracting the resulting bromo-ester with ether, we obtained 38 grams
of an oil which distilled over at 161 — 163° under 44 mm. pressure. On
analysis :
0-4166 gave 02798 AgBr. Br = 29-07.
CgHj^O^Br requires Br = 29*96 per cent.
There are two possible isomeric ethyl monobromomethylsuccinates,
namely, (a) CH8-OBr(C02Et)-CH5-COjEt, and
. (/3) CH,-CH(C02Et)-CHBr-C02Et.
If the oil obtained by the method first indicated had the formula (a),
then on condensing it with ethyl sodiocyanoacetate we should obtain
the cyano-ester of /3-methyltricarballylic acid ; on the other hand, if it
had the constitution (fi), it would under similar treatment yield the
cyano-esters of the a-methyltricarballylic acids.
On trying the experiment we obtained a 60 per cent, yield of an ethyl
methylcyanotricarballylate (b. p. 236 — 246° under 30 mm. pressure)
which, on hydrolysis with strong hydrochloric acid, yielded the two
a-methyltricarballylic acids, melting at 177 — 180° and 134° respec-
tively, but not a trace of any j3-methyltricarballylic acid. Hence the
ethyl monobromomethylsuccinate obtained when methylsuccinic acid is
brominated in the manner described has the constitution
CH,-CH(C02Et)-CHBr-C02Et.
The two a-methyl tricar bally lie acids obtained were analysed as
follows :
Add, m. p. 177—180°.
0-1921 gave 0-3174 COj and 00963 H2O. C - 440 ; H = 661.
0*3011 silver salt gave 0*1900 Ag. Ag = 63*10.
Acid, m. p. 134°
0*2614 gave 0*4228 OOg and 0*1280 HgO. 0 = 44*11 ; H = 6*44.
0*1991 silver salt gave 0*1266 Ag. Ag = 63*06.
C^HioOg requires 0 = 44*21 ; H = 6*26 per cent.
CyHyOgAgg „ Ag = 63*39 per cent.
Digitized by
Google
ALKYLTRICABBALLTLIC ACIDS. 49
oA'DifYUlhyhricarhailylic Acid,
Twelve grams of as-dimethylsuccinic acid, on bromination by the
Hell-Yolhard-Zelinsky method, yielded 20 grams of monobromo-ester
boiling at 1 59 — 1 64^ under 70 mm. pressure. On analysis :
0-2442 gave 0 1589 AgBr. Br = 27-69.
Cj^HiyO^Br requires Br =■ 28-47 per cent.
On condensing this with the calculated quantity of ethyl sodiocyano-
acetate, we obtained a 50 per cent, yield of ethyl oa-dimethylcyanotri-
carballylate boiling at 210 — 220° under 35 mm. pressure. When
hydrolysed with strong hydrochloric acid, this yielded aa-di methyl tri-
carballylic acid melting at 140 — 142°. On analysis :
0-2611 gave 0-4554 CO2 and 0-1408 H2O. C = 4759 ; H = 5-99.
0-2122 sUver salt gave 01308 Ag. Ag = 61-61.
CgHijOft requires C = 47 -58 ; H = 5-88 per cent.
CgH^OgAgg „ Ag = 61 -7 1 per cent.
We have also studied the bromination of ci«-9-dim6thyl8Uccinic acid
by methods similar to those already described. Many workers have
investigated the bromination of this and the isomeric ^an^-acid under
varying conditions and with widely different results. Hell and Roth-
berg {Ber., 1889, 22, 66) state that both acids behave normally on
bromination, yielding ci^-monobromodimethylsuccinic acid; Zelinsky
and Krapivin {Ber., 1889, 22, 390), Bischoff and Yoit {Bw., 1890, 23,
390), and Auwers and Imhauser {Ber.^ 1891, 24, 2^33), on the contrary,
assert that neither acid can be brominated under any conditions, and
that the substance which results is always the anhydride of pyro-
cinchonic acid (m. p. 95°), so that if any monobromo-anhydride (or acid) is
momentarily produced it must at once lose hydrogen bromide as follows :
CH3
CH3
;g-;> = HB. . -;:§:>.
Our own experience shows that when a mixture of ci^^-dimethylsuccinic
acid and amorphous phosphorus is treated with the quantity of dry
bromine required to form the dibromide of the monobromo-acid, bromina-
tion certainly takes place, for on pouring the product into alcohol, and
extracting and fractionating the resulting ester, we obtained a very
fair yield of a bromo-ester containing 26-86 per cent, of bromine
(CioHjYO^Br requires Br =28-47 per cent.).
On condensing this bromo-ester with ethyl sodiocyanoacetate, sodium
bromide was at once eliminated, but the product obtained was not a
cyanotricarballylic ester, and up to the present we have not been
VOL. LXXXr. E
Digitized by
Google
50 BONE AND SPRANKLINQ:
able to ascertain what really happened. The subject is still under
investigation.
In conclusion, we wish to state that one of us is investigating the
preparation and properties of tri- and tetra-methyltricarballylic acids.
The cost of the materials required for this investigation has been
largely defrayed out of grants from the Research Fund of the Society.
The Owrns Collbojs,
Manohestbr,
VI. — The Bromination of Trimethylsuccinic Acid and
the Interaction of Ethyl Bromotrimethylsuccinate
and Ethyl Sodiocyanoacetate.
By William A. Bone and Charles H. G. Spranklinq.
In connection with our investigations on the synthesis of alkyltricarb-
allylic acids, we have recently studied the bromination of trimethyl-
succinic acid, and the interaction of ethyl bromotrimethylsuccinate
and ethyl sodiocyanoacetate. Some years ago^ one of us, in conjunction
with Professor W. H. Ferkin, jun., unsuccessfully attempted the
synthesis of t-camphoronic (oajS-trimethyltricarballylic) acid by a
method involving this reaction, which is expressed by the following
equation :
(CH,)jC(OOjEt)-CBr(CHa)-COjEt + NaOH(CN)*C02Et- NaBr
+ (CH3)jC(CO2Et)-C(CH3)(C02Et)-0H(CN)-0O2Et.
As a matter of fact, a crystalline acid melting at 137° and quite
different from «-camphoronic acid was finally isolated from the hydro-
lytic products of the resulting cyano-ester, but the quantity obtained
was too small to allow of a satisfactory investigation of its properties
being made. The study of the subject was for the time being aban-
doned, partly on account of the difficulty experienced in preparing a
sufficient quantity of trimethylsuccinic acid by any method then
known, and also because Perkin and Thorpe succeeded in synthesising
i-camphoronic acid by another method in 1897 (Trans., 71, 1169).
SinoCi however, the preparation of large quantities of pure trimethyl-
succinic acid is no longer a difficult matter, we decided to reinvestigate
the subject, and, if possible, to ascertain the cause of the earlier failure
to synthesise camphoronic add.
In 1898, Gustav Komppa (Acta Soc. ScierU, Fenn», 24, 1 ; also Abstr.,
Digitized by
Google
THE BROMINATION OF TRIMETHYLSUCCINIC ACID. 51
1899, i, 419) tried to prepare bromotrimethylsuccinic acid by the action
of phosphorus pentabromide on the corresponding hydrozy-acid. He
was, however, unable to isolate any pure product from the complex
mixture of substances obtained, and his experiments indicated that the
three methyl groups in hydroxytrimethylsuocinio acid greatly hinder
the replacement of hydroxyl by bromine.
The rasults of our experiments may be briefly stated as follows :
(1) When trimethylsuocinic acid is heated with the calculated quantity
of bromine under pressure at 1 30^, it is quantitatively converted into
the characteristic white crystalline bromotrimethylsueciniG anhydruU
melting at 197 — 198° ; it is not possible to obtain the pxire bromotri*
methylsuccinic acid by dissolving this anhydride in hot water, since
partial decomposition, with loss of hydrogen bromide, occurs during
the process.
(2) If the bromination of trimethylsuocinic acid be carried out
according to the Hell-Yolhard-Zelinsky (phosphorus and bromine)
method and the product poured into alcohol, a mixture of bromo-
anhydride and ethyl bromotrimethylsuccinate results, from which it is
very difficult to obtain the latter substance in a tolerably pure state.
(3) Both the bromo-anhydride and ethyl bromotrimethylsuccinate
readily lose hydrogen bromide under the influence of an alkali. By
heating the bromo-anhydride with diethylaniline and subsequently
pouring the liquid into a solution of potassium hydroxide, we obtained
the potassium salt of methylenedimethylsucdnic cusid, C^H^qO^. The
ethyl ester of this acid very readily combines with hydrogen bromide,
forming a bromo-ester, C^H^^O^Br, which, so far as we have been able
to ascertain, seems to be identical with the ethyl bromotrimethyl-
succinate, (OH3)20(C02Et)'OBr(CH8)*C02Et, prepared directly from
trimethylsuocinic acid.
We would point out in this connection that Yincenzo Faolini
{GazxeUa, 1900, 30, ii, 497), by acting on ethyl hydroxytrimethyl-
sucdnate with phosphorus pentachloride, has obtained the ethyl ester
of an acid, O^HjqO^ melting at 153 — 154°. Since this acid neither
absorbed bromine or hydrogen bromide at the ordinary temperature,
nor decolorised cold alkaline permanganate, he concluded that its
molecule was not unsaturated, and described it as dimethyltrimethyl-
enedicarboxylie acid. The formation of such an acid he explained
by supposing that the ethyl chlorotrimethylsuccinate formed in the
first instance by the action of phosphorus pentachloride on the ester
of the hydroxy-acid at once loses hydrogen chloride, the elimination of
which takes place between the chlorine and a hydrogen atom of a
methyl group attached to the other carbon atom, so that 'ring-
formation ' occurs thus,
E 2
Digitized by
Google
52 BONE AND SPRANKLING :
oEt
COgEt ' CO.
•C-CHg , CHg-C-C
•C-OH ~^ CHg-C-C
iOgEt COj
His acid certainly appears to have properties quite different from
those of methylenedimethylsuccinic acid, and we are therefore led to the
interesting conclusion that the elimination of hydrogen bromide from
a bromotrimethylsuccinic derivative and of hydrogen chloride from a
chlorotrimethylsuccinic molecule may occur in two entirely different
ways. This is a point which certainly deserves further investigation.
(4) Ethyl bromotrimethylsuccinate reacts with ethyl sodiocyano-
acetate, yielding the cyano-ester of a tribasic acid, CgH^^O^, melting at
137 — 138°, and isomeric with i-camphoronio acid (m. p. 169 — 172°).
The formation of such an acid can be explained on the supposition
that ethyl bromotrimethylsuccinate loses hydrogen bromide, forming
ethyl methylenedimethylsuccinate, which at once condenses with the
ethyl cyanoacetate as follows :
(CH3)2C(C02Et)-C(C08Et):CH2 + CH2(CN)-C0jEt =
(CH8),C(COjEt)-CH(C02Et)-CHjj-OH(CN)-C02Et.
If this interpretation of the matter be correct, the acid, CgH^^O^,
obtained on hydrolysing the product with hydrochloric acid would be
aa-dimethylbfUan9-^pS-trie<»rb(>xyUc add,
(CH3)2C(COjH)-CH(COjH)-CH,-CHj-COjH;
the results of a ' potash fusion ' of the acid, which yielded acetic and
trimethylsuccinic acids, are consistent with this view of its constitution.
EXPEBIMENTAL.
Bromination of Trimethylsticcinic Acid. Formation qf BromotriiMtJiyl'
succinic Anhydride atid Ethyl BromotrimsthylsmDciruUe.
(1) ffdl-VoUiard'Zelinsky Method, — We have at various times carried
out experiments in which rather more than the calculated quantity
of dry bromine was slowly dropped on a well cooled mixture of trimethyl-
succinic acid and the theoretical amount of dry amorphous phosphorus.
In each case, a vigorous reaction ensued accompanied by a strong
evolution of hydrogen bromide, which only ceased after the mixture
had been heated on the water-bath in a reflux apparatus for 6 or 8
hours.
On dropping the resulting brown liquid into an excess of alcohol
(well cooled in ice) and afterwards pouring the alcoholic solution into
^ large excess of water, a heavy brown oil separated, which was
Digitized by
Google
THi. BROMINATION OF TRIMETHYLSUCCINIC ACID. 53
extracted with ether in the usual manner. After washing the ethereal
solution with dilute (5 per cent.) sodium carbonate solution, then
drying it over anhydrous sodium sulphate, and finally distilling off the
ether, there remained a heavy reddish-brown oil which appeared to
decompose when we tried to distil it under reduced pressure.
Analyses showed, however, that samples prepared at different times
invariably contained from 30 to 33 per cent, of bromine, or consider-
ably more than that required for ethyl bromotrimethylsuccinate,
OiiHijO^Br, namely, 27*12 per cent.
Bromotrimethylsuceinio Anhydride, — After the oil had stood for some
weeks in an exhausted desiccator over sulphuric acid, yellowish crystals
began to separate ; these were removed from time to time, and after
being pressed on a porous plate were recrystallised from hot benzene.
When qnite pore, they melted sharply at 197 — 198°. The substance
was insoluble in cold water or a cold solution of sodium carbonate, but
readily dissolved in a warm solution of potassium hydroxide without,
however, any formation of alcohol. ' It was, therefore, neither an
acid nor an ester ; the following analysis showed that it had a com-
position corresponding to that of bromotrimethylsuccinic anhydride,
and a further study of its properties showed it to be this substance.*
01691 gave 02368 COj and 0-0600 H^O. 0 = 38-19 ; H-3-95.
0-2364 „ 0-2036 AgBr. Br = 36-64.
C^H^OgBr requires 0 = 3806 ; H = 407 ; Br = 36*20 per cent.
Ethyl BromotrimUhylsuocinate, — The crude oil was kept for several
months until no further separation of bromoanhydride occurred : on
analysing the residual oil, we obtained, for two different preparations,
the following numbers :
(1) 0-5202 gave 0-3525 AgBr. Br = 2882 per cent.
(2)0-2528 „ 01713 AgBr. Br = 28-83 „
It, therefore, still contained 1-7 per cent, more bromine than that
* Assuming that the dibromide of bromotrimethylsaccmic acid is produced by
the action of phosphorus and bronine on trimethylsuccinic acid, the formation of
this bromoanhydride can only be accounted for on the supposition that when the
bromo-dibromide is dropped into alcohol, only part of it is decomposed, yielding
ethyl bromotrimethylsuccinate, and that the other part reacts with the alcohol
somewhat as follows :
(CHJ,C*COBr „ „ (CH8),C — CO^
<^> CH>r.COBr ^ ^^«=''°° = ci^iBr-CO^ "^ ^^'«'«' ^ ««°'
(CH,).C.COBr ^ ^ ^^^ ^ nCH.).C.CO.CH.-] _^
^' CHj-CBr'COBr L CH^-CBr •COBrJ
(CH3)jC — C0>^
CHa'CBr-CO-^
Digitized by
Google
54 SOKE AND SPKANKLlNa :
required for ethyl bromotrimethylsuccinate, an indication that there
remained a fair quantity of the bromo-anhydride in solution (a mix-
ture of 81 '3 parts of bromo-ester and 18'7 of bromo-anhydride would
contain 28*8 per cent, of bromine). With a view to the complete esteri-
fication of this bromo-anhydride, the oil was repeatedly heated with an
excess of ethyl alcohol containing 5 per cent, of hydrogen chloride ;
by this means, the bromine was reduced to 28*0 per cent. As we
subsequently found, however, that the bromo-anhydride, when treated
with alcohol and hydrogen bromide, forms only the monoethyl salt,
it was evident that the oil now consisted of a mixture of neutral and
acid esters. We therefore dissolved it in pure ether and extracted
the solution with a 5 per cent, sodium carbonate solution. On distil-
ling ofif the ether, we found that the residual oil could now be fraction-
ated under reduced pressure without undergoing any appreciable
decomposition. Under 20 mm., the greater portion of it distilled over
between 160° and 170° as a faint yeUow oil having a very pungent
odour. The following analysis indicated that it was practically pure
ethyl bromotrimethylsuccinate :*
0*2934 gave 01846 AgBr. Br « 26*76.
CjiHij^O^Br requires Br = 27*12 per cent.
(2) Action of Bromine on TrirMthyhuccinio Add <U 120 — 130° —
Bromotrimethylsuccinic anhydride may be most conveniently prepared
in quantity by the following method.
Five grams of trimethylsuccinic acid are heated with an equal
weight of dry bromine in a sealed tube at 120 — 130° for 6 to 8 hours.
The careful regulation of the temperature is important, since below
120° the bromination is not complete, and above 140° the contents of
the tube are liable to char. Great care should be taken in opening
such tubes after the heating, for the pressure in them is very great, and
since dense clouds of hydrogen bromide are evolved it is advisable to
carry out the operation in the open air. A solid with a slight orange
colour remains after the pressure has been relieved; sometimes it
swells up considerably during the escape of gas, and may occasionally
froth over out of the tube, and it is therefore advisable to have a
large beaker at hand in which to receive any that may be so forced
out. The solid should be washed with a cold dilute solution of sodium
carbonate, dried on a porous plate, and recrystallised from hot benzene.
The yield is quantitative.
* The preparation of this ester is best carried out by dropping the brominated
trimethylsuccinic acid into excess of ice-cold ethyl alcohol containing 5 per cent of
hydrogen bromide, heating the solution for about three hours on the water-bath to
convert the bromo-anhydride into the monoethyl salt, and subsequently removing
the latter by means of a cold 5 per cent, solution of sodium carbonate.
Digitized by
Google
THE BROMINATION OF TRIMBTHYLSUCCINIC ACID. 55
When pare, the bromo-anhydride melts at 197—198° ; it is qaite in-
soluble in cold water or a cold dilate solution of sodium carbonate.
We endeavoured to prepare bromotrimethylsuccinic acid by dissolving
the anhydride in warm water and evaporating the solution until, on
cooling, crystals appeared. In this way, colourless needles were ob-
tained which, however, melted indefinitely between 120° and 130°, and
contained only 31*1 per cent, of bromine; since the bromo-acid,
GYHjjO^Br, requires 33*5 per cent, of bromine, it was evident that
some decomposition had occurred during the solution of the bromo-
anhydride, and a subsequent careful examination showed that hydrogen
bromide is slowly liberated during the process.
Action of Alcohol and Sodium Ethoafide on ths Bromo-anhydride. —
On heating the bromo-an hydride with a molecular proportion of sodium
ethozide in ethyl alcohol, the liquid became neutral in about half an
hour without, however, any separation of sodium bromide. On passing
dry hydrogen chloride into the resulting liquid, sodium chloride separ-
ated, and as soon as the whole of the sodium had been thus eliminated
the liquid was filtered and the clear filtrate evaporated in a vacuum
over sulphuric acid. There finally remained a colourless, semi-solid
mass containing 30*3 per cent, of brominOj which exhibited all the pro-
perties of an acid ester (ethyl hydrogen bromotrimethylsuccinate,
CgHijO^Br, requires Br = 30*0 per cent.).
The same substance was obtained by heating the bromo-anhydride
with an excess of ethyl alcohol in sealed tubes at 160° and afterwards
distilling off the excess of alcohol on the water-bath. In neither of these
experiments were we able to detect the formation of any neutral ester,
and in each case the product instantly and completely dissolved in a
cold solution of sodium carbonate with evolution of carbon dioxide.
Our attempts to purify the substance by distillation under reduced
pressure were unsuccessful, since decomposition began at temperatures
below the boiling point.
We also made several unsuccessful attempts to prepare the silver
salt of this acid ester, but as soon as silver nitrate was added to its
aqueous solution neutralised with dilute ammonia, a copious yellow
precipitate of silver bromide appeared, and we were not more success-
ful in experiments in which freshly prepared silver carbonate was added
to the aqueous solution.
Action of Diethi/lcmiline on the Bromo-anhydride. Methylenedimethyl-
succinic Acid, C^HiqO^.
As bromotrimethylsuccinic anhydride showed a tendency to lose
hydrogen bromide on being boiled with water, we decided to study the
action of diethylaniline on it with the view of preparing the correspond-
ing unsaturated acid.
Digitized by
Google
56 BONE AND SPRANKLINQ:
Accordingly, a solution of 10 grams of the bromo-anhydride in
15 grams of diethylaniline was heated in a reflux apparatus on a sand-
bath for 10 hours, after which it was poured into a hot concentrated
solution of potassium hydroxide. After the diethylaniline had been
extracted with ether, the alkaline liquor was acidified, saturated with
ammonium sulphate, and again extracted with ether. In this way a
solid acid was obtained which was purified by dissolving it in excess of
sodium carbonate solution, extracting resinous matter with chloroform,
then boiling the solution with animal charcoal, finally acidifying and
extracting it with pure ether. The pure acid was thus obtained as
perfectly white crystals which melted at 140 — 141° On analysis :
0-2063 gave 0-4010 COa and 01290 HjO. 0 = 529 ; H = 6-51.
C7H10O4 requires 0 = 5316 ; H = 6-33 per cent.
MethylenediTnethylauccinic acid, ^ ^ * .X n/^^TT> ™®^^^ *^ ^^^ — 141°, is
OJij.O'UUaii
fairly soluble in cold water, and, like other succinic acids, gives an in-
soluble calcium salt when a solution of its neutral ammonium salt is
boiled with calcium chloride solution. Its aqueous solution instantly
decolorises alkaline permanganate and rapidly absorbs bromine in the
cold. The acid is readily estorified, and its liquid diethyl ester boils at
173—176° under 755—760 mm.
Action of Bromine on the Diethyl Ester. — ^Oq adding a solution of
bromine in chloroform to the diethyl ester, the halogen at once dis-
appeared ; as soon as no more of it was absorbed, the chloroform was
distilled ofE and the residual oil at once hydrolysed with hydrochloric
acid. On cooling, a white crystalline dibromo-acid separated, which
after recrystallisation from hydrochloric acid melted at 178—179°.
On analysis : -
0 3027 gave 0-3529 AgBr. Br = 496.
C^HjjjO^Brj requires Br = 50*0 per cent.
Action of Hydrogen Bromide on the Dietliyl Meter. — Ten grams of the
diethyl ester were mixed with an aqueous solution of hydrogen bromide
(saturated at 0°). Much heat was evolved, the bromide being very
quickly absorbed. The product was extracted with ether, and the
ethereal solution washed with dilute sodium carbonate solution and after-
wards dried over anhydrous sodium sulphate. On distilling off the ether
there remained a liquid diethyl ester of a bromo-acid which contained
an amount of bromine corresponding to that required for the empirical
formula C^iHigO^Br. Thus :
0-4165 gave 02650 AgBr. Br = 27-07.
OijHjgO^Br requires Br = 27*12 per cent.
Digitized by
Google
THE BROMINATION OF TRIMETHYLSUCCINIC ACID. 57
Most probably, therefore, this oil was ethyl bromotrimethylsuccin-
ate, (CH8)2C(COjC2H5)-0(CH3)Br-0O2C2H5, although it is just possible
that it was the isomeric ethyl a-methyl-S^bromobutane-/3y-dicarbozyl-
ate, (CH3)2C(CO2C2H5)-CH(0O,C2H5)-CH:2Br. On comparing the
action of the oil with that of the ethyl bromotrimethylsuccinate
obtained by the direct bromination of trimethylsuccinic acid (see p.* 5 4)
on ethyl sodiocyanoacetate (see next section), identical products were
obtained in the two experiments. We afterwards found that the
identity of these products does not necessarily imply the identity of
the two bromo-esters in question, so that which of the two foregoing
formulsB represents the constitution of the oil obtained by the action
of hydrogen bromide on diethyl methylenedimethylsuccinate is a point
we have not yet definitely established.
Interadion of Ethyl BromotrimethylsuccinaU and Ethyl Sodtocyano-
acetate,
(1) To a solution of 1*5 grams of sodium in 20 grams of alcohol
were added 7*5 grams of ethyl cyanoacetate and 19 grams of ethyl
bromotrimethylsuccinate ; much heat was developed, sodium bromide
separated, and the liquid became neutral after being heated for 3
hours on the water-bath. The product was extracted with ether and
fractionated under 20 mm. pressure. A fair proportion of it passed
over between 130° and 150*^, the temperature then rose rapidly to above
200°, and about half of the oil distnied between 210° and 215°. This
higher fraction was hydrolysed by boiling it with strong hydrochloric
acid for 24 hours. On cooling the liquid, no crystals separated, so it was
saturated with ammonium sulphate and thoroughly extracted with
ether.
In this way, a white crystalline acid was isolated which, after
recrystallisation from strong hydrochloric acid, melted sharply at
137 — 138°. This, it will be observed, is the same melting point as that
of the acid obtained by one of us and Professor Perkin some years
ago by the same series of reactions. The acid was therefore not
i-camphoronic acid (m. p. 169 — 172°), and, further, all attempts to iso-
late any camphoronic acid from the hydrolytic products by means of
its characteristic barium salt entirely failed, so we can only conclude
that none had been formed. Analysis of our acid, however, indicated
that it was tribasic and isomeric with camphoronic acid, Q^^fi^,
thus:
0*1364 gave 0*2500 CO2 and 00813 HjO. C = 49*87 ; H = 6*62.
0*1095 silver salt gave 00657 Ag. Ag« 600.
CgHi^Og requires 0 = 49*53 ; H = 6'42 per cent.
CoH^OeAga „ Ag= 60*10 per cent.
Digitized by
Google
58 THE BROMINATION OF TRIMKTHTLSUCCINIC ACID.
(2) In another experiment, 1 1 grams of the bromo-ester obtained by
the action of hydrogen bromide on ethyl methylenedimethylsuccinate
were added to the calculated quantity of ethyl sodiocyanoacetate sus-
pended in alcohol. Sodium bromide at once separated and on continu-
ing the experiment as described in the preceding paragraph, we finally
obtained a cyano-ester boiling at 230 — 240° under 40 mm. pressure.
This on hydrolysis with hydrochloric acid yielded the same acid,
CgHi^Oo, melting at 137°
Fusion of the Acid, CgH^^Og, with Potassium Hydroxide.
In order to obtain evidence as to the constitution of the acid, 5
grams of the substance were fused with a paste of 30 grams of
potassium hydroxide at 180 — 200°. A. vigorous decomposition ensued.
After being cooled, the fused product was dissolved in water, acidified
with dilute sulphuric acid, and the liquid then distilled with steam.
The distillate contained a fatty acid, the analysis of whose silver salt
showed it was acetic acid.
0-2021 silver salt gave 01303 Ag. Ag= 64-47.
CgHjOjAg requires Ag = 64*67 per cent.
On extracting the residual liquor with ether, a solid acid melting at
147° and in other respects identical with trimethylsuccinic acid was
obtained. (An analysis of the silver salt of this acid was made, but
the results have, unfortunately, been mislaid ; they agreed well with
the calculated numbers for silver trimethylsuccinate.)
These results are quite considtent with the view that the acid
OgHi^Og is aa-dimethylbutane-afi^lricarboxylic cuiid, and indeed it is
difficult to see what other constitution could be assigned to it. The
further investigation of its properties has, for the time being, been
stopped on account of lack of material, but will be resumed in the
near future.
We desire to state that the greater part of the materials required
for this research was purchased out of a grant from the Research Fund
of the Society.
The Oweks Colleoe,
Manchester.
Digitized by
Google
CONSnTITENTS OF THE ESSENTIAL OIL OF ASARUH CANADENSE. 59
VII. — The Constituents of the Essential Oil of Asarum
Canadense.
By Frederick B. Power and Frederic H. Lbeb.
The aromatic essential oil distilled from the underground portion,
rhizome, and rootlets of Aacirum canculense, commonly known as
Canada Snake-root, was first investigated by one of us a number of
years ago (Power, Inaug, Diss,, Strassburg, 1880 ; Froe, Amer. Fharm.
Assoc., 1880,28, 464). In that investigation, the following substances
were isolated. (1) A terpene, OjqK^^ b. p. 163—166°; (2) two
fragrant alcohols, differing considerably in their boiling points and
also in their odour, but both possessing the same empirical formula,
^10^18^- The alcohol of lower boiling point, 196—199°, was termed
asarol, and had an odour which recalled that of coriander, but was also
somewhat camphoraceous, whilst the alcohol of higher boiling point,
222—226°, had a rose-like fragrance ; (3) a fraction, possessing but
little odour, b. p. 254 — 257°, representing the largest single con-
stituent of the oil, which, upon ozidation with chromic acid, afforded
an acid of the composition OgH^oO^. This acid was subsequently
shown by Petersen {Bei\, 1888, 21, 1062) to be veratric acid and was
obtained by him by the ozidation of an analogous substance contained
in the oil of Asarwn eurapcteum, boiling at about 250°, which he
proved to be eugenol methyl ether ; (4) a fraction collected at 275 — 350°,
which contains a deep blue oil of undetermined composition ; (5) a
large amount of acetic acid, combined with the above-mentioned
alcohols in the form of acetic esters, together with a very small
amount of a less soluble, oily acid, which appeared to consist of, or at
least to contain, valeric acid.
In consideration of the advance in knowledge of the constituents of
essential oils since the period of the first investigation, and the
means which are now available for the more positive identification and
classification of these constituents by the preparation of well-defined
and mostly crystallisable derivatives, it has seemed desirable again to
subject the oil in question to a careful chemical examination.
Experimental.
The oil employed for this research, about 2 kilos, in amount,
was distilled by Messrs. Schimmel & Co. of Leipzig. Its density
at 15°/15° was 0-952, and its rotation aD= - 3°24' in a 100 mm. tube.
The oil was first shaken with a 5 per cent, solution of sodium
carbonate in order to remove the free acids, which were examined in
connection with the acids obtained by the subsequent hydrolysis of the
Digitized by
Google
60 POWER AND LEES: THE CONSTITUENTS OF THE
oil. It was then shaken three times successively with a 5 per cent,
solution of sodium hydroxide, and afterwards with water until the
washings were neutral. The combined alkaline liquids were shaken
twice with ether to remove any adhering oil, then acidified with
sulphuric acid, completely extracted with ether, and the ethereal
liquid dried with calcium chloride. After distilling off the ether, the
residual liquid was distilled in a vacuum. Under 10 mm. pressure, it
began to distil at 155°, rising rapidly to 250° and the last portion
was observed to solidify in^the condenser. When refractionated, there
were obtained :
I. A light coloured oil boiling at 172° under 35 mm. pressure.
II. A dark oil which boiled somewhat below 200° under 10 mm.
pressure and solidified on cooling.
The Phenol, CgHi202.
The first of the preceding fractions was distilled under the ordinary
(762 mm.) pressure and afforded :
(1) A light greenish liquid, becoming brown on standing, and
boiling below 245°.
(2) A light yellow liquid, boiling at 245—260°, which did not
solidify at -16°.
(3) A small residue, from which a little crystalline substance
separated on cooling.
Of these fractions, (1) and (2) were analysed.
(1) 0-1402 gave 0-3606 COg and 01008 HgO. C = 70*1 ; H = 80.
(2)0 1493 „ 0-3922 CO2 „ 0-1066 HjO. 0 = 716 ; H = 7-9.
These two fractions were then subjected to a final fractionation
under the ordinary pressure, with the following result :
(a) A few drops only distilled below 220°.
{b) From 220° the mercury rose rapidly to 245°
(0) The chief portion distilled between 245° and 255°, and was
fairly constant at 248—252°.
{d) Only a few drops distilled above 255°
Fraction (c) was then analysed.
01523 gave 0-3992 CO2 and 01073 HjO.' 0 = 71-5; H = 7-8.
OgHjjOj requires 0 = 71-1 ; H = 7"9 per cent.
These results indicate that the phenol contained in asarum oil has
the empirical formula CgH^fi^' ^^ ^^ ^ nearly colourless, oily liquid,
having an odour recalling, but more agreeable than, that of creosote.
In the process of liberating the phenol from its alkaline solution a
somewhat clove-like odour was developed, and this at first led us to
suspect that the phenol contained some eugenol. This, however, is
Digitized by
Google
ESSENTIAL OIL OF iJSARUM CANADENSE. 61
rendered highly improbable, both on account of the constancy of the
analytical results and the characteristic colour reaction which was
alEorded by all the fractions, but most strikingly by the principal
fraction (c). Thus a very small quantity of the phenol, when dissolved
in 90 per cent, alcohol, gives, with a trace of very dilute ferric chloride,
a beautiful violet colour which only gradually fades, whilst eugenol,
under the same conditions, gives a green. If, however, the phenol
from asarum oil be dissolved in absolute alcohol, and a trace of a
stronger solution of ferric chloride added, it affords a green colour,
whilst eugenol, under these conditions, gives a deep blue. The amount
of the phenol available did not permit of the formation of any deriva-
tives, but as it is not identical with any of the known phenols of the
formula indicated, it is evidently a new substance, and we shall endea-
vour to determine its constitution by some synthetical experiments.
IdetUificaiion qf PaZmitic Acid.
It was noted that in the first distillation of the phenol, a solid sub-
stance separated in the condenser, and that in a subsequent fractiona-
tion the higher fraction, designated as II (p. 60), solidified on cooling.
This was, therefore, brought upon a porous tile, and the substance
subsequently dissolved in hot light petroleum, from which, on cooling
it crystallised in colourless, pearly plates. lbs melting point was
60 — 61% and this remained unchanged on recrystallisation. On
analysis, it was identified as palmitic acid :
00874 gave 0-2382 OOg and 0-0982 H^O. C = 74-3 ; H = 125.
Ci^HggOg requires C = 75*0 ; H = 12-5 per cent.
Identification qf Pinene,
SeparcOion qf the Terpene, — ^The oil, which had been shaken with a
dilute solution of sodium hydroxide as previously described, was
washed with small, successive portions of water until the washings
were neutral, and dried with anhydrous sodium sulphate. It was
then distilled under diminished pressure and the portion collected
which boiled below 100° under 10 mm. pressure. After several frac-
tionations of this portion and drying with potassium carbonate, the
lowest fraction, which distilled below 85° under 10 — 16 mm. pressure,
was collected. Its density was 0*8566 at 18716°, which proved the
absence of any olefinic terpene. These liquids were then further
fractionated under the ordinary pressure, when the greater portion
finally distilled below 165°, (a), chiefly at 159 — 161°, and only exceedingly
small fractions were collected between 165° and 170° (fi) and from
170 — 180° (y). These were analysed, with the following results :
Digitized by
Google
62 POWER AND LEES : THE CONSTITUENTS OF THE
(a) 01252 gave 04003 COg and 0-1320 H^O. 0 = 87*2 ; H = 11-7.
(i8) 01526 „ 0-4873 COj „ 0-1606 HgO. 0 =- 871 ; H= 11-7.
(y) 0-1558 „ 0-4920 COjj „ 0*1628 HjjO. 0 = 86-1 ; H = 11-6.
CioHje requires 0 = 88-2 ; H = 1 1 -8 per cent.
FracUan helaw 165° — This fraction, boiling chiefly between 169°
and 161°, which is seen to consist of a nearly pure hydrocarbon,
amounted to about 2 per cent, of the original oil. Its physical con-
stants were as follows : d 16716° =0-8563. a^=^ +1°36'. It readily
formed a crystallisable nitrosochloride melting sharply at 103 — 104°.
From the latter, the nitrolpiperidide was prepared, which, after re-
crystallisation from methyl alcohol, melted sharply at 118 — 119°.
This fraction thus consisted of pineney and its low rotation indicates it
to be a mixture of the d- and /-forms. Petersen {B&r,^ 1888, 21,
1059) has previously recognised the terpene existing in the oil of both
the European and American species of Amrum as pinene, in the former
as the ^variety, but identified it only by the formation of an oily
monobromide and by its conversion into dipentene.
As it was possible that the very small fraction of our oil collected
between 170° and 180° might contain dipentene or limonene, it was
treated with bromine, but only an uncrystallisable, oily product was
obtained. After being carefully dried, a bromine determination was
made of this, with the following result :
0-2915 gave 0-3762 AgBr. Br = 54-9.
C^QH^gBr^ requires Br = 54*1 per cent.
This result serves to prove the absence of either dipentene or
limonene, both of which form crystallisable tetrabromides, C^oH^^Br^.
By a careful examination of all the fractions, no terpene other than
pinene could be detected in the oil.
Hydrolysis qf the OH
For further examination, all the oil boiling above the terpene frac-
tion was now hydrolysed by boiling with alcoholic potassium hydr-
oxide for about 2 hours in a flask provided with a reflux condenser.
After distilling off the greater portion of the alcohol from a water-
bath, the liquid was brought into a separating funnel and sufficient
water added to effect the separation of the oil. The latter was then
drawn off, the aqueous alkaline liquid shaken with successive portions
of ether, and the ether extracts mixed with the separated oil. The
latter was then washed several times with water, and these washings
added to the aqueous alkaline liquid. The ethereal solution of the oil
was quickly dried with calcium chloride, filtered, the ether distilled
off, and the residue finally subjected to fractional distillation, first
Digitized by
Google
ESSENTIAL OIL OF ASARUM CANADBNSE. 09
under diminished pressure, and then in part under the ordinary pres-
sure. The following fractions were eventually obtained : 195 — 203°
20a— 208^ 208—216% 2 1 6—222°, 222—236°, 236— 246°,and 246 —260°.
Identification o/LincUool.
Fraction 195 — 203°. — This was a large fraction which, when re-
distilled under the ordinary pressure, passed over mainly at 199° and
almost entirely at 198 — 202° under 768 mm. pressure. It is a colour-
less, fragrant liquid. It was analysed and its physical constants
were determined with the following results :
0a372 gave 0-3910 CO, and 0-1436 H^O. 0 = 77-7 ; H = ll-6.
CjQHjgO requires 0 « 77-9 ; H = 11-7 per cent,
d 15*5°/15° = 0-871 1. aD= + 10°48'ina 100 mm. tube; [a]D= -I- 12-4°.
When oxidised with chromic acid, it afforded citral, which was ob-
tained as a pale yellow liquid of strong, lemon-like odour, distilling at
110 — 115° under a pressure of 10 — 12 mm. The latter, by condensa-
tion with pyruvic acid and j3-naphthylamine, was converted into the
crystalline a-citryl-/3-naphthacinchoninic acid, melting at 196 — 198°.
The identity of this fraction with d-Hnalool is therefore definitely
established. It corresponds to the substance O^oH^gO (b. p. 196 — 199°),
which in the first investigation of the oil was designated aaarol.
Fraction 203 — 208°. — ^This fraction was too small for further ex-
amination, and evidently consisted simply of a mixture of the preced-
ing and the following fractions.
^Fraction 208 — 216°.— This was a small fraction, which distilled
mostly between 208° and 212°. It was analysed, and its physical
constants were determined, with the following results :
01368 gave 03866 00, and 0-1424 HgO. 0 = 771 ; H= 11-6.
0-1472 „ 0-4164 CO, „ 01624 HjjO. 0 = 77-0; H=ll-5.
Oi^HigO requires 0 = 77*9 ; H - 11 -7 per cent.
d 16-6°/15°= 0-911 ; a^^ -0°24' in a 100 mm. tube.
Identification of Bomeol.
The liquid had a camphoraceous and also somewhat rose-like odour.
When subjected to a temperature of - 10° for an hour, no crystalline
substance separated. As this fraction of the oil was relatively small,
and as its constituents were evidently contained to some extent in the
next higher fraction, the two fractions were mixed. A portion, how-
ever, of the higher fraction was reserved for special examination.
^This mixture of the two fractions was now gently oxidised with
Fittig's oxidation mixture {£er., 1886, 18, 3207) in the following
Digitized by
Google
64 POWER AND LEES: THE CONSTITUENTS OF THE
proportions: 10 parts of oil, 80 parts of potassium dichromate,
and 120 parts of sulphuric acid, the latter diluted with three times
its volume of water. The oxidising mixture was added in small
amounts at a time to the oil, which was kept cool by immersion of
the containing flask in water. After all the chromic acid solution had
been added, the mixture was heated on a water-bath for about 20 — 30
minutes. It was then distilled from a water-bath under diminished
pressure and the camphor which separated in the condenser and dis-
tillate was collected by filtration, and dned on a porous tile. A little
of the sublimed product was found to melt sharply at 1 75°. A deter-
mination of its specific rotation in 90 per cent, alcohol gave the follow-
result :
aD= -1°45'; /=0-5dcm. ; c = 8-684; [a]i>= - 40-3^
For further identification of the camphor, the oxime was prepared,
and found to melt at 115 — 116°. As camphor could not be detected
in the fraction of the original oil, its formation by the above method
of oxidation is conclusive proof of the presence of l-homeol in the oil.
The chromic acid liquor remaining from the distillation of the cam-
phor was subsequently shaken out several times with ether, the ethereal
solution washed with a little water, dried with calcium chloride, and
the ether removed by distillation. The residual light yellow oil, which
had a strong odour of acetic acid, was found to be not entirely soluble
in cold sodium carbonate solution. It was consequently redissolved in
ether and the ethereal solution shaken out several times with a dilute
solution of sodium carbonate in order to remove the acids. The ethereal
solution was then washed with a little water, dried with calcium
chloride, and the ether removed by distillation. The residue was a
light yellow oil possessing a coumarin-like odour, and on standing a
short time became a crystalline paste. This was drained on a porous
tile, when the substance was obtained quite white. After recrystallisa-
tion from dry ether it melted at 62°, and was insoluble in sodium
carbonate solution. On analysis :
0-1148 gave 02731 00^ and 00902 H2O. 0 = 649 ; H = 8-7.
^10^16^3 requires 0 = 65-2 ; H = 8*7 per cent..
This substance is undoubtedly identical with the ketoJacionBy
CiQHjgOg (m. p. 62 — 63°), which was isolated as a product of the
oxidation of terpineol by chromic acid by Wallach, and has been
further studied by him, as also by Tiemann and others {Annoden^
1893, 275, 153; 277, 118; Ber., 1895, 28, 1773, 1781).
The sodium carbonate solution from which the ethereal solution of
the above ketolactone had been separated was acidified with
hydrochloric acid, and shaken out several times with ether. The
Digitized by
Google
SSSENtlAL OIL Of" ASARUM CAKADENSE. 65
ethereal solution was washed once with water, dried with calcium
chloride, and the ether removed by distillation. The residue was a
light yellow syrup, which, on standing, deposited a crystalline acid.
The syrup was consequently diluted with ether, in which the crys-
tals appeared to be sparingly soluble, and from which they were
easily separated by filtration. After washing with dry ether, the
substance was finally recrystallised from bdiling ether. It melted
at 173 — 174° and dissolved with effervescence in a cold solution of
sodium carbonate.
0-1168 gave 0-2266 COj and 0*0695 Kfl. 0 = 52-9 ; H = 6-6.
CyHjoO^ requires 0 = 53-2 ; H = 6-3 per cent.
This acid is evidently identical with terebic acid, O^H^oG^ (m. p.
175°), which has been found as a direct ozidUion product of terpineol,
as also of the ketolactone, O^qHi^Oj (Tiemann and Mahla, Bm',y 1896,
29, 2621). The syrup from which the terebic acid crystallised was
not examined but probably contained terpanylic acid, which always
accompanies terebic acid when terpineol or the ketolactone is oxidised
with chromic acid mixture. It is thus shown that the fractions of the
oil which served for the identification of bomeol also contained a con-
siderable amount of terpineol.
In the first investigation of asarum oil, a considerable fraction was
collected at 222 — 226°, and as a portion of this was still available it
was thought of interest to examine it again. It was therefore oxidised
with a chromic acid mixture in the manner just described, and among
the products of oxidation there were isolated and identified : camphor
(m. p. 175°); the ketolactone, C^qR^qO^ (m. p. 62°); and terebic acid
(m. p. 173 — 174°). It therefore contained bomeol and terpineol, and,
apparently, a small amount of geraniol, as it had the characteristic
rose-like odour.
IderUiJicatton of Terpineol,
Fraction 216 — 222°. — This was a small fraction. It had a cam-
phoraceous and also a somewhat rose-like odour. It was analysed and
its physical constants were determined, with the following results :
01647 gave 0 4682 CO, and 01698 HjO. 0 = 775 ; H = ll-4.
0-1633 „ 0-4610 00, „ 01680 H,0. 0 = 77-0; H = ll-4.
Oj^jHjgO requires 0 = 77*9 ; H = 11 7 per cent.
d 15-6°/ 15° = 0-9267 ; a^- -8°26' in a 100 mm. tube.
The portion of this fraction which had not been used in connection
with the preceding one, as described under the latter, was employed
for the direct identification of terpineol. In view of the presence of
small amounts of other alcohols, the following method was employed.
VOL. Lie XXI. F
Digitized by VjOOQIC
66 POWER AND LEES : THE GOKSTITUENTS OF THE
The liquid was shaken with a concentrated solution of hydriodic
add (sp. gr. 2'0)» when a heavy, dark oil was formed. This was
separated from the aqaeoos layer, and shaken with a dilate eolation
of sodinm biBalphite to remove any free iodine. The oil was then
washed with water and allowed to stand, when after a short time
crystals began to form, and finally the whole became a crystalline
paste. This was spread on a poroos tile, when a small quantity of
nearly white needles was obtained, which, after recrystallisation from
light petrolenm (b. p. 30—40^), melted at 80^. This melting point
was identical with that of dipentene dihydriodide, OjoH^gl^, which, for
the purpose of comparison, we had also prepared from pure crystallised
terpineol, and when the two hydriodides were intimately mixed, the
melting point remained unchanged. The formation of this derivative,
and of the products of oxidation described in the preceding section,
proves conclusively the presence of terpinaol in the oil. The optical
rotation of the fraction from which it was obtained indicates it to be
the /'form.
Identifieatum qf Geraniol.
Fraction 222— 236°.— This fraction was collected within the above
limits, in view of the possible presence of both citronellol and geraniol.
It was relatively small in amount, and was analysed, and its physical
constants were determined, with the following results :
0-1544 gave 0*4370 CO^ and 0-1549 HjO. 0 = 77-2; H = 1M.
0-1420 „ 0-3999 CO2 „ 0-1433 HgO. C = 76-8; H=ll-2.
CjoHjgO requires C-77-9 ; H = 11-7 per cent.
d 15'5715° « 0-9340 ; aD= - 9°8' in a 100 mm. tube.
It possessed a camphoraceous and also a fragrant, rose-like odour.
Although its high density and rotation indicated that it contained a
considerable amount of terpineol, and its analysis also showed an
admixture with some of the next higher fraction, the small amount
of liquid precluded its further purification by simple distillation.
The odour of this fraction afforded such convincing evidence of the
presence of geraniol that Erdmann's method, which depends on the
formation of the crystalline geranioldiphenylurethane (m. p. 82°), was
resorted to for the identification of the substance («/. pr, Chem., 1897,
[ii], 66, 8). The oil was treated with diphenylcarbamio chloride in
presence of pyridine, as described by Erdmann ; the syrupy residue left
after distilling the product with steam was then purified by extraction
with ether, and the ethereal solution evaporated after extraction with
dilute hydrochloric add. The residual light brown oil was mixed
with a little alcohol, when it soon formed a crystalline paste, which
was drained on a porous tile. The substance was finally recrystallised
Digitized by
Google
ESSENTIAL 6lL OF ASARUM CANADENSE. 67
from a little alcohol, from which it separated in fine, glistening needles
melting sharply at 81 — 82°. On aoidysis :
0-1444 gave 0-4163 CO^ and 0-1015 HgO. 0 = 78-6 ; H = 7-8.
OjjHg^OjN requires 0 = 79-1 ; H = 7*7 per cent.
The fraction by gentle oxidation with chromic Acid afforded a little
citraly but although the amount of the latter was too small for con-
version into the naphthacinchoninic acid derivative, the evidence was
already sufficiently conclusive of the presence of g&rcmiol in this fraction
of the oil. There was, on the other hand, no indication of the presence
of citronellol.
It may be noted that in the first investigation of asarum oil by
one of OS, a fraction was obtained corresponding approximately in
boiling point (222 — 226°) to that just described, and that this, on
more energetic oxidation with chromic acid, afforded, besides acetic
acid, a small amount of a orystallisable acid. As a specimen of the
latter had been preserved, it has been re-examined and shown to be
a mixture of terebic and terpenylic acids.
Fraction 235 — 245°. — ^This was very small in amount, and was
evidently a mixture of the preceding and following fractions ; a little
of the crystallised geranioldiphenylurethane was obtained from it by
the method previously described.
. Identification qf Eugenol Methyl Eifyer.
Fraction 246 — 260°. — This constitutes the largest fraction of the
oil. On redistillation under the ordinary pressure it was easily
resolved into a large fraction, which was collected between 250° and
256°, bat distilled for the most part between 252° and 254°. It is a
colourless, nearly odourless liquid, and was analysed, and its physical
constants were determined, with the following results :
0 1648 gave 04522 COg and 01238 H^O, 0 = 74-8 ; H = 8-3.
^11^14^2 requires 0 = 742 j H « 7*9 per cent.
d 15°/16°= 1-0239 ; ttD = - 2°44' in a 100 mm. tube.
It has been shown by Petersen (Ber., 1888, 21, 1064) that the oil
obtained from the allied European species of Asofrum contains a sub-
stance of the same composition, boiling at about 250°, which on oxidation
affords veratric acid, and was fully identified as eugenol methyl ether.
In the first investigation of the oil of Asarum canadense by one of us, a
fraction was collected at 254 — 257°, which on oxidation with chromic acid
afforded a small amount of a crystalline acid, OgH^QO^, and this
Petersen has likewise found to be identical with veratric acid. The
same specimen of acid, after recrystallisation from water, we now
find to soften at 172°, and to melt completely at 177—178°.
F 2
Digitized by VjOOQIC
68 POWER AND lees: THE CONSTITUENTS OF THE
The confirmation of the identity of ihiu fraction with eugenol methyl
ether has now been effected by the preparation of the crystalline
bromoeugenol methyl ether dibromide, CgHjBr(OCH3)'C3H5Brj, which
melts at 78—79° (Wasserman, Conipt. rend., 1879, 88, 1206). This
was accomplished as follows : To the liquid dissolved in dry chloroform,
and cooled in a mixture of ice and salt, the requisite quantity of
bromine, also dissolved in chloroform, was added, drop by drop, and
any slight excess of bromine removed afterwards by shaking the
solution with a little sulphurous acid. The chloroform solution was
separated, dried, and filtered, and the chloroform removed by rapidly
drawing dry air through the solution. The residue was a thick syrup,
which, when dissolved in alcohol, deposited a quantity of glistening
crystals. These, on recrystallisation f ropi absolute alcohol, separated in
glistening, felt-like needles, which melted at 78 — 79°.
The optical activity of the fraction is due to admixture with a
small amount of a higher fraction, which it is difBcult to separate
completely by fractional distillation.
Seardi for isoEugenol Methyl EUier, — As it has been assumed by
Mittmann {Arcft. Pharm., 1889, 227, 543) that the substance con-
tained in asarum oil is not eugenol methyl ether but the isomeride, we
have thought it desirable to ascertain the correctness of this opinion.
For this purpose, a portion of the original oil which had been deprived
of terpene was fractionated under diminished pressure before being
subjected to hydrolysis. As eugenol methyl ether boils at 128 — 130°
(10 mm.) and uoeugenol methyl ether at 142° (10 mm.), fractions were
first collected at 130—140° and at 140-— 155° under a pressure of
about 10 mm. Further fractionation was conducted under 60 mm.
pressure, at which eugenol methyl ether was found to boil at 166°,
and Moeugenol methyl ether at 179°. A large fraction was thus
collected at 163—167° (60 mm.), and also a fraction at 175—185°
(60 mm.). For the differentiation of these two substances recourse
was had to bromination, as eugenol methyl ether in the cold yields the
bromo bromide, whereas tsoeugenol methyl ether under the same
conditions yields only a dibromide melting at 99 — lOl°(^0r., 1890,23,
1 1 67). On applying this test to the two fractions, only the crystUIine
derivative melting at 78 — 79° was obtained, which proves that the
original oil does not contain Moeugenol methyl ether.
Fraction boiling above 260°.
This fraction was distilled under reduced pressure, and after a large
number of distillations under 60 mm. pressure the following fractions
were obtained: Below 175°, 175—195°, 195—210°, 210—220°, and
220—230°.
Digitized by
Google
ESSENIIAL OIL OF ASARUM CANADENSE. 69
The characters of these fractions are shown in the following table :
Boiling
point
(60 mm.).
Below ITS'*
176-195
195—210
210—220
220—280 I
Analysis.
C=760; H= 9-3
0 = 78-4; H = 10-8
0=81-1; H = 10-6
Rotation in
100 mm.
tube.
ftD
=
-10M2'
ao
=
-41 40
Od
=
-100
Solubility in
70 per cent
alcohol.
Very freely soluble
Vtry freely soluble
Very freely soluble
Less freely soluble
Turbid
Oolour.
Slight
Light yellow
Bluish
Bluish
Qreenish
The fraction collected below 175° consisted chiefly of eugenol methyl
ether. The three subsequent fractions had an odour resembling that
of cedar wood, and when a few drops were dissolved in glacial acetic
acid and a drop of concentrated hydrochloric or sulphuric acid added,
an intense reddish- violet colour was prodticed. The fraction 220 — 230°
was very small in amount. The analysis of the principal fractions,
and particularly their ready solubility in dilute alcohol, proved that
they consisted of oxygenated compounds, and did not contain a sesqui-
terpene.* As the fraction 210 — 220° was the largest, this was again
carefully distilled, and the following fairly constant fraction obtained,
which was more fully examined.
Fraeiion 212 — 217° (60 mm.). — This is a thick, viscid liquid, having
a fine blue colour and an odour recalling that of cedar wood. It does
not solidify when exposed for some time to a temperature of - 18°. It
is very freely soluble in 70 per cent, alcohol and affords the same
colour reaction as the fraction from which it was obtained. It was
analysed^ and its physical constants were determined, with the following
results:
0-1069 gave 0-3133 CO2 and 0-1009 HgO. C = 79-9 ; H = 10 5.
d 15°/16°= 1-0063 ; od= -3°; Z=100 mm; c = 3-678; [o]i>= -81-5°
A molecular weight determination gave the following result : 0 4184
gram depressed the freezing point of 30-17 grams Qf phenol by 0*48°,
whence mol. wt. » 214.
This rei-ult would agree very well with a sesquiterpene alcohol of
the formula Oij^H^^O (mol. wt. =222), but the analytical figures do
not accord with those required for this substance (0 = 81*1 ; H = ll-7
per cent.). It is probable, therefore, that the fraction analysed still
consisted of more than one substance.
* The stotcment in '*The Volatile Oils," by Gildemeister and Hofi'niann (p. 123)
bat "Seinmler, in 1889, obtained from asarum oil a hydrocaibon, CjaH^ boiling
at about 265° " is an error of translation. It properly refers to the oil of Carliria
acaidia or Carline thistle (Qerman, Ehenimrz\ which is described on p. 690 of the
same work.
Digitized by
Google
70 POWER AND LEES: THE CONSTITUENTS OF THE
Treatment with Phoapharic^Oxide. — In order to obtain farther evidence
respecting the character of these bluish fractions, an attempt was nuide
to dehydrate them. The remainder of the fractions 195 — 210^ and
210 — 220^ (60 mm.)y about 5 grams of each, was separately dissolved
in dry benzene, phosphoric oxide added, and the liquids boiled for
about an hour, when they acquired a deep purple-red colour. After
distilling off the benzene, the residues were distilled under diminished
pressure.
I^aeUon 195 — 210° (60 mm.) afforded a liquid which distilled between
175° and 210° under 60 mm. pressure. It had a bright green colour,
a cedar-like odour, and was insoluble in 70 per cent, alcohol.
0-1068 gave 0-3198 COjj and 0-1000 H,0. C-81-7; H = 10-4 per cent
d 15715° = 0-975; [alo- -37°.
Fraetian 210—220° (60 mm.) afforded a liquid which distilled chiefly
between 200° and 220° under 60 mm. pressure. It had an olive-green
colour, a cedar-like odour, and was insoluble in 70 per cent, alcohol.
01 073 gave 0-3300 COg and 0*1006 H3O. C = 839 ; H = 10-4 per cent-
d 15°/15° = 0-985; [aj^^ -35-5°.
Both these liquids, when dissolved in glacial acetic acid and treated with
a drop of hydrochloric acid, afforded a purplish or red colour. Although
the insolubility of these products in alcohol and the increase in the per-
centage of carbon by the above treatment was evidence of the forma-
tion of a hydrocarbon, the substances themselves were not sufficiently
pure to admit of further identification. They were finally dissolved in
dry ether, the solutions saturated with dry hydrogen chloride, and
allowed to stand for several days, but from the very dark, oily residues
no crystallisable hydrochloride could be obtained.
Although several essential oils are known to afford high boiling
fractions of a deep blue colour, which have been designated as
ccerulein by Gladstone, and as azulene by Piesse, no properly charac-
terised compound has as yet been isolated from any one of them.
Adda obtained by tlie HydrcHyais of tlie Oil.
The strongly alkaline, aqueous liquid, separated from the hydrolyaed
oil and completely extracted with ether, as previously described, was
evaporated to a small bulk, then acidified with sulphuric acid and
distilled with steam. The first portion of the distillate was slightly
turbid, but it soon became clear. The entire acid liquid was then
made alkaline with sodium carbonate, and extracted several times
with ether. After distilling off the ether there remained a small
amount of a dark coloured, highly aromatic oil. This was insoluble
Digitized by
Google
KSSENnA.L OIL OF ASARUM OANADENSB. 71
in a cold solution of sodium hydroxide, but dissolved readily on
warming. The alkaline solution of the substance was shaken with
ether to remove any impurity, and then acidified with sulphuric acid,
which liberated the original oil. This was again taken up with ether,
the ethereal solution washed with a little water, dried, evaporated,
and the slightly coloured, oily residue finally placed in a vacuous
desiccator over paraffin to remove the last traces of ether, and then
analysed.
00463 gave 01292 CO, and 0 0403 H,0. C « 761 ; H « 97.
Ci^lS^O^ requires 0 » 76*4 ; H » 9*1 per cent.
This substance, to which we assign the provisional formula C14H20O2,
is evidently a lactone. Although existing in extremely small amount,
80 that we have not been able to examine it further, its powerful odour
indicates that it must have considerable inflaenoe on the odour of
the original oil. To it is also possibly due the somewhat dove-like
odour which was observed in the isolation of the phenol (p. 60).
After the lactone had been separated from the liquid which
had been made alkaline with sodium carbonate, this liquid was
concentrated, strongly acidified with sulphuric acid, and shaken four
times successively with ether. The ethereal solution was washed
twice with water, dried, and distilled. On fractionating the residue,
nearly all distilled between 110^ and 120^. A portion of this was
converted into the barium salt, and from the latter the silver salt was
prepared which gave the following figures on analysis :
0*085 gave on ignition 0*0550 Ag. Ag « 64*7.
CgHjOjA-g requires Ag=64*7 per cent.
This served to confirm the presence of 000^ add^ the previous invest-
igation having shown that esters of this acid were present in consider-
able amount in the oil.
Acids of Higher Bailing Point, — ^The residue from the distillation of
the acetic acid was very small in amount, and was therefore mixed
with a larger portion of acids obtained by shaking the original oil with
a solution of sodium carbonate. The whole of the acids of higher boiling
point, after standing over potash in a vacuous desiccator, was first
fractionated under 10 mm., and then under the ordinary pressure,
when the following three fractions were obtained.
(1) Below 240^; (2) 240—280^; (3) 280—300°.
The last fraction became solid on standing, and from the residue in
the fiask crystals were separated which, after recrystallisation from hot
light petroleum, melted at 57 — 58° ; these consisted apparently of pal-
mitic acid, which had been extracted by means of caustic alkali from
the original oil. The first two fractions were redistilled and the follow-
Digitized by
Google
72 ESSENTIAL OIL OF ASARUM CANADENSE.
ing fractions collected : A, 140—200; B, 200—230°; 0, 230—270°.
D was fraction (3) from the first distillation (b. p. 280—300°).
They were yellowish, oily liquids, nearly equal in amount, and were
present altogether to the extent of about 2 grains in a kilo, of the
original oiL They were first converted into ammonium salts, and then
fractionally precipitated by silver nitrate. The various silver salts
were spread on porous plates, and then dried at 80° for half-an-hour.
On analysis, they gave the following results :
A^. Ag « 46'1 per cent. ; A^. Ag » 48*4 per cent.
B,. Ag-451 „ B,. Ag=.471 „
Cy Ag«36-7 , Cy Ag = 39-5
D^. Ag=34-5 „ Dg. Ag = 39-2
C0H^jO2AgrequiresAg = 48*4. C^sH^jOgAg requiresAg» 35*1 percent.
It is thus seen that these acids of high boiling point constitute an
exceedingly complicated mixture, the amounts of silver found corre-
sponding to those required for salts of acids ranging from CgH^^^, to
^12^4^2* ^ further separation and identification ot them would
require a very much larger quantity of material than was available for
the purpose. It may also be noted that from the method by which
the chief portion of these acids was obtained, it is evident that they
exist in the oil in a free state, and not in the form of esters.
From the results of this investigation, the oil of Aearum eanadense
is seen to contain the following substances :
1. A {henol, CjjHuOg,
2. Finene, apparently a mixture of the d- and ^ forms,
3. i-Iinalool,
4. ^-Borneol,
5. /-Terpineol,
6. Geraniol,
7. Eugenol methyl ether,
8. A blue oil, of undetermined composition, consisting of oxygen-
ated substances of alcoholic nature,
9. A lactone, O^^B^qO^,
10. Palmitic acid,
1 1 . Acetic acid, and
12. A mixture of fatty acids intermediate between acetic and palm-
itic acids.
In order to ascertain approximately the amounts of the principal
constituents, the following determinations were made with the original
oil:
Digitized by
Google
POWER AND SHEDDEN: DERIVATIVES OP GALLIC ACID. 73
1. The eugenol methyl ether was deterrmned by Zeisel's method.
0*1898 gram of oil gave 0*1846 gram Agl, whence eugenol methyl
ether = 36*9 per cent.
2. The amountof esters, calculatedasCjoH^y'OgHgOj, is27'5 per cent.
3. The total amount of alcohols, C^QH^gO, free and as esters, after
acetylating the hydrolysed oil, was found to be 34*9 per cent., hence
the amount of free alcohols is 13*3 per cent. In reality, the amount
of free alcohols is somewhat larger than this, as it is known that lin-
alool and terpineol cannot be quantitatively acetylated.
4. As about 2 per cent, of pinene was found in the oil, the con-
stituents of high boiling point, blue oil, &c , would amount to some-
what less than 20 per cent.
The Wellcome Chemical Research Laboratories.
VIII. — Derivatives of Gallic Acid.
By Frederick B. Power and Frank Shedden.
In a paper entitled : ** The Chemical Character of so-called lodotannin
Compounds" {PJiarm, Jotum.^ 1901, [iv], 13rl47), the authors have
recorded the results of an investigation which was undertaken for the
purpose of ascertaining the character of the compounds prepared by
the direct action of iodine on tannic acid in the presence of water.
It was shown that under these conditions no definite compound of
either tannic or gallic acid with iodine could be formed, and it therefore
seemed of interest to ascertain whether iodine could be introduced into
the gallic acid molecule by indirect methods. With this object in view,
the following method of procedure was adopted. The well-crjstallised
ethyl gallate was converted into its triacetyl derivative, which, on nitra-
tion, yielded ethyl diniirodiacetylgallate. This substance, on hydrolysis
with sulphuric acid, was converted into ethyl dinitrogallate, and from
the latter, on reduction, ethyl monoaminogallate and ethyl diaminogcUlate
were obtained in the form of hydrochlorides. These hydrochlorides
were then separately diazotised, and the resulting solutions boiled
with potassium iodide in accordance with the well-known reaction.
Although many experiments were made, it was not possible to isolate
any product containing iodine.
As most of the substances required for the original purpose of this
investigation represent new derivatives of gallic acid, the method of
preparation and their characters are here described.
Digitized by
Google
74 POWER AND SHEDDEN: derivatives of GALLIC ACID.
Ethyl DinitrodiacelylgaUaU, C^{^O^UOfifi^\{OKyOOfi^}I^.
By the direct acetyUtion of ethyl gallate, the ethyl triacetylgallate
was first prepared, which has been described by Schiff (Beilstein's
Handlnteh, 2» 1922) as a thick yellow oil, which very slowly deposits
crystals. No difficulty was experienced, however, in obtaining it in
colourless crystals from either acetic acid or alcohol. It melts at 133^.
The triacetyl ester was nitrated in the following manner. One
hundred grams of the triacetyl ester were added to a cold mixture of
50 C.C. each of nitric acid (sp. gr. 1*42) and sulphuric acid, and 150
C.C. of glacial acetic acid. The mixture, while being kept cool, was
allowed to stand for five hours. The product was poured into a
litre of water, the yellow, crystalline precipitate filtered off, and the
nitro-compound separated from some unchanged ethyl triacetylgallate
by treatment with sodium carbonate, in which it readily dissolved.
The acid filtrate from the above-mentioned yellow precipitate was
extracted several times with ether and the ethereal liquid shaken
out with carbonate solution. This was mixed with the main sodium
carbonate solution obtained as above and the whole acidified with
hydrochloric acid. The separated yellow oil soon formed a crystal-
line cake, which was collected and recrystallised from chloroform.
It formed lemon-yellow needles which melted at 165^
01466 gave 0-2260 COj and 00486 H,0. C = 420 ; H = 37.
01974 „ 13-3 C.C. moist nitrogen at 11"^ and 744 mm. N«7-9.
^is^iaOiiNg requires C « 41 -9 ; H = 3*2 ; N = 7*5 per cent.
The substance is strongly acid, dissolving in sodium carbonate with
effervescence and forming an orange-red solution. It only dissolves
slowly in absolute alcohol, and the solution gives a bluish-green color-
ation with ferric chloride.
The ethyl triacetylgallate was nitrated in another manner with some-
what different results. One hundred grams of the substance were
mixed in a flask with 100 c.c. of nitric acid (sp. gr. 1*42), and, after being
kept cool for five hours, the mixture was worked up in the manner
already described. By this method, the product consisted of a mixture
of the dinitro-ester and the dinitrodiacetyl ester.
An attempt was made to form the sodium salt of ethyl dinitro-
diacetylgallate by dissolving it in alcohol and adding one atomic pro-
portion of sodium dissolved in a little alcohol. No precipitate was
produced, even after a portion of the alcohol had been evaporated off
in a vacuum. On standing for several days, an odour of ethyl acetate
was developed, and small, bright red crystals were deposited, which
consisted of the sodium salt of ethyl dinitrogallate.
Digitized by
Google
POWER AND SHEDDEN : DERIVATIVES OF GALLIC ACID. 75
Ethyl DinUrotriaoetylgaUaU, O^i^O^^i^Q^ILfi^^^COj:^^^.
This was prepared by the acetylation of the dinitrodiacetyl com-
pound, in the formation of which one acetyl group had become elim-
inated during the process of nitration. It was obtained in colourless
needles, which gradually become yellow, and melt at 145 — 146^
01716 gave 02704 CO, and 0-0504 H^O. C = 430 ; H = 3-3.
0-2086 „ 12*4 C.C. moist nitrogen at 16^ and 768 mm. N» 70.
CijjH^Oi^j requires C = 43 5 ; H = 34 ; N = 6*8 per cent.
The substance was insoluble in sodium carbonate. Its cold alcoholic
solution gave no reaction with ferric chloride, but on boiling, a bluish-
green colour was produced.
Ethyl DinUrogaUaU, OJi;j^0^j^OB.\'QOj^^U^.
This was prepared by boiling the dinibrodiacetyl compound with 50
per cent, sulphuric acid, when it crystallised out on cooling. The
ethyl radicle was not eliminated by this procedure.
It was obtained in the form of small, yellow scales, of a somewhat
deeper colour than the dinibrodiacetyl compound. When placed in
the melting point apparatus at 80 — 85^ it melted, but af ber drying at
a gentle heat it fused at 153^.
0-8040 air-dried substance lost 0-0486 H^O at 100"". B.f>=^^0.
Ce(N02)j(OH)s-COj02H5,H30 requires H20 = 5-9 per cent.
It was recrystallised by dissolving the dried substance in absolute
ether and adding an equal volume of light pebroleum. The crystals,
after drying for a few minutes in a water-oven, softened at 15 P and
melted to a clear liquid at 153 — 154°.
. 01432 gave 01990 CO, and 0-0382 HjO. C = 37 9 ; H = 3-0.
01662 „ 14-4 C.C. moist nitrogen at 20° and 759 mm. N = 9*9.
CgHgO^Nj requires C = 375 ; H « 2-7 ; N = 97 per cent.
The substance dissolves readily in absolute alcohol, and the solution
gives an olive-green colour with ferric chloride. It could not be
hydrolysed by heating in a sealed tube with hydrochloric acid at 125°
for six hours. It was also heated in a sealed tube with 50 per cent.
sidphuric acid at 155° for five hours without altering the melting
point or other properties. Ethyl gallate, on the other hand, when
heated in a sealed tub6 with 30 per cent, sulphuric acid at 150°, is
completely hydrolysed. On boiling with an excess of alcoholic
sodium hydroxide, the substance was destroyed. It was treated
with strong ammonia in the hope of forming the amide, but only
tarry products were obtained.
Digitized by
Google
76 POWER AND SHEDDEN : DERIVATIVES OF GALLIC ACID.
Reduction of Ethyl DinitrogaiUats. — The crude nitro-compound was
reduced by warming with tin and hydrochloric acid. After the reaction
was over, the liquid was diluted with water and the tin completely
removed by hydrogen sulphide. The clear liquid was distilled in a
vacuum, and, when it had become concentrated to a small bulk,
white, needle-shaped crystals began to separate out. The distillation
was then stopped and the crystalline precipitate filtered off and
washed with dilute hydrochloric acid. This substance proved to be
the hydrochloride of ethyl monoamtnogallate. The yield was about 12
per cent, of the original substance.
The filtrate and washings were evaporated to complete dryness in
a vacuum. The residue was a crystalline mass, which was purified by
dissolving it in hot methyl alcohol and diluting the solution with hot
chloroform. A light brown or nearly white, crystalline powder was
thus obtained, which consisted of the hydrochloride of ethyl diamino-
gallate. The yield of the latter was very variable, ranging from about
10 to 25 per cent, of the original substance.
The formation of the above monoamino-derivative by reduction was
at first thought to be due to the presence of a mononitro-ester in the
material used. This, however, was not the case, inasmuch as a pure
ethyl dinitrodiacetylgallate afforded the same yield of the monoamino-
hydrochloride. The conclusion may thus be drawn that the formation
of the monoamino-derivative is due to some change in the process
of reduction.
Hydrochloride of Ethyl MonoaminogcdUUe^
C,H(NH2)(OH)3-COAH5,HCl,H20.
This subbtance has the following characters. It is readily soluble
in water and the solution remains colourless. Its alcoholic solution
gives a dark green colour with ferric chloride. It melts at 210° with
blackening and frothing. When recrystallised by dissolving it in hot
absolute alcohol and adding chloroform to the solution, it still melted
at 210° and was quite white, showing no tendency to change on
keeping. When dissolved in a little water, it could be precipitated
by the addition of strong hydrochloric acid, and this reaction, besidecf
the other characters of the substance, distinguishes it from the diamino-
gallate. It may be heated in a water-oven without any change in weight.
01186 gave 01760 CO, and 0-0540 HgO. C = 40 5 ; H = 5-05.
01290 „ 0-1908 CO, „ 00586 HjO. 0 = 40-3; H = 5-05.
01724 „ 8*8 c.c. moist nitrogen at 24° and 753 mm. N»5*65.
01950 „ 01058 gram AgCl. 01=13-4.
0-5504 at 105° lost 0-0392 HjO. HjO = 7-1.
Cj,HiiOjN,HCl,H20 requires 0 = 404 ; H = 5-2 ; N = 5-2 ; 01 = 1 33 ;
HjO=6-7 percent.
Digitized by
Google
POWER AND SHEDDEN: DERIVATIVES OF GALLIC ACID.
W IHazogaOaU, g«H(OH),(CO,C,H,).|j
The ethyl monoaminogallate was dissolved in an excess of dilute
hydrochloric acid, and to the ice-cold solution a dilute solution of
sodium nitrite was slowly added until there was a permanent excess.
The liquid was heated on a water-bath for 20 minutes, and, after
cooling, the brown crystals were filtered off.
The product was almost insoluble in cold water, but dissolved in
boiling water, and on cooling the solution long, orange-yellow needles
were deposited. The solution was yellow, and gave a deep purplish-
brown colour with ferric chloride. The substance melted with sudden
decomposition at 182°. When recrystallised from dilute acetic acid, it
formed fine, reddish-brown needles melting as before at 182°.
0 2350 at 100° lost 00178 Hp. HjO = 7-6.
Q^Jd^^.'E.fi requires 'S,jd-*l'i per cent.
The dried substance was analysed with the following result :
01108 gave 0-1938 CO, and 00432 HjO. C = 47-7 ; H =■ 4-3.
0 0924 „ 10 c.c. moist nitrogen at 24° and 769 mm. N » 12*3.
CgHgOgNg requires C = 48 2 ; H = 3-5 ; N « 125 per cent.
0*1114 gram dissolved in 15*59 grams of pure phenol depressed the
freezing point by 0*263°. This corresponds to a molecular weight of
201. Mol. wt. of Cj^HgOgNj = 224.
One gram of the substance was heated with three times its weight of
water in a sealed tube at 220° for four hours, when complete solution was
effected. The dark brown liquid was filtered from a small amount of
a black residue, saturated with ammonium sulphate, and extracted
with ether. The ethereal solution, when washed with water, dried
with sodium acetate, and evaporated, left a yellowish, oily liquid which
became crystalline. The crystals, after washing with a little light
petroleum, melted at 139°, and were free from nitrogen. Their
aqueous solution gave a brown colour with alkalis and a bluish-black
one with ferric chloride. After recrystallising this from toluene, about
0*2 gram of the substance was obtained, and it then melted at 140°
without decomposition. It was dried in a water-oven and then
analysed :
01084 gave 0*2174 COg and 0-0504 H,0. C - 54*7 ; H = 5*2.
CgHj^Og requires C = 54'5 ; H = 5*l per cent.
The substance thus produced was therefore undoubtedly ethyl
gallate, the nitrogen having beeu completely eliminated by heating
the diazogallate with water.
Digitized by
Google
78 POWER AND 8HEDDEN: DERIVATIVES OF GALLIC ACID.
On treating the diazo-compound with stannous chloride in cold
hydrochloric acid, considerahle effervescence was produced and it
became completely dissolved. After removing the tin by hydrogen
sulphide, the products of the reaction |Were found to be ammoniuin
chloride and ethyl gallate. The production of ethyl gallate from this
diazo-compound will serve to explain the formation of the monoamino-
ester by the reduction of ethyl dinitrogallate (p. 76).
Hydrochloride of Ethyl JDiaminogallate,
C,(NHj),(OH)3-CO,0,H„2HCI.
* This substance, as obtained in the manner already described, forms
a fine, crystalline powder of a light brown colour. By redissolving it
in hot absolute alcohol (in which it is somewhat sparingly soluble)
and adding ethyl acetate it becomes much lighter in colour, but when
kept shows a tendency to darken. The alcoholic solution rapidly
assumes a pink hue. It is very easily oxidised. It dissolves readily
in water, but the solution almost immediately becomes blue, and, on
standing, blue flakes are deposited. The blue colour is intensified by
the cautious addition of ferric chloride, but is destroyed by adding an
excess. If the solution in water be acidified with hydrochloric acid,
the blue colour changes to pink (compare Ber., 1887, 20, 335 ; 1893,
26, 2184). Unlike ethyl monoaminogallate^ it cannot be precipitated
from its aqueous solution by hydrochloric acid. It melts with decom-
position at 197°. After drying in a vacuum, it was analysed.
0-1318 gave 0-1738 00a and 0-0670 H^O. C « 35-9 ; H = 4-8.
0-1330 „ 01746 COg „ 00538 Hp. 0 = 357 ; H = 4-5.
0-2129 „ 19-6 c.c. moist nitrogen at 23° and 764 mm. N= 10-4.
0-1588 „ by Oarius' method, 01520 AgOl. 01 = 23-7.
09H„05N2,2HC1 requires 0 = 35-9; H = 46 ; N = 9-3 j 01 - 23-6 per cent.
In order to ascertain the action of nitrous acid on this diamino-ester,
6*6 grams of the substance were mixed with an excess of dilute hydro-
chloric acid. The resulting dark coloured solution was cooled with
water and a dilute solution of sodium nitrite gradually added, which
caused considerable effervescence and the evolution of some nitrous
fumes. This liquid was extracted with ether, but on distillation the
latter left only a very slight residue. The remaining liquid was
heated on a water-bath, when there was considerable effervescence,
and, after this had ceased, a small portion of the liquid, when boiled
with potassium hydroxide, evolved ammonia. The remainder was
extracted with chloroform, then made alkaline with sodium carbonate
and again extracted with chloroform, but in neither case was any
definite product obtained.
Th£ Wsllcom£ Chemical Rksearch Laboratoeies.
Digitized by
Google
STEVENS: THIOCARBAMIDE HYDROCHLORIDE. 79
IX. — Thiocarbamide Hydrochloinde.
By Henry P. Stevens, M.A., Ph.D.
Thiocarbamide hydrochloride is stated by Glutz {AnnaUn^ 1870, 154,
40) to be obtained from thiocarbamide stannochloride by removal of
the tin and concentration of the resulting aqueous solution, as a crys-
talline substance which could be purified by recrystallisation from
alcohol. It was not analysed, nor was its melting point given, and the
only evidence brought forward by Glutz to show that it was thiocarb-
amide hydrochloride is the fact that with platinic chloride it gave a
double salt, and with silver nitrate a mixed precipitate of silver chloride
and the silver compound of thiocarbamide.
Eeynolds {AnnaUnj 1869, 150, 232) was unable to prepare the hydro-
chloride, whilst Claus {Annalen, 1875, 179, 131) states that, like
Eeynolds, he had been unable to -prepare the hydrochloride directij/f
but had often obtained the pure salt by Glutz's method, in spite of
which assertion, however, no analysis is given nor is the salt in any
way further characterised. It is therefore a matter of doubt whether
thiocarbamide hydrochloride has hitherto been isolated in a pure state.
On investigating the question, it was found that identical products
were obtained by Glutz's method and by bringing together aqueous
solutions of thiocarbamide and hydrochloric acid in the calculated pro-
portion.
The white^ crystalline substance produced on evaporation of the
solutions on the water-bath is very soluble in alcohol, and when frac-
tionally recrystallised from this solvent, it yields, in addition to some
unchanged thiocarbamide, well-formed, prismatic crystals having an ill-
defined melting point and containing an amount of chlorine too small
for such a compound as CS(KH2)2,HC1. Repeated crystallisation from
alcohol, instead of purifying the compound, lowered the percentage of
chlorine without, at the same time, yielding any free thiocarbamide.
Eventually the hydrochloride was obtained pure by the following
method. Thiocarbamide was dissolved in more than sufficient of the
most concentrated, warm, aqueous hydrochloric acid to convert the
whole into hydrochloride. On allowing the solution to stand, the
hydrochloride separated out in thick, massive crystals. The mother
liquor was poured off from the crystals, which were then redissolved
hy g^tly warming in the smallest possible quantity of hydrochloric
acid, from which, on standing, the greater part again separated. It is
difficult to dry the crystals without slight loss of hydrogen chloride ;
they may, however, be obtained in a pure state by pouring off the mother
liquor, washing them rapidly with cold alcohol on the filter pump, and
drying them over calcium chloride. On analysis :
Digitized by
Google
80 STEVENS: THIOCARBAMIDE HYDROCHLORIDE.
I. 0-1212 gave 01546 AgCl. 01 = 31-54.
II. 0-2374 „ 0-3017 AgCl. 01 = 31-42.
OHgNjClS requires 01 = 31-38 per cent.
As thiocarbamide ii^ aqueous solution has a neutral reaction with
respect to litmus, the whole of the chlorine can be accounted for as
hydrochloric acid by titration with decinormal ammonium hydroxide
solution. Thus, in analysis II, the substance was titrated, before pre-
cipitation, with diver nitrate and gave 01= 31*46 per cent.
The salt, when exposed to air, rapidly effloresces with the loss of some
hydrogen chloride, of which about one-half can be removed by prolonged
exposure in a vacuum over strong sulphuric acid and potassium hydr-
oxide. When heated, it softens gradually and melts below 100°. It
is extremely soluble in water or alcohol. If silver nitrate be gradually
added to a solution of the hydrochloride, the precipitate first formed
redissolves immediately in the excess of the hydrochloride solution, and
on allowing the clear liquid to stanti, fine, needle-shaped crystals separ-
ate out which melt at 172° and on oxidation with nitric acid yield
silver chloride. They appear to be identical with the compound
20S(NH5)2,Ag01 (m.p. 170— 171°) obtained by Reynolds (Trans., 1892,
61, 252) by dissolving silver chloride in a hot alcoholic solution of
thiocarbamide,
Thiocarbamide forms additive compounds with alkyl iodides and
bromides on standing in the cold or heating in sealed tubes (Olaus,
Annalen, 1875, 179, 145 3 Bernthsen and Klinger, Ber,, 1878, 11,
492, drc.) ; but no statement, however, is to be found with regard to
its behaviour with the alkyl chlorides. On treating a solution of thio-
carbamide in alcohol with ethyl chloride, freed from hydrochloric acid
by bubbling through water with calcium carbonate in suspension, no
appreciable action took place even on warming the solution. Never-
theless, it was possible that ethyl chloride, formed in the solution itself
by the action of hydrochloric acid on the alcohol, might prove more
reactive, and this was eventually found to be the case.
An alcoholic solution of thiocarbamide hydrochloride, prepared by
dissolving thiocarbamide in about ten times its weight of alcohol in
which the necessary amount of hydrogen chloride had been dissolved,
was boiled for several days in a reflux apparatus on a water-bath. The
solution was evaporated down twice with fresh quantities of alcohol to
remove any slight excess of hydrochloric acid. The product, a thick,
unpleasant smelling oil, solidified completely on standing and stirring
with a glass rod. Like thiocarbamide hydrochloride, it was extremely
soluble in water or alcohol, but insoluble in other solvents provided
they were dry, and on this account much difficulty was experienced
in finding a suitable solvent for its recrystallisation. Eventually the
product was dissolved by gently warming and shaking in glacial acetic
Digitized by
Google
METHOD FOR DETERMINING SMALL QUANTITIES OP CARBONATES. 81
addy a few drops of water or alcohol being added to promote solatioo.
Dry ether, insufficient in amount to cause any permanent precipitation,
was then added in small quantities at a time to the cold solution, and
the whole set aside to stand ; a crop of crystals formed which was
filtered off, and the mother liquor treated with more ether. In this
manner, by a process of fractional crystallisation, the new substance
was obtained in a state of purity. It is more soluble than thiocarb-
amide hydrochloride in the mixture of glacial acetic acid and ether, and
separates when pure from the same solvent in long, slender prisms.
It melts gradually just below 100^. The aqueous solution of the salt
is neutral to litmus, but the whole of the chlorine is precipitated as
silver chloride by silver nitrate in dilute nitric acid solution. Chlorine
estimations showed that it is an additive product of thiocarbamide and
ethyl chloride, or, from another point of view, that it is ethyl-^-thio-
carbamide hydrochloride.
0-2097 gave 02122 AgCl. Cl = 2502.
0-2183 ,. 0-2227 AgCI. 01 = 25-22.
CsH^NgClS requires 01 = 25 20 per cent.
This hydrochloride behaves similarly to the hydriodide obtained by
direct combination of thiocarbamide and ethyl iodide.
It follows, therefore, that thiocarbamide' hydrochloride cannot be
recrystallised from alcohol, as it reacts with it to give ethyl-^-thiocarb-
amide hydrochloride.
Chemical Laboratory,
St. -Thomas' Hospital, S.E.
X. — A Method for Determining Small Quantities of
Carbonates,
By Alfbeo Daniel Hall and Epwabd John Eussell.
The determination of small quantities of carbonates in material like
soil is attended with many difficulties, owing to the solubility of the
carbon dioxide in the acid used for decomposing the carbonate. When
the soil contains 2 per cent, or more of carbonates, calculated as
calcium carbonate, Schei bier's apparatus may be used, and the
empirical correction for solution of the carbon dioxide (Warington,
Chim, News, 1875, 31, 253) will not introduce a greater experimental
error than attaches to the natural variation of the sample for analysis.
But with small proportions of calcium carbonate, 05 per cent, and
VOL. LXXXI. G
Digitized by VjOOQIC
82 HALL AND RUSSELL: A METHOD FOR
below, Scheibler's apparatufi becomes unworkable, for all the gas pro-
duced remains in the reacting acid. Gravimetric methods, where the
carbon dioxide is either weighed directly or by difference, require very
refined manipulation when 100 grams of soil have to be attacked by
acid and the mixture boiled, <kc., to obtain quantities like 01 gram of
carbon dioxide. Working in a vacuum by ordinary gas analysis
methods, the carbon dioxide can be collected and measured, but there
are, again, difficulties due to bolution which render the process tedious
and susceptible of error.
The suggestion has often been made that the soil should be treated
with a known volume of standard acid, and the amount of calcium
carbonate present calculated from the acid neutralised. This process,
however, always gives results which are too high, owing to the
fact that various humates, silicates, and in some cases ferric oxide, are
also attacked by the acid without liberating any acid which affects
the indicator.
Stutzer and Hartleb {Zeit angew. Ch«m., 1899, 12, 448) have pro-
posed to distil the soil with a solution of ammonimu chloride ; the
calcium carbonate present forms ammonium carbonate by double de-
composition ; this dissociates, and the ammonia is caught by standard
acid and titrated. This method is open to all the sources of error
indicated above (compare Schutte, Zeit, angew, Chem,, 1899, 14, 854 ;
Woy, Chem, Centr,, 1899, ii, 847 ; and Immendorff, Zeit, angew, Chem.^
1900, 16, 1177).
In searching for a more workable process, the authors have devised
the apparatus described below, by means of which the main source of
error in determinations of carbon dioxide, its solubility, is eliminated.
The process is also reasonably rapid and requires no special skill in
manipulation.
The apparatus is figured on p. 83. (ii) is the reaction bulb, about
60 c.c. in capacity. It is connected from below with the small funnel
(B), carrying the stopcock (a), (ii) is connected to the rest of the
apparatus by a cup joint at (5). {C) is a second bulb, rather smaller
than {A) (in the apparatus actually used, its capacity was 42*5 c.c.) ;
on the tube connecting (C) to the rest of the apparatus is a stopcock
(c). The stopcocks and cup joint must be well ground and lubricated
so as to maintain a vacuum. (Z>) is a capillary tube 800 mm. long,
dipping into a small reservoir of mercury and serving as a manometer;
a third stopcock {d) is placed between the manometer and the pump.
The bulbs (ii) and (C) can be enclosed in a water-bath. Before the
apparatus is fixed on the stand, the capacity of the bulb (C) must be
ascertained with accuracy ; this may be done by filling the bulb with
mercury and then weighing the mercury when shaken out and
collected.
Digitized by VjOOQIC
DETERMININQ SMALL QUANTITIES OF CARBONATES.
83
Two to 10 grains of the Bubstance in a finely powdered state are
introduced into (A) and covered with water, the cup joint is wiped,
well lubricated, and (A) then joined to the rest of the apparatus. The
cup joint is sealed with a little mercury, a little also being poured into
the funnel (B), so that the bore of the tap is quite filled. The stop-
cock (c) is opened, and connection made to a good pump until approxi-
mately a vacuum is established inside the apparatus. Entire freedom
from air is not necessary, but when determining very small quantities
of carbon dioxide (1 to 5 c.c. from 10 grams of soil), the pressure
should be run down until the manometer indicates little more than the
To"poVv»|3 i
n
vapour pressure of water within the apparatus. When dealing with
larger quantities of gas, for example, 10 to 20 c.c, a mercury pump is
not necessary, it is sufficient to use a good water pump or hand air
pump that will establish an internal pressure of 50 — 60 mm. of
mercury.
The stopcock (d) is closed, the height of the mercury in (D) and the
temperature of the water-bath are read ; this readings ft. Stopcock
(c) is then closed, a well-boiled and cooled mixture of equal volumes of
sulphuric acid and water is placed in the funnel {B\ and a few c.c.
introduced into the reacting bulb. Since it enters from below, the
liquid and soil get well stirred up ; the mixture is left for a few minutes
G 2
Digitized by VjOOQIC
84* HALL AND RUSSELL: A METHOD FOR
to cool down to the temperature of the bath, then the apparatus is
shaken to expel the carbon dioxide present in excess in the solution,
and allowed to stand, with further occasional shaking, until the gauge
shows a constant reading.
The gas evolved causes an increase of pressure inside the apparatus,
and the manometer column is read again »R^. Communication is
now made with the bulb (C) by opening the stopcock (e\ the gas ex-
pands again into (0) and the mercury rises again in (Z>). A little
time and shaking cause the gas dissolved in the liquid in (A) to come
into equilibrium with the gas above at the new pressure ; the mano-
meter column is then read when constant, and « Ry
Assuming the temperature, t, of the water-bath has remained
constant, and calling d the difference in mm. between R aod R^^,
d! the difference in mm. between R and B2, and C the volume of
Cd,d*
the bulb (C), then the volume of gas evolved at N.T.P. = /^ _ vyiMs ^
273
-^— — * The operation amounts to finding an unknown volume of
273 + 1
gas in (A) by the change in pressure produced when it expands by a
known volume. The advantage of the method lies in the fact that the
volume of soil, liquid, <Sec., which may have been introduced into (ii)
is immaterial, and does not appear in the calculation, and especially
* The complete proof of the formnla given is as follows :
Let X = the p. v. of the carbon dioxide evolved at the given temperatnre.
A = the volume of the apparatus excluding ((7) and the liquid in {A).
G = the volume of ((7).
P s= barometer reading.
a = tension of aqueous vapour at the given temperature.
k =s the volume of carbon dioxide soluble in the liquid in (^) at unit
pressure.
J2, i^ and 22, = the readings as above.
At starting, the apparatus contains some air = (^+J?)(P- J2-a).
((7) is shut off and x of carbon dioxide evolved. Then :
x^AiJ-n-ck) = -4(P-J2,-(»)+ifc(5-J2,)
^ = ^+* .:....[!]
The gas is then allowed to expand into ((7), when
aj + (^ + (7)(P-i2-a) = (^ + C7)(/'-i?3-a) + ^«-i2a)
= A^GArh [2]
Combining [1] and [2]
a? _ G-Vx
whence x = ^(^-^Kf-^.) = O^ ^^^^^^
Digitized by
Google
t>ET£RMINtNQ SMALL QUANTITIES OF CARBONATES.
85
that the effect of the gas remainiog dissolved in the liquid in (A) is
also eliminated. The liquid is saturated by the gas, so that the gas
irithin and without the liquid is in equilibrium. When the volume is
increased by opening the stopcock to ((7), an amount of gas, propor-
tional to the reduction in pressure, escapes from the liquid. In brief,
the gas contained in the liquid of (A) obeys the same laws as the gas
above the liquid, and the liquid becomes practically only a portion of
the gas-filled space of (A),
It is necessary to have some solid particles like soil or glass beads
in (i), otherwise the liquid becomes, and remains, obstinately super-
saturated with carbon dioxide, nor can the excess be shaken out. This
tendency to sapersaturation forms the chief difficulty in working with
the apparatus ; the amount of substance taken should be such that
the pressure of the carbon dioxide liberated does not exceed 100 or
150 mm., or the time required to obtain equilibrium becomes very
great, and may even amount to 2 or 3 hours. The lower the pressure,
or, in other words, the smaller the amount of carbonate present, the
easier the determination is to carry out ; the limit is fixed only by the
accuracy with which the gauge can be read.
Appended are a few numbers obtained with the apparatus in the
case of pure sodium carbonate and Iceland spar, the bulb {A) being
half filled with glass beads :
Number.
Substance taken.
COaatN.T.P. (calc).
COaatN.T.P. (found).
1
2
8
4
6
6
7
0 000624 gram N8,C0,
000125 „ „
0 0026
0 005 „ „
0010 „ „
0 020 „
0 0508 „ CaCO,
018 c.c.
0-26 „
0-63 „
106 „
2-11 „
4-22 „
11-26 „
0-15 CO.
0-26 „
0-61 „
1-00 „
200 „
4-88 „
11-18 .,
The apparatus may be conveniently applied to auy reaction involv-
ing the measurement of a gas evolved from a liquid.
South Eastern Agricultural College,
Wye.
Digitized by
Google
86 MORGAN : INFLUENCE OP SUBSTITUTION ON THE
XL — Influence of Substitution on the Formation of
Diazoamines and Aminoazo-compounds.
By Gilbert Thomas Mobqan, D.So.
The action of a diazonium salt on an aromatic amine gives rise to a
diazoamine or an aminoazo-<!ompound, according as the diazonium
residue R'N^* remains attached to the aminic nitrogen or takes up
its position in the aromatic nucleus.
The former mode of combination occurs with the primary mono-
amines of the benzene series, these bases yielding diazoamines. The
mono-alkylated monoamines of the benzene series, on the other hand,
exhibit a tendency to form azo-compounds ; methylaniline, for example,
when treated with diazobenzenesul phonic acid, yields a mixture of the
diazoamino-compound, SOgH-CgH^'Ng'NMePh, and the isomeric
aminoazo-acid, SOgH'CgH^-Ng'OgH^'NHMe (Bemthsen and Gteske,
Ber., 1887, 20, 926, Bamberger and Wulz, Bm-., 1891, 24, 2082).
A somewhat heterogeneous class of bases including the following
sabstances : — diphenylamine, the naphthylamines and their mono-alkyl
derivatives, m-phenylenediamine and certain of its homologues and
substitution products, gives rise to aminoazo-com pounds without the
intermediate formation of stable diazoamines. Dimethylaniline and
a few other tertiary amines also yield aminoazo-derivatives, but
with these bases the*production of diazoamines is obviously impossible.
There is some reason for believing that the difference between
the behaviour of aniline and that of m-phenylenediamine towards
diazonium salts is due to the greater reactivity of the disubstituted
ring, so that substituent radicles find their way more readily into the
aromatic nucleus of the diamine than into that of the monoamine;
The introduction of chlorine and bromine by the action of hypo-
chlorous and hypobromous acids respectively is a case in point ; the
latter halogen enters the nucleus of the diamine so easily that
m-phenylenediacetyldibromoamine, CgH4(NBrAc)2, could not be
isolated (Morgan, Trans., 1900, 77, 1209 ; Chattaway and Or ton,
Ber.f 1901,34, 160), whereas phenylacetylbromoamine, Ph'NBrAc, is a
comparatively stable substance (Chattaway and Orton, Trans., 1899,
76, 1046; 1900,77,800).
Beasoning by analogy, it seems probable that the initial phase of
the interaction between m-phenylenediamine and a diazonium salt
involves the formation of an unstable diazoamine, thiB substance
immediately changing into the isomeric aminoazo-derivative. This
assumption is, however, not supported at present by any direct experi-
%l evidence. With the view of gaining additional information as
Digitized by
Google
FORMATION OF DIAZ0AHINE8 AND AMINOAZO-COMPOUNDS. 87
to the coarse of this reaction, the formation of azo-derivatives of the
homologues and substitution products of m-phenylenediamine has been
studied. 2 : 4<Tolylenediamine has long been known to be as reactive
towards diazo-compounds as m-ph^Qjlenediamine itself, and more re-
cently this was also shown to be true of l-chloro-2 : 4*pbenylenediamine
and its bromine analogue. These bases readily yield azo-compounds,
which, except for some slight difFerences in shade of colour, closely
resemble those derived from the parent base. The hydrochlorides of
chloro- and bromo-chrysoidines, for example, are obtained in crystals
very similar in shape and colour to those of the ordinary chrysoidine
of commerce.
The disubstituted m-diamines may be divided into two series with
reference to their behaviour towards diazonium salts. The first series
consists of the bases having the general formula
Y
1. x<;^ \nh„
whilst the second comprises all those disubstituted m-diamines which
contain one free para-ortho-position with respect to the ami no-groups.
The bases of the second series can be grouped together, because, pro-
viding this condition is fulfilled, the nature and position of the two
substituent radicles exercise very little influence either on the course
of the reaction or on the colour of the resulting azo-compound. An
amine of this series may possess any one of the following formulaa :
X Y
Y
Y
■< >
NH,< >
"«■<__>
NWg
X JUHg
X NH,
II.
III.
IV.
Diamines con*esponding with formulas II and III have been investi-
gated and the results compared with those obtained in the case of
their isomerides belonging to series I.
The first diamines to be examined from this standpoint were the
two diamino-m-xylenes described by Greviogk {Ber.y 1884, 17, 2426),
these bases having the following constitutions :
Me
Me NH^
NHg/ ^Me
NHj/ \Me.
NH2
The former is produced by the reduction of 2 : 6-dinitro-m-xylene,
the chief product of the nitration of fn-xylene; whilst the latter is
Digitized by VjOOQ IC "
88 MORGAN : INFLUENCE OF SUBSTITUTION ON THE
obtained from 2 : i-dioitro-mzylene, this nitro-compound being a bye-
product in the same operation.
Grevingk {loe. eU.) states that both these bases, when treated with
benzenediazonium chloride, give colouring matters of the chrysoidine
type, but he does not seem to have isolated any definite products. Witt,
however, after an unsuccessful attempt to prepare an azo-compound by
the action of diazobenzenesulphonic acid on the symmetrical base, main-
tains that this diamine does not yield chrysoidine derivatives (^er.,
1888, 21, 2419).
As a matter of fact, the two isomerides behave very differently
towards diazonium compounds. The consecutive base reacts just like
2 : 4-tolylenediamine, readily yielding a red azo-compound when treated
with benzenediazonium chloride in the presence of sodium acetate;
the symmetrical diamine, on the other hand, gives ris6 to a volumin-
ous, brownish-yellow precipitate, [which froths considerably, evolving
nitrogen, and finally becomes resinous, even while in contact with the
ice-cold mother liquor. This unpromising result was experienced with
other symmetrically disubstituted m-diamines and seemed to confirm
Witt's conclusion.
These failures could not, however, be accepted as conclusive evidence
that the diamines of the type indicated by formula I cannot yield
azo-compounds, inasmuch as the firm of Oehler & Co. has patented
the production of colouring matters derived from 2 : 4-tolylenediamine-
5-sulphonic acid (D.R.-P. 40905),
Me
'2N
a substance having a constitution similar to that of the bases
in question. Accordingly, further experiments were made with
different diazo-compounde, until, finally, it was found that these
bases would combine with primulin dyed and diazotised on the
cotton fibre. Under these conditions, the symmetrically disubstituted
diamines yielded azo-colouring matters possessing a yellowish-brown
colour, and differing altogether from the reddish-brown compounds
produced by similar means from the diamines belonging to the second
series. These results show unmistakably that the relative position of
the azo- and amino-groups is the most important factor in determining
the shade of colour produced.
With the experience gained in these experiments on diazotised
primulin, another attempt was made to prepare azo-oompounds
from 4 : 6-diamino-m-xylene and the simpler diazonium salts. The
product of reaction was allowed to remain in the ice-cold solution for
Digitized by
Google
^ORMATIOK OF DIA^OAMtl^ES AND AMINO AZO-COMPOI^NDS. 89
one or two hours and the tarry precipitate then washed, dried, and
carefully extracted with alcohol or benzene. In this way, a small
yield of aminoazo-compound was produced and the reaction was shown
to occur both with benzene- and p-toluene-diazonium salts.
The investigation was now extended to the symmetrical base,
5-chloro-2 : 4-tolylenediamine, simultaneously described by Reverdin
and Cr^pieux {Bw,y 1900, 33, 2507) and by Morgan (Trans., 1900, 77,
1209), and also to its isomeride, 2-chloro*3 : 5'tolylenediamine, prepared
by Nietzki and Rebe {Bw,, 1892, 25, 3005). The latter compound
contains one free para-ortho-position with respect to the amino-groups
and might be expectod to resemble 2 : 4-tolylenediamine and 2 : 4-
diamino-m-xylene in its behaviour towards diazouium salts. This
anticipation was completely confirmed ; the interaction resulted in the
immediate formation of an azo-compound, the yield being practically
quantitative.
The symmetrical isomeride behaved like the similarly constituted
4 : 6-diamino m-xylene j a brownish-yellow, voluminous precipitate
was again produced, which evolved nitrogen and speedily became
resinous. In this case, also, an azo-compound was extracted from the
tarry product, but the yield was even poorer than that obtained in the
experiment with the xylene base.
Since 4 : 6-dichloro-m-pheDylenediamine (Traus., 1900, 77, 1206)
combines with diazotised primulin, an attempt was made to condense
it with benzenediazonium chloride and its p- toluene homologue. In
these experiments, there was a considerable amount of frothing
and formation of resinous product, but the precipitate, on extraction,
yielded a large amount of unaltered base and did not furnish any azo-
compound. Although this result does not establish beyond doubt the
fact that an azo-derivative is not produced, yet, in conjunction with the
evidence obtained from the preceding experiments, it seems to indicate
that, with these symmetrically disubstituted m-diamines, the tendency
to form an azo-compound diminishes as the acidity of the molecule
increases. This increase in acidic character results from the gradual
replacement of methyl by chlorine, the pairs of substituent radicles in
the three diamines being respectively 2Me, ClMe, and 2C1.
The brownish-yellow precipitates, which evolve nitrogen and become
tarry^ are probably unstable diazoamines. This conjecture receives
additional support from the fact that under comparable conditions
diaminomesitylene yields a similar, readily decomposable product,
and in this instance the unstable substance cannot possibly be an azo-
derivative. Meldola has also noticed the formation of a labile inter-
mediate diazoamino-compound in the preparation of j[>-nitrobenzene-
5-azo-4-m-xylidine (Trans., 1883, 43, 428).
The aminoazo- bases derived from 4 : 6-diamino-m-xyIene and 5-chloro-
Digitized by
Google
90 MORGAN: INFLUENCE OF StTBSTlTtJTlON ON TflE
2 : 4-tol7lenediamine contain the azo-groups in the position contigaons
to the two amino-radicles, their constitution being indicated by the
general formula
X NH2
The azocompound derived from 2 : 4-diamino-tii-xylene has the
constitution
Me NH,
NH,^ ^N,Ph,
Me
since the position ocoupied by the azo-group is the only reactive
position available in the original diamine.
In the case of 2-chloro-3 : 5-tolylenediamine, the azo-group may enter
the ring in either the di-ortho-posibion or in one of the two para-ortho-
positions with respect to the amino-radicles. The ease with which the
azo-compound is produced in almost theoretical yield renders it in the
highest degree probable that the entrant radicle takes up the para-ortho-
position forming a colour base having the following configumtiouy
Moreover, the azo-compound produced on the cotton fibre from this
diamine and diazotised primulin has the reddish-brown colour charac-
teristic of the colouring matters having this constitution.
The naphthylamines and their derivatives containing hetero-
nucleal substituents belong to the class of amines yielding azo-
compounds without the intermediate formation of stable diazoamines,
and are thus distinguished from aniline and its homologues, the only
exception on record being ^-naphthylamine-S-sulphonic acid, which,
unlike its isomerides, gives a stable diazoamino-compound with
benzenediazonium chloride (Witt, Bw,y 1888, 21, 3483).
A similar difference has been noticed in the behaviour of the two
series of amines towards formaldehyde, jS-naphthylamine yielding
derivatives containing the methylene carbon atom attached to the aro-
matic nuclei (Trans., 1898, 73, 536), whereas aniline and its homo-
logues give rise to intermediate compounds of the methyleneaniline
and methylenedianiline types, containing methylene united with the
nitrogen of one or two amino-groups. The investigation of these
methylene compounds (Morgan, Trans., 1900, 77, 814) also showed
Digitized by
Google
fORitATION OJ" DlAZOAMlNES AND AMlN0A:20-C01tP0UNDS. 91
that the presence of a substituent radicle in the a-position contiguouB
to the amino-radiole of j3-naphthjlamine prevented the transference of
methylene into the ring.
Inasmuch as in their reactions with aromatic amines, formaldehyde
and diazoniom salts attack similar positions in the basic molecule
it might be expected that a j3-naphthylamine derivative substituted
in the manner indicated would yield a diazoamine but not an azo-
compound.
One compound of this type, namely, 2-diazoamino-l-chloro-4'bromo-
naphthalene, has already been obtained by Meldola and Streatfeild
(Trans., 1895, 67, 911) by the action of nitrous acid on l-chloro-4-
bromo-/3-naphthylamine. If the production of this diazoamine is deter-
mined by the presence of the chlorine atom in the a-position conti-
guous to the amino-radicle, then it should be possible to obtain similar
compounds from l-chloro-j3-naphthylamine. The experimental results
amply confirmed this assumption. The action of nitrous acid (1 mol.)
on this amine (2 mols.) gave rise to 2-diazoamino-l'ChloronaphthaIene,
CI 01
CO'''™00 •
a well defined diazoamine resembling the compound described by
Meldola and Streatfeild.
Mixed diazomines also were produced by the interaction of various
diazonium salts and l-chloro-j9-naphthylamine. ;7-NitrobeDzene-2-
diazoamino-1-chloronaphthalene, prepared by the action of ^^-nitro-
benzenediazonium chloride on this base, was also produced by the con-
densation of l-chloro-2-naphthalenediazonium chloride on j9-nitroaniline ;
this result indicates that Eekul^'s rule relating to the formation of
mixed diazoamines is applicable to those containing both naphthalene
and benzene nuclei.
These diazoamines do not show any tendency to change into amino-
azo-compounds containing the azo-group attached to the naphthalene
nucleus. Here also, as in the case of the methylene derivatives, the
directing influence of the amino-radicle in ^-naphthylamine seems to
be exerted only in one direction, and accordingly the substituent
radicles readily shift into the contiguous a-position, but do not replace
the hydrogen attached to the adjacent j3-carbon atom.
Digitized by
Google
92 morgan: influence of sxibstitution on tMeI
EXP£BIHENTAL.
Action of Diazonium Salts on tlis m-Diaminomxi/lenes,
Preparation of the Diamines, — vn-Xylene is nitrated in quantities of
250 grams by slowly adding a well-cooled mixture of concentrated
nitric and sulphuric adds (2*3 parts HNO3 ^^ ^P* S^' 1*52 to 2 parte
HgSO^) to the hydrocarbon surrounded by ice and salt. When the
hydrocarbon is added to the acid, a considerable amount of trinitro-
xylene is formed at the commencement of the operation. The mixture,
after remaining for several hours at the ordinary temperature, is
warmed for two hours at 40 — 50° and then poured on to ice. The
crude, viscid nitro-compounds are drained from oily products and the
solid residue is crystallised from alcohol, the last operation being
repeated two or three times. The crystals obtained in this manner
consist of 4 : 6-dinitro-m-xylene[melting at 91 — 92° (Grevingk, loc cU.,
gives m. p. 93°). The alcoholic mother liquors, when united and
allowed to evaporate spontaneously, deposit an oily substance which is
withdrawn as soon as the separation of crystalline product commences.
This second crop of solid nitro-compound is crystallised repeatedly
from alcohol, and the final product consists, very largely, of 2 : 4-dinitro-
m-xylene, crystallising in rosettes of hard, well-defined, flattened
needles which melt somewhat indefinitely at 58 — 61°. Grevingk
gives 80° as the melting point of the pure compound. As repeated
crystallisation does not raise the melting point, the substance is reduced
without further purification. The alcoholic mother liquors obtained
by working up 1250 grams of m-xylene yield about 120 grams of the
partially purified nitro-compound. The final mother liquors furnish a
further quantity of oily nitro-compound. These oily products, obtained
at various stages of the operation, when united and reduced give rise to
impure 4 : 6-diamino-m-xylene.
The diamines are obtained by reducing the respective dinitro-
xylenes with iron, 100 grams of the nitro-compound being treated
with 130 grams of iron filings, 800 c.c. of water, and 12 c.c. of con-
centrated hydrochloric acid. The whole of the water is not added at
the commencement of the operation, but about 300 c.c. are introduced
gradually duiing the reduction in order to moderate the reaction, which
sometimes becomes very violent. The product, rendered alkaline with
8 grams of sodium hydrogen . carbonate, is filtered from iron oxide ;
the filtrate acidified with acetic acid is treated with excess of acetic
anhydride (about 60 grams). The precipitated diacetyl derivative is
collected after 12 hours ; the filtrate, when concentrated and treated
with a further quantity of the anhydride, yields a second crop of
diacetyldiaminoxylena After crystallisation from glacial acetic acid,
Digitized by
Google
FORMATION OP DIAZOAMINES AND AMINO AZO- COMPOUNDS. 93
the diacetyl compound is hydrolysed with concentrated hydrochloric
acid and the resulting diaminozylene hydrochloride crystallised from
water and then decomposed with the calculated amount of concentrated
potassium hydroxide solution.
The 4 : 6-diamino-m-zylene separates in the solid form and is finally
purified by crystallisation from water; it then melts at 104^, the
melting point being identical with that given by Grevingk. The
isomeric 2 : 4-diamino-m-zylene separates as an oil ; this, after separa-
tion from the potassium chloride solution by means of ether, is dis-
tilled under diminished pressure. After three distillations, a viscid
oil is obtained boiling at 170 — 174^ under 13 mm. pressure ; this sub-
stance, when cooled to — 10% gradually solidifies to a mass of crystals
melting indefinitely at about 17°. The yield from 1250 grams of
xylene is about 26 grams.
These bases have been characterised by means of their acyl deriva-
tives, as these latter are well defined substances easily prepared by
the ordinary processes.
Di/ormylA : Q-diamino-m-xylene crystallises from water in colour-
less, slender, flattened, silky needles and melts at 182 — 183°.
0-21 22 gave 27-1 c.c. moist nitrogen at 18° and 758 mm. N = 14-71.
CiQHigOjNg requires N = 14*58 per cent.
The diacetyl derivative is very sparingly soluble in alcohol, but
dissolves more readily in glacial acetic acid ; it crystallises in lustrous,
silky needles and melts above 260°.
Tlie dibenzayl derivative, Of^'S^'M.e2(i^^'CO•GQTl^)2, crystallises from
alcohol or ethyl acetate in smaU, lustrous plates and melts at
252—253°.
0*1570 gave 11 '5 c.a moist nitrogen at 18° and 758 mm. N»8*43.
CggHj^OjjNg requires N = 8*13 per cent.
IHformyl-2 : ^^dicmino-m-xylene crystallises from water, alcohol, or
ethyl acetate in rosettes of colourless needles and melts at 219 — 220°.
0*1148 gave 14*4 c.c. moist nitrogen at 19° and 759 mm. N = 14*29.
CjqHijOjNj requires N« 14*58 per cent.
The diacetyl derivative is sparingly soluble in alcohol and crys-
tallises from glacial acetic acid in colourless, felted needles ; it melts
above 260^.
The dibenzoyl derivative crystallises from alcohol in felted needles
and melts at 232°
0-1458 gave 10*8 c.c. moist nitrogen at 18° and 769 mm. N=« 8-66.
CgjHj^OjNj requires N = 8-13 per cent.
Digitized by
Google
94 morgan: influence of substitution on the
Action of Benzenediazonium Chloride on 2 : 4:'Diamino-m-xylene,
— A solution of benzenediazonium chloride, prepared from 5 '4 grams
of aniline hydrochloride, is added to a dilute ice-cold solution of 5
grams of the diamine acidified with 3 c.c. of concentrated hydrochloric
acid. The solution remains clear until excess of crystallised sodium
acetate (10 grams) is added and then a gelatinous red precipitate is
produced which after remaining for a few hours is collected. The
substance is purified by crystallisation from methyl alcohol.
Benzm&^'az0'2 : i^iamino-m-xylene crystallises in yellowish-brown
needles and melts at 208— 209"^.
00308 gave 6*2 c.c. moist nitrogen at 18° and 769 mm. N = 23-54.
Cj^HigN^ requires N = 23*33 per cent.
The azo-compound is distinctly basic and dissolves in dilute hydro-
chloric acid ; the hydrochloride, however, is amorphous and separates
in masses of red filaments. The platinichloride is a brick-red, amor-
phous, insoluble salt.
Benzene'5'az0'2 : i-dicuxtyldiamino-m-xtflenef
CeH5-N2-OflHMe2(NH-CO-CHj)j,
prepared by warming the crude azo-base for a few minutes with a
mixture of glacial acetic acid and acetic anhydride, crystallises from
alcohol in orange plates melting above 260°.
01952 gave 28-8 c.c. moist nitrogen at 19° and 769 mm. N = 17*16.
CigHggOgN^ requires N=* 17*28 per cent.
A comparative experiment made with 2 : 4-tolylenediamine shows
that the two bases behave in a precisely similar manner towards
diazonium salts.
Beiizene-5-azo-2 : 4-tolylenediamine (compare Stebbins, Ber., 1880,
13y 717) crystallises in orange-brown needles or leaflets and melts at
161°.
Benzene-6-€tzo-2 : 4:-diacelyliolylenediainine,
CJ6H5-N3-C«H2Me(NH-CO-CH3)„
crystallises in flattened, orange prisms and melts at 216 — 217°.
0*1756 gave 27*4 c.c. moist nitrogen at 20° and 769 mm. N= 1806.
CjyHigOgN^ requires N = 18*06 per cent.
Action of Dicizonivm Salts on 4 : ^-Diamino-m^cylene, — ^The same
proportions of diamine and benzenediazonium chloride are employed
as in the preceding experiment. On adding the sodium acetate to
the clear solution containing the other reagents, a bulky, brownish-
yellow precipitate is formed which rapidly darkens and becomes
resinous. After 2 hours, the product is collected, washed, dried, and
extracted with alcohol. From this extract, henzene-bazo-Ai : MLxamino-m-
Digitized by
Google
FORMATION OF DIAZOAMINES AND AMINOAZO-COMPOUNDS. 95
xylene, G^'K^'N^'OQWNLe^{KEL2)2f separates in deep red, rhombio plates
which after two crystallisations melt at 182^183^. The compouad
is not decomposed on long boiling with alcoholic hydrochloric acid.
It develops a deep orange coloration with concentrated sulphuric acid,
and readily yields acyl derivatives by the ordinary processes.
0-0832 gave 0-2U0 CO^ and 0-0503 Kfi. C =- 7014 ; H = 6-70.
0-1030 ,, 20-3 C.C. moist nitrogen at 18° and 767 mm. N = 22-98.
Cj^Hi^N^ requires C = 700 ; H = 6*66 ; N « 23*30 per cent.
The dtacetf/l derivative, CjH5-N2-CgHMe2(NH-CO'CH3)2, obtained
from the preceding base by the action of acetic anhydride, crystallises
from alcohol in brownish-red needles and melts above 260°.
0-1390 gave 20*8 c.c. moist nitrogen at 18° and 759 mm. N = 1710.
^i8^'^^2^4 '^©qw'^J^GS N = 17*28 per cent.
^Toliiene-5-azo-4: : Q'diaminO'm'Xylene, C^^^M.e''N^'GQH.M.e2(NH^2*
obtained by substituting an equivalent amount of ^-toluidine for the
aniline employed in the preparation of the preceding azodiamine,
crystallises from alcohol or benzene in deep red, rhombic plates and
melts at 165 — 166°. It closely resembles its homologue in chemical
and physical properties.
01817 gave 34-5 c.c. moist nitrogen at 21° and 754 mm. N = 21*43.
^15^18^4 req^"^^ N = 22 04 per cent.
Action of Diazonium Salts on b'CIdoro-2 : ^-tolylenediamine and
2'ChloroZ : 5iolylenediamine.
5-Chloro-2 : 4-tolylenediamine can be readily obtained in large quan-
tities by the author's method (Trans , 1900, 77, 1209), and it has been
further characterised by the prep^iration of a series of its diacyl
derivatives.
Di/ormyl'6'ehlar0'2 : Uolylenediamine, CftHgMeOKNH- OHO)^, pre-
pared by heating the base for 3 hours with 2 — 3 parts of concentrated
formic acid, is obtained as a dark brown precipitate on treating the
product with dilute ammonia ; it is purified by three crystallisations
from water in the presence of animal charcoal, and finally separates
from this solvent in colourless, silky needles melting at 166°.
0-1218 gave 02370 CO2 and 00520 HgO. 0 = 51-72 ; H = 4-74.
0-2076 „ 24 C.C. moist nitrogen at 18° and 768 mm. N= 13-49.
01318 „ 00869 AgCk 01 = 16-31.
OaHoGNaCl requiresO = 5082 ; H = 423 ; N = 1318; 01 = 1670 per cent.
The diacetyl derivative melts above 260°, and not at 240° as pre-
viously indicated ; it is obtained free from the monoacetyl compound
Digitized by
Google
96 MORGAN: INFLUENCE OF SUBSTITUTION ON THE
by heating the base with excess of acetic anhydride, and crystallising
the product from glacial acetic acid ; it is sparingly soluble in methyl,
ethyl, or amyl alcohol, separating ^from any of these solvents as a
microcrystalline powder ; it crystallises from pyridine or acetic acid in
small prisms.
01405 gave 0-0842 AgCl. CI = 14'83,
C10H13ON2CI requires 01 = 14-76 per cent.
The dihenzoyl derivative, prepared by the Schotten-Baumann method,
crystallises from alcohol in colourless, acicular lamellsB and melts
at 205°.
0-2209 gave 0-0890 AgCl. CI = 9*96.
01576 „ 10-5 c.c. moist nitrogen at 19° and 7^1 mm. N =« 7-70.
CaiHj^ONjCl requires CI = 9*74 ; N = 7*68 per cent.
Benzen$-Z-aaO'^'chUyi*o-2 : i-tdyUnediamine, CgH^'Ng' CgHClMe(NH2)2.
— ^The brownish-yellow precipitate produced by adding an excess of
sodium acetate to a dilute hydrochloric acid solution containing equi-
valent quantities of benzenediazonium chloride and 5-chloro-2 : 4-tolyl-
enediamine, is allowed to remain in contact with the mother liquor for
2 hours and then collected, dried, and extracted with alcohol. The
filtered extract slowly deposits a crop of dark brown crystals con-
taminated with tar ; the crude benzene-3-az(h5'ehlorO'2 : i-iolyknedt-
amine is repeatedly crystallised from alcohol and finally obtained in
dark brownish-red plates melting at 147°. The compound is not de-
composed by prolonged boiling with alcoholic hydrochloric acid and
develops a deep yellowish-brown coloration with conceutrated sulphuric
acid ; it readily yields acyl derivatives when treated with the appro-
priate reagents.
The dihmzoyl derivative, CgHg'Ng-CeHMeCKNH'CO-CeHB)^, pro-
duced by the Schotten-Baumann method, crystallises from alcohol in
transparent, brownish-yellow plates and melts at 236 — 237°.
0-1692 gave 17-3 c.c. moist nitrogen at 19° and 765 mm. N= 11-83.
C^HjiOgN^Cl requires N = 11*95 per cent.
The acetyl derivative crystallises from alcohol in brownish-red,
flattened prisms, and melts at 225°.
i^T6luene''Z'azO'^'MorO'2 : AAolylenediamine^
C«H^Me-Nj-OeH01Me(NH2)2.
— This azo-com pound closely resembles its lower homologue and is pre-
pared in a precisely similar manner. In this case also there a considerable
evolution of nitrogen accompanied by the formation of much tarry
product, and the yield of crystalline base is small. The substance
crystallises from alcohol in dark brown plates and melts at 152°
Digitized by
Google
FORMATION OP DIAZOAMINES AND AMINO AZO-COMPOUNDS. 97
01196 gave 00600 AgCl. CI = 12-41.
0-1971 „ 35-2 C.C. moist nitrogen at 21° and 754 mm. N = 20-16.
Ci^HigN^Cl requires CI = 12*93 ; N = 2000 per cent.
A dilute hydrochloric acid solution of 4 : 6-dichloro-m-phenylene-
diamine and benzenediazonium chloride, when treated with sodium
acetate, yields a yellow precipitate, which, when crystallised from
alcohol, separates in brown leaflets melting at 137° The product
contains from 37 to 39*6 per cent, of chlorine and seems to consist
chiefly of unchanged diamine ; this base melts at 137° and contains
40'11 per cent, of chlorine, whereas the percentage of this element in
the required azoK^ompound would be 25 '26. A negative result was also
obtained with />-toluenediazonium chloride; in this experiment, 14*7
grams of the diamine were employed^ and 8*7 grams of the unchanged
base were recovered after recrystallisation, the other products of
reaction being tarry and indefinite.
On the other hand, the dichlorodiamine combines with diazotised
primulin, for a piece of cotton cloth impregnated with this diazo-
compound and placed in an aqueous solution of the base gradually
acquires a brownish-orange colour, similar in shade to the azo-colouring
matters produced under these conditions from 5-chloro-2 : 4-tolylene-
diamine and 4 : 6-diamino-m-zylene.
B0nzene'^-azo-2-ehloro-3 : d-iolylenediamme, — 2-Chloro-3 : 5-tolyIenedi-
amine(.B0r., 1892,25, 3005), is readily obtained from 3 : 5-dinitro-2-chloro-
tolnene, the nitration product of o-chlorotoluene, by the iron-filings
method of reduction. The iron oxide is freed from the diamine by
washing with alcohol ; the alcoholic and aqueous extracts are mixed
together, acidified with acetic acid, and treated with excess of acetic
anhydride. The acetyl derivative, crystallised from glacial acetic acid,
is hydrolysed with hydrochloric acid, and the free base purified by
crystallisation from water, in which solvent it is more soluble than its
symmetrical isomeride ; it separates in long needles melting at 74°.
The azo-compound, prepared in the ordinary way, separates im-
mediately as a flocculent, yellow precipitate on the addition of sodium
acetate to the hydrochloric acid solution of its generators ; it crystallises
from a mixture of benzene and petroleum in tufts of long, orange-red,
acicnlar prisms, these crystals being sometimes more than an inch in
length. The substance melts at 134°, yields a deep brownish-red
coloration with concentrated sulphuric acid, and is not decomposed by
prolonged boiling with alcoholic hydrochloric add.
0-1448 gave 26*2 c.c. moist nitrogen at 18° and 766 mm. N « 21 00.
0-1643 „ 0*0928 AgCl. CI -13-97.
CigHijjN^Cl requires 01 = 13*62 ; N = 21*49 per cent.
VOL. LXXXI. H
Digitized by
Google
98 MORGAN: INFLUENCE OF SUBSTITUTION ON THE
Benzen6'6'azO'2'Moro-3 : d-diaceiyltolylensdiamitie,
CeH5-Nj-CeHMeCl(NH-COCH3)„
produced by gently heatiog the azo-base with acetic anhydride, is readily
soluble in acetic acid or benzene, but dissolves only sparingly in alcohol
or ethyl acetate ; it crystallises in silky, orange needles and melts
at 261°.
0*1074 gave 14'6 c.c. moist nitrogen at 17° and 762 mm. Ns 15*68.
0*1631 „ 0*0663 AgCl. CI = 10*06.
Ci^HiyOgN^Cl requires CI = 10*30 ; N= 16*26 per cent.
The dibenzoyl derivative obtained by the Schotten-Baumann method
crystallises from benzene in orange needles and melts at 233°.
0*1162 gave 11*7 o.c. moist nitrogen at 18° and 758 mm. N^ 11*69.
00717 „ 0*0213 AgCi. Cl = 7*36.
CjyHjiOjN^Cl requires CI = 7*58; N« 11*95 per cent.
Action of Diazanium Salts on l-Ghloro-jS-naplUhylaminB.
2'Diazoamino-l'chhronaplUh€U&ney CioH^Cl-Nj'NH'CjoH^Cl, separates
as a light yellow precipitate on adding excess of sodium acetate to the
mixture formed by slowly dropping a glacial acetic acid solution of
1-chIoro-^-naphthylamine (1 mol.) into a hydrochloric acid solution of
l-chloro-2-naphthalenediazonium chloride (1 mol.) ; it crystallises from
benzene or chloroform in golden-yellow needles and melts at 162°.
The diazoamine may also be produced by adding sodium nitrite (1
mol). to an iCe-cold alcoholic solution of 1-chloro-^-naphthylamine
(2 mols.) acidified with hydrochloric acid, the precipitation ^of the
compound being completed by the addition of a saturated aqueous
solution of sodium acetate. The product obtained in this way is,
however, often contaminated with an amorphous, red substance, which
is not readily removed in the subsequent crystallisations.
0-1711 gave 17*1 c.c. moist nitrogen at 21 *6° and 769 mm. N = 1 1 *49.
0*2368 „ 01817 AgCl. CI = 19*19.
CjoHjgNgClj requires CI = 19*39 ; N= 11*47 per cent.
Although insoluble in alcohol, it readily dissolves in alcoholic
potassium hydroxide, yielding a deep orange-coloured solution, this
result pointing to the existence of a potassium derivative. The com-
pound is remarkably sensitive to light, and after a few weeks' exposure
its crystals, although retaining their shape, acquire a dark chocolate
colour ; it is decomposed on warming with hydrochloric acid, evolving
nitrogen and yielding l-chloro-/9-naphthyliuiine and resinous pro-
ducts.
Digitized by
Google
FORMATION OF DIAZOAMINES AND AMINO AZO -COMPOUNDS. 99
p-IRirobenzen6'2^iazo<KninO' 1 -chloranaplithalenef
NO,-CeH,-N2-NH-CioHeCl,
is obtained by adding a hydrochloric acid solution of />-nitrodiazonium
chloride (prepared from 3 grams of JE^nitroaniline) to a cooled alcoholic
solution of l-chloro-j3-naphthylamine (4 grams), the precipitation of
the diazoamine being completed by the addition of sodium acetate. It
may also be prepared by mixing together solutions containing equivalent
quantities of l-chloro-2-naphthalenediazonium chloride and JE^nitro-
aniline. The diazoamine produced by either of these processes is
obtained as a voluminous, yellow precipitate ; it is almost insoluble in
alcohol and only sparingly soluble in benzene, separating from the
latter solvent in brownish-yellow leaflets melting and decomposing at
197—198°
01342 gave 19*8 c.c. moist nitrogen at 21° and 759 mm. N» 1708.
0-1614 „ 0-0692 AgCl. 01 = 10-65.
Ci^HiiOgN^Cl requires 01 = 1087; N= 17-15 per cent.
The diazoamine is fairly soluble in hot chloroform, but when boiled for
some time with this solvent it partly decomposes. It is acidic in
character and its potassium derivative, produced by dissolving the
compound in an alcoholic solution of potassium hydroxide, yields a
a deep purple solution.
Ethyl Derivative qf ]p-Nitrohenzene'2'di€iZoamvnO'hehloronaphthalenef
NOj'CijH^'Nj'NEt'OiQHjOl. — An alcoholic solution of the potassium
derivative of the preceding diazoamine is boiled with a slight excess of
ethyl iodide until the deep purple coloration of the mixture changes to
orange. The crystalline product obtained on cooling the alcoholic
solution is purified by crystallisation from benzene, and separates
from this solvent in hard, orange-yellow, prismatic crystals melting at
193*— 194°. This compound does not develop a purple coloration with
alcoholic potassium hydroxide and on analysis gives numbers corre-
sponding with those required for an Myl derivative of the mixed
diazoamine.
0*2382 gave 32*1 c.c. moist nitrogen at 19° and 769 mm. N = 15*68.
0-1452 „ 0-0684 AgOl. 01=9-96.
OigHigOgN^Ol requires 01 = 10-01 ; N = 15-80 per cent.
A diazoamine resembling the preceding ethyl compound, but melting
at 182 — 1 83°, is produced by adding a solution of l-chloro-2-naphthalene-
diazonium chloride to an alcoholic solution of ethyl-/>-nitroaniline. The
study of these alkyl derivatives of naphthalenoid diazoamines is,
however, still incomplete owing to the difficulty experienced in
alkyUting l-ohloro-j3-naphthylamine and its analogues.
H 2
Digitized by VjOOQIC
100 MOIR: CYANOHYDROXYPYRIDINE DERIVATIVES
My best thanks are due to Miss F. M. G. Micklethwait for assistiDg
in the preparation and analysis of several of the compounds described
in this communication, and to Mr. E. Lodge for examining the
tinctorial properties of the two series of arainoazo-derivatives.
Royal Collegr of Scibnck, London.
South Kensington, S.W.
XII. — Cyanohydroxypyridine Derivatives from Dtaceto-
nitrile. New Derivatives of y^Lutidostyril.
By James Moir, M.A.» B.Sc., 1851 Exhibition Scholar of Aberdeen
University.
DiACETONiTBiLE was first prepared in 1889 by R. Holtzwart, in E. von
Meyer's laboratory, by the action of sodium on acetonitrile in the
presence of a diluent such as ether, which serves to keep the tempera-
ture below that required to form the termolecular polymeride cyan-
methine {J. pr. Clm%., 1889,39, [ii], 329).
While attempting to make the latter compound for another purpose,
I found that even if acetonitrile alone be used, diaoeUmitriU is almost
the sole organic product (instead of cyanmethine) if the sodium be
present in excess. During these experiments, in attempting to crystal-
lise the diacetonitrile from hot water, I noticed that ammonia was
evolved during the digestion of the solution on the water-bath, and
that subsequently a different substance crystallised from the liquid.
The formation of this substance, which is sparingly soluble in all
the ordinary solvents and beautifully [crystalline, had already been
observed by Holtzwart, who made an extensive study of diacetonitrile.
Although he analysed the compound and assigned to it the formula
CgHgONj, unfortunately he did not succeed in elucidating its constitu-
tion. The mechanism of the process by which it arises is, however,
not difficull to imagine, if it be remembered that, as Holtzwart has
shown, diacetonitrile has the constitution OH3*C(NH2)ICH*CN, and
that it is easily converted by hydrolysis into the isodynamic form of
cyanacetone, 0H3*C(0H)ICH*0N. If two molecules of the latter
compound lose one of water, a compound of the formula CgHgON, will
be produced. This formula is that of an anhydride of cyanacetone;
OHg-CICH-CN
Holtzwart therefore proposed to write the formula ]>0
OHj-CICH-CN
von Meyer subsequently suggested the alternative formula
CHg'C-CHg-CN
NC-8-00-CH3 '
Google
Digitized by ^
PROM DIACETONITRILB. 101
My experiments have led me to conclude that such formulsB afford
no adequate explanation of the extreme stability and inactivity of
the substance, and they seem to me to justify the conclusion that the
compound is in reality ^^ofno-^p-hUidostyrU :
CH,
^^f^^^S^A^^
or, more probably, a polymeride of that substance formed by a process
analogous to that by which benzonitrile is converted into cyanphenine.
Attempts to determine the molecular weight were frustrated by the
insolubility of the substance.
In preparing diacetonitrile and the compound under discussion, the
methods described by Holtzwart and by von Meyer were in the main
followed. As diluents of the acetonitrile, dry ether, beiizene, and
toluene were tried without much benefit. In all cases, the yield of
diacetonitrile leaves much to be desired. The best results are
obtained as follows. Forty grams of acetonitrile (distilled over phos-
phoric oxide or solid potash) having been covered with a layer of dry
light petroleum (to exclude air), 10 grams of sodium in thickish
slices are introduced gradually through the condenser. The action
is very violent until the surface of the metal becomes coated ;
finally, the flask is heated during four hours on the water-bath.
The mixture, having been transferred to a Buchner funnel, is thoroughly
stirred, to separate the product from the sodium, which can then be
mechanically removed. The solid — a mixture of sodium diacetonitrile
with sodium cyanide— is mixed with just enough water to dissolve it ;
diacetonitrile separates as an oil and may be completely recovered by
extracting with^benzene and then evaporating off the solvent.
To prepare Holtzwart's compound, the benzene extracts are digested
with about 20 parts of water : as the benzene evaporates, ammonia
18 given off and the liquid becomes brown ; eventually it deposits
needles of the condensation compound. The mother liquor, on digestion
with water, yields a farther quantity together with a red gum. The
product is obtained practically pure by one crystallisation from glacial
acetic acid. The loss caused by the formation of bye-products in this
double condensation is so great that the yield of the final product is
seldom over 8 per cent, of the acetonitrile used.
The substance so obtained agrees on the whole with Holtz wart's
description, forming bundles of small, short needles ; it has an intensely
bitter taste. It is equally soluble in boiling water and alcohol
to the extent of about 1 per cent. ; it is more soluble in boiling
glacial acetic acid, but in boiling benzene only to the extent of 1
Digitized by
Google
102 HOIB: CTANOHTDROXTPTRIDINE DERIVATIVKS
part in 600. It crystallkeB out easUy on cooling these solutions. I
have to add the correction that the purified substance melts
sharply at 293"^ (305'^ corr.), although it darkens somewhat above 280^.
It can be sublimed without much loss at a higher temperature.
Holtzwart states that his compound melted <*oberhalb 230°" — a
serious underestimate, as I have never observed a lower melting point
than 260°, even in the case of the crude preparation. Holtzwart's
formula was coufirmed by the following analysis :
01513 gave 035 96 CO, and 00762 H,0. 0-6482 ; H = 6-62.
00855 „ 13-7 c.c. moist nitrogen at 13-5° and 760 mm. N = 18-63.
OgHgONj requires 0 = 6479 ; H = 5-46 ; N = 18-96 per cent.
Despite the presence of two nitrogen atoms, the compound is not
basic and may be crystallised from aqueous acids ; it does not combine
with platinic chloride.
It is easily soluble in alkali hydroxides, metallic derivatives being
formed ; these can be isolated by adding excess of alkali, and crystallise
well from a mixture of absolute alcohol and ether, although very soluble
in spirit or acetone. The potassium derivative forms long, lustrous
needles ; the sodium derivative, short, opaque needles. That they are
pheuolic in character is shown by the fact that the addition of carbon
dioxide or of ammonium salts to their solutions causes a precipitate
of the original substance. Attempts were made to analyse these, but
the results were vitiated by the rapid absorption of carbon dioxide
during the drying ; the figures are too low in consequence.
0-2472 potassium derivative gave O'llOO K^SO^. K = 19-97.
OgH^ONjK requires K = 2101 per cent.
0-4923 sodium derivative gave 0-1804 Na^SO^. Na» 11*88.
0*1329 „ „ dried in a vacuum, gave 19*7 c.a moist
nitrogen at 13° and 746 mm. N = 17*2.
CgH^ONjNa requires N= 1647 ; Na = 13-54 per cent.
Holtzwart's compound is a substance of unusual stability, and is not
affected by
(1) Prolonged boiling with a 10 per cent, aqueous or alcoholic solu-
tion of sodium hydroxide.
(2) Prolonged boiling with methyl iodide and sodium hydroxide. .
(3) Prolonged heating at 120° with 70 per cent, sulphuric acid.
(4) Heating at 80° with fuming sulphuric acid.
(5) Prolonged boiling with acetic anhydride.
It had previously been shown by workers in von Meyer's laboratory
that it is not affected by acetyl chloride, hydroxylamine, nitrous
acid, &c. It gives no coloration with nitrososulphuric add. It is only
slightly attacked by boiling dilute nitric acid and by permanganate,
Digitized by
Google
FROM DIACETONITRILE. 103
aiul although it at onoe reduces a solution of chromium triozide in
acetic acid, nothing definite can be isolated. It is also scarcely affected by
boiling its solution in absolute alcohol with a large excess of sodium.
The first clue to the nature of the substance was obtained by heating
it with zinc dust ; a distillate smelling like pyridine was obtained, but
in too small a quantity for investigation.
The only attempt to hydrolyse the compound which has succeeded
was performed by heating it with concentrated hydrobromic acid
{d 1*47) in a sealed tube during 6 hours at 170°. A large yield of a
substance was obtained, which proved to be ^-lutidostyril, or 2 : 4-di-
methyl-6-hydrofypyridine, a substance first described by Hantzsch
(Ber.y 1884, 17, 2904), derivatives of which have frequently been ob-
tained by the interaction of ethyl acetoacetate or its derivatives and
ammonia {GazzeUa, 1886, 16, 449 ; Annalen, 1890, 259, 169 ; Trans.,
1895, 67, 220 ; 1897, 71, 299, Ac.).
It will be seen that the formula of i^-lutidostyri1, O7H9ON, may be
derived from that of the original substance, C3H3ON2, by displacing GN
by H, and that the latter may be regarded as a cyano-^-lutidostyril.
This was confirmed by the detection of carbon dioxide and ammonia
as bye-products of the interaction, which may be expressed as follows :
05HN(CHj),(OH)-CN + 2H30 + HBr = C6H^(OH8)2-OH + NH^Br +
CO^ The slightly charred contents of the tube were extracted with
water, filtered, and concentrated on the water-bath until the excess of
acid was removed. On redissolving in a little water and adding soda
until neutral, ammonia was freely evolved and the solution nearly
solidified owing to the separation of a mass of long needles. These
were filtered off and were found to be free from sodium and to melt at
171 — 173°. When heated in a test-tube, this product sublimed un-
changed, the sublimate melting at 176°, and after recrystallisation at
177—178° (179—180° corr.). It boiled at 303° (uncorr.). On adding
excess of sodium hydroxide to its concentrated solution, a sodium
derivative crystallised out in thin, lustrous plates. The substance is
therefore Hantzsch's ^-lutidostyril.
This was further established by directly comparing the product with
a specimen made by ColUe's method (Trans., 1897, 71, 299). On
bromination, it gave a product agreeing with Kerp's 3 : 5-dibromo-^-
lutidoBtyril, but melting and decomposing at 253° (corr.) (Kerp
gives 235°).
0*1058 gave 01405 AgBr. Br » 56-49.
CyHyONBr, requires Br =56-89 per cent.
OH3
This substance is therefore Brf NBr.
Digitized by
Google
104 moir: cyanohydroxypyridine derivatives
On nitratiog the ^lutidostyril, two compounds were obtained, one of
which was Collie's 5-nitro-deriyative melting at 254^ (corr.) ; the other
which crystallised in rosettes of short needles melting constantly at 196^
(corr.), also gave numbers on analysis agreeing with those required for a
mononitro-com pound, and was apparently a mixture of Collie's 5-nitro-
compound with the 3-nitro-compound (m. p. 260^ corr.), which I
have obtained in a different way (see p. 116). The sodium derivative of
the product melting at 196^ (corr.) was made, washed with ether, and
analysed :
00650 gave 00234 Na^SO^. Na= 11-68.
CyHyOgNjN* requires Na- 12
12' 12 per cent.
The free substance is therefore a nitro-^ -lutidostyril.
In the preliminary note (Proo., '1901, 17, 235), I described this
incorrectly as 3-nitrO'i^-lutidostyril itself. Both compounds give, on
reduction, the colour-reactions characteristic of 5-amino- ^-lutidostyril
(Collie, Trans., 1898, 73, 232).
If Holtz wart's compound be regarded as a cyano-^-lutidostyril,
it must be represented by one or other of the two following formulae :
CH3 CH3
Nc/\, (A) or 1^^0N (b),
CH I lOH ^ ' CKJ^ JOR ^ ^'
and, curiously enough, its formation from 2 mols. of '^tsocyan-
acetone " can be explained on either supposition, according as it is
assumed that either methyl or hydroxyl wanders in the process. The
follovring scheme will make this clear :
CH. 9H.
i^-g^ ^S>CH Ncc^Nqn
•g\N^
?■
!H,
a
NC-C-^ ^CH or NC-C^ N:JH
(A). (B.)
The compound represented by the formula (B) is already known,
and has been prepared in a manner which leaves no doubt as to its
constitution, namely, by condensing acetylacetone, ammonia, and
cyanacetic ester (that is, acetylaceconamine and cyanaoetamide),
Digitized by
Google
FKOM DIACETONITRILE. l05
V^s CH
i + ^^ -> NO-C<^ \CH
\NH|H NH^i ^N^ '
(Guareschi, ii«» R. Accad. Torino, 1892, 28, 330 ; 1898, 34, 27 ; C/iem.
Centr., 1893, ii, 648 ; 1899, i, 289).
As Guareschi's compound was stated to melt at 288 — 289°, whilst
mine melted at 293° (uncorr.), I found it necessary to prepare the
former substance for comparison with my product. The two com-
pounds exhibited a remarkably close resemblance, both physically and
chemically, and careful comparison was necessary to determine that
the two were in reality different ; indeed, it is only in their derivatives
that the difference is at all decided. Guareschi's compound forms
longer and more lustrous needles than mine, although possessing similar
sparing solubility in the usual solvents and the same alkaloidal bitter
taste. The melting point given in the literature is a corrected one '»
hence the difference between the isomerides in this respect is twelve or
thirteen degrees instead of four. [I found Guareschi's compound to
melt at 291° (corr.), whilst Holtzwart's melts at 305° (corr.). A
mixture of the two melts between 270° and 275°, but if this mixture
be recrystallised, the product is quite different in appearance from
either constituent, consisting of long, hair-like needles, which melt
at 236 — 242°.] The only other physical property in which the crystals
differ is their action on polarised light — Holtzwart's compound (m. p.
305°) causing a uniform extinction at about 50° to the axis, whilst
the crystals of Guareschi's isomeride (m. p. 291°) frequently produce
no effect, and when an extinction is observed it is confined to half the
breadth of the needles and is nearly parallel to their axis.
Chemically, too, Guareschi's compound resembles mine (1) in being
non-basic ; (2) in affording metallic derivatives (which are, however,
less soluble than those of my compound) ; (3) in giving ^-lutidostyril,
carbon dioxide, and ammonia when hydrolysed by fuming hydrobromic
add, the cyanogen group being directly displaced by hydrogen just as in
the case of the isomeride (p. 103) ; (4) in resisting the action of sodium
hydroxide, sulphuric acid, methyl iodide, &o.
This complete analogy between the two compounds leaves no doubt
that both are cyano-^lutidostyrils, and as Holtzwart's compound is
differmU from Guareschi's — which is 5-cyano-^-lutidostyril [formula
(B)] — it can only have the constitution represented by formula (A),
CH,
that is, it is Z-cyano-^p-luMoBtyril, q-^ | Iqu .
Digitized by
Google
106 MOIR: CYANOHYDROXYPYEIDINE DERIVATIVES
Such a compound should yield only mono-derivatives; this was
actually found to be the case.
BroniincUion qf Holtzwart's Compound, — A nearly saturated solution
of the substance in glacial acetic acid was mixed at 40° with a similar
solution of an amount of bromine just in excess of one molecular
proportion. Action soon set in, crystals separating from the solution.
The liquid was diluted to separate the part remaining in solution and
the product was digested first with a warm dilute solution of potass-
ium carbonate and then with a cold very dilute solution of sodium
hydroxide. The slight residue insoluble in alkalis was recrystallised
from boiling glacial acetic acid, from which it separated in minute
prisms, nearly insoluble in other solvents, melting at about 270° (280°
corr.), but decomposing. This substance contained 33*0 per cent, of
bromine. The amount obtained was very small and insufficient to
determine its nature.
On precipitating the alkaline solutions with acid, substances were
obtained which ultimately proved to be identical. The major product
was that extracted by sodium hydroxide ; this was purified by dis-
solving it in the least possible quantity of a solution of sodium hydr-
oxide and concentrating the liquid to the point of crystallisation.
Long, white needles of a sodium derivative were thus obtained, easily
soluble in water, and having a soapy feel. Before analysing this
substance, it was recrystallised.
0-2291 gave 01760 AgBr. Br = 32-69.
CgHgONjBrNa requires Br = 3208 per cent.
To separate the parent substance, the solution was precipitated with
acid ; the precipitate was well washed with boiling water, dried, and
analysed, as it could not be recrystallised. It consisted of minute,
white needles, which melted at 313° (327° corr.), but underwent
decomposition.
0-1929 gave 0-160 AgBr. Br « 35-3.
OgH^ONjBr requires Br = 35*21 per cent.
The amount of bromine found in the portion extracted by alkali
carbonate was 35*79 per cent.
There can be no doubt that the substance^ produced was the
CH3
o-bromo-compound^ P^t I \ow ^^^^^^^ '^ii}^ the j»-bromo-compound
(m. p. 261°) obtained by Guareschi.
Nitration qfH6liziD€vri*8 Gompouni, — This may be effected either with
fuming nitric acid and with a mixture of this acid with strong sul-
Digitized by
Google
FROM DIACETOMITRILE. 107
phuric aoid. No change occurs below 50°, and at a higher temperature
the action tends to be violent. To complete the nitration, the solution
was warmed on the water-bath during a few minutes, cooled, diluted
with ioe, and then supersaturated with sodium hydroxide. On stand-
ing, a sparingly soluble sodium derivative of the nitro-compound crys-
tallised out in orange rosettes. On recrystallisation, these formed
lo^gi yellow, lustrous needles, sparingly soluble in water, and quite
distinct, therefore, from the salt of Collie's 5-nitro-^-lutido8tyril-
carbozylic acid (Trana, 1898, 73, 234).
0-2591 gave 00865 Na^SO^. Na= 10-80.
OgHgOjNgNa requires Na = 10*71 per cent.
The colour is doubtless due to isomerisation to the quinonoid nitroate
CH ' ]*0 ' ^^ which the white, nearly insoluble, free
OH3
substance is the ^-acid, that is, J^f J^^^^ .
On acidifying the solution of the salt, the nitro-compound was pre-
cipitated as a nearly white mass of needles, which melted at about
240°, but decomposed. After several recrystallisations from boiling
water, it was obtained in long, opaque prisms which melted at 253°
(260° corr.).
As the product resembled Collie's nitro-acid, I determined nitrogen
in it ; although the nature of the substance prevented slow combus-
tion, the result shows that the cyanogen group is intact.
0-1985 gave 37-2 c.c. moist nitrogen at 85° and 753 mm. Na 22-59.
OgHyOjNj requires N-21-80 per cent.
The potassium salt of this substance closely resembles the sodium
salt, whereas the ammonium salt is deeper in shade, forming reddish-
brown prisms melting at 251° (corr.).
A further quantity of the nitro-compound was obtained by evapor-
ating the alkaline liquid, then acidifying, and extracting with alcohol.
No other product could be isolated.
An attempt to remove the cyanogen group with fuming hydro-
faromio acid led only to the destruction of the substance.
Hiiraiion qf Ouareaehi'a Compaundf CgHgON,. — This was carried out
as in the preceding experiment. The nitro-compound separates on
diluting the acid in pale green, lance-shaped crystals. These melt at
261 — ^263° and dissolve in a solution of potassium carbonate, forming
Digitized by
Google
108 . MOIR: CYANOH YDROXYPYRIDINE DERIVATIVES
an intensely yellow liquid, which, however, on evaporation, gives a
white solid. To remove traces of a coloured impurity, the solid was
washed with a little water ; the white potassium salt was then redis-
solved and the nitro-compound precipitated from the orange-ydlow solu-
tion by acid. After recrystallisation, it formed spear-like, oblique plates
melting at 263—264° (272° corr.). The sodium and ammonium salts
were also white in the solid state, but gave yellow solutions.
The colour phenomena manifested by the two isomeric nitro-deriva-
tives are obviously analogous to those given by o- and jthnitrophenol
respectively, to which they correspond in the relative arrangement of
the nitro- and hydrozy-groups.
On hydrolysing the nitro-compound by warming it with fuming
sulphuric acid at 100°, diluting, and boiling with a nitrite (Bouveault's
process), a new compound was obtained giving salts which were orange
in the solid state. The free substance melts at 282° (corr.). It-s
ammonium salt dissociates on drying.
The best direct evidence of the position of the cyanogen group in
Holtzwart's compound is afforded by the behaviour of the amino-
compound formed on reducing its nitro-derivative. A solution of this
substance gave very characteristic colour reactions, namely, (a) a cherry-
red colour on aerial oxidation in presence of ammonia ; (b) with ferric
chloride, a green colour, darkening to an intense indigo shade (very
sensitive). Precisely similar changes were observed by Collie to take
place in the case of his 5-amino-^4utidostyril and its carbozy-aoid
(Trans., 1898, 73, 232). There can therefore be little doubt that
Holtzwart's compound is, as previously argued, the nitrile of Collie's
acid.
To complete the series of reduction products, the nitro-derivative of
OH,
NO r iCN
Guareschi's compound — presumably pxr^f Jqa; — was boiled with zinc
and acid as before. The solution gave merely a ' dull brown shade
with ferric chloride, and on adding ammonia an intense blue fluor-
escence was developed, but no colour appeared in the liquid.
Much time was unsuccessfully devoted to attempts to establish a
direct connection between Holtzwart's compound and Collie's ^-lutido-
8tyril-3-carboxylic acid. The ester of this acid is obtained by condens-
ing ethyl /3-aminocrotonate under special conditions, an interaction in
every way analogous to mine (Trans., 1897, 71, 299) ; I am greatly
indebted to Dr. Collie for a specimen of this ester with which he
provided me when, at first, I had some difficulty in preparing it.
Attempts were made both to hydrolyse Holtzwart's compound to
Collie's acid, and also to transform the latter into the former. Although
Digitized by
Google
FROM DIACETONITRILE. 109
neither series gave positive results, the experiments are of interest as
exemplifying the stability of this class of compound. In the first
instance, a solution of the substance, in 80 per cent, alcohol, was
boiled during fifteen hours with potassium hydroxide in large excess.
The alcohol was then boiled off and a solution of ammonium carbonate
added; a copious crystallisation of the unchanged substance took
place. It was to be expected that if any carboxylic salt were formed
it would remain in solution ; but on acidifying the filtrate only a faint
turbidity was produced, and, as the expected acid (Trans., 1897, 71,
304) is practically insoluble in water, it may safely be asserted that no
hydrolysis whatever had occurred. This peculiar procedure was neces-
fdtated by the fact that both the expected acid and its nitrile have the
same melting point and general properties.
In addition to the methods already mentioned, heatiug with soda
under pressure and also fusion with potash were tried; both pro-
cesses, however, destroy Holtzwart's compound completely, although it
is attacked only at a high temperature. Again, the action of warm
fuming sulphuric acid (which hydrolyses Guareschi's isomeride) was
tried in vain, the substance being either unattacked, or sulphonated
to a minute extent.
The inverse experiments are of greater interest, as throwing light on
the probable cause of the resistance to hydrolysis of the nitrile group
in Holtzwart's compound ; for the same inertness is shown, in a lower
degree, by the carbethoxyl group in Collie's ester (m. p. 137°), and
this is doubtless the cause of the failure of my efforts to synthesise
the corresponding nitrile. In the first experiment, the ester was heated
with excess of strong ammonia during five hours at 155 — 160° ; prac-
tically no action occurred, the only new product being a very small quan-
tity of the ammonium salt of Collie's acid. This is very soluble in water.
No trace of an amide was observed. Similarly, the ester was quite
unaffected when heated with excess of zinc-chloride-ammonia. This
agent also did not act on the corresponding ethyl 6-chlorolutidinecarb-
ozylate obtained Ify Collie by the action of phosphorus pentachloride
on his ester (Trans., 1898, 73, 589).
In the remaining experiments, I started with the acid (melting at
300^corr.). In preparing it, time can be saved hj fusing the ester
with potash ; quite a high temperature is necessary, but the yield of
add is good, as it completely precipitated on acidifying the solution of
the product. The dry ammonium salt of the acid was first heated with
excess of phosphoric oxide at 300°, but on extraction with water, no
trace of Holtzwart's compound was left. On heating the ammonium
salt alone, it decomposed smoothly at its melting point (about 270°)
into ^-lutidostyril, carbon dioxide, and ammonia. In a final experiment,
the acid was heated with 2 mols.. of phosphorus pentachloride, and after
Digitized by
Google
110 MOIR: CTANOHYDROXYPTRIDINE DERIVATIVES
removing the ozyohloride the residue was heated with excess of solid
ammonium carbonate. On working up the product, a small quantity
of sparingly soluble needles was separated ; these, however, contained
chlorine and were not investigated.
These experiments exemplify the '' protective influence " of the two
o-methyl groups on every group which becomes imprisoned between
them in the ring. Several cases in which this kind of protection is
Br
observed in benzene compounds, for example, (^ ^CN , have been
,^Br
OH,
investigated by Sudborough and others. The cyanoxylene, \ ^|0N
(Noyes, Atmt. Chem. 7., 1898, 20, 792), is a particularly dose analogue
of Holtzwart's compound.
There remain to be mentioned two points in which my experience
has differed from Holtzwart's; the first has reference to the bye-
products formed in preparing the substance OgHgON, from diaceto-
nitrile, and the second to the action of phosphorus pentachloride on
this compound. By treating the distillate obtained in preparing his
compound with phenylhydrazine acetate, Holtzwart claims to have
obtained cyanacetonephenylhydrazone. I was unable to confirm this
observation, but as the liquid in the flask gives the hydrazone copiously,
it is possible that in Holtzwart's case some of this liquid may have
come over mechanically with the ammonia. In any case, the litera-
ture on cyanacetone is in a state of confusion, there being no less than
four claimants for the name. Of these, (1) that described by Glutz
(/. pr, Chem.y 1870, [ii], 1, 141) seems to be crude ^-lutidostyril ;
(2) Bender's sparingly soluble, beautifully crystalline compound, may
be Holtzwart's C^UfiN^ {Ber., 1871, 4, 518), whilst the oils and syrups
obtained by Matthews and Hodgkinson (Ber., 1882, 16, 2679)^ and by
James {Annalen, 1885, 231, 245), seem to be polym^rides of the true
cyanacetone of Holtzwart, a substance which, however, seems to have
but a momentary existence.
As to the action of phosphorus pentachloride on Holtzwart's com-
pound, the author states {loc, cit., 329) that the product is gummy,
but that he isolated from it a substance melting at 175° and giving
figures agreeing with those required for the formula OgH^N, [which
Beilstein enters wrongly as O^H^N, (ffandbuch, 3, 1455)].
In an experiment with a pure preparation of the substance, I found
it very difficult to cause any action to take place, but finally obtained
a small quantity of glistening needles melting at 165 — 166°, but am-
taining chlorine not removable by alkaUa, This substance is probably
Digitized by
Google
PROM DIACETONITRILE. Ill
CH,
»
the correspondiDg 2-chlorolutidine derivative, NOi^ > ^ op
OgH^NgCl, but the quantity obtained did not permit of an analysis
being made, I tried to synthesise it by the Sandmeyer method from the
corresponding amino-compound (see next part), but only obtained
Holtzwart's compound instead ; such abnormalities in the behaviour of
2-aminopyridines have been frequently observed.
It is evident from Holtzwart's description of this experiment that
he must have used a crude material, and I think that his compound
GgH^Nj owes its formation to some impurity. I found, for example,
that on boiling the crude compound with acetic anhydride, a small
quantity of a new compound crystallising in plates melting at 155^ was
obtainable, whereas the pure substance gave no trace of this product.
XL Th$ nonrexiat&nce of von Meyer' a " l8<ymeric CgHgONj."
By acting on diacetonitrile in ethereal solution with acetyl chloride
and then adding water, Holtzwart obtained a base of the formula
OgH^g, melting at 222° {J. pr. Chem., 1889, [ii], 39, 236). The
same compound was obtained by several other workers in von
Meyer's laboratory by acting on diacetonitrile with a variety of
reagents, such as ethyl chlorocarbonate, ethylene dibromide, alcoholic
hydrogen chloride, <fec., all of which act merely by removing ammonia
from two mols. of diacetonitrile and inducing condensation; thus,
I have found that the best yield of this compound is obtained by
simply heating diacetonitrile with zinc-chloride-ammonia until the
mass solidifies ; on dissolving in acid and supersaturating with soda,
the new compound is precipitated and may be filtered off.
By acting on this substance with nitrous acid, von Meyer obtained
a product of the formula CgHgON^, which may evidently be regarded
as the corresponding hydroxy-compound ; thus, CgH7N2*NH2 +
HNOj^OgHyNj-OH + Nj + HjO {J. pr. Chem., 1895, [ii], 62, 89).
This compound is described by von Meyer as melting at about 260°>
and he pronounced it to be different from the compound of the same
formula made by Holtzwart in his laboratory in 1889, the evidence
for this statement being the apparent difference in their melting
paints and certain differences in solubility.
I have repeated this work, and find that the two compounds are in
reality identicaL The solution [of the compound CgH^Ng in dilute
sulphuric acid was treated with a slight excess of nitrite and digested
for some time at 30 — 40% as it diazotises with some difficulty. On
Digitized by
Google
112 MOIR: CYANOHYDROXYPYRIDINE DERIVATIVES
boiling the solution, nitrogen was evc^ved ; the compound CgHgONj,
being non-basic, crystallised out on cooling, and after one crystallisa-
tion from water, melted at 278 — 282^ ; on recrystallising, the melting
point was raised to 291 — 292°, and under the microscope the crystals
were indistinguishable from those of Holtzwart's compound. The
melting point was not depressed by mixing the two.
To confirm this result, the product was nitrated by the method
described on page 107, and gave the golden needles of the sodium
* salt ' of 5-nitro-3-cyano-^-lutidostyril there described. On reduction
with zinc dust and sulphuric acid, the two colour reactions with
anmionia and with ferric chloride were obtained. In all these
particulars, von Meyer's product agrees with Hoi tz wart's compound
and no doubt can remain as to their identity. It is curious that von
Meyer, having both substances at his disposal, should have been led to
consider them different ; yet it is evident, judging from their melting
points, that his specimens must have been very impure, and hence
misleading data as to solubility, &c,, were given by them.
Von Meyer's "isomeric CgHgONg" (Beilstein, Handbuchy 4, 1161)
is thus 3-cyano-^lutido8tyril, and as it is obtained by the diazo-reaction
from the compound CgHgNg, the latter must be 3-cyano-6'amino-2 : 4-lut-
idine and its formation by the direct condensation of diacetonitrile
may be expressed as follows :
9H3 JH
^P ?'^\. -^ nc-c/Nh
CHj-C + qH
3- Cyan 0-2 : 4-dimethyl-6-ami]]opyridine,
CgHeN, (m. p. 222**).
From these data, probable constitutions can be assigned to the
obscure compounds obtained by von Meyer's students from diaceto-
nitrile with various agents. Thus, the compound OgHj^ON^ (m. p.
145^), from cyanamide, which on boiling loses carbon dioxide and
CH3
ammom'a, leaving Holtzwari's CgHgONj, must be /i^T iNH-CO-NfT
and one of the compounds CgHjoN^, from hydrazine, must be
CH,
NCj
N
CH,' JNH-NH,
Digitized by
Google
FROM DIACETONITRILB, 113
III. \p-Luiido8tyril-b'C<vi*h(}xylic Acid and some of its Derivatives*
As already mentioned, every attempt to hydrolyse Holtzwart's
CgHgONj (3-cyano-^-lutido8tyril) to the corresponding acid has failed.
On the other hand, I have succeeded in obtaining from Guareschi's
isomeride (5-cyano-i/r-lutidostyril) the corresponding amide and acid.
I may, however, first describe a number of experiments instituted to
ascertain the mechanism of Guareschi's interaction, which is character-
ised by the ease with which it takes place without a condensing agent.
The interacting substances are ethyl cyanacetate* and /3-diketones, in
the presence of a primary amine, and the reaction has been realised by
its discoverer in a large number of cases {Atti R. Accad, Torino^ 1893,
28, 330, 836 ; 1898, 34, 24 ; see also 1900, 86, 645). Of these, the
simplest is that leading to the compound CgHgON, (m. p. 289°
oorr.) from aoetylacetone, ethyl cyanacetate," and ammonia ; but
since the first two substances are both acted on by ammonia, forming
respectively acetylaoetonamine, CH3»C(NH2)ICH-CO-CHg, and cyan-
acetamide, NC'CH^'CO'NHg, these must be considered the true inter-
acting compounds. I found, in fact, that when the ammonia acts
beforehand on only one of the substances, the condensation does not
occur ; that is, mixtures respectively of acetylacetone with cyanacet-
amide, or of acetylaoetonamine with ethyl cyanacetate, do not con-
dense; whereas, if acetylaoetonamine and cyanacetamide are pre-
viously prepared free from ammonia, then the condensation occurs on
mixing their aqueous solutions and gently warming.
Now there are two possible explanations of this interaction,
CHa-CiO HJC-CN CHg-CiO ELiN
/ ' ' \ / ' \
CH + 00 or OH + C-OH,
% ^ % ^
OHj-C-jNH^ B^^^ CHj-O-jNH^ HiC-ON
of which only the former is a " methylene condensation." To decide
between them, the experiment of heating acetylaoetonamine with cyan-
acetmethylamide was performed. The sole product was the N-methyl-
derivative of Guareschi's compound, OgHgON^, and it appear sto me that
its formation is not explicable by the second of the two schemes, as in
this case there is no amino-group free from which water can be formed
with the adjacent carboxyl group. This condensation, therefore, occurs
as follows :
CHg-CO H^iO-ON
CH3
CH + bo — > ohI .J:o
»\v/
VOL. LZXXI.
CH,
I
Google
Digitized by
114 MOIR: CYAV0HT1>R0XYPYRIDINE DERIVATIVES
On the other hand, when the methyl group is introduced into the
other constituent of the reaction^ that is^ when acetylacetone-methyl-
amine is heated with cyanacetamide, the sole product is Guareschi's
compound CgHgON^, and not its N-methyl derivative. In this case,
methylamine, and not ammonia, is eliminated, and in both cases the
amine originally attached to the acetylacetone is the one which*[i8
expelled when the condensation takes place, the present reaction being
expressed as follows :
!H + CO -^ /\C
% / CH,I O^
L-O-iNHCH, HINH ^^^
Another experiment had the object of ascertaining whether the acid-
ifying influence of the cyanogen group is the determining factor in such
condensations, and this was found to be the case, for when malon amide
was substituted for cyanacetamide, no condensation with acetyl-
acetonamine could be induced, although the only variation is the sub-
stitution for the active CN group of the CO'NH^ group.
The methods of hydrolysis which proved successful with Guareschi's
compound were: (1) fusion with potash; (2) treatment with warm
fuming sulphuric acid. As both processes gave the same products, I
shall confine myself to the latter one, which gives a good yield.
If the solutioaof Guareschi's compound in the acid (10 percent. SO3)
be diluted after standing for some time at the ordinary temperature,
only unchanged substance separates; if, however, the solution has
been warmed at 100° for a short time, nothing separates on dilution,
but after several days a copious crystallisation of rosettes of needles
is obtained. These are sparingly soluble in water, melt at 209°
CH3
,CO-NHg.
(215° corr.), and consist of the sulphate of the amide, pxr I \fya
they are not alEected by acetic anhydride, and when treated with
ammonia or boiled with solution of potassium carbonate give the
amide which melts at 290—221° (227° corr.), is quite easily soluble in
water, and appears to be dimorphous, forming at first hard granules,
which on recrystallisation give small, flat needles with square ends.
Like the other substances of this class, it is easily soluble in caustic
alkalis, forming a phenolic * salt' crystallising in plates ; even on boiling
with potassium hydroxide, hydrolysis of the amide to the acid is slow,
as is also shown by its occurrence in the potash fusion. Unlike the
original product, the amide acetylates readily, and, curiously enough.
Digitized by
Google
FROM DIACETONITRILE. 115
the product after recrystallisation is so like Ouareschi's compound,
OgHgONg, that at first I thought it had been regenerated by the de-
hydrating action of the reagent. It forms long, white needles melting
witluHU darkening at 279—280° (290° corr.). That this substance is
different from the two compounds of the formula CgHgON^ was proved
by the method of mixed melting points and also by an analysis which
0-1075 gave 13'0 c.c. moist nitrogen at 18° and 756 mm. N » 13*89.
^10^18^8^8 'G^uiros N = 13'49 per cent.
It is remarkable that the amide should be so basic as to form stable
salts and an acetyl derivative, and for this reason I at one time thought
CHg
that it might be an amino-acid, namely, q^ \ J^t| » instead of
CH,
CH I JoH *' *^^ ^ decide this point tried a number of experi-
ments, such as the Monitrile test and the action^of nitrous acid followed
by alkaline /S-naphthol. The results were negative and the second
formula was then definitely proved by conversion of the substance,
by means of bromine and soda, into Collie's 5-amino-^-lutido8tyril,
CH,
CH 1 iOM ' ^^^^ gives extremely characteristic colour reactions
(see page 108, and Trans., 1898, 73, 232).
The substance (m. p. 227°) is therefore really the amide of ^-luUdo-
siyril'b'^uxrbaxylic acid. The next step was to obtain this acid. As
the hydrolysis of the amide is effected only slowly by acids or alkalis,
I tried the action of nitrous acid. On boiling the solution, nitrogen was
evolved and the carboxylio acid — which is very sparingly soluble —
was precipitated. This acid forms needles closely resembling its iso-
meride (Collie's i^-lutidostyril-3-carboxylic acid, m. p. 300 — 304°), but
melts at 244° (252° corr.), and, like its isomeride, decomposes into
^-lutidostyril and carbon dioxide when heated above its melting point.
Potassium Salt of tp-Lutidostyrilr^'Carhoxylic Aoidy m. p. 252°
(corr). — ^This was prepared by adding a solution of potassium carbonate
to the acid, evaporating to dryness, and crystallising from boiling
alcohol. It forms long, flat needles and was dried at 120°.
I 2
Digitized by VjOOQIC
116 CYANOHYDROXYPYRIDINE DERIVATIVES.
0-1564 gave 0-0668 K^SO^. K= 19-18.
CgHgOgNK requires K = 19*06 per cent.
3'Mtroil/ lutidostyril, — ^In preparing this compound, I followed Collie*8
description of the processes used in producing S-nitro-^-lutidostjril
(Trans., 1898, 73, 231 ). On nitrating the acid melting at 244"", I obtained
S-nUroil/lutido8ij/ril'5'Carb<xci/lic acid in the form of white, sparingly
soluble needles melting at 225 — 227^ (corr.), and giving intensely
orange salts. On reduction in acid solution, the solution gave the
same brown coloration with ferric chloride as its nitrile (the amino-
derivative of Guareschi's compound, see p. 108) gives.
On heating the above nitro-acid at 260° until the evolution of carbon
dioxide ceases, it is transformed into 3'nitro-\lf4utido8lyril^ which on
recrystallisation forms pale, shining leaflets moderately soluble in
water and melting at 260° (corr.), and on reduction gives a reddish-
brown coloration with ferric chloride. The analogy with Collie's
work in this field is brought out by the following scheme :
CHg CH3 CHg
Collie's acid, S-Nitro-^-lutidostyril- 6-Nitro-f-lntido-
m. p. 300 — 804* (corr.). S-carboxylic acid. styril.
OHj OH, CH,
/\C0,H N0,/NC0,H N0,/\
CH,kjj>rf CH.I^N/'orf CH3l^j^J0E
New acid, 8-Nitro-if^-lutidoRtyril- 3-Nitro-^latido-
m. p. 252° (corr.). 6-carboxylic acid. styril.
The Formation qfil/'Luiiidoatyril/rom Ethyl Acetoaoeiate,
Duisberg {AnnaUn, 1882, 213, 174), by heating ethyl acetoacetate
with excess of ammonia, evaporating, and heating the resulting gum
at 80°, obtained a compound decomposing at about 280° and event-
ually giving figures approximating to those required for the formula
CgHgON,.
Thinking that this might be Holtzwart's compound, I tried to obtain
it by heating ethyl acetoacetate with an equal bulk of strong ammonia
in a sealed tube during 2 hours at 150°. The product was an oil con-
taining crystals, but the latter were merely ammonium carbonate. On
evaporating the thick filtrate from these, a brown gum was left which
was kept on the water-bath for some time and then boiled with water
and excess of animal charcoal. On concentrating the pale filtrate, I
Digitized by
Google
THE DETERMINATION OF AVAILABLE PLANT FOOD IN SOILS. Il7
obtpained crystals which, after purification by repeated crystallisation
from water, melted at 173— 175° (177—179° corr.), and behaved in all
respects as ^fhhUidoatyrU,
This had evidently been formed from Modehydracetic acid, the first
stage in the condensation of ethyl acetoacetate.
CHj-(J-i6H EtjO^^^^^ CHg-C (:]H CHg-C (:|H
^H + (jJ-CHg-^ CH C-CH3-> CH (^-OHg
OO-jOEt HiO CO— O CO— N
Ethyl acetoacetate. -> "iwDehydracetic acid." -> if^-Lutidostyril.
Chemical Department,
Central Technical College, London, S.W.
XIII. — The Determination of Available Plant Food in
Soils hy the use of Weak Acid Solvents.
By Alfred Daniel Hall, and Fbangis Joseph Flymen.
In the analysis of soils, it has been customary of late years to employ
a weak acid solvent in order to extract those mineral constituents,
phosphoric acid and potash in particular, which are present in the soil
in such a state of combination as to be readily taken up by the crop.
The phosphoric acid and potash thus extracted have been termed the
'' available," as distinct from the total, amounts of the same substances
which can be extracted by hot, strong hydrochloric acid, or other
solvent, which completely breaks up the soil. It is claimed that better
indications of the comparative richness or poverty of the soil and of
the need or otherwise for special mineral manures can be obtained by
determinations of the available rather than of the total constituents,
the information supplied by the latter being often not in accord with
the results of cropping.
Although chemists are agreed generally about the value of weak
solvents in the analysis of soils, considerable diversity of opinion exists as
to the acid to use and the theoretical basis on which its action depends.
Dilute acetic acid, originally suggested by H. von Liebig, was used by
Deh^rain (Ann, Agran., 1891, 17, 445). An aqueous solution of carbon
dioxide has been worked with in America, by Gerlach (Landw.
Vermchs.'Stat., 1896, 46, 201) and by T. Schloesing (Campt rend.,
1900, 131, 149). Its adoption is obviously based upon the fact that
the natural soil water, by which much of the nutrient matter of
Digitized by
Google
118 HALL AND PLYMEN : THE DETERMINATION OF
the soil is conveyed to plants, largely owes it solvent power to carbonic
acid.
Petermann, in his examinations of Belgian soils (Recherckes de
Chimie et Phi/siologie, 1898, 3, 60), employs an ammoniacal solution
of ammonium citrate for the determination of available phosphoric
acid ; he regards it as " veritable reactif de groupe," distinguishing
between the mineral phosphate of lime and the precipitated phosphates
of lime, iron, and alumina which will rapidly come into action in
the soil.
Hydrochloric acid of various strengths has been used ; the American
Association of Official Agricultural Chemists has recommended a
solution of fifth-normal stiength; trials have also been made in
America with hundredth-normal hydrochloric acid.
Emmerling {Bied. GerUr,, 1900, 29, 75) has recommended a solution
of oxalic acid of 1 per cent, strength for the purpose of distinguishing
between phosphoric acid combined with the alkaline earths and that
combined with the sesquioxides.
Hoffmeister (Landw. Versuclis.-Stat, 1898, 50/363) suggests an
ammoniacal solution of humic acid for estimating the relative values
of different forms of phosphoric acid^ and Maxwell (J, Amer. Chem^
Soe.f 1899, 21, 415), in his examination of Hawaiian soils, used a
1 per cent, solution of aspartic acid, which was found to dissolve
"phosphoric acid, lime, potash, and other bases out of the soil in
almost the exact proportions that these elements have been found in
the waters of discharge and in which they are removed by cropping."
T. Schloesing, jun. (Campt. rend., 1899, 128, 1004), working with
dilute nitric acid of various strengths, found that as the strength of
the acid was increased, the amount of phosphoric acid dissolved first
increased, then remained stationary during a ceHain range, and then
began to increase again ; at which point, and not before, iron began
to appear in the solution. He concludes that this stationary pro-
portion of phosphoric acid indicates the amount of readily available
calcium phosphates and that the beginning of the attack upon the
ferric phosphate marks the point at which all the available phosphoric
acid has passed into solution.
But of all the dilute acids, none has been more widely applied to
the determination of "available" plant food than a 1 per cent,
solution of citric acid, as described by Dyer in a communication to this
Society (Trans., 1894, 66, 115), the 1 per cent, citric acid solution
being taken as approximating both to the nature and average strength
of the natural solvent, the root sap.
It is, however, doubtful if sufficient data exist upon which to base
any a priori decision as to the best acid and strength to employ ; the
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS. 119
state of oombination of the phosphoric acid and potash in the soil,
the nature of the root sap, and the part it plays in obtaining mineral
matter from the soil as compared with that which enters the plant
by osmosis from the natural soil water, are all too imperfectly known
to provide a theoretical basis for a method of analysis. In the present
state of our knowledge, these processes can only be tested by com-
paring the conclusions to which they lead with the results obtained
by cropping the soil ; indeed, the crop alone can measure the material
available in the soil.
It was in the hope of obtaining some critical results with regard to
the various acids suggested for determining the available constituents
in the soil that the authors have obtained a number of soils which
have been the subject of field experiments, and submitted them to
the action of certain of the acids indicated above. As a rule, abnormal
soils have been chosen, that is, soil which are markedly deficient in
available phosphoric acids or potash, as indicated by the large returns
which could be obtained by the application of one or other of these
substances in the shape of manure.
By the kindness of Sir J. Henry Gilbert, the authors further were
enabled to examine seven samples from the Broadbalk ^Field at
Bothflimsted, which had been under wheat and continually manured
in the same way for forty-two years. Sir Henry Gilbert was good
enough to furnish the authors with[material drawn from seven sharply
contrasted plots on this classic field, su£Bcient for duplicate* determina-
tions of both the phosphoric acid and potash dissolved by all the
solvents to be described later.
Determinations were made of both phosphoric acid and potash in
the Broadbalk soils and in four other cases ; the nine remaining soils
were only analysed for one constituent. Arising out of the work,
determinations were also made of the calcium carbonate and the
organic matter in each soil, and a few other determinations were
made to ascertain what degree of variation might be introduced by
the strength of the acid employed and the quantity of calcium car-
bonate present.
Tht SoUa Examined,
The soil samples from the Broadbalk Field, Bothamsted, were taken
in October, 1893 ; the plots had then grown wheat continuously for
Gfty years and the same manures had been applied to each plot year by
year, with one exception, for forty -two years {J, Roy, Agric, Soc. JEng.,
1884,20,391).
The following table shows the numbers under which the plots are
described in the Bothamsted Memoirs, the manures per acre per annum,
and the average yield of grain and straw :
Digitized by
Google
120 HALL AND PLTMBK : THE DBTKHMIHATIOH OF
Gnin. Stisw.
No. of plot. Mannre per acre per anniun. Ba^eU. Cwt.
2h Farmyard maniire, 14 tons 34| 32\
3 Uninantired continuotis] J 12| lOg
5 MineraUonly UJ 12^
0 Minerals + 200 pounds ammoniam salts 24 21i
7 Minerals + 400 pounds ammonium salts 32J 32}
ifa Minerals + 275 pounds sodium nitrate .. 34| 38|
1 6 Minerals + 800 pounds ammonium salts,
13 years — —
Unmanured, 19 years 27^ 28}
Minerals + 550 pounds sodium nitrate,
10 years — —
In the above table, '' minerals " stands for 200 pounds of potassium
sulphate, 100 pounds of sodium sulphate, 100 pounds of magnesium sul-
phate, and 3^ cwt. of superphosphate (37 per cent, soluble phosphate) ;
ammonium salts means equal parts of sulphate and chloride of ammon-
ium containing about 43 pounds of nitrogen, which is also that con-
tained in 275 pounds of sodium nitrate.
If the quantities given above are translated into pounds of phosphoric
acid and potash supplied and removed per acre per annum, the follow-
ing approximate figures are obtained. They are partially taken from
a recently published paper by Dyer on the phosphoric acid and potash
in wheat soils of Broadbalk Field, Rothamsted (Phil. Trans., 1901,
B. 194, 235 — 290), and are based on the manures supplied and the
analyses of the grain and straw removed :
Phospliaric (icid. Polash,
Plot. Supi)lied. Removed. Supplied. Kemoved.
26 78 26 235 50
3 0 9"3 0 15
5 65 14 104 23
6 64 17 108 33
7 62 22 107 51
9a 64 26 108 50
16 35 20 50 43
Of the other soils. No. 1 is a clay soil from Essex furnished by Mr. T. S.
Dymond. Home of the results obtained on this field in 1899 may be
(luoted as showing the response of the soil to dressings of phosphates :
Manure. Without lime. With lime.
Sodium nitrate, 2 cwt 33 82
„ ,, 4 cwt. superphosphate 17*8 25*4
Other results with phosphatic manures, both in this year and 1900,
Digitized by VjOOQIC
AVAILABLE PLANT POOD IN SOILS. 121
confirm the need for phosphates (see The Essex Field Experirneiits^
1901, I, 28).
Soil No. 2 is a Welsh soil from Cardigan, selected by Mr. T. Parry
as typical of the soils in that district which respond freely to dressings
of basic slag. The experimental plots in the same field showed
" astonishing results " for a dressing of 10 cwt. of basic slag, but,
being in pasture, no weights can be given.
Soils Nos. 3, 6, and 10, were indicated by experiments carried out under
the Bath and West of England Agricultural Society, in 1891, as likely to
be deficient in available phosphoric acid, and were kindly procured for us
by the occupiers, Mr. J. B. Till, of Park Farm, Thornbury, Gloucester-
shire ; Mr. B. W. Drew, of Crichel, Wimborne, and Mr. W. H. Tremaine,
of Trerice Manor, Grampoundj Road, Cornwall, from the fields which
had been under experiment. The following extract from the report on
the trials (/. Bath and West qf England Agric. Soc., 1891—1892,
[iv], 2, 264) shows the effect of phosphatic dressings on the mangold
crop :
4 cwt. niti-atc.
Plot Character of soil. With 4 cwt. uitrate. ,, superphosphate.
3 Gravelly loam 6 323
6 Deep loam on chalk 12*7 26
10 Stone rush 87 19-7
Soil No. 4 is from strong land on the Weald Clay, near Marden,
Kent ; the sample was taken from an arable field immediately adjoin-
a hop garden which has been under experiment since 1895 by the
South Eastern Agricultural College. The plots have always given
large returns for the application of phosphates, as will be seen from the
following table, giving the mean results 1895 — 1899 :
Mean of 5 years' crop,
Plot. Manure per acre per annum. cwt.
1 Nitrogen, potash, 6 cwt. phosphates 1 2*5
2 „ „ 8 „ 151
3 „ „ 10 „ 15-7
4 „ „ 15 „ 16-6
On the same soil, the omission of potash gave no consistent returns ;
on three occasions, the plot receiving nitrogen, phosphates, and potash
was superior by 9 per cent., 6 per cent., and 1 per cent, respectively;
on two occasions, it was inferior by 15 per cent, and 11 per cent. ;
hence we may fairly conclude that the soil can supply the potash re-
quirements of an ordinary crop (see J. Souih Eastern Agric. Coll., 1900,
No. 10, 33),
No. 5 is a sandy soil, resting on the Tunbridge Wells beds, near
Frant, and is also taken from a field adjoining a hop garden which has
been under experiment. In this case, phosphates above a certain
Digitized by
Google
122 HALL AND PLYMEN : THE DETERMINATION OP
point give little return, but potash salts produce a great increase in
the crop. The table sets out three years' results :
Plot. Manure per acre per annum. Mean crop.
1 Rape dust 15 cwt. ( = 70 lbs. nitrogen) + 0 15-2
2 „ „ + basic slag 5 cwt 15*1
3 „ „ „ 10 „ 16-1
4 M „ „ 15 „ 15-4
5 ,, „ „ 5 + potassium sulphate 5 cwt. 17*7
The only consistent increase in crop each year has been on the plot
receiving potash, where the effect has also been noticeable in the
character of the foliage (see J. SotUh Eastern Agric. CM. 9 loo, oii.).
Soil No. 7 was supplied by Mr. J. Alan Murray of the University
College, Aberystwyth, and was taken from grass land on a lights
alluvial loam at Falcondale, which has been under experiment for
8 years, and has given marked returns for dressings of phosphatic
manure. Taking the mean figures for 4 years, when phosphatic
manures were applied, the excess of hay produced as compared with
the plots receiving no phosphate was as follows (see Univ. Coll.
Aberyslwythy Annual Report on Field Experhnenta, 1900) :
For 112 lbs. superphosphate 336 lbs. per acre.
224 „ „ 518
336 „ „ 652
85 „ basic slag 364 ,,
170 „ „ 713
255 „ „ 777
Soils 8 and 9 were from the garden at Ham el's Park, Buntingford,
Essex, belonging to Mr. H. Shepherd Cross, M.P., a soil notable for
causing chlorosis in many species of plants grown there, especially in
laurels, fruit trees, and chrysanthemums. Applications of superphos-
phate had mitigated the onset of the disease, but it is by no means
certain that a deficiency in available phosphoric acid is the cause.
Soils 11 and 12 were from the experimental plots of the South
Eastern Agricultural College, at Wye ; the soil is a light loam resting on
the chalk and as a rule shows no particular need for mineral manures.
Soil 1 1 was from a plot which had for five consecutive years grown
barley without manure. Soil 12 had also grown barley, but had
received a general dressing of artificial manures, including 4 cwt. of
superphosphate containing 26 per cent, of soluble phosphate and
1^ cwt. of potassium sulphate.
The following mean figures obtained with barley, oats, and grass
in 1896 and 1897 serve to show the response the crop makes to
mineral manures ; the various crops are i-eduoed to a common standard
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS. 123
by calculating them on a basis of 100 for the plot with the complete
manure.
Plot. Manures per acre. Relative crop.
A. No manure , 73
B. Nitrogen + 2 cwt. superphosphate, no potash 93
D. Nitrogen + 1 cwt. potassium sulphate, no phosphoric acid ... 87
E. Nitrogen + 2 cwt. superphosphate, | cwt. potassium sulphate 100
Soil No. 13 was supplied to the authors by Mr. J. L. Duncan, B.Sc,
from his farm at Birgidale Knock, Rothesay, N.B. It is a deep, alluvial
loam, in good heart, but gave extraordinary returns for potash in some
experiments with turnips carried out by Professor J. Patrick Wright
in 1895.
Nitrogen and Nitrogeo, potash, +1 cvt.
Manure, nil. Phosphate only. phosphate. sulphate of potash.
Crop, nil. 5*9 8*9 198 tons.
(See Reports on Manuring, dsc, Glasgow aiid West qf Scotland
Technical College, 1896.)
27ie Dilute Acids Used.
Since the 1 per cent, solution of citric acid is so widely used, es-
pecially among chemists in this country, for the determination of
available phosphoric acid and potash, it was taken as the basis of
comparison, and the other acids, as far as possible, were reduced to
the same strength. This seemed preferable to using the other arbitrary
strengths which have been suggested, such as 1 per cent, acetic acid,
1 per cent, and one-fifth normal hydrochloric acid, especially as pre-
liminary experiments showed that the strength of the acid is a factor
in the amounts of phosphoric acid and potash dissolved. Citric acid
solution containing 10 grams of the pure crystallised acid per litre is
approximately one-seventh normal and is equivalent to a solution of
acetic acid containing 8*57 grams per litre and one of hydrochloric
add containing 5*2 grams per litre.
The ammonium citrate solution cannot be compared in strength with
the other solvents ; it is made up according to Petermann's formula,
and used for the estimation of phosphoric acid only : 1 litre contains
87*1 grams of ammonium citrate, rendered alkaline by 9*2 c.c. of strong
ammonia (sp. gr. 0*880) ; 500 c.c. are digested with 50 grams of the
soil for 1 hour at a temperature of 35 — 40% with constant shaking.
As a source of water charged with carbonic acid, recourse was had
to the " sparklet '' bottles of commerce ; one of the larger sized bottles
holds conveniently 50 grams of soil and 500 c.c. of water. Into this
a sparklet charged with liquid carbon dioxide was broken in the usual
Digitized by
Google
124 HALL AND PLTMEN ; THE DETERMINATION OP
way, the contents of the bottle were allowed to stand for a week and
shaken from time to time as with the other weak acids. The larger
sparklets were found by trial to contain about 4*5 grams of carbon
dioxide, so that the solution within the bottle would contain a little less
than 9 grams per litre, and be approximately 0*4 normal. After
opening the bottle, as soon as the first effervescence has subsided, the
solution must be rapidly filtered and the filtering completed before
all the free carbon dioxide has diffused out of the liquid.
When chalk is present in the soil, a strong solution of calcium bi-
carbonate is produced in the sparklet bottle, and precipitation df cal-
cium carbonate begins when the solution is brought into contact with
the atmosphere. Preliminary tests showed that solutions of acid cal-
cium phosphate and calcium bicarbonate can exist together until the
excess of carbon dioxide is expelled, when calcium phosphate is pre-
cipitated. However, the first portions of calcium carbonate precipi-
tated during filtering, although mixed with a little fine day, showed
no appreciable amount of phosphoric acid.
The three acids^ citric, acetic, and hydrochloric, of the same titre,
together with carbonic acid water, were used on the soils for the
estimation of both the phosphoric acid and potash.
Methods qf Analysis,
The soil samples were all air-dried, gently broken in a mortar with
a wooden pestle, and passed through a sieve having round holes
3 jnm. in diameter. The stones retained by the sieve were rejected,
the fine earth that passed the sieve was used for analysis without
any further preparation.
In the case of the soils from the Broadbalk Field, the samples
had already been put through a wire sieve with meshes ^ inch apart.
The 3 mm. round sieve took out a few more stones, amounting to
about 24 grams from each sample of 3 pounds, or, approximately, 1*8
per cent.
Except in the case of the ammonium citrate and the carbonic acid
solutions, 200 grams of the air-dried soil were put into a Winchester
quart bottle with two litres of the dilute acid, the bottle was kept
stoppered and shaken whenever convenient during 7 days at the ordi-
nary temperature of the room.
At the end of this period, the solution was filtered and an aliquot
part of the extract (generally 500 c.c.) was evaporated to dryness
and ignited. For the determination of phosphoric acid, the residue
was attacked with hydrochloric acid, evaporated to dryness, and ignited
very gently to render the silica insoluble. It was then taken up
with dilute nitric acid, a few grams of ammonium nitrate were added,
Digitized by
Google
AVAILABLE PLANT FOOD IN SOIIiS. 125
with 50 C.C. of a solution of ammonium molybdate, containing 60
grams of molybdic acid per litre. The volume of the nitric acid
solution was always brought to 50 c.c. before adding the ammonium
molybdate, in order that the work should always be carried out
under uniform conditions. The mixture was well stirred and allowed
to stand in a warm place, not exceeding 40^, for 24 hours. The
phosphomolybdic acid, after washing with ammonium nitrate solution,
was dissolved by ammonia into a tared basin, evaporated to dryness,
ignited gently over an Argand burner^ and weighed. The resulting
material was assumed to contain 3*794 per cent, of phosphoric acid.
In potash determinations, the ignited residue from the evaporated
solution was taken up with weak hydrochloric acid and the potash
determined by Tatlock's method as described by Dyer {loc. dt., p. 141),
the potassium platinichloride being sometimes weighed as such, and
.sometimes converted into metallic platinum.
The so-called '' total " potash and phosphoric acid were determined
on portions of the same soils that were ground until they would pass
through a woven sieve of 1 mm. mesh. Twenty grams of such soil were
extracted with 70 e.c. of strong hydrochloric acid containing 20*2 per
cent, of pure acid (that is, the acid which results on boiling the con-
centi'ated acid under ordinary atmospheric pressure) for 48 hours on a
water-bath in a loosely stoppered flask.
The amount of calcium carbonate is calculated from the amount of
carbon dioxide evolved on treating the soil with dilute acid by a
method described in another communication (this vol., p. 81).
Some of the carbon dioxide may be derived from magnesium car-
bonate, but as the factor that is wanted is the amount of '' base "
available in the soil, it is not necessary to attempt to differentiate
between calcium and magnesium carbonates.
All the figures given are calculated as percentages on the soil in an
air-dry condition ; the amount of water each soil loses at 100° is
also given.
I. Phosphoric Acid Results.
Soils /ram the Broadbalk Field.
In the table on p. 126, the results obtained by the action of the
various acids employed on the soils from the seven plots of the
Broadbalk wheat field are set out.
(1). A first inspection of the figures shows that in general citric acid
dissolves the most, ammonium citrate a little less, hydrochloric acid
comes next in order, then acetic acid, the carbonic acid charged water
dissolving least of all. This order of solvent power is preserved in
each plot. Taking the means of the quantities dissolved from the six
manured plots, 2b, 5, 6, 7, 9a, and 16, it will be seen that the citric
Digitized by
Google
126
HALL AND PLTMEN : THE DETERMINATION OF
Table I.
Plot.
Manuring.
Citric.
t
1
HCl. Acetic. COo.
1
Am-
monium
citrate.
0-0433
Strong
HCl.
2b
Dung
0-0477
0-0080
0-0510
0-0446
0-0402
0 0295
0-0208
1
00224 0-0166 00095
0-209
3
Unmanured
0-0021 ; 0-0011 0-0005 i 0*0069
00360 ' 0-0098 0-0058 0*0888
0-0264 0-0086 0-0031 0-0283
0-114
5
Minerals only
0-228
6
Minerals + 200 lb. ammon-
ium salts *.
0-195
7
Minerals + 400 lb. ammon-
ium salts
0 0243 ^ 0-0067
0-0030
0-0021
0-0266
0-0197
0*191
9ft
Minerals + 275 lb. sodium
nitrate
0-0070
0-0051
0 0082
0-0016
0-164
■1
Minerals + 800 lb. am->
monium sal ts. . . 1 3 years
Unmanured 19 years V
Minerals + 550 lb. sodium
nitrate 10 years ^
0-0011
0-0141
0-157
acid dissolves about ten times as much as the carbonic acid, about five
times as much as the acetic acid, and twice as much as the hydrochloric
acid (Table II).
In the case of the unmanured plot, the ratios are of the same
order.
Table IL
TaOj dissolved from
Solvent.
Citric acid
Ammonium citrate .
Hydrochloric acid .
Acetic acid
Carbonic acid
Six manured plots. . Unmanured plot.
0-0390
0 0080
0-0285
00069
0-0202
0-0021
0 0077
0 0011
0-0042
0-0005
(2). The ratios in which the various acids dissolve phosphoric acid
are not the same for each plot, as will be seen from a consideration of
the following table (III)» where the results are recalculated as per-
centages of the " total " phosphoric acid, that is, the amount dissolved
by strong hydrochloric acid from each soil.
It is now seen that as the total phosphoric acid in the soil diminishes,
so does the fraction which is soluble in any of the acids. Citric acid
dissolvesmore than 20 percent, of the total phosphoric acid in the soil from
the dunged plot and from the plots receiving minerals alone or minerals
and ammonium salts; the percentage drops to 13*3 in the soil from
plot 16, which had been for some time unmanured and at other times
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS,
Table III.
127
Plot.
' 2b. I
^ I !
r ,
Total phosphoric acid. 0*209 ' 0114 I 0-228
I
Percentages of total dis-,'
solved by : —
Citric acid 122
Ammonium citrate
Hydrochloric acid . .
Acetic acid
Carbonic acid
1 22-8
7-02 1
20-7
G05
10-7
1-84
7-92
1 0-965
4-53
1 0-439
22-4
17-0
15-1
4-30
2-54
0-195
22-9
14-5
18-5
4-41
1-64
0-191
9a. 16.
0164 0-157
21-1
18 0
13-9
120
12-7
4-28
8-51
1-95
1-57
1-28
13-8
8-98
3-25
1-02
0-701
drained of mlDerals by the use of heavy dressings of nitrogenous
manures, and still further drops to 7 per cent, in the soil from the
unmanured plot. With the other acids, the same progression is
observed. The crops first remove the more soluble portion of the
phosphoric acid within the soil, and on those plots where the phos-
phoric acid has been reduced by cropping, the residue is in a com-
paratively insoluble form, attacked with increasing difficulty by the
dilute acids employed.
(3). ] n order to compare the relative powers of attack possessed by
the acids on the different plots, it is convenient to take as a standard
for each plot the amount dissolved by the citric acid and reduce the
results given by the other acids to this basis. The following table is
thus obtained :
Table IV.
Amount dissolved by
Citric acid
Ammoniam citrate
Hydrochloric acid
Acetic acid
Carbonic acid
Plot 2b.
100
90-8
46-9
34-8
19-9
3. 6.
100 100
86-2
26-2
13-7
6-2
76-1
70-6
19-2
11-4
6.
7.
100
100
63-5
66-2
59-2
60-4
19-3 -
16-7
7-0
7-5
9a.
100
66-8
23-7
10-8
7-1
16.
100
67-8
24-5
7-7
5-3
It is clear that some difference exists between the actions of the
various acids ; if a given acid has twice the solvent power of another
in dealing with one soil, it does not follow that the same ratio will be
preserved on passing to a soil of a different type.
The solution of hydrochloric acid has about two-thirds the solvent
power of the citric acid in dealing with soil from the group of plots
5, 6, and 7, which receive minerals alone or with ammonium salts ; one-
Digitized by
Google
128
HALL AND PLYMEN : THE DETERMINATION OF
half with the soil from 2b, which contains much organic matter ; and
less than one-fourth with the soils from plots 9a and 16, where the
minerals have been accompanied by nitrate.
When compared with citric acid, acetic acid also dissolves a smaller
proportion of the phosphates in the soils from the nitrated plots 9a
and 16, but a higher proportion than usual when dealing with the
dunged plot 5.
Carbonic acid dissolves a fairly constant proportion of the phosphates
dissolved by the citric acid except in dealing with the dunged plot,
when its solvent powers are comparatively high.
The attack of ammonium citrate is relatively speaking at its
best in dealing with the dunged plot and with the continuously i\n-
manured plot.
(4). Turning now to the practical question, which acid yields results
most in accord with the past history of the plots, it will be convenient
to arrange the results in a fresh form. In the following table (Y),
the amount of phosphoric acid dissolved from plot 5 (minerals only)
will be taken as the standard of comparison, thus showing the varia-
tion caused by the plots in the case of each acid. Plot 5 is chosen for
the standard, as it has been continually manured with minerals,
and but scantily cropped owing to the absence of nitrogen ; it should
therefore contain the greatest amount of " available " phosphoric acid.
Table V.
Total.
100
Citric.
Ammonium
citrate.
HCl.
Acetic.
CO^
Plot 5
100
100
100
100
100
„ 2b ...
91-9
93-6
112
62-2
169
164
„ 6
85-5
78-0
87-4
78 '8
87-8
63-5
„ 7
83-8
68-8
78-8
67-5
68-4
61-7
„ 9a ...
71-9
57-8
60-8
19-4
32-7
86-2
„ 16
68*9
40-8
36-4
14-2
16-8
190
n 8
50-0
157
17-8
5-8
11-2
8-6
It is seen that all the weak solvents give more trustworthy
information about the soil than the strong hydrochloric acid does.
With the strong hydrochlorio acid, the variation in passing from the
richest plot, 6, continuously manured with superphosphate and very
scantOy cropped, to the poorest plot, 3, which has been cropped without
manure for 50 years, is only 100 : 50, whereas with other acids the ratio
varies from 00 : 17*8 to 100 : 5*8. With a few exceptions, each of the
acids would set the plots in the same order of fertility ; the ratios of
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS.
129
attack shown by citric acid and ammonium citrate are fairly similar,
those of acetic and carbonic acids are still more alike.
Acetic and carbonic acids and ammonium citrate rate 2b, the dunged
plot, as richer than 5, the plot which receives minerals only.
Hydrochloric acid rates the dunged plot very low, below 6 and
7, receiving mineral manures with ammonium salts ; hydrochloric acid
also rates 9a, the nitrated plot, very low, extracting less than one-
third as much from this plot as from plots 6 and 7, whereas citric
acid would make this plot almost as rich as 6 and 7.
With the variable factors introduced by the long-continued
use of dung, ammonium salts, and nitrate respectively, it would
be difficult to say which of these plots would be shown by crop-
ping as relatively the richest in phosphoric acid; the surplus of
the phosphoric acid supplied as manure over that removed in the crop
during the last 42 years gives some figures wherewith to form an
opinion, but one that does not take into account the different states
of combination into which the phosphoric acid has entered in the
soil.
The following table compares the surplus of phosphoric acid added
to the soil during the last 50 years with the amounts removed from
each plot by the various acids, assuming for the fine earth down to
the depth of 9 inches, an average weight of 2,500,000 lbs. per acre.
The figures are in pounds per acre.
Tablb VI.
Surplus P-Og
retained by
soil.
PgOg dissolved by
Citric.
Ammonium
citrate.
HCl.
Acetic.
COa.
Plot 5
„ 2b...
„ 6
» 7
„ »» ...
., 16
» 8
2582
2619
2356
1985
1885
765
-467
1275
1198
1115
1005
788
520
165
970
1082
707
665
492
852
172
900
560
660
607
175
127
52
245
415
215
167
80
40
27-5
145
237
77-5
75
52-5
27-5
12-5
The following table shows the calcium carbonate and the loss on
ignition of the soils under consideration. The loss on ignition includes
both organic matter and water of hydration, but as the latter is likely
to be constant in dealing with soils from the same field, the
variations in the loss on ignition represent pretty nearly the varia-
tions in the amount of organic matter present.
VOL. LXXXI K
Digitized by VjOOQIC
130
HALL AND PLTHEN : THE DBTEBMINATIOM OF
Table VII.
2b.
3.
6.
6.
7.
9a.
16.
Calcium carbonate
Loss on ignition
3-86
6*21
2-26
3-56
3-82
1-92
8-67
8-66
1-86
2-60
8-76
2-03
2-62
4-44
1-92
4-17
4-49
206
3-03
4*34
Hygroscopic water lost
at 100'
2-33
The amounts of either calcium carbonate or organic matter present
in the soils do not shed any consistent light on the different rates of
attack shown by the solvents employed. The amount of calcium
carbonate present is in no case sufficient to neutralise the acids, for
which purpose about 15 grams of the carbonate would be required.
Much of the calcium carbonate in the soil of plots 6 and 7 has been
removed by the continual use of ammonium salts, and this may ex-
plain why the hydrochloric acid dissolves far more from these plots
than from the nitrated plot 9a, which is richest in calcium carbonate.
00-6
0 04
^ 0-03
i
Phosphoric acid — BroadbaUc Field,
0-02
^ 0*01
I ^'O
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS.
131
But the acetic acid, the solvent action of which is little affected by
variation in the calcium carbonate present, also dissolves less from 9a
than from plots 6 and 7. On the other hand, the dunged plot is rich
in calcium carbonate and is comparatively resistent to hydrochloric
acid, yet it is the plot which yields the most to acetic acid.
It is noticeable that the citric acid and ammonium citrate solutions
contain considerable quantities of organic matter, silica and salts of
iron and aluminium. The same mineral materials are attacked by the
hydrochloric acid, but are not present to any appreciable extent in
the solutions in acetic and carbonic acids. The comparative action of
the various adds may be most clearly seen in the diagram on p. 130,
where the heights of the vertical columns are proportionate to the
amounts of phosphoric add dissolved in each case.
For purposes of comparison, the total phosphoric acid soluble in
strong hydrochloric acid is added, but plotted to the smaller scale of
one-tenth.
FhosphoHe Add Results on other Sails.
The following table shows the percentages of phosphoric acid dis-
solved by each of the adds from the soils 1 to 12 previously described,
arranged according to the total amount of phosphoric acid they contain :
Table VIII.
Soil.
Citric.
HCl.
Acetic.
Carbonic
Ammoniam
citrate.
Strong
HCl.
Total.
1
00056
0 0024
0-0007
0-0083
0-0080
0-078
2
0-0085
0 0013
0-0007
0-0018
0-0295
0-089
3
00100
0 0085
0-0016
0-0080
0-0128
0-089
4
0-0029
0-0021
0-0007
0 0017
0 0104
0 104
5
0-0082
00031
0-0011
0-0023
0-0099
0110
6
0-0033
0*0003
0-0008
0-0008
0 0122
0-112
7
00188
0-00435
0-0006
0-0011
0-0182
0-118
8
0-0210
0-0067
0 0016
0-0019
0-0210
0-121
9
00085
0-0040
0-0016
0 0022
0-0081
0142
10
0-0071
0-0022
0-0019
0-0022
0-0089
0-145
11
0-0240
0-0167
0-0056
0 0014
0-0166
0-152
12
0*0420
0 0860
0-0120
0-0089
0-0540
0-168
(5). It is at once seen that the order in which the soils are arranged
according to the total phosphoric acid is not Mie order of their relative
richness in *' available " phosphoric acid as judged by any one of the
dilute solvents. This is only to be expected, considering the very differ-
ent types of soil here brought together. The results generally afford
K 2
Digitized by
Google
132 HALL AND PLYMBN : THE DBTKBMINATION OP
strong confirmation of the practical value of dilute solvents in judging
of the need of a given soil for a phosphatic manure. With three excep-
tions, all the soils contain more than 0*1 per cent, of total phosphoric
acid, which has been regarded as sufficient for fertility ; yet the cropping
tests of these soils show that only two, 11 and 12, are at all
properly furnished with phosphoric acid. If, on the contrary, Dyer's
limit of O'Ol per cent, of phosphoric acid soluble in 1 per cent, citric
acid be taken as a criterion, the two latter soils are sharply dis-
tinguished from the rest, as containing 0*024 and 0*042 per cent, re-
spectively, and the others with two exceptions would be rated as in
need of phosphoric acid.
With acetic acid as a solvent and a limit of 0*0025 per cent, of phos-
phoric acid soluble, all the soils except the two, 11 and 12, known to
be provided with phosphoric acid, would be rated as in need of a phos-
phatic manuring.
(6). The action of the different acids can be best reviewed by plot-
ting them as before, and also by recalculating the results in terms of
the amounts dissolved by citric acid from each soil, compare Table IX
(p. 133) with Table IV (p. 127).
Table X. (p. 133) shows the calcium carbonate, the hygroscopic
moisture, and the loss on ignition for each soil.
In Table XI (p. 133) the soils 1 to 12 are arranged as the
Broadbalk soils in Table Y (p. 128) ; that is, one soil is taken as a
standard of comparison (in this table, No. 5, which is known to be
very slightly if at all in need of phosphatic manuring) ; the phos-
phoric acid dissolved by each acid from this plot is called 100, and the
amounts dissolved by the same acid from the other plots are reduced
to this standard.
An inspection of the diagram (p. 134) shows that citric, acetic, hydro-
chloric and carbonic acids agree, with one or two exceptions, as to the
comparative richness in available phosphoric acid of any plot. The
vertical columns representing the acids rise and fall together in pass-
ing from plot to plot, as was the case with the Broadbalk soils.
The ammonium citrate, however, gives results essentially different ;
it rates soil 2 as better than 3, the other acids make 3 distinctly
richer than 2 ; again, it rates 4 below 5, contrary to the relative
position assigned to these two soils by the other acids and by cropping
experiments.
From all the soils 1 — 8, 10, and 12, ammonium citrate extracts more
than citric acid, a result never obtained with any of the Broad-
balk soils. The high and irregular results given by ammonium citrate
as compared with the other acids may probably be attributed to the
comparative richness of these soils in organic matter and their poverty
in calcium carbonate. The soils, 2, 4, 6, 7, and 8, which are rated
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS.
133
1
=1
04 coop QO
00 rH 0» «p
o <^ w 00 lb
OOi ^ 1-1 <N
O Or-i A-a»
o o> -^
iF^ rHO
o o a»o> -^
fH rH r-
04 OOO
o o ^
O 04 00 -^ O
OC0t^M«0
rHOO
O ^ rH i-<
O ^04 04 00
O ^ ^ rH to
04
04»O fH OOOOO
rH -^ CO lO a» ^
tfS U3 rH O 00 rH
00 00 A 1-4 04 00
O CO OO Ob CO oo
04 rH»0 ^ rH
OO OO CO
O O»04
O »o -^
Is
So.-
O O CO 7H
« O O O «9 04
00 Ob t«> !>• a> CO
a»
06
0
C0 04C0 00 03O
kO f-l i-H 00 00 f-l
0404 04 rl 1-4
CO
04 «^0OO»
^
0 ^*^oo4
ud kO^<- ^ «3 "^
000 COCOt^Oi
iH
06
SSS|S5
04
OOpQO^
^ 00 04 t^^ 0
0 Ob^ioiooo
rH 04
00000 -^ rH
r«0 W» rH 04 «
CO 00 r« CO ><• CO
:s
^-6
^ § ? o g
Digitized by
Google
134 HALL AND FLYMEN: THE DETERMINATION OF
I
Photpharic oxide, PjOj, per cent.
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS.
135
comparatively high by ammonium citrate, are rich in organic matter, 2
and 7 being the only pasture soils in the group, and 8 an artificially
made soil. No. 5, which is rated low by ammonium citrate, is excep-
tionally deficient in organic matter.
The quantities dissolved by acetic and carbonic acids are very
similar ; it is to be noticed that acetic acid dissolved slightly less than
carbonic acid from the soils 1 — 10, which are, with the exception of 9,
short in calcium carbonate^ but that it obtained the larger amount of
phosphoric acid from soils 11 and 12 and from the Broadbalk soils
which contain more than 1 per cent, of calcium carbonate.
On close inspection of the figures many differences are evident in
the mode of attack of the various acids, which when followed up on
a number of soils will provide information as to the forms in which
the phosphoric acid of the soils is combined. The authors, however,
wish in this communication to confine themselves to the question of
which dilute acid yields results most in accord with the known
history of the soils, and is therefore most likely to be useful in judging
an unknown soil.
(7). A few figures may be here inserted showing the effect of varia-
tion in the strength of the acid used, and of additions of calcium
carbonate to the soil. Dyer {lac, cit.) has already given figures
showing that an increase in the strength of the acid results in more
phosphoric acid going into solution ; the authors' results are in the
same sense :
Table XII.
Solvent.
Percentage of P,Os
dissolved.
SoU7.
SoUA.
Citric aoid 0*2 noimal
0 0198
00188
00084
0*0424
,, 1 per cent
0 0349
,, O'l nonnal
0*0206
Soil A does not appear elsewhere in this paper, but was chosen as
one rich in phosphoric acid and calcium carbonate, but poor in
organic matter, and thus a complete contrast to soil 7.
Soil 7 was further mixed with varying amounts of calcium
carbonate, obtained by grinding Iceland spar to a fine powder, and
subjected to the action of citric, acetic, and carbonic acids, with the
following results :
Digitized by
Google
136
HALL AND PLYMEN : THE DETERMINATION OF
Table XIII.
Soil 7.
Phosphoric acid.
Potash.
Citric.
Acetic.
CO,,
Citric.
Acetic.
Soil only
, , +2 per cent, calcinm carbonate
II +0 II »» >>
,, +10 ,, ,, ,)
0-0188
0-0090
0 0066
0 0007
0-0006
0 0009
0 0006
0-0007
0 0011
0-0007
0-0009
0 0009
0 0148
0 0092
0 0092
0*00714
0-00706
0-00710
These trials were not pushed further; the citric acid as it was
neutralised by the calcium carbonate dissolved less and less phosphoric
acid, until with 10 per cent, of calcium carbonate (more than is
requisite for complete neutrality), the amount of phosphoric acid
dissolved approximated to that dissolved by carbonic acid only. The
solution effected by carbonic acid is independent of the calcium
carbonate present, and that effected by acetic acid approximately so,
because the liberated carbonic acid is an equally efficient solvent.
Review of Result;
(8). On reviewing the whole of the results, it seems very improbable
that any distinction of kind can be drawn between '* available " and
" non-available " compounds of phosphoric acid in the soil ; that is,
there is not a compound or group of compounds '' available," which
can be wholly removed by the plant or dissolved by an acid before the
remaining compounds are attacked. Were this the case, those soils
which contain only a limited amount of " available " phosphoric acid
would yield al] of it or none to a given solvent, and the strength of
the solvent would be without influence on the result when the time
limit is large.
On the contrary, the amount of phosphoric acid dissolved varies
with both the nature and strength of the acid. There is no reason
for regarding the phosphoric acid dissolved by the citric acid solvent
as the '* available " phosphoric acid in the soil rather than that which
is dissolved by the acetic acid.
A soil which contains much or little ** available " phosphoric acid
according to one acid would be rated in the same way by another acid,
even when the absolute amounts dissolved are ten times as great in one
case as in another. The individual acids possess a certain selective
power for different combinations of phosphoric acid and attack the
different types of soils with more or less vigour, but in the main the
relative action of all the acids on all the soils is alika
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS. 137
The phosphoric acid of a soil must not be looked on as existing in
certain compounds A, 6^ C, D, &c,f of which A and B are insoluble and
unavailable, C and D as ''available" ; rather A, B, C, D, &c,, repre-
sent compounds possessing in each soil a coefficient of solubility, vary-
ing with the acid and with their own physical condition. The latter
factor affects all the acids alike, and combined with the absolute
quantity of the phosphoric acid in the soil determines the
" available " phosphoric acid. The available phosphoric acid measured
by a given acid depends on the coefficient of solubility possessed by
the acid and the relative proportions of A, B, C, D, &c,, in the soil*
As soils of the same type contain A, B, C, D, <&c., in roughly the same
proportion, the latter factor is eliminated and the amounts of available
phosphoric acid from different soils as measured by any one of the acids
will be proportional to the phosphoric acid which is really '' available/'
so that all the acids will show roughly the same relations between the
soils.
Again, a soil may contain di- and tri-oalcium phosphates, ferric
and aluminium phosphates, and organic compounds of phosphorus
like nuclein and lecithin ; it would be no gain to discover a
reagent which would dissolve the di- and tri-calcium phosphates
only and leave the rest, for the physical conditions of these phosphates
may render them less " available " to the plant than the other com-
pounds of phosphorus present which happen to be in a favourable
physical or mechanical condition for solution.
On this view the hope must be abandoned of finding any particular
acid which will dissolve out the ''available" phosphoric acid and
leave the rest ; in the results obtained by any acid, the factors are too
numerous and variable to admit of exact discussion ; because of its
complexity, the method becomes empirical and the best acid is that
which most accords with experience.
(9). In forming a conclusion as to the most suitable solvent, three
things should be taken into account :
(a) The amount of phosphoric acid dissolved should show a wide
variation in passing from soil to soil, so as to discriminate sharply
between rich and poor soils. The largest quantity of phosphoric acid
dissolved by strong hydrochloric acid from any one of the soils examined
is 0*228 and the smallest 0*0727 per cent. ; other things being equal,
variations of this order would not discriminate so well between the
soils as the variations exhibited by citric acid, which lie between 0*051
and 0*0029, or of acetic acid, which lie between 0*012 and 0*0003 per cent.
(b) The afQount of phosphoric acid dissolved from normal soils
should be sufficient for exact estimation, so that the variations ex-
hibited may be of a different order of magnitude from the experimental
error, which is inevitably large.
Digitized by
Google
138 HALL AND PLTMEN : THE DETERMINATION OF
(c) The variations in the amount of phosphoric aoid dissolved should
so follow the known history of the soils that the reaction of an un-
known soil to phosphatic manures can be predicted from its analysis.
For this reason, the action of the acid should not be markedly affected
by other variable constituents in the soil, such as calcium carbonate
and organic matter.
Ammonium citrcUe fails to meet the last requirement ; although
when dealing with soils of one type, like the Broadbalk soils, its results
fall into line with those given by the other solvents, yet with the
■other soils the indications provided by the analysis do not agree with
experience, ^ils 2, 7, and 8 yield comparatively large quantities of
phosphoric acid to ammonium citrate solution and would be rated as
sufficiently, supplied with phosphoric acid, but 2 and 7 respond freely
to phosphatic manures. Soils 4 and 6 yield more phosphoric acid than
5, which is quite contrary to the crop results.
These discrepancies are due to the solubility of the humus containing
phosphorus compounds in the alkaline ammonium citrate solution,
thus introducing material of a different order of solubility, and as
the ammonium citrate solution offers no compensating advantages it
may be dismissed as unsuitable.
Hydrochloric acid presents many anomalies of attack ; it has very
little solvent power for phosphoric acid when dealing with soils 1 — 10
which are poor in calcium carbonate ; for example, it can only dissolve
0*0031 per cent, from soil 5, which is fairly provided with phosphoric
acid as judged by the, crop, whereas it can get 0*0021 percent, from
the unmanured plot at Rothamsted, and as much as 0*0167 per cent,
from soil 11, the poorish chalky Wye soil which had been unmanured
for 5 years. The Broadbalk plot 9a, which receives minerals and
sodium nitrate, is rated very low ; it yields only three times as much
phosphoric acid as the continuously unmanured plot, and less
than one-third as much as the corresponding plot 6, which receives
ammonium sulphate instead of sodium nitrate. The dunged plot is
also rated as inferior to the plots receiving minerals and ammonium
salts.
On the whole, the results obtained with hydrochloric acid are difficult
to reconcile with experience, and present no features which would
justify its recommendation in place of citric acid.
Water charged unth carbonic acid is so similar in its action to acetic
acid, both in the relative and absolute amounts dissolved from the
various soils, that the greater convenience of using the latter acid
would cause it to be preferred.
The choice thus becomes narrowed down to acetic and citric adds.
Of these two, acetic acid better satisfies the first condition laid down
above, the variations in the amounts dissolved are larger. With the
Digitized by
Google
AVAILABLE PLANT POOD IN SOILS. 139
Broadbalk soils they range from 169 to 11% against 100 to 15*7 for
citric acid (Table Y). On the other soils they range from 1053 to
26*3, against 512 to 35 for citric acid (Table XI).
As regards the second criterion, the quantities of phosphoric acid
dissolved by the acetic acid are very small, one-tenth to one-fifth of
the amount dissolved by citric acid. The limit to be taken as
indicating the need for phosphatic manuring would be about 0*002
per cent., which means the determination of only 0*001 gram of
phosphoric acid in the 500 c.c. of isolation commonly employed. On the
other hand, the acetic acid solution is the easier to manipulate, owing
to the absence of iron, alumina, silica, and dissolved organic matter ;
so that the experimental error is not likely to be greater than with
citric acid, less indeed in unskilled hands.
As regards the interpretation of the results, it is clear that all soils
deficient in calcium carbonate, as 1 — 8, are rated very low by acetic
acid. In such soils, much of the phosphoric acid is present as precipitated
ferric and aluminium phosphates, which are left practically untouched
by the acetic add, yet there is no evidence that such phosphates are
quite " non-available " for the crop. Soil 5 is a case in point ; acetic
acid dissolves only 0*001 per cent, of phosphoric acid, yet the crops on
this soil find no great need of phosphates. The Broadbalk soils are
very clearly differentiated by acetic acid, the doubtful point being the
comparatively low position attached to 9a and 1 6, the nitrate plots. The
position assigned to these two plots and to 5 in the other group makes
it difficult to accept acetic acid as the most '' critical " solvent*
Considering the results yielded by citric acid, some difficulty of
interpretation attaches to soils 2, 3, 7, and 8.
Taking the limit of 0*01 per cent, of phosphoric acid suggested by
Dyer, soils 7 and 8 are above the limit with 0*0133 and 0*021 per
cent, respectively ; soil 3 is on the limit, and soil 2 is a little below
with 0*0087 per cent. ; yet the field trials indicate a need of phosphates
on soils 2, 3, and 7, probably on 8 also, although as an exceptional soil
it is hardly comparable with the rest.
Of all the soils examined, soils 2, 7, and 8 show the greatest loss on
ignition ; 2 and 7 are old pastures, 8 is a made soil containing leaf
mouldy and as citric acid dissolves some of the organic matter of soils, it
is to this source that the high proportion of phosphoric acid yielded by
these soils may be attributed. Probably the superior limit of 0*01
per cent, of phosphoric acid, as indicative of the need of phosphatic
manuring, requires revision when dealing with pastures and other
soils rich in organic matter.
The results yielded by soil 5 also require a little explanation ; the
citric acid solution only dissolves 0*0082 per cent., yet the crops show
no exceptional response to phosphatic manuring. The soil is a very
Digitized by
Google
140 HALL .AND PLYMBN : THE DETEKMIKATION OF
light sandy loami typical of many of the soils derived from coarse,
ferruginous sandstones of secondary age. It contains very little calcium
carbonate (0*08 per cent.) and little organic matter (loss on ignition
3*08 per cent.). The phosphoric acid must be largely present in. this
soil as ferric phosphate, and although citric acid is a better solvent than
acetic acid in such cases, even the citric acid does not indicate all the
phosphoric acid that seems to be " available " for cropa Gerlach (Zoe.
cit.) has already indicated that typically sandy soils from which citric
acid dissolves less than 0*01 per cent, of phosphoric acid may give
little response to phosphatic manures.
As regards the Broadbalk soils, the results yielded by citric acid are
more in accord with our knowledge of the plots than those furnished
by acetic and the other acids ; in particular the plots receiving nitrate
9a and 16, though below all the others except the unmanured plot, are
shown as still high above the limit which may be taken to indicate
the need of phosphatic manuring.
Reviewing the whole body of results, the authors consider the 1
per cent, solution of citric acid gives results which are most in accord
with the known history of the soils. On soils well provided with cal-
cium carbonate all the acids tried give very similar relative results, but
this type of soil is rarely in need of phosphatic manuring, and the
practical question for which the analysis is performed, whether the
soil is in need of phosphatic manuring or not, usually arises in the case
of soils poor in calcium carbonate.
From these soils, acetic acid can extract so little that it reduces them
all to practically the same level, whilst citric acid is able to dissolve the
natural phosphates of iron and alumina in a manner more in accord
with the natural attack of crops.
II. Potash REsuLTa
Methods of analysis based upon the solvent action of weak acids
must be even more empirical, when dealing with the potash in soils
than with the phosphoric acid. Certain definite compounds of pho'B-
phorus, such as the organic residues, the phosphates of the sesqui-
oxides, the neutral and acid phosphates of calcium and magnesium,
exist in the soil, and are, to some extent, differentially attacked by
the various solvents, but the potash compounds are far more com-
plex and indefinite. In addition to more or less weathered silicates,
like felspar and glauconite, there are indefinite compounds formed
when humus and clay withdraw potash from the solution produced
by the weathering of potash minerals or the application of manures.
Even the amount of potash dissolved by strong hydrochloric acid
from a soil is a purely conventional figure, dependent on the strength
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS.
141
of the acid and the length of attack ; the Broadbalk soils, for example,
yield about 0*5 per cent, of potash to strong hydrochloric acid, but
the total potash contained in the soil from plot 5, as determined
after breaking up the soil completely with ammonium fluoride,
amounted to 2*26 per cent.
The tables below show the results yielded by the soils from the same
seven plots of the Broadbalk Field at Eothamsted, and by five other
soils previously described ; the results are also set out graphically on
p. 143 in the same manner (compare p. 131) as were the phosphoric
acid results.
Table XIV.
FoUuh — soils from Broadbalk Field,
Plot.
Manaring.
Citric.
HCL
Acetic.
Carbonic.
Strong
HCL
2b
Dung
0-0400
0-0048
0-0468
0*0822
0*0288
0 0272
00208
0-0684
0-0147
0-0622
0-0487
00464
0-0414
0-0421
0 0461
0-0082
0-0807
0-0271
0-0240
0-0287
0-0184
0-0380
0-0111
0-0216
0-0161
0-0091
0-0288
0*0146
0*463
3
Unmanured
0-380
5
Mineralfl only
0-468
6
Minerals 200 lb. ammonium
salts
0-680
7
Minerals 400 lb. ammonium
salts
0-600
9a
Minerals 276 lb. sodium
nitrate
0*440
16-
Minerals 800 lb. ammon- .
ium salts 13 years
Unmanured 19 years •
Minerals 660 lb. sodium
nitrate 10 years''
0-604
On examining the results yielded by the Broadbalk soils, it is
noticeable that the amounts of potash dissolved by the different
acids are very similar, much more so than with phosphoric acid.
Citric acid dissolves ten times as much phosphoric acid as the water
charged with carbonic acid, whereas hydrochloric acid, the most energetic
solvent for potash, dissolves only about three times as much as the weak-
est, which is again carbonic acid. On the whole, each acid leads to the
same conclusions with regard to the relative richness of the plots in
" available " potash, but citric acid shows the widest variation in passing
from plot to plot ; the ratio of 2b, the dunged plot, to 3, the un-
manured plot, is 9*3 : 1 for citric acid against 4*65 : 1, 5*5 : 1, and 3*4 : 1
for hydrochloric, acetic, and carbonic acids respectively.
The results with the Broadbalk soils would indicate that the citric
acid is the most ** critical " solvent for '* available " potash in the
8oiL
Digitized by
Google
142 HALL AND PLTMEN : THE DETERMINATION OF
ft.
Potmh, Kfi, per cmL
Digitized by
Google
AVAILABLE PLANT FOOD IN SOILS.
143
Table XV.
FoUuh dissolved by weak (kdds from other soils.
Soil
Citria
Ha
Acetic.
Carbonic.
■ Strong
HCl.
CaCO,.
Loss
on
Water
lost at
ignition.
100^
11
00060
0 0480
0-0104
0-0176
0-390
4-59
4-08
2-06
12
0 0098
0-0580
0 0156
0 0241
0-378
3-32
4 01
1-87
4
0 0250
00113
00058
0 0057
0-813
0-01
4-74
3-13
5
0-0110
0-0154
0-0069
0-0062
0-439
0-08
3-09
2-34
13
0-0085
0-0178
0 0058
0-0079
0-592
002
12-53
13-04
Of the other soils examined, 11 and 12 should be compared together
as soils freely supplied with calcium carbonate, whereas soils 4,
5, and 13 are notably deficient in this constituent. Soils 11 and 12
are from the plots, -side by side, on the same field, shown by experi-
ment not to be particularly in need of potash manuring. No. 11
had been cropped without manure for 5 years, during which time 1 2
had received each year a general manure containing 1^ cwt. per acre of
potassium sulphate. All the weak solvents show 12 as richer than 11
in '' available " potash, whereas the strong hydrochloric acid would
make them practically alike. The difference between them is most
sharply drawn by citric acid ; it is also noticeable that citric acid
shows both plots as comparatively poor in ** available ** potash, the
other three acids would rate them as comparatively rich.
Of the other three soils, field experiments have shown that 4, a
strong clay, is in no need of potash manuring, but 5 and 13 gave very
marked returns for potash dressings. Strong hydrochloric acid would
make both 5 and 13 much richer in potash than 4; it dissolves 0*592
and 0*439 per cent, respectively from 13 and 5, against 0*313 per cenU
frona soil 4. Dilute hydrochloric acid would also set soil 4 below 5
and 13 in "available" potash, acetic and carbonic acids would rate
them alike, the differences between the various results being of the
same order as the experimental error. Citric acid alone draws a
sharp distinction between the soils ; it dissolves 0*025 per cent, from
4, and only 0*011 and 0*0085 per cent, respectively from the other two
soils.
The results with these five soils afford most striking evidence of the
practical value of weak solvents as against extntction with a strong
acid in judging of the requirements of a soil for a potash manure ; at
the same time, they indicate it may be necessary in the light of extended
experience to adopt different limits for soils of different types, for
example^ soils rich or poor in calcium carbonate.
Digitized by
Google
144 THE DETERMINATION OF AVAILABLE PLANT FOOD IN SOILS.
Of the four weak acids employed, the authors regard citric acid as
furnishing results most in accord with the history of the soils exam-
ined.
Summa/ry,
The authors have compared the amounts of phosphoric acid that could
be extracted from nineteen different soils by a 1 per cent, solution of
citric acid, by equivalent solutions of hydrochloric acid and acetic acid,
by a saturated solution of carbonic acid, and by an ammoniacal solution
of ammonium citrate respectively. Seven of these soils were from plots
on the Broadbalk Field, Rothamsted, which had been continuously
manured in the same manner for forty -two years previously ; the r^
maining twelve were soils of very varied origin, which had been the
subject of crop experiments and whose reaction to phosphatic manuring
was well marked.
In the same seven soils from the Broadbalk Field, the authors deter-
mined the potash extracted by the same dilute solvents, with the exception
of ammonium citrate ; five other soils of different origin, whose response
or otherwise to potash manuring had been tested by experiment, were
also examined in the same way.
Determinations were also made of the phosphoric acid and potash
dissolved after long digestion with strong hydrochloric acid, of the loss on
ignition, and of the earthy carbonates present in each soil.
The authors conclude : — (1). That no sharp line of distinction can
be drawn between '< available " and non-available phosphoric acid and
potash in the soil, and that any process of determining the " available "
constituents is an empirical one, dependent on the strength and nature
of the acid used.
(2). That the weak solvents give information as to the requirements
of a given soil for mineral manures of a far more trustworthy nature
than that which is afforded by such a solvent as strong hydrochloric
acid.
(3). That of the acids examined, the 1 per cent, solution of citric
acid gives results most in agreement with the recorded history of the
soil, although there is evidence that the same interpretation cannot be
put on results obtained from all types of soil.
South Eastbkn Aorioultubal Golleob, Wyb,
Digitized by
Google
DOBBIE AND LAUDER: CORYDALTNE. PART VH. 145
XIV. — Corydaline. Part VIL ' The Constitution of
CorydcUine.
By Jambs J. Dobbie, M.A.y D.Sc., and Alexander Lauder, B.Sc.
The results obtained by oxidising corydaline with potassium perman-
ganate and nitric acid have been described in previous communications.
In the present paper, some additional experimental details are given,
and the whole of the results are discussed in their bearing on the con-
stitution of the alkaloid.
Attention has already been drawn to the resemblance which cory-
daline bears to berberine (Trans., 1899,75, 670). This resemblance is
not merely superficial ; the two alkaloids probably differ only in some
of the details of their structure. The comparison, however, must be
drawn, not between corydaline and berberine, but between corydaline
and tetrahydroberberine, or between dehydrocorydaline (which differs
from corydaline by 4 atoms of hydrogen) and berberine. The con-
clusion, based on the chemical investigation, that the two alkaloids are
closely related, has been confirmed by an examination of their absorp-
tion spectra, which we have found to be almost identical. The spectro-
scopic results will form the subject of a separate communication.
€k)rydaline has been analysed in recent years by various chemists,
with results practically identical with those which we published in
1892 (Trans., 61, 244 ; Freund and Josephi, AnnaUtij 1893, 277, 1 ;
Ziegenbein, Arch. Pkarm,, 1896, 234, 492 ; Martindale, ibid., 1898,
236, 214). F;:om the analytical results, we deduced the formula
C^^3L^0^^, and Freund and Josephi the formula OsgH^^O^N. The
latter is probably the correct formula.
By the action of mild oxidising agents such as dilute nitric acid or
iodine in alcoholic solution, 4 atoms of hydrogen are removed from
the corydaline molecule and an intensely yellow base, dehydrocorydaline,
GjsH^O^N, is produced, from which, by reduction, an optically in-
active modification of the alkaloid may be obtained (Ziegenbein, loc*
eU. ; E. Schmidt, Arch. Pharm., 1896, 234, 489 ; Dobbie and Marsden,
Trans., 1897, 71, 657). The ease with which corydaline can be oxi-
dised to dehydrocorydaline, and dehydrocorydaline reduced to cory-
daline, shows that these two substances are very closely related to one
another. It wiU be remembered that berberine, which is a yellow base
like dehydrocorydaline, and tetrahydroberberine, which resembles cory-
daline in being colourless, can also be readily converted the one into the
other.
When corydaline is heated with a concentrated solution of hydrogen
iodide, it is converted into a phenoiio derivative containing four hydr*
VOL. LXXXI. I.
Digitized by VjOOQIC
146 DOBBIE AND LAUDER: CORYDALINE. PART YII.
oxy\ groups, each molecule of corydaline yielding 4 molecules of
methyl iodide. The alkaloid has therefore all its four oxygen atoms
present in methoxyl groups (Trans., 1892, 61, 605).
By oxidising corydaline with potassium permanganate at the boiling
point, the chief products of oxidation are hemipinic and m-hemipinic
acids:
OMe
OMe/NcOjH OMe/NcOaH
I JcOjH OMel ICO2H
HemipiDic acid. m-Hemipinic acid.
The presence of two benzene nuclei in the molecule is thus established
(Trans., 1894, 66, 57; 1897,71, 657; 1899, 76, 670). Along with
the hemipinic acids, a small quantity of corydaldine is also obtained,
the yield of which is considerably increased by conducting the oxida-
tion at the ordinary temperature. Corydaldine has been shown to
CHj-O^ XO-NH
have the following constitution, p-rr . /y^CgHj^pxi- ./itr > which
proves the presence of an uoquinoline nucleus in the alkaloid (Trans.,
1899, 76, 670).
When nitric acid is used as the oxidising agent in place of perman-
ganate, dehydrocorydaline is first produced ; one of the bensene nuclei
is next destroyed, and the beautiful, yellow, dibasic corydic add,
O^gHj^O^N + ^HjO, formed. When corydic acid is in turn oxidised
with permanganate at the boiling point, it is resolved into an in-
soluble, colourless, tribasic acid, CiYH^^OgN, which we propose to term
corydilic acid, a methylpyridinetricarboxylio acid, and m- hemipinic
acid (Dobbie and Marsden, Trans., 1897, 71, 657). In the preeent
paper, it is shown that the methylpyridinetricarboxylic acid has either
the formula
CO,H 00,H
/\cOoH CO,H,/\
COjHI /'Me *''' COjHI JMe
Corydilic acid, on continued boiling with potassium permanganate, is
gradually split up into a mixture of the methylpyridinetricarboxylic
acid and m-hemipinic acid.
These results afford a basis for the discussion of the constitution of
corydic acid. This acid is derived from dehydrocorydaline by the
destruction of one of the benzene nuclei, and since it yields m-hemipinic
acid as one of its oxidation products, the nucleus which is destroyed
must be that from which hemipinic acid is derived. The 2-methyl-
pyridinetricarboxylic acid, which is also one of the oxidation pro-
ducts of corydic acid, contains 6 atoms of carbon^ exclusive of the
Digitized by
Google
THE CONSTITUTION 09 CORTDALINS.
14?
9 -a ,o
9 5
fi Co
I J
1
1
o
r
6< •
•Ms'*. I
I
1^
sa
•33-!:
■fsa
.a»!
-I
1
|c5^
I
L 2
Digitized by VJ OOQ IC .
148
DOBBIE AND LAUDER: CORYDALINB. PART YU.
carb6n atoms of the carboxyl groups. It cannot, therefore, be derived
from the pyridine ring of the t^oquinoline nucleus, since the investiga-
tion of corydaldine has shown that this pyridine ring has no side chain
attached to it. The 2-methylpyridinetricarbozylic acid represents,
therefore, a second ring to which the nitrogen atom, as in the case of
berberine, must be common. We thus arrive at the following formula
for corydic acid :
CO2H
i^
\
01&ef\^V
iCOjH
Me
OMe!
A>.
Fig. 1. — Corydic acid.
This formula accounts for the relation of the molecular formula of
corydic acid to that of dehydrocorydaline ; for the presence of the two
carboxyl groups, and for the formation, on oxidation, of corydilic acid,
the 2-methylpyridinetricarboxylic acid, and m-hemipinic acid. The
formation of the last-named acid establishes the position of the
methoxyl groups. There is no direct experimental evidence to prove
that the positions of the carboxyl groups are those which we have
assigned to them, rather than the positions 4 : 5, but we shall presently
state our reasons for introducing a direct] link between the carbon
atoms 2 and 5, which limits the carboxyl groups to the positions shown
in the formula.
The formula (2), which we have assigned to dehydrocorydaline
follows from that of corydic acid. Ferkin's formula for berberine is
placed side by side for comparison (Ferkin, Trans., 1889, 65, 63) :
. ll"\lMe
OMe;
OMe
IV
^^
Fia. 2. — ^Dehydrocorydaline.
^"
iiv>Me
IIOMe
OMe^V^^
HJ " JHMe
OMe'.
:K}-'
\.
Fio. 4.— Cotydaline.
Tetrahydroberberine.
Digitized by
Google
THE CONSTITUTION OF CORYDALINE. 149
Corydaline differs from dehydrocorydaline in containiDg four more
atoms of hydrogen. Having regard to the great ease with which
corydaline can he oxidised to dehydrocorydaline and the latter
substance reduced to corydaline, it may be assumed that we have to
do here with a group similar to that which exists in certain anthraoene
and acridine derivatives, and such as Perkin has assumed to be pre*
sent in berberine. The existence of a double bond between the carbon
atoms 5 and 6 and of a direct bond between the carbon atoms 2 and 5
in ring II of the formula for dehydrocorydaline (Fig. 2) would explain
the ease with which the one substance passes into the other.
The formula proposed for corydaline, (Fig. 4), explains the reac-
tions and accounts for the formation of all the derivatives of the
alkaloid which have been examined. By oxidation, *the rings, which
for convenience of reference we have numbered I and lY on the
diagram, would yield hemipinic and m-hemipinic acids respectively, and
ring II methylpyridinetricarboxylio acid. Oorydaldine, OiiH^gOgN
(Fig. 6), containing rings III and lY, would result from the oxidation
of corydaline in the same way as oHiminoethylpiperonylcarboxylic
anhydride (Fig. 7) results from the oxidation of berberine :
can
?^fT^S on.<^Y>S OM,V>'
H^ H, ^/\C0jH
Pig. 6.— Corydaldtne. Fio. 7.— w-Amuioethyl- Fio. 8.— Corydilic acid.
Corydic acid (Fig. 1) would be formed by the destruction of ring I,
and corydilic acid (Fig. 8) from corydic acid by the oxidation of
ring III.
If our formula are correct, they inoidentally prove that Ferkin's
formula, which we have quoted, is to be preferred to the alternative
formula suggested by him for berberine, in which the carbon atoms
2 and 5 are connected by a double bond, because, on account of the
presence of the methyl group in dehydrocorydaline, no double bond
is possible between the carbon atoms 2 and 5, and if a double bond
existed in berberine in this position the very close resemblance between
the two substances would not be satisfactorily explained.
When the decomposition products of berberine are compared with
those of corydaline, a close parallelism is observed between them. Both
alkaloids yield hemipinic acid as a derivative of ring I. From rings
in and lY, <o-aminoethylpiperonylcarboxylic anhydride is obtained in
the case of berberine, just as corydaldine is obtained from the corre-
Digitized by
Google
160 DOBBIE AND LAUDEB : COBTDALINE. PART VII.
sponding rings of corydaline. Ring IV of berberine yields hjdrastic
acid:
"=.<80?8:i.
the corresponding decomposition product of corydaline being m-hemi-
pinic acid. The oxidation product obtained from ring II is of special
interest in the case of both alkaloids. Weidel {Ber., 1879, 12, 410),
by oxidising berberine with strong nitric acid, obtained as chief oxida-
tion product berberonic acid :
CO,H
oo^O"-"
We have also obtained the same acid from a new derivative of ber-
berine, which is described in another communication. In discussing
the constitution of berberine, Perkin does not take into account the
occurrence of berberonic acid amongst its decomposition products. It
is clear, however, that its occurrence affords important confirmation of
the correctness of his formula, since it would result from ring II by
the oxidation of. the attached rings I and III, but could not result
from ring in, which would yield cincbomeronic acid. There is thus
direct evidence in the case of berberine, as well as in the case of
corydaline, of the existence of a fourth closed chain in the molecule
of the alkaloid. It is remarkable that both in the case of berberine
and of corydaline, ring II is the more stable of the rings to which the
nitrogen atom is common. From neither alkaloid has any acid corre-
sponding to ring in been obtained. A further instance of the com-
parative ease with which ring III in corydaline is broken up is afforded
by the formation of corydilic acid from corydic acid.
Whilst our formula for corydaline satisfactorily accounts for the
similarity between this alkaloid and berberine, it also explains the
absence from amongst the decomposition products of corydaline of
derivatives corresp<mding to berberal, Gg^Hj^O^N, berberilic aoidi
O^oHj^O^N, oxyberberine, CgoHj^OgN, &o., all of which have an atom
of oxygen attached to the carbon atom 2 of ring IL On account of
the presence of the methyl group in combination with the correspond-
ing carbon atom in corydaline, it would be impossible for an oxygen
atom to occupy this position in similar derivatives of corydaline. On
the other hand, the formation of corydic acid from corydaline suggested
that it might be possible to obtain a similar acid from berberine. We
show in a separate communication that by the oxidation of berberine
with dilute nitric acid such an acid is readily produced.
One further point remains to be dealt with, the stability of the
Digitized by
Google
THE CONSTITUTION OF OORYDALINE. 161
methyl group in ring II. With the exception of the pyridinetetra-
carbozylic acid (see below), all the oxidation products in which ring II
is present, so far examined by us, contain this group. This is not
remarkable when it is recalled that prolonged treatment with
potassium permanganate in alkaline solution is required for the
preparation of 2:3:4: 6-pyridinetetracarboxylic acid, either from
2 : 4 : 6-trimethylquinolinecarboxylic acid (Michael, Annalen, 1884,
225, 121) or from flavinol (Fischer and Tauber, Ber., 1884, 17, 2925).
When, however, large quantities of corydaline are oxidised it might
be expected that small quantities of a monocarboxylic acid should
be obtained. We believe that we have had such an acid in our hands.
In our earlier experiments, in which several hundred grams of cory-
daline were oxidised with potassium permanganate at the boiling
point, a small quantity (about 1*5 grams) of a colourless nitro-
genous acid which crystallised in tufts of delicate, silky needles and
melted sharply at 156'' (Trans., 1895, 67, 17) was obtained. We
were only able to make a slight examination of this substance. A
nitrogen determination gave a result agreeing with that required by
the formula C2iH2404N*C02H (nitrogen, found, 3*55 ; calculated, 3*50
per cent.). A determination of the methoxyl groups by Zeisel's
method showed that the four methoxyl groups present in corydaline
were also present in this acid, and the analysis of a silver salt showed
that the acid possessed a high molecular weight
We leave over for the present the full discussion of the relation be-
tween the constitution and the colour of some of the corydaline deriva-
tives. The further investigation of the products obtained by the
oxidation of corydic acid with potassium permanganate at the
ordinary temperature, described below, promises to throw further
light on this question. It may, however, be mentioned now that the
colour seems to depend on the presence of rings II and III, since only
the derivatives which contain these rings are coloured.
EXPFBIHBNTAL.
The oxidation of corydic acid with potassium permanganate (Dobbie
and Marsden, Trans., 1897, 71, 657) has been repeated on a larger
scale, and the results already published have been confirmed; the pro-
ducts of oxidation are corydilic acid, Ci2HgN(0'CH3)2(C02H)3, a methyl-
pyridinetricarboxylic acid, C^^HyOgN, and w-hemipinic acid.
Examination of Hie Afethf/lpyridineirtcarhoxylic Acid.
This acid can be obtained, not only by the oxidation of corydic acid
with permanganate, but also by the oxidation of corydaline with strong
nitric acid ip the manner followed by Weidel in the preparation of
Digitized by
Google
152 DOBBIE AND LAUDER: CORYDALINE. PART VII.
berberonic acid from berberine (Ber.^ 1879, 12, 410). The yield by
this method is, however, unsatisfactory. The analysis and general
properties of this acid have already been given (Trans., 1897, 71, 657).
The copper salt, obtained by adding copper acetate to a neutral solution
of the acid is blue in colouV, and not yellow, as previously stated. This
acid is undoubtedly a methylpyridinetricarbozylic acid, as is shown by
its analysis and the analysis of its salts, but it is not identical with
any of the known acids of this constitution. Freund and Josephi
{Annalen, 1893, 277, 10), from the similarity in behaviour of methyl-
corydaline and hydrohydrastinine, inferred that corydaline, like
hydrastine, contains a methyl group attached to the nitrogen atom.
By heating the acid with sodium amalgam, we failed to obtain any
evidence of the formation of methylamine, and concluded from this
that the methyl group was not attached to the nitrogen atom, as
Freund and Josephi suggested. This conclusion was confirmed by the
investigation of corydaldine which has no methyl group attached to its
nitrogen atom. Further, Herzig and Meyer {MonaUh,, 1897, 18, 385)
showed that there are only four methyl groups altogether in corydaline
which can be split off by the action of hydrogen iodide, and since we
have shown that there are four methozyl groups, there can be no
methyl in union with the nitrogen atem.
The methylpyridinetricarbozylic acid is an exceedingly stable sub-
stance and can be boiled for some time with a dilute solution of
potassium permanganate without undergoing any appreciable amount
of oxidation. When, however, it is dissolved in excess of potassium
hydroxide and a solution of potassium permanganate added, it slowly
undergoes oxidation, the operation requiring from eight to nine days at
the temperature of the water-bath for completion. Two experiments
were made, one with 3 grams and the other with 2 grams of the
acid. The excess of permanganate was reduced, the alkaline solution
filtered, neutralised with nitric acid and treated with calcium nitrate
to remove a small quantity of oxalic acid which had been formed.
After filtering from the precipitated calcium oxalate, the solution was
treated with lead acetate and the precipitate filtered off and washed.
On decomposing this precipitate with hydrogen sulphide, a strongly
acid solution was obtained, which on evaporation yielded a residue
very soluble in water and insoluble in alcohol. This residue contained
inorganic matter. Its solution was found to give an insoluble salt
with copper acetate which remained undissolved even when heated with
acetic acid. It was therefore precipitated with copper acetate with
the object of removing the inorganic matter, the blue copper preci-
pitate filtered, well washed first with strong acetic acid and then with,
water, and decomposed with hydrogen sulphide. The acid obtained
from the filtrate was still found, however, to be contaminated with a
Digitized by
Google
THE CONSTITUTION OF CORTDALINE. 158
small quantity of inorganic matter, from which by reprecipitation
we were unable completely to purify it. We were thus unable to get
an accurate determination of the melting point or a specimen of the
acid in a sufficiently pure state for analysis.
So far as the qualitative examination was concerned, the acid
showed all the properties and gave all the reactions of 2:3:4:6-
pyridinetetracarbozylic acid obtained by Michael (Anndhn, 1884,
225, 121) from 2 : 4 : 6-trim6thylquinolinecarboxylic acid, and by
Fischer and Tauber (Ber., 1884, 17, 2925) from flavinol. It agreed
with this acid in being very easily soluble in water and very
sparingly so in alcohol; in giving with ferrous sulphate a dark
cherry-red colour, and with ferric chloride a yellow precipitate.
With calcium chloride, the free acid gave no precipitate, but with
barium chloride a copious white precipitate. The copper salt, as
already mentioned, was insoluble even in boiling acetic acid. The
silver salt on ignition decomposed suddenly, swelling up and filling the
crucible with reduced silver which resembled a mass of tea leaves,
exactly as described both by Michael and by Fischer and Tauber.
Further information as to the identity of the oxidation product of
the methylpyridinetricarboxylic acid was obtained by boiling it with
strong acetic acid. When 2:3:4: 5-pyridinetetracarboxylic acid is
heated at 160°, 3:4: 5-pyridinetricarboxylic acid is obtained, and
2:3:5: 6-pyridinetetracarboxylic acid decomposes at 150° into 3 : 5-
pyridinedicarboxylic acid. In both cases, the carboxyl groups which
are eliminated are adjacent to the nitrogen atom. It was therefore
to be anticipated that, under similar treatment, the tetracarboxylic
acid obtained by the oxidation of the methylpyridinetricarboxylic acid
would yield cinchomeronic acid by the elimination of the carboxyl
groups 2 and 6, if we had rightly identified it. As a matter of fact,
we found that cinchomeronic acid was produced by boiling with acetic
acid, and identified without difficulty. The tetracarboxylic acid was
boiled for some time with strong acetic acid and the solution evaporated
to dryness. The residue was insoluble in cold and only dissolved with
difficulty in hot water. The aqueous solution deposited the acid on
cooling in colourless, prismatic crystals, which after purification by
recrystallisation melted at 260°. The acid was insoluble in chloroform,
almost insoluble in ether, and only very slightly soluble in alcohol. It
gave no reaction with ferrous sulphate or with ferric chloride. Silver
nitrate and lead acetate gave white precipitates when added to its
aqueous solution. Calcium and barium chlorides gave no precipitate
even on the addition of ammonia. The copper salt was more soluble
in cold than in hot water and was precipitated by warming a cold
aqueous solution ; the precipitate redissolved again on cooling. The
last reaction which is characteristic of cinchomeronic (pyridine-3 : 4-
Digitized by
Google
154 DOBBIE AND LAUDER: CORYDALINE. PART VII.
dicarbozylic) Acid, taken in conjunction with the melting point, solu-
bility, and the reactions above described, left no doubt as to the identity
of the acid which we had obtained. Cinchomeronic acid might be
formed either from pyridine-2 : 3 : 4 : 5-tetracarbozylic acid or -2 : 3 : 4 : 6-
tetracarboxylic acid, by the elimination of the carboxyl groups 2 and 5
or 2 and 6 respectively. It could not be derived from the 2:3:5:6-
acid. The tetracarboxylic acid which we obtained not only agreed in
every respect with the 2:3:4: 6-acid, but differed from the 2:3:4:5-
isomeride in giving no precipitate with zinc sulphate in neutral solu-
tion. The difficulty of removing inorganic matter from the tetra-
carboxylic acid which we obtained is characteristic of the 2:3:4:6-
acid.
The methylpyridinetricarboxylic acid from corydaline must there-
fore have one or other of the following formulae (Figs. 9 and 10) :
COjH
COjH
co,h(j^)S'^
co,h/\
CO,H^^>.
Fio. 9.
Fio. 10.
The position of one of the carboxyl groups must be adjacent to the
nitrogen atom, since it follows that, when the isoquinoline nucleus
is destroyed in the formation of methylpyridinetricarboxylic acid,
the carbon atom 1, next to the nitrogen atom, must have a carb-
oxyl group attached to it representing carbon atom 9, which is common
to the benzene and pyridine rings of the isoquinoline nucleus (see
Fig. 2).
The two remaining carboxyl groups must represent one of the rings
of the corydaline molecule which has been destroyed by oxidation and
must therefore occupy positions adjacent to one another.
The position of the methyl group is fixed by the following consider-
ations. It cannot occupy the position 4, because, in that case, the
only arrangement possible would be [CHg : (C02H)3 = 4:2:6:6]. This
acid is known, and is not identical with the acid under investigation.
The position 3 is likewise excluded, since, in that case, the tetracarb-
oxylic acid obtained on oxidation would be [ (COgH)^ = 2:3:4:5 or
2:3:5:6], having regard to the fact that two of the carboxyl radicles
represent a ring destroyed by oxidation, and must therefore be adjacent
to one another. By similar reasoning, position 5 is excluded ; the
methyl group must therefore occupy the position which is assigned to
it in the formula. It is shown earlier in this paper that the methyl-
tricarboxylic acid is probably [CH3 : (COjBQg = 2 : 3 : 4 : 6], but we have
no direct experimental evidence which enables us to decide between
this formula and [CH3 : (C0jH)8= 2:4:5:6].
Digitized by
Google
THE CONSTITUTION OF CORTDALINE. 165
Exatninaiian qfCarydUie Add, Oi,HjN(0'CHj),(C02H)g.
The analysis and description of this acid have abeady been published
(Dobbie and Marsden, Trans., 1897, 71, 657). Corydilic acid is obtained
along with m-hemipinic and 2-methylpyridinetricarbozylic acider when
corydio acid is oxidised with potassium permanganate at the boiling
point. From the former it is easily separated, but it is more difficult
than we at first supposed to free it entirely from the latter. Kepeated
recrystallisations are necessary to effect complete purification. This
explains why the specimens which we analysed gave results slightly
lower than the theoretical numbers. In addition to the reactions
already described for this acid, we have made the following observations.
Its aqueous solution gives no reaction with ferrous sulphate or ferric
chloride, and no precipitate with barium chloride, calcium chloride,
cadmium chloride, or copper acetate, even in presence of ammonia.
From alkaline solution, corydilic acid is precipitated by the addition of
excess of strong hydrochloric acid. If, however, the alkaline solution
is exactly neutralised with dilute hydrochloric acid, no precipitation
takes place, and a slight excess of hydrochloric acid may be added with-
ut causing the acid to separate. The solution so obtained has a faint
green colour, and on standing, sometimes deposits pale, greenish-
yellow crystals, which apparently consist of a hydrochloride of the
acid. The crystals are very unstable, and decompose on the addition
of water, leaving a residue of corydilic acid. Owing to its instability,
we were unable to get this substance in a fit condition for analysis.
Oxidation qf Corydilic Acid with Potaasium Permanganate,
Corydilic acid is very stable, but on heating for several hours with
potassium permanganate in alkaline solution it gradually undergoes
oxidation. The acid employed was carefully purified from every trace
of the methylpyridinetricarboxylic acid. About 6 grams of the pure
acid were oxidised in quantities of 2 grams at a time. After removal
of the manganese oxides^ the alkaline solution was concentrated and
precipitated with lead acetate. This precipitate, on decomposition
with sulphuretted hydrogen, yielded a mixture of acids, which, on separ-
Ation by fractional crystallisation, was found to consist of undecom-
posed corydilic acid, 9i»-hemipinic acid, and the 2-methylpyridinetri-
carboxylic acid. The two latter acids were compared with specimens
prepared directly from corydaline and found to agree in every respect.
It has already been shown that corydilic acid is tribasic, and that it
contains two methoxyl groups. The following formula explains its
Digitized by
Google
156 DOBBIE AND LAUDER: COBYDAUNE. PABT VII.
formation from corydic acid as well as all the facts connected with its
decomposition products:
CO,H
lOjH
OMeA/V*''
OMel
Fio. 11.
Oxidation of Corydic Acid with Potcusium Permanganate at the
Ordinary Temperature.
Corydic acid was suspended in cold water and about twice its weight
of potassium permanganate added in aqueous solution in small quanti<
ties at a time. The alkaline solution was precipitated with silver
nitrate and the precipitate decomposed with sulphuretted hydrogen in
the usual way. The filtrate, on evaporation, deposited a bright yellow
acid which, after purification by repeated recrystallisation from water,
melted at 212 — 215^. This acid is anhydrous, and differs from corydic
add in being more soluble in cold water and in giving a precipitate
with silver nitrate in neutral solution. It was dried at 100° and
analysed, with the following results :
0-2503 gave 06506 CO^ and 0-1207 H^O. 0 = 59-99 ; H = 5-36.
0-2086 „ 0-4621 CO, •„ 0-0990 H^O. C-60-42; H«6-27.
0-2748 „ 10-6 o.c. nitrogen at 16° and 758 mm. N-4-55.
CijH^^O^N requires C - 6018 ; H « 5-33 ; N = 439 per cent.
This acid is dibasic and forms both a normal and an acid silver salt.
Its precise relation to corydic acid is still under investigation.
We have limited our investigation of corydaline derivatives and de-
composition products to those substances which seemed most important
for the determination of the constitution of the alkaloid, as the ex-
pense entailed has been very heavy. For the same reason, our account
of some of the substances actually described is less complete than we
could have wished. We hope in a future paper to supplement the
information on some of the more important points which require
fuller elucidation.
We have to express our best thanks to the Society for the liberal
assistance granted to us from the Research Fund, and to Prof. W. H.
Perkin, jun., for kindly giving us specimens of the decomposition pro-
ducts of berberine for comparison with those of corydaline.
Univkbsity Collxqb or North Walks,
Banoor.
Digitized by
Google
THE RELATIONSHIP OP CORYDALINE TO BERBERINB. 157
XV. — The Relationship of Corydcdine to Berherine.
Berberidic Acid.
By Jambs J. Dobbie, M.A., D.Sc., and Alexander Laudeb, B.Sc.
Perkin (Trans., 1890, 57^ 992) has proposed the following alternative
formuln for berherine, expressing the opinion that (I) is the more
probable of the two :
i^NoMe i^\0Me
/I^JoMe ^yOMe
I. 11.
K
In the preceding paper, we have shown that the constitution of
corydaline can be represented by a formula similar to I, and assum-
ing the correctness of this formula for corydaline, that the absence
of compounds corresponding to berberal, C^oHj^O^N, berberilio acid,
GjoH^gOgN, tfec., from amongst the decomposition products of cory-
daline is explained. ''^
Whilst the absence of certain decomposition products is satisfactorily
accounted for, the similarity of the formulsB assigned to the two
^dkaloids suggested the possibility of obtaining from berherine an acid
corresponding to corydic acid, and, as a matter of fact, we found no
difficulty in preparing the expected acid by a method similar to that
used in the preparation of corydic add. For convenience of reference,
we shflkll provisionally term the substance so obtained berberidic acid.
Ten grains of berherine nitrate were suspended in two litres of dilute
nitric acid (1 in 20) and heated at the temperature of the water-bath
until completely dissolved. When the solution cooled, a small quantity
of the new acid was deposited as a yellow, crystalline precipitate. This
was filtered off, the solution neutralised with ammonia, concentrated,
and precipitated with silver nitrate. The silver precipitate was decom-
posed with sulphuretted hydrogen and the acid separated by fractional
crystallisation from a more soluble substance not yet examined, which
was formed along with it. In crystallising the acid, a considerable
amount of tarry matter 'Separated out. The acid was finally freed
from this and obtained in a pure state by dissolving in sodium hydr-
oxide and precipitating with hydrochloric acid. In later preparations,
* For f^irther oomparison of berberine with corydaline, see preceding paper.
Digitized by
Google
158 DOBBIE AND LAUDER : THE RELATIONSHIP OF
the purification was greatly facilitated by fractional precipitation with
silver nitrate, the first fraction carrying down most of the tar. The
subsequent fractions were light in colour and practically pure. The
yield of purified acid amounted to about 20 per cent, of the berberine
nitrate used. Berberidic acid crystallises from water in radiating tufts
of yellowish-brown, prismatic crystals, which have a pure yellow colour
when powdered. It contains no water of ci^ystallisation. When
heated in a capillary tube, it darkens at about 235° and remains
without further change, so far as can be seen^ until 285°, when it
melts with decomposition. It was dried at 100° and analysed, with
the following results :
0-2637 gave 0-5925 CO^ and 0-0866 H^O. 0 = 61-28 ; H = 3-61.
0;2831 „ 0-6344 COg „ 0 0917 11^0. C«61-12; H = 3-59.
0-3243 „ 13-0 c.c. nitrogen at 16° and 761-5 mm. N-4-75.
0-2808 „ 11-0 C.C. „ 13° „ 751 mm. N-4-63.
CijHiiOjN requires 0 = 61-34; H = 3-51 ; N = 4-47 percent.
Berberidic acid is insoluble in cold and only sparingly soluble in
boiling water. It is very sparingly soluble in boiling alcohol and in-
soluble in ether or chloroform. It dissolves easily in sodium hydipxide
to a dark blood-red solution, from which it is precipitated by hydro-
chloric acid.
Berberidic acid is dibasic. All its salts, with the exception of the
two silver salts, appear to be soluble. The normal silver salt is
obtained by precipitating a solution of the acid, which has been
neutralised with ammonia, with silver nitrate. A curdy, yellow pre-
cipitate is obtained, which darkens on exposure to light. This salt was
repeatedly prepared and analysed without exact results being obtained,
owing, apparently, to admixture with the acid salt.
The acid silver salt is prepared by precipitating an aqueous^ solution
of the acid with silver nitrate. The curdy precipitate so obtained is
filtered, washed, and purified by repeated recrystallisation from water.
It is finally obtained in stellate clusters of beautiful, yellowish-brown
needles. On heating, it decomposes suddenly with evolution of thick,
brown vapours. After being dried at 100°, it was analysed with the
following results :
0-2470 gave 0-0828 AgOl. Ag = 25 23.
0-2616 „ 00655 AgOl. Ag = 25-04.
• OigH^^OjNAg requires Ag = 25-71 per cent.
When berberidic acid is heated with concentrated hydrogen iodide
solution, no methyl iodide is evolved, a fact which proves that in the
formation of this add the ring of the berberine molecale containing
the methoxyl groups is destroyed.
Digitized by
Google
CORYDALINE TO BERBERIKE. BBRBERIDIO ACID. 159
OxicUUian of Berbmdie Acid wUh Poiaanum Permanganaie. — Five
grams of berberidic acid were boiled with a dilute solution of per-
manganate until the permanganate was no longer reduced. The
solution was filtered from the manganese oxide, concentrated, and pre-
cipitated with silver nitrate. The silver precipitate \^as decomposed
with solphoretted hydrogen and the filtrate from the silver sulphide
evaporated to dryness. The residue was repeatedly exhausted with
hot absolute alcohol, in which a considerable part of it dissolved. The
portion of the residue insoluble in hot alcohol dissolved readily in
boiling water, from which it separated on cooling in prismatic crystals.
The acid so obtained was decolorised by boiling with charcoal and
purified by repeated recrystallisation from water. It melted at 235^
or 242% according to the rate of heating. It dissolved with difficulty
in cold, but was readily soluble in boiling, water ; it was insoluble in
ether or chloroform. Its aqueous solution gave an orange-red colora-
tion with ferrous sulphate. The acid agrees in every particular with
berberonic acid, qq-o] J ^ > which was obtained by Weidel (5«r.,
1879, 12, 410) by the direct oxidation of berberine with concentrated
nitric acid. The melting point of berberonic acid is variously given at
238—242°.
The normal silver salt, which is almost insoluble in water, was pre-
pared by precipitating a solution of the acid, previously neutralised
with ammonia, with silver nitrate. After being dried at 100°, it was
analysed with the following result :
0-2978 gave 0-1798 Ag. Ag - 60-38.
OgHjOjNAgj requires Ag = 60*88 per cent.
The presence of hydrastic acid amongst the decomposition products
of berberidic acid has not yet been proved. By dissolving berberidic
acid in potassium carbonate and oxidising it with potassium perman-
ganate at the ordinary temperature, a small quantity of a substance
was obtained as a scum on the surface of the strongly alkaline solu-
tion. From its insolubility in potash, we suspected that this sub-
stance might be oi-aminoethylpiperonyl carboxylicanhydride, which
is insoluble in alkaline solutions. On examination, we found that
it agreed in every particular with the anhydride in its neutral
reaction, solubiHty, peculiar mode of crystallisation, and behaviour
with mercuric chloride. As the amount of substance obtained was
too small to admit of complete purification, the melting point observed
was slightly lower than that given by Ferkin.
Berberidic acid clearly bears the same relation to berberine that
Digitized by
Google
160 FORSTER AND MICKLETHWAIT : STUDIES IN THE
corydic acid bears to dehydrocorydaline. Since it contains no methoxyl
groups, it follows that the ring of the berberine molecule which yields
hemipinic acid is destroyed in its formation. The occurrence of
ot>-aminoethylpiperonylcarbozylic anhydride and berberonic acid amongst
its oxidation products proves that it contains the three remaining
rings and that its constitution may therefore be expressed by the
formula :
CO,H
<»
By oxidising berberidic acid with potassium permanganate at the
ordinary temperature, a yellow derivative is obtained like that obtained
from corydic acid by similar treatment.
As berberine, unlike corydaline, can be obtained at comparatively
low cost, we have undertaken a more thorough investigation of ber-
beridic acid, which we hope will throw further light on the constitu-
tion of both alkaloids, and espidcially on the relation between the
constitution and colour of some of their derivatives.
University Collkob op North Wales,
Bakoor.
XVL — Studies in the Camphane Series. Part VI.
Stereoisomeric Halogen Derivatives of a-Benzoyl-
camphor.
By Martin Onslow Forstkb and Frances M. G. Mioklkthwait.
In accordance with its unsaturated character, l-hydroxy-2-benzoyl-
camphene, the enolic form of a-benzoylcamphor, immediately decolorises
a solution of bromine in an indifferent solvent. At the same time
hydrogen bromide is eliminated, and. if one molecular proportion of the
halogen is employed, the crystalline residue obtained on evaporating
the liquid has the empirical formula of benzoylbromocamphor. There
is no difficulty, however, in resolving this product into two distinct
substances which, although isomeric and nearly alike in chemical
behaviour, are widely different in physical properties. The more
soluble constituent of the mixture crystallises from alcohol in six-
sided prisms, >elts at 114°, has [o]„ -10-0° in benzene, and [a]i>
Digitized by
Google
CAMPHANE SERIES. PABT YI. 161
+ 10*3° in chloroform; the isomeride is deposited from alcohol in
rectangular plates, melts at 214°, has [ajo -63*2° in benzene, and
[a]i> - 19-3° in chloroform.
The method of preparation, the fact that neither substance dissolves
in alkalis, and the transformation of both isomerides into l-hydroxy-2-
benzoylcamphene by the action of alcoholic potash, are droumstances
which point to the condusion that the compounds in questionare a-bromo-
deriyatives of a-benzoylcamphor, and that their physical differences are
the result of a difference in configuration. Theoretical considerations,
moreover, led us to expect the formation of two derivatiyes displaying
isomerism of the cuirans-type, as indicated by the following formula
(compare Lowry, Trans., 1898, 78, 572) :
It is evident that a similar explanation would account also for the
production of two isomerides from enolic benzoylcamphor in the event
of that substance being shown to have the alternative formula, namely,
that of phenylhydroxymethylenecamphor, CgHj^-OC^ * *, a
possibility which is not yet excluded.
Several instances of this form of isomerism in the camphor series
have now been established. Leaving aside the somewhat uncertain
cases of the monohalogen derivatives of camphor, there remain the
isomeric chlorobromocamphors, chloronitrocamphors, and bromonitro-
camphors investigated by Lowry (Trans., 1898, 73, 669 and 986),
and the benzylbromocamphors described by Haller and Minguin
(Compi. rend,, 1901, 133, 79). Up to a certain point, the case of the
benzoylbromocamphors resembles those of the four derivatives men-
tioned, the difference between the two forms being, however, greater
than has been observed hitherto ; but an important feature distinguishes
it from those already described.
Li deaUng with the isomeric chlorobromocamphors, Lowry records
unsuccessful attempts to convert a-chloro-a-bromocamphor into
a-chloro-a'-bromocamphor by the action of heat and of acids (Trans.,
1898, 73, 581). Neither in his subsequent communication nor in the
paper of Haller and Minguin {loc, cit,) is it stated that the chloronitro-
camphor, bromonitrocamphor, or benzylbromocamphor of lower melting
point can be transformed into the corresponding isomeride, and it is
VOL. LXXXI. H
Digitized by
Google
162 FOBSTER AND HICKLKTHWAIT : STUDIES IN THE
probable therefore that the change cannot be effected or it would
have been observed. It is in this respect that the benzoylbromo-
camphors differ from the foregoing disubstituted a-derivatives, for the
compound having the lower melting point is readily converted into
the isomeride by the action of hydrogen bromide.
This transformation is the first recorded instance of stereoiaomeric
change on the part of a disubstituted derivative of camphor in which
both substituents occupy the a-position. It has therefore a direct
bearing on the explanation given by Marsh in accounting for the
unstable character of a specimen of bromocamphor which is described
as melting at 61^ (Trans., 1890, 57, 832 ; compare also Lowry, Trans.,
1898, 73, 672). The validity of the explanation in question depends on
the formation of an intermediate isomeride, which represents Hie enolic
modification of the material transformed, and, in the case discussed by
CBr
Marsh, would have the formula CgHi4-«^^^p-.. The experiments
described in this paper have led us to consider this explanation
improbable. In the first place, it cannot be applied to derivatives of
camphor of the class to which benzoylbromocamphor belongs, and
secondly, there seems to be no need for any explanation so complex,
several cases of stereochemical transformation being known in which
there is no room for any structural change to occur.
In general features, the benzoylchlorocamphors resemble the
corresponding bromo-derivatives very closely, the two modifications
which melt at 88^ and 219° displaying similarity as regards solubility
and crystalline form when compared respectively with the bromo-
derivatives melting at 114° and 214°. It is noteworthy, however, that
we have been hitherto unable to convert one isomeride into the other.
Moreover, the action of sodium hypochlorite on enolic benzoylcamphor
gives rise to a preponderating quantity of the benzoylchlprocamphor
of the lower melting point, whilst the benzoylbromocamphor of the
higher melting point is the almost exclusive product when potassium
hypobromite is employed ; bromine dissolved in chloroform yields a
mixture of the isomerides in nearly equal parts, whilst bromine and
glacial acetic acid containing sodium acetate afford chiefly the benzoyl-
bromocamphor of lower melting point.
In describing the stereoisomeric halogen derivatives of ci-benzoyl-
camphor, we have adopted the convention suggested by Lowry {loc, eit.),
so that the nomenclature of the new derivatives may be uniform with
that of the unsymmetrical di-derivatives already prepared. Assuming
that benzoylcamphor, with [a]]> + 137*5° in alcohol, is an a-derivative,
it will be noticed that the optical influence of the benzoyl radicle
exceeds that of the chlorine atom, since a-cblorocamphor has [a]i> + 96°
in the same solvent; it may be concluded therefore that the di-
Digitized by
Google
CAHPHAN1E SERIES. PART YI. 163
derivatiye which has a specific rotatory power least remored from that
of camphor itself, is that which contains the benzoyl radicle in the
orpofiition. This modification is the one which melts at 219°, having
[a jo +26*2° in chloroform, and is accordingly termed a-benzoyl-a'-
chlorocamphor. In the case of the bromo-derivatires, it is not so easy
to decide which isomeride contains the benzoyl radicle in the orposition,
because the recorded values for the specific rotatory power of benzoyl-
camphor and of bromocamphor in alcohol are practically identical.
There is reason to believe, however, that the specific rotatory power
of benzoylcamphor at the moment of dissolution in alcohol is lower
than ]o + 137*5°, because the substance, dissolving somewhat slowly
in the cold solvent, suffers partial conversion into the enolic modifi-
cation, with [a]i> + 262°, before it can be examined in the polarimeter ;
chloroform, however, which dissolves the substance very readUy, yields
a solution having [ajo + 126°, and it is therefore probable that the
optioal influence of l^e benzoyl radicle is less powerful than that of
the bromine atom, because a-bromocamphor has [ajo +135°. If this
is the case, the modification which melts at 114° and has [a]i> + 10-3°
in chloroform must be called a-benzoyl-a-bromocamphor, whilst
the isomeride melting at 214°, having a specific rotatory power more
remote from that of camphor, must be regarded as having the bromine
atom in the a -position ;
[«]p m. p.
Camphor + 42° ( alcohol ) —
a-Ohlorocamphor ^. + 96 ( „ ) —
a-Bromocamphor +135 ( . „ ) —
a-Benzoylcamphor +125 (chloroform) —
a-Benzoyi-a-ehlorocamphor... > 28 ( „ ) 88°
a-BeuzoyW-chlorocamphor ... + 26 ( „ ) 219
a-Benzoyl-a-bromocamphor... + 10 ( „ ) 114
a-BenzoyW- bromocamphor... - 19 ( „ ) 214
From this table, it will be noticed that the benzoylchlorocamphor and
benzoylbromocamphor supposed to contain the halogen in the a-position
both melt at the lower temperature, whilst the less readily fusible
modifications are assumed to have the halogen substituted in the
a'-position.
EZPEBtMBNTAL.
» fi z /^TT ^CBrCO'C-Hg
.'BenzaiflbromocampkarSf O^a^^K^l^ • ® .
aa-
Twenty grams of l-hydrozy-2-benzoylcamphene were dissolved in
chloroform and cooled in melting ice. A cold solution of 12*4 grams
of bromine in chloroform was then added in small quantities at a time,
M 2
Digitized by VjOOQIC
164 FORSTER AND HICKI4ETHWAIT : STUDIES IK THE
and the pale red liquid, from which hydrogen bromide was being
evolved, transferred to a basin and allowed to evaporate spontaneously.
A previous experiment having shown that two compounds are produced
by this means, the crystalline residue was divided into four fractions
by extracting it successively with quantities of hot alcohol amounting
to 100 c.c. (twice), 200 cc, and 300 ac, and allowing the solutions to
cool.
Fraction I, weighing 6 grams, consisted of thin, transparent needles
melting somewhat indefinitely at 108 — 110^ ; a 2 per cent, solution in
benzene gave [a]o -15*4^; and in chloroform [a]i> +7'9^. A large
proportion being readily soluble in warm, light petroleum (b. p. 60 — 90^),
the whole fraction was extracted with this solvent ; the solution de-
posited large, thin, six-sided prisms melting at 114° and giving [a]]>
" 10*0° in benzene and [a]^ + 10*3° in chloroform. Reorystallisation
from light petroleum did not change the specific rotatory power.
Fraction EE, weighing 8 grams, consisted chiefly of needles, and
melted somewhat indefinitely at 109 — 111°; a 2 per cent, solution in
benzene gave [ajo - 19*6°, and in chloroform [a]j> + 2*9°.
Fraction III, weighing 2 grams, consisted of thin, rectangular plates,
beginning to shrink and to change colour at about 186° and melting
at 210°; a 2 per cent, solution in benzene gave [ajn —62*3°, and in
chloroform [a ]]> -18*6°.
Fraction lY, weighing 3 grams, consisted of thin, rectangular plates,
beginning to shrink and to change colour at about 190°, and melting at
214° ; a 2 per cent, solution in benzene gave [a]i> - 63*2°, and in chloro-
form [aji) " 19*3°. The properties of this fraction were not altered
by recrystallisation from boiling alcohol.
a-Benzoyl-a-bromooamphor is most conveniently prepared by dissolving
l-hydroxy-2-benzoylcamphene in glacial acetic acid containing Ij^ mols.
of sodium acetate and adding 1 moL of bromine dissolved in glacial
acetic add ; the white precipitate obtained on pouring this liquid int<o
water is then collected, washed, dried, and crystallised from light
petroleum. It is readily soluble in chloroform, benzene, alcohol, or
light petroleum, crystallising from the last-named in large, transparent,
six-sided prisms, and from alcohol in slender needles having the same
crystalline form:
0*1992 gave 0*1101 AgBr. Br<-23*62.
Oj-^HijOgBr requires Br = 23*88 per cent.
The substance melts at 114°, but fusion is not complete until the
temperature is raised to about 180°. A solution containing 0*6 gram
in 26 cc. of benzene at 21° gave aj, —24' in a 2 dcm. tube, whence
the specific rotatory power [a]i> — 10*0° ; 0*6029 dissolved in 26 cc. of
chloroform at 21° gave a^ + 26', corresponding to [a]i> -I- 10*3°.
Digitized by
Google
'CAMPHANB SERIES. PART VI. 165
a-Benzoyl-a-bromoeamfhor was obtained in the following manner.
One hundred grains of bromine were dissolved in an ice-cold aqueous
solution containing 150 grams of potassium hydroxide, and slowly
added to 20 grams of l-hydroxy-2-benzoylcamphene dissolved in dilute
potash. The sticky solid which inmiediately separated soon hardened,
and after an interval of 12 hours was collected, washed, and recrystal-
lised from boiling alcohol. The yield of aa-benzoylbromocamphor
obtained by this method is quantitative, and the product consists chiefly
of the variety of high melting point. It dissolves very readily in
chloroform, but only sparingly in cold alcohol or benzene, and is almost
insoluble in boiling light petroleum ; it crystallises from hot alcohol in
transparent, rectangular plates, begins to shrink and to change colour
at about 190^, and melts at 214^ to a pale brown liquid which evolves
gas. The substance may be crystallised from concentrated nitric acid
without undergoing change :
nuioui^ unaergomg cnange :
01353 gave 00756 AgBr. Br = 2377.
CifHi^OjBr requires Br » 23*88
per cent.
A solution containing 0*5015 gram in 25 o.c. of benzene at 21°
gave od -2^8' in a 2 dcm. tube, whence the specific rotatory power
[a]D -53*2°; 0*6451 gram dissolved in 25 c.c. of chloroform at 2P
gave a]> - 1^0', corresponding to [a]]> - 19*3°.
AMon rf Alcohdto Potassium Hydroxide an aorJBenzoylbromocamphor,
— A specimen of a-benzoyW-bromocamphor which melted at 210° and
gave [ajo — 18'5° in chloroform, was heated during 4 hours in a reflux
apparatus with potassium hydroxide (2 mols. ) dissolved in alcohoL The
liquid soon became dark brown, and on evaporation yielded a residue
which dissolved completely in water. A current of well washed carbon
dioxide was then passed into the aqueous solution until no further pre-
cipitation occurred, and the product, after crystallisation from alcohol,
I obtained in the pink octahedra characteristic of l-hydroxy-2-benzoyl-
uphene.
The same compound was obtained by reducing a -benzoyI*a-bromo-
camphor with alcoholic potassium hydroxide.
• AcUon qf Bromine on l'Benzoxy'24>enzoyleamphene,^^'Wheia a solution
of l-benzoxy-2-benzoylcamphene in chloroform is treated with bromine,
the colour of the halogen is not immediately destroyed, but after an
interval, action is found to have taken place.
Ten grams of the dibenzoyl derivative were dissolved in 100 cc of
chloroform and enclosed in a stoppered bottle with 4*4 grams (I mol.)
of bromine. After 24 hours, the colour of the halogen had almost
disappeared. On allowing the liquid to evaporate, a considerable
quantity of hydrogen bromide was liberated, and a crystalline residue
was obtained having the odour of ethyl benzoate. The solid product.
Digitized by
Google
166 FORSTER AND MIOKLBTHWAIT : STUDIES IH TSE
weighing 9 grams, was exhausted with 50 c.c. of hot alcohol, and the
solution deposited crystals melting at 1 10 — 115° and giving [ajo + 4*1°
in chloroform :
0*1598 gave 00884 AgBr. Br » 23*64.
CiyHijOjBr requires Br » 23*88 per cent.
The substance was evidently a mixture of the two aa-benzoylbromo-
camphors, and by repeated crystallisation from alcohol, a specimen of
the modification of higher melting point was obtained, giving [aJD
- 19*5° in chloroform.
Conversion of One Isomeride into the Other. — During the first at-
tempt to separate the isomerides from one another by fractional crys-
tallisation, a most unexpected change of the variety of lower melting
point took place. A specimen of that substance, which had been re-
crystallised twice from alcohol without altering the melting point,
melted at 111 — 112° and gave [a]i> —29*0° in benzene; it was dissolved
in hot alcohol, which on cooling deposited lustrous plates melting at
201 — 204°, and giving [ajn -51*0° in benzene. Although we have
not succeeded in reproducing the conditions of this experiment^ the
transformation of one modification into the other can be effected by
the agency of hydrogen bromide. A specimen of a -benzoyl-a-bromocam-
phor melting at 109 — 111° and giving [a ]d +2*9° in chloroform was
finely powdered and placed in a stoppered bottle with sufficient fuming
hydrobromic acid to convert it into a thin paste. The following morn-
ing, water was added and the solid product filtered and washed. The
substance, when dried in the desiccator, melted at about 200° and gave
[^Id -18*6° in chloroform, and when recrystallised from alcohol
yielded the lustrous plates characteristic of a-benzoyl-a'-bromocamphor.
Action of Bromine on a-Benzoyleamphor,
In. describing the a-substituted halogen di-derivatives of camphor,
Lowry (Trans., 1898, 73, 572) suggests that "the production of stereo-
isomerio di-derivatives is most readily explained by supposing that the
action of the halogen involves addition to the enolic form of the mono-
derivative." This explanation is a very probable one, and the follow-
ing experiment appears to give it direct support.
A specimen of ketonic a-bensoylcamphor, giving only a faint oolora-
tion with ferric chloride, was dissolved in cold glacial acetic acid
containing sodium acetate (1^ mols.) \ to this liquid, a solution of
bromine (1 mol.) in glacial acetic acid was added, when it was observed
that the colour of the halogen was immediately destroyed. Although
it must be remembered that a small proportion of the benzoyloamphor
is enolised by the solvent, it is still fair to say that the behaviour of
o^benzoylcamphor towards bromine exactly resembles that of the
Digitized by
Google
CAMPHANE SERIES. PART YI. 167
unsaturated enolic isomeride, and it is noteworthy that the identity
extends to the product of the change, which gives rise to a -benzoyl-a-
bromocamphor in both cases.
aa-Benzoylehlorocampkors, CgHi^-^JL^ * *,
Having found that potassium hypobromite converts l-hydroxy-2-
benzoylcamphene into a mixture of the benzoylbromocamphors, we
employed the corresponding method in preparing the benzoylchloro-
camphors in preference to treating the hydroxy-compound with the
free halogen. Ten grams were dissolved in dilute aqueous potassium
hydroxide, cooled with fragments of ice, and treated with 200 c.c. of a
solution of sodium hypochlorite containing 30 grams of available
chlorine per litre. A pink, dough-like solid separated and rapidly
became hard. After an interval of several hours, the product was
collected, washed, and extracted with 100 c.c. of boiling alcohol, thus
dividing the substance into two portions, of which the more readily
soluble melted somewhat indefinitely at 86 — 87^ and gave [a]^ - 20*6^
in chloroform, whilst the residual fraction melted at 219° and gave
[at +26-0°.
a-Benzayl-a-chlorooamphorf obtained by recrystallising the more
soluble fraction from alcohol and then from light petroleum, crystal-
lises from each solvent in prisms and melts at 88°:
01284 gave 0-0627 AgCl. CI = 1208.
C17H19O3CI requires Gl» 12*22 per cent.
It is readily soluble in alcohol and very freely so in chloroform, but
dissdlvefl only sparingly in light petroleum. A solution containing
0*4186 gram in 26 c.c. of chloroform at 21° gave a^ —66' in a
2 dcm. tube, whence the specific rotatory power [a]i> —27*9°:
a-Benzoyl^'-Moroeamphor remains after the mixture of the two
isomerides has been exhausted with a small quantity of hot alcohol ;
it crystallises from that solvent in plates resembling the corresponding
bromo-derivative and melts at 219°:
01324 gave 0*0668 AgOl. CI » 12*29.
(\^llyP^Q\ requires CI » 12*22 per cent.
It is freely soluble in chloroform, but dissolves only sparingly in
alcohol and is insoluble in light petroleum. A solution containing
0*8973 gram in 26 c.c. of chloroform at 21° gave a^ -f- 60' in a
2 dcm. tube, whence the specific rotatory power [a]D -1-26*2°
ROTAL COLLE6B OF SCIEKOE, LOTOON.
SOXTTH KZHSINOTON, S.W.
IlO
Digitized by
Google
168 DrxoN : the action of phosphorus
XVII.— The Action of Phosphoi^us Trithiocyanate on
Alcohol.
By Augustus Edwabd Dixon, M.D.
In a prelixaiDary note {J. pr. Cham., 1872, [ii], 7, 474), Lossner
records that he has obtained (1) by the actioo of phosphorus tri-
chloride on potassium thiocyanate in alcoholic solution, a substance crys-
tallising in fine needles, whose analysis leads to the empirical formula
CgHjgON^S^ ; and (2) from benzoyl chloride and alcoholic potassium
thiocyanate, a compound, CgH^ONS. No analytical results are given in
this note, which is very brief ; but the interaction in which benzoyl
chloride takes part is dealt with by Lossner at considerable length in
a paper published a couple of years later (ibid., 1874, [ii], 10, 237) ;
the compound CgH^ONS now appears as O^qHi^O^NS, that is, benzoyl
thiocyanate plus a mol. of ethyl alcohol, and is r^arded by him as
'benzoylethylozysulphocarbamic acid,' PhOO*NEt*00*SH; a paper
dealing wiUi the constitution of this substance and of certain of
its deriyatiyes has lately been published (Dizon, Trana, 1899, 76,
376).
No reference is made in L<>ssner's second communication to the
compound OgH^gON^S^ ; nor, in fact, so far as the author can ascertain^
is any description of it to be found in chemical literature. It is not easy
to understand how a substance of this composition could be formed
out of the materials used, unless through the occurrence of some profound
decomposition; with the view of ascertaining whether such a change
really took place, and more particularly since the interaction to be ex-
pected of these substances appeared to belong to the class of interactions
recently studied by the writer, in which phosphorus and phosphoryl
' thiocyanates' take part (Trans., 1901, 79, 641), it was decided to re-
examine Lossner's reaction.
Before doing so, and incidentally to the incipient study just
mentioned, some experiments were carried out in order to learn
whether "phosphorus thiocyanate," P(SCN)g or P(NCS)s, would unite
directly with ethyl alcohol so as to afford a phosphoretted thio-
urethane, thus :
P(NOS)g + 3CjHa-0H = P(NH-CS-OOjHj)g ;
although, in view of the great ease with which both this and the
corresponding phosphoryl derivatiye undergo hydrolysis, it scarcely
seemed probable.
The phosphorus compound was prepared as already described {loe.
eit.f p. 646), about 13 grams of phosphorus trichloride being used in
each preparation: on treating the benzene solution with absolute
Digitized by
Google
TRITHIOGYANATE ON ALCOHOL. 169
alcohol, there was marked evidence of chemical interaction^ the
temperature rising in three successive experiments in which it was
measured, by 47^ 46% and 69° respectively, whilst free thioojanic
acid was evolved. On concentrating the mixture at the ordinary
temperature, a yellow, cfystalline solid was deposited; the mother
liquor formed a clear brown syrup, intensely acid, reacting freely for
thiocjanic acid and phosphorus, and soon beginning to decompose
with evolution of mercaptan.
The solid product occurred in limited quantity, not more than a
gram, at most, being obtained for every 13 grams of trichloride used ;
it was insoluble in benzene, sparingly soluble in boiling water, and
moderately sd in hot alcohol, but did not crystaUise well 'from the
latter solvent. When recrystallised from much boiling water, it was
obtained in yellow, flexible, hair-like needles (on one occasion several
inches long and closely resembling Spirogyra in outward appearance) :
they began to darken and change at about 230% but were not melted
at 260*'.
The substance contains no phosphorus, and hence is not the
desired phosphorus trithiotriurethane. It is desulphurised by heating
in alcoholic solution with ammoniacal silver nitrate, or with alkaline
lead tartrate ; its aqueous solution is somewhat acid to litmus and
gives with lead acetate a bright yellow precipitate. Ferric chloride
yields practically no colour reaction, either when added to the aqueous
solution or to the mixture produced by first dissolving the solid in
warm alkali hydroxide and then acidifying the solution with hydro-
chloric acid. The substance dissolves readily in potassium cyanide
solution, and the resultant liquid, if acidified and treated with ferric
chloride, now gives the intense blood-red thiocyanic reaction.
From the properties just described, there could be little doubt
that the substance was nothing more than wopersulphocyanic acid,
C2H2N3S3, and the results of analysis showed this to be the case :
S found, 64-3 ; N found, 18*9 ;
CsH^NjSs requires S » 64 ; N » 18*7 per cent
The mechanism whereby this substance comes to be formed is probably
as follows : the '< phosphorus thiooyanate '' is decomposed in part by
the alcohol, yielding free thiocyanic acid :
P(S0N)3 -I- SOjHg-OH = P(0-0jHj)3 -h 3HSCN ;
whilst another portion, in like manner, yields phosphorous acid :
under the influence of this mineral acid, the former could afford
wopersulphocyanic acid, thus :
3HS0N = CjH jN^S, -I- HON.
Save the ifiopersulphocyanic acid, no other solid product was found ;
Digitized by
Google
170 ACTION OF PHOSPHORUS TRITHIOCYANATE ON ALCOHOL.
consequently, if the phosphoras trithiotriurethane is formed, or, at all
events, oontinaes to exist, under the above conditions, it must be as
one of the constituents of the acid, syrupy mother liquor, but the
foul smell of the latter rendered it so unpleasant to work with that it
was not examined further. However, as phosphorus trithiotriurethane,
if capable of existence under ordinary circumstances, would probably
be a solid substance more or less easily decomposable by moisture, it
is doubtful whether it could have been extracted from the liquor, even
if present.
Aj9 regards the interaction between alcoholic potassium thiocyanate
■and phosphorus trichloride, there was scarcely any reason to anticipate
that it would run a course materially different from that between phos-
phorus trithiocyanate and alcohol ; however, the experiment was tried,
with the following result.
On dropping phosphorus trichloride into a saturated solution of
potassium thiocyanate in 99*5 per cent, alcohol, violent action occurred,
.and potassium dbloride was precipitated ; on filtering this off and concen-
trating the filtrate by slow evaporation, thiocyanic acid escaped, and
yellow, crystalline material [separated in an oily, very acid, liquid ; the
former, when recrystallised from boiling water, proved to be identical
with the solid obtained from phosphorus thiocyanate and alcohol,
namely, wopersulphocyanic acid. In this case, as the liberated thio-
x^anic acid is in contact with much free hydrochloric acid proceeding
from the interaction between the phosphorus haloid and the alcohol, it
is a simple matter to account for the production of Mopersulphocyanio
acid. Aj9 in the preceding case, the quantity of this acid bears but a
small proportion to the amount of phosphorus chloride used. So far,
the writer has failed to identify any other substance in the solid pro-
duct, yet Lossner, strange to say, does not mention the occurrence of
Mopersulphocyanic acid at all.
It would seem, therefore, either that the interaction must have pro-
ceeded on different lines when conducted by this chemist, or else that,
through some accident, he must have attributed to Mopersulphocyanic
add, C^HsNgSg, the formula CgH^gON^S^. How this could happen it
is not very easy to see, considering that the percentages of sulphur are
64 and 40*8, respectively. It is conceivable, however, that some un-
suspected cause of error may have temporarily crept into his analytical
practice, more especially bearing in mind that his benzoyl chloride
product, above mentioned, which was stated in the preliminary note to
have, according to the results of analysis, the formula CgHgONS, turns
out to be really C^^B-jfi^l^S ; here the theoretical results are by no
means so widely divergent as in the preceding case, but still the figures
differ by nearly 4 per cent, for the sulphur, 3 per cent, for the nitrogen,
4ind so on«
Digitized by
Google
CONSTITUTION OP BSNZENEAZO-a-NAPHTHOL. 171
In the hope of possibly obtaining phosphorus tribensjltrithioearb-
mnate, a cold, somewhat dilute solution of " phosphorus thiooyanate '*
in benzene was treated with benzyl alcohol. Interaotion ocourred at
once, the temperature of the mixture rising by about 30^ ; but after
driving off the solvent and allowing the residue to stand, a mere trace
of white, solid matter was deposited, the amount being too small to per-
mit of identification. It crystallised well in white prisms from boiling
water, volatilised completely, on heating, without preliminary fusion,
gave no ammonia when heated with alkali, contained no phosphorus,
gave no colour reaction with potassium cyanide, hydrochloric acid, and
ferric chloride, and consequently was neither wopersulphocyanic acid
nor phosphorus tribenzyltrithiocarbamate. The mother liquor was
almost completely volatile in a current of steam; the distillate, a
yellowish oil, consisted partly of unchanged benzyl alcohol, and partly
of an unpleasant smelling oil which contained sulphur but no phos-
phorus, the latter being wholly retained in the trifling residue of the
steam distillation.
Chkmical Depabtkent,
Queen's Colleob, Cork.
XVIII. — The Relationship between the Orientation of Sub-
stituents in and the Constitution of Benzeneazo-a-
naphthol.
John Theooobb Hewitt and Saxubl James Manson Auld.
The question of the constitution of the ozyazo-compounds has aroused
a considerable amount of discussion, and in order to obtain further
information on this point, one of the authors of the present commun-
ication has, in conjunction with several of his pupils, made experi-
ments on the substitution derivatives of these substances. In all cases
so far studied, the results have given an unqualified support to the
oxyazo-formula, the phenolic nucleus always being first attacked by
dilute nitric acid or bromine in presence of an excess of sodium
acetate. The appearance of a communication by Mohlau and Kegel
{B&r.t 1900, S8, 2858), in which they ascribed a tautomeric formula to
benxeneazo-a-naphthol, Tendered necessary the further investigation
of the action of substituting agents on the benzeneazonaphthols. The
results obtained in the case of the azo-derivatives of j9>naphthol are
reserved for a future communication.
Mbhlau and K^el found that p^uinones and their derivatives
Digitized by
Google
172 HEWITT AND AULD: CONSTITUTION OF
generally reacted with benzhydrol and Michler's hydrol (tetramethyl-
diaminobenzhydrol) to form compounds of the type [R^O^H^ or
C,H,.N(CH,)J:
O
Hb^^ybn '
and extending the reaction to the so-called benzeneazo-a-naphthol
obtained substances in which the hydrol had behared as if the azo-com-
pound were quinonoid in type. Had a strong acid been present, such
a reaction would not have been surprising ; the condensation was, how-
ever, carried out in the absence of such a compound. Moreover, the
complicated azo-derivatives so obtained behaved, on acetylation, as
quinone-hydrazones, the acetyl group attaching itself to a nitrogen
atom. By the complete reduction of the acetyl derivative of benzene-
azotetramethyldiaminobenzhydryl-a-naphthol, Mohlau and Kegel ob-
tained acetanilide but could detect no aniline ; from these results, they
concluded that benzeneazo-a-naphthol, as well as the condensation
product with Michler's hydrol, had the constitution of quinone-
hydrazones. The condensation was, however, not incompatible with
the presence of both forms in equilibrium in solution, whilst the course
of the acetylation of the condensation product might be explained in a
similar way in conjunction with the undoubted steric hindrance which
might be experienced in the case of acetylating an ortho-substituted
a-naphthoL We therefore resolved to re-examine the acetylation of
benzeneazo-a-naphthol, and further to study the action of substituting
agents on the azo-naphthol itself. It may be mentioned here that the
results of all experiments made with nitric acid on benzeneazo-a-
naphthol were thoroughly unsatisfactory ; either reaction did not take
place or only tarry products were obtained.
Reduciian of Benzensazo^-naphthyl AceicUe.
Benzeneazo-a-naphthol was prepared by Witt and Dedichen's method
{Ber,, 1897, 30, 2657), and acetylated by boiling in a reflux apparatus
with excess of acetic anhydride and fused sodium acetate. The melt-
ing point of the product (128^) agreed with that given by Zincke and
Bindewald {Ber,^ 1884, 17, 3030). The complete reduction of this sub-
stance was effected in cold alcoholic solution, so that any possibility of
one or other product becoming acetylated during the process and thus
leading to erroneous conclusions might be obviated. Two grams of the
acetyl derivative were dissolved in 100 c.a of absolute alcohol and
treated with 5 c.c. of concentrated sulphuric acid mixed with
10 c.c. of alcohol. Zinc dust was now added and the solution well
Digitized by
Google
BENZEN£AZO-a-NAPHTHOL. 173
shaken until entirely colourless. The excess of zinc dust was re-
moved by filtration and the filtrate diluted with water, rendered alka-
line with sodium carbonate, and then twice extracted with ether. The
ethereal eactracts were united, the excess of ether evaporated, and the
residue distilled in a current of steam. The presence of aniline in the
distiUate was confirmed by its conversion into tribromoaniline. In
one experiment, the weight of tribromoaniline obtained was prac-
tically equal to that of the benzeneaionaphthyl acetate employed.
After the steam distillation, the residue in the flask was examined in
order to isolate the other product of fission ; the acetoxy-a-naphthylamine
eould not, however, be obtained in a crystalline form.
By partial reduction of benzeneazo-a-naphthyl acetate, a hydrazo-
compound is obtained, which, from its insolubility in dilute alkali,
evidently does not contain a free hydroxyl group. To obtain this
substance, 1 gram of benzeneazo-a-naphthyl acetate was dissolved in
alcohol, a small quantity of acetic acid added, and the*solution shaken
with zinc dust until colourless. The filtered solution deposited crystals
on standing, which were collected, washed, and dried. The substance
so obtained, although at first colourless, turned faintly yellow on dry-
ing; the melting point (160 — 165^) was far from sharp and the sub
stance reddened considerably on heating.
01686 gave 0-4580 00^ and 00866 H,0. C-7413 ; H = 6-6»,
0-1445 „ 11-9 C.C. nitrogen* at 16*^ and 754 mm. N = 9-51.
C^gH^^OjNj requires C=73'97 ; H-5*48; N-9-52 per cent.
These results absolutely confirm the constitution usually assigned
to benzeneazo-ornaphthyl acetate, namely, that it is an oxygen ester.
The possibility of the existence of an isomeric derivative was also
examined. Benzeheazo-a-naphthol, on treatment with mineral acids,
readily furnishes salts of a-naphthaquinone phenylhydrazone. Two
grams of benzeneazo-a-naphthol were added to glacial acetic acid which
had been saturated with hydrogen chloride and warmed in a flask
provided with a reflux tube down which 8 grams of acetyl chloride
were added in small quantities at a time. After half-an-hour's heat-
ing at 100°, the product was poured into water, the precipitate collected,
and recrystallised from glacial acetic [acid. The acetyl derivative so
obtained melted at 127° and when mixed with the acetyl derivative
prepared by acetylation with acetic anhydride and fused sodium
acetate did not depress its melting point. Hence salts of a-naphtha-
quinone phenylhydrazone furnished derivatives of benzeneazo-a-
naphthol on acetylation.
* Measoied over 60 per cent, potaesiom hydroxide solntioiL
Digitized by
Google
174 HEWITT AND AULD : CONSTITUTION OF
Action of Bromine on Benzeneazo-u-naphiM.
In acting with bromine on an oxyazo-compoond, a solution or sus-
pension of the latter in aoetio acid is best employed, and it is very
necessary to take especial care that hydrogen bromide is removed aa
quickly as it is formed. If this be not done, the hydrogen bromide
converts oxyazo-compounds into salts of quinone-hydrasones and sub-
stitution takes place in the nucleus free from oxygen (Hewitt and
Aston, Trans., 1900, 77, 712, 810). The bromination of benaeneazo-
a-naphthol has already been effected by Margary {OazzoOaf 1884, 1^
271), who took no precautions to avoid presence of a mineral acid.
The substance so prepared he regarded as /hbromobenseneaxo-a--
naphthol, stating that he obtained |^bromoaniline on reduction. Such
a result would not have been surprising were it not that the product
is described as occurring in two forms melting at 185^ and 197^
respectively, whereas the substance obtained synthetically by Bamber-
ger melted at 237—238^ (Bor., 1895, 28, 1896).
Bromination, if carried out in the following manner, furnishes a
product, melting at 196° which contains no bromine in the benzene
nucleus. Benzeneazo-a-naphthol, together with its own weight of fused
sodium acetate, is dissolved in 10 times its weight of glacial acetic
acid. The calculated quantity of bromine, diluted with twice ita
weight of acetic acid, is then added and the mixture allowed to stand
at the ordinary temperature in a closed flask until the odour of the
bromine has disappeared ; this frequently requires a week. The solid
matter is then' filtered off, washed with water, and recrystallised from
boiling glacial acetic acid, in which the substance is fairly soluble,
although the cold solvent dissolves it but sparingly. Analysis showed
that a monobromo-derivative had been produced ;
0*2040 gave 01 132 AgBr, Br » 2397.
0*2460 „ 01404 AgBr. Br =» 24-22.
0-2239 „ 17-0 C.C. nitrogen at 20° and 737 nmi. N»8-61.
GisHjiON^Br requires Br » 24*42 ; N»8*68 per cent.
The substance dissolves very easily in acetone, it is also dissolved
by alcohol, ether, carbon disulphide, or ethyl acetate, benzene dis-
solves it only sparingly^ whilst in light petroleum it is almost insoluble.
The solution in strong sulphuric acid has a much bluer shade than
that of the parent substance.
The reduction was effected by solution in alcohol and boiling with
an excess of tin and hydrochloric acid in a reflux apparatus for 1 hour.
After cooling, sodium hydroxide was added in excess and the mixture
distilled in a current of steam. The distillate was rendered alkaline
with soda, shaken with a small quantity of benzoyl chloride, and^the
Digitized by
Google
BBKZBN£AZO-a-NAPHTHOL. 176
precipitate collected and recrystallifled from benzene. Colourless plates
separated, which proved to be free from halogen and melted at 168^
(nncorr.). The substance was therefore bensanilide. It follows that
when benaseneazo-a-naphthol is brominated in presence of sodium
acetate, one atom of bromine enters the naphthol nucleus. The only
benzeneazobromo-o-naphthol hitherto described is the 8-bromo-4-benz-
eneazo-a-naphthol prepared by Meldola and Streatfeild {Trans. ^ 1893,
68^ 1058). It is probably not identical with our compound, although
its melting point, 197^ lies very near to that of the substance obtained
by direct bromination. To further characterise the latter, a number
of derivatives have been prepared and analysed.
The ethyl eiher was obtained by dissolving, successively, 0*1 gram of
sodium and 1*0 gram of the azoKsbmpound in 6 c.c. of ethyl alcohol
and heating with an excess of ethyl bromide for 2 hours at 120 — }30^.
The precipitate obtained on addition of water was recrystallised twice
from a miztiure of chloroform and alcohol ; the product melted at 220^
(unoorr.) :
01060 gave 00540 AgBr. Br » 22-61.
CigHi^ONsBr requires Br = 22*53 per cent.
The ethyl ether is a black powder, fairly soluble in acetic add and
somewhat readily so in chloroform. Most of the other usual organic
solvents dissolve it only sparingly in the cold.
The acetyl derivative was obtained by boiling in a reflux apparatus
for 2 hours a mixture of the azophenol with 1} times its weight of
fused sodium acetate and 3 times its weight of acetic anhydride. The
substance was isolated in the usual manner and recrystallised from
glacial acetic add; its melting point waa found to be 146^ (corr.) :
0'1441 i^ave 8*56 c.c. nitrogen at 8^ and 756 mm. Na 7'63,
CigHijOjNjBr requires N=:7'59 per cent.
To compare the product obtained by substituting bromine in
benzeneazo-a-naphthol with the three bromobenzeneazo-a-naphthols,
the latter were prepared and converted into acetyl derivatives.
I%e liomerio Bramobenzeneazo-w^fiaphthole,
o-BramoienzeTieazo-a-naphthol. — ^Pure o-bromoaniline (prepared from
o-nitraniline by Sandmeyer's reaction and subsequent redaction of the
o-bromonitrobenzene so obtained) was diazotised, the solution of the
diazonium salt added to the requisite quantity of a-naphthol dissolved
in methylated spirit, and an aqueous s<^ution of sodium acetate stirred
into the mixture. The product was collected, washed with dilute
alcohol, and recrystallised from glacial acetic acid, in which it is fairly
Digitized by
Google
176 CONSTITUTION OF BENZENEAZO-a-NAPHTHOL.
soluble on boiliDg, but only sparingly so when cold. It melted at 183^
(corr.):
0-2917 gave 21*0 c.c. nitrogen at 14"^ and 754 mm. N»8-49.
CjoHjiON^Br requires N => 8*58 per cent.
The (icetyl derivative, after recrystallisation from boiling glacial
acetic acid, melted at 123° :
0*1728 gave 10*3 c.c. nitrogen at 14'' and 754 mm. N» 7*03.
OjgHjgOsNjBr requires N=s7'59 per cent.
m-BrimMbrnzeneazo-a^naphihol, after recrystallisation from benzene,
melted at 211'' (uncorr.) :
0-2298 gave 16*8 c.c. nitrogen at 20"^ and 761 mm. N = 8*36.
OigH^jONgBr requires Na!8*58 per cent.
The acetyl derivative was prepared in the usual manner ; it melted
at 112°:
0*1252 gave 9*0 c.c. nitrogen at 23° and 744 mm. N^ 7*80.
CjgHig02N2Br requires N = 7*59 percent.
^J^ofnoberizenMzo-a-naphihol has already been described by Bam-
berger {Ber., 1896, 28, 1896). The melting point given by him is
237 — 238°; our preparation melted at 226° (uncorr.i the corrected
melting point would be about 233°). These melting points do not
differ materially, but are far removed from those given by Margary,
namely, 185° and 197° {loe. cU.). On analysis :
00572 gave 00328 AgBr. Br = 24-36.
CiflHiiONjBr requires Br » 24*42 per cent.
The aoeif/l derivative was also prepared in order to characterise the
substa9ce further. Prepared in the usual manner and recrystallised
from glacial acetic acid, it melted at 141° (corr.) :
0*1484 gave 0*3195 COa and 0*0499 HjO. 0 = 58-72 ; H«3*68.
OigHijO^NjBr requires 0 = 58-54 ; H«3*62 per cent.
The substance is easily soluble in benzene or chloroform, fairly so
in acetone or ethyl acetate, but only sparingly so in alcohol.
It is thus conclusively proved that in absence of strong acids, benzene-
azo-a-naphthol furnishes a substance which does not contain bromine
in the benzene nucleus. The position of the bromine atom in the
fl-naphthol nucleus has not been determined; it probably enters
position 2. So far, attempts at preparing the substance by the inter-
action of phenylhydrazine and Zincke and Schmidt's 2-bromo-l : 4-
naphthaquinone {Ber., 1894, 27, 2757) have been unsuccessful,
although from the production of benzeneazo-a-naphthol f rom a-naphtha-
Digitized by
Google
MAGNETIC ROTATION OF SOME POLYHYDRIC ALCOHOLS. 177
qainone and phenylhydrazine observed by Zincke and Bindewald, tbe
carrying out of sucb. a reaction appears easy of accomplishment.
Under the circumstances, we are compelled to leave the actual proof
that position 2 is occupied by the bromine atom to some future
occasion.
East London Technical Collieoe.
XIX. — The Magnetic Rotation of some ^ Polyhydric
Alcohols, Hexoses, and Saccharohioses.
By W. H. Pebkin, sen., Ph.D., F.R.S.
The remarkable changes in optical activity which many carbohydrates
show when in solution in water have engaged the attention of several
observers for a long period. To take an example, a freshly prepared
solution of glucose has a rotation of [a]© + 105*16°, but this gradually
diminishes and finally becomes constant after about six hours, the
rotation being then [aj^ +52*49° (Parous and ToUens, AnncUen,
1890, 257, 160). This phenomenon has been called bi-, multi-, or
muta-rotation, and it has been suggested by Tanret {Compt, rend.,
1895, 120, 1060) that the first form of glucose should be called
a-glucose and the second )8-glucose ; this method of distinguishing the
two modifications will be used in the present paper, not only in the
case of glucose, but in all cases where birotation has been observed.
A risumi of the views which have been entertained in reference to
birotation is given in a paper by Horace Brown and S. U. Pickering
"On the thermal changes attending change of rotatory power of
carbohydrates " (Trans., 1897, 71, 769). From this, it is seen that the
earlier attempts to explain the phenomenon of hi- or multi-rotation
were based on physical considerations. Subsequently, the probable
chemical aspect of / the matter came to be more fully discussed ; E.
Fischer, for example, has suggested that the remarkable birotation
shown by glucose may be due to the gradual assimilation of water and
conversion into the heptahydric alcohol, O^Hj^O^. This view has
latterly found considerable favour, and Brown and Pickering think
that the results of the heat determinations made by them are con-
sistent with it.
As the study of the magnetic rotations of the sugars might possibly
throw some light on this difficult subject, it was thought desirable to
undertake the examination of some of the more important of these
substances. Until lately, however, the measurements could not be
VOL. LXXXI. N
Digitized by VjOOQIC
178 PERKIN : THE MAGNETIC ROTATION OF SOME
made with any degree of accuracy, because stroug solutions t>£ these
sugars rotate the plane of polarisation through such large angles that,
as is well known, the impurities in the sodium light seriously affect
the appearance of the half-shadow disc of the polari meter, causing the
two sides to be very unequally tinted, so that useful numbers cannot
be obtained. Thus, a 50 per cent, solution of fitictose in a 100 mm.
tube has an optical rotation of about 50^, and this is the point at
which the magnetic rotation commences. Fortunately, after many
attempts, I have succeeded in finding a simple spectroscopic arrange-
ment by which this difficulty can be overcome, so that very large angles
may now be measured with considerable accuracy, and with this new
arrangement I have found it possible to determine accurately the
magnetic rotations of a number of carbohydrates. In a future com*
munication, I hope to give an account of this improvement and
also of the new apparatus which I am at present using for the deter*
mination of magnetic rotations.
Besides the sugars themselves, two of the polyhydric alcohols have
been measured, so that the magnetic rotations of this class of com-
pounds from the mono- to the heza-hydric are now known, with the
exception of that of the pentahydric alcohol, O5H12O5, which, however,
can be easily estimated. The examination of this series of alcohols
was important in order that a basis might be obtained from which to
calculate the probable rotation of the various sugars.
The nun^bers obtained for the magnetic rotation of this group of
alcohols may be briefly summarised as follows :
Mol. mag. lot.
Methyl alcohol H2(CH-0H) 1-640
Glycol Hjj(CH-0H)2 2-943
Glycerol Hj5(CH-OH)3 4-111
Eiythritol llJfill'OB)^ 5-230
Pentitol (missing) H2(OH-OH)5 6 300 est.
Mannitol HjjCCH-OH)^ 7-351
If the magnetic rotations of the alcohols actually examined be
plotted out, they form a regular curve, from which the rotation of the
missing pentahydric compound may be calculated ; also if the
curve be carried further, the rotations of the heptahydric and other
higher alcohols may be estimated, doubtless with considerable accuracy
(see diagram). From this curve, it will be at once seen that the
successive CH*OH groups have a smaller and smaller value as they are
repeated ; this, however, is not due to the group CH*OH as a whole,
but to the hydroxyl group which it contains, since in the homologous
series of paraffins, aliphatic acids, monohydric alcohols, and esters, it
has been conclusively proved that the value of each CH,, even in com-
pounds containing eighteen carbon atoms, is constant, namely, 1*023.
Digitized by
Google
POLTHYDRIC ALCOHOLS, HEXOSES, AND SAOOHABOBIOSES. 179
Attention has previously been directed to the diminishing influence
caused by suocessiye displacements of hydrogen by hydrozyl (Trans.,
1884, 45, 559) ; this diminishing influence is more clearly seen by
subtracting from the value of the polyhydric alcohol that of the corre-
sponding alcohol containing one hydrozyl less in its molecule. In
these cases, in which the magnetic rotation of the latter has not been
0-6
0-7
0-8
0-9
1-0
1-1
1-2
1-8
1-4
C Hj(CH-
OH)
\
\
•
>
^,(CH-
0R%
\
, H,CCH
0^\
V
"^
^UCH-
0H)4
OH),
Hj(CH-0
H).
c.
C4
XTte mofMiie ntatioM art found by adding the ordinate* to the cairbon Mmbon cf
the obieitKt.
directly determined, it can be obtained hj the addition of the value of
CH, to that of the next lower alcohol, thas ;
Glycol OjH«(OH),
Ethyl alcohol ... CjH5(0H)
Glycerol C,Hj(OH),
Less OHj+C^«(OH),
MoL mag. rot.
2-943 \
2-780 J
4-111 \
3-966 /
Digitized by
Diff. for OH
disp. H.
0-163
0-146
N 2
Google
Mol. mag. rot
5-230 ■
6-134.
Di£f. for OH
disp. H.
0096
6-300 \
6-263 J
0047
7-361 \
7-323 f
0-028
180 PERKIN: THE MAGNETIC ROTATION OF SOME
Erythritol C4H<j(OH)4
Less CH2 + 03H5(OH)3
Pentitol C6H^(OH)6
Less CHg + C4H^(0H)^
Mannitol CeH8(0H)g
Less CH2 + C5H7(OH)5
The influence of the bydroxyl group displacing hydrogen must, there-
fore, evidently become practically nil when the substitution has been
repeated seven or eight times. The results exhibited in the above
tables will be found to be very important in the calculation of the
probable rotations of glucose, fructose, &c.
Glucose is known to be an aldehyde. Now the difference between
the molecular magnetic rotations of an aldehyde and an alcohol, for
example, between those of heptyl alcohol and heptyl aldehyde^ is 0*438}
BO that the calculated rotation of glucose can be obtained by subtract-
ing this amount from that of mannitol.*
Mannitol 7-351
Less 0-438
Glucose 6-913
Fructose is known to be a ketone. The difference between the
magnetic rotation of a ketone and an alcohol, for example, between
that of Mc.octyl alcohol and of methyl hexyl ketone, is 0-495 ; this
subtracted from the value for mannitol should give the rotation of
fructose.
Mannitol 7-351
Less 0-495
Fructose 6-856
The actual determinations of the magnetic rotations of glucose and
fructose in aqueous solution have given almost identical numbers in
both cases, but the results are considerably lower than those calculated
above.
Glucose calc 6-913 Fructose oalc. 6-855
Found 6-723 Found 6729
Diff 0-190 Diff 0-126
* The actual comparison should, of course, be between glucose and sorbitol, but
the change of one asymmetric carbon atom in passing from sorbitol to mannitol
would have, if any, so little effect on the magnetic rotation that it may be
neglected.
Digitized by
Google
POLYUYDRIG iiLGOHOLS, HEX0S£8, AND SACGUAHOBIObES. 181
The queation then arises : Why are the actual magnetic rotations of
these sugars, determined in solutions which have undergone the usual
maximum change in optical rotation, lower than those calculated)
Is this due to the assimilation of water and the formation of a
heptahydric alcohol, 0^117(011)7, or must some other explanation be
found t
From the experimental part of the paper, it is seen that the magnetic
rotation of glucose in aqueous solution, obtained by subtracting the value
of llHjO from that of a solution of the composition O^Hj^O^yllH^O, is
found to be 6*723, and the same number is found in a similar way
from solutions of other concentrations. If, however, the glucose had
assimilated 1 mol. of water from the solution to form the heptahydric
alcohol OqH7(OH)7, the rotation of this compound will be obtained by
subtracting the value of only lOH^O from the result of the determin-
ation, that is to say, it will be 7*723. Erom the examination of the
curve (p. 179), it is clear that the rotation of the alcohol C^}i^(OK\
will be 8*380; if from thia we deduct the .value for OH^ (1*023), we
obtain 7*347 as the value of the alcohol O0H7(OH)7, a number which is
very different from that actually found, namely, 7*723. This evidence
therefore seems to show that glucose in solution is the anhydrous
substance OgH^j^a* ^^^ ^^ ^^^ combined with water to form the hepta-
hydric alcohol C^'H^iOR)^,.
Lowry (Trans., 1899, 76, 215), when referring to the subject of
birotation, suggests that the difference between glucose in the anhydrous
condition and in solution, after all change has taken place and the
optical rotation become constant, may be due simply to iaomeria change,
the aldehydic form I in the following table passing partly into one of
the isomeric modifications II or III. Of these expressions, formula U
was first proposed by ToUens {Ber., 1883, 16, 923), and afterwards
considered by E. Fischer as possibly, although not probably, repre-
senting the constitution of anhydrous glucose.
CHO CH-OH CH-OH
CH-OH ^ /CH-OH C-OH
CH-OH tfi-C
"^CH CH-C
CH-OH CH-C
CHo-OH CHo-
OH ^\^CH-OH CH-OH
^^H CH-OH
OH
OH
2 ^-^ ^"2
II. Ill
If, however, formula III be examined, it will be seen that it re-
presents an unsaturated compound, and this, according to the mag-
netic rotation, cannot be correct. The introduction of an ethylene
linking into the molecule of a saturated substance is known to raise
Digitized by
Google
182 pbrkin: the magnetic rotation of SOMfi
the magnetio rotation by 1*620^, and the value of glucose (calculated
from mannitol) would thus become 6*913 + 1*620 » 8*533, which is far
higher than the value actually found (6*723).
It has already been pointed out (p. 180) that the value for
glucose in solution (6*723) is lower by 0*190 than it should be if the
substance were an aldehyde, and the question then arises whether a
compound of the formula II would have a lower rotation than one
having the aldehydic formula I. That this will be the case can be
shown from the following comparisons between the values found for
glucose in solution, and those of ethylene oxide and the lactones, that
is, of substances which are constituted somewhat similarly to formula
n.
The value of ethylene oxide 0^X^^ calculated from that of gly-
col (2*943) by taking away 0*751 for the loss of the elements of water
(see p. 184) is 2*192, the value found was 1'935| making a difference of
0*247 (Trans., 1893, 63, 490). In the case of the lactones which have
been e2camined, namely, butyrolactone and valerolactone,
Butyrolactone. Valerolactone.
the following are the differences between values found and calculated in
a similar way: for butyrolactone - 0*230, and for valerolactone —0*195,
average, 0*212. Now the constitution represented by formula II agrees
best with that of the lactones, inasmuch as it contains a chain of four
carbon atoms closed by oxygen. If then glucose, when dissolved in
water, assumes to a greater or less extent this constitution, there is
good reason for believing that its rotation would be lower than that of
the aldehydic form, I, by about 0*2. This, it will be seen, agrees nearly
exactly with the number actually found, and there is therefore strong
support for the contention that, in solution, glucose has the constitu-
tion represented by formula II, or exists in some form analogous to
this. The solution would probably also contain a small quantity of
glucose in its ordinary aldehydic condition ; it is therefore possible
that the rotation of the p-iorm in the pure state may be a little lower
still than that found.
* The yalue for ordinary unsatnration with loss of H, is 1*112, bat as no
hydrogen la lost in this case, the value for unsaturation will be 1*112 + 0*508, the
value of Hg.
Digitized by
Google
i»OLYHTDBIO ALCOHOLS, HEXOSES, AND SACOHABOBIOSES. 183
If now formula II be slightly modified, an expression for the pos-
sible condition of fructose in solution may, in a Bimilar way, be obtained
wiiicli will be as follows :
C.OH
o/9H-oH ^
CHj-OH
What has been said about glucose applies equally well to fructose ;
the rotation is in both cases the same, and is lower than the calculated
value, although not quite to the same extent ; it is therefore probable
that fructose exists in solution, not as a ketone, but chiefly in a state
represented by the above formula, or by some other formula similar to
this.
If we now consider the relationship between the calculated mag-
netic rotation for glucose in its aldehydic form and that found for
galactose in solution, we have the following numbers :
Glucose, calc 6-913
Galactose, found 6-887
0-026
In considering these numbers, it should be noted that in optically
active compounds, difference in configuration only does not appear to in*
fluence magnetic rotation ; it is therefore probable that the magnetic
rotation of galactose as an aldehyde is the same as that of glucose as an
aldehyde. If, then, galactose in aqueous solution had been present
entirely in its aldehydic form, the number found should have been 6*913,
and the slight lowering observed in the value, namely, 0*026, appears to
show that, whilst present for the most part in its aldehydic form, galac-
tose has to some extent been converted into a modification similar to
that represented by formula II in the case of glucose in solution. It is,
however, remarkable that this small change appears to be accompanied
by so large an alteration in the optical rotation, since galactose, which
shows a rotation of approximately [ajo -t- 134*5° in freshly prepared
solutions, has a value of only [aj^ + 84*2°, when the solution has been
left to stand until the rotation has become constant, the formation of
the small amount of the substance of formula II being accompanied by
a fall in the optical rotation of no less than 50'3°. There is, however,
no evidence to show what the optical rotation of substances of the
type represented by formula II would be in the case of glucose,
Digitized by
Google
184 P£BKIN: THE MAGNETIC ROTATION OF SOME
fructose, or galactose. It is quite possible that such forms of the
sugars, although similar in general character, might have very widely
difEerent optical rotations, and this is evidently the case, since fructose
is IsBVorotatory in solution, whilst glucose and galactose are dextro-
rotatory in different degrees. Quite possibly a dextrorotatory alde-
hydic sugar might yield a Isavorotatory substance of the type repre-
sented by formula II on going into solution, and this might be so in the
case of galactose when it is entirely converted into its isomeric form.
We have in nitrocamphor a remarkable instance of this kind of
change, only of the reverse order; a-nitrocamphor, which is hevo-
rotatory, when changed into the isomeric ^-nitrocamphor, becoming
enormously dextrorotatory. Again, ir-bromonitrocamphor in its normal
condition has a rotation of [ajo —38% but in its pseudo-form has
[at +188° (Lowry, Trans., 1899, 76, 223).
The birotation of galactose is also much increased in amount by the
addition of lead acetate to its solution, the rotation falling by 53 per cent.
(Hanno Svoboda, Zeit. Ver. Eubenztieker.-Ind. Dtut. Beieha, 1896,
46, ff^t. 481, 29 pages; also Abstr., 1896, i, 406) I find also
that a cold solution of caustic alkali reduces the rotation very
considerably.
As sucrose represents glucose and fructose less 1 mol. of water,
its magnetic rotation can be easily calculated.
The decrease in magnetic rotation caused by the loss of the elements
of water when alcohol is converted into ethyl ether, acetic and
propionic acids into their anhydrides, &c., averages about 0*752 (Trans.,
1886, 49, 787), being in some cases a little less, and in others a little
more than this ; therefore when this value is subtracted from those
of the two sugars, the difference should approximately give the magnetic
rotation of sucrose thus :
a-Glucose + a-fructose, calc 1 3*768
lessHjjO 0-752
Sucrose calc 13-016
found 12-586
-0-430
From this it is seen that the experimental number is very much
lower than the calculated. If, however, the experimental numbers
of glucose and fructose in solution as /^-modifications be taken instead
of those calculated for the magnetic rotation of the anhydrous or
a-sugars, the following result is obtained :
Digitized by
Google
POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES. 186
j3-Gluoo6e +j3-fracto0e found 13*452
lessH^O 0-762
Sucrose 12-700
found 12-586
-0-114
As the difference between the numbers actually found and those
calculated in the above way is so small,^ it would seem that sucrose
is apparently built up of the isomeric or )8-form8 of glucose and fructose,
and not of the aldehydic and ketonic forms.
If, then, sucrose is built up of the isomeric forms of glucose and
fructose, it will probably have the formula :
^v 9H-0H Q/ / CH-OH
yn \(fH-OH '
CH-OH ^CH
CHj-OH CH^-OH
and its constitution in the dry state and in solution will most likely
be the same, since it does not exhibit the phenomenon of birotation.
The above formula for sucrose has already been proposed by E.
Fischer {Ber,, 1893, 26, 2406) ; it is a modification of that suggested
by Tollens (Ber.^ 1883, 16, 923), and clearly shows that when sucrose is
hydrolysed it should at first be resolved into the isomeric or )9-modifi-
cations of glucose and fructose :
</:
.9H-0H CHj-OH
CB'OH C'OH
^CH \ 6H'0H
i
H-OH ^CH
/3-Glaco8e. /B-Frnctose.
Maltose and Laeto$6.
These sugars differ in a marked manner from sucrose in that they
possess birotatory and cupric reducing powers ; there can therefore be
no doubt that they must have a structure essentially different from
that of sucrose.
* If, 88 supposed, the numbers found for these jB-compounds are a trifle high, on
account of the solution containing a little of the a-compounds (see p. 182), this
difference would be still smaller.
Digitized by
Google
186 PEBKIN: THE MAGNETIC BOTATIOK OF SOME
In order to account for this difference, E. Fischer {loc. cit.) suggests
the following formula for lactose :
</:
^Hg-OH CHO
H-OH CH-OH
H CH-OH
H-OH CH-OH
OH CH-OH
OH O
Galactose radicle. Glucose radicle.
In this, the galactose radicle is represented as in the /3- and the
glucose radicle in the a-condition, whilst if this formula be applied to
maltose, one glucose radicle will be in the ft- and the other in the
a-condition, On investigating this matter, it was at first thought that
the view of the difference in constitution between maltose and lactose
on the one hand, and of sucrose on the other, received some immediate
confirmation from the results of the magnetic rotations of the former,
which are rather higher than the value obtained for sucrose ; no doubt
this has a bearing on the subject, but it is doubtful whether any great
importance can be attached to this difference. Erom the fact,
however, that these carbohydrates contain a glucose instead of a
fructose radicle, their magnetic rotations should be about 0*057 higher
than that of sucrose. The rotations are as follows :
Maltose, found 12-690
Sucrose „ 12-686
+ 0-104
Lactose, found 12*714
Sucrose „ 12*586
+ 0128
If maltose be first considered, its magnetic rotation, on the assump-
tion Uiat its constitution is represented by the above formula, may be
calculated thus :
)8-Gluco8e, found 6*723
a-Glucose, calc 6*913
13-636
Less H,0 0-752
Calculated mol. mag. rot. of maltose... 12*884
Found 12*690
Diff.... 0194
Digitized by VjOOQIC
POLTHYDRIC ALCOHOLS, HBXOSBS, AND SAOCHAROBIOSES. 187
This difference is almost exactly the same as that observed between
a-glacose and jS-glucose, 0*190 (p. 180), and points to the probability
that the second or a-glucose radicle in maltose also undergoes con-
version, either entirely or in part, into the )3-modification when the
sugar is dissolved in water, and that the constitution of dissolved
maltose is :
!Hj-OH CH-OH
o/;
H-OH / CH-OH
^v 6h.oh
H-OH ^CH
V V.H-OH 6h-0H
^CH 0 CHj
The rotation, assuming that in a solution of maltose both glucose
radicles are in the )3*modification, may be calculated as follows :
Mag. rot. of 2 mols. )3-gluoose 13*446
Less HjO 0-762
12-694
Found 12-690
Diff 0004
The magnetic rotation of lactoHy as already stated, was found to be
12*714, and if this value be examined, it will be seen it also indicates
that lactose in solution contains both the galactose and glucose radicles
in the /3-condition. It has been seen that galactose when in solution
is chiefly in the a-condition ; if, however, it were principally in the
/3^x>ndition,its rotation, no doubt, would be similar to that of /3-glucose,
so that the rotation of lactose should be the same, or nearly so, as that
of maltose, and this is found to be the case, the difference being only
+ 0-020. In the dry state, it probably has the formula proposed by
E. Fischer, and this is, of course, equally true of maltose. Very prob-
ably these two carbohydrates, when in solution, always contain a little
of the glucose radicle in the a- or aldehydic condition.
EZPEBIHENTAL.
EryihfUol, C^H^oO^.
This substance was purified by recrystallisation from water. The
solutioA examined was supersaturated, containing 32*62 per cent, of
erythritol, it being found possible to measure its rotation before
crystallisation set in ; the composition of the solution was Gfi^fi^-^'
14H,0.
Density, d 16716°, M043; d 20°/20° M033.
Digitized by
Google
188 JERKIN: THE MAGNETIC ROTATION OF SOME
The average of three sets of determinations of the magnetic rotation
made at different times was :
t. Sp. rot. MoL rot of sol. Mol. rot. of C4H10O4.
15° 1-0220 19-230 5-230
Mannitol, C^H^^Oq.
This was recrystallised from water before use. As in the case of
erythritol, a supersaturated solution was employed ; it contained 18-176
per cent, of mannitol, its composition being CgH^^O^ + iOH^O.
. Density, d 15715°, 1-0752 ; d 20°/20°, 1-0746.
The average of four sets of measurements of the magnetic rotation
made at different times gave :
t. Sp. rot. Mol. rot. of sol. Mol. rot. of CeHi408.
17-5° 10154 47-351 7351
Glucose, O^jHjgOg.
Two specimens of this substance were examined, one obtained from
Kahlbaum, and the other, a very pure preparation, for which I am
indebted to Dr. Horace Brown. With the former, four sets of
measurements were made on different occasions and with solutions of
various strengths, the most dilute being represented by C0HJ2O0 +
2OH2O, and with that from Dr. Horace Brown also four measurements
were made, but with only one strength, represented by C^H^^O^ + 1 IHi^^O
and containing 47-619 per cent, of O^H^gO^. The products used were
the monohydrate dried over sulphuric acid in a vacuum :
The density of the solution C^Hi20g + IIH^O was d 15715°, 1*2147 ;
(2 20°/20°, 1-2135.
Magnetic rotation :
t
Sp. rot. •
Mol. rot. of sol.
Mol. rot of C,H,jO^
15°
1-0261
17-723
6-723
The average of the measurements made with E^hlbaum's specimen
was 6-715, which is very close to the above.
The permanent optical rotation of the solution containing 47*619 per
cent, of C^HjaOg was [a]i> 56*22° at^l6-9° This is a little higher than
that given for weak solutions.
If the magnetic rotations be calculated on the assumption that the
glucose has assimilated a mol. of water and thus become a heptahydric
alcohol, the solution will then have the composition CoH-^fi^ + lOH^O.
The calculation will be the same as the above, only the value of 10
instead of 11 mols. of water will have to be subtracted from the mole-
Digitized by
Google
POLYHYDRIC ALCOHOLS, HEXOSES, AND SACCHAROBIOSES. 189
cular rotation of the solution, and the rotation of the alcohol will thus
become 7*723.
Frucose, O^^fi^,
This was prepared from innlin and obtained from Kahlbaum. It
was dried over sulphuric acid and its composition checked by a
combustion ; it gave 0, 39*8, and H, 6*8, the formula C^H^^^a requiring
C, 40*0, and H, 6*7 per cent. Its solution was examined in one strength
only, containing 60 per cent, of fructose and represented by O^H.^fi^ +
lOH^O.
Density, d 16716^ 1*2226 ; d 20720°, 1*2211.
The average of five sets of measurements of the magnetic rotation,
made on different occasions, gave :
t. Sp. rot. Mol. rot of sol. MoL rot. of OeH^gOe.
15° 10227 16-729 6729
Optical rotation [o]d 96*19° at 16°
Galaetase, C^H^^O^.
This substance was examined in a very supersaturated solution,
from which it does not crystallise very quickly. It contained 60 per
cent, of the sugar, its composition being represented by OJS.^fi^^
lOHjO.
Density, d 16716°, 1*2311 ; d 20°/20°, 1*2299.
The average of four sets of measurements of the magnetic rotation,
made on different occasions, gave
t Sp. rot. Mol. rot. of sol. Mol. rot. of C^Hifi^
16° 1*0396 16*887 6*887
Optical rotation [a]^ 84-23° at 14-6°.
Sucrose, CijHjjOn.
The specimen used was ordinary sugar recrystallised from alcohol
(76 per cent.)r The composition of the solution used was ^presented
by G^HjjOjj + I9H2O, and contained 60 per cent, sucrose.
Density, d 16716°, 1*2327 ; d 20°/20°, 1-2316.
The magnetic rotation, determined on four different occasions, was :
t. Sp. rot. Mol. rot. of soL Mol. rot of C^aHaOu.
15° 1-0247 31-586 12-586
Optical rotation [a]© 66-61° at 17°.
VOL. LSXXI.
Digitized by
Google
190 MAGNETIC ROTATION OF SOME POLYHTDRIC ALCOHOLS.
Maltose, O^^B.^O^i.
For a very pure specimen of this compound, I am indebted to Dr.
Horace Brown. The solution employed contained 47*5 per cent, of the
sugar, its composition being represented by Cj2^2s^ii + ^OHgO.
Density, d 15°/15°, 1'2214 ; d 20720°, 1-2205.
The average of three sets of measurements of the magnetic rotation,
made on different occasions, gave :
t. Sp. rot. Mol. rot. of sol. MoL rot. of CiaHgO,!.
16° 10288 33-690 12690
Optical rotation [a]^ 137-0° at 16-7°
IfOCtOSB, O^^R^fi^v
This was obtained from Kahlbaum and was purified by fractional
crystallisation, the crop deposited during the first 12 hours being
rejected. It was dried in a vacuum over sulphuric acid. The solution
used was a supersaturated one containing 33*333 per cent, of the
sugar, and had the composition C13H22OJ1 + 4IH2O.
Density, d 16°/15°, 1-1413 ; d 20°/20°, 1-1406.
The magnetic rotation, as determined four times on different occa-
sions, was :
t. Sp. rot. Mol. rot. of soL Mol. rot. of Ci^HssOu.
18-4° 1-0213 53-714 12-714
Optical rotation [o]d 526° at 18°.
Summary.
The chief results obtained in this investigation go to show :
(1) That the influence of successive hydroxy 1 groups in polyhydric
alcohols on the magnetic rotations diminishes as they increase in
number, until about the seventh is reached, when it becomes almost
nil,
(2) That solutions of glucose and fructose, after all change has
taken place, give magnetic rotations which indicate that birotation
is not due to hydration, but that it is caused by a change in the
constitution of these substances.
(3) That galactose, when in solution, does not undergo isomeric
change to so large an extent as glucose.
(4) That sucrose is built up of the isomeric or /3-forms of glucose
and fructose by the elimination of the elements of a mol. of water.
Digitized by
Google
HTDROeTANIC, CYANIC, AND CTANURIC ACIDS. 191
(5) That maltose is fonned from 1 molecule of glucose in the
aldehydic or a-cond^tion and 1 molecule in the isomeric or /^-condition
by the elimination of the elements of a mol. of water and that lactose
is similarly derived from 1 molecule of a-glucose and 1 of /^-galactose,
both being constituted in a similar manner to that proposed by E.
Fischer for lactose, also that when in solution these sugars undergo
isomeric change, the a-portion becoming transformed, more or less,
into the /3-condition. This change accounts for the birotation and cuprio
reducing power of the two sugars.
XX. — The Constitution oj Hydrocyanic^ Cyanic, and
Cyanuric Acids.
By P. D, Chattaway and J. Mbllo Wadmorb.
Although the simplest cyanogen derivatives have been for more than
a century among the most familiar of carbon compounds, there is no
general agreement as to their constitution. They all contain a carbon
and a nitrogen atom associated together, and different opinions are
held as to the manner in which hydrogen or halogen atoms are at-
tached to this group.
As a rule, well-defined classes of alkyl derivatives corresponding with
each possible structure are known, the behaviour of which leaves no
doubt concerning their molecular arrangement, but the reactions of
the cyanogen acids, their salts and halogen derivatives, are contradic-
tory, and apparently equally well-established facts lead to opposite
conclusions.
Hydrocyanic acid, cyanogen chloride, cyanic and cyanuric acids, for
example, may have the following structures :
C-OH CO
H-C:N or H-NIC t/ \ in/ \tr
^•?o\ " '^n H0.8 Lh "' o? t
H*0*C:N or O.C*N'H \ >^ \ /
N KfH
The formulie generally adopted are those given first, the hydrogen,
halogen, and hydroxyl being regarded as attached to the carbon atom.
The knowledge which we have recently acquired of the strikingly
different behaviour of halogen when attached to carbon or to nitrogen
made it probable that a study of the action of halogens on the
o 2
Digitized by
Google
192 CHATTAWAT AND WADMORE: THE CONSTITUTION OF
cyanogen acids, and of the derivatives thereby produced, would afford
direct evidence as to their constitution. Speaking generally, imino-
hydrogen is more readily replaced by halogen than hydrogen attached
to carbon, and the imino-halogen compounds are characteristically
reactive, while the carbon halogen Mnkage is comparatively stable.
Cyanogen chloride, bromide, and iodide were among the earliest
discovered compounds of cyanogen, as they are formed with the
greatest ease by the action of the halogens on aqueous solu-
tions of hydrocyanic acid or its salts. A careful study of the be-
haviour of these compounds shows that they possess all the typical
and characteristic properties of compounds having halogen attached
to nitrogen.
They react, for example, quantitatively with solutions of hydriodic
acid, sulphurous acid, and hydrogen sulphide, hydrocyanic acid being
in each case reformed, while iodine, sulphuric acid, and sulphur are
respectively produced.
Taking cyanogen bromide, for example, the reactions are repre-
sented by the equations :
OIN^Br + 2HI = C:N-H + HBr + Ij.
CIN-Br + HjSOj + HjO = CIN-H + HBr + HjSO^.
C:N-Br + HjS = C:N-H + HBr + S.
This behaviour shows that the halogen is attached to nitrogen and not
to carbon in these compounds, and that, consequently, they must be
represented by the formulsB :
C:N-C1; CIN-Br and CIN-I.
The carbon is conventionally represented as divalent, and the nitrogen
as trivalent; no very different conception, however, would be ex-
pressed if the carbon were represented as tetravalent and the nitrogen
as pentavalent, since what is implied is that the carbon is attached
to the nitrogen by the resultant affinity which, under the circum-
stances, the atoms are capable of exerting.
The ease with which the cyanogen halogen compounds can be formed
from prussic acid and its salts, and again transformed into them,
makes it in the highest degree probable that these have the imino-
constitution, and hence should be represented by the formulae :
c:n-h c:n-k c:N-Ag.
This conclusion, moreover, is the only one which will satisfactorily
explain their whole chemical behaviour.*
The relations of the cyanides and cyanogen chloride to cyanic acid
* We have not thought it necessary to go into explanatory details as these can be
easily supplied.
Digitized by
Google
ffTDROCYANIC, CtANIC, AKD CYANlJRIC ACIDS. l9S
and itfl salts have been among the chief reasons which led to the
adoption of the hjdrozj-formul» :
n:c-oh n:c-o-k.
for these compounds.
Since, as we have just shown, the former are imino-compounds
these relations become reasons for adopting the alternative imino-
structure :
OIO-NH and 0:c:N-K.
For example, the production of potassium cyanate, when cyanogen
chloride is treated with aqueous potash, has been used as an argument
for the hydrozy-constitution, since if the chlorine in cyanogen chloride
is attached to carbon, it could be regarded as a normal case of
hydrolysis :
N:C-C1 + 2K0H - KCl + N:C-0-K + HjO.
Cyanogen chloride, however, has the imino- structure, and the reaction
becomes an argument in the other direction, for a comparison of this
behaviour with that of the analogous cyanogen iodide shows that it
must be regarded as a normal hydrolysis of a nitrogen chloride
followed by oxidation of the potassium cyanide first formed :
C:N-C1 + 2K0H =. C:N-K + KOCI + HjO.
« o:c:n-k + kci + Kfi.
Analogy with cyanuric acid also is in favour of the imino-structure.
The action of chlorine on a solution of potassium cyanurate is
precisely similar to its action on potassium cyanide, the potassium
atoms are replaced by chlorine, and a well-defined crystalline com-
pound is produced, thus :
C3K3N3O3 + 301^ = C3CI3N3O3 + 3KCL
The entire chemical behaviour of this substance shows that the whole
of its halogen is attached to nitrogen. It liberates chlorine when
treated with hydrochloric acid, iodine with hydriodic acid, and
oxidises sulphurous to sulphuric acid. Cyanuric acid is, in each case,
reformed, and the reactions are quantitative ; the action with hydro-
chloric acid, for example, takes place according to the equation :
C3CI3N3O3 + 3HC1 - C8H3N3O8 + 3C1^
It is hydrolysed by water or alkalis, yielding hypochlorous acid or
hypochlorites. It reacts explosively with a strong ammonia solution,
nitrogen being liberated, and also with a solution of hydrogen sulphide,
setting free sulphur. Cyanuric acid is in each oase reformed. The
compound must therefore be trichloriminocyanuric acid.
Digitized by
Google
19* CUATtAWAY AND WADMOBE: Tfi? CONSTITUTIOl* Of
Since cyanurates are so readily and completely converted into this
trichlorimino-derivative, and the latter in many reactions equally
readily and completely again into cyanuric acid, we are justified in
concluding that Hofmann was in error in assigning a hydroxy-
constitution to the acid and its salts, and that, on the contrary, they
have the imino-constitution, and assuming the correctness of the
cyclic structure that they must he expressed by the formulae
CO
CO
CO
^jf NH
00 CO
/\
/\
KN NK
00 CO
CIN NCI
OC CO
Vu
V^
NCI
Cyanuric add.
Potaariam cyannrate.
Trichlorimino-
cyanuric acid.
A similar study of the behaviour of cyanuric chloride and bromide
confirms Hofmann's conclusion that in them the halogen is attached
to the carbon and not to the nitrogen. They do not liberate iodine
from hydriodic acid or sulphur from hydrogen sulphide, nor do they
oxidise sulphurous acid, even when heated to 100° with these
reagents.
This constitution, however, was to be expected from the structure
of the cyanogen halogen compounds, from which they are produced by
polymerisation under the influence of halogen acids.
Cyanogen chloride and bromide, as we have shown, are chlorimino-
derivatives in which the carbon being unsaturated is able to combine
with two monad atoms. In the polymerisation, the halogen acid
in all probability first adds itself on forming molecules having the
constitution :
^>C:N-C1 or ^J>C:N*Br,
which, on coming into contact, unite into ring systems of normal
structure with elimination of halogen acid, thus :
Civ \c^ .CI C
CI J ^c/^ = fl \ + ^^^^•
.N<^ \C1 Cl-C C-Cl
/^„. ..
N
As Hofmann has pointed out, all the relations of the cyanogen
group can only be explained by assuming isomeric change to occur in
certain reactions ; the issue is as to where this takes place.
Digitized by
Google
HYDROCYANIC, CYANIC, AND CYANUIUC ACIDS. 195
Cyanuric chloride, as is well known, yields oyanuric acid and
hydrochloric acid on prolonged heating with water, the reaction being
more rapid if alkalis are present. This and the corresponding con-
version of cyanuric acid into cyanuric chloride by phosphorus
pentachloride are the chief grounds on which Hofmann assigned the
hydrozy-structure to the acid.
If, however, the views now put forward as to its constitution are
correct, these are the reactions where isomeric change occurs, and
analogous behaviour in other well-established cases renders this
probable. We must assume that in the hydrolysis of cyanuric
chloride normal cyanuric acid is first formed, but as in many cases
where we have the grouping _I , the configuration is unstable
O H
and passes into the stable arrangement _}A X- , so here we have an
intramolecular change, the stable imino-form of ordinary cyanuric
acid being the result :
COl C-OH CO
/\ /% /\
IC CCl HO-C C-OH OC CO
Cl(
NH N NH
\^ V ^-^
The action of phosphorus pentachloride on cyanuric acid is probably
analogous to its action on amides, the replacement of an oxygen atom
by two chlorine atoms being followed by the elimination of hydrogen
chloride :
CI CI
0 V
0 o
I'
/ \._ ^^ \.r / \
HN NH
NH NH N
EzrSBIHINTAI..
Cyanogen Chloride, CIN-Cl,
ThiB compound shows the characteristic behaviour of a nitrogen
chloride, although it reacts less readily than is usual with such
substances. When hydriodic acid is added to an aqueous solution
of cyanogen chloride at the ordinary temperature, very little iodine is
liberated ; the amount, however, increases slowly on standing, rapidly
Digitized by
Google
196 CHATTAWAY AND WADMORE : THE CONSTITUTION OF
on heating to near 100^, until it reaches about 80 per cent, of that
required by the equation :
c:n-ci + 2HI = c:n-h + hci + i^.
If the heating be prolonged, the free iodine slowly disappears,
probably owing to hydrolysis of the hydrocyanic acid, and oxidation
of the formic acid or ammonia produced.
When aqueous solutions of cyanogen chloride and hydrogen sulphide
are heated together to 100^, sulphur is set free in considerable quan-
tity ; the hydrocyanic acid formed is mainly hydrolysed, but a small
amount escapes decomposition and combines with some of the liberated
sulphur to produce thiocyanic acid.
Similarly, when solutions of cyanogen chloride and sulphurous acid
or sulphites are heated to 100^, the latter are oxidised while the hydro-
cyanic acid is destroyed, probably hydrolysed.
No liberation of chlorine can be detected ^hen a solution of cyanogen
chloride is heated with hydrochloric acid to 100^ ; the cyanogen chloride,
however, is completely decomposed at this temperature.
The production of potassium cyanate and chloride by heating cyano-
gen chloride with caustic potash is probably due to the normal hydro-
lysis which all chlorimino-compounds undergo, followed by a subse-
quent oxidation of the cyanide by the hypochlorite formed.
c:n-ci + 2K0H = c:n-k + koci -»■ h,o = o:c:n-k -h kci + h,o.
Cyanogen Bromide, CIN-Br.
Cyanogen bromide is much more reactive than cyanogen chloride.
At the ordinary temperature, it liberates iodine from hydriodic acid,
sulphur from hydrogen sulphide, and oxidises sulphui*ous acid or sodium
sulphite. All these reactions are quantitative, hydrocyanic and hydro-
bromic acids being formed in equivalent amount.
A weighed quantity of cyanogen bromide was added to an excess of
a solution of hydriodic acid made by dissolving 10 grams of potassium
iodide in 100 c.c. of a 5 per cent, solution of acetic acid ; hydrocyanic
acid and iodine were at once liberated, the latter being then estimated
by sodium thiosulphate :
0-2439 liberated I = 46 c.c. iV7l0 iodine. Br as IN-Br = 76-4.*
CIN'Br requires 76-43 per cent.
A weighed quantity of cyanogen bromide was added to an excess of
an approximately decinormal solution of hydrogen sulphide ; sulphur
* Throaghoat this paper the resalts are calculated on the asaninption that the
aubstancea under consideration react as typical nitrogen halogen compounds, the
nuQibera are then compared with the percentages calculated from the formdln.
Digitized by
Google
HYDROCYANIC, CYANIC, AND CYANURIC ACIDS. 197
was at onoe deposited and hydrobromic and hydrocyanic acids formed,
together with a little thiocyanic acid, produced by the action of the
solphar on the latter ; the excess of hydrogen sulphide was then esti-
mated by a solution of iodine :
0-4791 reacted with 907 c.c. iV/lO H2S/2. Br as :N-Br=-75-68.
GIN* Br requires 75'43 per cent.
A similar procedure was adopted in studying the reaction with
sulphurous acid. A weighed quantity of cyanogen bromide dissolved
in dilate acetic acid was added to an excess of a decinormal solution of
sodium sulphite, and then the excess of the latter estimated by a solu-
tion of iodine :
0-6050 reacted with 95-26 c.c. iT/lO l^(i^S0J2. Br as :N-Br = 75-41.
CIN'Br requires 75*43 per cent.
No bromine is liberated when cyanogen bromide is heated with a
solution of potassium bromide made acid with acetic acid, or when it
is heated with strong hydrochloric acid to 100^, although in the latter
case it is decomposed just as cyanogen chloride is when similarly
treated.
Cyanogen Iodide.
Cyanogen iodide is more reactive than cyanogen chloride or cyanogen
bromide, and behaves as a typical nitrogen iodide. It reacts with
hydriodic acid, liberating iodine, with hydrobromic acid liberating
bromine and iodine, with hydrochloric acid forming iodine monochloride ;
it oxidises sulphurous acid and sodium^ sulphite, forming sulphates and
sets free sulphur from hydrogen sulphide. Its behaviour towards several
of these substances was very carefully studied by E. von Meyer {J, pr,
Chem.f 1887, [ii], 35, 292). He showed that the reaction between
hydrocyanic acid and iodine is a reversible one, and that two molecules
of sulphur dioxide completely reduce two molecules of cyanogen iodide
to hydrocyanic and hydriodic acids. He, however, writes the formula
ICN, and concludes his paper by stating that it is the only oxygen-free
iodide soluble in water which shows the surprising behaviour of
liberating iodine under the action of reducing agents, but of remaining
unattacked by reagents which set iodine free from other iodides.
We have quantitatively studied the behaviour of cyanogen iodide in
order to compare it with that of the bromide and the chloride. A
weighed quantity of cyanogen iodide was added to an excess of a solu-
tion of 10 grams of potassium iodide in 5 per cent, acetic acid ; hydro-
cyanic acid and iodine were at once liberated, the amount of the latter
being then estimated by sodium thiosulphate :
0 2964 liberated I « 38-8 c.c. i^/10 iodine. I as IN-I » 83*02.
OIN-I requires 82*97 per cent.
Digitized by
Google
198 CHATTAWAT AND WADMOBE : THE CONSTITUTION OF
This result is exactly that required by the equation :
c:n-i + hi = c:n-h + i,.
When cyanogen iodide is dissolved in an excess of strong hydrochloric
acid at the ordinary temperature, very little action takes place, but on
warming to 20 — 30° the liquid becomes orange-coloured, owing to the
formation of iodine monochloride, and the colour deepens as the tem-
perature rises. There is no liberation of free iodine even on boiling
the solution. Hydrocyanic acid is also produced. In one experiment,
the iodine monochloride was distilled off into a solution of potassium
iodide, and the liberated iodine estimated. The amount of iodine
monochloride obtained was about 2 per cent, below that required by
the equation
c:n-i + Hci = c:n-h + ici.
The loss is probably due to the hydrolysis of a small amount of the
hydrocyanic acid and partial oxidation by the iodine monochloride of
the products.
When cyanogen iodide is similarly treated with hydrobromio acid,
both iodine and bromine are evolved, but, as with hydrochloric acid,
the amount falls somewhat short of that required by the equation
C:N-I + HBr « C:N-H + Brl,
probably from a similar cause.
When a solution of sulphurous acid is slowly added to cyanogen
iodide, iodine is liberated, hydrocyanic acid and sulphuric acid being
simultaneously formed ; if, however, the iodide be added to an excess
of sulphurous acid, no liberation of halogen occurs (compare Sbrecker,
Amialm, 1868, 148, 90).
A weighed quantity of cyanogen iodide was added to an excess of a
decinormal solution of sodium sulphite so that no iodine was set free,
and the excess of sulphite estimated by a dilute solution of iodine :
0-2960 oxidised 38-7 c.c. of NjlO NajS08/2. I as IN-I - 82-92.
CIN*I requires 82*97 per cent.
The action takes place according to the equation
2C:N-I + HjSOj + HjO = 2C:N-H + H2SO4.
Sulphur is set free and hydrocyanic and hydriodic acids are formed
when cyanogen iodide is added to an excess of a solution of hydrogen
sulphide. If the latter is slowly added to the iodide, iodine is also
liberated, owing to the action of the hydriodic acid first formed on
the unchanged cyanogen iodide. A little thiocyanic acid also is always
formed from the interaction of some of the hydrocyanic acid with the
sulphur.
Digitized by
Google
MTbROCTANlC, CYANIC, AND CYANUBIC ACIDS. l99
In the following experiment, cyanogen iodide was added to an ex-
cess of hydrogen sulphide, the amount of the latter remaining unacted
on being estimated by a solution of iodine :
0-2896 reacted with 37-9 c.c. JV/IO H^/2. I as IN-I = 83.
CIN'I requires 82 97 per cent.
The result is expressed by the equation
2C:N-I + HjS = 2C:N-H + 2HI + S.
Its behaviour towards a solution of potassium hydrate also shows
that in it the halogen is attached to nitrogen, and affords an explana-
tion of the apparently different action of the similarly constituted
cyanogen chloride and bromide.
When it is added to a boiling solution of caustic potash, it is at once
decomposed; among other products, a small quantity of potassium
iodate is formed. Cyanogen chloride and bromide, when similarly
treated, form no chlorate or bromate. All the cyanogen halogen com-
pounds, however, are readily decomposed by caustic alkalis, yielding
cyanates.
The nitrogen halogen linkage, as is well known, behaves in a char-
acteristic way on hydrolysis, the halogen being invariably replaced by
hydrogen and becoming itself attached to the residual hydroxy!,
thus :
•N-X -J- H-O-H = -N-H + X-O-H.
It is thus sharply distinguished from the carbon halogen linkage,
where the opposite happens, thus :
•C-X + H-O-H = -C-O-H + XH.
The formation of iodate in the reaction between cyanogen iodide and
potash shows that, at first, nitrogen halogen hydrolysis undoubtedly
takes place, thus :
c:n-i + 2K0H = c:n-k -»■ k-o-i + HjO.
A certain amount of the hypoiodite, on account of the ease with which
it is transformed into iodide and iodate, escapes reduction by the cyanide
simultaneously formed, a reaction which results in the production
of cyanate:
c:n-k + K-O-I = o:c-N-K + kl
In the cases of cyanogen chloride and bromide, the hypochlorite and
hypobromite, which must first be formed, do not transform so readily,
and consequently are wholly reduced.
Digitized by
Google
200 CflATTAWAT AND WADMORE : THB CONSTITUTION OF
Cyanuric Chloride and Bromide,
These compounds show none of the reactions characteristic of the
halogen nitrogen linkage.
Small quantities of each were taken and heated for 30 minutes at
100° in stoppered bottleSi air being excluded, with solutions of hydr-
iodic acid, hydrogen sulphide, and sodium sulphite. No iodine or sul-
phur was liberated, nor was the sulphite oxidised. This behaviour is
in agreement with Hofmann's view of their constitution, deduced from
altogether different reactions, and with the formula assigned to them
by him and generally adopted :
CCl OBr
/% /%
If ? ^5
CIC CCl BrC CBr •
\^ \^
N N
CO
/\
. CIN NCI
TriMoriminocyanuric Acid, qX qq '
\/
NCI
This compound is prepared with the greatest ease by dissolving
cyapuric acid in the theoretical quantity of a 5 per cent, solution of
caustic potash and passing a rapid stream of chlorine through the
liquid cooled to 0°.
Trichloriminocyanuric acid separates as a heavy, white, crystalline
powder which is obtained perfectly pure by washing a few times with
water and drying rapidly on a water-bath :
O3C3N3K3 + 3CI2 - O3C3N3CI3 + 3KC1.
Using about 3 grams of acid, a yield of more than 90 per cent, of the
theoretical is obtained. If a larger quantity than this be used or the
temperature be allowed to rise, the yield is much diminished and the
product is more or less impure.
Trichloriminocyanuric acid is a white, crystalline powder which,
under the microscope, is seen to consist of short prisms. It has a
characteristic odour resembling that of hypochlorous acid. It dis-
solves to some extent in water and glacial acetic acid on heating, but
the greater part undergoes hydrolysis; it is very slightly soluble in
chloroform, but insoluble in light petroleum. It melts at about
Digitized by
Google
HYDROCYANIC, CYANIC, AND CYANURIC ACIDS. 201
245°, Its behaviour is in every way that of a typical nitrogen
chloride. When added to strong hydrochloric acid, chlorine is
liberated, the halogen escaping rapidly with effervescence ; it liberates
bromine from hydrobromic acid, iodine from hydriodic acid ; it
oxidises sulphites to sulphates and sets free sulphur from hydrogen
sulphide, cyanuric acid in all cases being reformed. When added to
ammonia a violent action which may become explosive takes place,
nitrogen is evolved, and cyanuric acid reformed. When boiled
with water, dilute acids or alkaline hydroxides, it is hydrolysed,
cyanuric acid and hypochlorites or the products of their transformation
chlorides and chlorates, being produced.
The. percentage of chlorine was estimated by Carius' method, and
several of the reactions referred to above have been quantitatively
studied :
0-3280 gave 0-6062 AgCI. CI - 4562.
OgCjNjClg requires Cl = 45'76 per cent.
A weighed quantity was added to a solution of potassium iodide,
made acid with acetic acid, and the iodine liberated estimated by
thiosulpbate :
0-4177 liberated I = 107-9 c.c. iV^/10 iodine. CI as IN-Cl = 45-78.
requires 45*75 per cent.
(A )
A weighed quantity was dissolved in acetic acid, an excess of an
approximately decinormal solution of sodium sulphite was added, and
the excess afterwards estimated by a standard solution of iodine.
0-2783 oxidised 71-9 c.c. iV^/10 lifsi^0J2. CI as :N-C1«45-79.
requires 45*75 per cent.
(/%-c»).
A weighed quantity was heated with an excess of strong hydro-
chloric acid, in a current of carbon dioxide, in an apparatus with
ground glass joints {Chem. News, 1899, 85) and the evolved chlorine
passed into a solution of potassium iodide.
0-4689 evolved Cl = 1209 c.c. iV^/lO iodine. CI as IN-Cl = 45*7.
requires 45 75 per cent.
These actions are represented by the equations :
Digitized by
Google
202 HTDBOCTANIC, CYANIC, AND CTAKURIC ACIDS.
(/\ ) + 3HC1 = (/\ ) + SCly
\ ]^-Cl/, \ N-H/,
/ CO \ / CO \
( / \ ) + 6HI = ( / \ ) + sr, + 3H0I.
\ N-Cl/,
+ SH^Oj + 3H,0
{A )
2( / \ ) + SH^^ + 6HC1.
The reaction with hydrogen sulphide cannot he used to estimate
the amount of chlorine attached to nitrogen, as this sahstance, like all
nitrogen chlorides, oxidises a variable amount of the liberated sulphur
to sulphuric acid.
Action 0' Bromine on Potassium Cyanurate,
When bromine is added to a solution of cjanuric acid in the theor-
etical amount of a 5 per cent, solution of caustic potash, a pale
yellow substance separates from the liquid. This on exposure to air
rapidly decomposes, bromine being evolved; it cannot therefore be
freed from water and analysed. It liberates iodine from hydriodic
add an4 violently decomposes ammonia with evolution of nitrogen,
cyanuric acid being reformed in each case. When dried over sulphuric
acid i|i an atmosphere of bromine, a pale orange-coloured powder is
obtained which gives off bromine slowly at the ordinary temperature,
rapidly at 100^, leaving an orange powder having properties similar
to those of the original substance. We have not yet been able to
obtain a product which we could regard as a pure substance, the com-
position varying considerably with slight differences of procedure. A
very large number of analyses of different specimens seems to show that
the body first formed is a bromine additive product of a bromimino-
derivative of cyanuric acid, in which, however, all the imino-hydrogen
of the cyanuric acid is not replaced.
Action of Chlorine and Bromine on Potassium Cyanatc,
Attempts to prepare a chlorimino-derivative of cyanic acid have
hitherto been unsuccessful. When chlorine is passed into a cold solu-
tion of potassium cyanate, it is absorbed, gas is evolved, and a white,
crystalline powder separates, a very pungent odour, somewhat resem-
bling that of cyanogen chloride, being noticed during the reaction.
Digitized by
Google
PERKIN: MYRICETIN. PART II. 203
The white solid thus obtained is, however, cyanuric acid containing a
little (5 to 6 per cent.) trichloriminocyanuric acid ] on treating with
hydrochloric acid or ammonia to decompose the latter and recrystal-
lising from water, pure cyanuric acid is obtained.
The action of b^mine on a solution of potassium cyanate is similar
and results in the production eilher of cyanuric acid or of the pro-
duct already described as resulting from the action of bromine on a
solution of potassium cyanurate. If bromine be added to a 25
per cent, solution of potassimn cyanate, rapid evolution of nitrogen
and carbon dioxide (approximately in the proportion Nj : 200^) takes
place, and the temperature rises to about 80^, if the addition be con-
tinued until an excess has been added and this be then removed by
boiling ; cyanuric acid crystallises out on cooling. If the solution of
cyanate be cooled in a freezing mixture and the bromine be added
cautiously, ^similar effervescence takes place and a yellow solid separ-
ates nrhich in its composition and reactions resembles closely that
obtained by the action of bromine on potassium cyanurate; for
example, it liberates iodine from hydriodic acid and nitrogen from
ammonia, cyanuric acid being in each case produced.
We are at present engaged in a study of various other reactions of
tho cyanogen halogen derivatives which appears likely to throw
additional light on their structure.
St. Baktholomsw's HosrrrAL and Collsgb.
London, B.C.
XXI. — Myricetin. Part 11.
By Arthur Qeorge Perkin, F.R.S.E.
Mtricetin was first isolated from the bark of the Myrica nagi (Trans.,
1896, 69, 1287), and subsequently was found to be pi;^sent in the
leaves of Jihus cariaria, cotinuts, and metopium, in the Myrxca gale,
PUtachia lentiacus, and Hamatoxylon campeachianum. Its molecular
weight is represented by the formula CjgHjQOg ; it forms crystalline
acid compounds, an hexa-acetyl derivative, and by fusion with alkali
yields phloroglucinol and gallic acid ; these facts, together with the
similarity of its dyeing properties and those of quercetin, indicate that
it has the constitution of an hydroxyquercetin. The quantity of
colouring matter available for the above experiments was very small,
as the Myrica nagi contained but 0*27 per cent., and sumach only O'll
per cent. ; moreover, the stock of the former, a material not easy to
obtain, was soon exhausted. Although attempts to accumulate
Digitized by
Google
204 PERKIN; MYRICEtlN. PART 11.
sufficient myricetin were made from time to time, thej had to be
abandoued, and it is only lately, owing to the kindness of Professor
E. Joshitake, of Tokio, that an extract of the Myrioa nagi was
obtained, by means of which the following work could be carried out.
Experimental.
The extract, a brownish-black, brittle mass, was treated with ten
times its weight of hot water^ and when cold the clear liquid was
decanted, the residue again washed twice in a similar manner, and
drained on a porous tile. It was digested with boiling alcohol,
filtered from insoluble matter, and the filtrate evaporated until crystals
separated ; these were collected by means of the pump,* washed, first
with a little alcohol and then repeatedly with increasingly dilute
alcohol until the washings were almost colourless. The* yellowish-
brown residue was crystallised from dilute alcohol, then converted
into its acetyl compound, and the latter, when pure, decomposed with
acid in the usual manner. It was incidentally determined that the
melting point, 204 — 206°, previously given for acetylmyricetin, is
somewhat too low, and should be 211 — 212°. An analysis of myricetin
was again made :
0-1332 gave 02744 COj and 0-0389 H^O. C = 5618 ; H = 3*24.
CigHi^^Og requires C = 56 60 ; H = 3-14 per cent.
When crystallised from dilute alcohol, and allowed to dry, myricetin
has the formula C^gH^QOgyHjO, and this water of crystallisation is best
removed by heating at 160°, although it is almost entirely evolved at
100° :
0-5367 at 160° lost 00290 H,0. Found 5-40.
0-4686 „ 160° „ 0-0249 H^O. „ 5-31.
Theory requires H20 = 5-35 per cent.
Myricetin melts between 355° and 360°. Owing) however, to the
darkening of the tube, it was difficult to be certain to one degree,
although 357° is probably correct.
Bromine Compound, — By the action of bromine on myricetin sus-
pended in glacial acetic acid, a compound was previously obtained
which had the percentage composition of tetrabromomyricetin {Joc> eii,).
Owing to its soluble nature and peculiar dyeing properties, some
doubt as to its constitution was expressed, it being possible that
during the reaction a decomposition had ensued. To determine this
point, the bromine compound was digested for several hours with
boiling hydriodic acid and the product treated with sodium bisul-
♦ Filtrate A (see p. 207).
Digitized by
Google
PER^tN : MTBlOfitlK. PART IL 205
ptiitd soIutioD. The resulting yellow precipitate crystallised from
dilate alcohol in needles which had all the properties of myricetin
and gave a colourless acetyl derivative melting at 211 — 212^. The
compound in qnestion was thus without doubt letrabromomyrioetin.
Methylation qf Myricetin.
Four grams of myricetin, dissolved in boiling methyl alcohol con-
taining excess of methyl iodide, were treated drop by drop with a
solution of eight grams of caustic potash in methyl alcohol, the
addition extending over a day and a half. This procedure was
adopted with the object of preventing an oxidation of the myricetin,
which readily occurs in the presence of alkali. After removal of
unattached methyl iodide and the greater portion of the alcohol by
distillation, the residue was treated with water, extracted with ether,
and the ethereal solution washed with dilute caustic potash solution.
On evaporation, a semicrystalline product remained which was purified
by repeated crystallisation from alcohol :
01163 gave 02630 CO, and 00674 H,0. C«61-67; H-6-48.
0-1134 „ 0-2553 CO, „ 00549 HjO. 0 = 61-40; H = 5-37.
0-1000 „ 0-3040 Agl. OHg« 19-40.
CisHs^gCCHj)^ requires C«61-86 ; H = 5a5 ; CHg* 19-33 per cent.
Myricetin pentamethyl ether forms very pale yellow, almost colourless,
hair-like needles melting at 138 — 139% and is sparingly soluble in cold
alcohol. On acetylation in the usual manner, it gives an acetyl
derivative which crystallises from alcohol in colourless needles melt-
ing at 167 — 170^. Decomposition with acid indicated the presence of
one acetyl group :
0*4342 gave 0-3895 regenerated ether. Found 89*70.
Theory for loss of one acetyl group requires 90*23 per cent.
Myricetin thus contains one hydroxyl group which resists methyl-
ation, and is consequently in the ortho-position to a carbonyl group.
On treatment with alcoholic potash, the pentamethyl ether yields a
yellow potassium salt readily decomposed by water.
On digestion with alcoholic potash at 170^ for three hours, myricetin
pentamethyl ether was decomposed, and from the product of the
reaction an acid and a phenol were isolated. The acid crystallised in
colourless needles melting at 164 — 161% and was found to be gcilHo
acid trimethyl ether.
The viscous, readily soluble phenol yielded an azobenzene deriva-
tive which crystallised from a mixture of alcohol and acetic acid in
orange-red leaflets melting at 250 — 252^ This compound is identical
with that given by rhamnetin, quercetin tetramethyl ether (Proc,
VOL. LXZXT. P
Digitized by VjOOQIC
206 PfiRKWl: MYRICETlN. PART H.
1900, 181), and luieolin trimethyl ether nnder siinilar conditions and
18 consequently disazobenzme phlaroglucinol monomethyl ether. The
phenol is thus phloroglucinol monomethyl ether.
Eihylaiion of Ifyricetin,
Five grams of myricetin dissolved in a boiling mixture of alcohol
and ethyl iodide were treated during 12 hours with a solution of
9 '5 grams of caustic potash in alcohol, drop by drop. The product of
the reaction insoluble in alkali was purified by crystallisation from
alcohol :
0-1064 gave 0-2600 00, and 00682 H^O. 0 = 6664 ; H = 7-12.
0*1124 „ 0-2730 00, „ 0-0710 H,0. 0-66-24; H = 7-01.
^isHACOjHj)^ requires 0 « 6666 ; H =« 699 per cent.
It forms almost colourless needles melting at 149 — 15P, sparingly
soluble in cold alcohol. This compound does not contain a free hydr-
oxyl group, for after digestion with acetic anhydride and sodium
acetate, its melting point and percentage composition (found 0 « 66*56 ;
H = 7'01) were unaltered. Further, this product, on treatment with
sulphuric acid, sustained no loss, 0*4112 and 0*8174 yielding respec-
tively 0*4116 and 0*8173 gram of iinchanged substance. It is thus
without doubt myricetin hexaeihyl ether.
When decomposed with alcoholic potash at 170^, it yielded, like the
methyl ether, an acid and a phenol. The former crystallised from water
in colourless needles or leaflets melting at 111 — 112^. It was found
to be gallic acid triethyl ether :
0-1093 gave 0*2460 00, and 0*0717 H,0. 0 - 61 -38 ; H « 7*28.
Theory requires 0-61-41 ; H- 7*09 per cent.
The phenol dissolved in dilute sodium carbonate solution gave, with
diazobenzene sulphate, a bright yellow precipitate, which was collected,
washed, dried, and purified by several crystallisations from benzene.
It formed glistening, bright yellow needles melting at 163 — 165^, but
on account of its ready solubility in the usual solvents, sufficient was
not available for analysis. From analogy, however, it is probably
azohenzenephlaroglucinol diethyl ether.
When fisetin tetramethyl ether is decomposed with boiling alcoholic
potash, it yields veratric acid and fisetol dimethyl ether,
OH-0<jH8(OOH3)-00-OH,-OOH3,
a fact which enabled Herzig {MoncUeh., 1891, 12, 187) to determine the
constitution of fisetin. At this lower temperature, myricetin hezaethyl
ether is also decomposed, but the products were identical with those
given by alcoholic potash at 170^» and it thus appears likely that the
anticipated phloroglucinol derivative is too unstable to be produced by
Digitized by
Google
PKRKIN : MYBICETIN. PART II. 207
this method. The matter is worthy of further experiment, but owing
to lack of raw material this at present is impossible.
Manopotcutiwn M'yricetin, — ^When a boiling solution of myricetin in
absolute alcohol is treated with alooholic potassium acetate, an orange-
red, amorphous precipitate separates; if, however, a slightly dilute
alcohol be employed, the substance is obtained in a crystalline condi-
tion. It was collected and washed with alcohol, and dried at 100^,
when it assumed a dark green colour. When digested with boiling
water, it is decomposed, with separation of myricetin :
Found K» 11-96. Oi^H^OgK requires K=- 10*95 per cent.
Owing possibly to oxidation, the salt could not be obtained in a
chemically pure condition, but the result is suf&cient to prove that
myricetin reacts in an analogous manner to quercetin and the other
colouring matters of this group.
A Qlucoside of Myricetin,
The alcoholic filtrate (A, p. 204) from the crude myricetin, on stand-
ing overnight, became semisolid owing to the deposition of crystals.
These were drained from the black, tarry mother liquor, washed first
with a little alcohol and then with 50 per cent, alcohol until the filtrate
was nearly colourless. The product was dissolved in boiling water,
filtered from a small quantity of myricetin, and the crystals which
separated on cooling again treated in a similar manner. It was now
twice crystallised from alcohol, and again from water. Myricitrin, the
name proposed for this glucoside, crystallises from water in pale
yellow, almost colourless leaflets containing one molecule of water of
crystallisation; this cannot be removed at 100% but is completely
evolved at 160"":
1-0925 at 160^ gave 00420 HjO. Found 3-84.
1-1390 „ 160° „ 00476 H,0. Found 417.
Theory requires H2Oa3*60 percent.
01186at 100° gaveO-2205COj,andO 0620 HjO. C = 50-74; H = 4-87.
0-1098 „ 160 „ 0-2098 COg „ 00445 H^O. C = 6210; H = 4-50.
01160 ,,160 „ 0-2181 COj „ 00476 HgO. 0 = 51-72; H = 4-59.
Q^^Oy^Jlfi requires C» 50-40; H = 4-80 per cent.
C2,H^Oi3 „ C- 52-28 ;H- 4-56 „
When slowly heated, it sinters at 197° and melts at 1 99-— 200° and is
sparingly soluble in water and absolute alcohol. It dissolves in dilute
alkaline solutions with a pale yellow colour having a faint green tint
and this solution rapidly becomes brown on exposure to air. Aqueous
lead acetate gives a gelatinous, orange-yellow precipitate, and alcoholic
ferric chloride a deep, greenish-black coloration. In appearance, it
P 2
Digitized by VjOOQIC
208 PERKIN: IfYBlCETlK. PART It
cannot be distinguished from qnercitrin, and the dyeing properties of
the two substances are almost identical :
Chromium. Aluminium. Tin. Iron.
Quercitrin. Full brown-yellow. ^11 golden-yellow. Lemon-yellow. Deep olive.
Myricitrin. Full brown-yellow. Full golden-yellow. Lemon-yellow. Brown-olive
DeeompoaUion of tht GlueoMe, — One gram (approx.) of myricitrin,
dissolved in 500 c.c. of water, was treated with 1 c.c. of sulphuric acid
and digested at the boiling temperature for 45 minutes. Crystals of
myricetin separated out, and, after standing overnight, were collected
washed, and dried at 160^ :
1*0835 ail^dried glucoside gave 0*6915 CisH^o^s- Found 63*85.
11660 dried at 100° „ 0*7427 Oi^H^oOg. Found 63*69.
1*0380 „ ,,160° „ 0-6800 OisHioOg. Found 65*51.
CjiHjgOijjHjjO requires C^jH^^Og = 63*60 per cent.
C«H,,0i3 „ 0,,H,o08« 65-97 „
The free colouring matter had all the reactions of myricetin.
The Sugcw. — 'the acid filtrate from the myricetin was neutralised
with barium carbonate, filtered, and evaporated to a small bulk. The
residue yielded a crystalline osazone, which was collected, washed with
a little ether, recrystallised from alcohol, and finally from alcohol and
water. It formed yellow needles melting at 181 — 183°, and was
identical in properties with rhamnose oeazonCf a sample of which was
prepared for comparison from the pure sugar.
Myricitrin, on hydrolysis, thus gives myricetin and rhamnose, and
this reaction may be expressed as follows :
It is analogous to quercitrin which, in a similar manner, yields rham-
nose and quercetin.
Thbobbtioal.
The remarkable similarity between the reactions of quercetin and
myricetin, previously pointed out {loc* ciL), is enhanced by the above
results, and there seems no reason to doubt that myricetin is hydroxy'^
qusreetin.
Oh.
It is interesting that myricetin can so readily be fully etkylaied
with formation of a hexaethyl ether, whereas quercetin, alUiough con-
Digitized by
Google
PERKIN: MTRICETIN. PART II. 209
taixuDg five hydrozyl groups, gives but a tetraethyl derivative. This
distinction is not of importance, in view of the behaviour in this
respect of other members of the flavone group. Apigenin (Trans.,
1897, 72, 805), Ci5H70,(OH)3, has given a dimethyl and a diethyl-
ether; iuteolin (Trans., 1900,77, 1314), C„H^O,(OH)^ a dimethyl
ether of methylluteolin and tetraethylluteolin (Herzig, Ber., 1897,
30, 656), and campherol, Ci^H^02{OH^)f a dimethyl ether of methyl-
campherol (Testoni, Ocuaettay 1900, 30, ii, 327); such results are
thus evidently due to the presence or absence of certain hydrozyl
groups in these compounds, although in what manner they effect the
reaction is not at present ^clear. The resemblance between the dyeing
properties of quercetin and myricetin has been already alluded to,
but it is most interesting that quercitrin and myricitrin should behave
almost identically in this respect. These results indicate as probable
that in both compounds the sugar group is present in the same posi-
tion ; further, it is possible, from a knowledge of the dyeing properties
of some members of the flavone series, to indicate with some certainty
the locality of this in myricitrin at least. The shades produced from
fisetin, quercetin (Trans., 1896, 69, 1287) and myricetin
O OH O OH
Oh/\/\c Q>OH OH^^V^fi <^>>0n
OH 00 — ^^
Myricetin.
°tx>^^>^
CO
Fisetin.
are similar in strength and character, and the resemblance in this re-
spect between quercetin and rhamnetin (quercetin monomethyl ether,
O0Hg>-3) has also been pointed out. Consequently, it is evident
that the hydrozyls 3 and 1 do not appreciably influence the colouring
effect of quercetin or myricetin, the character of which is due to the
orthohydrozyls they contain in conjunction with that present in the
pyrone ring. Now, the dyeing properties of quercitrin and myricitrin
are almost identical with those of morin, the constitution assigned to
which (Trans., 1896, 68, 792) is very similar to that of myricetin,
from which the hydrozyl (4') has been removed. It is thus likely that
myricitrin has the constitution
Digitized by
Google
210 PEBKIN AND BRIQGS :
O OH
OHi
OH CO
and that that of quercitrin may be similarly expressed. The only alter-
native formula for myricitrin is that in which the sugar group has
the position (a); such a compound should, by analogy, dye like
luteolin, the shades of which {loc, cit.) do not widely differ from those
given by morin. Employing the monopotassium derivatives of quer-
cetin and myricetin (loo. cit,), experiments will be carried out with
the hope of preparing glucosides of these colouring matters.
The expense incurred during this work has been largely defrayed by
a grant from the Research Fund of the Chemical Society, and for this
the author desires to acknowledge his indebtedness.
OliOTHWORKBBS' RE8BA.K0H LABOBATOBT,
Dysino Dspabtment,
TOBKSHiaB COLLEOX.
XXII. — The Colouring Matters of Green Ebony.
By Arthur George Perkin, F.R.S.E., and Samuel Henry Clifford
Beiogs, B.Sc.
Green ebony is a yellow dyewood formerly employed to some extent
in this country, but now almost entirely replaced by other colouring
matters. It is native of Jamaica or West India, and according to
the << Treasury of Botany" (1884, p. 437) is obtained from Exco&oaria
glandulosa or JcuMranda ovalifolic^ but the botanical name of that
employed here is not certain, as this information could not be derived
from a specimen of the wood alone. The trunk of the tree is about
six inches in diameter, and the wood, which is very hard and of an
orange-brown colour, stains the hands yellow when freshly cut. Refer-
ences to this dyestuff are meagroi and it does not appear to have been
largely employed. Bancroft {** Philosophy of Permanent Colours," IL,
106, 1813) states, <*The wood known in England by the name of green
ebony, possesses a species of colouring matter very similar to that of
Morus iincioria in dyeing, and is sometimes employed in its stead," and
C, O'Neill (" Dictionary of Calico Printing and Dyeing," 1862) men*
Digitized by
Google
THE COLOURING MATTERS OF GREEN EBONY. 211
tions that it is " used principally in dyeing greens and other compound
shades." Until recently^ it had a limited sale in Yorkshire as a dye
for leather, particularly hat linings, but appears to have entirely
passed out of use as a woollen dyestuff. According to Sir Thomas
Wardle, it is little used in silk dyeing now, but was formerly employed
for greening blacks, or making the shade more jet coloured. It was a
good deal used for making shades less bright or flatter in tone, and in
France in the **avivago " to give a slightly yellowish tint if the shade
required it. The raw material was obtained by purchase from Messrs.
James Kichardson and Sons, of Leeds.
Experimental.
The rasped wood was extracted for six hours with ten times its
weight of boiling water, and the decoction strained through calico.
When cold, the orange-brown solution was saturated with salt, and
the resulting somewhat viscous precipitate (C) collected, drained
upon a porous tile, and allowed to dry. This product was ex-
tracted with boiling alcohol for four hours, the extract evaporated
to a small bulk, and poured into a large volume of ether, which caused
the separation of a dark-coloured, tarry mass devoid of tinctorial
property; this was removed by decantatiou and the ethereal liquid
repeatedly washed with water and evaporated to dryness. The viscous
residue was dissolved in boiling alcohol, alcoholic lead acetate added,
and the resulting orange-red precipitate repeatedly washed with boiling
alcohol and finally with boiling water. It was now suspended in cold
water, decomposed with a few drops of sulphuric acid, and the mixture
of lead sulphate and colouring matter collected, washed, allowed to dry,
and extracted with boiling alcohol. After evaporation to a small bulk,
the solution was poured into ether, the mixture washed with water
until a tarry impurity no longer separated, and after removal of the
ether, the residue was dissolved in boiling absolute alcohol. On
standing overnight, crystaljs of the colouring matter A, separated,
which were collected and washed with alcohol. From the orange-
brown, viscous filtrate, containing chiefly the resin A, on spontaneous
evaporation and resolution in a little cold alcohol, a further quantity
of the crystals was isolated. Different samples of the wood varied
considerably, 8000 grams of a good material yielding 3*16 grams
of crude colouring matter. A, whereas others gave as little as 1 gram,
or even less.
The alcoholic filtrate from the lead precipitate was evaporated to a
small bulk, poured into ether, and the mixture washed with water
until tarry matter no longer separated. After removing most of the
ether, chlorpfonn was added, and the crystals of the colouring matter
Digitized by
Google
212 PEEKIN AND BRIQOS:
B which separated were collected and washed with chloroform. The
filtrate, on evaporation, yielded the resinoos substance B. Eight
thousand grains of the wood usually gave about 17 grams of this
second colouring matter, but this again varied according to the
material employed. From the aqueous filtrate from the precipitate
G, by extraction with ether, approximately 0*3 gram could be isolated
for each kilo, of the wood employed, but this was eventually run to
waate on account of the costly nature of the operation.
As the yield of these crystalline substances was so small, it was
found more economical to collect a considerable quantity of the pre-
cipitate C and work this up in one operation.
The Cohuring Matter B.
The crude material was dissolved in boiling alcohol, the solution
evaporated to half its bulk, and, while hot, treated with an equal
volume of boiling chloroform. The crystals which separated were col-
lected on the pump, washed two or three times with a little cold ether,
and recrystallised in a similar manner :
0 1178 gave 02714 CO, and 00632 HjO. C«62-83 ; H = 601.
01180 „ 0-2720 COg „ 0*0480 H,0. 0 » 62-86 ; H - 4-52.
OigH^Oj requires 0 = 6290 ; H»4-84 per cent.
This substance, which it is proposed to name eoocoecarin, crystallises
in glistening, lemon-yellow needles, sparingly soluble in cold alcohol
or ether, insoluble in benzene or chloroform; when heated, it be-
comes orange-coloured at 210^ and melts with effervescence at
219 — 221°. It is soluble in aqueous and alcoholic alkaline solutions
with a beautiful, violet-red coloration, which is bluer in the latter
case, and in ammonia to form a red-brown liquid, and these solutions,
on exposure to air, are rapidly oxidised and assume a rich brown
tint. With alcoholic lead acetate, no precipitate is formed, and it
yields neither acid compounds with mineral acids nor insoluble salts
with alcoholic potassium or sodium acetates. Sulphuric acid dissolves
it with a brown, and nitric acid with an orange-yellow, tint, but
alcoholic ferric chloride gives no coloration.
Although it does not dye calico with or without mordants, it has a
weak, although decided, affinity for animal fibres, the best results
being obtained by employing 10 per cent, of the colouring matter in
conjunction with 5 per cent, of tartaric or oxalic acid. The shade
produced is a pure pale yellow, but the bath is not exhausted, and the
result, though scientifically interesting, is of no technical value.
When examined by Zeisel's method, it was found that excoecarin
does not contain a methoxy-group. When digested with acetic an*
Digitized by
Google
THE COLOURING MATTERS OF GREEN EBONT. 213
hydride and sodinm acetate in the usual manner, a viscous, readily
soluble product is obtained which could not be obtained in a crystal-
line condition.
The benzoyl compound was first prepared according to Baumann and
Schotten's method, but owing to the readiness with which the alkaline
solution of the colouring matter is oxidised, the yield was poor. The
substance was therefore heated at 175 — 1S5^ with ten times its weight
of benzoic anhydride for 5 hours and the mixture poured into alcohol.
A colourless precipitate separated overnight, and this was collected,
washed with alcohol, and purified by crystallisation, first from alcohol
and finally from a mixture of alcohol and acetic acid :
01 1 10 gave 0-2963 CO, and 00439 HjO. C = 7279 ; H - 4-39.
01106 „ 0-2959 COj „ 00445 HjO. C = 7303 ; H- 447.
01112 „ 0-2968 CO, „ 0 0445 H,0. 0 = 72-79; H = 444.
0^fi^(Q^llfi\ requires 0 = 72-85; H-4-28 per cent.
It consisted of colourless needles melting at 168 — 171% sparingly
soluble in boiling alcohol, more readily in boiling acetic acid, in-
soluble in dilute alkalis. In one operation, this compound was
contaminated with a trace of a more sparingly soluble product
melting at 209 — 211% probably a lower benzoyl derivative, for the
compound melting at 168 — 17P was not altered by further treatment
with benzoic anhydride.
Fution foUh Alkali, — Excoecarin was heated with twelve times its
weight of caustic potash and a little water at 200 — 220° for half
an hour. The rich brown fused mass was poured into water, neutral-
ised with acid, extracted with ether, and the extract evaporated to
dryness. The dark-coloured oil, which on standing became crystalline,
was dissolved in water, excess of sodium bicarbonate added, and
extracted with ether (A), and the aqueous residue neutralised with
acid and extracted with ether (B).
(A) yielded a crystalline residue, which was purified by crystal-
lisation from benzene with the aid of animal charcoal :
00779 gave 0-1928 00, and 00438 H,0. 0 = 6749 ; H - 6-24.
CyHgO, requires 0-67-74 ; H= 6-46 per cent.
It formed colourless leaflets melting at 121 — 123% readily soluble
in water. The alkaline solution becomes brown on standing in air.
Suspecting this to be a hydroquinone derivative, its aqueous solu-
tion was treated with ferric chloride and digested at the boiling
temperature for a short time. The solution was extracted with
ether, the extract evaporated, and the residue sublimed between
watch glasses. The product formed golden-yellow leaflets melting
at 68° and having the reactions of tolujuinone [OH3 : 0 : 0« 1 : 2 ; 5]
Digitized by
Google
214 PERKIN AND BRIQGS:
(Nietzki, Ber.y 1877, 10, 833). Substance A was therefore the
corresponding hydrotoiuquinane, for which Nietzki gives the melting
point 124°. B gave a dark-coloured, syrupy residue from which
crystals separated on long standing. This was dissolved in water,
the solution saturated with salt, filtered from the black, tarry pre-
cipitate, the filtrate extracted with ether, and the extract evaporated.
The residue, on crystallisation from benzene, yielded colourless needles
melting at 199 — 200% which, with aqueous ferric chloride, gave a
deep blue coloration. When distilled, a purple vapour was evolved
and a crystalline product condensed which melted at 160 — 165°; this
was dissolved in water, the solution treated with sodium bicarbonate,
extracted with ether, and the purified substance crystallised from
benzene with the aid of animal charcoal. It gave no coloration with
aqueous ferric chloride, melted at 166 — 168°, and was evidently hydrO'
quinone. Substance B was consequently hydroquinaneearhoxylie
(hydroxyscUicylie) cteid
[C02H:OH:OH = l:2:6].
It seemed probable that this acid was derived from the action
of the alkali upon the hydrotoluquinone first produced, and to de-
termine this, some quantity of the phenol prepared from o-toluidine
was heated with caustic potash under conditions similar to those
previously employed. Although a considerable proportion of the
substance remained unattacked, some hydroquinonecarboxylic acid
was produced, and its presence above, at least in part, must have
originated from the hydrotoluquinone.
Action of Bromine. — Attempts to prepare a bromine derivative of
excoecarin were unsuccessful, as decomposition so readily ensued, but
experiments in the following manner lead to the isolation of a new
product.
Two grams of the substance, dissolved in 25 c.a of a half-saturated solu-
tion of potassium acetate in absolute alcohol, were cooled in a freezing
mixture and bromine added drop by drop until the mixture, at first
green, developed an orange-red coloration. The resulting semisolid,
crystalline mass was collected on the pump, washed with a little
absolute alcohol, then with water, and finally with alcohol. This sub-
stance on standing over sulphuric acid gradually darkened, and on
drying at 100° became olive-green, but it was ascertained that this
colour change was not a decomposition. The yield was 1*65 grams. It
was finally purified by crystallisation from nitrobenzene :
0-1127 gave 0-2608 CO, knd 00435 H^O. C-63-10; H-4-28.
0*1081 „ 0-2506 CO, „ 0 0400 HjO. C-63-22; H-4-10.
OigHjoOg requires C= 63-41 ; H^4-06 per cent.
Thus obtained, it formed flat, copper-coloured needles or leaflets
Digitized by VjOOQIC
THE COLOUBINQ MATTEBS OF QREEN EBONT. 215
Bparinglj soluble in alcohol and melting with decomposition at about
250^. Caustic alkalis dissolve it with a brown tint which, on stand-
ing in air, becomes first olive-green, and finally brownish-black. This
substance, which it is proposed to name exeoecaroney has the composition
of ezcoecarin less 2H, and as the latter has been shown to contain a
hydroquinone group, the above reaction most probably consists in the
oxidation of this to a quinone nucleus. The following experiment
supports this view.
Reduction of Exeotcairone. — A boiling aqueous solution of the sub-
stance was treated with sodium bisulphite, and the digestion continued
until a sample of the liquid, on addition of alkali, gave a violet-red
coloration. It was then acidified, extracted with ether, the extract well
washed with water, and evaporated to a small bulk. On addition of
chloroform, yellow crystals separated, which, after purification, melted
at 219 — 221% and had all the properties of €XCoecar%n, *
Action qf Quiwme. — ^Addition of excess of quinone to a boiling
alcoholic solution of excoecarin caused the formation of a deep brown
liquid, which after a few minutes' digestion was allowed to cool.
Crystals of a deep green colour gradually separated, and these were
ooUected, washed with alcohol, and crystallised from the same solvent :
0-1126 gave 02609 CO, and 00420 H^O. C = 63-24 ; H = 414.
C^H^Oj-CigH^jOg requires 0 = 64-04; H«4-49 per cent.
Thus obtained, the product formed minute, green-coloured leaflets melt-
ing at 190^ with decomposition. Its alcoholic solution is deep brown,
and alkalis dissolve it with a coloration similar to that of excoecarone.
When digested with boiling sodium bisulphite solution, it yields excoe-
carin melting at 219 — 22P. Although complete proof is wanting and
could not be obtained owing to lack of raw material, it is considered
probable that this substance is of the nature of quinhydrone, the hydh>-
quinone nucleus of the excoecarin reacting with the quinone. Its simi-
larity in certain points to excoecarone, on the other hand, made it possible
that it consisted of this substance contaminated with a trace of some
impurity not readily removed by the methods employed.
Methylatum qf Excoecwrin, — Five grams of the substance were dis-
solved in 100 c.c. of methyl alcohol and treated with 20 e.c. of methyl
iodide. To the boiling liquid, a solution of 4'5 grams of caustic potash
in methyl alcohol was added, drop by drop, at intervals extending over
three days, this procedure having for its object to prevent as far as
possible the oxidation of the alkaline solution of the colouring matter.
The product was evaporated to a small bulk, poured into ether, and
the solution washed with dilute alkali. The pale yellow, ethereal
liquidi which had a strong green fluorescence, was evaporated, and the
viscous residue dissolved in carbon disulphide and left overnight-
Digitized by
Google
216 PERKIN AND BRIGQS :
Crystals gradually separated which were collected and recrystallised
two or three times with a mixture of benzene and carbon disulphide.
The mother liquor contained some quantity of an nncrystallisable
resin :
01126 gave 0-2694 CO, and 00600 H^O. C=»65-25 j H = 5-92.
01037 „ 01830 Agl. CH3=ll-26.
0-1077 „ 0-1890 Agl. CHg = ll-20.
'0-1222 „ 0-2116 Agl. CHg-11-04.
C,3Hio05(CHj), requires C = 65-21 ; H = 6'80; CH,- 10-87 per cent
This compound was very troublesome to prepare, at first nothing
but the resin being obtained, due evidently to the too rapid addition
of the alkali, but subsequently yields varying from 0*8 to 1*1 grains of
the crystalline ether were obtained.
It formed glistening, yellow needles melting at 1 1 7 — 11 9%very readily
soluble in alcohol, sparingly in hot carbon disulphide, and is character-
ised by the deep green fluorescence of its solutions. Alkaline solutions
do not dissolve it, and alcoholic ferric chloride gives no coloration.
With sulphuric acid, it gives a deep reddish-brown liquid^ which on
addition of a drop of nitric acid becomes blue-violet, and finally orange
coloured. When dissolved in nitric acid of sp. gr. 1 *64 and the solution
evaporated to a small bulk, a crystalline product separates on cooling,
consisting of oxalic acid and a compound readily soluble in benzene.
The latter crystallises in needles, and is apparently a nitro-compound,
but the yield was too small to permit of its investigation. As excoe-
carin contains three hydroxyl groups but yields only a dimethyl ether,
it would appear to contain one hydroxyl in the ortho-position to a
carbonyl group. Attempts to prepare an acetyl compound in a crystal-
line condition were unsuccessful, the product consisting of an orange-
yellow resin ; a similar result was obtained in an attempt to prepare a
benzoyl derivative.
Experiments carried out on the ethylation of excoecarin gave a
viscous substance, from which, after standing some weeks, a small
amount of crystalline matter separated; owing, however, to its
soluble nature, the yield of substance was too small to admit of
examination.
Moleouiar Weight qf Excoecarin. — For this purpose, the dimethyl
ether was employed with the following result :
0*4146 dissolved in 14*468 acetic acid depressed the freezing point
0*385°. Found, 287.
0*4090 dissolved in 13*380 acetic acid depressed the freezing point
0-420° Found, 283.
This corresponds with the formula Q^^^fi^{Q^^^^21^^ and excoe*
carin itself has thus the formula ^xfliP^*
Digitized by
Google
THS COLOUHINa MAKERS Of GR&&K SBOKY. !2l7
Summary. — The investigation of ezcoecarin shows that this colouring
matter contains three hydrozyl groaps, one of which is not methylated,
and that hy means of fused alkali it gives hydrotoluquinone and
hydroquinonecarhoxylic acid, the latter being derived in part, if not
entirely, from the former. Oxidation with bromine gives excoe^rone,
CigHi^Oj, which, by reduction, is reconverted into excoecarin, and
treatment with quinone yields a compound which is probably a quin-
hydrone compound of the formula O^gK^fi^, The latter results
indicate the presence in the colouring matter of free hydroquinone
hydroxyls. Any further clue to its constitution has not yet been
obtained, though this is probably of a nature not previously met with
among the natural colouring matters. Interesting results would, no
doubt, ensue by a study of the oxidation products of the dimethyl
ether, but the laborious operations involved in the preparation of any
quantity of this substance will render the work extremely slow and
difficult.
^r%0 Colouring Matter A (p. 211).
The crude product was purified by two or three crystallisations
from alcohol with the aid of animal charcoal.
01099 gave 02600 CO, and 0 0465 H^O. 0 - 64-61 ; H - 470.
^14^18^5 requires G « 64*61 ; H - 4*61 per cent.
This colouring matter, which it is proposed to name jaoarandinf
crystallises in glistening, yellow plates or leaflets which commence to
darken at 220^ and melt with evolution of gas at 243—245°. It is
sparingly soluble in alcohol and the usual solvents to form pale yellow
liquids having a green fluorescence, and soluble in sulphuric acid with
a deep orange coloration having a strong green fluorescence. With
caustic alkali solutions, it gives orange-red liquids; with alcoholic
lead acetate, a bright orange-coloured precipitate ; and with alcoholic
ferric chloride, a dark greenish-black solution. When examined by
Zeisel's method, it was found not to contain a methoxy-group.
It dyes mordanted fabrics good full shades, the following being
obtained with mordanted wool :
Chromiiuii. Alnminiom. Tin. Iron.
Dull yellow-brown. Oiange^brown. Bright golden-yellow. Deepolire.
These somewhat resemble those given by luteolin, but are rather
more orange. It has a slight affinity for animal fibres, pale yellow
shades being obtained, which are, however, feebler than those given
by excoecarin*
When treated with minetal acids in the presence of acetic acid, no
crystalline compounds separate, but with alcoholic potassium acetate
a potassium salt is formed. Analysis of this compound did not give
Digitized by
Google
218 P£RETN AND fiEIQOS :
concordant results, owing to the sparingly soluble nature of the
colouring matter, the small portion unattacked separating together
with the salt on cooling the mixture.
Acetyl compound, — This substance was digested with six parts of
acetic anhydride and one of anhydrous sodium acetate at the boiling
point for one hour. On pouring into water, a pale yellow, crystalline
precipitate separated which was purified byrecrystallisation from alcohol.
0-1166 gave 02667 00, and 0-0489 HjO. 0 = 62-37 ; H«4-65.
01211 „ 0-2775 COj „ 0 0535 H^O. 0 = 62*49 ; H = 4-86.
Oi4HioOfi(02H30), requires 0 = 62 79; H = 4-66 per cent.
It formed pale yellow needles sparingly soluble in alcohol and melt-
ing at 192 — 194^ An attempt to determine the acetyl groups in the
usual way, by decomposing the acetyl derivative with sulphuric acid,
gave 72*23 per cent, of regenerated colouring matter, but a slight de-
composition of the latter had ensued, as a resinous substance was also
present It has been found by one of us that some acetyl derivatives
are decomposed by digestion with a boiling alcoholic solution of potass-
ium acetate, and quantitative experiments are now being carried
out to determine if this method is of general application, not only
with acetyl, but also with benzoyl, compounds. Such a process would
be valuable with colouring matters which cannot withstand the action
of strong acids. As colouring matters of the nature of quercetin yield
in this manner the mono-potassium salt, this reaction was applied to
acetyl] acarand in, not only to determine the acetyl groups, but with
the hope of producing a potassium salt which would give some clue to
its molecular weight. A weighed quantity of the acetyl compound
was therefore dissolved in boiling absolute alcohol, treated with excess
of alcoholic potassium acetate, and the solution slowly evaporated. As
soon as crystals commenced to separate, the mixture was left for a
few minutes, the potassium salt collected, washed with methyl
alcohol, dried at 160^, weighed, and analysed. The filtrate and
washings were again evaporated, cautiously diluted with boiling
water containing a few drops of hydrochloric acid, and the crystals of
the colouring matter which separated collected when the mixture was
cold, and weighed. Three distinct preparations of the salt were made,
and two acetyl determinations carried out :
0-3857 gave 00607 K^SO^. K = 705.
0-4295 „ 00665 KgSO^. K = 6-94.
0 7496 „ 0-1180X2804. K-7-05.
OggH^OioK requires K = 6-99.
It formed glistening, yellow needles, insoluble in oold water, some-
what soluble in alcohol.
Digitized by
Google
tfitE dOLOUttllTa MATTEBS Of GB££K £B0NT. 219
Acetyl Determinatiant*
1*4140 gave 1*0682 colouring matter. Found 75*54.
1-8907 „ 1-4141 „ „ „ 74-82.
^14^10^5(^2^8^)2 requires Ci^HigOj- 75-68 per cent.
These results indicate that the molecular weight of jacarandin is
prohablyrepresented by the formula Oj^H^^^ft) ^^^ ^^^^ ^^ gives a diacetyl
compound, O^^K^QO^(C^B.fi)^. Its potassium salt, prepared in the
manner already described, and formed by the replacement of one
hydrogen in a double molecule of the colouring matter, will thus be
aniilogouB to those of rhamnetin and rhamnazin (Trans., 1899, 75, 433).
The benzoyl derivative was prepared by heating the colouring matter
to 180^ for 4 hours with ten times its weight of benzoic anhydride
and pouring the product into alcohol. Yellow, prismatic needles
slowly separated, and, after being left overnight, were collected and
purified by crystallisation from a mixture of alcohol and acetic acid.
It melted at 167—169° :
0-1121 gave 0-2952 00^ and 0-0428 H^O. 0-71 82 ; H = 4-24.
^14^10^6(^7^6^)2 requires C = 71*79 ; H := 4-27 per cent.
Experiments on the methylatum of jacarandin gave a large
amount of resinous matter and also a trace of a crystalline ether
too small in quantity for analysis. The latter forms bright yellow
needles, readily soluble in boiling alcohol, soluble in sulphuric acid,
forming an orange liquid which has a deep green-coloured fluorescence.
It melts at 154-^155°. The paucity of raw material did not permit
of further work in this direction, but it is possible that under altered
conditions a larger yield of the ether could be obtained.
Fusion of a trace with alkali at 200 — 220° yielded a very soluble
acid, which gave a green coloration with ferric chloride, but was not
identical with protocatechuic acid. The presence of a fatty acid,
possibly acetic acid, was also detected.
Unless a more plentiful source of this colouring matter be dis-
covered, but little insight can be obtained into its constitution. Of
the known natural colouring matters, it approaches in general
properties most closely to curcumin, but its molecular weight
(curcnmin, O^^HgoO^, Oiamician and Silber, Ber., 1897, 30^ 192)
does not accord with this supposition. Possibly, therefore, it is a
member of some group hitherto unknown.
The resin A (p. 211) was obtained as an orange-brown, transparent,
brittle mass, resembling jacarandin in many of its properties. It
contained some quantity of this colouring matter, and as no method
has been at present devised by which this can be entirely removed, it
Digitized by
Google
220 THE COLOUBINd MATrBRS OF OREEK KBONt.
has not been closely examined. It dyes mordanted calico similarly,
but not so strongly as jacarandin, but there can be no doubt that
to this substance the tinctorial properties of green ebony are mainly
due. The yield was approximately 0*1 per cent.
The reHn B (p. 212) in appearance resembles the resin A, but it
does not dye mordanted calico. Its reactions coincided with those of
excorearin, with which no doubt it was to some extent contaminated.
It is the main constituent of green ebony, and was present in one
sample to the extent of 0*4 per cent.
Dyeing FropertUa of Green Ehony.
These experiments corroborated those of Bancroft {he. eU,) in that
the shades given by green ebony are of a similar character to those
obtained with old fustic. The colours, however, from the former with
aluminium, tin, and copper mordants are browner, and with iron
greener and paler, than those yielded by old fustic. Employing
mordanted woollen cloth, the following shades were produced :
Ohromiam. Alaminiam. Tin. Copper. IroxL
Doll yellow-brown. Dall brown-yellow. Gk>lden-yellow. Pale brown. Olive-green.
With 40 per cent, of the dyewood, the iron mordant gives greener
and brighter tints than with larger amounts, in which case a browner
colour is produced. Possibly from this green shade and the extremely
hard and compact nature of the wood, the name " green ebony " has
originated. The sample of wood here eniployed possessed half the
colouring power of an average sample of old fustic.
In the earlier stages of this work, which has been in progress
for more than two years, considerable assistance was given by Mr.
B. Gloag Thomson, of Perth, to whom we are much indebted. The
authors also express their thanks ' to the Besearch Fund Committee
of the Chemical Society for a grant which has been in part employed to
cover the expenses of the research.
Olothworksbs' Rksearoh Labobatobt,
Dtkinq Dbpabtmbmt,
YOBKSHIBS GOLLBaS.
Digitized by
Google
ON BRAZIUC ACID AND THE CONSTITUTION OF BRAZILIN. 221
XXIII. — On Brazilic Acid and ilie Constitution of
Brazilin,
By W. H. Pebkin, jun.
In Part I. of thiff research (A. W. Gilbody, W. H. Parkin, jun., and
J. Yates, Trans., 1901, 70, 1401), it was argued that since trimethyl-
brazilin, on oxidation with permanganate, yields ^oarhoxy-^-methoocy'
phanoxyacetic acid and mitah^nipinie ctcid, the constitution of brazilin
must be represented by one of the following formulse :
I. II.
These two formulse are so similar that for a long time it was found
impossible to obtain evidence sufficient to afford even a clue as to which
was the correct one, but ultimately the detailed examination of brazilic
acid {loc, ciL, p. 1410) led to results which show clearly that formula I
is to be accepted as representing the constitution of brazilin.
Brazilic acid^ CijHjjO^, which is produced in a yield of only 0*7 per
cent, by the oxidation of trimethyl brazilin with permanganate, crystal-
lises from water in colourless needles and melts at 129° ; it is a mono-
basic acid, since its silver salt has the formula O^g^ii^S^e* ^^^ ^^
sodium salt the formula Cj^Hi^NaO^ ; when, however, its solution in
water is boiled with baryta, it yields a barium salt of the formula
Cj^HjoBaO^ and therefore, under these circumstances, it behaves like a
dibasic acid, a point which will be referred to later.
Brazilic acid contains 1 methoxy- group, as shown by Zeisel's method,
and when fused with potash, it is decomposed with formation of a readily
soluble acid which in aqueous solution gives an intense violet coloration
with ferric chloride ; it is therefore derived from the resorcyl nucleus
in brazilin.
On treatment with hydroxylamine, brazilic acid yields an oxime, and
with semicarbazide, a semicarbazone ; it therefore contains a carbonyl
group, and this group is evidently situated in the y-position to the
carbozyl group, because, when reduced with sodium amalgam, brazilic
acid yields dihydrobrazilic acid, C^^^^fi^ and this, when liberated from
its sodium salt, spontaneously loses water with formation of the lactone
These facts, although they throw much light on the constitution of
brazilic acid, are not sufficient to establish its formula, but the neces-
vou Lxxxr. Q
Digitized by
Google
222 PERKIN: ON BRAZILIC ACID AND THE
s.iry further information was ultimately obtained by the examination
of the behaviour of the acid with dehydrating agents, especially sul-
phuric acid. When brazilic acid is warmed with concentrated sulphuric
acid, it loses 1 molecule of water and is converted into anhydrdbrazilic
acid, a crystalline substance which melts at 197°, and differs sharply
from brazilic acid in being very sparingly soluble even in boiling water.
This new acid is monobasic and still contains a carbonyl group, since
with hydroxylamine it yields a crystalline ozime, OijHuOgN ; further-
more, a determination by Zeisel's method showed that it still contained
the methoxy-group which was present in the brazilic acid. It is also
an unsaturated acid^ because its solution in sodium carbonate at once
decolorises permanganate, and on investigating this oxidation it was
found that an almost quantitative yield of i^meihoxyaaticylio acid,
OMe
is formed if the oxidation is carried on at a sufficiently low temperature,
a fact which affords a valuable clue as to the constitution of the acid.
A further clue was obtained by the discovery that, when boiled with
baryta water, anhydrobrazilic acid is readily decomposed into formic
acid and a new acid, Cj^HijOg, thus :
C12H10O5 + 2H,0 = CiiHijOe + H-COjH.
This new acid crystallises from water in colourless needles and melts
at 155°; it is a monobasic, ketonic acid and its aqueous solution gives,
with ferric chloride, an intense violet coloration. Since anhydrobrazilic
acid gives no coloration with ferric chloiide, it was probable that the
elimination of formic acid had been accompanied by the formation of
a free hydroxy 1 group in the benzene ring. That this is the case was
proved by the fact that the methyl ester of the acid O^iHigOg, when
heated with sodium methoxide and methyl iodide, is converted into the
methyl ester of an acid, OjjHi^Og, which gives no coloration with ferric
chloride; the hydroxy-group of the former acid had therefore been
converted into a methoxy-group in the latter. The further considera-
tion of the properties of the acid CjiHijOj, taken in conjunction with
the fact that anhydrobrazilic acid on oxidation yields p-methoxysalicylic
acid, seemed to indicate that the acid CjjHj205 is 6-hydroxy-4-methoxy-
benzoylpropionic acid (1), and that its methyl derivative, CuHj^Oj, is
therefore dimethozybenzoylpropionic acid (II),
OMe/NoH OMe/NMeO
1. ' II.
Digitized by
Google
CONSTITUTION OF BRAZILIN. 223
In order to prove this point, it was decided to attempt the synthesis
of the dimethoxy-acid, and after a number of failures this was ulti-
mately accomplished (in conjunction with Mr. E. Ormerod) by treating
a mixture of dimethylresorcinol and the ester of the half-chloride of
succinic acid with aluminium chloride, thus :
OMer ^OMe ^ ciCO-CH^-CH^-OOjEt =
Kj
OMe/^OMe . xrp,
I JcO-CHj-CHj-COgEb "*■ ^^^•
The product of this reaction yielded, on hydrolysis, an acid melting at
147°, which was identical with the dimethoxy-acid, CigHj^Og, obtained
from brazilin, and the constitutions of this acid and of the hydroxy-
metfaoxy-acid, CjiHjjO^, from which it was obtained are consequently
proved to be represented by the formulas II and I given above.
Referring again to the conversion of anhydrobrazilic acid by hydrolysis
with baryta into hydroxymethoxybeuzoylpropionic acid and formic acid,
CijHioOj + 2HjO - OMe-OaH3(OH)-CO-CH2-CH2-COjH + H-OOgH,
we see that we have here a case of a decomposition which has repeatedly
been observed in the plieno-y-pyi'one series.
Thus fisetin, which is somewhat similarly constituted to brazilin,
when digested with alcoholic potash is decomposed into Gsetol and
protocatechuic acid.
0/H.OH VOH
Oh/Y _!. COaH./^OH
\/ \C0-0H,-0H \/
and many other similar examples might be given.
Arguing, then, from analogy, it is evident that the formula of
anhydrol»rasnlic add and its decomposition into methoxyhydroxybenzoyl-
propionic acid must be represented thus :
OMe/y^H +2H.0 -
Anhydrobrazilic acid.
OMe^Y^ + H-CO,H ,
^^^<:!0-CH,«CH,-COjH
Q 2
Digitized by
Google
224 perkin: on brazilic acid and the
and ibis formula is in accordance with all the properties of the
acid.
Since, then, anhydrobrazilic acid is produced from brazilic acid by the
elimination of 1 molecule of water and consequent formation of a
double-linking, it follows that there are only two formulse which can
represent brazilic acid, namely,
OMe/\/ \CH2 ^^^OMe/\/ \cH-OH
\/\qq/C(0H)-CH./C02H *'' '^/'n^q/CH-OH^-OO^H
I. II.
In formula II, the hydroxyl group is in the y-position in relation to
the carboxyl group, and an acid of this constitution should therefore
readily yield a lactone, whereas an acid, ^presented by formula I,
being a )3-hydrozy-acid, would not show any tendency to lactone
formation.
Since brazilic acid is not only stable at 100°, but even when boiled
with hydrochloric acid shows no tendency to pass into a lactone, its
constitution must obviously be represented by formula I.
It has been shown by the preparation and analysis of its sodium,
silver, and barium salts that brazilic acid is a well-characterised, mono-
basic acid, but it is also pointed out (p. 228) that when boiled with
baryta water it yields a very sparingly soluble barium salt of the
formula Oj^^j^O^Ba, which is quite different from the normal salt,
(C]2H^iOQ)2Ba, obtained by precipitating the solution of the sodium salt
with barium chloride. The formation of the salt C^jHi^BaO^ is evi-
dently due to the hydrogen of the hydroxy] group, as well as that of
the carboxyl group, being replaceable by barium, that is to say, the salt
has the constitution :
OMe/ Y ^9^2
^0-Ba^
That the hydrogen of this hydroxyl group should be replaceable by
treatment with caustic alkalis is not surprising in view of its
proximity to the CO group of the dihydropyrone ring.
When brazilic acid is reduced by sodium amialgam, it is converted
into the lactone of dihydrobrazilic acid, and the constitution of this
substance must therefore be represented thus :
ls^XCH^<^(OH).CK,
N) lo
Digitized by
Google
CONSTITUTION OF BRAZILIN. 225
If, now, the two formulse for brazilin given at the beginning of this
imper be examined, it will be seen that only formula I can yield brazilic
acid in a simple manner, and there seems, therefore, to be no reason to
doubt that the constitution of brazilin is :
OH
This formula accounts for all the known properties of brazilin in a
satisfactory manner, and of the four hydroxy 1 groups three are repre-
sented as phenolic and the other one as alcoholic ; it is thus obvious
that three only should be converted into methoxy-groups on treating
brazilin with sodium methoxide and methyl iodide, and this is actually
the case. The tri methyl brazilin so produced, which has been so largely
employed in this investigation, will have the constitution :
In the previous paper {ioc. cit, p. 1403), it was suggested that the
constitution of the dye-stuff, bi^aziUiriy Cj^HjgOg — which contains two
atoms of hydrogen less than brazilin and is produced from it by oxida-
tion— may probably be :
OH
the two atoms of hydrogen being removed, one from the CH(OH) group
and one from the OH group of the resorcyl nucleus. It must, how-
ever, be pointed out that it is quite possible that the two atoms of
hydrogen may be derived from the CHj group and the parahydroxy-
group of the catechol nucleus, and the formula of brazilein would
then be :
OHr^ \/ Nqh f^iIO
Hn i >0H
H
\/\CH/^^\CH^\/
!>
Since, however, brazilein yields such complicated, salt-like compounds
with sulphuric acid (A. G. Perkin and Hummel, Ber,, 1882, 15, 2343), it
Digitized by
Google
226 perkin: on brazilic acid and the
is not improbable that it may be derived from sereral molecules of
brazilin, and therefore have a constitution much more complex than
represented by the formulas given above. This view receives some
support from the fact that it has so far not been found possible to
reconvert brazilein into brazilin by reduction.
Brazilic Acid, O^jHjgO^.
This acid was mentioned in the previous paper (Trans., 1901, 79,
1411), and the method employed in isolating it from the products of the
oxidation of trimethylbrazilin with potassium permanganate was briefly
described. Two analyses were also given, the mean of which (C = 57'l ;
H = 5'0) agrees with the numbers required by the formula C^gH^gO^
(0 = 57*2 ; H = 4*8). The molecular weight of the acid has since been
determined by the cryoscopic method, when two experiments gave 265
and 276, whereas the molecular weight of O^JS.u'^q ^^ ^^^'
Brazilic add melts at 129 — 130^ and is only sparingly soluble in
cold water ; it dissolves readily in hot water and separates, on slowly
cooling, in long, colourless needles. It dissolves readily in alcohol,
ether, or acetic acid, but less readily in chloroform, and is almost
insoluble in cold light petroleum ; in hot benzene, it is readily soluble,
and separates, on cooling, in long needles. That brazilic acid is a
saturated substance is shown by the fact that its solution in cold
sodium carbonate does not decolorise permanganate, oxidation taking
place, indeed, only very slowly on warming. Bromine dissolved in
chloroform is also without action on the acid in the cold. An aqueous
solution of brazilic .acid gives no coloration with ferric chloride, but
when fused with potash the acid is readily decomposed, and, on acidi-
fying and extracting with ether, a syrupy substance is obtained
which is very soluble in water and gives, with ferric chloride, an in-
tense violet coloration.
When boiled with acetic anhydride, brazilic acid dissolves, forming
a yellow solution, but this rapidly becomes brown and then quite
black, decomposition evidently taking place. If a small quantity of
the dry ticid is heated in a test-tube, it chars and gives an oily distil-
late which has a strong odour of coumarin.
Brazilic acid contains one methoxy-group, as is shown by the follow-
ing determinations made by Zeisers method :
01429 gave 01370 Agl. 0CH3= 12'6.
01892 „ 01795 Agl. OCH3«12-5.
CijHjjOg, containing one OOHj, requires OCHj^ 12*3 per cent.
The residues from these methoxy-determinations were decolorised
by sulphurous acid and extracted with ether, when a substance was
Digitized by
Google
CONSTITUTION OF BRAZILIN. 227
obtained which crystallised from water in pale yellow crystals and
melted at about 178^ ; it was not further examined.
Salts of BrazUio Acid, — That this acid is monobasic was first proved
by titration with decinormal sodium hydroxide, using phenolphthalein
as the indicator.
0*21 gram, dissolved in warm water, required for neutralisation
8-1 c.c. of sodium hydroxide solution »0'0324 gram, whereas this
amount of an acid, CisH^j^g* ^^ monobasic, would neutralise 0*033 gram
NaOH.
The solution was then mixed with a further quantity of 11*9 c.c. of
the sodium hydroxide (making 20 c.c. in all) and boiled for 5 minutes,
when, on titrating back, it was found that the amount neutralised
was practically the same as before, namely, 0 034 gram.
The sodium salt, Ci^H^iO^Nsc, separates in glistening plates when a
hot solution of the acid is neutralised with sodium carbonate and then
allowed to cool. The salt was recrystallised from water, dried at
100^, and analysed :
01602 gave 00405 Na^SO^. Na = 82.
CjjHjiOgNa requires Na = 8'4.
This salt, which does not appear to contain water of crystallisation,
is comparatively sparingly soluble in cold water, although it dissolves
readily on warming. On account of the facility with which it crys-
tallises even when impure, it proved to be very valuable as a
means of isolating brazilic acid from mixtures with other acids and
resinous products. The corresponding potassium salt appears to be
readily soluble in water.
The silver salt, C^2^ii^^6^€>> ^^ obtained, on adding silver nitrate to
a neutral solution of the ammonium salt, as a white precipitate which
is very sparingly soluble in water. On analysis
01613 gave 0-2352 COg, 00496 HjO, and 00484 Ag. 0 = 398;
H-3-4; Ag = 30-1.
0-201 gave 0-2932 CO,, 00608 HgO, and 00605 Ag. 0 = 398;
H = 3-3; Ag = 301.
CjgHijOgAg requires 0 = 401 ; H = 3-l ; Ag = 30-1 per cent.
The harium salt, {Q^^^^O^^fiUjd, — When a neutral solution of
the ammonium salt of brazilic acid is mixed with barium chloride, an
amorphous, almost gelatinous, precipitate is at first produced, but this
rapidly becomes crystalline. This salt is readily soluble in hot water,
and separates, on cooling, in slender needles ; after draining on porous
porcelain and exposing to the ^ir for 3 days, it appears to contain
2 molecules of water of crystallisation :
Digitized by
Google
228 PERKIN : ON BRAZIUC ACID AND THE
0*2228 air^ried salt lost, at 100^ 0-0118 » 5*3.
(Cj2Hj|Og)2Ba,2H,0 requires 5-3 per cent, of water.
0 211, dried at 100° gave 00773 BaSO^. Ba = 21-6.
(Cj2HjiOg)jBa requires Ba=21-7 per cent.
A neutral solution of the ammonium salt of brazilic acid gives no
precipitate with calcium chloride, a pale blue precipitate with copper
sulphate, and a white, amorphous, very insoluble precipitate with lead
acetate.
The barium salt, C^j^^oO^Ba, the constitution of which is discussed
in the introduction to this paper, was prepared as follows.
A hot solution of brazilic acid was rapidly mixed with a large excess
of hot baryta water, when a white, granular precipitate rapidly
separated and increased, apparently, on boiling. After boiling for a
few minutes out of contact with air, the salt was rapidly collected on
the pump, washed repeatedly with boih'ng water, dried at 100% and
analysed :
0-4653 gave 02689 BaSO^. Ba = 340.
01449 „ 00824 BaSO^. Ba = 33-5.
CigHj^OgBa requires Ba= 35*5.
Oj2HjQOgBa,H20 requires Ba = 33*9 per cent.
These results, which agree better with the latter formula, clearly
show that brazilic acid, when boiled with baryta, yields a dibasic
barium salt. In order lo be certain that no change in constitution
had taken place during this treatment, the barium salt was ground in
a mortar with a little dilute hydrochloric acid, and the crystalline
precipitate collected on the pump, washed with water, and dried at
100°. It then melted at 129° and consisted of pure brazilic acid.
Oxime of Brazilic Acid, O^gHijOgN. — In preparing this oxime, the
pure acid (05 gram) was dissolved in dilute caustic potash (containing
2 grams KOH), a solution of 2 grams of hydroxylamine hydrochloride
was then added, and the whole allowed to stand for 24 hours. On
acidifying, a flocculent precipitate separated ; this was extracted with
ether, the ethereal solution dried over calcium chloride and evaporated,
and the syrupy residue left over sulphuric acid in a vacuum desiccator
'or several days, when it gradually solidified. As all attempts to re-
crystallise the substance were unsuccessful, it was analysed in its
crude form :
0-2376 gave 12-1 c.c. nitrogen at 22° and 756 mm. N«5-7.
C^jHigOgN requires N = 5'3 per cent.
This oxime dissolves readily in hot water and, on cooling, separates
M an oil.
tSeinicarbazone of BrastHic Acid, CjglTjjOgNg. — Brazilic acid appar-
Digitized by
Google
CONSTITUTION OF fiRAZILIN. 229
entlj combines with semicarbazide only wibh difficulty, as the follow-
ing experiment shows. About I gram of the acid was dissolved in
' hot water and mixed with 15 grams of semicarbazide hydrochloride
and 1'5 grams of sodium acetate and allowed to stand. After a few
days, a thick oil had separated, which, on vigorous stirring, soon
solidified to a mass of minute crystals ; these appeared to consist of
the semicarbazone mixed with small quantities of unchanged brazilic
acid, since the substance contained only 11 per cent, of nitrogen, in-
stead of 13 6 percent, required by the formula O^jH^gO^Ng. The fil-
trate from the crystals deposited, in the course of a few days, a hard,
compact crust of crystals, which were collected, washed well, and dried
at 100°. On analysis :
0-166 gave 184 c.c. nitrogen at 17° and 747 mm. N = 12-7.
CjgHjjNgOg requires N = 136 per cent.
The substance began to decompose at 125 — 126°, then at about
150 — 160° it became quite solid, and a few degrees higher it again
decomposed and was converted into a black mass. That this sub-
stance, although not quite pure, is the semicarbazone of brazilic acid
was proved by dissolviiig it in hot hydrochloric acid, when, on cooling,
the solution became filled with needle-shaped crystals of pure brazilic
acid.
Lactone of Dihydrohrazilic Acid.
Brazilic acid is readily reduced by sodium amalgam with formation
of the sodium salt of dihydrohrazilic acid, and on acidifying this the
y-hydroxy-acid at once loses water "with formation of its lactone.
The pure sodium salt of brazilic acid (0*5 gram) was dissolved in
water, the solution placed in a flat, porcelain dish cooled by running
water, and treated with 3 per cent, sodium amalgam (100 grams) in
small quantities at a time. On acidifying the product, an oily sub-
stance separated which was at first partly soluble in sodium carbonate
and therefore probably contained some hydroxy-acid. In order to
convert the whole into the lactone, the strongly acid liquid was
warmed for a few minutes on the water-bath.
After repeatedly extracting with pure ether, the solution was dried
over calcium chloride and evaporated to a small bulk, when, on stand-
ing, small, colourless, glistening crystals separated : these were col-
lected, washed with ether, and analysed.
01753 gave 03949 COg and 00831 HgO. 0 = 61 4; H = 5-3.
01052 „ 0-2354 COj „ 00498 HgO. C = 610; H = 5-2.
OjjHj^Og requires 0 = 61*0; H = 5*l per cent.
The lactam of dihydrobrazilic acid melts at 142—144° and is
sparingly soluble in dry ether ; it dissolves readily in warm water and
Digitized by
Google
230 PERKIN : ON BRAZILIC ACID AND THE
separates on cooling as an oil which, however, soon crystallines ; it is
readily soluble in alcohol, moderately so in chloroform and benzene,
and sparingly soluble in light petroleum. When heated in small
quantities in a test-tube, it decomposes to a large extent and gives an
oily distillate smelling of coumarin ; this solidifies on rubbing, and on
crystallising from water some of the lactone is recovered, showing that
it distils to some extent without decomposition.
The lactone dissolves in baryta water, yielding a readily soluble
barium salt, but is insoluble in cold sodium carbonate ; on boiling,
however, it dissolves, and the solution, if well cooled and acidi6ed,
remains clear and evidently contains the hydroxy-acid, since, if
heated to boiling and again cooled, the lactone separates. Concentrated
sulphuric acid colours the crystals an intense crimson, and on standing
a deep crimson solution is produced ; this, on warming, becomes at
first more intensely coloured, and then the crimson changes to dark
brown.
Anliydrobrazilio Acid,
This acid is obtained when brazilic acid is treated with sulphuric
acid under the following conditions. Pure brazilic acid (0*3 gram)
is dissolved in 5 c.c. of sulphuric acid and the test-tube containing the
solution plunged into boiling water for 1^ to 2 minutes ; the dark
brown solution is then cooled and mixed with 2 vols, of water when,
on rubbing with a glass rod, a pale yellow, crjstalline substance
quickly separates. The sparingly soluble precipitate is collected on
the pump, washed well, and dissolved in boiling water, a little purified
animal charcoal being added, when, on cooling the filtered solution, a
sandy, crystalline powder is deposited which consists of pure anhydro-
brazilic acid.
In the preparation of this acid, it was found best to always work
with the quantities given above ; if larger quantities are used, the yield
obtained is not nearly so good. From the sulphuric acid mother liquors,
small quantities of the anhydro-acid mixed with some unchanged
brazilic acid may be extracted with ether and the two acids may then
be separated by crystallisation from water.
The total yield of anhydrobrazilic acid obtained is, however, not
more than 50 per cent, of the brazilic acid used, the loss being
apparently due to some of the latter acid becoming sulphonated
during the process of preparation. On analysis :
0-1636 gave 0-3681 COj and 00645 HjO. C = 61 4 ; H = 4-4.
01472 „ 0-3319 00, „ 0-0580 H2O. 0 = 61-5; H = 4-4.
OjjHjqOj requires 0 — 61 5 ; H = 4-3 per cent.
Anhydrobrazilic acid is very sparingly soluble in water and melts at
197^. When heated in a test-tube, it decomposes and gives a brown
Digitized by
Google
CONSTITUTION OF BRAZILIN. 231
oily distillate which, on cooling, solidifieB and has an odour strongly
resembling that of coumarin. It dissolves readily in alcohol, but
is sparingly soluble in benzene, chloroform, or ether, and almost
insoluble in light petroleum. It behaves like an unsaturated acid,
since its solution in sodium carbonate rapidly decolorises permangan-
ata That it is a monobasic acid is shown by the results obtained by
titrating it with decinormal sodium hydroxide.
0*1536 required for neutralisation 0 0264 NaOH, whereas this amount
of a monobasic acid, C^^^^o^s' should neutralise 0*0263 NaOH.
Oxime of Anhydrohrazilic Acid, OigHj^NOg. — In preparing this oxime,
a small quantity of tlfe acid was dissolved in sodium bicarbonate, mixed
with an excess of a solution of hydroxylamine hydrochloride which
had been neutralised with sodium carbonate, and the whole allowed to
stand for 24 hours. On acidifying, the product deposited an oil, but
this rapidly solidified to a granular precipitate, which was collected,
washed well, and recrystallised from boiling water. It was thus
obtained as a sandy powder consisting of microscopic needles, which
melted with decomposition at about 175 — 180^. On analysis :
0-1533 gave 0*3260 CO2 and 00629 H^O. C = 579 ; H = 45.
01567 „ 7*9 C.C. nitrogen at 20° and 760 mm. N = 5-8.
CijHijNOg requires 0 = 57*8 ; H « 4*4 ; N = 56 per cent.
Oxidaiion qf Anhydrohrazilic Acid, Fomicttion of i^Methoxyaodicylic
Acid, OMe-CgH8(OH)-C02H.— The pure acid (0*5 gram) was dissolved
in dilute sodium carbonate and a cold saturated solution of perman-
ganate added drop by drop with constant shaking until the pink
colour remained permanent. After sufficient sodium sulphite had
been added to destroy the excess of permanganate, the whole was
boiled, filtered, and the filtrate and washings of the manganese pre-
cipitate evaporated to a small bulk. On acidifying, a colourless acid
separated which crystallised from water in colourless needles. On
analysis :
0-1 104 gave 0-2299 OOg and 0 0495 HjO. C = 568 ; H = 49.
CgHgO^ requires C = 57*1 ; H = 4*8 per cent.
This acid melted at 156°, and its aqueous solution gave, with ferric
chloride, an intense violet coloration. That it was /^-methoxysalicylic
acid was further proved by mixing it with an equal quantity of this
acidi when the mixture melted at 155 — 156°.
^'Hydroxy-^'ineihoxyhenzoylpropionic Acid,
OMe-OgH8(OH)-CO-OH2-CH3-C02H.
Anhydrohrazilic acid dissolves readily in warm barium hydroxide solu-
tion, and if the solution is boiled a thick, crystalline precipitate of the
Digitized by
Google
232 PERKIN : ON BRAZILIC ACID AND THE
barium salt of hjdroxymethoxjbenzoylpropionic acid rapidly forms, the
separation being complete after 15 to 20 minutes.
If the solution is dilute, the separation does not take place tin til it
is concentrated sufficiently, and then the barium salt separates in
stellate groups. The barium salt is collected on the pump, washed
with water, and dissolved in warm dilute hydrochloric acid, when, on
standing, the free acid separates in colourless needles, and by recrys-
tallisation from water is readily obtained pure, in long threads
somewhat resembling crystals of sublimed phthalic acid. On analysis :
0-17U gave 0 3702 00^ and 0-0840 Hfi. 0-58-9; H = 5-4.
01541 „ 0-3329 COg „ 00756 H^O. 0'=58-9; H = 5-4.
OiiHjgOg requires 0 = 58-9 ; H = 5-3 per cent.
^'Hydroxy- i-methoxyhenaoylpropionic acid melts at 155 — 156° and
dissolves readily in hot, but is rather sparingly soluble in cold, water ; its
aqueous solution gives an intense violet coloration with ferric chloride.
The barium salt, obtained by the action of barium hydroxide on
anhydrobrnzilic acid in the way described above, after drying at 100°,
was analysed with the following results :
0-3987 gave 0 2440 BaSO^. Ba = 36-1.
Oj^HjQ05Ba,H20 requires Ba = 36'4 per cent.
From this it would appear that the barium salt prepared in this
way has the formula 0||HjQOj^Ba,H20, that is to say, that the barium
has replaced, not only the hydrogen of the carboxyl group, but also
that of the phenolic hydroxyl group.
Such cases as thie have often been observed before ; thus salicylic
acid yields a barium salt, C7H^03Ba,2H20, which is very sparingly
soluble in water.
The filtrates from several preparations of the barium salt of hydr-
oxymethoxy benzoyl propionic acid were mixed and the excess of barium
hydroxide removed by passing carbon dioxide through the boiling
solution.
The filtrate deposited, on evaporation, a further small quantity
of the insoluble barium salt; this was removed by filtration, and
the concentrated solution, which contained a considerable quantity of
a barium salt, carefully tested for formic acid. This was easily proved
to be present, not only by the fact that the solution blackened silver
nitrate, but also because it readily reduced mercuric chloride to mer-
curous chloride. There can, therefore, be no doubt that anhydro-
brazilic acid is decomposed by boiling with barium hydroxide into
formic acid and hydroxymethoxybenzoylpropionic acid, according to the
equation
C^HioOb + 2H3O = H-00,1I + C„H,A-
Digitized by
Google
CONSTITUTION OP BRAZILIN. 233
Diimthoxyhenzoyljrroinonie Acid, Q^JipKQ\*Q0'QB.^^QU^^CO^>
It was necessary to prepare this acid, in order to compare it with
the acid obtained synthetically by the action of aluminium chloride on
the mixture of dimethylresorcinol and the ester uf the half-chloride of
succinic acid (see next section).
Sydroxymethoxybenzbylpropionic acid (I gram) was dissolved in
12 grams of methyl alcohol, 3 grams of concentrated sulphuric acid
added, and the whole heated to boiling in a reflux apparatus for
4 hours.
On adding water, a crystalline substance separated, which, since it
was insoluble in sodium carbonate, evidently consisted of the methyl
ester, OMe-OgH3(OH)-CO-0H,-CHj-CO2Me. In this condition, it
melted at about 85°, and its alcoholic solution gave an intense violet
with ferric chloride ; it was, however, not further purified, owing to
the necessity for using the whole quantity for further methylation.
The dry methyl ester (0*8 gram) was dissolved in dilute sodium meth-
oxide (containing 0*1 gram Na) and heated in a sealed tube with
3 grams of methyl iodide at 120° for 2 hours; the tube was then
opened, the same quantity of sodium methoxide and methyl iodide
again added, the tube resealed, and heated at 120—130° for 3 hours.
The product was poured into water, the oily precipitate extracted with
ether, the ethereal solution well washed with water, and then three
times with dilute soda in order to extract some unmethylated ester
which was present. The ethereal solution was then evaporated and
the residual oil hydrolysed by boiling for a few minutes with methyl
alcoholic potash. Water was then added and the clear solution evap-
orated until free from methyl alcohol, when, on the addition of hydro-
chloric acid, a crystalline acid was deposited, which was collected,
washed with water, dried on a porous tile, and recrystallised twice from
benzenor On analysis : *
01260 gave 02779 CO, and 00673 H,0. C = 606 ; H « 60.
Oi^Hj^Oj requires C « 60*5 ; H « 5-9 per cent.
Dimeihoxyhenzoylpropianie acid melts at 146 — 148°, with slight previ-
ous softening, and when kept a short time a few degrees above its
melting point, it becomes a deep violet colour. It is readily soluble
in alcohol and in hot benzene, but sparingly so in cold benzene ; its
alcoholic solution gives no coloration with ferric chloride.
Digitized by
Google
234 OK BRAZILIC ACID AND THE CONSTITUTION OF BRAZILIN.
Synthesis of Dimethoxf/benzoi/lpropionic Add.
[With E. Ormerod.]
In Bjnthesising this acid, a process was employed which is some-
what similar to that recommended by L. Bouveault {BtUL Soe. Chim.,
1897, [iii]) 17, 333) as suitable for cases of this kind.
Dimethylresorcinol (8*4 grams) was dissolved in nitrobenzene
(17 grams) and carbon disulphide (35 grams), and then mixed with
the ester of the half-chloride of succinic acid, Cl-CO-CHj-CHj-OOjEt
(10 grams). Powdered aluminium chloride''^ (8 grams) was then added
in small quantities and, after standing for some hours, the mixture
was gently heated on the water-bath for a few minutes.
The product was treated with water, the oily layer well washed
with water, and the carbon disulphide removed on the water- bath ;
the residue was then distilled in steam until the nitrobenzene and
small quantities of unchanged dimethylresorcinol had been removed.
The non-volatile oil was extracted with ether, the ethereal solution
evaporated, and the brown oil hydrolysed with a slight excess of
alcoholic potash. After evaporating off the alcohol, the residue was
dissolved in water and acidified, when a brown solid separated,
which was collected on the pump and purified by repeated recrys-
tallisation from benzene. On analysis :
01659 gave 03465 COj and 00834 Hfi. 0 = 606 ; H = 5-9.
OjjHj^Oj requires O = 60'5; H = 5*9 per cent.
The synthetical dimethoxybenzoylpropionic acid thus obtedned melted
at 146 — 148° with slight previous softening, and the melted mass be-
came a deep violet colour a few degrees above this temperature.
That it is identical with the acid of this composition obtained from
anhydrobrazilic acid as described above, was further proved by mix-
ing the two acids, when no alteration in the melting point could
be observed.
In conclusion, I wiah to express my thanks to Mr. J. Yates for his
valuable assistance in cai*rying out this investigation, and I wish also
to state that much of the expense incurred was defrayed by repeated
grants from the Royal Society Fund.
Thb Owens Collbge,
Manorestbr.
* Prepared from altiminium.
Digitized by
Google
BRAZILIN AND HEMATOXYLIN. PART III. 285
XXlV. — Brazilin and HcBmatoxylin. Part 111. The
Constitution of HcBmatoxylin,
By W. H. Perkin, jud., and J. Yates.
Hjematoxylin is the colouring matter of HcRmatoxylon campeehianum^
a tree which beloDgsto the family CcBsalpiniac«cs,SLnd the wood of which
appears to have been first imported into iilarope by the Spaniards
shortly after the discovery of America. The heart-wood of this tree,
known as logwood in this country, is still largely used in cotton and
wool dyeing for the production of blacks and greys, and it is also em-
ployed on account of its cheapness along with other colours, such as
indigo, for dyeing compound shades.
The importance of hsBmatozylin may be judged from the fact that it
has probably been, and perhaps is still, more largely used than any
other natural colouring matter, not even excepting indigo. The dye is
sent into the market either as a concentrated liquid extract or
in the solid form, the rasped wood, in either case, being extracted with
water and the extract evaporated in vacuum pans.
For dyeing cotton, or in calico printing, logwood extract is generally
used in conjunction with an iron, chromium, or aluminium mordant ;
the two first-named give intense black, whereas, with the latter, grey-
ish-violet shades are produced. Very large quantities of the liquid
extract are still used in wool dyeing for the production of blacks, the
wool being immersed alternately in the extract, and in a bath of potass-
ium dichromate and sulphuric acid. Again, in silk dyeing, logwood
extract is very largely used in conjunction with an iron mordant for
the production of blacks. It is worthy of note that hsematoxylin and
iron alum have long been used as one of the most important stains for
microscopical sections of animal tissues, but, quite recently. Professor
S. S, Hickson {Quart. J, Micros. JSci^ 1901, p. 469) has shown that
brazilin used with iron alum gives results which are even more
satisfactory.
The actual colouring matter of logwood was first isolated by Chevreul
{Ann. Chim. Phys., [ii], 82, 53, 126) in 1810, who obtained it by
extracting the wood with water, evaporating the extract to dryness,
and digesting the residue with alcohol, which dissolved the heematoxylin,
but not the other substances present.
After distilling 'off the alcohol, the residue was allowed to stand in
contact with water, when the hsematoxylin separated in crystals. Pure
hsematoxylin is now comparatively easily obtained from the dark
coloured crusts which gradually separate when the casks of concentrated
logwood extract are allowed to stand in a cool place. The crude mass
Digitized by
Google
236 PERKIN AND YATES: BRAZILIN AND HEMATOXYLIN.
is ground to a fine powder, extracted repeatedly with ether, the ether-
eal solution evaporated, and the residue left in contact with water
when dark coloured crystals separate, which, by recfystallisation from
water containing a small quantity of sodium bisulphite, may be obtained
colourless, and consist then of pure hsematoxyiin.
Hsematoxylin crystallises in tetragonal prisms with 3 molecules of
water of crystallisation, and is sparingly soluble in cold, but readily in
hot, water ; it dissolves in alkalis, forming an intense purple-coloured
solution. It is strongly dextrorotatory, a 1 per cent, aqueous solution
having a rotation of 1*85° in a 200 mm. tube; it also reduces Fehling's
solution and silver nitrate in the cold.
The first attempt to determine the composition of hsematoxylin is due
to Erdmann, who in 1842 {AnncUen, 44, 292) proposed the formula
C4QH17O5; this he subsequently altered to CjgHi^Og, for the an-
hydrous substance and this formula, which was confirmed by Hesse
{AnruUen, 1859, 109, 332) and by other workers, is now accepted as
correct.
From very early times experiments were made with the object- of
obtaining evidence as to the nature of this important dye-stuff, but the
results in most cases did not afford any clue to its constitution. The
literature bearing on the subject is, in fact, so extended, that it is only
possible in this paper to give a very brief outline of such of the work
as produced really valuable evidence of the nature of hsematoxylin.
Reim (^er., 1871, 4, 331) showed that wheu an ethereal solution of
hsematoxylin is mixed with a few drops of nitric acid, oxidation takes
place and haemate'in, CjgHjgOg, is produced. J. J. Hummel and A. Q.
Ferkin {Ber,, 1882, 15, 2337) subsequently obtained hsematein pure in
the form of reddish-brown crystals by leading air through a solution of
the dye-stuff in ammonia. The action of fused caustic potash on heema-
toxylin was also first investigated by Reim {loc. cit,, 332) and from the
fused mass he was able to isolate considerable quantities of pyrogaUoly
a result of great importance, since it showed that hsematoxylin is prob-
ably a derivative of this substance. R. Meyer (5er., 1879, 12, 1392)
submitted hematoxylin to dry distillation, and obtained a distillate
which he stated contained resorcinol, as well as pyrogallol, but the
qualitative tests by which he claimed to have recognised the former
of these were quite insufficient, and it will be shown in this paper that
resorcinol is not formed in this way. That hsematoxylin contains
several hydroxy! groups was clear from early times, and in order to
determine how many such groups were present in the molecule, Reim
{loc, cU,, 331) investigated its behaviour with acetyl chloride, and
obtained a substance which he considered to be hexa-acetylha3matoxylin,
CigHgOg(C2HjO)g, and as a result of this he suggested the first consti-
tutional formula for hsematoxylin, namely.
Digitized by
Google
PART III. THE CONSTITUTION OP HEMATOXYLIN. 237
OH OH
-OOO-'
OH OH
Erdmann and Schultz {Annal&n, 1883, 216, 234), by determining the
number of acetyl groups in this substance, subsequently showed that
Reim's substance was penUA-cbcetylkmrruOoxylin and had the formula
0^^^{Q^fi)fiQ, and therefore hsBmatoxylin contains only five
hydrozyl groups. This result was confirmed by the investigation of
the action of sodium methoxide and methyl iodide on hsBmatoxylin
when, under certain conditions (Herzig, MoncUsh., 1894, 16, 143),
tetramethylhsBmatoxylin, 0^^'BL^QO^{OM.e)^ is produced a substance
which still contains a hydroxyl group, since on treatment with acetic
anhydride it yields acetyltetramethylhsematoxylin,
0„HgO(OMe)4-OOjH30.
It is thus shown that hsematoxylin, like brazilin, contains an alcoholic
hydroxyl group which is not methylated by the action of sodium meth-
oxide and methyl iodide, and indeed, from the general similarity between
brazilin and hematoxylin, it has long been considered probable that
these two colouring matters are closely allied in constitution.
During the investigation of brazilin, we have also carried out a
number of experiments on hsematoxylin, the results of which, in our
opinion, not only prove the close relationship of these two colouring
matters, but also enable us to assign a formula to hematoxylin
which we believe correctly represents the constitution of this
substance.
In the present paper, we describe only the experiments on the oxida-
tion of tetramethylhsBmatoxylin with permanganate, and leave for a
future communication the description of the interesting substances
which have been obtained by oxidising tetramethylhaematoxylin with
chromic acid.
We have also carefully repeated B. Meyer's (loe, eit.) experiments on
the distillation of hematoxylin, and show that pyrogallol (but no re-
sorcinol) is produced in considerable quantity in this way. We are
thus able to confirm Reims' discovery {loe. eit) that hematoxylin is
probably a derivative of pyrogallol.
When tetramethylhematoxylin is oxidised with permanganate, it
yields m-hemipinie add, an important result, because it shows that
hematoxylin, like brazilin, contains the catechol nucleus. Briefly stated,
then, the molecule of brazilin is made up of a resorcinol and a catechol
nucleus, whereas hematoxylin contains a pyrogallol and a catechol
nucleus.
From the product of the oxidation of tetramethylhematoxylin, we
VOL. LXXXI. R
Digitized by
Google
238 PERKIN AND YATES : BRAZILIN AND HiSMATOXTLIN.
have also been able to isolate considerable quantities of a beautifully
crystalline acid, O^^H^i^.^, which melts at 215°.
This acid is dibasic, since it yields a silver salt, C^^'RjQO^A.g^ ; it is
also converted into the anhydride OijH^qOq by treatment with acetic
anhydride, and as it contains two methoxy-groups and yields pyrogallol
when heated in a sealed tube with hydrochloric acid, its constitution is
evidently represented by the formula :
OMe
OMe/No-CHa'COjH
JC0,H
^cc
that is to say, it is ^-carboxy-b : Q-dimelhoxyphenaxyaeetie acid. This acid
is therefore exactly similar to the 2-carhaxy-6'methoasi/phenoxy(ii€etie cund :
OMe/No-OHjj-COgH
which was obtained from trimethylbrazilin by oxidation with per-
manganate, from which, indeed, it only differs by containing an extra
methoxy-group. The close relationship between brazilin and hsema-
toxylin which owing to the general similarity in the properties of these
two colouring matters has so long been considered probable, is now
clearly proved an4 may be briefly expressed in the following way.
Trimethylbrazilin, on oxidation, yields 2-carboxy*5-methoxyph6noxy-
acetic acid and m-hemipinic acid, and therefore its constitutional
formula must contain the two nuclei :
and
'CC^
These two combined together lead, then, to the formula,
oh/Y ^<?h f^P^
which, as shown in the preceding paper, there is every reason to
believe represents the constitution of brazilin.
Similarly, tctramethylheematoxylin, on oxidation with permanganate,
yields 2-carboxy-5 : 6-dimethoxyphenoxyacetic acid and f»-hemipinic
acid, and therefore haematoxylia must contain the two nuclei :
OH
Digitized by
Google
PART III. THE CONSTITUTION OP HiEMATOXYLIN. 239
These combined together in the same way as in the construction of
the brazilin formula give the expression
OH ^
oh/Y \ch i^OH
as representing the constitution of hsematozylin, and as this accounts
in a satisfactory manner for all the known decompositions of this
colouring matter, there can be little doubt that it is correct.
Besides the acids mentioned above, another acid is formed in the
oxidation of tetramethylhsamatoxylin with permanganate in consider-
able quantities ; this melts at 180°, has the formula CjoH^qO^q, and has
been named hcBmcUoxylinie aeid.
This new acid, which it will be seen contains the same number of
carbon atoms as tetramethylhsematoxylin, is dibasic, since its silver salt
has the composition C^oH^gAgjO^Q.
On reduction with sodium amalgam, it is converted quantitatively
into an acid, O^oHgoOg, which is a monobasic lactonic acid, and it is
therefore evident that the latter is produced in two stages, thus :
^20^20^10 + 2H = ^20^22^10
HaBmatoxyliiiic acid. Dihydrohematoxylinio
acid.
^20^22^11 - ^2^ =* ^20^20^9
Lactone of dihydro-
hsematozylinic acid.
Hsematoxylinic acid corresponds in all its properties with brazilinic
acid, Ci^HigOg (Trans., 1901, 79, 1411), from which it differs only by
one methoxy-group, and it is extremely probable that its constitu-
tion is :
OMe
OMe/V^H f^OMe.
LA 6hO /^OMe
C0,H COjH
The lactone of dihydrohsematoxylinic acid would then be
OMe Q
OMe./'V^ \CH f^OMe,
\/vs'
COjH
or the GOgH of the catechol nucleus might of course take part in the
B 2
Digitized by
Google
240 PERKIN AND YATES : BRAZILIN AND HEMATOXYLIN.
lactone formation. To prove conclusively the constitutions of brazil-
inic and hsematoxylinic acids is a matter of great difficulty, and a
description of the large amount of experimental work which has been
accumulated with this object will, it is hoped, form the subject matter
of a future communication.
Oxidation of Tetramethi/lhcBmcUoxylin* with Permanganate.
This oxidation, like the oxidation of trimethylbrazilin, was carried
out under a variety of conditions, but the following method was the one
ultimately adopted in preparing the substances described in this paper.
Tetramethylhsematoxylin (10 grams) is ground up with a little
water into the finest possible paste, washed into a three litre flask
with 100 c.c. of water, and then 100 c.c. of a cold saturated solution
of permanganate added and the whole allowed to stand at the ordinary
temperature until the colour has disappeared. A further 100 c.c. of
permanganate are then added, and the operation continued until, after
standing for 10 — 12 hours, the liquid still remains pink. The excess
of permanganate is then destroyed by adding sodium sulphite and the
product, after heating to boiling, filtered on the pump ; the maganese
precipitate is then extracted twice with boiling water and the combined
aqueous solutions nearly neutralised with hydrochloric acid and
evaporated to a small bulk, but not to dryness.
On acidifying the residue with hydrochloric acid, a small quantity of
a black, tarry substance separates and is removed by filtration, the
red fitrate is then shaken with chloroform, which causes a small
quantity of a crystalline substance (A) to separate ; the filtrate from
this is then repeatedly extracted with chloroform. The chloroform
extract is washed with a little water, dried over calcium chloride,
when, on distilling off the chloroform, a semi-solid tarry mass remains
(B). The solution, after treatment with chloroform, is saturated with
ammonium sulphate and extracted at least 10 times with ether ; the
ethereal solution is dried over calcium chloride and evaporated and in
this way a brick-red, crystalline mass is obtained (0).
* The haBmatoxylia used in this research was obtaioed from Eahlbanm and was
always of excellent quality. The preparation of the tetrameihylhtematoxylin was
carried out almost exactly in the way described in tho case of trimethylbrazilin
(Trans., 1901, 79, 1403). On pouring the product of methylation into water and
allowing the whole to stand for a few days, in almost all cases, part of the tetra-
niethylheematoxylin separated in roseate groups of straw-coloured needles; these
melted at 66 — 70° ami contained water of crystallisation. After drying at 100* and
recrystallising from alcohol, the substance was obtained in an anhydrous condition
melting at 140 — 142**. The rest of the tetrametbylhsmatoxylin was extracted with
ether and purified as in the case of the trimethylbrazilin : but in tho oxidation
experiments, much better results were always obtained with the crystals melting at
65—70* than with the anhydrous substance.
Digitized by
Google
PART III. THB CONSTITUTION OF HiEMATOXTLIN. 241
The Substance A it 2-Carboxt/-6 : 6-dimethoxi/phenoxyaeetie Acid,
OMe
OMe/No-OHj-COoH
The substance A crystallises from glacial acetic acid in colourless
needles and so easily that even when much contaminated with resin,
as was frequently the case, it separates at once from this solvent in an
almost pure condition. After two crystallisations, the following
numbers were obtained on analysis :
01730 gave 0-3268 COg and 0-0753 H^O. 0 = 51-4 ; H = 4-8.
0-1153 „ 0-2179 COj „ 0-0495 HjO. 0 = 51-5 ; H = 4-8.
CiiHjgOy requires 0 = 51-6 ; H = 4-7 per cent.
2-Carhoxy'b : 6'dimethaxt/phe7U>xyacetic acid melts at 214 — 215° and
at about 225 — 230° decomposes with evoli^tion of gas. It is almost
insoluble in cold water, but is dissolved slightly by boiling water and
is deposited on cooling in needles ; it is sparingly soluble in cold acetic
acid, but dissolves readily on boiling.
The methoxy-groups were determined by Zeisel's method :
0-2703 gave 05148 Agl. MeO = 25-2.
002H-OeH2(MeO)20-OH2-COjH requires MeO = 24-2 percent.
The silver salt, O^^H^oO^Agj, is precipitated on the addition of silver
nitrate to a neutral solution of the ammonium salt as a white, gelatin-
ous precipitate which is difficult to wash. On analysis :
0-2251 gave 0-2343 00,, 00457 HgO, and 01035 Ag. 0 = 28-3 ; H = 2-2 ;
Ag«46-0.
OjiHi^OyAgj requires 0 = 28-1 ; H = 2-l ; Ag = 45-9 per cent.
A neutral solution of the ammonium salt gives a white, gelatinous
precipitate with calcium chloride, but with barium chloride no precipitate
is produced until the solution is boiled, and then a very sparingly
soluble, crystalline salt separates. When fused with potash, the acid
gives a brown mass, and the solid acid obtained from this by acidifying
and extracting with ether gives all the reactions of pyrogallol.
In order to clearly prove that this acid is a derivative of pyrogallol,
a small quantity was heated with hydrochloric acid in a sealed tube for
2 hours at 180°. On opening the tube, carbon dioxide escaped, and the
liquid contained small, black flocks in suspension ; these were removed
by filtration. The filtrate was saturated with ammonium sulphate, ex-
tracted five times with ether, and the ethereal solution, after drying
with calcium chloride, evaporated, when a brown oil remained
Digitized by
Google
242 PERKIN AND TAXES: BRAZILIN AND HAMATOXYUN.
which was distilled under reduced pressure from a small retort. The
distillate, which was a light brown oil, crystallised at once on rubbing
with a trace of pjrogallol, and, in contact with porous porcelain, the
dark miother liquor was rapidly absorbed, leaving a colourless, crystal-
line residue. After crystallising from benzene, this melted at 1 25 — 1 30°,
gave with ferric chloride a blue colour rapidly changing to brownish-
green, and with nitrous acid a yellow colour changing to brown, and
since when dissolved in potash and left exposed to air a deep brown
solution was obtained, there can be no doubt that the substance is
pyrogallol (m. p. 132°).
Anhydride of Garhaxi/dimethoxt/phenoxyacetic Acid.
When this acid is heated in a small flask under reduced pressure, it
first melts, then water is eliminated, and an oily distillate is obtained
which has a penetrating odour like that of formaldehyde, and which, on
cooling and rubbing, partly solidifies. When this was ground up with
ether, some dissolved, but a portion remained insoluble as a white, crys-
talline powder which softened at 150° and melted at 175°. This
substance dissolved only partly in cold dilute sodium carbonate, and
the solution, on acidifying, deposited a considerable quantity of the
unchanged acid melting at 214°. The residue consisted of the crude
anhydride of the acid.
It appears, therefore, that on distillation under reduced pressure the
acid is partly converted into the anhydride and partly distils un>
changed.
Subsequently, the pure anhydride was obtained by heating the acid
in a reflux apparatus with acetic anhydride for 20 minutes. The solu-
tion, on standing over potash in a vacuum desiccator, gradually de-
posited square plates from which, after drying at 100°, the following
numbers were obtained on analysis :
01488 gave 0-3013 COg and 00576 H2O. C * 552 ; H - 43.
CjjHi^O^ requires C = 55'4 ; H = 4-2 per cent.
This substance is therefore the anhydride qf earhaxydimethoxyphenoocy-
(icetic acid.
The Substance C is Metahemtpinic Acid, ^JJ J jc Jh"
The substance C, obtained from the ethereal extract (p. 240) of the
product of the oxidation of trimethylhaematoxylin with permanganate,
was a brick-red powder sparingly soluble in ether. In purifying it, it
was first washed on the pump with ether and then crystallised from
water with the aid of animal charcoal, and was thus obtained in almost
Digitized by
Google
PABT m. THE CONSTITUTION OF HiEMATOXTLIN. 243
coloorlesSy glistening needles which melted at 190° with decomposition
and consisted of pure mrhemipinic acid. On analysis :
01284 gave 02496 CO, and 0 0529 H,0. C = 530 ; H = 4-6.
CjqHiqOh requires C = 53*1 ; H= 4"4 per cent.
As it was most important to be certain of the identity of this acid,
it was next converted into the silver salt, which was obtained as a
granular precipitate on adding silver nitrate to a neutral solution of
the ammonium salt. On analysis :
0-1489 gave 0-1517 CO,, 00279 H^O, and 00728 Ag. 0 = 27-7;
H-21> Ag = 48-8.
^io^8^«-^^s requires C« 27*3 ; H = 1-9 ; Ag= 49*1 per cent.
Lastly, the characteristic ethylimide was prepared by dissolving the
acid in ethylamine, evaporating to dryness, distilling the residue, and
crystallising the distillate from alcohol. The yellow needles which
separated melted at 230° and consisted of pure m-hemipinethylimide.
On analysis :
01829 gave 9*6 c.a of nitrogen at 13° and 732 mm. K = 62.
CijHjjO^N requires N = 5-9 per cent.
There can therefore be no doubt that the acid is m-hemipinic acid,
a very important fact, since it proves that hsematoxylin contains the
catechol nucleus.
Hasmatoxylinic Acid, O^oH^o^io*
This acid is contained in the chloroform extract (B) of the product
of the oxidation of tetramethylhsematoxylio, and is evidently that
derivative of hsematoxylin which corresponds with brazilinic acid
(Trans., 1901, 79, 1411) obtained in a similar way from trimethyl-
brazilin. The tarry residue left after distilling off the chloroform
was boiled with a large quantity of water, when all dissolved except a
small quantity of dark coloured, resinous matter. The aqueous
solution was mixed with an excess of basic lead acetate, and the
pale yeUow, amorphous precipitate which separated was collected
and well washed with water. The precipitate was then ground
up with water to a fine paste and decomposed with sulphuretted hydro-
gen, first by passing the gas in the cold, and afterwards through the
boiling liquid. After filtering, the precipitate was again heated with
water and sulphuretted hydrogen, and the combined, nearly colourless
filtrates were then evaporated to a small bulk. The concentrated
aqueous solution, on standing, deposited nodular crystals ; these were
collected, washed with water, and recrystallised from glacial acetic
Digitized by
Google
244 PEBKIN AND TATBS : BBAZILIN AND H.£MATOXTUN.
acid, from which the new acid separated in colourless needles. On
analysis :
01717 gave 03582 CO, and 00714 H,0. C-56-9; H = 4-6.
01546 „ 0-3231 GO2 „ 0-0677 H,0. C-570; H = 4-8.
^20^20^10 requires 0 = 57-1 ; H = 4-8 per cent.
HcBmcUoxylinic acid melts at 180° without decomposition, and is very
sparingly soluble in water ; it dissolves readily in hot methyl alcohol
and glacial acetic acid, but is sparingly soluble in chloroform, very
sparingly so in benzene, and almost insoluble in light petroleum.
Hsematozylinic acid, like brazilinic acid, dissolves in concentrated sul-
phuric acid, producing an intense reddish-brown solution. That it is a
dibasic acid is shown by the following results, obtained by titrating
with decinormal sodium hydroxide.
0-2067 neutralised 0041 gram NaOH, whereas this amount of a di-
basic acid of the formula O^oH^^Ojo should neutralise 0*04 gram KaOH.
Salts of Hcematoxylinic Acid, — The silver salt, O^^H^gOiQAgj, is ob-
tained as a white, granular precipitate on adding silver nitrate to a
neutral solution of the ammonium salt. On analysis :
0-2201 gave 0-3030 00^ 0-0649 H^O, and 0-0704 Ag. 0 = 37-6;
H = 2-8; Ag = 33-6.
O^HigOioAgg requires 0 = 37-8 ; H = 2-8 ; Ag = 34-0 per cent.
The neutral solution of the ammonium salt gives no precipitate with
calcium chloride, barium chloride, or zinc sulphate, but on the addition
of copper sulphate a pale green, very sparingly soluble salt separates.
Lactone of DihydrohmrnaiooGylinic Acid, OjoH^oOq.
This lactone acid is formed by the action of sodium amalgam on
haematozyiinic acid, reduction and elimination of water taking place
simultaneously, as explained in the introduction (p. 239).
In preparing the lactone acid, pure hsematoxylinic acid was dissolved
in dilute caustic soda and left in contact with excess of 3 per cent,
sodium amalgam for 24 hours. On acidifying the strongly alkaline
solution, a gelatinous precipitate separated, which was collected on the
pump, washed with water, and purified by recrystallisation from glacial
acetic acid, from which it separated in beautiful, colourless needles. On
analysis :
0-1748 gave 0-377 00, and 0-0759 H^O. 0 = 588 ; H = 4-8.
01380 „ 0-2996 OOj, „ 0 0621 H3O. 0 = 59-2; H = 49.
Oj^^Og requires 0 = 59-4; H = 4-9 per cent.
The lactone of dihydrohain%toxi/linic acid melts at 192 — 193° with-
out decomposition, and is practically insoluble in cold water ; it dis-
Digitized by
Google
PART III. THE CONSTITUTION OF HiEMATOXYLIN 245
solves, however, slightly in boiling water, and the solution, on cooling,
deposits the acid in the form of peculiar gelatinous flocks. It is moderately
readily soluble in hot methyl alcohol, and separates on cooling in
microscopic needles resembling asbestos threads ; it is readily soluble
in hot glacial acetic acid, but alm(>st insoluble in chloroform, benz6ne,
and light petroleum. Concentrated sulphuric acid colours the crystals
a salmon-pink,* and on standing a pink solution is formed, the colour
of which becomes more intense on warming, and the solution on diluting
with water deposits a white, amorphous precipitate.
That this new acid is a lactonio acid is shown by the following
titration experiments :
I. 0*229 gram of substance, titrated with decinormal sodium
hydroxide in the cold, neutralised 0*0234 gram NaOH, whereas this
amount of a monobasic acid, C^qH-^O^j should neutralise 0'0227 gram
NaOH.
II. 0*2357 gram was boiled with excess of decinormal sodium
hydroxide for 10 minutes and the excess determined by titration with
decinormal sulphuric acid. It was found that 0*0478 gram NaOH
had been neutralised, whereas, on the assumption that the lactone
ring had been hydrolysed and the acid become dibcuiCf the amount of
substance taken should have neutralised 0*0467 gram NaOH.
ScdU of the lactone of Dihydrohcsmatoxylinic Acid, — The silver salt,
C^oH^gOgAg, was prepared by adding silver nitrate to a neutral solu-
tion of the ammonium salt. It is a white, amorphous precipitate very
sparingly soluble in water. On analysis :
0*1640 gave 0*2789 COj, 0 0657H,O, and 0*0342 Ag. 0 = 46*4
H = 3*8; Ag = 20*9.
Cj^Hij^OgAg requires 0 = 46*9 ; H« 3*7 ; Ag- 21 1 per cent.
The neutral solution of the ammonium salt gives no precipitate
with barium or calcium chlorides, even on boiling ; copper sulphate
gives a pale blue, very insoluble precipitate, and zinc sulphate a white,
caseous salt which melts in boiling water.
Diatillatum of HcBmatoxylin,
[With A. W. GiLBODY.]
It was pointed out in the introduction (p. 237) that our results
showed that hsematoxylin must contain a pyrogallol and a catechol
nucleus, and that the statement of R Meyer (^er., 1889, 12, 1392),
that this substance, on distillation, yields pyrogallol and resorcinol^ was
* ThlY colour reaction is much less intense than that shown in the case of the
corresponding lactone oP'dihydrobrazDinic acid, a sabstance which will he described
in a fntor^ commonicatioD.
Digitized by
Google
246 PEBKIN: on oa-DIMETHYLGIiUTACONIC ACID AND THE
not in harmony with our work, and we therefore felt it necessary to
repeat his experiments. In doing thifi, heematozylin (10 grams) was
distilled from a small retort under reduced pressure, and the brown,
semi-solid distillate from ten such experiments was dissolved in water
and filtered from a small quantity of insoluble resinous matter.
The solution, which reacted strongly acid, was precipitated with
excess of lead acetate and the voluminous lead salt, after washing
with water, suspended in water and decomposed by sulphuretted
hydrogen. The filtrate from the lead sulphide deposited, on evapora-
tion, a brown, semi-solid mass, which was repeatedly extracted with
hot benzene, when, on concentrating the benzene solution and allow-
ing it to stand, a mass of crystals of nearly pure pyrogallol separated.
After recrystallising from toluene with the aid of animal charcoal, the
following results were obtained on analysis :
0-2046 gave 0-4274 COa and 00878 HgO. C = 570 ; H-4'8.
CgH3(OH)3 requires C-57 1 : H = 4-8 per cent
The substance melted at 130 — 131° and showed all the reactions of
pyrogallol. The filtrate from the lead salt, which should contain any
resorcinol which had been formed, was acidified with sulphuric acid
and extracted ^ve times with ether. The ethereal solution was
washed with dilute sodium hydrogen carbonate, dried over calcium
chloride, and evaporated, when only a small quantity of a dark-
coloured oil remained, which was found to contain traces of pyrogallol
which had escaped precipitation by the lead acetate. We were not
able, in spite of every careful experiments, to detect even a trace
of resorcinol in this oil, and therefore conclude that the statement
that resorcinol is formed by the distillation of haematoxylin is in-
correct.
We wish to state that much of the expense which was incurred
during this long investigation has been met by repeated grants from
the Gk)vemment Grant Fund of the Royal Society.
Trb Owens Colusoe,
Manchester.
XXV. — On aa-Dwiethylglutdconic Acid and the Syn-
thesis of isoCamphoronic Acid.
By W. H. Pbrkin, jun.
moCamphoronic acid, C^H^^Og, has been obtained by the oxidation of
campholenic acid, camphoroxime, and other derivatives of camphor, and
it is also one of the products of the oxidation of pinene. For these
Digitized by
Google
SYNTHESIS OF ISOCAMPHORONIC ACID. 247
reasons the determination of the constitution of this acid hns always
been considered to be a matter of great importance, since, until this
is definitely proved, it is not possible to obtain a correct view of the
relationship which undoubtedly exists between the various members of
the camphor and terpene series. For a considerable time, two views as to
the constitution of t«ocamphoronic acid have been especially prominent.
Baeyer (^0r., 1896, 29, 2775), as the result of his classical researches
on pinene, came to the conclusion that this acid probably had the con-
stitution
COjH-CHj-CMe2*CH(C02H)-CHj-C02H,
whereas Tiemann {ibid., 2612), who prepared the acid from cam-
pholenic acid as well as from pinene, preferred the formula
COjH-CMe2-CH(CH3-C02H)2,
on account of the fact that wocamphoronic acid, when treated with con-
centrated sulphuric acid, is converted into terpenylic acid, and when
oxidised with permanganate yields dimethyltricarballylic acid, two
changes which are easily understood with the aid of his formula,
MeaC— 9H2— CH,
_ O— CO CO.
50jH CHj-COjH CO2H
MpjC CH CHj ^ "^^ ,.^^2^^
** T I T » Terpenylic acid.
"^ McgC CH CHj
CO,H COjjH OO2H
Dimethyltricarballylic acid.
but which are difficult to bring in agreement with Baeyer's formula.
In order to decide whether either of these two formulso represented
i^ocamphoronic acid, an experimental investigation on the synthesis of
the acids represented by these formulae has been in progress during
the last four years, and in a paper published some time since (Perkia
and Thorpe, Trans., 1899, 75, 897) a method was described by which
it was found possible to synthesise the acid having the formula which
Baeyer assigned to isocamphoronio acid. Briefly stated, this synthesis
is as f oUows :
Ethyl dimethylacrylate, CMealCH^COjEt, is heated with the
sodium compound of ethylic cyanoacetate, CN'CHNa-CO^Et, when
condensation takes place and the sodium compound of ethyl cyano-
dimethylglutarate, C02EfC(CN)Na-CMe2-CH3-CO,Et, is obtained.
When this is heated with ethyl bromoacetate, ethyl cyanodimethyl-
butanetricarboxylate is produced, and this, on hydrolysis with hydro-
chloric acid, yields the dimethylbutanetricar boxy lie acid, which is the
acid -represented by Baeyer's uocamphoronic acid formula,
CHj C,CN)-CMoj.CH, CH,-^CH-CMe,.CH,
COjEt OOjEt COjEt ^^^ CO,H OO^H C0,H"
^Google
Digitized by ^
248 PERKIN: ON aa-DlMETHYLGLUTACONIC ACID AND THE
The acid thus obtained did not crystallise, and differed from iso-
camphoronic acid in many of its other properties, and it was therefore
clearly proved that the latter cannot be a dimethylbutanetricarbozylic
acid of this formula. Since these results were obtained, a great num-
ber of experiments have been carried out with the object of synthesis-
ing the acid having the formula which Tiemann considered must
represent isocamphoronic acid, but until lately without success.
In the present paper, however, a method is described by which the
acid of this formula has been synthesised, and it is shown that the
synthetical acid is identical with Mocamphoronic acid, and therefore
Tiemann's view is the correct one, and i«aoamphoronic acid is
Moj
C CH CHg
CO,H CHj-COjH CO2H'
The synthesis of tsocampboronic acid may be briefly described as
follows. aa-Dimethylglutaric acid, COaH-CMej-CHj-CFIj'OOjH, is
converted into its anhydride, this is then brominated by treating it
with phosphorus pentachloride and bromine, and the product poured
into alcohol, when ethyl aybramoau'dimethylgltUarcUe is obtained as an
oil boiling at 170° (35 mm.). When this ester is digested with
alcoholic potash, decomposition takes place readily, and one of the
substances formed is a new dimetkylglutaiconic acid,
COgH-CMeg-CHrCH-COgH,
which melts at 172°. If, now, the ester of this acid is digested
in alcoholic solution with the sodium compound of cyanoacetic
ester, a condensation product is formed which, on hydrolysis with
sulphuric acid, yields i^ocamphoronic acid. These changes may be
represented thus :
COjEfCMejj-CHrOH-OOaEt -I- CN-CHNa-COjEt =
CMcj CH-CHo-OO^Et _ (j^Me^— CH-CHj-COjH
COjEt CNa(CN)-COjEt ~ CO2H CHj,-CO»H
The acid thus obtained melted at the same temperature as tiocam-
phoronic acid, and furthermore a mixture of equal parts of the syn-
thetical acid and of Mocamphoronic acid from pinene (which Professor
von Baeyer kindly sent the author) melted at exactly the same tem-
perature as the constituents. Finally, the synthetical acid, when
heated with sulphuric acid, is converted into terpenylic acid with
evolution of carbon monoxide, a reaction which, as Tiemann first
showed, is highly characteristic of mcamphoronic acid. There can
therefore be no doubt that the synthetical acid is the same as the
i80camphoronic acid obtained from camphor and from pinene.
The aa-dimeOti/lglutaconic acid, COjH-CMej-OHICH-OOaH, which
Digitized by VjOOQIC
SYNTHESIS OF ISOCAMPHORONIC ACID. 249
is formed from ethyl ai-bromo-oa-dimethjlglutarate by the action of
alcoholic potash as described above, is an acid of considerable interest,
for the following reasons :
It melts at 172° and is undoabtediy aa-dimethylglutaconic acid, since,
on oxidation with permanganate, it is quantitatively converted into
dimethylmalonic acid and oxalic acid :
COjH-CMeg-CHIOH-COjH + 40 = COjH-CMej-COjH + COjH-CO,H.
Two other aa-dimethylglutaconic acids have, however, been described,
namely, an acid melting at about J 33°, which Henrich (Monatsh,,
1899, 20, 559) obtained by heating the sodium compound of glutaconic
ester with methyl iodide and then again with sodium and methyl
iodide and hydrolysing the product.
I. COjEt-CHNa-CHIOH-OOaEt + Mel =
COjEt-CHMe-CHrOH-COjEt + Nal.
II. COjEfCNaMe-CHrCH-COaEt + Mel =
OOjEfCMej-CH:CH-COjEt + Nal.
III. COgEt-CMej-CHrCH-COaEt + 2HjO -
COjH-CMej-CHICH-COjjH + 2EtOH.
That this acid melting at about 133° is aa-dimethylglutaconic acid
is proved by the facf that, on oxidation with permanganate, it also
yields dimethylmalonic acid.
Conrad {Ber., 1899, 32, 137; 1900, 33, 1921) has prepared a
different acid melting at 150°, which he considers to be an aa-di-
methylglutaconic acid, the process he employed being briefly as follows.
Methyl bromodimethylacetoacetate was treated with potassium
cyanide and thus converted into methyl cyanodimethylacetoacetate ;
this, on hydrolysis with acids or alkalis, yields a crystalline sub-
stance melting at 214°, which he considers to be the lactone of
ajj^-dihydrozy-aa-dimethylglutario acid. He thinks it probable that the
formation of this substance takes place thus :
(a) CN-CH:CH0H)-CMea-C03Me + Hjd
Methyl cyanodimethylacetoacetate.
- CN-CH(0H)-CH(0H)-CMej-C02Me.
(6) CN-CH(OH)-CH(OH)-CMe2-C02Me + 2B.fi -k- HCl =
CO,H-^H.CH(OH).gMe, ^ ^^^^ ^ ^^^
By heating the lactone with hydriodic acid, he obtained a crystalline
lactone melting at 153°, which he concludes is the lactone of hydroxy-
dimethylglutaric acid,
CO,H-CH-CH,-CMe,
0 CO
Digitized by
Google
250 PERKIN: on aa-DIMEXHYLGLUTACONIC ACID AND THE
Lastly, the methyl ester of this lactonic acid, when heated with sodium
and Moamyl alcohol, undergoes molecular change, yielding an aa-di-
methylglutaconio acid,
COgH-CHIOH-CMea- COgH,
which melts at 150°. This acid is unsaturated, since its solution in
sodium carbonate readily decolorises permanganate, but it is not known
whether dimethylmalonic acid is or is not formed during this oxidation.
The author of the present paper is, however, of the opinion that this
unsaturated acid obtained by Conrad is not aa-dimethylglutaconic acid,
because the lactone melting at 153° from which it was obtained does
not appear to be the lactone of a^-hydroxy-aa-dimethylglutaric acid.
It is shown on p. 259 of this paper that the lactone of this constitu-
tion is produced in considerable quantities, along with aa-dimethyl-
glutaconic acid, by the action of alcoholic potash on elhylic bromodi-
methylglutarate.
COjEfCHBr-OHj-CMej . ^^ OOjH-CH-CHjj-CMej
and as thus obtained it is a substance which crystallises well from
toluene and melts at 83°, or nearly 70° lower than Conrad's lactone.
It must be left to further investigation to determine what the consti-
tution of Conrad's unsaturated acid is.
An acid of the formula of aa-dimethylglutaconic acid should, of
course, exist in cis- and ^an«-modifications corresponding with maleic
and fumaric acids, and these may be represented thus :
COgH-CAfe^-l^-COjH OOjH-CMeg-C-COjH
H-C-COjH' COjH-C-H '
cia. trans.
It is probable that the acid melting at 172°, obtained by the hydro-
lysis of ethyl bromodimethylglutarate, is the ^an«-modification, be-
cause of its high melting point, its sparing solubility in water, and the
fact that it is not easily attacked by bromine, and does not yield an
anhydride on treatment with acetic anhydride. There can also be little
doubt that the acid, of melting point about 133°, obtained by Henrich,
is aa-dimethylglutaconic acid, not only on account of the way in which
it is formed, but also because it yields dimethylmalonic acid on oxida-
tion. Possibly this is a mixture of the cis- and ^aTW-modifications
difficult to separate into its constituents by fractional crystallisation,
and experiments are at present being made in order to determine
whether this is the case.
Many attempts have been made to convert the acid of melting point
172° into the corresponding ct«-modification, but these have all failed,
Digitized by
Google
SYNTHESIS OF ISOCAMPHOBONIO ACID. 261
partly on account of the acirl not yielding an anhydride, but principally
because the acid is so readily decomposed on heating with elimination
of carbon dioxide.
This decomposition seemed so interesting that it was carefully in-
vestigatedy and it is shown on p. 256 that the elimination of carbon
dioxide is accompanied by the formation of an oily, unsaturated acid
of the formula C^HjoOj, which boils at 207—208^.
Oarbon dioxide may be removed from oa-dimethylglutaconic acid,
COsH-CMe'CHIOH'CO^H, in two directions, yielding the following
unsaturated acids.
CX),H-OMej-OH:CHj. OHMej-OH.'CH-COaH.
Vinyldimethylacetie acid, fi-isoPropylacrylic acid.
Since, however, the acid actually obtained is quantitatively converted
into a lactone on treating with 33 per cent, sulphuric acid, it cannot be
wopropylacrylic acid, which contains the double linking in the a^-position.
It is therefore vinyldimeihylaceti/i acid^ and the lactone formed by the
action of the sulphuric acid is the lactone of hydroxyeihyldirMthylaeetio
addf and is isomeric with the tffocaprolactone which has been obtained
by the distillation of terebic acid, ^X.pn«rTT ^ ' *°^ *^ other
ways.
CMej-CHj-CjJH, CMe,-CH3-CH2
00— 0 O CO •
Lactone of hydroxyethyl-
dimethylaceiic acid (b. p. 206°). isoCaprolacUme (b. p. 207').
The solution of vinyldimetbylacetic acid in chloroform instantly de-
colorises bromine with formation of dibromoethyldimethylctcettc acid,
OOsH'OMes'CHBr-CHjBr, and its unsaturated nature is also shown
by the fact that its solution in chloroform instantly reduces perman-
ganate ; but, on the other hand, the acid is not acted on by sodium
amalgam, a behaviour which has repeatedly been observed in the case
of other acids of similar constitution.
FrepanUion of aa'DinuthylgltUario Anhydride, 1^ ^ I *.
The method which was employed in preparing the large quantities
of this anhydride which were required for the research was as follows.*
Finely powdered ?>oIaurono1ic acid (100 grams) is mixed, in a flask of
1500 C.C, capacity, with 600 c.c. of nitric acid (sp. gr. 1-2) and cautiously
heated on the water-bath. As soon as the first violent reaction has
♦ Compare BIaug (Bull. Soc, Chim,, 1898, [ill], 19, 284).
Digitized by
Google
252 PERKIN: ON oa-DIMETHYLGLUTACONIC ACID AND THE
subfiided, the heating is continued for about 6 hours, with a further
addition of small quantities of nitric acid of the same strength if it is
found that the first quantity is not sufficient to dissolve the wolauronolio
acid completely. The clear solution is evaporated to a small bulk,
mixed with water, and the evaporation repeated with the constant
addition of small quantities of water, until nearly all the nitric acid
has been removed. The residue, after concentrating as far as possible,
is allowed to stand, when it gradually becomes converted into a semi-
solid mass, owing to the separation of dimethylglutaric acid. After
the mother liquor * has been removed as far as possible on the pump,
the residue is left in contact with porous porcelain until quite dry, and
then heated to boiling in a reflux apparatus with twice its weight of
acetic anhydride for 6 hours. The acetic acid and excess of anhydride
are then distilled off and the crude anhydride purified by fractionation
under reduced pressure, when, after two distillations, almost the whole
quantity distils at 175 — 180° (60 mm.), and on cooling sets to a mass
of crystals. These are transferred to the pump, and the residue left
in contact with porous porcelain ; the substance then consists of pure
oa-dimethylglutaric anhydride melting at 38 — 40°. On analysis :
01371 gave 0-2963 00, and 0-0896 H,0. C = 590 ; H = 73.
CyH^^Oj requires 0 = 59*1 ; H«-.7'l per cent.
Ethyl BroniodxtnethylgltaarcUe, OOsEt-OMe^-OH^-OHBr-OOsEt.
In preparing this bromo-ester, dimethylglutaric anhydride (14 grams)
is mixed with phosphorus pentachloride (22 grams) in a flask fitted
with a ground-in air-tube and the mixture heated to boiling for about
1 hour and until the whole of the pentachloride has dissolved. After
cooling, a slight excess of bromine (17 grams) is added and the whole
heated on the water-bath for 12 hours. The action of bromine is
unusually slow, and it sometimes happens that even after 12 hours the
whole of it has not been used up ; in such cases, the liquid is sealed up
in tubes and heated at 125 — 130° for 2 hours. The somewhat
brownish product is now poured in a thin stream into three times its
volume of alcohol, and after the vigorous reaction has subsided, and
the whole has become cold, it is poured into a large volume of ice and
water. The heavy oil is then extracted twice with ether, the ethereal
solution washed with sodium carbonate, dried over calcium chloride,
evaporated, and the residual oil rapidly fractionated under reduced
pressure, when almost the whole quantity passes over at 165 — 170°
* A farther quaiitity of crude dimethylglutaric acid may be obtained from thii
mother liquor by repeatedly eraporating with water and finally allowing the residue
to remain over sulphuric acid until nearly solid.
Digitized by
Google
SYNTHESIS OF ISOCAMPHORONIC ACID. 253
xmdar 35 mm. pressure and consists of nearly pure ethyl hromodimethyl'
ghUofraie, On analysis :
0-217 gave 0-1326 AgBr. Br = 26-1.
OiiHijjBr04 requires Br = 27 1 per cent.
Action qf Aleoholie Potash on Ethyl BromodimethylgltOaraie. Formation
of aa-DimethylglfUaconic Acid, COjH-CMe/CHICH-COjH.
When the bromo-ester is digested with a large excess of alcoholic
potash} hydrolysis and elimination of hydrogen bromide takes place
simultaneously with formation of dimetbylglutaconic acid and the
lactone of hydroxydimethylglutaric acid (p. 259). The process was
usually carried out in the following way.
Oaustic potash (35 grams) is dissolved in the least quantity of
boiling alcohol in a large flask fitted with a wide condenser, and
through the top of this the bromo-ester (50 grams) is run in rapidly from
a tap funnel, so that the reaction may be as vigorous as possible and yet
sufficiently under control to prevent loss through the liquid being
forced out of the condenser. The whole is then boiled on the water-
bath for 1 hour, diluted with water, and evaporated with successive
additions of water until it is quite free from alcohol ; the strongly
alkaline liquid is then mixed with excess of hydrochloric acid and
extracted 10 times with ether.
After washing with water, drying over calcium chloride, and dis-
tilling off the bulk of the ether, a point is reached in the concentration
when crystals begin to separate ; at this stage, the flask is well cooled
and shaken, and the crystals collected on the pump and washed with
ether.*
The colourless, crystalline mass consists of almost pure aa-dimethyl-
glutaoonic acid, and after once crystallising from water, needle-shaped
crystals were obtained which melted at 172^. On analysis :
0-1740 gave 0-338 CO, and 01008 H^O. C = 530 ; H = 6-4.
01581 „ 0-308 CO2 „ 00920 HjO. 0 = 531 ; H = 6-4.
C7H10O4 requires 0 » 53-2 ; H « 63.
aa-Dimethylglutaconic acid melts at 172°, and is very sparingly soluble
in cold water, but dissolves readily on boiling. It is readily soluble
in methyl alcohol, acetone, or glacial acetic acid, sparingly in
chloroform, ether, or toluene; even in boiling toluene it is only
dissolved to a very slight extent and in this respect it differs from
the aa-dimethylglutaconic acid prepared by Henrich (p. 249).
The molecular weight of the acid as determined by the boiling point
* For the description of the treatment of the ethereal mother liquors of these
cryetalB, eee p. 269.
VOL. LXXXI. S
Digitized by VjOOQIC
254 PERKIN : ON aa-DIMETHYLGLUTACONIC ACID AND THE
method was found to be 146 and 144, whereas the molecular weight
of CyHj^O^ is 158. That the acid is a dibasic acid was shown by
titration with decinormal caustic soda when 01763 gram neutralised
0-044 gram NaOH, whereas this amount of a dibasic acid, C^U^fi^,
should neutralise 0 044 gram NaOH.
a^P-Dtbromo-aa-dimethylglutario Acid, OOgH-CMeg-CHBr-CHBr-OOjH,
Dimethylglutaconlc acid is not readily attacked by bromine, as is
shown by the fact that when suspended in chloroform it does not de-
colorise bromine, but it may be converted into its dibromo-additive pro-
duct in the following way. The pure acid is ground to a fine powder and
mixed with a large excess of bromine, in which it dissolves ; the liquid
is left overnight in a closed vessel and then poured on to a watch
glass and exposed to the air.
After the bromine has evaporated, an almost colourless residue is
obtained which crystallises from glacial formic acid in needles and
melts at 217—219° On analysis :
01974 gave 0-231 AgBr. Br- 517.
C^Hj^O^Brj requires Br = 60*3 per cent.
a^P'JDibramo aordimeihylglutaric acid is almost insoluble in cold water,
benzene, chloroform, and light petroleum, but dissolves readily in
methyl alcohol. When the finely powdered substance is boiled with
sufficient water, it rapidly dissolves, much hydrogen bromide is elim-
inated, and the solution, when concentrated to a small bulk, deposits on
cooling beautiful, needle-shaped crystals. These, after recrystallising from
water, melt at about 168 — 170°, the fused mass giving off gas rapidly
at 180°, and becoming quite black. The crystals are readily soluble
in sodium carbonate in the cold, and the solution does not decolorise
permanganate in the cold, but on warming reduction sets in at once.
The substance contains much bromine and is evidently the laotone of
^-bromo-oj-hydroxy-aa-dimethylgiutaric acid,
I
Mejj-CHBr-CH-COjH
0 6
Ethyl Dimeihylgluiaeanaie, COsEt'CMe^'GHrCH-OOsEt.
In order to prepare this ester, the pure acid was heated in a reflux
apparatus with alcohol and sulphuric acid for 6 hours on the water-
bath. The product was mixed with water, extracted three times with
ether, the ethereal solution washed well with water and dilute sodium
carbonate, and after drying over calcium chloride, the ether evaporated,
Digitized by
Google
STKTHESIS OF ISOCAMPHORONIG ACID. 255
when a coloarless oil was obtained which distilled constantly at
195— 197° (200 mm.).
0-1679 gave 0-3797 00, and 0-1316 H^O. 0 = 61-7; H = 8-7.
^11^18^4 requires 0 = 61*7; H = 8 -4 per cent.
Ethffl dinuihylglutaconaU is a colourless oil which has a pleasant
but pronounced odour closely resembling that of pineapples. It was
noticed in its preparation that dimethylglutaconicacid is esterified with
great ease, and that only a trace of an acid ester was extracted from
the product by means of the sodium carbonate employed.
Oxidation of aa-D%methylglui(»eonic Acid. Formation qf Dimethyl-
nuUonic Acid and Oxailic Acid,
In carryix^ out this oxidation, pure dimethylglutaoonic acid
(4 grams) was dissolved in a slight excess of sodium carbonate, water
(500 CO.) was added, and after the solution had been mixed with
powdered ice, a cold saturated solution of potassium permanganate
was run in until the colour, which disappeared instantaneously at
first, remained permanent. During the whole operation, a rapid
stream of carbon dioxide was passed, and the liquid was also kept well
stirred by means of a turbine. The product was decolorised by adding
sodium sulphite, heated to boiling, filtered, and the filtrate and the
washings of the manganese precipitate were evaporated to a small bulk.
The concentrated residue was acidified and extracted ten times with
pure ether and the ethereal solution evaporated, when a syrupy
add was obtained which, on examination, was found to contain much
oxalic acid, as well as another acid, evidently dimethylmalonic acid,
since, on heating, carbon dioxide was evolved and an oily acid smelling
of ifobutyric acid produced. In order to separate these acids,
they were dissolved in water, the solution made strongly alkaline with
ammonia, heated to boiling, and then excess of calcium chloride added.
After filtering from the cadcium oxalate, the filtrate was concentrated
and allowed to stand, when a quantity of colourless, star-like crystals
separated, which consisted of the calcium salt of dimethylmalonic acid.
The crystals were collected, decomposed with hydrochloric acid, and
the solution extracted with ether. The ethereal solution was then
carefully dried over calcium chloride and evaporated nearly to dry-
ness, when, on standing, glistening, prismatic crystals separated
which melted at 190° and consisted of pure dimethylmalonic acid :
01723 gave 0-2887 00^ and 0-0956 H,0. 0 = 45'7 ; H = 6*2.
01206 „ 0-2016 00, „ 0-0669 H,0. 0 = 45-6 ; H = 61.
O^H^O^ requires 0-45-5 ; H = 6*1 per cent.
S 2
Digitized by
Google
256 PERKIN : ON oa-DIMETHTLaLUTACONIC ACID AND THE
As it was most important that there should he no doubt as to the
identity of this acid, it was heated in a small retort until the evolu-
tion of carbon dioxide had ceased, when an oily acid was produced,
which distilled constantly at 154° (748 mm.) and consisted of pure
iBobutyrtc acid :
0152 gave 03022 CO, and 0125 llfi. C = 54-2 ; H=9-3.
C^HgOj requires 0 = 54*5 ; H = 9*1 per cent.
By weighing the calcium oxalate formed, it was found that the
dimethylglutaconic acid had been converted almost quantitatively into
oxalic and dimethylmalonic acids by oxidation with permanganate
under the conditions given above.
Diatillatian qf DimethylgltUaconie Acid, Formation of TinyldimiUhyU
acetic Acid, COjH-CMej-CHICH,.
Dimethylglutaconic acid (3 grams) was heated in a small retort by
means of a metal-bath, when it first melted, and then, at about 200%
carbon dioxide commenced to come off in quantity. When the evolu-
of gas had ceased, the residue, which had a strong odour like that of
wovaleric acid, was distilled, and nearly the whole quantity passed
over at 195 — 210°, only a small quantity of a thick, dark-coloured oil
remaining. The distillate, on cooling, deposited a few crystals, con-
sisting probably of a trace of unchanged acid, but after again distilling,
a colourless oil passed over almost constantly at 207 — 208° (760 mm.),
and this showed no signs of crystallising. On analysis :
0-1186 gave 0-2767 00, and 00966 H,0. 0 = 63-6 ; H = 9-0.
O^HjoO, requires Ob63'2 ; H»8-8 per cent.
Vinyldimethylacetic acid is a colourless, unpleasant smelling oil,
which behaves like an unsaturated acid, since, when dissolved in
sodium carbonate, it at once decolorises permanganate at the ordinary
temperature ; its solution in chloroform also instantly absorbs bromine.
In order to investigate the latter reaction, 1 gram of the acid was
dissolved in chloroform, and after cooling in ice water, bromine was
added drop by drop until the colour just remained. On exposure to
the air in a watch-glass, it was noticed that the liquid gave off a little
hydrogen bromide, and after the chloroform had evaporated, a pale
yellow oil remained, which when left overnight became semi-solid. In
contact with porous porcelain, the dark-ooloured mother liquor was
rapidly absorbed and a colourless, crystalline mass was left. On
analysis :
' 0-0843 gave 01 141 AgBr. Br » 60*1.
C^HioO^Br, requires Br — 58*4 per cent.
Digitized by
Google
SYNTHESIS OF ISOGAHPHORONIG ACID. 257
This subsianoe is evidently dihromoethyldimUhyliKetic acid,
COjH-CMes'OHBr-CHjBr, and in its crude state it melted at about
100°. Unfortunately, the quantity was too small to allow of its being
reerystallised. Yinyldimethylacetic acid is apparently not acted on
by sodium amalgam, since 1 gram of it, after boiling with excess of
sodium amalgam for 1 hour, was recovered unchanged on acidifying
and extracting with ether.
Ltieione of ffydroxy$thyld%meihylaceiie Acid {iMa'lHmeihylhtUyrolacione)^
Yinyldimethylacetic acid dissolves in 33 per cent, sulphuric acid,
and if the solution is heated to boiling for a few minutes, an oil
separat'Ce. The product was cooled well, made alkaline by the careful
addition of potassium carbonate, and repeatedly extracted with ether.
The ethereal solution was dried over potassium carbonate and evap-
orated, and the residual oil distilled, when almost the whole quantity
passed over constantly at 205 — 206°. On analysis :
01568 gave 03607 CO, and 01271 HjO. C = 627 ; H = 9-0.
OjHi^Oj requires 0 = 63-2 ; H = 8-8 per cent
The laetone of hydraxyeihyldimethylaeetie acid has a rather pleasant,
fruity odour. It dissolves readily in water, but is reprecipitated
on adding potassium carbonate. It is closely related to the lactone
of y-hydroxyiffocaproic acid (Mocaprolactone),
CMej-CHjj-CH,
which boiht at 207°, and which Fittig and Bredt {Annaieny 1880,
200, 58, 259) first obtained by the distillation of terebic acid,
C,H,oO,.
SyvMeatB of iBoCamphanmio Add, CO^-OMe/GH(OHs*00^),.
In carrying out this synthesis, sodium (2 grams) was dissolved in alcohol
(30 grams), and the solution of sodium ethoxide thus obtained mixed
with ethyl cyanoacetate (13 grams) and allowed to stand for 10 minutes.
On adding ethyl dimethylglutaconate (20 grams), it was noticed that
there was very little rise of temperature, but the reaction seemed to
set in rapidly at 100°, as the sodium compound of ethyl cyanoacetate
soon passed into solution and the mass quickly acquired a yellow
colour. After heating for 15 hours in a reflux apparatus and allowing
to cool, a gelatinous mass was obtained which was mixed with excess
of dilute hydrochloric acid and extracted three times with ether. After
Digitized by
Google
258 PERKIN : ON aa-DIMETHYLGLUTACOKIC ACID AND THE
washing several times with water and evaporating off the ether, 30
grams of a yellow oil were obtained, which was not fractionated and
analysed but at once converted into wocamphoronic acid by the follow-
ing process. The oil is mixed with about twice its volame of 90 per
cent, sulphuric acid, in which it dissolves with development of heat but
without charring. After standing until cold, the whole is diluted with
1^ vols, of water and heated to boiling in a reflux apparatus for 6
hours, the hydrolysis being facilitated by removing the condenser
from time to time in order to allow the alcohol produced to escape.
The brownish liquid on standing overnight becomes filled with a
mass of crystals and these, after collecting on the pump and washing
with a little water, melt at 164° and consist of nearly pure itocam-
phoronic acid. A further considerable quantity in a less pure condition
may be obtained from the mother liquors by extraction with ether,
evaporating, and leaving the residue, which becomes semi-solid on
standing, in contact with porous porcelain.
The crystalline acid obtained in this way was not weighed, but the
yield was very considerable, and it is only necessary to recrystallise it
once from water in order to obtain it quite pure. On analysis :
01821 gave 0'3341 CO^ and 01064 HjO. C - 600 ; H = 6-5.
0-1661 „ 0-3013 COj „ 0-0972 HgO. 0 = 49-7; H = 6-6.
OgHj^Og requires 0 - 49'6 ; H = 6*4 per cent.
The synthetical acid, when rapidly heated, softened at 165° and
melted at 168°, whereas the melting point of Mocamphoronic acid is
given as 166°. A sample of pure isocamphoronic acid from pinene,
which Professor A. von Baeyer was good enough to send the author, was
mixed with the synthetical acid without any alteration in the melting
point being observable. This, and the fact that the synthetical acid
on treatment with sulphuric acid yields terpenylic acid, proves that it
is identical with the acid obtained from pinene and camphor.
Conversion of Synthetical isoCamphoronie Acid into Terpenylic Acid^
MejjC CH-0H,-00jH
0-00-CH,
In carrying out this decomposition, synthetical tsocamphoronic add
(1 gram) was dissolved in concentrated sulphuric acid (8 grams), and
the solution heated at 100°, when bubbles of gas were slowly given
off. After 6 hours, the liquid, which had scarcely darkened in colour,
was diluted with water and repeatedly extracted with ether, the ether-
eal solution was washed until free from sulphuric acid, evaporated,
and the oily residue mixed with a very little water and allowed to
stand. After about 3 days, the crystals which had separated were left
Digitized by
Google
SYNTHESIS OF ISOOAMPHORONIC ACID. 259
in contact with porous porcelain until dry, and then crystallised from
water^when glistening prisms were obtained, which melted at 59 — 60^,
and consisted of hydrated terpenylic acid. On leaving these crystals
over salphnrio acid in a vacuum desiccator, they soon became opaque,
and after two days the chalky mass melted sharply at 90^ and gave
the following results on analysis :
0-1634 gave 0-3332 CO2 and 01034 H,0. C = 55-6 ; H = 70.
CgHjjO^ requires C = 66-8 ; H « 7*0 per cent.
This acid is therefore terpenylic acid, which, according to Tiemann,
crystallises from water in well-defined, hydrated crystals melting at
56° ; these, when placed over sulphuric acid in a vacuum desiccator, lose
their water of crystallisation, giving a chalky mass which melts at 90^
Laetans of ai'ffydr<>Qcy'Cui'difmsthylgltUarte Acid,
CMeg-OHj-CH-OOjH '
The ethereal mother liquors of the crystals of dimethylglutaconic
acid (p. 253) yielded on evaporation a thick, brown oil, which, on
standing for several days and repeatedly stirring, partially crystallised.
The semi-solid mass was spread on porous porcelain and left for some
weeks until almost all the thick oil bad been absorbed and a yellow,
crystalline mass remained. This, after crystallising first from ether
and then from water, yielded a considerable quantity of pure dimethyl-
glutaconic acid (m. p. 172°).
The porous plates were crushed and extracted in a Sozhlet apparatus
with ether, the ether was then evaporated, and the dark brown oil,
from which nothing crystalline could be obtained directly, was esterified
by treatment with alcohol and sulphuric acid in the usual way (p. 254).
The large amount of ester thus obtained was distilled twice under
reduced pressure, and the fraction 215 — 217° (200 mm.), which was
about two-thirds of the whole, collected separately* On boiling
this fraction with 10 per cent, hydrochloric acid, it was readily hydro-
lysed, and on evaporating to dryness a very thick syrup was obtained,
which, when placed over sulphuric acid in a vacuum desiccator, soon began
to crystallise, and ultimately became almost solid. In contact with porous
porcelain, the mother liquor was slowly absorbed, and a colourless, crys-
talline residue was obtained, which consisted of the IcKUone of hydroxy-
dimethylgliUario add mixed with small quantities of dimethylglutaconic
acid.
The separation of these two substances is very tedious, and was
carried out by rubbing the crystalline mass with 50 per cent, hydro-
chloric acid, which dissolves the lactone, but in which the dimethyl
Digitized by VjOOQIC
260 SYNTHESIS OF ISOCAMPHORONIC ACID.
glutacoDic acid is practically insoluble. After filtering, the filtrate was
extracted with ether, the oil obtained after distilling oS the ether
allowed to solidify, and again treated with hydrochloric acid. Ulti-
mately, the lactone was further purified by recrystallisation from toluene
On analysis :
01631 gave 03177 CO, and 00960 H^O. 0 = 531 ; H=6-4.
01304 „ 0-2544 COj „ 0 0760 H^O. 0 = 532; H«6-6.
C7H1QO4 requires 0«63'1 ; H = 6'3 per cent.
The lactone of ai'hydroxy-aa-dtmeihylglutaric aoid melts at about 85^
but not quite sharply. It is readily soluble in water, and when the
crystals are placed on water they rotate rapidly, like camphor crystals,
and slowly dissolve. The lactone dissolves readily in hot toluene, but
is sparingly soluble in the cold.
That it is a monobasic lactonic acid is shown by its behaviour on
titration with decinormal caustic soda solution, when 0*1849 gram,
dissolved in cold water, neutralised 0*049 gram NaOH, whereas this
amount of a monobasic acid, O^H^oO^, should neutralise 0 047 gram
NaOH. A considerable excess of decinormal caustic soda was then
added, and the solution heated to boiling for 10 minutes, and the excess
of soda estimated by titration with decinormal sulphuric acid. It was
then found that the total amount of NaOH taken up was 0*0944 gram,
whereas 0*1849 gram of a monobasic lactone acid, CyH^^O^, on dissolv-
ing to form the salt of the hydroxy-dibasic acid, should neutralise
0094 gram NaOH. The first results of titration in the cold show that
the lactone ring is hydrolysed only to a very slight extent when the
lactonic acid is dissolved in cold water.
The hydroxy-dibasic acid, when set fr^e from its solution in caustic
alkali, is obtained on extraction with ether as a colourless syrup,
which does not rapidly pass into the lactone, since titration showed
that even after standing for a quainter of an hour it consisted princi-
pally of the hydroxy-dibasic acid. When, however, the syrup is placed
over sulphuric acid in a vacuum desiccator, it is rapidly converted into
the lactone and solidifies completely.
Silver Salt of the Lactone of HydrooDydimethylgluiarie Acid, O^H^O^Ag.
— When silver nitrate is added to a neutral solution of the ammonium
salt, there is at first no precipitate, but soon the silver salt begins to
separate in microscopic needles, and wh^n dried it has the appearance of
a silky mass of needles, which are readily soluble in hot, but sparingly
in cold, water. On analysis :
0-2310 gave 02691 CO,, 0*0714 H,0 and 0*0941 Ag. C=r31-8;
H = 3*4; Ag = 40*7.
Cy HjjO^ Ag requires 0 = 31*7; H = 3*4 ; Ag = 40*7 per cent.
Digitized by
Google
RUHEMANN AND STAPLETON : TBTRAZOLINE. PART II. 261
The peutral solution of the ammoDium salt gives no precipitate with
calcium or barium chlorides, or with copper sulphate or lead acetate.
The author wishes to express his thanks to Mr. J. Yates for his
valuable assistance in carrying out this investigation, and also to state
that some of the expense of the research was met by repeated grants
from the Research Fund of the Royal Society.
Ths Owens Collxob,
Makohbstku
XXVI. — Tetrazoline. Part II.
By SixoFRiSD RuHEKANK and H. E. Staplbton.
Hantzsch and Silbsrrad, in their interesting research on diazoacetic
ester {Bw.^ 1900, 83, 58), showed that one of the polymerides of
diazomethanes, described by Ourtius and Lang (/. pr. Chmn.^ 1888, [ii],
88, ,534) as trimethintriazimide, was hydrotetrazine (tetrazoline), and
they used for its synthesis the same method, namely, the action of
heat on monoformylhydrazine, which we had- published previously
(Tsans., 1899, 75, 1131). The remarkable properties of tetrazoline
induced us to subject it to a closer study. Although this is not yet
completed, we thought it advisable to give a record of the results
already arrived at in order to be able to continue the work undisturbed.
Our attention was mainly directed to the investigation of the action
of methyl iodide on tetrazoline. We expected that the basic character
of this substance, as indicated by the fact that it forms salts with
strong acids and, as stated in this paper, yields a thiourea with
phenylhydrazine, would also appear in its behaviour towards alkyl
iodides. We were, however, surprised to find that the action of
methyl iodide on tetrazoline was complicated and yielded several
products, two of which we were able to isolate. One of these is
rather unstable, crystallises in dark blue needles, and has the formula
CgHgN^Ty whilst the other forms colourless crystals, and has the
composition C3H7N4I. The formula of the latter compound would
correspond with the iodide of methyltetrazoline ; its behaviour, how-
ever, proves it to have another constitution, for, on attempting to
isolate the base from the iodide or the chloride by means of an alkali
or silver oxide, oxidation takes place, and a deep blue solution is
produced, whereas tetrazoline does not give a similar reaction. More
remarkable still is the nature of the other compound which is formed
from tetrazoline and methyl iodide. It is decomposed with separation
Digitized by
Google
262 RUHEMANN AND STAPLETON : TETBAZOLINE. PART II.
of iodine even by cold water, more readUj, however, on boiling ; this
property characteriBes it as a nitrogen iodida *
The continuation of this research, which one of us has undertaken,,
will most probably lead to the knowledge of the reaction between
methyl iodide and tetrazoline, and to the establishment of the con-
stitutional formula of the two compounds which are described in this
paper.
EXPEBIMENTAL.
FhenylUtrazolyUhiourea, CeH5-NH-CS-N<:^^*2>NH.
A mixture of tetrazoline and phenyl mustard oil, when cautiously
heated, yields an oil which, on adding a little dilute alcohol, sets to a
white solid. In order to remove the adhering mustard oil, it is washed
with ether and then dissolved in boiling alcohol, from which, on cool-
ing, it separates in colourless needles melting at 153 — 154^. On
analysis :
0*1457 gave 39'4 c.c. moist nitrogen at 16° and 774 mm. N» 32-15.
0-2040 „ 0-2219 BaSO^. S = 14-93.
CgHjjN^S requires N = 31 -96 ; S = 14-61 per cent.
Aetion of Methyl Iodide an Tetrazoline,
On heating tetrazoline with an excess of methyl iodide at 100° in a
closed tube for 3 hours, a dark red, viscous product is formed which is
freed from the unaltered alkyl iodide by evaporation on the water-
bath. After adding a few drops of methyl alcohol, the oil is allowed
to remain in a desiccator over sulphuric acid, when, in the course of a
day, a solid separates. This is filtered and dissolved in a warm mix-
ture of chloroform and alcohol, from which, on cooling, bluish-black
needles crystallise melting at 102 — 103°. The following analytical
data of the substances, dried in vacuo, correspond with the formula
0-2258 gave 00655 CO, and 00382 H,0. C = 791 ; H = 1 87.
0-2395 „ 00660 COj „ 00380 H,0. 0 = 7-51 ; H = 176.
0-2534 „ 24-5 c.c. moist nitrogen at 10° and 767 mm. K = 11*68.
0-2243 „ 0-3242 Agl. 1 = 7811.
0-2453 „ 0-3578 Agl. 1 = 78-81.
CgH^NJg requires 0 = 7-47; H= 1-87; N = 1162; 1 = 79 04 per cent
The substance is soluble in cold alcohol, and the solution gives with
silver nitrate a precipitate of silver iodide, which proves it to be an
iodide. Oarbon disulphide neither dissolves it nor is it coloured by it ;
the characteristic violet coloration, due to iodine, appears, however,
Digitized by
Google
BUHEMANN AND STAPLETON : TETRAZOLINE. PART II. 263
on adding a trace of water. This fact indicates that the compound is
a nitrogen iodide, and that water has to be excluded in its prepara-
tion. More readily than with cold water does the decomposition take
place on boiling. The substance melts, and iodine is separated which
distils with the steam. The dark solution which remains, and which
is almost neutral to litmus paper, is freed from iodine by extraction
with carbon disulphide, and, on concentration, yields a dark oil which
does not deposit crystals, even on standing for several days. The
viscous residue, when heated with potash, decomposes and evolves
ammonia ; we have not been able to isolate any other substance.
The yellowish mother liquor from the bluish-black needles contains
the second compound which is formed by the action of methyl iodide
on tetrazoline. It is evaporated on the water-bath, and the dark oil
which remains, on standing in vacuo over sulphuric acid, yields
slightly coloured crystals. A further crop can be obtained from the
syrupy filtrate ; finally, a viscous product is left behind from which
nothing crystalline separatea The compound, after recrystallisation
from water or methyl alcohol, appears in well developed, colourless
crystals which melt at 98 — 99% and on analysis yield data which
agree with the formula C^'EL^f'SJL.
0-2166 gave 0-1268 COa and 0-0605 H,0. C«15-96; H = 310.
0-2405 „ 61 c.c. moist nitrogen at 13^ and 750 mm. N = 24-71.
0.1597 „ 0-1634 Agl. I « 65-38.
0-2063 „ 0-2136 AgL. I = 55-93.
CjHyN J requires 0 = 1 592 ; H « 3-09 ; N » 2477 ; I = 66-19 per cent.
This compound is almost insoluble in chloroform, but dissolves with
the greatest ease in alcohol or water. It can be transformed into the
coiresponding chloride by boiling its aqaeous solution with an excess
of freshly prepared silver chloride. The filtrate is evaporated on
the water-bath, when an oil is left behind which solidifies on standing
in vacuo over sulphuric acid for a day. The chloride is extremely
soluble in water, less so in absolute alcohol, and crystallises from
it in deliquescent, colourless needles which melt at 130°. On
analysis:
0-3163 gave 0-3305 AgCl. Gl » 25*93.
CjH^N^Ol requires 01 = 26-39 per cent.
The concentrated aqueous solution of the chloride gives with
platinic chloride beautiful, red plates of the platinichloride which is
rather soluble in cold, readily in boiling, water, but is almost insoluble
in cold absolute alcohol, and melts at 175°. On analysis :
0-241 5 left, on ignition, 00775 Pt. Pt=.32-09.
(C3H^^)^,PiClg requires Pt- 32*11 per cent.
Digitized by
Google
264 FORSTER: studies in the CAMPHANE series, part VII.
The formation of a platinichloride leads to fche view that the sub-
stance C3H7N4OI is not the hydrochloride of methyltetrazoline, since
tetrazoline itself yields a compound of the formula (02H^N4)2PtCl4.
This conclusion is supported by the fact that solutions of the chloride
or the iodide in the presence of an alkali turn violet on exposure to
the atmosphere, but that such a reaction does not take place with
tetrazoline. On account of the solubility of the base in water,
and the ease with which it is oxidised, we have not succeeded in
isolating it, since silver oxide has not the desired effect; neither have
we been able, as yet, to obtain the blue oxidation product in a pure
state.
6ONVILLB AND CAITTS COLLBaE,
Oambridob.
XXVII. — Studies in the Ca/mphane Series. Part VI I.
Conversion of Hydroxycamphene into fi-Hcdogen
Derivatives of Camphor.
By Mabtin Onslow Fobstbs.
In a recent communication (Trans., 1901, 79, 644) I described, under
the name hydroxycamphene, an isomeride of camphor which exhibited
some remarkable properties. The method by which it was obtained
from camphor, although circuitous, leads by steps, apparently straight-
forward, through camphoroxime, hromonitrocamphane, nitrocamphene,
and aminocamphene. The substance itself behaves like an unsatur-
ated compound towards bromine and potassium permanganate, and
is readily converted into camphor by the agency of dilute mineral
acids ; it also contains the hydroxyl group. On the strength of these
facts, it was regarded as the tautomeric modification of camphor :
Hydroxycamphene. Camphor.
It was necessary to admit, however, that in some respects the be-
haviour of hydroxycamphene was dissimilar from that which is usually
associated with compounds of this type. Its unsaturated character,
and its ready conversion into the ketonio isomeride were certainly
in accordance with this view of its constitution, but its insolubility in
alkalis, the indifference of an alcoholic solution towards ferric chloride,
and the extraordinary stability of the substance on distillation, were
Digitized by
Google
FORSTER: studies in the CAMPHANE series, part VII. 265
points which demanded an explanation more complete than that which
was put forward at the time. It was suggested, namely, that the
abnormal behaviour of hydrozycamphene might be due to the fact
thaty unlike the enolio modifications of the ketonic esters which haye
received so much attention during recent years, it contains no acyl
substituent attached to the /3-carbon atom. This view received some
support from the observation (Trans., 1901, 70, 987) that a-benzoyl-
camphor, in which the acyl radicle occupies a ^-position relative to
the ketonic oxygen, is readily converted into an enolio modification
which has all the general properties of compounds belonging to this
class. On the other hand, it must be borne in mind that the con-
stitution of enolic a-benzoylcamphor is by no means assured, because
although, at present, the evidence is in favour of representing it as
l-hydrozy-2-benzoylcamphene, the matter is still under investigation,
which may produce facts in support of the alternative structure,
namely, that of phenylhydroxymethylenecamphor :
1 -Hydrox7-2-benzoyl- Phenylhy drozymethylene-
camphene. camphor.
With the object of gaining further evidence relating to the constitu-
tion of hydrozycamphene, the action of bromine on this compound has
been studied. It was anticipated that if the structure of hydrozy-
camphene is that which has been assigned to it, addition of bromine,
followed by elimination of hydrogen bromide, would yield a-bromo-
camphor, either alone, or mixed with the unknown a'-bromocamphor :
CsH.<^r
The product actually obtained by the action of bromine on hydrozy-
camphene (dissolved in glacial acetic acid containing sodium acetate
is a new bromocamphor, which melts at 78^, and when dissolved in
alcohol has [a]i> +19^; a-bromocamphor melts at 76^ and has
[a]o +135^ No addition to the list of mono-halogen derivatives
of camphor has been made since Kipping and Pope described
«^chlorocamphor and «^bromocamphor (Trans., 1895, 67, 371), pre-
pared by eliminating sulphur diozide from camphorsulphonic chloride
and bromide respectively. Since the new bromocamphor does not agree
in properties with those attributed to any of the isomerides previously
Digitized by
Google
266 forster: studies in the camphane series, part vii.
described, it seemed probable at first that it was a'-bromocamphor,
the unknown stereoisomeride of ordinary a-bromocamphor. Such a
substance, however, would be expected to yield a-bromocamphor on
treatment with the usual transforming agents, but the new derivatiye
resists the action of fuming hydrogen bromide and concentrated
sulphuric acid. Moreover, iJcoholic potash should convert it into
camphor, and although zinc and acetic acid reduce it to that substance,
showing it to be a true camphor derivative, alcoholic potash acts as a
hydrolytic agent, giving rise to a-campholenic acid :
C^oHisOBr + HjO = O^^U^fi^ + HBr.
Finally, bromine, which would yield a-dibromocamphor in the case of
a-bromocamphor, converts the new substance into /3-dibromocamphor.
These facts have only one explanation. The substance obtained by
the action of bromine on hydroxycamphene must be the unknown
/3-bromocamphor, bearing the relation to ^-dibromocamphor that
a-bromocamphor has to a-dibromocamphor. In the last-named sub-
stance both bromine atoms occupy the a-position, whilst in /3-dibromo-
camphor, which is more suitably named a^-dibromocamphor, one
bromine atom is attached to the a-carbon atom, the other occupying
the /S'position, of which the exact situation is unknown.
A bromo-derivative of camphor has been described by Marsh, who
refers to the substance as "j8-bromocamphor.'' Without disparaging
the claims of this compound to be regarded as an individual substance,
it may be questioned whether the bromine atom occupies the )3-po8ition
in the sense just indicated, it being nowhere stated that bromine con-
verts it into a/3 dibromocamphor, or that alcoholic potash resolves it
into o-campholenic acid.
It will be recognised that the production of /3-bromocamphor from
hydroxycamphene by a process so simple as addition of bronune
followed by elimination of hydrogen bromide, necessitates a recon-
sideration of the relationship between camphor and hydroxycamphene.
The last-named substance is indifferent towards hydroxylamine and
phenylhydrazine, but readily forms alkyl ethers ; its hydroxylio
character seems therefore assured, apart from the fact that it is pro-
duced by the action of nitrous acid on a primary amine. There is no
reason to doubt that the nitro-group in 1 : 1-bromonitrocamphane, and
consequently the nitro-group in nitrocamphene, and the amino-group
in aminocamphene, are attached to the carbon atom which, in camphor,
is ketonic, because bromonitrocamphane yields camphoroxime on re-
duction, and aminocamphene gives rise to camphor when heated with
acetic anhydride ; moreover, hydroxycamphene is converted so readily
into camphor that the action of nitrous acid on aminocamphtfie can
scarcely have been abnormal. The only uncertain step, therefore,
Digitized by
Google
FOBSTSB : STUDIES IN THE CAHPHANE SERIES. PART VII. 267
in the series connecting camphorozime with hydrorjcamphene, is the
removal of hydrogen bromide from 1 : 1-bromonitrocamphane. The
question arises, How does this change take place f
Adopting Bredt's formula for camphor as the basis of representation,
the constitution of 1 : 1-bromonitrocamphane is expressed as follows :
CHj-CH OHj
SHj-qCHj) CBr-NOa
The point to be decided, therefore, is the source of the hydrogen atom
which is eliminated in association with bromine when bromonitro-
camphane is converted into nitrocamphene. In view of the unsatur-
ated character of the latter, and the fact that, of the two carbon
atoms united with that to which the bromine is attached, only one
is combined with hydrogen, it was natural to conclude that nitro-
camphene is represented by the formula :
CH,-(jJH CH
L
I 9(CH,),
0H,-(!3(CHg)
.)— C-NO,
It is this expression which is now shown to be incorrect. It is,
however, the sole possible formula for a compound which is not
only produced by the removal of hydrogen bromide from 1 : 1-bromo-
nitrocamphane, but also contains an ethylenic link. The obvious
conclusion is that nitrocamphene, and consequently hydroxycamphene,
are compounds which owe their unsaturated nature to a trimethylene
or tetramethylene ring, which is resolved so easily into the original
form as to simulate an ethylenic structure.
Although, theoretically, this change might occur in each of six
ways, the production of a-campholenic acid by the action of alcoholic
potash on ^bromooamphor renders all improbable excepting two ; these
would involve the representation of hydroxycamphene by one of the
formula :
OH,.CH OH, CHj-CjJH OH,
I. I A(CH^, I II. I C(CH3), I
(JJH-C(CH3) (jJ-OH OH^'Ov yC'GH
the structure of )3-bromocamphor being then indicated by one of the
expressions :
-OH.
OH,— (jJH OH, OHj-CjJH-
I aOH.), I or I C(CH3), I
CHBr-qOH,) — UO CH,-C(CH,Br)-00
respectively.
Digitized by
Google
268 fobster: studies in the camphane series, part yii.
So far as the eyidence in this paper is conoerned^ it is scarcely
possible to conie to a de6nite conclusion, although the transformation
of )9-bromocamphor into a-campholenic acid is distinctly in favour of
the former representation, according to which the change in question,
although fundamental, would appear quite simple, as follows :
OHj
CH-9
I - 9(CH.), ^
CHBr-qCHj)— CO
iB-Bromocamphor.
,CE[2
?(CH3),
:H(CH3)H0'
CH,
1 9(CH3), ., I
CH(OH)-CH(CHJ HO-CO
— >
CH,
•CH-CHj
9(CH3),
2C02H
CH=C(CH,)
a-Campholenic acid.
Moreover, although the plane formula I represents union to have
occurred between two carbon atoms which appear to be at some
distance from one another, it is possible that they arc^ in reality
comparatively close, in accordance with the tetrahedral conception of
the combining power of carbon. This is illustrated by the following
representation of bromonitrocamphane :
B^/C-NO,
Simultaneously with the publication of a preliminary notice of
/3-bromocamphor (Proc., 1901, 243^ 245), Armstrong and Lowry
recorded the properties of the same substance obtained by the action
of heat on the camphorsulphonic bromide belonging to the Beychler
series. These authors have converted )9-bromocamphor by direct oxi-
dation into /3-bromocamphoric acid, and the investigation of the latter
should make it possible to determine which, if either, of the two
formuln suggested in this paper correctly represents the constitution
of ^-bromocamphor.
An examination of ^-bromocamphoroxime has already produced two
observations of some interest. In the first place, the oxime does not yield
the nitrile of an unsaturated' acid when treated with hot dilute sulphuric
acid, thereby differing completely from camphoroxime, which is con-
verted into a-campholenonitrile. This again may be construed in
Digitized by
Google
FQB8TER: STUDIES IN THE OAMPHANE SERIES. PART VII. 269
favour of the expression for )8-bromocamphor corresponding with
formula I for hydrorycamphene, as it accords with the simplest
possible explanation of the last-named reaction :
CH,-CH OH, OHj-^H-CHj-ON
-^ I WH3),
i-c:i}r CHriqcHg)
1 iip^s).. I
(J;h-c(ch3)-j-c:
Camphoroxiroe. a-Gampholenonitrile.
In the second place, although )9-bromocamphor is hydrolysed to
a-campholenic acid by alcoholic potash, this agent is without action
on )9-bromocamphoroxime, a result which seems to suggest that, in
the change undergone so readily by )8-bromocamphor, disruption of
the ring precedes replacement of bromine by hydroxyl, and the
frabsequent elimination of water. Indirectly, this also favours the
formula for ^-bromocamphor in which the bromine is represented as
replacing hydrogen in the ring, because if substitution had occurred
in a methyl group, it would be reasonable to expect the halogen to be
more easily removed.
EZPBBIMENTAL.
p'Bromoeamphor, CgHj jBr<OL *
Twenty-five grams of 1-hydroxycamphene were dissolved in 400 c.c.
of glacial acetic acid containing 26 grams of anhydrous sodium
acetate. To the well cooled liquid were added in small quantities 26*5
grams of bromine dissolved in 100 c.c. of glacial acetic acid; the
colour of the halogen was immediately destroyed, and heat being
developed, the temperature of the liquid was allowed to return to about
20^ before each addition of bromine. When the stated quantity had been
added, the pale yellow liquid was poured into a large volume of cold
water, which precipitated a colourless, crystalline substance. This
was filtered, washed with water, and dried in the desiccator, when it
weighed 38 grams.
/3-Bromocamphor dissolves very readily in warm petroleum, and
after being recrystallised twice from this solvent, melts at 78^, the
second crystallisation producing no change in the specific rotatory
power, and a very slight increase only in the fusion temperature. If
the substance, purified in this way, is then recrystallised from absolute
alcohol, the melting point is lowered about 1°.
0-2053 gave 01656 AgBr. Br » 3432.
C|^i50Br requires Br»34'63 per cent.
VOL. LXXXI. T
Digitized by VjOOQIC
270 FORSTER : STUDIES IN THE OAMPHANE SERIES. PART VII.
The substanoe dissolves very readily in chloroform, benzene,
or glacial acetic acid, also in warm petroleum or hot alcohol, crystallising
from the last solvent named in long, striated prisms ; when deposited
slowly by petroleum, it crystallises in large, well formed, transparent
prisms. It is readily volatile in steam, the vapour having a faint
odour of camphor.
A solution containing 0*6014 gram in 26 c.c. of chloroform at 20^,
gave ai) 39' in a 2-dcm. tube, whence the specific rotatory power
[oJd +16*2^; 0-6012 gram dissolved in 26 c.c. of absolute alcohol at
20^, gave ai> 46' in the same tube, corresponding with [a]^ +19'1^-
/3-Bromocamphor may be dissolved in fuming hydrobromic acid
(sp. gr. 1*83) or in concentrated sulphuric add without undergoing
any chemical change or alteration of rotatory power.
Canver8icn into Cctmphar. — ^Two grams of ^-bromocamphor were
dissolved in glacial acetic acid and treated with 6 grams of zinc
dust. After remaining on the water-bath during 2 hours, the
liquid was diluted with water, neutralised with sodium carbonate, and
distilled in a current of steam. Camphor was deposited in the con-
denser, and was converted into the oxime by the usual process ; the
product was crystallised from petroleum, which deposited the charac-
teristic crystals melting at 118^.
Canveraion into aP-Dibromoeamphar. — Five grams of the new bromo-
camphor were covered with one molecular proportion of bromine and
heated on the water-bath, when evolution of hydrogen bromide soon
occurred. When this gas was no longer liberated, the pale yellow
liquid solidified almost immediately on being withdrawn from the water-
bath. The product was washed with water, and recrystallised several
times from alcohol, until the melting point was constant at 114^, when
a mixture of the substance with a^-dibromocamphor melted at the
same temperature ; 1*0006 giam dissolved in 26 c.c. of chloroform at
21^ gave ttD 8^4' in a 2-dcm. tube, whence the specific rotatory power
[a]D -H00-8<'.
Aetion of AleohoHo Fotash on fi-Bramocamphar.
Twenty grams of recrystaUised )9-bromocamphor were dissolved in
100 c.c. of alcohol which had been distilled from caustic soda ; 16 grams
of caustic potash, dissolved in the minimum quantity of water, were
then added to the liquid, which was heated in a reflux apparatus during
8 hours. Alcohol was then removed by distillation, and a current of
steam passed through the residual liquid, a preliminary experiment
having shown that if an insufficient quantity of alkali has been em-
ployed any bromocamphor remaining unchanged can be removed
conveniently at this stage. The contents of the distilling flask were
Digitized by
Google
FOESTER: STUDIES IN THE CAMPHANE SERIES. PART VII. 271
then cooled, jost acidified with dilute sulphuric acid, and extracted
with ether, the extract being washed several times with water, dried
with calcium chloride, and evaporated on the water-bath. The residue
was then distilled, passing over completely at 256 — 267° under 767 mm.
pressure :
0-1670 gave 04087 COj and 0-1330 H,0. C = 7100; H = 9-41.
^o^iA requires C = 71*43 ; H = 9*62 per cent.
The substance is slowly volatile in steam, and has a very faint odour ;
it dissolves in sodium carbonate, forming a solution which immediately
decolourises potassium permanganate. It has a sp. gr. 0*9974 at 20720°,
and gives aj, 16°40' in a 2-dcm. tube at 20°i^ whence the specific rotatory
power [ a ]|> + 8*34°. The product was thus identified with a-campho-
lenic acid.
P'Bnnnoeamphoroximef GJS.^Br^^y^,^ .
OaJNUJtl
Ten grams of /3-bromocamphor were dissolved in 60 c.c. of alcohol
and heated with 5 grams of hydroxylamine hydrochloride and 6 grams
of dry sodium acetate in a reflux apparatus during 2 hours ; 3 grams
of the hydrochloride and the same quantity of sodium acetate were
then added, and the heating continued during 3 hours, a preliminary
experiment having shown that conversion into the oxime is incomplete
unless a considerable excess of the hydrochloride is employed. On
pouring the liquid into twice its own volume of cold water, a crystalline
precipitate was obtained ; this was filtered, washed, and recrystallised
twice from hot alcohol, which deposited lustrous, rhomboidal plates
melting at 166° :
0-2666 gave 01944 AgBr. Br » 32-23.
CjQHjgONBr requires Br = 32 -52 per cent.
The oxime is readily soluble in chloroform, benzene, or ethyl acetate,
but dissolves very sparingly in hot petroleum, which deposits it in
aggregates of rhomboidal plates.
A solution containing 0-5017 gram in 26 c.c. of chloroform at 21°
gave aj, - 2°67' in a 2-dcm. tube, whence the specific rotatory power
[a]o -73-5°
)3-Bromocamphoroxime dissolves in caustic alkalis and in 50 per cent,
sulphuric acid; it may be warmed with concentrated sulphuric acid
without undergoing conversion into a nitrite. It is also indifferent to
alcoholic potash, with which a specimen has been heated in a reflux
apparatus during 4 hours without becoming hydrolysed.
The benzoyl derivative is readily obtained by the Schotten-Baumann
method. It is very readily soluble in hot alcohol, from which it
T 2
Digitized by VjOOQIC
272 FORSTEB: STUDIES IN THE CAMPHANE SERIES. PART VII.
separates in tufts of white needles. Warm petroleum dissolves it less
readily, and it is scarcely soluble in the cold solvent, which deposits it
ii^ long, snow-white, silky needles melting at Tl-^TS^.
0 2695 gave 01437 AgBr. Br « 22*69.
Cj7H,^0jNBr requires Br » 22*86 per cent.
The substance dissolves very readily in chloroform, ethyl acetate,
or benzene. A solution containing 0*5040 gram in 25 c.c. of
chloroform at 21^ gave ai> - 1^35', whence the specific rotatory power
[a]D -39*2°.
fi-ChhrocamjphM-, CgHl3Cl<JL^
Twenty grams of hydroxycamphene were dissolved in 400 c.c. of
glacial acetic acid containing 20 grams of anhydrous sodium acetate.
The liquid was immersed in cold water and treated with a solution of
chlorine in glacial acetic acid, which was added in small quantities at a
time until, after an interval, the halogen was found to be in slight ex-
cess. As the precipitate obtained on pouring the liquid into a large
volume of water weighed only 16 grams, the filtrate was neutralised
with solid sodium carbonate, 6 grams more being obtained in this way ;
the product was then recrystallised three times from alcohol.
0*2064 gave 0 1565 AgOl. CI = 18*76.
CioHij^OOl requires Cl«1903 per cent.
The new chlorocamphor is very similar to the bromo-derivative in
appearance and properties ; it melts, however, at 132*5°. It is readily
soluble in chloroform, acetic acid, petroleum, benzene, or ethyl acetate,
but dissolves less freely in alcohol, from which it crystallises in slender,
striated prisms, sometimes exceeding an inch in length ; it is readily
volatile in steam, and the vapour has a. faint odour of camphor.
A solution containing 0*5012 gram in 25 c.c. of chloroform at
21° gave ai> 1°35' in a 2-dcm. tube, whence the specific rotatory power
[a]D +39*5°; 0*5019 gram dissolved in 25 c.c. of absolute alcohol at
19° gave an 1°38', corresponding to [ajn +40*7°.
p'ChlwocampIioraximey 0^'Bi^sO\<CX*
NOH'
The oxime of jS-chlorocamphor was obtained by the same process as
the bromo-derivative, which it closely resembles in appearance and
properties. It separates from alcohol in colourless crystals isomorphous
with those of )9-bromocamphoroxime, and melts at 134°.
0*2412 gave 0*1662 AgCl. 01 = 17*05.
C^oHiqONGI requires Cl» 17*62 per cent.
Digitized by
Google
F0R8TER: STUDIES IN THE CAMPHANE SERIES. PART VII. 273
It dissolves very readily in chloroform, ethyl acetate, or benzene,
but is sparingly soluble in boiling light petroleum, from which it
crystallises on cooling. A solution containing 0'5015 gram dissolved
in 25 c.c. chloroform at 21^ gave a^ - 2^20' in a 2-dcm. tube, whence
the specific rotatory power [a]^ - 58'1°. The benzoyl derivative was
prepared by the Schotten-Baumann method ; it dissolves very readily
in alcohol and in warm petroleum, crystallising from the latter in
lustrous, silky needles, which begin to shrink about 70^, and melt at
86°.
A solution containing 0*5012 gram dissolved in 25 c.c. of chloro-
form at 21° gave od - 1^9' in a 2-dcm. tube, whence the specific rotatory
power [o]d 28*7*'.
CH
P'Chloro-a'bromocamphar, CgHjjOI^I^
HBr
Five grams of )3-chlorocamphor were heated with 5 grams of
bromine in an open flask on the water-bath, when, after a short
interval, hydrogen bromide was evolved. Heating was continued
until the gas was no longer liberated, the pale yellow liquid solidify-
ing to a camphor-like mass on cooling ; this was washed with water
until free from acid, dried in the desiccator, and crystallised from a
small quantity of boiling petroleum, in which it dissolves readily.
After being recrystallised several times from this solvent, it was ob-
tained in tabular aggregates of white prisms, melting at 10 P.
0-1445 gave 01 790 AgOl + AgBr. 01 + Br = 4315.
OjoHi^OOlBr requires CI + Br = 43*50 per cent.
The substance dissolves very readily in cold ethyl acetate, benzene,
or chloroform ; also in hot alcohol, from which it crystallises in well-
formed, oblong prisms. A solution contaiDing 0*5021 gram in 25 c.c.
of chloroform at 20° gave ao 5°5' in a 2-dem. tube, whence the specific
rotatory power [ a]© -h 126*5°.
JSthers of l-Hyd/roxycamphene*
Further evidence of the hydroxvUc character of hydrozycamphene
was obtained by converting it into ethers, the method which has been
studied by Lander (Trans., 1900, 77, 729) having been found very
suitable for this purpose.
The methyl ether, C^qHi^-OOH,, was prepared by treating 15 grams
of hydroxycamphene, dissolved in 45 grams of methyl iodide, with 35
grams of dried silver oxide ; no apparent action took place in the cold,
and the mixture was therefore heated in a reflux apparatus on the
Digitized by
Google
274 forster: studies in the camphane series, part vii.
water-bath during 8 hours. ' The product was then filtered, and the
residue washed several times with dry ether, which was afterwards
removed by evaporation. The residual oil was heated with a further
quantity of methyl ^iodide (30 grams) and silver oxide (24 grams)
during a period of 8 hours, and subsequently, after removing excess of
methyl iodide and the ether employed in washing the silver residue,
distilled under the ordinary pressure. •
Tbe substance obtained in this manner is a colourless, limpid oil,
readily volatile in steam, and having a pleasant odour. It boils at
193—194'' under 766 mm. pressure, has the sp. gr. 0'9314 at 20720^
and gives ai> - 50^63' in a 2-dcm. tube at 20^, whence the specific
rotatory power [ajo —27*31°; a solution containing 0*5012 in 25 c.c.
of bensene at 20 gave a^ — 58' in the same tube, corresponding with
[a]D -24-5<'.
01621 gave 0-4743 COj and 0*1531 HgO. C-79-80; H = 10-60.
CiiHigO requires 0 = 79-52 ; H = 1084 per cent.
A solution of methoxycamphene in chloroform decolorises bromine
immediately in the cold.
The ethyl ether, GjQnj5*OC2Hjj, was obtained by the same prooess
from hydroxycamphene and ethyl iodide ; it boils at 203 — 204° under
a pressure of 760 mm.
An attempt to hydrolyse the substance with alcoholic potash was
unsuccessful, a specimen which was heated with that agent during
several hours in a reflux apparatus remaining unchanged.
Bromine converts ethoxycamphene into )3-bromocamphor. Five
grams were dissolved in glacial acetic acid containing sodium acetate
and treated with bromine in the same solvent ; on pouring the liquid
into water, a crystalline precipitate was obtained, and this, when
recrystallised from petroleum, melted at 78°.
An attempt to prepare the benzoyl derivative of hydroxycamphene
by the action of benzoyl chloride (2 mols.) on the hydroxy-compound
dissolved in pyridine (3 mols.) gave rise to a pale yellow, viscous oil,
which did not crystallise.
rotal collxoe of scibnoe, london.
South Kensington, S.W.
Digitized by
Google
RESOLUTION OF TRIMETHYLHYDRINDONIUM HYDROXIDE. 275
XXVIII. — Resolution of Trimethylhydrindoiiium Hydr-
oxide into its Optically Active Components.
By Fbsdebio Stanlbt Kipping.
Many attempts have already been made to resolve hydrindamine into
its optically active components by fractionally crystallising its salts
with optically active acids; the results of such experiments have
shown that, in the case of some acids, two partially racemic hydrind-
amine salts are obtained in very unequal quantities (Trans., 1900, 77,
861 3 Kipping and Hall, Trans., 1901, 79, 430), whilst in that of
others only one such partially racemic compound is formed (Trans.,
1901, 70, 370 ; Kipping and Hall, he. eU., and Trans., 1901, 79, 446) ;
in no case has it been found possible to obtain a salt which gives, on
decomposition, an optically active base.
Under these circumstances, it appeared interesting to study the
behaviour of externally compensated trimethylhydrindonium hydroxide
(Kipping and Hall, Trans., 1900, 77, 469). This base contains an
asymmetric carbon group very similar to that present in hydrind-
amine,
NMe,-
•OH NH,
Trimethylhydrindouiam hydroxide. Hydrindamine.
and might therefore be expected to show much the same behaviour as
the latter towards a given optically active acid.
Experiments showed that this was not so ; when trimethylhydrind-
onium bromocamphorsulphonate is submitted to fractional crystal-
lisation, it does not yield two partially racemic salts corresponding
with the hydrindamine bromocamphorsulphonates, but is gradually
resolved into the salts of the enantiomorphously related bases ; the salt
of the dextrorotatory base which is finally isolated gives, on decom-
position with potassium iodide, an optically active trimethylhydrind-
onium iodide. The behaviour of the quaternary base, therefore^ is
perfectly normal.
In attempting to account for the abnormal character of the hydrind-
amine salts, it wAs suggested that the observed isomerism might be
conditioned by the configuration of the quinquevalent nitrogen atom.
Now since the trimethylhydrindonium salts contain the group -NMe^X,
whic&, as regards any possible isomerism of this kind, would seem to be
identical with the group -NH3X, contained in the hydrindamine salts,
it might be concluded that, on the resolution of the quaternary salt,
such a view became untenable. This, however, is not a necessary or
Digitized by
Google
276 KIPPING: RESOLUTION OF TBIMETHTLHTDRINDONIUM
even probable consequence of the new fact ; the bromocamphorsnlphate
of the externally compensated quaternary base may consist of four
isomerides corresponding with those which, it has been suggested, are
formed from externally compensated hydrindamine and bromocamphor-
sulphonic acid, and the only difference in the two cases may be that in the
former the four isomerides do not unite in pairs to give partially raoemic
salts; on fractional crystallisation, therefore, under the conditions
employed, the most sparingly soluble salt is isolated, leaving a
mixture which would contain three other isomerides. It would be
useless at this stage to consider further the question of the isomerism of
nitrogen compounds ; experiments which are in progress may throw
some light on the subject.
EXPEBIMBNTAL.
The racemic trimethylhydrindonium iodide (Kipping and Hall,
Trans., 1900, 77, 470) which was required for this work, was prepared
by digesting an aqueous alcoholic solution of hydrindamine hydro-
chloride with methyl iodide and excess of potash until the solution no
longer gave a vapour alkaline to litmus ; on cooling, most of the
product separated in pale yellow crystals, and the portion remaining
in the mother liquors was obtained by evaporating to dryness with
dilute hydrochloric acid, extracting the residue with boiling alcohol,
and then conceentrating the alcoholic extract.
Trimethylhydrindonium Bromocampharsulphanaie,
CigHigNO-SOg-CioHi.BrO.
On adding silver bromocamphorsulphonate, dissolved in aqueous
alcohol, to a solution of trimethylhydrindonium iodide in the same
solvent, a heavy precipitate of silver iodide is immediately formed ;
after heating together molecular proportions of the two compounds for
about 30 minutes under these conditions to complete the interaction,
the filtrate, on evaporation; yields an almost colourless syrup which
soon begins to crystallise when kept over sulphuric acid. ^
The crude salt, when merely freed from oil with the aid of porous
earthenware, consists of an ill-defined, crystalline mass, and has no
definite melting point, liquefying gradually between about 155® and
165°; it is very readily soluble in water, alcohol, acetone, or
chloroform, but only sparingly so in cold ethylic acetate. During
damp weather, it liquefies or becomes pasty on exposure to the air.
In dilute aqueous solution, the salt gives the normal rotation, that
is to say, its molecular rotation is the same as that of bromocamphor-
Bulphonic acid.
Digitized by
Google
INTO Its OPTICALLY ACTIVE COMPONENTS. 277
P'25 gram dissolved in water, the sohition dilated to 25 c.c. and
examined in a 200 mm. tube, gave a +1^6'; hence [a]i> +55^ and
[M]d +267°.
A halogen determination was made with a sample dried over sul-
phuric acid :
0-2092 gave 0-0800 AgBr. Br- 16-3.
Calculated for O^sHg^NO^BrS. Br :» 16*5 per cent.
IsolcUian of d-lHmethylhydrindanium Bromocampharsulphonate, —
The best way of carrying out the fractional crystallisation of the
bromocamphorsulphonate is to dissolve the crude product in a small
quantity of chloroform, and then heat on the water-bath until nearly
all the chloroform has evaporated ; the salt is thus freed from traces of
moisture and is left in the state of a syrup, which is mixed with a
considerable volume of warm, dry ethyl acetate ; on cooling, crystal-
lisation occurs and the salt separates in small, nodular aggregates or in
the form of bulky masses of needles, the sohition often setting to a
paste which is difficult to filter. The deposit is separated, dissolved in
dry chloroform, and the liquid treated as before. As these operations
proceed, the melting point of ^ the most sparingly soluble fraction
gradually rises and becomes more sharply defined, until after some
10 — 20 crystallisations a uniform product, namely, c^trimethyl-
hydrindonium (^-bromocamphorsulphonate, is obtained. The rise in
melting point is accompanied by an increase in the specific rotation of
the most sparingly soluble fraction in aqueous solution ; thus after
four operations, the specific rotation was [aj^ +60*5°; after ten,
[a]o +67-4°; after sixteen, [a]j, +71-3°
The mother liquors from the first operations give, on evaporation,
deposits which can be easily crystallised from a mixture of chloroform
and ethyl acetate, but the melting points of such deposits vary in a
very irregular manner on recrystallisation,'sometimes rising and some-
times falling; for this reason, farther attempts to isolate the bromo-
camphorsulphonate of the ^base were given up, and the mother liquors
were directly employed for the preparation of the iodide of the ^base
(p. 279).
J-Trimethylhydrindonium d-bromocamphorsulphonate is obtained in
long, slender needles or prisms when its solution in chloroform and
ethyl acetate is slowly evaporated ; it melts at 199 — 200°, decom-
posing a little. It is very readily soluble in water, acetone, alcohol,
or chloroform, but almost insoluble in dry ether or ethyl acetate, even
on boiling.
The specific rotation was determined in aqueous solution with three
different preparations, of which II and III were successive crops of one
sample.
Digitized by
Google
278 KIPPING: RESOLUTION OF TRIMETHTLHYDRINDONIUM
I. 0*3445 gram dissolved in, and made up to 25 c.c. with, water,
examined in a 200 mm. tube, gave a + 1*97° ; whence [ajo + 71*5°.
II. 0-5088 gram under the same conditions ; a + 2-83° ; [a]n + 69-5^
III. 0-5040 gram under the same conditions; a +2*78^; [ajo +690°
The close agreement of the results obtained with II and III may
be taken as evidence of the purity of the salt ; the rather higher value
of that obtained with I is possibly due to the solution being more
dilute, determinations in alcoholic solution having shown that the
specific rotation is considerably greater in the latter, less dissociating,
solvent. That the salt is a pure substance is also shown by its
behaviour on decomposition with potassium iodide.
Taking the mean of the above results, the molecular rotation of the
salt is + 340°, and that of the acid ion being + 270°, the molecular
rotation of the base ion would be + 70° ; this value agrees well with
that calculated from the specific rotation of the iodide.
d'Trimethylhydrindoniwn Iodide^ CijHjgNI.
The iodide of the ef-base is so much more readily soluble in water
than that of the (i^-base that it is not precipitated on mixing concen-
trated aqueous solutions of c^trimethylhydrindonium bromocamphor-
sulphonate and potassium iodide ; it is best, therefore, to isolate the
salt by evaporating the mixed solutions to dryness and extracting the
powdered residue with boiling chloroform, in which the quaternary
iodide is readily soluble.
It crystallises from water in long, highly lustrous, transparent
needles, which, like the crystals of the racemic iodide, are anhydr-
ous, but are quite different from the latter in crystallographic
character, the racemic iodide forming compact, well-defined prisms.
The salt of the active base decomposes at 190 — 191°, giving, doubtless,
indene and trimethylamine hydriodide (Kipping and Hall, Trans., 1900,
77, 467) ; oh heating the optically active and racemic iodide simultan-
eously, a difference in decomposing point of about 5° or 6° can be noticed,
the racemic compound remaining unchanged until about 197 — 198°.
The active iodide is readily soluble in hot water, alcohol, or chloroform,
but practically insoluble in ether or ethyl acetate.
That the salt just described, also the bromocamphoreulphonate of
the (i-base, are pure compounds, was proved by fractionally crystallising
the iodide from warm water, and thus separating it into three portions ;
the whole was obtained in long needles, free from crystals of the
racemic iodide, which if present could be easily recognised, as was
proved by crystallising a mixture of the two salts.
As further proof of the purity of the iodide, the following deter-
minations of the specific rotation of the first two fractions may serve :
Digitized by
Google
INTO ITS OPTICALLY ACTIVE COMPONENTS. 279
I. 0*5333 gram dissolved in water and the solution diluted to 25
«c.c. gave on examination in a 200 mm. tube a +0*92°; whence
[a]|> +21-6°.
II. 0*3949 gram under the same conditions: a +0*70°; whence
[a]o +22-2.
Taking the specific rotation to be +21*9^, the molecular rotation is
[M]i> +66*3^, a value which agrees well with that deduced from the
molecular rotation of the bromocamphorsulphonata
PioraUa of the ExUmally Compensated and d-Bases.
The picrates of the optically inactive and of the active bases are
only very sparingly soluble in cold water and are easily obtained from
the iodides by precipitation. The racemic picrate crystallises in com-
pacty yellow prisms which melt and decompose at about 1 88^ ; the active
compound is more readily soluble in both water and alcohol^ from which
it separates in long, very slender needles melting at about 167°; the
differences between the picrates are therefore similar to those between
the iodides.
I'lHmethylhydrtndonium Iodide.
As the deposits of trimethylhydrindonium bromocamphorsulphonate
remaining after separating some of the salt of the (i-base were found
to behave as if they were complex mixtures, and it did not seem
possible to separate from them any other salt, the last mother liquors
were evaporated and the residue treated with potassium iodide as
already described; after separating the racemic iodide, which was
precipitated in large quantities, the nmother liquors were evaporated
to dryness and extracted with boiling chloroform. In this way, a
salt, obviously the iodide of the M>ase, was isolated without difficulty ;
it crystallised from water in large, fern-like masses, and its aqueous
solution was Invorotatory.
The author's thanks are due to Mr. L. L. Lloyd for assistance in
preparing some of the materials for this investigation.
Part of the expense incurred in carrying out this work was met by
a grant from the Government Qrant Fund of the Royal Society, for
which the author desires to express his thanks.
UxrVXBSITT COLLXQX,
Nottingham.
Digitized by
Google
280 SENIER AND GOODWIN: THE ACTION OF METHYLENE
XXIX. — The Action of Methylene Diiodide on Aryl-
and Naphthyl' amines : Diarylmethylenediamines,
AcridineSy and Naphthacridines.
By Alfbed Senieb and William Goodwin.
Ik a recent paper (Trans., 1901, 79, 254), we described two ethylene-
diaryldiaminea which resulted from the interaction of ethylene di-
bromide and zylidine and ^cumidine respectively. At the same time,
we mentioned that we had extended the reaction to the methylene
series and had obtained well crystallisable compounds. Methylene
diiodide was found to give better results than either the dibromide or
dichloride. We now submit the results of further study of those
compounds and of the reaction generally. It was expected that
these derivatives would prove to be methylenediamines or, at least,
to belong to one of the classes of compounds which are known to
result from the action of ethylenedi bromide on aniline. They might,
for example, be methylenediamines, OH2(NB[R')2, methyleneaniline
TO"!!'
homologues, CH^INR', methylenepiperazines, CHj^^j^r^^CHj, or
.^ CH,-NR'-CH, ^^. .
higher polymendes, X , -hryf^ This view is in agreement
with that of previous inquirers (LermontofiF, Ber,^ 1874, 7, 1252 ;
Gruuhagen, AnnaUny 1890, 256, 219 and 285 ; BisohoS and Nast-
vogel, ^er., 1890, 23, 2065).
Parallel with the reaction of arylamines and methylene diiodide is
the corresponding reaction between arylamines and formaldehyde or
methylene oxide, CH^IO. The experiments of Pratesi {Giizz^Ucky 1884,
14, 351), Tollens {Ber,, 1884, 17, 653), Wellington and ToUens (^ar.,
1885, 18, 3298), Eberhardt and Welter {Ber., 1894, 27, 1804), and
Grassi-Cristaldi and Schiavo-Leni (Gazzetta, 1900, 80, ii, 112) with
tbis latter reaction led to the discovery of the diamines and isomeric
methylenearylamines previously referred to. The diamines were also
shown to give rise by metastatic change to a series of methylene
benzidenes. Methylenediphenyldiamine, CH2(NHPh)2, for example,
. giving diaminodiphenylmethane, CH^(C0H4NH^)2.
The study of these reactions by the investigators mentioned, evi-
dently conducted with the greatest care, has led to very contradictory
results. The physical characters of the compounds described under
the same name differ enormously, and the analytical data are far
from satisfactory (compare Bischoff and Nast vogel). Working with
aniline and the toluidines, our experience has been the same. This
is due, no doubt, partly to the proved instability of the compounds
Digitized by
Google
DIIODIDE ON ARYL- AND NAPHTHYL-AMIN ES. 281
formed, and partly to their number and the difficulty of separating
them. With the increase in the number of alkyl groups, as in the
case of zylidine and ^cumidine, the reaction becomes more definite,
and with methylene di-iodide assumes a new character. With
^-cumidine, beautiful, yellow crystals are obtained, easily recrystal-
lised or sublimed, and in this way readily obtained of a constant
and definite melting point. Elementary analysis proved that it was
not a member of any of the classes of compounds alluded to above.
It gave C»86*8, Ha 8*41, Nb5'42 per cent., whereas methylene-
dioumyldiamine or its isomerides would require 0^ 80*85, Hs9'22,
N«9*93 per cent., and methylenecumidine or its isomerides would
require C»81'63, He8'84, N«9'52 percent. The relative proportion
of nitrogen found pointed to a condensation reaction whereby ammonia
or a derivative was evolved. This view found confirmation in the fact
that the odour of ammonia was observed during the course of the re-
action, and also that a notable quantity of ammonium iodide collected in
the tube of the reflux condenser employed in the experiment. The solu-
tions of the compound exhibited a marked and very beautiful fluor-
escence, and this, together with the high melting point, 221 — 222°,
also indicated condensation.
Consideration of the paper by Friedel and Grafts (Ann. Chim, Phys,^
1887y [vi], 11, 263) led to the hypothesis that the^compound was an
anthramine from which an amino-group had split off. These investi-
gators found that benzene and methylene dichloride condensed to
anthracene in presence of aluminium chloride thus :
2CeH, -H 2CH,Clj = C,H,<^^>OeH, + 2HCI -h H^.
The hydrogen formed, acting on an excess of methylene di-
chloride, gave, by a secondary reaction, methyl chloride. To make
certain of the right conditions, we repeated this experiment and
obtained anthracene. Oumidine was substituted for benzene and the
experiment repeated, but no reaction appeared to take place, and when
methylene diiodide replaced the dichloride in the last experiment, the
yellow, crystalline compound was obtained, but only to the same extent as
in the absence of the aluminium chloride. We next modified the hypo-
thesis by supposing condensation to take place at the position of the
liberated nitrogen group. In this way, an acridine might be formed,
thus:
20eHjjMe,NH2 -i- CHjIj = CeHMe3<^^> C^HMej -t- N H^I -t- HI -i- H^^.
The odour of methyl iodide was always noticed, which would explain
what becomes of the hydrogen written as free in the equation as in
the reaction of Friedel and Crafts. Such a hexamethylacridine would
Digitized by
Google
282 SENIER AND GOODWIN : THE ACTION OF METHYLENE
require €» 86*61, Hs8*06, N»5'33 per cent., which is in agreement
with the experimental numbers. A vapour density determination, using
boiling sulphur, was found to be possible by Victor Meyer's method :
0-1303 gave 11*8 c.a at 18° and 767*5 mm. Density «130.
C^gHjiN requires density =■ 131*5.
Further evidence in favour of this view was found in the experiments
of Bemthsen and Bender {Ber., 1883, 16, 1802), who prepared acridine
from diphenylamine and formic acid, for dicumylamine might well
result from cumidine and hydriodic acid, and formic acid is hydroxy-
methylene oxide. Similar analogies were found in the work of Mohlau
(Ber., 1886, 19, 2451).
Any doubt, however, that might remain as to the ^-cumidine com-
pound being hexamethylacridine was set at rest by the extension of the
reaction to the a- and /3-naphthylamines. These bases were found to
react with methylene diiodide in precisely the same manner as ^-cumid-
ine. The same odour of ammonia and of methyl iodide was noticed,
and the same condensation of ammonium iodide in the reflux condenser
tube. From each of the naphthylamiues we obtained yellow crystals
with well-defined melting points, and their solutions exhibited the
characteristic, well-marked fluorescence. On analysis, the numbers
obtained agreed with those of naphthacridine. Finally, the )3-compound
was identified by its melting point and other characters with /3-naphth-
acridine discovered by Beed {J. pr. Chem., [ii], 1886, S4^ 160; 1887»
36, 298) as a result of the action of methylal or of formaldehyde on
^•naphthylamine. This base has also been studied in an interesting
paper by Morgan (Trans., 1898, 73, 536). In a preliminary communi-
cation (Proc., 1898, 14, 132), Morgan announced that the result of
experiment^ on a-naphthylamine would be published later. Meanwhile,
we have obtained a-naphthacridine by the methylene diiodide reaction.
We find, then, that methylene diiodide yields methylene diamines
with aniline, as originally stated by Lermontoff , also with xylidine and
probably with the three toluidines. With ^umidine, however, and
with both a- and /S-naphthylamines, condensation takes place with the
formation of acridines. The acridines do not react with phenyl carb-
imide forming ureas, but the methylenediamines do so in the same
manner as their ethylene homologues (Trans., 1901, 79, 258X at least
so far as we have been able to complete their investigation.
The further study of a-naphthacrldine and of hexamethylacridine we
desire to reserve for another communication.
Digitized by
Google
DIIODIDE ON ARYL- AND NAPHTHTL-AMINES. 283
Diary Imethylenediamtnes.
Diphenylfnethylenediaminef CK^{NH.Fh\,
Methylene diiodide and aniline, in eqvuJ molecolar proportions, were
brought together with an excess of dry potassium carbonate in a flask
fitted with a reflux condenser. The flask was placed in a paraffin-bath
which was heated gradually to about 150°, when a violent reaction soon
took place. The flask was then removed from the bath until this
moderated, when it was replaced and the heating continued for 4 hours.
After cooling, the residue in the flask was treated with hot water to
remove potassium salts. The dark red, semi-solid mass remaining was
freed from aniline by distillation with steam. The resulting mass
solidified on cooling, and, when reduced to powder, was of a dark yellow
colour. It melted at 65 — 67^. After boiling with alcohol, the melting
point rose to 1 40°. Diphenylmethylenediamine, described by Eberhardt
and Welter, melted at 64 — 65°, and by boiling with alcohol it was
changed into methyleneaniline which melts at 139°. On analysis :
0-2502 gave 0-7148 00, and 01590 B,0. 0-77-92 ; H = 706.
C13H14N2 requires 0 = 78-75 ; H = 7*07 per cent.
Liphmplmethylenediamine Platinichloride, CB^^(SB.Vh)^IL^VtClfi,
separated as an olive-green precipitate on the addition of a solution of
platinio chloride to the base dissolved in dilute hydrochloric acid. The
salt was washed with water and dried in a vacuum over sulphuric a^id.
On analysis :
0-2318 gave 00736 Pt. Pt«31-75.
OjjHi^N^HjPtClft requires Pt = 32-06 per cent.
Diearh(mUiiodiphmyl^^ OHs(NPh*00-NHPh)s, was
obtained by treating the base with phenylcarbimide in the manner
described by us (Trans., 1901, 79, 258). The dosed tube was heated
at 120 — 124° for 3 to 4 hours. The residue was washed with dry
ether and recrystallised from glacial acetic acid. The melting point is
not definite. On analysis :
0-2750 gave 0-7397 CO, and 01668 H^O. 0 = 7335 ; H- 674.
CgyHj^N^Og requires 0 » 74*31 ; H « 5-50 per cent,
DUolylme^ylenediafniiiei, 013i^{^U*C^R^Ue)^.
Methylene diiodide (1 mol.) with toluidine (2 mols.), together with
potassium carbonate, were heated in the manner described for the
phenyl-homologue and the subsequent procedure was the same, except
that the distillation with steam was omitted. The o-tdyl derivative,
Digitized by
Google
284 SENIER AND GOODWIN: THE ACTION OF METHYLENE
probably di-o-tolylmethylenediamine, crjBtallised from alcohol in large,
colourless, foliate crystals. After repeated reciystallisation, these
melted at 156— 157^ On analysis :
0-2888 gave 08374 00, and 0-2124 H,0. 0 = 7908 ; H = 817.
0-3560 y, 38 c.c. moist nitrogen at IG"" and 756-5 mm. K » 12*41.
C^gHigNj requires 0 = 7964 ; H « 796 ; N = 1239 per cent.
From o-toluidine and methylene chloride, Grlinhagen obtained two
bases ; one was a liquid and the other melted at 135^. Using form-
aldehyde, Eberhardt and Welter prepared a base which melted at 52°.
The p-tolyl derivative, probably di-^htolylmethylenediamine, was ob-
tained in small, pale yellow crystals from solution in a mixture of
alcohol and chloroform. After several recrystallisations, it melted at
149_150o. On analysis :
0-2562 gave 07456 CO, and 01800 H^O. 0 = 79-37 ; H = 78.
01630 „ 0-4776 002 „ 01168 H^O. C=- 79-91 ; H = 7-96.
CigHjgNj requires 0 = 79-64 ; H - 796 per cent.
We were not successful in isolating a m-tolyl derivative in a definite
form. Yellow, amorphous compounds were obtained in two experi-
ments. The one melted at 78"", the other at 160"".
DixylylmMylenedtaminef GR^{^'EL'0^1I^yLe^)^
Methylene diiodide (1 mol.) and zylidine (2 mol&) were treated in
the same manner as in the corresponding reaction with the toluidines.
The base, recrystallised several times from rectified spirit, consisted of
pale yellow, large, foliate crystals, and melted at 127 — 128°. On
analysis :
0-2592 gave 07576 CO, and 02140 H^O. C« 79-71 ; H«9-17.
0-3743 „ 35 -2 c.c. moist nitrogen at 1 3° and 767 mm. N » 1 1 -2 1 .
Ci^H^Nj requires 0 = 80*31; H«8-67 ; N = 11-02 per cent.
Dixylylmethylenediamins PlatiniMoride,
CH,(NH-0eH3Mej)jpHjPt01e,
was precipitated when an alcoholic solution of platinic chloride was
added to an alcoholic solution of the base acidified with hydrochloric
acid. The precipitate, of a golden-yellow colour, was driM at 100°.
On analysis :
0-2076 gave 00620 Pt. Pt= 29-86.
Ci^HjjNj requires Pt = 29*36 per cent.
iTtfro-derivatives of dizylylmethylenediamine are produced readily
by direct nitration. Two experiments were made, but the products
were evidently mixtures. ^One contained Nb11«5 and the other
Digitized by
Google
DUOBIBE ON ABTL- AND NAPHTHTL-AMINES. 285
K» 12*7 per oent. A mononitro^erivative requires N=:9'82 and a
dinitro- N« 16*28 per cent.
l)ieafrhanUidooafh(mUidodi^
CHj;N(C«HgMe,)-CO*NPh-CO-NHPh]y
— When an excess of phenylcarbimide reacts with dizylylmethjlenedi-
amine, a urea derivative is formed from 4 molecules of the carbimide
instead of 2 molecules, as in other cases. The reaction described
(Trans., 1901, 79, 258) may be supposed to occur twice in each amino-
group of the diamine. This reaction is, indeed, only an extension of
Wohler's urea synthesis, for it may be a^umed that a methylene-
diunmonium dioyanate is first formed, which metastasises to urea. An
excess of phenyl carbimide was heated in a closed tube with the base
at 150^ for 6 hours ; the contents of the tube, freed from excess of
phenylcarbimide by evaporation, consisted of a yellowish-white sub-
stance resembling in appearance imperfectly bleached beeswax, this
was washed with dry ether, which effected no apparent change.
When rubbed in a mortar with successive small quantities of alcohol,
there was left a hard, white residue, which was dissolved in more
alcohol, and was recrystallised several times from that medium. Well-
formed, colourless needles were thus obtained which melt at 203°.
On analysis :
0*2766 gave 0*7438 00, and 0*1470 HjO. C - 73*33 ; H = 5*9.
0*2978 „ 281 C.C. moist nitrogen at lO^'and 754 mm. N» 11*21.
C^jH^gNgO/requires C = 73*97 ; H = 5*75 ; N - 1 1 -5 per cent.
'8<S^6l
Uexamethylacridiney C^HMej'^Jv^.^CgHMeg.
The constitution of this hexamethylacridine we reserve for further
investigation. If, however, it be supposed that the positions of con-
densation are neighbouring to the amino-groups, the positions of the
methyl groups are 1:2:4 and 5:6:8. ^-Cumidine, which melted at
67^ (2 mols.), with methylene diiodide (1 mol.), and an excess of pot«
assium carbonate were taken. The cumidine and potassium carbonate
were melted together and then brought into a flask fitted with a reflux
condenser, to which the weighed amount of methylene diiodide was
then added ; the flask and its contents were heated in a paraffin bath
until a temperature of about 160° was reached, when an energetic re-
action took place ; the flask was removed from the bath until the
violence had subsided, when it was replaced and the temperature con-
tinued at 150 — 160° for about 6 hours. The formation of a sublimate
in the tube of the condenser continued during the course of the reac-
tion, and its ceasing to form was an indication that the heating had
been sufficiently prolonged. This method was found to give better
VOL. L2XXI. U
Digitized by
Google
286 SENIER AND GOODWIN: THE ACTION OF HETHTLENE
results than when the diiodide was added in small quantities at a
time. No advantage, either, was found in omitting the alkali and
treating the product afterwards with potassium hydroxide. The con-
tents of the flask were then treated with hot water to remove potassium
salts, when there remained a yellow, crystalline mass; after being
washed with dilute alcohol, this assumed a greenish-yellow colour. It
was next dissolved and recrystallised from alcohol or, in some cases,
from acetone or light petroleum, when extremely fine, yellow crystals,
usually needles, were obtained. These crystals are slightly soluble in
alcohol or ether and more soluble in acetone, chloroform, benzene,
or light petroleum. They are insoluble in water. All the solutions
exhibit a beautiful, green fluorescence. Purified by repeated recrys-
tallisation or sublimation, the compound melts at 221 — 222°, and
sublimes easily without decomposition. On analysis :
0-2406 gave 0-7658 COj and 0-1823 H^O. 0 = 86-80 ; H = 8-41.
0-2622 „ 0-8290 OOj „ 0-1992 HjO. 0 = 86-23 ; H = 8-44.
0-2464 „ 0-7756 COj „ 0-1770 HgO. 0 = 86-20; H = 801.
0*3430 „ 16 c.c. moist nitrogen at 15° and 755*5 mm. N = 5*42.
O10H21N requires 0 = 86*69 ; H = 798 ; N = 5*32 per cent.
Using methylene dibromide instead of diiodide, the acridine was
obtained, but the yield was much less. With methylene dichloride,
no acridine was formed. In the latter experiment, a marked odour
of an i^onitrile'was noticed. «.
Heaxunethf/laeridine Piorate, CjHMej<^b>^6H^^®s'CeH2(N02)j-0H,
was prepared by adding an alcoholic solution of picric acid to an
alcoholic solution of the base, heating to boiling, and allowing to stand.
Yery fine, brown, plume-like crystals separated. These were collected
on a filter and washed with a small quantity of alcohol and again
with dilute alcohol and dried at 1 10°. They melted, not very sharply,
at 200—202°. On analysis :
0-3090 gave 30*5 c.c. moist nitrogen at 18° and 752*5 mm. N = 11-27.
0„H,iN,OeH3(N02)80H requires N= 11-38 per cent.
Dinilroliexamethylaeridtne, Oe(N02)Meg^jC_^^0^(NO,)Me8,waspro-
duced by treating the base with concentrated nitric acid, boiling well,
and pouring the solution into water; the yellow precipitate was
collected on a filter, well washed with water and dried on a porous
plate. It melted at 86 — 87°. It was dried in an exhausted desiccator
over sulphuric acid, and on analysis :
0-2064 gave 20*3 c.c. moist nitrogen at 18° and 761 mm. N = 11*37.
OjgH„(NOj)jN requires N= 11'89 per cent
Digitized by
Google
DnODII^* OK ARTL- AND NAPHTHTL-AMINES. 287
Trtbramolisxamethylacrtdiney C^BrMcj^Y ^^^CgBrMej, was obtained
bj ezposiDg the base to the action of bromine vapour for 3 days
in a covered vessel. The product was heated on a water-bath to
remove excess of bromine and then allowed to stand in a desiccator
over solid potassium hydroxide for 2 days. Finally, the red compound
was heated in an air-bath at 105^. On analysis by Carius's method :
0-31 16 gave 0-3466 AgBr. Br = 4733.
CjgHigBrjN requires Br = 47 98 per cent.
Hexameihylacndine NUraJe, CjHMe8<[V_J>CgHMe„HN08, was
formed by adding dilute nitric acid to an alcoholic solution of the
base, warming for a short time on a water-bath, &nd then setting aside
to crystallise. On standing, the solution, which had a deep orange
colour, deposited bright scarlet needles. An attempt was made to
further purify the compound by recrystallisation from dilute alcohol,
but this treatment was found to decompose the salt regenerating the
original base. The needles were washed with a little alcohol only and
partly dried by pressing between filter-paper. Finally, they were
dried at 100° and left over sulphuric acid in a desiccator. They melted
with decomposition at 163 — 164°. On analysis :
0-3366 gave 25-1 c.c. moist nitrogen at 14° and 770 mm. N = 8-92.
CigHj^NjHNO, requires N = 8-59 per cent.
UtxafMlhylacridiiM PkUinichloride,
(OeHMe3<g^C,HMe,),H,PtCl„
separated in the form of scarlet needles on the addition of a solution of
platinic chloride to a solution of the base in dilute hydrochloric acid.
Like acridine salts generally, it is decomposed by treatment with water
or boiling with alcohol. On analysis :
00577 gave 001 18 Pt. Pt = 20-45.
(CigHjiN)2,H2Pt01g requires Pt = 20-82 per cent.
Hexamelhylacridine Aurichloride, CgHMe8«^X_^O^HMe3,HAuOl4,
was obtained in a similar manner to the platinichloride. It is yellow,
and on analysis : •
01 1 10 gave 00358 Au. Au = 32-25.
(GiqHjiN)HAuC14 requires Au=: 32*70 per cent.
Htasamdhylcberidine MercuricUoride, CgHMeg-^X^^CgHMegjHgClj.
— An alcoholic solution of the base acidified with hydrochloric acid was'
Digitized by
Google
288 SENIER AND GOODWIN: THE ACTION OF UETHTLENE
mixed with an alcoholic solution of mercuric chloride and the mixture
allowed to stand. Dark red, glistening needles with a characteristic
yellow lustre separated. On analysis :
0-3076 gave 01272 HgS. Hg = 35-66.
(0„HaN)HgCIj| requires Hg = 35-11 per cent.
HexameAylacridine DiehromaU, CgHMej<^Y^^^CjHMe3,HjCrj07,
was formed on the addition of a solution of potassium dichromate to a
solution of the base in hydrochloric acid. It consists of deep red
crystals which were washed rapidly and dried at 100°. On analysis :
0-6536 gave 01378 CrjO,. Or = U-44.
(CijH„N)H,CrjOy requires Cr= 1400 per cent.
Other salts examined were the sulpkaiSf forming dark red crystals,
which do not melt below 285° ; the hydrochloride^ red crystals which
decompose on heating, and the niirite, a pale yellow powder obtained by
adding potassium nitrite to a solution of the base in glacial acetic add.
The nitrite decomposes on heating.
An ethylhsxamethylaoridinittm iodide was also prepared but not
further examined. The base was heated with excess of ethyl iodide
in a closed tube for 4 hours. It consists of bronze-coloured crystals
which, after drying on a porous plate, melted at 214 — 215°.
Naphthacridinea, (CioHeKS— >{0io^6)-
a-IfaphihcMridine.
a-Naphthylamine (2 mols.), methylene diiodide (1 moL), and an excess
of potassium carbonate were treated exactly as in the case of the corre-
sponding cumidine experiment. The contents of the flask were washed
with water, then with a small proportion of alcohol, after which the
residue was dissolved in and recrystallised from alcohol. More than
one compound appeared to be formed, and the acridine was best separ-
ated by sublimation and subsequent recrystallisation from alcohol.
It consists of pale yellow crystals which melt at 173°. The colour
becomes darker on exposure to light. It is soluble in chloroform,
ether, acetone, benzene, or light petroleum. The fluorescence exhib-
ited by its solutions is very noticeable. In solution in benzene, it is
red-green, in the other solvents it is violet or blue. On analysis :
01532 gave 0-5040 CO, and 00707 H,0. 0 = 8972 ; H-5-12.
0-1783 „ 7*6 c.c moist nitrogen at 9° and 764 mm. N-5-15.
C^^H^N requires C » 9032 ; H » 4-66 ; N » 502 per cent
The following are the only derivatives obtained so far, but the work
is being continued.
Digitized by
Google
DIIODIDE ON ARTL- AND NAPHTHTL-AMINES. 289
MoTumitro-anaphthc^oridinej C2iHi2(^^2)^*^^ precipitated when the
base was warmed with concentrated nitric acid and the mixture thrown
into water. After washing and drying, first on a porous plate and
then in an exhausted desiccator over sulphuric acid, it melted at
106—107° On analysis :
0*3248 gave 24*8 c.c. moist nitrogen at IG^'and 767*5 mm. N»8-85.
C2iHi2(N02)N requires N-8-64 per cent.
A second ni^o-derivative was prepared which, after recrystallisation
from alcohol and drying, melted at 185°, but was not further examined.
{ayNdphAacfidineFlatiniMoride.O^QB^^^^
was obtained in the form of small, deep yellow crystals on allowing a
mixture of alcoholic solutions of the base and platinic chloride, acid-
ulated with hydrochloric acid, to stand. They could not be subjected
to 100° without decomposition. Dried over sulphuric acid and analysed :
0-1676 gave 0-0322 Pt. Pt = 19-2.
(CjiHi3N)2,H2PtClg,2H20 requires Pfc- 19-41 per cent.
a-yaphthacridine Fiorate, CioH5<V_>Cj^Hg,C^Hj(NOj)3-OH, was
formed by mixing alcoholic solutions of the base and picric acid. It
separated in the form of scarlet needles. These were washed with a
little alcohol and dried by pressing between filter paper, and after-
wards by leaving them in an exhausted desiccator over sulphuric acid.
It melts at 176—178°. On analysis :
0*2312 gave 21 c.c. moist nitrogen at 10° and 771*5 mm. N»s 11*03.
C,iHi3N,CeH,{N02)jOH requires N= 11*02 per cent,
P-Naphthacridme.
The process adopted for the preparation of a-naphthacridine yields
its /3-isomeride With greater readiness and in better yield. After
sublimation and recrystallisation from alcohol, it consists of pale
yellow needles which darken on exposure to light. It melts at 215*5°.
It is soluble in chloroform, ether, benzene, or light petroleum, and
sparingly so in alcohol ; the solutions exhibit the characteristic blue
fluorescence. On analysis :
0*1882 gave 0*6172 CO, and 0-0858 HjO. C = 89-44 ; H = 5-06.
0-1990 „ 0-6540 CO, „ 0-0916 HgO. 0 = 8963; H = 5*11.
0-4260 „ 17-8 C.C. moist nitrogen at 13° and 776 mm. N=5*05.
CjiHigN requires 0 = 90*32 ; H = 4-66 ; N-5-02 per cent.
The compound is identical with the /3-naphthacridine discovered by
Reed, who gives its melting point as 216°. We obtained the following
derivative, which has not been previously described,
VOL. LXXXI. X
Digitized by VjOOQIC
290 SEKIEB AND WALSH : THE POLYMERISATION OF
p-I^apkihacridine FltUinicfdaridef
(CioHe<:?J[>CioHe)^H,RCl,,^
[Hrecipitated as a golden-yellow powder on the addition of alcoholic
solution of platinic chloride to an alcoholic solution of the base
acidified with hydrochloric acid. Dried over sulphuric acid and
analysed :
0-1396 gave 0-027 Pt. Pt - 19-34.
(C2iHi3N)2,H2PtClg,2H,0 requires Pt- 19-41 per cent.
QUXIN'S Ck>LLIOX.
Galway.
XXX. — The Polyifnerisation of Cyanic Acid : Oyanuric
Add, and Cyamelide.
By Alfred Senibr and Thomas Walsh.
When liquid cyanic acid is allowed to polymerise at 0% or just above
that temperature, it changes, as is well known, into a snow-white
solid, the 'insoluble cyanurio acid" or cyamelide of Liebig and
Wohler (Ann. Phya. [Chem., 1830, 20, 384). This solid is not, how-
ever, cyamelide only, as is generally supposed, but a mixture of the
two isomerides, cyamelide and cyanuric acid. The two compounds
are readily separated by treatment with water, in which cyamelide is
very sparingly soluble. Having prepared, in this way, cyamelide in a
state of purity not hitherto described, we made determinations of its
solubility and compared it with the solubility of cyanuric acid. We
also made numerous attempts to cause it to enter into reaction with
other compounds. The results, although negative^ are in some in-
stances interesting as bearing on the theory of its constitution.
Polymertmtion of Cyanic Acid. — The liquid cyanic acid employed
was prepared by distilling dry cyanuric acid in an apparatus made of
hard glass tubing similar to that used by vonBaeyer {ArMalen^ 1860,
114, 165). The horizontal sealed end of the tube containing the
cyanuric acid was heated in a short Hofmann combustion furnace, and
so arranged that the bend leading to the U-shaped condenser portion
was kept hot by the furnace. The condenser was kept at a tempera-
ture a few degrees below 0°. Liquid cyanuric acid, which was always
slightly turbid, collected. When the temperature was allowed to rise
to 0°, the liquid changed into a white solid. The polymerisation became
violent and was accompanied by loud reports when a higher tempera-
ture was employed. 0*839 gram of this white solid, which had been
Digitized by
Google
CYANIC ACID: CTANURIC ACID, AND CYAMELIDE, 291
obtained from some cyanic acid almost free from turbidity, was finely
powdered and treated with excess of hot water. The insoluble residue,
after drying, weighed 0*253 gram and was about 30 per cent, of the
substance taken. The washings on evaporation yielded crystals of
cyanuric acid which gave the characteristic pink^ ammonio-cupric salt
of W5hler.
Solubility qf Cyamdide ofnd qf Cycmwric Aeid. — Some of the less
soluble portions of the mixed polymerides were placed in a wash-bottle
made of a large test-tube ; boiling water was added and the contents
were agitated for 2 hours by a current of air ; the apparatus was
then placed for 12 hours in water kept at about 15°; afterwards
the contents were again agitated for an hour and the proportion of
solid in solution was determined. As the result of numerous experi-
ments, it was found that the percentage of solid dissolved decreased
after each successive treatment with water in the case of every speci-
men examined, until it attained to from 0*008 to 0*01 per cent., when
it became constant. The residues from the washings until the solu-
bilities mentioned were attained responded in 'all cases to Wohler's
test for cyanuric acid, but after that point was reached the residues
ceased to give that reaction. The solubility of cyamelide in water
may therefore be taken as 0*01 per cent, at 15^ Determined in the
same manner, as . the result of very many experiments, we find the
solubility of cyanuric acid to be from 0*145 to 0*160 per cent, at 15°
(compare Lemoult, Compt, rmd,, 1895, 121, 351). A specimen of
cyamelide having the solubility mentioned was submitted to ele-
mentary analysis :
0-2762 gave 02845 COg and 00678 H^O. C = 2809 ; H = 2*72.
(C0NH)3 requires 0 = 279 ; H = 2-32 per cent.
Further Experiments with Cyamelide, — Cyamelide was treated with
phosphorus pentachloride, but without any change occurring. Cyan-
uric acid yields, under the same circumstances, cyanuric chloride, and
this was verified by an experiment. It was suggested by Klason
{J.pr. Chem,, 1885, [ii], 83, 129) that cyamelide is itfocyanuric acid
related to the t«ocyanuric esters, just as cyanuric acid is related to
the normal cyanuric esters. It is interesting, therefore, as evidence
of this view that as the normal esters and normal cyanuric acid
yield a chloride with phosphorus pentachloride, so the tso-esters, and
now it is proved cyamelide, do not do so.
Attempts were made to prepare silver, bromine, and other deriva-
tives, but they were not successf ul,
QuEBN's COLLIOB,
Galway.
X
Digitized by
Google
292 PERKIN: MAGNETIC ROTATION OF RING COMPOUNDS.
XXXI. — Magnetic Rotation of Ring Compounds. Cam-
phor ^ Lim^nene, Carvene, Pinene, and som^e of their
Derivatives.
By W. H. Pebkin, sen., Ph.D,, F.R.S.
During the course of my experiments on the magnetic rotation of sub-
stances belonging to the aliphatic and aromatic series, very remarkable
differences have been observed in the values obtained for the two
series of compounds, but there is a class of substances which are, in a
sense, intermediate between the two series, namely, saturated closed
chain compounds, such as the derivatives of tri-, tetra-, penta-, and
heza-methylene ; these ha^e also received a certain amount of atten-
tion, but unfortunately the number of such compounds available has
been small. A comparison of the results obtained in the examination
of the latter with those of the aliphatic and aromatic series led me to
include in my investigations some of the members of the camphor and
terpene series and their derivatives, not only on account of the great
interest which always attaches to these important products, but also
on account of their relationship to the saturated closed chain com-
pounds mentioned abova
In previous papers on the magnetic rotation of substances belonging
to the aliphatic series, it has been repeatedly pointed out that un-
saturated members of this series differ from the corresponding
saturated compounds by a number which varies between 0*7 and about
1*112, according to the class of compound. This will be seen from the
following table :
Diff. for
nnflatnration.
Octylene CH3-CH:CH-[CHj4-CHj 9*436 1 . n-TiS
Octane CH,-[CHJg-CH3 8*692 J
Allylacetic acid ... OH^ICH-CHg-OH^-CO^ 6*426 \ . ngiS
Valeric acid CHj-CHj-CHj-CI^-COjH ... 5*513 J
Allyl alcohol CHjICH-CHj-OH 4*6821 . /..gu
Propyl alcohol ... CHg-CHj-CHj-OH 3*768 J
Ethyl crotonate. . . OHj-CHICH-OOgCjHg 7*589 \ . . ., t o
Ethyl butyrate ... CHj-OHj-CHj-COjOaHg 6*477 / +* ^^^
In the case of the hydrocarbons, the value for unsaturation is the
lowest, and the average for all the hydrocarbons of this class which
have been examined is 0*720 (Trans., 1895, 67, 261), whilst the
highest number is obtained in the case of the carboxylic esters, and is,
on the average, 1*112.
That this large plvs difference is due to the formation, from the
Digitized by
Google
PERKIN: MAGNETIC ROTATION OF BINQ COMPOUNDS. 293
scUuraied aliphatie substancei of a new class which we call unBcOuraUdf
and not merely to the removal of two atoms of hydrogen, is obvious,
since the value of two hydrogen atoms has been shown to be +0'508|
and their removal would therefore reduce the rotation.
The next point to be considered is that closed carbon chains have a
lower rotation than the corresponding open chain compounds, from
which they differ by two atoms of hydrogen. This will be clear from
a comparison of the magnetic rotation values of certain closed chain
compounds with those of the corresponding open chain aliphatic com-
pounds containing two atoms of hydrogen more :
Butyricacid C^HgO, ... 4-4721 _q.^^i
Trimethylenecarbozylic acid* O^H^O^ ... 4'141 J
Valeric acid C^K^fi^... 5-513 1 .0.455
Tetramethy lenecarbozylic aoidf C^Kfi^ ... 5 -048 /
Hexoioacid C^R^fi^.,. 65301 _q.q^^
Pentamethylenecarbozylic acidt C^K^qO^. . . 5*891 J
Octoicacid OgHi^Oj... 8-5801 ^.^^^
Methylhezamethylenecarbozylic aGid§ CgH^^Oj... 7*975 J
In this connection, it will be convenient to consider the irregularities
of these differences, which in the above table vary from about
0-33 — ^0*64. This peculiarity has hitherto been an enigma, but care-
ful study of the subject seems to show that this apparent anomalous
behaviour is due to the comparison having been made between wrong
members of the two series.
The study of the magnetic) rotation of the aliphatic compounds has
shown that the first two members of a series do not follow the same
rule as the succeeding ones. In the carbozylic acids, for instance,
formic acid yields a rotation very much higher, relatively, than
any other acid. The value for acetic acid is also, relatively, too
high, but to a less eztent, whilst all the other acids differ by a
constant value, as is seen in the diagram published in an earlier paper
(Trans., 1884, 45, 548).
Now it is evident that the trimethylene ring,
t
m*>CHj,
■^2
must be the first member of the series of ring compounds, the tetra-,
1*1*, the second, and the penta-, CHj^^ *1* , the third, and
so on, and it will be found, if these be compared with the first, second,
and third members of the corresponding saturated aliphatic series, and
♦ Tnma, 1896, 67, 117. t Trans., 1887, 61, 11.
t Tnms., 1894, 66, 99. § Tmu., 1888, 68, 209.
Digitized by
Google
294 [PERKIN : ICAGNSnC BOTATION OF BING GOMPOtTNM.
the usual allowance be made for additions of OH, groups, that these
anomalies in the value of ring formation due to loss of H, practi-
cally vanish, thus :
Mol. Diff. for
rot. ring formatioxL
Formic acid + (CH2X 3) = 1-671 +3069- 4-7401 .o-699
Trimethylenecarboxylic acid 4*141 J
Acetic acid + (CHjX 3) = 2-525 + 3069= 5-5941 _ 0-546
Tetramethylenecarbox jlic acid 5 -048 J
Propionic acid + (CH, x 3) = 3*462 + 3069 = ... 6531 1 _^ ^.^^^
Pen tamethylenecar boxy lie acid 6*891 J
Valeric acid + (CH2X 3) = 5-513 + 3-069= 8582 ) _o*607
Methylhexamethylenecarboxylic acid 7*975 J
From these comparisons^ it is seen that the numbers all approximate
closely to - 0*60.
If we compare the esters of the dibasic acids in the same manner,
analogous, although somewhat lower, differences are obtained, the
first comparison beginning, in this case, with diethyl oxalate, then
with diethyl malonate, and so on :
Hoi Diff. for
rot. ring formation.
Diethyl oxalate + (OH^ x 3) « 6*654 + 3*069 = 9*723 1 _ Q.g^y
Diethyl trimethylenedicarboxy late* 9*166J
Diethyl malonate + (CHj x 3) = 7-410 + 3*069 = 10*479 I _ Q.ggg
Diethyl tetramethylenedicarboxylatet 9 '940 /
Again, the ketones give similar results :
Mol. Diff. for
rot. ring formatioD.
Acetone + (CHjX 3) = 3-516 + 3069= 6-585 1 .o-684
Tetramethylene methyl ketone^ 5*901 J
Methyl ethyl ketone + (OHj x 3) = 4-452 + 3-069 = 7-521 ) ^ ^^^^
Tetramethylene ethyl ketone§ 6*911 )
Why tri- and tetra-methylene compounds should, throughout, be-
have in a manner analogous to the two first members in the aliphatic
seriesi it is difficult at present to understand, but the results of the
comparison are so constant and striking that there can hardly be a
doubt as to the correctness of the view advanced.
Taking the average of the foregoing, we get for the ring formation
with loss of Hjp a difference of 0*6 in acids, esters, and ketones.
In the case of hydrocarbons, this value is evidently much larger, as
is seen from the comparison of hexane and hexamethylene :
^^^^^■■■■- S1J!} -0-982
Hexamethylene 5 -664 J
♦ Trana., 1887, 61, 868. t Trana., 1887, 61, 4.
t Tnna, 1892, 61, 48. | Tiana, 1892, 61, 62.
Digitized by
Google
PnXJOf: MAGKEnO ROTATION OF BINO G0HP0X7KDS. 295
Although bezamethylene is the only substance of this class which
hasy so far, been examined, it will be shown in this paper that this
large value of 0*982 is probably correct in other similar cases.
If, now, we compare bezamethylene with tetrahydrobenzene, we
get
Tetrahydrobenzene ^'^^^14-0*734
Hexamethylene ' 6*658/
This + 0*734 is the difference for unsaturation in these ring com*
pounds, and is practically the same as that found in the case of open
chain hydrocarbons (0*720).
On chlorinating hexamethylene, a remarkable result is, obtained,
the product losing its very low rotation and showing the difference
for ring formation exhibited in the case of acids, esters, and ketones:
Propyl chloride 6*056
CH5X3 3069
8126
Chlorohexamethylene 7*489
Value for ring formation - 0*636
The second displacement by chlorine changes the character of
the rotation again, making the influence of the ring formation
still smaller, thus:
Propylene dichloride, OgH^Ol, ... 6*344
CB[jx3 3069
Calculated rotation of C^B^^jOl^ ... 9*413
Dichlorohexamethylene found ... 8*930
Yalue for ring formation ... — 0*487
This value is practically the same as the value of H^ la the
aliphatic series (0*508) (see Trans., 1884, 46, 672) ; this is remark-
able, as, besides the loss of H,, ring formation has also taken place.
Erom what has been said, it is seen that ring formation influences
the rotation, as follows :
In hydrocarbons , -0*982
„ monochloro-substitution products, ketones,
acids, and esters —0*600
„ disubstitution products of ring hydrocarbons
containing chlorine -0*600
The difference of rotation shown by unsaturated substances and
closed chain compounds is very remarkable, and has repeatedly proved
Digitized by
Google
296 PERKIN ; MAGNETIC ROTATION OF RING COMPOUNDS.
of the greatest value in deciding whether a new substance belonged to
one class or to the other.
Since, then, the formation of saturated ring compounds from
saturated substances containing two atoms of hydrogen more in the
molecule is associated with the above-mentioned reduction in rotation,
it became of interest to consider the case of the formation of so-called
double or 5nc^tf<]{ rings, in which the ring formation from the open
chain compound has taken place twice, and therefore with the loss of
fowr atoms of hydrogen.
The most convenient of such bridged ring compounds for this pur-
pose is ordinary camphor, C^qH^^O, since the results of extended
investigation leave but little doubt that this substance has the con-
stitution
CH2-9H OHjj
CHj-C(CH«)— CO
originally assigned to it by Bredt (^«r., 1893, 26, 3947).
The magnetic rotation of camphor was therefore determined and
found to be 9*265, and ffoln this number and the following consider-
ations we may obtain a fairly accurate value for the effect of bridged
ring formation in this case.
The saturated open chain ketone in the aliphatic series corresponding
with camphor would have the formula C^qH^qO, and its value may be
calculated thus :
Methyl hexyl ketone 8'509
OH5X2 2-046
Gale, rotation of ketone, G^qH^oO 10*555
Found rotation of camphor, G^qH^qO 9*265
Difference -1-290
This difference, due of course to the formation of the double or
bridged ring, is a minus one, and is, in fact, of a nature similar to that
observed in the formation of a single ring (see above) ; it is therefore
most interesting to note that by dividing the difference by 2, 0*645 is
obtained, or a number closely agreeing with the average value for the
influence of single ring formation in the case of ketones, &c., namely,
-0-60 (see p. 294).
A similar result is obtained from an examination of the magnetic
rotation of borneol. The constitution of this substance is represented
by the formula
GHg-CpH CH,
I 9(OHs)2 I
GHj-qGHj) CH-OH
Digitized by
Google
P«RltIN : MAGNETIC ROTATION OF RING COMPOUNDS. 297
and, as it is related to camphor in the same way as a secondary alcohol
is to a ketone, we should expect in the first place to find the rotations
of these compounds difiEering by about 0*525*.
The rotation of bomeol calculated from that of camphor would then
be camphor 9*265 + 0*525 » 9*790, a number closely agreeing with
9*807, which is the value actually observed.
If now we proceed, as we did in the case of camphor, to calculate the
effect qf bridged ring /armation, in passing from the corresponding
saturated compound, O^^^H^O, to bomeol we obtain the following
figures.
The value of an open chain saturated secondary alcohol of the
formula C^JBi^O may be calculated thus :
Mo.Octyl alcohol ..' 9*034
CHjx2 2 046
Value for OioHjjO calc 11080
Mag. rot. of borneol 9*807
Difference....* -1*273
^ This value for the effect of bridged ring formation agrees closely
with that deduced from camphor, namely, —1*290.
It is next of interest to study the case of menthol,
CH,.CH<g|«::^(gg)>CH.CH(0H3)^
because, although it is very similar in constitution to camphor and
borneol, it differs from them in being a single closed chain compound,
the bridged ring being absent.
The rotation of menthol was found to be 10*486, and if we deduct
this from the magnetic rotation of the corresponding saturated open
chain alcohol, C^qH^O, which, as shown above, is about 11*080, we
obtain 1 1 080 -10 -486 » 0*594, for the effect of a single ring form-
ation, but as menthol contains the i^opropyl group, the calculated
number should, perhaps, be about 0*100 higher ; this would make this
difference 0*694. This number is only a little higher than the
average value found in the synthetically prepared ring compounds,
namely, 0*60 (see p. 294).
Turning next to the case of the terpenes, C^qH^^, the magnetic rota-
tions of Mimonene, carvene ((l-limonene), camphene, and pinene, have
been determined and the results are given in the paper.
If we first consider the formula of camphene, the actual value of
* Thie is the difference between the yalne of the rotation of aecwiiyl alcohol,
CH,*[CHJb'CH(OH)'CH„ 9*034, and that of methyl hexyl ketone,
CH,-[CHJb*CO*CH„ 8*609.
Digitized by
Google
S298 PEBKIN : MAGNETIC fiOTATION OF BIKG OOHFOtJKBB^
bridged ring formation in a hydrocarbon of this class may be deduoed
in the following way.
It may be assumed that camphene is produced from an open chain
compound^ ^10^22^ ^7 ^^^ converting it into a six carbon unsaturated
ring» C|QH|g, and then subsequently forming the bridged ringi as
indicated by the following formulsB :
(CH3)CH(CH,) (CH3)CH(CB3)
CHj-CH-CHj CHg-CH CH
CH2-CH(CH3)-CH3 CHj-CH(CH3)-CH
OHj-CpH CH
I 9(CH3), II •
CH2-C(CH3) OH
The calculation of the value for the formula O^oH^^ is a simple one,
and may be carried out as follows :
Heptane 7*669
CHgx3 ;.: 3069
Calculated value of C^oH^ 10-738
The use of this calculation in the present case requires some explan-
ation. It has been shown (p. 297) that when CH^ is introduced
into an aliphatic hydrocarbon to form the next homologue, the normal
chain has a rotation which is lower, by about -0*1, than that
of the isomeric compound containing a side chain. Thus, starting
with n-pentane, CH3*CH2-CHj-CH2*CH3 (mag. rot. 5 -638), »-hexane,
CHa'CHj-CHj-CHj-CHj-OHg, has a roUtion of 6670, whereas the
rotation of Mohexane, (CH3),CH'CHj*CH2*CH3, is 6-769, a difiEerence
of 0 099. This behaviour is shown in all similar cases in the aliphatic
series.
On the other hand, in ring compounds the introduction of CHjp
whether it enters the ring or forms a side chain consistiog of a methyl
group, produces the same, or nearly the same, effect on the rotation,
thus:
(a) Pentamethylenecarboxylic acid .. Cn^<p^^'^ 6-891
CHj'CH'CO^H
(b) Methylpentamethylenecarboxylic ^xr mr nxr
acid CH,<^^*Y^^^ 6-914
^^CHj-CH*CO^
(c) Methylhexamethylenecarboxylic >^„ .nir •nr.niT
acid CHj-OHj-CpH-CH, ^.^^3
CH,-CH,»gH-COjH
/Google
Digitized by ^
Pfe&KlN : MAGNETIC ROfAtlOK OF RIKQ COMPOUNDS. 299
In the case of the fonnation of b, it may be assumed that the CH^
has entered into a to form a tids chain consisting of a methyl group^
and the result is an increase in the mag. rot. of 1*023. In the form-
ation of e from &, CH^ enters into the ring \ the increase in
magnetic rotation is 1*061, or about the same value -in both cases.
In calculating the value for a ring formula such as that of camphene,
it may therefore be taken that the differences for OH^ are about r023|
whether the CHg enters the ring or gives rise to the formation of a
methyl group.
The actual value of the formula 0^^^^ given above would evidently
be higher than 10*738 on account of the methyl side chains which it
oontainSy but as all of theae ultimately enter into the ring of cam-
phene, this extra influence peculiar to methyl groups in the open
chain compounds must be neglected.
Referring again to the calculation of the magnetic rotation of cam-
phene from that of the hydrocarbon O^^H^, the first process indicated
by the two f ormuUa,
CHj-CH-CHj CHj-CH-OHg
CH,-CH*CH, • —^ CHj-CH CH
0H,-CH(CH3)-CH, CHj-CH(CH3)-c!h '
can be easily followed, since the analogous relationships between
hezene and tetrahydrobenzene,
CHj-CHj-CHa CHa-OHg-CH '
have been investigated.
The magnetic rotation of hexane is 6'646, and that of tetrahydro-
benzene is 6*392 ; it follows, therefore, that the value for the form-
ation of the unsaturated six carbon ring is - 0*254. Applying this
to the case of camphene, it was shown on page 298 that the calcu-
lated value of the saturated hydrocarbon, C^QH^^yis 10*738; if, there-
fore, we deduct - 0*254 from this, we obtain 10*484 as the value of
the unsaturated six carbon ring, O^q^is* ^^^<^9 then, the observed
magnetic rotation of camphene is 10*135, it follows that the in-
fluenbia of bridging the ring in this hydrocarbon is represented by the
value 10*484- 101 35 = 0-349. The value of the bridged formation
in this compound must then be -0*982 + - 0*349 » -1*331, which
is not very different from that found in the case of camphor, - 1*290.
Having obtained this value, 0*349, for bridging the ring, it is possible
to estimate the probable rotation of terpenes such as limaneney
CHj-CrCHj
CpHj-OH 9H, ,
OHj*0(OH,):OH
Digitized by
Google
300 PERKIN : MAGNETIC ROTATION OF RING OOMPOUNDB.
and it will be at once seen that both limonene and camphene contain
their carbon atoms in the same order, and may be said to be de-
rived from almost identical saturated hydrocarbons of the formula
CjqHjj. Both camphene and limonene contain a six carbon unsat-
urated ring, and the essential difEerenoe between the two is that in
the former the group OH^'CH'GH, has lost two atoms of hydro-
gen on entering into this ring in a form of a bridge, whereas in
the latter this loss of two atoms of hydrogen has resulted in the
conversion of the group into the unsaturated group GH3*CHICH,,
which remains outside the ring.
In order to deduce the value of limonene from that of camphene, we
must deduct the value for bridging the ring (- 0*349), and add the
value for the conversion of the saturated group into the corresponding
unsaturated one ( + 0*720, see p. 296), thus :
Rotation of camphene 10*136
Deduct value of ring formation, ( - 0*349), ».«., add + 0*349
Add for nnsaturation +0*720
Calculated mag. rotation of limonene 11 *205
It is certainly interesting that the number thus ci^lculated coincides
nearly exactly with that actually found, as the average value of
/-limonene and c2-limonene (carvene) is 11*204.
The rotation of the limonenes may be calculated in another manner,
and with almost identical results, thus :
Calculated value of the hydrocarbon C^qH^j 10*738
Less single ring formation of hydrocarbon (see p. 295)... 0*982
9-766
Amount for nnsaturation outside ring 4- 0*7 20
„ „ inside ring +0*720
11*196
This is again almost identical with the number, 11*204, actually
obtained.
These considerations make it probable that the determination of the
magnetic rotation of a terpene of unknown constitution may prove of
considerable value in deciding its formula, as this will show clearly
whether it belongs to the camphene class (compare Marsh, Proc., 1899,
64, and Semmler, J?er., 1900, 33, 3420) and contains a bridged, un-
saturated ring, or whether, like limonene, it is doubly unsaturated and
the bridged ring is absent ; in the former case, the value would ap-
proximate to 10*133, in the latter to 11*204.
PfTMfM, although isomeric with camphene, is not so closely related to
Digitized by
Google
PXBKIH: MAGNETIC BOTATION OF RING COMPOUNDS. 301
Dphor, ainoe the camphor bridged ring is still intact in camphene
whereas in the ease of pinene, the bridged ring is of a somewhat dif-
ferent kind. The general similarity between camphene and pinene is,
however, indicated by the comparatively slight difference in the mag-
netic rotation, namely, +0*159.
CJdonh and BromthiubttUtUion DerivcUives qf Camphor.
The following have been examined :
a-Chlorocamphor OJB.i^<^\ ^
a-Bromocamphor CgHj^^ l^ ^
I
CO
aa-Dibromocamphor C^^^^^\ Br
Q^x-Br
a^-Dibromocamphor CgHijBr^ I ^
o^-Dibromo-a-chlorocamphor . . . CgHijBr^ i ^r
The first inflaence of chlorine and bromine when it displaces hydrogen
in camphor is as follows :
a-Chlorocamphor... 10*846 a-Bromocamphor... 12*761
Camphor 9*265 Camphor 9*265
CI displacing H ... 1581 Br displacing H ... 3*496
The values for the halogens displacing hydrogen in open chain com-
pounds fluctuate a good deal (see Trans., 1894, 66, 417, 418), but the
above numbers are abnormally high. The highest observed values in
open chain compounds for monochloro- and monobromo-derivatives are
1*470 and 3*424 respectively. When a further displacement of hydro-
gen by bromine occurs in camphor, the values are nearly normal
for the second substitution, as the following comparisons will
show:
aa-Dibromocamphor 15*994 a)9-Dibromocamphor 15*902
a-Bromocamphor ... 12*761 a-Bromocamphor ... 12*761
Br displacing H ... 3*233 Br displacing H ... 3*141
As will be seen, the a/3-compound gives the smaller number of the
two.
Digitized by VjOOQIC
302 PBRKIN : MAGNETIC ROTATION OF RING OOMPOUNDB.
Lastly, coming to a third displaoement, the result is again about
normal:
a)9-Dibromoa-chlorocamphor 17 *346
a)9-Dibromocamphor 15*902
CI displacing H 1-444
The high values found for the monochloro- and monobromo-compounds
were what might be expected, as it has been shown that the same thing
occurs in hexamethylenci only to a greater extent, the value for the
chlorine atom, which displaces hydrogen in the latter case, being no
less than 1*843. The second displacement by the halogen is, however,
about normal, just as when the second atom of hydrogen in camphor is
displaced by bromine.
a-Nitroeamphar, ]^ndoIiiirocamphar, Camphoryloasime, and Anhydro-
pseudont/rocam^Aor.
It will be convenient to consider the rotation of these substances to-
gether.
Nitrocamphor, compared with camphor, shows the following differ-
ence in rotation :
a-Nitrocamphor 9468
Camphor 9-265
NOjjdisp. H 0*203
This difference is similar to, although a little lower than, that noticed
between nitropropane and propane, which is 0*229 (Trans., 1889,65,729).
After obtaining this result, experiments were made with the view
of comparing the rotations of nitrocamphor and /w^utfonitrocamphor.
Dr. Lowry considers that these two substances are represented by the
formulffi :
Normal. Pseado.
As the latter cannot be isolated in a pure condition, it was necessary
to examine one of its stable salts, and, after trials with several of these,
the triethylamine compound was selected as most suitable. The
rotation of the />Mu<fonitro-compound, after allowing for the triethyl-
amine, was found to be 10469 (see p. 313), a number which is consider-
ably higher than that of nitrocamphor, thus :
^-Nitrocamphor 10*469
a-Nitrocamphor 9-468
1001
Digitized by VjOOQIC
PEBKIN: MAGNETIC ROTATION OF RIKO OOMPOUNDS. SOS
As it has been found that tlie magneido rotations of amides and
oximes are identical (unpublished results), it was of interest to compare
those of the psotMbnitro-compoond and camphorylozime, which are
related as follows :
.C=N-OH ^^^N-OH
C3H,/ >0 CH<T^6
NX) XX)
Camphoiyloxime. ^-Kitrocamphor.
and it was found that their values were very similar, that of the
ozime being 10-376.
As the relationship between the rotation of the acids and anhydrides
is known, it is possible to get a second value for the rotation of the
jM0iM2onitro-compound by determining that of the anhydride which is
stable^ and has the formula :
CsH„<^rS^03H„
Its rotation was found to be 19*712.
The difference between the rotation of the auhydride of a monobasic
acid aod that of 2 mols. of the acid it is derived from is 9 '752 ; if this
then be added to the rotation of the anhydride, the result will be the
rotation of 2 mols. of pMtM^onitrocamphor, and if this be divided by
2 it will give the rotation of this compound itself :
Anhydride of ^-nitrocamphor 19*712
Difference between anhydride and 2 mols. of acid ... 0*752
Rotation of 2 mols. ^-nitrocamphor 2/20*464
^-Nitrooamphor 10*232
This is only a little lower than that obtained from the triethylamine
salt. Considering the probable experimental errors connected with
this kind of comparison, the average of all will probably not be far
from the truth ; it amounts to 10*346, a value which is nearly identical
with that of its isomeride, campboryloxime.
It would appear, therefore, that the difference between the rotation
of the normal and the pMiKionitro-compounds amounts to about 1*00,
because the numbers for the nitro-compound are undoubtedly a little
high on account of the solution examined containing a small quantity
of the pseudo-form, as pointed out by Lowry. It is also interesting to
find that the rotations of the pseudo-compound and the isomeric oxime
are identical, and it affords another instance of the use of the magnetic
rotations in distinguishing between two forms of nitro-oompounds such
Digitized by
Google
804 perkin: magnetic rotation of rinq compounds.
The change of optical rotation which takes place when nitrocamphor
is converted into the pseudo-form is very remarkable. In ethylene
chloride solution, nitrocamphor has a rotation of [a]i> — 19*47^ whereas
the rotation of a-nitrocamphor, calculated from the aqueous solution
of the triethylamine salt, is no less than [a]D + 345*5^. Oazeneuve
found for the sodium salt [ajo +298'' {Bull. Soo. Chim., 1888, [ii],
49, 92).
There is one interesting point connected with the magnetic rotation
of pMt^onitrocamphor^ and that is, that it supports the formula
^io^u\ I ' ^^^ *^*^ substance and not the alternative one,
^10^14*^X0 * ^^^ essential difference between these formul»
is that the first represents the nitrogen as tervalent, and the second
as quinquevalent, and it has been shown that these two conditions of
nitrogen influence the magnetic rotation of substances containing
nitrogen very differently, the former by about 0*611 and the latter by
only 0*103, the variation being practically 0500 (Trans., 1889, 66,
737).
Now the rotation of /weiM^nitrocamphor, as seen above, being
identical or nearly so with that of its isomeride camphorylozime,
which contains tervalent nitrogen, the inference therefore is that this
/iMtM^nitro-compound must also contain it in that condition ; if it were
quinquevalent, the rotation should be about 0*50 lower than that of
•camphorylozime, which is a very large difference.
aa-Chloro- and aa-Bramo-niirooamphor,
O.H.4^ O.H„<^B, .
It has already been pointed out that in nitrocamphor the influence
of the substitution of the nitro-group for hydrogen is + 0*203, being
slightly less than the value found in the case of the conversion of
propane into nitropropane (0*229). In chloronitrocamphor, the NO^
group is slightly negative, and in bromonitrocamphor, negative to a
still greater extent. It has, in fact, been found that as the specific
rotation of substances increases, the influence of this group, which is
always small, diminishes ; the positive nature of the value in substances
of small specific rotation becoming negative in substances of large
specific rotation, thus, for example, in the case of chloroform, which
has a rather large specific rotation, nitropicrin, derived from it by thfi
Digitized by
Google
I>ERKIK: MAGNETIC ROTATION OF RING COMPOUNDS. 806
substitution of NO, for hydrogen, has a smaller rotation than chloro-
form itself, thus :
Chloroform 6*559
Nitropicrin 5*384
Difference -0-176
and this has been observed in several other instances, the particulars
of which have not yet been published. Ghloronitrocamphor having a
larger specific rotation than camphor, the introduction of the NO,
group yields a nitro-compound in which the influence of this group is
slightly negative, and again, bromonitrocamphor has a larger specific
rotation than chlorocamphor, and as a result, the nitro-compound gives
a value still more negative. The following table shows the rotation of
these nitro-compounds compared with that of the substances from
which they are derived :
C»>»Pl«»'' »-265) ^.0.203
a-Nitrocamphor 9*468 j
orChlorocamphor 10*846 ) .. 0*024
aa'-Ohloronitrocamphor ... 1 0*822 J
a-Bromocamphor 12*761 ) «o*041
aa'-Bromonitrocamphor ... 12*720 J
These results appear to be similar in character to those obtained
with mixtures of substances with greatly varying specific rotations, to
which I have previously drawn attention (Trans., 1896, 60, 1052).
CcmfhyhmwM^ OH
r^TT ^C CH OH,
2
2 NH, '
OHj*^
(Tiemann, jB«r., 1897, 80, 245).
The magnetic rotation of this base was found to be 11*770, and it
is evidently an unsaturated ring compound. The rotation of the
saturated aliphatic base corresponding with camphylamine is OiqHj,N;
its rotation may be calculated thus :
Propylamine, 0,HgN 4 563
OH2x7-1023x7« 7*161
Eotation of 0^^^ 11*724
The influence of unsaturation may be calculated from the difference
of rotation between allylamine and propylamine, namely, 5*687 - 4-563 •
1-024. If this be added to the above, the rotation of this basOi sub-
VOL. LZXXL T
Digitized by
Google
306 PERKIK : MAQNiSTIC ROtAllON OF RING COMMtTNl)S.
tracted from the rotation of camphylamine, will leave a value repre*
sentiog the influence of the ring formation, thoB :
O^oH^jN 11-724
For unsaturation 1*024
12-748
Rol tion of camphylamine 11*770
Ring t^. rmation 0*978
It is interesting to find that this number for ring formation ia
practically the same as that observed in the case of hydrocarbons
(0*982).
Bomyl Chloride {Pinene Hydrochloride), O^^B^^^GL
The magnetic rotation of this compound was found to be 11*058,
and if from this we deduct the rotation of pinene, 10*294, we obtain a
difference -■ 0*764. This difference, however, does not represent that
of hydrogen chloride merely, because pinene^ when converted into its
hydrochloride, becomes a saturated compound ; it is therefore neces-
ssary to add to this number (0*764) the amount which is lost by this
change, namely, 0*720, the value for unsaturation (see p. 295) ; this
gives 1*484. Since the value of fiOl is 1*987, the discrepancy of
- 0*503 points to the fact that pinene undergoes some further change
in structure when it is converted into the hydrochloride, and not
merely that resulting from its becoming a saturated compound in
union with hydrogen chloride. Wagner and Brickner (J^er., 1900, 32^
2325) have recently shown that pinene hydrochloride is not, as formerly
supposed, a derivative of pinene, but is bornyl chloride, moleculiur
change having taken place during the addition of hydrogen chloride,
thus:
CH-C(CH3)=0H CHj-0(CH3)=OH
|^C(0H3), I -^ I C(0H3),Cl
Pinene. Terpinyl chloride.
CH,-9(OH3) — CHCl •
I 9(OH3), I
CHg-CH OHj
Bornyl chloride.
That this view is in accordance with the results of the magnetic
rotation determinations may be shown as follows.
The rotation of M0.octyl alcohol is 9*004 and that of «dc.octyl
chloride is 10*248, showing that^ in passing from a secondary
Digitized by
Google
PEBKIN : MAGNETIC ROTATION OF RING COMPOUNDS. 307
alcohol to the corresponding chloride, there is a rise of rotation of
1*224. If, then, pinene hydrochloride is bomyl chloride, its rotation
should easily be calculated by adding to the rotation of the corre-
sponding secondary alcohol, bomeol (9*806), the number 1*244. This
gives 11*050 as the calculated rotation of bomyl chloride, which agrees
almost exactly with that found, namely, 11*058.
tTajaB-DipeTUene Dihydrochlaride^ C,oHj8»2H01.
The magnetic rotation of this compound, which, according to Baeyer
(fiar., 1893, 26, 2862), has the formula :
H,C OH,'
h,44h.
has been found to be 13*111. This formula represents the substance
G|0HjgG]2 as a dichloromethylMopropylhezamethylene, and its rota-
tion may be calculated thus :
Dichlorohexamethylene *. 8-906
OHgX 4 = 1*023x4= 4092
Addition for iso-group 0*103
Calculated rotation of Cj^jHigClj 13*101
It will be seen that the number found agrees closely with that
calculated in this way.
Refraction values of CamipkoT cmd its Compounds.
The results of these determinations, the particulars of which will
be found in the experimental part of this paper, are in most cases
lower than the calculated, as will be seen from the following table,
which gires the differences between the theoretical and observed
values for ^7 P-
d
Bomeol -0*409 o^-Dibromo-a-chloro*
Camphor +0*154 camphor '>0'560
a-Chlorocamphor - 0*150 a-Nitrocamphor —0*045
a-Bromocamphor -0*207 aa'-Chloronitrocamphor -0*522
aa-Dibromocamphor.... -0*436 aa'-Bromonitrocamphor -0*768
o^-Dibromooamphor... -0*808 Camphoryloxime +0*812
T 2
Digitized byVjOOQlC
308 PERKIN : MAGNETIC ROTATION OP RING COMPOUNDS.
The comparison of these results with the magnetic rotation deter-
minations is a matter of difficulty, but so far as it has been carried
out the variation in the two values seem to be of the same nature
except in the case of camphor.
There is a certain analogy between the refractive values of a-nitro-
camphor, aa'-chlorouitrocamphor, and aa'-brcmonitrocamphor, inasmuch
as that, whilst the nitro-compound has a refractive value nearly the
same as the calculated, the chloro-derivative has a considerably
lower, and the bromo-derivative a still lower value, these differences
being in the same order as the magnetic rotation. In the case of
camphorylozime, it is also found that its refractive value is higher
than that of its isomeride, a-nitrocamphor, a result similar to that
found in the case of their magnetic rotations.
EZPBBIMENTAL.
It will be seen from this part of the paper that as all the camphor
compounds are solids and not easily fusible, they had to be examined
in solution. A very good solvent for those examined is chloroform,
but owing to its volatility it is difficult to keep the solutions of constant
composition for any length of time. Ethylene chloride was therefore
tried and found to be an excellent solvent ; and as its boiling point is
83'5°, it was much more suitable than chloroform; moreover, its specific
rotation was found not to vary very largely from those of the camphor
compounds, experiment showing that it is a little higher than those of
camphor, its nitro- and chloro-derivatives, and a little lower than
those of the bromo-derivatives. The error, therefore, arising from this
cause would be small, it having been found that in the magnetic
rotation of mixtures a reduction occurs more or less proportional
to the difference between the specific rotations of the constituents.
A litre of ethylene chloride, purified by shaking with concentrated
sulphuric acid several times, was washed, and after being dried
first with potassium carbonate, and then with phosphoric oxide, was
filtered and distilled. It all came over within less than half a degree,
and when fractionated into three quantities, the densities of the first
and third only differed by 000018. The density at 15^ was 1*26197,
this is higher than that previously observed, both by Thorpe and my-
self, by about 00021 (Trans., 1884, 45, 528), and is probably due to
the greater purity obtained by treatment with sulphuric acid.
The magnetic rotation was determined on three occasions and gave :
t. Sp. rot. Mol. rot
18-7° 1-2564 5-496
With specimens previously examined, the mol. rotation obtained
was 5 '485, which is practically the same as that now found.
Digitized by
Google
PEREIX : MAQNETIG ROTATION OF RING COMPOUNDS. 309
Sohitions made with this solvent were preserved in bottles with
stoppers moistened with glycerine, and were found to undergo practic-
ally no change, even after being kept for months, as was proved by
redetermining their densities.
The solutions were made as concentrated as possible, but on account
of the varying solubilities of the camphor compounds, they could not
be made of similar strengths.
Menthol, CioHg^O.
This substance was examined in the fused state.
Density : d 40740°, 08909 ; d 45745°, 0-8888 ; d 60760°, 08868 ;
d 66766°, 0-8860; d 60°/60° 0-8835 3 d 65°/65°, 08820.
Magnetic rotation determined on three different occasions :
t. Sp. rot Mol. rot.
Average 46-2° 1-0764 10-486
Optical rotation at 47° [a]i,=- -49-88.°
Bomeol, C^^U^fi.
This was redistilled before use, b. p. 214° under 772 mm. pressure.
The strongest solution that could be used contained 23-728 per cent.,
the composition being C^qHisO + 6 mols. CgH^Cl^
Density : d 10710°, M819 ; d 16°/16°, M760.
Magnetic rotation determined on three different occasions :
t. Sp. rot. Mol. rot.
Average 15-8° 1-2122 37-232
Less 6 mols. CjH^Ola 27-425
Mol. rot. CioHjgO »... 9-807
Optical rotation [a]i> = + 35-22°
CamphaTf Ci^HjgO.
This substance was redistilled for examination. Two solutions in
ethylene chloride were used.
The first contained 60-682 per cent., representing in composition,
CioHjgO +1-60 mols. C^H^Clj.
Density : d 10710°, 10986 ; d 16716°, 1-0939.
Digitized by
Google
310 PEBKIN: MAGNETIC ROTATION OF RINQ COMPOUNDS.
Magnetic rotation twice determined :
t. Sp. rot. Molriot
Average ll-72« 1U83 17478
Less 1-5 mols. OjH^Olj 8-227
OioHieO 9-261
Optical rotation [a]D » + 51 '936^.
The second contained 67*18 percent., representing Oj0H^qO + O'76
mol. CgH^Clg. This was a practically saturated solution at the ordin-
ary temperature.
Density: d 10°/10^ 1-0523; d 16°/15°, 1^0489.
Magnetic rotation ; four times determined :
t. sp. rot. MoL rot.
Average 15'74° M171 13-393
Less 0-75 mol. OjH^Olj 4-113
OioHijO 9-280
Optical rotation [aj^- + 62-68°.
The magnetic rotations of both solutions being very olosCi the
average, 9-265, has been taken as correct.
aCMoroeamphor, OgHj^^Y
The solution employed contained 32*017 per cent, of a-chlorocamphor,
the composition being OjoH^gOOl + 4 mols. O^H^Ol,.
Density: d 10710°, 1-2154; d 15°/15°, 1-2102.
Magnetic rotation determined on three separate occasions :
t. • Sp. rot. MoL rot
Average 12-4° 1-2288 32-786
Less 4 mols. G^^Cl, 21-940
Mol. rot. OjoHigOCl « 10-846
Optical rotation [a]©- +106-6°. Lowry found (in alcohol) +96°.
\,'Bramoeamphor, OgHj^OT^ '.
The solution used contained 43*76 per cent, of this BubstancOi the
composition being CiJELifiBr + 3 mols. OJELfil^
Density: d 10°/10°, 1-3047; d 16716°, 1-2993.
Digitized by
Google
PKBKIH : ILLGNRIC BOTATIOK OF BING OOJCPOimBS. 311
Magnetic rotation determined on three ocoafiions :
t. Sp. rot MoL rot.
Average 122^ 1-2972 29*216
Less 3 mok. C^H^Ol, 16*455
Mol. rot. CioHijOBr 12761
Optical rotation [ a]i> » + 145*34''. Lowry found (in alcohol) + 135^
aa-lHbramocamphor, CgHj^^lL *.
The solution of this compound was not so concentrated as that of
the mono-derivative ; it contained 34*292 per cent., the composition
being Ci^Hi^OBr, + 6 mols. C^^Cl,.
Density : d 10710^ 1*3825 ; d 15^/15^ 1*3764.
Magnetic rotation determined on four separate occasions :
t. Sp. rot MoL rot.
Average 16*52° 1-3397 48*904
Less 6 mols. OjH^Clj 32-910
Mol. rot. OioHiPBr, =16-994
Optical rotation [a]i>- +42*067°. Lowry found (in chloroform)
+ 40°.
aP'Dibromoeamphor, CHBrvI^
HBr
This substance did not dissolve very freely ; the solution used con-
tained 25*812 per cent. Composition OioHi^OBr^ + 9 mols. OjH^Clg.
Density: d 10°/10°, 1*3388; d 15°/15°, 1*3528.
Magnetic rotation determined on five occasions :
t. Sp. rot Mol. rot.
Average 14-9° 1*3234 65*266
Less 9 mols. OgH^Cl, 49*364
Mol. rot. CioHi^OBr, -15*902
Optical rotation [o]d= +104*167°.
afi-Dtbromo^-chlarocamphor, OgHijBr^jL
The solution of this substance employed contained 36*668 per cent*,
the composition being OioHi8^^^^2 + ^ °^^^^* ^fifi^
Density : d 10°/10°, 1*4181 ; d 15°/16°, 1*4122,
Digitized by VjOOQIC
312 PEBKIN : MAGNETIC ROTATION OF RING OOMPOUNDa
Magnetic rotation determined on two oooasions :
t. Sp. rot. Mol. rot
Average IS-P 1-3633 50-213
Mol. rot. 6 mols. CjH^Cl, 32*910
MoL rot. CioHnOClBrj - 17-303
Optical rotation [aj^" +43'016°. Lowry found (in chloroform)
+ 44-6°
CcMiphylamine, CioHigN.
This Bubatance boiled at 205-5 — 206*5° corr. (b. p. given in Beilstein,
194—196°).
Density : d 15715°, 0-8729 ; d 20°/20°, 0-8697 ; d 25°/25°, 0-8669.
Magnetic rotation determined on three different occasions :
t. Sp. rot. Mol. rot
Average 16° 1-2078 11-770
This substance had a small + optical rotation.
.CH'NO,
a-NUrooamphory OgH^^v/v)
For this and the other camphor derivatives, I am indebted to Dr.
Lowry. This substance was examined in two solutions. In the second
solution, it was examined as soon as possible after being made.
The first solution contained 49*873 per cent, of nitrocamphor. Com-
position, OjQHigO'NOg + 2 mols. C^H^Olj. It was several days old
when examined.
Density: d 10°/10°, 1-209823 d 16°/15°, 1*20491.
Magnetic rotation determined on three different occasions :
t. Sp. rot. Mol. rot.
Average 12-14° 1-1246 20-438
Less 2 mols. CjH^Clg 10*970
Mol. rot. CioHijO-NOj 9*468
Optical rotation [a]© = 1947°.
The second solution contained 44-319 per cent, of nitrocamphor.
Composition, CioHi50'N02 + 2-5 mols. CjH^Cl,.
Density: d 10°/10° 1-2166; d 15°/15°, 1*2117.
Magnetic rotation once determined with fresh solution :
Digitized by VjOOQIC
PERKIN : MAGNETIC BOTATION OF RING COHPOUNDS. 313
t. Sp. rot Mol. rot
16-4« 1-1336 23-109
LeFB 2-5 mols. C^H^Clj 13712
Mol. rot. CioH„0-NOj ..• 9*397
a-NiiroocMiphoTy Trieth^lamine, and WcUer.
The compound of triethylamine and nitrocamphor was tused with
the view of getting the rotation of Twetuionitrocamphor. An excess of
triethylamme was employed to ensure all the nitrocamphor being in
combination. Experiments with piperidine as the base were also
made, but the product crystallised out too readily when the solutions
were strong.
The solution contained 36*532 per cent, of nitrocamphor. Gom-
position, CioHi50-NO, + l'25 mols. (C2H5)gN + 12 mols. H,0.
Density : d 15715°, 1-0468 ; d 20720°, 1-0444.
Magnetic rotation, determined on three occasions a month apart :
t Sp. rot Mol. rot
Average 15-5° 11569 33ll6 ;
Less 12 mols.HjO 12000
2M16
Less 1-25 mols. (0,H5),N 10-647
CioHijO-NOj 10-469
The optical rotation of this solution, calculated on the nitro-
camphor, was no less than [a]o +345-5°. Cazeneuve gives for the
sodium salt [a]]) + 298° ; this of course would be higher if calculated on
the nitrocamphor only, although not quite so high as the above
( + 331-2°).
Anhydro-^-nUroeamphorf CjoHjgOjNj.
This substance is not very soluble in ethylene chloride, the
strongest solution that could be conveniently used containing only
15*959 per cent and having the composition OsoH2g05N2 + 20 mols.
CjH^Olj.
Density: d 10710°, 1-2547; d 15715°, 1-2488.
Magnetic rotation, four times determined on different occasions :
t. Sp. rot MoL rot.
Average 14-9° 12348 129-412
Less 20 mols. CJELfil^ 109*700
Mol. rot. OjoH^OjN, -19*712
Digitized by VjOOQIC
31 PEBKIN : MAGNETIC ROTATION OF BING COMPOUNBB.
Optical j^otation [a]D» +132*84°. Lowry gives for eolation in
benzene + 187^ and in chloroform + 167°.
/C=N-OH
Camphoryloxime^ ^8^i4\ ^^
XJO
A 8oIuj}ion of this substance in ethylene chloride containing 44*319
per cent, was used. Composition, OiqHj503N + 2*5 mols. OjH^Clj.
Density: d 15716°, 1*2418 ; d 20°/20°, 1*2366.
Magnetic rotation, determined on three separate occasions :
t. Sp. rot. MoL rot.
Average 17*4° 1*2088 24*087
Less 2*6 mols. OjH^Clj 13*712
10*375
Optical rotation [ajo -■ 14*09°. Lowry obtained, in a 6 per cent,
solution in benzene, [a]]) «7*0°.
yO<^
^CO
NO.
"11 .
The solution examined contained 43*803 per cent, of this substance.
Composition, Ci^H,4001-NOj + 3 mols. OjH^Clj.
Density: d 10°/10°, 1*2641 j d 15°/15°, 1*2689.
Magnetic rotation, determined on two occasions :
t Sp. rot MoL rot
Average 12*86° M717 27*277
Less 3 mols. C^H^Clg 16 '465
Mol. rot. CjoHi^OCl-NOj 10*822
Optical rotation [ajo =6*924°. Lowry found (in chloroform) 5°.
.C<NO,
aa-Bromantirocamphar, CgHj^^ , ^Br *
^CO
The solution employed contained 31*724 per cent., the composition
being CioH^^OBr-NOj + 6 mols. CjH^Clj.
Density: d 10°/10°, 1*3288; d 16°/15°, 1*3231,
Magnetic rotation, determined on two occasioivs ;
Digitized by
Google
PEEKIN : MAQNETIO ROTATION OF RIKQ OOHPOUNDS. 816
t. Sp. rot. 'Mol. rot.
Average 127*^ 1-2516 45630
Less 6 mols. 0,H^01, 32*910
Mol. rot. OjoHi^OBr-NO, 12720
Optical rotation [a]]) B
I'Limonene, C^o^ie*
For this hydrocarbon, as well as for the terpene derivatives referred
to in this paper, I am indebted to Professor Tilden. It was purified
in the same manner as mentioned below in reference to carvene. Its
boiling point was 175*5 — 177^ (corr.).
Density: d 10°/10° 0-8549 ; d 15°/15°, 0-8514 ; d 20^20°, 0-8483 ;
d 25725**, 0-8453.
Magnetic rotation, determined on six different occasions :
t Sp. rot. MoL rot
15° 1-2578 11-162
Optical rotation [a]D« - 103-51°.
Carvene or d-JAtnonene, C^^H^q.
Obtained from ELahlbaum. It was purified by steam distillation
with alkali, then dried, and fractionated. The boiling point was
178-179° (corr.).
Density : d 4°/4°, 0-8576 ; d 10°/10°, 0-8532 ; d 15°/15°, 0-8498 ;
d 25°/25°, 0-8437.
Magnetic rotation, determined on four occasions :
t. Sp. rot. Mol. rot.
16-32° 1-2637 11-246
Optical rotation [a]|,- +122-7°.
Average mol. rot. of Mimonene and carvene, 11*204.
Pinene, O^JBL^^.
The specimen used had a boiling point of 158*5 — 159°.
Density: d 4°/4°, 0-8740; d 5°/5°, 0-8732; d 10°/10°, 0*8694;
d 15°/15°, 0-8658 ; d 20°/20°, 0-8624 ; d 25°/25°, 0-8595.
Magnetic rotation :
t. Sp. rot Mol. rot
20° 11750 10-294
Ojptioal rotation [ a Jo ai 15-47°
Digitized by
Google
316 perkin: magnetic rotation of ring compounds.
Bwnyl Chloride {PinefM ffydroeUaride), Oi^HiyCI.
^ For this and the following derivatives of this class, I am indebted to
Professor Tilden. This product was made from American turpentine.
Being a solid, it was examined in solution in carbon tetrachloride ; two
different solutions were employed.
Solution 1.— This contained 68*271 per cent, of CioH^eyHGl, and had
the composition OioHu,,HCl + 0-623 mol. CCl^.
Density : d 15715° M493 ; d 20720°, 1-1454 ; d 26726°, 1-1415.
Magnetic rotation :
t. Sp. rot Mol. rot.
17-4° M836 14-503
Less 0-523 mol. CCl^ 3-442
11-061
Optical rotation [ajn =- + 6-92°.
Solution 2.^This contained 68-40 per cent, of CioHie,HCl, and had
the composition CioHi^HCl + 0-517 mol. 001^.
Density: d 15716°, 11471 ; d 20°/20°, 1-1430 ; d 25°/25°, M393.
Magnetic rotation :
t Sp. rot. Mol. rot
14-8 1-1868 14-449
Less 0-517 mol. OCl^ 3-396
11-056
Average of both determinations 11 '068
JHpentene Dthydroohhride, O|oHi0,2HCl.
This was examined in the fused condition. F. p. 50°.
Density : d 46°/45° 1-0613 ; d 50750°, 10593 .; d 55°/65°, 1-0575.
Magnetic rotation :
t. Sp. rot Mol. rot.
54-8° 1-1942 13-111
Camphen$f CioH^q.
This was obtained from Schuchardt, and purified by pressure and
five fractionations ; it boiled at 157 — 157*6° under 750 mm. It was
examined in the fused condition.
Density : d 40°/40°, 0-8609 ; d 46°/45°, 0'8686 ; d 60°/60°, 0-8665 ;
d 66°/66°, 0-8644; d 60°/60° 0-8524.
Digitized by
Google
PERKCK: HA.QNBTIC ROTATION OF RINQ COMPOaKDS. 317
Magnetic rotation :
t. Sp. rot. Mol. rot.
45-8° M516 10-136
Refractive Values,
The refractive values of borneol and camphor and its derivatives were
determined. The measurements were made with the same ethylene
chloride solutions as were used for the magnetic rotations. It is
thought to be unnecessary to give all the details of the indices, &o,,
determined, the molecular refractions of the compounds calculated from
them being sufficient. The values obtained with these solutions would
be expected generally to be very slightly lower than the pure sub-
stances would give.
The measurements of the ethylene chloride used gave the following
numbers :
/*• d ~d~^ Calc
a 1*44758 0*35285 3493 34*94
P 1-45607 0-35955 35*56 —
y 1-46124 0-36363 35-99
i H^- Ha 1-06.
Diap.
a. $. y. Hy-H«,
Borneol 75791 77*175 77913 2-122
Calc 76-200
DifF -0-409
Camphor 74-354 75*690 76*517 2-163
Calc 74-200
DifF +0154
o-Nitrocamphor 84205 85*965 86*991 ;;;;^ 2*786
Calc 84-250
Diff -0045
aa'-Chloronitrocamphor 92*298 94*167 95*280 2*982
Calc ; 92-820
Diff. -0*522
oa-Bromonitrocamphor 97*572 99*711 ^ 100*982 3*410
Calc 98*340
Diff. -0*768
Digitized by
Google
818 HfiWITT AND MOORE:
A MODIFICATION OF ZlfiI8KL*8
Cii.mphoryloxime > . 85*017
Calc 84-205
Diff +0-812
a-Ohlorocamphor 82*620
Calc 82-770
Diff -0-150
a-Bromocamphor 88*083
Calc 88-290
Diff -0*207
aa-Dibromocamphor 101*944
Calc 102-380
Diff -0*436
aj8-Dibromocamphor 101*572
Calc 102*380
Diff -0-808
a/3-Dibrom(Hi-chlorocamphor 110*391
Calc 110-950
Diff -0-560
i3.
86^845
84*183
7-
87-912
Dispi
Hy - H*.
2*896
85046 2*426
89-870 90*919 2-836
104131 105*564
3-620
103-633 104*988 3-416
112-851 114-276 3-885
XXXII. — A Modification of ZeiseVs Method fw the
Estimation of Methoocyl Groups.
By J. T. Hewitt and T. S. Moorb.
The method described by Zeisel for the estimation of methozyl (and
ethoxyl) groups in organic compoands has freqaently proved of con*
siderable service. A disadvantage of this method is, however, the
time taken both in setting up the apparatus and in carrying out the
estimation. With the view of improving the process in this respect^
we determined, if possible, to replace the condenser of the original ap-
paratus due to Zeisel by a fractionating column, and this proved so
satisfactory that we were able to dispense, not only with the condenser
and the supply of water at constant temperature, but also with the
potash-bulbs containing water with red phosphorus in suspension kept
Digitized by
Google
Method for the estimation of METHoxtL groups. 319
at 40 — 60°. An efficient column not only returns the hydriodic acid
to the decomposition flask^ but also effectually holds back any iodine
which Yolatilises.
The apparatus we have used takes the form shown in the accompany-
ing figure, the arrangement of the bulbs in the fractionating column
being shown at (a).
Carbon dioxide is washed by silver nitra»te solution and led into the
decomposition flask;
not simply into the
neck as in Zeisel's
method. An ordinary
round-bottomed flask
of 150—200 c.c.
capacity is employed,
the inlet tube for car-
bon dioxide termi-
nating about 1'5 — 2
cuL above the sur-
face of the liquid. A
fractionating column
also passes through
the cork of the flask,
the pattern of the
one -we have used
being I due to Mr.
J. N. Tervet of this
laboratory. The
number of bulbs in
the column has been
seven or eight.
The method of pro-
cedure is as follows :
About 16 c.c. of
hydriodic acid (sp.
gr. 1-68—1-70) is
poured Jnto the de-
composition flask. This is fitted with the cork carrying the carbon
dioxide leading tube and the fractionating column. The bulb of the
flask is then immersed in a glycerine bath and heated at about 130°,
whilst a slow current of carbon dioxide is passed'through the apparatus
for 10 minutes.* The apparatus is then disconnected, the flask and
its contents cooled, and about 0*2 — 0*3 gram of the substance under
* In this manner, impurities (for example, phosphine) are removed from the
bydriodio acid. We do not know if this precaution has been previously employe4.
Digitized by
Google
320
A MODIFICATION OF ZEISEL'S METHOD.
examinatioii weighed into it. The apparatus is then joined together
again, and the tube from the fractionating colamn connected to one
passing directly into an aqueous alcoholic solution of silver nitrate
prepared according to Zeisel's directions. A further check flask is
also added, as Zeisel recommends.
The decomposition flask and its contents are heated at 130^^, and a
slow current of carbon dioxide passed through the apparatus. Satis-
factory results are obtained if during the operation the thermometer a^
the top of the column indicates a temperature of 20 — 25^. De-
composition is complete in 45 minutes (maximum time).
The following results demonstrate the satisfactory nature of the
method :
Substance.
Formula.
Weight
taken.
Agl pro-
duced.
CH,0
found.
CH,0
calcu-
Uted.
Blank
CeriJ(CH(f)(0H)(06H,)
CivH70N(0H)(0CH,)
Ci,HaON8(OCH,)
0-2196
0-2388
0-8917
0-2746
0-2065
none
0-2581
0-8598
0-2674
0-2191
0-14S2
15-50
20-80
9-06
10-62
9-51
%
Brucine . ... «.
1674
Vanillin
20-89
Codeine*
10-86
10-86
Quinine
9-56
* The codeine reunified on treatment with hydriodic acid. In the second <
tion, the substance was heated with hydriodic acid mixed with about its own volume
of acetic anhydride.
The method generally giving such good results, attempts were made
to estimate methoxyl in easily volatile substances. The results were not
very satisfactory :
Substance.
Formula.
Weight
taken.
^JcT
CH,0
found.
CH,0
calcu-
lated.
Methyl acetote ...
Methyl oxalate ...
6o,CH,
0-2159
0-2207
0-6142
0-8492
81-41
60-76
41-89
62-64
For ethoxyl compounds, the glycerine bath is heated at HO^, and
the current of carbon dioxide is made somewhat more rapid towardB
the end of the operation. The temperature at the top of the fractionate
ing column is about 27° :
Digitized by
Google
THE RADIOACTIVITT OF THORIUM COMPOUNDS. I. 321
Substance.
Formula.
Weight
taken.
Agl pro-
duced.
CjHjO
. found.
§■11
Tn-Chlorobenzene-
B2o-j7-phenetole
CeH^Cl-Na-CeHj-OCsHB
0-2660
0-2304
17-23
17-64
East London Technical College.
XXXIII. — T%e Radioactivity of Thorium Compounds. I.
An Investigation of the Radioactive Emanation.
By E. BuTHBRFOBD, M.A., D.Sc, Macdonald Professor of Physics,
and Fbedebick Soddt, B.A. (Oxon.), Demonstrator in Chemistry,
McGill University, Montreal.
The following paper contains a preliminary account of an investigation
into the property possessed by the compounds of thorium of giving a
radioactive emanation, and also into the nature of the emanation
itself.
It was shown by one of us {FhU. Mag,, 1900, [v], 40, 1, 161) that
the compounds of thorium, besides being radioactive in the same sense as
the uranium compounds, also continuously emit into the surrounding
atmosphere, under ordinary conditions, something which, whatever its
real nature may be, behaves in all respects like a radioactive gas.
This '* emanation,'* as it has been named, is the source of rays, which,
like the Hontgen and uranium rays, and the ordinary well-recognised
type of thorium radiation, will darken a photographic plate, and will
render a gas capable of conducting an electric current (that is, will
" ionise " it), but is sharply distinguished from them by the following
considerations. It can be moved from the neighbourhood of the
thorium compound by a current of air passing over it, or even by the
process of ordinary gaseous diffusion, and transported long distances,
80 that the characteristic photographic and ionisation effects appear in
the air far away from the original source of radioactivity. The
Rontgen and uranium rays, as is well known, travel in straight lines
from their source, and any object opaque to them interposed in their
path will sharply screen the space behind. But in the case of the
thorium radiation there is no such screening effect, because here we
have a case of a substance emitting, not only straight line radiation,
but also particles of a gas, itself radioactive, capable of diffusing
VOL. LXXXI. Z
Digitized by VjOOQIC
322 RUTHERFORD AND SODDT J
through the surrounding atmosphere around ohetaoles placed in its
direct path, and bo arriving and producing its effects at points com-
pletely screened from rays travelling from the thorium in straight
lines. It was shown in the original communication that these effeota
could not be ascribed to minute particles of thoria dust carried oS
mechanically, and all the subsequent work on the subject shows that
the hypothesis that the compounds of thorium emit a radioactive
gas is not merely the only one which will explain the facts, but
that it does so in every observed case in a completely satisfactory
manner.
Present State qfthe Subject /ram a Phyaical StandpainL
In the papers referred to, the general character of the phenomena in
question was presented, and a short resume will perhaps not be out of
place here. It was shown that the radiation from the emanation decays
rapidly, but at a perfectly defined rate, that is, the effects it produces
diminish with the lapse of time, falling to about one^half the original
value at the end of one minute. This " rate of decay," as will be
shown later, is of great value in identifying and distinguishing
between different types of emanation.
The emanation passes unchanged through cotton wool, weak and
strong sulphuric acid, and aluminium and other metals in the form
of foil, but not through an extremely thin sheet of mica.
The emanating power of thoria is independent of the surrounding
atmosphere, but is destroyed to a large extent by intense ignition,
and does not return when the substance is kept.
One of the most striking properties of the thorium emanation is
its power of exciting radioactivity on all surfaces with which it comes
in contact, that is, a substance after being exposed for some time in the
presence of the emanation behaves as if it were covered with an in-
visible layer of an intensely radioactive material. If the thoria is
exposed in a strong electric field, the excited radioactivity is entirely
confined to the negatively charged surface. In this way, it is possible
to concentrate the excited radioactivity on a very small area. The
excited radioactivity itself has a regular rate of decay, but different
from that of the emanation, its effect falling to half value in about
11 hours. There is a very close connection between the exdted
radioactivity and the emanation. It was shown that the amount of
the former produced under various conditions was proportional to the
amount of the latter, and if the emanating power of thoria' be de-
stroyed by intense ignition, its power to excite radioactivity oorre*
spondiugly disappears. Some apparent discrepancies which at first
stood in the way of too close a connection being inferred^have resolved
Digitized by
Google
tHK RADlOACnVlTt OF THORltTlt COMPOUNDS. I.
themselyes by recent work into strong confirmation of the view that
the two are related to each other as caase and effect.
Another remarkable property of the excited radioactirity is that it
18 solnble in sulphuric and hydrochloric acids, that is, a platinum
wire, rendered radioactive by being made the negative pole of an
electrid field in the neighbourhood of some thoria, will give up its
radioactivity to these acids. If the acid be then evaporated, the
radioactivity remains on the dish, whilst if left to itself the radio-
activity of the add solution decays at a rate identical with that of the
original excited radioactivity on the platinum wire; .
Simultaneously with the discovery of excited radioactivity due to
thoria, Curie showed that radioactive barium possessed a similar
property. Later, Dom {Ahk. der Naturfortch. Ou,fihr ffalU-dhS., 1900)
repeated the work quoted for thoria, and extended it to include two
preparations of radioactive barium compounds (radium) prepared by
P. de Haen, and a preparation of radioactive bismuth (polonium). He
found that radium gave out an emanation which was similar to that
from thoria, but which retained its radioactive power much longer.
The excited radioactivity from radium, on the other hand, decayed more
rapidly than that from thoria. The special property of emitting an
emanation is, however, confined to thecompounds of radium and thorium,
those of uranium and polonium do not possess it to an appreciable
extent.
An approximate determination of the molecular weight of the
emanation produced by radium has been carried out (Rutherford and
H. T. Brooks, Nature, 1901, 64, 157) by a diffusion method, taking
advantage of the slow rate of decay of the radium emanation. From
comparison of the rate of diffusion of gases of known molecular weight
into one another, it was found that the molecular weight probably lies
between 40 and 100.
It seemed probable that an examination of the phenomena by
ehemical methods might throw light upon its nature, and the emana-
tion produced by thoria was chosen as more suitable for the purpose
than that produced by radium, on account of the obscurity still sur-
rounding the chemistry of the latter, and the difficulty of producing
material of even approximate uniformity of properties. Thoria, on the
other hand, is an article of commerce, and specimens from different
sources show surprising uniformity in this respect.
During the progress of the work, the subject has acquired additional
importance and interest through the discovery by Elster and Geitel
{Fhy$. ZeU,^ 1901, 2, 590) that it is possible to produce excited radio-
activity from the atmosphere, without further agency, by simply
exposing a wire highly charged to a negative potential in the atmo-
sphere for many hours, and that this also possesses the property of
Z 2
Digitized by VjOOQIC
824 RUTHERFOEB AND SOHDY :
being dissolved o& by aciis, and of being left behind unchanged on
the evaporation of the latter. But here again the rate of decay is
different from that of the excited radioactivity produced by thoria,
which is evidence for assuming that the two are probably not identical,
although so strikingly analogous. However, the close connection
between excited radioactivity and the emanation established in the
case of thoria renders it probable that the excited radioactivity
obtained from the atmosphere is caused by the presence there of an
emanation or radioactive gas analogous to, although probably different
from, the Thorium emanation. The discovery is likely, as Elster
and Geitel point out, to have important bearings on the theory of
atmospheric electricity, and in our opinion renders a close study of
the thorium emanation the more imperative.
The Chemical Aspect qf the Quealian,
The foregoing furnishes a short review of the physical side of the
question at the present time. With regard to the chemical aspect,
this has so far not been studied. The photographic method, almost
the only one that has until now been used by chemists in the study
of radioactivity, is not one which allows of the recognition and differen-
tiation of an emanation as a component factor in producing the
phenomena. The photographic method is of a qualitative rather than
a quantitative character ; its effects are cumulative with time, and as
a rule long exposures are necessary when the radioactivity of a feeble
agent like thoria is to be demonstrated. In addition, Bussell has
shown that the darkening of a photographic plate is brought about
also by agents of a totally different character from those under con-
sideration, and, moreover, under very general conditions. Sir William
Crookes (Proc, Roy. Soc., 19G0, 66, 409) has sounded a timely note of
warning against putting too much confidence in the indications of the
photographic method of measuring radioactivity. The uncertainty
of an effect produced by cumulative action over long periods of
time quite precludes its use for work of anything but a qualitative
character.
Two or three chemists have studied the radioactivity of thoria, using
the photographic method, without, however, distinguishing between the
radioactivity due to the emanation and that due to the thoria itself.
Sir William Crookes (loc. cit.), who succeeded by an elegant method
in separating and isolating the radioactive constituent of uraniumy
also describes some experiments on thorium compounds with the same
object, but did not succeed in effecting a separation. A method based
on the fractional precipitation of the sulphate failed completely, but
another method, the fractional crystallisation of the nitrate, gave pre-
Digitized by
Google
THE RADIOACTIVITY OF THORIUM COMPOUNDS. I.
parations showing a difference in^their photographic actions in the
ratio of one to three. According to slight variations in the method
employed, as, for example, whether a glass or a card bottom was used
for the cell containing the substance to be tested (and both seem to
have been employed), the radiation from the emanation would or
would not contribute largely to the photographic action observed.
Debierne {Compt. rend,^ 1900, 130, 906), working on a very large
scale, obtained from pitchblende, by using reactions which would lead
to the separation of thorium, a material different in its chemical
properties from radium (barium) and polonium (bismuth), but consist-
ing in great part of thorium. This preparation was 100,000 times
more active than uranium, and he therefore assumed the existence of
a new element, ''actinium," therein. He hazarded the suggestion
that the radioactivity of thoria is due to the presence of the same
substance, and derived support for this view from the recent work
of one of us on the radioactivity of thoria, although on what grounds
is not clear.
In the course of their work on the atomic weight of thorium^ Brauner
(Trans., 1898, 73, 961) and Baskerville {J, Amer. Chem. Soo., 1901,
23, 761), have obtained evidence of the presence of a foreign
substance associated with thorium. The latter noticed that the
separation, as he interpreted it^ of this impurity reduced the photo-
graphic action considerably, and he concluded that the pure material
would be without photographic action. He employed a modification
of Crookes' photographic method, but it cannot be decided with
certainty from the description whether the radiation from the emana-
tion would be eliminated or not.
The present work is concerned primarily with the radioactive
emanation, although, of course, frequent occasion has arisen to examine
correspondingly the ordinary radiation also. The methods employed
are of an electrical character, based on the property generally
possessed by all radiation of the kind in question, of rendering a gas
capable of discharging both positive and negative electricity. These,
as will be shown, are capable of great refinement and certainty. An
ordinary quadrant electrometer is capable of detecting and measuring
a difference of potential of at least 10 ~* volts. With special instru-
ments, this sensitiveness may be increased a hundredfold. An average
value for the capacity of the electrometer and connections is 3 x lO"'^
microfarads, and when this is charged up to 10"' volts, a quantity of
electricity corresponding to 3 x 10~^^ coulombs is stored up. Now in
the electrolysis of water one gram of hydrogen carries a charge of
Ufi coulombs. Assuming, for the sake of example, that the conduction
of electricity in gases is analogous to that in liquids, this amount of
electricity corresponds to the transport of a mass of 3 x 10'^^ grams
Digitized by
Google
326 BUTHEBFO&D AND 80DI)7 :
of hydrogen, that is, a quantity of the order of 10'^' times that
detected by the balance. For a more delicate instrument, this amount
would produce an inconveniently large effect.
The effects under investigation, from the nature of their manifesta-
tion, may well be, and probably are, produced by quantities of matter of
the order of magnitude described, and therefore altogether beyond the
range of the balance. But to assume on that account that the
subject is beyond the pale of profitable chemical investigation is need-
lessly to limit the field of chemical inquiry. Although surpassing
the spectroscope as a detective agent, as a quantitative instrument
the electrometer is little inferior in accuracy to the balance. To take
as an example the case of thoria mixed with zirconia, the former
could be detected and accurately measured by means of its emanation
with an electrometer, even although it were only present to the
extent of one part in many thousands. A distinction must be made
here between emanation and rnnanatirig power. The quantity of the
former is what is measured by the electrometer. To express this in
terms of weight, the emanating power, that is, the quantity of emana-
tion produced by a given weight of the substance in question, must
be Imown. As will be shown, this value varies with the previous
history and present condition of the substance.
The electrometer also affords the means of recognising and differ-
entiating between the emanations of different chemical substances* By
the rate of decay, the emanation from thorium, for example, can be
instantly distinguished from that produced by radium, and although a
difference in the rate of decay does not of itself argue a fundamental
difference of nature, the identity of the rate of decay furnishes at
least strong presumption of identity of nature.
In the sense that has just been explained, the electrometer can be
said to supply the investigation of the property of emanation with
methods, so to speak, of quantitative and qualitative analysis which
are simple and direct, and there is therefore no reason why the
property in question, and even the nature of the emanation itself,
should not be the subject of chemical investigation.
Scope qf Work.
Of the great number of questions which immediately present them-
selves for answer in an investigation of this kind, the following are at
present claiming our more immediate attention.
1. Is the power of producing an emanation a specific property of
thorium, or is it to be ascribed to the presence of a foreign substance,
possibly in minute amount, associated with it and amenable to chemical
methods of separatiou t
Digitized by
Google
THE RADIOACTIVITY OF THOBIUM COMPOUNDS. I. 327
2. Oan the emanating power of *' de-emanated " thoriabe regenerated
by chemical means 9 It has been mentioned that thoria, when intensely
ignited, loses to a very great extent its power of giving an emanation.
If such de-emanated thoria be subjected to a series of chemical changes,
will it regain its emanating power or not f
3. Does the emanation or radioactive gas itself possess any property
which would associate it chemically with any known kind of gravita-
tional matter t
4. Is it possible to detect, by means of the balance, any loss in weight
corresponding to the continuous emission of the emanation or any gain
in weight of bodies rendered radioactive thereby 1
5. Does the chemistry of thorium present any peculiarity capable of
being connected with its almost unique power of producing an
emanation t
To interpret rightly the results obtained, a more or less complete
study of the effect of chemical and physical conditions on the eman-
ating power is necessary. The effect of the state of aggregation, the
presence or absence of water, the influence of light, temperature, the
nature of the surrounding atmosphere, the lapse of time since prepar-
atioa, &c., on the emanating power, as well as the differences in this
property exhibited by different compounds, have been investigated.
The present communication does not attempt^ a full answer to all
the above questions. The results so far obtained in answer to the first
three will be presented. The work on the fourth is in progress, whilst
the results of the investigation of the fifth question will be most
conveniently given later in a separate communication.
EkctirorMter Method qf rMtwwring Emanating Power cmd RcuiioacHvity,
The term radioactive is now generally applied to a class of sub-
stances, like uranium, thorium, radium, and polonium, which have the
power of spontaneously giving off radiations capable of passing
through thin plates of metal. The radiations are in some cases very
complex, but in the case of the substances mentioned, a portion at
least of the radiation is similar in all respects to easily absorbed
Rontgen rays. The characteristic and general property possessed by
these radiations is to produce, from the gas through which they pass,
positively and negatively charged carriers, which in an electric field
travel to the negative and positive electrodes respectively. In this
way, a small current is able to pass through a gas exposed to the radia-
tions, even with a very weak electric field, and the measurement of
this current by means of the electrometer affords a means of com-
paring the intensities of radiation.
4b bas been mentioned, compounds of thorium (and radium), in
Digitized by
Google
328 RUTHERFORD AND SODDY :
addition to radiations which travel in straight lines, emit radioactive
emanations, which behave in all respects like a temporarily radioactive
gas, and diffuse rapidly through porous substances, as, for example,
thick cardboard, which are completely opaque to straight line radiation.
Each particle of the emanation behaves as if it were a radiating centre,
producing charged carriers throughout the gas in its neighbourhood.
The emanation passes through plugs of cotton wool and can be babbled
through liquids without appreciable loss of radioactivity, whereas the
charged carriers, produced by the emanation in common with the
straight line radiation from radioactive substances, on the contrary,
completely disappear on passing through a plug of cotton or glass wool,
or by bubbling through liquids. The means of eliminating the effects
of the straight line radiation and of measuring the amount of the
emanation alone thus suggest themselves. Air passed over uranium
or other non-emanating radioactive substance will no longer conduct a
current after passage through cotton wool. The conductivity in the
case of thorium, however, will persist, and afford a measure of the
amount of emanation present.
Fig. 1 shows the experimental arrangement for comparing the
emanating power of substances. These are placed in the form of fine
powder in a shallow lead vessel inside the glass cylinder, C, 17 cm. in
length and 3 '25 cm. in diameter, provided with indiarubber corks. A
current of air from a large gas-bag, after passing through a tube con-
taining cotton wool to remove dust particles, bubbled through sulphuric
acid in the vessel, ^i. It then passed through a bulb containing tightly
packed cotton wool to prevent any spray being carried over. The
emanation mixed with air was carried from the vessel C through a plug
of cotton wool, 2>, which completely removed all the charged carriers
carried with the emanation. The latter then passed into a long, brass
cylinder, 75 cm. in length and 6 cm. in diameter. The cylinder insu-
lated on paraffin blocks was connected to one pole of a battery of small
lead accumulators, the other pole of which was connected to earth.
Three electrodes, E, F, H, of equal length were placed along the axis of
the cylinder, supported by brass rods passing through ebonite corks in
the side of the cylinder. The current through the gas, due to the
presence of the emanation, was measured by means of a Kelvin quad-
rant electrometer of the White pattern. The electrometer and the
connections were suitably screened by means of wire gauze connected
to earth. An insulating key was arranged so that either of the elec-
trodes E, Fy H, or all of them together, could be rapidly connected to
one pair of quadrants of the electrometer, the other two being always
connected to earth.
The insulation of the electrodes was first tested by sending a current
of air through the apparatus without any emanating material in C^
Digitized by
Google
THE RADIOACTIVITY OF THORIUM COMPOUNDS. I.
329
The rate of movement of the
electrometer needle was ac-
curately observed. On placing
the emanating substance in
G and continuing the air
current for several minutes at
a constant rate, the current
due to the emanation reached
a steady state. On separating
the quadrants of the electro-
meter, the deflection from
zero increased uniformly with
time. The time taken to pass
over 100 divisions of the scale
was observed with a stop-
watch. The number of di-
visions passed over per second
may be taken as a measure of
the current through the gas.
With this apparatus, the
emanation from 10 grams of
ordinary thorium oxide pro-
duces a current of 3-3 x lO""^^
amperes between the three
electrodes connected together
and the cylinder. With the
electrometer working at aver-
age sensitiveness, this corre-
sponded to a deflection of 100
divisions of the scale in 12
seconds, so that one-hundredth
part of this current could be
readily measured, that^is, the
emanation produced by one-
tenth of a gram of thorium
oxide.
An electrometer one hun-
dred times more sensitive than
this failed to detect the pres-
ence of an emanation or radio-
activity in the oxides of tin,
zirconium, and titanium, the
other elements of the same
group in the periodic table.
r*
o
5
1
Digitized by
Google
330 RUTHERFORD AND 80DDT
VaricUion of the Current with VoUage, — ^The current through the
gas observed with the electrometer at first increases with the voltage,
but a stage is soon reached when there is a very small increase for a
large additional voltage. This is one of the most characteristic pro-
perties of conducting gases. For small voltages, only a small pro-
. portion of the charged carriers reach the electrode, on account of their
recombination throughout the volume of the gas. When the electric
field is increased until all the carriers reach the electrode before any
appreciable recombination can occur, the current is at a maximum,
and remains constant for large increases of voltage, provided, of
course, that the electric field is below the value necessary for a
spark to pass. In the experimental case, a pressure of 50 volts
was found sufficient to give the maximum current between the
electrodes.
This property of conducting gases allows us at once to make sure
that the insulation of the electrodes is perfect at all stages of a long
experiment; 100 volts applied instead of 50 to the cylinder should
give the same current if the insulation is unaffected.
Rate of Decay of the Radiation from the Emanation, — ^The three
electrodes, E^ F^ H^ were used to compare the *' rates of decay " of the
radiations from the emanations of different substances. In the
previous papers quoted, it has been shown that the radiating power of
the thoria emanation falls to half its value in about a minute. In
consequence of this, the current observed for the electrode E is greater
than for electrode, H, Knowing the velocity of the current of air
along the cylinder and the respective currents to the electrodes Ey F,
J7, the rate of decay of the radiation can be readily deduced. If,
however, we merely require to compare the rate of decay of one
emanation with another, it is only necessary to compare the ratio of
the currents to the electrodes E^ F, H in each case, keeping the
current of air constant. If the ratio of the currents is the same we
may conclude that the radiating power of each diminishes at the same
rate. The comparison of emanation is thus rendered qualitative as
well as quantitative. In most of the experiments, the current to the
electrode E, was about twice that to the electrode H ; the velocity
of the current of air along the cylinder was thus about 0*8 cm. a
second.
Comparison of Emanating Power. — ^The experiments in all cases on
the amount of emanation from different substances are comparative.
The standard of comparison was usually a sample of 10 grams of
thoria as obtained from the maker, which gave out a conveniently
measurable quantity of emanation. Preliminary experiments were
made to find the connection between the weight of thoria and the
mnoiint of emiuiatiop as tested in the cylinder. The following
Digitized by
Google
THE BADIOACTIVITY OF THORIUM COMPOUNDS. I. 331
niimbers show that the amount of emanation is directly proportional
to the weight of the substance :
Divisions of
Weight of thoria. scale per second.
2 grams. 1*41
4 „ 2-43
10 „ 6-33
20 „ 13-2
This result shows that within the limit of accuracy desired we may
take the amount of emanation as directly proportional to the weight
of the substance. The determinations in the above table were made
with the three electrodes connected together with the electrometer,
and with a constant flow of air. The lead vessel in which the thoria
was placed was 7*4 by 3*5 cm. in area and 3 mm. deep. In the com-
parison of emanating power, the maximum current between the elec-
trodes for the standard 10 grams of thoria was first observed. This
was removed and a known weight of the specimen to be tested was
substituted, and the deflections again observed after the conditions
had become steady.
If d^^'No. of divisions per sec. for a weight, v^^, of thoria ;
then
Emanating power of specimen d^w^
Emanating power of thoria " djW2
The values c^ and d^ are corrected, when necessary, for natural leak-
age, that is, the current which passes under similar circumstances when
no emanating material is present. This current is chiefly made up of a
leakage due to conduction over the ebonite, as well as the current
produced by the excited radioactivity which has collected on the nega-
tive electrode during the course of the day's experiments. It is
generally very small, and the correction is only necessary when a
specimen of substance almost free of emanation is being tested.
An example taken at random from the note-book will serve to
illustrate the method of calculating the results, the emanating power
of the comparison sample being considered 100 per cent. :
Dec 7th, 11 a.m. — Natural leakage 10 divisions in 50""
0-20 „ 1"
6 grams comparison sample ThOg 100 „ 23*5"'
3-6 „ ThO, ignited 24 hours over
Bunsen burner in platinum crucible 50 „ 35*2"
(2^b4*25, corrected for nat. leakage^ 4*05
<^,= l-42- 1-22
^ « 0-42, ov 42 per oe»t,
Digitized by
Google
S32
RUTHERFORD AND BODDY :
Compariaan of InUnsity qf Straight Line Radiation.
It was frequently of interest to obtain information about the
i^ten8it7 of the ordinary radiation correspondingly with measurement
of emanating power. In order to do this rapidly and accurately, the
following method was used. Fig. 2 shows the general arrangement.
0*1 gram of the compound to be tested was reduced to fine powder and
uniformly sifted over a platinum plate 36 sq. cm. in area.
This plate was placed on a large metal plate connected to one pole
of a battery of 300 volts, the other pole of which was earthed. An
insulated parallel plate was placed about 6 cm. above it, and the whole
apparatus enclosed in a metal box connected to earth, to prevent
electrostatic disturbance. The shaded portions in the figure represented
Fig. 2.
Earth.
Hi-
Earth.
insulators. A door was made in the apparatus so that the plate could
be rapidly placed in position or removed. The current between the
plates is observed in the usual way with the electrometer. The ratio
of the currents for two substances is a comparative measure of their
radioactivity. It is only possible to compare together with certainty
substances of similar density and state of division, — a light, floury
material will tend to give lower values than a dense powder.
If a substance gives off an emanation, the current between the
plates increases with time. Under these conditions when the thoria
is exposed in thin layers with a maximum of radiating surface, all but
1 or 2 per cent, of the total effect is due to the straight line
radiation ; even when the effect due to the emanation has attained its
Digitized by
Google
TBfi HADIOACTIVITY OP THORIUM COMPOUNDS. I. 338
maximum, this constitutes a very small percentage of the whole. This
effect, however, may be to a large extent eliminated by taking the
current between the electrodes immediately after the material is
placed in the testing apparatus, or by passing a current of air
between the electrodes to remove the emanation, and prevent it
accumulating.
It is thus possible to compare intensity of radiations with an error
not exceeding 1 or 2 per cent., and with great rapidity, and in these
respects the electrical method is altogether superior to the photo*
graphic.
Camjpariaon of Emcmating Power, — The apparatus (Fig. 2) described
for the comparison of radiations, can also be quite well employed for
a comparison of emanating power. In this case, a thick layer of
thoria (several grams) is spread over the plate and covered with two
thicknesses of ordinary paper. This has been found almost completely
to stop the straight line radiation, whilst allowing the emanation to
pass through unimpeded. The current is now measured when a steady
state has been reached, due to the accumulation of * the emanation.
This takes some time, and draughts of air must be guarded against.
For this reason, it is less convenient than the method first described,
but the results obtained by the two methods are almost exactly the
same. Thus a sample of '^ de-emanated '' thoria which gave 12 per
cent, of the emanating power of the comparison sample by the first
method gave 13 per cent, by the second method, whilst a sample of
oxide prepai'ed from thorium oxalate gave 37 per cent, and 39 per
cent, by the two methods respectively. The close agreement in the
values by methods so completely different in character is a proof that
the indications of the methods are worthy of a great degree of
confidence.
The De-emanation of Thoria and the Regeneration qfthe
Emanating Power,
The emanating power of thoria, as has been stated, is destroyed to a
large extent by intense ignition. A closer study of this is the first
step in the investigation of the phenomenon. Previous experiments ^
had not succeeded in completely de-emanating thoria, although a
reduction to about 15 per cent, of its original value had been
accomplished. A sample of this preparation which had been kept for
two years had not altered from this value. An experiment was
performed in which thoria was heated to the highest temperature
which could be safely employed with platinum vessels : (1) in a thin
layer in a large platinum dish, and (2) in bulk in a small platinum
crucible placed inside the dish. The two were heated together by
Digitized by
Google
334 RttTHERFORl) AKD ^ODDY :
means of a powerful gasoline furnace for one hour. The temperature
was such that the fireclay walls fused, and the pipeclay of a triangle
showed signs of having been softened. It was found that the sample
that had been heated in a thin layer in the dish retained about 16 per
cent, of its original emanating power, whilst the other sample retained
about 8 per cent. There is thus no advantage in heating in thin
layers, in fact rather the reverse^ for the sample showing 16 per cent;
again heated to a slightly lower temperature for half-an-hour in a
small crucible was reduced to 12 per cent.
In another experiment, a small platinum crucible filled with thoria
was heated for half-an-hour in a small furnace by a large blowpipe and
powerful pair of bellows. Some asbestos wool had completely fused
on the outside of the crucible, and the temperature was probably but
little lower than in the previous experiment. This sample also
retained about 8 per cent of its emanating power. No further
attempt has yet been made to completely destroy the emanating
power.
A small quantity of thoria heated in a platinum crucible in the open
over an ordinary small sized blowpipe and bellows for five minutes
retained about 45 per cent, of its emanating power. The effect of time
as well as of temperature was studied by heating about equal quantities
in a platinum crucible over an ordinary Bunsen burner for different
periods.
Heated 10 minutes Emanating power « 61 per cent.
„ 1 hour „ - 69 „
,1 24 hours „ ^ 42 „
It thus appears that there is a large and practically sudden decrease
of emanating power for each temperature above a red heat, followed
by a very gradual decrease with time when the temperature is main-
tained ; thus five minutes on the blowpipe, whilst much more effective
than the same time at the temperature of the Bunsen burner, pro-
duced rather less effect than 24 hours at the latter temperature.
Effect of Moisture, — ^The next point to be examined was whether the
loss of emanating power could be attributed to a loss of water and
desiccation of the thoria by ignition. A sample of de-emanated thoria
(retaining about 14 per cent.) was placed in the middle of a Jena
glass tube, one end of which was closed and contained water, the other
end being drawn out to a jet. This was supported in a powerful tube
furnace in a sloping position, and the part containing the thoria
heated to the highest possible temperature, while a slow current of
steam from the water at the end was passed over it, escaping by the
jet. When all the water had evaporated, the jet was drawn off and
the tube allowed to cool in an atmosphere of steam 'free from air.
Digitized by
Google
THB fiADlOACnVlTt OF T^ORIttH COlCPOUi^0S. t ^S6
The thoria, on testing, was found to have been lowered in emanating
power to abont 7 per cent. The further heating had thus reduced the
emanating power without the steam having at all regenerated it.
In the next experiment^ the reverse process was tried. Two ezaetlj
parallel processes were carried out for ordinary thoria possessing the
normal amount of emanaticg power. In the first, it was heated in a
porcelain tube in the tube furnace for three hours, while about 500 0.0.
of water were distilled over it from a retort. In the second, another
quantity of thoria was heated in exactly the same way for the same
time, only a current of well dried air was substituted for the steam.
The result was conclusive : each sample had had its emanating power
reduced to exactly the same amount, that is, about 50 per cent, of the
original
These experiments prove that water vapour exerts no influence
either in de-emanating thoria or in effecting a recovery of its lost
emanating power.
The lUgeneralian qf the Etnanaiing Power by Chemiedl Fracesaes, —
The task of subjecting de-emanated thoria to a series of chemical
changes to see if it would recover its lost emanating power was then
undertaken.
It may first be mentioned that thoria which has been subjected to
ignition has changed very materially in chemical and physical proper-
ties. The pure white colour changes at temperatures corresponding
to the first stages of de-emanation to a light brown, and after subjec-
tion to the very highest temperatures to a pure pink. At the same
time, as has been observed before, the solubility of the substance in
sulphuric acid is greatly diminished. A part always obstinately
refuses to dissolve, even after long and repeated boiling with the con-
centrated acid, although this part is diminished by each successive
treatment and appears to be in no way different from the rest of the
substance. No difference, however, occurs in the readiness with
which chlorine attacks it when intimately mixed with carbon. The
formation of the chloride by this method is the easiest way of
dissolving ignited thoria.
Two quantities of the same de-emanated thoria were converted, the
one into chloride and the other into sulphate, by the usual methods,
and from each of these the oxalate was formed by precipitation of
the acid solution with oxalic acid. Parts of the oxalates were then
converted into oxides by heating over the Bunsen burner. In both
cases there was a very marked recovery in emanating power; the
oxide obtained from the sulphate had about 40 per cent., that from
the chloride about 55 per cent., whereas the de-emanated thoria
from which they were both produced had about 13 to 14 per cent, of
the emanating power of thoria. The oxalates from which the oxides
Digitized by
Google
836 RUTHERFORD AND SODDY :
were formed each had about 1 1 per cent, of the power, and in converting
them into oxides it was ascertained by a direct trial that too high a
temperature had been employed and the thorium oxide had suffered
partial de^manation. At this time also, it was beginning to be
realised that the emanating power was a quantity which varied, not
only with the nature of the chemical compound, but also for the same
compound very materially with its previous history. Thus the oxide
from the oxalate does not possess as a rule so great an emanating
power as that used for comparison, which would account for the above
result. The following two eicactly parallel experiments were therefore
made, the one with ordinary, and the other with de-emanated thoria
possessing 9 to 10 per cent, of the emanating power of the first.
Each was converted to chloride in the ordinary way, by mixing with
sugar solution, carbonising, and igniting the mixture of oxide and
carbon so obtained in a current of dry chlorine. Each sample was
then treated with water, the thorium precipitated as hydroxide with
ammonia, and the hydroxides washed and dried at 110^ The
hydroxide prepared from the de-emanated thoria possessed 128 per
cent., that from the ordinary thoria 108 per cent, of the emanating
power of ordinary thoria, when tested immediately after drying.
Now a sample of hydroxide previously obtained had shown no less
than three times the emanating power of ordinary thoria. The
specimens were therefore tested again after having been kept for four
days in loosely corked tubes. They now showed 157 per cent, and 139
per cent, respectively. The emanating power was thus increasing, so
both specimens were exposed side by side in open watch glasses under
a sheet of glass to keep off the dust. The result is again conclusive :
From de-emanated ThOj. From ordinary ThOj,
After nine days 253 per cent. 253 per cent.
After three more days 259 „ 259 „
Thus the process of de-emanating thoria by ignition does not irre-
trievably destroy the emanating power, for after solution and repreci-
pitation no difference whatever exists in the emanating power between
ordinary and de-emanated thoria.
The results also bring out another point, — the marked effect of
time and exposure to air in increasing the emanating power of thorium
hydroxide. This will be examined more fully later.
A fair conclusion from these experiments is that the cause of the
emanating power is not removed by ignition, but only rendered for
the time being inoperative.
Radioactivity of D^-muinaUd Thoria* — ^The '' straight line '' radiation
of thoria, de-emanated as completely as possible by ignition, was com>
pared with that of ordinary thorium oxide by the method described.
Digitized by
Google
THE RADIOACTIVITY OF TH0RIT7H COMPOUNDS. I. S87
It was found that within the limits of error no difference whatever
ooold be detected between them. This result serves to bring out the
fact that the power of thoria to give an emanation is independent of
its power to give a direct radiation.
I8 ike EmomaOng Fow&r a Speetfie Property of Thorium ?
Having shown that the de-emanation of thoria hj the processes
described consists rather in a temporary obliteration of the effect than
in a removal of the cause producing it, the next question to be con-
sidered is whether it is possible to remove from thorium compounds
by chemical methods any constituent to which the property of emanat-
ing power can be traced.
The thoria used in the investigation is that supplied by Messrs.
Eimer and Amend of New York, and is obtained from monazite sand
by a secret process. It, of course, does not consist of pure thoria,
although from superficial investigation it appears to be of excellent
quality. There is a small quantity of a substance present which can
be precipitated by sodium phosphate after removal of the thorium as
hydroxide by ammonia, the nature of which is at present under investi-
gation. The most noticeable impurity is about 1 per cent of thorium
«ulphate. Careful washing completely removed this impurity, and^the
emanating power of the washed sample was identical with the
ordinary. The impurity may therefore be disregarded for present
purposes.
Emanating power is not confined to thorium from any particular
source. Orangeite and thorite from Norway both possess it as well as
monazite sand from Brazil. A specimen of thoria prepared from
orangeite by the ordinary processes possessed about the same emanating
power as that obtained from monazite sand by the secret process.
A quantity of thorium oxide was converted into the anhydrous
sulphate and dissolved in iced water. The temperature was allowed
to rise and the hydrated sulphate precipitated in four fractions, a fifth
being obtained by evaporation of the mother liquor to dr3mess.
These showed no marked difference in emanating power among them-
selves. The first fraction was dehydrated and again submitted to
fractional precipitation as hydrated sulphate. The first fraction of the
new series— designated fraction AA — was then compared in the
following manner with the mother liquor fraction of the first series —
designated as fraction E. Both were dehydrated, dissolved in water,
and precipitated as hydroxide by ammonia, washed and dried under
the same conditions, and compared together at regular intervals with
a comparison sample of ordinary thoria.
VOL. LXXXI. A A
Digitized by
Google
338 RUTHERFOBD AND 80DDT :
Fraction AA. Fraction E.
At first 203 per cent. 200 per cent.
After 1 day 240 „ 249 „
After 13 days 316 „ 321 „
After 43 days 352 „ 372 „
The differences are too small to afford any indication of separation
of the emanating material.
The straight line radiations of the two fractions tested in the
apparatus (Fig 2) also proved to be identical.
It was obyiously useless trying any further fractionations by this
method. Since there was no appreciable difference in either property
in the fractions tried, there was nothing to be gained in a further re-
petition. These results completely accord with those of Sir William
Crookes {loo, ciC,), with which, however, we were not acquainted until
after our own experiments had been performed.
Another method for the puri6cation of thoria, employed by Dennis
{J. Amer. Chem. Soc, 1896, 18, 947), the precipitation of the hydroxide
by potassium azoimide, was next tried. The latter reagent was pre-
pared by Thiele's method (Annctlen, 1892, 270, 1) from diazoguanidine
nitrate. Hydrazoic acid partially neutralised with potash precipitates
thorium hydroxide from the boiling solution of a thorium salt. This
hydroxide, compared with a sample which had been precipitated wit?
ammonia in the ordinary way, showed similar emanating power.
These results, which fail to give any indication of a separation of the
emanating material by chemical means, taken in conjunction with
those already described in the preceding section on the regeneration of
the emanating power in de-emanated thoria, certainly seemed to point
to the conclusion that the power of giving an emanation is really a specific
property of thorium. Recent results, which will be given in the last
section (p. 343), put the question in a fresh light.
Effect qf Conditions upon Emanating Power,
Before any further work was undertaken, it was necessary to make
a close study of the influence of conditions upon the emanating power
of thorium compounds.
Effect of Temperature, — The effect of increase of temperature on the
emanating power of thoria has already been fully investigated by one
of us (Fhys. Zeit, 1901, 2, 429). The results stated briefly show that
an increase in temperature up to a certain limit, in the neighbourhood
of a red heat, correspondingly increases the emanating power. At
the maximum, this is between three and four times the value at the
ordinary temperature, and is maintained at this increased value for
Digitized by
Google
THE RADIOACTIVITY OF THORIUM COMPOUNDS. L 839
several hours without anj sign of diminution with time. When the
thoria is allowed to cool, the emanating power then returns to the
neighbourhood of the normal value. If, however, the limit of temper-
ature given is exceeded, de-emanation sets in, and even while the
high temperature is maintained, the emanating power falls rapidly
to a fraction of its former value. On cooling, the substance is found
to be more or less de-emanated. It is of interest that no increase
of emanating power is observed when de-emanation commences.
These experiments were extended to include the effects of cooling.
The platinum tube which contained the thoria was surrounded with a
felt jacket containing a mixture of solid carbon dioxide and ether. The
emanating power immediately fell to 10 per cent, of its former value.
On removing the cooling agent, it again rose quickly to nearly its
normal value.
In another experiment, some thoria was surrounded in a platinum
crucible with a mixture of solid carbon dioxide and ether, and
kept in a vacuum for several hours. On removing it and allowing its
temperature to rise, it possessed much the same value as an ordinary
sample, and after standing some time in the air it was again tested and
no difference could be detected between the two.
Thus changes in temperature produce very marked simultaneous
changes in emanating power, but between the limits of — 110° and an
incipient red heat no permanent alteration in the valu^ occurs.
Effect of Moisture, — Dorn {loo. dt,) had noticed that moisture pro-
duced a moderate increase in the power of thoria of giving an emana-
tion, and of exciting radioactivity on surrounding surfaces. We have
confirmed and extended his results by the following experiments.
Two similar weights of ordinary thoria were exposed in jars sealed
with wax, the one containing sulphuric acid and the other water, for
a period of 4 days. The desiccated sample showed 54 per cent, and
the sample exposed to water vapour 134 per cent, of the original
emanating power. The experiment was repeated and the samples left
for a week with much the same result : 70 per cent, and 141 per cent,
respectively. It was of interest to see if a more complete desiccation
would further reduce the emanating powen Five grams of thoria
were sealed up in a tube containing phosphoric oxid^, the two sub-
stances being separated by a plug of glass wool. Before sealing, the
tube was exhausted by a Topler mercury pump. After 26 days, the
end of the tube was connected with a closely packed phosphoric oxide
tube, the tip broken off inside the connection, and a slow stream of
dried air thus allowed to enter. The other end was connected to the
testing cylinder, and arrangements were made to send a stream of
air through into the cylinder. When all was ready, this end of
the tube was broken inside the connection, and the emanating power
A A 2
Digitized by VjOOQIC
340 BUTHEBIORD AHD SODDT :
maunred. A siiiiilar experiment made with an ordmarj sample of
ihoria, uring the same anangement^ showed that the desicoated
sample posaesaed 79 per cent, of the emanating power of the ordinaiy
sample tested under the same oonditiona.
A sample of thoria i^ra jed with water gave 125 per cent, of its
original emanating power. If completely flooded with water, howeTer,
the vmlne is much redoced, as would be expected from the redoetion of
sorfaoe.
Another trial was made, in which thoria was flooded with concen-
trated solphnrie acid« Hardl j an j emanation was observed so long as
the mixture remained nndistorbed, but when vigorooslj shaken it gave
nearl J one-half of the original emanation.
These experiments show that the presence of water, although pro-
ducing a marked increase, is not apparentl j essential for the produc-
tion of the phenomena. It must be mentioned, however, that thoria
only ceases to lose weight after prolonged ignition vdth the blowpipe^
that is, under conditions which nearly destroy its emanating power.
This, with analogous points, will be taken up, however, in a separate
communication on the more purely chemical side of the question.
The results of some experiments on the effects of other conditions
may be shortly tabulated. In each case the sample was exposed to the
conditions given for 4 days. The emanating power is that possessed
at the end of this period, compared with that of the first sample, which
is regarded as 100 per cent. :
1. Kept in sealed test-tube enclosed completely
in lead tube 100 per cent.
2. Taken from tightly stoppered stock bottle
containing the main quantity 1 00 „
3. Sealed up in test-tube and exposed to bright,
all-day sun 100 „
4. Exposed to the air of the laboratory in open
watch glass 105 ,,
5. Kept in a continuous stream of ordinary
air 88 „
The last experiment was made at a different time from the other
four, and therefore is not strictly comparable. The most useful
result attained is that thoria does not change in emanating power
when kept in closed vessels under different conditions, but when
exposed to the air the emanating power varies vrithin comparatively
narrow limits.
Thorium Hf^droxids^ — The effect of time on the emanating power
of the freshly prepared hydroxide already mentioned is one of the
most striking observations in this connection. The following addi-
Digitized by
Google
THE RADIOACTIVITY OF THORIUM COMPOUNDS. I. 341
tional experiments have been made on this point. A quantity of
hydroxide was prepared, and separate portions subjected to different
drying temperatures and subsequent conditions, as follows :
Emanating power.
1. Dried at 110^ and exposed some hours to
the air 264 per cent.
2. Dried as before at 110° and kept in desic-
cator until tested 226 „
8. Dried at 200° and kept in desiccator 220 „
4. Dried at 250° and kept in desiccator 219 ,,
From this, it appears that the additional loss of water caused by
exposure to increasing temperatures is without effect on the emanating
power.
A similar experiment to that described for thorium oxide was per-
formed with the hydroxide. Two quantities were exposed in closed
bottles to the action of moist air and of air dried with sulphuric acid re-
spectively, and showed, after 4 days, emanating powers of 394 per cent,
and 307 per cent. After having been exposed to the air for 24
hours, these samples showed 350 per cent, and 324 per cent,
respectively.
The next experiment was designed to include the effect of carbon
dioxide, which the hydroxide absorbs from the air to the extent of
2 per cent, of its weight. A quantity of hydroxide was tested
immediately after preparation, and possessed 140 per cent, emanating
power. A sample was sealed up in a test-tube, while another similar
sample was tested in the following manner. It was exposed to a
current of moist carbon dioxide for an hour, and then possessed an
emanating power of 156 per cent. It was then left exposed to the
air of the laboratory and tested at intervals :
After 2 days Emanating power 263 per cent.
if 6 J9 *••• >i 325 „
„ 10 „ „ 300
„ 11 if .« „ 341 „
>i 16 „ „ 362 „
On the last day, the sealed up specimen was opened and examined,
and was found to possess an emanating power of 298 per cent. These
experiments show that if the air is fundamental in producing the in-
crease of emanating power with time, a very limited quantity of it is
effective. For the present, it is perhaps better to consider it as an
effect of time simply, hastened no doubt by the presence of water
vapour.
Digitized by
Google
342 RUTHERFORD AND SODDT:
On the Chemical Nature of the Snianatian,
The folIowiDg work has reference to the emaDation itself and not to
the material producing it, and was designed to see whether the
emanation possesses chemical properties which would identify it with
any known kind of matter. It had been noticed at the time of its dis-
covery that it passed unchanged through concentrated sulphuric acid.
The same holds true of every reagent that has been investigated.
The effect of temperature was first tried. The air containing the
emanation, obtained in the usual way by passage over thoria, was led
through the platinum tube heated electrically to the highest attainable
temperature, and also through the tube cooled by solid carbon dioxide
and ether. The tube was then filled with platinum black, and the
emanation passed through in the cold, and with gradually increasing
temperatures, until the limit was reached. The effect of the intense
heat was to convert the platinum black completely into platinum sponge.
In another experiment, the emanation was passed through a layer of
red hot lead chromate in a glass tube. The current of air was replaced
by a current of hydrogen and the emanation sent through red hot
magnesium powder, and red hot palladium black, and, by using a
current of carbon dioxide, through red hot zinc dust. In every case,
the emanation passed without sensible change in the amount. If
anything, a slight increase occurred, owing to the time taken for the
gas current to pass through the tubes when hot being slightly less
than when cold, the decay en route being consequently less. It will be
noticed that the only known gases capable of passing in unchanged
amount through all the reagents employed are the recently discovered
gases of the argon family.
But another interpretation may be put upon the results. If the
emanation were the manifestation of excited radioactivity on the
surrounding atmosphere, then since from the nature of the experiments
it was necessary to employ in each case, as the atmosphere, a gas not
acted on by the reagent employed, the result obtained might bo
explained. Red hot magnesium would not retain an emanation con-
sisting of radioactive hydrogen, or red hot zinc dust an emanation
consisting of radioactive carbon dioxide. The correctness of this
explanation was tested in the following way. Carbon dioxide was
passed over thoria, then through a T-tube, where a current of air met
and mixed with it, both passing on to the testing cylinder. But
between this and the T-tube, a large soda-lime tube was introduced,
and the current of gas thus freed from its admixed carbon dioxide
before being tested in the cylinder for emanation. The amount of
emanation found was quite unchanged, whether carbon dioxide was
Digitized by
Google
THE HADIOACnVITY OF THORIUM COMPOUNDS. I. 348
sent over thoria in the manner described, or whether an equally rapid
current of air was substituted for it, keeping the other arrangements
as before. The theory that the emanation mAj oonsbt of the
surrounding medium rendered radioactive is thus excluded, and the
interpretation of the experiments must be that the emanation is a
chemically inert gas analogous in nature to the members of the argon
family.
It is perhaps early to discuss these results from a theoretical point
of view, although it appears certain that an explanation of the nature
of the emanation must precede, as a necessary step, any hypothesis put
forward to account for emanating power. The explanation already
advanced and disproved being left out of the question, two other
views of the origin and nature of the emanation are still possible. It
may be that one of the inert constituents of the atmosphere is rendered
radioactive in the presence of thoria and so constitutes the emanation.
The actual amount being probably extremely small, and air being a
constant impurity in all gases as ordinarily prepared, it is of course no
argument against this view that emanating power is independent of
the gaseous medium surrounding the emanating material. An
experiment is in progress, however, to ascertain whether emanating
power persists in a current of gas as free from air as present methods
of preparation allow. The other alternative is to look upon the
emanation as consisting of a gas emitted by the thorium compound.
It is not necessary that such should contain thorium, it might con-
ceivably be an inert gas continuously emitted in the radioactive state.
In the present state of knowledge, it would be premature to attempt
to choose between these two alternatives. But in any decision of this
point, the work already given on the regeneration of the emanating
power of thoria deromanated by ignition, the continuous loss of
emanating power by successive ignition at increasing temperatures,
and the increase in the chemical activity of thorium hydroxide with
time, must be taken into consideration.
Concentration oj the Radionctive Material.
Since the preceding account was written, developments have been
made in the subject whioh completely alter the aspect of the whole
question of emanating power and radioactivity. The first has
reference to thorium nitrate, which in the solid state hardly possesses
any emanating power. In a careful determination, using 20 grams of
the finely powdered commercial salt, this worked out to be only
1*8 per cent, of the emanating power of thoria. Dissolved in water,
however, and tested for emanation by bubbling a current of air
through it, it gives about three times as much emanation as thorium
Digitized by
Google
344 RUTHERFORD AND SODDY :
oxide. That is, solution in water increases the emanating power of
thoriam nitrate nearly 200 timea The emanating power, as in the
case of solids, is proportional to the weight of substance present, and
within the limits tried is not much affected by dilution, for a solution
of 10 grams made up to 25 c.c. in volume possessed a similar value
when diluted four times.
Solutions of thorium chloride also give a large amount of
emanation. *
In these experiments, the cylinder C (Fig. 1) is replaced by a
Drechsel bottle. A drying tube of calcium chloride is inserted
between it and the testing cylinder to prevent the moisture destroying
the insulation of the latter. In this connection, the method of testing
the insulation by varying the voltage is invaluable. The air current
under these circumstances cannot of course be kept so constant as
when working with solid substances, and the results are not strictly
comparable in consequence, but the arrangement works well enough
for a first approximation.
Simultaneously with this observation of the latent emanating power
of thorium nitrate, it was noticed that preparations of thorium
carbonate varied enormously in emanating power according to their
method of preparation. A sample prepared from the nitrate by
complete precipitation with sodium carbonate showed an emanating
power of 370 per cent, of that of the ordinary oxide, and this value
remained fairly constant with time. In another experiment, the
precipitated carbonate was partially redissolved in nitric acid, and the
redissolved fraction completely reprecipitated with ammonia as hydr^
oxide. The result was remarkable : the carbonate had an emanating
power of only 6 per cent., the hydroxide one of 1225 per cent of
that of the ordinary oxide. On repeating the experiments, both
fractions proved almost equally inactive, the carbonate showing 14 per
cent, and the hydroxide 19 per cent, of the emanating power of thoria.
An even greater difference between these two similar experiments was
observed in the efiPects of time on the different preparations. In the
first, the carbonate did not alter in value in 7 days, whilst the
hydroidde steadily decreased :
Hydroxide. Carbonate.
Original 1225 per cent. 6*2 percent.
Aftei* 1 day 1094 „ 8-4 „
After 4 days 696 „ 4*8 „
After 7 days 614 „ 4-7 „
After 14 days 473 „ —
In the second experiment, the emanating power of both the carbonate
and hydroxide had increased many fold when tested 11 days later, and
Digitized by
Google
THE RADIOACTIVITY OF THORIUM COMPOUNDS. I. 846
the former now possesBed 109 per cent., the latter, 273 per cent,
(originally 14 per cent, and 19 per cent, respectively).
The straight line radioactivity of the carbonate from the first ex-
periment which possessed such a low emanating power is of interest.
It proved to be similar to that of a specimen of hydroxide of normal
emanating power, which it resembled in density and state of division.
'After having been kept 7 days without showing any sign of re-
covering its emanating power, it was redissoWed in nitric acid, and
reprecipitated with ammonia as hydroxide. The latter now possessed,
when first made, an enianating power of 65 per cent., and after 24
hours 145 per cent., from which value it did not much alter.
These results throw a new light on the question of emanating
power. In the first experiment, which we have so far not succeeded
in repeating, by an accident in the conditions apparently, two frac-
tions were separated from thorium which varied in their emanating
power in the ratio of 200 to 1. The active fraction diminished to
nearly a third of its original value in 14 days spontaneously, whilst
the activity of the inactive fraction was, to a large extent, regenerated
by solution and reprecipitation, in an exactly analogous manner to the
behaviour of thoria de-emanated by ignition. Attempts to repeat this
result have so far led to the production of two more or less completely
de-emanated fractions, which, however, spontaneously increase in
activity with time, as in the second experiment, and this seems to be
generally the case, whether incomplete precipitation is effected as in
the experiment given by re-solution of the carbonate in acid, or by
using a deficiency of sodium carbonate in the first instance.
The production of preparations of such low emanating power led
naturally to an examination being made of the filtrates and washings
for radioactivity. It was found that these possess when concentrated
both emanating power and radioactivity in considerable amounts,
although from the nature of their production they should be chemically
free from thorium. The behaviour is quite general, a dilute solution
of thorium nitrate, after the thorium has been precipitated as hydroxide
with ammonia, shows when concentrated an emanating power of from one-
third to two-thirds that of the original nitrate in solution. It does not
matter whether the thorium is precipitated with ammonia directly, or
after preliminary partial precipitation as carbonate — either by adding
insufficient sodium carbonate in the first place, or by precipitating
completely and dissolving part of the precipitate in nitric acid — the
thorium-free filtrate invariably possessed emanating power, and when
evaporated to dryness exhibited straight line radioactivity also in
amounts very much greater than possessed by the same weight of
thoria.
The result of a careful chemical investigation of the active filtrates
Digitized by
Google
846 RUTHERFORD AND SODDY :
prodticed under the various conditions described was to show that these
contained no thorium, or at most only a minute trace, hut another
snhstance in very appreciahle quantities which can be precipitated with
sodium phosphate, and which, so prepared, is a white substance possess-
ing both emanating power and radioactivity, often many hundred-fold
greater than thoria. It has not yet been obtained in sufficiently large
quantities for an exhaustive chemical investigation, and it is impossible
at present to say what it may prove to be.-
We may at once state, however, that we do not incline to the view
that it is ThX, either in the sense of the radioactive or emanating con-
stituent of thorium. The evidence of a long series of experiments in
two directions, of which the final steps can only find place here, is quite
definite on this point, and in our opinion admits of only one conclusion.
There seems little doubt of the actual existence of a constituent ThX
to which the properties of radioactivity and emanating power of thorium
must be ascribed, but in all probability it is present in altogether
minute amount, and must therefore be possessed of these quidities to
a correspondingly intense degree.
But before the reasons for this view are put forward, it is necessary
to discuss more nearly the meaning of the experiments already given
on the emanating power. It has been shown that this is a most
uncertain quantity, similar experiments often giving preparations of
very varying value, as is clearly shown in the results given, as well as
in many others in the same direction. The most pregnant fact is that
although, as has been shown, precipitation with ammonia invariably
leaves behind considerable emanating material in the filtrate which is
lost, this seems to exert little influence on the emanating power of the
precipitates. These, prepared under different conditions, often by a
different number of precipitations, in which therefore varying amounts
of the emanating material are lost, show a surprising uniformity in
this property, especially after they have attained their maximum
power by keeping. It is only necessary to quote the experiment on the
almost completely de-emanated carbonate, which gained in emanating
power thirty times by conversion into the hydroxide, although during
the process much emanating material must have been lost, to show that
the value of the emanating power alone furnishes no criterion of the
amount of emanating material present.
It may safely be said that three things must be carefully dis-
tinguished between in considering the nature of the property possessed
by thoria of giving out a radioactive emanation. First, the nature of
the emanation itself, secondly, the nature of the emanating power,
and thirdly, the nature of the emanating material. The first, the
emanation itself, we have shown to possess the negative properties of a
chemically inert gas, whose radioactivity is unaffected by any con-
Digitized by
Google
ttill BADIOACtlVlTY OF THORItTM CX)MPOtJNDS. 1. 34t
ditions, apparently, except lapse of time. With regai^d to the seoond,
the emanating power or rate at which the emanation is produced per
anit weight of substance, it is certain that this does not depend only,
or even mainly, on the quantity of emanating material present. The
regeneration of the emanating power of thoria de-emanated by ignition,
the enormous variation with time in the emanating power of the hydr-
oxide and carbonate under certain conditions, and the comparatively
constant maximum which these substances ultimately attain, although
prepared under conditions where different amounts of the emanating
material are lost, make this point perfectly clear. These considerations,
taken in conjunction with the effect of temperature, moisture, &o., on
emanating power, and the nature of the emanation itself, make the
property appear rather as the result of a dynamical change, possibly in
the nature of a chemical reaction where the active mass of emanating
material is a constant, than as the property of a peculiar kind of
matter in the static state, additive with regard to mass.
It is, however, neither the emanation itself nor the emanating power
with which we are concerned in these experiments, but the third con-
ception, the emanating material, that is, the substance, whether thorium
or not which is responsible for the activity. It has been shown that
it is difficult to follow, by means of the value of the emanating power,
the progress of the removal of the active material. When this was
realised, attention was directed to the straight line radioactivity, which
is generally unaffected by these changes of conditions and previous
history which produce such profound alteration in the former property.
The two phenomena are undoubtedly connected. The intensely radio-
active preparations obtained from thorium in different ways always
show correspondingly great emanating power when the conditions are
favourable for the manifestation of the latter. Solution appears to be
the most generally favourable condition. The experiments we had
been engaged in were therefore repeated in a form which would allow
a close study of the total radioactivity, in the hope that this value
would prove a more suitable indication of the amount of active material
present than the emanating power alone.
Seventy grams of thorium nitrate were dissolved in four litres of
boDing water, and precipitated with ammonia added cautiously in
very dilute solution in excess. The filtrates and washings were
evaporated to about '60 o.c. and then possessed as much emanating
power as 146 grams of thoria. On evaporating the solution to dryness
and removing the ammonium salts by ignition the residue weighed
0*0583 gram. The emanating power of this residue in soltition was thus
about 2600 times that of ordinary thoria. In the solid state, how*
ever, the value fell to one-fiftieth. But its total radiation was equiva-
lent to at least 23*6 ^rams of thoria, that is, was about 400 times as
Digitized by
Google
348 RUTHERFORD AND SODDY:
great. It was diBsolved in hydrochloric acid, and ammonia added in
excess, when a precipitate weighing 0'0015 gram was thrown down.
This contained all the thorium present besides iron in appreciable
quantity which had beeen introduced during the evaporation. It
equalled in radioactivity 2*73 grams of thoria, the ratio in this case
being thus no less than 1800 times. Sodium phosphate precipitated
0*0225 gram of white substance the activity of which was equivalent to
4*4 grams of thoria^ that is, 200 times. The sodium salts freed from
ammonium still possessed a radioactivity equivalent to 3*6 grams of
thorium ozida In other experiments, however, these had been
obtained .quite free from activity, and this result is due to the solu-
bility of the phosphate in water, so that some was dissolved during
the washing (which the subsequent determination of the weight
rendered necessary) and appeared in the filtrate.
The radioactive residue obtained in the first place from the filtrate by
evaporation and ignition, before it was redissolved, had, however, been
tested to determine the penetrative power of the radiations emitted.
If the rays from various radioactive substances are made to pass
through successive layers of aluminium foil, each additional layer of
foil cuts down the radiation to a fraction of its former value, and a
curve can be plotted with the thickness of metal penetrated as
absciss», and the intensity of the rays after penetration as ordinates,
expressing at a glance the penetrative power of the rays being
examined (compare Rutherford, Phil. Mag,, 1899, [v], 47, 122)*
The curves so obtained are quite different for different radioactive
substances. The radiations from uranium, radium, thorium, each give
distinct and characteristic curves, whilst that of the last named again is
quite different from that given by the excited radioactivity produced by
the thorium emanation. The examination in this way of the pene-
trative power of the rays from the radioactive residue showed that the
radiations emitted were in every respect identical with the ordinary
thorium radiation.- In another experiment, the nature of the emana-
tion from a similar intensely active thorium-free residue was submitted
to examination. The rate of decay was quite indistinguishable from
that of the ordinary thorium emanation. That is, substances chemi-
cally free from thorium have been prepared possessing thorium radio-
activity in an intense degree.
The main quantity of thorium hydroxide in the last experiment was
redissolved in nitric acid, and the previous round of operations re-
peated twice, the filtrates from each operation being mixed and then
examined exactly as in the former case. The emanating power of the
concentrated solution was only equal to that of 8 grams of thoria in
this instance, and the radioactivity of the residue to that of 3 grams.
From this only a small quantity of the phosphate precipitate was
Digitized by
Google
THE BADIOACTIVITY OF THORIUM COMPOUNDS. I. 8*9
obtained (0*001 gram) the radioactivity of which was equal to that of
0-3 gram of thoria (ratio 200 : 1).
The emanating power of the main quantity of the hydroxide when first
so prepared was 73 per cent, that of thoria, that is, about one-half of its
usual value. The hydroxide was converted into oxide by ignition, and its
radioaetiviiy compared with that of the oxide from the original nitrate
prepared in the same way It was found to be only about one-third as
active, the exact ratio being 0*36 : 1.
Only one conclusion seems possible from this series of experiments.
There is no longer any room for doubt that a part of the radioactive
constituent ThX has been separated from thorium, and obtained in a
very concentrated form, in one instance 1800 times more powerful in
its actions. This result, taken into account with the reduction of the
radioactivity and emanating power of the main quantity of thorium
compound, and the identity of the radiations of the active thorium-free
preparations with those of the ordinary thorium radiation, warrant
the conclusion that ThX is a distinct substance, differing from thorium
in its chemical properties and so capable of separation therefrom.
The manner in which it makes its appearance, associated with each
precipitate formed in its concentrated solution, resembles the behaviour
of Orookes' UrX, which he found was dragged down by precipitates
when no question of insolubility is involved, and suggests the view
that it is really present in minute quantity. Even in the case of the
most active preparations, these probably are composed of some ThX
associated with accidental admixtures probably large in proportion.
These results receive confirmation from observations made in a
different method of separating ThX. The experiment was tried of
washing thoria with water repeatedly, and seeing if the radioactivity
was thereby affected. In this way, it was found that the filtered
washings, on concentration, deposited small amounts of material, with
an activity often of the order of a thousand times greater than that
of the original sample. In one experiment, 290 grams of thoria were
shaken for a long time with nine quantities, each of 2 litres, of distilled
water. The first washing, containing most of the sulphate already
referred to, was rejected, the rest concentrated to different stages, and
filtered at each stage. One of the residues so obtained weighed 6*4 mg.
and was equivalent in radioactivity to 11*3 grams of the original
thoria, and was therefore no less than 1800 times more radioactive.
It was examined chemically, and gave, after conversion into sulphate,
the characteristic reaction of thorium sulphate, being precipitated
from its solution in cold water by warming. N'o other substance than
thorium could he detected by chemical analysia, although of course the
quantity was too small for a minute examination. But the absence
of the substance precipitable as phosphate, noticed in the other experi-
Digitized by
Google
850 LUMSDEN: SOLUBILITIES OF THE CALCIUM
ments, confirms the opinion that this is an aooidental admixture without
influence on the qualities of radioactivity and emanating power. The
penetrative power of the radiation from this substance again estab-
lished its identity with the ordinary thorium radiation. In another
experiment, a small quantity of thoria was shaken many times
with large quantities of water. In this case, the radioactivity of the
residue was examined and found to be about 20 per cent, less radio-
active than the original sample..
There remains only one step to prove beyond doubt that the radio-
activity and emanating power of thorium are not specific properties
of the thorium molecule — the preparation of thoria free from these
properties — and on this problem we are now engaged. To sum up
briefly what has already been accomplished, two different methods
have effected a concentration of the activity many hundred-fold in one
fraction, and a corresponding diminution of activity in the remainder,
but in each case the character of the radiation is not thereby affected.
In one method, the active fraction appears to consist only of thorium,
60 far as examination has been possible, whilst in the other case
radioactivity and emanating power appear to be manifested indiscrim-
inately in all the products, without reference to their chemical
nature. The simplest explanation of this behaviour, on the present
view, is that so far the active constituent of thorium has only been
obtained in relatively minute quantity, and therefore does not answer
to any definite analytical reations.
Maodonald Physics Buildino.
Macdonald Chxmistrt and Mining Buildino.
McGiLL Univxbsitt, Montbbai..
XXXI V. — Solubilities of the Calcium Salts of the
Acids of the Acetic Series.
By John S. Lumsden, D.Sc, Ph.D.
It is well known that many calcium salts are more soluble in water at
low than at high temperatures. This is not in accordance with the usual
experience that the solubility of a solid increases with rise of tempera-
ture, yet the matter has never been submitted to careful investigation.
For the purpose of obtaining more information on the subject, it
seemed best to select a series of adds, to prepare the calcium salts in
a crystalline state, and to determine the solubility of each salt for a range
of temperature between (P and 100^.
The acetic series of organic acids was chosen and the calcium salts
Digitized by
Google
SALTS OF THE ACIDS OF THE ACETIC SERIES. 351
of the first nine normal acids and of tsobutyric and isovaleric acids
hare been prepared and examined.
In order to obtain accurate resaltB, the estimation of the solubility
of a substance must be done with very great care. The difficulties to
be overcome are : the maintenance of a constant temperature for
many hours, the production of complete saturation and the removal of
the solution from contact with the solid to another vessel without
change or loss.
The apparatus used to obtain a constant temperature is shown in
Fig. 1 (p. 352). The thermostat is an enamelled iron vessel supported
on bricks and heated from below by a luminous gas jet and a ring air
burner.
The temperature regulator consists of two bottles filled with water
or calcium chloride solution, one of which is shown at A connected
with an ordinary mercury gas cut-off, B. By means of Y glass con-
nections', the gas flame, the ring burner, and the fan jet, G, are supplied
with gas from the tap, D,
For high temperatures, the luminous gas flame connected with the
regulator does not produce sufficient heat and the ring burner is then
used, being set to keep up a constant temperature less than that required
and the gas flame then raises the temperature to the exact point and
regulates it.
The water level is kept constant by the syphon, E, When the
water sinks below the bell-mouthed tube, air enters and water flows
from the bottle until the level rises and shuts off the supply of air.
The stirrer, F, is mounted on a steel point running in a hole in a
glass stopper and moved by a wheel made of Japanese fans. At low
temperatures, the fan jet is required, but at high temperatures the
heated air from the apparatus causes a sufficiently rapid rotation.
The thermometer used reads to tenths of a degree, and for tempera-
tures between 20° and 90° the water in the thermostat can be kept
without variation of more than one^fifth of a degree for many hours.
Below 20°, constant temperatures were obtained by running a con-
tinual stream of water from the laboratory tap into the vessel. On
different days the temperature varied, but on the same day it was
sufficiently constant for six hours. Temperatures down to 6° were got
in this way. For estimations at the freezing point a mixture of ice
and water was used and boiling water was employed for the determin-
ations at 100°.
The apparatus used to obtain complete saturation of the solution
and to remove the saturated liquid is shown in Fig. 2 (p. 353), and on
a small scale in Fig. 1 (p. 352).
The object in view when this apparatus was devised was to carry out
all the operations under the surface of the water in the thermostat and
therefore exactly at the temperature of estimation.
Digitized by VjOOQIC
352 LUHSDEN: SOLUBILITIES OF THE CALCIUM
The Baturating vessel, &, is a bottle or small flask into which <
of finely powdered solid is placed along with the solvent. A glass stirrer
running in a glass collar and passing down the outer tube, H^ is driven by
Fio. 1.
a water motor as shown in Fig. 1. JT is the flask or bottle into which
the saturated liquid is to be filtered. The bulb, Z, is the filter and con-
tains a very small filter paper filled with cotton wool, which keeps the
Digitized by
Google
SALTS OF THE ACIDS OF THE ACETIC SERIES. 363
paper in position and also acts as a filter plug. Connection is made
with the saturating vessel by the tube i/, and the tube iT, rising above
the water level, is the suction tube. The filter flask is not attached to
the saturating vessel until a short time before filtering, since at high
temperatures water is liable to distil through the tube M, The end
of that tube is therefore plugged while saturation is taking place.
All the stoppers are of rubber except that which bears the neck of the
stirrer^ and this has a notch to prevent increase of pressure in the
bottle.
After saturation has been attained, which at low temperatures is
only after many hours but at high temperatures is complete in an
hour, the filter flask is attached, ten minuted
are allowed for heating to the temperature Fio. 2.
of the bath, then suction is applied by the
mouth to iV, and the liquid passes over and
filters quite clear into the flask. As much
as possible of the liquid is filtered, to reduce
any error due to absorption by the filter or
moisture in the connecting tube.
After filtration, the flask is quickly de-
tached, corked and immediately cooled to
the temperature of the laboratory by running
water on the outside. This quick cooling
prevents evaporation taking place and con-
sequent concentration of the solution.
It is necessary to know the exact com-
position of the solid of which a saturated
solution is being made, as the solid may
undergo change. An anhydrous salt, when
placed in water, may become a crystalline
hydrate and a salt containing water of
crystallisation may lose water in contact
with a hot solution. As a salt and its hydrates are essentially different
substances, they have different solubilities, and analysis is necessary to
determine the substance being dealt with.
The apparatus used for separating the solid at any temperature from
the saturated solution is seen in Fig. 3 (p. 354). The tube 0, used as
the saturating vessel, has a rubber stopper fitted to the lower end, and
through a hole in this passes a glass tube which makes connection
with the filter bottle. A rubber connection with a screw clip keeps
the tube closed until filtration is to take place. On the rubber stopper
rest two perforated porcelain plates with a filter paper between.
The Bolid is placed in the tube, the solvent added, and saturation
attained at the desired temperature. The clip is then opened, the tube
VOL. LXXXI. B B
Digitized byVjOOQlC
854 LUHSDEN: SOLUBILITIES OP THIS CALCIUM
P attached to a filter pump, and while suction is going on the solid id
rammed down tightly to express as much solvent as possible. The
apparatus is then quickly removed from the thermostat and the solid
placed on a warmed porous plate. Drying is speedily effected, and after
an hour in a desiccator the solid is analysed before any change can
take place. With all precautions, it is difficult in many cases to get
exact quantitative results : either the solid is insufficiently dried or water
of crystallisation is lost, but there is no difficulty in getting results
which shofv decidedly whether the solid is one or other of two hydrates.
The saturated solutions obtained in these determinations and the
solids containing calcium had to be analysed, and this is a tedious and
difficult process unless the following method is
Fio. 8. adopted.
It was found most exact to convert the
calcium compounds always into calcium sul-
phate. From 5 to 10 grams of the solution
were placed in a platinum crucible and weighed,
then somewhat less than the quantity of con-
centrated sulphuric acid required to complete
the change to calcium sulphate was added.
When the crucible was heated on a water-bath,
the organic acid rapidly evaporated^ and a hard
residue was obtained in an hour. This was
carefully ignited, and the dried mass, consist-
ing of oxide, sulphide, and sulphate of calcium,
was treated again with concentrated acid and
heated until fumes ceased to be evolved and
the residue was pure white.
The calcium salts employed in the following
determinations were prepared from the pure
acids obtained from Kahlbaum and from pre-
cipitated calcium carbonate. A quantity of the acid was largely
diluted with water, enough water being used in the case of the higher
acids, which are only slightly soluble, to form a complete solution, and
this was then poured on excess of calcium carbonate contained in a
large flask. After the first action ceased, heat was applied for some
time, and the liquid was then filtered.
This filtered solution was always slightly acid, and it was found that
when organic acids become very dilute they cease to act on calcium
carbonate. Since all the acids are volatile, the acidity disappeared
during evaporation of the liquid. The solid obtained by evaporation
over a water-bath was dissolved in water, and crystals obtained by
placing the vessel in a warm place, and occasionally removing and
drying those which peparated.
Digitized by
Google
SALTS OF the: ACIDS OF TfiE ACETIC SERIES. 35&
Most of these calcium salts are soluble in hot dilute alcohol, and it
was found easiest to obtain good crystals of the higher members from
such a solution.
SoluhUity qf CaleiuM, FormaU, (HC02)2Ca. Curve No. 1.
Calcium formate crystallises from solution in water or dilute alcohol
in anhydrous, hard, glass-like, rhombic crystals. Attempts to obtain
a hydrated salt by spontaneous evaporation of a solution in water at a
low temperature were unsuccessful :
0-9616 crystals gave 1 -0021 CaSO^. Oa = 307.
(CH:02),Ca requires Ca = 31-2 per cent.
The solubility curve of calcium formate between 0° and 100° is a
straight, upwardly inclined line, representing a steadily increasing
solubility with rise of temperature.
The weights of solid dissolved by 100 parts by weight of water are
as follows :
t
Porto.
t
Parts.
t.
Parts.
0°
1615
40°
17-05
80°
17-95
10
16-37
50
17-27
90
18-17
20
16-60
60
17-50
100
18-40
30
16-82
70
17-72
Formula for this range of temperature, 16*15 + 0-0225t°.
SohibUiiy qf Calcium Acetate, (CH8'C02)2Ca,2H:20 <md
(OH3-C02)2Ca,H20. Cm^e No, 2.
From a cold solution, calcium acetate crystallises in long, silky,
transparent needles which contain 2 mols. of water of crystallisatioD.
These effloresce readily, and the ordinary white solid contains only
1 mol. of water:
1-066 gave 0-8230 CaSO^. Ca« 22'71.
(C2H:j02)2Ca,Hp requires Ca = 22-73 per cent.
In solution, the change from the salt with 2H2O to that with IHjO
takes place at 84°.
The solubility at 0° is 37*40 parts in 100 parts of water; it then
diminishes until about 60°, when the solution contains only 32-70 parts,
then rises until the transition point is reached at 84°, when 33 '80
parts are dissolved. From 84°, the solubility of the new salt diminishes
rapidly to 100°, when the solution contains 29-65 parts.
The weights of solid, calculated as anhydrous salt, dissolved by 100
parts of water, are as follows :
B B 2
Digitized by
Google
356
LtTHSDEK: SOLTTBILITIES OF THE CALCIUM
(2H,0)
t.
Parts.
t.
Ports.
t.
PMto.
0°
37-40
(2B,0)
35°
• 83-50
(2H,0) 70°
32-98
6
36-65
40
33-22
75
33-22
10
35-98
45
33-00
80
33-60
15
35-32
50
32-82
T84
33-80
20
34-73
55
32-70
(HjO) 85
32-85
25
34-20
60
32-70
90
31-05
30
33-82
65
32-78
95
100
30-20
29-65
Solubility qf Calcium Fropumate, (CH3*CHj*COj)jCa,HjO.
Curve No. 3.
The crystals of calcium propionate are thin, glistening plates con-
taining 1 mol. of water of crystallisation^ which is only given off
above 100°
The solubility curve is markedly convex to the temperature axis.
At O^y as much as 42*80 parts of the salt dissolve in 100 parts of water,
but the solubility quickly diminishes until, at about 55°, only 38'20
parts are in solution. Above this temperature, the solubility in-
creases and at 100° 48*44 parts are dissolved.
To make sure that the solid in contact with, the solution was the
same on the descending and ascending parts of the curve, it was fil-
tered off at various temperatures and analysed. The results showed
that at all temperatures the solid contained 1 mol. of water :
At 12°, 0*1739 gave 0*1159 CaSO^. Ca- 19*61.
,,90°, 0*3895 „ 0*2591 CaSO^. Ca- 19*57
(Ca3H502)20a,HjO requires Ca— 19*61 per cent.
The weights of solid dissolved by 100 parts of water are as follows :
t.
Parts.
t.
Parts.
t
Parts.
0°
42-80
35°
38-75
70°
38-75
6
41-70
40
38-45
75
39-20
10
40-95
45
38-36
80
39-85
15
40-35
50
38-25
85
40-80
20
39-85
55
38-20
90
42-15
25
39-40
60
38-25
95
44-25
30
39-05
65
38-40
100
48-44
SohMlUy of Calcium ButyraU, [CHs-(CHj)2-COj20a,H20.
Curve No, 4.
Calcium butyrate crystallises by spontaneous evaporation in long,
ribbon*like leaves containing 1 mol. of water, and this is the compo-
Digitized by
Google
SALTS OF THE ACIDS OF THE ACETIC SERIES.
357
sition of the solid in contact with the solution at all temperatures
below 100^ :
0-4940 gave 0-2898 CaSO^. Oa = 17-26.
(Q^ILf02)fiAyB,fi requires Ca = 17*24 per cent.
The solubility of this substance has been determined by Hecht
{AnnaUn, 1882, 213, 65) and by Chancel and Parmentier (Compt. rand,
1887, 104, 474). Their results are slightly lower than those given
below.
The weight, calculated as anhydrous salt,* in 100 parts of water is
as follows :
t
Parts.
t.
PartB.
t
Parts.
0°
20-31
36°
16-70
70°
14-92
5
19-76
40
16-40
75
14-90
10.
19-15
46
16-00
80
14-96
15
18-65
60
15-70
85
15-10
20
18-20
56
15-40
90
15-25
25
17-75
60
15-15
95
15-50
30
17-26
66
15-00
100
15-86
SdvbUity qf Calcium Valerate, [OH3-{OH2)8-C02]aCa,HjO.
Curve No. 5.
Calcium valerate is obtained from solution in hot water as a micro-
crystalline powder, but if a cold solution is allowed to evaporate, fine,
interwoven crystals are obtained, and from dilute alcohol long, ribbon-
like plates readily separate out.
One mol. of water is always present below 100° :
0-4219 crystals gave 02210 CaSO^. Ca = 15*41.
(C5HjOg)2Ca,H20 requires Ca= 15-38 per cent.
The solubility for each 10° is as follows :
t.
Parts.
t.
Parts.
t.
Parts.
0°
9-82
50°
7 85
70°
7-80
10
9-26
56
7-76
80
7-95
20
8-80
67
7-75
90
8-20
30
8-40
60
7-78
100
8-78
40
806
The point of lowest solubility is 57°, when only 775 parts of solid
are in solution,
Digitized by
Google
358
lumsden: solubilities of the calcium
Solubility qf Calcifim Caproate (HexocUe), [CH8-(CH2)4-COj2Ca,H,0.3
Curve No, 6.
This salt separates from water in small, leafy crystals, but long,
thin plates are obtained from hot dilute alcohol. One mol. of water is
present up to 100° :
0-4580 gave 02162 CaSO^. Ca « 13-89.
(CgHjiOj)2Ca,HjjO requires Ca= 13-89 per cent.
The solubility curve, although very flat, distinctly shows a descending
and an ascending part, the point of lowest solubility being about 55°.
The solubility for each 10° is as follows :
t.
Parts.
t.
Parte.
t
Parte.
0°
2-23
40°
215
80°
2-30
10
2-20
50
210
90
2-45
20
218
60
215
100
2-57
30
217
70
2-20
SduhUity of Calcium (Enant/iate (Heptoate), [CB^*{CU^)^'CO^]^CA,lIfi.
Curve No. 7.
The crystals from dilute alcohol are long, monoclinic prisms forming
a fibrous, silky mass. They contain 1 mol. of water :
0098 gave 00424 OaSO^. Ca - 1352.
(C^Hi802)2Ca,H20 requires Oa- 12-88 per cent.
The solubility is as follows :
t.
Part.
t
Part.
t.
Part.
0°
0-95
40°
0-82
80°
0-98
10
0-90
50
0-80
90
110
20
0-86
60
0-82
100
1-26
30
0-84
70
0-90
Solubility of Calcium Caproate (Octoate)^ \(M^'{QB^^*Q0^J^2^,ILf>.
Curve No. 8.
As caproic and pelargonic acids are very slightly soluble in water, the
calcium salts were prepared by first making the ammonium salts and
then adding calcium chloride solution. The precipitated salts were
well washed and crystallised from dilute alcohol.
The crystals of calcium caproate are long, thin rhombic needles con-
taining 1 mol. of water.
0-3420 gave 0-1350 CaSO^. Oa- 11-61.
{CgHi502)2Ca,H,0 requires Ca« U-62 per cent.
Digitized by
Google
SALTS OF THE ACIDS OF THE ACETIC SERIES.
359
The greatest solubility is at 100^, when 0'5 part dissolves in 100
parts of water; the lowest is about 60°, where 0*24 part is in
solution :
t.
Part.
t.
P«rt.
t.
Part.
0°
0-33
40°
0-28
80°
0-32
10
0-32
60
0-26
90
0*40
20
0-31
60
0-24
100
0-60
30
0-30
70
0-28
Solubility of Caleium PdargonaU {NonoaU), [CH8-(CH2)7-C02]20a,HjO.
Cwrve No. 9.
Calcium pelargonate forms long, transparent, leafy crystals which,
massed together, resemble white satin in lustre. The crystals are
readily obtained from solution in hot dilute alcohol. One mol. of
water is present :
0-3208 gave 01 170 CaSO^. Ca= 10-73.
{Q^yfi^juSL,TLjd requires Ca = 10*75 per cent.
The solubility is now very small :
t
0°
10
20
30
Part.
0-16
015
014
014
t.
40°
50
60
70
Part.
0-13
013
012
0-12
t.
80°
90
100
Part.
015
0-18
0-26
SolubUUy of Calcium isoButyrate, [(CHj)2CH-COj]2Ca,5H,0 and
[(CH3)jCH-C02]2Ca,H20. Curve No. 10.
From solution at low temperatures, calcium t«obutyrate crystallises
in long, thick, prismatic needles containing 5 mols. of water of crys-
tallisation :
0 5648 gave 02516 OaSO^. Oa = 13-10.
{QJl>j0^j:^AfiB.fi requires Ca« 13-16 per cent.
The crystals filtered from a saturated solution above 80° contain
1 mol. of' water :
0-6062 gave 0-2964 CaSO^. Ca« 17-22.
{Q^^O^j:^^,B.p requires Ca= 17*24 per cent.
The transitioD point from one salt to the other is at 62-5°, and the
solubility is represented by two curves.
The weights of anhydrous salt dissolved by 100 parts by weight of
lyater are :
Digitized by
Google
360
(5H,0)
LtTMSDEN :
SOLUBILITIES OF THE CALCIUM
t.
Parts.
t.
PatU.
t.
Parte.
0°
20-10
(5H,0) 36°
24-55
(HjO) 65°
28-25
6
20-52
40
25-28
70
27-76
10
2110
46
2605
75
27-32
16
21-70
50
26-80
80
27-00
20
22-40
65
27-60
85
26-70
25
23-10
60
28-40
90
26-48
30
23-80
T62
28-70
96
100
26-28
26-10
The solubility of calcium t>obutyrate was determined in 1887 by
Chancel and Parmentier {Campt. rend., 1887, 104, 474). At low
temperatures, their values approximate to those now found, but
although they knew that above 80^ the salt contained only 1 mol.
of water, yet they show no transition point, but represent the
solubility by a simple curve, concave to the temperature axis. Possibly
by joining a few distant points, they missed the transition point.
Sohibilily qfCcUcium isoValerate, [(CHj)2CH-CH2-C02]2Ca,3H,0 and
[(CH3)2CH-CH2-C02l2Ca,H20. Curve No. 11.
Calcium taovalerate crystallises in two forms : in long, thick, well-
formed, pnsmatic needles containing 3 mols. of water, and from a
hot solution in thin plates containing 1 mol. of water :
0-2796 gave 0-1289 CaSO^. Ca= 13*56.
(C5Ha02)2Ca,3F20 requires Ca= 13*52 per cent.
The crystals filtered from a saturated solution above 80^ were
pressed on a porous plate and placed in a desiccator until they showed
signs of efflorescence :
0-8768 gave 04626 CaSO^. Ca = 15-53.
(C5F^Oj)gCa,H20 requires Ca = 15-38 per cent.
The solubility is shown by two well-marked descending curves,
the transition point being at 45 '5^.
The weights of anhydrous solid in 100 parts of* water at different
temperatures are as follows :
t.
Parts.
t.
Parts.
t.
Parts.
3H,0 0°
26-05
(3HjO) 40°
22-00
(H,0) 70°
17-40
5
23-75
45
22-30
75
1710
10
22-70
T45-5
22-35
80
16-88
15
2215
(Hfi) 60
19-96
85
16-75
20
21-80
55
19-00
90
16-65
25
21-68
60
18-38
96
16-68
30
21-68
65
17-85
100
16-66
36
21-80
1
Digitized by VjC
)OQle
SA.LTS OF THE ACIDS OF THE ACETIC SERIES. 861
Fio. 4.
Solubility cwrves of the calcium adUs of acida <if the acetic eerieg.
Digitized by VjOOQIC
362 CALCIUM SALTS OP THE ACIDS OF THB ACETIC SERIES.
The examination of the collected solubility curves (Fig. 4, p. 361)
shows the following points : — Calcium formate, the salt of the lowest acid
of the series, is peculiar ; it consists of anhydrous crystals, and shows a
simple ascending curve of solubility. All the other calcium salts
crystallise with water, and all with rise of temperature first diminish
in solubility then reach a minimum point, and thereafter the solubility
steadily increases.
Of the salts of the normal acids, only calcium acetate changes from
one crystalline state to another between 0* and 100^ while both
calcium ifobutyrate and calcium ivovalerate have double curves. It
will also be noticed that, with the exception of calcium formate, all the
salts which have been investigated, when in contact with their
saturated solutions at 100°, consist of crystals which contain 1 moL of
water.
Whilst it is difficult to compare a series of salts with respect to
solubility, since one may be more soluble than another at one
temperature, but less soluble at a different temperature, and different
hydrates are not truly comparable, yet in a general sense it may be
said of the calcium salts under consideration that those formed from
the normal acids increase in solubility from formate to acetate and
propionate, then decrease quickly with the growth in the number of
carbon atoms, and that the salts of the t»o-acids are more soluble than
those of the corresponding normal members of the series.
As the solubilities of these calcium salts with rise of temperature
diminish, reach a minimum, and then increase, each curve is convex to
the temperature axis, and it will be shown in the following paper that
this is the normal shape of a solubility curve. Calcium salts will be
found to be in no wise anomalous in diminishing in solubility with rise
of temperature ; they are simply peculiar in having the descending
parts of their curves within the range of temperature between 0° and
100°, whilst the curves of most other solids are the ascending parts of
convex curves, which would show a minimum and descending part if
the determination of the solubility could be made at a low enough
temperature.
UNIVBRSlTy COLLROE, DUNDKK.
St. Andrew's Uniyeiwity.
Digitized by
Google
THB EQUILIBRIUM BETWEEN A SOLID AND ITS SOLUTION. 63
XXXV. — The EquiUhnv/m between a Solid and its
Saturated Solution at Various Temperatures.
By John S. Luhsden, D.Sc., Ph.D.
In the foregoing paper, it was shown that the solubilities of the calcium
salts of the acids of the acetic acid series were represented by curves
convex to the temperature axis, indicating at first decrease then increase
of solubility with rise of temperature.
^As these salts, while undergoing no alteration in composition, yet
change iii solubility in such a way that at widely different temperatures
the action of the solvent produces solutions of equal concentration, the
factors which condition the equilibrium between the solid and the
saturated solution must undergo great variation with change of
temperature.
What these factors are and how their values alter will be considered
here, with the object of obtaining some reason for the shape of a
solubility curve.
Calcium propionate furnishes a typical convex curve of solubility,
and has been employed to obtain the experimental data used in this
paper.
The Factors whiohproduee Equilibrium in a Saturated SohUian.
A solid may be considered to be made up of particles which by their
thermal energy tend to separate, but which cohere because the attrac-
tion due to their mass, acting towards the interior, aided by the pres-
sure of the atmosphere, counteracts the externally directed force.
In a solid which is volatile at the ordinary temperature, the opposing
forces are nearly equal, but in most solids the tendency to ,pass into
vapour is much more than counterbalanced by the force binding the
particles together.
When, however, a solid is placed in a liquid, at the surface of contact
an action between solid and liquid takes place of such a nature that
the outwardly directed force is helped. This attraction, if strong
enough, enables the particles of the solid to pass into the liquid and so
form a solution, but if the attractive infiuence of the liquid is in-
sufficient, the solid will not dissolve. «
When solution does take place, the solid while being dissolved
becomes subjected to a gradually increasing pressure as the concen-
tration of the solution grows, due to the impact of the dissolved
particles on its surface. This pressure, the osmotic pressure, acts
against the forces promoting solution, and as the solid continues to
pass into the liquid and this pressure increases, the disintegrating
Digitized by
Google
364 LUHSDEN: THE KQU I LIBRIUM BETWEEN A SOLID AND ITS
action of the liquid on the solid lessens, the tendency of the particles
to leave the solid is diminished, and finally oesises.
This is not now a state of rest but of equilibrium, fcMr if at any point
a particle passes from the solid into solution, a similar particle will
leave the liquid and cohere to the solid. The solution has attained a
definite concentration, it is saturated, and excess of solid can have no
influence on this final state since the process of solution goes on only
at the surface of the solid and the counteracting osmotic pressure is the
same there at all points.
The equilibrium in a saturated solution in contact with the solid is
therefore conditioned by three forces, two of which promote sdution :
the thermal energy of the solid tending to drive its particles %)>art, and
the action between solid and solvent and one acting against and
balancing these two : the osmotic pressure of the dissolved particles.
Alteration in the value of any one of these factors will alter the
amount of solid dissolved, and change of temperature or pressure has
that effect.
Change of Equilibrium v)itk Change of Temperature.
The alteration produced on a saturated solution in contact with un-
dissolved solid by change of temperature is best realised by considering
the effect of change of temperature on each of the three forces which
together produce equilibrium.
The thermal energy of the solid must always be increased by heat ;
work is performed against the cohesion of the particles, and their kinetic
energy is augmented. This growth of disintegrating force may be pro-
portional to rise of temperature, but it is known toincreiseat agreater
rate than the temperature in the case of hydrated salts, especially as
the point at which dehydration takes place is approached. Increase of •
temperature will, therefore, always raise the value of this factor and
tend to facilitate solution.
The pressure exerted by the particles in solution on the undissolved
solid is also increased by heat. For a constant concentration, the in-
crease of osmotic pressure is directly proportional to the absolule
temperature, but according to Nernst, with solutions of very great con-
centration the increase of pressure is at a somewhat faster rate.
We have, however, to consider solutions in which the concentration
is not constant but changes slightly with the temperature, and it is
obvious that the pressure will increase less rapidly than the law
demands when the solid diminishes in solubility with heat, and more
rapidly when the solubility increases with rise of temperature. It is
also evident that expansion of the solution by diminishing the con*
centration will also lessen the osmotic pressure,
Digitized by
Google
SATlTRAtED SOLUTIOK AT VARIOUS TEMPERATURES. 365
Allowing, however, for these small disturbing influences^ increase of
temperature will always increase the pressure on the solid and retard
its solution.
The action at the surface of solid and solvent is probably chemical
in its nature, therefore the force of attraction will diminish as the
temperature rises, and the effect of heat will lessen the value of this
solubility factor.
Equilibrium is therefore the result of a force tending to produce
solution increased by rise of temperature, a force retarding solution
also increased by rise of temperature, and a force promoting solution
decreased by rise of temperature.
The resultant action of these three forces determines the amount of
solid in solution, and a curve which represents the varying weights of
solid, dissolved in a definite weight of -liquid, saturated at different
temperatures, will accurately represent the resultant effect of varying
the temperature of the system.
Such a curve cannot be made directly, because the total amount of
a solid which has passed into solution cannot be accurately found by
experiment. What is really estimated is, either the weight of solid in
a given volume or in a given weight of the saturated solution. In the
first case, the determination at different temperatures of the concen
tration or the weight of solid in a given volume of the solution : values
are not obtained proportional to the total weight of solid in solution
imless in the hypothetical case of a solid which dissolves in a given
volume of the solvent without that volume undergoingany change, and
no change of bulk takes place with rise of temperature. Now, since
a saturated solution has a greater volume than the solvent, and there
is always an expansion by heat, with rise of temperature there will be
proportionally greater amounts of substance dissolved than a curve of
concentration will show.
In the seoond case, the estimation at different temperatures of the
weights of solid in a given weight of the solution : the numbers obtained
do not bear any simple relation to the amount of solid dissolved, but if
from the figures obtained the weights of solid dissolved by a given
weight of the solvent be calculated, then values so found are accurately
proportional to the total weight dissolved at different temperatures.
This is clear from the following considerations. If a solid be placed
in 10 grams of a solvent, then at any temperature when the solution
has become saturated, the 10 grams, no matter what change in volume
has taken place, contain all the dissolved solid. One gram or any
other weight of the solvent will then always contain a quantity pro-
portional to the total amount dissolved, and even if during the
experiment some of the solvent evaporates, yet the result is not
Digitized by
Google
366 LUMSDEN: THE EQUILIBEUUH BETVTEEN A SOLID AND ITS
altered, since the oonoentration of the solution at aay temperature is
constant.
The ordinary soluhility curve showing the weights of solid dissolved
at different temperatures hy 100 parts by weight of solvent is therefore
the exact representation of the influence of temperature on the
equilibrium between the solid and its saturated solution.
The solubility curve being the resultant of the forces producing
equilibrium, it may be thought of as resolved into the component
Fig. 1. — Thti component factors of a sotubilUy curve.
Temperature •
curves representing the actions of each of the forces^ and it will be
seen how these severally influence the shape.
On Fig. 1 is seen the solubility curve of calcium propionate and
lines representing the actions of the three factors into which it might
be resolved. Above the central horizontal line the values denote
influences promoting solution, and below the line retarding forces.
The component due to the growth of thermal energy will be an
upwardly inclined straight line if the increase is directly proportional
to rise of temperature, and an upwardly inclined curve if the increase
Digitized by
Google
• SATURATED SOLUTION AT VARIOUS TESCPERATURES. 367.
is faster than the temperature. In the figure, a curve of the latter
kind is shown.
The retarding action of the osmotic pressure may with fairness be
represented by a downwardly directed line, and the affinity between
solid and solvent is drawn as a line with a high value at a low
temperature, decreasing in value as the temperature rises. The
resultant of the affinity and osmotic pressure is shown by a dotted
line, and the total resultant is the solubility curve.
A diagram of this kind is of great interest : by altering the values
of the factors, all varieties of convex curves can be obtained, even
one flattened to a straight line, but if the inference be true that the
affinity decreases and the thermal energy and osmotic pressure increase
by rise of temperature, no solubility curve concave to the temperature
axis is possible.
The Rdcaumtihip between the Heat qfSoltUian cmd the Shape of a
SclvJbilUy Cwrve,
When a given weight of solid is dissolved in a definite weight of
solvent, heat is evolved or absorbed, or under certain circumstances
there may be no change of temperature.
If the action between solid and solvent produces sensible heat, to
tlat heat there must be added the heat rendered latent by the fusion
of the solid and the heat of dispersion of the dissolved particles, in
order to get the total heat due to the action of the solvent on the
solid. Should the heat • rendered latent be greater than the heat
developed by combination, solution will cause a decrease of temperature,
and if these* two thermal effects are balanced, no change of temperature
will occur.
> The heat effect must, however, vary with the temperature at which
solution takes place, for, remembering that forces which condition a
development of heat are weakened by an increase of temperature, and
those which cause an absorption of heat are strengthened, it is obvious
that, with rise of temperature, the chemical action which is the cause
of the heat will be- weakened, whilst the physical actions will be
facilitated.
At a higher temperature, therefore, the heat required to fuse the
solid, and the amount of heat absorbed by the distribution of the
particles^ will be lessened, but if the chemical action between solid
and solvent is diminished to a much greater degree, it becomes
possible with the same substance to have heat evolved at a low
temperature daring solution, and at a higher temperature heat
absorbed. Thus, if a given weight of solid and a definite weight of
solvent be brought to the same temperature and mixed, heat being
Digitized by
Google
368 LtJMSDEK: THE EQUIUBRIUM BETWEEN A 80UD AKD ITS
evolved, at a higher temperature the physical changes may absorb all
the heat due to combination, and there will be no alteration of
temperature ; higher still, the physical changes will require more heat
than the combination gives, and a fall of temperature results, and it
is not unreasonable to suppose that at a temperature where there is
no action between solid and solvent, solution may take place by the
thermal energy being so increased' that the solid volatilises into the
solvent as camphor does into air.
To test experimentally whether the heat of solution diminishes with
rise of temperature, two estimations were made with calcium propionate
at temperatures where it is equally soluble, one being on the descending
part of the solubility curve at 10°, the other on the ascending curve
at 85° .
Ten grams of crystals of calcium propionate containing 1 mol. of
water were finely powdered and placed for 42 hours beside a vessel
containing 26 c.c. of water. The powder was then dissolved, and the
temperature rose from 10*2° to 19°. Complete solution occupied two
minutes, and 0*4° was added on for loss by cooling, making the total
increase 9*2°.
Next, 10 grams of the same solid were placed in a thin glass bulb
with a long stem, the bulb being weighted with a little mercury.
This bulb was lowered into a large test-tube containing 27 o.c. of
water, and the temperature in the calorimeter raised to 85°. When
the water had diminished to 25 c.c, the bulb was broken, and not
more than an increase of 1° was observed. It is so difficult to get
the correct heat of solution at a high temperature where the liquid has
to be stirred and evaporation takes place, that exact results were
impossible to obtain, but the experiment makes it sufficiently clear
that the heat of solution decreases as the temperature rises, and that a
negative value might even be observed at a higher part of the curve.
When a solution is nearly saturated^ the heat developed by dis-
solving a given quantity of solid in it is much less than in a more
dilute solution. The thermal effect of dissolving a given weight of
solid in a given weight of water, and secondly in a nearly saturated
solution of the solid made with the same weight" of solvent, cannot be
the same. In the latter case, although the heat developed by the
chemical action will be the same, yet because of the osmotic pressure,
the particle passing into solution will absorb more heat.
The following experiments show this clearly :
Four portions of 25 c.c. of water were taken, and 2|, 6, 7 J, and 10
grams of calcium propionate placed beside them. After standing
many hours at 12*4°, the several portions were mixed with the water,
and the alterations of temperature noted with a delicate thermometer.
The increases in temperature were respectively 3*6°, 6'2°, 7*6°, and
Digitized by
Google
SATimAtfiD SOLUTION- At VAtllOUS TBMPERATURBS. 369
8*8^. These values, as was expected, are not proportional to the
weights of solid dissolved.
The heat changes produced when the same weight of solid was
dissolved in solutiops of different concentrations made with the same
weight of solvent were next determined.
Twentj-five grams of water and solutions, made hy dissolving in
25 grams of water 2^, 5, 7^, and 10 grams of calcium propionate
respectively, were prepared. When all the materials were cooled to
12*4°, 2^ grams of solid were added to each vessel. Increases of 3*5^,
2% 1-4^, 0*9°, and 0*6° were obtained. These results are shown
graphically on Fig. 2.
Thus the more concentrated the solution, the less the amount of
sensible heat evolved when a given weight of solid is dissolved in it,
Fio. 2.— ir«rf ofsoltUian qfcaleit
\m propionate.
/
/
A
f
\
/
-.
10 20 30 40 «0 10-*' lO 204-10
Parts dissolved in 100 parts toater.
30^10 40^10
and if solution can be continued so that a supersaturated solution is
produced, absorption of heat may take place. Thus Reicher and
Deventer {Zeit. pkysikaL Chem.^ 1890, 6, 559) state that cupric chloride
dissolves in water with evolution of heat, but that in a completely
saturated solution it dissolves with reduction of temperature.
This heat of solution in a saturated solution can be experimentally
found, for if a supersaturated solution be made at a certain tempera-
ture, changed to another temperature, and the excess of solid separated
out by dropping in a crystal, the amount of heat absorbed by a solu-
tion supersaturated at a low temperature, or the amount of heat evolved
from a solution supersaturated at a high temperature, will be the same
38 that evolved or absorbed respectively by the solution of the separated
solid if the solution could take place in a saturated solution.
VOL. LXXXI. O O
Digitized by VjOOQIC
370 LUMSDEN: THS EQUILIBRIXTH BETWEEN A SOLID AND ITS
The solubility curve represents only the heat changes which will be
produced by dissolving a particle of solid in a solution nearly saturated.
Considering a curve like that of calcium propionate, and starting
from the point of least solubility, if the temperature is raised the
solution becomes unsaturated, solid passes into solution, and heat is
absorbed ; if the temperature is lowered, solid also passes into solution
and again heat is absorbed. A downward curve will therefore
indicate evolution, and an upward curve absorption, of heat when a
particle of solid finally saturates the solution.
When, however, a solution is dilute, even if the temperature is above
that of the least solubility, the addition of solid still causes an evolu-
tion of heat. The statement which is generally made that if a solid
when placed in water dissolves with evolution of heat it will diminish
in solubility with rise of temperature, is only partially true ; the solu-
bility curve indicates only the sign of the heat of solution in a nearly
saturated solution.
The Influence of Pressure an the Equilihrium in a Saturated Solution.
When pressure is applied to a saturated solution in contact with the
solid, the effect will be different according as the total volume of solid
and liquid is greater or less after solution than before.
If the solid plue the solution has a greater volume than the solid
plus the solvent, that is, if there is an expansion during solution, ex-
ternal pressure tending to diminish the volume increases the osmotic
pressure by increasing the concentration in a greater ratio than the
thermal energy of the solid is increased by compression, particles return
to the solid and the amount of substance in solution is diminished.
If, however, the solid plus the solution has a smaller volume than
the solid and solvent, the external pressure tending to diminish the
volume helps the solution of the solid and the concentration of the
liquid increases.
Thus the pressure helps most the forces which tend to cause a dimin*
ntion of volume and alters the equilibrium in the direction in which
these forces act.
Experimental proofs of these facts have been given by Sorby {Proe,
R<yy. Soc., 1863, 12, 538) and Braun {Ann^Pkye. Chem., 1887, SO, 250),
and in addition Braun showed from thermodynamical considerations
that with the same pressure the sign of the heat of solution would
influence the amount of solid dissolved.
When, therefore, the solubility curve is convex and the heat of
solution in a saturated solution changes from positive to negative with
rise of temperature, the effect of pressure considered apart from any
change of volume will be different at different parts of the ourvei
Digitized by
Google
SATtTRATBD SOLtJTK)N AT VARIOUS TEMPERATtJBES. 37l
No data are available for the volumes of solid and saturated solution
of a substance like calcium propionate at various temperatures, and the
true change of equilibrium by change of pressure cannot be followed.
The e£fects of small changes of pressure are, however, so slight that
in considering the shape of a solubility curve they may be neglected.
Compound ScUvhUUy Curves.
From the foregoing considerations, it is obvious that the forces at
work during solution slowly change their values with rise of tempera-
ture, and the resultant curve for the same substance must be an
unbroken line.
If a solubility curve shows one or more breaks, the points of flexure
are the meeting places of curves representing the solubilities of
different hydrates, which are physically different substances.
Experiment proves this to be the case : no broken curve is Obtained
for a substance which cannot change in composition with the tem-
perature. If an anhydrous salt is in contact with the solvent at a low
temperattire, the solubility curve will be unbroken ; if the solid contains
water of crystallisation the curve is broken or simple, according as the
substance in contact with its saturated solution loses water or not
in the range of temperature employed.
The point of flexure is neither the end of one curve nor the
beginning of the next curve ; it is simply a place of meeting. The
lower temperature ciirve may, under favourable conditions, be
continued above the transition point, and the higher temperature
curve can pass below it, but the equilibrium is then unstable, and it is
experimentally difficult to prevent change to the more stable solid.
The condition of equilibrium at a transition point is interesting:
two solids are there in equilibrium with the same solution, the sums
of the thermal energy and affinity in both are therefore equal, since
they are both subjected to the same osmotic pressure. From the point
of view of .the " phase rule," it is a non- variant equilibrium ; two
components are present : the solid and water, and ^there are four
phases : one liquid, one vapour, and two solid. There is no degree of
freedom, and any change of temperature or pressure destroys one of
the solid phases, producing a mono-variant system.
It is probable that all solubility curves which have been drawn,
showing a continuous bend concave to the temperature axis, should
be two curves. The observer, when joining the points found experi-^
mentally, has missed the transition point.
This is the case with barium acetate, the solubility of which was
determined by Tilden and Shenstone {PhU. Trans., 1884, 174, 23) and
the curve is figured in several books (Ostwald's Lthrhuch, vol. i.).
002
Digitized by VjOOQIC
372 EQUILIBBIUM BETWEEN A SOLID AND ITS SOLUTION.
I have found that there is no concave part on this curve, and tliat
the investigators by estimating the solubility only at points widely
apart, and neglecting to analyse the solid in contact with the solutioOi
joined points on two different curves and cut off the transition
point.
Another case is that of calcium wobutyrate, where Chancel and
Parmentier missed the meeting point of the two curves shown on
Fig. 4 in the preceding paper.
From theoretical consideiutions, it would se^m that the eaddtence of
solubility curves concave to the temperature axis was improbable, and
in all cases where concave curves have been carefully examined they
have been shown to be due to two substances, the exnct point of
mutual equilibrium of which had been missed.
Summing up, we have found that there are three factors which
condition the solubility of a substance : the affinity of solid and
solvent, the thermal energy of the solid, and the pressure of the
dissolved particles, and that the values of these vary with temperature
and pressure.
When equilibrium is attained, a solution is saturarted, and a curve
of solubilities is an exact representation of the resultant of the
solubility factors at varying temperatures. If the values of the
factors vary directly with the temperature, the solubility curve is a
straight line ; when the change is at a different rate, a curve is ob-
tained. No well authenticated case of a solid the solubility of jwhich
is represented by a curve concave to the temperature axis is known,
and the normal shape is either a straight line or a convex curve.
Many calcium salts have solubility curves which descen I with rise
of temperature, reach a minimum, and then ascend, but other sub-
stances have solubilities represented by portions of such a curve. It
will depend on the relative values of the solubility factor?, and on the
range of temperature over which the estimations have been made,
whether a curve will descend and rise, or descend only, or as is most
commonly the case be an ascending curve.
Uniybbsitt Collegx, Dundbb.
St. Andbbw's Univebs;ty.
Digitized by VjOO^ IC
bbown: enztme action. 873
XX;§!VI. — Enzyme Action.
By Adrian J. Bbown.
Introduction*
In a paper on the fermeDtative fanctions of yeast (Trans., 1892, 61,
380), the author described some experiments which showed that the
character of the action of fermentation differed in a very marked
manner from the character of the action usually attributed to enzyme
change.
The author's experiments indicated that during fermentative change
a constant amount of yeast decomposes an approximately oonstant
weight of sugar in unit time in solutions of varying concentration, and
that the velocity of fermentative action is therefore represented
gi*aphically by a straight line. On the other hand, the character of
the action usually attributed to an enzyme is that a constant amount
of the fermont changes in unit time a constant fraction of the reacting
substance present, and that the velocity of its action is represented by
the logarithmic curve of mass action.
At the time the author's work {loc. cit,) was published, the fermen-
tative power of the yeast cell was considered to be a life function
inseparable from the cell, and there appeared to be nothing specially
remarkable in the observation that fermentation, a life function,
differed in the velocity of its action from enzyme action. But since
the more recent work of Buchner has demonstrated that the phenomenon
of fermentation is caused by enzyme action, the question assumed
another aspect. If fermentation is now regarded as an enzyme action,
then, either the velocity of its action must be regarded as differing
essentially from that which is usually attributed to other enzymes, or
the experimental evidence on which the assumed difference rests must
be regarded as misleading.
It was with the intention of investigating this question that the
author commenced the work described in the following paper.
Velocity of the Action of Fermentation.
If the view is adopted as a working hypothesis that the supposed
difference in velocity of the actions of fermentation and ordinary
enzyme change does not exist, but that it is due to some misconception,
it is evident misconception may have arisen concerning either the
velocity of fermentation or that of ordinary enzyme change, and con-
sequently a re-examination of the experiments by which both velocities
.bay^ been determined appeared desirable.
Digitized by
Google
874 BROWN: ENZYME ACTION.
^s the author is responsible for the experiments by which the
velocity of fermentation has been determined, he commenced his in-
yest-igations by repeating them. It does not appear necessary, how-
ever, to give the results of this work, for the experiments were similar
to those described in the earlier paper {loc. cU.)y and the results fully
confirmed the conclusion that fermentative action does not proceed in
accordance with the law of mass action.
As the general character of the action of fermentation appeared to
be thus established, the author proceeded to examine the experimental
evidence from which the conclusion is drawn that the velocity of
enzyme action accords with the law of mass action.
Velocity of Enzyme Action.
The generally accepted view regarding the velocity of enzyme action
is based on the researches of Cornelius O' Sullivan and F. W. Tompson
on the action of invertase on cane sugar (Trans., 1890,57, 865). These
authors demonstrated the velocity of the action of invertase in the
following manner.
Invertase was caused to act in solutions of cane sugar, and during
the progress of the actions the quantities of sugar inverted during a
succession of time intervals were determined. By this means, obser*
vations were obtained from which time curves were constructed which
represented graphically the velocity of the action of inversion. When
these curves were compared with the curve representing simple mass
action, a very close agreement in shape was observed, which appeared
to indicate that they were of the same order, and from this close
agreement in shape, C. O'Sullivan and Tompson concluded that the
action of invertase instanced the operation of the law of mass action.
This conclusion has also received confirmation from the researches
of James O'Sullivan on the power of inversion of living yeast cells
(Trans., 1892, 61, 926), the experiments of this author indicating that
the velocity of action of the living cell is the same as that of the ex-
tracted invertase used by C. O'SuUivan and Tompson in their experi-
ments.
The evidence referred to is, so far as the author is aware, all
that has been brought forward to support the conclusion that the
velocity of enzyme action indicates the operation of a simple mass
action.
Hitherto, no doubt, the want of additional evidence has not been
felt, owing to 0. O'Sullivan and Tompson's experiments appearing
conclusive so far as invertase is concerned, and also to the fact that
the conclusion these, authors arrived at with regard to the character
of enzyme action iis one which there is every reason to anticipate.
Digitized by
Google
BROWN: ENZYME ACTION. 375
But when the character of the action of fermentation, . now very
generally recognised as an enzyme action, was found to differ essen-
tially from that attributed to invertase both in the free state and within
the living yeast cell,* it raised doubt in the author's mind regarding
the accuracy of C. O'SuUivan and Tompson's conclusion. Moreover, the
author found that he was not alone in his distrust, for it has already
been pointed out by Duclaux, in a criticism on C. O'Sullivan and
Tompson's work (Ann. Inst, Pasteur, 1898, 12, 96), that the logarithmic
curve representing the action of invertase, on which C. O'Sullivan and
Tompson founded their conclusion, may be shaped by other causes
than the supposed action of mass. For Duclaux maintains that
such a curve represents, not only a decrease in a changing substance,
but also, and equally well, an increase in the products of change, and
it is possible these products of change may act as the influence shaping
the curve and not the influence of mass action. No experimental
evidence sustaining this point is, however, brought forward by
Duclaux.
As C. O'Sullivan and Tompson's conclusion rests entirely on the
shape of the curve representing the action of invertase, the author
considered it advisable first to repeat the experiments from which the
curve was derived. Conditions of experiment similar to those used by
O. O'Sullivan and Tompson were employed, but the invertase used
was prepared in a different manner from the enzyme with which these
authors experimented.
C. O'Sullivan and Tompson worked with invertase obtained from an
extract of auto-digested yeast by precipitation with alcohol, and in so
doing encountered the difficulty that the action of invertase prepared
in this manner was very irregular unless it was associated with a
small quantity of sulphuric acid. Moreover, the amount of acid re-
quired to reach the point described by these authors as ** the most
favourable condition of acidity," at which point it was necessary to
work, varied in every experiment in a most remarkable manner.
It appeared very desirable to avoid this complication when repeating
0. O'Sullivan and Tompson's experiments, so the author employed in
his experiments an extract of invertase prepared from dried yeast by
digestion with water. An extract of invertase prepared in this
manner was quite suitable for the purpose of the experiments, and
the risk of modifying the activity of the invertase by precipitation
was avoided. That this method of obtaining a preparation of in-
vertase suitable for experiment was preferable to that employed by
O. O'Sullivan and Tompson was evidenced by the invertase being free
from the irregularities of action associated with the precipitated in-
* Presamably, inyertase within the wall of the living cell is in the same position
as zymaae with regard to its action as an enzyme*
Digitized by
Google
376
BROWN: ENZYME ACTION.
vertase used by these authors, and, in consequence, it could be em-
ployed without the complicating addition of sulphuric acid.
The author's experiments, like those of 0. O'SuUivan and Tompson,
consisted in the addition of a suitable amount of invertase solution to
a solution of cane sugar, and in determining, by means of a polari-
meter, the fractions of the sugar inverted during successive intervals
of time.
Instead, however, of expressing the velocity of the inversion change
by means of a curve, the author preferred to make use of the value k,
derived from the expression -^ogr: . This well-recognised means
of expression, usually adopted now to demonstrate such changes as
those of a mass action, has the advantage of avoiding certain difficulties
which attend the comparison of calculated and experimental curves.
The results of two series of experiments determining the velocity of
the action of invertase are given in Tables I and II :
Table I. — VelocUy of inversion change in 9*48 per cent, solution qfcane
sugar. 500 c.e. of solution of sugar and 25 e.c. of invertase solution
used. Tenip. 30°.
Duration of time interval
Fraction of sugar
*=N/ .
in minntes.
inverted in B,
e.
X.
e i-x
30
0-266
0-00446
64
0-609
0 00483
120
0-794
0-00671
180
0-946
0-00698
240
0-983
0-00787
800
1-003
During the course of a change proceeding as a simple mass action, it
is well known that the value k, determined for any point of the action,
is a constant. But in the experiments described in the above tables it
will be noticed that the value k increases in both experiments as in-
version proceeds, until the value at the termination of the experiments
is about 70 per cent, higher than at the beginning.*
Now, these results do not support the view that the action of in-
version instances a mass action, as 0. 0*Sullivan and Tompson believed,
for they differ very materially from the results these authors obtained.
But in order to emphasise more distinctly the difference between the
* It will be noticed that there is no indication of " reversion" in these inyeraion
experipients. An increase in the valae of k denotes an increasing velocity | r^-
venion would lead to a decreasing velpcity.
Digitized by
Google
BROWN : ENZYME ACTION.
877
Table II. — Velocity of inversicn change in 19'28 per cent solutioti of
ecme sugar. 500 o.c. of eugar solution and 25 c.c. of inverlaee
solution used. Temp. 30"".
Dnretion of time interval
Fraction of sugar
*-llog/ .
in minutes.
inyerted in 6.
S,
X.
9 \-x
80
0 180
0-00201
64
0-266
0 00201
120
0-464
0-00219
180
0-619
0-00282
240
0-788
0-00242
800
0-881
0-00267
860
0-890
0-00266
420
0-985
0 00283
480
0-961
0-00298
540
0-988
0-00827
681
0-990
0-00844
character of the action of inversion and that of a mass action, the
results of an experiment involving mass action are given in Table III
for the purpose of comparison. In this experiment, the author em-
ployed sulphuric acid to invert cane sugar, and thus obtained results
from a typical mass action, which are directly comparable with those
effected by invertase.
Table III. — Ydodttf of inversion chomge of cane sugar by acid. 600 cc.
qf a 20 per cent, solution of cane sugar^ and 36 cc. of normal
HjSO^ used. Temp. 48°
Duration of time interval
Fraction of sugar
7 ll 1
in minutes.
9.
inverted in 9,
X.
*=rt-.-
80
0166
0-00261
61
0-817
0-00271
90
0-488
0-00274
120
0-682
0-00276
160
0-617
0-00278
180
0-688
0-00281
248
0-786
0-00276
802
0 866
0-00279
862
0-902
0-00278
The above experiments show very clearly, when a true mass action
is followed under experimental conditions similar to those used when
determining; the character of the aption of invertase, that a very differ-
Digitized by
Google
378
brown: enzyme action.
ent result is obtained. The small and irregular variation in the value
of h^ is very different from the regular and well marked increase in
the value of h observed in the experiments with invertase given in
Tables I and II, and from this there can remain but little doubt that
the order of progression of inversion differs essentially from %hat of
a mass action. At present, the author has not attempted to determine
any expression for the order of progression of inversion under the con-
ditions of his experiment, because, for the immediate purpose of his
investigation, it is only necessary to show that inversion does not
proceed as a mass action, t
Although the author's experiments throw the greatest doubt on
the accuracy of the conclusion that inversion evidences mass action, they
cast very little light on the true character of the action of invertase,
and in order to obtain more knowledge, it became necessary to adopt
some method of experiment different from that which has already been
described.
* The TariatioDS in value of k are no doubt due to experimental error. Very alight
changes in temperature have a marked influence on the yeloeity of inversion change
by acid.
t Since writing the above, a communication from Victor Henri has been pub-
lished (Compt. rend,, 1901, 183, 891) on the velocity of inversion change. Tbis
author arrives at the conclusion that the action proceeds in accordance with the
expression 2ki=iAog .
9 1 — JB
On applying this purely mathematical expression of velocity to the inversion ex-
periments described in Table II, the following results have been obtained :
Table II.— Recalculated.
Duration of time interval
Fraction of sugar
2it,-W^^
in minutes.
inverted in 9.
9,
X,
9 1-x
30
0-130
0 00376
64
0-256
0 00856
120
0-454
0 00356
180
0-619
0-00346
240
0-738
0-00343
300
0-831
0-00353
360
0-890
000348
420
0-985
0-00351
480
0-961
0 -00354
640
0 988
0-00383
581
0-990
0-00895
It will be observed from the remarkably close agreement of the numbers in the
third column, representing the values 2k, that the author's experiments indicate a
very similar velocity for inversion change to that which is represented by Henri's
expression. This is of interest as further assisting to establish the fact that the
progress of an inversion change is not ordered in conformity with the law of :
ac^on.
Digitized by
Google
BROWN: ENZYME ACTION.
379
Such a method exists in causing a constant mnount of invertase to
act on varying amounts of cane sugar in constant volume of solution
for a constant brief interval of time. Under these conditions, the
variable in the actions is the amount of reacting substance (cane sugar)
present, and in a simple mass action under these conditions, the amount
of reacting substance changed in unit time is a eanBtani fraction oi the
reacting substance present.*
In Table lY, the results of a series of &ve experiments are given in
which a constant amount of inverta^ has acted on varying amounts of
sugar under the conditions just named :
Table IV. — Inversion changes by a conatant amount of invertase acting
in consUmt volume of cane sugar soltUions of varying concentrations,
in constant time, 1 c.c, of invertase solution added to 100 cc. of
cane sugar solution in each eseperiment Temp. 28^
No. of
experi-
ment.
Grams cane
sugar per
100 c.c.
Grams cane
sugar inverted in
60 minutes.
Fraction of cane
sugar inverted in
60 minutes = a?.
^■>^.-
1
2
8
4
5
4-89
9-85
19-91
29-96
4002
1-230
1-355
1-355
1-236
1-076
0-262
0-188
0 068
0-041
0 027
0 0(4210
0-00107
0-00051
0 00031
0 00020
\ When the law of mass action was evidenced by Ostwald's experi-
ments on methyl acetate {loc, cit.) under conditions similar to those
employed in the above experiments, he found that a constant /^action
of the methyl acetate present in each solution was hydrolysed in unit
time, and therefore, if the action of invertase is an instance of simple
mass action, a constant fraction of the cane sugar present in each of
the above experiments should be inverted. But it will be noticed that
instead of a constant fraction, a constant (or approximately constant)
weight of the cane sugar is inverted. The fraction inverted diminishes
in inverse proportion to the amount of cane sugar present in the
experiments, and, as a consequence, the value k, which is constant in
a true mass action, varies to a very large extent.
Those experiments,t therefore, confirm the conclusion derived from
the experiments given in Tables I and II, that the influence of
* For experimental confirmation of this necessary consequence of mass action, see
Ostwaldon the hydrolysis of muthyl acetate by hydrochloric acid ("Outlines of
General Chemistry/' p. 853).
t Daclaux (loc. eit.) quotes some experiments with invertase derived from Asper-
gillw niger wl^ich fully confirm the author's experiments described in Table iy»
Digitized by
Google
380 BROWN: ENZYME ACTION.
mass action does not' rule in inversion change. Moreover, when the
velocity of the action of inversion determined in the manner des*
crihed in Table lY is examined, it will be noticed that the action is
similar in character to that of fermentation, referred to at the com-
mencement of this paper. During alcoholic fermentation, a constant
amount of yeast decomposes in unit time an approximately eamtani
toeight of sugar in equal volumes of solution containing varying amounts
of sugar. Invertase is now found to invert approximately constant
quantities of cane sugar under similar conditions. Therefore the sup-
posed difEerence in character of the two actions of fermentation and in-
version which led the author to commence the investigation described
in this paper does not exist, for the action of both, if expressed graph-
ically, is represented approximately by a straight line. So far, there-
fore, the first object of the investigation is attained.
Although experiments carried out in the manner just described
show that the general character of the action of invertase resembles
that of fermentation, they do not explain the apparent paradox that
when the action of invertase is studied during a series of consecutive
changes in a single solution, the velocity of the action is then repre-
sented, not by a straight line, but by a curve, showing that there is a
decrease in the ahioluU amount of sugar inverted during each time
interval (see the experiments in Tables I and II and foot-note to p. 378
which indicate that, although the curve of the action is not so pro*
nounced as the logarithmic curve of mass action, it is still very
marked).
Apparently there are two causes which may lead to the production
of such a curve during the continued action of invertase in a solution
of cane sugar. Either it may be due to a natural weakening of the
invertase by continued work,"* or it may be due, as Duclaux has sug-
gested {loc, ct^.),to the action of invertase being influenced prejudicially
by the accumulation of its own products of inversion. From what is
known regarding the very large amount of cane sugar which is
capable of being hydrolysed by a very small amount of invei'tase, the
former cause appeared to be far less probable than the latter, so the
author turned his attention to the investigation of the possible
retarding influence of inversion products on the action of invertase.
Action qf Inversion Products on the Velocity of Inversion Change,
The method of experiment adopted by the author was to observe the
action of a constant amount of invertase during a brief interval of
time in equal volumes of solutions containing a constant amount of
* Some interesting experimcDts of Victor Henri {loc. cU*) indicate that aoataiiied
work does not weaken the action of inyertaae,
Digitized by
Google
brown: EK2TME ACTlO^.
381
cand sug^i* and mrying amount§ of invert sugar. The invert sugar
lued in the experiments was prepared by the action of invertase on a
concentrated solution of cane sugar until complete inversion was
obtained, the invertase being then destroyed by raising the temperature
of the solution to 90^.
The following table gives particulars of a series of experiments in
which different amounts of this solution of invert sugar were mixed
with a constant amount of cane sugar solution, the total volumes of
the solutions in the different experiments being made constant :
Table Y. — Influence qf invert eugar on the action qf invertase. Folume
jfeach experiment, 100 c.c. 1 c.c. of invertase solution used in each
experiment. Time qf change, 80 minutes. Temp. 30^.
No. of
experiment.
Grams cane sugar
present in 100 c.c.
Grams invert sugar
present in 100 c.c.
Grams cane su^r in-
verted in 80 minutes.
1
2
8
4
6
4-06
4 06
4-06
4-06
4-06
none
1-47
5*89
11-88
17-67
2-27
2-21
1-99
1-66
1-26
In these experiments, if the presence of insert sugar exerted no
inHuence on the action of invertase, the quantities of cane sugar
inverted in constant time would be constant, for the same quantities
of cane sugar and invertase were present in all the experiments. But
an examination of the table shows that the amount of cane sugar
inverted decreased as the quantity of added invert sugar increased,
until, in the last experiment (No. 5), the quantity of cane sugar
inverted has been reduced to nearly one*half in the presence of 17*87
grams of invert sugar.
The series of experiments indicate, therefore, that the presence of
invert sugar has diminished the activity of invertase, and that the
arresting influence has increased as the amount of invert sugar has
increased. But it is possible that the arresting influence of invert
sugar may be due, not to tbe presence of the sugar as such, but to the
increased viscosity of the solution containing it, for, owing to the
manner in which the experiments were conducted, the total amount of
sugars in the different solutions is an increasing one from the first to
the last experiment. In order to investigate this question, a series of
experiments was conducted in a similar manner to those described in
Table Y, excepting that lactose was used in the place of invert sugar.
Lactose is a sugar which is not changed by the action of invertase, but
its solutions possess a viscosity almost identical with that of solutions
Digitize'd by
Google
382
BBOWN: ENZYME ACTION.
of invert sugar of similar conoentrationa Therefore in a series of ex*
periments with lactose in place of invert sugar, the factor of increas-
ing viscosity is introduced apart from any special influence possessed by
invert sugar alone.
The results of a series of experiments with lactose are given
below :
Table YL — Infltbenee of Ictctose on the action qf invertaae. Volums in
each experiment^ 100 e.c. 1 e.c» of invertaee si^utian tued. Time
of change, 60 minutes. Temp, 28°.
No. of
experiment.
Grams cane sugar
present in 100 c.c.
Grams lactose
present in 100 c.c.
Grams cane sn^r in-
yerted in 60 minutes.
1
2
3
4
7-0
7-0
7-0
7 0
none
6 0
10-0
20-0
2-072
2052
2-052
1-893
The results given in this table show that the influence on the action
of invertase of the viscosity (or any other property) of the lactose
used in the experiments is comparatively insignificant. In' experiments
2 and 3, the retarding influence of 5 per cent, and 10 per cent, of
lactose lies almost within the limits of experimental error, and in i,
in which the large amount of 20 per cent, lactose is present, the
reduction in the amount of cane sugar inverted is only 9 per cent.
On the other hand, it has already been shown (Table Y, No. 4) that
17*8 per cent, of invert sugar under similar conditions reduced the
amount of cane sugar inverted to the extent of 45 per cent. The
major part of this reduction, therefore, is not due to viscosity, but
must be occasioned by the arresting influence of invert sugar as such.
When the arresting influence of invert sugar on the action of
invertase is thus established, there is then no difficulty in explaining
the apparent paradox that the true action of invertase, which is
indicated graphically by a straight line, is expressed by a curve when
the action is determined for a series of progressive changes in one
solution. Under the latter conditions, as the action of inversion
proceeds, the products of inversion accumulate, and these consequently
exert an increasing retarding influence on the action of inversion, and
thus compel the action to follow the course of a curve.
Digitized by
Google
BBOWN: BNZTME ACTION. 383
The Inversion Functions of Living Teaet CelU.
So far, whea discussing the action of invertase, the author has
referred more especially to 0. O'Sullivan and Tompson's experiments
and conclusion regarding the velocity of its action. It now remains
to discuss J. O'Sullivan's experiments, alluded to at the commencement
of this paper as supporting C. O'Sullivan and Tompson's conclusion.
It will be remembered that J. O'Sullivan, when studying the velocity
of the inversion change produced by living yeast in solutions of cane
sugar {J,oc. cU\ found that the value k derived from the expression
^log- , was constant^ or nearly so, during the progression of the
changes, and from this he concluded that the velocity of change
followed the law of mass action.
There is no doubt that J. O'Sullivan's 'determinations — like those
of 0, O'Sullivan and Tompson — indicate a velocity approximating to
that of mass action, when the progress of an inversion change is
followed in one solution ; but J. O'Sullivan has overlooked the fact —
rendered evident by his own determinations — that, although the
velocity in each separate change approximately follows the law, the
value k found for comparable experiments in which varying amounts
of sugar have been used, shows that there is no conformity with mass
action, but, on the contrary, indicates, that a conatcmi amount of sugar
is inverted — an action similar to that which has been shown for free
invertase.
For instance, in J. O'Sullivan's paper four comparable experiments
are described, in which equal amounts of yeast were used in equal
volumes of solution during equal intervals of time, the only variable
being the quantity of cane sugar present in the solutions. The results
of these experiments are given in the table on p. 384.
It will be noticed, on examining this table, that J. O'Sullivan
has determined the progression of inversion in each of the four
solutions at three time intervals, and the values of k for the changes
in each separate solution are fairly constant; but the values k
should also be constant for all foar of the solutions if the velocity of
change follows the law of mass action, because the solutions only differ
in containing varying quantities of sugar. On the contrary, however,
the value k varies inversely as the amount of sugar present, in a
similar manner to the value k in the author's experiments with
invertase, given in Table lY.
H A similar conclusion may also be derived from the numbers in the
oolomn in Table YII showing the fractions of cane sugar inverted
during the experiments. If the first numbers in each series are
Digitized by
Google
384
BROWN: ElNZYMBl ACttOK.
Table VIT. — Velocity of inversion hy living yemt cells {J, 0* Sullivan).
Grams cane
sugar per
100 c.c.
Grams of
yeast used.
Time of action
in minutes.
0.
Fraction of
sugar inverted.
'■>-ih-
5
0-5
30
60
120
01686
0-8164
0-5442
0-0025
0-0027
0-0028
10
0-6
80
60
120
0-1042
0-1544
0-2780
0-0016
0-0012
0-0012
20
0-5
80
60
120
0 0627
0 0860
0-1467
0-0009
0-0006
0*0006
80
0-5
80
• 60
120
0 0366
00495
0 0S62
0-0005
0-0008
0-0008
compared, it will be noticed that the fractions inverted are, ap>
proximately, in inverse proportion to the amounts of cane sugar
present in the solutions — or, in other words, the actual quantity of
sugar inverted is the same for all the experiments.
Thus J. O'Sullivan's experiments show that the velocity of action
of the inversion function of yeast falls into line with the action of free
invertase, and the action of fermentation,* previously demonstrated by
the author.
Time and Molecular Change.
It was stated at the commencement of this paper that the author's
object was to examine, and, if possible, bring together, certain con-
clusions regarding the nature of enzyme action which seemed to be
contradictory. Experimental evidence appeared to show that on the
one hand the action of invertase, both in the free state and confined
within the living cell, followed the law of mass action ; and, on the
other hand, that the action of the enzyme of alcoholic fermentation
followed a different law. The author has now shown that these
supposed differences in character of action do not exist, and that the
actions of both inversion and fermentation follow approximately
* It is interesting to note the agreement in character of action of the inversion
and fermentation functions of the living yeast cell, as it tends to strengthen the
conclusion that fermentation is a true enzyme action.
Digitized by
Google
BROWK: BN2YMB ACTION. 386
the same order of progression — an ^order which is not that of mass
action.
But this conclusion, that the actions of the two enzymes exhibit an
exceptional order of progression difEering from that of mass action,
introduces a question which requires explanation.
It appears impossible to believe that enzyme change, however
produced, is independent of mass action. According to our present
conception of matter and its mechanics, such an idea appears to be
inconceivable. Therefore, in looking for some explanation of the
exceptional character of the actions of inversion and fermentation,
the author concludes that the influence of mass in these actions, as
they have been studied so far, must be limited or concealed by some
other influence.
If such an influence is looked for, consideration shows that it may
be due to the existence of a time factor in certain forms of complex
molecular change.
When the law of mass action regulating simple chemical change has
been confirmed by direct experiment, the reactions investigated have
been changes such as the hydrolysis of methyl acetate by hydrochloric
acid (Ostwald, loc, cit,) and the inversion of cane sugar by acids. In
such experiments, the molecular change following collision of the re-
acting molecules takes place with extreme rapidity and the existence of
a time factor is not in evidence in experimental determinations of the
velocity of change. But it is quite conceivable, with regard to such
a change as that of enzyme action, that the time elapsing during
molecular union and transformation may be sufficiently prolonged to
influence the general course of the action.
There is reason to believe that during inversion of cane sugar by
invertase the sugar combines with the enzyme previous to inversion.
G. O'SuUivan and Tompson (loc, cit.) have shown that the activity
of invertase in the presence of cane sugar survives a temperature
which completely destroys it if cane sugar is not present, and regard
this as indicating the existence of a combination of the enzyme and
sugar molecules. Wurtz {Campi, r&nd., 1880, 91, 787) has also shown
that papain appears to form an insoluble compound with fibrin
previous to its hydrolysis. Moreover, the more recent conception of
E. Fischer with regard to enzyme configuration and action, also im«
plies some form of combination of enzyme and reacting substance.
Let it be assumed, therefore, that one molecule of an enzyme
combines with one molecule of a reacting substance, aad that the
compound molecule exists for a brief interval of time during the
further actions which end in disruption and change. Let it be
assumed also that the interval of time during which the compound
molecule of enzyme and reacting substance exists is 1/100 of a time unit.
VOL. LXXXI. D D
Digitized by VjOOQIC
386 BROWN: ENZYME ACTION.
Then it follows that a molecule of the enzyme may assist in effecting
100 completed molecular changes in unit time, but that this is the
limit to its power of change.
Again, let it be assumed that the number of molecular collisions be-
tween the active and reacting molecules which lead to their combination
bears some proportion to the number of possible completed molecular
changes in unit time. Let the number of collisions be 20, then there
may be 20 complete molecular changes ; if 40, there may be 40 changes.
In fact, the action of the mass law is observed, for other conditions
being equal, the average number of molecular collisions must depend
on the number of molecules, or mass, of the matter present.
But now assume that the mass of reacting substance is increased, so
that the number of molecular collisions in unit time exceeds 100 ;
let it be 150, 1000, or any other number larger than 100. Then,
although the number of molecular collisions may exceed 100 by a
number following the law of mass action, 100 molecular changes can-
not be exceeded, for the compound enzyme and sugar molecule is only
capable of effecting 100 complete changes in unit time.
It follows, therefore, that if, in a series of changes like the imaginary
ones described, a constant amount of enzyme is in the presence' of
varying quantities of a reacting substance, and in all cases the quan-
tity of reacting substance present ensures a greater number of
molecular collisions in unit time than the possible number of molecular
changes, then a constant weight of substance may be changed in unit
time in all the actions.
When invertase acts in solutions of cane sugar of varying con<ien-
trations, an approximately constant weight of sugar is inverted in unit
time, and the yeast cell, under similar conditions, ferments an approxi-
mately constant weight of sugar; it appears, therefore, that the ex-
ceptional character of these changes may be satisfactorily accounted
for by the theory advanced.
Experimental evidence may also be brought forward in support
of this theory.
In Table lY, the results of the author's experiments show that
approximately constant quantities of cane sugar are inverted in unit
time in solutions varying in concentration from 5 to 40 per cent. If
the results of these experiments are looked at in the light of the
author's theory, the number of molecular collisions in unit time in each
experiment must have equalled, or exceeded, the possible number of
changes by the compound molecule of enzyme and sugar. But this
has happened in solutions containing 5 per cent, and upwards of cane
Sugar. It must, however, be possible to make solutions of varying
quantities of cane sugar so dilute that the number of molecular col-
lisions taking place in unit time between the sugar molecules and a
Digitized by
Google
BftOWN: ENZYME ACTtON,
887
oonstant number of invertase molecules will fall below the possible
number of changes. Then, if the author's theory be correct, the pro-
gress of inversion in a series of these dilute solutions of cane sugar of
different concentrations will exhibit an action in accordance with
the law of mass action, for the time interval of change no longer
restricts its effect.
It seemed very possible when commencing the attempt to demonstrate
this experimentally that it might prove that the solutions of cane sugar
required for the purpose were too dilute to use for experimental pur-
poses. But when the attempt was made, it was found that the necessary
dilutions are within the limit of experiment, as the results given in the
following table (YIII) show :
Table YIII. — Vdooity of action of invertaae in very dikOe sohUtons qf
cane eugar, 100 c.c. qf cane 8^<vr solution and 1 ex. of diluted
invertase solution employed for each eacperimient. Time of change^
60 minutes. Temp. 3 P.
No. of
experiment.
Grams cane sngar
per 100 c.c.
Grams cane sngar in-
verted in 60 minutes.
e *l-a!
1
2
8
4
2-0
1-0
0-5
0*25
0-808
0-249
0129
0-060
0-00182
0-00219
0-00289
0-00228
The results given in the above table furnish very strong evidence in
support of the view that in the dilute solutions of cane sugar employed
the number of contacts of the sugar molecules with the invertase
molecules in unit time have been reduced to a less number than the
possible number of molecular changes. In experiment No. 1, in which
a concentration of 2 grams of sugar per 100 c.c. has been used, the
dilution appears to have been hardly sufficient to reach the desired
point. In Nos. 2, 3, and 4, however, the quantities of sugar inverted
in unit time are no longer constant quantities — as was found in
the experiments with concentrations of 5 per cent, and upwards,
given in Table lY., and decrease in direct proportion with the con-
centrations.
Moreover, the value k in these experiments is a constant number.
These observations indicate a change in accordance with mass action,
which, according to the author's theory, should be evidenced in solu-
tions of sufficient dilution^ There is, therefore, reason to believe from
the results of the above experiments that the exceptional action of
inversion in all but very dilute solutions of cane sugar is due to a time
Digitized by VjOOQI(
S88 BROWN AND GLENDINNINQ : TttE VELOCITY Olf
factor accompanying molecular combination and change which limits
the influence of mass action.
It has been shown in this paper that the action of alcoholic fer*
mentation follows approximately the same order of progression as that
of inversion, and the work of Ksistle and Loevenhart {Amer, Chem. J,f
1900, 24, 491) shows that the action of lipase progresses in the same
manner ; it therefore appears probable that both these enzyme actions
are regulated, like inversion, by a time factor accompanying complex
molecular change.
It will be noticed that the author's theory demands, not only the
formation of a molecular compound of enzyme and reacting substance,
but the existence of this molecular compound for an interval of time
previous to £nal disruption and change. Various speculations regard-
ing the conditions ruling such an effect suggest themselves, but the
author does not at present attempt to discuss this question.
The British School of Malting and Bbewing,
Univb&sitt of Bieminqham.
XXXVIL— 2%e Velocity of Hydrolysis oj Starch hy
DiastdsCy with some Remarks on Enzyme Action.
By HoBAOx T. Bbown, LL.D., F.R.S., and T. A. Glbndinkino, F.I.C.
The experimental work here described, on the rate of change during
the hydrolysis of starch by malt diastase, was completed more than four
years ago, but the results were temporarily put on one side, owing to
the impossibility of reconciling them with the views then current with
regard to the analogous changes induced by the action of invertase on
cane sugar.
The investigations of C. O'Sullivan and Tompson (Trans., 1890, 81
834) and of J. O'Sullivan (Trans., 1892, 61, 926) had led these ob-
servers to conclude that the time rate of change during the inversion
of cane sugar by the enzyme is the same as that observed in add hydro-
lysis ; that it conforms, in fact, to the logarithmic formula character-
istic of a unimolecular reaction.
Our own observations on the hydrolysis of starch by diastase had
convinced us that the rate of change certainly does not conform to this
simple law of mass action, but that there is a progressive increase in
the value of the " velocity coefficient " which appears at first sight to
differentiate the mode of action of diastase from that of invartase. The
recent work of Adrian Brown and of Y. Henri has shown, howeveTi
Digitized by
Google
HTDROLTSIS OF STARCH BY DIASTASE. 889
t^at in cane sugar hydrolysis, both as regards the time elements of
change and the influence of varying concentration, the action of the
enzyme differs materially from that of dilute acids.
This has led us to re-examine our experiments on starch hydrolysis,
with the result that they are found to be in complete accord with the
observations of the two last-mentioned chemists, thus rendering it
probable that there will be found one fundamental law expressing the
rate of change in all enzyme reactions which can be quantitatively
studied with sufficient accuracy.
If we desire to follow the course of a starch transformation as a
function of the time, there are two methods open to us, one based on
the diminution of the optical activity of the solution, the other on the
augmentation of the cupric reducing power of the mixed products of
change.
It has been frequently shown by one of us (Brown and Millar,
Trans., 1899, 76, 315) that when soluble starch is transformed by an
active diastase at temperatures below 60°, the hydroijsed solution
speedily attains definite optical and reducing properties corresponding
to a well-defined molecular decomposition of the original starch. This
point is reached when the mixed products of hydrolysis have attained
a specific rotatory power of [a]D 150°, and a cupric reducing power of
R SO} R being expressed in terms of maltose per cent, of the mixed
products. This stage of the reaction is so well defined and permanent
that it may be regarded, under ordinary conditions of starch trans-
formation, as an indication that hydrolysis is complete. If, therefore,
we ascertain either the total fall in rotation (which may be expressed ,
either in degrees or in arbitrary scale units), or the increase in cupric
reduction which a given solution of starch experiences in attaining this
final resting stage, then the rotation or reduction at any intermediate
stage will give us a measure of the amount of hydrolysis at that
particular moment of time.
We have employed both these methods during the investigation, but
since it is impracticable to use solutions of soluble starch of greater
concentration than from 3 to 4 per cent., the total fall in optical
activity is not large, and the errors of reading of the polarimeter con-
sequently assume a very sensible magnitude relative to the total fall.
Under these conditions, the optical method is far less accurate as a
measure of the progress of the reaction than that based on cupric
reduction, which is the one which latterly was exclusively employed.
The experiments we have cited in this paper were carried out as
follows.
A 3 per cent, solution of soluble starch, prepared by Lintner's acid
method, was maintained at a constant temperature in a thermostat,
and to a known volume of the solution was added a smaU, definite
Digitized by
Google
800 BEOWN AND GLENDINNING : THE VELOCITY OF
amount of a cold water eztract of an actively diastatic malt, the time
of this addition being accurately noted. In those cases in which the
transformation was to take place at an elevated temperature the malt
eztract was previously heated for an hour at that temperature, or a
little above.* The amount of malt eztract was arranged with due
regard to the temperature at which the transformation was to take
place, and was so adjusted that the reaction did not progress too
rapidly for the subsequent operations.
After the commencement of hydrolysis, and at carefully noted inter-
vals, portions of the liquid amounting to ezactly 25 o.c were quickly
taken out with a pipette, and were at once mized with about double
the volume of boiling Water, the temperature of the mizture being
sufficiently high to at once arrest further diastatic action. After boiling
for a few minutes, the solution was evaporated and accurately made up
to its original volume. The cupric reducing power was then deter-
mined gravimetrically with Fehling's solution under the standard
conditions laid down in a previous paper (Brown, Morris, and Millar,
Trans., 1897, 71, 94). Meanwhile a portion of the original starch
solution had been hydrolysed completely down to the final stage by
digestion for an hour at 50^ with a very active malt eztract added
at the rate of 5 c.c. or more per 100 c.c. of the solution.
This, of course, gave the mazimum reducing power of the fully
hydrolysed starch, and the ratios of the intermediate reductions to this
amount afforded a measure of the hydrolysis at each stage. It is scarcely
necessary to add that all the necessary corrections were introduced for
the reducing power and the volume of the malt eztract employed, for
the small initial reducing power of the soluble starch itself, and for
the changes of volume of the abstracted samples due to temperature.
From the ezperimental data so obtained, the ^ coefficient of velocity '
was either directly calculated or, as was generally the case, the results
were plotted out on a system of rectangular coordinates, representing
on the one hand equal time intervals, and on the other the proportion
of hydrolysable substance still left in solution. A perfectly even curve
was then drawn through the ezperimental points with the aid of a
flezible lath.
If we denote the cupric reduction of unit volume of a completely
hydrolysed starch solution as tm%, and take x as the ratio of the
reduction observed at any given time reckoned from the commencement
of the reaction, then, if the time curve representing the course of the
hydrolysis is logarithmic,
^logj-^.*,
* This is a necessary precaution, in order to avoid any slow changes in tl|«
activity of the extract after it has been added te the hot solution of starch,
Digitized by
Google
HYDBOLYSIS OF STARCH BY DIASTASE. 391
where A; is a constant representing the * coefficient of velocity * of the
hydrolysis, and 0 is the time elapsed.
The results of a few typical examples of experiments of this kind
are given in the Tables appended to the paper, the values of k deduced
from the above formula being given in the fourth column.
Table I records a transformation carried out at a temperature of
51 — 52° for more than 2^ hours, during which time, as will be seen
from the value of 1 — a; at the close of the experiment, 95 per cent, of
the full hydrolytic change had taken place.
It will be noted that throughout the reaction there is a steady
augmentation in the value of k. This is an invariable feature of all
our numerous experiments, and proves that the course of the
hydrolysis does not conform to the above logarithmic expression.
If we assume that the normal curve expressing the rate of change is
logarithmic, but that its form has been modified by secondary dis-
turbing causes, then the constant augmentation of k points to a set of
conditions which is producing a constant acceleration in the velocity
of change. In other words, within any given time interval there is
somewhat more of the residual substance hydrolysed than there should
Jbe according to the logarithmic formula.
In an early stage of the inquiry, when we still had reason to believe
that this was an exceptional instance of departure from the " mass
law," we were led by certain considerations to search for the disturbing
cause in the complex intermediate products of hydrolysis. We certainly
cannot account for the results by assuming that they are due to the
gradual accumulation of the products of change, with a consequent
tendency to chemical reversion, since any influence of this kind will
tend to diminish, not to increase, the velocity coefficient k.
Whatever difEerences of opinion may still exist as to the exact
nature of some of the intermediate products of starch transformation,
it is quite certain that, in course of its hydrolysis with diastase,
soluble starch, unlike cane sugar under the action of invertase or acids,
does not at once split up into its final products, maltose and a well
characterised dextrin. This stage is reached only through a series of
intermediate substances of the amylo-dextrin and malto-dextrin class.
Now it is quite conceivable, although it has never been proved, that
these intermediate substances may show a differential resistance to
hydrolysis. If the hydrolysis of the lower members of the series is
more readily effected than that of the products which stand in closer
genetic connection with starch itself, these differential properties of
the intermediate products would almost certainly show themselves by
some such augmentation of the velocity coefficient as we have observed.
We attempted to solve this question experimentally in the following
manner.
Digitized by
Google
BROWN AND GLENDINNING : THE VELOCITY OF
Two solutions, A and B, were prepared, containing the same amoant
of hydrolysable starch products, as measured by the amount of cuprio
reduction which the solutions gave on complete hydrolysis with an
activ^ malt extract. Solution A, however, contained starch products
which had been only very slightly hydrolysed, whereas B contained
more highly converted products, that is to say, a much larger
proportion of those intermediate substances which, on the above
assumption, ought to undergo hydrolysis more rapidly. To equal
volumes of these solutions were added exactly equal quantities of a
cold water malt extract, and the course of each reaction was followed
by the time method already described, care being of course taken to
keep all the conditions exactly the same for the two solutions. The
result certainly did not confirm the above hypothesis of the more ready
hydrolysis of the lower members of the series.
It was this failure to discover any reasonable explanation of the
augmenting value of k in starch transformations, and our consequent
inability to bring our observations into line with the analogous •
hydrolysis of cane '^sugar, which caused us to delay publishing our
results four or five years ago.
If we critically examine the results of the two starch trans-
formations given in Tables I and II appended to this paper, we see
that the time-curve expressing the rate of hydrolysis is approximately
represented by a straight line until from 30 to 40 per cent, of the
total hydrolysis is complete. This is well shown in the last columns
of Tables I and II, where we have given the increased amount of
oupric reduction for equal intervals of time. Within the limits
mentioned above, the amount of hydrolysis is approximtUelf/ pro-
parUanal io the time.
That the amount of transformation is, up to a certain point, a linear
function of the time, is also well shown in the following experiment,
where the observations have been restricted to a transformation of
only 15*5 per cent, of the total hydrolysable products :
Tran^armatian qfaS per cent, soluble starch solution with 1 e.c.qfnwU
extract per 100 c.c. Temperature 21^.
Time in minutes.
Amount iTansformed.
Total hydroly8iB=l.
Amount which should have
beenNJtTansformed if
hydrolysis is directly
proportional to the
time.
6
10
20
0-038
0-077
0-lW
0-076
0-162
Digitized by
Google
HYDROLYSIS OF STARCH BY DIASTASE.
398
In this oase, the linear expression is almost strictly correct, that is
to say, up to a point at which 15*5 per cent, is hydrolysed, we have
equal amounts of transformation in equal times.
We may go further than this, and ascertain the nature of the curve
after the linear elements of it have been passed. For this purpose, we
have taken the results given in Table I at the end of the paper, and
have commenced from the stage of the hydrolysis which was reached
in 40 minutes, when 0*4355 of the complete hydrolysis had been
attained, that is to say, when 0*5645 of hydrolysable substance was
left. Making this last value the new starting point, and taking it
therefore as unity, we obtain the following results on recalculation :
Old time units, miniite8.
Kew time units.
k.
40
60
10
0-00842
60
20
0-00881
70
80
0-00821
80
40
0-00887
90
50
0-00818
100
60
0-00807
110
70
0-00822
120
80
0 00840
130
90
0-00855
The values of k no longer show the steady increase which they do if
we take the commencement of the transformation as our starting
point, and their approach to equality shows that this part of the curve
is approximately logarithmie.
We are able therefore to analyse our time curves and divide them
into two parts, an earlier one which is Ztndor, and a later one which is
approximately logarUhmiOy the change from one expression to the
other not being abrupt, but gradual. In his paper on enzyme action,
Adrian Brown (this vol., p. 373) has brought forward a striking
number of facts showing that the time-curve representing the action
of invertase on cane sugar is net logarithmic, as had been previously
believed, but that the value of k steadily augments during the reac-
tion, just as we have found it do in the case of starch transformations.
He has also shown that with varying concentrations of sugar, all other
conditions remaining the same, approximately equal masses of the sugar
are hydrolysed in equal times, providing the comparisons are made at
an early stage of the hydrolysis ; in other words, the inversion up to
a certain point is a linear function of the time.
Still more recently, V. Henri {Gompt. rmd., 1901, 133, 891) has
also shown that the velocity of inversion of cane sugar with invertase
Digitized by
Google
894 BBOWN AND GLENDINNING : THE VELOCITY OP
increases more rapidly than is expressed by the ordinary logarithmio
formula of a unimolecular change.
He proposes the following empirical formula as more accurately re*
presenting the course of the reaction :
^log— - 2k,.
We have applied this formula to the results obtained in our experi-
ments on starch hydrolysis by diastase, and find that the values of k,
so obtained are much more nearly constant than those of k derived
from the original logarithmic expression, as will be seen on looking at
the values of k, given in the fifth column of the appended tables.
From what has been said, however, about the composite nature of the
time-curve, it is improbable that any single mathematical expression
will be found which is strictly applicable to all parts of the curve.
It is now perfectly clear that as regards the general course of the reac-
tions there is a close parallelism between starch hydrolysis with diastase
and inversion of cane sugar by invertase, and that the observed changes
in the velocity coefficient of starch hydrolysis are not necessarily con-
ditioned by the exceptional nature of the intermediate products. Both
reactions are linear in the early stages, and both are influenced by
variations in the concentration of the hydrolyte. The influence of de-
creasing concentration is to cause a larger proportion of the remaining
substance to be hydrolysed than would be expected from the applica-
tion of the ' mass law ' of a unimolecular change, provided always the
time units are reckoned from the commencement of the reaction. In
the case of starch transformations, the simple logarithmic formula is
fairly well applicable in dilute solutions when from 30 — 40 per cent,
of the hydrolysis is complete. The same will probably be found to
hold good with regard to cane sugar inversion.
Henri's formula is equally well applicable to both reactions.
Adrian Brown (foe. oU.) explains the results which he obtained in
his experiments with cane sugar by assuming that an appreciable time
elapses between the molecular union of the sugar and enzyme and the
actual hydrolysis, and that this time interval necessarily limits the
amount of work which the enzyme can perform, so that when the
ratio of sugar molecules to enzyme molecules is large, a certain
maximum amount of hydrolytic work is accomplished by the latter
which cannot be exceeded, and that no increased effect is conse-
quently produced by increasing the concentration of the sugar, all
other conditions remaining the same. He obtained apparent confirmap
tion of this by gradually decreasing the concentration of the sugar, the
enzyme remaining constant in amount. He then found, when a
certain low point of concentration had been reached, that the reac-
Digitized by
Google
HTDBOLYSIS OF STARCH B7 DIASTASE. 896
tion appioximately conformed to the mass law, that is, was inde-
pendent of concentration.
It appears to us that the time-curves representing enzyme action
can be explained in a somewhat different manner, and without pos-
tulating any differences in the time intervals between the succes-
sive stages of the reaction other than those due to variations in
the respective masses of the reacting substances existing in unit
volume.
It will simplify matters if we consider the ease of the inversion of
cane sugar, although the same argument, with a little modification, is
equally applicable to the hydrolysis of starch.
When cane sugar is inverted with dilute <ieid8 of different kinds,
but of the same molecular concentration, the velocity coefficients of
hydrolysis not only vary in the same order as the electric conductivi-
ties of the dilute acids, but there is also a remarkable numerical agree-
ment in the values representing the invertive action on the one hand,
and the electric conductivity on the other.
The striking general agreement between these two properties has
been emphasised by Ostwald, who has shown that it exists for a large
number of acids (see " Outlines of General Chemistry," p. 360).
This can be explained by assuming that the velocity coefficient of
inversion is a function of the number of molecules of electrolytically
dissociated ions per unit volume of the solution.
The number of free hydrogen ions in unit volume of the most dilute
acids which have hitherto been employed in such experiments must be
very large compared with the number of cane sugar molecules, and
under these conditions it might be expected that the course of the reaction
would be expressed, as it actually is, by the logarithmic formula of a
simple mass action, since for all practical purposes the cane sugar is the
only substance of^which the mass is changing. In the case of enzyme
hydrolysis, such a^ that effected by invertase, the apparent active
agent is a non-electrolyte, incapable of any appreciable dissociation,
and at first sight it would appear that notwithstanding the identity
of final products, there is some essential difference between acid and
enzyme hydrolysis. We believe this is only apparent, and that enzyme
hydrolysis is also brought about either by active water ions, or by
active water molecules dissociated from the inactive and large aggre-
gates of which the liquid mainly consists.
Pure water is in itself not a perfect non-electrolyte, but even if it
were, we are never dealing with pure water in such experiments, and
the remarkable influence of very minute but sensible amounts of acid
in intensifying the actidn of invertase and diastase is so well estab-
lished as in itself to suggest that water dissociation in some form or
Digitized by
Google
896 BROWN AND GLENDINNING : THE VELOCITY OF
other IB intimately bound up with the action of these and other
enzymes.*
The ordinary conditions of feeble acidity necessary for the complete
development of these enzyme actions are not sufficient to produce any
appreciable direct hydrolytic action on the cane sugar within the time
j>rdinarily occupied by such an experiment and within the limits of
temperature necessarily imposed. In such cases, however, thei*e can
be little doubt that hydrolysis is preceded by a combination of the
hydrolyte with the enzyme, and that this combination is much more
unstable and less able to withstand the action of the active ions or
dissociated molecules of the electrolyte than the original cane-sugar.t
According to this view these active ions are the true hydrolysts, not
the enzyme itself, which has only an intermediate function, and it is now
necesRary to consider how this way of looking at the facts will affect
our conception of the rate of change of the hydrolyte.
A solution of cane sugar undergoing hydrolysis by invertase must
contain the following substances :
A — the unaltered cane sugar.
a — ^the cane sugar in combination with the enzyme.
5— the added enzyme, a portion of which, h\ is at any one moment
in combination with a,
(a + V) — the combination of cane sugar and enzyme.
Oj — the products of hydrolysis of a.
X — the free water ions, or dissociated water molecules, which act as
the true hydrolysts.
It is assumed that, owing to favourable conditions, the number of
active ions, x, is very' large indeed compared with that of the other
reacting molecules, and further that there is no difference in the
respective velocities of the other reactions except those conditioned by
the varying masses of the reacting substances in unit volume.
If we wish to follow an inversion experimentally, it is always
* On the question of the influences of mere traces of acid on the action of diastase
and invertase, see Baranetzky, Die Stdrkeurribildenden FermenU in den P/Ujmxen,
Leipgig, 1878 ; O'Snllivan andTompson, Trans., 1890, 67, 855 ; Brown and Morris,
Trans., 1890, 67, 511 ; Fernbach, Joum. Fed. Inst, Brewing, 1896, 128.
The extreme sensitiveness of both diastase and invertase to minute quantities
of alkali in the solution becomes easy to understand if the hydrolytic action is con-
ditioned by the presence of electro-positive ions.
t There is a considerable amount of experimental evidence in favour of there being
a real combination of hydrolyte and enzyme prior to hydrolyaifi. The reoent work
of Emil Fischer, showing that there is some sort of stereochemical relationship
between hydrolyte and enzyme, seems to point to the possibility of the enzyme itself
holding the combined hydrolyte with some definite orientation, which facilitates the
action of the active water ionn upon it The enzyme may be regarded therefore as
the vice which holds the sugar molecule in a position favourable for the splitting
agent to act.
Digitized by
Google
HTDROLTSIS OF STABCH BTj DIASTASE. 397
necessary to refitiict its velocity^ and in-order ^to do t|iis the
ooncentration of b, the added enzyme, must be very small in relation
to the initial concentration A of the cane sugar. In the earlier stages
of the hydrolysis therefore, A in unit volume will be very large
compared with (a + 6'), the combination of cane-sugar and enzyme
present at any moment. But (a + &') must strictly speaking be
considered the starting point from which the hydrolysis conmiences,
the true hydrolyte, in fact, and the velocity of the inversion will
depend on the concentration of (a + b'). So long as the concentration
of the unaltered cane sugar A remains very large compared with
(a + &'), this latter will remain almost constant, and equal amounts of
inversion will take place in equal times : the time ' curve ' will in fact
be approximately a straight line. When, however, the concentration
of ^ is materially reduced and begins to approach that of (a + b') in
the order of magnitude, then, by the ordinary laws of mass action,
(a + b') will gradually get smaller, and the rate of inversion will more
nearly approach the logarithmic expression.*
This explanation accords very weU with all the known facts, as will
be seen from the following considerations.
We have seen from the results obtained by the hydrolysis of starch
by diastase that the first part of the time reaction is represented by a
straight line. This has also been found to be the case by Adrian
^rown as regards the inversion of cane sugar by invertase ; not only
do the first portions of the curves approach very nearly to straight
lines, but he also finds that, with equal concentration of invertase,
ponsiderable variations in the initial concentration of the sugar up
to a certain stage of the reaction do not materially affect the absolute
amount of sugar inverted in a given time. This we should expect
from what has already been said : when a certain point is reached, any
increase in il no longer sensibly influences the value of (a + b'), which
is really the regulating factor of the reaction, and so long as these
conditions exist we have equal amounts of hydrolysis in equal
Our hypothesis also affords an explanation of Kjeldahl's empirical
*Maw of proportionality," which may, in fact, be deduced from it.
Kjeldahl found that the relative diastatio power of two solutions is
expressed by the cupric reducing power produced in a given time when
the enzyme solutions act on the same weight of starch at the same
temperature, provided always the reducing power is not allowed to go
* We hare here, for the sake of simplicity, omitted any reference to the disturbing
influence produced by the accumulation of c^, the products of inversion. Towards
the close of a hydrolysis of a fairly concentrated solution this effect is unquestionably
noticeable, but we doubt whether it has much influence in determining the form of
th« curve of dilute solutions, such as we used for starch transformations.
Digitized by
Google
S98 teOWK AND OLEi^DINNING : THE VELOCflTY O^
beyond E 40 — 49.* This oorresponds to the hydrolysis of from 50 to
60 per cent, of the starch. It is clear, however, from an inspection
of Kjeldahl's own curves, expressing the amount of hydrolysis with
varying quantities of diastase, that the reaction is not by any means
rectihnear as far as R 40, although it may be sufficiently so for all
practical purposes of diastasimetry. It is, however, very nearly
straight as far as R 29, at which point about 36 per cent, of the full
hydrolysis is complete. This corresponds very closely with our own
observations, already referred to, on the straight part of the
curve.
The conditions of Kjeldahl's experiment amount to increasing the
concentration of b, and consequently of (a + V), leaving everything else
the same. Under these circumstances, the hydrolysis efEected in a
given time will be approximately proportional to the concentration of
(a + 6'), provided the reaction is not allowed to go so far that the con-
centration of A, the untransformed hydrolyte, is no longer able to
maintain the combination of the enzyme and hydrolysable substance
practically constant. This is guarded against by making the com-
parisons for cupric reduction within the time during whiish the course
of the reaction is practically rectilinear, and when, consequently, a
large amount of hydrolysable material still remains.
We can also make another important deduction from our hypothesis,
and predict that by largely increasing the relative amount of eniyme
to hydrolyte, that is to say, increasing the value of (a + &') in relation
to A, the more completely will the linear element of the time-curve be
eliminated, and the more nearly will the course of the whole reaction
be represented by the unimolecular formula jlog^ ^k.
In actual practice, it is not possible to go very far in this direction,
since the velocity of the action then becomes so great that the course of
the hydrolysis cannot be followed with sufficient accuracy. The re-
sults of Adrian Brown's experiments on the inversion of very dilute
solutions of cane sugar with fixed amounts of enzyme point, however,
to the correctness of this deduction, for he found under these condi-
tions that the "velocity-coefficient " was much less influenced by varying
concentration j in other words, there was a much nearer appoach, even
in the earlier stages of the reaction, to the simple logarithmic expres-
sion of a unimolecular change.
We must express our thanks to Mr. D. McCandlish for assistance in
the experimental part of this inquiry.
* In his original paper {MeddeUUer f, Carlsherg Lab., 1, French £Smm4^ p. 117)i
Ejeldahl has employed an older form of notation for expressing the reducing powers.
We have converted them into the usual form, which expresses the specific reduelBg
powers in terms of percentage of apparent maltose.
Digitized by
Google
ttYDROLYSIS OF STARCH BY DIASTASK.
S9d
In the following tables :
0sa the time units elapsed from the commencement of the experi-
ment,
ajssthe proportion of substance hydroljsed when 1 » total hydrolys-
able substance.
t = " velocity-coefficient " from formula ^^^gi ^ ^
0 1 —X
1 1 4-a!
^j« velocity-coefficient from Henri's formula -log ^2k..
Table I. — Tran^ormation of aS per cent, starch solution with 0*25 c.c
of malt-extract per 100 c.c of solution. Temperature, 51 — 52^.
e.
Amount of trans-
Time in
minutes.
X*
1-aj.
Je.
K
formation in equal
time intenrais.
10
0-1084
0-8916
0*00498
0-00472
20
0-2250
0*7760
000568
0-00497
0-1166
80
0*8850
0-6660
0-00590
0-00504
0-1100
40
0-4855
0-5645
0 00620
0*00506
01006
50
0-6850
0-4660
0*00650
0-00518
0*0995
60
0-6160
0*8850
0 00690
0-00518
0*0800
70
0-6800
0-8200
000706
0-00614
0*0650
80
0-7886
0-2615
0-00728
0-00514
0-0586
90
0-7800
0-2200
0-00780
0-00504
0-0416
100
0-8150
0'1850
0 00782
0 00495
0*0350
110
0-8600
0-1500
0-00749
0-00496
0-0860
120
0 8800
0-1200
0-00762
0-00497
0-0800
130
0-9030
00970
0*00779
0-00497
0-0230
140
0-9220
0-C780
0-00791
000497
00190
150
0*9400
00600
0-00814
0-00508
0-0180
160
0-9500
00500
0-00818
0-00492
0 0100
Tablb II. — Transformation of a 3 per cent, starch solution with 1 e.c. qf
maU-extract per 100 ex. of solution. Temperature, 21°,
e.
Amount of trans-
Time in
as.
1-x.
k.
^1.
formation in equal
time intervals.
minutes.
10
0-096
0*905
0*00438
0-00418
20
0191
0*809
0*00460
0-00419
0-096
80
0-286
0-716
0-00485
0*00424
0-094
40
0*378
0-622
0-00615
0*00431
0-098
60
0-466
0-686
0-00548
0*00487
0-087
60
0-648
0-452 .
0-00674
0-00446
0-088
70
0-628
0-872
0 00618
0-00457
0-080
80
0-708
0-292
^ 0*00668
0-00479
0 080
Digitized by
Google
400
BAKSB: THfi UNlOlf OP HTDROGXK AKD OXYQEK.
Tabls III.— ^^tw/brmo^ion qfaS per cent, eicureh sohuian with 1 e.e.
of malt-extrciet per lOO e.e. I\nfifMf«rffifi9, 21^.
9.
Time in
z.
1-x.
k.
*i.
minnteB.
10
0-081
0-919
000866
0-00852
20
0-108
0-887
0 00886
000857
80
0-288
0-762
0-00898
0-00861
40
0-808
0-692
0-00899
000846
44
0-884
0-666
0-00410
0-00842
50
0-878
0-622
0-00412
0 00845
60
0-440
0-660
0-00419
0-00841
70
0-506
0-494
000487
0-00845
Tabls IY. — Trar^farmatian qfaZ per cent, eta/rch aolutian with 1 ce.
qf mcUt-extraet per 100 c.c. Temperature, 21^
e.
Time in
X,
I'X.
h.
jfcx.
minutes.
5
0-084
0-966
0-00298
0*00296
10
0-072
0-928
0-00828
0-00818
20
0145
0*855
0 00889
0-00817
40
0-887
0-668
0-00446
0-00880
60
0-480
0-520
0 00478
0 00878
80
0-579
0-421
0-00469
000868
100
0-660
0-840
0-00468
0 00844
120
0-720
0-280
0-00460
0-00828
140
0-760
0-240
0-00448
0*00809
XXXVIII. — ITie Union of Hydrogen and Oocygen.
By H, Brsbbtok Bakeb, M.A.
The fact thatalarge number of chemical actions have been shown to be de*
pendent on the presence of aqueous vapour has led to many experiments
being made on the union of hydrogen and oxygen. Prof. H. B. Dixon
and Prof. Victor Meyer found no apparent diminution in the velocity
of this action when the gases were dried. I have myself obtained the
same result in a number of experiments carried on during the last 10
years. Recently, however, a method for the preparation of very pure
Digitized by
Google
bakeb: the union op hydeogen and oxtqen. 401
hydrogen and oxygen was devised by Mr. F. R. L. Wilson and myself.
Some years ago, Prof. Edward Morley pointed out the various im-
purities which were present in the gas produced in the electrolysis of
sulphuric acid and potassium hydroxide solutions. Lord Rayleigh
{Proe. Roy» Soc.y 1889, 45, 425), using the latter method, attempted to
minimise the amount of hydrocarbon impurity derived from the
presence of carbonate in the potassium hydroxide by the addition of a
small quantity of barium hydroxide. Since, however, the action of
barium hydroxide on potassium carbonate is a reversible one, and since
the potassium hydroxide is present in large excess, the precaution was
probably not a very effective one. The new method consists of the
electrolysis of highly purified barium hydroxide. It has been shown
that barium carbonate is insoluble in a solution of the hydroxide, and
that hydrocarbons are absent from the hydrogen. At the same time,
the oxygen is apparently quite free from ozone and hydrogen peroxide.
The use of this method suggested a new series of experiments on the
union of hydrogen and oxygen, in which a new precaution has been
found necessary. In a series of experiments, described later in the
paper, it is shown that the undried hydrogen and oxygen combined
very slowly in sunlight. It is, then, obviously necessary that the
drying must take place in the dark, since the rate of the continued
production of water in the light might conceivably be equal to the
rate of its absorption by the drying agent.
In order to find if moisture had any effect on the combination of the
dried gases, the following procedure was followed. It may be pointed
out that the reaction is a very sensitive one, and the omission of
any one of the precautions taken will almost certainly lead to the
failure of the experiment. The glass used has been either hard Jena
or hard Bohemian glass. Only one experiment has been made with
soft glass ; in it, combination took place at a low red beat. The tubes
were bent into the shape indicated by Fig. 1 (p. 402). They were then filled
with a mixture of nitric and chromic acids and boiled in the water-bath
for 24 hours. After this cleansing, the tube was washed out with dis-
tilled water and fitted to the tapered end of the platinum tube of a con-
denser. Purified water was then distilled through the tube for an hour.
After allowing it to drain, the tube was dried by heating it to redness,
while a current of air, dried by distilled sulphuric acid, was drawn
through it. A plug of distilled phosphoric oxide was introduced, the
upper end of the tube drawn off, and the lower end drawn out to a
capillary of about 0*6 mm. diameter. A small piece of fusible metal
was introduced and the tube was exhausted. It was then connected
with the electrolysis apparatus, the mixed gases being roughly dried by
passing them through a phosphoric oxide tube a foot long. The barium
hydroxide used had been recrystallised fifteen times ; it was found not
VOL. LXXXI. BE
Digitized by VjOOQIC
402 baker: the union op hydeogen and oxygen.
to be radioactive. When the tube was full of gas, the fusible metal
was melted and allowed to run into the capillary. When the wall of
the capillary is thick there is no fear of the glass being cracked by the
expansion of the metal. When the metal was cold, the outer portion of
Fio. 1.
>
Fio. 2.
lUI
Phosphoric oxide
Fusible metal
the capillary was sealed in order to prevent any possibility of leakage
round the metal. The tubes were then allowed to stand in the dark
for varying periods. Comparative tubes were made at the same time
from the same length of tubing and treated in precisely the same
way, except that no phosphoric oxide was sealed up in them.
Digitized by
Google
BAKER: THE UNION OF HYDROOEN AND OXYGEN. 403
After 10 days' drying, two such tubes were heated side by side in the
saine Bnnsen burner flame. In twelve experiments, the wet tube
exploded and the dry tube did not. In only one experiment has a dry
tube exploded, but in this case the tube had been carried for some
miles by hand, and most probably some of the phosphoric oxide had
been shaken into the part of the tube which was heated. In two
experiments, where only 2 days' drying had boQA allowed, water was
slowly formed in the dried tube, but although visible moisture was
present, no explosive combination took place, and a slow combination
only occurred. In each of the twelve experiments mentioned, the dried
tubes were opened under mercury, a very small contraction was ob-
served in some tubes, in most of them none at all. On addition of a
email quantity of distilled water, the contents of each tube exploded
on bringing them to a flame.
In order to see if increasing the temperature beyond the ordinary
temperature of explosion (600^, Y. Meyer and Krause, Armalmy 1891,
294, 85) produced any effect, a thin coil of silver wire was attached
to platinum wires by fusion, and the latter sealed through the walls of
a hard glass tube (Fig. 2, p. 402)u The tube was then dried as
before and filled with the explosive mixture. It was found that no
explosion took place, even when the silver was heated to its melting
point by a current passed through it. No contraction was observed
on opening the tube under mercury. The silver wire was drawn with
great precaution from silver prepared for atomic weight purposes by the
ammonium formate method.
Since a temperature of over 1000^ was insufficient to bring about
the union of the gases, a coil of thin platinum wire was substituted
for the silver in the last experiments. The coil was cleaned in situ
by nitric and chromic acids. It was heated by the current while the
tube was being dried. After the admission of the explosive mixture,
the tube was left attached to the mercury gauge for some hours.
Even in its very imperfectly dried state, the gases seemed to be able
to resist the catalytic action of the platinum. The tube was sealed as
before, and left for ten days. At the end of this time there was no
appearance of moisture on the surface of the phosphoric oxide. As
this is a very delicate test for the presence of traces of water, it may
be asBimied that cold platinum wire has no effect on the dried mixture.
The temperature of the platinum coil was raised by an electric current,
and just after reaching visible redness the mixture exploded. The
catalytic action of platinum is therefore apparently sufficient to bring
about the union of the dried and purified gases at a low red heat.
Some twenty eizperiments have been made in order to find if electric
sparks could be passed in the dried gas without explosion. In only
one case has this been done, and only with extremely small sparks.
E E 2
Digitized byVjOOQlC
404 baker: the union of hydrogen and oxygen.
These sparks were accidentally obtained in a tube fitted with a silver
coil, which, however, was not fused to the platinum wires, but only
allowed to hang on them by Its hooked ends. On passing a direct
current through the coil, and shaking the tube at the same time,
sparks were obtained at the points of contact of the two metals. No
apparent combination resulted, although the process was repeated often
in the course of the day. Next day the gases had become too dry to
allow of the very small spark discharge passing through them. Other
experiments in which sparks of less than 0*1 mm. from an induction
coil were used, have, so far, always resulted in the explosion of the
With regard to the explanation of the diminution of chemical
activity in dried gases, it has always been asserted by Dr. Armstrong
that, without an electrolyte, no chemical action is possible, and that
the effect of removing water is only the removal of the possibility of
an electrolyte being formed. This hypothesis is borne out in a very
striking way by the behaviour of the partially dried gases. When
they are heated, water is slowly formed, and although it is then
present in enormously larger quantity than is necessary to bring
about the action, no explosion takes place. It may be assumed that
the water formed by the union of the very pure gases is itself very
pure, and since pure water is not an electrolyte, then this water
should not cause the explosion of the gases.
It has always seemed possible that this theory of Dr. Armstrong's
should be pushed a step further by assuming that when water is
present in the gases of ordinary purity, union can only take place by
means of the ions produced by the dissociation of the gases. This
extension of the ionic theory to moist gases seems to fit the ex-
perimental evidence fairly well, but it may be considered, a priori,
that it is improbable that the very small quantity of water necessary
[less than 3 mg. per 1,000,000 litres, according to Professor E. Morley's
approximation (Amer. J, Set,, 1889, 34, 200)] could act in the same
way as liquid water. The fact, however, that feeble electric discharge
cannot pass through dried gases lends support to this view.
I attempted to put the question to experimental trial by finding if
any contraction in volume took place when a gas was dried from the
point at which its moisture ceases to exert any measurable tension to
the point at which the gas ceases to be chemically active. To effect
this, a long tube shaped like a syphon barometer (Fig. 3) was carefully
cleaned and dried as in the experiments described above. A sufficient
quantity of mercury and a thin tube of distilled phosphoric oxide were
introduced into the cylindrical bulb. The long tube was then ex-
hausted, the mercury being boiled by playing on the surface of the
tube with the flame of a Buns en burner. The end of the barometeir
Digitized by
Google
baker: the union op hydrogen and oxygen. 406
tube was then sealed, the lower cylinder was then exhausted and filled
with the gas, previously dried by a long phosphoric oxide tube. In
the case of the explosive mixture, the side tube, opening under mercury,
was made use of for introducing the gases, so that the mercury in the
narrow tube served to disconnect the main bulb from the part of the
tube which was heated in sealing. Six tubes were prepared in this
way containing (1) hydrogen, (2) oxygen, (3) nitro-
gen^(4) air, (5) a mixture of hydrogen and oxygen ; ^^®' ^*
the sixth tube contained air without phosphoric oxide,
and served for a standard. The tubes were rigidly
fixed to a heavy iron stand on which they were wholly
immersed in a deep tank with glass sides. A current
of water was passed through the tank for two hours
befoi*e the readings were taken, the water being ob-
tained from a supply cistern. The readings were
taken late at night, some 18 hours after the daily
filling of the supply cistern. The apparatus was
placed in a cellar without any outside walls, in
which the temperature is remarkably constant. A
thermometer reading directly to 0'01° was hung in
the moving water. Headings were taken nightly for
a month, the upper level of the mercury in the
barometer tubes being read off by a cathetometer
against a glass scale attached to the glass window of
the tank. A difference of 0*1 mm. could be easily
read, and by a simple calculation the difference^of
temperature is allowed for. Absolutely no difference
was noted for the first month, and afterwards weekly
readings were taken for six months, still with the
same result. Hence a gas, on drying it to the
utmost limit possible, does not change its volume by
l/7000th, which is about the limit of error. This
experiment does not, of course, prove that dissocia-
tion does not take place in moist or partially dried
gas. It only shows that, if it does occur, it does not
take place to the extent measured in the experi-
ments. There is a point which is worthy of note,
that dried hydrogen and oxygen do not combine to a measurable
extent at the ordinary temperature in the dark.
In order to find if sunlight had any effect on the union of these
gases, two tubes were prepared as in the last series of experiments.
One contained the dried, and the other the undried, mixture of gases
from the electrolysis of barium hydroxide. After reading, they were
exposed outside a south window for four months, from September to
Digitized by
Google
406 FOBSTEB AND KlCS:L£TBWAtT :
December. At the end of this time, readings were taken which
showed no alteration in volume of the dried gas, whilst the undried
mixture showed a contraction of one-twenty-third of its volume.
Hence the precaution mentioned in the earlier part of the paper of dry-
ing the mixed gases in the dark was shown to be necessary.
General Condttsiana.
1. The gases produced by the electrolysis of purified barium hydr-
oxide do not explode on heating to redness after drying with distilled
phosphoric oxide.
2. The gases can be heated to the melting point of silver without
combination.
3. If only partially dried, the gases unite slowly on heating, and
although visible water is present, no explosion takes place.
4. The undried gases unite slowly in sunlight at the ordinary tem-
perature, the dried gases do not»
5. There is no contraction observable during the thorough drying of
gases, so that the dissociation of gases in the undried condition, if it
exists, cannot be proved by volume measurements.
In conclusion, I wish to give my best thanks to Dr. Armstrong for
much encouragement and advice which he has given me during the
progress of the research.
DULWICH OOLLEQB, S.E.
XXXIX. — Studies in the Camphane Series. Part VIIL
TXi'Nitrobenzot/lcdmphor.
By Mabtik Onslow Forster and Frances M. G. Micklethwait.
Ketones of the aliphatic series and of hydrogenised cyclic compounds
have not been observed to form enolic modifications capable of separate
existence. The tendency to change into a hydroxylic isomeride is first
noticed when hydrogen^ combined with a carbon atom adjacent to the
carbonyl group, is replaced by an acyl radicle. During his investigation
of 1 : 3-diketones, Claisen found that this disposition to undergo enol^
isation is influenced by the nature of the acyl substituent^ becoming
more pronounced as the negative character of that radicle increaaoB
{AntuUm, 1896, 291, 37).
In view of this observation, we have prepared m-nitrobenzoylcam-
phor with the object of ascertaining whether that substance would
Digitized by
Google
SttTDIES IN THE CAMPHANS SERIES. PART VIIT. 407
form, with benzoyloamphor and camphor itself, a series exhibiting
the gradation of properties displayed by triacetylmethane, diacetyl-
benssoylmethane, acetyldibenzoylmethane, and tribenzoylmethane (loc*
Camphor. Benzoylcunphor. Kitrobenzoylcamphor.
On comparing benzoyloamphor with the nitro-derivative, it is found
that the difference between them is almost as great as that sub-
sisting between camphor and the benzoyl derivative. Camphor does
not change into the enolic modification ; benzoyloamphor, although
stable in the solid form, rapidly changes into the isomeride when
dissolved in chloroform, whilst m-nitrobenzoylcamphor shows so little
disposition to undergo this transformation that it has not been found
possible to obtain it in the ketonic form. That a small proportion
of enolic m-mitrobenzoylcamphor becomes converted into the ketone
is shown by a diminution in the specific rotatory power of a solution
in chloroform; this amounts to less than 6 per cent, but until the
neutral modification can be isolated it will not be possible to state
the percentage of ketonisation represented by this change in optical
activity.
Two methods suggested themselves for the preparation of a nitro-
benzoyloamphor. In the first place, sodium camphor might be treated
with nitrobenzoyl chloride, which would probably give rise to the nitro-
benzoyl ester of enolic nitrobenzoylcamphor ; this compound could be
converted by hydrolysis into a nitrobenzoylcamphor, in which the nitro-
group would occupy a known position. The alternative process consists
in subjecting aa-benzoylbromocamphor to the action of nitric acid, the
product, if a mononitro-derivative, being then reduced with alcoholic
potash ; the position of the nitro-group in the resulting nitrobenzoyl-
camphor could be ascertained by oxidation, which should give rise to
one of the nitrobenzoic acids.
On consideration, the latter method appeared the more promising, and
was therefore adopted. £ach isomeric aa-benzoylbromocamphor, when
treated with fuming nitric acid, yields a nitrobenzoylbromocamphor
which, on reduction with alcoholic potash, is converted into enolic
nitrobenzoylcamphor ; potassium permanganate is immediately reduced
by an alkaline solution of this compound, which is thus resolved into
camphoric and m-nitrobenzoic acids.
^»^"<§:0H^*^**^°' -^ C^,,<^«| + CO,H.C.H,.NO,.
The n»-nitrobenzoylbromocamphors described in this paper are related
to one another in the same way as the benzoylbromooamphors from
Digitized by
Google
408 FORSTER AND MICKLBTHWAIT !
which they are derived. As already stated, the ketonic modification
of m-nitrobenzoylcamphor has not been isolated, and consequently the
influence on rotatory power exerted by the m-nitrobenzoyl radicle
could not be compared directly with the effect produced when bromine
is the substituent. We therefore refer to the derivative of lower
melting point as a -m-nitrobenzoyl-a-bromocamphor ; this componnd
has [a]D +87*9^ in chloroform and melts at 93 — 94% whilst a-m-nitro-
benzoyl-a-bromocamphor has [ajo — 26'P, and melts at 101 — 102^. If
this nomenclature is uniform with that adopted in the case of the
benzoylbromocamphors, it may be inferred that the optical influence of
the m-nitrobenzoyl group is less powerful than that of the benzoyl
radicle (compare Forster and Micklethwait, this vol., 163). The effect
of the benzoyl radicle has been found to exceed that of a chlorine atom
occupying the same position {loe. cU.), and is in turn exceeded by that
of the bromine atom ; it is probable that the optical influence of the
m-nitrobenzoyl radicle is inferior to that of the chlorine atom, because
a-m-nitrobenzoyl-a-chlorocamphor, which melts at 72 — 74% h%s
[a]D +40-4% whilst the isomeride melting at 110° has [a]© +7-1°
It will be noticed that in yielding a m-nitro-deri vative, the aa-benzoyU
bromocamphors and oa-benzoylchlorocamphors conform to Crum Brown
and Gibson's rule (Trans., 1892, 61, 367), but although the m-nitro-
compounds are produced almost exclusively, a small proportion of the
o-nitro-derivative also is formed sometimes when the aa-benzoylbromo-
camphors are nitrated. We have been thus enabled to prepare enolic
o-nitrobenzoylcamphor, identified by oxidation with potassium perman-
ganate, which resolves it into camphoric and o-nitrobenzoic acids.
A comparison of the aa-nitrobenzoylbromocamphors with the corre-
sponding chloro-derivatives has revealed one point of some interest.
Whilst the first-named substances, on reduction with alcoholic potash,
yield enolic m-nitrobenzoylcamphor and potassium bromide,
the aa-nitrobenzoylchlorocamphors, on similar treatment, are resolved
into a-chlorocamphor and potassium m-nitrobenzoate ;
OA.<S"°°'*''"*""''*". -^ C.H„<g«*00,H.O.H,.NO,
This affords an interesting example of the relative success with which
derivatives of these two halogens resist the action of a reducing agent.
Digitized by
Google
STUDIES IN THE CAMPHANE SERIES. PART VIII. 409
EZPBBIHBNTAL.
aa-mrNUrobeTizaylbromoeampfior, OgHj^v^ JL. •
Owing to the readiness with which fuming hydrobromic acid converts
a -benzoyl-a-bromocamphor into the isomeride, we were not prepared to
find that the corresponding nitro-derivatives could be^ obtained by the
direct action of fuming nitric acid, expecting that the unstable benzoyl*
bromocamphor would yield the nitro-derivative of the stable isomeride.
The earlier experiments were therefore conducted with a mixture of
the two benzoylbromocamphors, but it was soon found that the product
was not an individual substance, being resolved by fractional crystal-
lisation into specimens of diiEFerent specific rotatory power.
The two isomeric substances were therefore nitrated separately. Ten
grams were covered with 30 c.c. of fuming nitric acid (sp* gr. 1*52),
which dissolved the compound and became warm ; after an interval of
about 20 minutes, the liquid was poured into a large volume of cold
water, the precipitated nitro-derivative being filtered, washed, and
crystallised from methyl alcohol.
a-m-Nitrobenzoyt-a-bromocamphor, prepared from a -benzoyl-a-bromo-
camphor and fuming nitric acid, separated in the form of a yellow,
sticky mass on pouring the acid liquid into a large volume of water ;
it did not harden on continued washing or after an interval of several
days. The substance was obtained in crystals by dissolving in the
minimum quantity of hot methyl alcohol, allowing the liquid to cool,
and decanting the clear solution from the yellow oil which separated ;
the deposit from this solution was recrystallised twice from hot methyl
aloohol, from which it separates in long, flat, prismatic needles, pale
yellow in colour, and melting at 93—94° :
0-2636 gave 0-4994 COj and 0-1120 B^O, 0 = 53-70; H-4-95..
0-2382 „ 0-1179 AgBr. Br = 21-06.
Oi^HjgO^NBr requires 0 = 53-66; H = 4-78; Br = 21-06 percent.
A solution containing 0*4906 gram dissolved in 25 c.c. of chloroform
at 21° gave a^ + 3°27' in a 2-dcm. tube, whence the specific rotatory
power [o]d +87-9°.
An attempt was made to convert the substance into a-m-nitro-
benzoyl-a'-bromocamphor by the agency of hydrobromic acid (sp. gr.
1 '83), but a specimen which was allowed to remain in contact with
the agent during 24 hours underwent no change in specific rotatory
power during this period.
a-m-NUrobenzoyl^' -bromocamphor separated as a hard, granular
precipitate wlien the solution of a-benzoyl-a'-bromocamphor in fuming
Digitized by
Google
4iiO FORSTER AND MICKLETHWAIT i
nitric acid was poured into a large volume of cold water ; after being
recrystallised twice from methyl alcohol, it was obtained in aggre-
gates of small, pale yellow needles melting at 101 — 102° :
0-5036 gave 15-6 c.c. nitrogen at 17° and 757 mm. N-3-45.
0-2288 „ 0-1124 AgBr. Br = 20-90.
Oj^HigO^NBr requires N«3*68 ; Br = 21 05 per cent.
A solution containing 0*4624 gram dissolved in 25 ac. of chloroform
at 21° gave aj, —58' in a 2-dcm. tube, whence the specific rotatory
power [ajo -261°.
Enolio m-NUrohmzoylcamphor, Q^ll^^<^{^^'^^^*^^\
Twenty-seven grams of nitrobenzoylbromocamphor were dissolved ia
150 c.c. of alcohol which had been distilled from caustic soda; the
solution was then heated in a reflux apparatus with 8*5 grams of
potassium hydroxide during 1^ hours, after which the alcohol was
distilled off and the residue dissolved in about 100 c.c. of water. The
clear, pale red solution was saturated with carbon dioxide, which
precipitated a pale yellow solid, slightly sticky at first, but rapidly
becoming hard. The product was washed several times with water
and crystallised twice from hot alcohol :
0-3037 gave 11-7 c.c. nitrogen at 16° and 764 mm. N = 4-60.
CiyHj^O^N requires N = 4-65 per cent.
The substance dissolves readily in alcohol, from which it crystallises
in long, pink, silky needles which melt at 106 — 107° ; it is insoluble in
cold petroleum, and not very soluble in the warm liquid, from which
it separates in rosettes of needles. Aqueous alkalis dissolve the com-
pound, which is precipitated from the solution by carbon dioxide.
Ferric chloride develops an intense purple coloration with alcoholic
solutions, and a green precipitate is formed with copper acetate. A
solution in chloroform immediately decolorises bromine, and potassium
permanganate is reduced by cold solutions in alkali hydroxide.
m-Nitrobenzoylcamphor has not been obtained in the ketonic modi-
fication ; a specinien was heated with boiling formic acid in a reflux
apparatus during 2 hours and then precipitated by water, but the
product dissolved readily in alkalis and developed an intense coloration
with ferric chloride. The only evidence of a tendency to undergo
transformation into the isomeride is the slight change exhibited by the
specific rotatory power of a solution in chloroform during an interval
of several hours.
0*4892 gram dissolved in 25 c.c. of chloroform at 20° gave ai> 8°12'
in a 2-dcm. tube, whence the specific rotatory power [a]i> + 209*5°,
Digitized by VjOOQIC
STlTDtlBS IN TMB CAMPltANB SERIBIS. PART VIII. 4ll
which in the coarse of Beveral days diminished to [ajo +200*1^,
remaining constant at that point.
The same slight reduction in specific rotatory power is immediately
effected by adding a single drop of piperidine to the solution.
The ctcetyl derivative was prepared by heating the substance with
acetic anhydride in a reflux apparatus during 2 hours ; the liquid was
then poured into a large volume of cold water, the viscous product
being thoroughly washed with water and treated with a small quan-
tity of cold alcohol :
0-2826 gave 103 c.c. nitrogen at 14° and 774 mm. N-4-36.
CigHj^OjN requires N = 408 per cent.
The substance is readily soluble in alcohol, and is most conveniently
crystallised from warm light petroleum, which deposits it in clusters of
pale brown needles melting at 127 — 128°. The alcoholic solution is in-
different towards ferric chloride.
Ganveriion into m-yUrobenzoi/lbromoeamphar. — ^The unsaturated
character of enolic benzoylcamphor is reproduced in the nitro-deriv-
ative, which is converted by bromine into m-nitrobenzoylbromo<
camphor. A solution of 3*6 grams of bromine in glacial acetic acid
was added to 6*1 grams of the nitro-compound dissolved in glacial
acetic acid containing 2*6 grams of dried sodium acetate in solution.
The colour of the halogen was destroyed, and on pouring the liquid into a
large volume of cold water a somewhat sticky precipitate was obtained ;
this was washed several times with water and crystallised from methyl
alcoholy which deposited pale yellow needles melting at 90 — 94° and
giving [aji) + 59 0° in a 2 per cent, solution in chloroform. The product -
was therefore impure a -m^nitrobenzoyl-a-bromocamphor.
Oxidation of m-NUrobenzoylcamphor, — An alkaline solution con-
taining 3*6 grams of enolic m-nitrobenzoylcamphor was treated with
300 C.C. of a 2 per cent, solution of potassium permanganate, which
was added in small quantities at a time to the cooled liquid. The
deep green solution was warmed, and treated with alcohol until the
manganate was completely reduced, the hydrated oxide being then
filtered. The liquid having been evaporated to a small bulk and the
crystallised potassium sulphate removed, dilute sulphuric acid wai^
added until no further precipitation occurred. On dissolving the mixed
acids in ammonia and adding lead acetate to the hot neutral solution,
a bulky, white precipitate was formed ; this was filtered and extracted
with boiling water until the washings gave only a faint coloration
with ammonium sulphide.
The precipitated lead salt was then treated with a boiling solution
of sodium carbonate, filtered from lead carbonate, evaporated to small
bulk, and acidified with dilute sulphuric acid. Camphoric acid was
Digitized by
Google
412 FORSTER AND MICKLETIIWAIT :
thus obtained, and after being crystallised twice from boiling water,
melted at 183—184''.
The filtrate from lead camphorate was united with the washings,
evaporated to a small bulk, boiled with sodium carbonate, filtered, and
acidified with dilute sulphuric acid, which precipitated m-nitrobenzoic
acid {uL p. 141°).
^nolie o-NUrobenzo^lcamphar, O^Tl^^<P\*^^'^^^^'^^^.
On reducing with alcoholic potash a specimen of nitrobenzoylbromo-
camphor obtained from evaporated mother liquors, it was noticed that
the pale yellow needles of enolic m-nitrobenzoylcamphor were associated
with pale brown, transparent prisms; the same substance was
occasionally produced in small and uncertain quantity from re-
crystallised nitrobeuzoylbromocamphor. After recrystallisation from
alcohol, it melted at 118°:
0*2524 gave 10*4 c.o. nitrogen at 18° and 771 mm. N»4*82.
CjyHjgO^N requires N =» 4*65 per cetit.
A solution containing 0*4647 gram in 25 cc of chloroform at 21^
gave ai> + 1°39' in a 2-dcm. tube, whence the specific rotatory power
[oJd +44*5°; after an interval of three days, the same solution gave
[a]o +60*5°, remaining constant.
At first it seemed probable that the substance just described was
the ketonic modification of m-nitrobenzoylcamphor, but it was soon
found that alcoholic solutions develop colour with ferric chloride and
yield a precipitate with copper acetate; moreover, it dissolves in
alkalis, and generally resembles enolic m-nitrobenzoylcamphor in
chemical behaviour. It was ultimately identified as enolic o-nitro-
benzoylcamphor by oxidising it with potassium permanganate under
the conditions which convert m-nitrobenzoyl camphor into camphoric and
m-nitrobenzoic acids. In this manner, camphoric acid was obtained
in association with o-nitrobenzoic acid (m. p. 148°).
.... ,., r ^„ ^9C!l*00-CeH,-N02
aa-mrlvttrobenzoi/khlorooamphara^ ^8-^i4^<v)
The benzoylchlorocamphors were dissolved in fuming nitric acid, the
solution, after .an interval, being poured .into a large volume of cold
water. As in the case of the corresponding bromo-derivatives, the
product from the benzoylchlorocamphor of the higher melting point
was much more granular in the crude state than the isomeride, which
exhibits a tendeocy to remain sticky.
Digitized by
Google
STUDIES IN THE CAMPHANE SERIES. PART VIII. 413
a-m-NitroUnzayl-a'^shlorooafnphcr crystalliBes from alcohol in aggre-
gates of pale yellow prisms and melts at 72—^74° :
0-2962 gave 11-3 c.c. nitrogen at 11° and 760 mm. N = 4-54.
0-2024 „ 00890 AgCl. 01=10-88.
CiyHigO^NCl requires N»417 ; 01 = 10-58 per cent.
A solution containing 0-6616 gram, dissolved in 25 c.c. of chloroform
at 20°, gave ai> + 1°49' in a 2-dcm. tube, whence the specific rotatory
power [a]D +40-4°.
am- NUrobenzoyl-a-chlarocamphor separates from alcohol in small,
nearly colourless needles melting at 110°:
0-2533 gave 9 '6 c.c. nitrogen at 13° and 755 mm. N»4*44.
0-1824 „ 0-0770 AgOl. 01-10-44.
Oi^HigO^NOl requires N « 417 ; 01 = 10-58 per cent.
A solution containing 0-4379 gram dissolved in 25 c.c. of chloroform
at 21° gave a^ + 15' in a 2-dcm. tube, whence the specific rotatory
power [a]D + 7-1°.
f Reduction of m-Niirobenzoylchloroeamphor with Alooholie Potash,
Thirteen grams of a -m-nitrobenzoyl-a-chlorocamphor were dissolved
in alcohol and heated in a reflux apparatus with 4 grams of potassium
hydroxide during 1^ hours ; alcohol was then removed on the water-
bath and the residue treated with water. Oarbon dioxide was then
passed through the liquid, and the precipitate, a portion of which had
appeared on first adding water, filtered, washed, and crystallised
twice from alcohol. It was then found to contain chlorine, but no
nitrogen :
01958 gave 01527 AgOL 01 = 19-29.
OioHjjOCl requires 01 = 19*11 per cent.
The substance crystallised in thin, lustrous, white plates melting at
92°, and gave [ajo + 95° in chloroform ; it v^as thus identified as
a-chlorocamphor.
The aqueous filtrate from chlorocamphor was acidified with hydro-
chloric acid, which precipitated f7»-nitrobenzoic acid melting at 139°.
Reduction of the Benzoylchloroccmphora with Alcoholic Potash,
The unexpected difference in the behaviour of the m-nitrobenzoyl-
chlorocamphors and m-nitrobenzoylbromocamphors towards alcoholic
potash led us to inquire whether the benzoylchlorocamphors are dis-
tinguished from the benzoylbromocamphors in the same respect. It has
beec^ already ascertained (this voL, 165) that the benzoylbromocamphors
Digitized by
Google
414
MELLOR AND ANDERSON : THE UNION OF
yield enolic benzoylcampbor when reduced with alcoholic potash. On
subjectiDg the l>enzoylcblorocamphor8 to this treatment we found that
a-benzoyl-a-chlorocamphor yields enolic benzoylcampbor, a-chloro-
camphor, and benzoic acid, whilst the two last^uamed substances alone
are produced when a-benzoyl-a'-chlorocamphor is reduced.
RoTAL College of Scienoe, London.
South Kensington, S.W,
XL. — The Union of Hydrogen and Chlorine. Part IV.
The Draper Effect.
By J. W. Mellob and W. B. Anderson.
In 1843, Draper* {Phil. Mag., 1843, [iii], 23, 403, 416) published a
very curious observation to the effect that if light from an electric
spark is allowed to fall upon a mixture of equal volumes of hydrogen
and chlorine gases, the volume of the mixture suddenly expands and
immediately returns to its original condition. In the first part of this
work, this phenomenon was called the Draper effect.
Priugsheim rediscovered the momentary expansion in 1887 and
considered that it was in no way analogous to the Budde photo«
expansion, since chlorine alone does not expand under the same
conditions.
The Draper effect is best demonstrated in the following apparatus :
Fig. 1.
Insolation vessel.
ThQ mixed gases are contained in a flat glass bulb, A, called the
insolation vessel. The lower part of the insolation vessel usually
contains some water saturated with the two gases. The capillary tube,
BG, contains a thread of liquid (ae) to serve as an index. Under
the influence of a flash of light, the thread of liquid (ac) is pushed
outwards to return immediately to its original position. Thus, a
travels to b and immediately returns to a, Bunsen and Rosooe's method
* After this, only historical references omitted in the earlier parts of this work
will be giren here.
Digitized by
Google
HYDROGEN AND CHLORINE. PART IV. 415
is the best way of preparing the gases. The spark from an induction
coil, intensified by means of a Leyden jar battery is the source of light,
which may be from 10 to 20 cm. away from the insolation vessel.
In 1897, Wild and EJArker {Electrician, 1897, 38, 690) found that
sparks from a Wimshurst machine were as active as those derived
from the coil, and that the magnitude of the effect varied directly as
the visual brightness of the spark.
Although Hertz has shown that chlorine and most coloured gases
and vapours partially absorb ultra-violet rays, Wild and Marker did
not succeed in detecting any action which could be attributed to the
presence of ultra-violet rays of light. The interposition of a layer of
any substance, like glass or mica, opaque to the ultra-violet rays made
no perceptible difference to the effect obtained. Wild and Harker heoce
conclude that the Draper effect is not due to the absorption of ultra-
violet radiations by the gaseous mixture. Fringsheim's negative
result with chlorine also confirms this conclusion.
Dixon and Baker (Trans., 1896, 69, 1308 ; compare Rzewuski,
Wied. Biebl, 1896, 20, 1016 ; Hemptinne, Zeit. phyaxkaZ. Chem., 1896^
21, 493; J. J. Thomson, Proc. Camb. FhU, Soc., 1901, 11, 90) have
obtained negative results with Eontgen radiations. J. J. Thomson
{loc, cU.) has found thorium radiations do not perceptibly influence the
magnitude of the Draper effect. He also failed to detect any free ions
in the gas under conditions which would have enabled him to observe
one in 10^^ of the molecules present.
In addition to these isolated observations, we have o bserved the
following facts:
Influence of VaricUion in the Composition of the Hlvminated Gas,,
We have been unable to detect the Draper effect when light is
flashed upon :
1. Chlorine gas under pressures varying from one to half an
atmosphere :
2. Chlorine gas at atmospheric or half the atmospheric pressure and
heated from 15° to 100° :
3. Dry chlorine, or chlorine saturated with steam : *
4. When chlorine is mixed with half its volume of nitrogen, air,
carbon dioxide, carbon monoxide, f or methane.]:
* One of ua has shown that hydrogen chloride is formed when a miztore of
chlorine and water vapour is strongly illuminated by an arc light.
t 'W'e once obtained a slight indication, but were never able to repeat it.
t Nothing very definite appears to have been published on the behaviourof mixtures
of methane and chlorine in sunlight. Gay Lnssac and Thenard record that a mixture
of 2 vols, of methane and 4 vols, of chlorine deposit carbon and form 8 vols, of
hydrogen chleride in daylight. Dumas says a mixture of 2 vols, of methane and
6 vols, of chlorine will explode in difluie daylight.
Digitized by
Google
416 MELLOR AND ANDERSON: THE UNION OP
5. Moreover, Professor Dixon (private communication) has failed to
obtain the Draper effect with a dry mixture of hydrogen and chlorine
gases, using concentrated sulphuric acid for the index fluid.
When chlorine is mixed with an equal volume of hydrogen confined
over water saturated with the two gases, the Draper effect is readily
obtained, but not if the proportions of the compounds differ by more
than 3 per cent, from equality. The best results are given by the
" sensitive mixture " of Bunsen and Roscoe.
It has also been observed that this mixture detonates most readily
when exposed to a magnesium light.
The Action of Different Sources of Light.
We have enclosed a magnesium lamp in a box fitted with a Thorn-
ton-Fickard time shutter, and tried the effect of a momentary exposure
of the mixture of hydrogen and chlorine gases to the magnesium
light. The shutter must be set to give an exposure from one-sixteenth
to one second duration. In this way, we have obtained better results
than with the Ruhmkorff coil (1^ in. spark) and Leyden jar (one
quart capacity). With commercial ribbon burning at the rate of 2
cm. per second, we have obtained displacements of the index up to
12 cm., whilst with a spark we have never had a greater displacement
than 2 cm. One of us manipulated the time shutter while the other
observed the index motion through a cathetometer.
The Draper effect has not been obtained by ultra-red or by ultra-
violet rays.
The experiments with the magnesium light show that the effect is
not due to an .electrical disturbance induced in the gas by the electrical
discharge. A coating of lampblack on the glass is sufficient to
prevent any sign of the Draper effect with either maguesium light or
an electric spark.
The more readily the gases detonate under the influence of light,
the greater the Draper effect. If the hydrogen and chlorine gases
are [not in suitable proportions to produce an explosion under the
influence of magnesium light, the Draper effect may sometimes be
obtained with magnesium light and not with the electric spark.
If certain non-explosive mixtures of hydrogen and chlorine are
exposed to magnesium light for two or three seconds, the index at a
(Fig. 1, p. 415) will immediately expand and then contract faster than
the eye can follow, right into the insolation vessel, owing to the formation
of hydrogen chloride at nearly an explosive rate.
When a large expansion is taking place, the index moves in a series
of rapid jerks, each about a centimetre in length. This appears to be due
to friction between the index fluid and the walls of the capQlary tube.
Digitized by
Google
HYDROGEN AND CHLORINE. PART IV. 417
Does Chemical Action talce Place during the Draper Effect !
It has hitherto been assumed that no hydrogen chloride is formed
daring the Draper effect because the index, after expansion, returns to
its original position. If hydrogen chloride were formed, it is believed
that the index would indicate a contraction in the volume of the mixed
gases owing to the removal of hydrogen chloride by the water con-
tained in the insolation vessel. We have tried to prove this in the
following manner.
Two Bunsen and Roscoe's actinometers were filled with the same
mixture of hydrogen and chlprine by leading the electrolytic gases into
Fia. 2.
the apparatus shown in Fig. 2, through the cock a and escape at d. The
whole was enclosed in a box so that the index of each actinometer was
screened from the rest of the apparatus by means of a dividing
partition. All the glass parts on the left side of the partition were
painted dead black with the exception of one-half of the bulb of one
actinometer as shown in the figure. After the whole had been filled
with a sensitive mixture of hydrogen and chlorine gases, the four
cocks were closed. Temperature, pressure, and index readings were
taken. The lid screwed on to the box and the whole was perfectly
VOL. LTXXI. F F
Digitized by VjOOQIC
418 THE UNION OF HYDROGEN AND CHLORINE. PART IV.
screened from extraneous light. The electrical connections were made
BO that sparks passed midway between the insolation vessels of the
two aotinometers. A clockwork arrangement was fixed so that a
spark passed between the terminals every half -hour.
One hundred and eight sparks gave a contraction of 3 cm., equiv-
alent to about 2 per cent, contraction. In another experiment, 120
sparks at intervals of one hour caused a similar contraction. The
darkened actinometer gave no indication of change.
It is necessary to allow some minutes to elapse between each spark,
because Pringsheim has shown that if the sparks succeed each other
after short intervals of time, the mixture will be carried through the
period of induction when combination ensues.
The result of a number of similar experiments is to show that 100
sparks will cause a contraction up to about 2 per cent., indicating that
hydrogen chloride is formed during the Draper effect, but in quantities
too small to be detected by other than cumulative methods even with
extra large insolation vessels.
The Draper effect may be likened to a very small explosion with in-
sufficient energy to propagate itself throughout the gas. We have
obtained effects of all magnitudes up to actual explosion, by varying
the intensity of the light and the time of exposure.
Condusiofis,
1. Hydrogen chloride is produced during the phenomenon called the
Draper effect.
2. The Draper effect is only produced by the luminous rays of light.
3. The Draper effect occurs with mixtures of approximately equal
volumes of hydrogen and chlorine, but not with chlorine alone, or
mixed with steam, air, nitrogen, carbon dioxide, carbon monoxide, or
methane.
4. The amount of combination depends not only upon the number,
but also upon the intensity of the sparks.
5. When the effect reaches a certain magnitude, depending on the
" sensibility " of the gas, explosion occurs. An explosion appears to
be a large Draper effect.
6. The motion of the index fluid which occurs when the insolation
vessel of Bunsen and Boscoe's actinometer is exposed to a flash of
light appears to be brought about by some disturbance in the gas
attending chemical combination.*
The Owbns Gollsok,
Manchbstbr.
* Malardand Le Chatelier have observed a "period da moTement vibratoiitt'
(not quite similar to the one under discussion) antecedent to the explosion of certain
gaseous mixtures. See Dixon's Bakerian Lecture, Phil. Trans,, 1898, A, 184» 97.
Digitized by
Google
CONDENSATION OF PHENOLS WITH ESTERS. 419
XLI. — Condensation of Phenols vnth Esters of
Unsaturated Acids. Part VII.
By Siegfried Buhemann.
Bbnzo-1 : 4-PTBONB (chromone) and its homologues, as shown in this
paper, have basic properties ; they dissolve in hydrochloric acid, and
these solutions give precipitates with cobalticyanic acid (Baeyer and
Yilliger, Ber., 1901, 34, 2679), as well as with platinic chloride. The
hydrochlorides and the platinichlorides are, however, much less stable
than the corresponding salts of dimethyl-y-pyrone, which can be
recrystallised from water (Collie and Tickle, Trans., 1899, 75,
710).
Previously, only the monohydric phenols have been used for the
formation of chromone and its homologues ; since then, however, ex-
periments have been made with the view of obtaining hydroxybenzo-
pyrones from polyhydric phenols, but as these have been unsuccessful,
I have attempted to prepare those compounds from the mono-ethers
of the polyhydric phenols.. One of these ethers, namely, guaiacol, had
some time ago (Ruhemann and Stapleton, Trana, 1900, 77> 1180)
been subjected to the action of ethyl phenylpropiolate, and the product
which was thus formed had been found to suffer a decomposition
analogous to that of the other aryl ethers of j8-hydroxycinnamic acid.
Guaiacol, as stated in this paper, reacts with ethyl chlorofumarate and
forms ethyl guaiacoloxyfumarate> which, on treatment with potash,
yields the corresponding acid. Its transformation into methoxy^
benzo-1 : 4-pyrone and the subsequent hydrolysis of the methyl ether
have not yet been carried out, as the study of the action of ethyl
chlorofumarate on the naphthols has lately occupied my whole
attention.
Whilst the ethyl ester behaves towards /9-naphthol in the same
manner as towards other phenols, and yields jS-naphthoxyfumaric
ester, it reacts with a>naphthol partly to form ethyl a-naphthoxy-
fumarate in the normal way, but chiefly to form two substances, one
of which is an ester of the formula C^qHjjO^, the other a compound of
the composition O^^'S^fi^,
The facts, on the one hand, that this reaction is accompanied by
the loss of alcohol, and, on the other, that the compounds have pro-
perties unlike those of the chromones, lead to the following view as
to the formation and the constitution of the ester O^JEL^fi^ :
Digitized by VjOOQIC
420 RUHEICANN: CONDENSATION OF PHENOLS WITH
+ COgEf COJIOH'OOjEt
6Na
,Q ^ + NaCl + CjHeO,
CrOH-COjEt
This compound may be derived from the tricyclic type, J^
s
which I call naphtharone, and, accordingly, may be termed Myl
naphtharonylaeetate.
The second substance, C24HJ2O4, is perhaps formed from the first,
according to the equation
SCi^HigO^ - Oj^HijO^ + COjEt-CHICH-COgEt.
The isolation of ethyl fumarate, figuring above, has not been effected
as yet, but experiments to accomplish this task are in progress. The
compound O^H^^^i' ^> most probably, to be represented by the
symbol
c:
and may be called bimaphtharonyL
EZPEBIMENTAL.
Formation qf Salts from Benzo-l : i-pt/roru and ita Homohgues.
The members of the chromone group dissolve in concentrated hydro-
chloric acid either in the cold or on slightly warming, and form salts
which, however, are unstable, since the benzo-1 : 4-pyrone8 are precipi-
tated on adding water, and are extracted by ether from their solutions
in the acid. Cobalticyanic acid gives white, crystalline precipitates
with solutions of the chromones in hydrochloric acid, and platinic
chloride, dissolved in concentrated hydrochloric acid, yields platini*
chlorides which separate in yellowish needles. These, however, are
readily decomposed by water with the re-formation of benzopyrones.
Digitized by
Google
ICSTEBS OF UNSATURATED ACIDS. PAHT VII. 421
The platinichlorides of chromone and some of its homologues have been
analysed ; for this purpose, they were washed with concentrated hydro-
chloric acid and dried in a yacuum over sulphuric acid and soda-lime.
Owing to the unstable nature of those salts, the analytical numbers
differ somewhat from those required by theory, as is shown in the fol-
lowing table :
Platinichlorides of —
Benzo-1 : 4'pyrone :
0-2606 left on ignition 0-0710 Pt. Pt - 27-21.
(C9flgO,)8,HjPtClg requires Pt = 27-71 per cent.
o-Toluo-l : 4-pyrone :
0-2690 gave 00715 Pt. Pt « 26-68.
(OioH80,)„HjPtClg requires Pt-26-66 per cent.
j>-Toluo-l :4-pyrone:
0-2836 gave 00726 Pt. Pt - 26-68.
(CioH80j)j,H,Pt01g requires Pt» 26-66 per cent.
6 : 8-Dimethylbenzo-l : 4-pyrone :
0-2937 gave 0 0732 Pt. Pt - 24-92.
(CiiHioOj)2,HjPtClg requires Pt« 26-67 per cent.
Action of the Sodium Derivaitive qf Ottaiaeol on Ethyl CMorqfumarate.
Ethyl OiMkiacoloxyfumarate, (j^'^^' ' §^o ^cj ^''
This compound is formed by adding ethyl chlorofumarate (1 mol.)
to the hot solution of sodium (1 at) in an excess of guaiacol. The
dark, viscous product, when cold, is agitated with dilute sulphuric acid
and ether ; the ethereal layer is then freed from the excess of guaiacol
by potash, the ether evaporated, and the remaining oil distilled in a
vacuum. It has a yellow colour and boils at 212 — 213^ under 15 mm.
pressure. On analysis :
0-1926 gave 04333 CO, and 01076 H,0. 0 =- 61-38 ; H - 6-20.
OijHigOj requires C- 61-22 ; H:-6-12 per cent.
ChaiacoloxufwMync Aeid, (ChJ-0)-C^H,-0-C(CO,H):OH-002H.—
The potassium salt of the acid separates on boiling the ethyl ester
with alcoholic potash for 2 hours. After evaporation of the alcohol,
the residue is dissolved in water and the solution treated with an
excess of dilute sulphuric acid, when an oil is precipitated consisting
of the organic acid and guaiacol. The latter compound, which is
Digitized by
Google
422 RUHBlf ANN : CONDENSATION OF PHENOLS WITH
formed along with guaiacolozyf umaric acid on hydrolysis of the ester,
is removed by shaking the ethereal solution of the oil with sodium
carbonate, adding dilute sulphuric acid to the aqueous layer^ and ex-
tracting the organic acid about sixteen times with fresh quantities of
ether. On distilling off the ether, a solid is left behind ; this readily
dissolves in boiling water and on cooling crystallises in yellowish
needles, which melt at 138^ with evolution of gas. On analysis :
0-2028 gave 0-4115 CO, and 0-0803 H,0. 0 = 56-34 ; H = 4-39.
CjiHioOg requires C = 5546 ; H = 4*20 per cent.
Guaiacolozyf umaric acid readily dissolves in concentrated 'sulphuric
acid, forming a yellowish solution which, after standing for 24 hours, is
poured into cold water. As no solid separates, the solution is extracted
with ether, and on evaporation of the latter, a yellowish solid remains
behind ; this crystallises from hot water in slightly coloured needles
which melt and decompose at 251^. This substance, which has not
yet been further examined, is probably o-methoxybenzo-1 :4-pyrone-
carboxylic acid.
/\/\ CH-CO--C,H.
Ethyl p'liaphthox^umaraie, f T 1 q— 8-C0,-0 H^ '
For the preparation of this compound, a method has been used
similar to that which served for the formation of ethyl a-naphthoxy-
cinnamate (Ruhemann and Beddow, Trans., 1900,77, 989). /3-Naphthol
(1 mol.) is added to a solution of sodium (1 at.) in absolute alcohol, the
alcohol is removed by heating the solution in a vacuum, first at 100^
and finally at 180 — 190°; ethyl chlorofumarate (1 mol.), dissolved in
toluene, is then added to the dry naphtholate, when a dark brown solu-
tion is produced with development of heat. This is boiled in a flask,
attached to a reflux condenser for 1 hour in order to complete the
reaction, and, when cold, is agitated with dilute sulphuric acid and
ether. The ethereal layer is freed from unchanged jS^aphthol by
shaking with potash, and, on evaporation of the ether and toluene,
yields a dark oil which is fractionated under diminished pressure.
Almost the whole quantity distils at 240 — 242^ under 12 mm. pressure
as a very viscous, fluorescent yellow oil. On analysb :
0-2145 gave 0-5398 CO, and 0-1 130 H,0. C « 68-63 ; H = 5-85.
CigHigOg requires 0 = 68-79 ; H-=5'73 per cent.
pnNaphihaxyfwmaric acid, CioHy*0-C(COjH):CH-00^, is obtained
on hydrolysis of the ethyl ester by means of alcoholic potash. On
mixing the reagents, a red coloration is produced, and, after a slunt
time, the potassium salt of the acid is precipitated as a yellowish solid.
Digitized by
Google
ISTKRS OF UNSATTTRATKD ACIDS. PABT VIL 423
After boiling on the water-bath for 1 hoor^ the alcohol is distilled off,
the residue dissolved in water, and the organic acid liberated by hydro-
chloric acid. For the removal of /3-naphthol, which is formed along
with )3-naphthozyfumaric acid, the precipitate is dissolved in sodium
carbonate, the solution filtered and mixed with an excess of hydro-
chloric acid. )3-Naphthoxyfumaric acid is insoluble in water, it dis-
solves with difficulty in ether, readily, however, in hot alcohol ; but on
boiling the solution for some time, decomposition takes place with
liberation of /9-naphthol. The acid crystallises in small, yellowish
plates, which darken at about 230° and melt at 236° with evolution of
gas. On analysis :
0-1773 gave 0-4223 COj and 0-0613 H,0. C = 64-96 ; H - 3-84.
Oi^HjoOj requires C-66-11 ; H = 3-87 per cent.
I am at present engaged in the study of the action of concentrated
sulphuric acid on )3-naphthoxyf umaric acid with the view of condensing
it to ^naphtha- 1 : 4-pyronecarboxylic acid, and hope to publish the
result shortly.
Aetum of Ethyl Chloro/wna/rate on a-N'apktKol.
This reaction is of especial interest since, as mentioned in the intro-
duction, a-naphthol differs from the other phenols in its behaviour
towards the ethyl ester, inasmuch as it yields two substances which
belong to a new class of cyclic compounds. The reaction is carried out
in a manner similar to that employed in the preparation of ethyl
^-naphthoxyfumarate ; namely, by adding a-naphthol (1 mol.) to an
alcoholic solution of sodium (1 at.), removing the alcohol by distillation
in a vacuum, finally at 190°, and mixing the dry naphtholate with
ethyl chlorofumarate (1 moL) dissolved in toluene. The action takes
place with development of heat and is completed by boiling in a flask
attached to a reflux condenser for 1 — 2 hours. The contents of the
vessel consist of a yellowish, crystalline substance and a brown liquid ;
theseare agitated with dilute sulphuric acid, and the crystals arecollected
and washed with ether. The solid represents the bisnaphtharonyl,
C^Hi^O^, referred to in the introduction, whilst the dark filtrate con-
tains ethyl naphtharonylacetate, O^^K^fi^, and ethyl a-naphthoxy-
fumarate,
Bienojohikaronyl i§ almost insoluble in all the ordinary solvents, hot
alcohol only dissolving traces and yielding a yellowish, fluorescent
solution. It dissolves readily, however, in boiling nitrobenzene, form-
ing A dark red solution with a deep green fluorescence, from which, on
cooling, it separates in orange needles. These are freed from the
adhering solvent by washing with hot alcohol ; they do not melt at
335°. On analysis :
Digitized by
Google
424 ruhemakn: condensatiok of phenols with
0-1787 gave 0-5171 00^ and 00548 Ufi. C » 78-92 ; H - 3-40.
0-1760 „ 0-5100 COj „ 0-0525 H^O. 0« 79-02 j H = 3-31.
C^H„04 requires 0 « 79-11 ; H = 8-30 per cent
Bisnaphtharonyl is insoluble in cold alkali, but dissolves when boiled
with concentrated aqueous potash for 2 hours in a flask provided with
a reflux condenser, forming a red solution. On adding hydrochloric
acid, a yellowish solid is precipitated, which is insoluble in alcohol or
acetic acid, but dissolves in nitrobenzene. From this solution, orange
needles separate, which are no longer dissolved by cold potash. The
properties of this compound point to its identity with bisnaphtharonyl,
and this conclusion is supported by the following analysis :
0-1863 gave 05391 OOj and 00690 Hfi. 0 - 78-91 ; H = 3-51.
C24H1JO4 requires C = 7911 ; H-3-30 per cent
This experiment proves that one or both rings, containing oxygen,
in bisnaphtharonyl, open up by the action of potassium hydroxide, but
under the influence of an acid close again to form the original com-
pound. This may be symbolised as follows :
<-^
KJEL V y KJU
In the hope of effecting a deep^eated decomposition of bisnaphthar-
onyl which would supply further evidence for its constitution, I have
subjected the compound to the action of fused potash, but found that
complete carbonisation takes place.
On adding concentrated sulphuric acid to the substance, it becomes
purple and then dissolves very slowly at the ordinary temperature,
but rapidly on slightly warming, to yield a red solution. This, when
poured into water, gives a yellowish solid which is soluble in water
with great ease. Most probably there is thus formed a sulphonio acid
of bisnaphtharonyl j this view seems to be supported by the behaviour
of nitric acid towards bisnaphtharonyl.
TetranUrobUnaphtharonyl, 0^'B^^O^O^\0^. — Bisnaphtharonyl dis-
solves in fuming nitric acid on warming slightly, and the red solution
gives, with water, a yellowish solid. This is insoluble in the ordinary
solvents, but dissolves in hot nitrobenzene, and on cooling crystallises
in minute, yellowish-brown prisms, which are freed from the adhering
Digitized by
Google
ESTERS OF UNSATURATED ACIDS. PART VII. 426
Bolvent by washing with hot alcohol. The compound does not n^elt at
325''. On analysis :
0*2690 gave 22-8 c.c. moist nitrogen at 19"^ and 778 mm. N « 10*36.
O^fifii^l^^ requires N* 10*29 per cent.
ISthyl IfaphthanmylacOaiB, \— <^ ^ICH-COj-C^H^
The dark mother liquor from bisnaphtharonyl, after being diluted
with ether, is agitated with an excess of potash in order to free it from
unchanged a-naphthol. The ethereal layer, which has a deep green
fluoresoenoe, is then dried with calcium chloride, the ether evaporated,
and the last traces of toluene removed by heating in a vacuum on the
water-bath. The dark oil which remains behind, partly solidifies on
standing. This product is dissolved in hot alcohol, and the solution,
on cooling, deposits long, yellowish needles which after recrystallisa-
tion froia the same solvent, melt at 146 — 147°. The substance is ethyl
naphtharonylacetate. On analysis :
0*1992 gave 0*5226 COj and 0*0806 H,0. 0 = 71*54 ; H = 4*49.
CijHijO^ requires 0*71*64 ; H:-4*48 per cent.
Ethyl naphtharonylacetate may be distilled in a vacuum, when it
passes over as a yellowish oil which quickly solidifies ; it is sparingly
soluble in cold alcohol or acetic acid, but readily so on boiling. The
hydrolysis of the ethyl ester l^ still under examination, but the
results^ of a few experiments undertaken in that direction may be
recorded here. On using alcoholic potash, a deep purple coloration of
the solution takes place, and the alkaline liquor, after removal of the
alcohol by distillation and solution of the residue in water, yields a
brown, gelatinous precipitate on treatment with hydrochloric acid.
This is insoluble in water, but dissolves in alcohol with the greatest
ease and separates from the solution in an amorphous state. If
aqueous, instead of alcoholic, potash is used for the hydrolysis, the
ethyl ester dissolves on boiling for 2 hours. On adding hydrochloric
acid to the reddish solution, a yellow precipitate is formed ; this is a
mixture of two or more acids which have not yet been separated.
^3-<^'
Ifajpkth(mmylacekmide,y-\ ^OrCH-CO-NHj. — The ethyl
eeter reacts with alcoholic ammonia on remaining in contact with
it at the ordinary temperature for some time. The needles of
Digitized by
Google
4^6 FENTdN AND RTFl'flL : MESOXALIC SEMI-AIDSHTDI^.
the ester gradually disappear, being transformed into a yeUowifihf
crystalline product, whilst the solution turns pink. After 2 days, the
precipitate is collected, washed with water, and then with alcohoL It
is insoluble in either of these solvents, but dissolves in glacial acetic
acid and crystallises from this solution in groups of yellowish needles
which blacken at 258° and melt to a dark liquid at 265°. On
analysis:
0*2395 gave 12 c.c. moist nitrogen at U° and 749 mm. N =- 5*88.
Cj^HjjOjN requires N«5'85 per cent,
EUtyl a-Nafhthoocyfumaraie^ y — <^ CH-COj'CjHg'
The dark alcoholic mother liquor from ethyl naphtharonylaoetate^
on concentration, yields a further crop of crystals of the same ester ;
finally, an oil remains which is fractionated under diminished pressure.
The greater part distils at 246 — 248° under 16 mm. pressure as a
viscous, yellow oil. This is ethyl naphthozyf umarate. On analysis :
01995 gave 0-5025 OOg and 0-1090 HjO. C - 68-68 ; H - 607.
CigHigOg requires C = 6879 ; H - 5- 73 per cent.
In conclusion, it may be mentioned that the action of ethyl chloro-
f umarate on a-naphthol takes place, also, at the ordinary temperature
on adding to the dry sodium derivative of a-naphthol a solution of the
ethyl ester in absolute ether ; heat is developed, the mixture turns
red, and deposits a solid. After standing for several days in a flask
provided with a dr3ring tube, the product is treated as in the former
case in order to separate the above-mentioned compounds.
GONVILLB A.ND CaIUS COLLEGE,
Cambridge.
XLII. — Mesoocalic Semi-Aldehyde.
By HsNBY JoHK HoBSTMAN Fenton, F.RS., and John Hbnkt Btffbl,
B.A., B.Sc.
Chlorinb and bromine have, as is well known, very little action on an
aqueous solution of tartaric acid at the the ordinary temperature ; it is
found, however, that in presence of ferrous iron the action is consider-
ably accelerated. If the solution be saturated with chlorine, the yellow
colour which is first produced soon disappears on standing, and after
Digitized by
Google
FEKtON AND KTFFEL : MESOXALIC SElfl-ALDEHTDE. 427
some hours the odour of chlorine is no longer perceptible. On
addition of phenylhydrazine acetate or hydrochloride to the solution so
obtained, a bright orange-yellow precipitate is produced.
In order to study the nature of this reaction, the following method
was adopted. Ordinary (i-tartaric acid was dissolved in five to ten
times its weight of water and miited with a small quantity of freshly
prepared ferrous tartrate (obtained by dissolving < ferrum redactum *
in a solution of the acid). A slow current of chlorine was then passed
into the mixture until saturation appeared complete. After standing
overnight, or until the odour of chlorine had disappeared, it was again
saturated with the gas, and this treatment repeated until a sufficient
quantity of the product had been formed, the progress of the change
being ascertained by the phenylhydrazine reaction. The liquid was
then concentrated to a small bulk by distillation under very reduced
pressure at about 50^, and was then allowed to stand, preferably in a
vacuum desiccator, until most of the unaltered tartaric acid had
crystallised out. The mother liquor from these crystals remains as a
thick syrup which refuses to crystallise. The product is very stable
at the ordinary temperature and keeps remarkably well without
apparent change, but boiling changes its character, the product then
giving a small quantity of a highly crystalline, brownish precipitate
when tested with phenylhydrazine.
Bromine, or hypochlorite, produces an effect similar to that of chlorine,
and a very simple method of demonstrating the change is to add a small
quantity of sodium hypochlorite to potassium hydrogen tartrate sus-
pended in water and mixed with a little ferrous salt. After allowing
the mixture to stand a short time and removing any excess of chlorine,
if necessary, by a current of air or by sulphurous acid, the liquid
gives an abundant orange-yellow precipitate with phenylhydrazine salts.
The proportion of iron used is (as in Various similiar oxidation pro-
cesses previously described by one of the authors) not a matter of
importance ; the merest traces have a marked effect. In these experi-
ments, about 1/5000 to 1/1000 part of iron to 1 part of acid was employed.
In the present case, a certain amount of action can be detected in
absence of iron, but the process is then very slow and the yield poor.
The syrupy product obtained in the manner described above still
contains tartaric acid, and in order to investigate its nature the action
of various reagents was studied.
Action of Phent/lhydrazine.
A dilute solution of the produce gives the above described orange-
yellow precipitate almost immediately in the cold with either the
acetate or hydrochloride of phenylhydrazine. The precipitation is
Digitized by
Google
428 FBNTON AND RYFPEL : MESOXALIC SEMI-ALDBHTDK.
accelerated by heating and is complete in half-an-honr or less. After
leashing with water and drying in the air, the precipitate dissolves
easily in alcohol and sparingly in hot benzene or chloroform. From
the latter solvent, it separates on cooling in masses of brilliant orange-
golden needles or prisms. These, when qidckly heated, melt nsoaUy
at 218° after the first crystallisation and on further recrystallisation
once or twice from the same solvent, melt constantly at 222 — 224°.
Analysis of the product dried at 100° gave the following result :
0-1181 gave 30-1 C.C. nitrogen at 18° and 749 mm. N- 19-74 per cent.
It dissolves in boiling sodium carbonate solution, and, on cooling, a
beautifully crystalline sodium salt is deposited, the aqueous solution
of which dyes silk, wool, &o., a bright lemon-yellow colour. Heated with
acetic anhydride, it yields a substance which crystallises from alcohol
in brilliant orange-red needles which melt at 160°.
The composition and properties of the product coincide in every way
exactly with those of one which has previously been described in
various former communications (Fen ton and Jones, Trans.| 1900, T7,
77, and 1901, 79, 91).
It was first obtained by the further oxidation of malic acid in pre-
sence of iron, secondly by oxidation of oxalacetic acid under similar
conditions, and thirdly by heating the phenylhydrazine salt of di-
hydroxymaleic acid with excess of p^nylhydrazine for some time on
a water-bath. (The same result is also produced by heating the salt
with water alone.) Analysis of the compound prepared in these ways
gave C»63'2, H«6*l, Ne20*l per cent, as a mean of several con-
cordant experiments. The nature of the compound was the subject of
much investigation, since it so closely resembled the osazone of hydroxy-
pyruvic acid, first obtained by Nastvogel from dibromopyruvic acid
(AnncUmf 1888, 248, 85), and subsequently by Will from collodion-
wool {Ber,, 1891, 24, 400 and 3831). The same osazone was afterwards
obtained by the action of phenylhydrazine on the product of oxidation
of glyceric acid in presence of iron (Fen ton and Jones, Trans., 1900,
77, 72). There remained, however, the very considerable discrepancy
in the melting point. The osazone of hydroxypyruvic acid melted at
201—203° (Nastvogel), 206° (Will), and 207° (Fenton and Jones);
whereas the product at present under discussion melts at 222 — 224°.*
In consequence of this difference, and in view of the fact that the
present product of higher melting point was always obtiained from
« Friedel and Combes (BuU, Soe. Chifn., 1890, [iii], 8, 770) state that by adding
phenylhydrazine to the product of electrolysis of tartaric acid, they obtained the
osazone of glyoxal, melting at 160°, and the osazone of glyoxalcarboxylic acid,
melting at 218^ They give, however, no details, analyses, or farther information
whatever.
Digitized by
Google
FENTON AKD BTFFEL : MBSOXALIC SBMI-ALDEHTDE. 429^
acids containing 4 carbon atoms in the molecule, other possible ex-
planations as to its nature were suggested (Fenton and Jones, Trans.,
1901, 79, 98). It might, for example, be the hydrazide-dihjdrazone
of dioxosaocinic acid, and this idea was apparently supported by the
fact that when heated with acetic anhydride it gives notable quantities
of carbon dioxide and aniline in addition to the phenylhydrazine-
ketophenylpyrazolone of Knorr. This result might, however, be other-
wise explained, and the question still remained open whether the
product of higher melting point is a derivative of a 3 or 4 carbon
acid. The calculated composition shows very little difference, Nast-
vogel's osazone, CigHj^OgN^, requiring C 63 82, H 4-96, N 19-86
per cent., and the 4 carbon derivative above suggested, C^jH^^OgN^,
requiring C 63*46, H 4*80, N 2019 per cent.
By a numerical coincidence, the same close similarity of composition
exists between the corresponding derivatives of various other hydr-
azines (nitrophenyl-, bromophenyl-, tolyl-hydrazine, ^.), so that it is
evident that the question cannot be settled by analysis of any of these
derivatives. It appeared probable, hpwever, that a careful comparison
of the properties of derivatives of these substituted hydrazines obtained
from dibromopyruvic acid with those from the tartaric acid oxidation
product now under discussion should give more definite information
as to the question, and the following experiments were therefore made
with this object.
(1) The product obtained from tartaric acid by action of chlorine,
described above, was mixed with excess of ^-bramophenylhydrazine
dissolved in acetic acid. The resulting orange-coloured precipitate
was washed, dried in the air, and recrystallised twice from a mixture
of absolute alcohol and benzene. The long, bright orange-coloured
needles so obtained melted at 245—246''.
Dibromopyruvic acid (1 mol.) was then dissolved in water and mixed
with a solution of p-bromophenylhydrazine hydrochloride (2 mols.) and
the mixture allowed to stand 3 to 4 hours at the ordinary temperature.
An orange precipitate exactly similar to that last mentioned was ob-
tained, and this, when recrystallised in a similar way, melted precisely
at the same temperature, 245 — 246^.
(2) The tartaric acid product was mixed with an excess of ^p^yl-
hydraxine hydrochloride in aqueous solution. An orange-red precipitate
soon appeared, and after 3 or 4 hours was filtered off, washed, dried,
and recrystallised from hot benzene. The orange-coloured needles so
obtained, when slowly heated, began to soften at 188^ and melted com-
pletely at 194 — 195^ Nastvogel (loc. eit,), by the action of p-tolyl-
hydrazine hydrochloride on dibromopyruvic acid, obtained golden
needles which melted at 186—188''.
Digitized by
Google
430 FENTON AND RTFFEL : MESOXALIG SEMI-ALDEHYDE.
(3) These results pointed so strongly to the identity of the prodnots
from the two different sources that it appeared desirable to prepare
the phenylhydrazine derivative from dibromopyruvic aoid exactly
according to Nastvogel's directions, and to compare it with the osazone
from the tartaric acid product.
The melting point of this osazone from tartaric acid, ozalacetic
acid, malic acid, and dihydroxymaleic acid has been determined a very
large number of times, and the results from all these different sources
are remarkably concordant ; it may, in fact, be taken as established
that the melting point is 222—224''.
The specimens from all these sources have, for analysis, been re-
crystallised from Morqform and the melting points determined by the
quick-heating method. But even slow heating does not lower the
melting point more than 3^ or 4^.
With regard to Nastvogel's osazone, this author does not mention
how the melting point was determined, but Will determined it by the
quick-heating method. But in all the methods by which this osazone
has hitherto been obtained (that is, from dibromopyruvic acid,
collodion-wool, and glyceric acid) the product was purified by crys-
tallisation from benzene,
Nastvogel's experiment was therefore carefully repeated: — 3'5
grams of phenylhydrazinc hydrochloride were dissolved in water and
a solution of 3 grams of dibromopyruvic acid was added, the mixture
being kept cold. A bright orange precipitate began to separate
almost at once, and after 3^ hours was collected, washed, dried in the
air^ and recrystallised three times from hot ehtorqfarm. The resulting
product resembled in every respect the osazone from tartaric acid,
kc,, and melted at 222—224^.
It is therefore evident that this is the true melting point of the
osazone, and it is remarkable that so many observers have obtained
the lower value, the explanation being, apparently, that chloroform
is the more appropriate solvent for its purification.
The above facts practically remove all doubt as to the identity of
Nastvogel's osazone with that at present under discussion. Further
evidence of this was obtained in the following way. The highly con-
centrated syrup from the oxidation product of tartaric acid was dis-
solved in absolute alcohol, saturated with dry hydrogen chloride,
allowed to stand overnight, and distilled to small bulk under
diminished pressure and the product again treated in a similar
manner. It was then poured into cold water, extracted with ether,
and dried over calcium chloride. After distilling off the ether, the
liquid contains some ethyl tartrate and the products are difficult to
separate, but on adding phenylhydrazine acetate and diluting with
Digitized by
Google
FENTON AND RTFFEL : MESOXALIC SEMI-A.LDEUTDE. 431
water a lemon-yellow precipitate slowly separaten. This, when re-
orystallisedj first from alcohol and then from benzene and a little
light petroleum, was obtained in yellow, transparent plates which
melted at 229 — 231". This substance dissolves easily in hot alcohol
or benzene, but is nearly insoluble in alkalis.
0-1889 gave 29'8 c.c. nitrogen at 19° and 751 mm. N=- 18-28.
OiyHigOgN^ requires N — 18*06 per cent.
This product is evidently identical with that which Will obtained
by the action of ethyl iodide on the osazone from collodion-wool, and
is, in fact, the ethyl ester of this osazone :.
CH(NjHPh)-C(N3HPh)-C0,Et.
The osazone, 'OH(NjHPh)-C(N8HPh)*002H, may obviously be
derived from (1) hydrozypyruvic acid, (2) the semi-aldehyde of tartronic
acid, or (3) the semi-aldehyde of mesoxalic acid. The product which
Will obtained from collodion-wool is considered by him to be hydrozy-
pyruvic acid from analysis of its metallic salts and from the fact that
it is not oxidised by bromine.
The substance obtained by Fenton and Jones by oxidation of glyceric
acid in presence of iron is in many respects similar, but it gives an
intense violet colour with ferric salts in presence of alkalis, a property
which presumably is not possessed by Will's product, since no mention
is made of it. The glyceric acid product, if it is not hydroxypyruvio
acid, may possibly be the tautomeric dihydroxyacrylic acid,
CH(OH):0(OH)-0OjH.
The oxidation product from tartaric acid at present under discussion
might, so far, be any of the above-named acids. It gives, however,
when its formation is complete (see below), no colour with ferric salts
in presence of alkali. It might conceivably be dichloropyruvic acid,
but this is excluded by the fact that if the solution be precipitated
with barium acetate, the well-washed barium salt obtained contains
no chlorine, and on decomposition with dilute sulphuric acid gives,
with phenylhydraisine, the same osazone as before.
Action of ffydraacylctmine.
If the substance under discussion is the semi-aldehyde of mesoxalic
acid, it would be expected that the action of excess of hydroxylamine
should give the dioxime, CH(NOH)'C(NOH)-OOjH, or dioximido-
propionio acid. This oxime was obtained by Soderbaum {Ber.^ 1892,
25, 904) by the action of hydroxylamine on dibromopyruvic acid, and
was shown to exist in two forms. The ' primary ' acid,
B-C C.CO,H
N'OH OH-N
Digitized by
Google
432 FENTON AND RTFFSL : MESOXAUC SEMI-ALDEHYDE.
ly^Q C-CQ TT
meltB at 141—143% and the • secondary ' add, Ji U.qH ' **
about 172°
If the primary acid is dissolved in the least possible quantity of
ammonia, well cooled by ice, and acidified with hydrochloric acid, it is
transformed into the secondary acid. Both forms give a blood-red
colour with ferric chloride, and with ferrous sulphate and a little
caustic soda they give an intense, but unstable, violet colour. Ouprio
acetate gives an olive-green precipitate.
The present tartaric acid oxidation product was concentrated as be-
fore and freed as much as possible from unaltered tartaric add ; it
was then dissolved in water, mixed with excess of an aqueous solution
of hydroxylamine hydrochloride, and neutralised, or nearly so, by the
gradual addition of solid sodium carbonate, the mixture being cooled
by ice. The solution turns dark red or violet, and on standing de-
posits a white sodium salt. It was allowed to stand for a few hours
at 0° and then for about 24 hours at the ordinary temperature. The
mixture was filtered with the aid of suction, the solid sodium salt
treated with excess of dilute sulphuric add, and extracted several
times with ether. The ethereal solution was distilled to a small bulk
and allowed to stand in a vacuum desiccator, when it solidified to a
reddish mass. This was dissolved in the least possible quantity of
ammonia, well cooled by ice, and acidified with strong hydrochloric
acid.
After standing a short time, a mass of long, colourless needles
separated ; these were purified by redissolving in ammonia and acidify-
ing with dilute sulphuric acid under similar conditions.
The resulting product, dried in a vacuum desiccator, when slowly
heated melted at 178 — 180°. It is somewhat sparingly soluble in
cold water and the aqueous solution behaves with ferric chloride,
ferrous sulphate, and cupric acetate exactly as above described with
Soderbaum's acid.
The vacuum-dried product gave the following results on analysis :
I 01571 gave 01551 OOj and 00454 Ufi. C-26-92 ; H«3-21.
11.01064 „ 18-8 c.c. nitrogen at 18° and 767 mm. N- 20-98,
OjH^O^Nj requires 0-27-27; H-303; N = 21-21 percent.
OoDidation to Meaaxalio Acid,
The action of hydroxylamine and of phenylhydrazine practically
settles the question as to the nature of the product under discussion,
that it is the semi-aldehyde of mesoxalic acid. It was considered, how-
ever, that it would be satisfactory to prove the matter conclusively by
oxidation^of the aldehyde to mesoxalic acid. ,
Digitized by
Google
IfSl^nO}!^ AKP tiTF^BL: MitSOlALIC dEMt-ALDl&HtDie. 483
Bromine or chlorine did not appear to be suitable agents for this
purpose, owing to the presence of unaltered tartaric acid in the sub*
stance to be operated on, so that cupric oxide was selected as the most
appropriate. A solution of the tartaric add oxidation product was
made alkaline with soda and mixed with excess of freshly precipitated
Qupric hydroxide. On warming to about 60°, a rapid redaction took
place, and when this appeared to be complete the mixture was filtered,
the liquid acidified with hydrochloric acid, and the copper removed by
hydrogen sulphide, excess of the latter being avoided. Addition of
phenylhydrazine, as acetate or hydrochloride, to this solution gave a
yellow colour, and after standing a short time^ a pale yellow precipi-
tate consisting of masses of fine needles. These, when recrystallised
from hot alcohol, melted at 171 — 172° and corresponded* in properties
exactly with the hydrazone of mesoxalic acid (compare Elbers, Anncden,
1885, 227, 341 ; Clemm, Bwy 1898, 31, 1451 ; Fenton and Jones,
Trans., 1900, 77, 71).
For analysis, the substance was prepared from the pure product
described below and phenylhydrazine hydrochloride. Thus obtained,
it was quite pure without recrystallisation. After being well washed
with water and alcohol and dried in a vacuum desiccator, two distinct
specimens, prepared on different occasions; melted at 173 — 174° and
gave on analysis the following results :
L 0*2243 gave 25*5 c.c. nitrogen at 14° and 765 mm. N== 1365.
n. 0-1910 „ 22-2 „ „ 17° „ 759 mm. N- 13-70.
OgHgO^Nj requires N« 13*46 per cent.
Olemm gives the melting point of mesoxalic hydrazone as 174°.
Mode qf Formation*
Theoretically there are, of course, several ways in which the semi-
aldehyde of mesoxalic acid might result from tartaric acid by oxida-
tion, and experiments were made with the object of ascertaining
which of these is more probable. The simple removal of one mol. of
carbon dioxide from ' anhydrous ' dihydroxytartaric acid or dioxosuccinic
acid would at once afford a direct explanation,
COjH-CO-CX)-CG,H - COjH-OO-CHO + CO,,
but as yet it has not been found possible to prepare the substance
from dihydroxytartaric acid ; heating the acid alone, with acid and
with iron salts all gave negative or unsatisfactory results. It was
observed, on the other hand, that in preparing the product by the
action of chlorine on tartaric acid in presence of iron in the manner
above described, the liquid in the first instance always gives a notable
reaction for dthydroxymaHeic acid when tested with ferric chloride and
VOL. LXXXI. G G
Digitized by VjOOQIC
434 FENTON ANI} RTF^Et : MESO^ALIC SElil-ALDBdYDli.
alkalis (compare Fenton, Ghem.NwB, 1876,33, 190; 1881,43, 110), but
after the completioD of the process, by farther action of chlorine and
subsequent distillation, this property no longer appears. The initial
change iVhich takes place may therefore consist in the formation of
dihydrozymaleic acid, and this, by futher oxidation and loss of carbon
dioxide, may give rise to mesoxalic semi-aldehyde,
C^H^Oe + O = CsHjO^ + HjO + COg.
In order to test the latter hypothesis, pure dihydroxymaleic acid
was subjected to oxidation under various conditions. It was previously
shown that this acid is oxidised almost quantitatively by bromine at
the ordinary temperature to dihydroxytartaric acid (Trans., 1894, 67,
48), and in trying modifications of the process it was often observed
that the presence of iron — the agent which was essential for the
formation of dihydroxymaleic acid from tartaric acid-^had a deleterious
effect. In view of this fact, it seemed not improbable that, in the
process at present under discussion, the ferric salt produced may be
the active agent in the oxidation of dihydrozymaleic acid to mesoxalic
semi-aldehyde. This supposition is entirely borne out by experiment,
and the oxidation is found to take place almost quantitatively when
carried out in the following way.
Crystallised dihydroxymaleic acid, GJlfiQ^^Ufi, is covered with
water or dilute alcohol and a solution of ferrio chloride or sulphate
gradually added. Each drop of the ferric salt produces an intense
violet-black coloration which quickly disappears, with a considerable
rise of temperature. The temperature of the mixture is allowed to
rise to about 40% being aided by warming if necessary, and the
addition of the ferric salt is continued until a further quantity no
longer produces the coloration mentioned. This point is reached when
the substances are present very nearly in the ratio of G^H^O^ : 2Fe.
The iron is entirely reduced to the ferrous state, and the end«point
may easily be demonstrated by the usual indicators. A brisk evolution
of carbon dioxide occurs during the process, and the change may be
expressed by the equation :
C^H^O^ + Fe2(S04)3 - OjHaO^ + 2FeS0^ + H^SO^ + 00^
The reaction appears to take place only with ferric salts of strong add
radicles and in presence of water ; if ferric acetate be used in aqueous
solution, or if alcoholic ferrio chloride be added to alcoholic solution of
dihydroxymaleic acid, the dark violet colour produced remains quite
persistent, at any rate for several hours.
In order to obtain the product free from iron, it is best to employ
ferric sulphate as oxidising agent, and, after concentration in a
vacuum desiccator, to precipitate the ferrous salt by alcohol and ether,
Digitized by
Google
t*ENTON AllD RYI'FKL : MESOXlLIC SElfl-ALDEHTDE. 435
the free sulphuric acid being neutralised by the calculated amount of
Bodium carbonate or baryta water. By repeating the concentration and
treatment with alcohol and ether, and removing the latter in a vacuum
desiccator, a syrup practically free from iron is left, which gives all the
reactions described above for the tartaric acid oxidation product.
Phenylhydrasdne acetate or hydrochloride gives the above-described
osazone melting at 222 — 224°; hydrozylamine gives the dioxime
identical in every way with that obtained from the tartaric acid
product, and oxidation with alkaline cupric hydroxide gives a large
yield of mesoxalic acid.
In order to avoid the large quantity of iron salts present .in the
above method, a very small quantity of a ferric salt may be employed,
and, as soon as reduction is complete, the resulting ferrous salt may be
re-oxidised by hydrogen dioxide, the addition of the latter being
continued until the change is completed. This method appears to give
good results except that a small quantity of dihydroxytartaric acid
may be formed at the same time, so that the use of a ferric salt only as
oxidising agent is the most reliable.
The behaviour of ferric iron in the reaction here described is of
much interest in throwing light upon certain processes of oxidation in
presence of iron where a ferric salt has been employed (compare
Eenton, Trans., 1900, 77, 1296). In these cases, there is little doubt
that a ferrous salt is first produced, possibly only in minute quantity,
and that this determines the oxidation in the usual manner ; ferric
salt appears to encourage the breaking down of the resulting product
with evolution of carbon dioxide.
The isolation of mesoxalic semi-aldehyde as above described leaves
now only four out of the eleven possible oxidation products of glycerol
which have not been obtained ; mesoxalic dialdehyde being known only
in the form of oxime, and tartronic semi^aldehyde and dialdehyde,
and hydroxypyruvic aldehyde being unknown.
Many interesting results may be expected from a further study of
this aldehyde-acid ; its aldehyde hydrate may be regarded as tautomeric
with the hitherto missing trihydroxyacrylic acid, which is of much
interest owing to its relationship to uric acid.
A considerable part of the eixpenses incurred in carrying out this
investigation has been defrayed by funds kindly supplied by the
Qovdtnment Grant Committee of the BoysJ Society* -
Digitized
I by Google
436 CROCKER: THE PlCRlMINOTHIOCARfiONlC ESTERS.
XLIIL — The Picnminothiocarbonio Esters.
By Jambs Codsington Cbooksb, B. A., Scholar of St. John's College,
Cambridge.
A SHORT preliminary note on the picriminothiocarbonic esters has
akeady appeared (Comb. Univ. Rep,, 1902, 82, 23, 549), but in view of
the publication of a paper dealing with the action of acidic thiocyanates
on alcohol (Dixon, Trans., 1902, 81, 1^8), it seems advisable to
publish the work in full. The experiments deal entirely with the
action of picryl chloride on thiocyanates and alcohols. The reaction
is a remarkable one, and takes place, in most cases, with ease, the
products being beautifully crystallised bodies. The first ease investi-
gated was the reaction of picryl chloride on ammonium thiocyanate in
ethyl alcohol solution.
Ethyl Ficryl Fionminothiocarbanate.
Ten grams of picryl chloride dissolved in hot absolute alcohol were
mixed with a hot alcoholic solution of 3*2 grams of ammonium
thiocyanate. Precipitation immediately took place, and a yellow oil
sank to the bottom and later solidified to a yellow solid. The whole
mass was filtered on cooling, washed with sJcohol, water, and lastly
with alcohol again. The yield was 10 grams, and the mother liquor
contained free hydrochloric acid. On crystallisation from ethyl alcohol
and glacial acetic acid (1 : 1), 8 grams of the pure substance were
obtained. It forms golden-yellow, compact prisms melting at 138^.
It is insoluble in water or ether, and sparingly soluble in alcohol, but
easily so in benzene or acetic acid. Analysis gave the following
results:
0-2049 gave 0*2566 CO^ and 00352 H^O. C>«3415; H-.1-9L
0-2152 „ 34*6 C.C. moist nitrogen at 17"^ and 755 mmu 17 ^ 18-53.
0*2107 „ 34*8 C.C. „ 20^ „ 759 mm. N = 18*87.
0*2001 „ 0-0910 BaSO^. S = 6-24.
CijH^OijNyS requires C-34-16; Hml*71; N-18-59; St« 6 07 percent.
The substance probably contains two picryl groups. On boiling it
with very concentrated aqueous potash, ammonia was evolved and
ethyl alcohol was found in the distillate ; it must thus contain an
ethoxy-group. This was shown quantitatively by the Zeisel method.
The whole apparatus was kept at 85 — 90^ and an additional bulb con-
taining dilute copper sulphate was interposed to retain the sulphur-
Digitized by
Google
CROCKER: THE PICRIMINOTHIOCARBONIC ESTERS. 437
etted hydrogen formed during the reaction. The result shows that
one ethozy-group is present.
0-2575 gave 0-1107 Agl. OEt = 8-23.
CjgH^Oi^N^S'OEt requires OEt-8'54 per cent.
The substance is only oxidised with difficulty, and is recovered
practically unchanged after boiling for two hours with a 10 per cent,
solution of chromic acid in glacial acetic acid. If boiled for two
hours with a mixture of glacial acetic acid and hydrochloric acid,
hydrolysis takes place, and after dilution and decolorisation of the
resulting liquor, picramide is deposited on cooling in bluish-yellow
crystals melting at 188^. The yield of picramide was about 25 per
cent, of the ethoxy-compound taken.
These facts gave a clue to the structure of the substance. It must
contain two picryl groups, one of which is attached to a nitrogen atom,
forming a picrimino-group, Pi*NI, the other being attached to the
sulphur atom. The reaction is doubtless connected with the remark-
able property which picryl derivatives possess of forming additive
compounds, since neither phenyl thiocyanate nor phenylthiocarbimide
reacts with picryl chloride and alcohol. The formation is explained
by the equations, picryl thiocarbimide, Pi'NICIS, being regarded as the
first product :
Pi-N:o:s + Pia = Pi-N:c<^J^
Pi-N:C<^. + EfOH = Pi-N:C<JpJ + HCl.
The alternative explanation, involving the formation of a pteudo-
thiourethane, Pi«N:C(OEt)*SH, by the addition of a mol. of the
alcohol to the thiocarbimide, is rendered improbable in view of the
fact that picryl chloride is not decomposed by boiling with water or
alcohol. Much less likely is it, then, to act on a SH-gronp. On the
other hand, ethyl chlorocarbonate is easily decomposed by water or
alcohol. This renders it very probable that the ethyl picrimino-
chlorothiocarbonate — which in addition contains two strongly acidic
picryl groups — will easily react with alcohols.
The hydrolysis to picramide is easily understood from the following
equation :
Pi-N:C<JJf + H,0 - Pi-NHj + CO<JJ\
When hydrolysed with potash, the picramide first formed is decom-
posed by the alkali into ammonia and potassium picrate.
The action of picryl chloride and ammonium thiocyanate has been
tried on other alcohols with similar results. It was somewhat
unexpected, however, to discover two isomeric methoxy-ocmpounds
Digitized by
Google
438 cbocrer: the picriminothiogabbonic bstsbs.
easily transformable one to the other. The more stable form, the ^* /3-
isomeride," is obtained in the usual way.
Methyl Picryl Ptonminothioearbanates,
Ten grams of picryl chloride were dissolved in hot methyl alcohol
and mixed with a solution of 3*2 grams of ammonium thiocyanate in
methyl alcohol ; immediate precipitation took place^ and on cooling the
crumbling mass was filtered o£F. The yield amounted to 9'2 grams^ and
the mother liquor contained free hydrochloric acid. The crude product
may be crystallised from a mixture of glacial acetic acid and methyl
alcohol (1:1); it is then obtained in fine, fluffy needles, melting
constantly at 169^, which are insoluble in water or ether, slightly
soluble in alcohol, and easily so in bensene or glacial acetic acid.
Analysis confirmed the view that the substance was methyl pioryl
pieriminothioearbonaU :
0-2027 gave 02437 OOj and 00319 H,0. C - 3278 ; H- 1-76.
0-2534 „ 40-8 c.c. moist nitrogen at U"" and 769 mm. N » 19-18.
0-2245 „ 0-1004 BaSO^. S = 6-14.
Ci^H^Oi3NySrequire8C«32-75; H=1-37;N-19-10; S-6-24 per cent.
The methoxy-group was determined by the Zeisel method in the
usual manner, except for the addition to the apparatus of a bulb of
dilute copper sulphate solution to absorb the sulphuretted hydrogen
formed. The result indicates that one methoxy-group is present in
the molecule :
0-2679 gave 01322 Agl. OCHg- 6-51.
OigH^OijNyS'OOHg requires OOH3=6 04 per cent.
The a-isomeride which was first discovered is prepared from the
)9-compound ; for this purpose, 5 grams of the )3*isomeride are crys-
tallised from 140 c.c. of a mixture of ethyl alcohol and acetic acid (1 : 1).
The crystallisation is carefully watched. Minute, compact prisms are
deposited at first. Immediately the jS-isomeride begins to separate
in fluffy aggregates, the mother liquor is poured off from the crystals,
which are washed with alcohol and then with ether. The yield of the
arisomeride obtained by this means amounts to 1-2 grama The sub-
stance consists of compact, golden-yellow prisms, so like the ethoxy-
compound in appearance as to be almost indistinguishable from it. It
melts at 158% is insoluble in water or ether, sparingly soluble in alcohol,
and easily so in benzene or acetic acid. It is perfectly stable in the dry
state and may be preserved for months unchanged, but in contaet
with solvents it is slowly but completely converted to the /S-isomeride.
The analyses indicate that the substi^nce is meth^fl pieryLpieriminathuh
carbtmaU ;
Digitized by
Google
CROCKER: THE PICRIMINOTHIOCARBONIC ESTERS. 439
0-2042 gave 02471 (JOa and 0'0291 HjO. C = 33-01 ; H = 1 -58.
0-2035 „ 0-2439 COj „ 0-0302 H,0. 0«32-68;,H=l-65.
0-1636 „ 25-3 C.C. moist nitrogen at 18° and 767 mm, N- 1921.
0-2160 „ 0-0912 BaSO^. S = 5-80.
aj^H^OjaK7Srequire8C = 32-75;H=l'37; N=1910; S = 6-24 percent.
The presence of a methoxy-group was shown quantitatively hj the
Zeisel method :
0-2106 gave 00962 Agl. OOU^ = 6-03.
CigH^OigN^S^OCH^ requires OOH3-6-04 per cent.
That the two substances are not polymeric was shown by the identity
of the molecular weights in benzene solution. Owing to the slight
solubility in cold benzene, only one reading could be taken in each
determination.
orlsomeride :
0*1693 gram lowered the freezing point of 21-67 grams of benzene
byO-075^ Mol. wt.=:510.
P'Isameride :
0'0823 gram lowered the freezing point of 16-73 grams of benzene
by 0-048°. Mol. wt.-502.
The calculated mol. wt. is 613.
These are probably not desmotropic forms, because they are perfectly
stable in the dry state. On the other hand, the ease with which they
can be transformed, one to the other, in solution shows that they are
very probably stereoisomerides of the oxime type. Two isomeric
forms are possible, represented by the expressions :
Pi-S-C-0-OHa ^ Pi-S-C-O-CH,
iJ.Pi ^'^^ Pi.il •
The fact that isomerides of the other homologues have not been
prepared is not surprising when the limited means applicable in the
particular case of these substances is considered. When melted, they
decompose, and even when kept at about 130° for some time they are
in most cases completely charred, owing to the high percentage of
nitro-groups contained in them.
n-Propyl Picryl Picriminothwcarbanate.
2*5 grams of picryl chloride are dissolved in 10 c.c. of n-propyl
alcohol and mixed with a hot solution of 0*8 gram of ammonium thio-
cyanate in 10 c.c. of n-propyl alcohol. A flocculent precipitate at once
forips. Op cooling, this is filtered from the mother liquor— which
Digitized by
Google
440 CROCKER: THE PICRIMINOTHIOCARBONIC ESTERS.
is again found to be acid — ^washed with alcol^ol, water, and lastly
with alcohol again. The yield amounted to 2 grams. The product
was crystallised from alcohol and acetic acid (1:1). The pure sub-
stance melts at 151 — 152^. It forms lustrous, goldeo-yellow leaflets,
insoluble in water or ether, sparingly soluble in alcohol, fairly easily
so in benzene, and easily so in acetic acid. Analysis indicated that the
substance was propyl picryl picriminothiocarbonaU :
01861 gave 02408 C^O and 00360 H^O. C = 3529 ; H « 2-16:
0-2323 „ 36-4 moist nitrogen at 16° and 745 mm. N- 17-89.
0-2424 „ 0-1049 BaSO^. S = 5'94.
^16^1 As^tS requires 0 = 35-49 ; H « 2-03; N = 18-1 1 ; S = 592 percent.
iaoBtUyl Picryl Picriminothioearbanate.
2 '5 grams of picryl chloride dissolved in 10 o.c. of wobutyl alcohol
are mixed with a hot solution of 0*8 gram of ammonium thiocyanate
in 20 c.c. of wobutyl alcohol, fhe mass becomes almost solid from
the deposition of minute plates. On cooling, the precipitate is filtered
off from the acid mother liquor. The yield amounts to 2*4 grams.
After washing with alcohol, water, Ac, the product is crystallised
from a mixture of acetic acid and alcohol (1 : 1). The new substance
melts at 173° and consists of golden-yellow, lustrous leaflets^ which
are insoluble in water or ether, but easily soluble in acetic acid.
Analysis showed it to be isobtUyl picryl picritninothiocarhonaie :
0-2288 gave 0-3080 CO, and 0-0448 H^O. C « 36-72 ; H = 217.
0-1971 „ 29'9 c.c. moist nitrogen at 16° and 765 mm. N = 17*80.
0-2194 „ 33-4 „ „ 15°,, 760 mm. N = 17-83.
Cj^HigOigN^S requires 0 = 36*77 ; H - 2*34 ; N - 17-66 per cent.
Similar compounds have been obtained with Mopropyl, allyl, and
benzyl alcohols, and the investigation of these is being proceeded with.
The ease with which this reaction takes place makes it probable that
it will be of use for the characterisation and identification of the
lower alcohols ; in most cases, it takes place easily, even when the
alcohol is diluted with benzene.
In conclusion, I have to thank Dr. G. S. Turpin for his kindness
in permitting me to start this work, which is the result of an observa-
tion made by him in 1891.
University Laboratory,
Cahbridox.
Digitized by
Google
ISOMERIC ADDITIVE COMPOUNDS OF DIBENZTL KETONE. 441
XLIV. — Isomeric Additive Compounds ofDihenzyl Ketone
and Deoocybenzoin with Benzylidene-ip-toluidine,
jxi'NitrobenzyUdeneaniline, and Benzylidene-m-
nitraniline. Part III.
By Francis E. Fbancis, B.Sc, Ph.D.
Experiments similar to those carried out with dibenzyl ketone and
benaylideneaniline (Trans., 1899, 76, 865 ; 1900, 77, 1191) were tried
with this ketone and the benzjlidene-o-toluidine and ptoluidine. As
previously mentioned, the former substance does not appear to give
additive compounds with dibenzyl ketone, whereas the latter readily
reacts with it and with deozybenzoin.
The dibenzyl ketone benzylidene-p-toluidines isolated were similar
to those obtained in the case of dibenzyl ketone benzylideneaniline,
but apparently much less stable ; the a-modification, melting at 163^,
is converted into what is presumably the /3-form melting at 174° by
recrystallisation from benzene containing traces of phenylhydrazine,
and although it may be recrystallised several times under these con-
ditions, its melting point falls to 163° if pure benzene is employed.
A similar change takes place on treating the modification melting at
174° with traces of sodium ethoxide, when a substance melting at
182° is obtained, but again the melting point of this falls to 163° on
recrystaUisation. With regard to deoxy benzoin benzylidene-jE>-tolu-
idine, great difficulty was experienced in obtaining more than very
small amounts of the a-product owing, apparently, to its instability.
A modification of high melting point and slight solubility was simul-
taneously produced from the mixture of deoxybenzoin and benzyl-
idene-jE>-toluidine, and if the temperature was high enough, this was the
only product formed. As it was desired, if possible, to obtain con-
firmation of the work that had been previously carried out, the
further investigation of these substances was abandoned.
m-Nitrobenzylideneaniline was next employed, but the additive
compound obtained with it and dibenzyl ketone turned out to be a
mixture of two substances which could only be separated in the pure
state in small amount and with considerable difficulty. The product
possessing the lower melting point was apparently similar to the
o-modifications previously described, the /3-form obtained from it in
the usual way by recrystallisation from benzene containing traces of
piperidine showed a melting point 31° higher ; this was regarded as a
satisfactory confirmation of previous results where the melting points
of the ^-modifications had never been more than 10 — 11° higher than
those of the a-forms. On the other band, traces of sodium ethoxide
Digitized by
Google
442 FBAKCIS: ISOMERIC ADDITIVE COMPOUNDS OF
which had preyiously given rise to the ^modifications of higher
meltiDg point than /S-forms had but slight effect on either the a- or /3-
forms, and no substance of constant melting point could be isolated.
The product having the higher melting point, produced with the
a^modification and separated from it by its lesser solubility in dilute
acetone, had the same molecular weight. Its melting point was
within a few degrees of that of the /3-form, but since its hydrochloride
was different and its* melting point was raised 10^ by recrystaHisation
from benzene containing piperidine, it was presumably a different
modification.
The difficulty found in obtainiug a sufficient quantity of these
substances led to similar experiments being tried with benzylidene-m*
nitraniline, and, unlike the isomeric m-nitrobenzylideneaniline, thisgave
a pure additive compound with dibenzyl ketone. The resulting
a-dibenzyl ketone benzylidene-m-nitraniline was a stable, reddish-yellow,
crystalline substance, and from it the )3-modification was obtained with
rather more difficulty than had been previously experienced, but when
obtained had strikingly different characteristics. Although its
molecular weight was identical with that of the a-modifioation, its
melting point was 43° higher, and its crystalline form and greenish-
yellow colour, together with its lesser solubility in the ordinary
solvents, sharply differentiated it from that modification. It was also
unaffected by heat, whereas most of these modifications are reconverted
into the more stable a-form. A very similar change to this takes
place when the a-modification is recrystallised from benzene containing
traces of sodium ethoxide ; the substance obtained was, in appearance,
solubility, and stability towards heat, very similar to the )3-form, and
only differed from it in having a slightly higher melting point,
182— 183°, as compared with 177—178°. In previous cases, it had
been possible to further identify the different modifications by means
of their hydrochlorides, but in this instance they were too unstable for
this purpose.
The correspondence between a- and ^ibenzyl ketone bensylidene-«»-
nitraniline and a- and ^ibenzyl ketone m-nitrobenzylideneaniline
is close, and the distinction between the a- and /^-modifications of each
very much more marked than between any of those previously
investigated, and sufficient finally to settle the individuality of
the so-called )3-forms. As regards the y-modifications, the
investigations described have not been so satisfactory ; however, if
the compound obtained from a-dibensyl ketone beniylidene-fn*
nitraniline corresponds to it, and it seems hardly likely that piperidine
and sodium ethoxide should give rise to the same substances, then the
previous statement about the fi- is equally true about the y-modifica- .
tions. Xt is hoped that fuptker investigation will also settle tbia
Digitized by
Google
DIBENZTL KSTONJB AND DEOXTBENZOIK. 4AS
point. The substances of higher melting point which have been pro*
duced in addition to the a-additive componnds in the case of benzyl-
ideneaniline and deoxybenzoin (but not further investigated for want
of material), benzylidene^toluidine, and deoxybenzoin, m«nitrobenzyl-
ideneaniline, and dibenzyltketone and without the a-modification in the
case of benzylidene-m-nitraniline and deoxybenzoin^ appear to be
similar in molecular weight and composition to the modifications
described in this and previous papers as a-, j3-, and yf orms. The inability
so far to obtain sufficient quantities of these substances makes it very
difficult to investigate them further or to be able to state whether or
no the addition of the benzylideneaniline or its derivatives to the
ketone has taken place in the same manner as with the other
modifications.
An attempt is now being made to prepare such a ketone as diphenyl-
acetophenone, for the following reason. According to Schiff, the
constitution of the substances obtained from ethyl acetoacetate and
benzylideneaniline may be formulated in the following manner :
NH C-OH
I
60jEt 64H5 COaEt xy^j^
KetoDio form. Enolio form.
Unlike ethyl acetoacetate, diphenylacetophenone, or a similar ketone
having one GH- but no OH,- group, should react only in one way, that
is, it should give rise to a ketonic modification as pole product,
It is hoped that this investigation may throw some light on the
substances described in this and previous papers, and indicate
whether or no they are similarly constituted to the additive products
obtained by Schiff from ethyl acetoacetate.
EZPISRIHENTAL.
Dibenzyl Ketone and Benzylidene-^tdluidine,
(i) a-Dibenzyl Ketone Benzylidene-^tohi/idine. — ^When molecular pro-
portions of dibenzyl ketone and benzylidenef>-toluidine are kept at 60°
for 48 hours, this substance separates out. It is purified by washing
with light petroleum and recrystallisation from boiling benzene ; the
final yield of pure product was small, It is a white, crystalline sub.
Digitized by
Google
444 FRANCIS: ISOMERIC ADDITIVE COMPOUNDS OF
stance, showing a coDstant melting point of 164% soluble in hot,
' and to a less extent in cold benzene, and recrystallising unchanged
from boiling alcohol, in which it is only slightly soluble. On analysis,
the following results were obtained :
Found 0 = 85-89 ; H = 6-89.
CggHj^ON requires 0 = 85*92 ; H « 666 per cent
a-ffydrochloridey CjqH^yONjHCI. — If the a-additive product is
dissolved in excess of benzene and dry hydrogen chloride passed in, the
salt separates out as a white, crystalline mass melting at 143% It is
dissociated by water, giving the a-base melting at 164% and, unlike the
corresponding hydrochloride of dibenzyl ketone benzylideneaniline,
which is partially converted into the y-base, it is dissociated by
absolute alcohol, yielding the unchanged a-base melting at 164°.
On analysis, the following result was obtained :
HCl found 8*05 ; calculated 8-26 per cent.
(ii) fi' and y-Dibenzyl Ketone Benzylidene-p-toluidine. — Unlike the pre-
viously described compounds, the a-form does not appear to be converted
into the j9-modfication by piperidine in benzene solution, but if the
a-form is recrystallised from benzene containing a trace of phenylhydr-
azine, a substance which appears to be the unstable )3-modification
melting at 174 — 175° separates out. This may be recrystallised from
benzene containing traces of phenylhydrazine and shows an unchanged
melting point, but if the substance be recrystallised several times
from pure benzene its melting point falls to 164% that is, it is recon-
verted into the more stable a-modification. The following results were
obtained on analysis :
Found 0« 85-89 ; H = 6-92.
C29H27ON requires 0-85'92 ; H=6-66 per cent.
If this modification, which is more unstable than any of those
previously described, is dissolved in benzene and treated with traces
of sodium ethoxide, it is precipitated unchanged by light petroleum
after standing for 5 or 6 hours, but if kept at a temperature of 50°
for 12 hours, the product then obtained by the same means shows a
distinctly higher melting point, namely, 181 — 182% but on recrystallis-
ing the substance from pure benzene this gradually drops to 164%
the melting point of the a-modification. There appears to be some
indication, therefore, of the existence of a ymodification, but both
this and the fi- were so unstable that fnrther work upon them wa9
abandoned.
Digitized by
Google
DtB^NZTL ^tOlTE AKD BSOXYBtlN^OiK. 445
DeoQsyh&nzoin and Benzylidene-^ioluidine,
When molecular quantities of these substances were allowed to
remain at 50^ for 2 days, great difficulty was experienced in separating
the additive compounds formed. A partial resolution into two fairly
pure modifications was effected by benzene, the substance possessing
the higher melting point being much less soluble than the other. The
latter, of which only small quantities were obtained, could be
recrystallised from benzene and light petroleum, and showed a fairly
constant melting point of 147^, but from its crystalline appearance it
was evidently impure ; it corresponds, apparently, to the a-modification,
and on melting appeared to be partially converted into the substance
of higher melting point. It gave the followmg results on analysis :
Found G» 85*21; H«6*58.
OggH^ON requires 0-85*93 ; H»6'39 per cent.
The substance of higher melting point can be obtained without the
a-modification by keeping the mixture of benzylidene-p-toluidine and
deoxybenzoin between 65° and 70° duriog the condensation, and on
recrystallisation from boiling toluene melted sharply at 191°. It can
be recrystallised unchanged from toluene containing traces of piperidine.
A determination of the molecular weight in pyridine solution gave 410,
as compared with 391, the calculated value. It evidently corresponds
to the similar substance obtained from benzylideneaniline and deoxy-
benzoin and to others described in this paper.
It gave the following results on analysis :
Found C» 85*58; H-6'65.
C^Hg^ON requires 0 =^ 85*93 ; H » 6*39 per cent.
Dibenzyt Ketone and mrNiirobenzylideneaniline.
(i) a'Dibenzyl Ketone m-yitrobenzylideneaniline, — When the ketone
alid base are mixed in molecular proportion and kept at a temperature
of 50 — 60°, this substance, together with another of much higher
melting point, slowly separates out. The removal of the a-modifica-
tion from this mixture is best effected by fractional crystallisation
from dilute acetone, or, with more difficulty, from chloroform and light
petroleum ; the yield of pure a-product melting at 147° is small. It
is a light yellow, crystalline powder, soluble in benzene or chloroform,
and best recrystallised from benzene or light petroleum. It is not
affected by recrystallisation from boiling alcohol, and may be rapidly
melted without change. It gave the following results on analysis :
Found 0 = 77*42; H-584.
C^Hj^OjN, requires C - 77*06 ; H •» 5*50 per cent.
Digitized by VjOOQIC
446 FRll^GIS: ISOMERtC ADDITlVB OOMPOtrKDS OP
It is slightly basic, bat judging from the instability of the hydro-
chloride obtained by passing dry hydrogen chloride into a benzene
solution of the base, the introduction of the nitro-group has weakened
the basicity. The hydrochloride is a white^ crystalline powder, melt-
ing between 136^ and 137^i and rapidly dissociating, even in dry air,
into the a-base and hydrogen chloride.
The substance possessing the higher melting point obtained, with the
a-modification as described above, can be obtained pure by recrystallisa-
tion from pure benzene. It melts at 179 — 180°, and is a yellow, crystal-
line mass, less soluble in benzene than the a-form. A determination
of the molecular weight in benzene solution gave 412 and 409 instead
of 436, the calculated value. It slowly melts if kept for 11 minutes at
147°, the melting point of the a-modification. If crystallised from
benzene containing traces of piperidine; its melting point rises to
188 — 189°. It gave the following results on analysis :
Found C- 77-39; H-5-78; N-6-40.
CjgHj^OgNj requires 0 - 7706 ; H = 550 j N « 642 per cent
The hydrochloride obtained in the usual way is a white, crystalline
powder melting at 14Q — 149°, and is dissociated by water, giving back
the base melting at 177 — 178°. It gave the following result on
analysis :
Found HCl- 7-40.
OggHj^OgNj requires HCl-7-72 per cent.
(ii) P'Dibenzyl Ketone m-Nitrobenzylideneaniline, — When thea-additive
compound is recrystallised from benzene containing traces of piperidine,
this substance separates out as a light yellow, crystalline mass. It
can be purified by recrystallisation from bensene or chloroform, and
shows a constant melting point of 178 — 179°, that is, 31° higher than the
a-modification. It recrystallises unchanged from acetone or absolute
alcohol, and may be rapidly melted without decomposition or change.
If kept at 147°, the melting point of the a-form, for 18 minutes it melts,
that is, the transformation required 7 minutes longer than in the case
of the substance described above, with a nearly similar malting point.
It gave the following results on analysis :
Found 0-77-06; H-5-61.
Cj^Hj^O^, requires 0-77*06 ; H-6*50 per cent.
The fi-hydroehloride, obtained as previously described, is a white,
crystalline powder melting at 158° and easily dissociated by water or
alcohol, giving back the )3-form melting at 178-^179°. It gave the
following result on analysis :
Found HOI -7-68.
^2B^A^s»^C^ requires HOI - 7*72 per cent
Digitized by
Google
l>lfiEN^YL ^ETOKE A^D DlfiOlCTBENZOIK. 447
(iii) When an attempt was ttiade to obtain the y-modification by
means of sodium ebhoxide, it was only found possible to raise the
melting point of the a-modification from 147° to 153°, and it appeared
impossible to change the aform completely into a substance with
constant melting point ; this was equally true of the /9-modification.
The following diagram illustrates the relationships between the
substances just described :
Piperidine
M. p. 178-180'. — 1 >. M. p. 188— 199'.
Hydrochloride, Not investigated*
Dibeiizyl ketone f ^^^^.-'-''^^ m. p. 148— 149'.
and J
»»i-NitrobenzyIidene-l ^.^ «• «. .,. ^*
aniline. [ ^\^ ^ ^ ,,8_,,,. ^P^^ jj ^ ^j,_^j,,^
Hydrochloride, Hydrochloride,
m. p. 137—138°. m. p. 157—168°.
Dihmzyl Ketone and Benzylidme-m-nitraniline.
(i) a-Dibenzyl Ketone Benzi/lidene-m-^itranUine.*-^The preparation of
this substance was carried out in the usual manner, but a much longer
time, between seven and ten days, at 40 — 50° was required before the
a-additive product separated out. It was purified by recry stall isation
from benzene and light petroleum, and the yield obtained was much
larger than with any of the substances previously described. It
crystallises in long, reddish-yellow needles melting at 134 — 135°, and
may be recrystallised unchanged from boiling alcohol or acetone ; it
may also be rapidly melted without decomposition taking place. De-
terminations of the molecular weight in benzene solution gave 459 and
440 instead of 436, the calculated value. The hydrochloride, a white,
crystalline mass mth indefinite melting point, is very unstable and
rapidly dissociates into hydrogen chloride and the free base.
On analysis, the following results were obtained :
Found C- 7715; H*5-82.
OjgHj^OjN, requires 0 « 77*06 ; H * 5*50 per cent.
(ii) P'Dibenzyl Ketone BenzyHdene-m-nUraniline, — 'If the a-modificap
tion is dissolved in excess of benzene and rather more piperidine
added than on previous occasions, the ^•^form separates out slowly in
two or three days. It is only very slightly soluble in cold benzene.
It is a greenish-yellow, crystalline mass melting at 177 — 178°, or 43°
higher than the a-modification. It remains unaltered when kept for
one hour at 134 — 135°, the melting point of the a-form, and on raising
the temperature melts sharply at 177 — 178°; it may also be rapidly
melted without decomposition or change taking place. A determination
of the molecular weight in benzene solution gave 483 as compared with
436, the calculated value. There is no evidence of the formation of
Digitized by
Google
448 ISOMERIC ADDlTlVfi COMPOt^NM Oi* DtBBlN^YL &BT0N£.
any molecular compound between the a- and /^-modifications, since on
mixing equal quantities of the two and dissolving in boiling benzene,
the )3-form separates out on cooling, and from the filtrate, addition of
light petroleum throws out the a-modification in a nearly pure state.
Like ail other substances of this class so far investigated, it gave no
colour reaction with ferric chloride.
The following results were obtained on analysis :
Found 0 = 77-16; H = 6-80.
CjgHg^OjNg requires 0 = 77-06 ; H = 6-50 per cent.
(in) y-Dihenzyl Ketone Benzylidene-m-nitrantline. — When the a-modi*
fication is allowed to stand for several days in benzene solution
containing traces of sodium ethozide, the substance which slowly
crystallises out resembles in appearance, crystalline form, and solu-
bility, the /^-modification. The melting point is 182 — 183°, compared
with 177 — 178° of the latter, and, up to the present, with the exception
of this, no other points of difference have been found between them.
The hydrochloride, which is as unstable as that of the /^-modification,
has too indefinite a melting point to serve as a method of distinguish-
ing between them. The additive product gave the following results
on analysis :
Found 0 = 77-51; H-5-81.
OjjgHg^OgN, requires 0 = 77*06 ; H = 650 per cent.
Deoxyhenzoin and Benzylidene-m-niiraniline.
Although molecular quantities of these substances were kept for
1 2 days between 30° and 40°, no additive product separated, but on
raising the temperature to 70° the mass slowly solidified. The product
was purified by recrystallisation from large quantities of boiling benzene
or from boiling toluene; it was a greenish-yellow, crystalline mass
melting at 208°. A determination of the molecular weight in benzene
solution gave 412 instead of 422, the calculated value. The following
results were obtained on analysis :
Found 0 = 7713; H = 5-51.
02^Hjj03N2 requires 0 = 7677; H = 5-21 per cent.
Judging from its high melting point, slight solubility, and the fact
that it can be recrystallised unchanged from benzene containing
piperidine, this substance appears to be analogous to the other
isomerides of high melting point which have been previously described.
Digitized by
Google
THE BASES CONTAINED IN SCOTTISH SHALE OIL. PART I. 449
The following table gives the results which have been so far obtained.
The difference between the reaction of the two ketones, deozybenzoin
and dibenzyl ketone, with benzylidene-m-nitraniline is remarkable :
Ketone.
Benzylidene-
aniline
derivative.
a.
Product of
high m. p.
obtained
with the
o-fonn.
fi.
r
Dibenzyl ketone
)>
>>
Benzylidencanilin e
r Benzylidene-p- ^
\ toluidine /
f m-Nitrobenzyl- 1
\ ideneaniline J
f Benzylidene-m- S
\ nitraniline /
164-6"
164
147
184—135
179—180"
174-5"
174—175?
178-179
177-178
182-5"
181—182
182—188
Deoxybenzoin
t»
91
Benzylideneaniline
f Benzylidene-p- "^
\ toluidine /
f Benzylidene-m- "\
\ nitraniline /
154"
147
nil
191"
208
164-5"
178—174"
?
My thanks are due to Mr. Ludlam for his determinations of the
molecular weights given in this paper. The method employed was a
modification of Landsberger's, which he proposes shortly to com-
mnnicate to the Society.
Ukivkesity Collbob,
Bbistol,
XLV. — The Bases contained in Scottish Shale Oil.
Part I.
By F&EDEBic Chables Garrett and John Armstrong Smtthe.
Although many workers have examined the basic substances con-
tained in coal tar, very few have investigated those found in the crude
oil obtained by the distillation of bituminous shale. Greville
Williams examined the tar from Dorsetshire shale in 1855 (Q. J, Chem,
Soe»f 7, 97), and G. Carr Robinson obtained some quinoline bases
from Scottish shale oil (Trans. Bay. Soc. £din., 1879, 28, 561 ; 1880,
29, 265 and 273), but in 1897, George Beilby (J. Soo. Chem. Ind.,
1899, 16, 886) pointed out that practically nothing is known on this
subject^ and that in view of the fact that from seven to ten million
VOL. LZXXI. H H
Digitized by VjOOQIC
450 OARRETT AND SMTTHE : THE BASES CONTAINED IN
gallons of '' bade tar " are obtained in Scotland alone during the year,
it is very desirable that this tar should be fully investigated.
The ** crude oil " obtained by distilling the shale undergoes a
second distillation in large iron retorts, and the distillate is divided
into two portions known as ''green naphtha" (the more volatile
portion) and '' green oil " (the less volatile portion) ; towards the close
of the distillation, a red heat is attained and a considerable quantity of
"stOl coke'' is left in the retort. The oils are then treated with
strong sulphuric acid (about two per cent.), which gives a thick, black
tar, and removes almost the whole of the nitrogenous compounds
from the oil ; after washing to remove the excess of sulphuric acid,
this tar is either burnt under the boilers or thrown away as rubbish.
In the extraction of the pyridine bases from coal tar, the best results
are obtained by washing the " light oil " with dilute sulphuric acid,
and we therefore asked Mr. D. R. Steuart of Broxburn — whom we
have to thank for the great trouble he has taken in order to supply us
with whatever material we have needed — to have some of the ** green
naphtha" treated in this way. Some 200 gallons of ''green
naphtha " were washed with weak sulphuric acid (one volume of acid
in nine of water), giving about 5 gallons of a thin, brown-red, foul-
smelling liquid of sp. gr. 1*13. This acid liquor was heated almost to
boiling and steam blown through for 6 to 12 hours to remove small
quantities of a dark oil having a most offensive smell ; it was then
made strongly alkaline by solid caustic soda (200 grams per litre),
and superheated steam blown through until all the volatile bases had
been driven over. The distillation proceeded rapidly at first, but slowly
afterwards, and a considerable volume of distillate, was obtained ;
from this, the basic oil was separated as completely as possible, and the
aqueous portion distilled until about one-fourth had passed over ; this
second distillate was then made strongly alkaline by caustic soda, the
bases removed and added to the first portion, and the whole dried over
caustic potash. The yield amounted to about 120 grams per litre of
acid liquor, about 3 kilograms being obtained in all. The mixture of
bases was then fractionally distilled, using a " rod and disc " still head
of twenty discs, the receiver being changed as a rule every five
degrees, and the whole quantity being worked over seven times.
The yield was as follows :
Below 120° 0-3 percent.
120—160 13-4
160—200 43-2
Above 200 43*1
100-0 „
Digitized by VjOOQIC
SCOTTISH SHALE OIL. PABT I. 451
It was thought desirable to compare the bases obtained from the
« green naphtha" with those from the ''green oil/' and therefore
20 gallons of each liquid were treated with dilute sulphuric acid (1 lb.
of vitriol in 1 gallon of water) and worked up as already desoribed|
the bases, however, being fractionally distilled only twice.
The ''green naphtha '* yielded 226*3 grams of base, of which 68 per
cent, boiled below 200^, whilst the "green oil" only gave 120 grams,
of which 40 per cent, boiled below 200''. We are indebted to Mc
Arnold Merrick and Mr. W. Saunderson for their assistance in making
this comparison.
Only those portions boiling below 164^ have been examined as yet,
the plan adopted being to ref ractionate each fraction two or three
times, and then to treat each with mercuric chloride, a solution of
the bases in hydrochloric acid being added to a hot 10 per cent, solu-
tion of mercuric chloride. It was generally found most convenient to
use 2 mols. of mercuric chloride to one of base, as the salts obtained
usually contain that amount of mercuric chloride, although occasion-
ally more complex salts were found. The mercury salts were then
recrystallised from water slightly acidified with hydrochloric acid
until their melting points became tolerably constant and analysis
showed them to be fairly pure. The base was next regenerated by
removing the mercury either by caustic soda or by hydrogen sulphide,
and if found to be impure the treatment with mercuric chloride re-
peated. There are very great discrepancies between the boiling points
of the bases recorded by different observers, partly because of the
difficulty of obtaining them in the pure condition, and partly because
of the great influence of variations in the height of the barometer ;
we have determined all boiling points with " short scale " thermometers
of very good quality, the column of mercury being completely immersed
in the vapour.
The following bases have been isolated :
Pyridine B. p. 116—116°
2-Methylpyridine (o-picoline) B. p. 129-5° (763 mm.)
2 : 6-Dimethylpyridine B. p. 142-5° (760 mm.)
2:4-Dimethylpyridine B. p. 159— 159-5°.
2 : 5-Dimethylpyridine B. p. 154 — 155°.
2:4: 6-Trimethylpyridine B. p. 170*5° (763 mm.)
PyridtTie,
Three attempts were made to isolate pyridine by Mohler's method
of precipitation with a strong solution of potassium ferrocyanide (Ber.,
1888, 21, 1015), but without success, as the base recovered from the
precipitate showed no constant boiling point, but distilled between
H H 2
Digitized by VjOOQIC
452 QABRETT AND SMTTHE : THE BASES CONTAINED IN
115^ and 130°. Apparently this method may he advantageously used
for freeing pyridine from small quantities of its homoiogues, hut not
for separating small quantities of the hase from larger amounts of
picoline, &c.
The hase (ahout 10 grams) recovered from these experiments was
therefore hoiled with excess of potassium permanganate solution, and
the liquid distilled; from the distillate, 3 grams of hase hoiling
almost entirely hetween 115° and 116° were recovered; it gave a
yellow, crystalline platinichloride melting at 228 — 230° without de-
composition. On analysis :
0-2665 gave 00915 Pt. Pt = 34-33.
(OBH5N)2,H3PtClg requires Pt = 34*33 per cent.
2-Metht/lpyridine (a-Picoline),
From the lower fractions was isolated a fair quantity of a mercuric
compound of 2-methylpyridine crystallising in small plates melting at
161°:
0-6330 gave 0-4390 HgS. Hg« 59-79.
0-5490 „ 0-3800 HgS. Hg = 59-67.
CgHyN,HCl,2HgCl2 requires Hg = 59-61 per cent.
This salt yielded a hase which was a colourless liquid, easily soluhle
in water, with a powerful pyridine-like odour, hoiling at 129*5° under
763 mm. pressure. Its platinichloride formed orange-red crystals
melting at 194° with decomposition, and on analysis gave 32*65 per
cent, of platinum (calculated 32-72 per cent.). The hase was oxidised
hy potassium permanganate, and after removal of manganese and
potassium salts, treated with copper acetate, when it gave the beauti-
ful violet-blue, crystalline copper salt characteristic of picolinio acid.
From this copper salt, the acid was obtained ; it crystallised from a
mixture of alcohol and ether in thin, colourless needles melting at
134°
2 : i-DiiMihylpyridine,
This base was isolated from the fractions boiling between 150° and
165° by means of its mercurichloride, and is a colourless liquid, easily
soluble in cold, but sparingly so in hot water ; it has a characteristic
cucumber-like odour ; its sp. gr. at 14° is 0*9380 and it boils at
159—159-5°.
Its mercurichloride forms fine needles melting at 1 27°. On analysis :
0-4870 gave 0*3305 HgS. Hg » 58-49.
0-6390 „ 0-6662 AgOl. 01 = 25-77.
0-3622 „ 0-3796 AgCl. 01 = 25-90.
OyH«N,H01,2HgOl2 requires Hg- 58-41 ; 01 = 25*84.
Digitized by
Google
SCOTTISH SHALE OIL. PABT I. 453
The pictate was obtained by direct precipitation as a yellow precipitate
melting at 178—180''.
With auric chloride, the base gave a yellow oil which rapidly
crystallised and melted at 94^ without decomposition. On analysis :
0-4454 gave 01958 Au. Au - 43-95.
O7H0N,HCl,AuC]3 requires Au«44'13 per cent.
The platinichloride crystallises extremely well in orange-red crystals,
which become dim on exposure to air without losing weight ; on slow
heating, they melt at 216^, and on rapid heating at 223^ with decom-
position. On analysis :
0-3020 gave 00951 Pt and 0-4126 AgCl. Pt = 31-49 ; 01 = 33-78.
(CyHgN)3,HjPtClg requires Pt«31-24; Cl«3400 per cent.
On oxidation, an acid was obtained crystallising in white, silky needles
containing water of crystallisation and melting at 235^ with decom-
position ; with ferrous sulphate, it gives a deep red colour ; with copper
acetate, no precipitate^even on boiling ; with silver nitrate, a white,
flocculent precipitate ; and with lead acetate a white precipitate soluble
in excess, the solution depositing crystals on standing.
When heated at 210^ for two hours, the acid decomposed, giving a
white sublimate and an infusible residue ; both the sublimate and the
residue sublimed without melting when heated at above 200°, and
were proved to be ^nicotinic acid by this fact, and also by their
behaviour with ferrous sulphate, silver nitrate, copper acetate, and
lead acetata
The acid is, therefore, lutidinic acid, and the base 2 : 4-dimethyl-
pyridine.
2 : 6'Dvmethylpyridin6,
This base (which had not previously been described) was isolated
from the fractions boiling at 150 — 165° by means of its mercuri-
chloride; this salt forms small, heavy crystals melting at 163°, and
contains 6 mols. of mercuric chloride. On analysis :
0-5260 gave 0-4120 HgS and 0-5472 AgOl. Hg = 67-51 ; 01 = 25-70.
d-6515 „ 0-4315 HgS „ 0-5746 AgOl. Hg = 67-42; 01 = 25-22.
0-9688 „ 1-0140 AgOl. 01 = 25-86.
07H^N,H01,6Hg01j requires Hg = 67-43 ; 01 = 25-86 per cent.
The base is a colourless liquid boiling at 154 — 155° and gives a
picrate melting at 151 — 152°.
The auricfUoride melts at 156 — lb7° without decomposition. On
analysis :
0-1681 gave 0-0738 Au. Au= 43-90.
OyH^N,H01,AuOl8 requires Au- 44-13 per cent.
Digitized by
Google
454 OARRBTT AND SMTTHE : THE BASES CONTAINED t^
The pkUiniMaride is of an orange^red colour and crystallifles well^
but very diverse melting points were observed with the same samples
on different days, the lowest being 195^ and the highest 216°; the
anhi/draui salt melts at 238° with decomposition. On analysis :
0-8671 hydrated salt lost 0-0472 H^O at 105° 2^0*5*44.
(C7Hj>N)2,HsPt01g,2HjO requires H,0 = 6-46 per cent.
0-3169 anhydrous salt gave 0-0970 Pt. Pt= 30-62.
0-2291 „ 0-2290 CO, and 0-9664 HjO. C«27-14; H^317.
0«2502 „ 0-2457 OOj „ 00740 H,0. 0« 26-79 ; H = 336.
(C7Hj>N)2,HjPt01e requires 0 = 26-93; H=>3-21; Pt = 31-24 per cent.
On oxidation, an acid melting at 237 — 238° with effervescence was
obtained. With ferrous sulphate, it gave a fine yellow colour ; with
copper acetate, a bright blue precipitate on boiling; with silver
nitrate, a white, gelatinous precipitate, and with lead acetate, a white
precipitate insoluble in excess. On treatment with methyl alcohol and
hydrochloric acid, it formed an ester crystallising in white needles
melting at 160°.
When heated for 2 hours at 210°, it decomposed, and the residue
sublimed completely ; that this residue was nicotinic acid was proved
by ,its melting at 229° and by its behaviour with copper acetate,
ferrous sulphate, and silver nitrate.
The acid is therefore Mocinchomeronic acid, and the base 2 : 5-di-
methylpyridine.
Since the above was completed, Errera {Ber., 1901, 34, 3699) has
obtained this base synthetically, but in too small quantity for purifi-
cation, and his description in no way agrees with ours, possibly
because his compound had not been thoroughly purified.
2 : 6'IHmMylp$^me.
From the fraction boiling between 140° and 145°, a considerable
quantity of the trimercuric salt of this base was obtained in colourless
plates melting at 160 — 161°. On analysis :
0-6725 gave 0-4900 HgS. Hg « 62-80.
0-5395 „ 0-3935 HgS. Hg- 62-86.
C7H^N,HCl,3HgCl, requires Hg-62-77.
A second merourichloride of this base was obtained from several of
the fractions of low boiling point, and forms small, lustrous scales
resembling cadmium iodide ; it melts at 186°. On analysis :
0*5034 gave 0-2817 HgS. Hg« 48-32.
C^HyNjHClfHgCl, requires Hg» 48*30 per cent.
The base itself was found to be a colourless liquid, fairly easily
Digitized by
Google
SCOTTISH SHALE OIL. PART I. 455
soluble in water, having a characteristic odour resembling that of both
pyridine and peppermint, and boiled at 142 '5° under 760 mm. pressure.
The platinichloride crystallised well and melted at 210° with decom^
position. On analysis :
0-422 gave 0-3176 Pt. Pt« 31-22.
(OyHj,N)2,H2PtClg requires Pt = 31-24 per cent.
The base was oxidised by potassium permanganate, and gave a good
yield of dipicolinic tueid (Ladenburg, Ber^y 1885, 18, 53, and Epstein,
ArvnaUn^ 1885, 231, 1). After jSltering off the precipitated oxide of
manganese, the solution was reduced to a small bulk, acidified slightly
with dilute sulphuric acid, and left to stand for some hours, when a
crystalline precipitate (apparently a mixture of the free acid and an
acid potassium salt) was obtained. This was warmed up with absolute
alcohol and a little hydrochloric acid, and some potassium chloride
filtered ofE ; from the filtrate the acid separated in rosettes of needles
melting at 232°, and these, on recrystallising from alcohol, formed
very small prisms melting at 226° (darkening at 225°), whilst from
water they gave long, hair-like crystals also darkening at 225° and
melting at 226°. Further recrystallisation produced no change in the
melting point, and on no occasion were we able to confirm Epstein's
observation of 236° as the melting point of this acid.
2:4: Q'Trimethylpyridine (B-CoUidine),
From several fractions between 150° and 180°, considerable quantities
of a mercuric compound were obtained which separated in large, fern-
like or saw-like crystals built up apparently of many small plates, and
melted at 154°. The mean of four determinations gave 57*16 per cent.
of mercury (maximum 57*3, minimum 56*97 per cent.), the percentage
calculated for C8HnN,H01,2HgCl2 being 57*22.
From 200 grams of this salt, 25 grams of base were recovered
and again worked up with mercuric chloride ; the salt formed after
being fractionally crystallised yielded 93 grams of crystals melting
sharply at 154° and yielding 13 grams of the base :
.0-483 gave 0-3215 HgS. Kg « 57*38 per cent.
The base was a colourless liquid with a not unpleasant odour ; its
sp. gr. at 20° was 0-917, and it boiled at 170*5° under 763 mm. or at
169*5 — 170° under 7^6 mm. pressure. Its platinichloride was easily
obtained in well defined, orange-red crystals melting at 223—224° with
decomposition.
On oxidation, an acid crystallising in feathery needles, darkening at
225° and melting at 228°, was obtained ; and this yielded, with abso-
Digitized by
Google
456 STEELE AND DENISOK :
lute alcohol, an ester melting at 127*5°. It is therefore trimesitic
acid, and the base must be the 2:4: 6-trimeth jlpyridine.
The authors have pleasure in stating that this work has been carried
out by the aid of a grant from the Besearch Fund Committee of the
Chemical Society.
Thb Durham Collbge of Science,
Nkwcastle-on-Tynb.
XL VI. — The Transport Number of very Dilute Solutions.
By B. D. Steblb, B.Sc, and R. B. Dekison, B.Sc. (1851 Exhibition
Scholars).
In discussing the constitution of certain salt solutions, it was recently
shown by one of us (Steele, Phil, Trans., 1902, 198, A, 105) that if the
change in transport number which undoubtedly occurs with change
in concentration for salts such as magnesium chloride, is assigned to a
variation of the specific velocity of the chlorine and magnesium ions
into which the salt is assumed to be ionised ; and that if u and v re-
present thd specific velocities of cation and anion at a given concen-
tration n, and if u^, and v^, represent the values of the velocities of the
same ions at the concentration n^, a similar relation to the following
holds for a large number of salts.
In the case of calcium chloride, between n — 0*01 and n» 5*0 the
anion transport number p varies between 0*58 and 0*74. In the dilute
solution, tt « —-t; « 0*723*, and in the stronger solution Wj « --Vj
58 74
B 0*350vj, and hence — » 2*06 — , or, assuming the velocity of the
anion to remain constant, that of the cation has diminished by more
than one-half.
It was also shown that, if the coefficient of ionisation is given even
approximately by the relation a; » — , where fi is the molecular con-
Moo
ductivity at the given concentration and /Iqq that at infinite dilution,
then we get the astonishing result that, as the concentration of the
calcium chloride increases, the velocity of the Ca ion is steadily
diminished, whilst that of the CI ion is correspondingly increased.
A far more satisfactory explanation of the change in question is that
first suggested by Hittorf , who assumed the existence of complex ions
Digitized by
Google
THE TRANSPORT NUMBER OF VERY DILUTE SOLUTIONS. 457
in such salt solutions; a solution of magnesium chloride of which the anion
transport number is found to increase with increasing concentration
would, according to this conception, consist of a certain proportion of
simple chlorine and magnesium ions, and in addition to these a certain
number of complex anions ; and since the proportion of the latter
increases with concentration, the transport number would, as in the
case of cadmium chloride, be greater than unity, provided sufficiently
concentrated solutions could be investigated (Kittorf, Pagg, Annalen,
1859, 106, 546).
For a solution containing complex anions of one species only in
addition to the simple ions, if u, v, and v' are the specific velocities of
the cation, anion, and complex anion respectively, and if e is the ionic
concentration of the cation, c' that of the complex anion, then e - c'
is the concentration of the simple anion, and the expression for the
anion transport number is
(c - c')v + cV ct? + c'(t?'— t?) ...
In determining p experimentally by Hittorf's method, the quantity
represented by the denominator is correctly measured by means of a
silver voltameter ; the numerator, on the other hand, is determined as a
concentration change brought about by the migration of the ions, and
hence the degree of complexity of the complex anion has to be taken into
consideration. If m is the number of monad anions into which the
complex would ionise if completely dissociated, then the increase in con-
centration at the anode is proportional to (c - c')v + e'mv\ and the above
becomes
_ cv + c'{inv' -v) _ t? + a(mt;' - 1;) .^.
^ ~ c{u + v) + c'{v''-v) "" u + t; + a(t;'-t;) * • * ^ ''
where a=>— , or the ratio of complex to total anions. If this expres-
sion is put into the form
V -h a{v - 1?) + av'{m - 1)
i; + a(«'-r) + w '
it is seen at once that in order that p should be >1, it is only
necessary that av(m-l) should be greater than u, a relation which
is fulfilled if either a or m is large. For the majority of salts, neither
of these factors attains a sufficient magnitude ; but for zinc chloride
and cadmium chloride,^ is greater than 1 for very concentrated solu-
tions, and the presence of complex anions in solutions of these salts
is universally recognised.
Equation 2 shows at once that no constant value for p can be
Digitized by
Google
458 StEHLS AND BEKISON:
obtained in solutions of a salt which forms complex ions, since a
would vary with the concentration ; for such salts, it is only at
dilutions at which a disappears that the equation takes the form
given by Hittorf, p- .
Since the value for the specific ionic velocity, which is given by the
relation v^pfiL^^ and u^(\'-p)ii.^i depends on the correctness of
Hittorf's equation, it is obvious that, in the case of ions which show a
tendency towards the formation of complexes, a constant value for
u ov V can only be obtained when p is determined for solutions of
such dilution that no complexes exist; for more concentrated solutions,
I'Moo °^v + a; and (1 — /))fboo ^u-x^il the change is an increase of p
with concentration, and vice versd.
If, however, p can be measured in sufficiently dilute solutions of
several salts containing a common ion, then the value for the
specific ionic velocity of the latter should be the same, whatever
the salt from the measurements of which it is calculated. The problem
is, in fact, the determination of the transport number at increasing
dilutions until it is found to remain constant.
For the calculation of u and v, it is therefore necessary that we
should know both the molecular conductivity at infinite dilution
and the '' constant" transport number. The former of these is
obviously not capable of direct experimental determination; but
from measurements at very great dilutions, which have been carried
out by Kohlrausch and others, it can be correctly obtained by ex-
trapolation.
The transport number, on the other hand, if we leave out of
consideration salts of the type of potassium chloride and nitrate,
for which it is practically constant at all concentrations, has not
yet been systematically determined at what may be called " constant *'
dilutions. To the large class of salts, for which, at ordinary con*
centrations, considerable variations in p are found to take place,
belong all salts of dyad and triad ions, and for only a few of these
has the '* constant " range of concentration been reached. This is due
to the fact that the determination of p for very dilute solutions is, for
several reasons, a matter of the greatest difficulty.
In all Hittorfian transport number determinations, it is necessary
that a certain portion of the solution between the electrodes should
remain unchanged in concentration. In the earliest of these deter-
minations, in order that this might be the case, an experiment could
only be carried on for a very limited time, as otherwise, by the migra-
tion of the H and OH ions developed at the electrodes, concentration
changes took place through the whole column of liquid. This difficulty
has been overcome in various ways, the method employed by Hittorf
Digitized by
Google
TttE tRANSPORT NUMBER OF VERY DILUTE SOLUTIONS. 4S9
being the use of a cadmiom anode to prevent the formation of H ions ;
and to prevent the development of OH ions, the cathode was ear-
rounded with a concentrated acid solution.
Hopfgartner {Zeit. p^iysikal, Chem.^ 1898, 25, 115) employed a zinc
anode, and at the cathode a concentrated solution of zinc chloride over
a mercury cathode.
A further source of error is that due to mixing by diffusion, on
account of the large concentration changes that take place during
electrolysis in the neighbourhood of the electrodes. This has been
eliminated quite recently by Noyes (Zeit. phyHkal. Chem,y 1901, 86, 63),
who adds to the cathode and anode solutions respectively in a W-shaped
apparatus, solutions of the acid and alkali of which the original
salt was composed, and of such a strength that the concentration
of the salt at the electrodes remains unaltered. By this means,
extremely accurate results were obtained by Noyes, but unfortunately
the method cannot conveniently be applied to extremely dilute
solutions.
The only measurements of such solutions which have been made,
until quite recently, are those of Bein {Z^L phyaikcU. Chem,, 1898,
27, 1). In all his experiments, however, the amount of matter which
was tiansported did not exceed 9 — 30 milligrams of chlorine, this
being determined as the difference between two very much larger
quantities of material. Accurate measurements of dilute solutions
have recently been made by Jahn's pupils (ZeU. physihal. Chem.f 1901,
37, 674) ; the method employed was one in which the development of
H and OH ions was prevented by the use of a cadmium anode and a
mercury cathode covered by a concentrated solution of copper salt.
A very high voltage was employed, and, in the analytical work, the
limit of possible accuracy was approached ; in some of the experiments,
a very large concentration change at the electrodes took place, appar-
ently without affecting the concentration of the intermediate portion.
The method is, however, not applicable for solutions more dilute than
about iV7150.
Jahn criticises the employment of any method which results in the
development of gas bubbles at the electrodes, remarking that this
gives rise to quite uncontrollable currents, which cause the whole
solution to become mixed. Noyes, on the other hand, obtained
perfectly concordant results by the use of a properly shaped apparatus,
and blank experiments have been carried out by the authors, which
will be described immediately, and which show that in an apparatus
of the shape of that used by Noyes, absolutely no disturbance of
the intermediate portion takes place by the gas development at the
electrodes even after 48 hours.
It is worth pointing out that the only salt the transport number
Digitized by
Google
460 STEELE AND DENISOK:
of which has been determined by both Noyes and Jahn is bariom
chloride, and for this salt the following values were foond for p.
N/50 p = 65-69, 65-81, and 56-84. Mean = 66-78 (Noyes).
i^/60 p«54-2, 54-4, and 54-3. Mean = 64-3 (Jahn).
These two values differ by about 3 per cent., or about ten times as
much as the extreme error in either "series. From the fact that by
Noyes' method no concentration changes occur during an experiment,
his figures should, perhaps, be of the greater value.
The object of the experiments that have been undertaken was two-
fold; firstly, to develop a general method by means of which it
would be possible to determine the transport number at dUutioua
comparable ¥rith those at which accurate conductivity measoremente
are made ; and secondly, to test the question as to whether, from
the results so obtained, constant values for the migration constant
of such an ion as Ca'^'^ would be found. The salts of calciom were
selected for the experiments, because good measurements of their
conductivities have been made at dilutions down to n» 0*0001.
The concentration of the solutions that have been measured varies
between n » 000529 and n » 00025.
It is probably not practicable to work at dilutions greater than
the latter on account of the conductivity of ordinarily purified dis-
tilled water and the practical impossibility of rigorously purifying
such large quantities as are required for the experiments.
In solutions containing so small an amount of salt as those under
consideration, it is obvious that, in order to get any considerable
quantity of salt carried by the current, it is necessary, either to
electrolyse a very large volume of solution, or, using smaller volumes,
to carry the experiment for so long a time that a very large change in
concentration is brought about. If the usual method is employed, the
former of these alternatives requires the use of an apparatus of
unmanageable size, whilst the latter is attended with the danger of
loss of the experiment on account of the backward diffusion and con-
sequent change in concentration of the middle portion.
The apparatus, shown in Fig.l (p.461),admitsof the possibility of elec-
trolysing an unlimited volume of liquid in a vessel of reasonable size.
The electrolysing vessel consists of two U-tubes, A and (7, of about
4*0 cm. diameter, one limb of each being bent away at right angles,
and the two sealed together at B ; two narrow glass tubes are sealed
in at E and £*, and two wider ones at D and />', the total length from
D to £ being about 25 cm. D and IX are connected by means of
rubber tubing to the T-piece, F, which in its turn is joined to the large
stoppered funnel & ; the small tubes, E and J^, are each connected with
pieces of long, narrow bore glass tubing which can be brought out
Digitized by
Google
THE TliANSFORT NUMBEK OF VBEY DILUTE SOLUTIONS. 461
over the edge of the thermostat, these are closed either by stopcocks
or by rabber tubes and pinchcooks. The U-tubes, A and (7, are
each supplied with an electrode vessel which is shown in Fig. 2 (p. 462).
This consists of the cup e, about 4 cm. in height and 2-0 cm. in diameter,
which is attached to the bent up piece of the broad capillary tube a,
the end of which passes through and projects for about 2 cm. into the
cup e ; the other end of the tube a, after passing through the cork, g,
is sealed to the bulb
6, and is provided Fio. 1.
with a stopcock h ;
through the cork g,
passes also the short
tube with stopcock
J, and the glass tube
/, to which is at-
tached the platinum
electrode, 0, which
surrounds the pro-
jecting piece of tub-
ing. The cork g] is
hermetically sealed
with sealing wax
into the open end of
the tube A (Fig. 1).
To prevent the
formation of H and
OH ions at the
anode and cathode
respectively, solu-
tions of alkali and
acid are added.
Working with such
dilute solutions, it
was not found con-
venient to add solu-
tions of sufficient
dilution to counter-
balance the concentration changes, as was done by Noyes, on account
of the very large volumes of such solutions that would be required ;
small quantities of half normal solutions were therefore used in
stead. The experiment is carried out in the following manner.
The apparatus is first placed in position in a large water-bath,
and the two tubes attached to E and E' are brought out over the
edge. All the stopcocks are then closed. G is next filled with
Digitized by
Google
462 STEELE AND DENISON :
the solution to be electrolysed, and this is allowed to ran into
the apparatus, more solution being added to 6? as required. When the
last bubbles of air in i^have been got rid of (by suction if necfessary),
the stopcocks, cZ, are cautiously opened, and the liquid allowed to
rise in the outer limbs of A and (7 to a point about 2 or 3 cm. above
the electrode vessel, o. The stopcocks, d^ are then closed. The
solutions of acid and alkali are next added. If the solution of both
electrodes is to be subsequently analysed, it is necessary to know exactly
the quantity of each solution which is added;
Fio. 2. in all the experiments tabulated, the cathode
solution only was analysed, and therefore the
acid only was weighed. This was done by means
of a glass weighing pipette of about 30 c.c.
capacity, which was weighed before and after
each experiment. After weighing the pipette,
about 1 c.c. of the acid is transferred to the bulb
b I all the stopcocks but d being closed, the aoid
is allowed to run in slowly by carefully opening
the stopcock h; the acid runs in through the
tube a, and since its density is very much greater
than that of the solution, falls over and around
the electrode e and lies at the bottom of the cup c,
A similar quantity of alkali is added in the same
manner to the anode. In order to judge when
the reagent becomes exhausted, small quantities
of an indicator are added simultaneously. In
the majority of cases, methyl-orange was the
indicator employed. From time to time during
the experiment, fresh quantities of acid and alkali
are added in the same manner when they are
shown to be required by the reaction of the in-
dicator. The electrodes are connected with the
terminals of the battery, a silver voltameter whose
platinum cathode had an area of about 1 sq. cm.
being placed in circi^t, and the experiment is
started by inserting the cathode. Under ordinary
circumstances, an experiment arranged in this manner could not
be allowed to run more than 60 to 80 minutes with a voltage of
60 volts without concentration changes reaching the portion B;
but in that time very small quantities of salt will have been trans-
ported. About every twenty minutes the portion of solution con-
tained in the inner limb of the U-tube is removed as follows. To
remove the portion from A, all the stopcocks being closed, H is first
opened, and then very carefully the stopcock attached to £ j the sola-
Digitized by
Google
THE TRANSPORT NUMBER OF VERY DILUTE SOLUTIONS. 463
tion then flows in the direction of the arrows marked (1), and is collected
in a suitable vessel and weighed ; on now closing E and opening If the
current flows in the directions indicated by the. arrows (2), and the
solution from C is thus removed. Unless it is desired to analyse this
portion also it is not collected, but is allowed to run off.
The duration of experiments varied according to the voltage used
from 7 to 36 hours; in the later ones, with a voltage of 170 to 180,
7 to 8 hours was usually sufficient, but with the earlier ones, using
70 volts only, the experiment was allowed to run overnight with
a low voltage, to prevent danger of mixing in B. To test the question
as to whether mixing occurs due to gas development at the electrodes,
the apparatus w^s arranged as described, a small quantity of phenol-
phthalein being added to the whole volume of solution. No liquid was
removed, but the circuit was closed for short periods every half-hour
or so, so as to start the convection currents in the outer limbs of the
apparatus. After a lapse of 48 hours, no trace of alkali could be
detected as having found its way into the portion B of the apparatus.
A similar experiment with litmus failed to detect either acid or alkali
at^.
After the three salts of calcium had been measured, the method as
above described was slightly modified, with the result that subsequent
experiments with potassium choride gave much more closely agreeing
figures. This is probably due partly to the fact that the modification
eliminates certain very small sources of error, and partly also to the
fact that chlorine is capable of far more accurate determination as
silver chloride than calcium as calcium sulphate. The modification
consists in the attachment of long capillary tubes to E and E\ so that,
instead of periodically removing the solution, it is allowed to flow
through the apparatus in a steady but very slow stream during the
whole course of the experiment. At the conclusion of an experiment,
all the stopcocks are closed, and the current is disconnected by
removing the cathode from the silver voltameter; the cathode is
immediately washed in distilled water and dried with alcohol. The
T-pieee, F, is then disconnected from the tubes, D and 2/, and the
solution from B, and to a depth of about 4 cm. in J, is removed
through D by means of a large pipette ; this portion is weighed for
analysis, as the middle portion. On opening now the stopcock attached
to ^, the solution from G is run out. The apparatus is then lifted from
the water-bath, and the remaining solution from the cathode limb. A, is
removed through E^ and finally the whole of this limb, the bulb, h,
and the cup, c, rinsed out several times with small quantities of the
original solution, all the rinsings being added to the cathode solution
for analysis. If it is required to analyse the solution from the anode,
the limb, (7, must of course be treated in the same manner.
Digitized by
Google
464 STEELE AND DENISON:
For analysis, the solations were in all cases evaporated to a small
bulk ; this was done as each portion of liquid was removed from the
apparatus.
The calcium salts were in all cases estimated as sulphate ; in the
case of the chloride and nitrate, the solution was evaporated over a
water-bath in a large porcelain evaporating basin, the evaporation
being continued until the solution to be analysed, which varied
between 2 and 3 litres, was reduced to about 100 ac ; it was then
transferred to an accurately weighed platinum basin of about 250 ac.
capacity, a very slight excess of sulphuric acid was added, and the
solution evaporated to dryness on the water-bath, the basin being
finally ignited at a dull red heat.
In the earlier experiments, the basin was covered during ignition
with a piece of ashless filter paper, which was afterwards burnt, and
the ash weighed separately, but as in no case was a weighable
quantity of ash so obtained, the basin, in subsequent experiments,
was simply covered during ignition with a clean clock-glass.
The analysis of the calcium sulphate solution could not be carried
out in the same manner on account of the separation of the calcium
sulphate, and the impossibility of transferring this without loss from
the porcelain to the platinum basin. It was therefore necessary to
evaporate in a vessel which could be afterwards weighed.
The evaporation of 3 litres of solution in a basin of 250 c.c. capacity
was an operation too slow to be attempted ; a large platinum basin
holdii)g about half a litre, and weighing 300 grams, was therefore
employed, and the solution after weighing was transferred directly to
this, and the solid calcium sulphate weighed in it. The difficulty of
performing analytical operations of great accuracy under these
conditions probably accounts for the greater variations found for the
transport number of calcium in calcium sulphate.
In the experiments with potassium chloride, the chlorine was
estimated as silver chloride, the solution being, as before, evaporated to
a comparatively small bulk, usually about 150 c.c*
The precipitate was collected in a Gooch crucible and weighed, after
thorough washing and drying, at a temperature of 180^. An attempt
was made to estimate the potassium chloride by direct evaporation
and weighing the residue, but this led in all cases to results far too
low, pointing either to volatilisation of the salt in steam from con-
centrated solutions, or, what seems more probable, to loss from
* If it Ib desired, as in traDsport nnmber determinations, to estimate the chlorine
to the nearest tenth of a milligram, this concentration is necessary ; for, since
the solubility of silver chloride in water at 18* amounts to 1*5 m^., corresponding
to 0*4 mg. chlorine per litre, the estimation in very large yolumes of dilute solu-
tions is attended with a constant error of considerable magnitude.
Digitized by
Google
THE TRANSPORT NUMBER OP VERY DILUTE SOLUTIONS. 465
spirting, even when no higher temperature than that of a water-bath
is employed.
The calculation of the results will be best rendered clear by the
following example, that of calcium nitrate. In this experiment, these
were passed through the cathode limb of the apparatus, weighed, and
evaporated, 2702*94 grams of solution, including 26*20 grams of nitric
acid, added to the electrode during electrolysis. Deducting the latter,
the weight of solution analysed was 2676*74 grams, which before the
experiment would yield 0*8932 gram of calcium sulphate. On
analysis afterwards, it was actually found to yield 1*0205 grams of
calcium sulphate.
The difference, 0*1273 gram, gives the actual amount of calcium,
calculated as sulphate^ which had been brought to the cathode by the
current. During the experiment, 0*4517 gram of silver had been
deposited on the cathode of the voltameter ; if 108 is the atomic weight
bf silver, and 68 the equivalent weight of calcium sulphate, the ex-
0*1273x108 , . . ^, ^ ,. 1. fi
pression — - — --____. = 1-^ gives at once the transport number of
Do X 0*4ul7
the calcium ion in CaSO^. In making the calculation for more con-
centrated solutions, a further correction is required for the volume
changes brought about by the movement of the ions during electro-
lysis; but at the dilutions dealt with in the present paper this
correction is quite negligible.
In the table are given the values found for the transport numbers,
and also the data from which these are calculated ; in the first column
are given under N the concentration of the solutions in gram equiva-
lents per litre ; the second column gives the weight of the solution
analysed, these are the actual weights of solution removed from the
cathode less the weight of acid added during the experiment. The
next two columns give the salt contents of this quantity of liquid,
under " original '' being tabulated the contents calculated from analyses
made on the original solution, and under " final " the actual weight of
salt found to be present after electrolysis. The difference between
these is given in the next column as salt transported. In the case of
the calcium salts, all the figures refer to calcium sulphate found on
analysis, reduction to weight of calcium nitrate and chloride for these
salts being unnecessary. In the case of potassium chloride, all the
figures refer to the weight of silver as calculated from the silver
chloride weighed. In the seventh, eighth, and ninth columns are
tabulated the results of the analyses of the middle portion for the
various experiments ; three of these were unfortunately lost.
A comparison of the results given above with those obtained by
previous investigators cannot be made for calcium sulphate and nitrate.
For calcium chloride, values for the transport numbers have been found
VOL. LXXXI. I I
Digitized by VjOOQIC
466 STEELE AND DENISON :
by Hittorfy Beio, and others, which are gathered together and tabu*
lated by Kohlrausoh as follows :
iVT* 10 5 2 1 0-2 005 002 0*01
!-;>- 0-21 0-263 030 0314 034 0-89 0-41 042
which approach with diminishing concentration the value found by ua
for iVT- 0-004.
Our value for potassium chloride is in agreement with all the best
determinations, and is confirmatory of Eohlrausch's fundamental
assumption as to the constancy of the transport number for this class
of salt with increasing dilution.
The Specific Ionic Velocittea.
From the figures given above, it is possible to calculate, by the aid
of the conductivity, the migration constants of the ions |0a, E, ISO4,
Gl, and NO3 ; the results so obtained are grouped together in the fol-
lowing table:
Salt. Moo • 1 -p» Ion« u, v,
CaSO^ 1220 0-441 JCa 53-8
ISO^ 68-2
Ca(N0g)2 115-5 0-450 JCa 52-0
NO3 63-5
CaClg 118-7 0-438 JCa 52-0
01 66-7
KCl 131-5 0-495 K 65-1
01 66-4
To which may be added, for purposes of comparison and discussion,
Noyes' figures for KjSO^ :
K2SO4 135-5 0-496 K 67-2
JSO^ 68-3
Of the above, the figures for E and 01 from KOI merely confirm the
accepted values. It is further seen at once, by a comparison of the
figures for 01 from calcium and potassium chlorides, that, at the dilu-
tions at which p has been determined for these salts, calcium chloride
dissociates in such a manner as to form chlorine ions having the same
velocity as those formed by the dissociation of potassium chloride;
the dissociation at these dilutions is therefore normal. Looking neact
at the velocity of the calcium ion, this is seen to be identical for solu-
tions of the nitrate and chloride ; very dilute solutions of these salts
seem therefore strictly to obey the law of the independent vrander-
ing of the ions and to be comparable with potassium chloride in
more concentrated solutions. If, however, these two salts are
Digitized by
Google
THE TRANSPORT NUMBER OF VERY DILUTE SOLUTIONS. 467
i
fH
5
s
1
o
b
Cation
namber.
CO
1
1
00
kO
»o
00 00
o
o
b
b
b b
b
b b
1
1
kO
«o
to 04
S
eo "^
1 ra
00
lO 1-4
00
00 '^
i
.s
1
s
1
o
lO t^
o o
**
<?
CO <>>
9 9
^
o
b
b b
o
© ' ©
s
J
i
■^
*i
i
C4
•*
*^ «o
04
** rH
&
£
1
00
1
•«* fH
4>
o
b
o o
©
b ©
s
o
•5
^
'S
® d
C4
^
O) «<)
1-i
©4 a
"S s
%a
^ o
^
n
^ 9
^1
^»
1
04
1
00
o
04
eo rH
2 S
1
.
^
si 1
1
s
IS
1
1
l-H lO
eo eo
1
lA
00 l^
O) rH
<o to
CO yt
(
o
O
b
b
o o
'1
b
b b
>
g
1
~ 1
04
00
l-H
00
r-l
oo a»
1
eo
© CO
o
9*
o
rH
rH
9> 9>
O
rH
1-1 r-i
o
o
b
b
o o
b
b ©
1
£
1
j
eo
to
s
s
1
»0 04
si
t^
CO to
SB S
«J0 ©
g
f'i
o
o
b
b
b b
1-i
b rH
1
-i
g
•*>
•2
s
s
s
00 M»
eo o4
00
CO 09
to CO
ts
" -a
K
eo
*«•
??
rH <0
s
00 O)
«-- 00
1
: °
o
b
b
b
b b
b
b b
•a
l§
t»
o
04
•«*
eo o
S
to 21
5P
w
"9
®
00 fO
«
7^ **
II
■^
^
t*. CO
a»
to b
1
l-H
8
s
i-H
CO kO
04
to K^
00
eo
eo to
kO
J2 S
»«*
»«
•«*
04
S ^
OO
eo Oi
^
i
3! S
9 9
o
b
b
b
b b
b
b b
Digitized by
Google
468 THE TRANSPORT NUMBER OF VERY DILUTE SOLUTIONS.
<X)
kO
1
CO
A
"♦
-*
S
b
b
•tl •
m
04
0
00
1-i
04
\a
oa
00
to
^
to
r*
eo
■^
QO
a
a»
a
O)
A
o»
A
A
A
0
'*
0
b
b
b
0
hH
b
b
H«<
b
b
b
b
eo
10
0
0
00
00
r*
In.
■
l-H
to
C4
00
oo
0
A
«o
1
<*
00
w
rH
CO
9
^
t*
1
00
s
0
0
p
0
p
p
p
9
1
9
p
§ ■*
0
0
b
b
0
b
b
b
0
0
b
1
i
•a
"cT
00
O)
«
00
»*
eo
t-
S4
1 .5
1-*
CO
04
«D
00
*«•
i-«
a»
r^
r^
"^
lA
»^
0
00
eo
«o
1
5c
QD
QD
•^
Ctt
«o
«o
t^
1
J9
•n
0
<?
P
p
P
0
9
9
9
p
p
0
0
0
0
"^
b
0
0
0
0
b
b
•^
f-^
-*
fH
I'*
0
eo
?<
fp
?>
,
P
p
<?«
lA
^
p
?*
0
•15
c«
I-H
^
S
00
00
eo
b
00
"^
1
•^
b
^
eo
©1
8
«o
«>.
00
*^
t*
t>.
rH
^g
<N
<N
M
l-l
r-l
r-l
r-*
rH
l-H
09
©4
0
V
an
^1 1
,|
•S
<N
«o
94
g
C4
rH
0
"^
04
4^
04
eo
0
SllVC
(leposi
in
oltamc
0
l-H
^
^
kO
00
l-H
00
Hit
09
t^
|§
1*
00
5
?5
s
0
U3
s
P
S3
kO
3
s
«)
0
0
b
g
0
b
b
0
b
b
0
b
0
-
>•
g
I
■^
" 1
-1
^
Q
1-i
fH
1-i
a»
^
a»
00
Q
rH
»«*
i>.
^
«
*«•
•«*
«N,
l-H
CO
^
6
S
04
0
to
2
3
CO
0
e
P
kO
r^
rH
^
iM
rH
04
iH
rH
r-*
1^
17I
s
0
0
b
b
b
b
b
b
b
b
b
b
0
a»
^
<%
A
to
tN>
10
r^
1
00
»-H
00
«o
r*
»^
• 5
^
C4
'^
^
^
1
3?
Oi
l-H
00
t«
00
1
p^
a»
t^
oo
^-
T**
p
00
9
p
p
9
K^
0
0
b
■ 0
0
»-H
b
0
rH
b
0
0
8
a
■M
13
0
ka
0
00
CO
-^
*a
kO
04
.2
I
c
r^
CO
^
HJI
00
T-t
t^
0
*a
:l
!:i
00
01
t*.
eo
0
r^
^
(N
t*.
0
«o
kO
kO
f^
10
I
c
00
«
t>-
00
p
OI
9
^
OI
p
w
p
0
0
0
0
b
0
b
I-«
b
f^
^
0
fH
b
a
"*
-"T
O)
00 -e
(N
T**
00
9
p
w
9
9
"?•
tf>
p
p
«3>
b
b
-^
o?
l-H
00
t^
»<*
b
HP
kO
•S S3
C^
t'-
»>-
a»
"*
a»
»A
CO
«D
rH
»*•«
eo
•^
CO
to
CO
f-H
t>.
rH
A
^g
(N
<M
M
CO
<N
04
<N
CO
04
"^"^
CO
04
o>
rl
o>
(N
©4
tfs
10
«
t«
"^
l-H
to
CO
4^
fe;
kA
"^
■^
"i*
CO
0
0
p
0
<?
§
s
s
S
§
§
§
g
^
0
b
0
b
b
0
b
b
b
b
b
0
Digitized by
Google
PERKIN AND ALLISON : RHAMNAZIN AND BHAKNSTIN. 469
considered in less dilute solutions, entirely different values are
obtained for u and v; thus, for N/dO calcium chloride, 1 — ^"-0*41,
and accordingly, u»487 and v = 700.
The complex present in this solution is therefore one which
diminishes the apparent cation velocity and increases the apparent
anion velocity, since, by the dissociation of the complex, the latter
cannot presumably be diminished below the value it has in salts such
as potassium and sodium chloride. The dissociation of the complexes
in calcium chloride, and also in the nitrate, appears to be complete.
The value for u, calculated from calcium, sulphate, however, is
found to be 3*46 per cent, higher than in either of the other salts.
This behaviour would appear to be in some way connected with the
presence of the sulphate ion, for the value for K, calculated from the
potassium sulphate figures, is also higher than that calculated from
the chloride or the nitrate, and, as a coincidence, it may be noted that
the percentage increase is about the same ; further, the value of v for
SO4, obtained from the potassium and the calcium salts, is practically
identical.
It is conceivable that this may be due to a very small amount of
hydrolysis occurring in very dilute solutions of sulphates, which at
very great dilutions becomes of sufficient magnitude to give too
high values for the molecular conductivity, and thus increase both
u and V, It is certainly difficult to see how otherwise such enhanced
values can be obtained.
It is our pleasant duty to express our thanks and indebtedness
to Professor Abegg for his kindness and assistance to us during the
course of our work.
Phtsico-Chbmical Section,
Chemical Institute,
Univbrsity of Bbeslau.
XL VII. — Rhamnazin and lihamnetin.
By Abthue George Pbbkin, F.R.S.E., and John Raymond Allison,
B.Sc.
Although rhamnetin has been shown to be a monomethyl ether of
quercetin (Herzig, Monatah,, 1888, 0, 548), the locality of the methoxyl
group has hitherto not been definitely ascertained, and either the (3)
or the 7-position might equally well be assigned to it. Similarly, in
rhamnazin (Trans., 1897, 71, 818), a quercetin dimethyl ether
although the position of one methoxyl group is known, that of the
Digitized by
Google
470 PEBKIN AND ALLISON: RHAMNAZIN AND RHAMNETIN.
second, as in the case of rhamnetin, is uncertain. Thus if rhamnetin
or rhamnazin be decomposed by digestion with boiling alcoholic
potash or by the aspiration of air through its alkaline solution, .
protocatechuic acid and vanillic acid respectively are produced
besides a syrupy phloroglucinol derivative; the yield of the latter,
however, is small, and being readily soluble it is difficult to identify ;
moreover, preparation of the raw material, especially rhamnazin, is
extremely laborious. For the phenols in question, three constitutions
were possible, namely, that of phloroglucinol monomethyl ether, or
those of the hydrozyfisetol monomethyl ethers,
OMef^OH 0h/\0H
IJcO-CHj-OH and i JoO-CHa-OMe ;
OH OH
these two being suggested by Herzig's {Monatah^f 1891, 12, 187) de-
composition of fisetin tetramethyl ether into fisetol and veratric
acid.
The study of the azobenzene derivatives of phloroglucinol (Trans.,
1897, 71, 189, 1154) and obher compounds derived from it, has shown
that such substances are sparingly soluble and readily crystallised,
and it appeared likely that the phenols from rhamnetin and rhamnazin
might be identified by these means.
Bhamnazin was decomposed by digestion for several days with boil-
ing alcoholic potash, but this is more rapidly accomplished (in about
two hours) by the aspiration of air through its solution in dilute
aqueous potassium hydroxide. The phenolic product of the reaction
which is the same in both cases, was isolated in the usual manner, and
its solution in dilute sodium carbonate treated with diazobenzene
sulphate until a precipitate no longer formed ; this, which is orange-
red, was collected, well washed, transferred to a dish, and dried on the
water-bath. Extraction with alcohol removed a resinous compound,
and the residue, after being crystallised two or three times from a
mixture of alcohol and acetic acid, gave an average yield of about ten
per cent, of the rhamnazin employed :
0-0919 gave 0-2208 00^ and 0-0377 H^O. C = 65-61 ; H = 4-54.
00844 „ 11-7 C.C. nitrogen at 16° and 754 mm. N= 1602.
(CeH5Nj)jOgH30j-OaH3requiresC = 65-51; H = 4-69; N = 1 609 per cent .
It formed glistening orange-red needles sparingly soluble in alcohol
and melting at 250—252"^.
EhamneHn was decomposed by the same methods, and the phenolic
product converted into its disazobenzene derivative. This melted at
250 — 252°, and was identical with that obtained from rhamnazin :
Digitized by
Google
PERKIK AND ALLISON: RHAMNAZIN AND RHAMNETIN. 471
0-1136 gave 02715 COg and 0-0475 ^,0. C-66-17 ; H«4-64.
0-1167 „ 15-6 O.C. nitrogen at 14° and 746 mm. N » 15*57.
0-1788 „ 0-1234 Agl. OCH3 = 4-40.
CigH^OjN^-OCHg requires 0 = 6551 ; H = 4-59 ; N^ 16-09 ; OCHj^
4*31 per cent.
This compound is therefore disctzobenzene pfdoroglucinol moncmeihyl
€ther, and the phenol obtained both from rhamnazin and from rham-
netin is phloroglucinol monomethyl ether. The relationship of these
colouring matters may, therefore, be indicated by the f ormulsB,
O OMe O OH
OH CO OH CO
RhamnaziD. Rhamnetin.
and the suggestion (loc, cit.) that rhamnazin was rhamnetin mono-
methyl ether is thus shown to be correct.
The shades given by these colouring matters upon mordanted wool
are as follows -.
Ghromiom. Alaminiam. Tin. Iron.
Rhamnetin Red-brown. Brown- orange. Bright orange. Olive-black.
Rhamnazin Golden-yellow. Orange-yellow. Lemon-yellow. Olive-brown.
These results show that the dyeing properties of rhamnetin are
identical with those of quercetin, and are interesting in that they
prove that in quercetin the hydroxyl (7) * has no effect on its dyeing
properties. On the other hand, the replacement of the hydroxyl (3')
by methozyl with production of rhamnazin (compare also taorhamnetin,
quercetin monomethyl ether [0Me=»3'] Trans., 1898,73, 267) has a
most marked effect on the dyeing properties ; this was to be expected,
as the compound does not then possess o-hydroxyl groups. A third
quercetin monomethyl ether has been shown to exist in minute quantity
in the Tamaria Afiricana (Trans., 1898,73, 380), and an attempt is now
being made to obtain sufficient substance for the location of its
methoxy-group in the above manner.
Qftercetin Tetrametht/l and Tetraethyl Ethers, — These compounds, as
Herzig has shown (Afanatah., 1888, 0, 552), when decomposed with alco-
holic / potash, give respectively protocatechuic acid dimethyl and
diethyl ether, and also phenolic compounds which are derivatives of
phloroglucinol ; to determine the constitution of the latter, their azo-
0 _0H
,,y\4/(5-OH
Digitized by
Google
472 FERKIN AND ALLISON: BHAMNAZIN AND RHAMNETIN.
benzene compounds were examined, with the result that the one derived
from quercetin tetramethyl ether was found to bo dUazobenzene phlaro-
gludnol manomethyl ether, m. p. 250 — 252^
The phenol from the tetraethyl ether gave a disazobenzene compound
crystallising in orange-red needles melting at 212 — 214° :
0-0828 gave 0-2015 COj and 0-0365 H^O. C « 66-36 j H = 489.
0-0771 „ 10-5 cc. nitrogen at 18° and 751 mm. N« 15-53.
(CeH5N,)2CeH(OHVOCjH5 requires C- 66-24; H«4-94; N = 16-47
per cent.
It was evidently disazobenzene jMorcgludncl rrumoeihyl ether,
Kampherol Monomethyl Ether, — A trace of this substance, recently
described by Testoni {Gazzeita, 1900, 30, ii, 327) as a constituent of
Gkilanga root {Alpinia officinarum), was available for examination. When
decomposed by the aspiration of air through its alkaline solution, it
yielded benzoic acid, (m. p. 121°), and a phenolic compound which gave
the phloroglucinol reaction. In the presence of sodium carbonate, this
gave an azobenzene compound, identified as trisazobenzene phloro-
glucinol, and consequently there was no methoxy-group in this portion
of the molecule. Adopting the constitution for kampherol suggested
by Kostanecki, it is evident that the above methyl ether must be re-
presented as follows :
l\y\^C-OMe
OH CO
This method of analysts has already been employed with the decom-
position products of the ethers of luteolin (Trans., 1900, 77, 1314),
myricetin (Trans., 1902, 81, 203), and genistein (Trans., 1900, 77,
1310), with the result that in all cases the ethers of phloroglucinol
were isolated. There is no reason to doubt that the corresponding
derivatives of chrysin and apigenin would by similar methods give a
like result, and it appeared unnecessary to undertake their preparation
for this purpose.
The authors are indebted to the Besearch Fund Committee of the
Chemical Society for a grant which has been in part employed to cover
the expense of this research.
Clothworkers' Resbaroh Laboratory,
Dyeixo Department,
Yorkshire College.
Digitized by
Google
ROBININ, VIOLAQUERCETIN, MYRTICOLOBIN, AND 08YRITRIN. 473
XLVIII. — Robinin, Violaquercetin^ Myrticolorin, and
Osyritrin.
By Arthur George Pebkin, F.K.S.E.
Eobimn.
Ik a previous communication (Trans., 1900, 77, 430), it was shown
that the leaves of Rohinia Faeitdaeacia contain acacetin, a monomethyl
ether of apigenin, an interesting point in view of the discovery by
Zwenger and Dronke (Anruden Suppl., 1861, 1, 263) that the flowers of
the same plant contain a quercetin glucoside, robinin. As these
flowers, however, are practically devoid of dyeing properties, it seemed
doubtful whether a quercetin compound was present, and reference to
the work of these chemists made it probable this suspicion was well
founded. At the time of Zwenger and Dronke's investigation, the
obstinacy with which these colouring matters retain water of crystal-
lisation was not fully appreciated, and results' accurate in themselves
frequently received a wrong interpretation. The analyses of their
" quercetin " were made with material dried at 100°, a temperature at
which it is rarely rendered anhydrous ; again, it was unlikely that
picric acid would result from the action of nitric acid on quercetin.
Some difficulty was at first experienced in procuring raw material
for this investigation, but ultimately this was overcome by the kind-
ness of Dr. J. van Rijn, of Maasstricht, who was good enough to
superintend the gathering and drying of some quantity of the flowers.
For isolating robinin, Zwenger and Dronke digested the flowers with
boiling water, subsequently evaporating the extract and treating the
residue with alcohol. The following method is more rapid, and is suit-
able for dealing with small quantities of raw material.
The flowers were digested with ten times their weight of boiling
alcohol for 4 hours, the mixture strained through calico, and the
residue well pressed and again treated in a similar manner. The pale
green extract, which deposited a wax on cooling, was concentrated by
evaporation, poured into water, and the mixture extracted with ether,
the alcohol contained in the aqueous liquid being removed by distilla-
tion; on standing overnight, this solution deposited crystals which
were collected and washed with a mixture of alcohol and chloroform
until the washings were colourless. The residue was then purified by
two or three crystallisations from water with the aid of animal char-
coal. One hundred and ninety grams of the flowers gave 1*76 grams,
or 0*82 per cent., of robinin in the crude condition, this being reduced
approximately one-half on purification.
As thus obtained, it consisted of extremely pale yellow needles sinter-
YOU LXXZI. K K
Digitized by
Google
474 PERKIN : EOBININ, VlOLAQtTEBCETlN,
ing at 190° and melting at 196—197° (Z. and D. 195°), and had the
general properties mentioned by these authors. When dried at 100° :
0-1103 gave 0-2123 CO, and 0-0550 Hfi. C-52-49 ; H = 5-54.
CjjH^jO^ requires 0 = 62-24 ; H = 5-54 per cent.
Zwenger and Dronke assign to robinin at 100° the formula OgsHg^O^^
(0»51-19; H = 5*ll), and with this their analytical figures closely
agrea This discrepancy suggested that our preparation might contain
a trace of free colouring matter which would raise the percentage com-
position. An alcoholic solution of a second preparation was therefore
poured into ether (in which the colouring matter is soluble, but not
the glucoside) and the precipitated product collected, again crystallised
from water, and dried at 100° :
0-1131 gave 0-2166 00, and 0-0575 H,0. O-52-20; H«6-64.
The glucoside was thus evidently homogeneous.
Determinations of the water of crystallisation contained in the
glucoside dried at the ordinary temperature, employing two distinct
preparations, gave the following results :
1-5790, at 100°, lost 02450 H^O. H,0 - 15*51.
1-6079, „ „ 0-2605 H,0. H,0 = 15-58.
^88^42^20'^^s^ requires H,0« 15*97 per cent.
C88H«0,o.7iH,0 „ H,0-1511 „
Air-dried robinin thus crystallises with SHfi. These results are
not in agreement with those of Zwenger and Dronke, who found 14-53
per cent, of water, agreeing with the amount required for the formula
^86^80^i6»^i^2^> or according to the above, G^K^fi^,7ILfi. This
discrepancy is curious, although the lower figures these authors give
for robinin dried at 100° suggest the possibility that their product still
contidned some water of crystallisation.
Decomposition toith Acid. — The dried glucoside was boiled with dilute
sulphuric acid * for at least 2 hours, and after standing overnight, the
liberated colouring matter was collected, washed, and dried at 100° :
0-6110, at 100°, gave 02330 colouring matter. Found, 38-13.
0-9804, „ „ 0-3745 „ „ 38-19.
1*2265, „ „ 0-4650 „ „ 37-92.
Zwenger and Dronke, on the other hand, found that air-dried
robinin gave 37*96 per cent, of '* quercetin " dried at 100°, an amount
considerably in excess of that given above. Thus, air-dried robinin,
^88^42^20-^-^s^' should give 33-70 per cent, of the colouring matter
OigH^oO^^HjO, or but 31*70 per cent, of the anhydrous substance. In
their paper, they state that robinin is '' extremely easily " deoom-
* 750 c.c. of 1 per cent. Bolntion for each gram of glaooaide.
Digitized by
Google
AiTRTtCOLORtN, AND OSTBtTRtN. 476
posed by boiling dilute sulphuric or hydrochloric acids, but with
acid of 1 per cent, strength this was not found to be the case, for]with
a short digestion, 39*54 per cent, of '* colouring matter " was obtained,
a fact suggesting that it was contaminated with unaltered glucoside.
Tht Colouring Matter, — For analysis, the product was recrystallised
from dilute alcohol and dried at 160° :
0-1122 gave 0-2589 COj and 0-0355 H,0. C- 62-93 ; H = 3-51.
CigHjoOg requires 0-62 93 ; H: = 3-49 per cent.
Zwenger and Dronke found that this colouring matter, dried at
100°, gave 0=59-31, n» 4*49, numbers in close agreement with those
required for the formula C^j^EL^QOf^^Hfi (0 = 59-2, H-3-94).
Prepared as above, it crystallised in slender, yellow needles melting
at 276 — 278°, readily soluble in boiling alcohol, and soluble in alkaline
solutions with a pale yellow colour. For additional proof that it was
not quercetin, comparative dyeing trials were carried out employing
woollen cloth mordanted with chromium, aluminium, tin, and iron.
Chromium.
Qnercetin Red-brown.
Robinin colooriDg matter Brown-yellow.
With mineral acids in the presence of acetic acid, it yielded crystal-
line compounds, and to confirm its molecular weight the sulphuric acid
derivative was analysed.
0-1556 gave 0-2657 00, and 0-0457 HgO. 0 =» 4657 ; H = 3-26.
OjgHioO^jHgSO^ requires 0 = 46-87 j H= 3-12 per cent.
Alcoholic potassium acetate yielded a monopotaasium salt, but
owing to lack of material this was not fully investigated. The acetyl
derivative, prepared in the usual manner, crystallised from methyl
alcohol in colourless needles :
0-1093 gave 0-2438 00, and 00400 H^O. 0 = 60-83 ; H = 4-06.
C^iT3:fif^{C^'Efi\ requires 0 - 60-79 ; H = 3*96 per cent.
. When fused with alkali, the colouring matter gave p-hydroxt/benzoio
acid (m. p. 208 — 210°) and phloroglucinol.
These facts, together with a comparative dyeing trial, conclusively
proved that the colouring matter derived from robinin is identical with
that isolated from the flowers of the Delphinum Consolida (Proc., 1900,
16, 182). This similarity was corroborated by the peculiar behaviour
of their acetyl derivatives when heated ; thus, while some preparations
melted at 180 — 182°, others became liquid at about 116°, resolidified
as the temperature rose, and finally melted at 180 — 182°. The colour-
K K 2
Alominiom.
Tin.
Iron.
Brown-orange-
Bright
Olife-black.
yellow.
onnge.
Full golden-
Lemon-yellow.
, Deep oUve-
yellow.
brown.
Digitized by
Google
476 PERKIN: ROBININ, VIOLAQUERCETIN,
iDg matter is in reality kampherolf first obtained by Gordin {Dug.
Berne) by the decomposition of its monomethyl ether kampheHde^ which
is contained in galanga root {Alpinia qfieinarum). This, as Kostan-
ecki suggests (Ber., 1901, 34, 3723), has in all probability the con-
stitution
and may be considered as the connecting link between apigenin and
quercetin. Kampherol, and not quercetin, is the colouring matter pro-
duced when robinin is hydrolysed with acid.
The Sugars. — ^The acid filtrates formed by the decomposition of
robinin were neutralised with barium carbonate, filtered, and
evaporated to a small bulk. The product yielded an osazone which
after three crystallisations from dilute alcohol was obtained as a
spongy mass of yellow needles sintering at 165^ and melting at
178 — 180^. As a further treatment in this manner did not yield a
homogeneous substance, it was dissolved in alcohol, the solution
poured into ether, and well washed with water. On slow evaporationi
a small quantity of crystalline matter separated, and this was collected,
washed with ether, and recrystallised from alcohol. It melted at
204 — 205^ and resembled glueoaazcne.
The filtrate which contained the main bulk of the osazone, on
spontaneous evaporation, deposited crystals which were extracted with
benzene, washed with traces of ether, and recrystallised from dilute
alcohol. In the preliminary notice (Proc., 1901, 17, 87), it was
considered probable that this substance was galactosazone, but fer-
mentation experiments kindly carried out for me by Dr. Turnbttll»
of the Leather Industries Department, did not corroborate this view.
Employing the sugar solution, as obtained in the above manner from
robinin, a slight fermentation did indeed occur, but the main bulk
was unattacked and gave an osazone meeting at 180 — 182^. This
was found to be identical with rhamnoactzone prepared from pure
rhamnose, thus harmonising with the results of Zwenger and Dronke.
Whether the trace of the glucose derivative simultaneously isolated is
an ingredient of robinin itself or is derived from a second glucoside of
kampherol also present in minute quantity must, although it appears un-
likely, remain a matter for conjecture. Very large quantities of raw
material would be required to decide this point. According to these
views, the formula of robinin is consequently either C^^H^fi^ or
C33H42O201 ftnd its decomposition by acid may be represented as
follows :
Digitized by
Google
MYRTICOLORIN, AND OSYRITRIN. 477
C88H44O80 + 4H,0 - OijH.oO, + 3CeHj,0a.
CS8H42O20 + ^HjO = CijHjoO, + 2C,H,,0e + C.H.gOe.
It is thus a most interesting glnooside, as it appears to be the first
known substance of this class to contain three sugar nuclei. The
above equations require respeotivelj a yield of 37*63 and 37*73 per
c^nt. of kampherol, whereas that actually obtained was 38*13,
38*19, and 37*92 per cent.
Dyeing experiments carried out in the usual manner with woollen
cloth showed that robinin is almost devoid of tinctorial properties ; this
was to be expected from the preliminary examination of the flowers.
Tht Idmvtity of Oayritriny Violaquercitrin and Myrticolorin,
In a previous communication (Trans., 1897, 71, 1132), it was shown
that the leaves of the Colpoon cmnpreasum contain a queroetin glucoside,
osyritrin, to which, dried at 130°, the formula O^^^ifivi ^^ assigned.
At that time, the raw material available yielded little more than
1 gram of the pure substance, but subsequently, through the kindness
of Mr. R. H. Goaton, of Wellington, Cape of Good Hope, a larger
supply was obtained, and it became possible to investigate it more fully.
Water of CrystaUiaation, — For determining the loss of weight which
the air-dried glucoside undergoes, it was exposed for about three
weeks over sulphuric acid in a desiccator. As the product on exposure
to the atmosphere rapidly assumed its original weight, the analysis
was performed indirectly in this manner.
1*3345, dried as above, gained 0*0770 H^O. Found 5*76.
C^H^QO^^ffiUfi requires HjO = 5'44 per cent.
At 130°, employing an oil-bath jacketed with amyl alcohol (b. p.
130—131°), it suffered a further loss of 0*0200 HjO, or 1*49 per cent.
(^HjO requires 1*43) and finally, at 160° (oil-bath jacketed with tur-
pentine), 0*0220 gram H,0 was evolved, or 1 '67 per cent. (^HjO = 1 *47).
Osyritrin, therefore, crystallises from water with 3 mols. of water of
crystallisation.
Anhydrous osyritrin is remarkably hygroscopic, thus, on exposure
for 1 hour in the air of the room, it completely regains its water of
crystallisation. On analysis :
0*1170 gave 0*2275 COj and 00537 H,0. C « 5303 ; H - 5*09.
Oj^HjgOi^ requires C - 53*28 ; H - 4*60 per cent
Decomposed with dilute sulphuric acid:
0*8400 gave 0*4137 O^^H^fi^. Found 49*25.
CjyHjgOig requires O^^'K^qO^ « 49*67 per cent.
Digitized by
Google
478 PEBKIN: ROBININ, VIOLAQUBRCETIN,
The osazone of the sagar melteKl at 204 — 205°, as previously founds
and was evidently gluoosazone.
Osyritrin dried at 160^ has thus the formula Cg^H^gOi^, and at 130°
(0^K2fiiQ)^r^2^* and not Og^Hi^O^^ as formerly suggested. Its hydro-
lysis with acid is analogous to that of rutin, which in a similar manner
gives quercetin and rhamnose (Schunck, Trans. ^ 1888, 63» 264) :
C27H280ia + SH^O = Ci^HjoO^ + 2C,H„0«.
Rutin.
ViolctqiAercUrin.
Mandelin (Jakreaber.y 1883, 1369) isolated this glucoside from the
flowers of the Viola tricolor variensis, and assigned to it the formula
^42^42^24* ^^ ^^ ^^^ previously pointed out that this is more
correctly O^lS^O^f^t as the true molecular weight of quercetin was not
known at that time.
For its preparation from the flowers, a method identical with that
employed for the preparation of osyritrin (loc. eit) gave an excellent
result:
01071 , dried at 130°, gave 0-2036 COg and 00510 H^O. C = 51-84 ;
H = 5-29 percent.
0-1144, dried at 160°, gave 0-2230 CO, and 0-0535 H,0. C = 53-16 ;
Has 5*19 percent.
As Mandelin employed water for the isolation of his substance, it
seems possible that the result he obtained was due to the contamination
of his product with a trace of quercetin.
Water of CryataUiaatum, — ^When dried over sulphuric acid, 0-9702
gained, on exposure to the atmosphere, 0*0580 H^O. Found SL^O =
5*64 ; 2H.fi requires 5*44 per cent.
1 -0280, at 130°, lost 0*070 H^O. Hfi = 6-80.
2-5HjO requires HjO = 6-79 per cent.
1 -0280, at 160°, lost 0-085 HjO. B.fi = 826.
2iH.fi requires HjO = 8-16 per cent.
Yiolaquercitrin thus crystallises with ZHfi,
When dried at 160° and decomposed with acid, it gave 49-35 per
cent.* of quercetin, which is in accordance with the following equa-
tion (49-67 percent.):
CgrHjgOie + SHjO = 0^,H^fi^ + 2Q^H^fi^.
The osazone of the sugar melted at 204 — 205° and was evidently
glucosazone. Yiolaquercitrin melts at 186° when slowly and at 190°
when rapidly heated, and is undoubtedly identical with osyritrin.
Digitized by
Google
MYRTICOLORIN, AND OSYRITRIN. 479
Mt^Hcolorin.
This quercetin glucoside was isolated by H. G. Smith (Trans., 1898,
73, 697) from the leaves of the Eucalyptui 'mctororhynoha^ who
assigned to it the formula 0^'S.^O^^^ and represented its decomposition
with acid by the equation
CjffHaOitf + 3H2O « CjjHioOy + SO^HjA'
Smith at first considered the sugar thus produced {foe. eii.) to be
galactose, but more recently, in a private communication, he informs
me that this is not the case, as he has obtained from it glucosazone,
(m. p. 204 — 205^). It is worthy of note that he {loo. cU.) calls atten-
tion in his paper to the remarkable similarity of myrticolorin to
osyritrin.
The raw material employed was some commercial myrticolorin, for
which I am indebted to the kindness of Mr. Smith. After purifica-
tion, a sample dried at 160° gave 0=» 53*03 and Ha 5*09 per cent,,
and on decomposition with acid 49*25 per cent, of quercetin :
Dried over sulphuric acid, on exposure to air it gained H^O = 5*76 percent.
„ „ and heated at 1 30"" it lost H^O « 1 '49 per cent.
Dried at IZ(P and heated at 160'' it lost H^O^^ 1*67 per cent.
Myrticolorin, therefore, crystallises from water with SH^O, and this
may be fractionally removed by methods identical with those employed
in the case of osyritrin. As the melting points, general reactions, and
dyeing properties also agree, there can be no doubt that the auhatcmoea are
ideniiealf moreover, it was previously shown (Trans., 1899, 76, 433) that
osyritrin, violaquercitrin, and myrticolorin give, by means of potassium
acetate, monopotassium derivatives (found K =3 6*21). There is no
doubt, therefore, that these also must be regarded as one of the same
compound, represented by the formula C^yH^O^^ (K»6*03 per
cent.).
Certain glucosides, as ruberythric acid and the purpurin glucoside
contained in madder are decomposed during the dyeing operation by
means of the mordant, which combines with the alizarin or purpurin
thus liberated. Such, however, I find is not the case with the known
glucosides of the quercetin group, which are dyestuffs of themselves,
and give shades diifering considerably in most cases from those yielded
by the colouring matters from which they are derived.* A simple
experiment with persian berries clearly illustrates this point. This
dye contains glucosides of rhamnetin, rhamnazin, and queroetini and
• Nietzki(" Chemistry of Organic Dye-etuflfs,*' 1892, 256) consideri it probable
that in dyeing the qnerdtrin splits up, and that the shades obtained are dqe to
tl^e fonnatlon of quercetin lakes,
Digitized by
Google
480 perman: the influence of salts on the
there is also present a ferment which at about 40° in the presence of
water hydrolyses these componnds ; if, therefore, the dye-bath be
raised slowly to the boiling point, this change occurs, and the resolting
shade is due to the free colouring matters and not to the glucosides. On
the other hand, if the berries be plunged into boiling water, the activity
of the ferment is at once destroyed; and the tinctorial property of this
extract is now due to the glucosides as the shade indicates. This
difference is very similar to that shown between quercetin and its
glucosides, which gave the following results with woollen cloth mor-
danted in the usual manner :
Tin. Iron.
Lemon-yellow. Deep olive.
Bright orange. Olive-black.
Lemon-yellow. Dull brown.
Lemon-yellow. Dall brown.
It is interesting to observe that rutin and osyritrin have identical
tinctorial properties, which points to the fact that the two sugar nuclei
of each (in the one casd rhamnose and the other dextrose) are similarly
attached to the quercetin residue. As regards their actual disposition,
satisfactory proof is at present wanting, but it is highly probable that
one at least is attached to the catechol group, as the dyeing properties
suggest the absence of o-hydrozyl groups. Their behaviour with
potassium acetate is an indication that they contain intact the hydr-
ozyl of the pyrone ring, for it is found that galangin, by this method,
forms a monopotassium salt. The properties as a whole would harmonise
closely with those of a compound containing both sugar nuclei attached
to the catechol group, but further evidence is needed before this conclu-
sion can be adopted.
Cloth workers' Research Laboratort,
Dtsing Department,
Yorkshire College.
Chromium.
Aluminium.
Quercitrin..
. Full brown -yellow.
Full golden-yellow.
Quercetin ..
Red-brown.
Brown-orange, in-
clining to red.
Osyritrin ..
Brown-yellow.
Full golden-yellow.
Rutin
Brown-yellow.
Full golden-yellow.
XLIX. — The Influence of Salts and other Substances on
the Vapour Pressure of Aqueous Ammonia Solution.
By Edgab Philip Pebman.
The author recently published a series of measurem.ents of the vapour
pressure of aqueous ammonia solution showing the effect of alteration
of concentration for certain selected temperatures (Trans., 1901, 70,
718). These results may, of course, be looked at in another way.
Digitized by
Google
VAPOUR PRESSURE OF AQUEOUS AMMONIA SOLUTION. 481
namely, as expressing the solubility of ammonia under varying pres-
sure, and it may be well to compare them with the results of other
investigators in this line of research.
Bosooe and Dittmar (Anncdm^ 1869, 112, 349) determined the
solubility at 0^ at pressures varying from 18 mm. to 1963 mm.
The pressures given are ** partial pressures" obtained by subtract-
ing the vapour pressure of water at the temperature of the experiment
from the total pressure. It may be pointed out that this method
of calculating partial pressure is quite inadmissible, for the ammonia
solution may be regarded as a mixture of two liquids, and the total
vapour pressure would be equal to the sum of the vapour pressures of
the liquids taken separately only if they were completely immiscible.
(The author is already engaged in determining the partial pressures
of the ammonia and the water in an aqueous ammonia solution.) On
adding the vapour pressure of water to the pressures recorded, the
few results available for comparison agree well with those given by
the author in his previous paper.
Sims {Annalen, 1861, 118, 345) found the solubility of ammonia in
water at 0^, 20^ 40% and 100°. The results are somewhat scanty, but
free use was made of graphic interpolation. The pressures given are
'* partial," as in the previously named paper ; when corrected to total
pressure, the numbers are much lower than those obtained by me.
Watts {Anrudm Suppl., 1865, 3, 227) made experiments at 0° and
20° with a mixture of ammonia and air ; when corrected to total
pressures, as before, the results fall almost exactly on the curves repre-
senting the results of my experiments.
Raoult {Ann. Ckim. Fhya., 1874, [v], 1, 262) made some single deter-
minations at temperatures from 0° to 28° at a '' partial pressure "
of 760 mm. They cannot well be compared with those under con-
sideration.
Konowaloff («/. Ruse, Phya. Chem. Sac., 1894, 26, 48), as the result
of experiments of which the numbers are not given, states that aqueous
solutions of ammonia do not follow Dalton's law at the ordinary temper-
ature, but with rise of temperature the disagreement becomes less,
until at 100° the numbers are in accord with this law. This result
was obtained approximately also by Sims. Konowaloff (ibid,, 1899, 31,
910) has also found the partial pressures of some solutions at 60° by
the dynamical method, but the solutions used were so weak that the
results are not comparable with mine.
The outcome of this comparison, so far as it can be carried out, is a
confirmation of the accuracy of the number^ given in the previous
paper, except in the case of the results obtained by Sims, which are
probably erroneous.
Much work has also been done on the vapour pressure of ammonia
Digitized by
Google
482 PERliAK : THE INFLUENCE OF SALTS ON THE
and gait solutions. Raoult (loo, cU.) investigated the solubility of
ammonia in solutions of potassium hydroxide, sodium hydroxide,
ammonium chloride, ammonium nitrate, sodium nitrate, and calciom
nitrate respectively, and his results, briefly summarised, are :
(1) Potassium and sodium hydroxides greatly lessen the solubility;
equivalent quantities have the same effect, and the effect of each is
proportional to the amount present in a given volume.
(2) Ammonium chloride also decreases the solubility proportionally
to the amount present.
(3) Ammonium and sodium nitrates have practically no effect on
the solubility.
(4) Calcium nitrate increases the solubility in proportion to the
amount present, but the effect of alteration of pressure on the solution
is nearly the same as on an ammonia solution.
(5) Experiments on the heat of solution of ammonia in solutions of
these substances appeared to show that they were approximately the
same as the heat of solution in water.
Konowaloff {J. Ruaa. Phys. Chem. Soo., 1899, 81, 910) found by the
dynamical method the effect of the presence of a number of salts on
the partial pressure of the ammonia. The decrease in the pressure of
the ammonia can be expressed by the formula
H « Hj (n - km) where
Hj » pressure of the ammonia in pure aqueous solution of less than
1^ normal strength.
H = pressure of the ammonia in solution containing the salt,
n s number of gram-molecules of ammonia per litre.
9» « „ „ salt per litre.
For silver nitrate k'^2, and for cadmium nitrate, zinc nitrate, nickel
chloride, copper nitrate, copper chloride, copper sulphate, and copper
acetate, the value of k approaches H. This is taken to indicate the
presence of complexes like AgN03,2NHg ; CuS0^4NH3 in solution.
Gaus (Z&U. anorg. Chem., 1900, 26, 236), by a very interesting
modification of the dynamical method, has determined the influence of a
number of salts on the partial pressure of the ammonia, the results
agreeing with those already mentioned. Ammonium nitrate and
barium chloride produce hardly any effect. Copper sulphate causes a
large depression which is proportional to the amount of salt present.
Quite recently this work has been largely extended by Abegg and
Riesenfeld {Z&it. phyMal. Chem.y 1902, 40, 84).
A consideration of the work of these investigators as well as that of
Dawson and McCrae (Trans., 1900, 77, 1239), seems to indicate clearly
the existence of metal ammonia complexes in solution, notwithstanding
the opinion of {Uioult concerning caldum nitrate, but the effect of salts
Digitized by
Google
VAPOUB PRBS8UBE OF AQUEOUS AMMONIA SOLUTION. 488
which would not he expected to have any chemical action has not yet
been explained.
The experiments about to be described were directed towards the
elucidation of (1) the effect of substances having no direct chemical
action on the ammonia, (2) the effect of change of temperature on the
copper sulphate ammonia solution, (3) the existence, or otherwise, of
hydrates in solution.
The method was very nearly the same as that described in a previous
paper {loc. oit., p. 725) except that the vapour jacket was replaced by
a water-bath with a toluene gas regulator ; the temperature did not
usually vary more than O'OP during an experiment. The height of
the mercury column, as well as that of the ammonia solution above it,
was read by means of a mirror scale placed in front of the bath at an
angle so that the reflection of the mercury could be easily read. With
this arrangement, parallax can be easily avoided by making the scale
line nearest the reading coincide with its reflection.
Influence of Substancei having no direct Chemical Action on Ammonia,
Substances having no direct action on ammonia are hard to flnd ;
those employed as likely to have no action were urea, mannitol, potass-
ium sulphate, and ammonium chloride. A preliminary experiment
was made to discover whether urea undergoes any decomposition when
heated with ammonia solution ; 20 grams of urea were heated in a
sealed tube with an aqueous ammonia solution of about four times
normal strength at 40^ for 6 hours. Be/ore heating, 10 c.c. of the
solution required (1) 41*98 cc, (2) 4207 cc, and afier heating (1)
41*90 cc, (2) 41*90 cc. of normal sulphuric acid solution for neutralis-
ation.
To determine the vapour pressures, a solution was made containing
a weighed quantity of urea (or other substance) and the ammonia then
estimated by titration ; after the vapour pressures had been measured,
the ammonia was again estimated, and in every case the amount
agreed well with the previous result. After the solution had been
transferred to the vessel for measuring the pressures, the air was
driven out of it by repeatedly lowering the pressure (by lowering the
open tube containing mercury) and then driving out the liberated air
by raising the mercury quickly and opening the stopcock at the top
for a moment.
The pressures for a corresponding ammonia solution have been found
by reading off the values for the particular strength on the curves
given in the previous paper. These give values of pressure for 0^,
20^ 34"4° 46*4°, and 61*3°. From these, a curve was constructed
w)iich may be called an ifoeihen or line of equal strength, and from it
Digitized by
Google
484
PERHAN: THE INFLUENCE OF SALTS ON THE
the pressures for the temperatures employed were read oCE. The
strength of the corresponding ammonia solution has been calculated in
two ways (I) as a solution containing the same proportion of ammonia
to water, (II) as a solution containing the same mass of ammonia in
unit volume of the solution. Both are given where possible in the
following table&
UreaSdtUumL—l^H^ 16-36; CO(NHj),, 10-43; 11,0, 7321 percent.
Density, 0*9691 at 22^. Strength of corresponding NH, solution, (I)
18-26, (n) 16-73 per cent. :
Aqueous NH, solution.
Temperature.
Pressure in mm.
Pressure I in mm.
Pressure II in mm.
24 -SS*
256-4
258
221
29-89
312-0 ■
318
276
85-20
408 0
405
354
40-38
496-8
500
438
45-59
611-4
611
582
54-43
856-1
854
738
59-07
1014-4
1008
880
C/rea Solution //.— NHj. 1722; C0(NH,)3, 5-29 ; HjO, 77-49 per
cent. Density, 0*9425 at 21^. Strength of corresponding NH, solu-
tion (I) 181 8, (II) 17-38 per cent. :
Aqueous NH, solution.
Temperature.
Pressure in mm.
Pressure I in mm.
Pressure II in mm.
25 -05*
260-4
263
240
29-58
816-6
322
294
34-96
400-3
402
373
39*68
485-1
485
452
45-90
621-8
618
669
50-07
726-2
722
659
64-45
856-9
854
775
68 00
978-7
970
886
When the vapour pressure of the urea-ammonia solution is compared
with that of an aqueous ammonia solution having the same ratio of
ammonia to water, the agreement is remarkable, but when it is
compared with that of a solution containing the same amount of
ammonia in a given volume, the apparent effect of the urea is to
decrease the solubility of the. ammonia. This may be due simply to
the increase in volume of the solution on addition of the urea.
Digitized by
Google
VAPOUR PRESSURE OF AQUEOUS AMMONIA SOLUTION. 485
Mannitcl SohUt(m.^lilK^, 12-27; mannitol, 4-56; HgO, 83-17 per
cent. Density, 0*9636 at 16°. Corresponding NHg solution, (I) 12 86,
(II) 12*43 per cent. :
NHg solution.
Temperature.
Pressure in mm.
Pressure I in mm.
Pressure II in mm.
22-92'
143-8
160
141
29-77
197 '7
296
196
86-78
271-2
276
262
42-74
848-8
366
332
60-61
479 1
484
463
* 67-79
629-6
636
696
When the vapour pressure of the mannitol-ammonia solution is
compared with that of an aqueous ammonia solution, a small but
regular decrease is shown, due, possibly, to the formation of a com-
pound of mannitol and ammonia. Mannitol is known to form stable
compounds with the alkaline earths.
Fotasnum Sulphate ^o/«<ww.— NH^, 749 ; KjSO^, 3-05 ; H^O, 89-46
. per cent. Density, 0*9826 at 35°. NH^, 7-73 per cent, of NH, and
Temperature.
40-42'
46-70
61-14
68-51
Pressure in mm.
208-6
266-2
320 0
427-6
Pressure of 7-78 per cent.
NH| solution in mm.
192
260
302
407
It was difficult to make a suitable solution owing to the slight
solubility of the potassium sulphate, and with the solution employed
the measurements could not be begun below 40°. The large increase
in pressure may be due to a displacement of ammonia molecules by salt
molecides, or possibly to the formation of hydrates producing a concen-
tration of the solution. Potassium and soditlm chlorides were found by
Gaus to increase the pressure largely, and the hydroxides still more.
The increase in pressure must be ascribed to the change in the nature of
the solvent, and the relation between the two changes reserved for
further investigation.
Ammonium Chloride Sdulion /.— NH3, 16*85 ; NH^OI, 5-27 ; H^O,
77*88 per cent. Density, 0*9472 at 14°. Corresponding NH3 solution,
(I) 17*79, (II) 17*07 per cent. :
Digitized by
Google
486
pbbman: the influsnck of silts on thb
Temperatare.
PresBore in mm.
KH| Bolaiion.
1
Pressure I in mm.
Prennre II in mm.
19-or
(182-1)
1
188 169
2604
261-1
263
246
82-93
848-0
866
838
89-24
449-8
467
430
49-26
667-2
676
622
67-88
918-2
938
860
Ammonium Chloride Solution //.— NHg, 12-90 ; NH^Ol, 1026 ; H,0,
76-84 per cent. Density, 0*9724 at 16"^. Corresponding NH, solution,
(I) 14*37, (II) 13-24 per cent. :
NH, solution.
Temperature.
Pressure in mm.
Pressure I in mm.
Pressure II in mm.
28 •12''
168-6
171
164
80-89
280-6
248
228
88-26
818-6
884
801
45-87
428-8
447
402
61-18
648-2
660
602
68-14
707-6
723
661
If the results are compared with those obtained with ammonia solu-
tion I, the ammonium chloride is seen to produce a decrease in pressure^
and the decrease is approximately twice as much in solution II as in
I (except at the low temperatures, where there seems to be some
experimental error in solution I). This suggests the formation of
some compound of the ammonia with the ammonium chloride. Since
the completion of these experiments, the author has found that com-
pounds such as NH401,3NH3 and NH^C1,6NH, have been isolated
(Compt. rend,, 1879, 88, 578), and it is probably due to the formation
of compounds like these that the vapour pressure of the solution is
diminished.
Gaus found barium chloride and ammonium nitrate to have only a
very slight effect on the ammonia pressure, and Raoult found am-
monium nitrate and potassium nitrate to have none ; Baoult's experi-
ments, however, were carried out with such strong solutions that they
are hardly comparable with those now described.
From the results hitherto obtained, it seems impossible to predict
what effect will be produced by introducing into the solution a sub-
Digitized by
Google
VAPOUR PRESSURE OF AQUEOUS AMMONIA SOLUTION. 487
stance which would be expected to have no direct chemical action on
the ammonia ; the influence, hpwever, would appear to be small except
in the case of salts of the alkalis. If, however, a substance, as, for
example, silver chloride or zinc sulphate, is known to form a deflnite
compound with ammonia, the pressure of the ammonia is invariably
diminished, but the effect of the ammonia going into combination may
be complicated by other efEects not yet understood.
Effed qf Change qf TrnipercUure on a Solution qf Cupri-ammonium
StilphaU.
It was found necessary to use weak ammonia solutions, otherwise a
sufficient amount of copper sulphate could not be kept in solution at
the lower temperatures. An attempt was made to measure the vapour
pressure at 0^, but without success ; the height of the mercury column
could not be read, for the solution stuck to the glass and the mercury
rose inside it. This phenomenon, which is presumably a surface-
tension effect, did not occur at about 20^«
Copper Sulphate Ammonia Solution /.— NH3, 14-65; CuSO^, 2-68;
Hfi, 82-67 per cent. Density, 0-9652 at 15''. Corresponding NH,
solution (I) 16-05, (II) 15*01 per cent. :
Temperature.
Pressure in mm.
Pressure NH,
solution in mm.
Pressure of NH, solution
in mm.
18-88"
26-87 •
86*08
43*64
47-49
60-81
66-54
138-5
197*6
806*9
421-6
492-6
559-9
692-6
145
212
824
446
518
588
726
188) subtracting SNH,
197 j-mols. for each mol.
808 J CUSO4,
419)
555 * '♦ SNHj „
/686.
\700 „ INH3 „
The copper sulphate produces a large decrease in the pressure, in all
probability due to the formation of a cupri-ammonium compound.
If the copper sulphate in the solution combines with ammonia, forming
the compound CuS04,4NH2, the amount of free ammonia remaining
in the solution can be calculated by subtracting 4 mols. NH3 for every
OuSO^ mol. ; an aqueous ammonia solution calculated of this strength
has lower pressures than any found for the copper^ammonium sul-
phate solution at the same temperatures. Subtracting 3 mols. NH3,
the pressures coincide at 19° and 26°; subtracting 2 mols. NH3, the
pressures are nearly coincident at 44° and 47°. There is evidently
progressive dissociation, until at 56*5° only about 1^ mols. NH3 for
Digitized by
Google
488
PERMAN : THE INFLUENCE OF SALTS ON THE
every CuSO^ mol. remain in combination. Unfortunately, a great
assumption has to be made in calculating out these numbers, namely,
that the copper sulphate has no influence on the pressure other than
that caused by the ammonia going into combination with it The un-
certainty in this matter makes it useless to develop the theory of the
decomposition from the numbers obtained. The compound at first in
solution is probably SOj^q.jth^.nH*^^"' which dissociates into
S02<g;JJgs>Cu, and very possibly SO,<g'^^»'gg»>Cu, is also
formed.
On evaporation of the ammonia, or on dilution of the solution, de-
composition takes place, probably thus :
Ou(NH3)jS04 + 2H,0 - Cu(0H)2 + (NH^,SO^.
These compounds will also be more or less ionised.
Copper Sulphate Ammonia Solution II.—l^U^, 654 ; OuSO^ 3 94;
HjO, 89-52 per cent. Density, 1-010 at 15° Corresponding NH,
solution, (I) 6-81, (II) 680 per cent. :
Temperature.
Pressare in mm.
Pressare NH,
solution in mm.
Pressare NH, solution
in mm.
80-57
36-97
.41-61
46-85
5261
57-68
91-8
126-2
155-1
194-7
250-1
808-8
112
148
180
227
294
867
93.
124
154
197
251
8I3J
-Subtracting 8NHj
In this case, the amount of ammonia held in combination by the
copper is apparently the same as before, but no dissociation takes
place, no doubt owing to the much smaller proportion of water with
regard to the copper sulphate. After each experiment was completed,
the copper solution was tested for mercury, but only a trace was found
in each case.
The Exietence qf Hydraies in Solution.
The experiments on potassium sulphate before described have an
interesting bearing on this question. The author has shown in a
recent paper (Trans., 1901, 79, 725) that anhydrous sodium sulphate
has a great effect in raising the ammonia pressure, but it is now found
that potassium sulphate has a similar effect. The experimental data
available for the comparison are given in the table on page 489.
Fromthepenumbers wefind that the ratio ^^n^ber mole. Na^^^^.^^^
number mols. K^SO^
Digitized by
Google
VAPOUB PBffiSUBK OF AQUBOUS AXMOHIA 80LUTI0K. 489
Composition of solatioiL Tempentuve.
Pmnireof
oomspondiiig
NHjSolatioii
in mm.
Percent.
• of
preesore.
[HjO 89-4«J
40-42*
4670
208-5
265*2
192
250
5-99
6-08
|N^0. 4-25
[H,0
I 4-25]
9-90}-
85-85J
40-42
4670
276
360
246
820
12-2
12-6
whilst the ratio per cent, increaae of press, caoaed by theNa^O,__g.^g
per cent, increase of press, caused by the K^SO^
(2-04 at 40-42'' and 2*06 at 46-70°).
These ratios are sufficiently near to each other to show that
molecular proportions of potassium sulphate, which crystallises
without water, and of anhydrous sodium sulphate have approximately
the same effect. There is no difference, such as would be caused by
one forming a hydrate in solution and the other not. Supposing the
sodium sulphate to take up 10 mols. of water, the concentration of
the solution thereby produced would cause an increase in pressure of
about 10 per cent.
The increase of pressure caused by the potassium sulphate is 6 per
cent., and the number of molecules of sodium sulphate is 1*7 times as
great as the number of potassium sulphate molecules ; the effect of the
sodium sulphate should therefore be (supposing molecular proportions
to have the same effect) an increase of pressure of 6 x 1*7 » 10*2
per cent. : adding this to the 10 per cent, for the increase in con-
centration of the solution, we obtain a total increase of 20*2
per cent. This, divided by the increase for the potassium sulphate
(6 per cent.), gives 3-37 for the ratio !°^"^^ ^J ^f ^^ instead of
increase by K^SO^
2*05, the number found, a difference far beyond the limits of ex-
perimental error.
There is little reason for supposing potassium sulphate to form a
hydrate in solution, so that these experiments seem to indicate the
non-existence of a hydrate of either sodium sulphate or potassium
sulphate in solution.
The experimental part of the work here described was carried out
at the Physikalisch-chemisches Institut, Leipzig, and the author
wishes to acknowledge the great facilities there afforded him.
UNIVXaSITT Ck)LLEGB,
CABDurr.
VOL. L2XXI. L L
Digitized by VjOOQIC
490 OBTON: THE NITRATION OF S-TBIHALOGEN ANILINES,
#
L. — The Nitration of Q-Tmhalogen Anilines.
By K. J. p. Orton.
Ik the hope of possibly throwing farther light on the process of sub-
stitution in anilines, I have studied the carefully regulated action of
nitric acid on those anilines in which chlorine or bromine occupies the
positions 2, 4, and 6 in the benzene nucleus relatively to the amino-group ;
namely, those positions into which a substituting group most readily
finds its way. That the nitro-group is capable of displacing bromine,
at least, from these positions in phenols has been shown by Armstrong
and Harrow (Trans., 1876, 30, 448), who obtained 2 : 6-dibromo-4-nitro-
phenol and 2-bromo-4 : 6-dinitrophenol from «-tribromdphenol. Thiele
and Eichwede (^nna^en, 1900, 311, 363), by the action of amyl nitrite
on this phenol, replaced, not the para-, but an ortho-bromine atom,
2 : 4-dibromo-6-nitrophenol being thus formed. Further, from
Mribromoaniline by the action of nitric acid, Losanitsch {Ber,, 1882, 15,
474) obtained 2 : 6-dibromo-4-nitroaniline.
In their behaviour * with nitric acid (diluted with acetic acid), the
anilines investigated divide themselves sharply into two classes ; (1)
anilines with a bromine atom in the para-position relatively to tiie
amino-group, «-tribromoaniline, 2-chloro-4 : 6-dibromoaniline, and
2 : 6-dichloro-4-bromoaniline, (2) anilines with a chlorine atom in the
para-position relatively to. the amino-group, 4-chloro-2 : 6-dibromo-
aniline, 2 : 4-dichloro-6-bromoaniline, and Mrichloroaniline. The
crystalline aniline nitrate initially formed dissolves on heating, pro-
ducing solutions of characteristically different colours in the two claases ;
when a bromine atom is in the para-position, the solution is finally
orange-yeUow ; when a chlorine atom is in the para-position, the
solution is finally crimson (see experimental part). From anilines of
the first class, bromine is evolved, and there is obtained a product which
possesses a nitro-group in the para-position instead of an atom of
bromine. The amount of the dihalogen-nitroaniline represents about
75 per cent, of the original Mrihalogen aniline. From anilines of the
second class, neither chlorine nor bromine is evolved, nor is an aniline
obtained in which a nitro-group has replaced the j>chlorine atom.
Further, in no case was the replacement of an o-bromine atom
observed. In this respect, the behaviour of these anilines with
nitric acid is analogous to their behaviour with acetylchloroamino-
* In these experiments, an inyestigation of the firuU prodacts of tbe action of
nitric acid on these anilines was not intended. This has been done in the case of
9-tribromoaniline by Losanitsch {loc, cit), dibromodinitropiethane, tetrabromo*
benzene, bromanil, oxalic and picric acids being isolated. Doubtless, these similarly
oonstitated substances undergo degradation in much the same manner.
Digitized by
Google
obton: the nitration of s-triiialogen anilines. 491
benzenes, when only the p-bromine atom is replaoed by chlorine (Trans,
1901, 70, 822). From 2:3:4: 6-tetrabromoaniline in which one meta-
position is also occupied by bromine, only 2:3: 6-tribromo-4-nitro-
aniline is formed, although replacement of an o-bromine atom might
be expected from analogy with the action of chlorine, bromine, and
nitric add on meta-substituted anilines, and anilides, which always
yield a considerable proportion of the ortho-derivative.
Lastly, in no case was hydrogen in the meta-position replaced by the
nitro-group under the conditions employed.
During the period of heating of the acetic acid solution of the aniline
and nitric acid (aniline nitrate), there is present a small amount of the
nitroamine of the aniline, thus from s-trichloroaniline is obtained
l-niProamino-^ : 4 : ^-Prichlcrobenzens^ C^H^Clg'NH'NOj. These nitro-
amines^ would appear to be formed in this reaction by the elimination
of water from the aniline nitrate, just as aniline acetates on heating
lose water and become acetylamino-derivatives.* At no period of the
experiment did the nitroamine amount to more than 5 — 10 per cent, of
the aniline used. As under the conditions, namely, heating in acetic
acid solution in the presence of a mineral acid, the nitroamines them-
selves undergo change, no great quantity can at cmy one time be present .
When a nitroamine in which a bromine atom is in the para-position
relatively to the amino-group is dissolved in acetic acid to which one
or two drops of sulphuric acid have been added, the nitro-group is
transferred to the nucleus and displaces the />-bromine atom ; t whereas a
nitroamine with a chlorine atom in the para-position, under identical
treatment, yields no nitroaniline but gives a deep crimson solution,
similar in appearance to that obtained directly from the aniline and
nitric acid. From the crimson solutions from both sources, red nUh
etanoea can be isolated in very small amount ; these have not yet been
obtained in sufficient quantity for a thorough analysis or investigation.
Experiments are now in progress having as their object the pre-
paration of these compounds.
Up to the point of the formation of the nitroamine from the aniline
nitrate, there is a complete parallelism in the action of nitric acid on
these two classes of a-trihalogen anilines (namely, the one class with a
p-bromine atom and the other with a p^hlorine atom); as the
nitroamines in each class are under an identical influence, the tendency
* So far as I am aware f-trisubstituted nitroaminobenzenea have not hitherto
been prepared ; nor have nitroaminobenzenes been obtained by the direct action of
nitric acid on anilines. These substances were prepared by Bamberger, Pinnow,
and others, by oxidation of alkaline solutions of benzenediasot&tes, by the action of
nitrogen pentozide on anilines in chloroform solation, and by addiog the dry aniline
nitrate to acetic anhydride.
t By a similar means, Bamberger brought about the transformation of nitro-
aminobenzenes into 0- and j^-nitroanHines.
L L 2
Digitized by
Google
492 OBTON: THE NITBATION OF S-TRIHALOGEN ANILINES.
in each case must be for the nitro-group to replace the p-halogen
atom. It was suggested {loc. eit.) that in the transformation of the
acetylchloro- and acetylbromo-aminobenzenes (L), into o- and p-
chloro- and bromo-acetanilides (Ill)f an iminoquinone (II) formed a
transient intermediate stage.
-N-X -N -N-H
_^V _A_ A
I J i]_ _> II J ii_ -^ III I II
H A X
From this point of view (which was , originally foreshadowed by
Lapworth, Trans., 1898, 73, 450), it would be excepted that both with
a />-bromo-, and a />-chloro-aniline the nitroamine would pass into an
iminoquinone in which both the nitro-group and the bromine atom are
attached to the same carbon atom ; in the one case, the bromine is
eliminated, and ap-nitroaniline produced ; in the other case, the chlorine
is not eliminated, but some derivative of the iminoquinone type is
formed. It is possible that the red substance above mentioned is such
a derivative.
It is open to doubt whether, in the ordinary nitration of anilides,
the stages observed in the action of nitric acid on these s-trihalogen
anilines actually occur; the nitroamino-derivatives of the anilides
have never been obtained; it is possible that the nitrating agent
reacts directly with an acyliminoquinone (compare Lapworth, Trans.,
1901, 79, 1267 ; Thiele, Annalen, 1899, 306, 87).
EXPEBIMBVTAL.
Bectctum of Nitric Acid with B-Tribromocmiline and the B-Dihalogenrp-
bromoanilinoB,
S'Tribromocmiline, — Losanitsch(^oc.ci^.) heated 6-tribromoaniline with
nitric acid diluted with acetic acid and obtained 2 : 6-dibromo-4-nitro-
aniline, but he does not state the proportion of nitric acid used, or
give the details of the experiment.
Five grams of the aniline were covered with 50 c.c. of glacial acetic
acid (m. p. 15 '2^) a quantity insufficient to completely dissolve this
base at the ordinary temperature. To the solution containing some
solid in suspension were added 7 — 8 c.c. of nitric acid (sp. gr. 1 *5),
which was colourless and free from nitrous acid* ; a crystalline precipi-
tate of the aniline nitrate immediately separated. The mixture was
* In the presence of nitrons acid, diazotisation took place to a laige extent,
confosing the direct interaction of the aniline and nitric acid.
Digitized by
Google
ORTON: THE NITRATION OF S-TBIHALOOBN ANILINES. 493
now heated on the water-bath, when the nitrate dissolved, forming a
solution which rapidly became deep orange, and then lighter in tint
and more yellow than orange in colour. After about 20 — 30 minutes,
the evolution of bromine had become very obvious, and on cooling 2 : 6-
dibromo-4-nitroaniline (3*5 grams) separated in a nearly pure state.
It melted at 206°; its acetyl derivative melted at 235^ and the
acetylchloroamino-derivative at 109 — 110°; the last named contained
CI as INC1»9*33 per cent, (instead of 9*51 per cent.). A careful search
was made in the mother liquors for 2 : 4-tribromo-6nitroaniline,
and f or 2 : 4 : 6-tribromo-3-nitroaniline. The mother liquors were pre-
cipitated by water, and the solid thus obtained extracted with
aqueous sodium carbonate (to remove any nitroamino-derivative), and
then distilled in steam. The fact that the distillate was colourless
indicated the absence of any o-nitroaniline (this vol., p. 496). Small
amounts of «-tribromoaniline were alone found.
l'NUroamino-2 'A : Q'ttHbramobenzenef C^H^Brg-NH-NO,. — In order
to isolate the nitroamine, the acetic add solution at any time during the
period of 5 to 15 minutes from the beginning of heating was poured
on to ice ; the yellow solid which separated was collected and ^washed
free from acid. It was then extracted with cold dilute aqueous sodium
carbonate. The remaining yellow solid, consisting mainly of un-
changed «-tribromoaniline and 2 : 6-dibromo-4-nitroaniline, was fdtered
from the alkaline liquor, from which, on addition of a mineral acid,
the nitroamine separated as a white precipitate. From 10 grams of
aniline, 0*4 gram of nitroamine was obtained. It is readily soluble in
all organic solvents except petroleum ; from dilute acetic acid or dilute
alcohol, it crystallises in flesh-coloured needles. In cold water, the
nitroamine is insoluble ; 1500 c.c. of hot water are required to dissolve
1 gram ; from this solution, it crystallises in slender, long, often curved,
flesh-coloured needles, melting and decomposing at 143 — 144° with
evolution of oxides of nitrogen :
0-1634 gave 0 246 AgBr. Br - 64-04.
C^HsO^NgBrs requires Br » 63*98 per cent.
This substance is .acid to litmus ; on addition of aqueous sodium
hydroxide to an aqueous solution of its sodium salt, the latter separ-
ates in pearly white plates :
0-3444 gave 0053 NajSO^, Na =p 562.
C^HjOjN^BrgNa requires Na-5*79 per cent.
When the nitroamine is dissolved in acetic acid to which a drop of
sulphuric acid has been added, the colour of the solution becomes
orange, and after some time 2 : 6-dibromo-4-nitroaniline separates.
2-CA/bro-4 : ^-dibi'amoanUine was treated in a perfectly similar
Digitized by
Google
494 ORTOK : THB NITRATION OF S-TRIHALOGEN ANILINSS,
manner with nitric aoid. 2-Chloro-6-bromO'4-Qitroani]ine was the main
product ; it melted at 177°, yiel4ed an acetyl derivative melting at 22P
and was in every respect identical with the synthetical product (ibis
vol., p. 496). The nitroamine obtained in this case melted and decom-
posed at 137°; as only a small quantity of the aniline was used, there
was not sufSicient of the nitroamine for analysis. From 2 : ^<liehlar(h
^-bramoaniUtnSf 2 : 6-dichloro-4-nitroaniline (m. p. 188°) wsa obtained ;
its acetyl derivative melted at 210 — 211°. The nitroamine was only
obtained in very small quantity, and melted and decomposed at
136--137°
Reaction of NUrio Aeid wUh B-TriMoroanUinB and with B-Dihalogonr^
ehloroanilinet.
B-TricUoroaniline. — Five grams of the aniline were dissolved in
50 c.c. of acetic acid and 8 c.c. of pure nitric acid added, whereupon the
nitrate of the base immediately crystallised out. On heating the
mixture on the water-bath, the solution became purple as the nitrate
dissolved, and then rapidly changed to magenta, which more slowly
became crimson. No chlorine was evolved. After 30 minutes' heating,
a very small quantity (0*05 gram) of hezachloroazobenzene separated ;
on recrystallisation from chloroform and alcohol, it was obtained in
long, lustrous, red needles, melting at 188°, identical in melting point,
solubility, <fec., with the specimen previously prepared by the action
of acetylcbloroamino-2 : 4Kiichlorobenzene on «-trichloroaniline (Trans.,
1901, 70, 467).
I-I^itroamino'2 : 4 : ^-triehlorobenzme, C^HjOlj-NH'NO,, was isolated
in a manner completely similar to that used for the tribromo-
derivative, and resembles the latter very closely in appearance and
properties. It crystalUses from much hot water in long, flesh-
coloured needles, melting and decomposing at 135°:
0-1402 gave 14*2 c.c of moist nitrogen at 17° and 775 mm. N^ 11-93.
01486 „ 0-266 AgCl. CI «: 44-26.
C^HgOgNjClg requires CI- 4406 ; N - 1163 per cent.
Its sodium salt crystallises in glistening plates, and is very soluble in
water or alcohol.
A solution in glacial acetic acid to which one drop of sulphuric acid
has been added soon becomes crimson in colour. At the ordinary
temperature, the nitroamine changes only slowly, but on heating on the
water-bath it rapidly decomposes, the colour of the solution quickly
deepening.
To isolate any substance or substances which may have been formed,
the crimson solution obtained either from the nitroamine, or directly by
Digitized by
Google
S-CtiLOttOBBOHONtTROANILINES AND THBIB DKBITATIVlSS. 495
the action of nitric acid on an acetic acid solution of the aniline, was
poured on to ice ; the red solid was collected, washed free from acid^
and extracted with aqueous sodium carbonate to remove any nitro-
amine. The solid was now fractionally crystallised from petroleum
(b. p. 50—80^) ; finally, a small amount of a substance was obtained,
which crystallised in elongated, brilliant red plates melting and decom-
posing at 143^ and very soluble in the usual solvents. It dissolved in
concentrated sulphuric acid with a reddish-brown coloration; on
addition of water and on warming, the solution became colourless. A
solution in acetone was immediately reduced by zinc dust and acetic
acid with the production of a colourless substance, crystallising in
needles from petroleum melting at 188^ Sufficient of these substances
(0*2 gram from 20 grams of aniline) could not be obtained by the
above method for a complete investigation.
2 : A'DiMonh^-bromoanUtne and 4-cA2or(^-2 : B-dibromoaniline behave
in a completely analogous manner ; from each, nitroamines were ob-
tained, decomposing at nearly the same temperature, 137 — 138^.
Azobenzenes were formed in small amount. From the red solutions,
obtained as in the case of t-trichloroaniline, no attempt was made to
isolate the corresponding red substances.
St. Babtholomew's Hospital and Colucgx.
London, E.O.
LI. — Some s-Chlorohromonitroanilines and their
Derivatives.
By K. J. p. Orton.
Fob the purpose of recognising the products possibly obtainable by
the action of nitric acid on the «>trihalogenanilines (see preceding
communication), the anilines hereafter described were prepared and
their derivatives investigated.
As in the case of all di-o-substituted anilines, monoacetyl derivatives
are only with difficulty prepared from these bases — a difficulty which
is intensified if a nitro-group occupies one ortho-position. Only after
many hours' heating with excess of acetyl chloride and sodium acetate
is the acetyl derivative obtained. When quite free from the respec-
tive bases, the monoacetyl derivatives of the anilines possessing a
nitro-group either in the ortho- or the para-position (for example,
2 : 4-dibromo-6-nitro- and 2 : 6-dibromo-4-nitro-anilines) are quite colour-
less ; but when obtained from the aniline as just mentioned, even after
Digitized by
Google
496 OBTOK : SOME S-OHLOROBROMONITROANILINES
repeated recrystallisation, they are generally of a yellow tint owing
to the presence of traces of the latter.
All di-o-halogen acetanilides (as «-tribromoacetanilide) dissolve fairly
readily in 10 per cent, aqueoas sodium hydroxide, and are precipitated
unchanged by acids; these alkaline solutions may be heated for a
considerable time without effecting any appreciable hydrolysis.
The acetyl derivatives of 8-trisubstituted anilines, with a nitro-group
in an ortho- or a parsrposition, dissolve readily even in iiT/lO cold
aqueous sodium hydroxide ; the anilides with an o-nitro-group form an
orange-coloured solution, which becomes markedly redder on warming ;
those with a />-nitro-group form a canary-yellow solution which does not
much deepen in tint when warmed. From these alkaline solutions,
acids precipitate the anilides as perfectly white solids. By taking
advantage of the solubility of the anilides, they may be obtained
completely free from the anilines after acetylation. Although these
«-trisubstituted nitroanilides are more difficultly hydrolysed than even
di-o-halogen anilides (^-tribromoacetanilide, dsc.) by boiling with sul-
phuric or hydrochloric acid and alcohol, they are very easily converted
into anilines when their solutions in excess of 10 per cent, aqueous
sodium hydroxide are heated for a short time (compare Kleemann,
Ber., 1888, 19, 336).
The diacetyl derivatives of #-trisubstituted anilines are formed when
the aniline is boiled or heated under pressure with acetic anhydride
for some hours. In this operation, no formation of monoaoetylated
compound takes place, although some of the aniline frequently remains
unchanged, when the heating has not been sufficiently prolonged. The
diacetanilides are very rapidly converted into the monoacetyl deriva-
tives by aqueous alkalis.
To recognise a small amount of an o-nitrodihalogen aniline in the
presence of a much larger quantity of the para-isomeride, it is best to
distil the mixture in steam. Although both anilines pass over, the
ortho-isomeride distils more readily and gives a canary-yellow dis-
tillate, whilst the para-derivative gives a colourless distillate. By
this means, not only can the presence of^the ortho-compound be recog-
nised easily in a mixture containing less than 2 per cent., but 0*02
gram can be readily detected.
EXPEBIMBMTAL.
2-CWoro-6-6romo-4-ni«roantf»iM, N0,'C^H,01Br-NHy was prepared
from 2-bromo-4-nitroaniline ; the latter was obtained by adding a
solution of bromine (1 mol.) in glacial acetic acid to a hot solution of
p-nitroaniline (1 mol.) in the same solvent. On cooling, a little
2 : 6-dibromo-4-nitroaniline separates. The acetic acid solution was
Digitized by
Google
AND THEIfi DSBIVATTVES. 497
precipitated by water, and the mixture of anilines thus obtained
dissolved in alcohol, to which was added a weight of sulphuric acid
equal to that of the /7-nitroaniline used. On now adding water, only
2-bromo-4-nitroaniline separates, whilst any p-nitroaniline remains
dissolved in the acid. The yield is about 75 per cent, of the calculated
amount.
2-Bromo-4-nitroaniline was chlorinated by passing dry chlorine
into its solution in dry benzene ; chlorination is completed when the
solid which separates settles easily to the bottom of the flask. This
solid, which is the hydrochloride of 2-chloro-6-bromo-4-nitroaniline,
was collected and well washed with petroleum. On digesting it with
water, the aniline is obtained. The latter is only slightly soluble in all
solvents, and crystallises from alcohol in long, prismatic, bright
yellow needles melting at 177° :
0-2636 gave 0-345 AgCl + AgBr and 0-2248 Ag, Cl-14-02; Br = 31-58.
CgH^OjNjOlBr requires 01 = 1409 ; Br = 31-78 per cent.
2'Chlor(h6-l>r(mo-Arniiroacetan%lide, NOj-CeHjClBr-NH-CO-CHg, was
prepared by adding acetyl chloride in some excess to a warm solution
of the aniline in glacial acetic acid. The mixture was then boiled for
half-an-hour. From alcohol, in which it is only slightly soluble, the
anilide separates in lustrous, white prisms melting at 221 — 222° :
0-1916 gave 0-2163AgCl + AgBrandO-1406Ag. 01 = 11-86; Br- 27-66.
OgHgOjNjOlBr requires 01= 12-08 ; Br = 27*24 per cent.
AoetylM<>roaminO'2-chl(}ro-6^<miO'i'nitrobens;ene,
NOj-OeHjOlBr-NOl-OO-OHg,
prepared from the anilide in the manner previously described, crystal-
lises in white, lustrous prisms, melting at 84 — 85° :
01727 liberated I = 105 c.c. iVVlO iodine. 01 as INOl - 10-77.
OgHgOsN^OlgBr requires 01 as :N01= 10-81 per cent.
2'CMaroQ-bromO'A-nitrodiacetantlide, NO,«OgH20lBr-N(00-OH8)2,
was prepared by boiling the aniline (1 gram) with acetic anhydride
(10 grams) for 10 hours ; some unchanged aniline still remained. The
diacetyl derivative was readily soluble in all solvents, and crystallised
from petroleum in long, four-sided, white prisms melting at 133 — 133-5°
It was not analysed.
2'ChlorO'i'bromO'6-niiroanUine, NO^'O^H^OlBr-NH^, was prepared
from the anilide ; the latter (2 grams) was dissolved in 20 c.c. of a
10 per cent, solution of sodium hydroxide; the solution was then
heated on the water-bath; after 10 minutes, the aniline began to
separate, and hydrolysis was complete in half-an-hour. The aniline
was crystallised from alcohol, in which it is far more soluble than
Digitized by
Google
4d8 ORTON: some S-CHLOROfiROMONlTROANILlKEB
the isomeride just described. It forms silky, yellow needles meltiDg
at 114°:
01860 gave 02432 AgCl + AgBr and 0-1584 Ag. CI = 1401; Br = 31-84.
CgH^OaNjClBr requires 01 = 14-09 ; Br ==31 '78 per cent.
2'Chlor(>-i-bramO'6-niiroaeetanUtde, NOj'CgHjClBr-NH'OO-CH,, was
prepared from 2-chloro-4-bromoacetanilide j the latter (10 grams) was
added in small portions to 50 c.c. of ice-cold nitric acid (sp. gr. 1*5).
The acid solution was thrown on to ice, and the solid which separated
recrystallised from alcohol. The anilide crystallises in white needles
or prisms melting at 194° :
01306 gave 0-1466 AgCl + AgBr and 0-0956 Ag. CI - 12-07; Br = 26-96.
CgHgOgNjClBr requires CI = 12-08 ; Br = 27*24 per cent.
AcetylcM(n'oaminO'2<hl(^(h4rbr(mio-6'nitrob$nzene,
NOj-C^HjClBr-NCl-COCHj,
was prepared in the usual way from the anilide. It crystallises in long*
pale yellow, lustrous prisms melting at 56 — 57° :
01706 liberated 1 = 10-6 c.c. iT/lO iodine. CI as :NC1 = 11-02.
CgHgOgNgCljBr requires CI as :NC1 = 10*81 per cent.
This substance is readily soluble in petroleum. It is noteworthy that
this chloroamino-derivative is yellow, although obtained from a colour-
less anilide ; in its low melting point it differs from the chloroamines
of other di-o-substituted anilides.
4rChlor(h2-bramO'6^UroainUine, NO^'CoH^ClBr^NH,, was prepared by
hydrolysing the anilide in the manner just described. It crystallises
from alcohol i^ yellow, silky needles melting at 114 — 115° :
0-1428 gave 0-1876 AgCl + AgBr and 0-1224 Ag. Cl = 14-14; Br«31-52.
C^HPgNjClBr requires Cl = 1409; Br = 31-78 per cent.
4-CWoro-2-ftromo-6-ni^aflK»tontfufo, NOg-C^HjClBr-NH-CO-CH,, was
prepared by nitrating 4-chloro-2-bromoacetanilide. It crystallises from
alcohol in colourless needles or flattened prisms melting at 207° :
0*1824 gave 0-2048 AgCl + AgBr and 01338Ag. Cl = 12-29; Br-26*65.
CgH^OsN^ClBr requires CI « 1208 ; Br :^ 27*24 per cent.
Aoeti/lMaroamin(h2 : ^-dihroma-ir^itrobenzene,
NOj-CeHjBrj'NCl-CO-CH,,
This hitherto undescribed chloroamine was prepared from the anilide.
It crystallises from petroleum (b. p. 50 — 80°) in small, four^sided prisms
with domed ends, melting at 110 — 111° :
0*488 liberated I ^ 25-5 c.c. N/IO iodine. CI bs INCl » 9*27.
OgHjOjNgOlBrj requires CI as :NC1-9*51 per cent.
Digitized by
Google
AKD mOK DKRIYATTVU. 499
2:4r'lMrvmth^^miirvdiae6Umai4U, N02-C<B[,6r,-N(00-CH^y was
prepared by boiling the aiiiline (1 gimm) with aoeiic aiihjdride (^^
for several hourB. After evaporafcioii oi the exeeas of acetic anhydride,
the solid zemdne waa reeryatalliaed &om petrc^enm. It forms aggregates
of ooloorless, riK»nbie prisms melting at 96 — ^97® :
0-1847 gave 01840 AgBr. Br « 4239.
CioHgO^NjBr, requires Br » 42*07 per cent
2:3:4: e^Teirabrcmoaeeiamlide, C^HBr^-NH-GO-CH,, was prepared
by acetylating the aniline with acetyl chloride in the manner above
described It crystallises from benzene in silky needles melting at
228 — 229^y and is fairly soluble in alcohol and acetic acid :
— -^ "«' f — —^ «w mmmm^M.j w««*w»w mam w*w«w« mmm.
01 762 gave 02784 AgBr. Br » 6722.
CgH^ONBr^ leqnires Br»67'S
r'35 per cent.
2:3:4: e'Teiraiframodiaeeianaid&, C0HBr4-N(GO-CH,)y prepared in
the nsiial way from the aniline and acetic anhydride, crystallises
from petroleum in transparent, four-sided prisms melting at 164^,
and is very soluble in all solvents except petroleum (b. p. 50 — 80^) :
0*11 gave 01674 AgBr. Br » 64-76.
C^oRfi^^TBr^ requires Br = 6488 per cent
2:3: ^Tribromo-l^itroaniline, NOj-C^HBrs-NHj.— This aniline
was prepared from 3-bromo-4-nitroaniline (m. p. 17C^), by the action
of bromine on an acetic add solution of the latter. It crystallises
from alcohol, in which it is only slightly soluble when cold, in pale
lemon-yellow needles melting at 155 — 155-5^ :
01775 gave 0-2672 AgBr. Br » 6404.
CgHsO^s^r, requires Br » 63*98.
This aniline distils slowly in steam, and as in the case of other
j>-nitroaniline8, the aqueous distillate is colourless, whilst the distil-
late of the isomeric o-nitroaniline is bright yellow.
2:3:4-7W6romo-6-nt<roam^tns, NOg-CeHBrj-NH,, prepared from
3-bromo-6-nitroaniline (m. p. 151°), crystallises in orange-yellow
needles melting at 165*5 — 166° :
0-1266 gave 0*1908 AgBr. Br « 64*12.
O^H^OsN^Brg requires Br* 63*98 per cent.
The acetyl derivative orystalliBee in flattened, white needles melting
at 221°.
The 3-bromo-4-nitro- and the 3-bromo^6-nitro«niline8 just men-
tioned were prepared by nitrating m-bromoanilina f7»-Bromoaniline
(1 part) was dissolved in concentrated sulphuric acid (10 parts), and
to the ice-cold solution the caloul ated quantity of nitric acid (sp. gr.
Digitized by
Google
600 ORTON: THE NITRATION OP
1'5) dissolved in sulphurio acid was added. The add solution was
poured on to ice, and the nitroanilines precipitated with ammonia. They
are separated by distilling in steam from an equal weight of solpharic
acid ; d-bromo-6-nitroaniline distils over whilst 3-bromo-4-nitroaniline
remains in the distilling flask. After one recrystallisation from
alcohol, each possesses the correct melting point (compare Glaus and
Scheulen, J, pr. Chem., 1891, [ii], 43, 201 ; Glaus and Wallbaum,
iW<^., 1897, [ii],66, 54).
St. Bartholomew's Hospital axd Oollbgb.
London, E.C.
LII. — The Nitration of a-Trihalogen Acetanilides.
By K. J. R Obton.
In a previous paper (this vol., p. 490), it was shown that in the
action of nitric acid on the «-trihalogen anilines («-tribromoaniline, &c,\
a bromine atom in the para-position relatively to the amino-group was
replaced by a nitro-group, whilst this is not the case with a /^chlorine
atom or an o-bromioe atom (at least, under the same conditions).
Further, it appeared very probable that at least one intermediate pro-
duct, a nitroamine, existed between the nitrate of the aniline first
formed and the final product (or products). It seemed desirable to
ascertain what action nitric acid had on the acetyl derivatives of the
9-trihalogen anilines, and whether, in this case also, a j^bromine atom
was displaced.
The action of nitric acid on «-tribromoacetanilide has been studied
by Bemmers {Ber.y 1874, 7, 351), who obtained an anilide which is
described as crystallising in yellow needles readily soluble in alcohol ;
no melting point of the compound is given, and the numbers found in an
estimation of bromine correspond with those required for the formula
NOj'G^HBrj'l^H'GO'GH,. By hydrolysis with ammonia, an aniline
was obtained from it which crystallised in insoluble, yellow needles
melting at 214 — 215°; a bromine determination indicated that the
aniline was represented by the formula NOj'GoHBr^'NH,. This sub-
stance could not be 2 : 4 : 6-tribromo-3-nitroaniline, which was prepared
by Korner {Jahresher,^ 1875, 347) from nv-nitroaniline and melts at
102'5^ It was possible in Bemmers' experiments that the nitro-group
had displaced a bromine atom either from the ortho- or the para-
position, the liberated bromine entering the aniline molecule in the
meta-position. This transformation, sufficiently improbable in itself,
does not produce either 2:3: 6-tribromo-4-nitroaniline or 2:3: 4-tri-
Digitized by
Google
S-TRIHALOGEN AOBTANILIDKS. 501
bromo-6-iiitroaniline, which, the author has shown, melt respectively
at 155° and 166° (this vol., p. 499). The other possible aniline has
not hitherto been described.*
Bentley {Amer. Chem, J,, 1898j 20, 472) repeated Bemmers' ezperi-
mentp, but was unable to obtain any evidence of the substance,
described by him, and only succeeded in isolating the final products of
the action of nitric acid on this anilide, such as bromoanil, &o,
I have therefore reinvestigated the action of nitric acid on »-tri-
bromoacetanilide. When diluted with acetic acid, nitric acid has
little or no action on this substance ; when dissolved in cold fuming
nitric acid (sp. gr. 1*5), the anilide is only slowly attacked, but when
this solution in nitric acid is kept at a temperature of 50 — 55° for 15
minutes, the compound is largely converted into 2:4: 6-tribromo-3-
nitroaoetanilide. No bromine is evolved and no other substances are
produced.
Under sin^ilar treatment, 4-chloro-2 : 6-dibromoacetanilide yields
4-chloro-2 : 6-dibromo-3-nitroacetanilido. Both these anilides have
been converted into the corresponding anilines, namely, 2:4: 6tri-
bromo-3-nitroaniline and the hitherto undesoribed 4-chloro-2 : 6-di-
bromo-3-nitroaniline. These bases have also been prepared respec-
tively ivoin. fn-nitroaniline and 4-chloro-3-nitroaniline, and a compari-
son has been made of the anilines and their acetyl and acetylchloro-
amino-derivatives obtained from both sources.
It would appear from these results that the acetyl derivatives of
these «-trihalogen anilines behave, with nitric acid, in a manner
initially different from the s-trihalogen anilines themselves ; there is
not only no sign of the formation of nitroamino-derivatives and of the
products of their transformation, namely, replacement of the p-bromine
atom, but there is also no indication that the nitric acid reacts with an
iminoquinone derivative, as suggested in the previous communication,
in which case the /^-bromine atom would probably also be eliminated.
The directing influence of the amino-group as thus defined has been at
least partially suppressed on acetylation (compare Armstrong and
Horton, Proc., 1901, 17, 246). An analogous case is the nitration of
aniline sulphates in the meta-position relatively to the amino-group in
the presence of a large quantity of sulphuric acid (Nolting and Collin,
Bw.^ 1884, 17, 226). The behaviour of «-tribromophenol towards
nitric acid may be quoted to illustrate a similar phenomenon when an
oxygen, instead of a nitrogen, atom is the directing agent in substi-
tution ; this phenol yields 2 : 6-dibromo4-nitrophenol (Armstrong and
Harrow, Trans., 1876, 30, 477), the p-bromine atom being replaced by
* As Remmers obtained his substituted auiline (m. p. 214 — 215°) by heating the
anilide nnder preasare with ammonia, it is possible that a profound change took
place, and not a simple hydrolysis.
Digitized by
Google
502 OBTON : THB NTTRATIOir OF
a nitro-groap, whereas the ethyl ether, l-ethox7-2 : 4 : 6tiibromo-
benxene yields l-ethoz7-2 : 4 : G-tribromo-d-mtrobenzene (compare
Guareechi and Daocomo, Bm'.^ 1885, 18, 1175, and F^ust^ ilfmobn,
1869, 149, 152).
EXPBBIMBNTAL.
NUraiion of s-Tribromoaceianilide,
Five grams of «-tribiomoaoetamlide (m. p. 23S^) were added to 5 c.c.
of fuming nitric acid ; the anilide rapidly dissolves, and the solution is
cautiously warmed to about 55^ for 15 minutes. It is then poured
on to ice ; the nearly colourless solid which separates is washed free
from acid and dried. It is readily soluble in alcohol (compare Remmers,
loe. eU.)f and dissolves in a 10 per cent, solution of sodium hydroxide,
forming a colourless solution, whereas a nitroacetanilide, in which the
nitro-group is either in an ortho- or the para-position, dissolves in sodium
hydroxide with a marked yellow coloration (this voL, p. 496). Ke-
crystallised from benzene or dilute acetone, it forms colourless plates
melting at 197^, and, after repeated fractional crystalUsation, at 203^.
An analysis of this product gave numbers which indicated that it was
a mixture of a-tribromoacetanilide and a tribromonitroaoetaniUda
It was found that these two anilides could be separated from the
mixture (m. p. 197°) by converting them into the chloroamino-deriva-
tives. The mixture was dissolved in warm glacial acetic acid and
some excess of a solution of bleaching powder (0*8 normal, HCIO) was
added. The oil which at ^t separated solidified on cooling; the
solid was dried on a porous tile and then dissolved in a boiling mixture
of chloroform and petroleum. On cooling, crystals separated whidi
melted at 148°; on recrystallisation from petroleum (b. p. 50 — 80°),
an acetylchloroamino-derivative separated in characteristic translucent,
lustrous prisms melting at 159°, which is the melting point of acetyl-
chloroamino-2 : 4 : 6-tribromo-3-nitrobenzene (see below) :
0-1485 liberated I = 6-4 c.c. iVyiO iodine. 01 as :NC1= 7-64.
NOj-OgHBrg'NCl-CO'CH, requires 01 as :N01 = 7-85 per cent.
From the first mother liquor (a mixture of chloroform and petroleum),
impure acetylchloroamino-2 : 4 : 6-tribromobenzene (m. p. 100° instead
of 110°) was obtained ; by treatment with alcohol, this was converted
into «-tribromoacetanilide.
From the acetylchloroamino-2 : 4 : 6-tribromo-3-nitrobenzene (m. p.
159°), the corresponding anilide, 2:4: ^-trtbromO'S-nitraacetanUidey was
prepared by the action of warm alcohol ; it crystallised from dilute
acetone in tufts of silky needles melting at 217° :
0-201 2 gave 12*2 c.c. moist nitrogen at 15° and 761 mm. N » 6*93.
CgHjOjNjBrj requires N = 6*73 per cent
Digitized by
Google
8-TRIHALOGEN ACBTANIUDES. 503
The oorresponding base, 2:4: 6-tribromo-3-mtroaniline, was obtained
by heating the anilide (I part) with sulphurio aoid (4 parts) and
alcohol (4 parts) for se^^ral hours. On adding water, the aniline
separated and crystallised from dilute alcohol in small, pale yellow
needles melting at 102'' (Korner, loo. eU., gives 102*6'').
This base was also prepared from m-nitroaniline ; from this specimen,
on boiling with acetyl chloride in the presence of glacial acetic acid,
2A:6-irUn'(>mo-3'7iUroaoetanilide, NOj-C^HBrj-NH-CO-CHg, was ob-
tained, which crystallised in needles from dilute acetone, and in
flattened prisms or plates from benzene, and melted at 216 — 217'' :
0-148 gave 0-2004 AgBr. Br = 57-59.
CgHjOgNjBr, requires Br = 57*54 per cent.
2:4: e'Trthromo-Z-nitrodiacetanilide, N02-CeHBr3-N(00-OH3)2.— The
base or its monoacetyl derivative (1 gram) was boiled with acetic
anhydride (8 grams) for six hours. The anhydride was evaporated on
the water-bath, and the solid residue dissolved in chloroform. Large
lustrous, perfectly transparent rhombs separated melting at 175 — 176° ;
these were readily soluble in alcohol, less so in acetic aoid, and only
slightly so in cold benzene or chloroform :
0*158 gave 0*1946 AgBr. Br » 52*4.
CjQHy04N2Br3 requires Br = 52*28 per cent.
Ac6iylchloroamino-2 : 4 : ^-tribromo-S'^itrobenzenef
NO/CgHBrg-NCl-CO-OHg,
prepared from the anilide, crystallised in lustrous prisms from a mixture
of chloroform and petroleum and melted at 159" :
0-366 liberated I » 15-6 c.c. iVr/lO iodine. 01 as INOl » 7*55.
OgH^OjNjOlBrg requires 01 as :NC1 = 7'85 per cent.
Nitration of i-Chloro-2 : Q-dibromoacetcmUide. — ^The nitration of this
anilide was carried out in a manner exactly similar to that employed
in the case of 8-tribromoacetanilide. The mixture of unchanged anilide
and nitrated product melted at 19P ; it was dissolved in glacial acetic
acid and treated with a solution of bleaching powder; the chloro-
amines thus obtained were dissolved in a hot mixture of chloroform
and petroleum, and from the solution crystals separated which melted
at 126 — 130°. • These were recystallised from petroleum. Pure acetyl-
Moroami7ioA-chlorO'2 : Mibromo-Z-nitrohenzene^
NOj-OeHOlBrj-NCl-CO-OHj,
separated in aggregates of short prisms which melted at 134 — 135° :
01973 liberated I « 10-0 o.c. i^/10 iodine. 01 as :N01*8'98.
CgH^jNgCljBr, requires 01 as :N01-8-71 per cent.
Digitized by VjOOQIC
504 DIVEBS AND OGAWA : PREPARATION OF
From the first mother liquor, a mixture of chloroform and petroleum,
impure aoetylchloroamino-4-chloro-2 : 6-dibromobenzene (m. p. 102^
instead of 110°) was obtained which was reconverted into the anilide ;
after recrystallisation, the latter melted at 226°.
4-Ghloro-2 : e^bramo-S-nitroaeetanilide, NOg-CeHClBrj-NH-CO-CHa.
— This anilide was prepared by acetylating 4-chloro-2 : 6-dibromo-3-
nitroaniline with acetyl chloride. It crystallised from dilute alcohol
in tufts of silky needles melting at 224°, and was moderately soluble
in alcohol, acetic acid, or benzene :
01442 gave 0-201 AgCl + AgBr and 01253Ag. 01 = 9-49; Br = 42-99.
CgHjOsNjjClBr, requires 01 = 9*52; Br = 42-93 per cent.
It was also obtained from the corresponding chloroamine, prepared
from the product of nitration of 4-chloro-2 : 6-dibromoacetanilide.
This specimen crystallised in silky needles melting at 224° :
0-2022 gave 139 c.c. of moist nitrogen at 17-2° and 763 mm. N » 7*80.
CgHgOjNjClBr, requires N«7-64 per cent.
i-Chloro-2 i^dibromo-^-nitroaniline, NOj'C^HOlBrj'NH,, was pre-
pared by brominating 4-chloro-3-nitroaniline (m. p. 103 — 104°), and
it was also obtained by hydrolysis of the product of nitration of
4-ch]oro-2 : 6-dibromoacetanilide. It crystallises in small, pale yellow
needles melting at 103°, and is fairly soluble in all the usual organic
solvents :
0-140gaveO-2195AgCl + AgBrand 01368Ag. 01 = 10-73; Br = 48-39.
O^HjGjNjOlBrj requires 01 = 1 0*62 ; Br = 48-43 per cent.
St. Babtholoiibw's Hospital and Colleoe.
London, E.C.
LIIL — Preparation of Sulphamide from Ammonium
A midosulphite.
By Edwabd Dive&s and Masataka Ogawa.
Sulphamide occurs among the products of the spontaneous decom-
position of ammonium amidosulphite. That this appeared to be the case
was mentioned in the paper describing this salt (Trans., 1900, 77, 324).
It had then been isolated, not*only in too small a quantity to admit of its
purification and full analysis, but in a way that rendered its identity
almost doubtful. The decomposed amidosulphite had been extracted
with 95 per cent, alcohol, the residue from the evaporated voluminous
Digitized by
Google
SULPHAHIDE FROM AMMONIUM AMIDOSULPHITE. 505
solution extracted with undried ether, and the again very voluminous
solution evaporated. Half a gram of crystalline residue from about
150 c.c. of the ether solution was thus obtained, answering the tests for
sulpbamide, but melting much above 81% tasting not bitter, and yield-
ing a little too much sulphur on analysis. Then, too, we had failed to
get silver sulpbamide from the aqueous solution of the decomposed
amidosulphite, owing, as we afterwards found, to our having used
ammonia in excess. A 11 these points differed, or seemed to difEer, from
Traube's description, and caused us to hesitate in pronouncing the
substance to be sulpbamide. Since then we have obtained it in larger
quantity and pure, and thus become certain that sulpbamide is a
little soluble in absolute alcohol and even very slightly so in dry
ether, that it melts at 91% and that its silver derivative is insoluble
in ammonia alone, but soluble in ammonia in presence of the ammonium
nitrate which its mother liquor always contains. The publication of
Hantzsch and HoU's important contribution to the knowledge of sulph-
imide and sulpbamide (Ber,^ 1901, 34, 30), in whioh Traube's account
of sulphamide {Ber.^ 1893, 26, 609) is amended, affords welcome con-
firmation, so far as it goes, of the correctness of our own observations.
Hitherto, as is well known, sulphamide has only been got from
sulphuryl chloride and ammonia, a mode of preparing it which
Hantzsch and HoU have shown to be most laborious and unprofitable,
and the difficulty of getting it in this way has quite recently induced
Ephraim to try to obtain it from sulphuryl chloride by means of
urethane, but without success (Ber,, 1902, 36, 776). Sulphuryl chloride
is stated to give only 1 — 2 per cent, of pure sulphamide, whilst am-
monium amidosulpbite, by a process not unduly troublesome, yields 10
per cent, of its weight, and probably much more by skill and care.
In order to prepare the ammonium amidosulpbite and decompose it
afterwards, ammonia in excess and sulphur dioxide are led into a
closed flask, fitted with a thermometer and an exit-tube dipping in
mercury. To absorb the heat caused by the combination of the gases,
the flask is held in a bath of brine and crushed ice, which is more
effective when the flask contains some ether and is kept in motion,
because then the salt does not stick to the walls of the flask as a waxy,
badly-conducting coating. The rate of flow of the gases is to be regu-
lated by the operator's ability to prevent the temperature in the flask
from rising much above 10°. The inside of the apparatus, the gases,
and the ether are all to be dried before use.
When as much amidosulphite has been formed as may be wanted or
be convenient to prepare, the cooling mixture in the bath is replaced by
water, and a slow stream of dry hydrogen passed through the flask,
whilst the temperature of the water is slowly raised to about 70° and
then kept at that point for five or six hours or so long as ammonia
VOL. LXZXI. M M
Digitized by
Google
506 DITBBS AKD OOAWA: PRSPABATION OF
continues to oome off in any quantity. During thia operation, the
ether, if used, also evaporates. The sulphamide is all formed at
temperatures not higher than 30 — 35°, and a higher temperature is
hei-e employed only for the purpose of destroying as much as possible of
the thionic compounds which are formed along with the sulphamide and
would at a later stage consume much silver nitrate and undesirably
produce much ammonium nitrate, if present. The employment of a
higher temperature than 70^ to destroy all the thionic compounds, is
not possible, because then the sulphamide itself would be decomposed.
When the flask has cooled down, enough ice-cold water is poured in
to dissolve all its contents other than the considerable quantity of
sulphur left by the destroyed compounds. To the yellow, unfiltered
solution, which has been poured into an open vesael, barium hydroxide
is added in quantity a little more than sufficient for the salts it preci*
pitates, among which are sulphate, imidosulphite, and thiosulphate.
In order to lessen the dilution of the solution of sulphamide, the
barium hydroxide is used in mixed solution and crystals, as obtained
by rapidly cooling a hot, concentrated solution. The precipitate is to
be filtered off, although it is not very easy to get a bright filtrate,
and, even when this is accomplished, the filtrate soon becomes turbid
again, owing to further production of sulphate by the decomposing
salts present in it. This does not matter, however, and to the turbid
filtrate silver nitrate is added just so long as it continues to give a
precipitate. The barium hydroxide will have liberated much ammoniay
but a good deal of this will have evaporated during the time taken up
in filtration, especially if the precipitation has been carried out in an
open vessel. What remains of it, interferes only temporarily with
the silver precipitation, and does not usually need external neutralisa-
tion, for so much acid is formed as the result of a very nipid de-
composition of the precipitated silver salts (in which they change from
white to black) as to be more than enough to neutralise the anmumia
remaining in the solution, and also to dissolve up any silver sulphamide
that may have been thrown down at first. When the mother liquor
has become thoroughly acid or, exceptionally, has been made so by
adding nitric acid, and still holds silver in solution, it is filtered
from the black precipitate and just neutralised with ammonia. Any
slight precipitate then formed is also filtered off and rejected; it
contains no trisulphimide. The filtrate holds little else than sulphamide
and ammonium amidosulphate, and if evaporated over sulphuric acid
would yield both these substances in characteristic crystals. But
to isolate the sulphamide, it is to be precipitated from this solution by
silver nitrate and ammonia, that is, by Traube's method. The silver
sulphamide, thus obtained, is almost pure, there being no such add
matter present as is met with when the sulphamide has been prepared
Digitized by
Google
SULPHAHIDE FBOM AMMONIUM AMIDOSULPHITE. 607
from sulphuryl chloride. In that case, a viscid silver salt accompanies
the silver sulphamide, and according to Hantzsch and Holi, can only
be removed from it hy a process entailing the destruction of much of
the Bulphamide. Even in the present case, however, the amidosulphate
left with the sulphamide may cause a little difficulty, unless care be
taken.
On referring to the memoir on amidosulphuric acid in the Transac-
tions of the Society for 1896 (pp. 1647—^1649, " amidosulphonic acid "),
it will be found there stated that when an alkali is added in suitable
quantity to a solution of mixed silver nitrate and potassium amido-
sulphate, a bright yellow, amorphous salt is precipitated, which is
very probably AgHN'SOjK, and is soluble in, and ultimately
decomposed by, excess of alkali. It is now found that, in precipitating
silver sulphamide in presence of anunonium amidosulphate^ as in the
present case, a very small quantity of a bright yellow substance,
probably ammonium argentamidosulphate, is apt to accompany the
silver sulphamide, and that in order to circumvent this liability and
at the same time to avoid loss of the silver sulphamide through its
solubility in ammonia in presence of ammonium nitrate, precipitation
should be carried out in the following way. Having added more silver
nitrate, dilute ammonia is dropped in, slowly and with stirring, until
the solution is slightly alkaline. The precipitate is quickly filtered
off and washed free from mother liquor. The solution is again treated
with silver nitrate and ammonia, as before, in order to see whether any
more sulphamide is thrown down. This being quickly filtered off, if
it be desired to obtain a sight of the yellow compound, a few drops or
more of ammonia may be added, and dilute silver nitrate very
slowly dropped in, when it will be produced.
The silver sulphamide, perhaps a little yellow after all, is to be
dissolved in dilute nitric acid, ammonia added to slight alkalinity, as
before, and then two or three drops of silver nitrate. After a
repetition of this treatment, the precipitate is treated with exactly
enough dilute hydrochloric acid to decompose it^ just as Traube
directs. The filtrate from the silver chloride, which must not be acid,
gives the sulphamide in good crystals, when it is evaporated in the
desiccator. It is to be recrystallised. Since it is exceedingly soluble,
the mother-liquors must be worked up, if a good yield is wanted.
We are imder obligations to Mr. Tojiro Suzuki for material
asflistanoe in the experimental work of this paper.
M M 2
Digitized by VjOOQIC
508 TILDEN AXD BTTRROWS : TH£ COMSTtTlTTION OF LnCKmir.
LIV. — The Constitution of Limettin.
By W. A. TiLDEN and H. Buiuows.
LnmriK is a substanoe which occurs in the pericarp of the lime fruit
and is deposited from the essential oil which has been extracted by the
sponge process or by pressure. It was shown by Tilden (Trans., 1892,
61, 344) that its composition is expressed by the molecular formula
^11^10^4' and its constitution by the formula C^HJify^QB^^^OJELO^
The same substance has been since shown by E. Schmidt {Api>ih, ZeiLf
1901, 16, 619) to be present in oil of lemon, and this observation is
confirmed by the experiments of Burgees (Proc., 1901, 17, 171).
From what follows, it appears that the constitution of limettin is
similar to that of coumarin, and that the group CgHO, has the struc-
ture of an unsaturated lactone ring, corresponding to that of ooumarin«
Limettin is, in fact, 4 : 6-dimethoxycoumarin, isomeric with dimethoxy-
iBsculetin and daphnetin,
0-OH,
This view of its constitution is supported by the facts, recorded
in the previous paper, that limettin is soluble ift strong alkali solu-
tions and is reprecipitated unchanged by acids, and that the products
of fusion with potash are phloroglucinol and acetic acid.
The dibromo-derivative, when treated with alkali, readily yields
up one atom of bromine in the formation of the corresponding
monobrominated coumarilic acid,
(OH,-0)AHBr<^^' -> (OH,-0),OjHBi<^>C-CO,H.
Di- and tri-chlorolimettin behave in a similar manner.
Owing to its extreme solubility, the sodium salt of limettin can>
not be isolated from its aqueous solution, but it is precipitated by
the use of an alcoholic solution of sodium ethoxide. This com-
pound could not be methylated by the action of methyl iodide, un-
changed limettin being the only product. From the corresponding
silver salt, however, a small quantity of a homologue was formed,
together with a relatively lai^e proportion of limettin. From con-
siderations which appear later, the homologue has the following
structure :
/Google
Digitized by ^
TILDEN AND BURROWS: THE CONSTITUTION OF UMETTIN. 509
An attempt was made to synthesiee limettin by proceeding from
the aldehyde from phloroglucinol, described by Gkittermann and Kobner
(Ber., 1899, 32, 280), by condensation with acetic anhydride, but it was
found that under the conditions tried the saponified product was
unchanged by the action of methyl iodide.
EXPBBIMBNTAL.
The material required in the following experiments was prepared
from the same source as the original specimen, namely, the deposit
which is filtered from commercial oil of limes. As described in
the first paper {loc, eit,), it melts at 145^ and exhibits a distinct
blue fluorescence, but whs never obtained perfectly white.
Bromo-i: : 6'dirnethoxyeauma/riiio Acid.
Dibromolimettin, boiled with a 10 per cent, solution of aqueous
potash, readily dissolved, and the solution, after cooling, gave a
copious precipitate when acidified. The substance, recrystallised
from acetic acid, exhibited acid properties and melted at 239^.
Dibromocoumarin, when similarly treated, behaves in exactly the
same way, and yields monobromocoumarilic add.
Representing dibromolimettin by the formula
.OHICBr
(OH,.0),C,HBr<:^rjQ ,
the corresponding coumarilic acid is (CHj-0)jOeHBr<^_>C'CO^ :
0-3062 gave 0-4900 CO, and 0-0786 H^O. C = 43 -79 ; H - 2-86.
0-3396 „ 0-2100 AgBr. Br = 26-31.
CiiHjjOjBr requires C - 43-86 ; H = 299 ; Br- 26-67 per cent.
The potassium salt crystallises from dilute alcohol in fine, white
needles :
0-4872 gave 01238 K^jSO^. K = 11-6.
C^iHgOgBrK requires K- 11-6 per cent.
A solution of the above salt in methyl alcohol, heated with
methyl iodide, gave an easily saponified ester in the form of white
needles melting at 181^ :
0*1784 gave 0-1068 AgBr. Br « 26-48.
Cj^H^O^Br requires Br » 26-39 per cent.
Digitized by
Google
610 TILDfiN AND BURROWS: THE CONSTITUTION OF LIMBITIN.
^CHICBr
4 : Q-Dtaoetyltribr<mioe(minarin, {C^Hfi^\Cf^Br^<^Q Xq .
The ultimate bromiDation product of llmettin is obtained by heat-
ing the dibromo-compound with bromine under pressure.
Five grams of dibromolimettin, 5 grams of bromine, 2 grams of
iodine, and 3 c.c. of water were heated in a sealed tube at 110^ for
3 hours. The resulting garnet-red crystals dissolved in cold alkali,
but on heating the solution they decomposed. Nitrobenzene was
found to be the only available solvent, and analysis showed the
recrystallised substance to be still impure. On boiling these crys-
tals, presumably consisting of the tribromodihydrozycoumarin, with
acetic anhydride and allowing the solution to cool, white prisms
were deposited, which crystallised from acetic acid and melted at 244° :
0-2432 gave 0-2812 COj and 0-0306 HjO. C-31-66 ; H- 1-39.
0-2272 „ 0-2670 AgBr. Br = 48-14.
CjjH^OjBrg requires 0 = 3126; H = 1 40 ; Br = 48 09 per cent.
AfonoMarolimetUn, (CH3-0)2CeHCl<^^^^ .
This compound is obtained when, in the preparation of trichloro-
limettin, the current of chlorine is stopped as soon as the precipitate
first formed has reached a maximum and re-solution begins. The
precipitate crystallises from acetic acid in needles which melt at
242°. Long-continued boiling with 10 per cent, aqueous potash
solution or with alcoholic potash has no effect on the halogen atom:
0-1462 gave 0-2934 CO, and 00446 HjO. 0-64-71 ; H = 3-39.
0-1954 „ 0-1168 AgOl. 01=14-79.
OjjHjjO^Ol requires 0 = 6488 ; H = 374 ; 01 = 1476 per cent.
DichlorolimetHn, (OH,'0)aOffH01<Q ^ .
The final product of the action of chlorine on limettin appears to be
the trichloro-derivative, whether in the presence of a " carrier " or
not. The precipitate first formed when a few crystals of iodine are
added has, however, very different properties from the monochloro-
compound. It is much less soluble in acetic acid, and one atom of
chlorine is removed from it by alkali It melts at 276° :
01368 gave 0*2396 00, and 00324 H,0. 0 = 4812 ; H - 2-65.
01646 „ 0-1692 AgOl. 01 = 25-48.
O^HjO^Ol, requires O*»48-0 ;fL^i'9; 01 « 26-81 per o^t,
Digitized by
Google
TXLDBN AND BURROWS : THE CONSTITUTION OF LIMETTIN. 511
The tnanoehhroeoumanlic acid, {ClIj^*0\G^B.Ol<^^C*CO^T[, ob-
tained from dichlorolimettin by the action of alkali, melts at 189^ :
01278 gave 0-2410*OOj and 00434 H^O. 0«6M1 ; H = 3-77.
0-1410 „ 00796 AgCl. a-13-96.
C^Hj^OgOl requires C = 51 46 ; H = 35 ; CI = 1384 per cent.
The diehlorocaumarUio ^acid, (0Hs-0)2C^Cl,<^^C'C0jH, is
formed by the action of alcoholic or aqueous potash on trichloro-
limettin. It melts at 269^ :
0-2249 gave 03760 COg and 00590 H^O. 0 - 46-59 ; H - 2-91.
0-2244 „ 0-2188 AgCl. Cl-24-11.
G^iHgOsCl, requires G« 45-36; H»2-74; 01 - 24*4 per cent
JHsodiiim^iQ-dmethoxf/caumarate, (CH,-0),C^H^<^5^^^'^^2^*.
The sodium salt of dimethoxycoumaric acid separates out on heating
an alcoholic solution of limettin with excess of sodium ethoxide on the
water-bath. It is insoluble in absolute alcohol, but extremely soluble
in alcohol containing a very small quantity of water. Washed with
absolute alcohol, the salt exhibits a strongly alkaline reaction to litmus
paper, owing probably to adhering ethoxide ;
0-8968 gave 0-4902 Na^SO^. Na - 1 7-60.
^11^10^6^*2 requires Na = 17-16 per cent.
The silver salt is obtained by adding an excess of silver nitrate solu-
tion to the sodium salt dissolved in water. It is insoluble in water
and the usual organic solvents :
0-3022 gave 0-1976 AgOl. Ag« 49-21.
C^^H^oO^Agj requires Age 49-31 per cent.
4 : Q'JHfnMofty'a'fnethyleoumarin, (OHg'Q)gO^Hg<CQ Xq .
The sodium salt of limettin, when heated with absolute alcohol and
methyl iodide on a water-bath, gave only regenerated limettin. The
same result was also obtained by heating at 100° in a sealed tube, but
by heating the silver salt suspended in methyl alcohol with methyl
iodide for^ hours, evaporating the alcohol off, and recrystallising the
residue from acetone, white tufts of needles were deposited which
melted at 189°. The mother liquor contained limettin. The new sub-
stance is soluble in boiling alkali, but is precipitated unchanged by
acids. ^fxnXjBeB showed the substapce tp be a hopiologue of limettin ;
Digitized by
Google
512 DAWSON AND OBANT: A MSTHOD OF DETEBXINIKO THE
0-2294 gave 05474 00, and 01124 H,0. 0 = 6498 ; H«5-44.
01368 „ 0-3278 00, „ 00664 H,0. 0 = 66-34; H- 6-39.
OjjHj^O^ reqoiiBS 0 = 65*45 ; H=5*45 per cent.
Difn^thoxyHi-methyl-P^omo(xmmarin, (OHft*0),OitH^'^^ *!lL *.
A solution of bromine was gradually added to homolimettin, both
dissolved in acetic acid, until the colour of the bromine was permanent.
The solution poured into water gave a precipitate which ciystallised
from acetic acid in the form of needles melting at 260° :
01340 gave 0-2362 00, and 00422 H,0. 0 = 4807 ; H - 35.
01882 „ 0-1172 AgBr. Br=26-5.
^12^1 ABr requires 0 = 481 6 ; H = 3-7 ; Br = 26-75 per cent.
JHrnethoxy-a-rMthyl-fi-hydraxi/caumarinf (OHg'O)gOflH^'^^ ^*Y^ ',
The bromine atom in the monobromo-compound is readily removed
by heating with alkali. The product then crystallises from dilute
alcohol in minute rosettes of needles and melts at 248°. Heated with
acetic anhydride, the acetyl compound formed melts at 134°. On
saponification, the original substance was obtained with unaltered
melting point :
0-1172 gave 0-2612 00, and 0-0542 HgO. 0 = 6078 ; H = 5-13.
Oi,Hi,05 requires 0 = 6101 ; H = 508 per cent.
As there is no formation of a coumarilic acid when the substance is
treated with alkali, the halogen in this compound must occupy the
/9-position relatively to the 00-group, being in this respect unlike the
hajogen derivatives of limettin previously described.
Royal College of Science, London.
South Kensington, S.W.
LV. — A Method of Determining the Ratio of Distribu-
tion of a Base between Two Acids.
By H. M. Dawson and F. E. Qbant.
Several methods have been employed for determining the ratio of
distribution of a base between two competing acids. These may be
divided into two classes, methods belonging to the one being designated
Digitized by
Google
RATIO OF DISTRIBUTION OF A BASE BETWEEN TWO ACIDS. 613
as chemical, and those to the other as physical. Where the system
under investigation is homogeneous, that is, where the substances are
all in solution, the physical methods only are capable of application,
and chemical analysis can furnish no information in regard to the con-
centrations of the reacting compounds.
Any physical changes which accompany the chemical phenomena, or
physical properties which depend on the chemical arrangement of the
reacting components, can be used for the determination of the equili*
brium in the solution. The thermal changes accompanying the
chemical transformations have been utilised by Thomsen ; Ostwald has
made use of the variations of the specific volume and specific refractive
power, these three well known methods being of wide application and
jgiving fairly concordant results. Jellet {Trans. Ray. Irish. Acad,,
26, 371) made use of optical rotatory power, and Wiedemann {Ann^
Phys. Chem., 1878, [iii] 6, 46) of magnetic qualities, to ascertain the
equilibrium relationships, the range of applicability being in these cases,
however, very small. Golorimetric and photometric measurements have
also been employed, the former by Gladstone {Fhil, Mag.f 1865, [iv], 0,
636), and by Muller (Foggendarff's Ergdnsungshand, 1873, 6, 123), and
the latter by Settegast {Ann. Phys. Chem., 1879, [iii], 7, 242). The
colorimetric method appears, however, only capable of yielding results
of a qualitative nature.
The theory of the method used by us can be described in a few words.
An aqueous solution containing two different acids and a quantity of a
base insufficient for neutralisation may be regarded as containing the
foUowing components :
HA MA MA HA'
HA and HA' representing the two acids and MA and MA' the
corresponding salts. For the sake of simplicity, the electrolytic
dissociation of the reacting components is left out of account in this
formulation. If this aqueous solution is shaken up with a liquid with
which it is practically immiscible, and if this liquid is capable of taking
up one, and one only, of the four reacting components from the aqueous
solution, say the acid HA, we have at once a means of determining
the concentration of each of the components in the aq\ieous solution if
the quantities of the two acids and the base in the original solution are
known. A knowledge of the ratio of distribution of HA between
water and the second non-miscible liquid will enable us to calculate
the concentration of the free acid HA in the aqueous solution for any
given concentration in the second liquid, and since the original quanti-
ties of the two acids and the base in the aqueous solutions are known,
the concentrations of the other components can be easily calculated.
As the result of preliminary experiments, it was found that when
Digitized by
Google
514 DAWSON AND GRANT: A METHOD OF DETERMINING THE
aqueous solations of tartaric, malic, citrio, and raocinic adds are shaken
up with chloroform, the acid remains entirely in the water. Solntiona
of acetic acid, on the other hand, give up an easily measurable propor-
tion to the chloroform layer. Each of the first mentioned acids can
therefore be combined with acetic acid and the ratio of distribution of
a base between these can be determined by extracting the aqueous
solution with chloroform.
Distribuium qf Aeetie Acid hehoemi Water and CUatriform.
From the preceding description of the method, an accurate know-
ledge of the ratio of distribution of acetic acid between water and
chloroform is requisite. The mode of experimentation was the follow-
ing. Sixty C.C. of acetic acid solution and 60 c.c. of freshly distilled
chloroform were introduced into a cylindrical separating funnel, which
was immersed in a thermostat maintained at 20°. At intervals of
about five minutes during a period of three-quarters of an hour, the
funnel was removed and the contents violently agitated. The chloro-
form and aqueous solution were then allowed to separate completely
by leaving the funnel in the thermostat for half-an-hour, after which
the two layers were run off into separate stoppered tubea Measured
portions were then titrated, the aqueous layer with i\r/4, and the
chloroform layer with iV/40 sodium hydroxide, free from carbon
dioxide. The results obtained are given in the following table, the
concentrations being expressed in gram-equivalents per litre :
Concentration of aqneous
Concentration of chloro-
C^
layer.
form layer.
''t
Cv
0^
1-586
0-2277
6-74
0-9084
0-08904
10-21
0-6089
0-04567
18-87
0-8178
0-01676
20-17
0-2696
0-01222
22-07
0-2615
0-01109
22-68
0-1946
0 007686
25-66
0-1691
0-005608
28-36
0-1269
0-004061
81-23
0-09594
0-002848
38-68
0-06446
0 00174
87-04
It is evident from the numbers in the last column that the ratio of
distribution of the acetic acid varies very considerably with the con-
centration, and that the proportion of acid in the aqueous layer
increases as the absolute amount decreaaes. This is perhaps due to
the acetic acid dissolved in the cblorofonn consisting largely of dopbio
Digitized by
Google
BATIO OF DISTRIBUTION OF A BASE BETWEEN TWO ACIDS. 615
molecalesy whereas in water, neglecting the small amount of electrolytic
dissociation, the acid is present in the form of simple molecules. The
C
increase of the ratio j^ with increasing dilution would then correspond
to a gradually increasing dissociation of the double molecules in the
chloroform into simple molecules.
In order to be able to determine the concentration of acetic acid in
.water for any given concentration in chloroform, the graphic re-
presentation in the figure has been made use of. In this diagram the
concentration of the acetic acid in the chloroform C^ is represented on
the abscissa, and the distribution ratio q on the ordinate. The diagram
contains only the data for the smaller concentrations. Multiplication
of any value of G^ by the corresponding value of q gives at once the
concentration of acetic acid in the water layer.
If^fiuence of Dissolved Adda and Salts on the Distribution qf Acetic
Acid between WcUer and Chlorqform,
In the actual experimental determination of the ratio of distribution
of sodium hydroxide between acetic, and, say, tartaric acids by the
method indicated above, it must be remembered that the aqueous
layer contains free tartaric acid and sodium salts of the two acids.
The presence of these may influence the ratio of distribution of the
acetic add between the water and the chloroform, in which case a
correction would have to be made in the value of q given by the curve
before the concentration of the free acetic acid in the aqueous layer
could be calculated for a given concentration in the chloroform. To
ascertain the order of magnitude of this iDfiuence, experiments have
been made on the quantity of acetic acid extracted from aqu€i6us
solutions containing (1) sodium acetate, (2) tartaric acid. It is
obviously impossible to determine experimentally the influence of
sodium tartrate on the distribution of the acetic acid between the two
liquids, for acetic acid would partially displace the tartaric acid from
the salt. It may be assumed, however, that its influence will be ap-
proximately the same as that of sodium acetate.
The experimental data are collected in the following table, the efiEect
of the sodium acetate or tartaric acid on the distribution of the acetic
acid being clearly seen by a comparison of the values of q' with those
of q^ the latter being taken from the curve on p. 516. The corre-
sponding values of ^ and q are the ratios of distribution for the same
concentration of acetic acid in the chloroform when the aqueous layer
contains either tartaric acid or sodium acetate, or is free froqi thes^
$ubstapc^s ;
Digitized by
Google
516 DAWSON AND GRANT: A METHOD OF DETERMINING THE
Concentration of foreign
substance in aqueous
solution.
Concentration
of acetic acid
in aqueous
solution, Ci,
Concentration
of acetic acid in
chloroform, C^
9-
0*80 jY sodium acetate
0-25 „
0-20 „
0-10 „
0-25 iVtertaric acid
0-8012
0-2505
0-1991
0-1008
0*2868
0-01482
0-01069
0007654
0*002988
0-01021
21-08
28-46
26*86
84-81
28-16
20-9
23-0
25-8
88 8
23 'S
The differences between the values of q' and q indicate that the
sodium acetate and tartaric add in the aqueous solution have some
38
\
9A
\
O*
\
\-
QA
\
\,
S
\
on
\
^
X
oo
"^
^-
0-002 0*005 0*010
Concentration in chlorqform {gram-equivtilents per lUr»).
0-014
influence on the ratio of distribution of acetic acid between water and
chloroform. In all the experiments the influence exerted is, however
very slight, and since the chief object of this paper is to describe
new method of determining the distribution of a base between two
acids rather than to establish more accurate values, we have assumed
the influence of dissolved salts and second acid on the distribution aa
negligible.
Digitized by
Google
BATIO O** DIStRIBTTTION OF A BASE BfiTWEBN tWO ACtDS. 617
DetemUnaiion qf the Acetic Acid exti'octed by Chlorqform /rom Aqueous
SohUiane containing Equivalent Qtumiitiee of Acetic Acid, Sodium
HydroQcide^ and a Second Acid,
Solations of the various acids and of sodium hydroxide of as nearly
as possible normal concentrations were prepared. The acids were
carefully purified, and the solution of alkali was made from metallic
sodium. If equal volumes of normal solutions of sodium hydroxide,
acetic acid, and, say, tartaric acid, are mixed, and the resulting solution
is shaken up with chloroform, then since the latter extracts a certain
amount of the acetic acid from the aqueous solution, this will no longer
contain the base and the two acids in equivalent quantities. To com-
pensate for the acetic acid thus extracted by the diloroform, this acid
was added in slight excess to the original solution, so that the latter after
being shaken up with chloroform would contain the acetic acid in
amount equivalent to the base and the other acid present. The
quantity of acetic acid thus required in excess of the other acid was
determined approximately by consideration of the curve.
The exact method of experimentation and calculation will be seen
best if one of the determinations is considered in detail.
Twenty c.c of i\r-tartaric acid, 20 ac. of iV-sodium hydroxide, and
20*7 C.C. of i^-acetic acid were introduced into a graduated flask and the
mixture diluted to 100 c.c. ; 60 c.c. of this solution were then shaken up
as previously described with 60 c.c. of chloroform. Titration of the
aqueous layer gave 0*1990, that of the chloroform 0*005632 gram equi*
v&lents per litre. From the first titration number, we may write down
the concentrations of the acids and base in the aqueous solution. In
gram-equivalents per litre they are 0*200 of tartaric acid, 0*200 of
sodium hydroxide, and 0*199 of acetic acid. From the curve, the
distribution ratio of acetic acid between water and chloroform at a
concentration in the latter solvent of 0 005632 gram equivalents per
litre is found to be 28*3. The product, 28*3 x 0005632 « 0*1593, gave
the concentration of the free acetic acid in the aqueous layer. The acetic
acid present as sodium acetate is (0*199 • 0*1593) « 0*0397 gram-equiva-
lents per litre, and this number represents also the conoentration of
the free tartaric acid. Finally, the tartaric acid in combination with
the base is (0-200- 00397) » 0*1603 gram-equivalents per litre. The
ratio of the amount of tartaric to that of acetic acid in combination
0*1603
with the base is therefore A.A3Q7 ^ ^'Oi.
Similar experiments on the distribution of sodium hydroxide between
acetic and tartaric acids to that just described have been carried out
at other concentrations. In the strongest solution, the conoentration
Digitized by
Google
518 DAWSON AND GRANT : A METHOD OF DETBBBClNIHa THS
of each of the two acids and the base was 0*3 I/', in the next 0*25 If,
in the third, 0*2 iT, and in the most dilate solution, 0*1 If.
The same concentrations were employed in corresponding series of
experiments, in which the tartaric acid was replaced succes^iyely bj
citric, maJic, a^nd succinic acids. The tables on p. 519 contain the
data calculated from the experiments. The first column gives the
concentrations of the two acids and the base in the aqueous solution
after shaking up with chloroform ; the second gives the concentration
of the acetic acid in the chloroform layer ; the third, the corresponding
value of the distribution coefficient q. In the . fourth, the concentra-
tions of the free acids in the aqueous solutions are given, the fifth
containing that of the combined acids, all concentrations being ex-
pressed in gram-«quivalents per litre. The last column contains the
ratio of the amount of combined tartaric (citric, malic, succinic) acid
to that of the combined acetic acid.
The accuracy attainable in the determination of the ratio given in
the last column is smaller the smaller the concentration, the amount
of acetic acid in the chloroform layer decreasing rapidly with
the concentration of the solutions and becoming very small when
the acids and base are present in 01 iiT-concentration. For this reason,
experiments with 0*1 iV-solution were not carried out in the case of
malic and succinic acids.
An assumption, not yet mentioned, has been made in connection
with the calculation of the above results, namely, that no sensible
change of volume takea place in the aqueous solution when the latter
is shaken up with an equal volume of chloroform. Hers {B&r., 1898,
31, 2669) found that, at 22"", 100 volumes of water dissolve 0*42
volume of chloroform, whilst 100 volumes of the latter are capable of
taking up 0*152 volume of water. We may conclude that the aqueous
solution on being shaken up with chloroform undergoes a slight altera^
tion of volume. To ascertain the order of magnitude of this alteration
under our conditions of experimentation, 0*25 i\r-citric add solution
was shaken up at 20% with an equal volume of chloroform and the
concentration of the acid in the aqueous solution compared with that
in thb original solution. The concentrations were :
Original solution 0*2519 gram-equivalents per litre.
After shaking with chloroform 0*2508 „ „ „
The volume of the aqueous solution then undergoes a slight increase
after shaking with the chloroform, but for the reason previously stated
this small alteration has not been taken into account in the above
calculations.
Although an attempt was made to obtain solutions containing the
acids and base in exactly equivalent quantities, .inspection of the
Digitized by
Google
RATIO OF DISTRIBItTION OF A BASE BUTVITEBN TWO ACIDS. 619
Oompoeition
of aqueous
solution.
Concentia*
tion of
acetic acid
in
chloroform.
Concentration of^
free acids in
aqueous solution.
Concentration of
combined acids
in aqueous
solution.
Ratio of
quantities
of
combined
acids.
0-800 tartaric
■ 0-800 KaOH
0'8081 acetic
0-250 tartaric
• 0-250 NaOH
0*9502 acetic
f 0-200 tartaric
i 0-200 NaOH
1 0*199 acetic
f 0-100 tartaric
-1 0-100 NaOH
10-1017 acetic
0*8015 citric
0-SOO NaOH
.0-3049 acetic
'0-2612 citric
0-250 NaOH
0*2522 acetic
'0*201 citric
0-200 NaOH
0-1988 acetic
'0*1005 citric
0-100 NaOH
0*1004 acetic
0*800 malic
0-800 NaOH
0-2998 acetic
0-250 malic
0-260 NaOH
0-250 acetic
'0*200 malic
0-200 NaOH
0*1999 acetic
0*0108
000822
0-005632
0 00224
22-9
25-0
28*8
84*9
/Tartaric 0*0558
\ Acetic 0-2478
/Tartaric 0*0447
\ Acetic 0*2055
/Tartaric 0-0897
\ Acetic 0 1598
/Tartaric 0-0235
\ Acetic 0*0782
Tartaric
Acetic
Tartaric
Acetic
Tartaric
Acetic
Tartaric
Acetic
0*2441\
0-0558/
0-2058\
0-0447/
0-1608\
0-0897/
0*0765\
0-0285/
Ae&iie and Citric Acids,
000700
26-4
•
0-005185
29-1
0-008682
82-0
0 001516
87-9
0*008204
25-0
0-006046
27-7
0-004386
80*5
/Citric
L Acetic
0 1216
0-1848
Citric
Acetic
0-1799\
0-1201/
/Citric
\ Acetic
0-1040
0-1494
Citric
Acetic
01472\
0*1028/
/Citric
L Acetic
0*0815
0-1178
Citric
Acetic
0-1195\
0-0806/
/Citric
\ Acetic
0-0485
0-0574
Citric
Acetic
0*0570\
0-0480/
md Malic Acids
.
/ICaUc
\Acetic
0-0942
0*2051
Malic
Aoetic
0-2058\
00942/
/MaUc
\ Acetic
0*0825
0*1675
Malic
AceCic
0-1675\
0-0825/
/Malic
LAcetic
0*0662
0*1887
Malic
Acetic
0-1888\
0-0662/
4-88
4-59
4*04
3*26
1-50
1-43
1*48
1-88
2*08
2-08
2*02
Acetic and Succinic Adda.
0*8006 succinic
• 0 800 NaOH
0*8027 acetic
0*005594
28-4
/Succinic 0-1444
\Acetic 0*1589
Succinic 0*1562\
AceUc 0*1488/
1*09
0*800 succinic
« 0-800 NaOH
0*8084 acetic
0-005781
28-1
/Succinic 0-1410
VAcetic 0*1624
Succinic 0-159 \
Acetic 0*141 /
1*18
0-2505 succinic
0*250 NaOH
0-2585 acetic
0*004852
30-6
/Succinic 01218
\ Acetic 0-1827
Succinic 0 -12921
Acetic 0-1209/
1-07
f 0-2004 succinic
0-200 NaOH
0-1986 aoetio
0*008078
88-4
/Succinic 0*0964
I Acetic 0-1026
Succinic 0*1040\
Acetic 0*0960/
1-08
Digitized by
Google
620 BATIO OF DISTRIBUTION OF A BASE BETWEEN TWO ACIDS.
numbers in the first oolumn will show that this is only approximately
the case. The greatest deviations are found in the first and third
experiments with citric acid. It is possible by applying the law of
mass action to the equilibrium in the solution to correct for these
deviations from exact equivalency. This has been done in the case
of the experiments mentioned, but the corrected values (1*51 and I -47)
differ to an inappreciable extent from the numbers (1*50 and 1*48)
actually obtained. The differences are well within the errors of
experiment and the numbers given in the last column may accordingly
be taken to represent the ratios of the combined acids when these ar*)
present in equivalent quantities.
It may be of interest to compare the above numbers with the values
of the ratio of distribution obtained by other static methods. The
results obtained by the thermochemical method are scarcely accurate
enough to admit of a rigorous comparison. Ostwald's measurements
of volume and specific refractive power were made at approximately
JV/3 concentration, and those on the solution of calcium oxalate by
different acids at N and i^/10 concentrations. The following table
contains the data which may be regarded as comparable, the older
numbers having been recalculated so as to make acetic acid the basis
of comparison :
Tartaric
Malic ..
Citric ..
Snccinic
Acetic ..
New method.
Concentratioii
O'SN.
4*88
208
1-60
111
1-00
Volume method
(Ostwald).
Concentratioii
0-88 JV.
4*28
2*80
1-18
1-00
New method.
Concentration
01 N.
8-26
1*88
1-00
Calciom oxalate
method
(Ostwald).
Concentration
0-1 a:
1*93
1-64
l-9«
1-27
100
The first two series of numbers agree very well, but at the smaller
concentration considerable deviations are found. This is to be expected
since the two series of values given by Ostwald can scarcely be reconciled
with each other. Such large changes in the relative values are scarcely
probable for such a small change of concentration.
Finally, it may be pointed out that the above method of experimental
tion will serve equijly well to determine the ratio of distribution of an
acid between two bases. The necessary condition is that some liquid^
not miscible to any extent with water, can be found, which will
extract in conveniently measurable quantity one of the four reacting
components. It can be shown that even if two of the components,
say the two acids, are extracted by the second liquid, yet under
Digitized by
Google
. THE MOLKCOLAB COMPLSXITT OF ACETIC ACID. 521
oertain conditions the ratio of distribution of the base in the aqueous
solution can be measured.
Suppose that of the components HA, MA^ MA' and HA\ present
in the aqueous solution, HA and HA' are taken up by the second
liquid. Let q be the ratio of distribution of HA^ and g[ that of HA
between water and the second liquid. Suppose in any experiment,
the total concentration of acid in the second liquid =< G^t then 0^ »
Cha -^OffA't and the concentration of free acid in the water ^C^t
then C^ ^qpHA -^^^HA'* From the values C^ and {7,, determined
experimentally, it will obviously be possible to calculate the concen-
trations of each of the two acids in the second liquid, and therefore the
concentrations of the free acids qGaA ^"^^ 4^ha' ^^ ^^ aqueous
solution, provided the values of the distribution ratio q and q[ are
sufficiently different from each other. In the case where q^(i^
then the individual concentrations of the two acids cannot be deter-
mined at all. It is supposed, in this description of the method, that
the presence of the dissolved salts and second acid in the water, and
that of the second acid in the chloroform, has no effect on the dis-
tribution of the first acid and mot versd.
Tbb Tobkshirb Collbob,
LXBDS.
LVI. — The Molecular Complexity of Acetic Acid in
Chloroform Solution.
By H. M. Dawsok.
The experimental results communicated in the previous paper show
that the ratio of distribution of acetic acid between water and chloro-
form increases very considerably with increase of the dilution. It has
been there suggested that this is due to the gradual splitting up of the
double acetic acid molecules, which are supposed to be present in
solvents not containing the hydroxyl group, into simple molecules.
With the object of testing this assumption, the equation of the dis-
sociation isotherm has been applied to the acetic acid dissolved in the
chloroform^ the experimental data in the previous paper being utilised
for the purpose of calculation. As will appear from what follows, the
concentrations of acetic acid in chloroform which have been used
correspond to a region of almost complete molecular transformation.
As before, let e| and e^ represent corresponding concentrations of
acetic acid in water and in chloroform. If a is the degree of dissociation
of the acid in the aqueous solution at this concentration, then 0^(1 - oi)
YOL. LXXXI. N N
Digitized by
Google
622
DAWSON : THE MOLECULAR COMPLEXITY Of
is the concentration of the ondissociated molecules. If r is the ratio of
distribution of the simple molecules between water and chlorofcnrmy
0,(1 -a)
— will be the concentration of the simple acetic acid molecules
in the chloroform^ and c, - -^ that of the associated molecules.
- Applying the law of mass action, the result obtained is :
«.(i-»>
or
f^e^ - »'Ci(l - o)
.E,
where f is a constants
By taking the second and the seventh experiments (see first table
in previous paper) as a basis, and assuming that the experimental data
for these give the same value for K, the value of r is calculated to be
42*9.
For each of the eleven experiments on the distribution of acetic
acid the value of the above expression for the constant K has beea
calculated by inserting the experimental numbers for e^ and c^ the
value for r « 42*9, and that for a obtained from Ostwald's dilution
formula which, in the case of acetic acid at 26**, is ,^^— 0*000018.
' , '1 -a
The following table contains these values of ^ as well as the con-
centrations of the simple and double acetic acid molecules in the
chloroform, these concentrations being expressed in gram^uivalents
per litre :
Concentration of
Concentration of
Concentration of
nndiBsooiated acetio
Bimpla molecnles
in CHCl,.
doable moleculea
K.
acid in water.
in CHCl,.
1-530
0-03565
01920
0-0066
0*9048
0-02109
006796
0*0066
0-6062
0-01411
0*08146
0-0063
0'S156
0-007355
0-008405
0-0064
0-2675
0*006234
0-005986
0*0066
0-2496
0*005816
0-005274
0-0064
0*1929
0*004495
0*003091
0*0066
01574
0 008669
0*001939
0-0069
01254
0*002922
0-001139
0*0076
0-09469
0002207
0*000641
00076
006386
0-001477
0*000263
0-0083
The constancy of the values of K in the last column leads to the oon-
elnsion that the assumption of a gradual dissociation of the double acetic
acid molecules in the chloroform into simple molecules with increasing
Digitized by
Google
ACETIC ACID IN CHLOROFORM SOLUTION. 523
dilution is correct. The maximum and minimum values of K for the
first eight experiments are 0*0069 and 0*0063 respectively. The
limiting concentrations of acetic acid in the chloroform which give
this constant value of K are 0*23 and 0*006 gram-equivalents per
litre, the ratio of these concentrations being approximately 40 to 1.
For the most dilute solutions which have been investigated, the
deviations of K from the mean value are more considerable and increase
with the dilution. At these very small concentrations, however, it
must be remembered that the denominator in the first expression given
above for K^ representing as it does the concentration of the double
acetic add molecules, is the small difference between two gradually
decreasing and approximating quantities, and a small error in the
estimation of the amount of acetic acid in the water and chloroform
layers would explain this gradually increasing value of K> To show
that this is the ca^, we may, for the experimental data at the smallest
concentration investigated, assume the true value of iT to be the mean
of the values found at the higher concentrations, namely, 0*0065, and
calculate inversely the concentration of the acetic acid in the chloroform
layer. We have then in the expression :
^o^«(l-a)« ^ Q.0065,
c^ a- 0-06445, 01 » 0*017, r«42*9, from which the value of e,, the
concentration of the acetic acid in the chloroform, is calculated to be
0*00181, whereas the number found by experiment is 0*00174. The
difference is less than 4 per cent., and since in this experiment less
than 4 c.c. of i\740 sodium hydroxide were required for the titration
of 50 c.c. of the chloroform solution, it is obvious that an error of
about 0*15 C.C. in the titration would completely account for the
discrepancy between the numbers.
It is of interest to note the considerable difference in the relative
proportions of the simple and double acetic acid molecules within the
limits of concentration investigated. At the highest concentration, the
acetic acid present in the form of double molecules is more than
five times as large as that present in the form of simple molecules,
whilst at the lowest concentration the proportion is less than one-fifth.
The value 42*9 calculated for the ratio of distribution of the simple
acetic acid molecules between water and chloroform at 20°, when
compared with the highest ratio of distribution determined experi^
mentally, 37*04, would indicate that at concentrations only slightly
less than the smallest actually investigated, the acetic acid in the
chloroform would consist practically completely of simple molecules.
It was not considered advisable to carry out experiments at higher
concentrations, for the addition of acetic acid to a mixture of chloro*
N N 2
Digitized by
Google
524 DAWSON AND GAWLEB : THE EXISTENCE OF
form and water increases the mutual solubility of these liquids, and
this increase is only terminated by complete miscibility.
A similar dissociation of the doable molecules of acetic acid into
simple ones has been observed in benzene solutions (Nernst, Z$U.phy9%ioeU,
Chenk, 1891,8| 110 ; Hendrizson, ^eii. anorg. Cham, ,1897, 13, 73). Such
results are interesting from the fact that it is generally assumed that
substanced containing the hydrozyl group, when dissolved in liquids
which do not contain this group, are polymerised. It would appear, how-
ever, that such a polymerisation is essentially dependent on the oonoen-
tration, and that if the latter is sufficiently decreased these substances
containing the hydroxyl group will assume the simplest molecular con-
dition. Further interest is attached to the results in view of the recently
expressed opinion (Walden and Centnerszwer, ZeU, phyMal, Ch&tn,,
1902,39, 513) that substances enteringinto solution have quitegeneraliy
the tendency to form polymerised molecules or complex associated mole-
cules in which one or more molecules of the solvent are contained.
It may be noted, finally, that the distribution method is par-
ticularly well suited for the investigation of molecular dissociation
phenomena, which only take place at high degrees of dilution. The
freezing point and boiling point methods, which are convenient for
more concentrated solutions, are in such cases quite useless.
The Torkshius Oollbob,
Lbbds.
LYIL— The Existence -of Polyiodides in Nitrobenzene
Solution. L
By H. M. Dawson and R. Gawlbb.
In the course of some experiments on the ratio of distribution of
iodine between nitrobenzene and a solution of potassium iodide, one of
us obtained some apparently very peculiar results, and the inquiry
into the cause of these abnormalities has furnished the material
contained in the following paper.
The nature of the observations which served as the starting point of
our investigation may be stated very briefly. A series of experiments
was carried out in which a mixture of 20 c.c. of nitrobenzene and 30
c.c. of i^/10 potassium iodide solution was shaken up with gradually
increasing quantities of iodine, and the amount of iodine in each layer
determined by titration with sodium thiosulphate. The numbers
obtained are given below, the concentrations being expressed in grama
per litre :
Digitized by
Google
POLTIODIDES IN NITROBENZENE SOLUTION. I.
525
AqneoQfl solution.
0-3890 gram
0-7117 „
0-9872 „
1073 „
Nitrobsniene.
14-79 grams
30-89 „
49-0 „
56-63 „
Aqueous tolntion.
NitrobanseiM.
1-161 grams
82*83 gruns
1157 „
140-4 „
0-7731 „
168-5
0-7667 „
187-4 „
It will be observed that as the amount of iodine added to the
mixture increases, the iodine concentration in the aqueous solution at
first increases, passes through a maximum, and then decreases. The
only possible explanation of this peculiar phenomenon seemed to be
that the nitrobenzene extracts the potassium iodide from the aqueous
solution, thereby diminishing the solvent power of the latter for
iodine. Distillation of some of the nitrobenzene solution at once
proved the correctness of this assumption, for a considerable quantity
of potassium iodide was left behind.
Experiments were then carried out to ascertain how the quantity of
potassium iodide thus taken up from a given volume of the aqueous
solution by the nitrobenzene depends on the volume of the latter
and on the quantity of iodine added. In the first series, the volume
of the nitrobenzene was constant, namely, 20 cc, this being shaken
up at constant temperature (20^) with 50 cc. of iT/lO potassium iodide
solution after addition of an accurately weighed quantity of iodine
After the attainment of equilibrium, a measured portion of the
aqueous solution was evaporated, and in this manner the quantity of
potassium iodide remaining in the latter was determined. The numbers
obtained are collected in the following table :
Iodine added.
Potassium iodide in
aqneons solution.
Potassium iodide
in nitrobenzene.
1*00 grams
1-50 „
2-00 „
2-50 „
8-00 „
8*50 „
(5-00) „
0*8817 gram
0*5584 „
0*4497 „
0*3561 „
0*2729 „
0-2064 „
0*1549 „
0*0680 „
0*2716 gram
0*8808 ,.
0*4789 „
0*5571 „
0*6286 „
0-6751 „
0-7620 „
In the last experiment, the iodine added to the system was not com-
pletely brought into solution, although the liquids were shaken for a
very considerable time. The numbers given for the potassium iodide
in the nitrobenzene are obtained by subtraction (50 cc. of JV/IO
potassium iodide solution containing 0*83 gram of potassium iodide).
From the above table, it is dear that addition of iodine results
in the transference of potassium iodide from the aqueous solution
Digitized by
Google
626 DAWSON AND QAWLEB : THE SXI8TENGS OF
to the nitrobenzene. As the amount of iodine increaseg, the qnanr
tity of potassium iodide extracted from the aqueous solution also
inoreases, although less rapidly, as is seen by the fact that the first
gram of iodine causes the removal of 0*27 gram, the 8ec<Hid of 0'2O,
and the third only of 0* 16 gram of potassium iodide. The last experiment
shows further that almost the whole of the iodide in the aqueous
solution can thus be removed by the nitrobenaene if a sufficient
quantity of iodine is added.
In a second series of experiments, the same quantity, namely, 2 grams,
of iodine was added to the 60 c.c. of iT/lO potassium iodide solution
(containing 0*83 gram of potassium iodide), whilst the volume of the
nitrobenzene was varied ; the numbers are tabulated below :
Yolnmeof nitrobeiusenein CO. 20. 80. 40. 50.
Oram of potassium iodide in
60 c.c. aqueous solution after
shaking with nitrobenzene... 0*3661 0*3346 0*3220 0*3130
The numbers indicate that the amount of potassium iodide extracted
from the aqueous solution by the addition of a fixed quantity of iodine
increases with the volume of nitrobenzene employed. Tlus increase
is, however, relatively small, and it is evident that the most important
factor in the determination of the amount of potassium iodide extracted
from the aqueous solution is the quantity of added iodine. We may
now give the results of a third series of experiments similar to those
already described, in which^ however, the distribution of the iodine, as
well as that of the potassium iodide between the two liquids, was
determined. Approximately weighed quantities of iodine were added
to a mixture of 20 c.c. of nitrobenzene and 60 o.c. of Njh potassium
iodide solution. After shaking thoroughly at 20% the concentration
of the potassium iodide in the aqueous solution was determined by
evaporation, and that of the iodine in both layers by titration with
sodium thiosulphate. The iodine in the nitrobenzene could be de-
termined quite accurately by means of sodium thiosulphate solution
and starch paste if well shaken up in a stoppered bottle. The numbers
contained in the table on p. 627 represent the concentrations in gram-
molecules per litre.
The concentration of the potassium iodide in the nitrobenzene was
determined by difference, neglecting the changes of volume which
take place in the two solvents ; in the case of the more highly
concentrated solutions, these may, however, attain a considerable
magnitude, and the numbers in the third column are certainly too
low. The table shows clearly that the apparently abnormal varia-
tion of the iodine concentration in the aqueous solution which is
seen in the naiiibe:r9. in tl^e third ^lunin is due to th^ removal of thfi
Digitized by
Google
F0LTI0DIDB9 IN NITBOBBMZENE SOLUTION. I.
627
Iodine added
in grama.
Concen-
tration
Klin
aoneona
aolntion.
Concen-
tration
I, in
aaneona
Botation.
Concen-
tration
Klin
nitrobenxene.
Concen-
tration
Lin
nitrobenxene.
Ratio j^j
in
nitrobenxene.
108
1-71
1-91
2-6
8-08
5-02
7-08
01618
0-1408
01847
01192
0 1087
0-0618
0-08287
0 006188
0-007468
0008474
001
001117
0-01226
001084
0*09648
0-1494
0-1682
0-2019
0-2407
0-8466
0-4178
01866
0-8046
0-8894
0-4411
0-6461
0-8887
1-164
1-964
2-088
2-079
2186
2-268
2-428
2-764
potassium iodide from the aqueoos solution by the nitrobenzene^
whereby the solvent power of the former for iodine is diminished and
that of the latter increased. The experimental observations which
formed the basis of the investigation are therefore in all probability
explained by the great solvent power of nitrobenzene for certain
polyiodides which are formed, this solvent power being even greater
than that of water.
It is perhaps possible to subject the above experimental results to «
more minute analysis. By means of the equation which determines the
equilibrium between the iodide, tri-iodide, and free iodine in the aqueous
solution (Jakowkin, ZeU, physikal. Chem.y 1894, 13, 539 ; 1896, 20,
19 ; Dawson, Trans., 1901, 79, 238), the concentration of the free
iodine in the aqueous solution can be calculated from the potassium
iodide, and total iodine concentrations determined experimentally.
Multiplication of this by the value of the ratio of distribution of iodine
between nitrobenzene and water, which in two experiments at concen-
trations of 18 and 35 grams of iodine per litre of nitrobenzene was
found to be 166*4 and 187*2 respectively, gives us the concentration of
the free iodine present in the nitrobenzene. That portion of iodine
in the nitrobenzene which is in chemical combination with the potass-
ium iodide can thus be determined. On account of the complicated
nature of the relationships, we have, however, refrained from
speculations based on these calculations.
Since the distribution experiments do not give in a simple manner any
definite information in regard to the existence of any particular poly-
iodides in the nitrobenzene solution, we have attempted to ascertain the
nature of these by solubility determinatiqps. Whilst potassium iodide is
practically insoluble in nitrobenzene, a preliminary experiment showed
that if the nitrobenzene contains iodine it readily dissolves a considerable
quantity of potassium iodide. Before such solubility determinations
could be carried out, it was necessary to have a method of analysing the
solutions of iodine and potassium iodide in nitrobenzene. The iodine
Digitized by
Google
623 DAWSON AND QAWLEB ! THE EXISTENCE OF
is readily determined with sodium thiosulphate by shaking in a
stoppered bottle. Several methods were tried for the estimation of the
potassium iodide before a sufficiently accurate and trustworthy process
was discovered. The method finally adopted consists in distilling a
measured volume of the nitrobenzene solution in steam. Ten c.c.
or less of the dark brown solution are introduced into a hard
glass flask, a little water added, and a current of steam passed
through the mixture, when the iodine and nitrobenzene distil over.
In about twenty minutes, the whole of the iodine and practically all
the nitrobenzene have passed over and an aqueous solution of potass-
ium iodide is left in the flask, the solution being usually slightly yellow
in colour owing to a little nitrobenzene which remains dissolved. This
aqueous solution, after cooling, is extracted four or five times with
carefully purified carbon disulphide to remove the nitrobenzene,
evaporated to dryness in a platinum basin on the water-bath, and the
residue gently heated over a free flame. The residue obtained in
this manner is perfectly white and consists of the potassium iodide
contained in the nitrobenzene solution subjected to distillation.
Blank experiments with known quantities of potassium iodide and
variable quantities of iodine and nitrobenzene showed that this method
of analysis is capable of yielding good results.
SclvbUUy of Iodine and Potoisiwn Iodide m Nitrobenzene.
The solubility determinations were made in a stoppered bottle
attached to the circumference of a rotating wheel driven by a small
hot-air engine. The temperature was that of the room, and was there-
fore not very constant ; but for the object of the investigation this
was of no consequence. The time of rotation ranged from ten to thirty
hours. As previously stated, potassium iodide is insoluble in nitro-
benzene, whilst two determinations of the solubility of iodine gave
60*71 and 50*53 grams per litre at 16 — 17°. The mean of these is
50*62, or 0*200 gram^molecule per litre. The experiments in which
solutions were obtained containing both iodine and potassium iodide
can be divided into two series. In the first series, the solubility of
potassium iodide in nitrobenzene containing different quantities of
iodine was investigated, the resulting solutions being saturated with
regard to potassium iodide, but not so with reference to iodine. The
experiments in the second series relate to the solubility of iodine in
nitrobenzene containing potassium iodide, these solutions being
saturated only with regard to iodine. Tables I and II contain the
solubility numbers, the concentrations being expressed in grams and
gramomolecules per litre :
Digitized by
Google
POLTIODIDKS IK MITBOBENZENE SOLUTIOK. I.
629
I. SoluMUy qfPotasswim Iodide in Niiirchenzme containing Iodine.
Concentration of iodine.
Concentration of potassinm
iodide.
Ratiir U^^^
Ki moia.
Grams
Gram-mola.
Grama
Gram-mols.
per litre.
per litre.
per litie.
per litre.
18-02
00710
12-81
0-0741
0-986
29-97
01181
19-71
0-1186
0-994
88*88
01581
25-18
0-1614
101
58-26
0-2296
88-70
0-2881
0*986
71-61
0-2828
45-41
0-2786
1-08
144-2
0-6684
81-98
0-4989
1-16
144-8
0*669
80*88
0*484
1-18
210-8
0-881
111*4
0*671
1*24
411-2
1-621
164-0
0-988
1*64
697-8
2*867
211-9
1-276
1*86
658*8
2-696
221-0
1-881
1*96
695-8
2*748
286-7
1*420
1*93
It is interesting to note what relatively large quantities of iodine and
potassium iodide can be dissolved in nitrobenzene when both of these
substances are present together. Although potassium iodide itself is
insoluble, and iodine only dissolves to the extent of 60 grams per litre,
yet the most concentrated of the above solutions contains about
240 grams of potassium iodide and 700 grams of iodine. The action
of chemical affinities, resulting in the formation of polyiodides, is
obviously the cause of these remarkable solubility phenomena. The
last column in the table furnishes us with a means of ascertaining the
nature of the complex compounds. In the first five experiments, the
molecular ratio of iodine to potassium iodide is constant and equal to
unity ', in other words, at low concentrations one molecule of potassium
iodide dissolves in the nitrobenzene for each molecule of iodine present.
The conclusion to be drawn from this is that the triiodide, KI3, is
formed. The amount of potassium iodide which dissolves when the
nitrobenzene contains more than about 0*3 gram-mol. of iodine per
litre is, however, less than the molecular equivalent, and with increas-
ing iodine concentration the molecular ratio of iodine to potassium
iodide continually increases, attaining a value nearly twice as great as
that which it has in the dilute solutions. There is here undoubted
evidence of the formation of a higher polyiodide (or polyiodides), but
its composition is not determinable from the available data.
The data in the table on p. 630 refer to solutions saturated with regard
to iodine, but notj to potassium iodida Under these circumstances,
it will be observed that a solution can be obtained ;_ containing
Digitized by
Google
530
DAWBON AND QAWLEB : THE BXISTKNCK OF
n. SdvbUUy of Iodine in Nitrobenzene oorUaining Potassium Iodide,
Concentration of iodine.
Conoentiation of potaasiom
iodide.
Eatioii5^
K mola.
Grama
Oram-mola.
Grams
Gram-mola.
per litre.
per litre.
per litre.
per litre.
1127
0-4439
12-86
0-0744
6-97 (8-28)
179-8
0-7066
24-86
01467
4-82 (8*46)
2967
1-166
46-66
0-2746
4-26 (8-62)
4047
1-696
66-66
0-896
404
6117
2-017
82-60
0-4976
406
698-2
2762
116-8
0-698
8-94
861*0
8-866
1481
0-862
8-89
941-8
8718
166-6
0-948
8-94
948-8
8719
166*2
0-986
8-98
approximately 950 grams of iodine and 160 grams of potassium iodide
per litre. On comparing one of these solations with one of those in
the first solubility table containing approximately the same quantity
of iodine, it is found that the corresponding quantities of potassium
iodide are very different. For this purpose, we may take the solutions
containing 700 grams of iodine per litre. When this solution is
saturated with reference to potassium iodide, it contains 236 grams of
the latter per litre ; but when it contains 700 grams of iodine, and is
saturated with regard to this substance, then it contains only
116 grams of potassium iodide. The last column in the table shows
that the molecular ratio of iodine to potassium iodide at first decreases
rapidly with increasing concentration, and then becomes practically
constant. Within the errors of experiment, the value of this
molecular ratio for the last six solubility determinations is equal to
four. It must be pointed out that the theoretical treatment of the
solubility data in this case where the solutions are saturated with
regard to iodine, is not so simple as in the case where they are
saturated with reference to potassium iodide. The reason of this is to
be found in the fact that, whereas potassium iodide is practically in-
soluble in nitrobensene, iodine at the ordinary temperature dissolves to
the extent of 0-2 gram-mol. per litre. Before it is possible, therefore,
to draw conclusions with regard to the nature of the chemical com-
plexes, the formation of which is the cause of the greatly increased
solubility of the iodine, we must take account of the quantity of iodine
which is present in the nitrobenzene solution in the free and un-
combined condition. To do this quantitatively is a matter of some
difficulty. It may safely be said that the free iodine concentration.
Digitized by
Google
P0LTI0DIDE8 IN NITROBENZENE SOLUTION. I. 531
which corresponds to the condition of saturation with regard to solid
iodine, will diminish as the solution becomes more concentrated
with respect to the other components of the solution, namely, the
poljiodides, and in the most concentrated solutions represented in
the previous table is probably much smaller than in pure nitro<
benzene. A continually decreasing quantity must therefore be sub*
tracted from the total iodine concentration as the solution increases
in concentration, in order to obtain the quantity of iodine which
has entered into chemical combination with the potassium iodida
With the data at disposal, it is, however, impossible to determine' the
quantity of iodine which must in this manner be subtracted. The
numbers in brackets in the last column are the values of the
molecular ratio of iodine to potassium iodide obtained by subtracting
from the total iodine concentration the quantity representing the
solubility of iodine in pure nitrobenzene. These numbers are all less
than four, and gradually increase with increasing concentration of
the solution, whilst those representing the ratio of the total iodine to
potassium iodide are greater than four and gradually decrease. It is
therefore possible that a knowledge of the true values of the free
iodine concentration in the various solutions would lead to values of
the molecular ratio of combined iodine to potassium iodide for the
dilute solutions approximately equal to four. Assuming that this is
the true ratio of the chemically combined halogen and potassium
iodide, the concentration of the free iodine in the various solutions
can be calculated from the experimental data. In the first, second,
and third solutions, these concentrations are respectively 0*146, 0*120,
and 0*068 gram-moL per litre, whereas in pure nitrobenzene the
concentration of the saturated solution of iodine is 0*20 gram-mole-
cule per litre. The sequence of these numbers is obviously that
' which could be theoretically anticipated, and their order of magnitude
indicates that the correction which would have to be introduced for
the presence of the free iodine in the more concentrated solutions
becomes almost negligible. The conclusions to be drawn from the
relationships exhibited by the solubility data contained in the previous
table must therefore be based on the following resulta In all the
concentrated solutions the ratio of iodine to potassium iodide is equal
to four. Taking account of the concentration of the free iodine
present in the solution, this is probably the value of the molecular
ratio of combined iodine to potassium iodide in the dilute solutions.
The correction factor for the free iodine becomes very small in the
case of the concentrated solutions, and therefore the yalue of the
ratio of total iodine to potassium iodide is at once a measure of the
ratio of combined iodine to potassium iodide. Four molecules of
iodine are thus found to be chemically combined with one molecule of
Digitized by
Google
532 DAWSON AND QAWLER : THE EXISTENCE OF
potassium iodide, and we conclude that the nitrobenzene solution
contains the polyiodide KIg.
The very concentrated solutions of iodine and potassium iodide in
nitrobenzene which have been obtained by us are viscous, dark-brown,
almost black, liquids. Attempts to isolate the polyiodides by cooling
down the concentrated solutions in a mixture of ice and salt, and by
the addition of other liquids, such as benzene, carbon disulphide, and
carbon tetrachloride, have not been successful. Apparently, the only
effect of the addition of these liquids is to precipitate potassium iodide
from the solution. It has been observed that the most concentrated
solutions are very deliquescent, this property being probably character*
istic of the polyiodides existing in the solution. The crystals of potass-
ium triiodide obtained by Johnson (Trans., 1877, 31, 249) were found
to be extremely deliquescent. In some cases, the specific gravities of
the solutions were determined. The knowledge of the specific gravity,
combined with that of the concentration of the iodine and potassium
iodide in the nitrobenzene solution, completely determines the relative
proportions of the three components. These measurements were made
with the view of ascertaining whether any simple molecular ratio exists
between the nitrobenzene and the other components in the concentrated
solutions; but the function of the nitrobenzene is, apparently,
simply that of a solvent, for the numbers indicate no simple molecular
relationship. The most concentrated solution examined had a sp. gr.
200.
JBkdrioal Conductivity of J^itrob&nzene Solutions containing Iodine and
Fotcusiwn Iodide.
From the high value of the dielectric constant of nitrobenzene,
namely, 36*2 (Turner, Zeit, physikal. Chem., 1900, 86, 403), it might be
expected that this solvent would have a considerable electrolytic dis-
sociating power. This should be the case, at any rate, if nitrobenzene
does not form an exception to the Nemst-Thomson rule. The only
experiments which have been made in this direction are those of
Kahlenberg and Lincoln, who investigated the electrical conductivity
of nitrobenzene solutions containing ferric chloride, stannous chloride,
bismuth chloride, and antimony trichloride. These solutions were
found to have a comparatively small conducting power («/*. Phyeioal
ChMi., 1899, 3, 12).
We have investigated the conductivity of nitrobenzene solutions con-
taining iodine and potassium iodide, and find that the solutions are
remarkably good conductors of electricity. The method employed was
that of Kohlrausch, in which an ordinary Arrhenius conductivity vessel
was used, the resistance capacity of which was 0-1424, and the temper-
ature 18^ Two solutions were prepared each containing approzimaMy
Digitized by
Google
POLYIODID£S IN NirROB£NZ£N£ 80LUTI0N. I.
533
25 grams of potasaium iodide per litre, one of them being saturated
with reference to iodine, the other with reference to potassium iodide.
When the solutions had become saturated, the solid matter was allowed
to settle, and 20 cc of the solution were introduced into the conduct-
ivitj vessel. The influence of dilution on the conductivity was deter-
mined in each case by the successive removal of 10 cc. of the solution
and addition of 10 cc. of nitrobenzene. The experimental results are
given below, the first column giving the concentration of potassium
iodide in gram-molecules per litre (c), the second the specific conduct-
/1000^\
ivity (Z), and the third, the molecular conductivity ^ — - — I, the
conductivities being expressed in terms of the new unit (Kohlrausch and
Holborn, Leitvermogen der Eldctrclyte).
Solution 8<Uur€tted with regard to Iodine,
Composition of original solution: 24*60 grams « 0*1482 gram-mol.
potassium iodide per litre; 175*7;[grams« 0-6926 gram-mol. iodine per
litre.
Molocnlar concentration
of potassium iodide.
Specific condnotiyity.
Molecular conductivity.
lOOOJT
c
01482
00741
0-08706
0*01862
0-002814
0-001662
0*0008476
0*000466
18*99
20*94
22*88
24-67
Solution scUurcUed with regard to Potaeaiwia Iodide.
Composition of original solution : 25*00 grams « 0*1506 gram-mol.
potassium iodide per litre; 38*39 grams »> 0*1513 gram-mol. iodine per
Utre.
of potassium iodide.
Specific conducti?ity.
Molecular conductivity.
0 1606
0*0768
008766
0-01882
0*00941
0D0470
0 00286
0 003161
0 001764
00009276
0-0006014
0 0002662
0 0001898
0*0000726
20*92
28*29
24*68
26-68
28*29
29*61
80-91
Digitized by
Google
534 EXtSTfiKGE OP POLlriODIDES IN NITB0BBK2ENB SOLITTIOK
The values of the molecular conductivity increase regularly with
increasing dilution, but in neither case do these show much indication
of approximating to a limit which would correspond to complete
electrolytic dissociation of the polyiodide. The increase in the
observed molecular conductivity may, however, not only be due simply
to increased electrolytic dissociation of the polyiodides, for it is poesible
that on dilution the components of the solutions undergo dissociation
of a non-electrolytic character. That changes do take place on
dilution is shown by the experiments on the amounts of potassium
iodide extracted from a given volume of aqueous solution by different
quantities of nitrobenzene (p. 526) on addition of a constant quantity
of iodine. Under the specified conditions, the amount of potassium
iodide extracted by 50 c.c. of nitrobenzene is about 8 per cent, greater
than that extracted by 20 cc., which difference is probably due for the
most part to changes undergone by the polyiodide in the nitrobenzene.
A comparison of the conductivities of solutions having the same
potassium iodide concentration, the iodine concentrations of which are,
however, in the ratio of 4*5 : 1 (corresponding solutions in the above
two tables), shows that the solutions containing the smaller quantity of
iodine conduct approximately 10 per cent, better than those containing
the larger quantity. This difference must be partly due to the
difference in the viscosity of the two solutions and is scarcely sufficient
to permit of any conclusions being drawn relatively to the nature of the.
components of the two solutions. Comparing the specific conductivities
of these nitrobenzene solutions with aqueous solutions of binary
potassium salts of the same concentration, we find that the former
conduct approximately one-fifth as well as the latter, and must be
considered as good conductors.
Freezing Faint DetermincUicna,
In our attempts further to elucidate the character of the solutions,
some freezing point determinations were made, the results of which may
be stated here. On account of experimental difficulties in the determina-
tions, no quantitative conclusions can be drawn from the observations.
One portion of nitrobenzene was distilled three times in a vacuum, another
was purified by freezing out three times and distillation in a vacuum,
but in each case the liquid obtained did not give a constant value for the
freezing point in successive determinations. Under these circumstances,
only very approximate measurements could be made. Iodine was dissolved
in the purifiiod nitrobenzene, the resulting solution containing about 40
grams per litre. A portion of this iodine solution was then shaken
with potassium iodide until saturation was attained, and the solution thus
obtained containing potassium iodide and iodine in molecular proportion
Digitized by
Google
THS SLOW OXIDATION O^ MBtHAKS AT LOW TEMPSEATUBES. 635
was poured off from the excess of solid matter. The concentration of
the iodine and of the potassium iodide in this solution was 0*15 gram-
tnoleoule per litre. The freezing points of the pure nitrobenxene, the
iodine solution, and the iodine-potassium iodide solution were then
measured. The difference between the freezing points of the second
and third solutions, taking the mean result of two series of deter-
minationS) was found to be about four^fifths of the differeoce between
the freezing points of the pure nitrobenzene and the iodine solution.
This relationship corresponds to a considerable dissociation of the
triiodide which we suppose to be present in the solution.
Although these measurements do not completely determine the
nature of these interesting solutions, yet the solubility deter-
minations indicate clearly the existence of complex periodides in
the nitrobenzene solution. The high electrical conducting power
points to electrolytic dissociation of these complex compounds existing
in the solution, and the freezing point measurements indicate that
dissociation takes place to a very considerable extent.
Tbb Yorkshire Collsor,
Lbxds.
LVIIL- — The Slow Oxidation of Methane at Low
Temperatures.
By WiLUAM A. Bone and Richard V. Whbblbb.
Tbx mode in which a hydrocarbon burns in a supply of oxygen in-
sufficient to completely oxidise it to carbon monoxide (or dioxide) and
steam has been the subject of much controversy. The view, at one
time generally held, that under such conditions the hydrogen bums in
preference to the carbon, can hardly be reconciled with the results of
experiments on the explosion of ethylene or acetylene with less than
an equal volume of oxygen (Lean and Bone, Trans., 1892, 61, 873 ;
Bone and Cain, Trans., 1897, 71, 26). These results indicate that the
main reaction in the explosion wave may be represented by such
equations as
CjH^ + Oj = 2C0 + 2Hj.
OjH, + On = 2C0 + Hj.
From these and other similar experiments, it is sometimes argued
that in a limited oxygen supply the carbon of a hydrocarbon bums
preferentially to the hydrogen; it should, however, be pointed out
that the evidence supporting this contention is wholly derived from
Digitized by
Google
536 BONB AND WHBELER: THE SLOW OXIDATION OF
inyestigations of the oxidation of hydrocarbons at the high temperatnreB
of the explosion wave. The present paper records the results of experi-
ments on the slow combustion of methane in an oxygen supply just
sufficient to oxidise the carbon to carbon monoxide, at temperatures
where the velocity of the reaction is just appreciable. This line of
inquiry is, we believe, entirely new.
The reasons for the selection of methane as the most suitable hydro-
carbon for these studies were as follows : (1) it is the simplest saturated
hydrocarbon, and its molecule contains on^ly one carbon atom, (2) pre-
liminary experiments showed that it. can be. maintained at 480° to 500°
(temperatures considerably higher than those afterwards employed in
the oxidation studies) for many days without undergoing the least
change, and (3) a mixture of two volumes of methane with on^ of
oxygen is non-explosive. This was the mixtui^ used throughout our
experiments.
In a preliminary series of experiments, the object of which was to
determine the most suitable experimental methods for the inquiry,
the mixture of methane (2 vols.) and oxygen (1 vol.) was circulated
through a tube containing fragments of unglazed porcelain maintained
at a constant temperature (between 400 and 450°) in a Lothar Meyer
furnace. The apparatus involved, however, proved too complicated for
experiments which necessarily extended over several days continuously ;
the results, it may be stated, indicated that a portion of the methane
was burnt to carbon monoxide, carbon dioxide, and steam without
any liberation of free hydrogen or carbon. We also tried maintaining,
the mixture at 300° to 350° in contabt with palladium Uack, but the
* catalytic ' effect of the metal introduced complications which made it
difficult for us to follow the real cause of the reaction.*
We finally resorted to the simple expedient of sealing the mixtures of
methane and oxygen, under atmospheric pressure, in cylindrical bulbs
of boro-silicate glass with capillary ends ; the bulbs were afterwards
maintained at constant temperatures (between 300° and 400°) for
several days in an air-bath, until the whole or a part of the oxygen had
disappeared. The cooled bulbs were subsequently opened under mercury,
any change in volume {always a contraction) noted, and the residual
gas withdrawn for analysis. We were thus able to examine the gases
at various periods during the oxidation of the methane at any given
temperature, and we could hardly fail to detect the formation of a
product at any stage of the oxidation which afterwards disappeared
before the process was completed.
The interaction of two such gases as methane and oxygen in glass
vessels at low temperatures being a surface phenomenon, the temperature
* We desire to acknowledge oar indebtedness to Hr. John Wilson, of the Batter-
sea Polytechnic, for mach valuable help in these preliminaiy experiments.
Digitized by
Google
METHANE AT LOW TEMPBRATUBE8. 537
at which its velocity is jast appreciable, as well as the yelociiy at any
other given temperature, will, to a certain extent, depend on the character
of the surface. In our experiments, 300° was the lowest temperature at
whioh any interaction could be detected after the lapse of two or three
weeks. At 325°, however, the velocity was much greater ; at 350°, in
some instances, the whole of the oxygen disappeared within three or
four days, whilst at 400° the oxidation was always completed in a single
day. But we have so far been unable to measure the relative velocities
of the reaction at different temperatures on account of the fact that,
except at 400°, when the velocity is considerable, the influence of the
' surface factor ' may be very different even in two bulbs of the same
size, shape, and material. Thus, between 325° and 350°, an appreciable
amount of oxidation always occurred within three days, but of a series
of similar bulbs filled with the same mixture and heated in the same
bath, some exhibited a greater amount of oxidation in two or three
days than others did in a week or more. Nor does the 'surface
factor ' of a given bulb remain constant over two or three successive
experiments with the same mixture ; whether it would finally become
constant after a long series of experiments is a point we are now
investigating.
It is of course impossible, by any means at our command, to determine
the precise manner in which the methaiie molecule is attacked by the
oxygen, but we are able to discover what is the first distinct stage of
the reaction. This first stage may, obviously, involve one of three
things, namely :
(1) Selective combustion of hydrogen, thus,
(a) OH^ + Og = O + 2H3O; or
{b) 2CH^ + 02 =* 2a + 2HjO + Hj,
(2) Selective combustion of carbon, thus,
(a) 2CH^ + 0j - 2C0 +4Hj;or
(6) OH^ + 02 - CO2 + 2HS.
(3) Simultaneous oxidation of carbon and hydrogen, thus,
(a) CH4+ O2 « CO +H2O + H2; or
(6)»2CH4 + 30, = 2C0 +4H,0; or
(c) CH4 + 2O3 - COg + 2H30.
The primary oxidation products (which it will be seen may include
carbon, hydrogen, carbon monoxide, carbon dioxide, and steam) may,
however, react with each other, or with the original constituents of the
mixture. We had therefore, as an important part of our inquiry, to
investigate the possibilities of such reactions occurring in our bulbs at
temperatures between 300° and 400°. We will now briefly discuss the
evidence so obtained.
VOL. LXXXI. O O
Digitized by
Google
538 BONE AND WHEELER: THE SLOW OXIDATION OF
The possible secondary reactions may be classified as follows :
(A), those in which free carbon may be involved, for example :
(1) C + HgO-CO + Hj; (2) 00^ + 0-200; (3) 20 + 0,-200.
The evidence obtained shows that none of these reactions begin at
temperatures between 300^ and 400°.
(B) OO + HjjO ;=l OOj + Hj.
A series of experiments with bulbs similar to those used in the
oxidation studies have shown that steam and carbon monoxide can be
maintained at 325° for a fortnight without the slightest change occur-
ring. At 350°, no action could be detected within a week, but after
ten days some 1 '7 per cent, of carbon dioxide had been formed ; at
400° about the same amount of change occurred in a week. On the
other hand, mixtures of equal volumes of hydrogen and carbon dioxide
showed no signs of change when kept at 325° or 350° for a fortnight,
or at 400° for a week.
The results of these experiments show that no complication arising
from the interaction OO + H^O ^^ OOg + H, enters into any of our
experiments on the oxidation of methane.
(C) 20O + O2-2COa
We have found that moderately dry carbon monoxide and oxygen do
not react between 300° and 400° ; the formation of between 0*7 and
r7 per cent, of carbon dioxide could usually be detected when the maiti
gases were maintained at 325°, 350°, or 400° for a week. The effects
of this possible secondary change in the methane experiments are
therefore practically negligible.
(2)) 2H2 + Og=-2H,0.
In 1895, Y. Meyer and Raum {Ber., 28, 2804) published the
results of an investigation on the combination of the elements of
electrolytic gas in glass bulbs, very similar to those used by us, at
temperatures between 300° and 518°. At 300°, the formation of water
could just be detected after 65 days ; in the case of four out of five
bulbs maintained at 350° for 5 days, a very small amount of combina-
tion occurred (between 0*5 and 1'9 per cent, ouly of the original gas
had disappeared), whilst in the fifth bulb as much as 16*4 per cent, of
the gases had combined.* At 448°, the combination was still very
slow. We have carefully repeated these experiments, using bulbs
which had previously been employed for heating the mixtures of
* Judging from onr own experliaents at this temperature, we are indined to
attribute this relatively large formation of water to some roughness of the inner
surface of the bulb used either present originally or caused by a partial de?itrifica
tion of the glass during the heatiug<
Digitized by
Google
METHANE AT LOW TEMPE^TURES. 539
methane and oxygen. At 335°^ we have never been able to detect the
slightest formation of water from electrolytic gas within a week ; at
350^, in six bulbs no combination took place in a week, although in the
case of a seventh bulb, in which the glass had become devitrified at one
end, the formation of water could be distinctly seen. At 400°, three
bulbs exhibited no signs of change after three days ; after a week,
water could be distinctly seen in one bulb (nearly 40 per cent, of the
mixture had combined), but none could be detected in the other two.
Professor Dixon informs us that some years ago he maintained glass
bulbs filled with electrolytic gas at 350^ for several weeks, but was
unable to detect any formation of water. Our own experience shows
that hydrogen and oxygen, even when mixed in combining proportions
and undiluted with other gases, do not within a week or two combine
at 350^ to any appreciable extent provided the glass surface with
which the gases are in contact remains perfectly smooth. At 400^,
however, we are on the border line where the formation of water may
occasionally be recognised within a week, but hardly within three
days; in our experiments with methane at this temperature, the
whole of the oxygen always disappeared within a single day.
(E) We have also found that the following pairs of gases have no
mutual action at temperatures between 350° and 400° (within a week
or two). Methane and carbon dioxide ; methane and steam ; carbon
monoxide and hydrogen.
It may therefore be taken for granted that no appreciable com-
plication arising from possible secondary changes enters into our
experiments on the oxidation of methane, the results of which may now
be discussed.
We find that between 300° and 400° ioiethane combines with
oxygen with an enormously greater velocity then does hydrogen
itself under the. same conditions. We have followed the course of
the oxidation in at least thirty cases, and although our mixtures Jiever
contained more than 1 volume of oxygen to 2 volumes of methane,
in no due hetve toe been akle to detect the liberation of free hydrogen or
free carbon at any atage^ from beginning to end, qf the process. The
disappearance of oxygen was always accompanied, in the cooled pro-
ducts, by a corresponding diminution in volume, due to the formation
of water. This formation of water could always be detected even
in the initial stages of an experiment, when less than one-tenth of
the oxygen originally present had disappeared; the amount of it
increased as the oxidation proceeded, until the inner surface of the
cooled bulb was completely wetted. Since these phenomena occurred
at temperatures below that at which the elements of electrolytic gas
combine (in similar bulbs) witk any appreciable velocity, and since
earbon dioxide and hydrogen do not t«a<st eten at 400°, the natural
e o 2
Digitized by
Google
540 BONE AND WHEELER: THE SLOW OXIDATION OF
inference is that tocUer is one of the primary products qf the partial
oxidation qf methane at these temperatures.
At all stages of the oxidation, the gases in the bulbs contained
(besides water yapour) carbon monoxide, carbon dioxide, unchanged
methane, and, in some cases of course, oxygen. At no period baye we
been able to detect the formation of such products as methyl alcohol,
formaldehyde, or formic acid; the carbon of that portion of the
methane burnt always appeared in the products as carbon monoxide
or dioxide. There was, however, no regularity in the ratio GO, : GO
at any given period of the oxidation, except towards the end ; during
the earlier period, it usually varied between 1 : 7 and 1 : 3, but finally
it approximated to a value between 1 : 2 and 1:1. Since the influence
of the reaction GO + H^O^GOg + H, between 300"^ and 400^ has been
shown to be negligible, and that methane and carbon dioxide have no
mutual action, our experiments point to the conclusion that tits first
stage in the "partial" combustion qf mathane at low temperatures is a
simultaneous oxidation of carbon and hydrogen to carbon monoxide and
steam, thus :
2CH^ + 30j«2GO + 4H30.
One curious feature about our results is the unexpectedly large pro-
portion of carbon dioxide found in the gases at each stage, but especially
towards the end of the oxidation. Now the rate at which carbon
monoxide combines with oxygen, or reacts with steam, between 300^
and 400° has been shown to be negligible ; therefore we cannot ex-
plain the formation of any considerable quantity of carbon dioxide in
an experiment on the supposition that carbon monoxide actually
liberated during the primary oxidation is afterwards gradually trans-
formed into the dioxide through the agency of steam and oxygen, or
steam alone. The largest proportions of carbon dioxide were found in
bulbs in which the mean rate of oxidation had been, for any particular
temperature, fast ; the two circumstances are, we are inclined to think,
not unconnected. Further, we should perhaps state that our experi-
ments suggest that the ' oxidation velocity ' in a given case is acceler-
ated after the first portions of the oxygen have disappeared, but this
is a point which requires fuller investigation. We are inclined to
take the following view of the formation of so much carbon dioxide.
When the methane molecule is burnt, carbon monoxide and steam
simultaneously come into being in an atmosphere containing oxygen.
At the moment of their formation, these new molecules would be in
an extremely labile and reactive condition, and it is probable that
during this transitory ' labile period ' a much more frequent exchange
CO I OS. I
of oxygen would occur in the system qq qd* I 0,, than under ordin-
ary circumstances. It is also conceivable that the ' lability ' of such
Digitized by
Google
METHANE AT LOW TEMPERATURES.
541
a syBtem would be influenced by the rate at which the methane
succumbs to the attack of the oxygen.
£ZPBBIMBNTAL.
I. — Thi Freparcaian o/Fure Methans.
The preparation of pure methane by Qladstone and Tribe's method
is always a tedious process ; if, however, an aluminium-mercury couple
be substituted for the zinc-copper pair, the result is much more satisfac-
tory. In fact, the action of the aluminium-mercury couple on a mixture
of methyl iodide and alcohol is so energetic that the reaction vessel must
be well cooled by immersion in iced water at the outset of the pre-
paration, or otherwise the evolution of the gas becomes too rapid, and
is difficult to control. In addition to being contaminated with the
vapours of the alcohol and iodide, the gas almost invariably contains
a small quantity (2 or 3 per cent.) of hydrogen. This may be easily
Fio. 1.
removed, after the other Impurities have been eliminated, in one of
two ways, namely : (1) by passing the gas through a layer of ' oxidised '
palladium black at 100% or (2) by liquefaction of the methane in a
bulb immersed in liquid air, when, of course, the hydrogen passes on.
The details of the method are as follows :
The central bulb of the Wohler U-tube A (Fig. 1) is filled with
clean aluminium foil, cut into pieces about one-eighth of an inch square.
A few cc. of a mercuric chloride solution are poured on to the foil
which, in the course of a minute or two, becomes coated with mercury.
The liquid is then drained off, and the couple rapidly washed (twice)
with methyl alcohoL About 30 grams of methyl iodide are then
poured on to the couple. The tap funnel, B, containing a mixture of
2 to 3 parts by weight of methyl iodide and one part of methyl alcohol,
is quickly inserted in one limb of the U-tube by means of a rubber
cork. The other limb is immediately connected with the arrangement
for washing the gas, consisting of (1) the worm 0 surrounded by iced
Digitized by
Google
(8).
(8).
(4).
44-40
66-26
67-0
89-2
130-0
112-0
44-6
66-3
66-3
542 BONE AKD wheeler: THE SLOW OXIDATION Of
water, (2) worms D and E containing a atrong solution of aodiam
methozide in methyl alcohol, and (3) worm F containing strong
sulphoric acid. The lower part of A should be immersed in iced
water ; on dropping the mixture from B on to the couple, a good
evolution of methane is obtained. The air is swept out of the appar-
atus through the vertical branch of the three-way tap, G, before con-
nection with the apparatus for the elimination of hydrogen is made.
When oxidised palladium sponge is used for this purpose, it ia placed
in the small U-tube, H, wliich is immersed in boiling water. The gas
is afterwards dried by passing it through the worm, K, containing
sulphuric acid. Where palladium sponge is not available, the removal
of traces of hydrogen can be effected by means of liquid air, the
methane condensing to a colourless liquid as fast as it is evolved.
We have frequently tested the purity of the gas so prepared by
explosion analyses ; the following are the results in the cases of four
typical samples :
(1).
Yohime of gas taken ... 58*00
Contraction 0 115-8
Absorption A 57*95
n. !l%e Action qf Eeat an Methane.
Since in the subsequent 'oxidation' experiments, mixtures of
methane and oxygen were maintained at temperatures between 300°
and 400° for, in son^ cases, as long as two or three weeks together, it
was necessary to ascertain whether methane itself undergoes any
change at temperatures at all near these limits. Accordingly, about
3 litres of pure methane were kept continuously circulating for six
days and nights through a hard glass tube (about 0*75 metre long),
packed with fragments of well-dried, unglazed porcelain, maintained
at 480° in a Lothar Meyer furnace. In the circuit was a glass spiral,
which, during the experiment, was kept surrounded by a freezing
mixture. On examining the tube after the experiment, no blackening,
or even discoloration of the porcelain, could be detected, nor could any
liquid be seen in the glass spiral The volume of the methane remained
constant throughout the experiment, and analysis showed that it had
undergone no change. Further experiments showed that only at
650 — 700° does methane begin to decompose, and even then very
slowly.
III. — The Oxidatian EsopervmmJte,
The mixtures of methane and oxygen (the latter prepared by heating
recrystallised potassium chlorate in hard glass bulbs, and afterwards
washing it through a strong solution of potassium hydroxide) were
Digitized by
Google
MIfiTHAKE At LOW TEMPS&ATl^RES. 643
tnade in a graduated glass holder over pure strong sulphuric aoid.
The methane in each case was determined analytically hy an explosion
method, and the oxygen by absorption with strongly alkaline pyro-
gallol (freshly prepared). Altogether five mixtures were used ; the
percentage composition of each, leaving altogether out of account the
small amount of nitrogen present (namely, 1*0, 2*16, 0*95, 3*6, and
2-67 per cent, respectively),* is given below :
Mixture. A. B. C. D. E.
Methane 69*0 66*66 67*44 66*80 66*90
Oxygen 31*0 33*34 32*56 32*20 33*10
Filling of the Bulbs, — ^The cylindrical Jena boro-silicate glass bulbs
used in these experiments terminated at each end in a long capillary
tube (1 mm. bore). They had a capacity of between 60 and 70 c.c.
(length about 10 cm., diameter 3*5 cm.), except a few, somewhat
larger (capacity 70 to 80 cc), used during, the later stages of the
research. Boro-silicate glass is eminently suited for the making of
vessels in which gases are to be heated for long periods under pressure ;
in our own experiments the bulbs, filled at atmospheric pressure and
temperature, could be maintained at 400^ for, if necessary, many
weeks without showing change in shape or capacity, or signs of devitri-
fication. Similar bulbs made of ordinary soft or hard glass generally
devitrified or burst when subjected to the same treatment. Further,
boro-silicate glass possesses another great advantage over other kinds
in that it withstands sudden changes of temperature.
Before being filled with the mixture under investigation, the inner
surface of each bulb was thoroughly cleaned with hot strong nitric
acid, and afterwards with distilled water. Each was subsequently
dried in a current of hot air (dust free). A number of such clean, dry
bulbs were connected in series, on the one hand, with the holder
containing the mixture of methane and oxygen, and, on t^he other,
through a drying tube to a Sprengel pump and manometer. When
nearly vacuous, each bulb was strongly heated with a Bunsen burner
and tiie exhaustion completed.. As soon as the bulbs were cold, the
mixture was admitted from the holder until the pressure in the appar-
atus was 2 — 3 mm. below the atmospheric; the capillaries of each
bulb were then successively drawn out and sealed off in the blow-pipe.
ffecUing of the Bulbs, — ^The bulbs were heated, in batches of from
2 to 6 as required, in a special air-oven, the gas supply of which
passed through a Stott's governor, and then through a Lothar Meyer
regulator.. The temperature was registered by a thermometer reading
* In Older to make the results of the yarious experiments strictly comparable, we
propose to adopt this coarse throughout the paper in tabulating the compositiqn of
the various gaseous mixtures under discussion. Since the nitrogen in each case was
taken ' by diff!erenee,* the tabulated results will always add up to 100.
Digitized by
Google
644 BONE AND WHEELER: THE SLOW OXIDATION OF
up to 500^, and we had no diffictiltj in keeping it practicallj constant
for weeks together. The extreme yariations on either side of the
desired mean temperature were about 5° in the longest experiments ;
in the case of experiments extending over a few days only, the varia-
tions in temperature were even less.
JSxaminattan qf the Bvlhs qfur HeaUng, — As soon as each bulb was
withdrawn from the air-bath, it was quickly cooled in water. The
most superficial examination of the cold bulbs revealed two significant
facts connected with the partial oxidation of the methane, namely, (1)
no carbon had been deposited, and (2) the formation of a dew on the
inner surface of the vessel. On standing suc^ a bulb in a vertical
position, the dew soon collected in the capillary as a colourless liquid.
The liquid was tasteless and odour-
^^®- 2- less, it did not contain hydrogen per-
oxide, formaldehyde, or formic acid;
it solidified in a freezing mixture, and
the solid melted at 0°. It was, in
short, water.
The bulbs were subsequently opened
under mercury, and the gases with-
drawn as follows :
A deep scratch having been made
near the end of one of the capillaries
of the bulb A (Fig. 2), the glass was
nipped off under the surface of mercury
contained in the cylinder, B. In each
case, the mercury entered the bulb,
partially filling it. After the lapse of
half-an-hour, the level of the mercury
inside the bulb (a) was marked (the
levels inside and outside having been
previously equalised). The open end
of the capillary was then forced into a stout rubber joint connected
with the U-shaped capillary tube, C; the whole of this rubber
joint and capillary had been previously filled with mercury as far
as the tap D, through which connection was made with a Sprengel
pump. All connections having been thoroughly exhausted, the tap, D,
was opened, and the gas in A drawn off through the pump into tubes
over mercury. Finally, the bulb, A, was detached, and its total
volume, as well as the volume up to the mark a, determined. In this
way, the contraction, due to the formation of water during the
oxidation, was approximately measured ; in all cases, it corresponded
to the volume of oxygen which subsequent analysis showed had
disappeared.
Digitized by
Google
METHANK AT LOW TEMPERATURES. 545
AncdysU of the Residual Gaeee. — ^The analysis of the residual gases
coDstituted the most important part of the work, for on their accuracy
depends the proof that no free hydrogen was formed during the coarse
of the oxidation. The apparatus used was that described by one of us
at a meeting of the Society in 1898 (Proc., 1898, 14, 154), and a long
experience of its working has shown that it admits of a high degree of
accuracy.
In addition to a large amount of unchanged methane, the gases
always contained carbon dioxide, carbon monoxide, and in some cases
oxygen also. The last three were removed and estimated in the
following order, namely (1), carbon dioxide by means of a strong
solution of potassium hydroxide, (2) oxygen by means of a freshly
prepared and strongly alkaline solution of pyrogallol, and (3) carbon
monoxide by means of freshly prepared ammoniacal cuprous chloride,
the gas being afterwards treated with dilute sulphuric acid before
remeasurement. When the gases contained more than 5 per cent, of
carbon monoxide, they were subjected to a second treatment with a
fresh portion of the cuprous chloride solution. A series of trial
experiments satisfied us that practically the whole of the carbon
monoxide in such a mixture can be removed in this way ; certainly
never as much as 0*5 per cent, remains unabsorbed.
After removal of the three foregoing constituents, a measured
portion of the residual gas was exploded with a large excess of oxygen,
and the contraction in volume (C7), and the absorption {A) when the
products of explosion were treated with potassium hydroxide, deter-
mined. It is essential to the accuracy of such an analysis that the
explosive mixture (CH^-f 20}) should be largely diluted with excess
of air, in order to avoid the oxidation of any nitrogen present, or
deposition of carbon in the explosion vessel due to shock. In the case
of these mixtures, we always added at least 100 volumes of diluents to
every 50 to 60 volumes of the explosive mixture, and exploded the
gases under half an atmosphere pressure. These precautions ensured
the necessary conditions of accuracy.
n
From the ratio -j obtained in any case, we were able to determine
whether the gases exploded contained hydrogen in addition to methane.
This ratio for pure methane is, of course, 2*0, and for mixtures of
methane and hydrogen would be greater (thus it would be 2015
for a gas containing 99 per cent, of methane, and 1 per cent, of
hydrogen, and trial experiments have shown that this quantity of
hydrogen can be detected with certainty). An examination of the
results of twenty-five different experiments detailed in the following
paragraphs will show that in sixteen cases the ratio fell between 1*99
and 2-00 ; in seven other cases, it lay between 1*95 and 1-99, whilst in
Digitized by
Google
546 BONE AND WHEKLSR: THE SLOW OXIDATION Of
two cues only (2*007 and 2*005 respectiyely) was it higher )^han 2*00.
These numbers, therefore, prove the absence of free hydrogen in the
oxidation products at all stages of the process.
Experiments at 300^
The rate at which methane and oxygen combine at this temperature
is very slow, so that it is hardly ever possible to detect any change
unless the heating be continued over two or three weeks. The results
of two experiments, in both of which the formation of water could
be distinctly seen, are given below. The mixture employed was A
(methane = 69 0 ; oxygen = 31 '0) :
Composition of dry gases after
U days. 21 days.
Carbon dioxide 0-70 240
Carbon monoxide 3*85 6*40
Oxygen 2710 2140
Methane 6835 6980
-J for residual gas exploded 1*97
2-00
Experiments at 325°.
First Series. — Four bulbs filled, at 18^ and 758 mm., with mixture
D (methane a 66 '8; oxygen •» 33 2) and heated for 30 hours, 3, 6, and
1 1 days respectively. All showed the formation of water ; it will be
seen that more oxygen had disappeared in the bulb heated 3 days than
in the one in which the action had continued over 6 days. The
analytical results are given below :
Duration of heating. 80 hoars. 8 days. 6 days. 11 days.
Carbon dioxide nil 1 66 0*90 230
Carbon monoxide 107 7 24 435 7-45
Oxygen 3160 2410 2845 2325
Methane 6733 6700 6630 67*00
^ for residual gas 1 ^,^ g.^^ ^.^ 2^
A exploded J
Second Series. — Four bulbs filled, at 21° and 755 mm., with mix-
ture E (methane a 66-9 ; oxygen* 33*1) and heated for 3, 7, 14, and
21 days respectively. Singularly enough, by far the greatest amount
of interaction occurred in the bulb heated for 3 days ; indeed, the
order according to 'rate of oxidation' is nearly the reverse of that
Digitized by
Google
METHANE AT LOW TEHPERATUBES. 647
according to the duration of the heating. These experiments afford a
good illustration of the point emphasised in the introduction, namely,
that the influence of ' surface factor/ in the case of several exactly
similar vessels, upon the velocity of a gaseous reaction is often enor-
mously different. The analytical results are as follows :
Daration of heating. 21 days. 14 daya 6 days. 8 days.
Carbon dioxide 2-07 1-40 2-00 12-86
Carbon monoxide 5-60 600 616 1630
Oxygen 26-60 26-70 2410 (2-00)»t
Methane. : 66-83 67-90 67-76 (68-86)
2 ^^'Sir^.^l 2*^^ ^'^^^ ^'^^^ ^'^^^
JSxperimmta at 360°.
First Seris8.— Three bulbs filled with mixture E (methane » 66-9 ;
oxygen » 33'l)at 22° and 747 mm.,and heated for 24 hours, 66 hours, and
a week respectively. Analyses of residual gases gave following results :
Duratioii of heating. 24 hours. 66 hours. 7 days.
Carbon dioxide 0-36 0-38 14-0
Carbon monoxide 2-06 2-90 16-8
Oxygen 31-38 3036 09
Methane 66-22 66-37 68-8t
Q
J- for residual gas exploded 1-98 1*99 2*00
Second /SMw.— Three bulbs filled with mixture E at 21° and
766 mm., and heated for 1, 3, and 7 days respectively. We would
draw attention to the fact that the bulb heated 7 days exhibited a
much smaller amount of oxidation than the one heated for the same
time in the previous experiment. The analytical results for the
residual gases were as follows :
Daration of heating. 1 day. 8 days. 7 days.
Carbon dioxide 0-40 2*80 2-26
Carbon monoxide 216 8*96 4-90
Oxygen 31-40 2180 26-60
Methane 6606 6646 66-26
-jT for residual gas exploded 200 2-00 1 -996
* TMs number for oxygen is only approximate ; the same therefore applies to
that given for methane. t See footnote, p 548.
Digitized by
Google
548 THE SLOW OXIDATION OF METHAKB AT LOW TBMPERATURCS.
ThirdySeriss.Six balbs filled with mixture B (methane -66 -6 ;
oxygen » 33*3) maintained at 350^ for 13 days. In each case, practi-
cally the whole of the oxygen had disappeared, and a large quantity
of water was produced. The contractions^ on opening tbe bulbs under
mercury, amount to between 30 and 33 per cent, of the original
volume. The products of four of the bulbs were analysed, with the
following results ; attention is drawn to the large quantity of carbon
dioxide produced in each case :
(1). (2). (8). (4).
Carbon dioxide 14-8 10'2 10-0 lOO
Carbon monoxide 160 23*6 21*5 22*2
Oxygen nil nil 0*3 nil
•Methane 692 662 68-2 678
C
-J for residual gas exploded 2*007 1*994 2-006 1-966
JSa^oerimmte at 400"".
At this temperature, methane is rapidly oxidised, and differenoea
due to surface factor in a series of similar bulbs tend to disappear.
We have always found that the oxygen in our mixtures disappeared
within a single day, and that the phenomena were altogether more
regular than at lower temperatures. We would in this connection
again point out that we have maintained bulbs containing electro-
lytic gas at 400° for 3 days, and in many cases a week, without
finding any appreciable formation of water. The following are the
analytical results from an experiment in which five bulbs filled
with mixture C (methane » 67*44 ; oxygen « 32*56) at 20"^ and 750 mm.
were heated for If 2, 4, and 11 days (two bulbs) respectively. The
contraction observed on opening these bulbs under mercury amounted
to as nearly as possible one-third of the original volume :
Duration of heating. 1 day. 2 days. A days. 11 days. 11 days.
Carbon dioxide 11-4 12-4 113 106 11-9
Carbon monoxide 18-4 170 17-6 18-9 18-6
♦Methane 70*2 70-6 71-1 70-6 69-6
ft
jfor residual gas exploded 196 1-99 2*00 2-00 2-00
* The percentage of methane in these products, and also in nearly all other
experiments in which a large proportion (over 10 per cent.) of carbon dioxide was
produced, is rather higher than it should be. We have reason to believe tihat thia
is caused by a small absorption of carbon dioxide by the alkali in the glass.
Digitized by
Google
DERIVATIVES OF a-AMINOCAMPHOROXIME. 549
We are continuing this work and extending it to other hydro-
carbons.
In condasion, we desire to state that the expenses of this research
have been defrayed out of grants received from the Government Qrant
Committee of the Royal Society.
Thb Owbns OoLusas,
Manohxstxr.
LIX. — Derivatives of a-Aminocamphoroxime.
By Abthub Lapwobth and Alfred William Habvet.
It is a point of some theoretical interest that most a-substituted camphors
are not capable of affording oximes, although when the substituents are
present in other positions, that inability is not observed. Hitherto,
exceptions to this general rule have been noticed with a-hydroxy camphor
(Mana8se,^«r., 1897, 30, 668), a-tsonitrosocamphor {Ber., 1894,06, 243),
and Beychler's camphorsulphonic acid {Bvll. Soo. Ghxm,^ 1898, [iii], 19,
120). Substituted camphoroximes have also been prepared from ir-bromo-
camphor (Kipping and Revis, Proc., 1896, 12, 77) and from )3-bromo-
and )3-chloro-camphor (Forster, frans, 1902, 81, 268).
We have found that oraminocamphor constitutes another exception
to the above-mentioned rule, as it yields an oxime without difficulty if
proper precautions are taken.
The study of camphoroxime, which has engaged the attention of
several chemists, has presented points of more than ordinary interest
in consequence of an extremely labile character of the molecule which
the substituted camphoroximes do not appear to possess. a-Amino-
camphoroxime, however, might be expected to exhibit certain peculiar-
ities of its own, more especially as the oximes of the type ^'(^v-.ryET
are known to yield a variety of interesting ring compounds on treat-
ment with anhydrides, aliphatic aldehydes, d^s.
a-Aminocamphoroximeis much more susceptible to alkaline hydrolysis
than is camphoroxime itself, and is not easily prepared from a-amino-
camphor by the use of an alkaline solution of hydroxylamine. By
employing hydroxylamine acetate in presence of a large excess of
or a ' condensatioii ' of carbon dioxide on the glass surface. Such an absorp-
tioB, or condensation, would make the methane appear proporlUmaUly higher
in the products, since it was \jk all cases taken as the '.difference ' between the total
volume of (nitrogen free) gas analysed, and the sum of the carbon dioxide, carbon
monoxide, and oxygen found.
Digitized by
Google
550 LAPWORTH AND HARVET :
sodium acetate, however, the oxime may be obtained without difficulty
in large amount.
The Bubstanoe has the properties both of an amine and of an oxime ;
it dissolves readily in dilate acids and in excess of alkalis, forms definite
salts with adds, and yields crystalline metallic derivatives. It affords a
beautiful, crystalline dibenzoyl derivative, CgHj^'O^.^ , and
reacts with one mol. of potassium cyanate to give a well-defined
carbamide. In accordance with the fact that it contains both an amino-
and a hydroxyl group, it combines with two mols. of phenylcarbimide.
When warmed with benzaldebyde, the amino-group reacts in the
normal manner and a monobensylidene derivative,
CH-N:CHPh
is produced.
Aminocamphoroxjme shows no tendency to become converted into a
campholenonitrile (compare Forster, Trans., 1902, 81, 268), and on
treatment with acetic anhydride is converted into a well-defined
crystalline compound, which is perhaps the analogue of the compound
obtained under similar circumstances from benzenylamidioxime, that is to
say a ring compound of the azoxime type.
It was originally the intention of one of us to investigate these
products and also to determine whether a-amino- and a-hydroxy-
campholenonitriles would break down with alkalis as camphanonitrile
was found to do (Lapworth and Lenton, Trans., 1901, 70, 1292), but
as we hear privately from Dr. Forster that he has for a long time con-
templated an extensive examination of substituted camphoroximes, the
work has been abandoned in his favour.
EXPSBIHKNTAL.
CH'NH
a'Aminocam^f^ioroQsime, CgHj^^I.-^^^ *•
Ordinary aminocamphor may be converted into its oxime by the use of
hydroxylamine in the presence of a large excess of cold concentrated
sodium hydroxide, and the amount of oxime produced may frequently
be considerable. This method was the one which we at first employed,
but in very many cases in which we attempted to deal with more than
5 grams of material at a time, it was found that the yield of oxime
was very poor, and we did not succeed in discovering what were the
necessary conditions. In some cases, the experiments were perfectly
satisfactory, whilst in others, in which the conditions and concentrar
tions employed appeared to be exactly similar, the reverse was the oase^
Digitized by
Google
DERIVATIVES OF a-AMINQfCAKPHOROXIME. 551
However, by using an aqueous solution of hydrozylamine hydro-
chloride containing a large excess of sodium acetate, the results were
more consistent. The following is the mode of treatment which we
have finally adopted, and by means of it the oxime may be prepared
in considerable quantities at a time.
Aminocamphor hydrochloride (5 grams), hydrozylamine hydro-
chloride (3 grams), and sodium acetate (12 grams) are dissolved in water
(20 C.C.) and heated on tha water-bath for 4 — 6 hours. The liquid is
then cooled and separated by filtration from any insoluble matter which
may have appeared. As the precipitation of the impure oxime from an
acid solution such as this frequently causes its deposition as a viscid
oil, it is better, at this stage, to pour the liquid, with constant stirring,
into excess of 10 per cent, sodium hydroxide solution (about 50 c.c). To
the resulting clear liquid, dilute hydrochloric acid is added drop by drop
until the white material which separates no longer increases in amount.
After 10 minutes, the solid matter is separated by filtration and well
washed with water. A further quantity of oxime may usually be
obtained from the mother liquor by suitable treatment with acids or
alkalis. The yield is about 50 — 60 per cent, of the theoretical. The
dried oxime, which presents the appearance of a balky mass of minute
plates, may be purified by crystallisation from hot benzene. Oo
analysis :
0-1475 gave 0-3563 COj and 0-1344 H^O. 0= 659 ; H= 10-1.
0-2270 „ 31'1 c.c. moist nitrogen at 20° and 761 mm. N = 15-9.
CioHigONj requires C«65-9 ; H = 9-9 ; N= 15-4 per cent.
The oxime dissolves somewhat readily in ethyl or methyl alcohol,
ethyl acetate, acetone, benzene, chloroform, or carbon tetrachloride,
but is practically insoluble in cold water or light petroleum. It crys-
tallises badly from most 'organic solvents, with the exception of
benzene, from which it separates in thin, flaky plates or in flattened
prisms. It melts sharply at 144 — 145°. The pure, dry substance,
when slightly warmed, becomes very easily electrified by friction.
The smaller crystals are well-formed, six-sided plates, which belong,
in all probability, to the rhombic system. When carefully heated
on a glass slip beneath a cover-glass, the oxime sublimes slightly in
similar forms ; after melting, it solidifies rapidly to aggregates of well-
formed, elongated plates, the surfaces of which are, for the most part,
parallel to the axial plane, but here and there, in convergent polarised
light, the axis of a biaxial figure of wide angle may be observed ; here,
the double refraction is negative.
The oxime and its derivatives are optically active. A 1 per cent,
solution of the oxime in absolute alcohol had [a]j> 60*5° at 18°*
In order to observe the rotation of the ion in aqueous solutions of
Digitized by
Google
552 LAPWORTH AND HARVET :
the salts of the ozime, a 1 per cent, solution of the ozime in dilute
hydrochloric aoid was examined ; excess of hydrochloric acid produced
no marked effect on the rotation. The solution had [a]i> 36'7° at
16°, giving for the ion [a]^ 36*5°.
The substance has both acidic and basic properties, dissolves readily
in dilute mineral acids and in a large quantity of alkali. On adding
acid to its solution in alkalis, the oxime is almost completely precipi-
tated, whilst the liquid remains strongly alkaline. It forms sodium
and potassium salts, but these are only obtained crystalline in presence
of a large excess of strong alkali, and have not been closely examined.
The compound, when gently heated above its melting point, emits
a faint odour resembling that of camphoroxime. At still higher
temperatures, profound decomposition occurs, and water and gases
with an ammoniacal odour are evolved, whilst a waxy substance,
with an odour resembling that of camphenone, distils.
Aminocamphoroxime hydrochloride, C^QH^gON^tHClyHjO, is prepared
by dissolving the oxime in the requisite quantityjof hot 15 per cent,
hydrochloric acid. It crystallises, as the solution cools, in flat,
rectangular prisms, is very readily soluble in ethyl or methyl alcohol,
and still more so in water. The water of crystallisation could not be
determined directly owing to the instability of the substance when
heated. It contains IH^O, as the following facts indicate :
0-3071 contained 0-0460 CI. CI = 15-0.
CioHi80Nj,HCl,H,0 requires CI =- 150 per cent.
Moreover, its specific rotation in 2 per cent, aqueous solution was
[ajo 27*6°, whilst from the observations above recorded the calculated
number for the monohydrated salt is [a]o 28*3°.
The majority of the other salts of the oxime, such as the sulphate,
nitrate, picrate, ferro- and ferri-cyanide, and the aurichloride, are very
soluble in water and are not very characteristic.
The plcUintehloride, {Oj^f^H-^fiiN^^t^^^tGl^t is precipitated in the form
of minute, yellowish-grey needles on mixing strong solutions of the
hydrochloride and chloroplatinic aoid. It is readily soluble in water,
dissolves slowly in hot alcohol, and may be precipitated from its
alcoholic solution by addition of ether as long, yellow needles which
melt and decompose at 209 — 211°. The extinction in polarised light
is straight; the double refraction is strong, and the directions of
greatest elasticity and length are at right angles :
0-1639 gave 00398 Pt. Pt = 243.
{CioHi30N,)„H^tCl^ requires Pt = 25-2 per cent.
.CH-NH-CO-C.H-
Dtbenzoylamtnocampharoxtme, Cg^w^^' v.o-CO»C H • — '^^^ ®"^'
Digitized by VjOOQIC
DERIVATIVES OF a-AMINOOAMPHOROXIME. 553
stance is formed when aminocamphoroziine is dissolved in a consider-
able excess of 10 per cent, sodium hydroxide solution, and subjected
to the action of benzoyl chloride in the usual manner, the temperature
of the mixture being kept low by the addition of powdered ice. The
white, oily mass which is deposited may be collected by means of a
glass rod, and, after drying, triturated with absolute alcohol, when it
rapidly becomes pulverulent, and may be drained on porous earthen-
ware and crystallised from absolute alcohol. On analysis :
0-2708 gave 07364 00, and 0-1656 H,0. 0 = 740 ; H - 6-8.
^iO^^fii^i r«lttire8 0 = 74*2 ; H = 6*7 per cent.
The compound is somewhat readily soluble in methyl or ethyl
alcohol, ethyl acetate, or acetone, much less readily so in ether or
benzene, and is nearly insoluble in light petroleum. It is best crys-
tallised froin ethyl alcohol, from which it separates in brilliant, trans-
parent prisms, which are often of considerable size and probably
belong to the rhombic system. Eapidly recrystallised, it forms
minute, compact prisms. It melts sharply and without decomposition
at 146 — 147°, and solidifies on cooling very slowly, unless the tempera-
ture is kept at about 120°.
When crushed fragments of the crystals are observed in convergent
polarised light, a biaxial interference figure may occasionally be seen,
The double refraction is strong and positive in sign.
Th6 rotatory power was determined in absolute alcohol. 0*3317 gram
dissolved in 25*1 c.c, gave ai> 2*77° in a 2-dcm. tube at 18°, hence
[aJD 104-8°
Attempts to prepare an acetyl derivative of the oxime were unsuc-
cessful. Treatment with acetyl chloride converts the substance into
a mixture, of which a part is soluble in water and the remainder
is oily and has resisted all attempts at purification. Acetic anhy-
dride acts violently on the oxime, affording a small quantity of
a crystalline compound which evolves acetic acid when warmed with
sulphuric acid.
The carbamide, C^n^^<^'f^*^^*^^^.—This is easily prepared
by adding a solution of potassium cyanate to one of the hydro-
chloride of the oxime, and warming the mixture on the water-bath
for 15 minutes, when it separates as a bulky mass of minute needles
which may be purified by crystallisation from methyl alcohol. A
specimen was analysed :
0-2641 gave 0*5664 00^ and 0*2004 H^. 0 « 58*2 j H = 8*4.
Oil 10O2N3 requires =58*6; ss 8*4 per cent.
The compound is readily soluble in acetone or ethyl alcohol, less
VOL. LXXXL P P
Digitized by
Google
554 LAPWOBTH AND BAXtltt t
readily bo in ethyl acetate or methyl alcohol, and dissolves only
sparingly in chloroform, carbon tetrachloride, benzene, or light
petroleum. It is most conveniently crystallised from methyl alcohol^
from which it separates in two entirely different forms according to
the conditions. If the hot concentrated solution is cooled rapidly, the
compound frequently appears as a bulky or flocculent mass of slender
needles, whilst if the solution is allowed to cool very slowly, or if
dilute ethyl alcohol is the solvent, the substance is usually deposited
in the form of large, transparent, six-sided plates, but it is not always
possible to ensure the deposition of either form at will. The melting
point of the two forms appears to be the same, namely, 203 — 204^,
when complete decomposition ensues; on one occasion, however, a
specimen of the plate-like form was seen to fuse at 158 — 159°, and,
solidifying immediately, melted once more at 203 — 204° ; this would
appear to indicate that, on heating, one form undergoes conversion into
the other, but we have not been able to obtain satisfactory confirma-«
tion of this view.
A solution of the compound in hot benzene forms a transparent
jelly on cooling and could not be made to deposit crystals.
The rotatory power was determined in absolute alcohol, 0*2333 gram
being dissolved in 25 c.c. of the solvent, and was found to be [ajo
409° at 14°.
The pkenylcarhamide of camphoroximepheniflcarhamctiet
CH-NH-CO-NH-CeH,
— Aminocamphoroxime (3*8 grams) is suspended in absolute ether and
treated with phenylcarbimide (5 grams) dissolved in the same liquid*
Immediate combination occurs with formation of a white, very
sparingly soluble material, which may be separated by filtration and
purified by crystallisation from a large bulk of methyl alcohol. On
analysis :
0-2266 gave 05716 OOjj and 0 1381 HjO. C«68-8; H = 6'8.
Og^HjgOgN^ requires C « 68-6 ; H = 6-7 per cent.
CiyH,30^3 ., 0 = 67-7; H- 7-6 „
The substance, which, as analysis indicates, is evidently the pro-
duct of union of two molecules of the phenylcarbimide with one of
the amino>oxime, is readily soluble in ethyl acetate, acetone, or ethyl
alcohol, less readily so in methyl alcohol or benzene, and is very
sparingly soluble in ether or light petroleum. It crystallises from hot
absolute alcohol in long, slender prisms, and from hot methyl alcohol
in bulky masses of asbestos-like needles. It melts at 175 — 177° and
does not solidify on cooling. The crystals from absolute alcohol have
straight extinction in polarised light, the directions of greatest elasticity
Digitized by
Google
bERIVATIVKS OF a-AUINOCAMPHOBOXIME. 555
being coincident with their length ; the double refraction is strong.
For the determination of its optical activity, 0*2603 gram was dis-
solved in 25 c.c. of absolute alcohol, and this solution, examined in a
2-dcm. tube at 15°, gave ao-l'lS^ whence [a]D-56-6°.
The benzf/ltdene compound, CgH^^^I, ' * ^ — Molecular pro-
portions of the ozime and benzaldehyde are dissolved in a small
quantity of alcohol and warmed on the water-bath for 4 hours. A.t
the end of this time, the mixture is cooled and the crystals separated
and crystallised once or twice from hot methyl alcohol; a further
quantity of material may be obtained from the mother liquor by
dilution with water. On analysis :
0-3290 gave 0-9085 COj and 0*2390 HgO. C = 75*4 ; H - 8-1.
CjyHjgONjj requires 0 = 75*6 ; H * 82 per cent.
The dompound dissolves somewhat readily in ethyl or methyl
alcohol, ethyl acetate or acetone, less readily in benzene, chloroform,
or carbon tetrachloride, and is insoluble in light petroleum. It
separates from alcohol in brilliant, transparent, apparently hemi-
hedral prisms, closely resembling those of magnesium sulphate. It
melts at 153 — 154° and on cooling solidifies very slowly, melting
afterwards at the same temperature.
The crystals are probably rhombic, have straight extinction in
polarised light and are strongly doubly refractiva Melted on a
glass slide beneath a cover-slip, the substance slowly sets to a mass of
well-formed plates, which usually show straight extinction ; through
some, however, the optic axis of a biaxial interference figure of wide
angle emerges obliquely through the field.
The compound is nearly insoluble in dilute sodium hydroxide
solution, but dissolves readily in dilute hydrochloric acid, being precipi- .
tated from the solution on addition of alkali. When warmed for a
few minutes with moderately concentrated acid, it suffers hydrolysis,
yielding bensaldehyde and the original amino-oxima
The authors' thanks are due to the Research Fund Committee of
the Ohemical S6ciety for a grant defraying a portion of the cost of
material used in this work.
Goldsmiths' Inbtituts,
Nsw Cboss, Londok, S.E,
Digitized by
Google
556 HARTLEY; THE ABSORPTION SPECTRA OF
LX. — The Absorption Spectra of Metallic Nitrates.
By Walter Noel Hartley, D.Sc, F.R.S.
A BAND of absorption was discovered by Soret in nitric acid and
potassium nitrate, but in ethyl nitrate examined by Soret and Billiet
it did not appear. In 1887, I made an extended examination of the
absorption spectra of those nitrates of simplest constitution, such as
nitric acid, and potassium, sodium, silver, and thallium nitrates,
chiefly with the object of ascertaining definitely whether there was
any difference in the absorption spectra transmitted by these salts
both in strong and in dilute solutions. They were selected because
the band which is characteristic of them is situated in the ultra-violet
region ; they are anhydrous, are derived from monad metals, and there
are great differences in the atomic masses of the metals which enter
into their composition, for instance, H = l, Na = 23, K = 39, Ag=108,
Tls=204. The results which immediately followed are stated under
the head of Xst Series, In 1898, I returned to the subject and the
results obtained are in part described under 2nd, 3rd, ith, and 5th Series.
The method of examination was precisely the same as that pursued
with organic substances. As many, if not all, of the photographs taken
at that time are still in existence, I have recently re-examined them and
find that in the earlier stages of the work the simple cadmium spark
spectrum was used as the source of rays, but that the mode of examina-
tion recorded in detail in ''The Absorption Spectra of the Alkaloids"
was afterwards adopted (see Fhil. Trcms,, Part II, 1885). Subsequently
the lead, tin, and cadmium alloys were modified and improved by the
introduction of bismuth. There were thus photographed from these
electrodes 153 sharply defined metallic lines of nearly the same
intensity, distributed fairly equally throughout the whole spectrum,
which was 3 '95 inches in length, and, if necessary, capable of being
enlarged to 25 diameters. In addition to the metallic lines, which are
definite fixed points in the spectrum, there was a background of
continuous rays filling up the spaces between them. For the extreme
ultra-violet about wave-length 2000, electrodes of indium and of copper
were used for the purpose of filling up the spaces between the cadmium
lines.
Seven different series of experiments have been made on different
metallic nitrates, chiefly on those which afford colourless solutions.
The term colourless is here used in the ordinary acceptation of the
term, that is, no colour could be perceived by the eye when viewing
the solutions in the usual manner.
Digitized by
Google
METALLIC NITRATES 657
Ist Series, — A decigram-molecular weight of each of the four follow-
ing substances, nitric acid, potassium, silver, and thallium nitrates
was made up to a given volume with pure distilled water. The three
former were made up to a volume of 40 c.c, but thallium nitrate,
being less soluble, was contained in 400 c.c. The solutions were
examined in cells of a thickness of 25 mm. down to 1 mm., in the
case of thallium nitrate from 50 mm. to 2 mm.
The solutions were then successively diluted down to 1 decigram-
molecule in 5000 c.c, at which dilution it was expected that the co-
efficient of extinction of absorption would be attained.
As the thallium solution is 1/1 0th the strength of the others, 50
c.c. were reckoned as equivalent to 5 c.c. of them.
A description of the absorption spectra at different stages here
follows, the measurements being expressed both in oscillation-fre-
quencies and wave-lengths.
The diagram on p. 558 shows the most characteristic portion of the
curves, which were plotted in the usual manner as in the investigation
of organib compounds.
In explanation of the diagram, the following notes are appended.
Standard of volume 40 c.c. The weight of substance contained in
40 C.C. when viewed through different thicknesses which are specified,
or in the case of thallium nitrate, which is less soluble, through
equivalent thicknesses, is shown below :
1^^ Series.
I 5 mm.
1st dilution i ^
2nd dilution 1 mm.
Molecular proportions
With regard to 40 cc, it may be remarked that it was found by
experience to be a convenient standard to work with, since in the
cells which were used it is a volume which very nearly occupies a
cube, and consequently linear measures of the layers of liquid
examined represent proportional weights or molecular proportions of
the dissolved salt in the path of the rays.
These curves are such as I have previously described as curves of
molecular vibrations.
HNOg.
KNOj.
AgNOj.
TINO,.
grams.
grams.
grams.
grams.
1-26
202
3-38
5-32
0-252
0-404
0-672'
1064
0-0504
0-0808
0-1344
0-2128
63
101
169
266
Digitized by
Google
558 UABTLST: THE ABSORPTION SPECTBA OF
OieiUcUien/requeneia,
^1
I
th
1/5
th
1/25
th
1/126
mm.
5
mm.
8
mm.
2
mm.
1
4
8
2
1
• 3000 1 a s ♦ j
(1 7 • •♦OOOi
L
■u
A
i"*
^4
1
1
t
J^
.-U
t -.^
"H
1
r
"1"
'.'Xt -
-dt
- t
-it
- X
-it
X
.__Lt
-li
>c ^
TJ4
AjP
I
3
J^
i i
7^
utt -,
■ " \ii
'^'^ - .-
..-* 1
Jl
—J
X Kitrio acid and potatftium nitrate.
— — Thallinm nitrate.
■■ ■■ Silver nitrate.
Digitized by
Google
METALLIC NITBATES.
669
04
CI
09
04
i
«8 s«
9
^
to
O cog geo o
« T1 T .e
cq
1
I
1
.9
■■a
§
.r
QQ^
5
I
•i
03
I
Digitized by
Google
560
HARTLEY: THE ABSORPTION SPECTRA OF
Ui Series.—Silver Nitrate. A-gNOg, 16-9 grama in 40 ex.
Thickness of laver of
liquid '
6 mm.
4 mm
8 mm
2
mvm
1 mm
Description of spectrnm.
il
i|
it
if
if
II
n
1*
ii
as
5£
if
o ^
If
Spectrum extends to
Absorption band
indicated.
Ray transmitted at
2941
3408
2941
8408
2941
8408 2941 8408 8008 8824
a strong a strong
line at line at
3647 2740 3647 2740
Spectrum extends to..
Hay transmitted at ..
AgNOg, 16*9 grama in 200 e.e.
3008
8647
3324
2740
3896
2564
8896
256413896
256418896
2564
AgNOg, 16-9 grama in 1000 c.c.
Spectrum extends to 18896
2564 4028
248214028 2482
14028 248214028
2482
The transmission of an isolated strong line indicates the position of a feeble band
of absorption lying between it and the continuous spectrum, also that the absorp-
tion is greatly weakened and is beginning to fade away. This remark applies to
the descriptions of other spectra which will follow.
AgNOg, 16*9 grama in 5000 e.e.
Spectrnm extends to 14028
248214066
245914066 2459
Lo66 245914851
2299
Digitized by
Google
METALLIC NITRATES.
561
o
o
I
to
1
"S
1
i
o
S
I
'sapaonbaij
'Bdionanboj}
noT)V[|i9so
'eaioaonbdi;
udpBjiioBO
'8q;9a9[
-9A»^
'SQioneubejj
nopvixpsQ
■9A»^
*soiDa3nbajj
aopBiiioBQ
•8q!^8uoi
-9A«jVV
*Boioii9nbajj
nopv||T98o
'8q)9a9i
■9AB^
*S9i3a9nb9jj
*8q^a9i
•9A«^
*89pn91lb9JJ
nop«[ip80
U
C
msa^
"CO N«
MM
"W
o99
3 00 o^^
si
o
to
OH
s
a
s
a
K>
iH
U»
O
•41
^
<o
«>
a
as
6
kO
09
>o
o
-*
^^UMX
04
00
s
;s
a
00
00
^
04
00
a
s
a
00
•♦
3
dM
e
00
0«
Digitized by
Google
562
HABTLET: THE ABSORPTION SPECTRA OF
It will be remarked that the absorption carves of nitric acid
and potassium nitrate are identical in every particular, whilst those
of silver and thallium nitrates are somewhat different^ not only from
that of nitric acid, but each differs from the other. The salts with
high molecular weights are those which exhibit the greater extent of
absorption, and moreover the absorption band extends further
towards the less refrangible rays. The loading of the molecule pro-
duces a similar effect in organic substances as is now well known.
2nd Series, — Normal solutions of nitrates were made and examined
through a uniform thickness of 200 mm., the solutions being diluted
successively through 1/2, 1/4, 1/20, 1/100, 1/200, 1/300, 1/400, 1/500,
1/600, 1/700, 1/800, 1/900, and 1/1000 volumes of water. The final
dilution may be expressed as a gram-molecular weight of the substance
contained in 1000 litres.
It was necessary to carry the dilutions thus far in order to ascertain
whether or no a second absorption band occurred in the region of rays
more refrangible than those absorbed by nitric acid, since a band
might be introduced by the metallic element of the salt. The follow-
ing measurements are recorded in oscillation frequencies only, which
are the reciprocals of the wave-length numbers.
2nd Series. — NOrie Aoid.
Normal solution ; 63 grams per litre. Column of liquid 200 mm.
Spectrum continaons
Dilution.
to
Vx.
1/1
2884.
1/2
2884.
A line at 2988.
1/*
2884.
1/20
Line at 2988. Absorption Band.
8004. V^-
Line feeble 8069. 8069 to 8842.
Bays transmitted.
Line at 8842.
lAOO
8155. 8155 to 8504.
8504—8902.
1/200
8155. 8155 to 8504.
8504>-8940.
1/800
The same as 1/200, bat absorption band less
marked.
Rays transmitted to
4050.
1/400 to 1/900
The same.
Spectrum ends at
4050.
Digitized by
Google
METALLIC NITRATES.
663
2nd S&riea, — Potassium NUreUe,
Normal solution ; 23*1 grams in 200 c.o. Yarious dilations ; 200 mm.
Dilution.
1/1
1/2
1/4
11/20
1/100
1/200
1/800 to 1/900
Spectrum continuous
to
2ddi.
2884.
Line very feeble 2988.
2884.
Line stronger 2988.
2884.
Lines yery feeble Absorption Band. Just ylsible 8886.
8004-8062 V^-
8155. 8155 to 8504. 8504 to 8906.
Lines very feeble 8808 and 8841
8062 Bays feebly transmit- 8504 to 8894.
ted to 8504.
The same, a'weakness seen between 8062 and 8471.
Spectrum ends at 4034.
2nd Series. — Silver NitnUe.
Normal solution ; d3'98 grams in 200 c.o. Yarious dilutions; column
of liquid, 200 mm. in thickness.
Dilution.
^'}
1/2
V*
1/20
1/100
1/200
1/800
1/400
1/500 to lAOOO
Spectrum continuous
to
VA.
2818.
2884.
2988.
8004. Absorption band.
Line feeble 8069. V^*
8149. 8140 to 8886. Line visible at 8886.
8149. 8140 to 8504. 8495 to 8940.
8149. The same. 8495 to 8940.
The band has disappeared, rays continuous to 4084.
The same, rays continuous to 4120.
This solution was brilliantly clear and it remained five days exposed
to diffused sunlight without becoming opalescent or discoloqr^ \n
the slif^htest degree.
Digitized by
Google
564
HARTLEY: THE ABSORPTION SPECTRA OF
2nd Series, — Silver Nitrate.
Normal solutioa dilated. Columo of liquid, 200 mm.
Dilation.
^1}
1/2
V*
1/20
1/100
1/200
1/800
1/400
1/500 to 1/900
1/1000
Spectrnm cx)iitinaoiiB
to
2818.
2884.
8004.
A line at 2988. Absorption band.
8149. V^- Li°« visible 3886.
8149. 3140 to S604. Rays transmitted
The same. 8495 to 3940.
8149. Absorption complete. 8495 to 3940.
4084 rays continnous, absorption band ceases.
The same, with the spectrum becoming stronger.
Rays continuous to 4120.
Water transmits to ^/\ 4555, a line may be seen even beyond thin
about Va 4668 ?
It is worthy of remark, as showing the purity of these solutions, that
the silver nitrate has retained absolute freedom from deposit or any
sign of coloration after being exposed to light for two years. It is
also quite neutral to litmus.
Zrd Series, — In this series of observations, advantage was taken of
the fact that the nitrates as a rule show their most characteristic
absorption spectra between the dilutions of l/20th to 1/lOOth of a
normal solution.
This is to be observed by examining the 1^^ and 2nd Series, The
meaning to be attached to the expression l/20th normal is that an
equivalent weight of a salt in grams is contained in the volume
of twenty litres of water, and so on.
I
^d Series, — Nitric Acid.
Normal solution ; 63 grams per litre. Column of liquid, 200 mm.
Dilution.
Spectrum continuous
to
Va.
Ray just visible at
1/20
8069.
Absorption band.
8886.
1/80
8069.
VA.
Rays transmitted.
8652 to 3894.
1/40
8084.
S084toS652.
1/50
8084-8166.
8156 to 8586.
8586 to 3894.
1/60
8084-8156.
8156 to 8545.
8546 to 8940.
1/70
8156.
8156 to 8407.
8497 to 8940.
1/80
8156.
8156 to 8497.
3497 to 3940.
1/90
8156.
Lines exceedingly
8504 to 3940.
feeble are trans-
These lines are
mitted here.
not of normal
intensity.
Digitized by
Google
METALLIC NITRATES.
565
3rd Series.— Lithium NitraUy LiNOg.
Normal solution 3 69 grams per litre or 13*8 grams in 200 c.c. Ck)lumn
of liquid, 200 mm.
DilatioD.
Spectnim continuooB
to
VA.
Absorption band.
Rays transmitted.
Very feeble.
VA.
1/20
8069.
8069 to 8766.
8766 to 8900.
1/80
8076.
8076 to 8679.
8679 to 8900.
1/40
8149.
8149 to 8405.
8496 to 3900.
1/60
8149.
8149 to 8495.
8496 to 8940.
1/60
8149.
8149 to 8405.
Rays very feebly
transmitted.
8496 to 8940.
1/70
3149.
The same.
Absorption band.
VA.
8267 to 8459.
The same.
1/80
8267.
3469 to 8940.
Very feeble.
1/90
The same.
Absorption band.
still strong, but there
are some rays feebly
visible in it, which
8940.
indicates that it is about
to be weakened.
It was found subsequently that this salt contained water in the
proportion indicated by three molecules of water to four of lithium
nitrate. This requires 82*5 grams instead of 69 granis per litre.
Zrd Series, — Silver Nitrate,
Normal solution. Column of liquid, 200 mm.
Dilution.
Spectrum continuous
to
1/20
Va.
8004—8069.
Rays just visible at
1/80
3069.
Absorption band.
8886.
1/40
8069.
VA.
Rays transmitted.
8069 to 8579.
8579 to 8900.
1/60
8148.
8148 to 8579.
8579 to 8900.
1/60
8149.
8149 to 8495.
3496 to 3900.
1/70
8149.
8149 to 8405.
8496 to 8900.
1/80
8149.
8149 to 8405,
but rays very feebly
transmitted from
8267 to 8800.
8496 to 3900.
1/90
3149.
Rays very feebly
transmitted.
8496 to 8900.
Digitized by
Google
666
HARTLEY: THE ABSORPTION SPECTRA OF
ith Serisa. — ^A fourth series was taken in which normal solutions .
were not diluted, but examined in columns of different lengths, com-
mencing with 200 mm. and decreasing thus : 100 mm., 50 mm., 40
mm., 30 mm., 25 mm., 15 moL, 10 mm., 5 mm., 4 mm., 3 mm., 2 mau
As through 200 mm. the absorption effect is caused by one gram-
molecular weight of the salt, in 2 mm. it is the effect of 1/lOOth of
this weight. This remark is equally applicable to the dyad metals to
which the term 'normal' must be applied in the usual sense of
NO3/IOOO employed in volumetric analysis.
Comparison of differences in length of column of liquid of normal
strength with the corresponding dilutions :
Normal
200 mm.
100
1/2
60
l/4th
40
l/5th
30
3/20th8
26
l/8th
20
1/lOth
16 mm.
8/40th8 ^
\on
10
2/40th8
»
6
l/40th
»
4
l/60th
•>
3
S/200th8
n
2
1/lOOth
II
1
l/200(h
M
Lithium nitraU. 69 grams of salt per litre or 13*8 grams in 200 c.c«
of solution. This was intended to be a normal solution, but after it
had been made up, I determined the quantity of water in the salt
and found it to be in the proportion of ZKfi to 4LiN0y, but it is not
to be supposed that this was in the nature of a crystalline hydrate,
the salt being somewhat hygroscopic. The solution of lithium nitrate,
intead of being normal, was only 5/6th normal, as 83 parts should have
been dissolved instead of 69.
iik Series.— Lithium NiWaU. From 200 mm.
to 2 mm.
Thickness.
Spectrum continuous
to
mm.
200
2884.
100
2988.
60
2988 feebly to 8000.
40
8000.
80
8000.
25
8000 feebly to 8062.
15
8000 feebly to 8076
Abson>tioii band.
^ Va. ^
Bays transmitted.
10
8076.
8076 to 8766.
8766 to 8900.
6
8076 feebly to 8149«
8149 to 8766.
8766 to 8900.
4
8149.
8149 to 8488.
liinee rery feeble seen
about 8267 and 8808.
8488 to 8940.
8
8149.
The same.
8488 to 8940.
S
8149. Bays weak, but continuous to
8940.
Digitized by
Google
Metallic nitrates «
667
ith Series.— Sodium ifitraU, NaNO^.
Normal solution ; 85 grams per litre> or 17 grams per 200 c.c.
Thickness.
Spectrum continaons
to
Tnm.
Va.
200
2988.
100
2990.
60
2990.
80
8004.
20
3004-8069 very feeble.
16
8004—3069 feeble.
Absorption band.
Rays transmitted
V^.
rery feebly^
10
8069.
Isolated
Unes.
8069 to 8766.
8766 to 8884.
6
8069—8166.
8155 to 8579.
8679 to 8900.
4
3069—8165.
8155 to 8585.
6886 to 8900.
8
8069—8166.
8155 to 8504.
8604 to 8900.
2
8069-8166
8155 to 8440.
Rays here transmitted
very feebly.
8274 to 8808*
8440 to 8940.
Uh Series.— PotaeHum NitraU, KNOg.
Normal solution.
Thickness.
Spectrum continuous
to
mm.
Va.
200
2884.
100
2940.
"
60
2940—2988.
40
2988—3004.
80
8004.
Line rery faint at
20
8047—8069.
Absorption band. 8886.
16
8047—8069.
Va. 8886.
10
8069.
8069 to 8760. 8760 to 8900 yery
feeble.
6
8076—8166.
8155 to 8579. 8679 to 8900.
8155 to 8504 8604 to 8940.
4
8156.
8
8190.
8190 to 8459. 8469 to 8979.
Rays about 8274 and
8376 feebly trans-
mitted.
2
8190.
Rays transmitted from 8267 to 8979, but
rery imperfectly between 8341 and 8469.
Digitized by
Google
568
HABTLRT : THE ABSOBPTION SPECTBA Ot
Uh J^fertes.—nalUum NUraie.
Ov^ing to the sparing solubility of this salt, the solution was made
l/4th normal. 200 mm. to 2 mm.
Thickness.
Spectrum conUnuous to
mm.
Va.
200
2946.
100
8004.
50
8004 to 8069 feeble.
40
8069. Absorption band.
Very feebly
transmitted.
80
8069. 8069 to 8652.
8652 to 8886.
25
8069 feebly to 8155. 8155 to 8585.
8585 to 8894.
16
8069 „ 8155. 8155 to 8585.
8585 to 8894.
10
8155. 8155 to 8440.
Rays very feebly
transmitted about
8267 and 8375.
8440 to 8940.
5, 4, 8, and 2
Rays continuons and strong to 3267 and feeble from 8267 to
8979.
There is a deceptive appearance when curves are drawn from spectra
taken only at the above points, so that, for instance, thallium nitrate
appears practically the same as potassium nitrate when allowance is
made for differences in the strengths of the respective solutions.
Oompare the curves of the let Series and it will be seen that there
is a great difference between them.
4<A Serie$,—Magneeium NitraU, lA.g(^O^^fiJlfi.
Normal solution ; 256*4 grams per litre, or 51*3 grams per 200 cc.
The spectra of the dyad metals are not strictly comparable with
those of the monad group.
Thickness.
Spectrum continnons to
mm.
Va.
200
2798.
100
2884.
50
2988.
40
2988.
Isolated
line.
85
2938—2990.
80
2990-8004.
20
2990—8004.
15
2990-8004.
An isolated
10
8004— 806^ very faint
line rery faint.
6
8069. Absorption band.
8842.
Va.
4
8069.
8069 to 8772.
8772 to 8900.
8
8069—8149
fMbfe.
8149 to 8652.
8652 to 8900.
2
8069—8149
very
feeble.
8149 to 8504.
8504 to 8940.
Digitized by
Google
METALLIC NITfiATES.
569
4<A Seru3.—Calc%wm NUraUy Ca(N03)j,4HjO (Ordway).
Normal solution ; 47*2 grams per 200 c.c.
Thickness.
mm.
200
100
50
40
80
26
20
15
10
5
4
3
Spectrum continaoas
to
Va.
2988.
2938.
3004.
8004.
8004.
8069. ,
8069. Absorption band.
3069. Va.
8069—8149 reiy faint 8149 to 8772
8069—8149 very faint
8076—8149.
Absorption band. Line indicated faintly.
8842.
8842 yery feeble.
8772 to 3905 feebly
yisible.
8149 to 8660. 8660 to 3905 feeble.
8140 to 8604. 8504 to 8916.
Rays feebly
transmitted.
4£& Serie$.—Zino NUraU, Zn(N03)2,6H^O.
Normal solution ; 297 grams per litre, or 59*46 grams per 200 c.c.
Thickness.
Spectnim continnons
to
mm.
Va.
200
2884.
100
2884.
50
2988.
40
2988—2976.
80
2976.
20
2975—8002.
15
8002—3076.
10
8076.
Absorption band.
6
3076.
Va.
Rays transmitted.
4
8076.
8076 to 8646.
8646 to 8894 very
faintly.
8646 to 8894
8
8076—8156.
8156 to 8646.
feeble.
2
8155-3828.
8828 to 8426.
8426 to 8940
feeble.
VOL. tXXXI.
Q Q
Digitized by
Google
— »to
HiJttLEY: fHE AdSOaPTtOlf SPtiCTlU Of
1
4fA SwiM* — Sofrium NUreUe^
l^hifl salt being sparingly soluble, a solution l/4th normal was used ;
66*25 grams per litre ; 200 mm. to 2 mm.
Thickness.
Speotram oontinuoiu
to
tntn^
Va.
200
2088.
100
2988 to 8004 feeble.
60
8004 to 3047 feeble.
40
8004 to 8047 feeble. Absopption band.
Vx.
Bays truismitted.
8772 to 8906.
SO
8047. 8047 to 8772.
26
8047. 8047 to 8772.
8772 to 8906.
16
8047 to 8166 rwj faint 8166 to 8662.
8662 to 8906.
10
8047 to 8166 feeble. 8166 to 8686*
8686 to 8940.
6
8204. 8204 to 8686.
Rays fully transmitted
8686 to 8986.
here. Partial absorp-
Strong and con*
tion.
tinuona.
4
8204. Absorption band still
Tiflible although rays
8471 to 8986.
fully transmitted.
Rays oontinnoua
8
Continuous spectrum to 8986, but enfeebled between 8204 and
8471.
2
Absorption between 8204 and 8471 almost inyisible.
4M S&riM.—Lead NUrcUe.
A sparingly soluble salt. The solution was made seminormal ; 165*45
grams per litre } 200 mm. to 2 mm.
Thickness.
Spectrum continuous
to
mm.
Va.
200
2884.
100
2884 feebly to 2988.
60
2938 feebly to 2982.
40
2990 feebly to 8007.
80
8007.
26
8007 to 8047 feebly.
20
8007 to 3062 feebly.
Ab8on>tion band.
16
8062.
Va.
Rays transmitted.
10
8062 to 8148 feebly.
8148 to 8760.
8760 to 8836.
6
8062 to 8148 feebly.
8143 to 8760.
8760 to 8886
4
8148 to 8272 feebly.
8274 to 8686.
stronger.
8686 to 8894.
Digitized by
Google
MBtALLtC NItBATBS. 671
Thei^ is a slight difference in the carves drawn from the measure-
ments of the magnesium, calcium, and zinc nitrates when normal
solutions of each are examined in columns of 200 mm. long and shorter
columns down to 20 mm., also in smaller thicknesses varying hetween
15 mm. and 2 mm.
The variation is in the extent of the continuous rays transmitted in
the first instance^ and in the form of the curve in the second. The
barium and lead nitrates present considerable differences with respect
to absorption Of the rays in small thicknesses. These salts are, how-
ever, less soluble than the former, and the results, in consequence, are
not quite strictly comparable without corrections for the barium nitrate
and lead nitrate, their solutions being one-quarter normal and semi-
normal respectively.
Qram-molecular weights of the following nitrates contained in
1000 c.c, show a complete absorption of all rays beyond the wave-
lengths indicated, when a column 200 mm. in length is photographed.
Va a
Mg 2798 357
Oa 2938 340
Zn 2884 346
Tip 2946 340
Bai**^ 2938 340
Pbi 2884 346
Er 2905 ' 343
Thl/100«» 2883 346
Series 5. — Comparison of the spectra of nitric acid, potassium
nitrate, and silver nitrate, when normal solutions are examined in
cells of different thicknesses, and when these solutions are diluted in a
corresponding ratio and contained in a tube of 200 mm. in length.
Thus 10 mm. of a normal solution were compared with 200 mm. of a
solution l/20th normal, and 4 mm. of a normal solution with 200 mm.
of one l/50th normal. It may be remarked here that measurements
of spectra cannot be made with such accuracy through 200 mm. as
through 10 mm. or any smaller thickness, the boundary of the absorp-
tion bands being not so sharp, hence some difference in the reading
may be anticipated. The refraction of the length of the column of liquid
tends towards a diffusion of the absorbed rays.
•A
X
H
.. 2884
346
Li
.. 2884
346
Na ...
... 3938
340
K
... 2884
346
Ag ...
... 2816
356
Q Q 2
Digitized by VJ OOQ IC
672
HARTLEY: THE ABSORPTION SPECTRA OF
6ih Series, — Nitric Acid,
Bays continaoiiB to
Absorption band.
Rays tiansmitted.
mm.
200— l/20th
10 — normal
200— l/60th
4 — normal
8084
The same feeble
8084—8166
The same
Va Va
8084 to 8772
The same
8165 to 8504
The same
FaUuHum Nitrate,
Va Va
8772—8784—8906
The same
8604 to 8940
The same
200— l/20th
10— normal
8069
The same feeble
8069 to 8766
The same
8766, 8782, 8836
The same
200— l/50th
4 — normal
8069—8166
The same
8155 to 8510
The same
8610, 8640, 8906
The same
Silwr Nitrate,
10 — normal
6 — normal
4— normal
8004
8069
8069
8069 to 8585
8069 to 8540
none
8686, 8662, 8706
the hist barely visible
8604, 8686, 8662
200— l/20th
200— l/40th
200— l/60th
8004-8069 feeble
8069
8148
8069 to 8579
8148 to 8579
none
8679 to 8900
8679 to 8900
From another Bolution of silver nitrate and another series of photo-
graphs :
10 — normal
4 — ^normal
200— l/20th
200— l/60th
8002
8067
8067
8067—8160 feeble
8067 to 8581
8150 to 8581
none
8681 to 8688
8681 to 8888
It will be observed that in nitric acid and potassium nitrate, 200 mm.
of l/20th normal solution yields the same spectrum as 10 mm. of a
normal solution ; and 200 mm., l/50th normal, the same as 4 mm. of
the normal solution. ut with silver nitrate there is a considerable
difference in this respect, inasmuch as the more dilute solutions trans-
mit rays more freely through thicknesses 20 and 50 times as great as
that of the solution which is of normal strength, although no less than
the same quantity of nitrate is contained in each corresponding
solution.
Digitized by
Google
METALLIC NITRATES. 573
This difference I believe to be due to the action of light, although I
do not propose at this stage of the investigation to formulate anj views
of the actual change in the constitution of the solution, whether due to
ionisation or otherwise, but there is reason to believe that the action
of radiant energy on dilute solutions is of more frequent occurrence
than is usually admitted or taken into account ; in other words, there
are several other salts which are affected besides those to which such
action is usually admitted to be a cause of change in constitution. I
will merely mention three, namely, cobalt, nickel, and manganese ; to
these, I think, may also be added lithium. In the three former, we
must not lose sight of the possible effect of atmospheric oxygen, although
whatever the effect of oxygen may be, it is light which brings about
the change. It is not only on the surface, but within the liquid and
upon that side of the containing vessel which is exposed to the light
that the action is most vigorous. This fact was observed in nickel
bromide, independently of any observations of mine, by Mr. J. A.
Cunningham, B.A., who was assisting me at the time with some work
on the properties of solutions {Sci. Trans, Roy. DM, Soe.^ 1900, [ii],
7, 263).
It is worthy of note, as tending to dispel all doubts on the subject
of the purity of the silver nitrate solutions, that they were perfectly
neutral, and that they have been kept for more than 2 years in white
glass bottles freely exposed to the light of day without the slightest
change in their appearance. This is a proof that no organic matter or
any impurity has affected them.
The absorptive power of the thorium salt is very astonishing. On
the first plate that was exposed there was little of a spectrum to be
seen. The solution was then diluted 100 times, and through 200 mm.
a continuous spectrum was transmitted as far as wave-length 346.
These measurements serve to explain what has been recorded by
Liveing (Trans, Comb, Phil. Soe,, 1900, 18, 298), namely, that strong
solutions of the nitrates have a general absorption of the more refrang-
ible rays, and in this respect these salts differ from the chloride and
acetate of didymium and from the chloride of erbium.
At a future date, an account will be given of some solutions which,
owing to their mode of preparation, may be deemed to be more strictly
comparable with each other, and an opportunity may then occur of
discussing the whole of the results.
liFote. — Since the above was written, G. P. Drossbach (J5«r., 1902,
35, 91), working with 10 per cent, solutions of colourless metallic
nitrates and a thickness of 20 mm. of liquid, finds that there is a
general absorption of the rays beyond X 340. Absorption bands
become visible when the solutions are diluted.
I have purposely excluded from the present communication an
Digitized by
Google
674 KIPPING AND HUNTER : THE RESOLUTION OV PHENO-
account which had been prepared of the finely-marked absorption
bands exhibited in the ultra-violet region by nickel, oobalt, and uranium,
in the form of bromides, chlorides, and acetates. The bands are
characteristic of the metallic elements, and they lie in a region of rays
less refrangible than that of the bands of the nitrates. In some cases,
one band is affected by the other, so that the question which is in
course of investigation would become complicated.
LXI. — The Resolution of Pheno-a-aminocjcloheptane
into its Optical Isomerides. Tartrates of Pheno-
a-aminocyoloheptane and of Hydrindamine.
By Frederic Stanley Kippinq and Albert Edward HuirrsR.
PHENo-a-AMiNOCYCLOHEPTANE (Kipping and Hunter, Trans., 1901,
79, 602) and pheno-a-aminocycZopentane (a-hydrindamine), being very
similar in constitution, especially with regard to the nature and
position in the molecule of the asymmetric carbon group, as shown in
the following formulsB,
^6H4<55|nH J- GKp" ^e^*<CH(NH J>»
Pheno-a-aminocyc^heptane. Pheno-a-aminoeye^opentane.
it seemed possible that the two bases would behave in a similar manner
towards a given optically active acid ; if this actually proved to be the
case, the study of the compounds of pheno-a-aminoc^c^heptane
might throw some light on the nature of those hydrindamine salts
which have recently been investigated (Kipping, Trans., 1900, 77,
861 3 1901, 79, 430) and appear to be altogether abnormal.
The experiments which have so far been made with the e^fdo-
heptane derivative have shown, however, that the two bases behave
quite differently; whereas (^/-hydrindamine hydrogen tartrate is not
changed by fractional crystallisation from water, the corresponding
compound of c2/-pheno-a-aminoo^c/oheptane with d-tartaric acid is
readily resolved into the salts of its enantiomorphously related com-
ponents.
Of these two salts, the more sparingly soluble is that of the /-base,
which is thus easily isolated ; it is almost optically inactive, owing to
the molecular rotation of the base being approximately equal to that
of the acid, but of opposite sign.
The active base does not show the least tendency to undergo racem-
sation, and when liberated from its salts or when submitted to
Digitized by
Google
a-AMlMOCTGLOHBPTANE INTO ITB OPTICAL ISOMSRIDES. 6T6
distiUation in steam, its optical properties seem to remain absolately
nnchanged ; further, when treated with benzoyl chloride and sodium
hydroxide^ it yields an'optically^active benzoyl derivative, and although
the sign of rotation is changed, there is no evidence of even partial
raoemisation having occurred.
These observations might be quoted in support of the view that the
existence of the hydrindamine salts already referred to cannot be
explained by assuming that the optically active hydrindamines
Immediately undergo complete racemisation when liberated from their
salts ^(Trans., 1900, 77, 878, 909); but considering that in the one
of other bases, almost as closely related to pheno-a-amino<^«2ahept8n6
in constitution, as, for example, in that of tetrahydro-)9-naphthyl-
amine (Pope and Harvey, Trans., 1901, 79, 74), partial racemis&tioh
occurs when the base is set free from its salts and when it is benzoylated
by the Schotten-Baumann method, it would seem that arguments based
on analogy have little, if any, value as regards such reactions. It would
follow, therefore, that even if racemisation of amino-compounds, in which
the amino-group is directly united with the asymmetric carbon atonr,
is due to tautomeric changes, >OH*NH2 :;:^ >>CINH3, as pre-
viously suggested, such changes are not necessarily intermediate
steps in the formation of salts or in the production of alkyl derivativM.
The relationship between the various salts which have been obtained
from cf^pheno-OFaminoc^c^oheptane and <i-tartaric acid is not without
interest, and of the five (or six) compounds which might be prepared^
namely,
I. dAl{^ II. dijJI III. dA,lB IV. dilj^l V. dA4Bi
all have been isolated and studied.
The first of these compounds, a normal salt, is deposited from
neutral and also from acid solutions ; it may be regarded as a partiAlly
compensated compound, but may possibly be more complex in char-
acter and consist of a partially racemic compound of the two saltsj
The normal salt of the M>ase (11) cannot be obtained directly by
orystallisiDg solutions of the cB-base in aqueous tartaric acid, but watf
prepared from the pure ^base ; it is more readily soluble in trater
than the normal salt (I) of the (2/-base, and also more readUy soluble
than the hydrogen salt (III).
The hydrogen salt of the ^base (III) is obtained by crystallising a
solution of the dM>ase in a large excess of aqueous d-tartaric acid,
and is consequently more sparingly soluble than the corresponding salt
gf the (i-base (Y) ; it contains 3 mols. H,0. ^he i^ormal salt of th^
Digitized by
Google
576 KIPPING AND hunter: THE RESOLUTION OF PHENO-
d-h&8% (TV) can be isolated from the mother liquors obtained in the
oryBtallisation of the preceding compoand ; it is anhydrous. The
hydrogen salt of the c^-base (Y) was obtained by crystallising the
normal salt from aqueous tartaric acid.
The tartrates of <U-hydrindamine behave differently from the
corresponding salts of the cyc/oheptane base. The hydrogen tartrate
is easily obtained in crystals from a solution of the base (1 mol.)
in aqueous (^tartaric acid (1 mol.) ; it is readily soluble in water,
and is unchanged on repeated fractional crystallisation. The normal
tartrate is still more readily soluble, and does not yield salts of
enantiomorphously related bases when fractionally crystallised.
The bromocamphorsulphonates of pheno-a-aminoc^oheptane and
of hydrjndamine, like the tartrates, fail to show the similarity in
properties which might have been expected ; whereas the salt of dl-
hydrindamine crystallises very readily and can be resolved into unequal
quantities of the isomerides already described (Kipping, lac. cit,) ; that
of the dl-ef/clohepixjie base is usually deposited as an oil which
solidifies very slowly and does not lend itself to fractional crystal-
lisation. The bromocamphorsulphonate of the /-base, on the other
hand, crystallises with great facility even at about 100^, so that the
opposite behaviour of the dU-ealt is obviously due to the fact that it is
a mixture.
The examples afforded by the above salts may be employed to give
precision to the term * partial racemism ' or ' partially 'racemic ' which,
first used by E. Fischer (compare Ladenburg, Ber,^ 1898, 31, 938), has
come to be employed rather in a wide sense to denote salts of a d- or ^
acid with a dl-haLse, or of a c2- or ^base with a cl^-acid, which may be
very different in character.
According to Ladenburg (loe. ett.), a partially racemic compound is
** eine Yerbindung zweier Eorper, die nur theilweise Spiegel bilder sind,
so dass also durch die Yerbindung nur eine theilweise Aufhebung der
optischen Activitat stattfindet und der racemische Eorper noch
optische Activit&t besitzt," and he applies this definition more par-
ticularly to the salt of (2Z-roethylglutaric acid with quinine and to the
salt of (i^tartaric acid with strychnine (Ber,, 1898, 31, 1969) ; these
salts, however^ he seems to regard as derivatives of racemic acids, for
he says further that the quinine salt '^ kein Oemisch der Salze von c^
und /-S&ure, sondern ein einheitliches Salz ist, also nur das Sals der
racemischen Saure sein kann," and he also represents the strychnine
salt by the formula rC^H«Oe,2C2iH8jO,Nj,6JHjO.
Wow it is rather difficult to understand Ladenburg's point of view in
this matter, since a racemic acid is merely a crystalline combination of
the d- and /-isomerides, and has noexistenoeezceptin such a form ; when
dbmbined with a base, a racemic acid ceases to exist, and there ar^
Digitized by
Google
a-AMINOCTCLOHEPTAKE INTO ITS OPTICAL ISOMEBIDES. 677
formed salts of the d- and ^acid% which may behave in varioas
ways.
Considering this point more generally, an externally compensated
dibasic acid and an opticaUy active mon-acid base may unite in solation
, D and lA \ . d» analogous
compounds being produced, namely, dA^ dB, and I A, dB, if the acid be
monobasic or if, being dibasic, the hydrogen salts are formed.
The two salts thus produced (1) may be different in physical proper-
ties and may be separable by fractional crystallisation as is frequently,
if not usually, the case ; (2) they may unite crystallographically when
deposited from solution, giving a substance differing in crystalline form
and in other physical properties from one or from both its components ;
(3) they may possibly be deposited from solution side by side in equal
quantities forming a mere mixture 3 (4) they may possibly form a
crystalline intercalation, somewhat similar to a pseudoracemic substance
(Kipping and Pope, Trans., 1897, 71, 989) as regards its crystallo-
graphic relationship to its components.
An optically active dibasic acid and an externally compensated base
may unite in solution to form three normal salts, namely, (a) dA \ jj^
(b) dA I ^^ and (c) dA< j^; on crystallising, the salt containing
both bases in one molecule may alone be deposited as a definite com-
pound, but if not, the mixture of the two salts (b) and (c) may possibly
behave in any one of the four ways already enumerated.
Corresponding possibilities are of course to be expected when, in any
of the above cases, the dibasic acid and mon-acid base are exchanged
for a di-aoid base and a monobasic acid.
Now in all the above cases, except in that in which a separation
occurs on fractional crystallisation and in that in which only one
kind of molecule (a) is produced, the salt conforms to Jjadenburg's
definition unless the meaning of ' Yerbindung ' be interpreted as a
crystallographic union, the result of which is to give a product differing
from at least one of its components in crystalline form, and conse-
quently in other properties ; if this limitation be not made, the term
* partially raoemic ' would include a number of salts of different types
in much the same way as did at one time the term racemic (Kipping
and Pope, loe. oit»)
Before classing a salt as a partially racemic compound, it is necessary
therefore to compare its behaviour with that of its component salts in
much the same way as is necessary in characterising a racemic com-
pound ; according to Roozeboom {ZeiL physikal. Chem.^ OS, 1899, 494),
the method based on solubility determinations with mixtures of equal
Digitized by
Google
578 KIPPIHO AND HUNTER : THB BESOLUTION OF PHSKO-
and of tmeqnal quantities of the two component salts may be made nse
of, but no experiments of this kind seem to have been made.
Betnrning to the tartrates described above, it would seem that the
partially racemic hydrogen salt of pheno^-aminocyeloheptane does
not exist, as the mixture of the two components dA/lB, dA^B, shows
the behaviour noted under (1). The normal tartrate of the <2^base,
which differs crystallographicaUy from the normal salt of the ^base^
may be either a partially racemic salt or consist of identical moleenles
^^\ 1^9 ^ cl'^M & compound of the latter type as partially racemic
would be obviously incorrect, and opposed to the whole meaning of the
word racemic as now used ; as, therefore, there are no means of dis-
tiDguishing between the two pofisibilities, the salt may, for the present,
be classed as a partially compensated substance.
The existence of partially externally compensated salts which are
mere mixtures of equal quantities of their components, and which t
nevertheless, cannot be resolved by fracfcional crystallisation, seems to
be doubtful, although, possibly, (i^hydrindamine (l-mandelate is an
example of such a substance (Kipping and Hall, Trans., 1901, 79, 443).
Experimental.
dl-PAtfno-a-amtnocycloAtfpto^M TarlraU,
When (tt-pheno-a-aminooye^heptane, partly dissolved and partly
suspended in water, as obtained by distilling in steam, is neutralised
with ^tartaric acid and the solution then concentrated and allowed to
cool, a salt separates in highly lustrous needles or prisms, and further
quantities of the same compound are obtained on again concentrating
the mother liquors.
This salt has a neutral reaction to litmus, and is the normal salt of
the c2/-base; it has the composition C4H0O0,2CiiHjgN. It is only
sparingly soluble in cold water, but dissolves fairly easily on boiling ;
it melts and decomposes at about 235^, but the rate of heatiog
influences the result very considerably, temperatures ranging from
230^ to 240° being observed, according as the salt is slowly or rapidly
heated.
The specific rotation of the ci^salt was determined in aqueous
solution with the following result :
0*25 gram of air-dried salt dissolved in water and the solution
diluted to 25 c.c. gave in a 200 mm. tube a +0*26°, whence [a]o 13^
The molecular rotation of the salt is therefore [M]o +60°, a value
which agrees fairly well with that calculated from the s|)ecifip
Digitized by
Google
a-AMINOCTGLOHBPTANE INTO ITS OPTICAL ISOMERIDES. 579
rotations of the normal metallic tartrates, namely, QMJd +67°
(compare Landolt, Dm optische Drehungsvermogen).
Fractional crystallisation of this normal tartrate from water does
not seem to resolve it into salts of the d- and Z-bases, and the salt may
be regarded as a partially compensated compound having the com-
, D ov <^^\ ^jD + ^-^ \ i£f *^® attempts to resolve it
into different fractions by crystallisation were not carried beyond two
or three operations, owing to the occurrence of hydrolytic dissociation
and consequent loss of base.
Salts of l-Fhona-a-cunmocyolc^ptime.
' l-PA0no-a-ami7U>cyclo^<and Hydrogen TartrcUe, — ^The mother liquors
remaining after the separation of several crops of crystals of the
normal salt just described give, finally, deposits consisting of very
slender needles, which form felted masses quite*unlike the prisms of the
normal salt; these very slender crystals consist of the hydrogen
tartrate, as was proved by titrating a dried and weighed sample with
sodium hydroxide solution. The presence of this salt is not due to
excess of tartaric acid having been added, originally, but to the fact
that the normal salt is partially dissociated hydrolytically in aqueous
solution, and on boiling or evaporating on the water-bath the base
volatilises with the steam.
In order to prepare the hydrogen salt in larger quantities, the
normal salt was mixed with one molecular proportion of cf-tartario
acid and the mixture dissolved in a considerable quantity of hot water I
on cooling, however, the normal salt was deposited unchanged and the
mother liquors gave only small quantities of the hydrogen salt together
with the tartaric acid which had been added, and most of which had
remained uncombined. This rather unusual behaviour is doubtless due
to the fact that in solutions containing the hydrogen salts of the d- and
Z-bases, the normal partially compensated compound is more sparingly
soluble than either of the hydrogen tartrates, and on crystallising a
change occurs which may be presented as follows, dA,dB-\-dA,lB^
On adding a considerable quantity of (i-tartaric acid to a solution
saturated with the normal dl-salt and containing a few crystals of the
latter in suspension, the crystals slowly dissolve, but on concentrating
the solution and allowing it to cool, the normal salt is again deposited
in the first fractions, apparently in a pure condition ; if, however, the
proportion of (2-tartaric acid present in solution be increased, the first
fractions consist of a mixture of the normal dZ-salt and of the hydrogen
Digitized by
Google
680 KIPPING AND HUNTER: THE RESOLUTION OF PHENO-
salt, until at last, by using a very large excess of the acid, crystals of
the hydrogen tartrate are deposited almost, but not quite, free from
those of the normal salt.
The pheno-a-aminoc^c^oheptane hydrogen tartrate which is thus
obtained is the salt of the /-base ; after one recrystallisation from
water, it was obtained in needles melting at about 175° and having a
specific rotation of about [a]j) + 4°. On repeated crystallisation, it
seems to be completely separated from any normal salt and from the
hydrogen salt of the (2-base, and is thus obtained in lustrous needles
or prisms very similar in appearance to the crystals of the normal
desalt. It melts and decomposes at 181 — 182°, the rate of heating
having very little effect on the melting point.
Unlike the normal salt, it contains water of crystallisation :
0-19361o8t0-0282 H^O at 100° HgO^ U-5.
0-2711 „ 0-0370 HjOoversulphuricacidandOOOlSat 100°. H^O-U^.
C^HgOgjCiiHijNjSHjO requires HjO-U'S per cent.
It seems to be rather more readily soluble in cold water than the
normal partially compensated salt, but it is only sparingly soluble in this
liquid and also in ethyl alcohol ; it crystallises from water unchanged
even in absence of tartaric acid.
As this salt is formed by the combination of a dextrorotatory acid
with a IsBvorotatory base, and as the two compounds have approximately
the same molecular rotations, but of opposite signs, solutions of the
salt are almost optically inactive.
Two different samples were examined with the following results :
0*4560 gram of dehydrated salt dissolved in water, the solution
diluted to 25 c.c. and examined in a 200 mm. tube, gave a - 0*03° ;
[a]„ -0-8°.
0 4974 gram of dehydrated salt under the same conditions gave
a -0-10°; [a]i> -2-5°.
The readings in these experiments being so small, the agreement
may be considered as fairly satisfactory, and the molecular rotation,
calculated from the mean value, may be taken as [MJdb - 5°.
As the molecular rotation of <i-tartaric acid in the metallic hydrogen
tartrates is [M]d +42°, that of the Isavorotatory base would be
[ajo —47°, a value which agrees very well with that obtained from
observations made with its hydrochloride.
l'Pheno-a-aminocjc\oh^l4ane Tartrate. — On adding an alcoholic solu-
tion of (i-tartaric acid to an ethereal solution of /-pheno-a-amino-
cye/oheptane, keeping the base in excess, the normal tartrate is
deposited in crystals which melt fairly sharply at 215 — 217°,efferve8C<
ing and turning slightly brown.
Digitized by
Google
a-AMlNOCYCLOHEPTANE IN1*0 ITS OPTICAL ISOMBRIDBS. 581
The salt thus obtained was quickly dissolved in hot water and the
solution rendered alkaline with the ^base ; from this solution it wa s
deposited in flat, transparent, triangular plates, verj different in appear*
ance from the crystals of the partially compensated normal salt, and
apparently anhydrous. It was comparatively easily soluble in hot
water, much more so than the hydrogen tartrate, but its solution on
boiling gave off half of the base, and on subsequently cooling the
hydrogen tartrate (m. p. 181 — 182°) was deposited. On adding
powdered tartaric acid to a cold saturated solution of this normal salt,
a heavy precipitate of colourless needles of the hydrogen tartrate was
produced.
I'Pheno-a-CMninocjclohepteme Hydrochloride, — This salt was prepared
by decomposing the hydrogen tartrate of the ^base with sodium
hydroxide, distilling in steam, and evaporating the distillate with excess
of hydrochloric acid. It separates from water in lustrous, striated
needles or prisms and shows no signs of melting when heated at 250° ;
it is moderately easily soluble in boiling water, sparingly so in cold
water, and very sparingly so in ethyl alcohol.
In order to ascertain whether partial or complete racemisation had
occurred in the formation of the hydrochloride from the hydrogen
tartrate in the above manner, the halogen salt was fractionally
crystallised from water and thus separated into two crystalline
portions, leaving only a very small quantity in the mother liquors.
The two fractions were then dried and examined optically.
Fraction I, 0*272 gram dissolved in water, the solution diluted to
25 C.C., and examined in a 200 mm. tube gave a — 0'52°, whence
[a]„ -24°.
Fraction IL 0*264 gram under the same conditions gave a - 0*50°,
whence [a]D - 23-7°.
It seems from these results that racemisation does not take place
either on liberating the active base from its salts or on heating it at
100°, and this conclusion is confirmed by experiments described later.
Taking the mean of the above values, the molecular rotation of the
base would be [M]]> —47°, a result which agrees very well with that
deduced from the observations made with the hydrogen tartrate.
BtfiizoyUVjiheno-^aminoojclohoptaflMy Oj^Hi^N'OO'CgHg.
This compound was prepared by treating the ^base with benzoyl
chloride by the Schotten-Baumann method. It was immediately precipi-
tated in crystals, and the crude product, when merely washed with
water and dried, melted at 174 — 175° It crystallised from alcohol in
long, lustrous needles, very similar in appearance to the crystals of the
benzoyl derivative of the dlrhe^^ (Kipping and Hunter, loc. cU,) ; the
Digitized by
Google
582 kippiKg aKd huktbr: t^e resolution of PHENO-
two compounds^ however, differed iiH melting point, the derivative of
the ^base melting at 176 — 176^, the optically inactive compound at
171—172°, and a mixture o£ the two from 160—166°
The benzoyl derivative of the ^baee is dextrorotatory in methyl
alcoholic solution ; it is insoluble in cold water, but moderately easily
soluble in cold chloroform.
Picraiea of dl' and of l-Pheno-a-amtnoojclohepiane.
The picrate of the dl-hsise is obtained as a precipitate, mixed^
apparently with a little normal tartrate, on adding a hot aqueous solu-
tioh of picric acid to a similar solution of the cK^-tartrate' ; it crystal-
lises from water in yellow prisms and decomposes at about 205°.
The picrate of the Z-base, prepared from a solution of the hydrogen
tartrate of the Z-base in a similar manner, crystallises from water in
well-defined, yellow prisms and decomposes at about 185°.
The decoktiposing points of both these picrates depend greatly on
the rate of heating, and the compounds are consequently of little use
for purposes of identification ; they are both moderately easily soluble
in methyl alcohol, but only very sparingly so in cold water.
Salts qf d'Phmo-a-ammocjcloh^tane.
d'Pheno-a'aminocycloheptane TaHraie. — ^As a rule, it is not very
easy to isolate both the bases by crystallising the product of the com-
bination of a dl'hoBe with an optically active acid, as, after separating
the more sparingly soluble salt of one of the bases, there remains a
mixture which usually does not lend itself to further fractional crys-
tallisation. In the case of ciZ-pheno-a-aminocycJoheptane, however,
owing to the unusual behaviour of its tartrates, salts of both bases
can be obtained in a pure condition by one series of crystallisations.
Starting with an aqueous solution of the cZZ-base in a large excess of
tartaric acid, the hydrogen tartrate of the ^base is first isolated as
already described ; the mother liquors from this salt give, on evapora-
tion, a deposit which seems to consist of a mixture of both the
hydrogen tartrates with free tartaric acid, but on separating this
deposit from the remaining aqueous solution of the acid and again
dissolving in water, crystals of the normal salt of the dl-hsae are ob-
tained, because the relative quantity of free acid in the solution has
been diminished. If, now, the normal, partially compensated salt be
separated as far as possible, the mother liquors yield, on evaporation,
ft mixture of tartaric acid and the normal or hydrogen tartrate of this
d-base ; this normal salt is finally obtained in a pure condition by
further fractional crystallisation.
This salt crystallises from water in concentrically grouped needbi
Digitized by
Google
A^AMnfOCTCLOBKFtAHK UtfO ITS OPTICAJL ISOMKRIDKS. 583
mr fnumsy has a nentnl nrntHoa to litBraa, and is anhydrous ; ii melts
and denmiposes at abont 216 — 217^, bat unless carefollj dried its
melting point falls to about 21(f, It is less readily soluble in water
than the hydrogen salt of the dAmab, and for this reason it is obtained,
and ean be erystallised, from scdntions oontaining free tartaric add ;
in this case, theref ore, the beharionr is similar to that of the tartrate
of the dl-YmBe^ and the formation of this normal salt^ on ciystalliBing
such acid solutions, may be represented as foUows : 2 dA^dB^
d'Pheno-a-aminocydoheptane Hydro^n TwrtraU, — This salt is de-
posited when excess of tartaric acid is added to a hot solution of the
normal salt of the c^base and the solution concentrated if necessary ;
it can also be obtained by b<nling a solution of the normal salt, when
hydrolytic dissociation occurs and half the base volatilises with the
steam. It forms long, slender needles, melts and decomposes at
205 — 206°, and seems to be more readily soluble in cold water than
the other salts, except^ perhaps, the normal tartrate of the ^base.
d'Fheno-a-aminocj^ohapkme ffydroMaride, — That the ^two salts
just described are really those of the cl-base was proved by decompose
ing the normal salt with sodium hydroxide, distilling in steam, and
evaporating the distillate with hydrochloric acid ; the hydrochloride
thus obtained was identical with that of the M>ase in ordinary pro-
perties, but its solution in water was dextrorotatory.
0*241 gram dissolved in water, the solution diluted to 25 cc. and
examined in a 200 mm. tube, gave a +0*42°, whence [a]]> +21*8°.
Ck>nsidering the unavoidable experimental errors in dealing with
such low specific rotations, this result agrees satisfactorily with that
obtained in the case of the corresponding salt of the ^base.
Tartrates qf dl-ffydrindamine.
The tartrates of c^hydrindamine, as already stated, show little, if
any, analogy with those of pheno-a*aminocye^heptane. The hydro-
gen tartrate is easily obtained by mixing the base with an aqueous
solution of one molecular proportion of the acid and then evaporating ;
it forms masses of small needles or prisms, and is moderately soluble
in cold water but only sparingly so in cold alcohol :
1*5285 lost 0-0905 HjO at 100° H,0-5'9.
OgHuNjC^H^O^HjO requires HjO - 5*98 per cent.
The anhydrous salt melts and decomposes at 168 — 169°.
The principal points of difference between this salt and the corres-
ponding derivative of pheno-a-aminoqyo?oheptane are, firstly, that this
Digitized by
Google
584 PHENO*a-AMINOCYCLOHEPTAKE.
compound crystallises unchanged in absence of tartaric acid, whereas
the mixture of d- and ^pheno-a-aminocyc^oheptane salts gives the
normal salt of the cU-haae, and, secondly, that fractional crystallisation
fails to resolve the c^-hydrindamine salt into its components. A con-
siderable quantity of the hydrogen salt was crystallised from water a
great number of times, and the two end fractions were then
examined ; they were identical in melting point and in outward
properties, and on examination in the polarimeter were also found to
be identical optically.
0-3955 gram of air-dried salt dissolved in water, the solution
diluted to 25 c.c. and examined in a 200 mm. tube, gave a + 0*44°,
whence [a]D +13-9°.
The molecular rotation, calculated for the anhydrous salt, is there-
fore [M]d + 42°, and as that of c?-tartaric acid is [M]i> 42° in the
hydrogen metallic tartrates, it is obvious that the salt contains equal
quantities of the d- and Abases.
The normal tartrate of c^-hydrindamine, prepared by neutralising an
aqueous solution of the acid with the base, and allowing this solution to
evaporate spontaneously, crystallises in plates and melts at about 200° ;
it is much more readily soluble than the hydrogen salt and does not
crystallise very well. When its aqueous solution is boiled, hydrind-
amine volatilises, and subsequently the hydrogen salt is deposited.
I'Fheno-a-aminocjclohepktne Bromocamphoraulphanate.
Since it is with o{-bromocamphorsulphonic acid that (2^-hydrindamine
forms two well-defined partially externally compensated salts, it
seemed possible that the c^o2oheptane base would yield analogous
isomerides. On neutralising an aqueous solution of the acid with
dlrh&ae and allowing the solution to crystallise spontaneously, the
deposits were always of an oily consistency, and only very partially
solidified, even in cold weather. Such solutions were therefore not
examined further, but a separation of the bases with the aid of tartaric
acid having been obtained, the ^base was combined with bromo-
oamphorsulphonic acid.
The compound thus obtained crystallised with great facility from
aqueous solutions, even on the water-bath, and was deposited in highly
lustrous, striated prisms melting at 216 — 217°; it was only sparingly
soluble in cold water, but readily soluble in alcohol.
The specific rotation was determined by dissolving 0*45 gram in water,
diluting the solution to 25 cc, and examining it in a 200 mm. tube ;
a -1-1*72°, whence [ajo -1-47*8°. The molecular rotation is therefore
[M]i> +225*6°; taking that of the bromo-acid to be +270°, the moler
cular rotation of the base is [M]d - 44*4°. This result confirms those
Digitized by
Google
OOLOUAIKO MATT£B FBOM DELPHINIUM CONSOUDA. 585
previously obtained, and affords additional evidence of the stability of
the base as regards racemisation ; fractional crystallisation of the
bromocamphorsolphonate failed to reveal the presence of any
isomeride.
This work has been carried out with the aid of a Grant from the
Government Grant Committee of the Boyal Society; the authors
express their thanks for this pecuniary assistance.
Uniybbsitt Collsgb,
nottikgbam.
LXII. — Colouring Matter from the Flowers of Delphinium
Consolida.
By Abthur Gbobob Febkin, F.R.S.E., and Edward John Wilkinson.
In a previous communication (Trans., 1898, 73, 267) it was shown
that the flowers of Ddphinium zalU contain as glucosides iso-
rhamnetin, quercetin, and a trace of a third colouring matter. With
the desire at the time to study the latter compound more closely,
attention was directed to Delphinium CansoUda^ a plant more readily
procurable, in the hope that it might contain the same constituents.
Experiment showed that the blue flowers contained a moderate quantity
of yellow colouring matter which differed in composition from that
present in Delphinium zalil. This, judging from the melting point of
its acetyl derivative, appeared to be a mixture, and as attempts to
effect a separation of the substances failed, the subject was laid aside
for some time. Recently it was found that the colouring matter was
a single substance, and its investigation was therefore proceeded
with. Delphinium Caneolida is a common European plant belonging to
the Larkspur family ; its name refers to its powers, real or imaginary,
of healing or consolidating wounds.
Experimental.
The flowers themselves were first employed, but subsequently, to
economise time and labour, an extract was obtained by purchase from
Merck of Darmstadt.
Four hundred grams of this product dissolved in 4^ litres of boiling
water were digested at the boiling point with 30 cc- of sulphuric acid,
causing the separation of a light coloured, viscous deposit and a con-
siderable quantity of calcium sulphate. The hot liquid decanted from
this was treated with 100 c.o. of sulphuric acid, again boiled for an
VOL. LXXXL R R
Digitized by
Google
586 PEBKiy AND WILKINSON : CX>LOURINa MATTER FROM THS
hour, and allowed to stand overnight. A dark brown, resinous precipi-
tate containing the colouring matter was thus formed, and this ynM
collected on calico, washed with water, extracted with boiling aloohol,
and the extract evaporated to a small bulk. Addition of ether caused
the deposition of a tarry product, and the ethereal solution was
continuously washed with water until no further impurity separated
in this manner. On evaporation, a semi-crystalline mass of the emde
colouring matter was obtained, which was collected, well washed with
water, and crystallised from dilute alcohoL In this way, 500 grains
of the flowers gave 5*13 grams of crude colouring matter or approxi-
mately 1 per cent. The product was now converted into its acetyl
derivative, and this, when colourless, reconverted into the colounn^
matter in the usual manner :
0-1128 at 160° gave 0-2614 00, and 00319 HjO. 0-63-20; H = 3-14.
01086 „ 0-2503 OOj „ 00345 H,0. 0 = 6285 ; H= 3-53.
^16^10^6 '©quires 0« 62-93 ; H = 3'49 per cent.
It consisted of pale yellow needles resembling queroetin in appear-
ance, melted at 276 — 277°i was readily soluble in boiling alcohol and
dissolved in alkaline solutions with a pale yellow colour. Alcoholic
lead acetate gave an orange precipitate, and alcoholic ferric chloride a
greenish-black coloration. It contained no methoxy-^roup. Addition
of potassium acetate to its boiling alcoholic solution caused the separ;
ation of minute, orange-yellow, prismatic needles, which were collected
with the aid of the pump, washed with alcohol, then with a little water,
and finally with alcohol, and dried at 160"" :
0-3265 gave 0-0875 K^SO^. K = 12-01.
0-5085. „ 0-1345 KjjSO^. K= 12-04.
CijHgOgK requires K= 12-03 per cent.
This monopoUuHum salt closely resembles those of queroetin, morin,
&c. (Trans., 1899, 76, 433), and is decomposed by water with separ-
ation of the free colouring matter.
On adding sulphuric atiid to the substance suspended in boiling
acetic acid, a tulphate was obtained crystallising in orange-red, glisten-
ing needles :
01 159 gave 0-1985 OOj and 0-0354 H^O. 0 - 46-70 j H = 3-39.
^i6^io^6»^aSO^ requires 0 = 4687 ; H = 3-12 per cent.
The hydrochloride and hydrobromide were similarly prepared, but
were not analysed, as they suffer decomposition at 100°. The hydr^
iodide was somewhat more stable :
0-1148 gave 0-1863 00^ and 0-0310 H^O. 0 = 44-25 ; H = 3-00.
OigHioOoiHI requires 0 - 43-48 ; H « 2-65 per cent.
Digitized by
Google
FLOWERS OF DELPHINIUM CONSOLIDA. 587
Adion qf Bromine, — ^To the colouring matter gospended in aoetie
acid sufficient bromine was added to convert it into a tetrabromo-
derivative, rise of temperature being avoided. After standing for 3
days, the product was drained on a porous tile, washed with alcohol,
and purified by crystallisation from acetio acid with the aid of animal
charooal:
0-1177 gave 01492 CO, and 00190 HjO. 0 - 34-57 ; H « 1-79.
01526 „ 01944 00, „ 0 0185 H,0. 0-3476; H = 134.
O^HyO^Br, requires 0» 34-41 ; H- 1-34 per cent.
The frtftromo-compound crystallises in pale yellow needles, sparingly
soluble in boiling acetic acid and soluble in idkaline solutions with an
orange-yellow coloration. It melts at 275 — 277°.
The colouring matter has thus the formula O^^H^^O^, and from its
reactions is probably a member of the quercetin group.
The (usetyl derivative, prepared by heating the colouring matter with
one part of anhydrous sodium acetate and three of acetic anhydride,
crystallised from methyl alcohol in colourless needles. These, when
heated, commenced to melt at 116°, became completely fluid at 120°;
on further heating, however, gradual solidification ensued, and the pro-
duct subsequently melted at 181 — 182°. When ethyl alcohol was
employed, this preliminary liquefaction was not so pronounced, and
only a slight sintering at 120° was observed. Investigation showed
that no loss in weight occurred during the preliminary fusion, and the
resolidiGed product was not viscous as would be expected if an impurity
of low melting point were present. The amount of impurity, if present,
must have been infinitesimal, and all attempts to eliminate it were
unsuccessful ; moreover, there is evidence to show the product was a
pure substance :
0-1065 gave 0-2368 00, and 00400 H,0. 0 = 60-64 ; H = 4-17.
C^H^fiiQ requires O = 60-79 ; H = 3-96 per cent.
Acetyl determinations in the usual manner gave the following re-
sults, indicating that the compound was a tetracetyl derivative :
0-9453 gave 0-6018 Oi^Hi^O^,. O^jHioO^j = 6366.
4-2595 „ 2-6560 OijHjoOe. Oi^HioOej- 62-36.
0„HgOe(0,H30)4 requires OijHjoO^ = 63-00 per cent.
Fuaum with alkali a.t 200—220° gAve phlorogludnol (m. p. 210°) and
a crystalline acid melting at 208 — 210°. The identity of the latter
with p'hydrooDybenzoio add was confirmed by analysis :
0-1114 gave 02480 00, and 00394 H,0. 0-6071 ; H-3-93.
O^H^Os requires 0 « 60-86 ; H - 4*34 per cent.
Dyeing Propertiee, — In investigating these, mordanted woollen cloth
B B 2
Digitized by
Google
588 PEBKIN AND WILKINSON : COLOURING MATTEB FBOM THE
wafl employed, and the shades obtained are described^ together with
those of morin, for the sake of comparison :
Chromium. Alaminiam. Tin. Iron.
Morin OHve-yeUow. DuU yellow. Bright yellow. /^**^^]^*'
In this respect, therefore, these colouring matters are almost iden-
tical, morin, however, being stronger to a slight extent.
The reactions of the colouring matter C^sHj^O^ harmonised closely
with those given by Gordin {Disi., Berne) to kampherol, which he
prepared from the monomethyl ether, kampheride, contained in galanga
root {Alpinia officinarwn). Thus the melting points of these compounds
and those of their acetyl derivatives are practically identical, and the
decomposition products are the same in both cases. To kampherol,
Kostanecki has given the following constitution as probable {Ber.^
1901, 34, 3723) :
O
OH^ V >i C >0H
it being thus the connecting link between apigenin (Trans., 1897, 71,
805) and quercetin. An unsuccessful attempt was made to procure a
copy of Gordin's dissertation, but an extract evidently taken from it is
given in Der Chemie der ncUHrliehen Fa/rhsioffe^ 1900, p. 80. In this
description, the only distinction of importance between the colouring
matter under discussion and kampherol is to be found in the account
given of the dyeing properties of the latter, regarding which the
following sentence occurs : " K&mpferid sowohl wie Kampferol forben
Thonerdebeize schwach gelb an." As this is not in harmony with the
above results, it was necessary to prepare some kampheride, and from
this kampherol.
In isolating the constituents of galanga root, its ethereal extract
was treated according to the method given by Giamician and Silber
(^er., 1899, 32, 861). Possibly the quality of the root varies, for the
process employed by these authors was not entirely satisfactory. Thus
the extract diluted with benzene gave a semi-crystalline precipitate,
devoid of kampheride, which, after purification, melted at 292 — 295^,
gave an acetyl compound melting at 175 — 176^, and was identical
with the galangin monomethyl ether described by Testoni (0d«Mtta,
1900, 30, ii, 327-7-329). The filtrate, on addition of light petroleum^
deposited a brown tar, and this, on solution in hot chloroform, gave, on
cooling, a precipitate containing kampheride and galangin, which wart
Digitized by
Google
FLOWBHS OF DSLPHINIUM CONSOLIDA. 589
separated by fractional crystalliBation from alcohol. The kampheride
thus obtained melted at 223 — 226°, and the kampherol prepared from
it at 271 — 272°. The latter possessed well marked dyeing properties,
identical in all respects with those of the colouring matter of
Ddphinivm CoMoUda^ and there can be no doubt that the two are
identical
A study of the acetyl deriyative of kampherol from kampheride
showed that this, when crystallised from methyl alcohol, had the same
double melting point as that described above; further, the acetyl
compound of kampherol obtained from robinin (this vol., p. 475)
behaved similarly. It thus appears that this is a definite property of
the substance.
Molisch and Goldschmiedt {MancUah,, 1901, 22, 679) have recently
described a colouring matter, scutellarein, which exists in the form of
its glucoside, scutellarin, in the Scutellaria cUtistima. This has the same
formula and general reactions as kampherol, and yields the same
decomposition products, but melts at above 300°. To be certain that
the melting point of kampherol here given was correct, a sample was
treated in numerous ways; for example, with hydriodic acid, then
converted into its sulphate and into its potassium compound, then
crystallised from acetic acid, ^., but the melting point was practically
unaltered, the final product melting at 276 — 277°. Scutellarein and
kampherol, therefore, cannot be identical.
Attempts to isolate the glucoside of kampherol which exists in the
Delphinium Canaolida have not yet been successful, but the results show
that this compound is not robinin.
The Dyeing Propertiee of same Member $ of the IClavane Oraup,
Being in possession of the galangin prepared as above, we studied
its dyeing properties, with the object of gaining further insight into
the functions in this respect of the various hydroxyl radicles contained
in the compounds previously studied. Woollen cloth similarly
mordanted (mordanted calico, so frequently employed by others, is of
little service for comparing the members of this group) was employed
in each case, and the experiments were carried out, as far as possible,
in an identical manner. The positions of the hydroxyl radicles are
indicated by the numbers in the following formula :
Digitized by
Google
690 OOLOUBINQ MATTBR FROH DSLPHINIXm OONSOLmA.
Ghromiam.
Alnininiain.
Tin.
Lran.
Chryain, CigHgO,(OH)a
Yellow, faint
orange tint.
Pale yellow.
Undyed.
Pale
chocolata
Apigenin, Ci,HA(OH), ..
[6:7:4']
»>
Pale yellow,
alightly
atronger.
tt
Chooolate-
brown.
LuteoUn, CuH,0,(0H)4 ...
Brown-
orange.
Orange-
yellow.
Bright
yellow.
OUye-black.
GalaDgin, C„HeO,(OH), ...
[6:7:8]
Olive yellow.
Yellow.
Lemon-
yellow.
Deep olive.
Kampherol, C,BHeO^OH)4.
Brown-
yellow.
It
Lemon-
yellow.
Deep olive-
brown.
Uor^^B.^0^^1.
Olive-yellow.
Dull yellow.
Bright
yellow.
Deep olive-
hrown.
QnerMtin, CuH,0«(OH), ...
[6j7:8:8':4']
Red-hrown.
Brown-
orange.
Bright
orange.
OUve-black.
"^"*l6^'=*?^:5hl
t%
»»
Bright red-
orange.
OUre-black.
[6:8:8^4']
19
>>
Bright
orange.
Deep olive.
Fiaetin, (\,Bfi^On),
»>
Reddiah.
brown-orange
Bright red-
orange.
Olive-black.
Certain other colouring matterv, as apigenin monomethjl ether
(Trans., 1900, 77, 430), kampheride, galangin monomethjl ether, tio-
rhamnetin, rhamnazin, Ac,, could be added to this list, but are omitted,
as the above form a complete series. Only five of these colouring
matters, luteolin, fisetin, rhamnetin, quercetin, and myricetin, con-
tain two hydrozyl radicles in the ortho-position relatively to one
another, a condition which Liebermann and Kostanecki noted to be
essential in the case of the anthraquinone dyea]^u&.
Although this factor, doubtless, exercises a strengthening effect in
the dyestuffs of the flavone series, it is not a necessity, as is specially
evident in the cases of morin and kampherol, the former of which,
occurring as it does in old fustici is still most extensively employed
in the dyeing industry. An interesting point is the marked increase
in tinctorial property associated with the presence of hydroxy] in
position 3, as is at once observed between chrysin. and galangin,
fipigenin and kampherol, and no doubt exists between lotoflavin,
Gi5H«0,(OH)4 [5:7:2': 4'] (Dunstan and Henry, Froo. Roy- Soe.,
1901, €^ 374), and morin; this is again to be seen in the case of
luteolin and quercetin, for its presence in the latter gives strength
and redness to the shades. It will be noted from the above examples
that a multiplication of hydroxy] radicles in the flavone series does
not by any means exert such a marked tinctorial effect as is observed
in the anthraquinone group (compart", for instance, alizarin, anthra-
gallol, and alizarin-oyanine R) ; tliis is evident by the almost identical
character of the sliades given by apigenin and clirysiD, by fisetin,
Digitized by
Google
BlTNtHlSIS OlP IMINO-BTHBRS. N-ARTL6ENZIMIN0-BTHEBS. Sdl
qtieroetin, rhamnetin, and myrioetin, and by galangin, kampherol, and
morin. The introduction of hjdrozyl in the positions 4' and 4/ : 2',
in galangin, with the respective formation of kampherol and morin,
certainly exerts some strengthening effect, but on the other hand
fisetin, quercetin, and myricetin do not materially differ in strength
of shade. It is likely, however, that hydrozyl radicles in ortho-positions
relatively to one another in the nucleus I (p. 589) would exert a more
marked effect on the shade, for the benzylideneanhydroglycogallol,
OH
OHf^
'Qc^^-^=0
(Friedlander and Rudt, Ber., 1896, 09, 878), a member of a closely
allied group, yields powerful and characteristio colours. The dyeing
properties of these and other phenolic compounds are intimately con-
nected with their property of forming monosubstituted salts (Trans.,
loo, eU,) ; this question will be discussed in a further communication,
which it is hoped will be laid before the Society at an early date.
A preliminary investigation of the flowers of Frunus tpinoaa has
shown that these contain an apparently new colouring matter, which
in its general reactions resembles kampherol. This will be further
investigated.
The authors express their thanks to the Research Fund Com-
mittee of the Chemical Society for a grant which has been in
part employed to cover the expenses of this research.
GLOTHWoiuaRs' RxsxABOH Labobatobt,
Dtxing Dxfabtment,
TOBKSHiaX GOLLBOB«
LXIII. — Synthesis of Imino-ethers. ^'ArylhenziminO'^
ethers.
By G. D. Lander.
Although aryl substituted benzimino-ethers may be prepared from the
benzoylated amines by means of silver oxide and alkyl iodides (Trans.,
1901, 79, 698), the yields are poor, partly owing to the relatively
sparing solubility of the benzoyl compounds. It therefore became
desirable to find a more economical process, by which imino-ethers of
this class can be prepared in larger quantities for the purpose of
Digitized by
Google
692 LANDfiB: SYNTHESIS OF DilNO-fiTHBItS.
studying some of their properties, and especially a process in whieh
the chance of the simultaneous formation of both N- and O-ethers,
which seems to be characteristic of methjlation by means of silver
oxide, is either absent or at a minimum.
This ' molecular rearrangement in the course of synthesis ' of methyl
ethers by means of silver oxide and methyl iodide is not so character-
istic of the benzoylated as of the acetylated arylamines. Whilst the
methylation of aceto-o-toluidide by this process leads to the formation
of about equal proportions of the isomeric ethers, and that of aceto-|>-
toluidide to the almost exclusive formation of the N-ether, with
benz-(^toluidide scarcely any N-ether is formed, and with benz^
toluidide about equal amounts of the two isomerides are produced.
The possibility of preparing imino-ethers from imide chlorides and
sodium alkyloxides, in accordance with the general reaction,
R'CCIINR" + NaOR'" = R'C(0R'"):NR" + NaCl,
appeared most promising; moreover, on a priori grounds, it was
probably a reaction of simple replacement, and consequently one in
which molecular rearrangement would not occur. The application of
this mode of synthesis is not new. . It has been used, among other
instances, by Hantzsch {Ber., 1893, 26, 927) in the preparation of
N-phenylbenzimino-phenyl ether, and by Lengfeld and Stieglitz {Amar.
Chem. J., 1895, 17, 98) in the formation of ethyl Mocarbanilide from
carbodiphenylimide monohydrochloride (the imide chloride corres-
ponding to diphenylurea). I have found that the reaction is very well
adapted to the synthesis of aryl substituted benzimino-ethers, and that
methyl ethers are obtainable with as great ease as the ethyl derivatives,
no isomerisation appearing to take place.
Our knowledge of the imide chlorides is due mainly to Wallach
{Anruden^ 1876, 184, 1), who' has shown how they are derived from
substituted amides by the action of phosphorus pentachloride, the
group -CO-NH- giving first -C01,-NH- and then -CCKN-. The use
of imide chlorides derivable from amides, as distinguished from
compounds such as CeHg-NICCl-NH-O^Hg, C^Hj-NICCl^ and N:C'Br,
which, although belonging to the same category, are not usually
prepared directly from an amido-compound, is restricted as a practical
meUiod of imino-ether synthesiB, apparently, to the benzoylated aminea
Its application to the preparation of phenyl, o-tolyl, and p-tolyl
substituted benzimino-ethers iB described in this paper.
Acetanilide imide chloride is difficult to prepare, as it passes at
about 50° into the chlorinated base Cj^Hj^N^Cl, which is resolved by
alcohol into two mols. of acetanilide (Wallach, loe. ct<., 86).
In the hope that sodium ethoxide might cause the resolution of
this compound into molecular proportions of acetanilide and N-phenyl-
Digitized by
Google
N-ABTLBXNZIMINO-ETHBBS. 593
aoetimino-ethyl ether, I caused a solation of sodium in ethyl alcohol
to act on a benzene solution of the base ; the main product was,
however, diphenylethenylamidine, probably formed by the conversion of
the chlorinated base by the action of the ethoxide into the chlorine-
free base Gi^Hj^Nj, and the subsequent change of this into the amidine
(compare Wallach, loe. cU,).
The imide chloride of t-benzoylphenylhydrazine (Pechmann, j^er., 1894,
27, 322) gave, on treatment in benzene solution with: alcoholic sodium
ethoxide, a deep brownish-red, tarry substance, possibly a formazyl
derivative, which was not further examined.
EXPBBIMBNTAL.
, I. N'Phenylhenzimin<he there.
Formation of N-Phenylbenzimtno-eihyl Eth&r from BenzanUids Imide
Chloride
In the earlier' experiments on the formation of this imino-ether,
the benzanilide imide chloride, prepared as directed by Wallach (loc.
eii.), was purified by distillation in a vacuum, a procedure subsequently
found to be unnecessary. In the first instance, the action of dry sodium
ethoxide on a solution of the imide chloride in light petroleum was
tried. The liquid acquired a dark brown colour, but after several
hours' boiling the reaction was incomplete. Most of the solvent having
been distilled off, the residue was extracted with water, and on distil-
lation in a vacuum, a small quantity of a yellowish-brown liquid
having all the properties of N-phenylbenzimino-ethyl ether was ob-
tained. The employment of an alcoholic solution of the ethoxide leads,
however, to the production of imino-ether, both rapidly and in excel-
lent yield. The method of preparation finally adopted may be des-
cribed.
Fifty grams of benzanilide and 51 grams of phosphorus pentachloride
were fused together, warmed gently until the evolution of hydrogen
chloride had ceased, and the phosphorus oxychloride distilled off in
a vacuum. By extraction of the residue with light petroleum and
filtration, a somewhat opalescent but practically pure solution qi the
imide chloride was obtained.
A solution of 5*5 grams of sodium in the requisite amount of ethyl
fdcohol was added to this solution, which was cooled, the addition
requiring only a few minutes, and the resulting product having an
alkaline reaction. The sodium chloride was extracted by water, the
petroleum solution filtered from some benzanilide, dried with calcium
chloride, and the product distilled in a vacuum. By this means,
41*5 grams of N-phenylbenziminu-ethyl ether, boiling at 175 — 177^
Digitized by
Google
594 LANDER: SYNTHESIS OF IKINO-ETHERS.
under 16 mm. pressure, was obtained. The boiling point is 172^
under 15 mm., 168 — 170^ under 14 mm., and 220 — 230^ under from
20 to 30 mm. pressure :
0-128 gave 7 c.c. moist nitrogen at 16^ and 762 mm. N»6*37.
O15H15ON requires N=s 6*22 per cent.
The compound showed all the properties of this imino-ether, readily
undergoing hydrolysis by dilute mineral acids into aniline and ethyl
benzoate^ and giving a sticky hydrochloride which evolved ethyl
chloride on being warmed or kept, leaving a residue of benzanilide.
A small quantity of a substance of high boiling point was invari-
ably formed in this as in other similar reactions. It remained as a
glassy mass after the imino-ether had been distilled. Unless in a
state approaching purity, it was exceedingly difficult to crystallise, in
spite of its high melting point. The formation of this bye-product was
considerably increased at the expense of the imino-ether by adding
alcohol drop by drop to a boiling solution of the imide chloride, to which
the calculated amount of sodium wire had been added, until the latter
was dissolved ; but thus prepared, the substance resisted all attempts
at crystallisation. The united residues of several imino-ether prepar-
ations were crystallised from a mixture of chloroform and light petrol-
eum, forming dense, pale yellow, prismatic crystals melting at
170 — 172^ without decomposition. On analysis :
01486 gave 0-4515 OOg and 00798 H,0. 0 = 82-86 ; H - 6 96.»
0-1278 „ 0-3874 OOg,, 0-0660 H3O. 0 = 82-67 ; H = 5-73.
0*1560 „ 10 c.c. moist nitrogen at 13° and 768 mm. N = 7*64.
01330 „ 8-4 c.c. „ 12° „ 768mm. N = 7-56.
A cryoscopic determination of the molecular weight in acetic add
solution gave the value 372.
These data agree best with the formula O^'H.^ON^* which requires
0-82-97; H-6-32; N = 744 per cent.
This is the composition of henzoyldiphenylbenzenykumdine,
PhO:NPh-N(OOPh)Ph, with which the substance proved to be identical.
By boiling it with moderately concentrated sulphuric acid, it was re-
solved into benzoic acid and diphenylbenzenylamidine. The benzoyl
derivative of the latter compound, prepared either by allowing benzoyl
chloride and the base to interact in benzene solution, or by the Schotten-
Baumann method, possessed the same crystalline iippearance, colour,
and melting point as the compound from the imide chloride, and gave
Nb7'56, instead of the calculated 7*44 per cent.
It is exceedingly probable that the formation of benzoylated
amidine from imide chloride takes place in the following manner. The
♦ Analysed by Mr. G. Clarke, A.I.C.
Digitized by
Google
N-ARYLBENZIMINO-ETHERS. 595
.alcohol used as a solvent for the sodium ethozide causes a partial
decomposition of the benzanilide imide chloride into benzanilide,
PhOClINPh -> PhC01(0Et)-NHPh -> PhOO-NHPh.
By the action of the sodium ethoxide, sodiobenzanilide would be
formed and yield the benzoylated amidine thus,
PhNIOClPh PhNICPh
PhCO-NNaPh "^ PhCO-NPh'
N-Bhanylb&nziminO'msthyl Ether.
By the interaction of a solution of the required amount of sodium in
methyl alcohol and a light petroleum solution of benzanilide imide
chloride from 25 grams of benzanilide, 18*5 grams of the imino-ether
boiling at 157 — 158^ under 12 mm. pressure were obtained :
0-1858 gave 108 c.c. moist nitrogen at 18° and 770 mm. N«6*79.
^14^13^^ requires N = 6-63 per cent.
The compound was resolved by dilute mineral acids into aniline and
methyl benzoate, and by anhydrous ethereal hydrogen chloride into
benzanilide and methyl chloride.
This iinino^ther has been prepared by Wislicenus and Gk>ldschmidt
{Ber,, 1900, 33, 1471) by an application of Lossen's method (Annalwk,
1891, 266, 138). They give 145—150'' under 8 mm. pressure as its
boiling point.
Benzoyldiphenylbenzenylamidine was a bye-product of the synthesis,
and in addition a small quantity of diphenylbenzenylamidine, melting
at 143 — 145% appeared to have been formed.
MethykUion of BrnzanUide by Methyl Iodide and Silver Oxide. — ^This
methylation was carried out for the purpose of comparison with the pro-
cess of alkylation of acetylarylamines by the same method. The product
obtained from 10 grams of benzanilide, 35 grams of silver oxide, and
42 grams of methyl iodide in 50 c.c. of benzene after 3 hours' boiling,
consisted of 5 grams of N-phenylbenzimino-methyl ether boiling at 163°
under 11 mm. pressure, and identified in the usual way, some unaltered
benzanilide, and between 3 and 4 grams of benzoylmethylaniline. The
last named compound was present in the part of the product which
boiled between 180° and 190° under 9 mm. pressure, and was not
obtained solid. It was freed from imino-ether by steam distillation with
dilute hydrochloric acid, the unaltered oil separated, hydrolysed with
moderately concentrated sulphuric acid, and the methylaniline thus
obtained identified by means of its acetyl derivative.
Digitized by
Google
596 LAKDEB: SYNTHESIS OF IMINO-BTHEB8.
N-Phenylbenzifnifno-prapyl Ether.
Tkui compound was prepared by the action of a solutioii of 3 grams of
sodium in fft-propyl alcohol on the light petroleum extract of the imide
chloride from 25 grams of benzanilide. It boiled at 180 — 182^ under
13 mm., and at 177 — 179^ under 11 mm. pressure. After three dis-
tillations, 10 grams of the desired product, displaying the usual
reactions of an imino-ether, were obtained :
01872 gave 9-2 c.c. moist nitrogen at 9° and 760 mm. N«5-89.
CjgHjyON requires N = 5*85 per cent.
11. N-o-Tolylhenzimino-ethers.
N-o-TclylbenzMnino-eihyl Ether.
A light petroleum solution of the imide chloride from 20 grams of
benz-o-toluidide was treated with 3 grams of sodium dissolved in ethyl
alcohol in the manner already described. After three distillations, H
grams of imino-ether boiling at 179 — 180° under 15 mm. pressure
were obtained :
0*2344 gave 11*8 c.c. moist nitrogen at 9"^ and 749 mm. ^ = 5*95.
Ci^HiyON requires N — 5-86 per cent.
As in the similar reaction between sodium ethozide and benzanilide
imide chloride, a residue of high boiling point remained after distillation
of the imino-ether. It displayed, however, less tendency to crystallise,
and in view of the very probable similarity between it and the product
already dealt with, was not further examined. The same remark is
true of the products obtained f r6m the imide chloride of benz-p-toluidide.
N-o-TclylbenzimivuMMthyl Ether, ■
The imide chloride from 20 grams of benz-o-toluidide^ by interaction
with the requisite amount of sodium in methyl alcoholic solution, gave
15 grams of the imino-ether boiling at 173° under 15 mm. pressure :
01495 gave 04375 OOj and 00925 H,0. 0-7981 ; H-6-86.*
CijHijON requires C = 8000 ; H = 666 per cent.
MethykUion qf Benzo-toluidide by Methyl Iodide and Silver Oxide.
— ^Interaction between 35 grams of silver oxide, 42 grams of methyl
iodide, and 15 grams of benz-o-toluidide dissolved by the aid of 60 c.c
of benzene, resulted in the formation of 8 grams of N-o-tolylbenzimino>
methyl ether boiling at 170 — 171° under 11 mm. pressure. Nearly all
• Analysed by Mr. O. Olarke, AI.O.
Digitized by
Google
N-ARTL6ENZIMIN0-ETHERS. 597
the remaining product was unaltered toluidide, but after this had been
removed as completely as possible by crystallisation from aqueous
alcohol, a small quantity of a thick oil remained. After separation,
hydrolysis with sulphuric acid and liberation of the base, a sufficient
quantity was obtained to give Liebermann's reaction, thus showing
that benz-o-methyltoluidide was a product of the reaction.
The influence of orientation on the properties of the substituted
benzimino-ethers is notably less well defined than in the cases of the
analogous substituted acetimino-compounds, indeed, scarcely any in-
fluence can be detected by the methods applicable to the latter sub-
stances (compare Trans., 1901, 70, 693). When cold dilute hydro-
chloric acid is added to the o-tolylbenzimino-ethers, solution occurs, but
the almost instantaneous appearance of a turbidity indicates the
commencement of the characteristic hydrolysis, and a platinichloride
cannot be prepared, even by the addition of alcoholic platinic chloride
to an alcoholic solution of the base. The hydrochlorides could not be
prepared by means of anhydrous ethereal hydrogen chloride.
I take this opportunity of making a correction and addition to the
description of the substituted acetimino-ethers formerly described
{loc, dt,). When freshly distilled acetyl chloride is added to a light
petroleum solution of N-o-tolylacetimino-ethyl ether, an oil is precipi-
tated, which solidifies on rubbing to colourless needles, fusing with
copious evolution of gas at 109 — 110° and leaving a residue of aceto-o-
toluidide. These needles are the hydrochloride of the imino-ether in a
purer state than the amorphous salt melting at 90 — 91° previously pre-
pared from the base by means of ethereal hydrogen chloride. On
analysis, the crystalline salt gave 01 = 16*42 instead of 16 '62 per cent.
Moreover, by the same mode of treatment N-phenylacetimino-ethyl
ether, the hydrochloride of which I have not formerly prepared, gives
that salt in a crystalline state ; it melts evolving gas at 100°. This
result of the action of acetyl chloride is easily explicable, bearing in
mind the difficulty of freeing this substance from traces of hydrochloric
acid. The reagent, however, has not proved suitable for the prepara-
tion of the hydrochlorides of substituted benzimino-ethers.
The only respect, then, in which orientation seems to influence the
formation of the aryl substituted benzimino-ethers is displayed in the
extent to which the isomeric N-methyl homologues are formed in
methylation by means of silver oxide (loc, oit.y 696).
III. N-i^Toylhenzimino^thera,
'!^-i^Tolylhenzim%n(h€thyl Ether,
The imide chloride corresponding to 20 grams of benz-p-toluidide
gave, on treatment with sodium ethoxide and four distillations of the
VOL. LXXZI. S S
Digitized by VjOOQIC
600 SILBERRAD: POLYMERISATION PBODUCTS FROM
Preparaiian of Ethyl DiazoaceMe.
The following method will be found most convenient where large
quantities of the ester are required. Five grams of sodium acetate are
dissolved in 2 litres of water in a 10 litre separating funnel ; to this
solution, 1 kilogram of the finely powdered hydrochloride of ethyl amino-
aoetate (for the preparation of this compound see Hantzsch and
Silberrad, loc. eU., p. 70) is added, and then 750 grams of sodium nitrite.
The mixture is shaken until the temperature has fallen to about 0^
Five c.c. of 10 per cent, sulphuric acid and ^ litre of ether are then added
and the whole again well shaken. During this period, the gradual solution
of the still undissolved salts cools the mixture and prevents the
reaction from becoming too violent. Ab soon as the action slackens,
the ethereal solution of ethyl diazoacetate is run off, fresh ether added,
and 10 per cent, sulphuric acid run in from time to time in small
quantities until red fumes are evolved. The ethereal solution is
then run off, added to that already obtained, washed with small
quantities of dilute sodium carbonate solution until the washings
assume a deep yellow colour and have an alkaline reaction. The
ethereal ^solution is dried by shaking with fused calcium chloride, and
freed from ether on the water-bath. The yield amounts to 770 grams
or 94*7 per cent, of the calculated quantity.
Iminoctzoacetamide (PaeiMiodiazocusetamide),
nh:c(conh,)*n:n-c(conHj):nh.
Iminoazoacetamide is obtained as a clear yellow, crystalline powder
when the purified ammonium salt (see below), dissolved in the least
possible quantity of water at 0^, is precipitated with an excess of
80 per cent, acetic acid also cooled to 0°. The precipitated amide, after
being washed with water and dried on a porous plate in a vacuum,
decomposes violently at 135 — 136°. On analysis, the following results
were obtained :
Found: 0 = 2844; H-3-70; N = 49*59.
(OjHjONs)^, requires 0 - 2823 ; H = 3-53 ; N = 49*41 per cent.
Theam»noniwnsalt,NH:0(0ONH,)-N:N-C(0ONHj):N(NHJ.— After
many experiments, the following method of preparation was found to
give the best results. One hundred and twenty grams of methyl diazo-
acetate are added to 1 litre of aqueous ammonia, saturated at 0°, and
allowed to stand in a well-closed bottle for 14 days at -15°. The
separated crystals are then filtered off and the mother liquor again
allowed to stand for 14 days at — 15°, when a further supply of crystals
Digitized by
Google
DIAZOACETIO ACID. 601
separates out. These are reduced to the finest possible powder and
well washed with small quantities of ice-<sold water, whereby traces of
diazoacetamide and ammonium bisdiazoacetate are removed. The
resulting product, on drying over sulphuric acid, loses no ammonia,
and melts with decomposition at 155 — 157^. On analysis, the following
figures were obtained :
Calcalated for Found. Calculated for
(C^HjONjjjNHj. Curtius and Lang. SUbcrrad. (CsH30Ns),2NHj.
0 = 25-67 26-49 2601 25*00
H= 4-81 4-92 4-98 6-19
N = 52-41 52-97 52-45 53-28
These results agree more closely with the formula (02H3ONs)^NH3
than with (0,H30N2)32NH8, a fact previously noticed by Curtius (Ber.^
1885, 18, 1291) ; he was, however, under the misapprehension that all
the polymerides of diazoacetic acid were termolecular, a mistake for
which an erroneous molecular weight determination by E. Wiedemann
{J. pr, Chem,^ 1888, [ii], 38, 541) is chiefly, responsible.
The Mver salt, 3NH:C(00NH2)-N:N-C(00NH2):NAg + AgNOj.—
The silver salt is obtained as a voluminous yellow precipitate by the
addition of silver nitrate to a solution of the ammonium salt. The
precipitate is extremely sensitive to actinic rays, and can be obtained
pure only by working entirely by red light ; on warming, it blackens
immediately.
After drying in a vacuum over sulphuric acid in the dark until the
weight became constant, the following results were obtained on analysis :
Calculated for Found. Calculated for
Ci,HiB09NMAg4. Curtius and Lang. Silberrad. C,HN4(NAg),(CONHa)„lJHaO.
0=14-39 — 14-29 14-52
H- 1-50 — 1-78 2-02
N- 26-57 25-58 26-00 25-40
Ag3c43-16 43-57,43-81 43-62 43-56
It will be seen that the sUver and the nitrogeu, which were the only
constituents estimated by Ourtius and Lang, agree equally well with
either formula ; the carbon and hydrogen, however, leave no doubt as
to which is correct.
This complex formula evidently represents a double salt, thus :
^wHifiOigNg Ag^ - 3NH:0(OONH3)-N:N-0(OONH2)NAg + AgNOg,
which is strictly analogous to other polymeric products of diazoacetic
acid (Hantzsch and Silberrad, loc. cit, p. 67).
Aetion qf Nitrogen Trtoodde (a). Frepofratum c^f Bisazoxi/aeetic Acid, —
Twenty grams of the finely powdered amide were placed in a dry flask,
cooled to 0^, and subjected to the action of nitrogen trioxide (prepared
Digitized by
Google
662 SILBEHRAD: POtYMERISATiON PRODttCTS J'ROM
from nitric acid of sp. gr. 1*38 and arsenious acid) until a sample of the
product dissolved completely in cold-water. The contents of the flask,
which had assumed a carmine-red colour, were placed on a porous plate
over caustic potash in a vacuum. On heating, the product decomposed
below 100^, and on analysis proved to be very impure (found N=«30'4;
calculated for C^fiQl^^, N = 27*4 per cent); Consequently the product
was converted into the sodium salt by neutralising it, suspended in ice^
cold water) with dilute caustic soda also cooled toO^« To the crude sodium
salt so obtained) a few drops of dilute nitric acid and then silver nitrate
solution were added, when the characteristic dark green silver salt of
bisazoxyacetic^acid was precipitated. This, on analysis, proved still to
be contaminated with a substance richer in nitrogen than bisazozy-
acetic acid. The product was therefore suspended in water and satur-
ated with sulphuretted hydrogen at 0°, following the method described
by Hantzsch and Lehmann {loc, cit, p. 3674) for the reduction of
bisazoxyacetic acid to bisdiazoacetic acid.
The product, freed from sulphur by extraction with cold absolute
alcohol, was shaken with excess of ammonia, filtered, and the filtrate
cooled to 0° and precipitated with dilute sulphuric acid. The pre-
cipitate, after drying on a porous plate, melted at 150°, and on analysis
proved to be bisdiazoacetic acid :
Found N « 2701. 0,H2N4(C02H)2,2H20 requires N = 2692 per cent.
Hence it is evident that the product of the action of nitrogen tri-
oxide on dry iminoazoacetamide was principally bisazoxyacetic acid.
(b) Freparatian qf Triazole, — Five grams of iminoazoacetamide sus-
pended in ice-cold water were saturated with nitrogen trioxide. The
solution assumed a carmine-red colour (indicative of the formation of
bisazoxyacetic acid), and on evaporation on the water-bath yielded a
yellowish, crystalline mass ; this, after repeated recrystallisation from
absolute alcohol, gave a product melting at 138°, which on analysis
proved to be triazole nitrate :
Found, 0 = 18-33; H = 3-22; N = 42-30.
OjH8N8,HN08 requires 0 « 18-18 ; H = 3-03 ; N = 4242 per cent.
Action qf Caustic Soda, (a) Preparation of Sodiwm Salt (1). — ^Dilute
cold caustic soda solution decomposed the ammonium salt with evolu-
tion of ammonia, giving the characteristic intense yellow colour of
the salts. On evaporation in a vacuum, decomposition occurred, hence
the sodium salt could not be isolated.
(5). FornuUion of BiBdiazoacetic Acid, — Ten grams of the finely
powdered substance were thrown, little by little, into a solution of 10
grams of caustic soda in 25 c.c. of water previously heated to 90°. At
each addition, considerable rise of temperature and violent evolution
Digitized by
Google
DUZOACETIC ACID. 603
of ammonia occurred. The contents of the flask assumed a deep
yellow colour and became semi-solid owing to the separation of sodium
bisdiazoacetate.
The pi'oduct was washed with alcohol, recrystallised from water, and
decomposed with dilute sulphuric acid, whereby bisdiazoacetic acid was
obtained. This was identified by its melting point, 152°, the colour
reaction of its silver salt (Hantzsch and Silberrad, loc. cit,^ p. 73), and
by a nitrogen determination :
Pound N = 26 -99; CjHjN4(COjH)j,2HjO requires N = 26-92 per cent.
Action of Barium ffydroonde. Preparation of Iminoazoaeetie
Acid. — ^The addition of barium acetate to a concentrated solution of
ammonium iminoazoacetamide produced no precipitate. If, however,
the solution of the ammonium salt be warmed with baryta water, a
pale yellow precipitate forms and ammonia is evolved.
The product is probably barium iminoazoacetate, fi.p/-M-TT\.pn*^^**
After washing and drying over sulphuric acid, it gave the following
result on analysis :
Found N= 18-36; Ba = 44-82.
C^HjO^N^Ba requires N- 18'28 ; Ba =: 44-52 per cent.
Attempts to isolate the free acid have not yet been successful.
Dilute sulphuric acid appears to bring about decomposition ; its action
was studied as follows : To 2 grams of the barium salt suspended in
ice*cold water, a very slight excess of dilute sulphuric acid was added.
On evaporation, only oxalic acid and hydrazine could be detected.
The oxalic acid was separated as the calcium salt, which, after drying
at 100°, gave the following result on analysis :
Found Ca« 27-11 ; CgO^CajHjO requires 27'40 per cent.
The hydrazine was isolated as benzylideneazine, which after crys-
tallisation from alcohol melted at 93° and gave the following result on
analysis :
Found N = 13-50 ; Cj^HijjNj requires 13*46 per cent.
Action of Ammonia. Preparation of Diazoacetamide a9id isoDiazo-
acetamide, — ^Five grams of the ammonium salt were dissolved in 100 c.c.
of 10 per cent, ammonia and the solution allowed to evaporate over
soda lime in a vacuum. A very porous brown product resulted, which
on exposure to air rapidly shrank to a dark brown, resinous mass. If,
however, the porous substance be treated with dilute acetic acid (with-
out opening the desiccator), complete solution occurs with evolution
of nitrogen (showing thereby the absence of any unaltered imino-
azoacetamide). On addition of benzaldehyde to this solution, a crys-
Digitized by
Google
604 SILBEREAD: POLTMEBISATION PRODUCTS FROM
talline magma rapidly forms. After recrystallisationy the product
melted at 93% and on analysis proved to be benzylideneazine :
Found C = 80-66; H = 600; N= 13-54.
Ci^H^^j requires C = 8077 ; H = 577 ; N = 13-46 per cent.
CoTUtUution of iaoDictzoacetamide. — ^The stability of the tetrazine
ring disarms any suggestion that this readily decomposable product is
a tetrazine derivative. Indeed, its ready solubility in water points to
its being an isomeride rather than a polymeride of diazoacetamide.
Further, the readiness with which it yields hydrazine indicates that
it is not a true derivative of diazoacetic acid, so that it appears
probable that the compound is Modiazoacetamide, or, more likely, its
ammonium salt, CONH,-C^^^"^*^*.
The evolution of nitrogen above referred to seems to indicate the
presence of derivatives of normal diazoacetic acid. The product of
the action of ammonia on a further quantity of iminoazoacetamide
was therefore exposed to the air, the resinous mass extracted with
cold absolute alcohol, and the yellow solution allowed to evaporate
in a vacuum, whereby beautiful, yellow, prismatic crystals were ob-
tained. After recrystallisation from warm absolute alcohol, the
compound melted at 114° with violent decomposition and proved to
be diazoacetamide.
Found N = 49-00 ; C^HjONj requires N - 49-41 per cent.
The yield is small, being about 0*4 gram from 10 grams of iminoazo-
acetamide ; the chief product of the reaction is evidently Modiazo-
acetamide.
Bisdiazoctceiamide,
The method described by Curtiusand Lang {J, pr, Chem,^ 1888, [ii],
S8, 543) from ethyl bisdiazoacetate (Hantzsch and Silberrad {loc. eU.^
p. 72) is by far the most convenient for the preparation of the sub-
stance. Unlike iminoazoacetamide, bisdiazoacetamide forms no
salts.
Aetion of Nitrogen Trioxide. — Treated as described in the case of
iminoazoacetamide, this substance gave rise to bisazoxyacetic acid
both when dry axkl in aqueous suspension. The product was Identified
* Since the oompletion of this part of the work, a paper has appeared by Hantzach
and Lehmann {Ber., 1901, 84, 2610), in which they describe a precisely similar
product as resulting from the action of ammonia on ethyl isodiazoacetate. This
appears to be a strong argument in favour of the above constitution ; it must, how>
ever, be stated that normal ethyl diazoacetate can also be almost quantitatively
converted into the same product by the direct action of ammonia. This will be
discussed in a subsequent communication.
1
Digitized by
Google
DIAZOAOETIO ACID. 605
by its characteristic green silver salt, which, however, could not be ob-
tained pure from this source, and was therefore further identified by
reduction with sulphuretted hydrogen, whereby bisdiazoacetic acid was
obtained melting at 152°. On analysis :
Found N- 26-89 ; 02H2N^(COjH)2,2HjO requires N = 26-92;per cent.
Neither in aqueous suspension nor in the dry state could any triazole
nitrate be obtained, although 50 grams of the substance in aqueous
suspension were treated with nitrogen trioxide.
On evaporation on the water-bath, large quantities of ammonium
nitrate separated. The product was dissolved in water and calcium
acetate added, whereby a precipitate of the oxalate was obtained.
Found Oa« 26*87; CjO^OajH^O requires Oa = 27*40 per cent.
The filtrate, after removal of calcium, was rendered neutral and
treated with copper acetate ; no precipitate of copper triazole formed,
although the solution was allowed to stand for several weeks.
Action of Alkalis, — As already observed by Cortius and Lang {J. pr,
Ch&m,, 1888, [ii], 38, 344), bisdiazoacetamide is easily saponified by
aqueous alkalis in a perfectly normal manner. Aqueous ammonia is
without action on the amide.
B'l^'Dihydrotetrazinedicarhoxylamide.
The instructions given for the preparation of TMlihydrotetrazinedi-
carboxylic acid (Hantzsch and Silberrad, loe. cit., p. 77) are not clear
owing to the accidental omission of an important sentence. The
directions should be as follows : 100 c.c. of ethyl diazoacetate are run
into a solution of 160 grams of caustic potash in 120 c.c. of water, as in
the preparation of potassium bisdiazoacetate {J. pr. Chem., 1888, [ii],
38, 634). A solution of 57 grams of caustic potash in 93 c.c. of water
is then added, and the resulting semi-solid mass well shaken and heated
on the water-bath in an open fiask until the greater part of the alcohol
formed by the saponification of the ethyl diazoacetate has been driven
off. The flask is then attached to a reflux condenser, and the heating
continued until an almost colourless, thick liquid results, a process
which requires about 48 hours. This product is shaken with alcohol
until the bulk of the caustic alkali has been removed. The thick, oily
residue is then allowed to stand in the cold, when the greater part
of the N-dihydrotetrazinedicarboxylic acid separates out as the potass-
ium salt. The crude product is then freed from adhering traces of
potassium bisdiazoacetate by recrystallisation from water.
For the preparation of the ester, the potassium salt is converted into
the silver salt, which after being washed successively with water, ab-
solute alcohol, ether, and finally benzene, is treated with the calcu-
Digitized by
Google
606 SILBEftRAD: BOLYMEftlSATlON PRODUCTS FROM
lated quantity of a 25 per cent, solution of methyl iodide in ben2ene.
The mixture is heated on the water-bath for 4 hours, filtered, and freed
from benzene at 100° in a vacuum.
The ester so obtained is a thick, syrupy , almost colourless oil) 'readily
soluble in benzene, alcohol, or ether. On standing, it shows signs of
becoming crystalline. This compound, together with certain some-
what abnormal derivatives to which it appears to give rise, is still
under investigation.
For the preparation of the amide, further purification is unnecessary.
One hundred grams of the crude ester are therefore dissolved hi
500 c.c. of absolute alcohol and treated with an equal volume of satu-
rated alcoholic ammonia. On standing, the solution slowly ' deposits
the amide as a white, crystalline precipitate consisting of tiny needles,
which after recry stall isation from warm water melt at 278® with de-
composition. The compound is almost insoluble in the ordinary organic
solvents, but dissolves fairly readily in hot water, from which it crystal-
lises in needles or long prisms which show a beautiful play of colours.
On analysis, the following results were obtained :
Found 0 = 28-11; H«3-72; N- 48*97.
C^HgOjNg requires 0 = 28-32; H«3-53 ; N« 49-40 per cent.
Unlike iminoazoacetamide, this compound gives no precipitate
with silver nitrate or even with silver acetate in the 'presence of
ammonium acetate; copper acetate behaves similarly. Mercuric
chloride, on the other hand, produces a pure white precipitate, soluble
in boiling water, from which it again separates on cooling. Both
hydrochloric and acetic acids dissolve the precipitate readily.
Mercurous nitrate gives nse to a white precipitate, soluble in acids,
^hich blackens on boiling. Nessler's reagent produces a white pre-
cipitate which turns pale canary-yellow on boiling. These compounds^
which N-dihydrotetrazinedicarbozylamide forms with mercury, aro
exceedingly stable ; even after boiling for half-an-hour with a 25 per
cent, solution of caustic soda in the presence of an excess of Nessler's
reagent, ammonia could not be detected, although the amide itself is
easily saponified on warming with a few drops of dilute caustic soda
solution. This, together with the fact that neither copper nor silver
salts produce precipitates, seems to indicate that in these mercury
salts it is the hydrogen of the OONH, group and not that of the
imino-group which is displaced, as in the case of iminoazoacetamide.
Baryta water in the cold has no effect, but on warming, ammonia is
given off and a thick, white precipitate of barium n-dihydrotetrazine*
dicarboxylate is produced. This salt was identified by the liberation of
the acid which melted at 287^ and contained K^^ 32*30 per cent
[O8H,N^(0O5H), requires N - 32-56 per cent.].
Digitized by
Google
DtA^OACETtC ACID. . 60?
Action of J^ttrogen Trioocide. — On treatment with nitrogen triozide,
no bisazoxyacetic acid could be obtained. Triazole is always the
principal product of the reaction. Its isolation and identification were
conducted precisely as described above under iminoazoacetamide,
the yield is 85 — 90 per cent, of the theoretical. The product melted
at 138°. On analysis, the following results was obtained :
Found N = 42-21, CgHjNgjHNOg requires »" = 42-42 percent.
Caustic potash, either in aqueous or alcoholic solution, readily
causes saponification of the amide. One gram of the amide was
warmed with a slight excess of a 2 per cent, solution until all odour of
ammonia had passed off ; the solution was then evaporated on the
water-bath until the potassium salt crystallised out, this, after re-
crystallisation from warm water, gave the following result on analysis :
Found N = 22-22; K = 31-34.
C^HjO^N^Kg requires N « 22*59 ; K = 31*45 per cent.
Constitution of Iminoazoacetamide.
From the analysis of its salts and the readiness with which the
compound is converted into tetrazine derivatives, it is obviously a
bimolecular polymeride of diazoacetamide, CHNg'CONHj. Its yellow
colour indicates the presence of an azo-group. The action of baryta
water shows the compound to be the true amide of an acid quite dis-
tinct from both bisdiazoacetic acid, C02H*CH<^:^>OH-CO,H, and
N-dihydrotetrazinedicarboxylic acid, COgH'C^v^^^o^C'COjH.
Further, the preparation and investigation of n-dihydrotetrazine-
dicarboxylamide has shown that the imino-groups in the tetrazine ring
possess properties totally distinct from those of iminoazoacetamide,
whilst the production of wodiazoacetamide on treatment with
ammonia indicates that iminoazoacetamide is a chain compound,
and does not contain the stable tetrazine ring at all. In short, the
only constitution which explains these apparently contradictory proper-
ties is that of iminoazoacetamide, NH:0(C0NH2)N:N-C(C0NHj):NH.
The action of ammonia is rendered clear by the following equation :
CONH,-C<^(^^lL^?J>C'CONHj -->
N- N *
The ease with which derivatives, of both C- and N-dihydrotetrazine
can be obtained is thus elucidated.
Digitized by
Google
608 POLYMERISATION PBODUCTS FBOM DIAZOACETIC ACID.
Baryta water gives rise to the corresponding acid,
nh:c(conHj)-n:n-c(conHj):nh + Ba(OH),
N-C(NH)CO,^
whilst warm caustic soda yields bisdiazoacetic acid,
CONH,-C^££5>C-CONH3 + 2NaOH
= COjNa-CH<2:5J>CH-00,Na + SNHg,
in a manner precisely analogous to its production from ethyl iao-
diazoacetate,
C02EfC<J^ + S5>C-00-Et + 2NaOH
* JN IS *
=. C0jNa-CH<^:^>CH-C02Na + 2EtOH.
The action of nitrogen trioxide affords additional confirmation of this
view, as by its means derivatives of either C- or N-dihydrotetrazine can be
obtained according to the conditions. The formation of bisazozyacetic
acid is indicated by the following scheme :
CONHa'C<^lLZ^>C'CONHa + Nfi^ ~>
COjH-CH<§_^>CH-COjH.
The formation of triazole necessitates the following steps (compare
Trans., 1900, 77, 1188, for formation of triazoles from «-dinitro8o-N-
dihydrotetrazine and its derivatives) :
(a) Formation of i;-dinitros(v-N-dihydrotetrazine,
OONH,'C<gJE-Jlg>C'OONH, + Sfi^
{h) Formation of triazole nitrate,
In conclusion, I wish to express my thanks to the Government
Grant Oommittee of the Royal Society for pecuniary assistance in
carrying out this work.
Digitized by
Google
609
ANNUAL GENERAL MEETING,
Maech 26th, 1902.
Professor J. Emebson Reynolds, M.D., Sc.D., V.P.R.S., President,
in the Chair.
The Pbesident declared the ballot open for the election of Officers
and Council for the ensuing year, Dr. Henbt and Mr. Ramaqe being
appointed Scrutators. He then said ; —
I have the agreeable duty to-day of congratulating the Fellows
on the continued and even increasing prosperity of the Society, as
indicated by considerable additions to its list of Fellows, by its con-
sequent gain in the means of promoting its objects, and by the extent
and character of the work published during the past year.
The numerical strength of the Society was 2335 on March 28th,
1901. Since that date 163 Fellows have been elected, and 3 have been
reinstated by the Council, making a gross total of 2501. Of these,
32 have withdrawn, 25 have been removed for non-payment of two
annual subscriptions, and 28 have died.
The actual number of Fellows to date is therefore 2416, the
highest number yet reached, and the number of Foreign Members is
32.
The names of those removed for non-payment of subscriptions
are : —
B. W. Allsom, W. D. Bohm, C. F. Branson, T. H. Coleman, E. D.
Ewen, F. G. Fuller, C. H. Field, W. G. Fraser, A. T. Gailleton, W. H.
Grieve, T. C. Hepworth, A. F. Hogg, J. Harger, W. Ince, R. S. Ladell,
R. D. Littlefield, D. C. Mackenzie, A. Mackay, F. L. Overend, M.
Pearson, S. Smith, A. H. Turton, C. W. Vincent, W. H. Walker, A.
Walton.
The following have withdrawn : —
J. M. Arnot, W. H. Barr, F. Belton, H. D. Berridge, M. Cochran,
W. Collingridge, J. Craig, H. L. Dampier, M. J. R. Dunstan, G. Evans,
H. P. FitzGerald, P. H. Grant, W. H. Greene, H. W. Gough, A. B.
Griffiths, J. B. Guyer, H. M. Hastings, E. S. Hayward, A. H. Mac-
donald, C. J. S. Makin, J. Maudsley, J. McLeod, H. C. Myers, G. A.
Parkes, L. G. Patterson, J. C. Quinn, A. Schloesser, C. Thompson,
E. A. Wates, J. I. Whimster, B. W. Winder, S. Wood.
The following have died : —
T. H. Aquino, F. J. Beale, J. H. Beckett, Henry Bird, Sir J. H.
Gilbert, F.R.S., A. Hartridge, Alexander Hay, Lawrence Hislop, Robert
Irvine, David Johnson, N. Leonard, H. G. Madan, William Martindale,
Digitized by
Google
610 ANNUAL GENERAL MEETING.
Dr. Ira Moore, Dr. G. Harris Morris, John Paul, Prof. H. von Pech-
mann, W. B. Randall, W. Shapleigh, Louis Siebold, Prof. Maxwell
Simpson, F.R.S., W. T. N. Spivey, W. Terrill, Andrew Thomas, John
Thomson, J. L. W. Thudichum, E. A. Warmington, G. F. Wilson, F.R.S,
The most important test of the prosperity of the Society is, how-
ever, to be found in the record of its work. In this respect also I
have a favourable report to make. Since the last anniversary, 181 com-
munications have been made to the Society. Abstracts of all these have
appeared in the Prootedings^ and 139 have already been published in
the Traneaciiona, I venture to think that the quality, generally, of
the work presented is as high as in any previous year, and clearly
indicates the continued enthusiasm and activity of the Fellows.
The Traruaetions for 1901 contain 146 memoirs, occupying 1411
pages; and the volume of the preceding year 127 memoirs occupying
1334 pages.
The volumes for 1901 contain 3754 abstracts of papers, published
mainly in Continental journals, occupying 1496 pages, arranged as
follows :
Paet I.
Pages. No. of Abatracts.
Organic Chemistry 784 1530
Part II.
General and Physical Chemistry 403
Inorganic Chemistry 376
Mineralogical Chemistry 169
Physiological Chemistry 363
Chemistry of Vegetable Physiology and Agri-
culture 306
AJialytical Chemistry 607
712 2224
Total in Parts I. and II 1496 3764
The volume for 1901 contains a Memorial Lecture giving an account
of the life-wo^k of Rammelsberg. A set of the Memorial Lectures
which had appeared up to the end of 1900 was issued in September
last in a separate form.
The use of the Library by the Fellows continues to show their ap-
preciation of it. Eight hundred and eighty books have been borrowed,
as against 810 during the corresponding period of last year. A large
number of these have been Journals issued by post to Fellows resident
in the country, and the Library Committee invite special attention to
this development of the Society's usefulness. The additions to the
Digitized by
Google
ANNUAL GENERAL MEETING. 611
Library comprise 153 books, 441 volumes of periodicals, and 33
pamphlets as against 95 books, 327 volumes of periodicals, and 30
pamphlets during the corresponding period of last year.
It is desirable that the Society should have in the Library one copy,
at least, of every work printed in English on chemical subjects to the
end of the 18th century, and I would invite the co-operation of the
Fellows in making the Library complete in this direction.
Li the preparation of the new Library catalogue, the printing of
which has been decided upon, advantage has been taken of the opportunity
to constructanew and convenient Card Catalogue, which, it is believed,
will materially assist readers in making use of the Library.
It is my privilege to offer, on behalf of the Society, our warm con-
gratulations to Dr. Schunck, to Mr. Lloyd Bullock, and to Dr. Francis, who
this year have reached their sixtieth anniversary of admission to the
Fellowship of the Society. It gives me pleasure to add that Mr. Buckton,
F.R.S., Mr, F. Claudet, and Mr. Darby have reached their jubilee,
and to them I would also convey our sincere congratulations.
Last year our illustrious senior Foreign Fellow, M. Berthelot,
celebrated the fiftieth anniversary of his first scientific publication, and
all countries united in expressing their admiration and respect for the
veteran chemist. On behalf of the Society, in company with Dr. Gladstone
and Professor Bamsay, I presented a congratulatory address to M.
Berthelot at the imposing function which was held in the Sorbonne on
November 24th, 1901. That address has already been printed in the
Frooeedings,
During the year, the Society had joined in the celebration of the
450th anniversary of Glasgow University, and in the jubilee of Owens
College, Manchester.
Considering the large number of Fellows now in the Society, the
mortality is small ; nevertheless, this year I regret to say we have lost
21 Fellows. This melancholy list includes the name of Sir Henry
Gilbert, Past President of this Society, and one ever devoted to its
welfare. His immense work, carried out with Sir John Lawes, laid
the scientific foundation of British Agriculture and serves as the model
on which all future researches must proceed. The Society was fully
represented on the sad occasion when Sir Henry Gilbert was interred,
and its representatives laid a wreath on his grave, whilst later on the
Council passed a vote of condolence with his mourning relatives. In
a short time, I hope, a full obituary notice will be published by
one far more competent to undertake it than I am. Dr. Maxwell
Simpson is another of those passed away from amongst us full of
years, leaving memories of good work well done, especially in synthetic
chemistry, and of him and of the other Fellows whose life-work has
elosed records will also be shortly published.
Digitized by
Google
612 ANNUAL GENERAL MEETING.
Considerable discussion has taken place within the Society on the
question of altering the day and hour of the Ordinary Meetings, which
was raised in the address of my predecessor, Dr. Thorpe, last year.
There have been two Extraordinary Meetings on this subject, and the
outcome is the experiment now in operation, of trying alternate
evening and afternoon Meetings until the end of the present session.
Until that experiment has been fairly made, the best course, obviously^
is to suspend judgment on the question.
Grants amounting to £250 have been made from the Research Fund
in aid of chemical investigations.
This is one of the occasions on which, as I venture to think, it is
not only permissible, but desirable, to consider some general question,
even of a speculative order, in the light of any new facts discovered in
recent years. I therefore propose to discuss as briefly as possible a
question of this character, and one which possesses high interest for
us, namely, whether the many and important additions to our know-
ledge of the chemical elements made during the last decade or so
have given us any clue to the nature of the relations existing between
them.
That the elements are related as a whole is now an axiom^ and
underlies all modem classification ; equally axiomatic is the statement
that periodicity can be traced to a large extent between the atomic
weights and properties of the elements.
The recognition of this periodic principle by Newlands in 1864,
when more fully interpreted by Mendel^eff in 1869, and by Lothar
Meyer on somewhat diiSerent lines, marked an important advance in
science. It served the highly important purposes of correlating a
large number of the facts then known, of stimulating research, and of
inducing closer scrutiny of the atomic weights. The " periodic law "
formulated by MendeUeff in asserting that " the properties of ^the
elements are periodic functions of their atomic weights " claims an
attractive universality, which gained for it much recognition.
Mendel^eff's valuable and interesting tabular classification of the
elements, arranged in accordance with the " law " and supported by
very ingenious and often cogent reasoning, contributed much to its
wide acceptance. Later on, the verification of Mendeleeff's brilliant
predictions confirmed confidence in the principle. Nevertheless, closer
scrutiny revealed difficulties in detail which gradually led to doubts as
to the general validity of the 'Maw,'' and these doubts have been
accentuated in recent years by the discovery of the non-valent elements
of Bayleigh and Bamsay, for which there seemed at first to be no
place in Mendel^eff's classification. It is true that the law neither
predicted nor excluded the existence of such elements, and that very
Digitized by
Google
Digitized by
Google
I
Journ, Chem. Soc.y
I
ts.
0 10^
hii|uii 1 1
NOTE TO BINDER.
This diagram is isfined to take the place
of the one on p. 613 (in the Journal for
Jane), which ia to be cancelled and re-
moted.
When the Journal is bonnd the present
diagram should, be pasted at the back of
the asterisks, and aflBzed to page 612, so
as to face p. 618.
Orthoperiodio L'Be
Macropenodic
ITodal He^nm
J^-:
Digitized by
Googk--*"'"
ANNUAL GENERAL MEETING. 613
doubtful positions have since been assigned to them. More serious
difficulties are, however^ to be found in the anomalous position given
to hydrogen ; in the fact that the atomic weight of argon places it
between potassium and calcium ; that tellurium has a higher atomic
weight than iodine, contrary to theory, and in other details which
will be specified later on. In consequence, the law is seen to be an
empirical rule which, in so far as it properly applies, is of considerable
assistance, but is probably only part of a wider generalisation than
any we have yet reached.
The difficulties above referred to are masked in the well known
table used in illustrating the connection between the atomic weights
and the chemical properties of the elements. That this table includes
much that is true in reference to the comparative properties of the
elements is undoubted, but it has of late been rather used as a con-
venient system of pigeon holes for elements which are docketed, as it
were, with special atomic weights and put away, without much refer-
ence to the relations in which they stand to other elements. The fact
is there is something beyond the table and beyond the '* law " of which
we must endeavour to form some conception, if we are to explain
existing difficulties. We must therefore get back to the axioms
mentioned at the outset, and see whether it is possible to form such
a mental picture of the relations subsisting between the elements as
shall gire us some clue to the nature of those relations, and serve to
explain the partial truth of the " periodic law."
The rough outlines of a picture of this kind were, in fact, sketched
before 1886,^ but, being rather cumbered by the more rigid notions of
valency which then prevailed, the result was not sufficiently clear.
Since then, however, important details have been filled in and the
discovery of the argon group of elements has given consistence to the
whole. The idea underlying the curve which I published in 1886 was
that of a vibrating system, based on our knowledge of the properties
and relations of the vibrations which physicists term "stationary
waves " ; but before entering into particulars it is well to define the
facte to be kept in view in working out the scheme.
First, and obviously, the atomic weights, on which so much excellent
work has been done in recent years, represent the prime group of facts,
and those used are the numbers given in the revised table which is
printed in the January number of the current volume of the Berichte,
and are based on the scale of 0 » 16. I have indicated these numbers by
dots, crosses, or dashes under the scale of equal parts given at the head
of the annexed diagram. This is instructive, as it indicates at a glance
the more obvious and important numerical relations of the atomic
* "On a Method of Illustrating the Periodic Law.*' By J. Emerson Reynolds,
Chem. News, 1886, 54, 1.
VOL. LXXXI. T T
Digitized by VjOOQIC
614 ANNUAL GEl^ERAti MfiETtl^a.
weights or active masses of the elements. As you know, several
'* laws " have been deduced from the study of these numbers, of which,
probably, the best example is Johnstone Stoney's *' logarithmic law/'
but we aire not directly concerned with any of these at present.
Secondly, there are the well known relations which may be shortly
summed up in the following way. The best defined elements from
lithium on> are naturally divisible into sections or ** periods/^ each includ-
ing seven members ; in these periods there is one member, the fourth,
which more or less distinctly subdivides it into two parts. Further^
within these periods the first, second, and third members in order of
atomic weight are essentially electropositive, and in order of valence 5
the fifth, sixth, and seventh members are essentially electronegative,
and of more variable valence ; while the fourth, or unique member,
marks a transition, and is generally tetravalent.
Thirdly, we have the very significant fact that, as comparing period
with period, and similarly placed elements within the periods, we find
the cUtemaU members are those which are most closely related in
chemical and physical properties. In order to mark this in the diagram
the alternate periods are distinguished by different signs— dots for one
set, crosses for the other.
Fourthly, in transition from one period to another there is abrupt
change of sign, in certain cases, from strongly electropositive to
strongly electronegative, as from fluorine to sodium, chlorine to
potassium, bromine to c»sium, and so on, and it is about such points
that we find the atomic weights of the singularly indifferent or neutral
elements of the argon group. On the other hand, where there is no
such abrupt change of sign from period to period, as from manganeee
to copper and from molybdenum to silver — ^that is, in the alternate
positions — instead of single non>valent elements, there are groups d
three individuals differing little in atomic weights, and all exhibiting
high valence comparable with that of the third, fourth, and fifth
members of the regular periods. These are the triplets, such as iron,
nickel, and cobalt; ruthenium, rhodium, and palladium, dEo., which are
marked on the scale in the diagram by dashes, and are included in
Mendel^eff's unsatisfactory '* eighth group."
So far these are matters of fact which are generally admitted, and
are stated above independently of any ** law '' or theory except the
atomic theory.
At a very early stage in teaching on the lines of the periodic law, I
came to the conclusion that the latter, even when aided by the tabular
classification, gave but an imperfect representation of the facts known
at the time. One of the clearest deductions from the evidence seemed
to be that this peculiar connection between properties and atomic
masses must be the outcome of something in the nature of a vibratory
Digitized by
Google
AKXtJAL Q£N£RAX MESTIKG. 615
relationship, for only in some sach way could all the facts he explained*
I therefore sought for some physical phenomena involving periodic
change which could help the imagination to form some corresponding
picture of the elemental relations, even though no real analogy in the
ordinary sense were traceable.
Of all the phenomena suggested for the purpose those vibrations
which are known as '' stationary waves " seemed the most suitable!
They can, as you know, be set up in a light cord fixed at one end and
attached at the other to one of the rapidly moving' limbs of a large
tuning fork, kept in regular vibration by an electro-magnetic arrange-
ment. The motion of the fork is transmitted to the cord with the
well known result of establishing a beautiful system of apparently
rigid loops and nodes. Each particle of the cox'd, except at the nodes,
travels in a circular path at right angles with the axis of the whole
system, the amplitude of the motion being, of course, greatest at the
crest of each loop or antinode.
The diagram shows twelve of these small loops which were plotted in
the following manner. The atomic weights of the unique, or carbon^
silicon j group of elements were marked off alternately at either side
of, and at equal distances from, the axial scale ; the points were then
connected by right lines and the intersection of the scale by each line
was taken to mark the node ; the loops were then drawn between the
nodal points so found. The two loops between 152 and 196 were
necessarily obtained by a kind of interpolation, which cannot, however,
be much out, as the result is checked by the lead period beyond. The
numbers given at the foot of the diagram show the relative lengths of
the loops from node to node. These increase up to the sixth loop and
then diminish again, just as would happen if the vibrations took place
in a medium of unequal density, or the density of the cord was
greater nearer the middle of its length than at either end. The
axial scale serves to mark off the positions of the atomic masses,
which latter are shown by dots and crosses placed on the curve
formed by the cord in one phase of vibration. Each dot and
cross rapidly rotates in its limited circle round the axis of
the whole system and the areas of the circles described increase from
the first to the fourth dot and, of course, diminish to the seventh, the
direction being the same as that in which the chemical properties of
the elements vary. At the nodes there is apparent rest.
Again, as the loops are compared, it is seen that those which are
adjacent are in opposite phases at any given moment, while the alter-
nate loops are in the same phase, just as similarly placed members of
alternate periods are found to be most closely allied in properties.
So far as our knowledge extended in 1886, this served as a good
T T 2
Digitized by
Google
616 ANNUAL QENEBAL MEErriNQ.
illustration, save in one particular which I shall presently deal with ;
but at that time, as I have already said, we were rather hide bound
by our conceptions of valency, and the zigzag curve I then published
shows this. Even in that form, however, I am glad to know it proved
useful in teaching, and that it served^ when somewhat modified in
form by Sir William Orookes, as a basis on which he has reared his
most interesting theory of the genesis of the elements.*
The unsatisfactory particular to which I referred in the last para-
graph was the difficulty in placing the members of the " eighth group.''
These triplets have atomic weights which place them about the
nodal points of the cUterfKUe periods, but, as you know, they are far
from being non-valent. On the contrary, they exhibit properties cor-
responding to the polyvalent members of the regular periods. It is
true that their compounds are, generally speaking, easily reduced, and
the elements themselves by no means active in the free state, but the
fact remains that they all exhibit high valency. The idea suggested by
these and other considerations was that they are members of another
series of elements harmonically related to the first series, somewhat as
shown on the diagram by the larger loops. From that point of view
they are *^ interperiodic," but in a new sense; and in one, moreover,
which is consistent with their appearance only about the alUmaU
nodes of the regular periods. Whether or not these lai-ge vibrations
should be represented as taking place at exactly half the rate of the
smaller ones, was a subject often discussed with my valued colleague,
the late Professor G. F. Fitzgerald, but we agreed that it is sufficient
for the general purpose in view to take the simplest relationship be-
tween the two sets of vibrations as shown in the diagram. The
difficulty about the hypothesis was that we had then no evidence of
the existence of any other elements which could be supposed to belong
to the special series, and the picture remained incomplete until the
discovery of argon and its allies supplied the necessary links. It had
already been foreseen that any other elements of the slow moving
triplet series which might exist would probably exhibit much feebler
chemical activity than those of any of the smaller periods, but the
existence of non-valent elements was not anticipated.
Once the latter were discovered, however, they were seen to be just
of the kind required to complete the picture. Their atomic weights
placed them at or very near to each of the nodes which is apparently
common to both vibrating systems, and their non-valence in contrast
with the antinodal triplets sufficiently accounted for the absence of
intermediate elements of the same series.
* Address to the Chemical Section of the British Assodfttion at the Birmingham
meeting in 1886, and Trans., 1888, 58, 487.
Digitized by
Google
ANNUAL GENERAL MEETING. 617
I must DOW invite your attention to another matter of interest
which bears on the disputed question of the position of hydrogen in
relation to the rest of the elements.
I have already described the means adopted for deducing the nodal
points of the minor periods from the atomic weights of the members
of the carbon-silicon group. But between carbon (12) and zero, the
symmetry of the particular loop necessarily guided its continuation
and fixed its node at 4 on the scale, or slightly in advance of the
number which was subsequently found to represent the atomic weight
of helium. At the time, it was difficult to understand this exceptional
shortening of the loops, but on making the experiment with a cord
some 4 metres in length thrown into " stationary waves," I found
that when partially stopped with the finger so as to compel the forma-
tion of a node at some 4 units from the vibrator, the nearest loops
were found to be shortened relatively and, approximately, in the pro-
portions indicated on the diagram, a result which naturally increased
confidence in the value of the picture as a whole. In thus definitely
fixing the node at 4 units of the scale, the curve between that point
and the vibrator necessarily assumed the form shown on the diagram
which represents the "cut-off" end of the loop. Now, hydrogen
must find its place here at 1*008 of the scale; therefore, from our
point of view, hydrogen seems to be the last member of a period
rather than (as supposed by Mendel^eff) the first of a seven-member
period, of which six are still unknown. On general grounds also,
the facts now known support this view to which we have been led
as to the position of hydrogen, though I am inclined to think that
the rdle of that element in nature is far more important than that
of a typical halogen ; but I shall return to this point shortly.
Having now worked in most of the details of the picture, we
can next consider its general effect. Before doing so, however, it
is convenient to designate as :
Orthopertodic. — The members of the twelve minor periods.
Afacroperiodic. — The triplets represented about the antinodes of the
greater periods.
Nodal, — ^The elements of the argon type.
Broadly speaking, this scheme represents the atomic masses as form-
ing a dual vibrating system, the two parts of which exhibit apparently
simple harmonic relations. Whether or not these relations are quite
so simple as they appear to be, there is no doubt that the system as a
whole also gives some evidence of still wider periodicity, as the lengths of
the middle loops are somewhat greater than those on either side.
From our point of view, hydrogen, so far from being limited in its
analogies as the end element of a period, seems rather to be the
Digitized by
Google
618 ANNUAL GENERAL HEBTINa,
elemental material which is a type of hoih atomio series, and therefore
cannot be claimed as the first term of any single groap of elements.
Of the nodal elements, helium marks the node common to the two
vibrating systems, and the strong contrasts between flaorine and
sodium account for the presence of the nodal neon at 20, instead of
the maoroperiodic triplet, which analogy would otherwise lead us to
expect to find about this point. The next element of the nodal dasSy
argon, brings us at once to a case which is at variance wHh the
periodic law.
No difficulty would arise if argon had an atomic weight just below
that of potassium, but Bamsay does not admit a lower value than
39*9, and this we are bound to accept. The result is that the
neutral argon is placed between the two strongly positive elements,
potassium and calcium, which differ from one another only in de-
gree. In the cases of helium, neon, krypton, and xenon, the ele-
ments occur in neutral positions between strongly contrasted sub-
stances, positive and negative. Argon is, therefore, quite exceptional,
and breaks the order required by the <'law." From our present
point of view, we have, however, the choice of two hypotheses*
One is that argon is not strictly nodal, and is to be placed a little
beyond the common node on the major loop, so that its position be-
tween potassium and calcium is apparent rather than real, but this
implies the possession of very feeble chemical properties. The other
was suggested by Prof. Fitzgerald, namely, that the nodes of the minor
and major loops do not quite coincide, and that of the latter is at 39*9.
It is probable that at least two other nodal elements remain to be
discovered. One of these should have an atomio weight about 174
and the other near 218.
Turning now to the Macroperiodio elements, I have already stated
the reason for thinking that no triplets are likely to exist with atomic
weights round 20.
The iron, palladium and platinum sets of triplets, forming Men-
del^eff's '' eighth " group, fall naturally into their places on the major
vibrations. Their neighbourhood disturbs the symmetry of the minor
periods, and more especially do they appear to influence the properties
of the elements nearest in atomic weight so far as they are known.
On the other hand, as Mendel4eff ingeniously seeks to show, the ad-
jacent elements retain in some degree their <* group " characters, hence
he justifies the classification of chromium with oxygen and sulphur^
and of manganese with the halogens. Nevertheless, he assumes that
each set of triplets, together with the elements near in atomic weight,
form a single '' long period " ; but, in view of the preceding contention,
this really implies the admission of periods within periods. It appears
to me that the nature of the relations of the triplets to the periods
Digitized by
Google
ANNUAL GENEBAL MEBTING. 619
between which they occur is more consistently represented as harmonic
in character.
I need scarcely say that our knowledge of the elements with atomic
weights between 140 and 195, that is, the ''rare earths/' is still very
unsatisfactory ; but the didymium elements^^praseodymium, neodym-
ium, and samarium — possess some characters which seem to give
ground for supposing that they may be the macroperiodic triplets
wanting between 140 and 150, and they are shown in the diagram in
this position. The two first- named elements each give two oxides,
unlike most of the other rare earths ; all three elements afford coloured
salts ; their solutions exhibiting characteristic absorption spectra ; and
the oxides have been found by du Bois and Otto Liebknecht ^ to show
high paramagnetic susceptibility, only second to that of the members
of the iron triplet. I venture to make the suggestion that they belong
to the maoroperiods with all reserve, as our knowledge of these sub-
stances is still limited.t
The next point of interest to be considered principally affects the
Orthoperioddc series. I refer to the question of the relative positions
of iodine and tellurium. According to the rigid form of the periodic
law, the atomic weight of tellurium should be just below that of
iodine, as selenium is below bromine, sulphur below chlorine, and
oxygen below fluorine, All the best determinations of the atomic
weight of tellurium, however, seem to leave no doubt that it is higher
than that of iodine by nearly a unit. There is here direct conflict
between the *' law " and the fact, in which the latter must prevail.
From our present point of view the explanation is simple enough.
Iodine is the element which has the highest atomic weight of any
known number of the '' halogen " group. One of the most character-
istic properties of members of the group is their power of combining
with hydrogen to form the acids of the type HX. Now hydrofluoric,
hydrochloric, and hydrobromio acids are exothermic compounds, the
heat of formation of hydrobromic acid being the lowest, Hydriodio
acid, on the other hand, is an endothermic compound, as energy must
be supplied in its formation. In this particular, therefore, iodine has
almost lost the important group characteristic, while its power of
forming fairly stable compounds with oxygen is much greater than
that of its lower homologues. But iodine departs still further from
the halogen rdle in its power of forming with phenyl, and other similar
radicles, basic substances of the hydroxylamine type — the iodonium
compounds discovered by Hartmann and Victor Meyer in 1894. In
* Ber., 1899, 82, 3344. Compare Stefan Meyer, Monatsh, 1900, 20, 209.
t Note added May 17th.—-lt appears from Nature of May 15th that Brauner ha3
jQSt published a paper with the Russian Chemical Society adopting a similar ?iew
as to the positions of the didymium elementa
Digitized by
Google
620 ANNUAL GENERAL MEETING.
the compound IPb^I, derived from IPhg'OH, trivalent iodine must be
recognised simulating nitrogen, so far as the capacity of the latter is
concerned for holding unozidised radicles and affording basic products,
notwithstanding the fact that iodine does not form a simple hydrogen
compound of the same order as ammonia. In this respect, then, iodine
draws near to antimony, so that, whatever the conditions were which
prevailed at 'Hhe birth of the elements," it would seem that the
increase in mass of the somewhat indifferent iodine grouping was
checked by the pull of its more positive neighbour, while tellurium
was free to gain instead.
With the rare elements which seem to form only sesquioxides, little
can be done at present. So far as definite atomic weights have been
assigned to them they are marked on the diagram, but with all reserve.
There are obviously several elements of the higher periods still
unknown.
We have no certain knowledge of elements of higher atomic weight
than bismuth, save thorium and uranium ; but the radioactive sub-
stances radium, polonium, and actinium, if elements in the usual sense,
probably have high atomic weights also, and may ultimately be found
to fill some of the gaps in this neighbourhood.
Substances of the radium class are known to constantly give off
kathodic radiations which can perform^ definite chemical work on a
sensitive plate, and, according to Professor J. J. Thomson, they must
have emitted similar radiations for millions of years. Such a steady
distribution of energy must be balanced by a supply ab extroy just as
the moving particles of the cord receive theirs from the vibrator ; but we
have not as yet any clue to the source or sources from whence radium
and its allies draw their supplies. All that we definitely know is that
the active substances — whether simple or compound — are comparatively
massive molecules, which serve for the collection of energy and its
distribution partly, at least, in radiant forms. This rather
suggests the idea that the less massive atoms of the other elements
may also act in varying degrees as energy transformers into different
orders of chemical activity.
Dr. Gladstone, F.II.S., proposed a vote of thanks to the President,
coupled with the request that he would allow his address to be printed
in the Transactions.
Dr. Thorpe, C.6., F.R.S., seconded the motion, which was carried
by acclamation.
The Pbesident having returned thanks,
Prof. Tilden, F.B.S., the Treasurer, in giving an account of the
Balance Sheet which he laid before the Society, duly audited,
said : —
Digitized by
Google
ANNUAL GENERAL KEETINO. 621
The receipts had been : — By admission fees and subscriptions,
£4532 ; by sale of Journal and advertisements, £835 II9. 2d, ; and by
dividends on invested capital, £476 128. 6(2. The total receipts
from all sources amounted to £5884 Is, Sd. The expenses had been :
— On account of the Journal, £3233 Oa, 2d,; on account of the
Proceedings, £20 1^. 6d, ; on account of the Library Catalogue,
£42 lis. id.; on account of the Library, £429 I6s. 6d,; House
expenses, £239 I9s, Id, ; the total expenditure being £4913
10s, ed.
The Treasurer, in concluding, proposed a vote of thanks to the
auditors, which was acknowledged by Mr. Chapman.
Prof. H. B. Dixon, F.R.S., proposed a vote of thanks to the
Treasurer, Secretaries, and Council.
Dr. Hewitt seconded the motion, which was unanimously adopted.
Prof. Meldola, F.B.S., responded.
The Scrutators having presented their report to the President, he
declared that the following had been duly elected : —
President: J. Emerson Reynolds, Sc.D., M.D., V.P.R.S.
Vte&-Presidents who hamJUled the office of President : Sir F. A. Abel,
Bart., K.C.B., D.C.L., F.R.S. ; H. E. Armstrong, Ph.D., LL.D.,
F.RS. ; A. Crum Brown, D.Sc., LL.D., F.R.S. ; Sir W. Crookes, F.R.S. ;
J. Dewar, M.A., LL.D., F.RS. ; J. H. Gladstone, Ph.D., D.Sc., F.R.S.;
A. G. Vernon Harcourt, M.A., D.C.L., F.RS. ; H. Muller, Ph.D.,
LL.D., F.RS. ; W. Odling, M.B., F.R.S. ; W. H. Perkin, Ph.D., LL.D.,
F.RS,; Sir H. E. Roscoe, LL.D., F.RS.; W. J. Russell, Ph.D.,
F.RS. ; T. E. Thorpe, C.B., LL.D., F.R.S. ; A. W. WilliaiAson,
LLD., F.RS.
Vice-Presidents: E. Divers, M.D., D.Sc., F.R.S.; P. F. Frank-
land, LL.D., F.R.S.; H. McLeod, F.R.S.; R Meldola, F.R.S.;
H. A. Miers, D.Sc, F.RS. ; T. Stevenson, M.D.
Seoreiaries : W. R Dunstan, M.A., F.RS. ; A. Scott, M.A., D.Sc.,
F.RS.
Foreign Secretary : W. Ramsay, LL.D., F.R.S.
Treasurer: W. A. Tilden, D.Sc., F.RS.
Other Members qf Council : H. B. Baker, M.A. ; F. D. Chattaway,
Ph.D., D.Sc. ; F. Clowes, D.Sa ; J. J. Dobbie, M.A., D.Sc. ; A. E.
Dixon, M.D. ; M. 0. Forster, Ph.D., D.Sc. ; A. Harden, M.Sc., Ph D. ;
J. Lewkowitsch, Ph.D. ; J. E. Marsh, M.A. ; S. XJ. Pickei-ing, M.A.,
F.RS. ; J. A. Voelcker, Ph.D. ; J. Walker, D.Sc., F.R.S.
Digitized by
Google
622 ANNUAL GENERAL MEETINQ.
THE TREASURER IN ACXX)UNT WITH THE CHEMICAL
Dk.
S i. d. S ». d, M B, d.
BaJanoe at Bank, Maroh 84th, 1900 1818 13 7
„ InhandaofTreaBorer ^^ «. ^ 0 1 S
Receipts by Life CSompoeitlons, AdmlMfon Fees and Subaerlp.
tlons from Maieh 2Srd, 1901, to Maiob 22iid. 190S :—
Life CompoBitiont— 3 at £30. 1 Bal. at J28, 1 at JSO, 1 at £li,
1 at ilO. 1 Bal. at £8 168 0 0
168 AdiniBston Foes ^ 658 0 0
4 8abMriptioii8forl899 „ £8 8 0 0
1 Sabseriptlon for 1900 „ £1 10 0
181 Subscriptiona for „ „ £8 ^ 862 0 0
6 „ „ 1901 „ £1 6 0 0
782 „ „ ., „ £2 UH 0 f
6 „ „ 1902 „ £1 5 0 0
998 .. ., ,. „ £2 1966 0 0
1913 13 10
Sale of Journals ^ 689 IS S
M Proceedings 18 18 0
„ General Index IS 4 8
„ Bale of Memorial Lectures 27 10 6
ProceedsofAdvertiiements In Journal ^^ 83 4 9
BubeeripUon firom the Society of Chemical Industry to June, 1901 ... 8 8 0
M „ ., Publlo Analysts to January 1st, 1902 11 11 0
„ „ Physical Society to January 1st, 1902 19 19 0
Repayment of Income Tax 28 14 0
Tear's Diyldends on £6,7S0 Metropolitan Board of Works 8| per
oent. Stock 222 6 2
„ „ £1,060 London and North-Westem Railway
Debenture Stock 29 14 7
„ „ £1,520 14«. 8d. Gardiif Corporation Stock 42 19 4
„ „ £1.400 India 8^ per cent Stock 33 0 0
,, ,. £2.858 Midland 2( per cent. Preference 55 12 8
„ „ £2,400 Bristol 2| per cent Debenture ». 56 10 0
Interest on Bank Deposit 13 16 0
4582 0 0
835 11 8
89 18 0
476 12 6
-— 5884 1 $
StlimaUd
Assets. y^^'
March 28rd, 1001. £ f. d.
Balance at Bank (Current Account) ~ 2034 0 0
„ „ (on Deposit) 1000 0 0
,. in hands of Treasurer 4 8
£6,780 Metropolitan Board of Works 3} per oent. Stock 7201 2 0
£1,050 London and North-Westem Railway Debenture
Stock 1065 16 0
£1,520 14f. Sd. Cardiff Cori)oration 8 per cent. Stock... 1466 16 0
£2,858 Midland Railway 2^ per cent. Preference 1861 0 7
£8.400 Bristol Corporation 2i per centDebenture Stock 1992 0 0
£1,400 India 2i per oent. Stock 1197 0 0
£17807 18 7
£7197 Mi 6
Digitized by ^
ANNUAL GENERAL MEETING. 623
SOCIETY, FROM Mabch 23ed, 1901, to Maboh 22nd, 1902.
Ob.
Expenses on Account of the Journal and Prooeedings,
S i, d, £ i. d. £ $. 4.
Salary of Bditor S50 0 0
„ Sub-Editor 200 0 0
„ Indezer ^ 80 0 0
Editorial Postages 13 16 H
Abstractors' Fees 359 19 8
Printing of Journal..... 1068 15 0
Printing of Advertisements 29 16 8
Printing Wrappers „ 184 6 0
Distribution of Journal by Printers » 884 15 6
M M Society , 15 6 5
Authors' Copies , „ „,.,..« 98 1 0
Illustrations for Journal 8 5 0
Printing of Prooeedings 156 0 1
Distribution of Proceedings...., 50 1 4
Publishers' Commission 73 4 9
Advertising Agents' Commission 18 0 9
8283 0 2
806 1 5
86 14 6
Expenses on Account of Collective Iihdex 1893—1902.
Salaries 65 11 1
P^tty expenses « 1 17 11
Index Slips 5 9 0
Expenses on Acoount of the Library Catalogue.
Salaries SO 10 0
Petty expenses 0 14 4
Csse for Catalogue 1110 0
E3q>enses on Account qf the Library,
Salary of Library Assistant 49 18 0
Books and Periodicals 811 6 11
Binding « 75 15 6
79 18
49 14 4
429 16 5
House Ea^penses,
Providing Reflreshments 22 0 10
Lighting the Building (Gas, £20 Si. U. ; Electric Light,
£20 18i. Id.) « « 41 8 8
Heating the Building (Coals) 20 19 6
Cleaning 15 0 0
Repairs.. 41 5 8
Petty House Expenses «... 86 8 1
House Porter's Wages 65 0 0
f, „ Uniform 5 19 0
Annual Fee to Gate Porter 8 8 0
Inhabited House Duty 0 6 8
289 19 1
Salary of Assistant Secretary 800 0 0
Pension to Mr. Hall 180 0 0
Miscellaneous Printing 78 14 4
Stationery 14 1 8
Addresses to M. Bertlielot and the Owens College 8 12 0
Indexing for International Catalogue 80 6 «
Bxpenses on account of Anniversary Dinner 87 7 6
Legal Charges 16 6 0
Show<case for Medals and Framing Portraits, Ac IS 4 0
Memorial Lectures, Binding of ^ 18 7 8
Bank Charges 0 9 9
Treasurer's Petty Cash Disbursements 0 2 9
„ Assistant 10 0 0
Postage Account: Office and Secretarial Postages, £9 i§. 7d ;
Postal Cards and Stemped Envelopes (Clay), £29 2«. 2d. ;
Embossed Stamps, £12 10< 50 16 9
4913 10 6
Transferred to Deposit Account » 250 0 0
Balance at Bank 20St 0 9
„ in hands of Treasurer 0 4 8
£7197 15 6
_ C. CHAPMAN.
found correct. f}^ ^ THOBNl.
pud Mvmh, 1902,
/Lt. r
Digitized by
Google
624
ANNUAL GENERAL MEETING.
in
D
<
g S
^ s
;^
J CO
3 g
O
H
04
o
o
o
CO
"W oooooooo
•6* oooooooo
ooooooooooo
ooooooooooo
illl
30-3 12 ei«.
: '2 : : : 5*
r
PQ
^
-.i^^
§«
Oil
O C4
t^ CO O
O i-H rH
09 I
«o (^
^2
o o
O 00
10 00
(N 00
00 "2
00^
Or
«^
11:1!
II
' ^
J4 P.fe*^j5
■ - 00 °
5
ot2o SIS'*
ii
Si
^
S^'o 00 P*5oo
«•! «•!
Ok ^
-^5
Digitized by
Google
OBItUAKY. 625
OBITUARY NOTICES.
Sir Joseph Hemby Gilbert, Ph.D., M.A., LL.D., Sc.D., F.R.S.,
born August 1st, 1817, died December 23rd, 1901.
Ok May 18th, 1841, when the Chemical Society was barely three
months old (it having been founded on February 23rd of that year),
there was elected to the Fellowship a young man of twenty-four, who
was destined to play an important part in the history of the Society
and in that of the progress of chemical science and investigation, and
whose name, in conjunction with that of his fellow-worker and patron,
Sir John Bennet Lawes, was to be for ever associated with the benefits
which chemistry has conferred on the industry of agriculture. This
young man was Joseph Henry Gilbert, born at Hull of parents well-
known in the literary world, and himself lately returned from Giessen,
where he had been studying under Liebig. His earlier years had been
impaired, and his future career threatened, by an accident which
deprived him of the use of one eye, but his subsequent record forms
a striking instance of triumph over physical disability. He had
worked at Glasgow University under Prof. Thomas Thomson,
Stenhouse being there, as also at Giessen, his fellow student. At
Giessen, whither he was attracted by Liebig's fame, he had not only
Stenhouse, but also Playfair, as companions, and here took his degree
of Ph.D. Returning to England, he worked at University College,
London, and became assistant to Dr. A. T. Thomson, meeting here
also J. B. Lawes, with whom he was afterwards to be so closely
identified. It was at this time that, as stated, he became a Fellow of
the Chemical Society, and so was almost one of its original members.
Forty-one years later (1882) he was elected to the Presidential
Chair, serving in this capacity during the session 1882-3, and sixteen
years later he formed one of the group of six Past-Presidents whose
fifty years' continuous membership of the Society was celebrated by
the remarkable and unique gathering of November 11th, 1898, when
a banquet was given in honour of the distinguished veterans in
science — Gilbert, Frankland, Odling, Abel, Williamson, and Gladstone.
Of this group, Gilbert was the senior, Playfair, who would have
formed the seventh member and the only original member of the
Society among the seven — for other original members were then
living — having passed away only a few months previously.
In reviewing Gilbert's work one cannot do better than recall some
of the remarks which were made at the notable banquet referred to,
as they apply with special force to Gilbert's character and aims. The
then President, Prof. Dewar^ spoke of the guests thus honoured as
Digitized by
Google
626 OftlTltAttY.
*' men whose one idea has been, with steady aim and vigilant eye, to
labour on with that sole incentive of scientific work, the triumphant
hope of making an advance. These men have laboured for half a
century in our interests, and they have added enormously to our
knowledge of the science .... We are still able to go back to a
man, who sits on my right hand (Gilbert), who worked in the laboratory
of ThomasThomson, who has seen Dalton and the beginning of the atomic
theory." Then, addressing himself to Gilbert, he said, <* The work of
Gilbert, as we know, was early differentiated into that most complex and
mysterious study, the study of organic life. For the last fifty years
he has devoted his attention to the physiology of plant life in every
phase of its development. With a skill that has been unprecedented,
he has recorded from year to year the variations ^in the growth of
every kind of nutritious plant. He has examined into the meteoro-
logical conditions, the variations of climate, of soil, and of mineral
agents, of drainage, and of every conceivable thing affecting the pro-
duction and development of plant growth. These memoirs are
admitted throughout the world to be unique in their importance.
Wherever the chemist or the physiologist, the statistician or the
economist has to deal with these problems, he must turn to the results
of the Eothamsted experiments in order to understand the position of
the science of our time. These results will be for ever memorable ;
they are unique and characteristic of the indomitable perseverance and
energy of our venerated President, Sir Henry Gilbert."
These words most aptly describe the life-work of Gilbert, and when
it is remembered that for another 3 years after this he laboured on
unremittingly and died positively "in harness," some idea can be
formed of the devotion of Gilbert to the pursuit of the science he has
so largely enriched.
Gilbert's first contribution to the Chemical Society's Memoirs was
the translation of {a paper by Bedtenbacher and Liebig on *'The
Atomic Weight of Carbon " — this was previous to his formal admis-
sion to the Society. But it was in 1843 that the important step was
taken by him which shaped his future career, for it was then that he
was invited by Lawes, who had been his fellow student at University
College, London, to assist him in the agricultural investigations which
he had just begun on coming into possession of his country property
at Bothamsted, Herts. From that time began the unbroken collabor«
ation which, for its duration, intimacy, and results on agriculture, has
had no parallel.
The name '< Bothamsted," from being merely that of a country
gentleman's seat, has come to designate a storehouse of knowledge,
and a centre from which the efforts of two distinguished men, each
working on his own line but combining their powers in the setting-out
Digitized by
Google
OftlTUARt. 627
of their joint conclusions, have been put forward for the elucidation
of nature's secrets and the benefit of the great industry of agriculture.
It is hard, and it would be undesirable, to attempt to separate the
work of two men so closely associated, for " Lawes and Gilbert '' is
the fitting expression for this unique collaboration. Still, it may not
be amiss to point out how great was, in the case of both men, the
influence of the scientific method in investigating the problems of
practical agriculture and in framing the conclusions to be drawn.
Oilbert was remarkable for the complete conscientiousness of his work,
the extreme care and patience which he displayed, the scrupulousness
with which he verified his results by frequent repetition, the pertin-
acity with which he maintained the continuation of the experiments
when once set on foot and defended the conclusions drawn from them,
and for his untiring energy, his life-long devotion to the work in the
thorough conviction — ^which we can now share — that what he was
engaged in would remain as a monument and an example for the
future. No one could be brought into contact with Gilbert without
feeling that it was for Rothamsted that he lived, his one aim was
not his own aggrandisement (for there was little of this in his
case), but the making of Rothamsted a centre of usefulness and the
home of agricultural research. Few men have there been who have
set before them such single-minded purposes as Gilbert, his own posi-
tion, his own advancement, it can be truly said of him, were put aside
for higher considerations, and what honours and distinctions were
accorded to him sprang from the initiative of an appreciative public
outside, and were none of his seeking.
To every worker who cared to apply to him, Gilbert was ever ready
to give help, and sincere is the gratitude which the younger generation
of agricultural chemists feel towards him for what he has done for
them, and for the way in which he endeavoured to elucidate any
point in the Rothamsted experiments.
Gilbert was, in fact, the exponent of the Rothamsted work and
the one to familiarise its results and lessons to the scientific world.
In this capacity he frequently read or wrote papers for the
Chemical Society, the Royal Agricultural Society, and other bodies,
and in pursuit of the same idea he would attend the meetings of the
British Association, or visit the Continent, or make journeys to
Canada or the United States.
None of those who were privileged to know him well will ever for-
get his demonstrations in the Rothamsted laboratory, his exposi-
tions in the field, and the infinite care that he took to make
everything clear and to emphasise it finally by reference to that
vast series of tables with which all students of Rothamsted have
become familiar.
Digitized by
Google
628 OBITUARY.
Not even increase of years seemed to dim the energy or activity,
still less the enthusiasm, of Gilbert, and it was not until dlter the death
of Sir John Lawes in August, 1900, that he showed any signs of failing
health. Even then he could not be persuaded to put his work aside
or take a needed rest, but continued, as before, to send in his regular
reports to the Lawes Trust Committee, and to plan out work for future
development.
In disposition, Gilbert was most friendly and amiable. Everyone
had a good word to say for him. Jealousy of anyone else seemed to
have no place in his nature, and desire of personal gain was an element
foreign to his character. His devotion to work was shared in and
ably aided by his wife — the present Lady Gilbert — who survives him,
and to whose unremitting care is largely due that he was so long
spared to continue his labours.
To detail Gilbeit's work would be to write the account of the
Bothamsted experiments — a task beyond the limits of the present
notice — and it is only necessary to recall the fact that it dates from
the time of the inception of the " mineral theory " of Liebig — which
made way for the *' nitrogen theory " of Bothamsted — to the com-
paratively recent questions of soil organisms and assimilation of
atmospheric nitrogen by plants.
Gilbert was made a Fellow of the Royal Society in 1860, and in
1867 he^ with Lawes, received a Royal medal of that Society. Many
other distinctions from learned societies, public bodies, and universities
followed the prolific work emanating from Rothamsted. Oxford Uni-
versity made him Sibthorpian Professor of Rural Economy in 1884,
the Albert Gold Medal of the Society of Arts was awarded him, and
the conclusion of 50 years of Lawes and Gilbert's joint labours was
marked by a presention to him on July 29, 1893, and the inscription
of his name, together with that of Lawes, on a granite monolith
erected in front of the Rothamsted Laboratory. As a fitting sequence,
Gilbert was, to universal satisfaction, given, by her late Majesty Queen
Victoria, the honour of knighthood.
He died at his own house, at Harpenden, which adjoins the scene
of his labours, and now he rests in the village churchyard there beside
his life-long associate, Lawes. Many may well envy the encomium on
Gilbert which the Spectator expressed when it said of him " He
achieved the rare distinction of gaining a world-wide reputation
mthout being known to the man-in-the-street." J. A. Y.
Hbnbt Gbobos Madan, who died on the 22nd of December, 1901,
was bom on September 6th, 1838, and after being educated at a
private school near Bath and at Marlborough College gained a classical
Digitized by
Google
OBITUARY. 629
exhibition at Corpus Christi College, Oxford, in 1857. He obtained
a second class in Moderations and a first class in Natural Science, and
in 1862 was elected to a Fellowship at Queen's College, which he held
for the remainder of his life. From 1863 to 1869 he was demon-
strator in chemistry in the University. In 1869 appeared the first
edition of the well-known JSocerdses in Practical Chemistry, written by
Mr. A. G. Yemon Harcourt and himself. In the same year, he was
appointed First Science Master at Eton College. He held this post
for twenty years, and, though not an inspiring lecturer or one able to
obtain a mastery over unwilling boys, he was an admirable teacher for
those who desired to learn ; by them his labours as a schoolmaster will
always be appreciated and remembered with real gratitude.
Among his published works are to be mentioned a new edition of
Wihon^a Inorgamc Chemistry (1871), Leaaona in Elementa/ry Dynamics
(1886), An Elementa/ry Treatiae on Heat (1889), Tahlea of Qimntitative
Analyaia (1881), and several papers and notes on optical experiments
and their demonstration. His name is not associated with much
original work, but the observations on the remarkable potassium
chlorate crystals afterwards investigated by Stokes are noteworthy.
Towards the close of his life he devoted much attention to optical
research and to the British Association bibliography of spectroscopy.
A few years before his death, he was crippled by an accident which
deprived him of bis right arm and disabled one leg, and doubtless laid
the seeds of his fatal illness. To a man of his extraordinary bodily
activity who devoted himself, even in advanced years, to rowing as an
exercise and to dextrous mechanical work as a pastime, the loss must
have been heartbreaking ; but it only gave him the opportunity of
displaying his unyielding and strenuous character; he had always
steeled himself against the exhibition of feeling ; and with indomi-
table energy he returned as far as possible to his scientific work.
Mr. Madan possessed a keen and rigid intellect ; he was an able
experimentalist and a fine mechanician, and used to say that he never
employed a machine which he could not make for himself. He was
also an excellent scholar ; in this connection it will be remembered
that he suggested the Homeric names, Deimos and Fhobos, for the
satellites of Mars.
His scientific apparatus, much of which was made by himself, he
distributed before his death, partly to his College, and partly to the
Mineralogioal Department of the University Museum as a gift to one
of his old pupils. H. A. M.
Mr. W. B. Randall, of Southampton, who died on March 14th,
aged 81, was one of the oldest of our Fellows, having been elected as
an Associate in April 1843. Originally an apothecary in the old
sense of the word, he finally adopted pharmacy as his profession.
VOL. LXXXI. Digitized byM350gIe
630 OBITUARY.
Mr. Kandall was a private pupil of Mr. R. WariDgton, the first
Secretary of the Chemical Society, and studied Ohemistry at University
College under Thomas Graham, taking the Silver MedaJ in the session
of 1843. Mr. Randall was a J.P. for Southampton, and had held
office on the City Council and as Chairman of the School Board.
Savillb Shaw was horn at Ardwick, Manchester, on December
22nd, 1864. He received his early education at the Manchester
Commercial School, and while at school displayed that keen interest
in experimental science which he retained through life, and which, no
doubt, led to his pursuing his education at the Owens College, where
he became a student in 1880. During the period he was at Owens
College, Shaw devoted himself mainly to the study of chemistry nnder
Sir Henry E. Rosooe, and after completing his course spent a year in
the "Doctor's" private laboratory. It was at this time that be
conducted the investigations on pentathionic acid and the penta-
thionates, the results of which were embodied in a paper published in
the Transactions of this Society for the year 1883.
In January, 1884, Saville Shaw, who had just entered on his
twentieth year, came to Newcastle as assistant in the Chemical
Department of the Durham College of Science, with which institution
he was associated until his untimely death, a period of upwards of
seventeen years. At first with the duties of this post those of lecture
assistant were combined, in the discharge of which Shaw displayed
great care and ingenuity, and a resourcefulness which the writer of
this notice has had occasion repeatedly to admire and thankfully to
appreciate. Relinquishing this work after a time, he became lecturer
in chemistry, devoting himself to the teaching in the laboratory and
the lecture room, in both of which his quiet manner, his knowledge,
and his keen appreciation of the difficulties of the student contributed
to make him popular as a teacher. Nor was this alone the opinion of
the College students, for under the auspices of the County Councils of
Northumberland and Durham he gave several courses of lectures in
various parts of these two counties, and by schoolmasters and miners
alike his lectures were highly valued and appreciated.
When, in 1890, the College in enlarging its sphere of work added
a Department of Metallurgy, Mr. Shaw, who had always shown a
special leaning to inorganic chemistry and the chemistry of the metals^
was elected to take charge of this department, and in this capacity
commenced an investigation on the micro^tructure of alloys. Of the
results of this investigation he has only published a short note, which
appeared in Nature^ August 11th, 1898, accompanied by reproductions
of two photographs of sections, which show well his admirable skill in
this class of work. Amongst his effects have been found a large
Digitized by VjOOQIC
OiiTTTABY. 6S1
number of photographs and sections and also the notes relating thereto,
which Mr. Stead has undertaken to edit for publication.
Despite the arduous duties of his post and a by no means robust
constitution, Shaw was not unmindful of the fact that in an industrial
community there are many problems the scientific investigation of
which may serve to advance knowledge and help at the same time to
the proper understanding by the layman of the advantages of a
scientific training. Thus, he not only lent valuable assistance to the
writer in an investigation of the cause of the explosion of an air-receiver
at one of the neighbouring collieries, but also was one of the most regular
attendants at the meetings of the Flameless Explosives Committee
formed by the North of England Institute of Mining and Mechanical
Engineers, and in the conduct of the numerous experiments of that
committee gave invaluable aid and contributed to the report a special
: 3tion dealing with ** Gases and Gaseous Mixtures."
S vllle Shaw took a great interest in the literature of his own
countiy, he had a facile pen and a keen and somewhat caustic wit, to
which he occasionally gave liberty in verses which were much enjoyed
by a small circle of friends.
For some years he acted as Hon. Sec. and Treasurer to the New-
castle Section of the Society of Chemical Industry, for which body he
laboured most successfully, his organising power and ability finding
expression in the great success of the general meeting of that society
which was held in Newcastle in 1899. To mark their sense of appro-
elation of his work, the members of this section have subscribed to
found a medal, which is to be awarded as a prize to students in
chemistry or metallurgy, so that his memory may be always associated
with the science to which he was so devoted.
His unassuming modesty and kindly nature endeared him to his
colleagues, to whom and to his relations his sudden death on
November 5th, 1900, came as a great shock.
In 1896, the University of Durham, in recognition of his work in
connection with the College of Science and the extension lectures in
the county, conferred on him the degree of M.Sc. P. P. B.
Maxwell Simpson, who died in London on February 26th last, was
a son of Mr. Thomas Simpson, of Beech Hill, co. Armagh, Ireland, and
was the youugest of a family of nine. He was born on March 15th,
1815, and thus nearly completed his eighty-seventh year. He was
educated at a private school in Newry, which was well known at the
time, kept by Dr. Henderson, and from this school he passed into
Trinity College, Dublin.
In his early boyhood he was much in the company of the brilliant
and witty Charles James Lever, the novelist and physician. Suscept^
ible as he was throughout his life to personal influences, the conversa-
DigitizecSyCSogle
632 OBITUART.
tion of Lever, which often turned on physiological problems, led the
young Simpson to look to the profession of medicine as his future
career. But with this directing impulse towards science there was
also a natural restlessness, a seeking after new ideas, new impressions,
which may well have been intensified by the example of the author of
*' Charles O'Malley," himself a wanderer, who died in a distant land.
Thus, after taking his A.B. degree, be did not then proceed to the
M.B., but travelled to London, from which place he made several
journeys to Paris.
During one of these visits to Paris he attended the lectures of
Dumas. This was a great event in his life. Within the charm of
this great investigator and teacher, of the personality so well
remembered by many of us, the Irish student found what he had
dreamt of, what he had longed and sought for. He found satis-
faction for that restless discontentedness with a conventional carec
which had led him to wander from home, that healthy discontent to
which the world is indebted for so many great men. He bu'lt for
himself in his imagination an ultimate object of life, an ideal ; he saw
the possibility of molecular structures of never-ending wonder and
beauty, and he pledged himself to endeavour to realise them. To
this ideal he remained fervently true throughout his life. Maxwell
Simpson was a man of wide culture and never-failing humour^ and a
personality as kind as it was true ; but no one really knew him who
did not know of his ideal. Like the knight in DUrer's picture, he rode
on 3 all other things which most men prize — wealth and worldly
recognition — he brushed aside; he had one single object only, his
ideal, which meant the pursuit of chemistry.
On his return to London he decided to go through a preliminary
training in chemistry, and for that purpose entered Graham's
laboratory in University College, where he remained two years. After
this, in 1845, he settled in Dublin and married a daughter of Mr. Samuel
Martin, of Langthome, co. Down. Mrs. Simpson entered with an en-
thusiasm which never relaxed into her husband's ideal and made it her
own. Much of his success was due to her womanly tact and counsel. In
1847 he became lecturer in chemistry at the Park Street Medical
School, called later the Ledwich School of Medicine, and in order
to hold this position he took his M.B. degree at Dublin TJniversity.
He retained this lectureship until 1857.
In 1851 he longed once more for that atmosphere of scientific
research which was found then, as now, in its best development in the
universities of the Continent. He obtained leave of absence from the
Medical School and went with his wife and family to Germany, where
he remained three years. The happy scientific and social intercourse
of those years always remained fresh in the memories of Dr. and Mrs.
Simpson, and in after years nothing pleased them better than to relate
Digitized by VjOOQIC
OBITUARY. 683
to their friends their pleasant experiences of " plain living and high
thinking" in the university towns on the Neckar and the Lahn,
Like other foreign students who ai*e privileged to enjoy both Crerman
living and Crerman thinking, Maxwell Simpson became a missionary
of the culture and industry of the Germans to his native land.
After his training in Graham's laboratory he was prepared to begin
original inquiries, and in the laboratory of Bunsen in Heidelberg,
and of Kolbe in Marburg, he made good progress. In these
laboratories he had among his fellow workers and friends many who
have since become well known in the annals of science. His first
published paper emanating from Heidelberg was on improved methods
of organic analysis: '^TJeber neue Methoden zur Bestimmung des
Stickstoffs in organischen und unorganischen Yerbindungen" (Annalenf
1855, 06, 63). These methods are especially applicable to compounds
which bum with difficulty, and a full account of them will be found in
Fresenius's " Quantitative Analysis " and in Boscoe and Schorlemmer's
"Treatise on Chemistry,"
Maxwell Simpson returned in 1854 to his medical classes in Dublin,
but they seem to have interfered too much with the pursuit of his ideal,
for in 1857 he resigned the lectureship and proceeded with his family
once more to the Continent. This time it was Wurtz, in Paris, who,
by his discovery of the glycols, attracted him, as he attracted other
students, to his laboratory. Here he worked for upwards of two years
enjoying the same friendly intercourse among French chemists that he
had experienced before among the Germans across the Rhine. His
inquiries were directed to the elucidation of the theory of polyhydric
alcohols. At this period he published " Note concernant Taction du
Br6me sur I'lodide d'Ald^hyd^ne " {Compt. rend., 1868, 66, 467);
" On the Action of Acids on Glycol," two papers {Proc, Roy, Soc., 1859*
0, 725; 1860, 10, 114); "Action du Chlorure d'Acetyle sur I'Ald^-
hyde" {Compt, rend., 1858, 47, 874) ; "Sur une Base nouvelle obtenue
par r Action de I'Ammoniaque sur le Tribromure d' Allyle " {Compt.
rend., 1858, 46, 785).
Returning once more to Dublin in 1860, he fitted up a laboratory in
his house in Wellington Road where, for seven years under the most
difficult conditions, he pursued his inquiries, and where his greatest
successes were achieved. In the back kitchen of this house he obtained
for the first time synthetically succinic and other di- *and tri-basic
adds. That alkyl cyanides on hydrolysis yield monobasic acids was
known ; but he was the first to apply the reaction to the cyanides of
the dyad and triad radicles, obtaining di- and tri-basic acids. Thus,
he prepared succinic acid from ethylene dicyanide, pyrotartaric acid from
propylene dicyanide, tricarballylic acid from tricyanopropane, and also
many hydroxy-acids from corresponding cyanides. The following con-
tributions were the result of work done in Wellington Road : "^ a Com-
.digitized by Google
634 OBITUAEY.
pound of Dibromallylammonia and Chloride of Mercury " (Phil, Mag.y
1859, [iv], 17, 194); "On Cyanide of Ethylene and Succinic Add'*
{Proc, Roy. Soe., 1860, 10, 674); *'0n the Synthesis of Succinic and
Pyrotartaric Adds" {Phil. Trans., 1860, 61); "On the Action of
Chloride of Iodine on Iodide of Ethylene and Propylene Gas " {Prac
Boy. Soo., 1862, 11, 590); "On the Synthesis of Tribasic Acids"
{Proc. Roy. Soo., 1863, 12, 236; this Joum., 1866, 18, 331); "On
the Direct Transformation of Iodide of Allyle into Iodide of Propyle "
{Proo. Boy. Soc, 1863, 12, 533); "On the Acids Derivable from the
Cyanides of the Oxy-Radicles of the Di- and Tri-atomic Alcohols"
{Proe. Roy. Soc, 1864, 13, 44) ; "On the Action of Chloride of Iodine
upon Organic Bodies" {Proo. Roy. Soo., 1864, 13, 640); "On the
Formation of Di-iodacetone " {Laboratory, 1867, 1, 79); "On the
Direct Transformation of Chloride of Ethylidene into Glycol " {Phil.
Mag., 1868, [iv], 35, 282); " On some New Derivatives of Acetone"
{Proc. Roy. Soc., 1868, 16, 364).
In 1867 he again visited Paris and continued his researches in
Wurtz's laboratory. These inquiries resulted in two memoirs, the
one " On the Formation of Succinic Acid from Chloride of Ethylidene "
{Report Brit. Assoc., 1867, 42), and the other in conjunction with
Arm. Gautier, " Sur une Combinaison Directe d' Aldehyde et d' Acide
Cyanhydrique " {Compt. rend., 1867, 66, 414). He then resided in
London for a few years, when he acted as Examiner at Woolwich,
Coopers Hill, and for the Indian Civil Service. He examined also in
Materia Medica for the Queen's University in Ireland.
Maxwell Simpson, at the age of 67, received his first important
appointment, and thereby became entitled to a regular professional
income for the first time. On the death of Dr. John Blyth in 1872,
he was appointed Professor in Queen's College, Cork. He took to the
Cork College the prestige of a scientific inquirer, well known in the
laboratories of England, France, and Crermany. He held the chair for
nearly twenty years, when he retired in 1891. The behaviour of
iodine chloride in organic reactions had always interested him, and
this and other subjects attracted his attention while in Cork. He
published the following papers : " On the Bromiodides " {Proc. Roy. Soc.,
1873, 22, 51) ; " On the Determination of Urea by Means of Hypo-
bromite of Soda," in conjunction with Mr. C. O'Keeffe (this Journal,
1877, i, 538) ;'" Chemical Notes " {Proc. Roy. Soc, 1878, 27, 118) ; " On
the Formation of Chlor-Iodide and Brom-Iodide of Ethylidene " {Proc.
Roy. Soc, 1878, 27, 424). After retirement from the Cork chair in
1891 he resided in London.
The discoveries of Maxwell Simpson without doubt place him among
the great chemists who laid the foundations of organic chemistry in
the last century. But it may be that still greater than his actual
achievements was his life-long devotion to the prosecution of
digitized by V^
OBITUARY. 635
goience, to the ideal he formed for himself under the influence of
Dumas.
Maxwell Simpson became a Fellow of the Chemical Society in 1857,
served as a Member of the Council 1864 — 1870, and Vice-President,
1872—1874; Fellow of the Koyal Society, 1862 ; Honorary Fellow of
the King's and Queen's College^ of Physicians of Ireland, 1865 ;
Senator of the Queen's University in Ireland, 1873 — 1882 ; President
of the Chemical Section of the British Association, 1878; Fellow of
the Royal University of Ireland, 1882—1891. He received the
degrees of M.D., 1864, and LL.D., 1878, from Dublin University, and
D.Sc, 1882, from the Queen's University, honoris causa, A. S.
William Thomas Newton Spivey.— On October 9th, 1901, a serious
accident occurred in the University Chemical Laboratory, Cambridge,
which a fortnight later led to the death of Mr. W. T. K Spivey, and
robbed the University of one of the most promising and popular
members of its chemical school, adding one more name to the roll of
those who have lost their lives in the pursuit of science.
Mr. Spivey was born on the last day of the year 1868, and received
his early education at Elmfield College, York, where he at once gave
evidence of the all round thoroughness which characterised his later
work. In 1883, while still at Elmfield, he gained distinction in
almost all the subjects of the Cambridge Local Examination, taking
the first place in mathematics and chemistry, and thus winning a
scholarship at the High School, Newcastle-under-Lyme. From 1883 to
1887 he worked at Newcastle, chiefly at science and mathematics, and
while there, passed the London Matriculation Examination, taking the
first place on the Honours List, and thus gaining the Exhibition.- About
the same time, he was elected to a sub-sizarship at Trinity College,
Cambridge, for science and mathematics, and came into residence in
October, 1887.
His career as an undergraduate was marked by success, the result
of diligence and enthusiasm in his work in the laboratories. In 1889,
he was placed in the first class in Part I of the Natural Science Tripos.
In the same year he was elected a Scholar of his college, and in
1891 he gained a First Class in Part II of the Natural Science Tripos,
with chemistry as his chief, and physics as his second, subject, and pro-
ceeded to his B. A. degree. He continued to reside in Cambridge, giving
much of his time to teaching both as a junior demonstrator in the Uni-
versity Chemical Laboratory, and as a private tutor, and during these
years proceeded to his M.A. degree at Cambridge, and his London B.Sc.
In 1896, he left Cambridge to take the post of Science Master at
Epsom College, where he at once showed himself to be a successful
teacher. Before the end of the year, however, he returned to Cam-
bridge, being invited by Professor Dewar to become Jacksonian
Digitized by V^
636 OBITUARY.
Demonstrator in the University, which post he continued to hold until
the day of his death.
As a lecturer, he at once attained great popularity, and all his courses
of lectures were largely attended. He was especially successful in his
lectures on advanced organic chemistry, in which subject his very
wide reading and his extraordinarily methodical mind enabled him to
arrange the vast amount of information in his lectures in an orderly,
logical, and systematic manner which was of the greatest benefit to
his pupils. As a demonstrator, his kindly, genial manner, and his
earnestness and enthusiasm for his subject, marked him out as the
man to whom the student in difficulty always turned for help, with
the certainty of receiving the most carefuland courteous explanation,
however trivial or however wide the question might be.
His great popularity among the students was shown by the success
which attended his effort to reorganise the Students' Chemical Club,
of which he was secretary at the time of his death, and which he
left in a thoroughly flourishing condition.
With so large a proportion of his time ungrudgingly given up to
teaching, Mr. Spivey's contributions to chemical research were not so
numerous as they might have been had he used for his own work
the time he gave so willingly to helping others.
The formation of closed carbon chains by condensations starting
from acetylene tetrabromide first claimed his attention, but he was
unfortunately anticipated in the publication of his results. He
then undertook with two of his colleagues in the University Chemical
Laboratory an investigation on the active principle of Indian hemp,
which^ together with a number of terpenes and other compounds
occurring in hemp, formed the subject of two papers in the Transac-
tions of the Society. Concurrently with this work he was engaged in
the synthesis of various terpeue derivatives, but again had the mis-
fortune to be anticipated.
It was while he was preparing material for the synthesis of
cannabinolactone, a derivative of cannabinol, the narcotic principle
of Indian hemp, that he met with the sad accident which ultimately
caused his death. A flask containing a considerable quantity of
carbon disulphide, which he had been using as a diluent in £tard'8
reaction for preparing aromatic aldehydes, accidentally broke in his
hand after the reaction had been completed. Some of the disulphide
saturated his clothes, the rest vaporised, and the explosive mixture of
vapour and air became ignited. The explosion caused several wounds
but the most serious injuries were the burns due to the ignition of
the disulphide with which his clothes were satui*ated. All appeared to
be going well with him for a week after the accident, when pneumonia,
which so frequently follows severe bums, supervened and caused his
death after a second week's painful illness. T. B. W.
Digitized by VjOOQIC
NITROGEN CHLORIDES CONTAINING THE PROPIONYL GROUP. 637
LXV. — Nitrogen Chlorides containing the Propionyl
Group.
By F. D. Chattaway.
Anilides containing the propionyl group have been comparatively
little studied, and their derivatives are entirely unknown. In the
course of an investigation of the effect produced upon the properties
of substituted nitrogen chlorides and bromides by a change of acyl
group, a number of such substances have been prepared.
Aniline and the chloroanilines, when they are heated with propionic
anhydride or in the case of symmetrical tri-derivatives with pro-
pionyl chloride, readily yield propionanilides. These resemble closely
the corresponding acetyl compounds, but differ from them in the
prevailing habit of the crystals, and in a more ready solubility in
organic solvents. All the propionanilides, on treatment with hypo-
chlorous acid, yield substituted nitrogen chlorides. Hypochlorites are
probably first formed, the nitrogen developing its higher valency and
water being subsequently eliminated thus :
^01
^^<^0.OH,.CH3 ^ ^^0 .
The action appears to be a reversible one, for on placing a nitrogen
chloride in water the opposite change takes place to a small extent
until a position of equilibrium is reached.
These nitrogen chlorides show all the characteristic reactions of the
group, a noticeable feature of such reactions being the invariable
replacement of the halogen by hydrogen.
Some of these changes are reversible, the following equation, for
example, expresses what takes place when chlorine is passed into a
solution of an anilide, or hydrogen chloride into a solution of the
corresponding nitrogen chloride :
CI CI CO-CHjj-CHj
CI CI CI
CI
CI
CH,-CHg ^ jj^j
/Google
Digitized by ^
638 CHATTAWAY: NITROGEN CHLORIDKB
The direction in which action takes place depends on the relative
masses of the various substances in the system. In presence of salts
of weaker acids, such ad propionic or acetic, which remove the
hydrogen chloride as fast as it is formed, the nitrogen chloride is
produced, whilst if excess of hydrochloric acid be added, or the chlorine
withdrawn as it is liberated, the anilide is re-formed.
When hydrogen is attached to the phenyl nucleus either in the
para- or the ortho-position, these nitrogen chlorides readily undergo
isomeric change.
When an unsubstituted phenyl group Is present^ the chlorine atom
attached to the nitrogen and a hydrogen atom, either in a para« or an
ortho-position change places, and a chloropropionanilide results.
When either of these positions is occupied by halogen, exchange
takes place into the one still occupied by hydrogen, whilst when both
the para- and one ortho-position have been taken up, the halogen passes,
although somewhat less readily, into the remaining unfilled ortho-
position. No transference of halogen from the nitrogen to a meta-
position has been observed. The following scheme shows the direction
of transformation :
NHPr NClPr
NClPr ^
Nk
NClPr
Y
These intramolecular rearrangements take place on heating and
especially readily under the influence of chlorine or reagents, such as
hydrogen chloride, which can cause the liberation of chlorine.
When nitrogen chlorides containing phenyl "groups with an ortho-
or para-position unoccupied by halogen take part in the group reactions
previously referred to, ^the primary action often plays quite a sub-
ordinate part owing to the rapid transformation, induced by the
reagent or by some product of the change, destroying the nitrogen
chloride before it can enter into reaction. The behaviour of propionyl
phenyl nitrogen chloride with hydrochloric acid or with alcohol may be
^iven as an example.
Digitized by
Google
CONTAINING THE PROPIONYL GROUP. 639
These isomeric changes, which follow exactly the course of direct
substitution, strengthen the view that in the chlorination and bromin-
ation of anilines and anilides the halogen becomes first attached to the
nitrogen, and then, during the periodic motions of the molecule, passes
into the more stable oon6guration where it is attached to the ring. In
the presence of an amino-group or of an acyl imino-group, the other
atoms or groups present exert a comparatively inappreciable infiuence.
EXPEBIMENTAL.
Fropianyl Phenyl Nitrogwi Chloride, CgHg-NCl-CO'CHj-CHj.
This compound is best prepared by adding a large excess of a solu-
tion of sodium hypochlorite containing potassium bicarbonate to an
alcoholic solution of propionanilide, both cooled to zero. The nitrogen
chloride slowly separates in small, hard plates. It is very readily
soluble in chloroform or benzene, and easily in warm petroleum of low
boiling point. It crystallises from the latter solvent in colourless,
transparent, glistening plates, apparently rectangular with domed ends,
and melts at 77° :
0-2054 liberated I = 22-3 c.c. N/IO iodine.* 01 as IN-Ol = 19-24.
' CgHioOINCl requires 01 as IN-Ol = 19*31 per cent.
On being heated rapidly above its melting point, this compound
undergoes transformation suddenly at about 180 — 190° with consider-
able development of heat to a reddish, slightly impure, mixture of
p- and o-chloropropionanilide. This transformation is also brought
about by heating the nitrogen chloride under a little water, or by
passing into its solution in any solvent a few bubbles of hydrogen
chloride. In the latter case, the isomeric change is very rapid, and the
solution boils violently from the heat developed. It is slowly and quanti-
tatively transformed with a very slight development of colour when
dissolved in chloroform to which a few drops of propionic acid have been
added and the solution allowed to stand for some days. About 90 per
cent, of the p- and 10 per cent of the o-chloropropionanilide are
produced.
^'Chlaropropiananilide, OeH^Ol-NH-OO-OHg-OHj.
This can be easily obtained by once crystallising from alcohol the
transformation product of propionyl phenyl nitrogen chloride. On
account of the inevitable slight waste in purification and the cost of
* All the nitrogen chlorides and bromides described in this paper were analysed
in the same way. A weighed quantity waa dissolved in dilnte acetic acid, excess of
potassiam iodide added, and the liberated iodine estimated by a decinorm^l solntion
of sodium thiosnlphate.
Digitized by
Google
640 CHATTAWAY: NITROGEN CHLORIDES
propionic anhydride, it is more economical to prepare it from j^^ihloro-
aniline. This base, when mixed with the equivalent quantity of pro-
pionic anhydride, readily reacts with considerable development of heat,
and the operation is completed by heating for an hour to 120^ in an oil-
bath, the solid product being recrystallised first from alcohol and
finally from chloroform. It is sparingly soluble in petroleum of low-
boiling point, but readily so in acetic acid, alcohol, or chloroform.
From the latter solvent, it separates in small, glistening, four«ided,
apparently rhombic plates, and melts at 141° :
0-2005 gave 0-1562 AgCl. 01 = 19-26.
CgHioONOl requires 01 = 19-31 per cent.
Propumyl ]^Chlorophmyl Nitrogen Chl<mdey CeH^01-N01-00-CH2-OHj.
This compound was prepared and purified by the method just
described. In solubility and general properties, it closely resembles
the phenyl compound. It crystallises from petroleum of low boiling
point in glistening, colourless, apparently rectangular plates with
domed edges, and melts at 55°. When heated above this temperature,
it begins to darken in colour at about 150° and at about 210° undergoes
transformation into 2 : 4-dichloropropionanilide with considerable
development of heat. When treated as described on p. 639, this trans-
formation takes place quantitatively :
0-2069 Uberated 1 = 19 c.c. iVyiO iodine. 01 as :N-01 = 16-28.
OgH^^OOl-NOl requires 01 as :N-01 = 16-26 per cent.
Propianyl ^-Chlorophmyl NUrogm Brofnide, OgH^Oi-NBr-OO-OHj-OHj.
This, like all the nitrogen bromides described in this paper, was
prepared by shaking the corresponding propionanilide dissolved in
chloroform for about an hour with a solution of hypobromous
acid* containing a little potassium bicarbonate. The chloroform solu-
tion was well washed with water, and finally with very dilute alkali to
remove any free bromine, dried, and the solvent evaporated off on a
water-bath in a current of dry air. When the chloroform was com-
pletely removed, the yellow oil thus obtained solidified on cooling to a
pale yellow solid, which was best purified by recrystallisation from
petroleum of low boiling point. This, like the nitrogen bromides
described later, is a yellow, well crystallised solid very readily soluble
in benzene or chloroform, and moderately so in petroleum.
Propionyl j9-chlorophenyl nitrogen bromide crystallises in groups of
long, glistening, bright yellow, transparent rhombic prisms, apparently
four-sided with domed ends, and melts at 71° :
- * Made by shaking mercuric oxide saspended in water with bromine.
Digitized by
Google
CONTAINING THE PKOPIONYL GROUP. 641
0-3564 liberated I = 27-1 cc. iT/lO iodine. Br as IN'Br « 30-4.
CgHgO0i:NBr requires Br as :N-Br=30-45 per cent.
It is readily transformed into 4>chloro-2-bromopropionanilide when
its solution in chloroform containing a few drops of acetic acid is
allowed to stand for some days in a sealed tube.
o-ChloropropionanUide, C^H^Cl-NH-CO-CHj-O Bg.
This substance is best prepared by heating o-chloroaniline with
propionic anhydride. Considerable heat is evolved on mixing these
substances, and the action is complete after an hour's heating at 120°.
The product is very readily soluble in ordinary solvents, and crys-
tallises well from petroleum (b. p. 80 — 100°) in groups of colourless,
glistening, needle-like plates melting at 91° :
0-1831 gave 0-1421 AgOl. Cl= 19-19.
CgHioONCl requires 01 « 19*31 per cent.
PropUmyl o-Chhrophsnyl Nitrogen Chloride, CgH^Ol-NCl-OO-OHj-OHg.
This compound crystallises from light petroleum in groups of glisten-
ing, colourless, transparent rhombic plates with domed edges and
melts at 57° :
0-1969 liberated I = 179 cc. iVyiO iodine. CI as :N-C1 - 1611.
CgHgOClINCl requires CI as IN-Cl- 16-26 per cent.
It is more stable than the corresponding j7-chloro-derivative, but is
transformed quantitatively into 2 : 4-dichloropropionanilide if heated
at 150° in a sealed tube with a few drops of propionic acid. When
heated above its melting point, it decomposes at about 220°, but the
product is dark in colour and contains only a little of the isomeric
anilide.
Propionyl o-Chlorophenyl Nitrogen Bromide, CgH^Cl-NBr'CO-CHj-CHj.
This nitrogen bromide crystallises from petroleum of low boiling
point in elongated, four-sided rhombic prisms, often 1 — 2 cm. in length,
of a very pale yellow colour, and melts at 106°. It is much less
soluble in all solvents than the other nitrogen halogen compounds
described in this paper :
0-3625 liberated I = 27-5 cc. iT/lO iodine. Br as IN-Br = 30-33.
CgH^OClINBr requires Br as :N-Br = 30-45 per cent.
When slowly heated in a sealed tube with a few drops of propionic
acid at about 140°, it is transformed into 4-chloro-6-bromopropion-
amlide.
Digitized by
Google
642 CHATTAWaY: NITROaiEN CHIiOBiDlBd
2 : i-Dichlaropropionanilide, CgHgClj-NH-CO'CHj-CH,.
This compound can be prepared by the direct chlorination of
propionanilide dissolved in acetic acid, but is most easily obtained by
beating 2 : 4-dichloroaniline with the equivalent quantity of propionic
anhydride at 160° for 3 hours. It crystallises from alcohol in long,
colourless, flattened prisms and melts at 121°. It is easily soluble
in chloroform and crystallises well from this solvent in long, silky
needles :
01484 yielded 0-1946 AgCl. CI = 32-42.
CgH^ONCls requires 01-32*52 per cent.
Prapionyl 2 : ^-Dickloropkenyl Nitrogen Chloride,
CgHaOlj-NCl-CO-CHj-CHs.
This is most easily prepared by dissolving 2 : 4'dichloropropionanilide
in a little cold glacial acetic acid, and slowly pouring into this an
excess of a strong solution of bleaching powder. A pale yellow oily
liquid is thrown down. This is extracted with petroleum of low
boiling point, washed well with water and a dilute solution of potassium
bicarbonate, and dried. On allowing the solvent to evaporate in a
vacuum, the nitrogen chloride separates in glistening, colourless, trans-
parent, flat rhombic plates, apparently rectangular with domed edges,*
and melts at 64°. It is extremely soluble in chloroform, but cannot
be crystallised from this solvent, as a mere trace of it prevents solidi-
fication.
This compound can also be easily prepared by dissolving propion*
anilide, jo-chloropropionanilide, or 2 : 4-dichloropropionanilide in cold
glacial acetic acid containing an excess of sodium acetate, and passing
chlorine into the cooled solution until it is no longer absorbed. On
adding water to the product and extracting with petroleum as above, a
theoretical yield is obtained :
0-3204 liberated I = 25-4 c.c. iT/lO iodine. 01 as :N-01 = 1405.
OgHgOOlg'.NOl requires 01 as :N-01 = 14-04 per cent.
Propionyl 2 : 4-dichlorophenyl nitrogen chloride is very stable, and
can be heated at 100° in a sealed tube for a long time without change.
It undergoes transformation, however, into 2 : 4 : 6-trichloropropion*
anilide on heating for a short time at 150° in a sealed tube with a
* Most of the hydrogen halogen deriTatires containing the propionyl group crys*
tallise from petroleum in modifications of this form and are thus distinguished from
the acetyl derivatires, which, as a rule, separate from this solrent in prisms with
domed ends.
Digitized by
Google
CONTAINING THfi PttOPlONtL GROtJP. 643
little propionic acid. When heated alone above 100°, it slowly darkens
in colour, and at about 220 — 230° decomposes, giving off a little
chlorine, and leaving a dark coloured, tarry mass. When heated with
water, it is slowly hydrolysed, 2 : 4-dichloropropionanilide and hypo-
chlorous acid being formed. When it is treated with cold hydro-
chloric acid, chlorine is rapidly evolved, and 2 : 4-dichloropropionanilide
regenerated.
Fropionyl 2 : i-Dichlarophenyl NUrogen BromidSy
O^HjOlj-NBrCO-OHj-CHs.
This crystallises from petroleum at low boiling point in glistening,
transparent rhombs of a very pale yellow colour, and melts at 66° :
0-4374 liberated I = 29-5 c.c. iVyiO iodine. Br as :N-Br=- 26-96.
CjHgOClalNBr requires Br as IN-Br = 26*92 per cent.
It is readily transformed into 2 : 4-dichloro-6-bromopropionanilide
when heated for a short time at 140° in a sealed tube with a few drops
of propionic acid. Heated alone in an open tube, it begins to darken
in colour at about 110°, and is transformed with some decomposition
at 160—170°.
2 :'4 : Q-TrieUoropropionanUide, OgFjOlj-NH-CO-OHj-CHg.
This was obtained by heating together 2:4: 6-trichloroaniline with
a slight excess of propionyl chloride until hydrogen chloride was no
longer evolved, the temperature being finally raised to about 180°. It
was recrystallised from alcohol, in which it is readily soluble, and from
which it separates in long, colourless, flattened prisms, melting at 161° :
0-1473 gave 0-2604 AgOl. CI - 4203.
O^HgONOls requires 01 « 42-12 per cent.
Propionyl 2:4: ^'Triehlorophenyl Nitrogen CUoride,
C^HjOla-NOl-CO-CHj-CHj.
This substance is best prepared by adding an excess of a solution
of bleaching powder to a solution of the anilide in glacial acetic acid.
It separates at first as a yellow oil which after some time solidifies.
It crystallises well from petroleum of low boiling point in clusters of
small, elongated plates which have a pearly appearance when pressed
together ; these melt at 80°. When heated above its melting point,
it decomposes at about 230° :
0*1618 liberated 1-11-3 c.c. NjlO iodine. 01 as :N-01 = 12-38.
OgHyOOVNCl requires 01 as :N*01- 12-35 per cent.
Digitized by
Google
644 RIy : DIMERCURAMMONIUM NITRITE AND ITS
Fropionyl 2:4: 6-Triehlarophenyl NUrogen Bromide,
CeHaClg-NBr-CO-OHj-OHg.
This compound crystallises from petroleum (b. p. 80 — 100*^) in
bright yellow, rhombic plates and melts at 106°. When heated above
its melting point, it decomposes about 180°, giving off bromine :
0-2660 liberated 1=16 c.c. i\r/10 iodine. Br as IN-Br = 24-05.
CoH^OCljINBr requires Br as IN-Br =24-12 per cent.
It may be noted that the nitrogen bromides derived from chloro-
propionanilide and from 2:4: 6-trichloropropionanilide, which have a
symmetrical structure, are of a bright yellow colour, whilst those
derived from o-chloropropionanilide and 2 : 4-dichloropropionaniUde,
which have an unsymmetrical structure, are very pale yellow in
colour.
St. Bartholomew's Hospital and College,
London, E.C.
LXVI. — JDimer(mr(mimonium Nitrite and its Haloid
Derivatives.
By Fbafulla Chandra RAy, D.Sc. (Edin.).
SiNCB the publication in the Proceedings (1901, 17, 96) of a note on
the subject of the present paper, I have, after fuller consideration
of the facts, arrived at somewhat different conclusions from those
I had then come to; when, therefore, any statement in this paper
differs from what is found in that note, it is to be taken as
expressing my later view of the matter. The experimental results
remain unaltered.
It has already been pointed outr by me {Zeit anorg. Chem,, 1896,
12, 365 ; Trans., 1897, 71, 337) that mercurous nitrite is partly
decomposed by wat^r into mercury and a solution of both mercurous
and mercuric nitrites, and that when this solution has been prepared
by triturating the salt with cold water, about 22 per cent, of the salt
dissolves without decomposition.* The proportions of the two nitrites
* In my paper on mercarous nitrite in the Annalen (1901, 816, 250), it is stated
(p. 262) that much the greater part of this salt dissolyea in water unchanged ; bat
that is a clerical error, made in translating. Another error in that paper is the
reference (p. 250) to the notice of this salt published by Divers and Haga in the
Transactions. The reference to the statement of these chemists about merourons
Digitized by
Google
HALOID DERIVATIVES. 645
in solution may therefore be formulated by 4Hg(N02)2 + (HgNOg)^.
The addition of sodium chloride precipitates the mercury of the mer-
curous salt and leaves in solution mercuric and sodium nitrites, no
doubt as one or more double salts, similar to that described recently
by Rosenheim {Zeit. cmorg, Chem,, 1901, 28, 171). It is probable
that the solution of the two mercury nitrites contains them as double
salts also, mercuric-mercurous nitrites, firstly, because neither of the
two is obtainable in solution by itself ; secondly, because mercurous
nitrite dissolving in a solution of sodium or potassium nitrite is wholly
decomposed into metal and mercuric nitrite combined with the alkali
nitrite, and, thirdly, because, as pointed out in earlier papers, the solu-
tion of the mixed nitrites is neutral to litmus, whereas even mercuric
chloride is acid in absence of an alkali chloride. On the other hand,
however, it should be mentioned that during the spontaneous evapora-
tion of this solutioD, hydrated mercurous nitrite crystallises out
(Trans., 1897, 71, 340). The non-acidity of these mercuric-mercurous
nitrite and mercuric-sodium nitrite solutions, as well as their stability
on dilution, distinguishes them from that of either of the mercury
nitrates, and other differences are to be found in their behaviour with
urea and with sodium sulphate, neither of which precipitates them
(Trans., 1897, 71, 1103).
Feeling that the formation of stable and neutral double salts of
mercuric nitrite gives support to the view that nitrites have a consti-
tution allied to that of haloid salts, as distinct from that of ozylic
salts such as the nitrates, it occurred to me that further insight into
the matter might be gained by a study of the action of ammonia on
mercuric-sodium nitrite. Ammonia yields with the mercuric-sodium
nitrite solution an insoluble compound having the composition ex-
pressed by the formula NHg^NO,) together with a little less than
half a mol. of water, of which it is partly deprived when exposed in
a desiccator; it is slightly decomposed in the steam-oven. Since
mercuric nitrate, similarly treated, gives also a somewhat
hydrated precipitate, NHgjNOg, which, according to Pesci, becomes
anhydrous when thoroughly washed with boiling water, a treatment
hardly ' possible in the case of the nitrite, the composition of the
ammoniated nitrite fails to throw the expected light on the consti-
tution of nitrites.
The ammoniated mercuric nitrite and its haloid derivatives have,
however, a special interest in the support they a£Eord to the dimercur-
ammonium theory propounded by Eammelsberg in 1888 (/. pr, Chem,,
[ ii ], 38, 558), and since extended to all ammoniated mercury com-
nitrite la also erroneoualy given in the footnote to my paper on this salt in the
Transactions (1897, 71, 387). In both cases it should have been to Proc., 1886, 2,
250, and Trans., 1887, 51, 49.
VOL. LXXXI. X X
Digitized by
Google
646 rIt: dimebct^rammonium nitrite and its
pouDds by Fesci (Gazzetta, 1890, 19, 509, and 20, 485). The
anhydrous nitrate is known (Rammelsberg, Pesci), and the anhydrous
bromide (Pesci), as well as the impure anhydrous chloride (Wehl,
Rammelsberg), but in the rest of the ammoniated mercury compounds
the dimercurammonium salt assumed to be present is so either in
combination with water, or with ammonium salt, or with mercury
salt. The double salts, however, behave in every respect as such
(Rammelsberg, Pesci), familiar instances of which are the two ' white
precipitates,' NHgj01,NH4Cl and NHg2Cl,3NH4Cl. But hydrated di-
mercurammonium salts — Millon's, or oxydimercurammonium, salts —
do not yield water until the salt itself decomposes. This water is,
however, readily displaceable by ammonium salts (Pesci, Andr6, and
others), whilst its fixity to heat before decomposition of the salt may
be no more than what is seen in the like behaviour of hydrated
aluminium chloride, hydrated magnesium chloride, and many other
salts. K. A. Hofmann and Marburg, it is true, deny the existence of
dimercurammonium salts {Annalen, 1899, 305, 191), but Pesci's reply
to them (Zeit cmorg. Chem,^ 1899, 21, 361) is an experimental refuta-
tion of most of their objections.
Now, the ammoniated nitrite, having at most only half the hydra-
tion necessary to constitute the oxydimercurammonium nitrite, must
be at least half dimercurammonium salt. Not only is this the case,
but it has yielded me, in a way to be described presently, a dimercur-
ammonium chloride and a bromide, each also with only half the hydra-
tion demanded by the oxydimercurammonium constitution. Finally,
this half mol. of water can seemingly be displaced by a half mol. of
either mercuric bromide or chloride. The half-hydrated dimercur-
ammonium chloride had already been obtained by Andr6 in 1889
{CompL rend,, 108, 1164), although unknown to me when I was
examining it. Andre's method of getting it is perfectly definite, and
consists in treating mercuric chloride in dilute solution with ammonia
equivalent to one-fourth of its chlorine in presence of potassium hydr-
oxide equivalent to the rest, thus ensiu*ing the non-formation of any
ammonium chloride. My success was gained in essentially the same
way, for I had present only just enough ammonia to supply the nitrogen
of the dimercurammonium. Having prepared from the ammoniated
nitrite the new double mercuric-ammonium salts, 2HgC]2,NH^Cl and
2HgBr2,NH^Br, I treated them each with potassium hydroxide, and
in this way secured the conditions I have mentioned. The interaction
occurs according to the equation :
2(2HgCl2,NH401) + 8K0H = 2NHg2Cl,HjO + 8KC1 + 7Rfi,
as I proved by finding always four-fifths of -the chlorine in the mother
liquor. By using a little less potash, I have sometimes succeeded in
Digitized by
Google
HALOID DERIVATIVES. 647
getting the salt 2NHg2Br,HgBrj{, and, in an impure state,
2NHg2Cl,Hg01g, which salts, it is evident, oan only be written down
as dimercurammonium compounds.
EXPSBIMENTAL.
Dimerew'ammormim Nitrite, — In order to prepare this salt, merourous
nitrite is acted on by water so as to obtain a solution of mercurio-
merourous nitrite, which is then changed to one of mercuric-sodium
nitrite by careful addition of just enough sodium chloride. As in the
case of silver nitrate and sodium chloride, the end point is hard to hit
off, and with every care a minute but unimportant quantity of either
mercurous salt or of chloride must be left in solution. Another and
better way of preparing mercuric-sodium nitrite solution is to dissolve
mercurous nitrite in a concentrated solution of sodium nitrite, when
half the mercury separates as metal ; but with this I only became well
acquainted too late for use in the work of this paper. Solutions of
mercuric-sodium nitrite can also be prepared, by adding either mercuric
(Eosenheim) or mercurous nitrate to sodium nitrite solution, but
these, of course, contain sodium nitrate, which, for this investigation,
would be an undesirable, although probably inactive, constituent of the
solution. To the mercuric-sodium nitrite solution, filtered from the
mercurous chloride, dilute ammonia is gradually added until in slight
excess. A cream-coloured, flocculent, but somewhat dense precipitate
is produced, which is washed on a filter by aid of the suction-pump.
Being slightly decomposed in the steam-oven, it has to be dried over
sulphuric acid, and as it then collects into compact lumps, these are
broken up and again exposed in the desiccator in order to facilitate the
drying. The dry salt is pale yellow.
Nitrogen was determined by combustion with oxide of copper and
chromate of lead. Although the analyses of the salt are not sufficient
to decide the point, there is strong presumption in favour of the salt
being anhydrous, as there is no guarantee that all moisture was removed
from the preparations analysed :
Mercury. Nitrogen.
2NHg2N02,H30 requires 83-68 5-85
/84-13 5-99
84-64 6-02
84-87 5-92
85-30 600
85-94 —
.86-02 —
NHgjNOg requires 86-96 608
When heated in a bulb-tube, the salt decomposes without fusing,
X X^
Digitized by VjOOQIC
Found .
648 rIy: dimercurakmonium nitrite and its
giving o£E nitrous fumes, mercury, and water, and leaving a yellowish
residue, mostly mercuric oxide.
Hofmann and Marburg find that * infusible white precipitate ' yields
the whole of its nitrogen as ammonia when it is heated at 130° for
half-an-hour or more with a 20 per cent, solution of potassium hydr-
oxide. I have repeated their experiments and can confirm their
experience, having obtained 5'36 per cent., calculation giving 5*57.
Dimercurammonium nitrite, however, could not be made to yield more
than 2 '5 per cent, of nitrogen by this treatment, whilst theory requirejB
3 per cent.
New Mercuric Ammonium Chloride and Bromide. — Like other salts
of its kind, dimercurammonium nitrite dissolves readily in warm
hydrochloric acid or hydrobromio acid ; nitrous fumes escape, and the
solution leaves, when evaporated, a white, crystalline mass which
volatilises slowly, even on the water-bath. It proves to be a new mercury
and ammonium salt, chloride, or bromide, according to the acid used,
having the composition shown by the formula 2HgOl2,NH4Cl or
2HgBr2,NH46r. The double chloride fuses and sublimes at a gentle
heat, and is very soluble in water. The double bromide is decomposed
by water into its constituent salts, and when the sparingly soluble
mercuric bromide has been dissolved again by stirring it with the
boiling solution of ammonium bromide, it separates as the solution
cools in white, nacreous crystals. The presence of free acid seems to
be necessary for the formation of these double salts, and I have, in
fact, been able to prepare the double bromide by dissolving the single
salts in the indicated proportions in presence of hydrobromic acid and
evaporating the solution, but I have not succeeded in a similar way in
getting the double chloride. The double bromide forms prisms and
tablets, probably triclinic. The results of analyses of the two double
salts are as follows :
Mercury. Chlorine. Nitrogen.
Gale, for 2HgCl8,NH^01 6717 2981 235
Found 67-92 2900 196
67-94 29-20 —
67-20 29-42 —
Mercury. Bromine. Nitrogen.
Calc. for SHgBrjjNH^Br 4890 4890 1'71
Found 49-29 48-31 166
49-38 47-86 —
48-66 — —
Production of Dimercurammonium Chloride and Bromide from the
above-deaeribed Double Salts. — ^When the double chloride is stirred
with excess of dilute solution of potassium hydroxide, it gives a
pale yellow precipitate having the composition of, and being apparently
Digitized by
Google
HALOID DERIVATIVES. 649
identical with, Andr6's half-hydrated dimercurammonium chloride,
2NHg2Cl,H20, according to the results of analysis, which were :
Mercury. Chlorine. Nitrogen.
Calc 87-25 7*74 3-05
Found 87-10 7-54 2*90
„ 87-94 — —
„ 87-06 — —
Tt may be heated to 150^ without appreciably losing water. The
filtrate from this precipitate contains exactly four-fifths of the chlorine
of the double salt; found : 24-10 and 23*98, instead of 23-85 » 4/5 of
29-81 per cent.
When treated with excess of potassium hydroxide, the double bromide
yields a deep red precipitate which has the composition expressed
by 2NHg2Br,H,0, after drying at 100°. There is left in the filtrate
39 per cent, of the bromine of the double salt, which is 4/5 of the
total, 48*9. Analysis of the precipitate gave :
Mercnry. Bromine. Nitrogen.
Oalc 79-52 15-91 2*78
Found 79-82 16*66 2-60
„ 79-97 16-10 2-54
New Mercuric^mmonium Chhrobramide. — When dimercurammonium
bromide is dissolved in hydrochloric acid and the solution concentrated,
the double salt, 2HgCl2,NH4Br, is obtained. Dimercurammonium
bromide is re-obtained quantitatively by mixing this salt in solution
with excess of dilute potassium hydroxide, all the chlorine, and no
bromine, remaining in solution. Thus, 22-37 and 23-02 per cent, of
chlorine were found in solution instead of 22*19, given in the subjoined
table :
Mercury. Chlorine. Bromine. Nitrogen.
Calc 62-50 22-19 12-50 2-19
Found 63-01 20*89 12-71 1-97
63-5 _ _ _
„ 62-89 _ _ -^
It is somewhat remarkable that the compound dimercurammonium
chloride, when dissolved in hydi'obromic acid and the solution con-
centrated, does not yield the bromochloride, 2HgBr2,Nn^01, but the
double bromide, 2HgBr2,Nn^Br ; in other words, the chlorine is re-
placed by bromine. This reaction has its analogy in Field's experiment
(Trans., 1858, 11, 234).
Dimu'iyurcmimonium'fnercuric Bromide and Chloride, — By adding
potassium hydroxide, gradually and not in excess, to a boiling solution
of the double bromide of mercury and ammonium, I have obtained the
Digitized by
Google
650 MORGAN: INFLUENCE OP SUBSTITUTION ON THE
salt 2NHg2Br,HgBr2, which is, however, not always easy to get. The
analyses of three distinct preparations are given below :
Mercury. Bromine. Nitrogen.
Calc 74-19 23-74 2'07
Found 73-55 24-30 2-36
72-65 22-82 —
73-44 21-99 —
I have not succeeded in getting the corresponding chloride in a pure
state, but a yellowish-white precipitate, which can hardly be anything
else but impure dimercurammonium-mercuric chloride, is obtained
when, to a dilute solution of the mercuric-ammonium chloride, above
described, potassium hydroxide is very gradually added with care, to
avoid using it in excess. The calculated percentage composition of
2NHg2Cl,HgCl2 is mercury, 85*5; chlorine, 12*1; and nitrogen, 2 '4,
whilst the mean numbers of some half-dozen fairly concordant analyses
of as many different preparations of this precipitate are, respectively,
84-5, 11*3, and 3*0. The same numbers would express the percentage
composition of the following mixture, which, it will be seen, consists
mainly of dimercurammonium-mercuric chloride, 2NHg2Cl,f HgClj +
JNH,Cl-H§H,a
In conclusion, I wish to express most cordial thanks to Professor
Edward Divers, F.R.S., for some valuable suggestions, which have
greatly aided me in putting this paper together.
Chemical Labor atort,
Presidbnct Colleob,
Calcutta.
LXVII. — Influence of Suhstitiition on the Reactivity of
the Aromatic m-Diamines.
By GiLBEBT Thomas Moboan, D.Sc.
It has already been shown in a recent communication (this voL, p. 86)
that the reactivity of the aromatic fTi-diamines towards diazonium salts
may be very considerably modified by the introduction of snbstitnent
radicles into certain positions in the aromatic nucleus. It was found
that the symmetrically disubstituted m-diamines furnished only small
quantities of azo-compounds, whereas their isomerides containing a
free para-position with respect to an amino-group gave rise to aio-
derivatives in theoretical quantities.
Digitized by
Google
ttEAGTlYITY OF Tfi£ AROMATIC MBTADI AMINES. 661
The behaviour of the homologues of m-phenylenediamine on methyl-
ation has now been investigated, and the results indicate that the
diminution in reactivity, brought about by the introduction of radicles
into the para-ortho-positions with respect to the nitrogen atoms, is
equally noticeable when the diamines are subjected to the action of
alkylating agents.
The influence of substitution on the methylation of primary aromatic
amines has already been studied hj Finnow {Ber,y 1897, 30,3110;
1899, 32, 1401 ; 1901, 34, 1129), who treated the hydrochlorides or
hydrobromides of these bases with methyl alcohol at 145 — 190°. Under
these conditions, a monoamine in which the two ortho-positions
adjacent to the nitrogen atom are both unoccupied yields a mixture of
tertiary base and quaternary salt. If, however, the amine contains a
substituent radicle in one of these positions, it gives rise to a dialkyl
derivative only. It remained uncertain as to whether this rule holds
for the diamines, for although the methylation of benzidine and the
three phenylenediamines had been investigated (Pinnow, loc, ot^.), the
research was not extended to the homologues of these diamines.
An examination of 9n^tolylenediamine (2 : 4-diaminotoluene) and
4 : 6-diamino-m-xylene from this standpoint shows that Pinnow's
generalisation may be extended to the m-diamines, the former of these
bases furnishing a mixture of 2 : i'tetramethyldiaminotolii&nd and the
quaternary hydrochloride or hydrobromide, the latter yielding 4 : 6-
tetramethyldiamino-mrxi/lene as the sole product.
A comparative experiment made with 97i-phenylenediamine led to
the production of the tertiary base and the quaternary salt together
with fluorescent, tarry impurities, the alkylation taking place more
energetically, but less smoothly, than with the homologous diamines.
Diaminomesitylene, when treated in this manner, remains practically
unchanged.
These results indicate that the successive introduction of methyl
groups into the three positions (X, Y, Z) adjacent to the two nitrogen
atoms of m^phenylenediamine,
NHgY
— Z
is attended by a well-marked diminution in the reactivity of the
diamine towards methyl chloride or bromide, the interaction being
prevented when the substitution of the three contiguous hydrogen
atoms is complete (compare E. Fischer and Windaus, B&r,, 1900, 33,
345 and 1967).
The tertiary diamines obtained in this investigation exhibit a
gradation in properties similar to that observed in the case of the
Digitized by
Google
652 MORGAV: INFLUENCE OF SUBSTITUTION ON THE
corresponding primary bases; their reactivity diminishes as the
homologous series is ascended. When distilled under diminished
pressure, these tertiary bases are obtained as almost colourless, oily
liquids ; tetramethyl-m-phenylenediamine, however, rapidly darkens on
exposure to light and air, whilst tetramethyl-4 : 6-diamino-m-zylene
remains unchanged even after a year. The intermediate homologue,
tetramethyl-2 : 4-diaminotoluene, becomes darker on keeping, but
much less rapidly than the first member of the series.
The two lower homologues readily react with diazo-compounds, and
the azo-colours obtained by the action of diazotised primulin closely
resemble those produced from the corresponding primary diamines.
2 : 4>Tetramethyldiaminotoluene readily condenses with j9-nitrobenzene-
diazonium chloride, giving rise to ]^UrohenzenS'6-{izO'2 : i-tetramethyl-
diaminotoluenef
NMe.
Me
NOg-OeH^-Nj^ )>NMe2 ,
in theoretical yield, and similar results are obtained with other
diazonium salts. 4 : G-Tetramethyldiamino-m-xylene does not combine
with diazotised primulin, and when treated with a solution of j9-nitro-
benzenediazonium chloride containing sodium acetate, it remains
entirely unchanged.
This inactivity of the symmetrically disubstituted tertiary diamine
should be contrasted with the behaviour of 4 : 6-diamino-m-xylene
under similar conditions (this vol., p. 88), for it furnishes additional
evidence in support of the view that the readily decomposable initial
product of the action of a diazonium salt on the primary base is an
unstable diazoamine which subsequently undergoes transformation into
tarry products and a small amount of azo-compound,
MeNH-NINX Me NH,
<Z> - CD"---
Me NHa Me NHj
In the present instance, however, where the initial attachment of
the diazo-residue to one of the nitrogen atoms is rendered impossible
by the complete alkylation of the amino-radicles,
MeNMcj
<
Me NMe,
the production of an o-azo-derivative is altogether prevented.
This hindrance to the formation of azo-derivatives, due to alkylation.
Digitized by
Google
REACTIVITT OF THE AROMATIC METADIAMINES. 653
although so strikinglj manifested in the case of the preceding tertiary
diamine, does not obtain among the tertiary amines containing a free
para-position with respect to nitrogen. As already indicated, 2 : 4-
tetramethyldiaminotoluene yields azo-compounds quite as readily as
2 : 4-tolylenediamine itself, in this respect resembling dimethylaniline,
which is employed in the production of several well-known azo-
colouring matters.
The difference in reactivity between 2 : 4-tetramethyldiaminotoluene
and 4 : 6-tetramethyldiamino-97i-zylene is also demonstrated by their
behaviour towards formaldehyde. The lower homologue readily
interacts with this reagent, yielding 2:4:2': 4i''OctamethylUtraminodi'
tolyl-b : ^'-meihane^
Me Me
NMe,/ VcHg-/ NnMoj .
NMcj, NMe,
The disubstituted tertiary diamine, on the contrary, is not affected by
formaldehyde, even when the experiment is carried out under pressure
at high temperatures in the presence of excess of acetic anhydride.
Experimental.
Methylation of m'Tolylenediamine.
fn-Tolylenediamine hydro bromide and an excess of alcohol (6 — 7 mols.)
were heated in sealed tubes at 180° until the product on cooling no
longer deposited a crystalline hydrobromide, this result being usually
obtained in 8 to 10 hours. The pressure was released after 4 hours'
heating in order to prevent the tubes from bursting. The product,
after being heated on the water-bath to remove the unaltered methyl
alcohol, was treated with excess of potassium hydroxide solution, the
oil which separated being extracted with ether. The liquid of high
boiling point remaining after the removal of the ether distilled at
254 — 259° under 757 mm. pressure, the greater portion boiling between
255° and 256°.
2 : i-Tetramethyldiaminotoliiene, thus obtained as a pale brownish-
yellow oil, was further purified by repeated distillation under reduced
pressure and finally boiled at 148 — 150° under 24 — 26 mm. pressure ;
it has a sp. gr. 09661 at 24° and does not solidify at - 10°.
The platinichlaridej OiiHigNjjHaPtCl^, separates from aqueous solu-
tions in stellar aggregates of slender, transparent, yellow prisms ; it
rapidly darkens when left in contact with the mother liquor, and is
decomposed by boiling water, evolving formaldehyde and yielding a
dark red solution. The salt when dried at 80 — 90° gave the following
numbers on analysis :
Digitized by VjOOQIC
654 MORGAN : INFLtTENCE OP SUBSTITttTION ON Tflfi
0-2642 gave 0*0840 Pt. Pt = 33-04.
C^iHjoNgOlgPt requires Pt = 3316 per cent.
The picrate crystallises from ethyl acetate in hard, transparent,
yellow prisms, and melts at 162 — 163^ ; it is formed by the union of
picric acid and the diamine in molecular proportion :
0-1479 gave 22-4 c.c. moist nitrogen at 16° and 760 mm. N = 17-66.
Ci^HjjOyNg requires N = 17-20 per cent.
The methol>r<mide, NMej-OgHgMe-NMegBr. — ^The solution of potass-
ium hydroxide employed in liberating the tertiary base assumed a
dark brown colour, and when saturated with the solid reagent yielded
a heavy brown oil which rapidly solidified to a mass of brown, acicular
crystals. This substance, which contained bromine, was identified as
the quaternary bromide by conversion into the corresponding chloride
and subsequent formation of the platinichloride. A solution of the
bromide was triturated with moist silver oxide and filtered, the filtrate
being then treated with hydrochloric acid and platinic chloride.
The platinichloride, NMeg'C^HgMe'NMeg.HPtClQ, which separates
out may be crystallised from hot water without decomposition ; it is
deposited on cooling in transparent, brownish-yellow prisms and is
far more stable than the corresponding salt of the tertiary base :
0-2269 gave 00736 Pt. Pt « 32-43.
0-1861 „ 0-0603 Pt. Pt = 32-40.
Ci2H25NjClgPt requires Pt = 32-39 per cent.
The remainder of the methobromide was converted into the tertiary
diamine either by heating with ammonia under pressure (Pin now, loe,
cit.)j or by treating the quaternary salt with moist silver oxide and
evaporating down the filtered solution of the quaternary ammonium
hydroxide.
The methylation follows the same course when 97»-tolylenediamine
hydrochloride is employed, the product in this case being a mixture
of the tertiary base with the methochloride.
Methylation qf 4 : ^-Diamino-m-xylene.
A mixture of 4 : 6-diamino-m-xylene hydrochloride and excess of
methyl alcohol was heated in sealed tubes at 180° until it remained
viscid on cooling ; the product, after being heated on the water-bath
to remove excess of methyl alcohol, was rendered alkaline with strong
potassium hydroxide solution and extracted with ether. The ethereal
extract, when dried over potassium hydroxide, and subsequently dis-
tilled, yields an oily base boiling at 243 — 245° under 757 mm. pressure.
4 : ^TeWamethyldiamiruhm^cylenef a pale brownish-yellow oil boiling
Digitized by
Google
REACTIVITY OF THE AROMATIC METADIAMINES. 655
at 124 — 126° under 12 mm. pressure, remains liquid at - 10° and has a
sp. gr. 0*9434 at 18°; it does not darken perceptibly when kept for
12 months in diffused light.
The platiniddoride, 0^^e^(^Me^^JI^tO\^ separates from an
aqueous solution of its generators in transparent, orange-yellow, acicular
prisms \ it is moderately soluble in water, melts indefinitely above 200°,
and is far more stable than the corresponding salt of 2 : 4-tetramethyl-
diaminotoluene ; it may be crystallised from hot water without decom-
position, and separates in reddish-orange needles often more than an
inch in length :
0-2682 gave 0-0861 Pt. Pt = 32-10.
OijHjjjNgCl^Pt requires Pt = 32-39 per cent.-
The puyraU, O^Ei^Q^i^MQ^^,Q^(^0^^*OILy obtained by mixing
together alcoholic solutions of its generators, crystallises from ethyl
acetate in hard, transparent, yellow, rhombic prisms, sparingly soluble
in the alcohols, but dissolving more readily in acetone ; when rapidly
heated, it melts at 202—203°, but slowly decomposes when maintained
at 193—195°:
022 gave 30-5 c.c. moist nitrogen at 14° and 756 mm. N« 16*22.
Oi8Hjg07Ng requires N« 16*62 per cent.
4 : 6-Tetramethyldiamino-l : 3-xylene is produced with equal readi-
ness from 4 : 6-diamino-m-zylene hydrobromide and methyl alcohol, the
tertiary base being the sole product whether the hydrochloride or the
hydrobromide is employed.
Diaminomesityleno hydrochloride, when heated with methyl alcohol,
does not undergo methylation, the greater part of the primary diamine
being recovered on working up the product.
In preparing this diamine from dinitromesitylene (m. p. 86°) by the
action of iron filings and water acidified with hydrochloric acid, it was
found that only one of the nitro-groups became involved in the
reduction, so that nitromesidine crystallised out from the filtered
solution in golden-yellow needles melting at 70°. The complete reduc-
tion was accomplished by means of excess of tin and hot hydrochloric
acid, but even with this reagent the action was very slow.
Comparative experiments made on the methylation of TTirphenylene-
diamine showed that the action is less under control than in the case
of the homologous diamines, the production of the tertiary base and
quaternary salt being attended by the formation of fluorescent, tarry
impurities.
The tertiary base, tetramethyl-m-phenylenediamine, after treatment
with acetic anhydride to remove any secondary amines, was purified
by repeated rectification under reduced pressure, and finally ob-
Digitized by
Google
656 MORGAN: INFLUENCE OF SUB8TITCJTI0N ON THE
tained as a brownish-yellow oil rapidly darkening on exposure,
boiling at 151 — 157^ under 26 mm. pressure, and having a sp. gr.
0-9934 at 27° (compare Romburgh, Eev. Trav. Chim,, 1888, 7, 3 ;
Wurster and Morley, Ber., 1879, la 1814).
Tertiary m-DuKninea and Diasaonium Salts,
When a piece of cotton cloth impregnated wifch diazotised primulin
is introduced into an aqueous solution of tetramethyl-m-phenylene-
diamine or its tolylene homologue, a brownish-red azo-colouring matter
is developed on the fibre, the colour of which closely resembles that
obtained by the use of the corresponding unalkylated m-diamina
4 : G-Tetramethyldiamino-m-xylene, on the other hand, does not com-
bine with this diazo-compound.
]P'Nitrohenzen6'5-azo-2 : i-tetramethyldiaminotohiene, — A. solution of
p-nitrobenzenediazonium chloride, when added to 2 : 4-tetramethyl-
diaminotoluene dissolved in cold dilute hydrochloric acid, produced a
dark-red precipitate the formation of which was completed by the
addition of excess of sodium acetate. This insoluble product was
crystallised from alcohol, separating from its solutions in dark-green
leaflets with a bronze lustre ; the yield was quantitative.
^Nitrobenz6ne-5-azo-2 : 4-tetramethyldiaminotoluene is sparingly
soluble in cold alcohol, dissolving more readily in ethyl acetate and
crystallising from this solvent in leaflets melting at 126 — 127°. Its
solutions in the organic solvents have Sr deep purple colour and the
azo-compound dissolves in cold concentrated sulphuric acid to an intense
brownish-red solution :
0*1808 gave 33*4 c.c. moist nitrogen at 15° and 756 mm. N»21'52.
^i7^2i^2^fi requires N = 21*41 per cent.
The hydrochloride and platinichloride are red precipitates, the nitrate
crystallises from alcohol in deep red leaflets, the picrate separates from
the same solvent in red needles.
When the preceding experiment was repeated with 4 : 6-tetramethyl-
diamino-m-xylene, the introduction of the jp-nitrobenzenediazonium
chloride produced no effect on the solution of the tertiary diamine.
After remaining for 24 hours, the mixture deposited a tarry pro-
duct, due to the gradual decomposition of the diazonium base set free by
the sodium acetate, but the filtered solution stiU contained the un-
altered diamine. This base was completely recovered by treating the
solution with excess of potassium hydroxide and extracting with ether,
being precipitated in the form of its picrate by the addition of an
alcoholic solution of picric acid to the ethereal extract. The yield of
Digitized by
Google
EEACTIVITY OP THE AROMATIC METADIAMINfiS. 657
recrystallised picrate obtained in this operation was 90 per cent, of
the theoretical.
Terti(vry m-Dxamines and Formaldehyde.
Six grams of 2 : i-tetramethyldiaminotoluene dissolved in 20 grams
of acetic anhydride were treated with 2 c.c. of 40 per cent, formalde-
hyde solution and allowed to remain for 20 hours ; the mixture was
then heated to boiling, subsequently cooled, and rendered alkaline with
excess of ammonia. The oily product which first separated rapidly
solidified, and, after drying on a porous tile, was crystallised from
ethyl acetate or alcohol.
2:4:2': i''0ctamethi/ltetraminoditolyl'6 : b'-methane separates in well-
defined, transparent, colourless prisms, either obliquely truncated or
terminated by pyramids ; the crystals often exhibit external twinning
and melt at 86° :
01473 gave 04037^002 and 01322 Hfi: C = 74-74 ; H = 9-97.
0-1833 „ 24-6 c.c. moist nitrogen at 18° and 764 mm. N = 15-58.
CgjHjjjN^ requires C = 7500 ; H « 978 ; N = 1521 per cent.
The tetramine is readily soluble in most organic solvents with the
exception of light petroleum ; it dissolves in dilute aqueous solutions
of the mineral acids, and yields a colouring matter of the acridine
series when heated with hydrochloric acid at 150°. It is not an
analogue of the leuco-base of malachite-green, for it does not give
rise to a colouring matter of this type when oxidised with lead peroxide
and acetic acid.
The picrate separates from ethyl acetate in spherical aggregates of
hard, transparent, yellow crystals and melts at 147 — 148°.
4 : 6-Tetramethyldiamino-m-xylene, when dissolved in acetic anhy-
dride and treated with formaldehyde solution, remains unaltered even
when an excess of the reagents is employed in sealed tubes at 150 — 170° ;
the unchanged base recovered from the mixture boils at 243 — 248°
(corr. 245°), and yields the characteristic picrate (m. p. 202 — 203°).
Royal College of Science, London.
South Kensington, S.W.
Digitized by
Google
658 HENDERSON AND PRENTICE: THE SPECIFIC ROTATIONS OF
LXVIII. — The Influence of Certain Acidic Oxides on the
Specific Rotations of Lactic Acid and Potassium
Lactate.
By George Gerald Henderson and David Prentice, PI1.D.
In the course of an investigation of the compounds produced by the
action of certain acidic oxides on metallic salts of hydroxy-acids in
aqueous solution, which has formed the subject of previous communi-
cations to the Society (Trans., 1895, 67, 102 and 1030 ; 1896, 69,
1451 ; 1899, 75, 542), we endeavoured to prepare arsenio- and antimonio-
lactates by dissolving arsenious and antimonious oxides respectively in
hot aqueous solutions of various metallic lactates, but did not succeed
in isolating any definite compounds, although it was found that the
solutions were capable of dissolving considerable quantities of the
former oxide. Similar experiments with other acidic oxides were
equally unsuccessful, except in one instance, where a crystalline potass-
ium molybdilactate, Mo02(C3H403K)29 was obtained. It was unfor-
tunate that no other derivatives of lactic acid could be prepared in a
sufficiently pure form to justify us in assigning formula to them,
because, as that was the only hydroxy-acid used in our experiments
which contained one alcoholic hydroxyl and one carboxyl group, the
determination of the composition of its derivatives was of importance
with regard to the bearing of the results on our general conclusions
concerning the constitution of the '' tartar emetic " class of derivatives
of the other hydroxy-acids (malic, tartaric, citric, and mucic) which
came within the scope of the research. However, another method of
investigation is available in the case of optically active acid:«, by means
of which it is possible to draw conclusions regarding the composition
of derivatives formed in solution, even although these cannot be
obtained in a form suitable for analysis.
It is well known that the dissolution of such optically inactive com-
pounds as boric acid or antimonious oxide in solutions of the optically
active modifications of malic and tartaric acids, or of their salts, pro-
duces more or less marked alterations in the specific rotations of the
solutions. There can be no doubt that these alterations are due to
the formation of new optically active compounds by interaction of the
inactive oxides and the active acids, because in many instances the
substances of which the existence is indicated by the rotations of their
solutions can be obtained in the solid state. This is true, for example,
of many boro-, arsenio-, and antimonio-tartrates. Moreover, Rosenheim
and Itzig {Ber,<, 1900, 33, 707) have recently confirmed in this way
the existence in aqueous solution of the alkali molybditartrates of the
Digitized by
Google
LACTIC ACID AND POTASSIUM LACTATE. 669
"type Mo02(04H^OgM')2, which were prepared by one of us some time
agOy and Itzig {Ber,, 1901, 34, 1391) has found that the maximum
rotation of a solution of molybdenum trioxide in aqueous ammonium
hydrogen malate is obtained when the substances are present in the
proportions necessary for the formation of the salt Mo02(C4H^05«N 11^)2,
which was formerly isolated and described by one of us. Therefore,
with the hope of gaining the desired information by this method, we
have begun to examine the influence of various acidic oxides on the
rotations of solutions of optically active lactic acid and its potassium
salt. The present paper contains the results of the experiments with
arsenious, antimonious, and boric oxides.
Any compound formed by the interaction of one of these oxides and
a metallic lactate would probably be of the type CHg-CH0(R0)-C02M'
(Trans. 1899, 75, loc, cit,), No such compound is yielded by anti-
monious oxide, which was found to be almost insoluble in hot solu-
tions of potassium lactate. Arsenious oxide, on the contrary, is very
readily dissolved by hot aqueous potassium lactate, and as the quantity
of oxide in the solution is increased, the rotation becomes greater and
reaches a maximum when the solution contains 1 mol. As^O^ to 4 mols.
C3H5O3K, that is, when the substances are present in the proportion
requisite for the formation of an arseniolactate of potassium of the
formula CH3«CHO(AsO)«002K. It is true that the change produced
in the rotation of the solution of potassium lactate by dissolving
arsenious oxide in it is not great, but then, on the other hand, the
oxide is dissolved in much greater quantity than by water, and on the
whole the conclusion seems justified that the solution contains a new
compound. When boric acid is dissolved in a solution of potassium
lactate, a marked change in the rotation is produced. If increasing
quantities are added to a solution of the dextrorotatory salt, the
rotation diminishes, until finally, when the boric acid and the salt are
present in the quantities required for the formation of a borolactate
of potassium of the formula GH3'CIIO(Bo)*002K, the solution is
almost equally strongly laevorotatory. At first sight, the change in the
rotation would appear to be due simply to the liberation of lactic acid,
but then the rotation of the solution is much greater than one of lactic
acid of equivalent strength, and, moreover, boric acid has also a marked
effect on the rotation of a solution of lactic acid. It is probable
therefore that borolactic acid or its potassium salt is formed when boric
acid is dissolved in a solution of lactic acid or potassium lactate.
For the preparation of d- and Mactic acids, we adopted Furdie's
process for the resolution of the inactive acid, namely, crystallisation
of the double zinc ammonium salts (Purdie, Trans., 1893, 63, 1142;
Purdie and Walker, Trans., 1895, 67, 616). The process, although
necessarily somewhat tedious, gave excellent results in the hands of
Digitized by
Google
660 HENDERSON AND PRENTICE: THE SPECIFIC ROTATIONS OF
several workers when the requisite precautions were strictly observed,
and our work was much lightened by Professor Purdie's kindness in
providing us with specimens of the pure active zinc ammonium salts
for starting the crystallisations, and in giving many useful hints on
details of procedure. In addition to the method prescribed by Purdie
for preparing the active acids from their pure zinc salts (conversion of
the zinc into the calcium salt, decomposition of the latter with the
calculated quantity of sulphuric acid, and extraction of the lactic acid
with ether), we employed the following process. A solution of the
pure zinc salt was saturated with hydrogen sulphide, the precipitated
zinc sulphide removed by filtration, and the filtrate concentrated at
the ordinary temperature in a vacuum desiccator containing sticks of
caustic soda until all the sulphuretted hydrogen had disappeared,
which required about forty-eight hours. The solution was finally
evaporated to a small bulk on the water-bath and filtered. Judging by
its 'rotation, the acid thus prepared was quite as pure as that obtained
by the other process.
The solutions of which the rotations were to be examined were all
prepared in practically the same way. For convenience, the solutions
of lactic acid and potassium lactate were made up of normal and semi-
normal strength. The required quantity of the solution of the lactate
was measured into a small flask, the calculated weight of the oxide
added, and the liquid kept boiling gently for about an hour. The
solution and the washings of the flask were then transferred to a
graduated flask, if necessary through a small filter, and the liquid
made up to the mark. All the polarimetric observations were made
at the same temperature, 20°. The length of the tube of the polarimeter
was 200 mm.
Experiments with Antimonious Oxide.
The experiments with antimonious oxide showed that no antimonio-
lactate is formed, at least under the conditions observed. Antimonious
oxide was found to be almost insoluble in solutions of potassium
lactate, even after prolonged boiling ; thus, ior example, 25 c.c. of a
normal solution of the salt dissolved only 2*5 mg. of oxide, and the
same quantity of a semi-normal solution only 3 mg. The rotations
of the solutions were practically unaffected by the presence of a trace
of the oxide.
Experiments with Arsenums Oxide,
In the first set of experiments we used a semi-normal solution of
y potassium lactate (c = 6*405), from which solutions containing arsenious
oxide in the proportions of \ mol., \ moL, f mol., and 1 moL As^Og
Digitized by
Google
Lactic acid and potassium lactate.
661
irespectively to 1 moL CgH^OgK were prepared in the way already
described. The solution was found to be saturated with arsenious
oxide when the latter was present in the proportion of 1 moL of oxide
to 4 mols. of lactate, for although a still larger quantity was soluble
in the hot liquid, the excess separated in the crystalline state as the
liquid cooled.* The saturated solution contains the substances in the
proportion necessary for the formation of potassium arsenio-lactate,
(AsO)C3H^OgK, and the maximum rotation was found at this point.
However, as will be seen from the following figures,t no very marked
change was produced in the rotation of the lactate, which is given for
comparison :
Weight of
QbHbOjK in
Molecular pro-
Weight of
portions of
o*^.
[air
25 C.C. of
A840e added.
CjHeOgK and
% •
solution.
As^Ofl.
1-6013
+ 1-86"
+ 10-64**
it
0-3094
4:}
1-405
9-19
II
0-6188
4:i
1-42
7-99
II
0-9282
4:i
1-44
7-11
II
1-2376
4:1
1-47
6-47
Two other sets of observations were made with solutions of different
strength. In the first, in which a normal solution of potassium lactate
(e= 12-81) was used, readings of the rotation were only taken up to
the point at which f mol. As^O^ was present for each 4 mols. C3H5O3K,
because although the solution was capable of dissolving larger quanti-
ties when hot, part of the oxide always crystallised out on cooling.
At this point, the rotation of the solution had increased from + 2 '58°
to + 2*77°. In the last series of observations, the strength of the
lactate solution was on^fourth normal (c=> 3*2025). In this case* it
was found that, although larger quantities were dissolved on heating,
the maximum quantity of arsenious oxide which was permanently
retained in solution corresponded with that necessary for the formation
of the compound (AsO)03H40gK. At this point, the rotation had
increased from +0*65° to +072° ([a]^ +6*33°).
For purposes of comparison, observations were also made of the
rotations of solutions of arsenious oxide in semi-normal lactic acid
(c«4'5). As will be seen from the following figures, the effect of the
* In some cases, the excess of arsenious oxide did not separate until the liquid had
been kept for several days at the ordinary temperature.
t Some of the experiments were made with d-t others with Mactic acid, or their
salts, but for simplicity the results haye been stated as if ^lactic acid, or potassium
^lactate, had been used in each case.
VOL, Lzxxi. y Y
Digitized by
Google
66i ROTATIONS OF LACTIC AClD AND POTlSSldM LACTATE.
presence of the arsenious oxide was slightly to diminish the rotation
of the solution :
Weight of
CjHeO, in
26 C.C. of
solution.
Weight of
As^Oe added.
Molecular pro-
portions of
CgHaO, and
AsA-
«r
[«]?■.
1-126
it
ft
0-6188
1-2876
4Tj
4:1
-0187'
008
0 07
--2 or
0-57
0-87
Experiments with Boric Add,
For the experiments with boric acid, a normal solution (c» 12*81)
of potassium lactate was employed* The addition of boric acid, in the
quantities given in the following table, produced marked changes in
the rotation of the solution, which reached a maximum when the
boric acid and lactate were present in the proportion (1 mol. : 1 mol.)
necessary for the formation of potassium borolactate (BO)C3H^02K.
At this point, at which the solution was saturated with boric acid, the
solution, which was originally dextrorotatory, had become laavo-
rotatory :
Weight of
CgHBGjK in
Molecalar pro-
Weight of
portions of
oW.
[«]r.
26 CO. of
B(OH), added.
CoHjOaK and
••p
solution.
B(OH),.
8-2026
+2-68'
+ 10-06'
f>
0-3876
1-.;:
+ 1-54
+ 6-86
»f
0-7760
1 :
+ 0-473
+ 1-48
»9
1-1626
l:f
-0-27
-0-77
>>
1-6600
1:1
-1-076
-2-83
It might appear that these changes were due to the liberation of
lactic acid. In order to determine this, the observations on p. 663
were made on the effect of boric acid on the rotation of a normal
solution of lactic add (c*9-0).
At this point, the solution was saturated with boric acid. Another
reading was therefore taken with a semi-normal solution of lactic acid,
which contained, in 25 c.c, 1*125 grams O^Tlfi^ and 0*7750 gram
B(0H)8 (1 mol. : 2 mols.). The specific rotation was found to be — 2*87°.
The examination of the effect of molybdic, tungstic, and other acidic
oxides on the rotations of lactic acid and potassium lactate is at present
in progress.
Digitized by
Google
OXONItJM BALT8 OF B'LUORAN AND ITS DMIVATIVES.
Weight of
CfiH^O, in
26 c.c. of
solution.
Weight of
B(OH), added.
Molecular pro-
portions of
«r.
[•r-
2-26
it
i»
0-3876
07760
1-1625
1 ^i
1 :i
l:i
-0-34*
0 73
0-92
1-07
-1-88''
3-46
8-80
3*92
We take this opportunity of expressing our thanks to Mr. James
Prentice for much assistance given in the course of this work*
Chbmioal Laboratory,
Qlasqow and Wjest op Scotland
TbCHNIGAL GOLLEOEk
LXIX. — Oxonium Salts of Fluoran and its Derivatives.
By J» T. Hewitt and J. N. Tebvet.
Two years ago (Proc., 1900, 16, 3, and Zeit. phyMal, Chem., 1900, 34, 1),
one of the authors of the present communication proposed a theory
connecting the fluorescence of certain organic compounds with their
constitution. Briefly stated, this was as follows^ If a substance can
be converted into a tautomeric modification of greater free energy by
two equal displacements in opposite directions, the molecules will
vibrate between the two extreme positions, and radiant energy of a
particular wave-length taken up by the molecules will be emitted with
a different wave-length. It was shown that generally the fluorescent
dye-stuffs conformed to this type of constitution ; the molecules of
fluorescein in solution may be compared with a pendulum swinging
between two extreme positions :
O
0
XSO'OH
H H(Y'%''^Y^'^
0
0«H,^
MX)-
OH
The parent substance of fluorescein, namely, fluoran (B. Meyer,
Ser., 1891,24^1412; 1893, 25, 1385), or phenolphthalein anhydride
Y Y 2
Digitized by
Google
664 fiEWITT AND TERVBT: OXOKltJM SALTS OF
(Baeyer^ AnruUen, 1882, 212, 349), is colourless, and gives colourless
non-fluorescent solutions in neutral solvents.
If, however, it be dissolved in concentrated sulphuric acid, a yellow
solution exhibiting a brilliant green fluorescence is produced, and to
explain this phenomenon, the formation of salts containing quadrivalent
oxygen was assumed :
SO.H H SO.H SO.H
0
00-OH
A similar assumption made in the case of xanthhydrol and allied
compounds has been abundantly confirmed (Werner, Ber., 1901, S4^
3300; Hewitt, Ber., 1901, 34, 3819). The existence of salts derived
from fluoran has, however, been regarded with suspicion in R. Meyer's
Jahrhuch der Chemie (1901, 10, 438). This is, perhaps, the more
surprising, seeing that Nietzki and Schroter {Ber,, 1896, 28, 56)
obtained a yellow additive product from the lactonic diethyl ether of
fluorescein (that is, diethoxyfluoran), to which, however, they assigned
the constitution :
I
5«H^-00-0H
To settle the question, attempts have been made to isolate salts
of fluoran and its derivatives with strong mineral adds, and the
results obtained are here communicated.
Fluoran Nitrate, O^'H.^fi^B.l^Oy — Fluoran was warmed with excess
of a mixture of colourless nitric acid (sp. gr. 1 *36) and acetic anhydride.
The substance went into solution with a yellow colour ; on cooling,
small, pale yellow crystals were deposited. The excess of acid was de-
canted, and the crystals dried on porous earthenware over sulphuric acid.
For analysis, the substance was decomposed by water, the fluoran
collected, and dried until the weight was constant. The loss in weight
represents the nitric acid with which the fluoran had combined :
01292 lost 00226 HNO,. HNO, - 17-49.
0^Hij03,HN0, requires HNO, « 1 736 per cent.
Digitized by
Google
FLUORAN AND ITS DERIVATIVES. 666
Fluoran Sulphate, O^M^fi^tlEL^O^. — Fluoran dissolved when ground
up with a mixture of strong sulphuric and glacial acetic acids. Careful
addition of water produced a yellow precipitate, which was dried on
porous earthenware ;
orous eartnenware :
0-3677 gave 02060 BaSO^. SO^ = 231 7.
CgoH^OyHgSO^ requires 80^ = 24*
12 per cent.*
We were unahle to isolate a hydrochloride of fluoran.
Dimethylfluoran NiiraUy Q^^^fi^.Wl^O^, — Dimethylfluoran has
already been described by Drewsen (Annalen, 1882, 212, 340). We
prepared the material for our experiments by his method from ^resol,
phthalic anhydride, and sulphuric acid. The nitrate was obtained by
finely grinding the substance with a mixture of nitric acid and acetic
anhydride, and drying on porous earthenware. From a hot solution
in a mixture of acetic anhydride and nitric acid, the substance separates
in small, well-formed, yellow prisms, terminated obliquely and fre-
quently twinned :
0-2628, on washing, lost 00407 HKO3. HNOg- 15-46.
Cg^HigOjjHNOg requires HNOg- 1611 per cent.
DimethylfiuarcM Stdphate, C^II^qO^,1I^BO^. — ^This salt was prepared
analogously to the fluoran sulphate. lb forms a yellow, crystalline
powder consisting of minute, short prisms :
0-3463 gave 0-1679 BaSO^. SO4- 1983.
C^H^fi^yn^O^ requires 804 = 22-53 per cent.
Dimeihylfluonm DisulphtUe, C2jHjgOg,2H2S04, was obtained when
dimethylfluoran and acetic acid were ground to a paste, concentrated
sulphuric acid added, and the salt allowed to separate without the
addition of water :
0-3670 gave 0-3286 BaSO^. SO^ - 36-89.
0-3850 „ 0-5257 BaSO^. 804 = 34-86.
0,2Hi^Og,2H2S04 requires 804 = 36-64 per cent.
Attempts to prepare a hydrochloride failed, although dimethylfluoran
assumes a yellow colour when treated with concentrated hydrochloric
acid.
Mttoreacein ffydrochloride, OjqHjjOjjHOI. — Fluorescein was ground
to a paste with cold acetic anhydride and then exposed to a current of
dry hydrogen chloride until all excess of acetic anhydride was removed.
An orange powder was obtained. For analysis, the substance was
dissolved in hot sodium carbonate solution, the fluorescein precipitated
* Considering the unstable nature of the compounds produced, it Is not stu-prisiDg
that in certain cases the analyses have only given approximate results.
Digitized by
Google
666 MORBELL AND CROFTS : ACTION OF
by dilute nitric acid, and, after removal of the Auoreecein, the chlorine
estimated gravimetrically in the filtrate :
0-0900 gave 00326 AgCl. HCl - 9-21.
CjoHijOg^Cl requires HCl = 9*16 per cent
Fluorescein Sulphate, CgoHjjOjjHjSO^.— Fluorescein heated with
sulphuric add at 100^ furnishes a substance to which the formula
CaoH^jOgiSOg has been assigned by Baeyer {Annalen, 1876, 183, 27),
The fact that the substance so obtained is immediately decomposed by
water into fluorescein and sulphuric acid would suggest that the
substance was a sulphata Fluorescein, we find, gives a disulphate,
^20^i3%2HjSO4, when treated in the cold with a mixture of acetic
and strong sulphuric acids. On addition of water (in small quantities)
to the mixture of fluorescein and the acetic and sulphuric acids, a
yellow powder was obtained :
0-3858 gave 02044 BaSO^. SO^ = 22-05.
CgoH^aOjjHgSO^ requires 804 = 22*33 per cent.
Fluorescein Disulphate, C^B.^fi^y2B.^S0^y forms microscopic, light
yellow prisms :
0-2740 gave 0-2315 BaSO^. 804 = 34-81.
0-3444 „ 0-2882 Ba804. 80^ = 34-48.
C^QlIjfi^,2B.^S0^ requires 804 = 3636 per cent.
Fart of the expense incurred in carrying out these experiments '
defrayed by a grant received from the Government Grant Committee
of the Royal 8ociety.
East London Tecbnioal CoiiIOKJe.
LXX. — Action of Hydrogen Peroxide on Carbohydrates
in the Presence of Ferrous Sulphate. IIL
By BoBERT 8elbt Mobbell and James Murray Obofts.
The specific action of hydrogen peroxide in the presence of ferrous
sulphate was first demonstrated by Fenton in the oxidation of tartaric
acid to dihydroxymaleic acid (Trans., 1894, 65, 899), and later in the
oxidation of polyhydric alcohols to aldoses (Fenton and Jackson, Trans.,
1899,75, 1). Cross, Bevan, and Smith (Trans., 1898,73, 500) have investi-
gated the action of these two reagents on deictrose, and found that in
addition to tartropic acid a substance was formed which gave an xai*
Digitized by
Google
HYDROGEN PEROXIDE ON CARBOHYDRATES. III. 6Q7
mediate precipitate with phenylhydrazine acetate, but they were unable
to decide the nature of the substance. In communication with these
authorSy we have continued the investigation, and have shown that
dextrose, IsBVulose, arabinose, and rhamnose are transformed by this
peculiar action of hydrogen peroxide in the presence of ferrous sulphate
into osones, which were recognised by their power to react with sub-
stituted hydrazines at the ordinary temperature, yielding osazones
(Trans., 1899, 76, 787; 1900, T7, 1219).
Mannose and galactose on oxidation would be expected to yield
ogones. In the case of mannose, an osone is formed giving, with
phenylhydrazine, phenylglucosazone, but galactose behaves differently,
and we are \inable as yet to determine to what extent this sugar is
oxidised by hydrogen peroxide in the presence of ferrous sulphate.
Fischer (Ber.t 1894, 37, 2031) states that galactose and c^talose are
acted on by ferments with greater difficulty than dextrose and mannose,
and in explanation of this difference he refers to the arrangement of
the CH'OH groups, which in galactose and drtsAose is very unlike that
in dextrose and mannose. We consider that the internal compensating
positions of the hydrogen atoms and hydroxyl groups may account for
galactose behaving differently from other hexoses when oxidised by
hydrogen peroxide in the presence of ferrous sulphate. We have
attempted the preparation of glucosone from dextrose and Ispvulose, and
have obtained a white, amorphous solid which gave analytical numbers
agreeing with those required for the formula O^H^jOg or CgH^QOg ; it is
not easy to decide between the two formula from the percentages
of carbon and hydrogen. The white solid reacted immediately with
phenylhydrazine acetate at the ordinary temperature and gave a good
yield of phenylglucosazone. From its optical activity, we are led to
the conclusion that it was contaminated with a small quantity of the
parent carbohydrate, since the osone obtained from dextrose was
slightly dextrorotatory, whilst the osone from Isavulose had a levo-
rotatory power less than that of Isavulose. Glucosone, obtained from
phenylglucosazone by E. Fischer, is feebly IsBVorotatory {B^., 1889,
22, 89).
We have fermented solutions of glucosone from dextrose and
l^vulose and found that at the end of the fermentation tl^e solutioT^
was slightly Itevorotatory and still reacted immediately with phenyl-
hydrazine acetate.
In order to test further the truth of the statement that osones are
formed when carbohydrates are oxidised under the conditions given
above, we have tried the action of bromine on aqueous solutions of
glucosone obtained from dextrose and IsBVulose and have obtained good
yields of salts of an acid which is not gluconic acid, but trihydr-
ox^bnt^c acid ideptdofd with that obtained from ^rythritol by oxidation
Digitized by
Google
668 MORRELL AND CROFTS: ACTION OF
with nitric acid (Lamparter, Annalen, 1865, 134, 260), or from c^ery-
throse by the action of bromine (Huff, Ber., 1899, 32, 3678). The barium
and calcium salts of (2-erythronic acid, obtained from dextrose or
IsBvulose, have been reduced with hydriodic acid and phosphorus, and
gave 9»-butyric acid which was easily identified by means of its silver
salt. It has not been shown, so far as we know, that trihydroxy butyric
acid is an oxidation product of dextrose, although Iwig and Hecht
(Ber.f 1886, 19, 169) have obtained a trihydroxybutyric add in small
quantities from mannitol by oxidation with potassium permanganate.
Trihydroxybutyric acid has also been obtained from Isevulose by the
action of mercuric oxide and baryta, or by the action of bromine
(Bbrnstein and Herzfeld, Ber„ 1886, 18, 3354; Herzfeld, Annalen,
1888, 244, 291 ; Euff, Ber., 1899, 32, 3680). RufiE states that the
yield of trihydroxybutyric acid obtained from Isvulose is exceedingly
small, but he was able to prepare a brucine salt of this acid which was
identical with the brucine salt of (2-erythronic acid. We have found
that the amount of calcium trihydroxybutyrate obtained from lievulose
by the direct action of bromine was not more than 1 per cent, of the
weight of the Isevuloso taken. The transformation of dextrose and
Isevulose into glucosone, trihydroxybutyric acid (c^erythronio acid), and
into n-butyric acid can be expressed in the following manner :
(jJHO CHj-OH CHO
3H-0H CO CO
CH,-
CH-OH (JUO Co
^H-OH 9H-0H (j)H-OH CO-OH CO-OH
(j3H-0H ^^ C^H-OH "^ 9H-0H "^ CH-OH "^ CH^
CH-OH CH-OH CH-OH ^H-OH . 6h,
CHg-OH CHg-OH CHj-OH CH^-OH CHj
Dextrose. Lsovnlose. Qlucosone. c^-Exythronic n-Batyric
add. acid.
We are engaged on the further investigation of the properties of
osones obtained by the action of hydrogen peroxide on carbohydrates
in the presence of ferrous salts, and hope to communicate at an early
date the results of the action of bromine on a solution of rhamnosone
which would point to the formation of a methylglyceric acid — a result
which may have an important bearing on the constitution of rhamnose.
EXPEBIMENTAL.
OxidaUon qf MannoM,
Forty grams of seminose syrup were transformed into mannose
hydrazone by treatment with phenylhydrazine acetate. The hydr-
azone was decomposed by benzaldehyde and the purified mannose ob-
tained as a syrup, which was found to contain about 8 grains of the
Digitized by
Google
HYDROGEN PEROXIDE ON CARBOHYDRATES. III. 669
sugar. One hundred and fifty c.c. of a 5 per cent, solution of mannose
were treated with hydrogen peroxide of 20-yolume strength in the
presence of 1 gram of ferrous sulphate. The amount of hydrogen per-
oxide used was such as to yield 0-66 gram of oxygen, which was the
calculated quantity required for the transformation of 7*5 grams of
mannose into its osone. The oxidiser was added slowly in tenths of
the required quantity, as in the oxidation of dextrose and Isavulose.
An aqueous solution of the osone was obtained in the manner
described in a former paper (Trans., 1900, 77, 1219). On treatment
of the solution of the osone with phenylhydrazine at the ordinary tem-
perature, a yellow precipitate formed immediately. After being
allowed to stand for a few hours, the precipitate was filtered, washed
with alcohol and ether, and dried in a vacuum. The yield of glucos-
azone was 5 grams. On recrystallisation from alcohol, the glucosazone
melted at 204° with decomposition and gave the following numbers on
analysis :
0-1670 gave 0-3789 COjj and 01017 H^O. C = 601 ; H = 6-6.
00678 „ 9-1 c.c. moist nitrogen at 12° and 750 mm. N = 1568.
CigHgjO^N^ requires 0 = 60-3; H = 6-2; N = 15-65 percent.
Twenty grams of seminose, which had not been purified by the
method given above, yielded, after oxidation with hydrogen peroxide
in the presence of ferrous sulphate, 4 grams of glucosazone, which»
on recrystallisation, melted at 199° with decomposition and oontained
15*98 per cent, of nitrogen.
Preparation of Glucoaone from Dextrose or LcbvuIom,
An aqueous solution of glucosone, prepared from Isevulose (Trans.^
1900, 77, 1219), was concentrated to a syrup in a vacuum at 50°. The
syrup was poured into warm absolute alcohol and the alcohol solution
filtered and concentrated in a vacuum on the water-bath and finally
poured into dry ether. A white, amorphous solid separated, which
was dissolved in hot absolute alcohol and the solution, after filtra-
tion and concentration, poured into dry ether. The yield of the
white, amorphous solid from 60 grams of IsBvulose amounted to
3 grams. An analysis of the substance dried in a vacuum gave the
following numbers :
0-1782 gave 0-2608 CO, and 0-0949 HjO. 0 = 39-92 ; H-5-87.
CgHjoOfl requires 0 = 40-40; H = 5-61 per cent.
The glucosone, which was practically free from ash, reduced Fehling's
solution without being warmed.
Qlncosone from dextrose was prepared in the same way, except that
Digitized by
Google
670 HORRELL AND CROFTS; ACTION OF
hydrogen peroxide of 10- volume strength was used instead of 20«volnme
strength. An analysis of the substance dried in a vacuum showed it
to be less pure than the glucosone obtained from Isevulose :
0-2002 gave 0-2906 COj and 0-1162 HgO, C - 39*57 ; H = 6-4.
It was not free from impurity, since 0*2 gram of substance, after igni*
tion, was found to contain 0*9 mg. of ash. The glucosone reduced Fehling's
solution at the ordinary temperature. Both the samples of glucosone
reacted immediately with a cold solution of phenylhydrazine acetate.
One gram of glucosone from dextrose dissolved in 10 o.c. of water was
treated with 2 grams of phenylhydrazine dissolved in 2 cc. of 50 per
cent, acetic acid and diluted with 4 cc. of water. A precipitate was
formed immediately. After standing for some hours at the ordinary
temperature, the glucoeazone was filtered o£P, washed with water, and
dried in a vacuum. The weight of the osazone amounted to 0*5 gram.
Half a gram of glucosone from Isavulose dissolved in 70 cc. of water
and treated with 1 gram of phenylhydrazine in 60 per cent, acetic
acid gave 0*23 gram of gluoosazone. The liquid was kept quite cold,
and under these conditions both dextrose and Isevulose do not give a
precipitate with phenylhydrazine acetate unless the solutions are
allowed to stand for several days. On recrystallisation from alcohol,
the osazone melted at 203^ with decomposition, and its identity was
further established by a nitrogen determination :
0*0663 gave 84 cc moist nitrogen at 12° and 760 mm.. N= 15-2.
OjgHjjjO^N^ requires N = 16*66 per cent.
The determination of the optical activity of the glucosone from
dextrose and Isevulose gave results which were unsatisfactory, and we
concluded that small quantities of the parent sugar were present as
impurity, especially in the case of the glucosone from dextrose, for the
preparation of which 10- volume hydrogen peroxide was used. The
glucosone in each case was dissolved in water :
0'604 gram of glucosone from l^vulose had [a]]> —64°;
0*6247 „ „ dextrose „ [ajn +13-6°; whereas
0-6164 „ Isevulose dissolved in water had [a]i> -89°.
E. Fischer states that glucosone prepared from phenylglucosazone is
feebly Isevorotatory {Ber., 1889, 22, 89).
We have fermented solutions of glucosone from dextrose ai^d Isevulose
with yeast in order to remove these carbohydrates. A solution of
glucosone from dextrose had a decided dextrorotatory power, but after
two days' fermentation at 30° became feebly Isevorotatory. The fer-
mentation was continued for another day, but there was no change in
thp rotfttory |)ower. Afte^ the |:emovftl of t)ie inorganip impurities bv
Digitized by
Google
HYDBOQBN PEBOXIDV ON CABB0HTDBATB8. III. 671
concentration of the aqueous solution of the glucosone in a vacuum on
the water-bath and treatment with absolute methyl alcohol, a syrup
was obtained which did not crystallise, but reacted immediately with
phenylhydrazine acetate at the ordinary temperature.
A solution of glucosone prepared from IsBVulose was treated in
exactly the same manner. The IsBvorotatory power became less on
fermentation and after two days the fermentation stopped. The solu^
tion was found to be slightly Isdvorotatory and reacted immediately
with a cold solution of phenylhydrazine acetate.
Oxidation of an Aqueous Solution oj Gluoosone prepared/rom Dextrose.
An aqueous solution of glucosone prepared from 30 grams of dextrose
was heated with 12*5 grams of bromine at 40^ for 12 hours (the
volume of the solution was 350 cc). The excess of bromine was
removed by a current of air and the yellow liquid was treated with an
excess of lead carbonate. After standing for 24 hours, the liquid was
filtered from lead carbonate and lead bromide, and sulphuretted hydro-
gen was passed in. The filtrate from the lead sulphide was concentrated
to a small bulk in a vacuum at 60^ on the water-bath. It was found
to be impossible to remove the last traces of hydrobromic acid by silver
carbonate as the solution contained a reducing substance. The residual
liquid was boiled with calcium carbonate until it was neutral, decolor-
ised with animal charcoal, filtered, evaporated in a vacuum on the
water-bath to a small bulk, and poured into absolute alcohol. The
calcium salt was obtained as a granular precipitate which was washed
with absolute alcohol and ether and the ether removed in a vacuum.
The yield of the salt varied between 25 and 30 per cent, of the weight
of dextrose taken. It was purified by treatment with the calculated
quantity of oxalic acid, and after boiling with calcium carbonate and
decolorising with animal charcoal was reprecipitated by absolute;
alcohol. The yield of the purified salt was never less than 12 — 15 per
cent, of the weight of the dextrose used. Analysis of the calcium salt
gave the following numbers :
0-1995, air dried, gave 0-2072 COg and 00872 H^O. C = 28-38 ; H = 4-85.
0-1958 dried at 110° ,,0-223 OO3 „ 00735 H2O. 0 = 31-0 ; H«4'2.
0-4170 „ 110° gave 01785 CaSO^. Ca = 12-6.
0-315 „ 110—130° „ 01405 CaSO^. Ca = 13-1.
0-4253 of the air-dried salt „ 0*1 602 OaSO^. Ca = 11 08.
(C4H705)2Ca,2H30 requires Oa = l 1-56 j 0 = 27-74; H = 5-2 per cent.
{GJ1^0^)^CsL „ Ca = 1 2-9 ; C = 30-96 ; H = 4-5
These numbers show that the salt is most probably calciun^ tri«
bydrox^but^rrate. -^
Digitized by
Google
672 MORRELL AND CROFTS: ACTION OF
Lead Err/thronate, {C^'H.fi^)Fb. — The lead salt was prepared by add-
ing normal lead acetate to a solution of the purified calcium or barium
salt. The white precipitate so obtained was well washed with hot
water, and after being dried at 130° was analysed :
0-1628 gave 01448 PbSO^. Pb = 60-8.
0-1550 „ 0-1660 PbSO^. Pb = 60-8.
{C^Bfi^)Fb requires Pb«=60-7 per cent.
This salt was obtained by Lamparter from erythronic acid {loe. eU,),
and is insoluble in dilute acetic acid. It has been considered to be a
characteristic salt of w-trihydroxy butyric acid (Fischer, Ber., 1889,
22, 110; Fenton, Trans., 1899,76, 7). The lead salt obtained by us
was insoluble in dilute acetic acid.
Barittm EryihronaU, B9^{Q^0^^filELfi. — For the analysis, the
barium salt obtained by boiling the solution containing erythronic acid
with barium carbonate and precipitating with alcohol must be freed
from iron compounds, which are present in small quantity. This was
done by dissolving the barium salt in a small volume of water, adding
alcohol in slight excess, filtering, and washing the undissolved residue,
which was brown in colour, with a little cold water. The filtrate
contained the barium erythronate, which was treated with dilute sul-
phuric acid, the liquid boiled with barium carbonate, and the pure
barium salt reprecipitated by alcohol. On analysis :
0-1803 lost, at 130°, 0-01450 H^O. HjO* 804.
01653, dried at 130°, gave 0-0938 BaSO^. Ba = 33-31.
0-1645, „ „ 0-0943 BaSO^. Ba= 33-76.
Ba(04H^05)2,2HjO requires H2O^8-03 per cent.
Ba(C,H70g), „ Ba = 33-66
Reduction of CcUcittm Erythronate to Butyric Add. — Fifteen grams
of calcium erythronate, obtained from dextrose, were heated for
8 hours with 130 c.c. of hydriodic acid (b. p. 127°) and 5 grams of
amorphous phosphorus in a flask fitted with a reflux condenser on a
sand-bath. The brown liquid was diluted with an equal volume of
water and shaken six times with ether. The ether was distilled oS
and a brown oil was left which possessed a strong odour of butyric
acid. To complete the reduction, the oil was heated with dilute sul-
phuric acid and zinc for an hour in a flask fitted with a reflux con-
denser. After being allowed to stand for several hours, the excess of
zinc was filtered off and a little more sulphuric acid was added to the
filtrate. A current of steam was passed into the liquid and the distil-
late, which was strongly acid, was neutralised with calcium carbonate.
After filtration and concentration on the water-bath, 1-2 grams of a
salt crystallised out. This salt had the characteristic properties of
Digitized by
Google
&YD&OGEK PEBOXIDE ON CARBOHTDBATES. III. 673
calcium bntyrate, being less soluble in hot water than in cold, and
having the peculiar odour of the acid. To confirm the formation of
butyric acid, the silver salt was prepared from the calcium salt, dried
in a vacuum, and analysed :
0-2345 gave 0*1345 Ag. Ag = 55*4.
0-1708 „ 0-0948 Ag. Ag = 55-5.
C^H^OjAg requires Ags55-4 per cent.
The total yield of calcium butyrate from 15 grams of calcium
erythronate was about 15 per cent, of the theoretical amount.
Oxidation qf cm Aqueous Solution of Glueosone prepared from LcBmdoee.
The action of bromine on an aqueous solution of glucosone prepared
from IiBvulose was exactly the same as in the case of glucosone obtained
from dextrose, and it is unnecessary to repeat the details of the separ-
ation and purification of the calcium or barium erythronate. From 30
grams of Isvulose in each case, the yield of calcium salt amounted to
15 grams or, 50 per cent, of the theoretical, and of barium salt to 11
grams, or 30 per cent, of the theoretical.
The calcium salt, dried at 110% was analysed :
0-2050 gave 0-2360 COj and 0-0845 Bfi. C = 31-3 ; H = 4-6.
0-2875 „ 0-1285 CaSO^. Ca= 13-15.
0-2875 „ 0-1280 CaSO^. Ca = 13-1.
(04H705)2Ca requires C = 31 0 j H « 4-5 ; Ca = 12-9 per cent.
The bctrium salt, dried at 1 30% was analysed, with the following
results:
0-2255 lost 0-0190 H,0. Kfi^S^.
0-2065 gave 0-1196 BaSO^. Ba = 33-8.
(C4H705)8Ba,2H20 requires HjO = 8-0 per cent.
(04H^05),Ba „ Ba= 33-66 „
The lead salt of erythronic acid was prepared from the barium salt.
The yield amounted to 1 gram from 3 grams of barium salt, that is,
about 25 per cent, of the calculated quantity :
01648, dried at 130°, gave 0-1465 PbSO^. Pb = 60-5.
{O^Kfi^)Tb requires Pb = 60-77 per cent.
It is not necessary to dry the salt at so high a temperature as 160°
(Lamparter, loe, cit.).
Strychnine Salt, — When the calcium salt was treated with the calcu-
lated quantity of oxalic acid and any excess of oxalic acid removed by
lime water, a solution of the free acid was obtained, which, on con-
centration, yielded a syrup which did not crystallise. From the
Digitized by
Google
674 ACTION OP HTDROOEN PEROXIDB ON CARBOHYDRATES. III.
solution of the acid, by boiling with strychnine, filtering and concentralr-
ing, and treating with absolute alcohol, the strychnine salt was pre-
cipitated, which, after recrystallising twice from a small quantitj of
water, gave, on analysis, the following numbers :
0-1777,driedatl30°,gave0-4137CO3and0-0986H5O. 0-«63-6; H-6-1.
OaSl 8, dried in a desiccator, gave 0*4083 CO, and O'Ol 86 H^O. 0 - 61-24;
H-6-57.
CjiH^OgN^C^HgOfi requires 0 = 638 ; H«615 per cent.
C^^fi^^C^nfi,;Rfi » 0*61-47; H-6-55 ,,
The strychnine salt crystallised in needles from water, and contained
one mol. of water of crystallisation which was not expelled under 130^.
Reduction qf Erythronic Add, obtained from Lasvulow^ to Butyric
Acid. — The details of this reduction are identical with those described
under the reduction of calcium erythronate from dextrose. The yield
of calcium or barium butyrate was about the same as in the reduction
of the calcium salt from dextrose* The calcium or barium butyrate
was transformed into the silver salt» which crystallised from water in
white needles.
The aUver salt was analysed :
01814 gave 0-1003 Ag. Ag - 56'3.
O^H^O^g requires Ag«=55'4 per cent.
Oxidation qf Erythritol by Nitric Acid,
Lamparter prepared erythronic acid from erythritol by the action of
strong nitric acid on a hot concentrated aqueous solution of erythritol
{loo. cit). Przibytk (J. Ruas. Phy$. Chem. Soc., 1881, 12, 208) statea
that dilute nitric acid transforms erythritol into oxalic acid and
mesotartaric acid.
Ten grams of erythritol were oxidised by 25 c.c. of nitric acid of
sp. gr. 1 *2 at 40^ for 28 hours, the liquid diluted with water, and the
nitric acid removed by distillation in a vacuum at 50^. A syrup was
obtained which contained oxalic and erythronic acid. The calcium
erythronate was prepared in the usual way and precipitated by means
of alcohol. The salt was purified by treatment with oxalic acid and
calcium carbonate and was precipitated from its aqueous solution with
alcohol, and washed with ether. The yield of the purified salt was
7 '5 grams :
0*2432, dried at 110°, gave 01085 OaSO^. Ca- 13-1.
{G^13LfO^)fi8L requires Oa* 12*9 per cent.
This salt is evidently identical with the calcium erythronate already
Digitized by VjOOQIC
4-lSOPBOPYLDlHYDBORBS80RCII^. 676
desoribed, and from it a lead salt was obtained insoluble in dilute
acetic acid.
The authors desire to express their thanks to the Government Qrant
Committee of the Boyal Society for funds which have enabled them to
carry out this investigation.
GONVILLB kJXD CaIUS COLLEGE LABOR ATORT»
Cambridqe.
LXXI, — Preparation and Properties oj A-iaoPropyl-
dihydroresorcin.
By Arthur William Crosslet.
Amonq the many interesting changes which substituted dihydro-
resorcins undergo is one which they suffer on treatment with barium
hydroxide. For example, Yorlander has shown {ArmcUen, 1897, 294^
317 ; 1899, 308, 188) that when 4-phenyldihydroresorcin is boiled
with barium hydroxide, hydrolysis takes place and the ring is broken
with formation of )9-phenyl-8-ketohexoic acid (j3-phenyl-y-acetylbutyric
acid),
,CH, OOv /CH.-CO
OeH,-CH< -\CH -^ C^Hj-OH/ \CH, .
XIH,-0(OH)'^-. N)Hj-00,H
Exceptions are found to this reaction in the cases of 4 : ^-dimethyl-
dihydroresorcin and 3:4: 4-trimethyldihydroresorcin (compare Ber,f
1897,30, 1801, and Trans., 1901, 70, 139),
(CH3),0<g^— g>>OH and {OR,)Mi^^o^OR ,
neither of which substances behaves in the above-mentioned manner.
The explanation is evidently to be sought in a difference of constitution.
These two substituted dihydroresorcins differ in one point, and one
point only, from those which have been found to hydrolyse on treat-
ment with barium hydroxide : they contain two alkyl groups attached
to the same carbon atom (marked with a + ), whereas all those which
have been found to undergo hydrolysis contain only one alkyl group in
this position.
In order to obtain further evidence on this point, 4H«opropyldihy*
droresorcin has been prepared, and its behaviour towards barium
hydroxide investigated.
Digitized by
Google
676 CROSSLE!Y: PBEPABATtON AND PB0PI!BT1ES OF
The starting point of the preparation was Mobutylideneaoetone,
(0H3)20H'CH:CH*C0'CH8, prepared according to the directions of
Franke and Kohn {Afonat$h,, 1899, 20, 876), and as this substance
condenses with ethjl sodiomalonate to give ethyl ^hsopropyldihydro'
resarcylcaey (OB^)jaK^OIL<^^^y^^>OIL, a characteristic re-
action of aj3-unsaturated ketone, it affords a further proof, if such
were needed, of the formula assigned to this ketone by Franke and
Kohn.
When the above ethyl salt is hydrolysed with either sodium car-
bonate or potassium hydroxide, A-isopropyldihydroresardn (formula I
below) is obtained, which crystallines with IH^O. Its constitution is
proved by its method of formation, and also by the fact that on
oxidation with sodium hypobromite it gives rise to /J-Mopropylglutaric
acid (compare Trans., 1899, 76, 772 ; 1901, 79, 139) :
(CH.),CH.CH<^^^~^OH -> (OH3),OH.OH<^52;00,H
XIH,-C(OH)< ^K)H,-CO,H
When boiled with barium hydroxide solution, wopropyldihydro-
resorcin is hydrolysed, as expected, with production of j8-isopropyW-
ketohexaic acid :
yOH, OOv yOH,-00
(CH3)jCH-CH< .\0H -^ (CHAOH-OH< \0H,.
XIH,-0(OH)^ XJH^-COjH
Apparently, then^ it is the presence of two alkyl groups attached to
one and the same carbon atom which differentiates between substituted
dihydroresorcins hydrolysed by barium hydroxide and those which
are not.
The ketonic nature of i^opropylketohexoic add was proved by
the preparation from it of a semicarbazide and an oxime ; and its
constitution follows from the oxidation with dilute nitric acid, when it
yields pimelic (t^opropylsuccinic) acid,
whereas t^opropyldihydroresorcin, under similar conditions, gives ^S-tfO
propylglutaric acid.
The following comparison of the properties of uopropyldihydro*
resorcin and Mopropylketohexoic acid leaves no doubt as to the non*
identity of the two substances :
Digitized by
Google
4-ISOPBOPTLDIHTDBORESORCIN.
677
4'-MoPropyldihydro
resoTcin
/3-woPropyl-8-keto-
hexoic acid
M. p. or
b. p.
} ''° {
y87' under
/ 16 mm.
FejCU.
Violet
colour.
No
colour.
Water.
Cgll j^Og, HgO,
ni. p. 67-6'.
Appreciably
soluble.
Oxidation
product.
MoPropylglutaric
acid
MoPropylsuccinic
acid
M. p. of
oxime.
146**
98—94°
In previous commuDications (Trans., 1901, 79, 138 ; Proc, 1901,
17, 172), substituted dihydroresorcins have been called substituted
dihjdroresorcinols or diketoc^c/ohezanes ; but the author no longer re-
tains these latter names, for although the substances behave apparently
as diketones towards hydroxylamine, there can be no doubt that their
most usual form is the ketoenolic constitution, represented by the
following formula :
8 3
(OH.),OH.CH<«g«— g^H.
The positions of the various substituting groups are indicated by
numbering the carbon atoms as above.
For these reasons, and also for brevity's sake, the name dihydro-
resorcin will in future be adopted to designate this class of substances.
EXPEBIMENT1.L.
JSthyl 4-iso PrapyldihydroresorcylcUe-S,
Twenty-three grams of sodium were dissolved in 275 c.c. of absolute
alcohol, 170 grams of ethyl malonate added, and, after cooling, 112
grams of Mobutylideneacetone (Franke an[d Kohn, Monatsh.y 1899, 20,
876). The mixture became reddish-pink and much heat was evolved.
On shaking, the whole set to a faintly-yellow, semi-solid mass, which
was heated on the water-bath for 4 hours to complete the reaction ;
water was then added, the alcohol evaporated, and the alkaline liquid
extracted with ether, which treatment removes some unaltered
material. After acidification with dilute sulphuric add, the whole
was again extracted with ether, the ethereal solution washed with
water, dried over calcium chloride, and the ether evaporated, when
230 grams of a thick, yellow liquid were obtained, which, on standing,
became semi-solid. The mass was spread on porous earthenware, thus
giving 100 grams of a white solid, and on extracting the porous plate
with 6ther,l98 grams of a thick, yellow liquid were obtained, which
VOL. LXXXI. Z Z
Digitized by
Google
678 CR0S8LET : PREPARATION AND PROPERTIES OF
showed no signs of crystallisation after many months' standing, and is
at present being more fully investigated.
The white solid was purified by crystallisation from a mixture of
benzene and light petroleum, and analysed :
01107 gave 0-2572 COj and 0-0804 H^O, C = 63-37 ; H = 807.
^12^18^4 requires 0 = 63-71 ; H = 7*96 per cent.
Ethyl Mopropyldihydroresorcylate dissolves slightly in water and
readily in the ordinary organic solvents. It crystallises from the
above-mentioned mixture in stellar aggregates of small needles melting
at 100*5 — 101^. Its aqueous solution gives a deep violet colour with
ferric chloride solution.
i^isoPropyldthydroreaarkn, (CH3)2CH-CH<^§a;;rj^>CH.
Ethyl isopropyldihydroresorcylate was hydrolysed by boiling for
2 hours with an equal weight of pure caustic potash dissolved in
alcohol. (Hydrolysis may also be effected by boiling for about 15
hours with sodium carbonate solution.) Water was then added, the
alcohol evaporated, and, after acidification with sulphuric acid, the
whole extracted with ether and the ether evaporated. As this sub-
stituted dihydroresorcin crystallises with IH^O, it is of advantage not
to dry the ethereal solution before evaporation. The solid residue,
obtained in nearly theoretical yield, was purified by crystallisation
from dilute methyl alcohol and analysed :
0-1446 gave 03315 CO, and 01214 H^O. 0 = 6252 ; H = 9-33.
Oj>Hj40j,HjO requires 0 == 62-79 -, H = 930 per cent.
t^oPropyldihydroresorcin is sparingly soluble in wator or light
petroleum, but readily so in the ordinary organic solvents; its
aqueous solution is intensely acid towards litmus paper and gives a
deep purple colour with ferric chloride. It crystallises from dilute
methyl alcohol, with IH^O, in stout, flattened needles melting at 67*5°.
When dried in a vacuum over sulphuric acid :
0-8860 lost 0-0940 H^O. Up = 10-61.
OjjHi^OjjHjO requires H^O- 10-46 per cent.
The dried substance, when heated in a capillary tube, melts at 82°,
and at about 100° a red film is formed above the substance, a phen-
omenon noticed in the case of 4 : 4-dimethyldihydroresorcin (Trans.,
1899, 75, .773). On analysis of the dried substance, the following
numbers were obtained :
0-1197 gave 0-3073 00, and 00980 H^O. 0 = 7001 ; H = 9-10,
CjHi^Oj requires 0- 7013 ; H = 9-09 per cent. -
Digitized by
Google
4-ISOPBOPYLDIHYDBORESOBCIN. 679
The Hlver salt, OgK^fi^g^ prepared in the usual manner^ is a
white, insoluble precipitate :
0-1761 gave 0*0727 Ag. Ag- 41-28.
CgH^gOgAg requires Ag — 41'38 per cent.
The diQxime was obtained by adding the calculated quantities of
hydroxylamine hydrochloride and sodium hydroxide dissolved in the
smallest possible quantity of water to an alcoholic solution of the
ketone. On standing, the solution became violet, and gradually
deposited crystals, which were separated, purified by crystallisation
from dilute methyl alcohol, and the nitrogen estimated :
0-1186 gave 15*6 c.c. moist nitrogen at 16° and 764 mm. N» 15-41,
C^H^^OjNj requires N« 16*22 per cent.
The dioxime is insoluble in benzene or chloroform, but readily soluble
in methyl or ethyl alcohol on warming, and crystallises from dilute
methyl alcohol in clusters of stumpy needles. Its melting point is
indefinite. On heating in a capillary tube, it runs together at
145°, forming a nearly clear jelly, which sticks to the side of the
tube. On more strongly heating, it becomes cloudy, and at 165°
decomposes and gives off gas, *
The eOi^l ether, C3Hy-CH<^]^.^,Q^^^9>CH, prepared by heat*
ing the dry silver salt with ethyl iodide in dry ethereal solution,
is a clear, faintly yellow, oily liquid boiling at 284° at 762 mm. :
0-1204 gave 0-3190 COg and 0-1086 HgO. C = 72-26 ; H = 1002.
CjjHigOa requires C = 7252 ; H - 9*89 per cent.
When hydi'olysed with alcoholic potassium hydroxide, it is quan^-
titatively reconverted into i^opropyldihydroresorcin.
l''BrornoA-i8oprapyldihydrare8ordn,{CK^)20K*OK<^
was prepared by adding a solution of bromine in chloroform to one
of dried t^opropyldihydroresorcin (m. p. 82°) in chloroform until the
colour of the former just remained permanent, when hydrogen bromide
was evolved and a heavy oil separated, which soon solidified. This
was filtered off, purified by crystallisation from dilute alcohol, and
the bromine determined :
0-2076 gave 0-1692 AgBr. Br = 34-68.
CgHigOjBr requires Br = 34-33 per cent.
Bromoisopropyldihydroresorcin is insoluble in water, readily soluble
in alcohol, acetone, or ethyl acetate on warming, and crystallises
in nacreous scales. Its melting point depends on the rate at which
it is heated ; when heated in the ordinary way, it melts sharply at
169° to a deep red liquid, which at once decomposes and gives off
DigitizecftyC%)OgIe
680 crosslet: preparation and properties of
gas. Other determinations in which the substance was heated more
and more slowly gave 166^ 162°, and finally as low as 152^ In each
case, the substance melted sharply at the given temperature with de-
composition and evolution of gas.
Oxidation of irlBoPrapyldihydraresarcin,
I. With Potassium HypohromiU. — Thirty-four grams of bromine
were poured into 400 o.c. of water cooled to 0° and a strong solution
of sodium hydroxide slowly added until the colour of the bromine had
disappeared. A solution of 7 grams of t^opropyldihydroresorcin in
sodium hydroxide (10 grams NaOH in 50 c.c. of water) was then
poured in and the whole allowed to stand 4 hours. The solution,
after separation from bromoform and carbon tetrabromide, was acidi-
fied with hydrochloric acid, evaporated to about one-fourth of its
original bulk, and repeatedly extracted with ether. The ethereal solu-
tion, after drying over calcium chloride and evaporation of the ethw,
yielded 5 grams of a white solid, which proved to be /3-t«opropylglu-
taric acid (compare Howies, Thorpe, and Udall, Trans., 1900, 77, 942).
Thus, it crystallised from water in stout needles melting at 100 — 100*5%
and gave a crystalline anilic acid melting at 121°. On analysis of the
acid, the following numbers were obtained :
01422 gave 0-2861 CO, and 0-1028 H,0, C = 54-87 ; H = 803.
CgHj^O^ requires C = 65-17 ; H = 804 per cent.
II. With Nitric Acid. — Two grams of Mopropyldihydroresorcin were
heated to boiling with 20 c.c. of dilute nitric acid (1 : 1), when after a
few moments oxidation took place vigorously. The whole was evap-
orated on the water-bath, with repeated addition of water, when
1-7 grams of a white solid were obtained melting at 100° and giving
an anilic acid melting at 121°, thus proving it to be )9-i9opropylglutario
acid.
p'iBoFropylB'ketohexoic Acid, CH3'CO-CH2-CH(C8Hy)'CH3-CO,H.
i^oFropyldihydroresorcin (1 part) was heated with barium hydr^
oxide (4 parts) and water (20 parts) for 36 hours in a flask attached
to a reflex condenser. The whole was then evaporated to about one-
third of its original volume, acidified with hydrochloric acid, and ex-
tracted with ether. The ethereal solution, after drying over calcium
chloride, was evaporated and the residual liquid purified by repeated
distillation in a vacuum, and analysed :
0-1206 gave 02776 COj and 0-1014 HjO. 0 = 62-77 ; H«9-34.
CgHj^jOg requires C -62-79 ; H = 9-30 per cent.
tffoPropylketohexoic acid is a perfectly colourless, thick liquid witl\
Digitized by VjOOQIC
4-ISOPROPTLDIHTDRORE80RCIN. 681
a sharp, but not unpleasaot odour. It boils at 187^ under 15 mm.
pressure, and requires to be distilled as rapidly as possible, for on slow
distillation, especially in air, it decomposes. The silver salt,
C^HjgOjAg, prepared in the usual way, is a white, insoluble pre-
cipitate :
0-2184 gave 0*0852 Ag. Ag»3901.
CgHjjOjAg requires Ag= 38'71 per cent.
The aemicarbazidef O^QHigOgNg, was obtained by adding to an
alcoholic solution of the ketonic acid the calculated quantities of semi-
carbazide hydrochloride and sodium acetate dissolved in the smallest
possible quantities of water and allowing the alcohol to evaporate
slowly. The solid which separated crys^llised from dilute alcohol in
small, transparent plates melting at 144° with slow evolution of gas :
0*1146 gave 18*4 c.c. moist nitrogen at 18° and 756 mm. N= 18*44.
CiQHjgOjNg requires N = 18*34 per cent.
Oxime. — Five grams of hydroxylamine hydrochloride and 3 grams
of sodium hydroxide dissolved in the smallest possible amount of water
were added to an alcoholic solution of 5 grams of the ketonic acid and
the solution heated to boiling for 4 hours. The pasty mass left on
evaporating the solvent was warmed with absolute alcohol, filtered
from sodium chloride, and the alcohol again evaporated. On stirring
the residue with a few drops of very dilute acetic acid, it gradually
solidified. It was spread on a porous plate, purified by crystallisation
from ethyl acetate, and the nitrogen determined :
0*2064 gave 13*5 c.c. moist nitrogen at 19*5° and 755 mm. N :- 7*45.
CgHjyOjN requires N = 7*49 per cent.
The ozime crystallises slowly in stellar aggregates of compact
needles melting at 93 — 94°. It is practically insoluble in light
petroleum, but ready soluble in the ordinary organic solvents, especially
on warming.
OxidcUion qf AcetyliBopropylbtUyrtc Acid, — 8*5 grams f)f the ketonic
acid were heated with 90 c.c. of nitric acid (sp. gr. 1*15) for half an
hour in a flask attached to a reflux condenser. The oxidation is not a
very violent one. After adding water and evaporating off the nitric
acid, 6 grams (calc, 7*3 grams) of a white solid were obtained, which
crystallised from water in stout, transparent needles melting at
115 — 116°. This is the melting point of pimelio (i^opropylsuccinic)
acid (compare Trans., 1898, 73, 22). To prove further the identity of
this acid, a portion was converted into the silver salt, and analysed :
0*2598 gave 0*1498 Ag. Ag« 57*66.
CyHi^O^Ag, requires Ag — 57*75 per cent.
Digitized by
Google
682 INNHS: tHB INFLUENCE OF TEMPBRATdRE ON
The acid'also gave an anhydride boiling at 255° at ordinary atano^
spheric pressure, and from this an anilic add crystallising from dilate
alcohol in shining scales melting at 135°.
As iBopropf/lsuccinanilic acid does not appear to have been previously
described, a specimen was made for comparison. The pimelic acid
employed was obtained by the fusion of camphoric acid with pot-
assium hydroxide (Tranp^, 1898, 73, 22). It was first converted
into the anhydride boiliug at 255° under ordinary atmospheric pres-
sure, and this, on treatment with aniline in benzene solution, gave
the anilic acid, crystallising from dilute alcohol in beautiful, nacreoos
scales melting at 135°. It is fairly soluble in benzene and chloro-
form on warming, and readily so in ether, ethyl acetate, acetone, or
alcohol in the cold :
0-3006 gave 16-4 c.c. moist nitrogen at 20° and 766 mm. N =• 5-91.
OigHj^OjN require^ N»=5'95 per cent.
The author's thanks are due to the Research Fund Committee of
the Chemical Society for a grant defraying the cost of the materials
used in this investigation. ^
Chemical Laboratort,
St4 Thomas's Hospital.
LXXII. — The Influence of Temperature on Association
in Benzene Solution, and the Value of the Molecular
Rise of Boiling Point for Benzene at Different
Temperatures.
By William Ross Innes, M.Sc. (Vict.), Ph.D. (Heidelberg).
Substances containing hydroxyl groups give, as is well known, ab-
normal molecular weights in hydrocarbon solutions by both the
cryoscopic and ebullioscopic methods. A large number of hydroxy]
compounds have been investigated in benzene by Beckmann, Auwers,
Paternb, and others, and it has been shown that they may be divided
into two classes according to their behaviour with increasing con-
centration : carboxylic acids and oximes have in general the normal
molecular weight in dilute solution ; as the concentration is increased,
the molecular weight increases, at first rapidly, then more slowly, until
it reaches double the normal value j further increase of concentration
affects the value but little. Alcohols and phenols also give the normal
Digitized by
Google
ASSOCtATtON IN Bte*IZ2Nfc SOLtJTTO!?. 683
molecular weight in dilute solutions, but the molecular weight found
increases regularly with the concentration and does not seem to reach
a limit. The behaviour of acids and oximes with increasing concen-
tration is qualitativlBlj similar to that of an associating gas such as
NO^ when its pressure is increased, and in some cases the association
follows Gtildberg and Waage's law with sufficient closeness, in oihetB
ttie agreement is not satisfactbry. It is generally assumed that the
apparent increase in molecular weight in the case of the alcdhols and
phenols is due also to association taking place with increasing con-
centration.
Although the influence of concentration on association has been
largely investigated, the effect of temperature has not, as yet, been
measured. Molecular weight determinations have been carried out with
a few substances by both the boiling and freezing point methods in
benzene solution, and it might be assumed that the difference between
the values found is due to the difference of temperature. It has yet
to be shown that the increase in cryoscopic molecular weight is not
due, in part, to separation of dissolved substance with the solid benzene.
Until this has been done, it is not justifiable to compare the results of
the two methods in the case of the alcohols and phenols. As most of
the acids give almost the same molecular weight over a considerable
range of concentration, the results obtained in this way with them are
more trustworthy.
Several methods might be used to determine the influence of tem-
perature on association in solution, the most promising being the
variation of vapour pressure with temperature, and the boiling point
method at different pressures. The latter method was chosen in the
present research, partly because it has been more fully worked out,
and partly to elucidate some other points about the boiling point
method.
The value of the molecular rise of boiling point may be calculated
in a number of ways. Arrhenius has shown that it may be calculated
from the heat of vaporisation, lOOr^RI^/L, where R is the gas
constant, L the heat of vaporisation of one gram of the solvent, and
T the absolute temperature. The total heat of vaporisation of benzene
has been determined by Regnault {Memoires de V Institute 26, 881),
and Schiff {Annalen, 1888, 244, 344) has determined the specific heat
at different temperatures. The heat of vaporisation of benzene was
calculated by means of their formulee at intervals of 10% and the
values so obtained substituted in van't Hoff's equation.
By means of the latent heat equation, we can substitute L in
IOOt^ BT^/L; we thus obtain IOOt'- Mp/ {dp/ dt) (Nernst and Roloff,
ZeU. physikxa. Chem., 1893, 11, 24)*.
The same formula may be derived from the lowering of vapour
Digitized by
Google
684 INNES: THE INFLUENCE OF TEMPEBATURE ON
pressure equation, p—p'/p = nl^ (Ostwald, GrundiHss, 203, 1899
edition). The equation, 100t = MpKdpjdt), enables us to calculate the
molecular rise of boiling point from the rate of change of vapour
pressure with temperature, provided the substance has the normal
molecular weight in the state of vapour at the boiling point, and the
method is quite independent of the state of association of the liquid
solvent.
The molecular rise of boiling point at atmospheric pressure haa
been determined in this way for a large number of solvents by Beck-
mann and Fuchs {Zeit. physUeal. Chem., 1895, 18^ 492). The results
obtained show a satisfactory agreement with those obtained by the
direct method.
In the present case, the value of the molecular rise of boiling point
has been calculated from Ramsay and Young's determinations of the
vapour pressure of benzene at different temperatures (Ramsay and
Young, PhU, Mag., 1887, [v], 23, 61 ; Young, Trans., 1899, 66, 501).
The results so obtained are compared with those of determinar
tions of the rise of boiling point, using phenanthrene, benzophenone,
and in three cases benzil as dissolved substances, at pressures ranging
from 31 to 109 cm. The pressures were chosen so as to give differences
of about 10^ in the boiling point of the benzene.
Determinations were also carried out at various pressures with
typical abnormal substances. The substances chosen had to be solids,
as it would be exceedingly difficult to introduce a liquid into the
apparatus without disturbing the pressure ; it was also necessary that
they should be easily soluble and have very little vapour pressure at
the highest temperature used. Benzoic acid, o-bromobenzoic acid,
/3-benzilmonoxime, and dimethyl tartrate were the substances used.
The value of the results with benzoic acid may be partly vitiated by
its volatility.
The Method.
In order to carry out the experiments, it was necessary to maintain
a very constant pressure in the boiling point apparatus for a consider-
able time.
The arrangement of the apparatus,* for pressures less than the
atmospheric and its method of working will be readily understood on
reference to Fig. 1.
The Beckmann apparatus is seen to the right ; both the boiling point
tube and the vapour mantle are connected to the large bottle, B, and
the pressure in the apparatus may be found by reading the manometer,
* An apparatus for maintaining a constant pressnre naar that of the atmosphere,
similar in principle to that used, has been described by A. Smit [{ZeU. physikal,
Chem., 1900, 2S, 88).
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
685
H^ and subtracting from the barometric height. The whole apparatus
is connected with a water vacuum pump. iV is a drying tube contain-
ing calcium chloride. The syphon barometer tube, U^ has a platinum
wire fused through at K and is connected to the pump through the
tap t. The tube L has a platinum wire fused through its lower end, w^
electrical connection is made with the copper wire lead by a little
mercury. Imagine the apparatus to be working and the tap to be
open, the pressure in the apparatus falls and the mercury in the right
Digitized by
Google
686 INNES : THfe INFLUENC* OF TfekPERATtlRE Ol*
hand limb of U rises until it touches the platinum wire at w ; the
circuit of the relay, R, is then complete and the arm, a, is raised, this
breaks the circuit of the magnet, M, and the tap t is closed by the
spring Si, Owing to leakage in the apparatus, the pressure gradually
rises and the mercury falls until the contact at w is broken ; the arm, a,
then falls, completing the circuit of M, and the tap is opened. The
pump was worked at its full power in all the experiments. To prevent
too great a rush of air when the tap opened, a capillary tube, c, was
placed between the pump and apparatus, and a final adjustment given
to the rate at which the air was pumped out by the screw clip, d. It
is evident that when the tap opens the pressure in U falls much more
rapidly than in the bottle, JS', consequently the mercury rises and the
tap is closed before the pressure in ff has time to fall much ; the
mercury in U then falls almost to its original height owing to equal*
isation of pressure in U and B*, The large bottle, j8, was connected to
^ by a capillary tube (c'). The pressure in B, therefore, only follows
the changes in ff slowly ; it is obvious that if the pressure in ff varies
rapidly by small amounts about a mean pressure, the pressure in B
will be practically constant and will be the mean pressure in ff. The
natural leakage in the apparatus was not sufficient to keep the tap
opening and shutting quickly enough to give the most constant pres-
sures ; the whole apparatus was so tight that, working at 109 cm., the
pressure only fell 2 cm. in an hour with the tap closed. A small flask,
Fy containing a little sulphuric acid was therefore connected to Bf^
and a stream of air, which could be regulated by means of a screw
clip, allowed to bubble through the acid at a convenient rate. With
the tap opening 20 to 30 times per minute, no motion at all could be
observed, even with a magnifying telescope, in the mercury manometer,
H\ using a water manometer for pressures near the atmospheric, only
a slight motion, about 1/5 mm., was visible. To alter the pressure, the
tube, Z, is raised or lowered to the necessary amount. The tube slides
in a piece of rubber pressure tubing and can easily be adjusted with
sufficient accuracy in the required position. The surface of the mercury
at W was covered with a little alcohol.
The magnet, if, was kindly designed for me by Dr. D. K. Morris so
as to give as equal a pull as possible over a considerable range. An
iron plug {y) was connected to the keeper and this moved in the core
of the magnet. The bottom and sides of the magnet were encased in
iron. An ordinary glass tap was used for <, this was fitted, at a suit-
able angle, with a brass arm held on by plaster of paris.
The Beckmann boiling point apparatus was arranged in the usual
way. A metal vapour jacket was used y this was about one-third filled
with benzene. The boiling point tube and condenser were made in
one piece, and the mouth of the boiling tube contracted so that a small
Digitized by
Google
ASSOCIATION IN BENZENB SOLUTION. 68*?
cork ooald be used. It, is of course, most important that there should
be DO leak in the boiling tube, as benzene would be swept out of it by
the escaping air. A rubber stopper could not be used to hold the
thermometer, as it was liable to absorb considerable quantities of
benzene : a good, well softened cork was found to be perfectly tight.
The loss of benzene after 3 or 4 hours seldom exceeded 0*1, and never
0*2 gram. The liquid in the boiling tube boiled quietly both under
reduced and increased pressure, even the platinum wire usually fused
through the bottom of the boiling tube was unnecessary. The beads
were placed in the tube in the way described by the author (Trans.,
1901, 79, 261), and platinum clippings placed over them. The plati-
num cylinder was not used as the boiling temperatures were not high
enough to make its use of advantage. The space between the boiling
tube and vapour jacket was packed at top and bottom with asbestos
paper. A gas regulator was used to keep the gas pressure constant.
The flames were protected from draught by pieces of zinc fitting closely
to the Beckmann stand, and the whol^ apparatus was surrounded by
a zinc screen as high as the top of the boiling tube. An electrical
tapper was used to tap the thermometer.
In the experiments at increased pressure, the capillary (c) was con-
nected to a large metal reservoir into which air was forced by a large
bicycle pump. A mercury manometer was in connection with the
reservoir. The flask F was removed, and the delivery tube from jff
joined to a tube drawn out to a point and dipping into water. The
rate at which the air escaped could then be readily regulated by a
screw clip. The relay was cut out and the leads from U connected
directly to the large battery and M. It will be seen that the tap now
opens when the pressure falls below that fixed upon and closes when
it rises above it The tap t was held in by a spring which pressed
gently against it It was found that more satisfactory results were
obtained if the pressure in the air reservoir was kept considerably
above that in the apparatus ; an excess pressure of at least one-fourth
of an atmosphere was used.
In the earlier experiments the pressure was allowed to vary con-
siderably. In series (16) the pressure in the reservoir changed by
about an atmosphere from time to time. In the other series at 79 cm.
the variation was about one-fourth of an atmosphere. The experiments
at 109 cm. were all carried out with a pressure in the reservoir which did
not vary more than 2 cm. about the mean, and the thermometer read-
ings were taken when the mercury in the manometer stood in the
mean position. It will be seen that very concordant results were
obtained, however much the pressure changed ; the only reason for
giving more attention to the pressure in the later experiments was
that it seemed safer to work under as constant conditions as possible.
Digitized by
Google
688
INNES: THE INFLUENCE OF TEMPERATURE ON
r After the liquid had hoiled about an hour, the temperature be-
came as constant as it is at atmospheric pressure under favourable
conditions.
When the temperature had become constant, the clip cT was closed,
the tube S removed, and the weighed pastille placed in it. The tube
was then replaced, the clip removed, and S pushed well down into the
rubber tube ; on tapping gently, the substance fell into the boiling tube.
The tube was then partly withdrawn and the clip replaced, when
everything was ready for the addition of the next portion of substance.
Working at increased pressures, the rubber tube was wired to JS before
the clip was opened, otherwise there was danger of JS being blown
out. As the rubber tube is fully distended in this case, the substance
falls in without difficulty.
The benzene used in the experiments was carefully purified and
was dried over sodium. The substances, with two exceptions, were
purchased from Kahlbaum, and were pure. The /3-benzilmonozime
was made according to Meyer and Auwers' instructions, and melted
at 113^ to 114°. The dimethyl tartrate was prepared by Frankland
and Aston's method, and purified by distillation and precipitation from
benzene by light petroleum.
The determinations at about 80° were carried out under atmospheric
pressure.
Determination of the Molecules Rise of Boiling Point at DifferenJt
Temperatures,
Column 1 gives the number of the series.
Column 2 „ the pressure in cm. at which the experiments were
carried out.
Column 3 „ the corresponding temperature.
Column 4 „ the weight of solvent.
Column 5 „ the weight of substance.
Column 6 „ the observed rise of boiling point.
Column 7 „ grams of substance per 100 grams of solvent.
Column 8 „ 1/100 gram-molecules of substance per 100 gramff
of solvent.
Column 9 ,> the molecular rise of boiling point.
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
Table I.
2.
3.
4.
6.
DiphenylamvMy CigHjiN^lGQ.
24*4
48 0
21-65
0-2696
0-139
0-6984
0-865
1 692
0-840
2-468
1-270
3-384
1-678
1-28
3-29
7-63
11-62
160
PJienanihrene, Oi^Hjq^ITS.
81-1
63-7
31-7
64-3
20-18
0-3040
0-184
0-706
0-427
1-270
0-765
1-989
1-166
3-127
1-789
19-81
0-2424
0147
0-751
0-466
1-465
0-875
1-977
1-175
2-973
1-717
1-54
3-57
6-42
10-06
15-8
1-26
3-87
7-49
10-19
15-32
Benzophenane, Ci3H^qO = 182.
31 0
53-7
21-81
0-2600
0-135
0-501
0-268
0-936
0-505
1-578
0-835
2-345
1-282
8-360
1-733
1-24
2-40
4-48
7-52
11-21
16-07
Benzil,C^^Il^fii = 2\0,
85-8
57-6
22-19
0-2794
0-128
0-454
0-211
0-771
0-867
1-157
0-634
1-702
0-774
2-401
1068
1-28
2-08
3-53
6-30
7-81
1103
0-764
1-95
4-45
6-88
9-47
Mean..
0-863
200
3*61
5 66
7 05
Mean..
0705
2-17
4-21
5-73
8-61
Mean..
0-683
1-32
2-46
4-13
6-16
8-83
Mean..
0-611
0-99
1-68
2-63
3-72
5-24
Mean..
Digitized by
18-4
18-7
19-6
19-2
18-4
19-0
21-3
21-3
20 9
20-6
20-1
21-0
20-9
21-4
20-8
20-5
19-9
20-9
19-8
20-4
20-6
20-2
20-0
19-6
20-2
21-0
21-8
21-2
21-1
20-8
20-3
21-0
Google
690 INNES: THB INFLUENCE OF TEMPEBATUBE ON
J
Table I. (continued).
1.
2.
3.
4.
6.
6.
7.
8.
9.
1
Fhenanthrene.
6
43-6
68-8
2418
0-2320
0135
0-982
0-551
24-5
0-5180
0-284
2-19
1-23
22-6
1-661
0-895
7-03
3-95
22-7
2-314
1-239
9-79
6-50
22-5
8-474
1-785
14-70
8-26
21-6
4-516
2-275
19-11
10-74
Mean...
21-2
22-6
7
43-5
63-3
24*99
0-2240
0-112
0-910
0 512
21-9
0-4968
0-268
202
1-13
22-7
0-9260
0-472
8-76
211
22 S
1-546
0-787
6-29
3-53
22-8
2477
1-240
10-07
6-65
Mean...
21-9
22-22
Benzophmone,
8
43-48
63-3
22-7
0-2450
0-128
1-08
0-593
21-6
0-5804
0-808
2-56
1-40
21-6
1-160
0 604
5-11
2-81
21-6
1-919
0-973
8-45
4-64
21-0
2-974
1-490
13-10
7-20
20-7
3-847
1-912
16*94
9-31
Mean...
20 5
21-40
9
43-3
63-1
21-63
0-2914
0-170
1-37
0-754
22-6
0-7286
0-406
3-43
189
21-5
1-4308
0-793
6-94
8-71
21-4
2-168
1-171
10-16
6-47
21-4
8-226
1-703
16-18
8-36
Mean...
20-4
21-72
Fhenant^^rene.
10
60-3
72-8
22-19
0-2490
0-159
114
0-642
24-8
0-608
0-371
2-79
1-57
28-7
1-217
0-742
5-58
814
28-7
2162
1-289
9-92
6-57
281
3-247
1-878
14-89
8-37
22-4
4-224
2-404
19-38
10-89
Mean...
22-1
23-6
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
691
Table I. {co7Uinued),
11
61-28
12
61-3
5.
Phenanihrene (continued).
78-2
22-19
0-2230
0-137
0-4970
0-304
0-929
0-573
1-507
0-917
2-175
1-304
2-940
1-738
Benzoplunone,
73-2
13
75-7
14
75-8
Phenanihrene.
80 -J
Benzophenone.
80-0
102
2-28
4-26
6-91
9-98
18 '49
22-19
0-3202
0-174
0-6920
0-401
1-182
0 681
1-946
1-130
2-723
1-563
22-19
0-1866
0-113
0-3884
0-260
0-6824
0-447
1-184
0-782
1-648
1-062
2-610
1-647
2219
0-3614
0-225
0-8166
0-490
1-2330
0-740
1-688
1-019
2-724
1-622
1-47
3-18
5-42
8-93
12-49
0-856
1-78
313
5-43
7-56
11-97
1-66
3-75
5-66
7-75
12-50
0-575
1-28
2-39
3-89
5-61
7-58
Mean..
0-807
1-74
2-98
4-91
6-87
Mean..
0-481
1-00
1-40
305
4-25
6-73
Mean..
0-91
2-06
3-11
4-26
6-87
Mean.
23-8
23-7
23-6
23-6
23-3
22-9
23-6
21-6
23-0
22-8
23-0
22-8
22-95
23-5
25'0
25-4
25-6
25-0
24-5
25-25
24-7
23-7
23-7
23-9
23-6
28-8
Digitized by
Google
692
INNBS: THE INFLUENCE OF TEMPEBATUIUB ON
Table I. (continued).
1.
2
8.
4.
6.
6.
7.
8.
9.
BmzU
15
76-5
80 1
22-19
0-2556
0124
1-17
0-658
22-2
0-5626
0-282
2*58
1-23
22-9
0-960
0-498
4-41
2-10
28-6
1-767
0-902
8-11
8-86
23-4
2-975
1-479
13-6
6*50
22-7
4-354
2-126
20-0
9-49
Mean...
22-4
23-8
FhenarUht'ene.
16
79-7
82-4
22-19
0-1886
0123
0-848
0-474
25-9
0-8978
0-257
1-83
1-08
25-0
0-741
0-476
3-40
1-91
24-9
1-060
0-691
4-86
2-73
25-8
1-660
1056
7-62
4-28
24-7
2-251
1-416
10-83
6-80
Mean...
24-4
25-0
17
79-2
82-0
22-19
0-2094
0-183
0-959
0-538
247
0-4496
0-285
2-06
1-16
24-6
0-857
0-550
3-94
2-21
24-9
1-446
0-925
6-64
3-73
24-8
2-213
1-870
1016
6-70
24-0
8-046
1-894
13-98
7-86
24-1
4-281
2-541
19-4
10-9
Mean...
28-8
24-8
Benzophenane.
18
79-2
82-0
22 19
0-2290
0-135
1-05
0-577
28-4
0-4706
0-278
2-16
1-19
28-0
0-767
0-446
3-52
1-93
281
1-166
0-669
5-86
2-94
22-8
1-672
0-967
7-67
4-22
22-9
2-488
1-403
11-4
6-27
22-4
3-327
1-894
15-3
8-39
Mean...
22-6
28*0
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
693
Table I. (eontintted).
1.
2.
8.
4.
5.
6.
7.
8.
9.
J 1
Phmanthrene,
19
108-2
92-8
22 19
0-2800
0-199
1-28
0-722
27-6
0-6670
0-402
2-60
1-46
27-5
1-001
0-708
4-59
2-58
27-4
1-631
1-070
7 08
8-95
27-1
2-268
1-566
10-4
5-85
26-8 .
8-167
2-188
14-6
8-17
26-2
4-271
2-819
19-6
110
Mean...
26-6
27-8
Benzophenone,
20
109 0
98-1
22-19
0-2956
0-197
1-86
0-745
26-4
0-6452
0-427
2-96
1-63
26-8
1-076
0-717
4-94
2-71
26-4
1-599
1078
7-84
4-03
26-6
2-169
1-405
9-95
5-47
25-7
2-871
1-848
18-2
7-24
Mean...
25-5
26-4
BenziL
21
109-0
93-1
22-19
0-3026
0-184
1-39
0-661
27-8
0-618
0-352
2-81
1-34
26-8
0-998
0-579
4-58
2-18
26-6
1-548
0 876
7 10
3-88
25-9
2-556
1-408
11-7
6-58
26 1
8-618
1-940
16-6
7-90
Mean...
24-6
26-2
The values of the molecular rise of boiling point given in the pre-
ceding tables, as well as the molecular weights to be given subsequently,
are all calculated with 0*4 gram less solvent than was actually
taken, to allow for the solvent adhering to the upper parts of the tube
and for that in the state of vapour.
Three series of determinations were carried out with diphenylamine
at 48^. Two of these gave values for r which rose or fell considerably
with the concentration. The pump could only further reduce the
pressure slowly, this may perhaps account for the error. All the
experiments carried out, with the exception of the above-mentioned
VOL. LXXXI, 3 A
Digitized by VjOOQIC
694
INNES : THE INFLUENCE OF TEMPERATURE ON
series, and of two in which the tap failed to act for lack of sufficient
grease, are given in the above tables.
On examination of the numbers, it will be noted that for concentra-
tions of less than 6/100 gram-molecules per 100 grams of solvent, the
Fio. ^.—Phsnanthrene,
128456789 10
Eundredtk-gram molecules per 100 ffranu of benzene.
11
values for r change but little. These values were used for calcu-
lating the mean molecular rise, and the first determination was not
included in the mean if the rise of boiling point was less that 0*1.
The agreement in the double series carried out with phenanthrene
and benzophenone is most satisfactory. The mean values for r never
Fio. 8. — Benzophenone,
28
26
24
22
20
18
— *
X.
^
"***-* J
W
73'
~" — }<
M-W^
■ — .
-—
63-
.
^
T H'
V
*-
..
^—
—■
u—
^S^!S^
sssj^
— ^
"^
128456789
ffundredth-gram molecules per 100 grams of benzene.
10
11
differ by more than 0*3, and for solutions of similar concentration the
agreement for single determinations is, in many cases, within 0-1.
The regularity of the results obtained at different temperatures is
well shown by the curves got by plotting r against the concentratioii
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
695
expressed in 1/100 gram-moleoules of substance per 100 grams of
solvent (Figs. 2 and 3). The curves are practically parallel ; those for
phenanthrene show a slight but distinct downward tendency, whilst
the benzophenone curves are almost horizontal.
The variation of r with temperature is shown clearly by plotting
the values of r for each substance at different temperatures against
the temperature. This is done in Fig. 4. The curve for r, calculated
Fio. 4.
29
27
20
46° 50* 65°
Temperature.
vo luO
from the latent heat (calculated A) and from Ramsay and Young's
determinations of the vapour pressures (calculated B), are also given.
Below is a table of the data from which these were calculated :
Table II.
Temperature.
Heat of vaporisation in
gram cal.
RT^IL.
50°
60
70
80
90
100-4
98 3
96 0
93-7
91-2
21-1
22-6
24-4
26-5
28-8
3 A 2
Digitized by \J OOQ IC
696 INNES: THE INFLUENCE OF TEMPEBATUBE ON
Table II (continued).
Temperature.
Pressure in
cm.
dp/dt.
Mean
pressure.
M.
Mp
dp/dt
40''
60
60
70
80
90
100
110
18-02
26-88
38-85
64-82
75-60
100-8
133-5
178-9
0 881
1-202
1-596
2 068
2-630
3-27
4 04
22-42
32-84
46-83
65-16
88-1
1171
153-7
78
19-85
21-31
22-89
24-58
27-17
27-96
29-68
Mean Values qf the Molecular Rise of Boiling Point at Different
Temperatures,
54^
58".
63°.
73°.
80°.
93°.
Calculated A
2xa
22-2
23 0
25-0
26-6
29-5
Calculated B
2105
21-6
22-4
24-26
25-6
27-8
Phenanthrene
21-0
—
22-4
23-65
25-25
27-3
Benzophenone
20 17
—
21-6
22-95
23-8
26-4
Benzil
—
21-0
—
—
23-3
26-2
General mean for
)heDanthrene and
)eDzophenone ...
20-6
220
23-25
24 1
26-9
The molecular rise of boiling point, calculated from the variation of
vapour pressure with temperature, forms, when plotted against the
temperature, a very regular curve from -5° to +145% the
values increasing somewhat more rapidly than the temperature.
The only value which does not lie well on the curve is that at 85**.
This deviation is accounted for by the fact that the vapour pressures
below and above 80° were determined in two separate researches (by
Hamsay and Young, and Young respectively). In Young's paper, the
vapour pressures calculated by means of Biot's formula and constants
calculated from his own measurements are given. The calculated vapour
pressure at 80° is smaller than that found by direct experiment^ and
is larger at 90° ; although the differences are small, dp/dt is decreased
considerably. If r be calculated from the pressures given by Biot's
formula for 80° and 90°, the value obtained falls well on the curve.
This is shown by the dotted portion of the curve.
The molecular rise found at the various temperatures using phenan-
threne as dissolved substance, agree closely with those calculated from
the vapour pressure, the greatest difference being a little more than 2
well 8tf<. The molecular rise with benzophenone is considerably smaller
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
697
than with phenanthrene. If smoothed curves be drawn, it will be seen
that the difference is nearly the same at the different temperatures
and amounts to about 0*8. It is well known that different values are
obtained for the molecular rise at atmospheric pressure according to
the substance dissolved, even though the substances cannot be con-
sidered as abnormal in the ordinary sense. That this difference is not
due to association in the case of benzophenone is shown by the fact
that the molecular rise does not decrease with concentration. It is
interesting that the two closely related substances, benzophenone and
benzil, give curves which agree closely with one another.
The molecular rise calculated from the heat of vaporisation is greater
than that calculated from the variation of vapour pressure with tem-
perature. The molecular rise found by direct experiment agrees much
better with that calculated in the latter manner : a result which was
hardly to be expected.
Abnormal Subatancea,
Columns 1 to 8 have the same meaning as in Table I (p. 688),
column 9 gives the molecular weight found. The molecular weights
are calculated with the mean molecular rise for phenanthrene and
benzophenone :
Table III.
22
23
24
6.
31-0
Benzoic Acid, C7Hg02=122.
48-3
61-3
537
24-21
01516
0 059
j
0-3373
0135
1
0-5636
0-213
1
1-2550
0-460
1-616
0-585
2-262
0-825
3 001
1-083
63-1
22-19
0-1210
0-066
0-2910
0-143
0-5294
0-254
0-926
0-416
1-601
0-655
2-2724
0-984
73-2
2219
0 1806
0-106
0-3680
0-202
0-6402
0-333
0-905
0-465
1-324
0-660
1
1-813
0-889
2-527
1-208
8-286
1-688
0-637
1-42
2-37
5-27
6-78
9-50
12-o0
0-555
1-88
2-43
4-25
6-89
10-42
0-829
1-69
294
4-16
6-07
8-32
11 60
151
0-622
1-16
1-94
4-32
5-56
7-78
10-33
0-455
1-09
1-99
8-48
5-65
8-54
0-679
1-38
2-41
3-40
4-98
6-82
9-50
12-86
229
223
236
243
248
246
247
185
206
210
226
281
233
182
194
205
207
214
218
223
228
Digitized by
Google
698 iNNfis: THE Influence ob* TBMPBRATtjRB on
Table III. (contintted).
1.
2.
8.
4.
5.
6.
7.
8.
9.
1 1 1 1
BenzotG Acid, C7HgOj = 122
(continued).
25
75-6
80-1
2219
0-8430
0-209
1-57
1-29
181
07016
0-403
8-22
2-64
192
1-026
0-663
4-71
3-86
201
1-891
0-743
6-88
5*23
206
2062
1-070
9-42
7-72
211
2-886
1-441
13-0
10-7
217
4-287
2-088
19-7
16-1
226
5-924
2-823
27-2
22-3
281
26
109-0
93-1
2219
0-2894
0-181
1-33
1-09
194
0-6898
0-369
2-71
2-22
200
0-948
0-560
4-35
3-67
206
1-426
0-809
6-54
5-36
215
2196
1-199
10-08
8-26
223
3-252
1-706
14-9
12-24
282
4-947
2-436
22-7
18-61
247
Benzoic Acid.
The results obtained are graphically expressed in Fig. 5. The tem-
perature at which each series was carried out is shown. The curve for
Fia. 6.—Ben2oic add, CyRfi^. Mol, ii^.sl22.
2 4 6 8 10 12
Hundredth-gram molectiles per 100 grams qf tolverU,
16
benzoic acid in benzene (Beckmann, ZeiL fihysikod, Chem., 1888, 2, 729)
is also given. The freezing point of benzene is 5*4^. The valaes
obtained by the boiling point method at 54° agree closely with those
by the freezing point method. The curve at 63° lies much lower than
Digitized by
Google
ASSOCIATION IK BENZENE SOLt^TION.
699
that at 54P, the difference being greatest in dilate solution, the differ-
ence between the curveB decreases up to 80^. The 93° curve lies higher
than that for 80°. The agreement in the molecular weights at 6*4°
and 54° does not necessarily show that there is no change of association
between these temperatures. It is well known that benzoic acid
volatilises considerably at temperatures below 100°, and it boils at 134°
under 12 mm. pressure. If the benzoic acid had a vapour pressure
of between 3 and 4 mm. in a solution containing 6/100 gram-mole-
cule per 100 grams benzene, this would raise the apparent molecular
weight about 10 per cent., that is, more than 20 units. The apparent
decreasing effect of change of temperature on the association as the
temperature rises may be due to increase of vapour pressure of the
benzoic acid with temperature, and the fact that higher values were
obtained at 93° than at 80° might be due to the same cause.
Although the unknown influence of the vapour pressure detracts con>
siderably from the value of the results with benzoic acid, it may safely
be said that increase of temperature brings about a decrease of associa-
tion between 54° and 80°, since the change of vapour pressure of the
benzoic acid would tend to bring the molecular weights at different
temperatures nearer together :
Table IV.
1. 2.
3.
4.
5.
6.
7.
8.
9.
1
1
o-Brom
\obmzoic
Aeid,C
^HgOjBr
»201.
1
22 1 85-8
57-8
2219
0-3872
0-124
1-78
0-884
807
0-885
0-262
4-06
2-02
832
1-341
0-382
6-15
8-06
845
1-914
0-521
8-78
4-87
861»
23
75-85
80-0
22-19
0-5388
0-193
2-47
1-23
307
1-074
0-364
4-93
2-45
826
1-599
0-529
7-34
3-66
333
2-382
0-758
10-7
6-44
346
8 176
0-989
14-6
7-25
364
4-225
1-278
19-4
9-64
864
24
109-0
93-1
22-19
0-3782
0-167
1-71
0-851
271
0-682
0-288
8-13
1-56
288
1-216
0-475
5-58
2-78
811
1-786
0-668
8-20
408
825
2-600
0-910
11-93
5-93
848
3-748
1-268
17-2
8-66
860
6056
1-660
23-2
11-54
878
6-776
2158
81-1
16-6
382
* The substance apparently dissolved completely, but farther addition of snb-
stanoe caused no increase of boiling point.
Digitized by
Google
700
INNES : THE INFLUENCE OP TEMPERATURE ON
o-Bromobenzaic Add.
Three series of determinations were made with this substance at
58°, 80°, and 93° respectively. The o-bromobenzoic acid was not
sufficiently soluble to make a series of determinations at the freezing
point. From the curves (Fig. 6), it will be seen that the molecular
Fig. 6. — o-Bromobcmoie acid, 07H50sBr. Mol. tr<.=201.
370
850
830
310
290
270
^
A
^
^^
^
}
y
<^
^y
^
/a
•*<
^
^
<
/4
_i^
V
12 8 4 5 6 7
Hundredth-gram moleetiles per 100 grams of hewune.
weight at 80° is, for similar concentrations, considerably smaller than
at 58°, and the difference increases with the concentration. In dilate
solutions, the 93° curve lies considerably below that for 80°, with
increasing concentration, the curves approach one another and finally
become practically parallel :
Table V.
25 I 35*8
4.
5.
fi'Benzilvtonoximet Cj^HjjOjN = 225.
57-8
21-96
0-3852
0-750
1116
1-576
2u33
2-643
0 185
0-320
0-441
0-591
0-716
0-880
1-79
3-48
6-17
7-31
9 43
12-26
0-794
231
1-65
269
2-80
280
3-24
296
4-19
814
6-46
882
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
701
Table V. (continued).
1.
2.
3.
4.
6.
a.
7.
8.
9.
i 1 1
P-BenzUvionoxime, C^^Hj^OgN
= 225
(continued).
26
75-5
80 0 22-19 0 1680
0 072
0-762
0-339
249
1 ' 0-339
0-151
1-56
0-692
243
1 1 0-520
0-233
2-39
106
242
1 ! , 0-897
0 389
4-12
1-83
350
1 1 ' 1*^27
0-693
6-55
2-91
261
! ' 2-019
0-812
9-26
412
269
27
109 0 I 93 1
22'19 1 0-3670
0-180
1-68
0-748
249
'
0-718
0-350
8-30
1-46
251
, 1-034
0-488
4-74
2-11
259
1
1 1-610
0-723
7-39
3-28
273
1 ' 2097
0-896
9-54
4-24
284
' 2-700
l-i97
12-39
5-51
301
P- BenzUmanaxime.
Determinations were carried out at 58°, 80°, and 93°. In the table
of curves, the molecular weights found by the freezing point method in
benzene (Auwers, Zeit, physikal, Chem,, 1893, 12, 701) and in naph-
thalene (Innes, Incmg. Diss. Heidelberg, 1896) are also given. De-
terminations with more concentrated solutions in benzene could not
be carried out at the freezing point, because of the small solubility of
the substance. j3-Benzilmonoxime crystallises with benzene of crys-
tallisation, the formula of the compound is 2C;i4H|^O3N,C0H0. The
abnormality in benzene might be considered to be due to the formation,
in part, of this compound. It is impossible to decide in the present
state of our knowledge whether this is really the case ; it seems more
probable, however, that the combination with the solvent does not
affect the molecular weight to an appreciable extent in this instance.
From the curves (Fig. 7, p. 703), it will be seen that the association
of /S-benzilmonoxime decreases considerably with rise of temperature
up to 80°, the association then seems to increase, the 93° curve lying
somewhat higher than that for 80° ; the greatest difference is about
4 per cent. :
Digitized by
Google
702
INNES: THE INFLUENCE OF TEMPERATUBE ON
Table VI.
1.
2.
3. 4.
6.
6.
7.
8.
9.
1
Dimethyl TartraU, C^U^qO^^^
178.
28
81-0
68-7
24-74
0-8196
0-147
1-31
0-74
186
0-8606
0-337
3 -58
1-99
218
1-470
0-520
6-04
3 -89
242
2 057
0-662
8-46
4-76
266
2-898
0-821
11-91
6-69
802
4-000
1-000
16-48
9-28
842
29
43-8
68-1
2219
0-1704
0-088
0-984
0-653
207
0-4402
0-214
2 02
1-13
208
0-816
0-867
8-74
210
225
1-366
0-566
6-22
8-49
246
2-086
0-764
9-57
6-88
276
2-917
0-957
13-39
7-52
807
8-982
1-156
18-04
10 14
848
5 122
1-844
28-51
13-20
886
30
61-28
78-2
22-19
0-2966
0-162
1-86
0-962
196
0-6744
0-839
8-09
1-74
212
1-207
0-567
5-54
8-11
227
1'886
0-794
7-61
4-22
247
2-894
1-091
11-84
6-66
288
4 184
1-871
16-90
9-60
822
6-218
1-571
21-31
11-98
864
6-288
1-761
26-69
14-44
388
81
78-6
79-2
22 19
0-2670
0-169
1-28
0-69
185
0-5528
0-301
2-64
1-42
202
0-990
0-500
4-54'
2*65
218
1-662
0751
7-68
4-88
244
2-528
1-029
11-60
6-61
271
8-919
1-371
18-0
11-2
815
5-026
1-592
28-1
14-86
848
82
109-0
98-1
2219
0-4280
0-266
1-96
1-10
198
0-8590
0-492
8-94
2-216
215
1-459
0-780
6-69
8-76
230
2-002
0-997
9-18
6-16
247
8-103
1-351
14-24
8-00
282
4170
1-618
19-14
10-76
818
Dimethyl Tartrate.
With increasing dilution the molecular weight of dimethyl tartrate
tends towards the same value at the various temperatures, the mole-
cular weight found being in every case near the normal (Fig. 8, p. 705).
In more concentrated solution the molecular weight decreases with rise
of temperature between 54^ and 79°, the decrease increasing with the
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION.
703
temperature. The molecular weight at 93^ is lower than that at 79^
in solutions of moderate concentration, but the curves cross at a
concentration of 8-7/100 mols. The curves are all slightly concave
downwards, with the exception of that at 93°, which is almost straight.
Fio. 7. — fi'B&nzilmonoxime.
12 8 4 6 6
Hundredlh-gram moUeuUs per 100 grama of benaene.
The degree of concavity increases with the temperature up to 79°. An
attempt was made to carry out a series of determinations of the
molecular weight by the freezing point method. Only one determina-
tion could be made ; at higher concentrations the substance sometimes
separated out :
Digitized by
Google
704
INNES: THE INFLUENCE OF TEMPERATURE ON
Freezing Point MetJiod,
Weight of
benzene.
Weight of
substance.
Lowering
of f.p.
Substance
per 100 grams
solvent.
Mol8./100
per 100 grams
solvent.
Mol. weight.
10-0
0-197
0-189 1 1-97
1
1-11
271
The association of all the substances examined decreased with rise
of temperature up to 80°.
Thermodynamical reasoning shows that association may either
increase or decrease with temperature, according as the heat of dis-
sociation is negative or positive. Further, the heat of dissociation
may change its sign with increase of temperature; if this change
is from positive to negative, the dissociation increases up to the
temperature at which the heat of dissociation is zero, then decreases.
Instances of such a change have been observed for ionic dissociation
by Arrhenius, and it is probable that among gaseous substances car-
bon dioxide, COj, silicon hydride, SiH^, and selenium hydride, SeH^,
behave thus at high temperatures. The observed molecular weight
of benzoic acid and of j3-benzilmonoxime is greater at 93° than at 80°
for all concentrations ; that of o-bromobenzoic acid and of dimethyl
tartrate is smaller in dilute solutions, in more concentrated solutions,
the molecular weight of the former is the same at both temperatures,
whilst that of the latter is greater at the higher temperature. This
apparent decrease of dissociation with rise of temperature may possibly
be due to a change of sign in the heat of dissociation. It does not,
however, seem probable that this is the true explanation. If a change
of sign actually occurred, we should expect this to take place at a
different temperature for each substance. Further, increase of tem-
perature brings about an increasing amount of dissociation of dimethyl
tartrate up to 80°; that a further increase of 14° in the temperature
should cause a decrease of dissociation is extremely improbable. The
fact that the experiments at 93° were carried out at increased pressure,
whilst those at the other temperatures were at reduced or atmospheric
pressure, suggests that the high molecular weights at this temperature
may be due to an error in the method. The constancy of the value of
the molecular rise of boiling point at different concentrations for each
of the three substances examined, as well as the position of the
molecular rise of boiling point at this temperature on the temperature
curve (Fig. 4), seems to show that the method is as accurate at in-
creased as it is at reduced or atmospheric pressure. It does not seem
probable that the increase of molecular weight at 93° is due to the
influence of the solvent.
Digitized by
Google
ASSOCIATION IN BENZENE SOLUTION. 705
The heats of dissociation (Q) of jS-benzilmonozime, benzoic acid, and
o-bromobenzoic acid from double to single molecules were calculated by
means of the equation :
1 ^ 1 a:? Q/l 1\
Molecular loeigJU.
in which x^ is the degree of dissociation at the absolute temperature T^
T,
V^ is the volume occupied by one gram-molecule of the sub-
stance calculated as double molecules, at the temperature Ty^
Digitized by
Google
706
ASSOCIATION IN BENZENE SOLUTIGN.
in which V^ is the volume occupied by one gram-molecule of the sub-
stance calculated as double molecules, at the temperature T^
„ V WAS taken as the volume of benzene in litres in which one
gram-molecule of the substance, calculated as double mole-
cules, was dissolved, and V^ as equal to V^
„ X was calculated by a slight modification of the equation used
to calculate the degree of dissociation of a gas from its
density. x = M is the molecular weight of the
double molecule, m the molecular weight found.
The following data were used :
Mols..
100
173
8-00
2 00
3 00
Wlj.
wig.
Xy,
x^
278
331
331
331
336
331
353
853
353
853
V,
Q.
)9-Benzi]inon- (
ozime [
356
264
344
332
220
264
249
328
318
195
0-258
0-704
0 169
0-211
0117
0-704
0-807
0-232
0-264
0-261
13
14
818
12-3
8-18
14600
19900
o-Bromobenzoic (
acid \
Benzoic acid
7500
5300
23000
The heat of dissociation of gaseous N^O^ (ei = 26'7°, <j = lll'3°) is
12,900 calories ; that of iodine vapour, 28,600 cal. ; acetic acid, 20,000
cal., and dimethyl ether hydrochloride, 8600 cal. It thus appears that
the heat of dissociation of a substance in solution is of the same order
as that usual for a vapour. The above result adds another instance
to the many already known of the close analogy between the behaviour
of a substance in solution and in the state of vapour.
The above experiments were carried out in the Chemical Laboratories
of the University of Birmingham. I should like to take this oppor-
tunity of thanking Professor Percy Frankland for his kindness in
supplying most of the apparatus required for the experiments.
Univkrsity College,
Liverpool.
Digitized by
Google
PREPARATION OF ABSOLtJTE ALCOHOL FROM STRONG SPIRIT. 707
LXXIII. — The Preparation of Absolute Alcohol Jrom
Strong Spimt.
By Sydnby Young, D.Sc., F.RS.
Owing to the fact that ethyl alcohol, like nrpropyl, tsopropyl, and tert,-
butyl alcohols, although miscible with water in all proportions, forms
with it a mixture of constant boiling point which distils without
change of composition at a temperature lower than either component,
it is impossible by any process of fractional distillation to separate the
pure alcohol from a dilute solution. All that can be done is to separate
the mixture of constant boiling point on the one hand, and water on
the other, from the solution. In the case of the other alcohols
referred to, if we remove a portion of the water from the mixture of
constant boiling point by some other method, then by fractional distil-
lation of the stronger alcohol with a very efficient still-head we can
effect its separation into the pure alcohol and the mixture of constant
boiling point.
Pure ethyl alcohol, however, boils less than 0*2^ higher than the
mixture of minimum boiling point, and such a separation is therefore
impracticable.
The method which has invariably been employed to prepare absolute
alcohol is to treat the strongest spirit obtainable by distillation with a
dehydrating agent, and the action of such agents has been studied by
many investigators, notably by MendeUeff and by Squibb.
In his classical paper on this subject, Mendel^eff {Ann, Phys, Chem,^
1869, [ii], 138, 230) discusses the behaviour of different dehydrating
agents and concludes that freshly ignited lime is the only substance
capable of giving good results, and that even when lime is employed
special precautions must be taken.
The results obtained by Mehdel^eff with four different specimens
show excellent agreement, the greatest difference from the mean sp. gr.
at 0°/4^ 0-806254, being only O'OOOOll, but Mendel^eff himself states
that when the alcohol was freshly distilled over lime he noticed an
ethereal odour which, however, disappeared on repeated distillation,
although the sp. gr. remained unchanged. The value 0*80625 has
been almost universally adopted as the correct sp. gr. at 0^/4°, but
Squibb (/. Amer. Ghem. Soc., 1893, 15, 126) has obtained even lower
values by slow percolation through lime. In an earlier investigation,
Squibb had found the sp. gr. 0*79350 at 15-6715-6^ which corresponds
to 0*80581 at 074"", but in the paper referred to he states that after
long contact with lime and subsequent percolation many times through
an improved apparatus, alcohol was obtained with a somewhat higher
Digitized by
Google
708 YOUNG : THE PREPARATION OF
specific gravity. The results were not perfectly concordant, the mean
of the best being given as 0*793562 at 15*6715*6'^, corresponding to
0-80587 at 074"". Squibb states that in his opinion absolute alcohol
had not yet been obtained.
When two liquids of different chemical type are distilled together, a
definite mixture of minimum boiling point in many cases distils over
first, the last portion of the distillate consisting of that component
which was originally present in excess. Ethyl alcohol may be regarded
as a derivative of water and as belonging to the water type H-O-C^H^,
or as a derivative of ethane, or, more generally, of a paraffin, OjHg'OH.
In its properties, it exhibits analogies in some respects to water and in
others to the paraffins. Most dehydrating agents which react or com-
bine with water behave in a somewhat similar manner towards the
alcohols, though to a less degree, and to a diminishing extent as the
molecular weight increases, and this accounts for the unsatisfactory
results obtained with them. Thus phosphoric oxide gives phosphoric
acid with water, and a mixture of ethyl hydrogen pliosphates with
ethyl alcohol; with barium oxide, water forms barium hydroxide,
whilst ethyl alcohol forma, according to Forcrand, a compound
SBaOyiCjH^O ; sodium acts in precisely the same way on the alcohols
as on water, but the intensity of the action diminishes rapidly as the
complexity of the alkyl group increases; calcium chloride forms a
crystalline hexahydrate with water, and a crystalline tetra-alcoholate
with methyl or ethyl alcohol; the behaviour of anhydrous copper
sulphate is striking ; it dissolves rapidly in water, and, on evaporation,
crystals of CuSO^jSH^O are deposited ; in methyl alcohol, it dissolves
slowly, but to a considerable extent, giving a blue solution from which,
according to Forcrand, greenish-blue crystals of CuS04,CH^0 may be
obtained ; anhydrous copper sulphate is, however, quite insoluble in
ethyl alcohol, and will extract some water from strong spirit, but it is
not a sufficiently powerful dehydrating agent to remove the whole.
If we compare the homologous normal primary alcohols together, we
find that in other respects also, as the molecular weight rises, the alkyl
group has increasing, and the hydroxyl group diminishing, influence,
and that the properties recede from those of water and approach those
of the corresponding paraffin. The table of boiling points on p. 709
shows this clearly.
Thus methyl alcohol boils only 35*3° lower than water, but 228-7°
higher than methane, whilst cetyl alcohol boils 244° liigher than water,
but only 56*5° higher than the corresponding paraffin. A^in, whilst
methyl, ethyl, and propyl alcohols are miscible in all proportions with
water, butyl alcohol is only partially miscible, and cetyl alcohol is
practically non-miscible with water.
Lastly, while a mixture of methyl alcohol and water distUs normally.
Digitized by
Google
ABSOLUTE ALCOHOL FBOU STBONO SPIRIT.
709
Number of
Boiling points.
carbon
atoms.
Paraffin.
A.
Alcohol.
A.
Water.
1
-16^"
+ 228-7*'
+ 64-7-
-36-3'*
100'
2
-93
171-8
78-3
-21-7
—
3
-45
142-4
97-4
-2-6
—
4
+ 1
116-0
117-0
+ 17-0
—
6
36-8
101-7
138 0
38 0
—
6
69-0
88-0
167-0
67-0
—
7
98-4
77-6
176-0
76-0
—
8
126-6
70-4
I96 0
96-0
—
16
287-5
56-5
344-0
244-0
—
both components being eaaily separated, ethyl and propyl alcohols form,
with water, mixtures of minimum boiling point, and the behaviour of
butyl alcohol and water approaches that of two non-miscible liquids.
We cannot well study the miscibility of the alcohols with the
corresponding paraffins, because the first four are gaseous at the
ordinary temperature and most of them are difficult to prepare in a
pure state. Normal hezane, however, can be obtained without much
difficulty, and we find that, although the lowest alcohols are miscible
with this hydrocarbon in all proportions, they form with it mixtures
of minimum boiling point. Benzene is much more easily obtained than
hexane, and behaves in a similar manner. The behaviour of mixtures
of benzene with the lower alcohols has been studied by Miss Fortey
and myself, and is fully considered in a separate paper ; it will be
sufficient here to state that whilst methyl, ethyl, i^opropyl, n-propyl,
fer^. butyl, and t^obutyl alcohols form mixtures of minimum boiling
point with benzene, isoamyl alcohol does not.
Thus ethyl alcohol forms mixtures of minimum boiling point, both
with water and with benzene (or hexane), whilst benzene and water
are practically non-miscible and distil over together at a temper-
ature lower than the boiling point of either pure liquid ; it
seemed reasonable to expect that a particidar mixture of all three
liquids would boil constantly at a still lower temperature. The original
mixture would, in that case, tend to separate on distillation into three
instead of two fractions : — (1) a definite mixtuie of all three liquids
boiling at a lower temperature than any of the three components, or
than any mixture of any two of them ; (2) a mixture of two com-
ponents boiling at a lower temperature than any single one ; (3) that
component which was originally in excess.
When aqueous alcohol is distilled with a dehydrating agent, the
water is more or less completely retained in the still, the dried alcohol
passing over as the distillate ; if, however, a mixture of ethyl alcohol,
VOL. LXXXI. 8 B
Digitized by
Google
710
YOUNG: THE PREPARATION OF
benzene, and water behave as suggested above, it should be possible to
reverse the process, the water being carried over in the first part of
the distillate and pure alcohol coming over last. Such a method would
be advantageous for this reason, among others, that it is almost always
easier to separate the least volatile component of a mixture in a pure
state by distillation than the more volatile components. These antici-
pations were fulfilled, and it was, in fact, found possible to eliminate
the water from strong spirit by distillation with benzene.
In all the experiments carried out in connection with this investiga-
tion, I have had the advantage of the able co-operation of Miss E. C.
Fortey, B.Sc.
The following table gives the boiling points of ethyl alcohol, benzene,
and water, and of the mixtures of constant boiling point that they can
form, also the composition of these mixtures.
The benzene-water values were calculated from the known vapour
pressures of benzene and water, these two liquids being practically non-
miscible, and their correctness was confirmed by actual experiments ;
the other values were determined experimentally :
Liquid of constant boiling
Boiling
point.
Percentage composition.
point
Alcohol.
Water.
Benzene.
1. Alcohol, water, and benzene
(W.A.B.)...
2. Alcohol and benzene ...(A.B.)...
8. Water and benzene (W.B.)
4. Alcohol and water (A.W.)...
6. Alcohol (A.)...
6. Benzene (B.)...
7. Water (W.) .
64-85°
68-25
69-25
78-15
78-8
80-2
100-0
18-5
82-41
95-57
100
7-4
8-88
4-43
100
74-1
67-59
91-17
100
•
It will be seen that the lowest boiling point is that of the ternary
mixture (W.A.B.), so that whatever mixture of the three liquids is dis-
tilled— ^unless one constituent is present in relatively very small quantity
— this ternary mixture will come over first. If there is more than
sufficient benzene to carry over the whole of the water, and if the
alcohol is present in excess, the ternary mixture will be followed by
the binary ( A.B.) mixture, and the last substance to come over will be
alcohol. This is the case, for instance, if we distil a mixture of equal
weights of benzene and, say, 93 per cent, (by weight) alcohol wiUi a
very efficient still-head. The distillate is at first turbid, and on
standing separates into two layers, although the original mixture is
quite clear ; the temperature remains constant at 64*85° for a long
time, it then rises slowly, but with increasing rapidity, to the middle
Digitized by
Google
ABSOLUTE ALCOHOL FROM STRONG SPIRIT. 711
temperature, 66'55^» between 64*85° and 68*25°, when the distillate
ceases to be turbid ; the temperature then rises more and more slowly,
and becomes nearly steady for some time at or a little below 68*25°,
when the binary (A.B.) mixture comes over. Then the temperature
rises again with increasing rapidity, and very rapidly indeed as it passes
the middle temperature, 73*3°, between 68 25° and 78*3°; afterwards,
the rise becomes slower and slower until the boiling point of alcohol is
reached.
It should thus be theoretically possible to carry over the whole of
the water in the first fraction and to remove the whole of the remain-
ing benzene in the second, leaving pure alcohol in the last. It will,
however, be noticed that the difference between the boiling points of
the ternary (W.A.B.) mixture and the binary (A.B.) mixture is only
3*4°, so that the separation is a difficult one and in practice it was found
that even when the mixture was distilled through an 18 column
Young and Thomas dephlegmator at the rate of 1 drop per second,
the alcohol in the final distillate, though containing the merest trace
of benzene, still retained about 1 '4 per cent, of water aa compared
with 7*4 per cent, in the original alcohol.
By redistilling the partially dehydrated alcohol once or twice with a
further quantity of benzene, the water could, however, be finally
eliminated.
It is convenient to collect the distillate in four fractions : —
I. From 64*85° to about 67*5°. This consists mainly of the ternary
(W.A.B.) mixture.
II. From 67*5° to about 73°. This consists chiefly of the binary
(A.B.) mixture.
III. From about 73° to 78*3°. The distillate should be collected in
III for a little time after the temperature has become constant at
78*3° to remove the benzene as completely as possible. This fraction
is much richer in alcohol than II ; it should be relatively very small
in amount.
lY. The dehydrated alcohol ; it is not essential that this should be
distilled, it may simply be run off from the still
Fraetion /, bailing at 64*85° to about 67*5°.
This distillate is turbid and separates into two layers, the smaller
(and usually but not necessarily the lower) layer consisting of water,
a good deal of alcohol, and some benzene, the larger layer consisting of
benzene with a good deal of alcohol and a little water. On adding
more water, shaking, and allowing to stand, two layers are again
formed ; the lower one, A, contains most of the alcohol and water with
very little benzene ; the upper one, B, contains nearly all the benzene
3 B 2
Digitized by
Google
712 YOUNG: THE PREPARATION OF
with very little alcohol and water. The two layers should now be
separated and the benzene washed once or twice with water to remove
the alcohol (and also the water, since alcoholic benzene dissolves more
water than pare benzene) more completely ; the water may be added
to A.
The two liquids, A and B, should now be distilled, preferably with
an efficient still- head.
Diatillation of A. — On distillation, this liquid tends to separate into
three fractions : (1) a very small quantity of the ternary (W.A.B.)
mixture boiling at 64'85° ; (2) the binary (A. W.) mixture boiling at
78*16^; (3) water. The whole of the benzene comes over below 78*15°,
and this small fraction may be added to other quantities of the ternary
mixture. The rest of the distillation consists simply in the recovery
of strong spirit from dilute alcohol.
Distillation qf B. — On distillation, B tends to separate into three
fractions : (1) a minute quantity of the ternary mixture, but this may
be absent if the benzene has been very thoroughly washed with
water; (2) the binary (W.B.) or possibly (A.B.) mixture, also;
exceedingly small in amount ; (3) pure benzene.
There is no advantage in keeping the two first fractions separate,
they may be collected together and added to the ternary mixture from
other distillations. After the temperature has reached 80*2°, the liquid
in the still consists of pure benzene and there is no necessity to distil it.
Fraction II, boiling from about 67*5° to abotU 73°.
This distillate is clear and consists chiefly of the binary (A.B.) mix-
ture, but it contains a little water. Dilute alcohol and benzene might
be obtained from it by addition of water, but it is more advantageous
to add it to the next mixture of strong spirit and benzene that is to be
distilled. If equal weights of strong spirit and benzene are again
taken and the (A.B.) mixture is added, the alcohol obtained will be
drier and the quantity a little larger. The fractions obtained in
this case will be the same as before, but fraction II (b. p. 67*5 — 73°)
will be larger. If at any time the quantity becomes too large to
be made use of in this way, water may be added to a portion and
pure benzene and strong spirit recovered as described under I.
Fraction III, boiling at about 73° to 78*3°.
This fraction should be relatively small ; it consists chiefly of
alcohol with some benzene and very little water. It is not worth
while to redistil the fraction from a single operation, but the frac-
tions from a series of distillations should be added together an4
stored until the quantity is large enough to be redistilled.
Digitized by
Google
ABSOLUTE ALCOHOL FROM STRONG SPIRIT. 713
On distillation, fraction (1) will be absent or very small ; fraction
(2) will be fairly large ; fraction (3) about the usual small quantity ;
fraction (4) large. Thus an additional quantity of partially dehydrated
alcohol will be obtained.
Fraction IV, bailing at 78*3°
This fraction consists of alcohol containing no more than a trace of
benzene and nearly free from water. It is impossible to state definitely
what percentage of water will remain in the alcohol ; the more efficient
the still-head, the slower the distillation and the larger the amount of
benzene originally added, the drier will the alcohol be.
In this process, the whole of the dehydrating agent, benzene, is re-
covered except the small amount lost by evaporation. There need also
be hardly any loss of alcohol. As there is no chemical reaction, there
is no possibility of introducing any impurity into the alcohol except,
perhaps, a minute trace of benzene.
A considerable number of distillations with an 18 column Young
and Thomas dephlegmator were carried out by this method at the rate
of 1 drop per second. The alcohol employed was obtained from Kahl-
baum ; its sp. gr. at 0°/4° was 0-82907, and it therefore contained 7*4
per cent, of water by weight ; it was quite free from other impurities.
After the temperature had reached 78*3°, the residual alcohol was
collected in fractions and the sp. gr. of the first and last were, as a
rule, determined. The results obtained were as follows :
(i) A mixture of 325 grams of 92 '6 per cent, alcohol (sp. gr. 0*82907)
and 325 grams of benzene (dried with sodium) was distilled. After
the temperature had reached 78*3°, the following fractions were col-
lected :
Percentage of water
by
weieht from
Weight.
Sp. gr. at O74-.
Mendel^ra data.
1
20-9 grams.
0-81176
1-85
2
85-6 „
—
—
3
. 35-7 ,.
0-80976
MO
Residue...
22-8 „
—
—
1650 „
It is very probable that the first fraction was in this case collected
a little too soon, and contained a perceptible amount of benzene, which
would raise its sp. gr.
(ii) Similar to the first, but the benzene was added in three portions.
The weight of each liquid was 200 grams, and that of the dehydrated
alcohol 89*5 grams. The sp. gr. of the first and last fractions
Digitized by
Google
714 young: the preparation of
were 0*81043 and 0*80946, corresponding to 1*40 and 1*03 per cent,
of water respectively.
(iii) A mixture of 254 grams of alcohol of sp. gr. 0*81033 (water
1*37 per cent.) was distilled with 169 grams of benzena At 78*3^,
the results were as follows :
Weight. Sp. gr. at 074^ Water per cent.
1 27*8 grams. 0*80741 0*38
2 48*3 „ . — —
3 56-6 „ 0*80683 0*13
Kesidue... 11*5 „ — —
144*2 „
(iv) The mixture distilled consisted of 282 grams of 92*6 per
cent, alcoholi 282 grams of benzene, and 270 grams of binary (A.B.)
mixture from previous distillations :
Weight. Sp. gr. at 074'. Water per cent
1 23*6 grams. — —
2 33-6 „ 0*80822 0*62
Kesidue... 108*7 „ — —
165*9 „
(v) To 275 grams of 92*6 per cent, alcohol and 275 grams of
benzene, 379 grams of a previous distillate collected between 66*5°
and 78*3^ was added :
Weight Sp. gr. at 074'. Water per cent
1 18'6 grams. — —
2 23*7 „ 0*80789 0*50
Residue... 121*3 „ — —
163*6 „
(vi) The 92*6 per cent, alcohol was first redistilled to remove a
little water. A mixture of 275 grams of redistilled alcohol, 275
grams of benzene, and 254 grams of the fraction coUected from the
previous distillation between 67*35^ and 78*3° was distilled :
Weight. Sp. gr. at 074*. Water per cent
1 17*3 grams. — —
2 22*0 „ 0*80761 0*47
Residue... 1320 „ — —
171*3 „
(vii) The mixture distilled consisted of 289 grams of alcohol, o<m-
Digitized by VjOOQIC
ABSOLUTE ALCOHOL F&OM 8TB0NQ SPIRIT. 715
taining a little more than 0*3 per cent, of water and 192 grams of
benzene :
Weight, Sp. gr. at 074*. Water per cent.
1 12*4 grams. — —
2 24-2 „ 0-80660 about 0*1
3 67-1 „ — —
4 630 „ 0-80638 trace
Residue... 16*0 „ — —
ITW „
(viii) A mixture consisting of 99*6 per cent, alcohol, (A.B.) frac-
tions from previous distillations, and benzene, the whole estimated to
contain 370 grams of alcohol and 310 grams of benzenoi was dis-
tilled :
Weight Sp. gr. at 074^ Water per oent.
1 20-8 grams. — —
2 22-9 „ 0-80673 about Oil
3 94-4 „ — —
4 24-9 „ 0-80634 dry
Residue... 39*0 „ — —
202-0
(ix) The whole of the driest alcohol, 371*6 grams, was distilled with
249 grams of benzene :
Weight Sp. gr. at 074^
1 26-4 grama 0*80646
2 102-3 „ —
3 29*6 „ . 0-80636
Residue 26*3 „ —
183-6
(x) The alcohol from the last distillation, 182*3 grams, was distilled
with 90 grams of benzene :
Weight Sp. gr. at 074'.
1 24*6 grams. 0*80644
2 20*4 „ —
3 36*0 „ 0*80634
Residue 140 „ —
950 „
In the earlier distillations, the sp. gr. of the first portions of alcohol
that came over at 78*3^ was distinctly higher than that of the later
portions ; weaker alcohol was, in fact, being partially separated from
stronger alcohol. In later distillations, when the dehydration was
Digitized by
Google
716 PREPARATION OF ABSOLUTE ALCOHOL FROM STRONG SPIRIT.
more complete, this difference of sp. gr. was considerably less, and in
the last two it was very small, although still noticeable.
The sp. gr. of the last and best fraction was nearly the same in the
last four distillations, 0*80638, 0*80634, 0*80636, 0*80634; meao,
0-80635.
The only effect of the eighth distillation was to reduce the diSee-
ence between the sp. gr. of the first and last fractions of alcohol, and the
redistillation of this alcohol with more benzene produced no further
change. It seemed probable, therefore, that the whole of the water
was removed, and that the slight difference in sp. gr. between the
first and last fractions might be due to a very small amount of residual
benzene, which, like water, would raise the sp. gr.
As already stated, n-hexane forms mixtures of constant boiling point
with alcohol and with alcohol and water, and may be used for dehydrat-
ing alcohol. It possesses the advantage that the hezane-alcohol miztare
can be separated very readily by distillation from the dehydrated
alcohol. Again, when hezane and benzene are distilled together, the
hezane will carry down more than 10 per cent, of benzene without
any rise of temperature, a mixture of minimum boiling point — ^a few
hundredths of a degree below that of hezane — containing about 5 per
cent, of benzene, being probably formed.
It seemed possible, then, that if alcohol dehydrated with benzene
were distilled with hezane, any minute residual quantity of benzene
would be carried down in the hezane-alcohol fraction, and that the
hezane itself could be completely eliminated. If, however, any hezane
remained, the first fraction should have a lower sp. gr. than the last,
since hezane is much lighter than alcohol.
A fresh quantity of alcohol was dehydrated with benzene, aiid 127*7
grams of this alcohol (sp. gr. 0*80638 at 074°) were distUled with
128*4 grams of 9»-hezane freshly distilled over phosphoric oxide. The
temperature remained quite constant at 58*65° for a long time, and
when it changed, the rise to 78*3° was exceedingly rapid. The alcohol
was, as usual, collected in fractions, with the following results :
Weight. Sp. gr. at 074".
1 22*3 grams. 0*80629
2 13*7 „ —
3 30*0 „ 0*80627
Residue 8*8 „ —
74*8 „
The difference between the two sp. gr. was now practically within
the limits of experimental error, although the tendency seemed to he
still in the same direction. The value 0*80627 may, I think, be taken
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 717
as very close indeed to the true sp. gr. of ethyl alcohol at 0^4°. It
agrees very well indeed with the sp. gr. observed by Mendel^eff,
0*806254, and the very low values observed by Squibb appear to be
due to some chemical action of the lime on the alcohol, probably to
the presence of a little ether.
My thanks are due to the Gk>vernment Grant Committee of the
Koyal Society for a grant by means of which a great part of the
expense of this investigation and of those described in this series
of papers was defrayed.
University Gollboe,
Bbistol.
LXXIV. — The Properties of Mixtures of the Lower
Alcohols ivith Water,
By Sydney Young, D.Sc, F.K.S., and Emily C. Fortey, B.Sc.
Methyl and ethyl alcohols are commonly regarded as exceedingly hygro-
scopic substances from whicH the last traces of water can only be
removed by means of the strongest dehydrating agents. Ethyl, n-propyl,
Mopropyl;i.and ^er^ butyl alcohols, which are miscible in all proportions
with water, are known to form mixtures of constant boiling point with
that substance, distilling at lower temperatures than the pure alcohols.
In the last three cases, these mixtures, which have a definite composition
when distilled under a given pressure, have been regarded as hydrates
of the alcohols. The experiments of Konowaloff {Ann, phys, Chem,,
1881, [iii], 14, 34), however, on the vapour pressures of mixtures of
the lower alcohols with water, seem to indicate that it should be
possible to separate pure methyl alcohol from its aqueous solution by
fractional distillation, and they throw great doubt on the existence of
definite hydrates of the alcohols. KonowalofE found, in fact, that the
vapour pressures of mixtures of methyl alcohol and water are always
intermediate between those of the components, the curve showing the
relation between vapour pressure and percentage composition at
constant temperature lying fairly evenly between the extreme points.
In the case of ethyl alcohol and water, there was far greater curvature,
and the vapour pressure curve for mixtui-es of n-propyl alcohol with
water was found to resemble that representing the behaviour of a
partially miscible pair of liquids, such as isohntjl alcohol and water.
From this, it appears improbable that the formation of mixtures of
constant boiling point can be due to increasing attraction between the
molecules of the alcohols and those of water.
Digitized by
Google
718 YOUNG AND FOETEY : THE PROPEBTEBS OF
Further reasons against the assumption that a definite hydrate of
n-propyl alcohol exists were brought forward in a paper read by one of
us in conjunction with Dr. Ramsay (Proc, 1888, 4, 101), and, more
recently, Thorpe (Trans., 1897, 71, 920) has shown the incorrectness
of the usual statement that there are three definite hydrates of wo-
propyl alcohol.
A more complete investigation of the nature and properties of mix-
tures of the alcohols with water than had yet been carried out seemed
to be desirable, and the results of this investigation are described in
the present paper.
I. Methyl Alcohol and Water.
The older determinations of the boiling point and sp. gr. of methyl
alcohol are very discordant and are invariably too high. This has
been attributed to insufficienb purification and especially to the
presence of water, which, it is stated, can only be removed with great
difficulty. All the more recent observers have used dehydrating agents
such as lime, sodium, copper sulphate, or potassium carbonate.
The investigation of Dittmar and Fawsitt (Trans. Roy. Soe, Edin.,
1888, 33, 509) is generally regarded as the most complete, and their
values of the sp. gr. of the alcohol and of'^ixtures of the alcohol with
water are considered to be extremely accurate. The sp. gr. of the
alcohol observed by Bamsay and Young (Fhil. Trans., 1887, 178, 313)
agrees very well with that of Dittmar and Fawsitt. The boiling
point, calculated from Biot's formula, was 64*9°, that of Dittmar and
Fawsitt being 64*97^ but the boiling points actually observed by
Bamsay and Toung were somewhat lower, the final values being 64*7^
and 64-75°
In both cases, dehydrating agents were used, baryta, lime, and anhy-
drous copper sulphate by Dittmar and Fawsitt, sodium (repeatedly) by
Bamsay and Young.
About 550 grams of Kahlbaum's purest methyl alcohol, free from
acetone, were distilled through an 18 column Young and Thomas
dephlegmator and the sp. gr. of the first fraction (92*4 grams) and of
a later fraction (60 grams), collected after 405 grams had come over
and when^the temperature had risen 0*1°, were determined. The
distillation was completed, the temperature at the last rising to 66'4°|
and the residue in the still was then distilled from a small balb|
when the temperature rose from 69*3° to the boiling point of water.
The sp. gr. at 074° of the first fraction was 0*81003, and of the later
one 0*81017. Thus, although the original alcohol was proved to contain
a little water, the sp. gr. of the first fraction was actually lower than
that of any specimen hitherto obtained by the action of dehydrating
agents. On redistilling the first fraction through the 18 column
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 719
dephlegmator, the sp. gr. of the distillate was very slightly lower, the
value observed being 0*81000.
It appeared then that water could be completely removed by frac-
tional distillation, but to confirm this conclusion it seemed advisable
to fractionate a dilute alcohol. A mixture of 200 grams of pure
alcohol with 40 grams of water was distilled through the same
dephlegmator, and it was found that only a little water was carried
over, for the sp. gr. of the first and of a later fraction were 0*81013
and 0*81025 respectively. Redistillation of the first fraction reduced
its sp. gr. to 0*80998, the same value as before within the limits of
experimental error.
The redistilled alcohol was then distilled over sodium, when it was
found that both boiling point, 64*71^ and sp. gr. 0*81001 were
unaltered.
The observed boiling points and sp. gr. of pure methyl alcohol are
tabulated below.
The value of dp/dt at the boiling point is 29*6 mm. per degree.
Boiling points.
Temperature Corr. to
Pressure. observed. 760 mm. Sp. gr. at O*/^'.
768*75 mm. 64*95° 64*65° —
772*6 „ 65*13 64*70 —
739*5 „ 64*02 64*71 0*81003
749-5 „ 64-42 64*77 0*81000
740-7 „ 64*04 64-69 0-80998
738*75 „ 63*99 64*71 0*81001
Mean 64*70 Mean 0*81000
From these results, it is clear that, as was anticipated from Kono-
walofE's curves, fractional distillation with an efficient still-head, with-
out the use of dehydrating agents, is sufficient to eliminate the last
traces of water from dilute methyl alcohol and, indeed, it appears
doubtful whether perfectly satisfactory results can be obtained by the
use of dehydrating agents alone.
II. Ethyl Alcohol and Water.
It was noticed by Mendel^ff {Ann. Phya. Chem., 1869, [ii], 138,
230), and earlier by Sommering, that when ethyl alcohol containing
but little water, say about 2-5 per cent by weight, is distilled, the first
portions of the distillate contain rather more water than the later ones,
but they apparently did not recognise the fact that the boiling point of
the weaker alcohol is really slightly lower than that of pure alcohol.
It is well known, from results obtained on the large scale, that by no
Digitized by
Google
720 YOUNG AND FORTEY : THE PROPERTIES OF
system of fractional distillation of weak spirit can alcohol of much
more than 95 per cent, by weight be obtained.
Le Bel (Compt. rend., 1879, 88, 912) observed that alcohol contain-
ing a little water has a slightly higher vapour pressure than pure
alcohol.
Linebarger (Ghem, News, 1894, 70, 52) determined accurately the
boiling points of various mixtures of alcohol and water, his concentra-
tions being expressed in parts of water in 100 of alcohol. He observed
a minimum boiling point with about 4*5 parts of water, the actual
temperature given for 4*497 parts of water being 77*990^ under a
pressure of 7^6 mm., that of alcohol, with 0*241 part of water, being
78*091^ and with 9*222 being 78*088. This would give a difference
of a little more than 01° between the boiling points of pure alcohol
and of the mixture of constant boiling point.
Quite recently, Noyes and Warfel {J, Amer, Chem, Soc,, 1901, 23,
463) have determined the boiling point curve for aqueous alcohol, and
find that 96 percent, alcohol (97*45 per cent, by volume) has a minimum
boiling point 78*174°, whilst that of absolute alcohol, and also of
90*7 per cent, alcohol, is 78*30. This would give a difference of 0*126°
between the two boiling points.
This difference is so small that it seemed hopeless to attempt a
separation of either pure alcohol or the mixture of constant boiling
point by fractional distillation of spirit containing more than about
96 per cent, of alcohol, but it was repeatedly observed that a partial
separation could be effected. As the difference in boiling point between
the mixture of constant boiling point and water is considerable, it
seemed not impossible, especially as similar mixtures of water with
other alcohols were very easDy obtained, that fractional distillation of
weak spirit would yield, not only pure water, but also the mixture of
minimum boiling point without much difficulty.
The question, however, whether the two components of a mixture —
either pure substances or mixtures of constant boiling point — can
be easily separated by fractional distillation does not depend solely, or
indeed chiefly, on the difference between the boiling points. The form
of the curve representing the relation between the boiling points and
molecular composition of various mixtures must be taken into account,
and in this case the curve is exceedingly flat near the minimum boiling
point. This fact, which is more fully discussed in the paper on frac-
tional distillation as a method of quantitative analysis (this vol.,
p. 752), explains the difficulty actually experienced in separating the
mixture of constant boiling point.
As a matter of fact, it was found on distilling a quantity of 92*6
per cent, alcohol that, although the strength rose rapidly at first,
the improvement became slower and slower, and even after seven
Digitized by
Google
MIXTURES OP THE LOWER ALCOHOLS WITH WATER. 721
distillations with the 18-column dephlegmator^ the percentage of
alcohol in the first fraction did not exceed 95*3 per cent, by weight,
although, as will be seen later, the mixture of minimum boiling point
really contains 95*57 per cent, of alcohol.
That the final point had not been reached was indicated by the fact
that even in the seventh fractionation the strength of the alcohol
steadily, although slowly, diminished as the distillation proceeded.
It was therefore hopeless to attempt to separate the mixture of
constant boiling point, or, indeed, to determine its exact composition
by fractional distillation. Moreover, the curve representing the relation
between the boiling points and the composition of mixtures of ethyl
alcohol and water is of such a form that, although the minimum boil-
ing point is easily readable, only a rough approximation to the com-
position can be obtained.
With regard to the difference between the boiling points of pure
alcohol and of the mixture, the results of Linebarger and of Noyes
and Warf el are not in perfect agreement, but it may be concluded that
this difference lies between 0*1° and 0*15°.
Our own experiments were not carried out with a view to the
accurate determination of this difference, and a thermometer, graduated
in whole degrees, was, in fact, employed ; yet, owing to the very large
number of determinations of the boiling points of alcohol-water mix-
tures that were made at various times, it was found possible to construct
a curve which would give a fair approximation to the value under dis-
cussion. The mean of the eight observations nearest to the point of
minimum temperature given in the table below would show a difference
of 0*2°, although the value we should have been inclined to adopt,
from our curve was 0*16°, which, at any rate, does not differ seriously
from the more accurate determinations referred to.
Taking the boiling point of pure alcohol as 78*3°, the observed boil-
ing points of mixtures of nearly constant boiling point were as follows :
Temperature
Sp.gr.
Pressure.
observed.
corr. to 760 mm
0-8196
758-9 1
nam.
78-00°
78-04°
0-8185
762-0
78-22
78-16
0-8194
763-6
78-18
78-06
0-8193
763-6
78-27
78-16
0-8195
756-4
77-95
78-07
0-8195
766-4
77-97
7809
0-8195
756-4
77-98
78-10
0-8173
762-3
77-85
7810
Mean 7810
The value of dp/dt at the boiling point == 30*2 mm. per degree.
Digitized by
Google
722 YOUNG AND FORTEY :* THE PROPERTIES OF
Composition of ths Mixture of ConstarU Boiling Point, — In addition
to systematic fractionation of the alcohol, numerous single distillations
of alcohol of various strengths were carried out with the 18 column
dephlegmator, and the sp. gr. at 0^/4° of the first and last fractions
were in each case determined. The results are shown in the diagram,
the sp. gr. being plotted against the total weights of distillate.
Thus, in No. X, three fractions were collected, weighing respectively
23*6, 73*4, and 27*6 grams. The sp. gr. of the first and third were
determined and were taken to represent those of the distillate at the
moments when 23*6/2 and when 23*6 + 73*4 + 27*6/2 grams respectively
had come over.
150
llOO
I-
60
^0-806 0-810 0*814 0'818 0822 0*826
Specific gravity.
As in the majority of cases there were only two determinations of
sp. gr., a straight line has been drawn through the points, although
the actual results should correctly be represented by a slightly curved
line.
Starting from the left of the diagram it is evident that the curve,
if obtained, would at first be represented by a vertical straight Ime
showing constancy of composition during distillation ; that the curves
then actually show an increasing slope towards the left, reaching a
maximum in the neighbourhood of sp. gr. 0*814 ; that the slope again
gradually diminishes, vanishes at about sp. gr. 0*8194, then changes
its direction and again shows a gradual increase. The actual experi-
ments have not been continued, but it is clear that the slope would
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 723
again pass through a mazimum, decrease and finally vanish at sp. gr.
0*99987, the liquid being then pure water.
The first two points at which the curves would become vertical
straight lines represent the sp. gr. of pure alcohol and of the mixture
of minimum boiling point. The method of preparation of pure alcohol
has been described in the previous paper and its sp. gr. is there
referred to.
By distilling various known mixtures of alcohol and water and
observing whether the curves sloped to the right or left, it would be
possible to get a closer and closer approximation to the mixture of
minimum boiling point which distils without change of composition. It
will be seen that curve X slopes to the left and curve XI to the right,
and that the sp. gr. of the mixture of constant boiling point must
therefore be between those (0*81936 and 0*81946) of the first fractions
in these distillations.
The lines are of so nearly equal slope that it may be assumed that
the required sp. gr. is 0*81941, the mean of the other two. If the
two lines are produced they meet at a point between 0*81941 and
0*81942.
The sp. gr. of the mixture of constant boiling point of ethyl alcohol
and water is therefore 0*81941, and the percentage of alcohol by weight
that it contains is, according to Mendel^ff's data, 95*57.
III. U'Propyl Alcohol and Water,
The fact that the addition of small quantities of water to n- propyl
alcohol has the effect of lowering the boiling point was stated in 1869
by Chancel {Compt, rend,, 68, 662), who, from the observation
that a particular mixture distilled at a constant temperature without
change of composition, concluded that a definite hydrate, CgHgOjHjO,
was formed.
The statement that this hydrate exists is to be found in some text-
books, but as a rule the point is not mentioned.
The vapour pressures of mixtures of n-propyl alcohol with water
were investigated by Konowaloff {loo. cU,), who states that at 88° a
mixture containing 75 per cent, by weight of the alcohol has a
maximum vapour pressure, but that the composition of the mixture
which exerts the maximum vapour pressure differs slightly at different
temperatures. Erom these results, he concludes that no hydrate of
propyl alcohol exists, and this conclusion seemed to be supported by
the fact observed by Chancel that anhydrous potassium carbonate
easily abstracts water from the mixture of constant boiling point.
Further evidence confirming Konowaloff's view was brought for-
ward by Ramsay and Young, in 1888, in the paper already referred to.
Digitized by
Google
724 YOUNG AND FORTEY : THE PROPERTIES OF
The yapour pressures were determined by both the dynamical (Ram-
say and Toung) method and the statical (barometer tube) method.
The differences were slight, but the pressures by the statical method
were invariably a little higher. If the substance were a hydrate, or
if the composition of the mixture remained constant at all tempera-
tures, no such difEerence should be observed ; but if the composition
varied with the temperature, partial fractionation would take place by
the dynamical method, somewhat higher temperatures being registered
than those corresponding to the original mixture of constant boiling
point obtained by distillation under the ordinary pressure.
KonowalofE states that the mixture which exerts a maximum vapour
pressure is slightly poorer in alcohol at low temperatures than at high.
Experiments were carried out to find whether the composition of the
mixtures boiling constantly under different pressures showed a similar
variation. It was found that when the mixture of constant boiling
point, obtained by distillation under the ordinary pressure, was redis-
tilled under 198*7 mm. pressure, a residue of nearly pure propyl
alcohol was left. Konowaloff's result was thus fully confirmed.
Yapour density determinations above 130° in a modified Hofmann's
apparatus, in which pressure and volume could be altered at will, gave
the value 18*14, and, assuming that there is no combination in the
state of vapour^ the composition corresponding to this density would
be:
n-Propyl alcohol 71*95
Water 28*06
10000
Again, if no contraction took place on mixing liquid propyl alcohol
and water, the volume of a gram of a mixture of this composition would
be 1*1587 c.c. at 0°. The volume of a gram of the mixture of constant
boiling point, obtained by distillation under the ordinary pressure, was
found to be 1*1362 c.c. at 0^ If the assumption that there is no com-
bination in the state of vapour is correct, there must be a contraction
on mixing the liquids of 0*0225 c.c. per gram of the mixture, equiv-
alent to 1-942 per cent, by volume ; it was known that some contrac-
tion, at any rate, does occur.
In any case, it was clear that since the composition calculated from
the sp. gr. of the liquid would differ but little from that derived from
the vapour density, there could not be more than a very small amount
of combination in the state of vapour.
The conclusion that the hydrate of propyl alcohol does not exist was
not accepted at the time, and the paper did not appear in the
Transactions. Further confirmatory evidence has now been obtained.
If the hydrate of propyl alcohol existed at the ordinary temperature,
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 725
it is certain that it would not be less stable at lower temperatures ;
yet it was found possible to separate about two-thirds of the water by
cooling the mixture of constant boiling point to - 40°. Partial crys-
tallisation took place, and the sp. gr. of the residual liquid was
0*84145, corresponding to 9*3 per cent, of water, as against 0*88004,
corresponding to 28*3 per cent.
The propyl alcohol employed in this investigation was obtained
from Kahlbaum ; it was carefully fractionated, and the boiling point
and sp. gr. were determined, with the following results :
Temperature
Pressure. observed. corr. to 760 mm. Sp. gr. at O*/*".
771*2 mm. 97*70° 97*31° 0*81931
759*5 „ 97-30 97*32 0*81934
The value of dp/dt at the boiling points 28*85 mm. per degree.
Afterwards the propyl alcohol was recovered from the mixture of
constant boiling point by distillation with beuzene after a portion of
the water had been removed by means of potassium carbonate. Both
the boiling point and the sp. gr. were slightly lowered by this treat-
ment:
Temperature
Preswure.
oSserved.
corr. to 760 mm.
Sp. gr. at O'H'.
767-55 mm.
97-44°
97-18°
0-81923
763-6 „
97-32
97-20
These values are taken as correct.
Mixture of Constant Boiling Paint of n-Propyl Alcohol and Water, —
In order to find whether the mixture of constant boiling point could
be obtained with quite constant composition by distillation, mixtures
richer both in alcohol and in water were fractionated, and the boil-
ing points and sp. gr. of the first fractions determined.
From mixtures richer in alcohol :
TempeTatnre
Vieuxae.
observed.
corr. to 760 :
mm.
Sp. gr. at 074'
756-9 mm.
87-59°
87-70°
0-88000
759-7 „
87-69
87-70
0-88002
744-2 „
87-17
87-72
0-88004
mixtures richer in water :
747-6 mm.
87-27°
87-70°
0-88008
763-2 „
87-88
87-77
Me
—
Mean 87-T2
lan 0-88003
The value of dp/dt at the boiling point » 28*7 mm. per degree.
▼OL. LXXXI. 3 C
Digitized by
Google
726 YOUNa AND rORTEY : THE PBOt^ERTtES Ol"
These results agree well with those previously obtained bj Bamsay
and Young, namely, boiling point, 8775° ; sp. gr. at 0°/4% 0-88015.
Spedfio QrcmtieB of Mixtwe» qf n-Frapt/l Alcohol and WcUer. —
Known weights of pure propyl alcohol and water were mixed together,
and the sp. gr. were determined with the following results :
of alcohol bjr
weight
Sp. gr. at 0"/4*.
94-95
0-83203
89-97
0-84307
84-87
0'86362
79-96
0-86360
74-93
0-87365
71-69
.0-88004
The last sp. gr. in this table is practically identical with that of the
mixture of constant boiling point obtained by distillation ; the per-
centage composition may therefore be taken as :
^-Propyl alcohol 71*69
Water 28-31
10000
This agrees very well indeed with the necessarily less accurate re-
suit, calculated from the vapour density, observed by Kamsay and
Young :
w-Propyl alcohol 71*95
Water 28-05
10000
and the conclusion previously arrived at, that there is no combination
in the state of vapour, is fully confirmed. The contraction that takes
place when this and other alcohols are mixed with water will be
referred to later on.
IV. iBoPropi/l Alcohol and Waters
Threedefinitehydrate8ofwopropylalcohol,2CgHgO,HjO,3CjH80,2HjO,
and dOjHgOyHgO are stated in ail text^books of organic chemistry to
exist, and a fourth hydrate, O^^fi^Hfi, has also been described by
Ruhemann and Carnegie (Trans., 1888, 63, 427).
Thorpe, however (Trans., 1897, 71, 920), has brought forward
evidence which renders the existence of any of these hydrates very
doubtful. Thorpe purified his alcohol by long continued treatment
with anhydrous copper sulphate, fractionating the alcohol from time
to time, and finally distilled it after boiling with lime for some time
in a flask provided with a reflux condenser. The sp. gr. of two fractions
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 727
were then found to be identical, 0*7982 at 47^°, and the boiling point
was 80*9 — 81*4° at 738 mm. The boiling point under normal pressure
adopted was 82 'P.
Thorpe then made mixtures of the purified alcohol and water in the
proportions required to form each of the four hydrates. Three of
these mixtures were allowed to evaporate over sulphuric acid until the
bulk of liquid was reduced to about one^^half, and through the fourth
a current of air was passed until the bulk was reduced to about the
same extent. In all cases, the residual alcohol was considerably
stronger than at first, as shown by the diminution in the sp. gr. From
these results, Thorpe concludes " that the so-called hydrates are not
definite chemical compounds, but are merely mixtures of the alcohol
and water/'
The four ** hydrates " are described by their discoverers as having
the following boiling points, taking the order already given : (1) about
80°, (2) 78—80°, (3) 81°, (4) 78—79°.
As these boiling points are all lower than that of the pure alcohol, it
seemed to us to be probable that Mopropyl alcohol really behaves like
n^propyl alcohol in forming with water a mixture of minimum boiling
point which would distil without change of composition.
About 200 grams of tsopropyl alcohol, obtained from Kahlbaum,
were distilled through a five-column "evaporator" still-head. It
contained, besides water, a very small amount of impurity of high
boiling point and a rather larger amount of one of low boiling point.
After eliminating these, relatively large quantities of two liquids, A
and Bf were obtained, boiling constantly at 80*4° and 82*4° respectively.
The first proved to be the mixture of constant boiling point of alcohol
and water ; the second was nearly pure alcohol.
After the boiling points had become nearly constant, the sp. gr. of
these substances were determined at the end of each fractionation with
the following results :
Sp. gr. at074^
Ko. of fractionation* A. B»
3 — 0-80199
4 — 0*80159
5 0*83331 0*80144
6 0-83356 —
7 0*83361 —
It will be seen that the sp. gr. of A shows a gradual rise, that of JS
a gradual fall, both tending finally to become constant.
The fraction B was boiled for 2} hours with barium oxide in a
reflux apparatus and was then distilled. Its sp. gr. at 0°/4° was now
0*80137. This agrees well with Thorpe's value, which, reduced by
means of Zander's formula, would be 0*8014 at 0°/4°.
3 c 2
Digitized by VjOOQIC
728 YOUNG AND PORTEY : THE PROPERTIES OF
It will be seen from these results that although the boiling points
of the two substances differ by only two degrees it was found possible
to separate the alcohol in a very nearly pure state, and it was later
found that the composition of A was identical with that of the mixture
of constant boiling point obtained by the distillation of more dilute
Mopropyl alcohol.
The alcohol was subsequently recovered from the mixture of con-
*stant boiling point by distillation with benzene, and it was evidently
obtained still drier by this process than by heating with barium
oxide, 'for the sp. gr. at 074°, 0-80120, was slightly lower, and the
boiling point, 82 -44^ was slightly higher. These values are taken
as correct.
Seventy grams of the mixture of constant boiling point {A) and 20
^rams of water were distilled together through the « evaporator" still-
head, antfjthPw^. gr. of the first fraction was found to be 0 '83434 at
0°/4° That port'ion^bflFe^narllalu uhicli uiZu^mi^^ ^^^^ ^°*
stant temperature was redistilled and the sp. gr. of th?|P* ''*^^*^'^
was found to be 0-83361, which is identical with that of -4.^|^^® .
composition of the distillate was the same whether the origir^^ ^
ture was richer in alcohol or in water.
The sp. gr. of mixtures of t«opropyl alcohol and water have
been determined by Thorpe {loo, dt,) at 15°/15°. The sp.
mixture of constant boiliDg point at 14*674% as determined
0-82116, which at 16715° would become 0-82153.
From Thorpe's tables, the percentage composition corresponding to
this sp. gr. would be
Alcohol 88-15
Water 11-85
ngiriw
lave alreSey J
p. gr. of tho^
i by us, was \
100-00
A mixture of this composition was then made, using the alcohol of
sp. gr. 0-80137 at 074°, and its sp. gr. at 074° was found to be
0-83297, corresponding to 0-82091 at 15715°.
Assuming that the alcohol used for the mixture was dry, the com-
position corresponding to the sp. gr. 0*82153 at 15715° would be
Alcohol 87-9 /
Water 12-1
100-0 /
or allowing for the traces of water present the percentage o^tm^^^^
would be about 87*85. j^ \
It will be seen that the percentage of water ia the mix |Sme ^^'
stant boiling point is much lower than in the case V. of two fra«*^^^^'*
ut is higher than in that of ethyl alcohol.
Digitized by
Google
MIXTURES OP THE LOWER AXCOHOLS WITH WATER.
729
The following are the observed boiling points of taopropyl alcohol
and of its mixture of constant boiling point with water ; the alcohol
was that dried by distillation with benzene :
AlcohoL
Temperatnre
Pressure.
761*5 mm.
771-6 .,
observed.
82-46°
82-85
corr. to 760 mm.
82-42°
82-46
Mean 82*44
Mixture qf Constant Boiling Point with Water,
From
excess of water.
From
excess of alcohol.
Temperature
Tempeiatnro
*"
corr. to
^
corr. to
Pleasure.
observed.
760 mm.
Fressnre.
observed.
760 mm.
746-7 mm.
79-96°
80-41°
751-86 mm.
8011°
80-38°
764-75 „
80-56
80-39
763-6 „
80-61
80-39
761-75 „
80-46
80-40
Mean 80-40
732-0 „
79-43
80-37
743-4 „
79-87
80-43
760-7 „
80-36
80-34
770-25 „
80-66
80-31
753-9 „
80-11
80-32
Mean 80*37
Value of d/p\dt at the boiling point : Alcohol, 30*0 mm. per degree ;
alcohol-water mixture, 29*7 mm. per degree.
Y. \AV^,BvMfl Alcohol and Water,
It was stated by Butleroff {Annalen, 1872, 162, 229) that ^-^.butyl
alcohol, which, like those already considered in this paper, is miscible
with water in all proportions, forms a definite hydrate of the formula
2G^B^fi,B.fi, The hydrate was described as a liquid boiling at 80°
and having the sp. gr. 0*8276 at 0°.
Two hundred and fifty grams of ^er^. butyl alcohol were obtained
from Kahlbaum. At the temperature of the room, 22°, the substance
consisted of fine, large crystals with a good deal of liquid. The liquid
was poured off aud fractionated five times with the 5-column " evapor-
ator " still-head and a good quantity was obtained boiling quite con-
stantly at 82*45° under a pressure of 758*4 mm., corresponding to
82*5° at 760 mm. The melting point of the best fraction was 25*25°.
Digitized by
Google
730 YOUNG AND FORTBY : THE PROPERTIES OP
Meanwhile the atmospheric temperature gradually rose from about
22° to above 25°, and small additional quantities of liquid were formed
by the partial fusion of the crystals, and were drained off from time
to time. Finally, about half the total quantity of residual solid was
melted and poured off ; its melting point was 25*20°, while that of the
residue was 25*41°.
As it was clear that fractional crystallisation gave better results
than fractional distillation, the former process was applied to the whole
quantity of nearly pure alcohol, and the melting point was finally
obtained as high as 25*43°. In carrying out the fractional crystallisa-
tion, the conditions were peculiarly favourable, for the temperature of
the room, after having fallen again, gradually rose in the course of
about a week from about 22° to about 25*5°.
Later on, the alcohol was recovered from the mixture of constant
boiling point, which it was found to form with water, by distillation
with benzene. As in the case of ethyl, n-propyl, and tsopropyl alcohols,
the purest specimen was obtained by this method, the melting point
being 25*53° as against 26*43°, and the boiling point 82*60° under a
pressure of 761*4 mm., corresponding to^82'55° at 760 mm,, as against
82*5° at normal pressure.
The sp. gr. could not be determined at 0°, owing to solidification ;
that of the recrystallised specimen was found to be 0*78560 at 20°/4°,
whilst that of the specimen dried by distillation with benzene was
0*78553 at 20°/4°
The sp. gr. of the pure alcohol and of mixtures with water were
actually determined at temperatures as near as convenient to 20° and
25°, and the values obtained were corrected to these temperatures on
the assumption that the sp. gr. is a linear function of the temperature
over this small range :
Sp, gr, of Mixtures of teri.Butyl Alcohol and Water.
Percentage of alcohol,
by weight.
100
"ft
20°.
0-78563
5'- -"
26*.
0-78056
97-36
0-79128
0-78663
94-24
0-79878
0-79415
90-68
0-80718
0-80268
8600
0-81820
0-81364
80-42
0-83146
0-82703
73-26
0-84832
0-84406
Mixture qf Constant Boiling Point with Water. — Numerous distil-
lations of alcohol-water mixtures were carried out, in some cases the
alcohol, in others the water, being in excess. The results are given in
t\^e f pUo^^ing table :
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 731
From mixtures richer in water :
Temperature Specific gravity
, « , , * ,
Pressure. observed, corr. to 760 mm. at 074'. at 2574'.
757-36 mm. 79-86° 7994° — —
762-2 „ 80-00 79-93 — -^
758-15 „ 79-84 79-90 — -^
758-7 „ 79-88 79-92 — —
757-5 „ 79-83 79-91 — —
746-45 „ 79-46 79-92 — —
747-7 „ 79-51 79-93 083046 0-80831
748-45 „ 79-53 7992 ^ ~
Mean 79-92
From mixtures richer in alcohol :
At 2074*
767-5 mm. 7983° 79-91° — —
748-45 „ 79-53 79-92 0-83041 —
758-1 „ 79-80 79-86 0-83041 0-81275
Mean 79-90
The value of d/p\d^ at the boiling point was 29*7 mm. per degree.
The composition of the mixture of constant boiling point was
ascertained in the following manner. Ourves were drawn showing the
relation between the percentage molecular composition and the sp. gr.
at 20° and 25° respectively. From each of these curves, the composition
was read off corresponding to the sp. gr. actually observed in the case
of the mixture of constant boiling point.
The percentage molecular composition was thus found to be
Alcohol. Water.
(a) From sp. gr. at 20° (distillate from mixture richer
in alcohol) 64-56 35-44
(6) From sp. gr. at 25° (distillate from mixture richer
in water) 64*62 35-38
Mean 64-59 36-41
The percentage composition by weight calculated from the mean is
Alcohol 88-24
Water 21-76
100-00
(er^.Butyl alcohol shows great similarity in many respects to i«opropyl
alcohol, as will be seen from the following table ;
Digitized by
Google
782 YOUNG AND FORTBY : I'HE PROPERTIKS OF
tertJBvLtyl alcohol. isoPropyl alcohoL
Boiling point at 760 mm 82-55° 82-44°
Sp. gr. at 2074° 0-78553 about 0-7859
Hixture of Gonstcmt Boiling Point with Water,
Boiling point at 760 mm 79-91° 8037°
Sp. gp. at 0°/4° 0-83043 083361
Percentage of alcohol by weight 88*24 87*90
VI. i&oButyl Alcohol and W<tter.
As is well known, this alcohol is miscible with water only within
limits, and the behaviour of a mixture with water, on distillation, is
similar to that of any other partially miscible pair of liquids.
About 350 grams of wobutyl alcohol, obtained from Kahlbaum, were
distilled through the 18-column dephlegmator. After four fractiona-
tions, a fair quantity was obtained , with constant boiling point, but
the sp. gr. of the fractions showed a slight rise. On repeating the
fractionation twice, the sp. gr. fell a little on the whole, and there was
still a slight rise from fraction to fraction (0-81705 to 0*81723 at 0°/4°).
The fractions within these limits were mixed together and redistilled,
when the sp. gr. at 0°/4° was 081698; at 16-35°/4°, 0-80459.
The boiling point was determined at various times, and the follow-
ing values were obtained :
Temperature
Preasnra.
observed.
corr. to 760 mm.
746-2 mm.
107-51°
108-00°
742-0 „
107-44
10808
738-75 „
107-31
108-07
763-26 „
107-86
10810
763-6 „
107-86
10809
756-4 „
107-94
108-07
766-4 „ -
108-26
108-03
Mean 10806
The value of dp/dt at the boiling point was taken to be 28*0 mm.
per degree.
Compoaition and Boiling Point oftlie Alcohol-Water Mixture qf Con-
stant BoUing Point, — A mixture of Mobutyl alcohol and water of
known composition was distilled and the weight of distillate below the
middle temperature was ascertained in ordec to find the composition
of the mixture of constant boiling point by the method described in a
separate paper (p. 752). The water was in excess :
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER. 733
Oompofiition of mixture
^ . . • «, . , , of constant boiling point.
Compoaition of Weight below ^ ^ ^
mixture taken. middle point. Uncorrected. Corrected.
Alcohol 49-6 Observed 73'95 Alcohol 67*1 66*8
Water 447 Corrected 74-26 Water 32-9 332
94-3 100-0 1000
Boiling Points.
Temperature
Pressure. observed, corr. to760mm.
T3» _ . X • 1- • X f 764-5 mm. 89-95° 89-80
From mixture richer m water... \ '"* *; "^ * °^ ^^ "^ "
1766-0 „ 90 03 89-82
alcohol. 757-75 „ 89-75 8983
Mean 89*82
dp/dt at the boiling point » 28*4 mm. per degree.
The sp. gr. of a series of mixtures of isohutjl alcohol and water
were determined at 0°/4° The results are given in the table below :
Percentage of iwbutyl alcohol by weight. Sp. gr. at 074°.
97-72 0-82251
95-09 0-82823
91-79 0-83492
88-53 0-84125
86-76 0-84470
84-81 0-84829
The last mixture was just saturated with water at 0°.
YII. isoAmyl Alcohol and Water,
About 450 grams of amyl alcohol, obtained from Kahlbaum, were
fractionated with the 18-colamn dephlegmator, benzene being added
at first to facilitate the removal of lower homologues.
It is well known that amyl alcohol obtained by fermentation is a
mixture of the isomeric tsoamyl alcohol, (CH3)2CH*CH2*CH2*OH, and
active amyl alcohol, CH3-OHj"CH(CH8)-OH2-OH.
The fractionation is not yet completed, but a sufficient quantity of
one of the isomerides, presumably i^oamyl alcohol, boiling very nearly
constantly at 132*05^, has been obtained. The remainder comes over
at lower temperatures.
Composition qf Mixture qf Constant Boiling Point with WaUer, — ^From
mixture richer in water :
Digitized by
Google
734
YOUNG AND FORTBY : THE PROPERTIES OF
Composition of mixture
of constaxit boiling point
Mixture taken. middle point.
Uncoriected.
Corrected,
Alcohol 38-8 Observed 76-4
Alcohol 50-8
50-5
Water 69-5 Corrected 76 9
Water 49-2
49-5
1083
1000
1000
From mixture richer in alcohol :
Alcohol 68-3 Observed 85-65
Alcohol 601
50-3
Water 42-7 Corrected 85-95
Water 49 9
49-7
1110
100-0
100-0
The mean of the two corrected values, which agree very well
together, may be taken as correct :
Alcohol ...
50-4
Water ...
Boiling Poinh
• 49-6
1000
Temperature
Pressure.
observed.
corr. to 760 mm.
755*85 mm.
950°
95-15°
756-3 „
950
96-15
Mean 9516
YIII. GeneraltscUitms from Eesuils.
In a previous paper by one of us (p. 708), it is pointed out that
as the molecular weights increase the properties of the alcohols diverge
from those of water and approach those of the corresponding paraffins.
Since the higher alcohols are practically insoluble in water, whilst
the lowest of them, methyl alcohol, dissolves in water in all propor-
tions with considerable contraction and evolution of heat, it may be
inferred that the attraction of the alcohol molecules for those of water
diminishes with rise of molecular weight, and it is natural to assume
that the change would be a gradual one.
This conclusion is confirmed by the results ^veu in the following
table ;
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER.
785
Specific Gravitiea, BoUing Points, and Composition of Mixtures qf
ConstwU Boiling Point of the Alcohols with Water,
Boiling point (760 mm.).
Sp. gr. 074*.
Percentage of
alcohol in
mixture.
Alcohol.
Mixture.
A.
Alcohol.
Mixture.
weight.
95-57
87-90
88*24
7169
}66-80
50-40
Mole-
cular.
Methyl alcohol ...
Ethyl „ ...
woPropyl „ ...
ter^.Butyl „ ...
ii-Propyl „ ...
iwButyl „ ...
iwAmyl „
64-70
78-30
82-44
82-55
97-19
108-06
18205
7?16
80-87
79-91
87-72
89-82
96-16
016
2-07
2-64
9-47
18-24
86-90
0-81000
0-80625
0-83120
0-81923
0-81698
0-81941
0-83861
0-83043
0-88004
r two
Xlayers
it
89-48
68-64
64-59
48 17
32'86
17*21
Comparing the three Dormal primary alcohols, methyl, ethyl, and
propyl, it will be seen that the first does not form a mixture of con*
stant boiling point with water at all, whilst the difference in boiling
point between the alcohol and the mixture of constant boiling point is
much smaller in the case of ethyl than of propyl alcohol. Again, the
percentage of alcohol in the mixture of constant boiling point is much
higher in the case of ethyl than of 7i~propyl alcohol.
The table further brings out the influence of constitution on the
properties of the alcohols.
That the boiling point of a paraffin or alkyl derivative containing an
iso-group is lower than that of the normal isomeride is well known, as
is also the fact that a tertiary alcohol boils at a much lower tempera^
ture than the isomeric primary compound.
The alcohols are arranged in the. table in the order of their boiling
points, and it will be seen that the same order is followed both as re*
gards the difference in boiling point between the alcohol and its water-
mixture of constant boiling point and also as regards the molecular
percentage of alcohol in the mixture.
IX. Contraction on Mixing unth Water.
The contraction that takes place on mixing the various alcohols
with ¥rater was calculated from the sp. gr. of a series of mixtures of
known composition in each case. ^
The required data for methyl, ethyl, and wopropyl alcohols wer©
already available [methyl alcohol, I)ittmar and Fawsitt {loc. cit.) ;
Digitized by
Google
736
TOUNG AND FORTEY : THE PROPERTIES OP
ethyl alcohol, Mendel^eff {loo, cit) ; t^opropyl alcohol, Thorpe {loc, cit.)],
and determinations were therefore only made with n-propyl, tert.hvitj],
and t«obutyl alcohols.
In the case of methyl and of ethyl alcohol, data are given at 0^/4^, but
for i^opropyl alcohol only at 15^/15°. In order to find whether the
contraction differed much at different temperatures, the data given by
Dittmar and Fawsitt for methyl alcohol at 15*67^° were also made
use of. It will be seen from the following table that the contractions
are nearly the same at the two temperatures, and it was therefore not
a matter of importance that the comparison had to be made at different
temperatures.
The plan adopted was to calculate the volumes of a gram from
the observed sp. gr., and also the theoretical volumes of a gram,
supposing that no contraction occurred on mixing (see p. 737). The
percentage contractions are given in the following table :
Percentage Cantraetion,
Mole-
cular per-
Methyl alcohol.
Ethyl
alcohol.
taoPropyl
alcohol.
teH.Butyl alcohol.
n-Propyl
alcohol.
ifoBatYl
alcohol.
centage
of
alcohol.
15 ^e'.
0'.
0'.
15".
20'.
25".
0'.
0*.
100
0 00
000
0-00
0 00
0 00
0-00
0-00
000
95
0-44
0-44
0-38
013
0-06
0-07
0-27
0-15
90
0-88
0-84
0-74
0-36
0-17
0-19
0-43
0*28
85
1-28
1-28
1-06
0-57
0-30
0-32
0-62
0-40
80
1-68
1-68
1-38
0-79
0-41
0-47
0-80
0-61
75
2-04
2 04
1-69
101
0-55
0-59
0-97
0-60
70
2-40
2-39
1-99
1-28
0-71
074
1-12
0-72
65
2-69
2-70
2-29
1-43
0-85
0-90
1-25
0-81
60
2-97
2-98
2-56
1-65
102
1-05
1-42
0-90
55
3-19
3-22
2-83
1-85
1-21
1-25
1-57
—
50
3-42
8*45
3 09
2-05
1-40
1-42
1-71
—
45
3-54
3-60
3-34
2-20
1-60
1-63
1-85
—
It will be noticed that of the six alcohols examined, four are primary
— methyl, ethyl, and n-propyl alcohols being derived from normal
paraffins, t«obutyl alcohol from an t^oparaffin ; of the other two, iso-
propyl alcohol is a secondary alcohol derived from a normal paraffin,
whilst tert,hvLtjl alcohol is a tertiary alcohol derived from an iso-
paraffin.
In studying the properties of these alcohols, account must therefore
be taken both of the constitution of the paraffins from which they are
derived, and also of the position of the hydroxyl group. Of these two
factors, the latter seems to have the greater influence so far as con-
traction on mixing with water is concerned, for, ¥nth the four primary
Digitized by
Google
MIXTURES OF THE LOWER ALCOHOLS WITH WATER.
737
71 -Propyl woButyl
alcohol. alcohol.
11
-1
CQ
'«« 0» "^ 0» MS O TJ« 00 c^ . , ,
<N r-t r^ O O O Oi C» 00
C<9 C4 (N Cq Cq (N rH .p4 rl 1 1 1
33
t^cq»o«ocoeOiAOso><Noo»o
Ot^eOOJiOOiOOiCOl-^OirH
(Nr-lr-lOO©0i0000t^«0«0
9JCqO^<NOacqrlrHrHrHrHr-l
t^\OCOO®Oi-lOrHOOOOO
o<^ooc4kOO«Gqot^ooa)o
<Ni-«OOOaoooot>.«)u3'^<<*«
<J1 W (N <N rH .-^ r-. rH r-. rH rH .-^
1 alcohol
o '
II
r-«»000«0<NCOO>0>COOt^"^
rHtN.COO»iaOTI«aO(N^(Or«
<N<NCq(N<NCq(NCq(N(NC^O^
«
00t-t^<0>O»0'*C0Cqr-lO0>
W (N <N <N CI « (N (N (N « (N rH
i
O
<N
31
C404e4C<1C4Cq<NC4<N(N<NC^
o 1
O00t^r-lC000t^OWiiH00»O
oooocooo<N»ooOf-ic^eocqrH
(NGqcqC4CqcqC4C>4GqCq<N74
i-Jr-lr-lr^t-Hf-HrHi-lf-Hr-lr-Ji-H
MoPropyl
alcohol.
r-i
11
l«<N00O0JC00^»0{NC0'^00
«Or-l«OC<l«Oi-«»OOOi-ICO'^-^*«
wcq9i(NOiic>^o^cqo^cq<N,-H
-1
o 1
»o«ocoo>ooooo.-ioot^o
«Da6C<l"^t>.OOOr-1i-HOOSOO
«0»0»0"^90<MCqi-«OOSt>.«0
<N (N <N <N <N <N « <N <N 7. .ji rH
'3
1
31
eouaeooooicoookOtooQOt^
OW>OT|«OOCq»ftOOOCq(N(N
C<JCqcqcqcqC<ICqr^r-«.-lrHr-i
^1
o 1
CO00C^00r-i.-H00rH.-i»«OOS
OO.-Hi-lC^Cs|rHi-lO00»C0
<«rso<Ni-(00>ootN.cO'^OOC4
{NCqCqCqC^rHrHrHrHrlf-l^
1
f
li
eo«ococ<i»0'^Oioo-«t«»o.-Hco
<NCqC<IO«CqrHrHrir-lr-ti-HrH
it
eoMOr^coowaoiOf-H-^co
TtiCNOt^tACOOCOIOOQOaO
OOCqrHO>00«^«OTjiCO(NrHOa
li
lArHIO^OrlOiOOOObOOr^
C4CqC4C<l(N(N<Nr-lr-lr-ir-4rl
o 1
iocococD'^eooao»o<^aoa»
^OCOOt^-^rHOO^rHOO-^iH
»«TjiCOr-»OOSt^«OlOeOOilrH
s
cular per-
centage
of
1
J
OtOOtOOlOOlOOlAOtO
oa»a»oooot^t^<D<oioto<<i
Digitized by
Google
738 MIXTURES OF THE LOWER ALCOHOLS WITH WATER.
alcohols, the percentage contraction decreases in the same order as the
boiling points rise and as the molecular percentages of alcohol in the
mixtures of constant boiling point diminish ; but with the secondary
and tertiary alcohols examined, the order depends on the concentration,
and it is only with large quantities of water that wopropyl alcohol
takes its proper place, whilst for mixtures rich in uopropyl alcohol, and
for all the ^<. butyl alcohol mixtures witbin the limits of the table,
the percentage contraction is very low.
Xi HetU CJiangt9 on Mixing v>ith Wctttr*
In order to carry out a complete investigation of these changes, a
much larger amount of material would be required than we had at our
disposal except in the case of methyl alcohol. It was thought worth
while, however, to make an approximate comparison of the heat
changes produced by mixing the alcohols with water, taking the same
molecular proportions in each case.
The alcohol and water were mixed together in a small, round-bottomed
flask in the proportion of 60 mols. of alcohol to 40 of water, the
total weight in each case being about 30 grams.
The initial temperatures of the alcohol and of the water, which were
identical, or nearly so, and the final temperature of each mixture are
given in the table below, together with the difference between the final
temperature and the mean of the initial temperatures.
Temperatures,
Water.
Aloohol.
Mean.
Mixture.
A.
Methyl alcohol
21-7"
22-2
26-25
22-7
22-8
21 7'
22-85
26-25
22-7
22-8
21 -7'
22-3
26*25
227
29-56'
25-25
25-20
+ 7 -85*
Ethyl ,
+ 2-95
tert.Butyl ,
-1-06
rt-ProDvi ■
-1-15
woButyl ,,
22-3 ' 10-15
-8*15
No experiment was made with t^opropyl alcohol, as the quantity at
our disposal was very small, or with woamyl alcohol on account of its
slight solubility, but it will be seen that in the case of the other
alcohols the evolution of heat diminishes, or the absorption of heat
increases, as the boiling points of the alcohols rise.
With a large excess of water, evolution of heat was observed in
every case, and rough experiments were made to find the maximum
rise of temperature for each alcohoL The results are appended^
Digitized by
Google
MlXtUHES OF THE LOWER ALCOHOLS WITH BENZENE. 739
Approximate mazimnm
rise of temperatnre.
Methyl alcohol 8*5°
Ethyl „ 6-5
tertButyl „ 40
w-Propyl „ 40
ifioButyl „ 1*0
Here again the order is the same as that of the hoiliug points.
With regard to the general question whether the mixtures of constant
boiling point of the alcohols with water are definite hydrates or not,
in addition to the special arguments against the existence of such
hydrates advanced in the case of n-propyl and of isopropyl alcohols, it
seems sufficient to point to the fact that in none of the cases examined
do the number of molecules of alcohol and water in the mixture bear
a simple ratio to each other.
TTniyersitt Collbob,
Bristol.
LXXV. — The Properties of Mixtures of the Lower
Alcohols with Benzene and with Benzene and Water.
By Sydney Young, D.Sc, F.R.S., and Emily C. Fobtby, B.Sc.
In a previous paper (p. 708), it has been pointed out that since the
monohydric aliphatic alcohols may be regarded on the one hand as
hydroxyl derivatives, CnH2n+i(0H), of the paraffins, and on the
other as alkyl derivatives, (CnH2n+i)0*n, of water, their properties
should approach those of the corresponding paraffins and recede from
those of water as the molecular weight increases. The second point
has been considered in the last paper (p. 717), and we now propose
to describe experiments which bring out the relationship of the alcohols
to the paraffins and other hydrocarbons.
Owing to the fact that the lowest paraffins are gaseous under
ordinary conditions, it is not possible to deal experimentally with mix^
tures of these paraffins with the corresponding alcohols. There are
also very few of the paraffins which can easily be obtained in a pure
state. 7»^Hexane might, indeed, be employed, as when prepared from
pure propyl iodide and sodium it is readily purified, but it is too
expensive a material to be used in large quantities. Benzene, how-
ever, was found to behave in a similar manner to Tt^hexane, and
it was therefore made use of in the majority of the experiments*
Digitized by
Google
740 YOUNG AND FORTEY : MIXTURES OF THE LOWER ALCOHOLS
In order to obtain satisfactory results, it is necessary to employ a
very efficient still-head. In all the experiments described in this
paper, with the exception of those with ethyl alcohol, a 5-eolumn
'^ evaporator " still-head was used. For ethyl alcohol mixtures, the
18-column Young and Thomas dephlegmator was employed.)
A few experiments were made with n-hexane and ethyl alcohol, the
results of which are given below. Mixtures of the carefully dried
alcohol and hexane were distilled together and the following boiling
points were observed :
Temperature
Pressnre.
observed.
corr. to 700 mm
740-1 mm.
57-91°
58-64°
748-86 „
58-33
58-74
764-4 „
58-83
58-67
Mean 68-68
dp/cU at the boiling point = 27*1 mm. per degree.
In two cases, known weights were taken, the alcohol being in ex-
cess, and the weights of distillate below the temperature midway
between the boiling points of the mixture of constant composition and
of alcohol were ascertained. From these results, the composition of
the mixture of constant boiling point was calculated by the method
described in the next paper, and was found to be :
Alcohol 21*0 per cent.
Hexane 79-0 „
100-0 „
When ethyl alcohol, n-hexane, and water are distilled together, a
mixture of the three substances comes over at a constant temperature.
One determination of this temperature was made, and it was found to
be 56*60^ at 760 mm., or 2*08° lower than the boiling point of the
binary alcohol-hexane mixture, 4*95° lower than that of the hexane-
water mixture, and 21*55° lower than that of the alcohol-water mix-
ture. These results are not very different from those observed with
benzene, but from the point of view of the separation of water
from alcohol they are less favourable.
I. Met^iyl Alcohol
Methyl Alcohol and Benzene. — Known weights of pure methyl alcohol
and benzene were distilled together ; in one case, M, the alcohol, and
in the other, N, the benzene, was in large excess over that which was
found to be required for the mixture of constant boiling point The
Digitized by
Google
WITH BENZENE AND WITH BENZENE AND WATER. 741
weight of distillate below the middle temperature between the boiling
point of the mixture of constant composition and that of the pure
liquid in excess was ascertained in each case, in order to calculate the
composition of the mixture of constant boiling point by the method
described in the next paper (p. 752).
Bailing poitUs:
Composition of mixture of
Temperature constant boiling point.
. ' . . * .
Pressure. obs. con. to 760 mm. M N Mean
M. 765-05 mm. 58-53° 5835° Alcohol 39-5 39*6 3955
N. 762-3 „ 58-41 5833 Benzene 605 60-4 60-45
100-0 100-0 10000
dp/dt at the boiling point = 28*1 mm. per degree.
Methyl Alcohol, Benzene, and Wetter. — A mixture of the three sub-
stances was distilled, but the distillate was clear and the boiling point
was practically identical with, and certainly not lower than, that of
the alcohol-benzene mixture. In this case, therefore, no ternary mix-
ture of minimum boiling point is formed, and the water, instead of
coming over, as in the case of ethyl alcohol, in the lowest fraction,
accumulates in the still.
II. Ethyl Alcohol.
The application of the results of this investigation to the preparation
of absolute alcohol from strong spirit have been described in the paper
which forms the first of this series (p. 707).
Mhyl Alcohol cmd Benzene, — Three determinations of boiling point
were made, the original mixture in two cases, A and B, containing
excess of alcohol, in the third, C, excess of benzene.
Temperature
Pressure. observed. corr. to 760 mm.
A. 745-3 mm 67-71° 68-26°
B. 753-3 „ 67-93 6818
C. 765-2 „ 68-49 6829
Mean 68-24
dp/dt at the boiling point « 26*6 mm. per degree.
In the mixture C, the benzene was only in slight excess, and the
sp. gr. of the distillate was determined in order to ascertain the com-
position of the mixture of constant boiling point.
The observed sp. gr., 0*86740, at 0°/4° corresponds to 32-36 per cent,
of alcoholy assuming that no change of volume occurs on mixing the
VOL. LXXXI. 3 D
Digitized by VjOOQIC
742 YOUNG AND FORTBY : MIXTURES OF THE LOWER ALCOHOLS
two liquids. In order to test this point, the sp. gr. of a mixture con-
taining 7*850 grams of alcohol and 17'269 grams of benzene, or 31*26
per cent, of alcohol, was determined, and was found to be 0*86851.
The volume of a gram corresponding to this sp. gr. is 1'1514»
and that calculated from the composition on the assumption that
there is no change of volume on mixing is also 1'1514. It may
therefore be assumed that there is no appreciable contraction or
expansion.
The composition was also determined by distilling a known mixture
containing a large excess of benzene, and weighing the distillate which
came over below the middle temperature. As will be seen, the agree*
ment is excellent :
From sp. gr. By distillation.
Alcohol 32-36 32-45
Benzene 6764 6755
100-00
10000
Ethyl Aloohol, Benzene, and Water. — In the course of the preparation
of pure alcohol by distillation of strong spirit with benzene, a large
number of determinations of the boiling point of the ternary mixture
were made. In these cases, alcohol was of course the final product,
but later on a few determinations were made with mixtures from
which (1) water, (2) benzene was finally left in the still. The results
are given below :
JReeidue in Still,
Alcohol.
Temperature.
Press.
Obs.
Corr.
ram.
7647
65-02-
64 85'
753-7
64-62
64-85
768-8
64-62
64-85
767-85
64-77
64-87
762 8
64-95
64-85
760-2
64-85
64-85
758-1
64-62
64-87
764-6
64-65
64-85
764-8
66-02
Mean...
64-84
64-85
Water.
mm.
761-2
762-6
763 -8
Temperature.
Obs. Corr.
64-94'
64-97
65-02
Mean..
64 ^O'
64-87
64-88
64-88
Benzene.
Press.
mm.
768 2
764-5
Temperature.
Obs. Corr.
64-99'
65-00
Mean..
e4-87'
64-84
64-85
Final mean 64*86
dpjdt at the boiling point =27*8 mm. per degree.
Digitized by
Google
WITH BENZENE AND WITH BENZENE AND WATER. 743
The ternary mixture is turbid, and separates on standing into two
layers, of which the lower, consisting of dilute alcohol with a small
amount of dissolved benzene, is by far the smaller. When the
alcohol is in large excess in the original mixture and the benzene
more than would be sufficient to take over all the water if the
separation were complete, the liquid tends to separate into the
three fractions: (1) benzene-aloohol-water, (2) benzene-alcohol, (3)
alcohol.
In this case, the distillate remains turbid until the temperature mid*
way between the boiling points of (1) and (2) is reached. After the
temperature begins to rise above 64*86^, the distillate becomes richer in
alcohol and poorer in water, and the alcohol-water layer therefore
becomes lighter. If the original mixture is poor in water, and
especially if the temperature of the room is low, it may happen that
the alcohol-water layer has the same, or even a lower, specific gravity
than the benzeue layer, and we have, in fact, observed the two layers
change places as the temperature altered.
Composition qf the Te/rMuty Miocture, — A direct determination of the
composition of the ternary mixture was made in the following manner.
A considerable quantity, obtained by distillation of mixtures rich in
alcohol was redistilled, and 141*3 grams of the mixture were placed in
a separating funnel, and water was added to separate the benzene
more completely. The aqueous alcohol, containing a minute quantity
of dissolved benzene, was run off, and the residual benzene was
washed three times with small quantities of water to extract com-
pletely the dissolved alcohol, the washings being added to the main
quantity of dilute alcohol. The washed benzene was poured into a
tared flask and was found to weigh 104*3 grams. The dilute alcohol
was then distilled until the temperature rose to 78^, and the turbid
distillate was transferred to the 'separating funnel, which still con-
tained a very small amount of benzene. This small quantity of
turbid mixture was then treated with water as before, and the
washed benzene separately weighed and found to amount to 0*3
gram.
The washings from this benzene were added to the original
quantity of dilute alcohol, and the whole was distilled until the tem-
perature rose to 100^. The distillate was no longer turbid even at
first, and could not have contained more than a trace of benzene. The
weight of distillate was 68*0 grams, and its specific gravity at 0^/4°
was 0*95201, corresponding to 38*5 per cent, of alcohol ; the weight of
alcohol was, therefore, 26*2 grams. The weight of water, calculated by
difference, was 10*5 grams. The percentage composition of the ternary
mixture is therefore :
3 D 2
Digitized by
Google
744 TOUNG AND FORTEY: MIXTURES OF THE LOWER ALCOHOLS
Benzene 74*1
Alcohol 18-5
Water :.. 7-4
100-0
The distillation method may be employed to determine the com-
position of the ternary mixtm-e if that of the original mixture and
that of the binary mixture of constant boiling point are known.
As there are three possible binary mixtures, and we may have
alcohol, water, or benzene left in the still at the end of the distillation,
it follows that when all three fractions are actually formed there are
six possible methods of separation, according to the relative quantities
of the components in the original mixture.
In practice, five of these cases are available for the determination of
the composition of the ternary mixture, the sixth being excluded on
account of the very small difference between the boiling point of the
alcohol-water mixture and that of pure alcohol.
The other five cases were actually investigated, and details of the
results are given in the separate paper on *' Fractional Distillation as
a Method of Quantitative Analysis " (p. 752). It will be sufficient
here to give the mean percentages, calculated from the ^ye deter-
minations, which agree very well with those obtained directly :
By distillation. Directly determined.
Benzene 74-3 74-1
. Alcohol 18-2 18-6
Water 75 74
1000 100-0
III. isoPropyl Alcohol,
iaoPropf/l Alcohol and Benzene. — MoPropyl alcohol and benzene form
a mixture of minimum boiling point the composition of which was
determined by distillation of a known mixture. The details are as
follows :
Composition of raiztiUQ
Weight below middle of constant boiling
Miztnre taken. point.* point. Corrected.
Alcohol 21-3 Observed 63*8 Alcohol 33*3
Benzene 65*35 Corrected 64*1 Benzene 66*7
86*65 100-0
* By "middle point" is to be understood the temperature midway between the
boiling points of the two liquids, whether pure substances or mixtares of constant
/Google
Digitized by ^
WITH BENZENE AND WITH BENZENE AND WATER. 745
The benzene was in excess in the original mixture..
Four mixtures of tgopropyl alcohol and benzene were distilled, three
containing alcohol and one benzene in excess. The boiling points are
as follows :
Alcohol in excess. Benzene in excess.
Pressure. Temperature Pressure. Temperature
/ * ^ , * ,
obs. corr. to 760 mm. obs. corr. to 760 mm.
763-5 7201° 71-88° 765-4 72*11° 71-91°
733-3 70-91 71-91
772-2 72-43 71-97
Mean 71-92 General mean 71 -92°.
d/p\dt at the boiling point = 26*6 mm. per degree.
i&oPropyl Alooholy Benxene, and Water. — Having only a small
quantity of pure alcohol at our disposal, the alcohol-water mixture of
constant boiling point was employed, and its composition determined as
follows :
Mixture taken. Corresponding to
Alcohol-water mixture 27*45 Alcohol 2413
Benzene 86-05 Benzene 86*05
Water 2*95 Water 6*27
116-45
116-45
The benzene was in greatest excess, and the distillate tended to
separate into (1) the ternary mixture ; (2) the alcohol-benzene mix-
ture ; (3) benzene :
Composition of ternary
Obs. Corr.
Weight below first middle point
Weight below second middle point
Obs. Corr.
mixture.
Corr.
83-3 83-6
Alcohol
18-7
25-4 25-5
Benzene
73*8
Water
7-5
100-0
It will be seen that the composition of the alcohol-benzene and of
the alcohol-benzene-water mixtures of <sonstant boiling point is nearly
the same in the case of ethyl and of i^opropyl alcohols.
For the determination of the boiling point, five mixtures were dis-
tilled ; in one case benzene and in the other four alcohol was the
final product left in the still.
boiling point, into which the original mixture tends to separate, or, in the case of
a more complex mixture, the temperature midway between the boiling points of
any two consecntiye fractions of constant boiling point
Digitized by
Google
746 YOtJNG AKD FOKTEY: MIXTURES OF TttE LOWER ALCOfiOLfl
Residue in still :
Benzene. Alcohol.
Pressure. Temperatnre Pressure. Temperature
obs. corr. to 760 mm. obs. corr. to 760 mm.
747*0 mm. 65-99° 66-47° 752-6 mm. 66-29° 6655°
757-05 „ 66-44 6655
General mean 66*51 751*1 ,, 6617 66-50
753-9 „ 66-24 66-46
Mean 66-51
dp/cU at the boiling point » 27*0 mm. per degree.
IV. iert.Btayl Alcohol.
teTt,Butyl Alcohol and Benzene. — ^^.Butyl alcohol and benzene
form a mixture of minimum boiling point the composition of which
was determined by distillation of a known mixture. The details are as
follows :
Composition of mixture
Weight below middle of constant boiling
Mixture taken. point point. Corrected.
Alcohol 33*8 Observed 92*0 Alcohol 36*6
Benzene 120*0 Corrected 92*3 Benzene 63*4
153*8 100-0
The benzene was in excess in the original mixture.
For the determination of the boiling point, two mi2diuresof forf. butyl
alcohol and benzene were distilled ; the benzene was in excess in both
cases:
Pressures. Temperature
, * ,
observed. corr. to 760 mm.
755-0 mm. 73-77° 73*96°
760*26 „ 73*95 73*94
Mean 73*95°
dp/dt at the boiling points 26*0 mm. per degree.
tert,BtUyl Alcohol, Benzene, and Water, — ^As in the case of Mopropyl
alcohol, the alcohol-water mixture of constant boiling point was used,
but no additional water was required. The composition was deter-
mined as follows :
Mixture taken. Corresponding to
Alcohol-water mixture 117*5 Alcohol 103*7
Benzene 145*0 Benzene 145*0
Water 13*8
262*5
262*5
Digitized by VjOOQIC
With benzene and With benzeke akd water 747
The alcohol was in largest excess, and the distillate tended to
separate into (1) the ternary mixture; (2) the alcohol-benzene mixture;
(3) alcohol :
Obs.
Weight below first middle point 169*3
Weight below second middle point 39*9
1000
For the determination of the boiling point, two mixtures were
distilled ; in both, the alcohol was the final product, but in one the
second fraction consisted of the alcohol-water mixture, in the other
of the alcohol-benzene mixture :
Temperatare
r ' -^
Pressure. observed. corr. to 760 mm.
756-3 mm. 67*16° 67-30°
757-55 „ 67-20 6730
Composition of ternary
Corr.
mixture. Corr.
169-7
Alcohol 21-4
40-0
Benzene 70*5
Water 8-1
Mean 67 30
d'pldi at the boiling point =» 27-0 mm. per degree.
V. n-Propyl Alcohol,
n-Propyl Alcohol and Benzene, — Known weights of pure ^-propyl
alcoh6l and benzene were distilled together; in one case, M, the
alcohol, and in the other, N, the benzene, was in excess over that
which was found to be required for the mixture of constant boiling
point. Details are given in the paper which follows, and it will be
sufficient here to give the corrected values :
Composition of mixture of constant boiling point
M. N. Mean.
Alcohol 16-9 16-9 16-9
Benzene 83-1 83*1 83*1
1000 1000 100-0
Boiling points :
Alcohol in excess. Benzene in excess.
Temperature Temperature
Pressure. obs. * corr. to 760 mm. Pressure. obs. corr. to 760 mm.
769-6 mm. 77*55° 77-17° 7664 mm. 77-39° 77*13°
761-0 , 77-10 77-06
Mean 77-11 General mean 77-12°.
dp/di at the boiling point « 25*0 mm. per degree.
Digitized by
Google
748 TOUNO AND FOBTET: MIXTUBES OF THE LOWER ALCOHOLS
n-Fropyl AhoM^ B&nzene^ and WaA&r. — ^The mizture was made in
the same manner as with tar^.butyl alcohol, and its composition deter-
mined as follows :
Mixture taken. Corresponding to
Alcohol-water mizture... 30*0 Alcohol 21*55
Benzene 1201 Benzene 120*10
Water 8*45
1501
16010
The alcohol was in greatest excess, and the distillate tended to
separate into (1) the ternary mizture, (2) the alcohol-benzene mizture,
(3) alcohol :
Composition of ternary
Obs. Corr. mixture. Corr.
Weight below first middle point... 98-1 98*4 Alcohol... 9*0
Weight below second middle point 46*9 47*0 Benzene... 82*4
Water ... 8*6
100*0
The boiling points of the above mizture and of another which
separated into similar fractions were determined :
Temperature
Pressure. obs. corr. to 760 mm.
770-25 mm. 68-81° 68*42°
761-2 „ 68*60 68*55
Meao 68*48
d]^\di at the boiling point » 26*3 mm. per degree.
YI. x^oBvJtyl AlcohoL
iaoBiUyl Alcohd and Benzene, — ^The difference between the boiliDg
points of the mizture of constant composition and of benzene was too
small to admit of the distillation being stopped accurately at the
middle point, and it was therefore only feasible to determine the com-
position by distilling a mizture containing ezcess of alcohol :
Composition of mixture of
Weight below middle point constant boiling point
Mixture taken. Obs. Corr. Corr.
Alcohol 35*35 108*1 108*5 Alcohol 9*3
Benzene 98*45 Benzene 90*7
100*0
Digitized by
Google
WITH BENZBNE AND WITH BENZENE AND WATEB. 749
Boiling point :
Temperature
Pressure. obs. corr, to 760 mm.
762-26 mm. 7993° 7984°
dp/dt at the boiliog point = 24*0 mm. per degree.
iooBtUyl Alcohol, Benzene, and W(Uer,— On distilling a mixture of
alcohol, benzene, and water, the benzene-water mixture of constant
boiling point came over first. In a quantitative experiment, 6*5 grams
of water were added to about 130 grams of the alcohol-benzene mix-
ture, the proportions being such that the liquid should separate into
(1) benzene-water, (2) alcohol-benzene, (3) alcohol. The boiling point
of the first fraction was as follows :
Temperature
calc. from
. Pressure. observed. corr. to 760 mm. vapour pressures.*
746-35 mm. 68-78° 69-31° 6925°
Composition of first fraction :
Weight below first middle point.
Percentage of water in first fraction.
calc. from vapour pres-
obs. corr. corr. sures and vapour densities.
73-05 73-35 8-86 883
The results obtained proved conclusively that no ternary mixture is
formed in this case, and that the mixture of minimum boiling point is
identical with that obtained by distilling merely benzene and water
together.
VII. isoAmyl Alcohol.
ISO Amyl Alcohol and Benzene, — A mixture of 26 '6 grams of isoamyl
alcohol with 85*7 grams of benzene was distilled. The temperature
rose at once to 80-27°, or slightly higher than the boiling point of
pure benzene. It. appeared, therefore, that no mixture of constant
boiling point was being formed, and the distillate was collected
below the middle temperature between the boiling points of benzene
and isoamyl alcohol :
Percentage composition of mixture.
Weight below middle point (106-1'). Found. Taken.
23-7
76-3
obs.
corr.
85-55
85-86
Alcohol
Benzene
uncorr.
corr.
23-8
23-6
76-2
76-4
100-0 1000 1000
It is thus clear that no mixture of minimum boiling point is formed
* Benzene and water being practically non-miscible.
Digitized by
Google
750 YOUNO AND I'ORTfiY : MIXTITRES OF THE LOWER ALCOHOLS
and that both pure benzene and pure t^oamyl alcohol may be separated
from the mixture by fractional distillation.
isoAmyl Alcohol, Benzene, and Water, — As in the case of isohntjl
alcohol, no ternary mixture is formed, but the mixture of benzene
and water of constant boiling point comes over first.
VIII. Tabulation qf Dala.
For the sake of convenience, the boiling points and the composition
of the alcohoUbenzene and of the alcohol- benzene- water mixtures of
constant boiling point are tabulated below :
Alcohol'Benzene,
Name.
Boiling point (760 mm.).
Percentage of alcohol In
mixture.
Alcohol.
Mixtore.
1
By weight. | Molecular.
Methyl alcohol
64-70'
78-80
82-44
82-55
97-19
108-06
182-05
58-34'
68-24
71-92
78-95
77-12
79-84
89*56 61 '4
Ethyl „
82*86 ' 44-8
MoPropyl ,,
38-8 I 89*3
fer^.Butyl ,
86-6 87-7
n-Propvl
16 '9 1 20-9
MoButyl ,,
9*3 i 9-71
iffoAmyl ,,
i
It will be seen that the boiling points of the mixtures and the
molecular percentages of alcohol in them follow the same order as the
boiling points of the pure alcohols.
Alcoliol-Benzene- Water,
Boiling
point at
760 mm.
Composition of mixture.
Name.
"Weight percentage.
Molecniar percentage.
Alcohol.
Benzene.
Water.
Alcohol.
Benzene.
Water.
Methyl alcohol
Ethyl
is&Propyl „
tert.BvLtY\ „
n-Propyl „
i^oButyl ,,
isoAmyl ,
••
34-86"
66-51
67-30
68*48
18-5
18-7
21-4
9-0
74 1
78-8
70*5
82-4
7-4
7-6
8-1
8-6
22-8
18-6
17*6
8-9
63-9
66-5
55-0
628
23 8
24-9
27-5
28-8
Digitized by
Google
With benzenIs and With benzene and Wateh. 7S1
Referring first to the table of alcobol-water boiling points in tbe
previous paper (p. 735), it will be seen tbat as tbe boiling points of tbe
alcobols fall tbere is a rapid and fairly regular diminution in tbe differ-
ences between tbe boiling points of tbe alcobols and of tbe respective
alcobol-water mixtures of constant boiling point, and as tbis difference
baa fallen in tbe case of etbyl alcobol to 0*15^ it is clear tbat witb metby 1
alcobol, if a binary mixture of constant boiling point could exist, it
would boil somewbat bigber tban pure metbyl alcobol, but, of course,
a mixture of constant boiling point could not be formed under sucb
conditions.
As regards tbe alcobol-benzene mixtures of constant boiling point,
passing from metbyl to isobutjl alcobol tbe boiling points ajre in tbe
same order as tbose of tbe alcobols, tbose of isobutyl alcobol and benzene
being very near togetber. If iao&myl alcobol and benzene formed a
mixture of constant boiling point, it would boil bigber tban benzene ;
it cannot tberefore exist, and it may be stated tbat no alcobol wbicb
distils at a bigber temperature tban Moamyl alcobol would form a mixture
of constant boiling point witb benzene.
By reference to tbe table of tbe alcobol-water boiling points in tbe
previous paper and to tbe tables of alcobol-benzene and of alcobol-
benzene-water boiling points given above, it will be seen tbat taking
tbe four alcobols, etbyl, Mopropyl, ^<. butyl, and ?i-propyl, tbere is a
rise of 9*57^ in tbe case of tbe alcobol-water mixtures, 8'88° witb tbe
alcobol-benzene mixtures, but only 3*62^ in tbe case of tbe ternary
mixtures.
From tbe last table, it would appear that tbe boiling point of tbe
metbyl alcobol ternary mixture, if it existed, would be not lower tban
60% but tbe boiling point of tbe metbyl alcobol-benzene mixture is
68*34° and as tbis is lower tban tbat of eitber tbe bypotbetical ternary
mixture, tbe binary benzene^water mixture, or tbe alcobol itself, tbe
alcobol-benzene mixture is tbe one wbicb must come over in tbe first
fraction.
Witb regard to tsobutyl alcobol, tbe ternary mixture, if formed,
would evidently bave a boiling point above 69° and probably above
69*25°, tbat of tbe water-benzene mixture of constant boiling point. As tbe
boiling points of tbe alcobol-benzene mixture, 79*84°, and of tbe alcobol*
water mixture, 89*82°, are botb bigber tban 69*25°, it was a question
wbetber tbe first fraction would consist of tbe ternary mixture or of
tbe benzene- water mixture. Tbe result of tbe distillation sbowed tbat
tbe ternary mixture is not formed and it may be concluded tbat no
alcobol witb a bigber boiling point tban tbat of i^obutyl alcobol would
form sucb a mixture, and tbat in all such cases tbe first fraction would
consist of benzene and water. Tbis was actually found to be tbe
case witb isoamyl alcoboL
Digitized by
Google
752 TOUNO AND FOBTEY: FRACTIONAL DISTILLATION AS A
In conclusion, it may be pointed out that the behaviour on distillation
of a mixture of a saturated monohydric aliphatic alcohol with benzene
and water is closely related to the boiling point of the alcohol.
In the case of the lowest member of the series, methyl alcohol, the
first fraction consists of the binary alcohol-benzene mixture; with
tfobutyl alcohol and alljothers of higher boiling point, the first fraction
is the binary benzene-water mixture; as regards the alcohols of
intermediate boiling point, the first fraction in the case of the four
alcohols examined is the ternary mixture, but for the three alcohols,
^ec.butyl alcohol, dimethylethylcarbinol, and ttfrt.butylcarbinol, the
boiling points of which are intermediate between those of n-propyl
and isohutyl alcohols, the question whether the first fraction would
consist of the ternary mixtui*e or the benzene-water mixture can only be
decided by experiment. It seems, however, not unlikely that the ternary
mixture would be formed, at any rate, with Mcbutyl alcohol.
Univissitt Collbgb,
Bbistol.
LXXVL — Fractional Distillation as a Method of
Quantitative Analysis.
By Sydney Youno, D.Sc, F.R.S., and EmiltC. Fortey, B.Sc
Im a paper by one of us entitled '* Experiments on Fractional Distilla-
tion " {J. Soc. Chem. Ind., 1900, 19, 1072), it was pointed out that the
composition of a mixture of homologous substances could in many cases
be ascertained with a fair degree of accuracy from the results of two
or three fractional distillations with an efficient still-head, or in the
case of a mixture of two components which are not difficult to separate,
from the result of a single distillation.
The method depends on the following facts: — ^Taking first the
simplest case, that of a mixture of two liquids, it is found that the
weight of distillate which comes over below the middle point* is almost
exactly equal to that of the component of lower boiling point, even
when the separation is very far from complete.
If the original mixture contains more;than two, say n, components, the
weights of these components will be very nearly equal respectively to
(1) the weight of distillate below the first middle point, (2 to n - 1) the
* By middle point is to be nndeistood the temperature midway between the
boiling points of the two liqaids, whether pure substances or mixtures of constant
boiling point, into which the original mixture tends to separate ; or, in the case of
a more complex mixture, the temperature midway between the boiling points of any
two consecutiyo fractions of constant boiling point
Digitized by
Google
METHOD OF QUANTITATIVE ANALYSIS. 763
weights of distillate between the successive middle points, (n) the
weight abovethe last middle point.
Only two mixtures of substances which are not homologous — methyl
alcohol-water and Moamyl alcohol-benzene — and which separate
normally into the two components, have been examined, but a con-
siderable number of cases in which mixtures of minimum boiling
point are formed have been investigated, and it has been found that
such mixtures of constant boiling point behave like pure
liquids. Thus, if the composition of the mixture of minimum boiling
point is known, that of the original mixture may be calculated from
the weight of distillate below the middle point, and, on the other hand,
if the composition of the original mixture is known, that of the mix-
ture of minimum boiling point may be calculated. The same remarks
would apply to binary mixtures of maximum boiling point, such as that
of formic acid and water, but so far we have not examined any such
mixture.
It is obvious that there must be some loss of liquid by evaporation,
which makes the weight of distillate somewhat too low. This loss will
be greater as the initial boiling point is lower, and as the temperature
of the room is higher. It is not proportional to the amount of liquid
distilled, for a great part of the loss is caused by the saturation of
the air in the flask and still-head while it is being expelled by the
rising vapour. Under otherwise similar conditions, the loss is there-
fore roughly proportional to the volume of air in the still and still-
head, that it is advantageous to use as small a flask as possible and
to employ a still-head of as small capacity as is consistent with efficiency.
A plain, wide still-head or one with spherical bulbs is the least satis-
factory, but the '' pear *' still-head, owing to the diminished capacity of
the bulbs and the increased efficiency, gives much better results. Of
all forms, the ** evaporator " still-head is the best, because the capacity
is relatively small, and the amount of condensed liquid in it is smaller
than in any other of equal efficiency, and because almost the whole of
the liquid returns to the still at the end of the distillation.
With a liquid of low viscosity, like one of the paraffins, the quantity
of liquid left in the still-head is almost inappreciable, and in other
cases it may be reduced to a very small amount by disconnecting the
apparatus while still hot from the condenser, shaking out any liquid
remaining in the funnels, and tilting the tube from side to side to
facilitate the flow of the residual liquid back to the still.*
When the liquid left at the end of the distillation was n-hexane,
hardly a trace was visible in the still-head after cooling, even when left
* In anew form of "evaporator" still-head, which will be described later, the
little fnnnels are done away with, and the tnbe merely requires to be tilted while
still hot
Digitized by
Google
764 YOUNG AND FORTEy : FRACTIONAL DISTILLATION AS A
in position^ whilst with benzene as the final tiquidy the amount could
certainly be reduced to 0*1 gram by taking down the still-head while
hot, as described.
For these reasons, the " evaporator *' still-head was used for ail deter*
minations, except when ethyl alcohol, benzene and water, ethyl alcohol
and hezane, or ethyl alcohol and water were distilled together.
As an example of the estimation of loss by evaporation, we may cite
the distillation of mixtures of methyl alcohol and benzene, one with
benzene, the other with methyl alcohol in excess over the amount
present in the mixture of constant boiling point.
In each case, the distillation was stopped when the middle point
was reached, and the liquid left in the still after cooling was weighed.
The results obtained were as follows :
Benzene in Alcohol in
excess. excess.
Weight of distillate 128-7 132-0
Weight of liquid in still 249 27*2
Total 163-6 159-2
Weight of mixture taken 164-2 160-1
Loss by evaporation and in still-head ... 0*6 0*9
When the benzene was in excess it is quite certain that the weight
of it left in the still-head was not greater than 0*1 gram, and the loss
by evaporation was therefore taken to be 0*5 gram, and in the calcula-
tion of composition this amount was added to the observed weight of
distillate.
Allowing the same amount, 0*5 gram, for loss by evaporation
in the second distillation, that would leave 0*4 gram as the weight of
liquid — mostly methyl alcohol — in the still-head, an amount which
appears quite reasonable, for this more viscous liquid did not flow back
nearly so freely, and there was a visibly much larger amount left in
the still-head.
When the "evaporator" still-head was employed, the correction
applied for loss by evaporation was usually 0-4 or 0*3 gram. With
the 18-column dephlegmator, which was used for the other distillations,
the loss was certainly greater, but could not be accurately estimated.
Experimental.
Methyl Alcohol and Water,
Two determinations were made to test the value of the method,
mixtures of known composition being distilled, in one of which the
alcohol, in the other the water, was in large excess.
Digitized by
Google
METHOD OP QUANTITATIVE ANALYSIS.
755
Methyl aJcohol in Iwrge excess.
Boiling points : Methyl alcohol, 64*7°; water, 100-0°; middle point,
82-35°.
Weight helow
middle point.
Percentage composition of mixture.
Mixture taken.
Found.
Taken.
Uncorrected.
Corrected.
Alcohol 90-9
Water 24*4
Observed 90-6
Corrected 90-8
Alcohol 78-5
Water 21 -6
100-0
78-7
21-3
78-8
21-2
116-3
100 0
100-0
Water in large excess.
Weight below
middle point.
Percentage composition of
mixture.
Mixture taken.
Found.
Taken.
Uncorrected.
Corrected.
Alcohol 89-7
Water 161-6
Observed 83-9
Corrected 34 -2
Alcohol 16-9
Water 83-1
17-0
83-0
100-0
19-7
80-3
201-2
1000
100-0
This second result is apparently unsatisfactory, but it is always
difficult to separate the more volatile component of a mixture when
present in relatively small amount, and, in such a case, a second distil-
lation is usually necessary. The first distillation was therefore
continued until the temperature reached 100°, and the whole of the
distillate, weighing 66*8 grams, was then redistilled, and the double
correction for loss by evaporation was applied. The weight below the
middle point was now 38*9, corrected 39*5, giving the percentage
composition :
Uncorrected. Corrected. Taken.
Alcohol 19-3 19-6 19-7
Water 80-7 80*4 80-3
100-0
1000
100-0
\
It will thus be seen that, by repeating the distillation, the result was as
satisfactory as that given by a single distillation when the alcohol was
/Google
Digitized by ^
766 YOUNO AND FORTSY: FRACTIONAL DISTILLATION AS A
in excess. Even withonfc correcting for loss by evaporation, the agree-
ment is fairly good, but it is much improved by introducing the
correction.
iaoAmyl Alcohol and Bemene,
Boiling points : benzene, 80*2° ; isoamyl alcohol, 132*05° ; middle
point, 106-1°.
Weight below
middle point.
Percentage composition of mixture.
Mixture taken.
Found.
Taken.
Uncorrected.
Corrected.
28-6
76-4
Alcohol 26-6
Benzene 86-7
Observed 86-66
Corrected 86 -86
Alcohol 28-8
Benzene 76 -2
23-7
76-8
112-8
100-0
100-0
100 0
The separation is here an easy one and the agreement is exceedingly
good.
Mixtures of Constant Boiling Point,
The first two experiments serve to show that the quantity of a
mixture of constant boiling point may be estimated by the distillation
method in the same way as a pure substance. The methods of ex-
periment and of calculation are similar in all respects.
For the sake of brevity, a mixture of constant boiling point of two
components will be referred to in this paper as a '' binary " mixture and
a mixture of constant boiling point of three components as a '' ternary "
mixture.
iBoFropyl Alcohol and Water,
Boiling points : binary mixture, 80*37°; water, 100*0°; middle point,
90*2°.
Weight below
middle point.
Percentage composition of mixture.
Mixture taken.
Found.
Taken.
Uncorrected.
Corrected.
Binary I57.7
mixture/ ' '
Water 20*1
Observed 67-3
Corrected 67 -6
^""f^ I73-66
mixture/"'"*'
Water 26-85
74 05
25-96
74-15
26-85
77-8
100-00
100-00
100-00
The agreement in this case is perfectly satisfactory.
Jigitized by
Google
METHOD OF QUANTITATIVE ANALYSIS.
757
teTt.Butyl Alcohol and WcUer.
Boiling points : binary mixture, 79-91° ; water, 100-0°; middle point,
89-95°
Woicht below
middle point.
Percentage composition of miitture.
Mixture taken.
Found.
Taken.
Uncorrected.
Corrected.
mixture /*^° °
Water 29-6
Observed 68*2
Corrected 68 '5
^?°f^ 165-9
mixture/
Water 84*1
100-0
66-26
88-76
66-6
88-4
88-8
100-00
1000
Here the agreement is not quite so good, but may be regarded as
fairly satisfactory.
In the following experiments, the composition of the mixtures of
minimum .boiling point was calculated from the results of the dis-
tillations.
n-Propyl Alcohol and Water.
For this distillation, a mixture of dry n-propyl alcohol with water
was employed.
Boiling points: binary mixture, 87-72°; water, 100-0°; middle point,
93-85°.
Weight below
mldale point.
Percentage composition of binary mixture.
Mixture taken.
By distillation.
From specific
gravity.
Uncorrected.
Corrected.
Alcohol 76*6
Water 60-0
Observed 106-4
Corrected 106-7
Alcohol 72-0
Water 28 0
71-8
28-2
100-0
71-69
28-81
126-6
1000
100-0
The calculation is based on the following considerations. If the
fractionation were complete, the original mixture would separate into
(1) the binary mixture containing the whole of the alcohol, (2) the
excess of water. The weight of the binary mixture is given by the
corrected weight of distillate below the middle point, and thus the
weight of alcohol in the binary mixture and that of the binary mix-
ture itself are ascertained.
YOL. LXXXI. 3 E
Digitized by VjOOQIC
768 YOUNO AND FORTEY: FRACTIONAL DISTILLATION AS A
The statement may be made, generally, thus : The ratio of the
weight of the component not in excess in the original mixture to the
corrected weight of distillate is equal to the proportion of that com-
ponent in the binary mixture.
Thus, in the actual distillation : Weight of alcohols 766 grams;
weight of binary mixture = corrected weight of distillate below middle
point =106*7 grams. Percentage of alcohol in binary mixture »
76-6 X 100
^067 ^^®-
The determination of the composition of the binary mixture by
means of the sp. gr. has already been referred to in the paper on the
properties of mixtures of the lower alcohols with water.
It will be seen that the agreement is very satisfactory.
iso^m^^ Alcohol and Water,
1. Water in excess.
Boiling points: binary mixture, 96 '16°; water, 100*0^; middle
point, 97-6°.
Mixture taken.
Weight below
middle point
Percentage composition of binary
mixtnres.
Uncorrected.
Corrected.
Alcohol 88-8
Water 69*6
Observed 76 "4
Corrected 76-9*
Alcohol 60-8
Water 49*2
50-6
49-5
108-3
100 0
100-0
* The correction introduced is slightly larger than usual on account of a minute
loss of vapour during the distillation.
2. Alcohol in excess.
Boiling points: binary mixture, 95*15°; alcohol, 132*05°; middle
point, 113-6°
Mixture taken.
Weight below
middle point
Percentage composition of binary
mixture.
Uncorrected.
Corrected.
Alcohol 68-8
Water 427
Observed 85*66
Corrected 85 '95
Alcohol 60-1
Water 49*9
100-0
60-8
49-7
111-0
100 0
The agreement between the corrected values is very satisfactory.
Digitized by
Google
M^tttOD OF QttAHTITATlVE AKALYBIS.
769
Melhyl Alcohol and Benzene.
These deierminations have already been referred to in estimating
the loss by evaporation ; full details are given below :
L Benzene in exeeai.
Boiling points: binary mixture, 58*34^; beD2ene, 80*2^; middle
point, 69-25°
Mixture takeHi
Weight below
middle point
Percentage composition of binary
mixture.
Uncorrected.
Corrected.
Alcohol 61-2
Benzene 108 0
Obaerved 1287
Corrected 129-2
Alcohol 89-8
BeDEene 60*2
89-6
60-4
164-2
100-0
100-0
2. Methyl cdcohol in exoeee.
Boiling points : binary mixture, 58*34° ; methyl alcohol, 64*7° ;
middle point, 61*5°.
Mixtare taken.
Weight below
middle point.
Percentage compoiition of binary
mixture.
Uncorrected.
Corrected.
Alcohol 79-9
Benzene 80*2
Obeerved 1820
Corrected 182-5
•
Alcohol 39-2
Benzene 60*8
89-5
60-5
1601
100-0
100-0
Ethyl Alcohol and Benzene,
Only one determiuation was made, the benzene being in excess, but
the composition of the binary mixture was also ascertained by a deter-
mination of its sp. gr. as described in the previous paper (p. 741).
Boiling points: binary mixture, 68*24°; benxene, 80-2°; middle
point, 74-2°.
3 £ 2
Digitized by VjOOQIC
760 TOUNC^ AKD B'ORTEYt FRACTIONAL DIStlLLATlOK AS A
Weight below
middle poi^.
Percentage composition of binaiy miztnie.
Mixture token.
By distillation*
From
specific
gravity.
Uncorrected.
Corrected.
Alcohol 26-7
Benzene 78 '6
Observed 78-9
Corrected 79-2
Alcohol 82-6
Benzene 67*4
100-0
82-45
67-66
32-86
67-64
104-2
100-00
100-00
- Ethyl Alcohol and ffex<me.
Two distillations were made, the mixture in each case containing
excess of alcohol, but both the relative and actual quantities differed
considerably. The 18-column dephlegmator was employed.
Boiling points: binary mixture, 58*68'^; alcohol, 78*3^; middle
point, 68-6°
Miztare token.
Weight below
middle point.
Composition of binary miztare.
Uncorrectod.
Corrected.
I. Alcohol 127-7
Hezane 128-4
Observed 161-1
Corrected 162'1
Observed 52-8
Collected 53-4
Alcohol 20-3
Hezane 79 '7
20-8
79-2
256-1
II. Alcohol 72-5
Hezane 42-0
100-0
Alcohol 20-6
Hezane 79-6
100-0
21-3
78-7
114-5
100 0
100-0
The agreement between the two corrected values is fairly satisfac-
tory ; the percentage of alcohol was taken to be 21*0.
n- Propyl Alcohol and Benzene.
In this case, two distillations were carried out, one of a miztare
with benzene in excess, the other with alcohol in excess.
1. Benzene in exoesi.
Boiling points: binary mixture, 77*12°; benzene, 80*2°; middle
point, 78-65^
Digitized by
Google
METHOD OF QUANTITATIVE ANALYSIS.
761
2. Aloohcl in excess.
Boiling points: binary mixture, 77-12°; alcohol, 97*19°; middle
point, 87-16°
Mixture taken.
Weight below
middle point.
Percentage composition of binary
mixture.
Uncorrected.
Corrected.
Alcohol 26-95
Benzene 163 '25
190-2
Alcohol 40-2
Benzene 80*0
120-2
Benzene in excess.
Observed 168 '8
Corrected 159*2
Alcohol 17-0
Benzene 83 0
100-0
Alcohol in excess.
Observed 96*0
Corrected 96*3
Alcohol 16-7
Benzene 83-3
100-0
16-95
83 05*
100-00
16-95
83 05
100-00
Here there is perfect agreement between the two corrected values.
Composition of a. Ternary Mixture,
When a mixture of three liquids gives rise, on distillation, to the
formation of a ternary mixture of minimum boiling point, the separ-
ation may, theoretically, take place in twelve different ways, and, in
addition to these, if the original mixture had the same composition as
the ternary mixture, its behaviour on distillation would be precisely
that of a pure liquid.
Determinations were actually made with only one set of three
liquids, and we may take this case, that of a mixture of ethyl alcohol,
benzene, and water, as a typical one. For the sake of convenience,
we will use the initial letters. A., B, and W, to represent the three
components.
The possible cases are as follows :
Fint fraction.
Second fraction.
Besidne.
1.
A.B.W
A.W
W
2.
»>
B.W
w
3.
»
A.W
A
4.
i>
A.B
A
Digitized by LjO
762 YOUNG AND FORTEY: FRACTIONAL DISTILLATION AS A
Tint fhuition.
Second fraction.
Renidae.
6.
A.B.W.
B.W
B
6.
»
A.B
B
■7.
>»
—
A
8.
>l
—
B
9.
It
—
W
10.
>9
—
A.B
11.
»
-—
A.W
12.
n
—
B.W
13.
if
—
.^
The first six canes, and, on redistillation of the first Traction, the
last, would be those commonly met with. For cases 7, 8, and 9, the
relative quantities of two of the liquids would have to be precisely the
same as in the ternary mixture, the third liquid being in excess. With
regard to cases 10, 11, and 12, the composition of the original mixture
would have to be exactly such as would be obtained by mixing together
the pure ternary mixture with any one of the three pure binary mix-
tures, although, of course, any proportion of these two mixtures might
be taken. Even if these conditions as regards cases 7 to 12 were
fulfilled, it is doubtful whether, owing to imperfect separation, the re-
sults specified in the above table would be actually attained, but the
matter has not been examined experimentally.
Of the first six cases, the third, when ethyl alcohol is employed, is
unrealisable in practice, owing to the very small difference between
the boiling points of the second fraction (AW) and the residue (A).
Mixtures, however, tending to separate in the other five ways speci-
fied, were employed for the determination of the composition of the
ternary mixtura
In order to calculate this, it is necessary to know, not only the com-
position of the original mixtures, but also that of the binary mixture
forming the second fraction. The composition of each of the three
binary mixtures is now known, and, for convenience of reference, the
boiling points of all possible fractions, and the percentage compositioa
of the binary mixtures, are given below :
Percentage composition.
Boiling points. A. B. W.
W 100-0° — — 100
B 80-2 — 100 —
A 78-3 100 — —
A.W. 7815 96-57 — 4'43
B.W 69-26 — 91-17 8-83
A.B 68-24 32-36 67-64 —
A.B.W , 64-86
Digitized by
Google
METHOD OF QUANTITATIVE ANALYSIS.
763
In making up the original mixture, the materials employed were (1)
99 '5 per cent, (by weight) alcohol, (2) pure benzene, (3) pure i;^ter,
(4) the binary A.B. mixture.
Eractions : A.B.W. ; A.W. ; W. Middle points 71 •SS^ and 89*1°
Mixture taken.
Weights below middle
points.
Composition of ternary
mixture.
Obseryed.
Corrected.
Uncorrected.
Corrected.
A. 66*0
B. 74*2
W. 50-6
(1) 99-5
(2) 61-7
99-9
61-8
A. 16-7
B. 74-6
W. 8-7
100-0
16-6
74-3
9-2
190-7
100-0
In making the calculation, it is assumed, as before, that the corrected
weights of the two distillates are equal to those of the ternary and
binary mixtures respectively, which would be obtained if the separ-
ation were perfect.
That being so, in the above case, the weight of benzene in the
ternary mixture is simply that of the benzene taken ; the weight
of alcohol in the ternary mixture is the weight taken less that in the
binary mixture, which can be calculated j the weight of water is given
by difference.
II a and II 6.
Two mixtures were distilled in this case.
Fractions : A.B.W, ; W.B. ; W. Middle points 67 05° and 84-6°.
Mixture taken.
Weights below middle
points.
Composition of ternary
mixture.
Observed.
Corrected.
Uncorrected.
Corrected.
A. 18-4
B. 120 0
W. 621
(1) 94-9
(2) 64-0
IIOw
95 -S
64-1
A. 19-4
B. 74-6
W. 6-0
19-8
74-2
6-5
190-6
100-0
1000
Digitized by
Google
764 YOUNG AND FORTEY : FRACTIONAL DISTILLATION AS A
Mixture taken.
Weights below middle
point.
Composition of temaiy
mixture.
Obeeryed.
Ck)iTeoted.
Uncorrected.
Corrected.
A. 18-6
B. 90 0
W. 64-3
(1) 96-8
(2) 20-7
116.
96-7
20-8
A. 19-4
B. 74-6
W. 6 0
19-8
74-8
6-4
162-8
100-0
100 0
The agreement between the two results, notwithstanding the different
relative quantities in the original mixture, is remarkable.
IV. Fractions : A.B.W. ; A.B. ; A. Middle points 66-55° and 73-3®
V. „ A.B.W.; B.W; B. „ 6705 „ 747
VI. „ A.B.W.; A.B.; B. „ 6655 „ 74-2
Mixtore taken.
Weight below middle
point.
Composition of temftry
mixture.
Observed.
Corrected.
Uncorrected.
Corrected.
IV.
A. 76-0
B. 108-0
W. 7-6
(1). 100-6
(2). 47-6
101-0
47-6
A. 17-1
B. 76-4
W. 7-6
17-6
76 1
7-4
190-6
100 0
1000
V.
A. 18-6
B. 160 1
W. 12-1
(1). 97-1
(2). 62-6
97-6
62-6
A. 19-1
B. 78-2
W. 7-7
19-0
78-4
7-6
190-7
1000
100-0
Digitized by
Google
METHOD OF QUANTITATIVE ANALYSIS.
765
Mixture taken.
Weight below middle
point.
Composition of ternary
mixture.
Observed.
Corrected.
Uncorrected.
Corrected.
A. 86-0
B. 148-8
W. 7-6
(1). 111-6
(2). 42-6
VI.
112 0
42-7
A. 19 0
B. 74-2
W. 6-8
18-9
74-3
6-8
190-9
100-0
100 0
The composition of the ternary mixture was directly determined, as
described in the previous paper. The results of this determination
and the mean of the results obtained by distillation, taking Ila and
lib as a single determination, are given below :
By distillation
Direct ^ ^ ^
determination, uncorrected. corrected.
Alcohol 18-5 18-3 18-2
Benzene 74-1 74*4 74-3
Water 74 73 7-6
100-0 100-0 1000
The agreement may be regarded as extremely satisfactory, although
some of the individual values, especially those of alcohol and water in
I, differ somewhat widely from the means. The explanation of the
rather large errors in the first distillation is given on p. 767.
Cases to whioh the Distillation Method is inapplicable.
Of the numerous mixtures investigated, two only, n-hexane-benzene
and ethyl alcohol-water, have given unsatisfactory results.
The relation of boiling point to molecular composition is very similar
for both these mixtures ; in both cases, the addition of moderate quan-
tities of the less volatile component has very slight effect on the
boiling point, but whereas with ethyl alcohol and water there is un-
doubtedly a definite mixture of minimum boiling point, the experimental
results do not indicate with certainty whether benzene and n-hexane
behave in this way, although it is extremely probable^hat such a
mixture, boiling less than 0*05^ lower than n-hexane, is actually formed.
The boiling points of various mixtures of benzene and hexane have
Digitized by
Google
766 YOUNG AND FORTEY: FRACTIONAL DISTILLATION AS A
been determined by Jackson and Young (Trans,, 1898, 73, 922), and
a curve was drawn to represent the relation between the molecular
composition and the boiling points. This curve is well reproduced by the
formula t = 68-97° - 0-0134wi + 0-001366m« - 0-0 J360m» + 0-0 J248mS
where t is the boiling point of the mixture and m the molecular
percentage of benzene.
In the following table are given the molecular percentages of benaene
and the boiling points, observed and calculated :
Boiling pointa.
Boiling points.
Molecular
1
Molecular
percentage
percentage
of benzene.
Observed
Calcu-
lated.
A
of benzene.
Observed
Calcu.
lated.
A
0
68-95"
68-97'*
+ 0-02''
67-1
72-70'
72-64°
-0-06'
6 0
69 00
68-94
-0 06
69-4
78-01
72-97
-0-04
7-1
68-96
68-94
-0 02
79-6
74-67
74-71
+ 0-04
12-6
69-00
69-00
0
80-4
74'75
74-87
+ 0-12
18-6
69-14
69-12
-0-02
86-4
76-12
7619
+ 0-07
29-8
69-47
69-60
+ 0-03
89-8
76-91
77-05
+ 0 14
33*8
69-72
69-72
0
90-9
77-20
77-85
+ 015
41-9
70-17
70-19
+ 0-02
92-7
77-75
77-85
+0-10
49 1
70-70
70-72
+ 0-02
94-5
78-49
78-38
-0-11
49-9
70-70
70-79
+ 0-09
95-7
78-80
78-74
-0-06
55-4
71-42
7128
-0-14
100-0
80-20
80-17
-0-03
The formula appears to represent the results with sufficient accuracy
to permit of its being employed for the calculation of the minimum
boiling point. The calculated boiling points for mixtures containing
1, 2, . . . 10 per cent, are given below, and it will be seen that the
minimum boiling point is 68*935°, or 0'036° below the calculated
boiling point of 7i-hexane.
Molecular percentage
of benzene.
Boiling point
calculated.
Molecular percentage
of benzene.
Boiling point
calculated.
0
1
2
8
4
5
68-970-
68-958
68-949
68-942
68-987
68-935
6
7
8
9
10
68-986*
68-939
68-944
68-951
68-960
It would appear from the above table that the mixture of minimum
boiling point contains about 5 mols. per cent, of benzene, but the
actual composition is somewhat uncertain owin^ to the flatness of the
cqrve m this neighbourhood.
Digitized by
Google
METHOD OF QUANTITATIVE ANALYSIS. 767
The estimation of the composition of a mixture by distillation is
possible when the separation of the components, whether single sub-
stances or mixtures of constant boiling point, is practicable. In the
case of ethyl alcohol and water, where the separation would be that of
alcohol-water from water, we found that repeated fractional distil*
lation with the most efficient still-head failed to give the pure mix-
ture of constant boiling point. From a mixture of benzene and
n-hexane it is impossible to separate either pure hexane or a mixture
of constant composition. The following facts, bearing on this point,
may be noticed about the benzene-hexane mixtures. (1) Benzene
must be added until the mixture contains about 16 mols. per cent,
before the boiling point rises 0*1^ above that of hexane ; (2) the boiling
point of a mixture containing equal molecular proportions is 70 '8^ or
the rise of temperature is only 1'85^ out of 11*25°, the difference be-
tween the boiling points of the pure components; (3) the mixture
which has the boiling point 74'6°, midway between those of hexane and
benzene, contains 79 molecules per cent, of benzene.
The ethyl alcohol-water curve, constructed from the data given by
Noyes and Warfel {J. Amer, Chem, Soc, 1901, 23, 463), is very similar
in form to the n-hexane-benzene curve, but cannot be represented by
so simple a formula. The following facts may be stated. (1) Water
must be added until the mixture contains about 25 molecules per cent,
before the boiling point rises 0*P above that of pure alcohol ; (2) the
boiling point of a mixture containing equal molecular proportions is
about 79*8°, or the rise of temperature is only 1*5° out of 21*7°, the
difference between the boiling points of the pure components ; (3) the
mixture which has a boiling point 89*15°, midway between those of
alcohol and water, contains about 93*5 mols. per cent, of water.
On distilling miictures containing from 16 to 25 per cent, by weight
of water through the 18-column dephlegmator and calculating the per^
centage of water in the mixture of constant boiling point in the usual
way from the weight of distillate below the middle point, values from
7*6 to 8, instead of 4*43, per cent, were obtained, showing that too
much water was carried down. Referring back to the calculation of
the composition of the ternary ethyl alcohol-benzene-water mixture
from the first distillation, if we take 7*8 as the percentage of water in
the binary W.A mixture, the calculated composition of the ternary
mixture would become :
A 18-2
B 74-3
W 7*5
100*0
which agrees very well with thftt directly observed,
Digitized by
Google
768 TOUNQ : THE VAPOUR PRESSURES AND
. General Conelusions.
From the foregoing results, it will be seen that the distillation
method, provided a very efficient still-head is used, may in the
great majority of cases be safely employed for the determination of
the composition of a mixture. But it must be borne in mind that from
a mixture of two liquids it is almost always more difficult to separate
the more volatile than the other component, and, therefore, if the
original mixture contains a relatively very small amount of the more
volatile component, a second distillation may be necessary, and a large
quantity of the original mixture would be required in order to give a
sufficient amount of distillate for a second operation. In such a
case, the best plan is to continue the distillation the first time until
the, boiling point of the less volatile constituent is reached. No
separation into fractions is necessary, but the whole of the distillate
should be employed for the second operation, and the weight below
the middle point then ascertained. A double correction for loss by
evaporation must be introduced.
As regards the separation of three or more substances from a mix-
ture, it may be pointed out that, as a general rule, the prder as regards
facility of separation is as follows : (1) the least volatile component,
(2) the most volatile component, (3) the intermediate components.
It appears to be only when the curve representing the relation be-
tween boiling point and molecular composition is exceedingly flat at
either end, as is the case for ethyl alcohol- water and for r^hezane-
benzene when the more volatile component is in large excess, that the
method is inapplicable.
Uniyersitt College,
Bristol.
LXXVII. — The Vapour Pressures and Boiling Points of
Mixed Liquids. Part I.
By Sydney Young, D.Sc, F.R.S.
It is well-known that when two non-miscible liquids are placed
together in a vacuous space, the pressure exerted by the vapour is
equal to the sum of the vapour pressures of the two substances when
heated separately to the same temperature. If the two liquids are
distilled together, the boiling point will be that temperature at which
the sum of the two vapour pressures is equal to the barometric pressure.
It will therefore be lower, frequently much lowery than the boiling
point of either pure substance.
Digitized by
Google
BOILING POINTS OF MIXED LIQUIDS. PART I. 769
If the two liquids are miscible within limits, the vapour pressure
will be lower than the sum of the vapour pressures of the components
but will, in general, be higher, and the boiling point will be lower, than
that of either pure component.
In the case of liquids miscible in all proportions, the vapour pressure
and the boiling point may lie between those of the compounds or thej
may be higher or lower than those of either component ; it is therefore
not possible to make any general statement as to the relation of the
vapour pressures or of the boiling points of such mixtures to those of
their components.
The question, however, what should be regarded as the normal
behaviour of two liquids miscible in all proportions has been discussed
by several investigators. Guthrie (Fhil> Mag.y 1884, [v], 18, 517)
concluded that if we could find two liquids showing no contraction,
expansion, or heat change on mixing, the vapour pressures should be
expressed by a formula which reduces to
y./>, + (100-p)P.,
^" 100
where P, P^, and P^ are the vapour pressures of the mixture and of
the two components A and B respectively at the same temperature, and p
is the percentage by weight of the liquid A. In other words, the
relation between the vapour pressure and the percentage composition
by weight should be represented by a straight line.
Van der Waals {Proc, Roy, Acad, Amsterdam, 1900, 3, 170) con-
siders that if the critical pressures of the two liquids are equal, and if
the relation suggested by Galitzine and by Berthelot, ai^=> Joijo^,
holds good (oj^ represents the matual attraction of the unlike mole-
cules, Oj and a, the attractions of the like molecules), the relation
between vapour pressure and mcleGular composition should be repre-
sented by a straight line ; or the equation already given should hold
good if p is the molecuUvr percentage of ^.
Kohnstamm {Inaug, Diss. Ameterdam^ 1901) has determined the
vapour pressures of various mixtures of carbon tetrachloride and
chlorobenzene, the critical pressures of which, 34180 mm. and 33910 '
mm., are nearly equal, and finds that the curvature, in this case is not
very marked. At the temperature of experiment, the maximum
deviation from the straight line amounted to a little over 6 mm. on a
total observed pressure of 93*7 mm., or about 6*6 per cent.
It seems reasonable to suppose that the molecular attractions a^, a^^
and a^.j should be most nearly equal, and the relation a^,^ = Ja^'a^
most likely to be true in the case of very closely related chemical com-
pounds, such as the halogen derivatives of benzene, for which I have
shown that many simple physical relations hold good. Thus the
Digitized by
Google
770
Young : tab vapocTb piusssitBBS akD
critical pressures are equal or very nearly so ; the ratio of the boUing
points on the absolute scale is the same at all equal pressures, and
consequently, dp/dt.Th&R the same value for both snbetanoeB at all
equal pressures ; again, the ratio of the actual to the theoretical density
at the critical pressure, and at other '* corresponding," and therefore in
this case equal, pressures, is the same for both.
Up to the present time^ however, although the vapour pressures of
many pairs of liquids have been determined, the substances are, so far
as I know, with one exception less closely related than the hal<^en
derivatives of benzene, and, probably with that exception, the relation
between the vapour pressures and molecular composition is not repre*
sented by a straight line.
!rhe exceptional case referred to is that of ethyl bromide and
ethyl iodide, mixtures of which were investigated by Guthrie
(foe. eU.).
The vapour pressures of ethyl chloride and ethyl bromide have been
determined by Regnault up to about 5000 mm., and of ethyl iodide up
to more than 500 mm., and within these limits, at any rate, the boiling
points of any two of the liquids on the absolute scale bear a constant
ratio to each other at all equal pressures, and it seems not improbable
that, like the halogen derivatives of benzene, their critical pressures
may be equal.
In the following table are given the vapour pressures at
16*7° observed by Guthrie, and those calculated from the formula
F « ^^-^ ^ ^^' ^, taking p, firstly, as percentage by weight,
and secondly, as molecular percentage oi
' ethyl bromide :
Vapour pressures.
Percentages of
ethyl bromide.
Calculated for p = percentage.
Obseryed.
1
By weight.
Molecular.
By weight.
A Molecular.
A
mm.
mm.
1
mm. mm.
mm.
100
100
462-2
462-2
0 462-2
0
90
92-79
428-2
423-3
^ 4-9 481-4
+ 8-2
80
86-12
406-2
894-4
-10-8 1 409-2
+ 40
70
76-94
380-4
366-6
-14-9 , 386-6
+ 61
60
68-21
360-9
336-6
-24*8
860-8
- 0-6
60
68-87
382-3
307-7
-24-6
833-4
+ 11
40
48-81
306-4
278-8
-27-6
304-2
- 2-2
80
38-00
276-4
249-9
-26*6 273-0
- 8-4
20
26-34
246-9
221-0
-26-9
239-8
- 73
10
13-71
214-8
192-1
-22-7
202-8
-120
0
0
163 2
163-2
0
168-2
0
Digitized by
Google
BOILlKa POINTS OF MIXED LIQUIDS. PARt I. 77l
There can be no question that the formula in which p » molecular
percentage represents the results very much better than the other, and,
with this meaning, although there is an individual deviation of 12 mm.,
the agreement between the calculated and observed pressures is fairly
satisfactory, especially about the middle of the table, where the greatest
differences would be expected.
Considering how difficult it is to obtain a mixture of known com^
position quite free from air, the deviations from the calculated values
may perhaps be regarded as within the limits of experimental error.
Some allowance may perhaps also be made for small fluctuations of
temperature. As regards volume change on mixing, Guthrie's specific
gravities are given only to two places of decimals, and therefore throw
no light on the question whether expansion or contraction occurs.
Experimental.
The experiments described in this paper were carried out in con jane*
tion with Miss Fortey.
Fresh quantities of chlorobenzene and bromobenzene, obtained from
Kahlbaum, were fractionated with a 12 bulb "pear" still-head, and
the sp. gr. at 0^/4^ were determined. The agreement of the new
results with those obtained by one of us in 1889 (Trans., 56, 487) is
satisfactory :
Sp. gr. at 074°
New results. Old results*
Chlorobenzene 112806 1*12786
Bromobenzene 1-52178 ' 152182
A mixture of the two liquids in nearly molecular proportion,
11*2717 grams of chlorobenzene and 15*6960 grams of bromobenzene,
was then made, and its sp. gr. at 0^4° was determined. It was
noticed afterwards that the chlorobenzene from which a portion had
been removed to make the mixture became slightly turbid when cooled
to 0°, and therefore contained a trace of moisture, and it was possible
that the mixture might also have been slightly moist. The sp. gr. of
the moist chlorobenzene was therefore determined, and was found to be
1*12787 at 074°. The sp. gr. of the mixture was then calculated on
the assumption that no change of volume occurs on mixing, taking the
chlorobenzene in the mixture to be (a) dry, {b) moist :
Sp. gr. qf mixture at 074^
Observed 1*32798 Calculated a. 1*32804
I. 1*32793
Digitized by
Google
772
YOtJKa: THE VAPOUR PRESSURES AND
A second determination was made, again with nearly molecular
quantities of the liquids :
Weights taken.
CeHjCl 11-4040
CeHgBr 169930
Sp. gr. at 074*,
Observed.
1-32860
CEtlculated.
1-32873
From these results, it appears that there is no perceptible change of
volume on mixing the two liquids.
An experiment was next made to find whether any heat change
occurs when the two liquids are mixed together in molecular propor-
tion. The bromobenzene, 15*6 grams, was weighed in a small, round-
bottomed flask, and the chlorobenzene, 11*2 grams, in a small beaker.
The temperatures of the two liquids were determined, and the chloro-
benzene was then poured into the flask, the mixture shaken, and the
temperature again read. The results obtained are as follows :
Temp.
Chlorobenzene.
1706°
Bromobenzene.
17*00°
Miztore.
17-02°
There is therefore no perceptible evolution or absorption of heat on
mixing the two liquids.
These results indicate that it would be impossible to find two liquids
more likely to behave normally than chloro-
^^®* ^* benzene and bromobenzene.
Vapour Preaswrea and Boiling Points o/MixtureM
of Chlorobenzene and Bromobenzene.
Owing to the difficulty which is always ex-
perienced in obtaining a mixture of two liquids
in known proportion entirely free from air and
moisture, it was decided to employ the dynamical
and not the statical method. The apparatus
used is shown in Fig. 1. It consists of a
bulb of about 156 c.o. capacity with a wide
vertical tube, to which is sealed a narrow side
tube cooled by water to act as a reflux con-
denser. The upper end of the side tube is con«
nected with an exhaust and compression pump
and a differential gauge* The wide vertical
tube is provided with a well-fltting cork through
which passes a rather narrower thin walled
tube, which has a hole blown in it just below the
cork. This narrower tube is also fitted with a cork, through which
passes the thermometer.
Digitized by
Google
BOILING POINTS OF MIXED LIQUIDS. PART I. 773
It is important that the volume of vapour should be as small as
possible relatively to that of the liquid, and the bulb is therefore filled
to about two-thirds of its capacity with liquid. The thin walled tube
is pushed down until the bottom of it is about 3 mm. above the surface
of the liquid when cold, and the bottom of the thermometer bulb is
about level with the bottom of the thin walled tube.
This arrangement possesses the following advantages : (1) The
liquid that returns from the reflux condenser cannot come near the
thermometer, and the amount of liquid which condenses on the thermo-
meter and on the inner walls of the thin walled tube is exceedingly
small ; on the other hand, with the large quantity of liquid which is
present and the small flame that is required there is no fear of the vapour
being superheated. (2) It is possible to take readings of the tem-
perature both of the vapour and of the boiling liquid without altering
the position of the thermometer, for whea the burner is directly below
the centre of the bulb, the liquid boils up into the thin walled tube
well above the thermometer bulb, but when the burner is moved a
little to one side, the surface of the liquid immediately below the thin
walled tube remains undisturbed and the liquid does not come in con-
tact with the thermometer bulb.
The apparatus was tested with pure chlorobenzene with the follow-
ing results :
Temperature.
, * ^ Calculated from
Pressure. Vapour. Liquid. A Biot's formula.
743-05 131-U° 131-54° 040 13116°
728-20 130-37 130-77 040 130-41
The temperatures of the vapour agree well with those calculated
from the constants for Biot's formula (Trans., 1889, 55, 487), and the
agreement is even better if we calculate from the obsexved boiling
point of the new sample of chlorobenzene. The liquid is evidently
somewhat superheated.
The boiling points of three mixtures of chlorobenzene and bromo-
benzene were then determined through a range of about 100 mm.
The temperatures of both vapour and liquid were read in each case,
and it was found that for each mixture there was an almost constant
difference between the two. The mean difference for each mixture
(0-50°, 0-61°, and 0-57° respectively) was subtracted from the tempera-
tures of the liquid, and the mean of the value so obtained and the tem-
perature of the vapour was taken in each case as the true boiling
point.
The experimental results are given in the table below :
VOL. LXXXI. 3 F
Digitized by VjOOQIC
774
TOUNQ: THE VAPOtJB PRESSURES AND
BaUing PoirUs of Mixtures of Chlorgbenzens and Brcmobenzene,
Molecular percentages of bromobenzene.
2501.
5000.
78-64
PresBure.
Temperature.
Pressure.
Temperature.
Pressure.
Temperature.
mm.
mm.
mm.
685-0
182"86°
687-4
188 -49'
685-4
144-14°
694-4
188-84
692-6
188-78
697-2
144-81
708-1
188-85
699-0
189-07
706-9
145-88
714-4
184-89
702-9
189-26
7190
145-99
725-2
134-97
708-6
189-55
728-1
146-51
784-4
185-48
715-5
189-90
788-4
147-02
789-8
185-72
721-1
140-18
745-5
147-41
740-5
185-76
728-0
140-55
755-5
147-94
741-5
185-80
786-6
140-98
765-7
148-44
749-8
186-24
742-6
141-27
776-5
149-00
758-4
186-69
751-4
141-72
788-1
149-60
768-8
187-20
758-4
142-10
779 0
187-70
765-7
142-48
789-0
188-18
775-4
142-87
~~
—
From these data, curves were constructed and the boiling points were
read off at 700, 730, 760, and 790 mm. pressure.
Pressure.
Molecular
percentage of bromobenzene.
25-01
50-00
78-64
mm.
700
780
760
790
188-66°
185-22
186-75
138-22
189-10*
140-67
14216
148-60
144-97°
146-59
148-16
149-67
If, now, for each pressure we plot the temperatures against the
molecnlar percentages of bromobenzene, including the boiling points
of the pure substances (j9»0and 100), the four isobars bo obtained
will be obviously curved, but the number of points is not sufficient to
enable us to draw the curves accurately.
In order, however, to find whether the results are in agreement with
the proposition that the isothermal s, representing the relation between
vapour pressure and molecular composition, are straight lines, we may
proceed in either of two ways.
Digitized by
Google
BOILING POINTS OF MIXED LIQUIDS. PAKT I.
775
1. From the formula P-
p.P^ + (100 -p)P,
100
we get p'
lOO(f.-P)
where p is the molecular percentage of bromobenzene and 2*^ and F^ are
the vapour pressures of pure bromobenzene and chlorobenzene respec-
tively at the boiling point of the mizture, and F is the pressure under
which the mixture is boiling. We may thei) calculate from the formula
the percentage of bromobenzene for the values 700, 730, 760, and 790
of JP at a series of temperatures, and so obtain the theoretical isobars.
It can then be seen whether the observed temperatures fall on them.
For the vapour pressures of bromobenzene, from 130° to 156°^ the
Fio. 2.
156
IKO"
'
/
rU6'
^
1^
^
^
J^^
^
140**
^"^
^A
r
186''
<^^
^
.'-"'^"^
180"
10 20 80 40 50 60
Molecular percentage of CfHsBr.
70
80
values given by Ramsay and Young (Trans., 1885, 47, 640) have been
employed ; the vapour pressures of chlorobenzene for the same temper-
atures have been calculated from the constants for Biot's formula
(Young, he. eit,), though for these nearly related substances, since the
ratio of the boiling points on the absolute scale is a constant at all
equal pressures, it would be sufficient to know the value of this ratio
and the vapour pressure of one of the two substances.
The theoretical isobars are given in Fig. 2, and the experimental
values are indicated by circles ; it will be seen that the agreement is
very good.
2. We may calculate the vapour pressures of the two pure liquids at
3 F 2
Digitized by
Google
776 VAPOUR PRESSURES AND BOILING POINTS OF MIXED LIQUIDS,
the boiling points of the mixtures, and then, from the equation
B
p,P^ + (100 -p)P^
100
, we may calculate the theoretical values of P.
In the following table are given the vapour pressures of chlorobenzene
and bromobenzene at the boiling points of the mixtures, and the
values of F calculated by means of the above formula :
Molecular
percentage
of CjHsBr.
t.
Vapour preesures at f.
P calc.
Observed
pressure.
A
CgHsBr.
CeHBCl.
25*01
50-00
73-64
183*66'»
13910
144*97
135-22
140-67
146-59
186*75
142-16
148-16
138*22
143-60
149-67
414-6
483-7
568*05
433-6
605-2
593-25
452-85
526-25
618*4
471-85
547-35
643-55
796-05
917-5
1065-4
Mean...
828-8
955-35
1109-3
Mean...
862*95
992-3
1153 0
Mean...
896-65
1029-25
1196-5
Mean...
699-9
700-6
699 15
699-9
729-95
730*3
729-3
729-85
760-4
759*3
759-3
700-0
>f
If
730*0
>>
>i
760*0
ft
»»
790-0
>>
-0*1
+ 0*6
-0*85
25-01
50-00
78-64
-0*1
-0-05
+ 0*8
-0*7
25 01
50-00
73-64
-016
+ 0*4
-0*7
-0-7
25-01
50*00
73*64
759*7
790*4
788*3
789-3
-0*3
+ 0*4
-17
-0*7
789-3
-0*7
The differences between the pressures calculated from the formula
and those under which the mixtures boiled are quite within the limits
of experimental error ; the experimental results therefore prove that
for the two liquids, chlorobenzene and bromobenzene, which are
chemically so closely related, and which, in many ways, exhibit such
simple physical relations, the vapour pressures of a mixture at any
temperature are accurately expressed by the formula
p^p.F^ + {100 -P)P3
• Too '
or, in other words, the relation between the vapour pressures and the
molecular composition of mixtures at any temperature is reprosented
graphically by a straight line.
It is proposed to extend the investigation to mixtures of other liquids,
Univbbsity Colleos,
Bristol.
Digitized by
Google
CORRECTION OF THE BOILING POINTS OF LIQUIDS. 777
LXXVIII. — Correction of the Boiling Points of Liquids
from Observed to Normal Pressure.
By Sydney Youno, D.Sc., P.RS.
In papers published jointly with Dr. Ramsay (Brit, Assoc. Eep,, 1885,
928 ; Fhil. Mag., 1885, [v], 20, 515), it was pointed out that (1) the
values of dp/dt. T (where dpfdi is the rate of increase of vapour pres-
sure per unit rise of temperature, and T is the boiling point on the
absolute scale) are approximately the same for different substances
when compared at the same pressure, but that (2) the differences are
real and are not due to errors of experiment, for they preserve the
same relative value whatever the pressure at which the comparison is
made, at any rate within the limits of the actual experiments. It was
further pointed out that for two closely-related substances, the boiling
points on the absolute scale bear a constant ratio to each other at all
pressures, and that for other substances {JPhU. Mag., 1886, [v], 21, 33)
the relation between the boiling points may be expressed by an equa-
tion which can be written thus :
where Tj^' and T^ are the boiling points of the two substances on the
absolute scale at a pressure, p\ T^^ and T^ the boiling points at a
pressure, p, and c is a very small constant.
In the light of these generalisations, Crafts (Bw., 1887, 20, 709)
devised a convenient method for the correction of the boiling points
of substances from observed to normal pressure.
Crafts gives the boiling points (absolute temperatures), the values of
A</A/7 between 720 and 770 mm. pressure, and the quotients
-yff^H for 25 substances, and points out that when the boiling
point of any liquid is to be corrected, the constant, H, for that com-
pound in the table most closely related to it is to be taken and is to be
multiplied by the approximate absolute boiling point of the substance
under examination, in order to find the value of A^/Ap for a barometric
variation of the 50 mm. between 720 and 770 mm.
Nemst {Thsoretische Chemiej 1893, p. 55) reproduces this table,
but divides the constants, H, by 50, so as to give the mean varia-
tion of temperature per mm. change of pressure between the same
limits.
■s. Digitized by
Google
778 TOUNQ: CORRECTION OF THE BOILING POINTS OF
There are a few misprints in Crafts' table which reappear in the
modified table given by Nernst, and since 1887 the vapour pressures of a
considerable number of additional compounds have been accurately
determined ; it may therefore be useful to give a revised and enlarged
table for reference.
Instead of calculating the mean value of the constant for a range of
pressure between 720 and 770 mm., I have thought it better to give
cU 1
the actual value of -r- • — « (7 at 760 mm., and therefore the constants
dp T
in this paper differ slightly from those in Nemst's table ; in most
cases, the new values are lower by 2 in the last place, and this may be
taken as the average difference between the value of C at 760 mm. and
the mean value between 720 and 770 mm.
In calculating the value of dp/dt at the boiling point, Biotas formula,
logp =^a + ba^ + c^, was employed when the constants for this formula
had already been ascertained ; in other cases, for which the vapour
pressures have been determined through a sufficient range of temper-
ature, constants for the simpler formula, logp=^a + ba^, were calculated
from the vapour pressures at three temperatures 20^, 30^, or 40° apart,
the middle temperature being near the boiling point under normal
pressure. This method was adopted, for example, in the case of the
29 esters of the methyl formate series for which accurate data are
available.
For the elementary gases, and for carbon monoxide and methane,
the vapour pressure data are not sufficient for this method to be
employed, and the formula, '^, = ;=;' +c(2'^'-^^), was made use of,
benzene beiog taken as the standard substance. The constant, c^ was
first calculated from the boiling point and critical temperature of the
gas and the boiliog point of benzene under normal pressure and at a
pressure equal to the critical pressure of the gas, the assumption being
made that Biot's formula might be employed for benzene for some
little distance above its actual critical point. It is probable that the
boiling points and critical constants of the gases are less accurately
known than the vapour pressures of the majority of substances
included in the tables that follow, and as the method of calculation of
djp\di is also less direct, the values of C are probably less accurate. The
constants determined in this way are marked with an asterisk.
Digitized by
Google
LlQtnDS I'BOH OBSERYISd TO NORMAL PRBSSltRE.
779
dp
, <a 760 mm.
Name of subeUnce.
Oxygen *
Nitrogen i.... *
Argdn *
Krypton *
Xenon *
Chlorine
Bromine
Iodine
Mercnry
Snlphur
Carbon monoxide... *
Snlphur dioxide
Ammonia
Carbon disulphide ...
Boron trichloride
Phosphorus trichlor-
ide
Silicon tetrachloride
Stannic chloride ...
Methyl chloride ...
Chloroform
Carbon tetrachloride
Methane
fi-Pentane
f^Uexane
nHeptane
n-Octane
tfoPentane
Diifopropyl
Diwbutyl
eycZoHexane
Benzene
Toluene ,...
Naphthalene
Anthracene
"m-Xylene
Diphenylmethane .
Triphenylmethane .
Flttorobenzene
Chlorobenzene
Bromobenzene
lodobenzene ,
Bromonaphthalene .
Ethylene dibromide
Methyl ether
Ethyl ether
Acetone
Boiling
point
(abs.temp).
90 -S*
77-6
86-9
121 S
163*9
239-4
881-76
458-3
629-8
721-4
83-0
262-9
240-1
819-25
291-25
846*85
829-9
887 1
249-35
883-2
849-75
1090
68 2
809-3
25 8
841-95
23-9
871-4
22-8
898-8
21 1
800-95
26-2
881-1
24*3
382 1
20 9
353-9
22-7
368-2
28-3
883-7
21-8
491
17-1
616
16 0
412
21 1
588
16-1
626
14-8
358-2
288
406-0
20-5
429*0
19-8
461-46
18-0
658-46
16-76
405
20-8
249*4
807-6
880
dpjdi
76-9
89-0
88-2
61-8
46-9
88-2
25-2
18-76
18-4
42*2
81-8
88-7
87 7
24 7
26-8
28 46
24-0
21-4
81 9
26-2
28-26
82-0
26-9
26*4
C.
0000146
0-000146
0-000138
0 000138
0 000138
0 000126
0*000120
0-000116
0 000118
0000114
0-000148
0-000118
0000110
0 000127
0 000128
0 000128
0 000126
0*000121
0*000126
0*000119
0-000128
0000186
0-000126
0-000122
0-000121
0*000119
0*000127
0-000124
0-000125
0-000124
0-000122
0*000120
0*000119
0*000108
0*000115
0 000123
0*000108
0*000120
0 000120
0-000120
0 000120
0*000115
0*000119
0000126
0-000121
0000116
Obeenrer.
Olszewski.
Baly.
Ramsay and Travers.
Enietsch.
Ramsay and Young
Young.
Regnault.
Olszewski.
Regnault.
Young.
Olszewski.
Young.
Young and Fortey.
Young.
Young and Fortey.
Crafts.
Young,
Ramsay and Young.
Crafts.
Regnault
Ramsay and Young.
Crafts.
Digitized by
Google
780 TOUNQ : CORRECTION OF THE BOILING POINTS OF
Values of 0^1} €Ui 760 mm. (continued).
Name of snbstance.
BeD zophenone
Dibenzylketone
Anthraquinone
Aniline
Qainoline
Methy] formate
Ethyl fonnate ,..
Methyl acetate
Propyl formate
Ethyl acetate
Methyl propionate ..
MoButyf formate
Propyl acetate ,
Ethyl propionate
Metnyl butyrate....
Methyl isobutyrate..
Amyl formate
iffoButvl acetate
Propyl propionate ..
Ethyl butyrate
Ethyl ifobutyrate ..
Methyl valerate
iffoButyl propionate
Propyl butyrate
Propyl Mobutyrate ...
MoPropyl Mobutyrate
Ethyl valerate
Amyl propionate
ifo Butyl butyrate ..
isoButvl iffobutyrate
Propyl valerate
Amy! butyrate ,
Amyl ifobutyiate .
t«o Butyl valerate
Methyl alcohol
Ethyl alcohol
Propyl alcohol
Amyl alcohol
Phenol
Acetic acid
Phthalic anhydride.
Salphobenzide
Water
Boiling
point
(abs.temp).
678-8'
603*55
650
457-4
510 5
304-9
327-8
3301
353 0
350 15
352-7
870-85
374-55
372-0
375-75
365-8
396-25
889-2
895 15
892-9
383-1
389-7
409-8
415-7
406-9
393-75
407-8
483-2
429-9
419-6
428-9
451-6
441-8
441-7
837-9
851-8
370-4
403
466
891-5
559
652
873 0
dpjdt.
15-8
16-2
18-6
19-6
17-0
28-8
26-6
26-8
24-6
25-1
24*9
23-45
23-5
23-7
23-8
28 8
21-8
22-6
22-8
22-8
22-5
22 4
21-4
20-9
21-8
22 0
21-4
20-4
20-6
20-6
20-6
19-6
19-6
19 9
29-6
30-85
28-8
26-8
20-6
28-9
16-0
15-2
27-2
0-000109
0-000109
0000113
0-000112
0-000115
0-000114
0-000115
0-000118
0 000116
0-000114
0-000114
0-000115
0-000114
0-000118
0 000114
0-000115
0-000115
0-000114
0-000114
0-000114
0 000116
0000114
0 000114
0-000115
0-000118
0-000116
0-000115
0 000113
0000113
0-000116
0-000114
0-000118
0 000115
0 000114
0 000100
0 000094
0-000094
0-000098
0-000107
0-000107
0-000112
0-000101
0-000099
ObMTrer.
Crafts.
Young.
Orafte.
Ramsay and Toung.
Young.
Young and Thomas.
Schumann.
Young and Thomas.
Schumann.
Young and Fortey.
Schumann.
Ramsay and Young.
Crafts.
Young.
Crafts.
Regnault.
It will be seen that the values of (7 ( x 10«) vary from about 146 in the
case of oxygen, nitrogen, and carbon monoxide to 99 fop water and 94
or ethyl and propyl alcohols.
Digitized by
Google
LIQUIDS FROM OBSERVED TO NORMAL PRESSURE. 781
The following relations are clearly observable :
1. In most groups of similar substances the greater the molecular
complexity the lower is the constant. This is well seen in the case of
the normal paraffins ; of benzene, naphthalene, and anthracene ; of
methyl and ethyl ether, and of acetone and benzophenone or dibenzyl
ketone.
2. Kise of molecular weight without increased complexity either
causes a similar change or has no influence ; thus, for the halogens
C diminishes from chlorine to iodine ; but, on the other hand, it has
the same value for the four halogen derivatives of benzene.
3. Comparing isomeric substances, it is seen that C is almost always
lower for normal than for t«o-compounds, and lower for these than for
compounds that contain two iso-groups.
4. By replacement of hydrogen by a halogen, the constant is
lowered; thus benzene, 122, chlorobenzene 120; naphthalene 119,
bromonaphthalene 115 ; ethane, above 125, ethylene dibromide, 119.
5. When there is association of molecules in the liquid state, the
value is low ; this is the case with water, the alcohols, phenol, and
acetic acid, all of which contain a hydroxyl group.
6. In the case of organic hydroxyl compounds, the influence of the
hydroxyl group diminishes as the complexity of the organic radicle
increases, and thus the depression of the constant diminishes, but, on
the other hand, the constant tends to fall as the complexity of the
molecule increases. There are thus two opposing influences, and in
the group of alcohols the constant shows considerable irregularity.
In the case of the esters, there is evidence of some molecular asso-
ciation in the liquid state and at the critical point, and we find similar
opposing influences, with the result that the variation in the constant
is very small ; indeed, for the whole 29 esters the extreme values are
113 and 116. Still, if we confine our attention to the higher esters con-
taining, say, 5 or more carbon atoms, (1) the influence of molecular
complexity, and (2) of constitution, is, with few exceptions, to be
observed, and in the table on p. 782 the results are summarised to
bring out this point :
In order to correct a boiling point from observed to normal pressure,
At is to be added, where
Ae=(760-j9)(273 + 0'C'.
The constant C is to be ascertained from the preceding tables,
taking account, if necessary, of the relations (1) to (6) ; j9 is the
observed pressure and t is the boiling point as observed or^ better,
roughly corrected.
In a series of papers {Phil, Mag,, 1892, [v], 34, 503; Trans., 1893,
63, 1254 ; PhU. Mag,, 1894, [v], 37, 1 ; ibid., 1900, [v], 50, 291), I
Digitized by
Google
782
CORRECTION OF TflE BOILING POINTS OF UQXTlDS.
Fonnala
of
Nonnal.
One Mo-group.
Two ito-gronpB.
ester.
C.
Mean.
a
Mean,
a
Mean.
C,HA
0-0001139
0 0001139
C,H.O,
0-0001 149\
0-0001180/
0-0001151)
0-0001140
—
—
—
—
C4H3O,
0-0001138 ■
0-0001136
0-0001142
—
-^
—
-=■
o,n,fi.
0 0001142
0-0001134 -
0 0001 136 J
0-0001137
0-0001150\
0-0001160/
00001165^
00001150
—
—
CeHijOa
0-0001137 \
0-0001141/
0-0001139
0-0001189 1
0 0001158 (
0-0001 143 J
0-0001140^
0-0001149
■"•
C*Hi.Oo
0-0001148
0-0001148
00001147 -
0-0001138
00001156
0-0001155
O-OOOII27I
0-00011291
CaH,«Oo
—
0 0001135 ■
0-0001183
0-0001166
0-0001156
0 0001 136 J
C^HisOa
—
0-0001134
0-0001134
0-00011521
0 •0001138/
0-0001145
Mean...
0-0001140
0-0001142
0-0001150
have shown that, excluding compounds which contain a hydroxyl group,
the ratio, y^, of the actual to the theoretical density for a perfect
gas at the critical point is approximately constant, the mean value
being about 3*76, but that the deviations are real ones and are
related to the molecular weight and constitution of the substances.
I have also shown that Cailletet and Mathias' law of the "dia-
meter " is not in most cases absolutely true, although very nearly so,
and that the relation of the mean density to the temperature should be
expressed by the formula I)t = I)Q + at + pfi, where a is always negative,
and p changes from a very small positive value, through sero, to a
very small negative value as the ratio — f increases.
Again taking the approximate formula 2>t = Z)^ + ai, originally given
by Cailletet and Mathias, it has been proved independently by
Mathias and myself that, if the generalisations of Yan der Waals were
strictly true, — should be a constant for all substances ; but I have
shown that there ore real though not large deviations in this case also,
the value of — S rising in general as —-^ increases*
Digitized by
Google
YOUNG AND FORTEY : ISOPROPYL ISOBUTYRATE. 783
It is of interest to observe that, as a rule, Cl = -, i at normal
V dpT)
pressure diminishes as -p rises.
So far, then, as the 26 substances which I have myself examined
are concerned, there is clearly a connection between the four constants
%, A ^;. and C.
The critical densities of the gaseous elements are not known with
sufficient precision to allow of any very definite statement being made
regarding the value of --p, but it is interesting to note that D.
Berthelot points out that for carbon dioxide this constant is lower
(3*6) than for any of the substances I have investigated, whilst for
oxygen and nitrogen it appears to be about 3*5.
The constant C ( x 10^) for the two last named gases is about 145,
and if a diagram is constructed to show the relation between 0 and -p
and these values, (7^' 145 and ^s 3*5 are mapped with the others, in-
eluding also those of ethyl and propyl alcohols, the connection between
C and 7p becomes quite evident, the former falling as the latter rises.
Methyl alcohol and acetic acid have low, but not proportionately low,
values of the constant C.
UNIVBRBmr COLLBOB,
Bristol.
LXXIX. — Va'pour Pressures and Specific Volumes of
iaoPropyl isoButyrate.
By Sydney Youno, D.Sc., F.R.S., and Emily C. Foetey, B.Sc.
In the preparation of diwopropyl by the electrolysis of potassium iao-
butyrate, it was found that the yield of hydrocarbon was exceedingly
poor, but that, on the other hand, a considerable quantity of t«opropyl
tsobutyrate was formed.
The t«obutyric acid was carefully fractionated before its potassium
salt was electrolysed, and the i«opropyl tsobutyrate was therefore
easily purified ; after three fractionations, it boiled quite constantly at
120*75^ under normal pressure.
A second specimen of the ester was prepared by saturating a mix-
Digitized by
Google
784
YOUNG AND FORTEY : VAPOUR PRESSURES AND
ture of pure t«obutyric acid and t^opropyl alcohol with hydrogen
chloride in the usual manner. The boiling point was found to be
identical with that of the ester obtained by the electrolytic method.
The sp. gr. of both specimens were determined at 0^, and that of
the second specimen at 21*35°, with the modified Sprengel tube. The
weighings were reduced to a vacuum. As the ester, when even slightly
moist, rapidly acquired an acid reaction owing to hydrolysis, it was
found necessary to remove the free acid from time to time by means
of sodium hydroxide, and the free alcohol and water by means of
phosphoric oxide. The sp. gr. and boiling point were determined after
each of these operations.
Specific gravities.
Ester obtained by electrolyslB. Edter obtained from acid and alcohoL
At 0° 0-86867 At 0° (1)
0-86874
(2)
0-86874
At 21-36°
0-84708
Mixture at 0° (1) 0-86874
(2) 0-86873
Mean sp. gr. at 0° = 0-86872
Boiling points.
Ester obtained by electrolysis. Ester obtained from acid and alcohoL
Tei
opera ture
corr. to
Temperature
corr. to
Pressure. obs.
760 mm. Pressure.
obs. 760 mm.
753-0 mm. 120-45°
120-77° 756-15 mm.
120-6° 120-77°
— —
— 763-0 „
120-9 120-72
— —
- 755-4 „
120-55 120-76
Mean 120 75
Mixture of the esters.
Temperature.
Pressure.
obs. corr. to 760 mm.
761-0 mm
120-80°
120-75°
771-3 „
121-32
120-81
763-9 „
120-9
120-72
738-8 „
119-77
120-73
747-6 „
120-27
Mean
120-83
120-77
General mean « 120-76°.
The value of dp/dt^ calculated from Biot's formula, is 22*0 mm. per
degree.
Digitized by
Google
SPECIFIC VOLUMES OF I80PR0PYL ISOBUTYRATE.
785
Vapour Pressures at Low Temperatures,
For pressures up to 170 mm., the method of Ramsay and Young
was used ; from 150 mm. to the atmospheric pressure, a modified dis-
tillation bulb with a refiuz condenser was employed.
Pressure.
Tem-
Pressure.
Tem-
Pressure.
■ 1
Tem-
Pressure.
Tem-
perature.
perature.
36-7
perature.
91-06
perature.
5-8
7-86''
14-3
28-5'*
40-7*
60 ^^
6-56
10-86
16-56
26-7 '
41-8
43-66
102-66
63-6
7-85
12-96
19-06
28-3
47-65
46-3
115-25
66-15
8-85
14-95
21-56
80-4
54-85
49-4
129-75
69-0
10-0
171
24-8
83-2
62-25
52-0
145-0
71-86
11-26
]9-2
; 28-1
36-66
70-6
54-8
163-0
74-86
12-8
21-4
; 82-15
88-35
80-1
57-66
172-0
76-1
148-4
721
' 246-0
86-55
411-8
100-4
645-9
115-2
168-65
78-9
272-7
88-28
449-9
103-3
703-9
118-15
176-4
76-6
806-7
91-75
498-1
106-2
770-7
121-3
106-86
79-26
889-0
94-8
541-4
109-2
—
—
220-8
82-5
378-6
97-46
590-8
112-1
I
—
—
Vapour Presstvres at High Temperatures.
The vapour pressures at high temperatures were determined with
the pressure apparatus employed in previous researches. It would
have been of interest, if possible, to ascertain the critical constants of
t^opropyl idobutyrate, on account of the presence of two i^o-groups in
the molecule, but, unfortunately, it was found that the ester began to
decompose at about 230°, and that at 260° the decomposition was so
rapid that it was quite impossible to obtain trustworthy results. The
data above 230° are therefore not given, as they would only be mis-
leading.
The observed pressures from 130° to 230° (the mean of four read-
ings in each case), together with those read from the curve constructed
from the observations at low temperatures, are given in the following
table:
Vapour pressures :
Dynamical method
Temperature.
from curve.
Temperature.
SUtical method
10°
6-65 mm.
130°
994 mm.
20
11-95 „
140
1290 „
30
20-9 „
150
1649 „
40
35-05 „
160
2096 „
60
66-9 „
170
2612 „
Digitized by VjOO^
786
YOUNG AND FORTEY: ISOPROPYL ISOBUTYRATE.
Dynamical method
Temperature.
from carve. Temperature.
Statical method.
60°
89-0 mm
180°
3216 mm.
70
1360 ,
190
3916 „
80
201-1 ,
200
4742 „
90
289-4 ,
, 210
5700 „
100
406-0 ,
220
6785 „
110
653-4 ,
230
8058 „
120
743-9 ,
Volumes of a Gram qf Liquid,
These were determined in the pressure apparatus. At 220^ and
230^, they were calculated from observations of the volume of vapour
and the total volume of liquid or vapour (Trans., 1893, 03, 1200) ;
the remaining volumes were read directly :
Volumes qf a Gram and Moleetdar Volumes qf Liquid.
Yolumes of a gram
Mole-
Volumes of a gram
Mole-
Tem-
m c.c.
cular
Tem-
m cc.
cular ]
perature.
volume
perature.
volume
Observed.
From
curve.
m c.c.
Observed,
From
curve.
in cc.
0'
11511
1-1511
149-81
120'
1-8540
1-8589
176-62
10
1-1646
151-06
130
1-8769
1-8767
178-67
20
1-1788
152-90
140
1-4010
1-4005
181-66
30
1-1989
1-1934
154-79
150
1-4266
1-4257
184-93
40
1*2092
1-2085
156-75
160
1-4521
1-4628
188-44
50
1-2240
1-2243
158-80
170
1-4814
1-4818
192-14
60
1-2411
1-2405
160-91
180
1-5125
1-6128
196-22
70
1-2675
1-2573
168 08
190
1-5460
1-5462
200-66
80
1-2746
1-2749
165-87
200
1-6841
1-6846
205-52
90
1-2923
1-2934
167-76
210
1-6270
1-6278
211 08
100
1-8128
1-8126
170-26
220
1-6828
1-6765
217-46
110
1-3325
1-8328
172-88
230
1-7827
1-7331
224-80
The volume of a gram of saturated vapour was only determined
satisfactorily at 230^. The result obtained was as follows :
Volume of a gram 21*05
Molecular volume 224-75
University Collbge,
Bristol.
Digitized by
Google
INFLUENCE OF THE METHYL GROUP ON RING FORMATION. 787
LXXX. — Influence of the Methyl Group on Ring
Formation,
By A. W. GiLBODY and C. H. G. Spbanklinq.
This research was undertaken with the twofold object of studying the
antipyretic effects obtained by successively introducing methyl groups
into some substance already possessing the property of producing such
effects, and of ascertaining what influence these methyl groups would
have on the stability of the mother substance.
Pyrantin (/>-ethoxyphenylsuccinimide) was chosen as a basis because
it had already been shown to be an excellent antipyretic (Piutti, Ber.y
1896, 29, 85), and by using methylsuccinic acids or their anhydrides
instead of succinic acid itself for the preparation, methyl groups could
readily be introduced into the imide ring.
The first study was unsuccessful because the pyrantin derivatives
were almost insoluble in weak (0*75 per cent.) sodium chloride solu-
tion at 30^, the solubilities ranging from 1 : 713 for pyrantin to
1 ; 1272 for trimethylpyrantin.
As another source of antipyretic substances, it was then attempted
to prepare the sodium salts of the intermediate methyl-substituted
j7-ethozyphenylsuccinamic acids containing an open chain, as for
example, ct9-«-dimethy]-j9-ethoxyphenylsuccinamic acid (cid-«-dimethyl-j9-
ethoxysuccinanilic acid), but it was found that these sodium salts are
unstable in aqueous solution when several methyl groups are pre-
sent, owing to the great tendency to ring formation and consequently
the study of the antipyretic action of the methyl group was abandoned.
The second object of the research was more successful, as the authors
were able to ascertain the stability constants of the methylpyrantins
which they had prepared according to the method devised by Miolati
and his colleagues for the alkyl-, phenyl-, tolyl-, and xylyl-succinimides,
where the radicle mentioned was substituted for the hydrogen atom
attached to the nitrogen of the succinimide ring (compare Miolati and
Longo, Atii E. Accad, Lincei, 1894, [v], 3, 515 ; 1895, [v], 4, 351 ;
Miolati and Lotti, 1896, [v], 6, 88). The results of the experiments
were then calculated from the formula used by Hantzsch and Miolati for
the measurements of the stability of the oxazolone ring, CJ^H^"^ *\l
1 X
namely, jj • . _ =Ac {Zeit. phyaikaL Cheney 1893, 11, 748).
The reactions by which these substituted pyrantins were prepared
took place in two stages, the open chain succinamic acid being first
Digitized by
Google
788 GILBODY AND SPRANKLING : INFLUENCE OF THE
prepared by the action of j9-pheDetidine upon the methylsaccinic
anhydride, using benzene or toluene as the solvent :
CHR-CO _ 9HR.C0.NH-C,H,.0Eb
where R = a methyl group.
The substituted succinamic acid thus obtained was then heated
alone above the melting point, when water was readily 'split off
according to the equation :
CHR-CO-OH CHR-CCr « *^ ^ «
Theory requires that some of the substituted succinamic acids
should exist in two isomeric forms, as, for example, methyl-/? ethoxy-
phenylsuccinamic acid (methyl-;^ethoxysuccinanilic acid),
CHj-CH-CO-NH-OgH^-OEt CHj-CO-NH-OgH^-OEt,
CHg-COjjH ' CHj-CH-CO^H
but in the authors' experiments only one acid has been obtained, and
they are unaware that any such isomerides have been prepared even
in the case of the corresponding methylsuccinanilic acids.
Pyrantin and cM-dimethylpyrantin can only exist in one form, but
the other methyl py ran tins which have been prepared, namely, methyl-,
tran8,8{f)-dimeihj\'y a<-dimethyl-, trimethyl-, or a-tsopropyl- pyrantin >
should exist in two forms.
The authors have, however, been unable to determine experimentally
which of the two possible formulas is correct in the case of the methyl-
/>-ethoxyphenyl6uccinamic acid obtained^ but as Blaise {Compt, rend,^
1898, 126, 753) states that a tertiary carboxyl radicle is much
more difficult to esterify than a primary one, it has been decided to
adopt the formula in which the primary carboxyl is united with the
amide group of the aromatic portion of the formula, and the secondary
or tertiary group is free ; thus formula II is taken as being correct
for methyl-/>-ethoxyphenylsuccinamic acid in the absence of evidence
to the contrary.
The pyrantins were all obtained in a pure crystalline condition by
removing any trace of colouring matter with a little glacial acetic
add and then crystallising from alcohol.
The modification in the method of measuring the stability constant
consisted in using a pure alcoholic solution of the alkylpyrantin, on
account of its insolubility in water, but aqueous standard solutions of
caustic soda and hydrochloric acid were retained.
The values obtained for the pyrantins were then converted into those
for the corresponding methylphenylsuocinimides in aqueous fiolution,
Digitized by
Google
METHYL GROUP ON RING FORMATION. 789
the detailed results, together with the conclusions deduced^ being given
after the experimental portion of the paper.
Experimental.
I. Preparation and Broperties of the Succinamic Acids,
i^Ethoxyph^ylauccinamic Acid (p-Ethoxt/euecincmilic Acid),
COaH-CHg-CHg-OO-NH-C^H^-O-CgHg.
Slightly more than 1 mol. (15 grams) of ^phenetidine was added
to 1 mol. (10 grams) of finely powdered succinic anhydride dissolved
in toluene by digestion on the water-bath with a large volume of the
solvent until solution was complete. After standing, much heat was
developed and in a short time crystals separated which were filtered
off, dried on a porous plate, and recrystallised from water, or better
still from dilute alcohol.
The acid separated from alcohol in colourless, rectangular leaflets
of pearly appearance which melted at 166 — 167°. After two or
three recrystallisations, the melting point had not altered. On
analysis :
0-2494 gave 0-5646 00^ and 01458 Hfi. 0 = 60-66 ; H « 6-50.
0-2416 „ 12-4 C.C. moist nitrogen at 13° and 756 mm. N = 6-03.
CijHijO^N requires C - 60*76 ; H = 6-32 ; N = 5 -90 per cent.
/>-Ethoxyphenylsaccinamic acid has also been prepared indirectly by
Piutti (jBer., 1896, 29, 85) f rom |>-ethoxyphenylsuccinimide (pyrantin)
and caustic potash. The substance he obtained in this manner could
hardly have been pure, as the melting point was stated to be 160 — 161°.
Piutti's acid also differed from ours in its behaviour towards chlorine
water, for although we were unable to obtain any coloration whatever,
Piutti states that his acid gave a violet coloration with this reagent. We
therefore conclude that, by the action of caustic potash on pyrantin,
Piutti produced some impurity which he failed to remove entirely.
The acid is soluble in alcohol or acetic acid, but only sparingly so
in dry ether. It can be precipitated by the addition of phenylhydrazine
to its solution in dilute acetic acid and heating for a short time.
The sodium salt (soluble pyrantin) was also prepared, and Piutti' s
previous statement that it is remarkably soluble in water confirmed.
Pearly leaflets separate from its aqueous solution on the addition of
ammonium sulphate. Its aqueous solution immediately gives a white,
curdy precipitate on the addition of silver nitrate and lead acetate, and
a green, amorphous precipitate with copper sulphate. If barium nitrate
be added, crystals slowly separate, and a precipitate also gradually
comes down after the addition of calcium chloride.
VOL. LXXXI. 3 G
Digitized by VjOOQIC
790 aiLBODY AND SPRANKLING : INFLUENCE OP TSIl
Melhyl-^ethoxyphenylsuccinamic Acid {Methyl-^p-^ihoocyaucoinanUie Acid),
C02H-CH(OH3)-CHj-00-NH-OeH4-0-02H5.
This acid, together with the higher homologues described in this
communication, were prepared in a similar manner tcr /^ethozyphenyl-
succinamic acid with the exception that benzene was used as the
solvent instead of toluene. As methylsuccinic anhydride is liquid at
the ordinary temperature, muck less solvent was required than in the
former case.
The acid separated from a 50 per cent, alcoholic solution in groups
of microscopic, colourless needles melting at 149 — 150°. On analysis :
01906 gave 04340 COg and 01156 H^O. C = 6210; H = 6-74.
0-3028 „ 13-6 c.c. moist nitrogen at 20° and 763 mm. N = 5'57.
CijHi^O^N requires C = 6215 ; H = 6-77 ; N=:5-67 per cent.
The sodium salt is readily soluble in water. It differs somewhat
from the previous salt in its action towards reagents. Amorphous
precipitates were obtained with'silver nitrate and copper sulphate. Lead
acetate gave a copious precipitate, which dissolved in large excess, but
came down again on long standing. No precipitate was obtained with
barium nitrate or calcium chloride.
8kBrDim$thyl-]p-ethoxyphenyl8uccinamiG Acid {eiS^Dimeihyl'^p^oasytuecin-
cmUic acid), QO^B.'0{(m^)^'QIL^'00'^^*Q^YL^'0'Qfi^.
This acid was obtained from freshly-prepared liquid cw-dimethyl-
succinic anhydride. It crystallised readily from alcohol in beautiful,
small, colourless prisms melting at 160 — 16P, with a very slight
evolution of gas. On analysis :
01799 gave 0-4170 00^ and 0*1184 H,0. 0 = 63-21. H = 7'31.
0-4120 „ 19-9 C.C. moist nitrogen at 16° and 74 mm. N = 5-36.
^uHiflO^N requires 0-63*39; H = 717i N = 6*28 per cent.
The acid, which is only sparingly soluble in hot water, separated on
cooling in thick, stumpy prisms.
The sodium salt was prepared in the same manner as the salts of
the other acids by neutralising on the water-bath as nearly as possible
with a pure normal caustic soda solutioUi It remained as a thick^
transparent gum on heating for about 3 hours in the air oven at
140 — 160° On cooling in a desiccator, it formed a transparent|
brittle, hygroscopic, resinous mass. It was powdex^ while hot,
transferred to a glass-stoppered vessel as rapidly as possible, and
obtained as a white powder which became quite sticky when left for
a few minutes in the open air.
Digitized by
Google
MBTHTL QBOtrP ON BXSQ J'ORMATION. 791
Its aqueous solution gave copious precipitates with silver nitrate
and copper sulphate, and more slowly with calcium chloride. Lead
acetate gave a precipitate which dissolved in a large excess of the
reagent. Barium nitrate only produced a slight turbidity.
tTMXB'&'IHmeihyl-p^th<^phenyl8tAcciifiam%c Acid (trans-s-Z>$m6t^^^p-
ethoaci/sficcinamlic Acid),
COaH-CH(OH8)-OH(OH8)-CO-NH-CeH4-0-aaH5.
This add was prepared from ^(m«-«-dimethylsuccinic anhydride and
^phenetidine. It crystallises from alcohol in needles melting at
184 — 185^ with slow decomposition, and is sparingly soluble in hot
water, separating from the solution on cooling in prisms. On
analysis :
0-2100 gave 0-4873 00, and 01376 H^O. 0 = 63-28; H = 7-28.
0*3206 „ 16 C.C. moist nitrogen at n'' and 760 mm. N = 5'81.
C14H19O4N requires 0-63-39; H = 717; N = 5-28 per cent.
The sodium salt gave colloidal precipitates with silver nitrate and
copper sulphate. The precipitate given by lead acetate dissolved in a
large excess of the reagent, but separated out again on stand-
ing. No precipitate was obtained with barium nitrate or calcium
chloride.
ciB'Dimethf/l'j^hooeyphenf/huccinamic Acid {ciarDimetkyl-jp-ethoxysuocin'
cmUic Add), 00,H-OH(OH3)-OH(OH8)-00-NH-OftH^-0-OjH5.
This compound crystallised from alcohol in warty groups consisting
of microscopic needles which did not readily separate from the solvent
on the filter pump and melted at 165 — 156°. On analysis :
0-1500 gave 0-3479 OOj and 00989 H^O. 0 - 63-26 ; H - 7-33.
0*2688 „ 12-2 CO. moist nitrogen at ll'' and 768 mm. N « 6*39.
^14^1 AN requires 0 -* 63-39 ; H « 717 ; N = 6-28 per cent.
The pure sodivm salt was not obtainable under the conditions
adopted. On attempting to neutralise the hot solution of the acid
and evaporating down somewhat, a small quantity of crystals separated*
After filtration, the aqueous solution had a slightly alkaline reaction
and no further attempt to isolate this salt was undertaken.
The crystals which separated were collected, washed^ and dried on a
porous plate. They then melted at 112 — 113°. On cooling the hot
aqueous solution, they separated in crystalline groups of sandy appear-
ance» which under the microscope were seen to consist of ball-shaped
groups of needles,
3 G 2
Digitized by VjOOQIC
792 aiLBODT AND SPBANKLING : INFLUENCE OP THE
As ciB-dimethylpyranHn melfcs a1» 114 — 1 15° and crystallises similarly,
there can be no doubt that under these conditions a portion of the salt
condenses to the ring compound in preference to remaining as the
sodium salt of the open-chain acid.
So far as the observations went, this salt differs from the trans-
sodium salt, as with the latter no crystals of a ring compound separated
on evaporation of the aqueous solution of the salt.
IHine^yl-p-MooM/phenyUucinnamie Acid {IHrnethyl-^ethooGysuecinaniUe
ilcW), 00,H-0(CH8)8-OH(OH8)-00-NH-CeH^-0-CjH5.
This acid crystallised from alcohol in warty groups of colourless
plates or prisms melting at 128 — 129°, which were sparingly soluble
in water. On analysis :
0-1462 gave 03448 COj and 01012 H^O. C-64'31 ; H-7-69.
0-3921 y, 17-95 cc. moist nitrogen at 18° and 753 mm. N -» 5*21.
CjgHjiO^N requires 0= 6451 ; H=s 752 ] N = 502 per cent.
On allowing the mother liquors from the above recrystallisations to
evaporate spontaneously, a further considerable quantity of the
crystalline substance separated. This differed from the compound
which had already separated in that it melted at 86 — 88°, and was
obtained in the form of needles. After recrystallisation from acetic
acid, it again melted at 87 — 88°, and was found to be trirMthyl-
pyrcmtin.
As in the case of the cM-dimethyl-^ethoxyphenylsuccinamic acid, a
pure sodium salt could not be obtained under the conditions employed.
When the acid was neutralised and the solution evaporated, crystals of
trimethylpyrantin melting at 87 — 88° were obtained^ and the solution
became alkalina When the solution was evaporated down, and the
residue dried at 140 — 150° and digested with water, a crystalline mass
remained behind, and the solution was found to be strongly alkaline to
litmus. On recrystallising this insoluble portion from dilute acetic acid,
it separated in needles melting at 86 — 88°, and was therefore nothing
more or less than trimethylpyrcnUin. Under these circumstances, no
further attempts were made to prepare the sodium salt.
hoPrapyl^^^-ethoxyphmyliuceincmie Acid {isoPropyl-ihethoxy'
aticcinanilio Add), 00,H-0H[0H(CH3)J-0H,-00-NH-0jH^-0-C8B[4.
This acid crystallised from alcohol in beautiful, transparent leaflets,
which had a pearly lustre when dry and melted at 151—152°. It
gave a faintly acid aqueous solution. On analysis :
Digitized by
Google
MKTHTL GBOUP ON RING FORMATION. 798
01901 gave 04478 OOj and 0-1306 Kfi, 0 = 6426 ; H = 763.
0*2819 ,, 12*6 CO. moist nitrogen at 17^ and 763 mm. N = 5*16.
C^jHjjO^N requires C- 64-61 ; H=7-52 3 N-502 per cent.
The pure sodium salt was not prepared, as on attempting to
neutralise the acid on a water-bath and to crystallise the product a
large amount of iaopropylpi/rairUin was obtained.
II. The Pyrantina.
Fyraniin {^Ethoxyphmyhueeinimide), Ag -CO^^'^*"^*'^*^^^*'
When heated aboTe its melting point, jp-ethozyphenylsuocinamic acid
readily parts with 1 mol. of water, the general method of preparation
of pyrantin and its homologues being to warm the corresponding acid
at about 180° in a sulphuric acid bath for about 40 minutes or until
the evolution of aqueous vapour ceased. The product, which in most
cases had darkened slightly in colour, was poured into a dilute solution
of sodium carbonate to dissolve out any unattacked acid, and the
resulting nearly colourless and brittle product filtered off, washed with
water, and recrystallised from a little glacial acetic acid to remove
small traces of colouring matter, and then from alcohol, from which it
separated readily in colourless needles melting at 165°.
This substance has been previously prepared by Piutti {loo. eU.) by
the direct condensation of p-aminophenetole hydrochloride or phen-
aoetin with succinic acid, and its melting point given as about 155°.
We also prepared this substance and its homologues by the direct
interaction of succinic or an alkyl-succinic acid and p-phenetidine at
temperatures of about 160 — 180°. In all cases, the reaction was very
vigorous, and the product readily solidified on cooling. To purify it,
the substance was dissolved in the least possible quantity of glacial
acetic acid, the solution cooled, and the crystalline mass spread on a
porous plate. The coloured matter was thus absorbed and a second
treatment, if necessary, followed by a recrystallisation, always yielded
a colourless product.
As obtained in this way, pyrantin melted at 155° Its solubility was
determined in a 0*76 per cent, solution of sodium chloride at 30° as
best representing conditions in the human system to which the drug
would have to be subjected if employed internally. It was found that
only 0-0701 gram dissolved in 60 c.c. of the solution, which gives the
ratio 1 : 713.
The colourless crystals (m. p. 155°), after being carefully dried for
about 20 minutes at 97°, were analysed, with the following results :
Digitized by
Google
794
GILBODT AND 8PRANKLING : INFLUBNCE OF THE
0-1996 gave 0-4771 CO3 and 0-1050 H^O. 0 «= 66-20 ; H - 6-90.
0-2922 „ 16*6 C.C. moist nitrogen at 2P and 765 mm. Ne6-52.
OijHjgOjN requires C« 65-75 ; H-5'93; N = 6-39 per cent.
The measurement of the stability of the sucdnimide ring was carried
out by a slightly modified form of the method employed by Hantzsch
t.
SB.
A-x.
X
A'X
At,
1
(1.) ii = 10-20.
12
5-61
4-69
1-2222
0-1018
16
5-67
4-63
1-2080
0 0802
18
6-66
3-64
1-8022
o-iooi
21
6-81
8*89
1-6221
0-0772
24
7-88
2*84
2-6916
0-1079
27
7-65
2-66
{2.) A'^10'2
2-8492
0.
0-1055
Heftn 0-0984
6
4*20
6 00
0-7000
0-1166
9
4-97
5*28
0-9508
0-1056
12
6-88
4-87
1-0945
0-0912
15
613
4-07
1-5061
0-1004
18
6*24
8-96
1-5757
0-0876
21
6-78
8-47
1-9895
0-0923
24
7-00
8-20
2-1875
0 0911
27
7-44
2 76
2-6956
0-0998
Hean 00981
(3.) ii = 9-4
0.
12
4*90
4-5
1-0888
00907
14
4-90
4-5
1-0888
00777
16
6-80
4 1
1-2927
0*0808
18
610
3-8
1-8484
0-1027
. 20
6-20
3-2
1-9875
0-0969
214
6 40
8 0
2 1833
0-0992
Mean 0 0913
The mean of the three series of determinations is ^e= 0-0949.
and Miolati (^. eit). Alcohol had to be used as the solvent on
account of the sparing solubility of the homologues of pyruitin in
water.
0*2190 gram of pyrantin was dissolved in 190 c,e. of alcohol j the
Digitized by
Google
METHTL GROUP ON RING FORMATION. 795
solution was allowed to stand in the thermostat for about a quarter of
an hour at 25^ and then 10 c.c. of a iT/IO caustic soda solution were
added. Twenty c.c. were then withdrawn at regular known intervals
and titrated with ^/lOO hydrochloric acid, using phenolphthalein as
indicator. Under these conditions, 10 c.c. of iT/lOO standard acid
solution should correspond exactly to the amount of caustic soda which
would be present in the 20 c.c. withdrawn for titration if no opening
of the ring had taken place.
1 X
The results (p. 794) were calculated from the formula Ac = -z.. _
already used for phenyl-, tolyl-, and xylyl-succinimides by Miolati and
Longo (loc. eit.), where A represents the amount of substance present
at the commencement of the reaction, x is the amount of the changed
substance present after time t (in minutes), and o is a constant. In
this and the estimations which follow ^ = 10 nearly,
Meihylpyrantin (Methyl-^^thaxi/phenylsitcoinimide),
was crystallised first from glacial acetic acid and then from alcohol.
It separated from the latter solvent in small clusters of fiat prisms
melting at 105 — 106°. Crystals having the same melting point were
also obtained by melting the crude substance in alcohol, adding hot
water to the warm solution until it became slightly turbid, and then
allowing it to cool, but the methylpyrantin obtained in this manner
had a slight greyish tinge. On analysis :
0-2109 gave 0-5176 COj and 01270 Ufi. C = 670 ; H = 6-7.
0-4106 „ 21-9 C.C. moist nitrogen at 18° and 754 mm. N = 6-09.
CigHujOgN requires 0 = 66-96; H = 6-43 ; N = 6-00 per cent.
It is only moderately soluble in hot water and separates in needles
on cooling ; 0-0694 gram dissolved in 50 c.c. of 0*75 per cent, sodium
chloride solution under the conditions given above for pyrantin gave
the ratio 1 : 720.
For the determination of the stability constant, as described above,
0*2330 gram was required to make A — 10.
Digitized by
Google
796 GILBODY AND 8PRANKLING : INFLUENCE OF THE
L
X.
Ax,
(1.) A = 10
X
Ac,
A-x
•20.
8
4-32
5-88
0-7347
0-2549
6
5-79
4-41
1-8129
0-2188
9
6-20
4-00
1-5500
0 1722
12
6-80
8-40
2-0000
0-1666
15
7-24
2-96
2-4459
0 1681
18
7-76
2-44
8 1808
0-1767
21
7-78
2-42
8-2149
01581
24
8-06
2-16
8-7542
0 1560
27
8-42
1-78
4-7808
01752
Mean 0-1807
(2.) A =10-20.
4
4-94
5-26
0-9391
0-2848
7-5
5-79
4-41
1-3129
01750
10
6-09
4-11
1-4817
0-1482
12-6
7-80
2-90
2-5172
0-2013
16-6
7-86
2-84
2-5915
0-1672
18
7-60
2-60
2-9281
0-1624
21
814
2-06
3-9514
0-1881
28-5
8-27
1-98
4-2850
0-1828
26-6
8-65
1-56
5-5806
0-2106
Mean 0-1855
The mean of the two series of experiments is ^6=0 1881.
&8'Ditnethylpyrantin {sLa-Dimethyl-j^ ethaxyphenylsuoeinimide).
The aa-dimetbyl-p-ethozyphenylsucoinamic acid used in the prepara-
tion of tiw-dimethylpyrantin effervesced very strongly as it melted,
and as the product seemed to darken rather rapidly and the reaction
appeared to be complete after about 20 minutes' beating at 180^, tbe
viscid liquid was at once slowly poured into a dilute sodium carbonate
solution. It did not crystallise out at all readily from alcohol until
a crystal of the substance bad been added ; colourless needles
were then obtained melting at 73°. This pyrantin was also pre-
pared by the direct method of heating together the equivalent amounts
of c»-dimethylsuccinic acid and /7-phenetidine. The specimen obtained
melted at 70—72° On analysis :
Digitized by
Google
METHYL GROUP ON RING FORMATION.
797
0-U70 gave 0-3668 00, and 0-0926 H,Q. 0 = 67 88 ; H-7-00.
0-4002 ,, 20-5 o.c. moist nitrogen at 1 9^ and 760 mm. N » 6*80.
Oj^Hj^OgN requires 0 = 6803 ; H-6-88 ; N-6-67 per cent.
0-0602 gram of the compound dissolved in 60 o.o. of 0*75 per cent,
sodium chloride solution giving the ratio 1 : 996. It is only very
moderately soluble in hot water, from which it separated in needles.
For the determination of the stability constant, 0*2470 gram was
required to make A » 10.
t
X.
A-x.
a;
A-x
Ac.
(1.) ii = 10-20.
6
4-31
6-89
0-7817
0-1219
9
4*68
6-62
0-8478
0-0942
12
5-16
6-04
1-0288
0-0863
15
6*44
4-76
1-1428
0 0762
18
6-08
4-12
1-4614
0-0806
21
6*29
8-91
1-6087
0 0766
24
6-61
8*69
1-8412
0-0767
27
7-00
. 8-20
2-1876
0-0810
Mean 0-0866
(2.) ii = 10-2
0.
9
4*86
6-86
0-9066
0-1007
12
6-60
460
1-2174
0-1014
16
6-76
4-45
1-2919
0-0861
18
6 02
4-18
1-4402
0-0800
21
6-24
8-96
1-6767
0-0760
24
6-81
, 8-39
2-0088
0-0887
Mean 0-0878
The mean of the two series is Ae^0*0S72.
dS'S-Dimeihylpyrantin (cis-s-2>tflM^^^p-«<Aoa^^p^tfnyb1MeitMmu20),
After having been crystallised from acetic acid, ct9-«-dimethylpyr-
antin separated from alcohol as a mass of small needles, which had a
silky appearance when dry and melted at 114 — 115°. It was also
obtained in small quantity during the attempts to prepare the sodium
salt of a«-«-dimethyl-p-ethozyphenyl8uocinamio acid (p. 791). On
analysis :
Digitized by
Google
798
QILBODT AND SPBANKLINa : INFLUENCE OF THE
0-1841 gave 0-4953'COj and 01130 H,0. 0-68-42 ; H- 68-3.
0*2806 „ 14-35 c.c. moist nitrogen at 11° and 762 mm. N « 5*81.
O14H17O3N requires 0^6803 ; H;;=6-88; N^5-67 per cent
0*0543 gram of the substance dissolved in 50 c.c. of 0*75 per cent,
sodium chloride solution, giving a solubility ratio of 1 : 920. It is
only moderately soluble in hot water, separating out in needles on
cooling.
For the determination of the stability constant 0*2470 gram
required to make A = 10.
t.
a;.
A^x.
X
A-x
Ac
(1.) ^ = 10*20.
6
6 09
6-11
0-9961
0-1660
9
677
4-43
1-3026
0-1447
12
6-40
8-80
1-6848
0-1404
16
6-76
8-44
1-9661
0-1310
18
7*16
8-08
2*8443
01302
21
7-41
2-79
2-6668
0-1265
24
7-79
2-41
8-2323
0-1347
27
8-03
2-17
3-7006
01871
80
8-19
2 01
4-0746
0-1358
Mean 01886
(2.) il = 10'2
0.
6
6-11
6-09
1-0039
01678
9
5-69
4-61
1*2126
0-1336
12
6-49
3-71
1-7499
01468
16
6-81
3-39
2-0088
01339
18
7-21
2-99
2-4113
0-1839
21
7-51
2 69
2-7918
0-1828
24
7-69
2*61
80687
01276
Mean 0*1898
The mean of the two series is ^=0'1S89.
tTAaa^Dime^ylpyranlin (trtma^Dimethyl-p-elhoxyphimi/laiuseinimide),
OH,.
'S(0H3).C0>N-^<'^«-<^*^«^-
As ^ron^tf-dimethyl-fT-ethozyphenylsuccinamicacid meitsat 184 — 185^,
a temperature of 190 — 195^ was employed in the condensation, and main-
tained until the evolution of aqueous vapour had ceased. After the
Digitized by
Google
METHYL GROUP ON RING FORMATION. 799
usual treatment with acetic acid and alcohol, colourless needles melt-
ing at 104 — 105° were obtained, a temperature which is 10° lower
than that at which the ci8-<K>mpound melts. A couple of recryatal-
lisations did not alter the melting point.
^a7iff-a-Dimethylsuccinic acid is easily converted into the ow-modifi-
cation by simple heating ; it might therefore be expected that in an
analogous manner a cM-pyrantin deriyative, or at least a mixture of
the eis' and ^rofM-compounds, would be obtained, and the low melting
point also lends colour to this possibility. As, however, a determina-
tion of the stability constant gave figures which are completely different
from those of the «i9-«-dimethylpyrantin, the results are published here,
and the definite solution of the problem as to whether the substance
was pure <fa?M-a-dimethylpyrantin or a mixture of the two modifica-
tions must be postponed to a future occasion. The stability constant
obtained with the material we had at our disposal gave a mean for one
series of experiments of 0*1839.
TrifMihylpyrcmtin {IHmethyl'p-ethoxyphmylmccinimide)^
Trimethylpyrantin has been isolated in several ways, as the ring
compound is formed with remarkable readiness. Thus it is obtained
(1) by heating trimethyl-fT-ethoxyphenylsuccinamic acid to 180° ; (2) in
large quantity in the alcoholic mother liquors from the recrystallisa-
tion of the acid, especially on standing ; (3) from the alkaline solution
produced on attempting to prepare sodium trimethyl-j9-ethoxyphenyl-
succinamate; (4) on dissolving trimethyl-^thoxyphenylsuccinamic
acid in dilute caustic soda and allowing the solution to stand, when
beautiful leaflets gradually separate out, and (5) by heating together
equivalent quantities of ^phenetidine and trimethylsuccinic acid.
After recrystaUisation from acetic acid followed by alcohol, it separates
in colourless needles melting at 87 — 88°. On analysis :
0-1802 gave 0-4662 OOj and 01218 HgO. 0 = 6906; H = 7-51.
0-1992 „ 9-4 C.C. moist nitrogen at 22° and 746 mm. N =^ 5*50.
OjgHigOgN requires 0 = 68*96 ; H « 7*28 ; N - 5*36 per cent.
It is only very sparingly soluble in 0*76 per cent, sodium chloride
solution, 00394 gram dissolving in 60 c.c, which gives the ratio
1 : 1272. In boiling water, it is almost insoluble, and the little that
dissolves crystallises out again on cooling in hair-like needles.
For the determination of the stability constant, 0*2610 gram wae
reqKired to make A cs ]0,
Digitized by
Google
800
OILBODT AND SPRANKUNO: INFLUENCE OF THE
L
r.
A-x.
X
A'X'
Ac
1
(1.) i( = 10'20.
9
8-86
6*86
0-4890
• 0-0548
IS
8*65
6-66
0-5673
0-0429
17
4-08
612
0*6666
0-0392
21
4-90
6-80
0-9246
0-0440
26
6-00
5-20
0-9616
0-0385
29
6 04
5 16
0-9767
0-0837
88
6-88
4-82
1-1162
3-0888
87
6-22
3-98
1-5628
0-0422
41
6-86
8-84
(2.) il = 10-2
1-6562
0.
0-0404
Mean 0-0410
10
3*60
6-60
0-5454
0-0546
20
4-60
6-60
0-8214
0-0411
SO
5-20
6-00
1-0400
0-0347
40
6 -86
4-86
1-8448
0-0336
65
7-80
2-90
(3.) A'^lO'i
2-5172
10.
0-0458
Mean 0-0419
11
8-80
6-40
0-6987
0-0540
16
4*25
6-96
0*7148
0-0476
19
4-62
6-68
0-8280
0*0486
23
6-89
4-81
1-1206
0-0487
27
6-82
4-88
1-0901
0-0404
31
6-54
8-66
1-7870
0*0676
86
6-51
3-69
1-7642
00504
Mean 0-0489
'
(
4,) ii«10-2(
).
9
3-87
6-83
0-4984
0 0548
12
8-70
6-60
0-6602
0 0474
15
4-22
5-98
0-7056
0*0470
18
4-58
6-67
0-7989
0-0444
21
4*92
6-28
0-9818
00473
24
6-28
4-92
1-0781
00447
27
6 '89
4-61
1-1206
00415
Mean 0 0467
The mean of the four series is ^fes 0-0446.
Digitized by
Google
HBTH7L GBOirP ON RING FORMATION.
801
iBoFropylpyrcmtin {iBoFropyl-p'ethoosyphenylsuceinimide)^
This compound crystallised from alcohol in colourless, glistening
needles melting at 98 — 99^. In addition to the ordinary mode of pre-
paration, it was obtained in large quantity during the attempts to
prepare a pmre sodium salt from the wopropyl-jp-ethoxyphenylsuccin-
amic add (p. 793) ; it then melted at 97^« On analysis :
0-1980 gave 0-4988 CX)^ and 01338 HjO. 0 = 6870 ; H = 7-61.
0*4208 ,, 20*4 c.c. moist nitrogen at 20^ and 763 mm. N ==^ 5*56.
CijHi^OgN requires 0-68 96 ; 'H=-7-28; N«5-36 per cent.
It is very sparingly soluble in theO'75 percent, salt solution, 0*0451
gram dissolving in 50 cc, thus giving the ratio of 1 : 1110. It is only
moderately soluble in boiling water and separates out in long needles
on cooling.
For the determination of the stability constant, 0*2610 gram was
required to make As 10.
t.
X.
A-x,
X
A-x
, Ac.
(1.) ^ = 10*2
0.
8
6*42
4-78
11389
0-1417
11
5*90
4*80
1-3721
0-1247
14
6-90
8*80
2-0909
0-1493
17
7-28
2-92
2-4981
0 1466
20
7-44
2-76
2*6966
0-1848
23
7-81
2 89
3-2678
0-1421
26
8-27
1-98
4-2860
0*1646
Mean 0 1484
(2.) ^ = 10*2
0.
6
6*17
608
1-0278
• 0-1718
9
6-78
4-43
1-8077
0-1468
12
6 12
4-08
1-6000
0 1250
15
6-94
3-26
2-1288
01419
18
728
2-97
2-4848
0-1862
21
7-54
2-66
2-8346
0 1860
24
7-92
2-28
8-4787
0-1447
27
8-17
2-08
4-0246
0-1491
Mean 01484
The mean of the two series is Ae=iO'lid^.
Digitized by
Google
802
QILBODY AKD SPRANKLING .* INFLUENCE OF THE
CH •CO
FhmyUucGmimide, ^j^.Q(^^*(^fp6'
This compound was prepared in the usual manner^ and its stability
constant was determined in both alcoholic and aqueous solution.
Three series of experiments were carried out in alcoholic solution
with the following results :
L
X,
A'X,
X
A-x
Ac.
1
(1.) ii = 9-66.
7
8-34
6-22
0-6370
00767
9
3-68
6-88
0-6269
0-0696
12
4-46
610
0-8746
00729
16
4-86
6-20
0-8887
0-0524
18-6
4-72
4-84
0-9762
0-0527
21
4-64
6 02
0-9044
0-0431
24-6
4-88
4-78
1-0212
0-0417
28
4-96
4-61
1-0738
0-0383
82
6-86
3-70
1-6840
0-0496
l£ean 00662
(2.) A^lO'i
20.
9
3-66
6-64
0-6696
0-0622
12
4-20
6-00
0-7000
0-0683
16
4-66
6-66
0-8068
00637
18
614
6-08
1-0168
0-0664
21
6-23
4-97
1-0628
0-0601
24
6-64
4-66
1-1888
0-0497
27
6-31
8-89
1-6196
0-0599
Mean 0-0657
(3.) ii-9-5
5.
6-75
2-86
6-70
0-4269
0 0742
10
3-60
6-96
0-6040
00604
12-76
8-91
6-66
0-6920
00643
16-26
4-22
6 84
0-7903
0 0618
19
4-76
4-80
0-9917
0-0622
21-76
6-11
4-46
11483
0 0628
26-26
6-21
4-86
1-1977
0 0474
29-6
6-00
8-66
1-6864
0-0671
Meui 0-0663
The mean of the three aeriea ia ^e= 0*0667.
Digitized by
Google
METHYL OnOXJP ON Bmo FORMATION.
803
To obtain a means' of comparing the resultfl obtained aboTe for tits
mtthylpyrantins with the tolyl- and zylyi^aocji^iiiiidM, it wa«
neeeisary to know the stability constant of pbenylsiiociAiiAiciQ.JK
aqaeouB solution. This determination has already been made by
Miolati and Longo {AUi E. Accad. Lincei, 1894, [v], 3, 601), but unfor-
tonately there is an error in the calculations. These authors give the
mean of three series of experiments as 2*27, but if the mistake in the
first series be rectified, the mean works out to be 2*18.
Two series of experiments were also carried out with phenylsuccin-
imide in aqueous solution, with the following results :
t.
aj.
A'X.
X
Ac.
A-x
(1.) il-9-6(
5.
8-3
8-09
1-47
5*486
1*5848
4-4
8-76
0-8
10-950
2-4886
5-5
8-96
0-6
14-983
2-7140
6-6
8-91
0-66
18-708
2 0770
9-9
916
0-4
22-900
2-3131
Mean 2*2551
(2.) 4 = 9-61
5.
8 0
8-26
1-8
6-354
21180
4-26
8-76
0-8
10-950
2-6765
6-25
8-81
0-76
11-747
2-2375
6-25
8-91
0-66
18-708
2 1938
7-26
9-08
0-60
18120
2*4965
9
9-14
0-42
21-762
2*4180
11
9-20
0-86
25-555
2-8232 ;
Mean 2-8376
The mean of the two series is ^e=2'29.
If the mean of Miolati and Longo's and of the authors' experiments
be taken as probably the most correct figure, namely ilc = 2-23) the
values for the substituted pyrantins in alcoholic solution can bo first
converted into the corresponding value8*for the respective substituted
phenylsuccinimides in the same solvent, and then the value of the
latter compounds converted from the stability constant, Ac^ in alcoholic
to that in aqueous solution.
The experimental data necessary to accomplish this have been deter*
mined above with pyrantin in alcoholic solution and with phenylsncdn-
imide in both alcoholic and aqueous solution^
Digitized by
Google
804 QILBODT AND SFEANKLING : INFLUJBKCE OF THE
The pyrantin was found to have a gtability constant, ule« 0*0949,
and phenylsncripi^zaide one of iioa 0*0567, so although the ethozyl
vrrotipis'so far away as the pararpoeition to the nitrogen in the suodn-
imide ring, it has a very great effect in decreasing the stability of the
fatty nucleus.
The two measurements give a factor ■ ,. ., (or divide by 1*704) for
0*0949 J /
the conversion of the values of the substituted pyrantins into those of
the corresponding phenylsuccinimides.
In Alcoholic Solution.
Ac foand. * Ac calc.
Pyrantin 00949 Phenylsuccinimide 0*0557
Methylpyrantin 0*1831 Methylphenylsuccinimide 0*1075
oa-Dimethylpyrantin ... 0*0872 o^-Dimethylphenylsuccin-
imide 0*0512
ct«-«-Dimethy Ipyrantin ... 0*1389 ct9-«-Dimethylphenylsuc-
cinimide 0*0815
^an8-«-Dimethylpyrantin (0*1839 )) ^an«-a-Dimethylphenyl-
suocinimide (0*1079?)
Trimethylpyrantin 0*0446 Trimethylphenylsuccin-
imide 0*0262
woFropylpyrantin 0*1432 woPropylphenylsuccin-
imide 0*0840
The stability constant, Ac, for phenylsuccinimide in alcoholic solu-
tion being 0*0557, and in aqueous solution 2*23, we get the second
factor for the conversion of the value in alcoholic to that in aqueous
solution, namely, 40*03, since 0*0557 x 40*03 = 2*23, giving the follow-
ing results :
Constant calcnlated
^ * ^
for Alcohol ic for aqueoiui
solution. solution.
Phenylsuccinimide 0*0557 (found) 2*23 (found)
Methylphenylsuccinimide 0*1075 4* 30
(M-Dimethylphenylsuccinimide 0*0512 2*05
ci9-9-Dimethylphenylsuocinimide 0*08 1 5 3*26
ttYm«-«-Dimethylphenyl8Uccinimide . . . (0*1079 X) (4*32 1)
Trimethylphenylsuccinimide 0'0262 1 05
woPropylphenylsuccinimide 0*0840 3*36
The values so obtained, in conjunction with the experimental data
of Miolati and his colleagues, enable a comparison to be made of the
Digitized by
Google
METHYL GROUP ON RING FORMATION. 805
result of an introduction of the methyl group into almost any position
in the pbenylsuccinimide ring,
S:S>''0
and such comparison is afforded in the following list :
Ac, Ac.
tran8-8'T>imeihylphenj\'
succinimide (4*32 t) Xylylsuccinimide 1*145
Methylphenylsuccinimide 4*30 /^-Tolylsuccinimide 1*12
ttfoPropy Ipheny Isuccin*
imide 3*36 m-Tolylsuccinimide 110
et8-«-Dimethylphenylsuc-
dnimide 3*27 Trimethylphenylsuccinimide 1 '05
Phenylsuccinimide 2*23 2 : 5-Xylylsuccinimide 0*88
a«-Dimethylphenylsuccin-
imide 2*05 o-Tolylphenylsuocinimide... 0*856
3 : 4-Xylylsuccinimide 1*27 2 : 3-Xylylsuccinimide 0*815
2 : 6-Xylylsuccinimide, 0*16.
III. Coneluaiona.
The general conclusions to be drawn from the data obtained in this
research would seem to be as follows for phenylsuccinimide and simi-
larly constituted substances.
(1) Methyl groups introduced into an aromatic ring which is linked
to an imide ring by means of the nitrogen atom cause tbe latter to become
more stable ; also the nearer the methyl group is to the nitrogen atom
the more stable is the ring (compare work of Miolati and his
colleagues).
(2) A methyl group introduced into the imide ring renders that
ring less stable, but on the introduction of more methyl groups
the stability increases, and after a certain number have been intro-
duced, the stability becomes greater than in the case of the ring
which does not contain the methyl group.
(3) For a corresponding number of methyl groups, a substance con-
taining constituents in the aromatic ring is much more stable than
one with substituents in the imide ring.
(4) The introduction of an ethozyl group into the para-position in
the aromatic ring causes a great decrease in the stability of the imide
ring.
The deductions could not be carried further without completing the
experimental study of the effect of introducing methyl groups in both
the aromatic and imide rings at the same time. One or two other interest-
VOL. LZXXI. 3 H
Digitized by VjOOQIC
806 ORTON: THE PREPARATION OF
ing points had also to be left in an incomplete state owing to the authors
leaving Owens College, where the practical part of the work was
carried out ; it was therefore decided to publish the results obtained.
BbADFORD TeOHNIOAL COLLBQB. GoYERNMENT LABORA.TORT,
AntiguAi Leeward Islaioms.
LXXXI. — The Preparation of Highly Substituted
Nitroarainohenzenes.
By K. J. P. Obton.
In the action of nitric acid on «-trihalogen anilines (this vol., p. 490),
the nitroamino-Mrihalogen benzenes are always formed in small amount.
These nitroamines, under appropriate conditions, yield products identical
with those obtained by the direct action of nitric acid on the anilines.
It was suggested (}oc» cU,) that possibly the following series of changes
took place :
NH2,HN08 NH-NOj
Y
NH .
w
— Br NO,-
NH,
NO,
The final transformation only occurs when a bromine atom is in the
para-position relatively to the amino-group. When a chlorine atom
is in the para-position, the iminoquinone, if formed, undergoes change
in some other manner.*
In order to study more minutely the transformations of these
nitroamino-0-trihalogen benzenes, it was necessary to devise means of
readily prepaHng them in quantity. The methods at present known
(Bamberger, Ber., 1893, 28, 471, 485 ; 1894, 27, 584 ; 1895, 28. 401)
of preparing nitroaminobenzenes are (1) oxidation of alkaline solutions
of diazotates by potassium ferricyanide or permanganate ; (2) addition
of dry aniline nitrate to acetic anhydride ; (3) treatment of the aniline
in solution in ether or chloroform with nitrogen pentoxide. Owing
* It is donbtful if in the nitration of anilines and anilides it is justifiable to assume
that the nitroamino-deriTatiyes necessarily occur as an intermediate stage. The
course followed by the reaction must depend on whether, under the given conditions,
the velocity of the formation of the nitroamino-deriyative exceeds that of the direct
action of the nitric acid with the aniline (or anilide) or with its tautomeric (imino-
quinone) form. The same considerations apply to the chlorination and bromination
of anilines and anilides.
Digitized by
Google
HIGHLY SUBSTITUTED NITEOAMINOBENZENES. 807
to the feeble basic characters of the ^-trihalogen anilineSi the first two
methods could only be applied with difficulty, and the use of nitrogen
pentoxide is attended by many obvious disadvantages.
I have found that the «-trisubstituted anilines (^-tribromoaniline,
3 : 5-dibromotoluidines, <&c.) can be converted with ease and practically
quantitatively into nitroamino-derivatives when a solution (or suspen-
sion) of the aniline in glacial acetic acid is treated successively with
some excess of nitric acid (free from nitrous acid) and acetic anhydride.
The nitrate of the aniline, first formed, rapidly goes into solution,
being converted nearly completely into the nitroamine. There is
some development of heat in the reaction. The method is also applic-
able to anilines in which one ortho-position is occupied by hydrogen
(for example, 2 : 4-dichloroaniline); in this case, the conversion of nitrate
into nitroamine does not take place so easily, and some nitration of the
aniline in the ortho-position also occurs. With p-halogen anilines, the
formation of the nitroamine from the nitrate is still more difficult, and
is accompanied by more nitration.
The nitroamino-9-trihalogen benzenes, which have a bromine atom
in the para-position relatively to the imino-group, are largely con-
verted into dihalogen-p-nitroanilines (compare this vol., p. 491), when
their solutions in acetic acid, to which a drop of sulphuric acid
has been added, are allowed to stand for several hours. With
2-nitroamino-3 : 5-dibromotoluene and l-nitroamino-2 : 4 : 6-tribromo-
3-nitrobenzene, the replacement of the jp-bromine atom by the nitro-
group scarcely appears to take place under these conditions.
When added to concentrated sulphuric acid, the majority of the nitro-
aminobenzenes here described give a deep purple or violet solution, with
development of heat; if the sulphuric acid contains a little water,
the solution is much paler and magenta-coloured. So long as a rise of
temperature is avoided, only a small quantity of the oxides of nitrogen
is evolved. Bromine is given off if a p- bromine atom is present. When
a nitroamine containing an ohydrogen atom is dissolved in sulphuric
acid and the purple solution poured on to ice, the colour changes sharply
to yellow, and a yellow precipitate of the isomeric nitrated aniline is
thrown down; thus nitroamine- 2 : 4dichlorobenzene is nearly entirely
converted into 2 : 4-dichloro-6-nitroaniline in this way. When the
purple solution of a nitroamino-^trihalogen benzene containing a
|>-bromine atom (^-tribromoaniline, &o,) is poured on to ice, the colour
changes to red and a red precipitate is thrown down ; this precipitate
contains a small amount of the dihalogen-;^-mtroaniline and a well
crystallised, red substance, probably nearly related to the substances
obtained by the action of nitric acid on, «-dihalogen-p-chloroanilines
{loc, eit,). If the nitroamino-«-trihalogen benzene contains a jp-chlorine
atom, its behaviour resembles that of the last case, except that no
3 H ioogle
808 orton: the preparation of
chlorine or bromine is evolved and no /^-nitrated aniline formed. The
products, whether «-dihalogen nitroanilines or red substances, obtained
by pouring these purple solutions on to ice, dissolve in concentrated
sulphuric acid ; they no longer give a deep purple coloration, but pro-
duce respectively a yellow or reddish-brown solution. Further, as
above mentioned, the nitroamino^erivative of 2 : 4-dichloroaniline under
this treatment yields 2 : 4-dichloro-6-nitroaniline nearly quantitatively.
Hence, it seems highly probable that the purple solution corresponds to
an intermediate stage between the nitroamine and the nitrated aniline.
In the ordinary nitration of anilines in the presence of a large excess
of sulphuric acid (for example, in the nitration of m-bromoaniline, this
vol., p. 499), a purple solution is formed, which changes to yellow
on pouring the acid liquor on to ice.
Concentrated nitric acid (sp. gr. 1'6) readily dissolves the nitro-
amines, producing a solution which is more or less transiently of a
purple colour ; the nitroamines derived from tetra-substituted anilines
(2:3:4: 6-tetrabromoaniline, (fee.) give only a brown solution.
EZPEBIMENTAL.
UNitroamino-2 : 4 : etnbromobmzene, OgHjBrj-NH^NOj.— Thirty
grams of finely powdered ^tribromoaniline were suspended in 300 c.a
of cold glacial acetic acid* (m. p. 15-5^). Twenty-five to 30 c.c of
nitric acid (sp. gr. 1*5) free from nitrous acid were then added, and
30 c.c. of acetic anhydride slowly poured in, while the mixture was
kept cold and well stirred. The solid slowly and completely dissolved,
forming a solution of a pale reddish or magenta colour. The liquid
was now poured on to about 600 grams of ice and water ; a copious
buff precipitate appeared, which was collected, washed free from acid,
and then extracted with a cold solution of 10 grams of sodium
carbonate in 150 c.c. of water. The solid nearly completely dissolved,
leaving 2 grams of a yellow substance, consisting mainly of Mribromo>
aniline together with a little 2 : 6-dibromo-4-nitroaniline. The colour-
less alkaline solution, after filtration, was diluted to 2 or 3 litres,
warmed to about 80°, and then slowly acidified with a slight excess
(about 100 0.0.) of 10 per cent, hydrochloric acid. The hot liquid,
from which the nitroamine separated in small, slender needles, was
rapidly cooled and the solid collected and dried. It weighed
31 grams.
Although the needles appear to be colourless in suspension in the
mother liquor, they are found to be flesh-coloured when in mass.
They can be obtained free from all colour by carefully acidifying a
* It is InadTisable to use sufficient acetic acid to completely dissolve this sparinglv
soluble aniline.
Digitized by VjOOQIC
HIGHLY StJBSTtTUTED NITROAMINOBENZENES. 809
dilute lukewarm solution of the barium salt. The nitroamines
derived from ^-trisubstituted anilines seem to have a greater tendency
to become coloured than other nitroaminobenzenes (see nitroamino-
2 : 4-dichlorobenzene), and their solutions in organic solvents deposit
coloured crystals.
This substance, thus prepared, melted and decomposed at 143^, and
was in every respect identical with the specimen previously obtained
by the action of nitric acid on ^tribromoaniline (this vol., p. 493). It
can be titrated by N/IO potassium hydroxide in the presence of
phenolphthalein :
0-3368 neutralised 9-18 c.c. iV/lO KOH. Equivalent = 366 -8.
O^HgOjNjBrg requires equivalent » 376.
The tUver salt is a white, insoluble powder insensitive to light.
The barium salt is prepared by mixing warm solutions of barium
chloride and the ammonium salt, when it immediately separates in
lustrous plates. On analysis of the air-Kiried salt :
0-9684, at 100°, lost 0019 H,0. H^O = 1 -96.
0-4934, over sulphuric acid, lost 0-0112 H^O. HjO-2-27.
0-5242 gave 01312 BaSO^. Ba = 14-73.
Ba(CeHj02N3Br8)2,HgO requires H3O = 200 ; Ba« 15-21 per cent.
On attempting to crystallise the salt from boiling water, it is
observed that at first short, transparent, four^sided prisms separate as
the liquid cools, and then plates. These crystals are not pure barium
salt, but contain a little of the nitroamine, as on analysis the numbers
found for barium are too low (about 0*7 per cent.) and those for
bromine are too high (about 1 per cent.). The salt is dissociated
hydrolytically to a small extent in water, and owing to the great
insolubility of the nitroamine, the latter separates on cooling the
hot solution of the salt. This was found to be the case with the
barium salts of all the insoluble nitroamines, whilst those of the
soluble nitroamines (l-nitroamino-2 : 4-dibromo-6-nitrobenzene and
l-nitroamino-2 : 4-dichlorobenzene) exhibit no such peculiarity.
Methyl Ethers of Nitroamino-B-tribramobenzene. — ^As Bamberger,
Franchimont, and others have found, the sodium salts of nitroamines,
on treatment with methyl iodide, yield mainly ethers, ^NMe'NO,, in
which the methyl group is attached to nitrogen, whereas the silver
salts yield mainly the isomeric oxygen ethers, 'NjO'OMe.
Two grams of the sodium salt were dissolved in 15 c.o. of methyl
alcohol and a slight excess of methyl iodide was added. The mixture
was left for 24 hours and then heated to gentle ebullition for 8 hours.
On cooling, clumps of long needles separated ; addition of water threw
down a further quantity of solid. This substance is practicaUy a pure
Digitized by
Google
810 OBTON: THE PREPARATION OF
methyl derivative; it is readily soluble in chloroform, benaene, or
acetone, sparingly so in boiling petroleum, and can eafiily be crystal-
lised from boiling alcoholi from which it separates in lustrous, long
bands, or in transparent, four-sided prisms from very dilute solutions,
or in needles, when the hot solution is rapidly cooled. These various
forms melt at 95*5°. There was no indication of the formation of the
isomeric o-methyl derivative :
01683 gave 0-2438 AgBr. Br = 61 -63.
CyHjOgNgBrj requires Br = 61*69 per cent
In concentrated sulphuric acid, the ether dissolves very slowly with
a magenta coloration ; in fuming nitric acid, it dissolves immediately
without giving any coloration and is apparently nitrated.
When the silver salt of nitroamino-«-tribromobenzene is suspended
in methyl alcohol and treated with methyl iodide, silver iodide is
rapidly formed. After filtration, the solvent was evaporated, and the
residual oil dissolved in chloroform and extracted with aqueous sodium
bicarl)onate. An oil was thus obtained which did not solidify at - 20^,
but after standing for some weeks deposited a few crystals* These
were separated and dissolved in hot alcohol, from which short prisms
crystallised melting at 55 — 56^ ; they are easily soluble in the lightest
petroleum.
l'N%troamino-2 : 4 : e4riohlorohenzen$, CeHjOlg-NH-NOj.— Fifteen
grams of Mrichloroaniline were dissolved in 150 c.c. of glacial acetic
acid ; 12 C.C. of nitric acid were added,* and then 10 c.c. of acetic anhy-
dride ; 16*5 grams of the pure nitroamino-derivative were obtained It
crystallised from a mixture of petroleum and chloroform in slightly
coloured prisms, melting and decomposing at 135^, and was identical
with the substance previously obtained from Mrichloroaniline and
nitric acid (^e. cit). The barium salt crystallised in plates and was
sparingly soluble in cold water :
0-413, at 100% lost 0-016 H2O. H^O = 2-8.
Ba(CgH802N,Clj)2,HjO requires H20-2-83 per cent.
l'MtroamvnO'2 : Q-dicMoro-ibromobenzene, CgHjCljBr'NH^NOj, was
prepared from 2 : 6-dichloro-4-bromoaniline ; it crystallised in slender, .
curved needles melting and decomposing at 137^ :
* B'Trichloroaniline nUrcUe. — The solid obtained by adding nitric add to a
solution of 9-trichloroaniline in glacial acetic acid was collected and washed with dry
ether, which dissolved a considerable quantity ; the remainder consisted of a felted
mass of needles and was immediately decomposed by water into ^-trichloroaniline
and nitric acid.
0*225 neutralised 8*1 c.c. N/IO Na^COs when titrated in the presence of methyl
orange, instead of 8*7 c.c, the calculated amount for CeHaCl,'NHa,HNOa.
Digitized by
Google
HlGflLT STTBSTlTUTm) NIT^ROAMINOBfiNZEKES. 811
0-1286 gave 0*2122 AgCfl + AgBr. AgOl + AgBr « 166.
CeHgO^jCl^Br requires AgCl + AgBr = 166 per cent.
In solution in acetic acid, to which a drop of sulphuric acid has been
added, this compound is transformed into 2 : 6-dichIoro-4-nitroaniline.
\-Niiroamino-2-cMoro-^ : ^•dihrcmohenze'Mf OgHjClBrj-NH-NOg, was
obtained from 2-chloro«4 : 6-dibromoaniline ; it crystallised in slender,
curved needles melting and decomposing at 137° :
01886 gave 0-2969 AgCl + AgBr. AgOl + AgBr = 157-4.
CjHBOjNjClBrj requires AgCl + AgBr = 1571 per cent.
From this nitroamine, 2-chloro-6-bromo-4-nitroaniline can be obtained,
but the yield is not so good as in the case of the preceding compound.
hUFUroamino-i-chloro-^ : Q-dibramohenzene, O^HjClBrj-NH-NOg, pre-
pared from 4-chloro-2 : 6-dibromoaniline crystallised in slender, curved
needles melting and decomposing at 137°:
0-2158 gave 0-339 AgOl + AgBr. A^Ol + AgBr = 157*09.
CgHgOaNgClBrj requires AgCl + AgBr =157-1 per cent.
l-iVt<roamtno-2 : 4rdichloro-6-bromobenzen$, CgH^OljBr'NH'NOg, pre-
pared from 2 : 4-dichloro-6-bromoaniline, crystallised in slender, curved
needles melting and decomposing at 137° :
0-1816 gave 0-2992 AgCl + AgBr. AgOl + AgBr =164-75.
O^H^OjNjOljBr requires AgOl + AgBr =166-0 per cent.
l-I^itroamin(h2 : i-dibroTna-^-nitrohmzene, NOg-CgHgBrj-NH-NO^. —
In this case, no precipitation occurred on pouring the acetic acid solu-
tion containing the nitroamine into water. The yellow acid liquid was
therefore repeatedly extracted with small quantities of chloroform and
the dark brown chloroform extract shaken up with aqueous sodium
carbonate, into which the nitroamine passed and from which it
separated on acidifying the solution with sulphuric acid. It dissolves
moderately both in hot and cold water, and crystallises in yellow
plates from its aqueous solution when a little sulphuric acid is added.
From a mixture of chloroform and petroleum it crystallises in well-
formed, lustrous, orange prisms melting at 91 — 92°. With acetone, it
forms an oil which is not decomposed by water. Its solution in con-
centrated alkali hydroxides may be boiled for many hours without
suffering any decomposition. In this behaviour, it offers a marked
contrast to o-nitroacet- or o-nitroform-anilide, which are hydrolysed
very readily by alkalis :
0-1426 gave 0-1580 AgBr. Br = 47-15.
OgHjO^NgBrj requires Br = 46*9 per cent.
The barium salt crystallises in yellow plates moderately soluble in
Digitized by
Google
812 orton: the prepabation op
cold, and readily so in hot water. An air-dried specimen was analysed
with the following results :
0-3304, over sulphuric acid, lost 0-0054 H,0. H,0= 1-63.
0-3304 gave 0091 BaSO^. Ba = 16-23.
BA{Pf;afi^lil^Br^)^,Kfi requires H3O-2I5; Ba= 16-4 per cent
This nitroamine could not be converted into 2-broino-4 : 6-dinitro-
aniline by the means previously described.
l'Mtraamin&-2 : 3 : 4 : 6-<6^a^oww)6enawie,CgHBr^-NH-NOj, prepared
from 2:3:4: 6-tetrabromoaniline, crystallises in plates with a silvery
lustre melting and decomposing at 136^ :
0-234 gave 0-3869 AgBr. Br - 70-36.
CjHjOjNjBr^ requires Br= 70*46 per cent.
This nitroamine dissolves in nitric acid with a brown and not a
violet coloration ; with concentrated sulphuric acid, the violet solution
is more slowly formed than with the majority of the nitroamines here
described.
l-IirUroaminO'2 : 4 : ^-tribramo-d-niirobenzene, NO^'CgHBrj-NH^NO,,
is a little difficult to isolate, as it does not separate well on pouring
the acetic acid solution on to ice ; it is best to extract the mixture with
chloroform. It crystallises from a mixture of chloroform and petroleum
in short, transparent, dull yellow, four-sided prisms melting and
decomposing at 108—109^ :
0-3516 gave 0-4702 AgBr. Br = 5692.
O^HgO^NgBrg requires Br — 67*14 per cent.
This substance neither dissolves in sulphuric nor in nitric acid with
the characteristic purple coloration.
l'MtroaminO'2 : irdichlarobenzene, CgHjClj-NH-NOj.— Five grams of
2 : 4-dichloroaniline were dissolved in 40 c.c. of glacial acetic acid and
4 cc. of nitric acid added, whereupon the nitrate of the aniline
separated as needles. The mixture was cooled to 10° and 5 c.c. of
acetic anhydride added slowly. The temperature slowly rose and the
nitrate gradually dissolved. It is best to keep the temperature between
20° and 25° ; below 20°, the nitrate is very slowly attacked, and above
25° there is danger of acetylating the aniline. The solution, which
was of a reddish-purple colour, was poured on to 150 grams of ice and
the turbid liquid extracted three times with ether. The ethereal
solution was shaken with water and then evaporated at a low
temperature. The oil thus obtained was poured into 200 c.c. of hot
containing 3 grams of barium hydroxide, the alkaline solution
filtere^Mipm a little solid (2 : 4-dichloro-6-nitroaniline), and then
exactly neutftiUsed with acetic acid. On cooling, the barium salt
Digitized by
Google
HIOHLT SUBSTITUTED NITROAMINOBENZENES. 813
crystallised in aggregates of plates, which were quite pure after re-
crystallisation :
0-4532, at 100^ lost 0-0466 H,0. HjO = 10-07.
0-4632 gave 0-174 BaSO^. Ba » 22 -69.
Ba(CgH30sNsOl2)s,3^H20 requires HjO- 10*26 j Ba=:22-43 per cent.
NitroaiDino-2 : 4-dichlorobenzene is prepared by extracting an acidified
aqueous solution of the barium salt with chloroform. The residue left
after evaporating the chloroform is dissolved in boiling petroleum,
from which the nitroamine separates in lustrous leaflets which may
have very irregular edges, or be elongated into bands. It is moderately
soluble in water, and crystallises therefrom in needles. It melts at
55 — 66° to a coloured liquid :
01834 gave 0-2524 AgOl. CI = 34-03.
CgH^OjNgClj requires 01 = 34-25 per cent.
This nitroamine is easily transformed into the isomeric 2 : 4-dichloro-
6-nitroaniline ; its aqueous solution slowly becomes yellow and de-
posits crystals of the aniline ; in solution in acetic acid, the same
change takes place, and is much hastened by the presence of a mineral
acid. From the purple solution which it forms with concentrated
sulphuric acid, water throws down the nitrated aniline.
^-Mtroamino-S :6-dibr(motoluene, CHg* CgHgBrj-NH'NOy— This sub-
stance is very easily prepared from 3 : 6-dibromo-p-toluidine by the
method described above for the nitroamino-«-trihalogen benzenes, which
it closely resembles. It crystallises in slender, curved needles melting
and decomposing at 122 — 123° :
0-2332 gave 0-282 AgBr. Br « 61 -44.
C^H^OjNgBrj requires Br = 61*59 per cent.
The barium salt crystallises in plates and is far more soluble in
water than the other barium salts which have been prepared :
0-288, at 100°, lost 0-0055 H,0. H^O =1-91.
0-288 gave 0-0828 BaSO^. Ba = 16-92.
B8L{0^B.fiJli(^Br^)^,lIfi requires H20 = 2*33 ; Ba = 17-76 per cent.
2-Nitroamin(h3 : Q-dibromotoluene^ CHg' OgH^Brj-NH-NOj, was pre-
pared from 3 : 6-dibromo-o-toluidine and crystallised in the usual
needles which melted at 112° and decomposed with evolution of gas
at 122° :
0-1355 gave 0-163 AgBr. Br = 61 19.
OyHjOjNoBr, requires Br =51 '59 per cent.
The harivm salt crystallises in sparingly soluble plates :
Digitized by VjOOQIC
814 CHATTAWAT : NITROGEN BEOMmsS CONTAINlNa
0-5234, at 100°, lost 00122 Hfi. H20= 2-33.
05234 gave 0-151 BaSO^. Ba - 16-9.
{Q^Ufi^'S^Br^)^Ba,,ILfi requires HgO=2-33 ; Ba= 17-76 per cent
Although a solution of this nitroamine in acetic acid containing
sulphario acid gives off bromine, a very small amount of a nitrated
base could only be obtained, and it was not possible to demonstrate
satisfactorily that this product was 3-bromo-5-nitro-o-toluidine.
The method here described affords a means of obtaining the nitro-
amino-derivatives of aromatic diamines, substances which have not
been hitherto prepared ; these compounds will be described in a
subsequent paper.
St. Babtholomew's Hospital and Collbge,
London, E.C.
LXXXII. — Nitrogen Bromides containing the Propionyl
Group.
By F. D. Ohattawat.
Nitrogen bromides are more difficult to prepare than nitrogen chlor-
ides, partly because hypobromous acid is less easily obtained and less
stable than hypochlorous acid, and partly because they undergo trans-
formation and are hydrolysed much more readily.
Except in colour they closely resemble the corresponding nitrogen
chlorides, and undergo similar isomeric changes when an unsubstituted
or partially substituted phenyl residue is also attached to the nitrogen.
Propionanilide and the bromopropionanilides readily react with
hypobromous acid, the iminic hydrogen atom being replaced by bromine.
A hypobromite is' probably formed as an intermediate product thus :
The action is a reversible one, for when nitrogen bromides are
placed in water the opposite change takes place, until a position of
equilibrium is reached. The ready hydrolysis with evolution of brom-
ine which nitrogen bromides undergo when heated with water depends
Digitized by
Google
THE PHOPlONirL GROUP. 81 5
on this and on the rapidity with which hypobromous acid passes into
bromine and bromic acid :
15H0Br - lOHBr + SHBrOg = GBr^ + BH^O + 3HBrOg.
These nitrogen bromides show the characteristic behaviour of the
nitrogen halogen linking, reacting readily with hydriodic add, sul-
phurous acid, hydrogen sulphide, alcohol, or potassium cyanide.
When hydrogen is attached to the phenyl nucleus, either in the
para- or ortho-position relatively to the nitrogen, the nitrogen brom-
ides undergo isomeric change more readily than the corresponding
chlorides. These changes are brought about in the same way as the
latter by heating alone or with water or an acid. Owing to the
readiness with which nitrogen bromides are hydrolysed, a little of
. the original anilide is generally re-formed in t^e process if water be
present.
The following scheme shows the directions of transformation, but
whilst the general course of the transformations is similar in the two
cases, it is important to note that the ^xira-derivative is the sole
product of the transformation of propionyl phenyl nitrogen bromide :
NBrPr NHPr NBrPr
\/ \ NHPr NBrPr NHPr
Br Br
All nitrogen bromides when treated with an excess of a solution of
hydrobromic acid are decomposed, bromine is liberated, and the corre-
sponding aniline is re-formed. On the other hand, when an anilide is
treated with an excess of bromine suspended or dissolved in water, a
nitrogen bromide and hydrobromic acid are produced. These changes
are the related parts of a reversible action, the direction of which is
determined by the conditions. An addition of halogen acid or halogen
to the nitrogen probably takes place thus : —
Br Br yPr Br
Br<3>N<^ + HBr^Br<^N^g^ + Br,.
Br Br ^H Br
A large excess of hydrobromic acid or the continuous removal of the
free bromine causes the anilide to be re-formed, whilst on the other
Digitized by
Google
816 CHATTAWAY: NITROGEN BROMIDES CONTAINING
hand the action proceeds in the opposite direction and a nitrogen
bromide is produced when a salt of a weaker acid, as an acetate or
borate, which can remove the hydrobromic acid, b added.
The reversible nature of the reaction is only seen clearly when
the nucleus is already fully substituted, as otherwise isomeric change
and substitution into the ring may take place.
EZPEBIMBNTAL.
Propionyl Phenyl Nitrogen Bromide, CaHj-NBr-CO-CH^'CHj.
This compound is prepared by shaking a solution of propionanilide
in chloroform with a little more than the calculated quantity of a
dilute solution of hypobromous acid* containing an equivalent weight
of potassium bicarbonate, the temperature not being allowed to rise
above 0°. The solution is separated, thoroughly dried over fused
calcium chloride, and the chloroform completely evaporated off in a
current of air, when a yellow, viscid mass is left which solidifies if
cooled below 0° and stirred with a little petroleum of low boiling point.
The pale yellow solid is dissolved in warm petroleum from which, on
cooling, the nitrogen bromide separates in slender, very pale yellow
somewhat irregularly grown pyramids. It melts at 88°.
0-3684 liberated I- 32-2 c.c. i^/10 I.t Br as :N-Br-= 34-94.
OoH^oONBr requires Br as :N*Br» 36*05 per cent.
When rapidly heated above its melting point, it darkens somewhat
in colour, and at about 125° is transformed almost explosively, with
considerable development of heat, into a reddish-brown mass consisting
mainly of />-bromopropionanilide. When melted and kept for some
minutes at its melting point, it quietly changes into />-bromopropion-
anilide, and solidifies to a white, crystalline mass, which, on further
heating, melts at 145—146°, or about three degrees below the melting
point of the pure isomeride. This transformation takes place slowly
on allowing the compound to stand exposed to mobt air, or rapidly on
heating it under water or on adding to its solution in chloroform a
little hydrobromic or hydrochloric acid ; some hydrolysis^ however,
takes place under these circumstances. Although a very careful search
has been made, no o-bromopropionanilide has been isolated from the
product obtained by the transformation of propionyl phenyl nitrogen
* The solutions of hypobromous acid referred to in this paper are easily made by
shaking bromine with about three times its weight of precipitated mercaric oxide
suspended in from 10 to 100 times its weight of water.
t All the nitrogen bromides and chlorides described in this paper were analysed
in the usual way. A weighed quantity was dissolved in chloroform and shaken with
an excess of a solution of potassium iodide acidified by acetic acid. The liberated
iodine was then estimated by a solution of sodium thiosulphate.
Digitized by
Google
THE PROPIONYL GROUP. 817
bromide ; j>-broniopropionanilide is apparently the exclufiive produot of
the change, and can be obtained pure by once crystallising from
alcohol.
^Br(mopropumanilid&, OgH^Br-NH-CO-OHj-OH,.
This is most economically prepared by mixing j9-ohloroaniline with
the calculated quantity of propionic anhydride, when mach heat is
developed, and heating for an hour at 120^. It is readily soluble in
alcohol or chloroform, and from the latter solvent it separates in
brilliant, colourless, rectangular plates with two domed edges. It
melts at 149°.
0-2147 gave 0-1766 AgBr. Br = 36.
O^HjoONBr requires Br=«35-05 per cent.
Propionyl ^Bramophenyl Nitrogen Bromide, O^H^Br-NBr-OO-CH^'CHj.
This substance can be obtained either from propionanilide,* or better,
from j>-bromopropionanilide by shaking a chloroform solution with a
strong solution of hypobromonsacid containing a little potassium bicar-
bonate and a little precipitated mercuric oxide. On filtering and
treating as described under propionyl phenyl nitrogen bromide, it is
obtained in bright yellow, transparent, glistening prisms and melts
at 78°:
0-3008 liberated 1 = 19-6 c.c. iV/lO I, Br as IN-Br = 26-05.
OgH^ONBrj requires Br as IN'Br- 26-04 per cent.
When heated above its melting point, it darkens in colour and is
transformed with development of heat at about 150 — 160° into a dark
coloured mass, from which 2 : 4-dibromopropionanilide can be isolated.
The transformation takes place with some little hydrolysis when the
compound is heated under water to 100°, and is best effected by dis-
solving it in chloroform, adding a drop of propionic acid, and heating
in a sealed tube for a short time at 1 00°.
Propionyl ^Bromophenyl Nitrogen Chloride, CgH^Br-NCl-CO-OHj^CHj.
This, like all the nitrogen chlorides described in this paper, was pre-
pared by shaking for a few hours a solution of the anilide in chloroform
with a half normal solution of potassium hypochlorite containing an
excess of potassium bicarbonate. The chloroform solution was separ-
ated, dried over fused calcium chloride, and the solvent evaporated. The
pale yellow oil thus obtained was cooled and treated as previously
* When an excess^of a strong solation of hypobromoos acid is used, the transforma-
tion of th3 propioQjl phenyl nitrogen bromide first formed takes place at the
ordinary temperature, even in presence of potassium bicarbonate and mercuric oxide.
Digitized by
Google
818 CHATTAWAY: NITROGEN BROMIDES CONTAINING
described until it solidified, when it was separated and recrystallised
from light petroleum.
Propionjl jo-bromophenjl nitrogen chloride crystallises in colourless,
transparent, four-sided plates, apparently flattened rhombs, and melts
at 59°:
0-4208 liberated 1 = 32-1 c.c. iV710 I. 01 as :N-C1 = 13-52.
OgHgONClBr requires 01 as :NC1 = 13-6 per cent.
When dissolved in chloroform, to which a drop of propionic acid
has been added, and heated slowly in a sealed tube to 100°, it is
transformed into 2-chloro-4-bromopropionanilide. .
o-Bromop7*opionanilide, OgH^Br'NH'CO'CBTg'CHj.
This is conveniently obtained by heating o-bromoaniline for two
hours to 120° with the equivalent quantity of propionic anhydride.
It is best recrystallised fi^m dilute alcohol, and finally from petroleum
(b. p. 60 — 80°), in which it is readily soluble. lb forms slender,
colourless prisms and melts at 93° :
0-1996 gave 01646 AgBr. Br = 35-09.
CgHj^GNBr requires Br = 35-06 per cent.
Propionyl o-Bromophenyl Nitrogen Bromide, OgH^Br^NBr'CO'CHj-CHg
This can be prepared from o-bromopropionanilide by the method
previously described. It ip, however, not necessary to prepare hypo-
bromous acid, as a freshly made, cooled solution of bromine in caustic
potash, to which excess of a solution of potassium bicarbonate has been
added, may be used. The latter may often be employed instead of
hypobromous acid in the preparation of those nitrogen bromides
which do not undergo transformation very readily.
Propionyl o-bromophenyl nitrogen bromide is very soluble in
chloroform, but only sparingly so in petroleum (b. p. 60 — 80°). From
the latter solvent, it separates in very pale yellow, long, transparent,
glistening prisms and melts at 117° :
0-2296 liberated I - 14-9 c.c. NjlO I. Br as :N-Br= 25-94.
CjHgONBrj requires Br as :N-Br = 26*04 per cent.
When slowly heated in a sealed tube to about 150° with a little
propionic acid, it is transformed into 2 : 4-dibromopropionanilide.
Propionyl (y-Bromophenyl Nitrogm Chloride, OeH^Br'NOl-OO-OHj-CHa.
This substance crystallises from petroleum (b. p. 60 — 80°) in
glistening, transparent, flattened rhombs and melts at 59° :
0-3392 liberated I « 25*8 c.c. iVyiO I. 01 as .'N-Ol = 13-48.
OjH^GNOlBr requires 01 as :N-01 = 13-5 per cent.
Digitized by
Google
THE PROPIONYL GROUP. 819
When slowlj heated in a sealed tube to about 150^ with a little
propionic acid, it is transformed into 2-bromo-4-chloropropionanilide.
2 ; A'Dihromopropumcmilide, CeHsBrj-NH-CO-OHg-CHg.
This is easily obtained by heating the aniline with the equivalent
quantity of propionic anhydride for 3 hours at 150^ It is readily
soluble in alcohol or chloroform, moderately so in petroleum, and crys-
tallises from alcohol in fine, long, colourless, silky needles which melt
at 136^ These become electrified on rubbing, and, when dry, the
particles fly apart on drawing a platinum spatula over them :
0-1670 gave 0*2046 AgBr. Br- 52-14.
OgHgNOBrj requires Br = 52*08 per cent,
Propwnyl 2 : i-JDibramophenyl Nitrogen Bromide,
CeHjBrg-NBr-CO-CHa-OHj.
This compound can be obtained from 2 : 4-dibromopropionanilide in
the ordinary way, and also by shaking for several hours a solution of
the anilide in chloroform with bromine suspended in water at 0^ con-
taining an excess of sodium acetate or borax ; the chloroform solution
is separated, thoroughly washed with water, dried over calcium chloride,
and treated as usual. It crystallises from petroleum (b. p. 60 — 80^)
in glistening, transparent, very pale yellow rhombs and melts at 87° :
0-3253 liberated 1 = 16-8 c.c. JV7IO I. Br as :N-Br = 20-64.
OgHgONBrj requires Br as :N-Br=20-71 per cent.
When heated slowly in a sealed tube with a drop of propionic acid
to about 120°, it is transformed into 2:4: 6-tribromopropionanilide.
Propionyl 2 : i-Dihromophenyl Nitrogen Chloride^
OgHgBrjj-NCl-CO-CHg- CHj.
This crystallises from petroleum in transparent, flat, four-sided,
apparently rectangular plates and melts at 71° :
0-6435 yielded 1 = 37-8 c.c, iVyiO I. CI as N:C1 = 10-41.
OgHgONClBrj requires CI as N:C1 = 10-38 per cent.
When heated in a sealed tube with a little propionic acid to about
120°, it is transformed into 2-chloro-4 : 6-dibromopropionanilide, which,
however, is somewhat impure, probably owing to the formation of a
little 2 : 6-dibromo-4-chloropropionanilida
2:4: ^THhromopropionomilide, CeHjBrj-NH-CO-CHj-CHj.
2 ; 4 : 6-Tribromoaniline was suspended in chloroform in which was
dissolved the calculated quantity of propionyl chloride, an equivalent
Digitized by VjOOQIC
820 NITROGEN BROMIDES CONTAINING THE PROPIONYL GROUP.
amount of pyridine wa43 then added, when at ottce the solution became
warm and the aniline dissolved. The liquid was boiled for an houi
and then poured ii}to water and heated until the chloroform had
volatilised. The white solid which separated was filtered off and
recrjstallised from alcohol, in which it was readily soluble. It separated
from this solvent in long, colourless, flattened prisms and melted at 203^:
0-1880 gave Q-2744 AgBr. Br - 62-1 1.
C^HgONBr, requires Br » 62*14.
Fropionyl 2:4: ^'Tribromaphenyl Nitrogen Bromide,
OeHjBrj-NBr-CO-CHj-OHj.
In the preparation of this substance from 2:4: 6-tribromopropion-
anilide, a solution of hypobromous acid or a solution of bromine in
caustic potash or bromine suspended in excess of a solution of sodium
acetate or borate can be employed. The last method shows the
reversible character of the reaction of a nitrogen bromide with hydro-
bromic acid.
This nitrogen bromide is readily soluble in chloroform, moderately so
in petroleum (b. p. 60—80°), and crystallises from the latter in clusters
of slender, flattened, bright yellow prisms melting at 82° :
0-3762 liberated 1-16-1 c.c. NJIO I. Br as :N-Br= 17-11.
CjHyONBr^ requires Br as :N-Br = 17-19 per cent.
When heated above its melting point, it decomposes at about 190°
with evolution of bromine, leaving a dark coloured residue from which
2:4: 6-tribromopropionanilide can be isolated.
Propionyl 2:4: ^-TSribrom/ophenyl Nitrogen Chloridey
CgHjBrg-NCl-CO-OHj-CHj.
This compound crystallises from petroleum in clusters of small,
colourless^ flattened prisms and melts at 75° :
0-2828 Uberated 1-13-5 c.c. N/IO 1. CI as IN-Cl » 8*46.
C^H^ONOlBrg requires CI as :N*C1»8-43 per cent.
It may be noted that, as in the case of the nitrogen bromides of the
corresponding chloropropionanilides, propionyl p-bromophenyl nitrogen
bromide and propionyl 2:4: 6-tribromophenyl nitrogen bromide are
bright yellow in colour, whilst propionyl o-bromophenyl nitrogen bromide
and propionyl 2 : 4-dibromophenyl nitrogen bromide are of a very pale
yellow. Propionyl phenyl nitrogen bromide itself is, however, very
pale yellow.
Chemical Laboratobt,
St. Baktholomew's Hospital and Colubgb, £.0.
Digitized by
Google
Digitized by
Google
Digitized by
Google