(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Colorimetric analysis"

Google 



This is a digital copy of a book that was preserved for generations on Hbrary 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 maiginalia 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 liave 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 fivm 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 attributionTht GoogXt "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 liabili^ 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/| 



OJLa-w^ lO 0^ M.I 



I 



l^at&atti College 3.tbiars 



GEORGE HAYWARD, M.D. 

(CIiu of 1809} 



SCIENCE CENTER LIBRARY 




COLOEIMETEIC 

ANALYSIS 



F. D. SNELL 



Il,I,VSTKATBn 



NEW YORK 

D. VAN NOSTRAND COMPANY 

Eight Wahreh Street 

1921 



Q„^_ \o,<^\,-^\ 



OCT I 1921* ) o 



Copyright, 1 92 1 
By D. Van Nostrand Company 



Printed in thk United States of America 



PREFACE 



It has been the endeavor of the author so far as is 
possible to avoid needless repetition of material in this 
work. Therefore all the fine points as to theory, appa- 
ratus and its use and the figuring of results which will 
apply to practically every test given, have been gathered 
together into three chapters at the beginning of the book. 
From this it will be seen that a knowledge of these 
chapters, particularly Chapters Two and Three, is neces- 
sary to the success of anyone except an experienced oper- 
ator, in the use of the methods of determination given on 
the succeeding pages. It has been the endeavor of the 
author to mention that a particular type of test was not 
applicable to those cases where the choice of methods is 
more limited than usual. 

This book is, in essence, an attempt to combine in one 
volume for ready reference, all the oolorimetric tests 
which experience has shown to be at all practical. Much 
material will be found here which has never been in print 
outside of the scientific magazines as well as the stand- 
ard tests which may be found by reference to any com- 
plete work on chemical analysis. In going over the 
material, it has been evident to the author that little has 
been done with regard to the determination of the range 
of greatest accuracy of many of the tests. Whenever 
such information is available it has been incorporated, 
so that the operator may know what degree of accuracy 
he may expect from a sample after he knows the ap- 
proximate amount of test substance contained in the 

sample. 

••• 

lU 



ly OOLOEIMETEia ANALTSIB. 

It is hoped that this book will furnish a source to 
which students may refer for a knowledge of the colori- 
metric methods practical for use and a source to which 
practical workers may refer for information as to the 
methods of performing the particular tests with which 
they are occupied. In all cases, weights of substances 
to be dissolved as standard have been given to four 
places, altho that by no means implies that absolute 
accuracy is necessary in the last two places, since the 
dilution will minimize any error made there. The giving 
of these accurate weights will, however, show the operator 
the direction in which the weight should tend in these 
last two places and thus can assist to greater accuracy 
without increasing the labor of weighing to any great 
extent. 

The author wishes to particularly acknowledge the 
splendid assistance rendered him by the Library of the 
Univei^ity of the State of New York in placing at his 
disposal their files of magazines and reference works. 
Acknowledgement is also made of the courtesy of Eimer 
& Amend, C. J. Tagliabue Co. and Arthur H. Thomas 
Co. in the furnishing of cuts. Acknowledgment is made 
of the services of C. A. Tyler and W. H. Pearce in the 
reading of proof. 

All factors have been figured by the 1918 table of 
atomic weights. 

F. D. S. 

New York, N. Y. 
July, 1921. 



CONTENTS 



CHAPTER SUBJECT PAGE 

I. Conditions of Use of Colorimetric 

Methods 1 

II. Apparatus Used and Methods of Using It. 6 

III. Figuring of Results 27 

IV. Determination of Iron 31 

A. By Potassium Sulfocyanate 31 

B. As the Chloride in Concentrated 

Acid 35 

C. By Potassium Ferrocyanide 36 

D. By Salicylic Acid 37 

E. As the Sulfid 39 

F. By Acetylacetone 41 

G. By Dimethylglyoxime 42 

V. Determination of Copper 44 

A. By Ainmonia 44 

B. As the Chloride in Concentrated 

Acid 46 

C. By Salicylic Acid 47 

Z>. By Potassium Ferrocyanide 48 

F. As the Sulfid 50 

F. By Potassium Ethyl Xanthate 51 

G, As the Bromide 53 

VI. Carbon in Steel 55 

VII. Lead, Bismuth and Arsenic 60 

Lead as the Sulfid 60 

Bismuth as the Iodide 62 

A. Arsenic by the Stain Produced on 

Mercuric Bromide Paper by Arsin. 64 

B. Arsenic by Silver Nitrate 67 

V 



Vi COLOEIMETEia ANALYSIS. 

CHAPTER SUBJEOT PAOS 

VIIL Aluminum and Chromium 68 

Aluminum by Alizarin-S 68 

A. Chromium as the Chromate 70 

B. Chromium by Disodium 1.8 dihydroxy- 

naphthalene 3.6 disulfonate 71 

IX. Nickel, Cobalt, Manganese and Zinc 73 

A, Nickel by Potassium Thio-carbonate . . 73 

B. Nickel as the Chloride in Concentrated 

Acid 74 

A, Cobalt as the Chloride in Concentrated 

Acid 76 

B. Cobalt by a-Nitroso-b-Naphthol 76 

A. Manganese as Permanganate, Oxida- 

tion by Persulf ate 78 

B. Manganese as Permanganate, Oxida- 

tion by Periodate 80 

Zinc by Eesorcinol 81 

X. Potassium and Magnesium 83 

A. Potassium by Determination of the 

Potassium Platino Chloride by Re- 
duction with Stannous Chloride . . 83 

B, Potassium as the Chlorplatinate by 

Potassium Iodide 85 

Magnesium by Determination of the 

Phosphate as Phosphomolybdate . . 86 

XI. Gold 89 

A. Method of Dowsett 89 

B. Method of Prister 90 

C. Method of Doring 91 

D. Method of Rose 92 

E. Method of Cassal 93 

F. By Decomposition of the Cyanide by 

Potassium Bromide and Sodium 

Peroxide 93 

G. By Decomposition of the Cyanide by 

Ammonia 94 

H. By Metaphenylenediamine 94 



CONTENTS. vii 

CHAPTER SUBJECT PAGE 

XII. Titanium, Vanadium and Tungsten 95 

A. Titanium by Hydrogen Peroxide 95 

B. Titanium by Thymol 97 

A. Vanadium by Hydrogen Peroxide ... * 97 

B. Vanadium by Strychnine 100 

Tungsten as the Oxide in Colloidal 
Suspension 101 

XIII. Fluorine, Chlorine and Perchlorates . . 102 

Fluorine by Estimation of Its Bleach- 
ing Action on an Oxidised Tita- 
nium Solution 1 02 

Chlorine by o-Tolidine 105 

Perchlorates by Methylene Blue 106 

XIV. Nitric and Nitrous Acids, and Ammonia . . 107 

A. Nitric Acid by Brucine Eeaction 108 

B. Nitric Acid by Diphenylbenzidene .. 110 

C. Nitric Acid by Phenolsulfonic Acid . . Ill 
/>. Nitric Acid by Pyrogallol 113 

A, Nitrous Acid by Sulfanilic Acid and 

a-Naphthylamine 114 

B. Nitrous Acid by Metaphenylenedi- 

amine Reaction 115 

G. Nitrous Acid by Zinc Iodide Starch 

Solution 116 

D. Nitrous Acid by a-Naphthylamine Hy- 

drochloride 117 

A. Ammonia by Nessler's Reagent 118 

B, Ammonia by Phenol 119 

XV. Phosphorus, Silica and Boron 121 

A. Phosphorus as the Phosphomolybdate. 121 

B. Phosphorus, Separated by Precipita- 

tion as Magnesium Phosphate 122 

Silica by Ammonium Molybdate in 
the Presence of Phosphorus 124 

A. Boric Acid by Curcumin 126 

B. Boric Acid by Turmeric Paper 127 



viii OOLOEIMETBIO ANALYSIS. 

OHAPTBR SUBJECT PAGE 

XVI. Oxygen and Hydrogen Peroxide 129 

A. Oxygen by Cuprous Chloride 129 

B. Oxygen by Adurol 132 

Hydrogen Peroxide by Its Oxidising 

Action on Ferrous Iron 133 

XVII. Sulfur, Hydrogen Sulfid and Selenious 

Acn> 135 

A. Sulfur by Paraphenylenedimethyldi- 

amine 135 

B. Sulfur by the Action of Hydrogen 

Sulfid on Arsenious Oxide Paper . . 136 
Hydrogen Sulfid by Methylene Blue. . 137 
Selenium as Selenious Acid by Potas- 
sium Iodide 138 

XVIII. Salicylic Acid and Cyanides 140 

Salicylic Acid by Fehling's Solution. . 140 
Cyanides by Changing to Sulfocya- 
nates and Coloring with Iron 141 

XIX. Color of Water, Oils and Dyes 144 

Color of Water 144 

Color of Oils 145 

Color of Dyestuffs 145 

XX. Nephelometry 147 



COLORIMETRIC ANALYSIS 



CHAPTER ONE. 

CONDITIONS OF USE OF COLORIMETRIC METHODS. 

CoLORiMETRic Hiethods of analysis consist of treating 
a solution of the substance to be tested with a reagent, in 
such a way as to produce a color which is proportional 
in intensity to thelTmount of the substance being tested 
for, present in the solution. The methods are not only ap- 
plicable to the determination of many metals but also to 
acid radicals and compound radicals to a limited extent. 
The unlmown is therefore spoken of in this general dis- 
cussion as the test substance. The color having been pro- 
duced, the solution containing an unknown amount of 
test substance is then estimated by comparison with a 
standard solution by one of the following four methods: 

1. The sample diluted to a definite volume may be 
compared with a series of standards of the same volume, 
previously prepared, the amount of test substance in 
which is known and the value of the unknown then taken 
to be that of the standard to which it conforms most 
nearly. In this way the amount of test substance present 
is obtained without figuring since, if the volume and color 
of the unknown and the standard are the same, their 
contents of test substance will also be identical. 

2. The standard and sample may be placed in similar 
graduated tubes and the darker diluted until its color 
when observed horizontally thru the tube is the same as 
that of the other. When this point is reached then each 

1 



2 OOLOBIMETBIC ANALYSIS. 

C.C. of one solution must contain exactly the same amount 
of test substance as each c.c. of the other and the amount 
in the unknown is to the amount in the known directly 
as their volumes. This is the method of dilution. 

3. A definite sample of unknown may be placed in one 
flat bottom graduated tube and amounts of the standard 
poured into another similar tube until their colors, when 
observed vertically thru the length of the columns of 
liquid, are identical. The amount of test substance in 
each tube will then be the same and since the amount per 
c.c. in the standard is known the total amount in the 
standard may be computed, which amount is identical 
with the total amount in the sample. To state it more 
briefly, their concentrations are inversely proportional to 
their depths. This is the method of balancing. 

4. The sample may be made up to a definite volume 
and nearly that volume of water in a similar container 
be treated with the same reagents for bringing out the 
color of the solution as were used with the sample. A 
concentrated solution of standard is then run into the 
blank from a burette, adding it drop by drop when the 
end point is near. The volume of the blank may be 
brought up by the addition of more water until the two 
colors and volumes are identical. The amount of stand- 
ard used in the making of the duplicate is then a measure 
of the amount of test substance in the sample. This is 
the method of duplication. 

In order for the colorimetric test to be accurate the 
color produced by the action of the given reagent on the 
test substance must be the only color present in the solu- 
tion. Therefore a colorimetric test is not possible if the 
original solution is colored before the test is begun, un- 
less that color is produced by the test substance or will 
be removed by the reagent used. In some particular in- 



CONDITIONS OP USB. 3 

stances a slight contamination of one color may be allowed 
and equalized by contaminating the sample and standard 
equally. Neither is a colorimetric test possible if the 
solution to be tested contains anything other than the test 
substance which will give a precipitate or color with the 
given reagent. It is essential for the success of colori- 
metric methods that the color produced be reasonably 
permanent and that the conditions under which it was 
produced be such that they can be duplicated without 
great difficulty. Under certain conditions tests are made 
with colors which fade after short exposure to the light. 

Colorimetric tests may be classified into two classes 
according to the reason for their use. Some find their 
popularity because they are fast, and the accuracy of the 
test is sacrificed for speed in obtaining the final result. 
The second class find use because they furnish a method 
of determination of small amounts of substances with 
greater accuracy than is possible by gravimetric or volu- 
metric methods. The first class may all be made in a 
short time but it may often happen that the method of 
preparation of the second class of tests may take hours 
to insure the accuracy desired. 

In spite of the limitations placed on its use by the 
previous requirements the colorimeter is coming into use 
more and more every day because of its answer to the 
demand in nearly every laboratory for speed. Colori- 
metric methods, used because of their speed, give results 
in five minutes to one hour from the time the test is 
begun which is in all cases less than half the time similar 
tests could be made by other methods. A leading bra&s 
manufactory of the country obtains an analysis of its 
brass from the laboratory within forty-five minutes after 
the delivery of the sample. Of the five constituents 
determined, four are determined by colorimetric methods. 



4 COIiOBIMBTBIC ANALYSIS. 

Thus the main thing to be said for colorimetry is that 
its methods are rapid and reasonably accurate. 

A broad field of usefulness of colorimetry is the 
analysis of salts or substances easily soluble in water 
or acid for the amounts of various impurities present. 
The methods are very delicate and accurate, some being 
so delicate as to detect one part in one hundred million 
parts of water, but they are seldom such as to permit 
of the determination of quantities greater than one per 
cent without resorting to aliquot parts and using a por- 
tion of the solution of the sample instead of the whole. 
The methods of accurate dilution are discussed in the 
next chapter; the only drawback to their use is that the 
use of a factor for determining the final result multiplies 
the error made in the comparison, often as much as a 
thousand times. 



CHAPTER TWO. 

APPARATUS USED AND METHODS OF USING IT. 

As previously outlined there are four types of deter- 
minations made by colorimetric methods, each of which 
uses a moi^ or less specialized form of apparatus. 

In cases where a series of permanent standards are 
made up these are usually placed in either round or 
square glass bottles. Such a series of standards should 
be placed in a row on a shelf, with sufficient space between 
each two for setting in a similar bottle. The sample should 
be treated in a similar bottle and, after making up to the 
same volume as that of the standards, it may be set into 
various of the gaps between the standards until the place 
is found where one standard is higher than the sample 
and the one on the other side lower, as estimated from 
the intensities of their colors. The position of the 
sample relative to these two known quantities can then 
be estimated. In some cases when permanent standards 
are used one bottle may be made to serve as two stand- 
ards. This is accomplished by the use of a rectangular 
bottle, twice as long as it is broad. The sample is then 
compared in a similar bottle and results are as follows: 
If the long way of the sample bottle compares with the 
long way of the standard the amounts of test substance 
in the two are the same. If the narrow way of the 
sample compares with the long way of the standard the 
sample contains twice as much test substance as the 
standard. If the long way of the sample compares with 
the narrow way of the standard the sample contains half 
as much test substance as the standard. Altho no fur- 

5 



6 OOLOEIMETEIC ANALYSIS. 

ther specific reference will be made to their use, this use 
of these special bottles is applicable in all cases where 
permanent standards are to be made up, except those 
where the actual color rather than the depth of color 
changes with the concentration. Care should be taken 
to see that the bottles used in all cases are clear in color 
and free from flaws. 

In many cases the color produced by a reaction fades 
in a short time in the light. In such a case either a 
standard must be prepared at the same time as the 
sample and test made by one of the other methods or 
permanent artificial standards must be prepared. These 
are conveniently prepared from solutions of inorganic 
salts, combining the colors in such a way as to give the 
desired color. The solutions which find the greatest 
favor for this purpose are half normal solutions of the 
nitrates or chlorides of cobalt, iron and copper.^ These 
three solutions may be combined in such a way as to 
form any color desired except the deep blues and reds. 
To remedy this it has been found that some of the miss- 
ing colors can be obtained by using a solution of potas- 
sium dichromate instead of the iron solution and the 
remaining colors are obtained by combinations of potas- 
sium dichromate and potassium permanganate. As an 
example it is found that a mixture of the first series 
suggested, cobalt three parts, to iron nine parts, diluted 
with water, corresponds to the color given by Nessler's 
reagent in reaction with ammonia. Varied dilutions of 
this may be used for a series of standards, making each 
standard to correspond to the result obtained by the use 
of that known amount of ammonia treated with the 
reagents necessary to bring out the color. In a similar 
manner permanent standards may be prepared for any 

1 J. Franklin Institute, 180, 200. 



APPABATtJS USED. 7 

test, keeping in mind the fact that every standard must 
be tested against a I^nown amount of the test substance 
to make sure of its accuracy. 



Pm. 1. Back of Straight Graduated Tubes. 

For the production of nonpermanent standards plain 
test tubes are ordinarily used. The solution may be 
quickly and conveniently emptied from these and the 
loss in case of breakage is not so great as from the use 
of graduated tubes such as Nessler tubes or from the use 
of bottles. 

The apparatus for tests by the dilution method con- 
sists essentially of a pair of graduated tubes. There is, 
however, a very convenient device (Fig.l4) manufactured 
for use with this method known as the colorimetric 
camera. This is a light-proof box, painted black inside, 
with holders for the two tubes near one end. The end 
near the tubes is fitted with a ground glass screen; the 
other end is fitted to the face of the observer so that no 
side light may enter. By the use of this a more accurate 
judgment as to the colors of the two tubes may be mad© 
than if they are compared out in the open. One advan- 



8 OOIiOBIMBTEIC ANALYSIS. 

tage of this apparatus is that it does not tire the eyes 
of the operator so quickly as do some of the types of 
colorimeter used for the balancing method. In all cases 
possible a colorimeter should allow the use of both eyes 
for making the comparison so as to eliminate fatigue. 
Sometimes the comparison of the two tubes is made by 
holding them in the hand and looking at them against 
a sheet of white paper. 

Whether or not the camera is used for the comparison 
the operation of the dilution method is always the same. 
The standard and sample are in two graduated tubes or 
in two Nessler tubes. The former are used when the 
camera is used for the comparison, the latter are useful 
when no further apparatus is used. The colors of the 
two solutions should be nearly alike to begin with. The 
experienced operator soon learns to choose his sample 
so that the resultant color will be nearly the same as that 
of the standard which he is using. Water is then added 
to the solution of the darker, carefully mixing after each 
addition, until the colors of the two solutions when ob- 
served horizontally thru the tubes appear to be identical. 
When this point is reached the concentration of the test 
substance in each solution must be the same and their 
contents are then to each other as the volumes. Since 
the amount of test substance in the amount of standard 
used is definitely known, it is easy to figure the test sub- 
stance in the sample by a direct proportion. Care should 
be taken that the tubes used are clear and free from flaws 
and that their thickness of glass, internal diameter and 
graduations are identical. A special type of tube (Fig. 
13) manufactured for carbon determinations is applicable 
to all dilution tests. This has a bend about two inches 
from the top so that the tube may be lightly shaken to 
mix the water added without danger of the contents being 



APPARATUS USED. 9 

spilled. Note that the standard may as well be diluted 
to match the sample as the sample to match the standard, 
provided only that the original volume of the standard is 
recorded and that the final colors are not too deep or too 
light to allow of accurate measurement. 

Determinations by the method of duplication are 
usually carried out in graduated Nessler tubes. The 
sample is first diluted to some convenient, definite volume. 
The same reagents as were added to the sample are then 
added to a volume of water amounting to about half 
the volume of the sample, this will usually be specified 
for the particular experiment since the amount of water 
used for the blank varies according to the concentration 
of the standard to be used. To the blank containing the 
same reagents as those used for the sample there is now 
added a standard solution, carefully mixing after each 
addition, until the color of the sample is duplicated by 
that produced in the blank by the addition of the stand- 
ard. The color of the blank having been made to bal- 
ance that of the sample they will still differ in volume. 
It is simple enuf to figure at this point by proportion the 
amount of standard that would be necessary if the 
volumes were equal but, to eliminate any chance of error 
in that, it is best to duplicate the volume as well as the 
color of the standard. This is accomplished by the addi- 
tion of water and standard alternately until the two solu- 
tions are identical in both color and volume. As said, a 
little simple mathematics will tell an operator of compara- 
tively little experience the exact amounts of standard 
and water to be added after the colors have once been 
duplicated. The same cautions apply to this test as to the 
previous one insofar as apparatus is concerned ; the tubes 
used must be of the same size, thickness of glass and 
internal diameter and the graduations on the outside 

2 




10 COLORIMETRIG ANALYSIS 

must correspond in height. That is, the ten c.c. mark on 
each must be at the same distance from the bottom, other- 
wise the contents are not the same or the tubes are not 
accurately marked. 

The apparatus for the balancing method is the most 
elaborate of all and the use of it, the simplest. The 

various instruments used include 
Hehner cylinders, Lovibond Tin- 
tometer and various types of colori- 
meters. The Campbell - Hurley, 
Duboscq, Schreiner, Stammer, Say- 
bolt, and White types are described. 
Other specialized forms are used 
in some special applications but 
their use is mostly in physiological 
chemistry. 

Fia 2 Hehner ^^ ^^^^^^ ^^^^^^^ ^^^ ^^ 

Cylinders. . *^ , 

simplest of the instruments for 
this use. They consist of two glass tubes with flat 
bottoms and each has a side tube with stop-cock, about 
three inches from the bottom. For the carrying out of 
a determination the solution of the sample is placed in 
one tube and the other tube is partially filled with the 
standard so that the depth of color seen by looking verti- 
cally downward thru the length of the column of liquid 
is deeper than that seen by similarly looking down thru 
the sample. An amount of standard is then drawn oflf, 
such that the colors of the two tubes when observed in 
this way are the same. The tubes are then said to be 
balanced and the readings of the volumes taken. The 
amount of standard used and its content of the test sub- 
stance per c.c. is known. The test substance in the sam- 
ple is then inversely proportional to the volume if the 
content per c.c. is desired, or the total amount of test 



APPAEATOS USED. 



PiQ. 3. Campbell-Hurley Colorimeter. 



12 COLOEIMETRIO ANALYSIS. 

substance in the two solutions is identical. For the con- 
venient balancing of the two liquids several mechanical 
appliances have been devised for the purpose of increas- 
ing the speed of the determination or of increasing its 
accuracy. 

The Campbell-Hurley type of colorimeter illustrated 
is the common form of such an instrument. This instru- 
ment^ is a modification of the Kennicott- Sargent type of 
apparatus and is sometimes known as the Kennicott 
colorimeter. The following is the method of operation 
as illustrated by the diagram. The unknown solution to 
be tested is placed in the tube A and, since the volume 
can be readily governed so as to come to some even 
graduation, these are only placed at five c.c. intervals. 
The standard solution is placed in the right-hand tube 
B which, because of the fact that the accuracy of the test 
depends upon the careful reading of the volume in this 
tube, is graduated to single c.c. The readings of tubes 
A and B have been given as in c.c. but, for such read- 
ings to be accurate, tubes A and B must be of the 
same identical bore. The same effect is obtained in less 
accurately made instruments by having the graduations 
to cm. rather than c.c. The proportion may then be 
obtained between the heights of the two columns, the same 
as tho the units used were c.c, since the columns will 
have colors proportional to the depths regardless of 
their amplitude. The method used for graduation must 
be taken into account in figuring the results of the test 
and will be discussed under that head. Tube B is per- 
manently connected by a glass tube with the reservoir C 
in which the glass plunger D works, so that the level of 
the liquid in B may be readily controlled by raising or 
lowering the plunger. As the tube B and the reservoir 

I J. A. C. S., ss, 1112. 



APPABATUS USED. 



13 



C are made in one piece the standard comes in contact 
with glass only, thus preventing the possibility of chem- 



E 



K 



1. ^ t ^ ' ^.n * ^ m.ti^^t.t.m.t.m* ^.ti.!.^^^ ^ ^r^ 








=— ": =A 



c-^l 



-B 












J 



2i. 



C 









Fig. 4. Campbell-Hurley Colorimeter. 

ical change due to coming into contact with the container. 
The plunger D is sometimes provided with a rubber 
collar E to prevent it from coming into contact with the 



14 COLORIMETEIO ANALYSIS. 

bottom of the reservoir and the resultant breakage pos- 
sible. The tubes A and B and reservoir C rest on wooden 
supports with holes under A and B for the passage of 
light. All glass parts are held in place by spring clips 
which allow for the easy removal of the parts for cleaning. 

For operation the colorimeter is turned with the back 
toward a window, preferably a north one, and the mirror 
G is so adjusted as to reflect skylight upward thru tubes 
A and B. By this anangement the back of the colori- 
meter serves as a screen to cut oflf all light except that 
reflected upward from G. The light, passing upward 
thru the tubes A and B, impinges on the two mirrors 
H and / cemented to brass plates sliding in grooves cut 
at an angle of 45° in the sides of the wooden box /. 
This box has a loosely fitted cover so that it may be 
removed for the cleaning of the mirrors. The mirror 
// is cut vertically and cemented in such a position as to 
reflect one half of the circular field of light coining thru 
the tube A. The light, passing upward thru B^ is re- 
flected horizontally by the mirror / thru a hole in the 
brass plate supporting the mirror H. One half of the 
circular field of light from the tube B is cut off by the 
mirror ZT, the vertical edge of which acts as a dividing 
line between the two halves of the circular field. The 
image of one half of the tube B is then observed in juxta- 
position to the opposite half of the image of the tube A. 

The juxtaposed images are observed thru a tube K^ 
2.5 cm. in diameter and 16 cm. long, lined with black 
felt and provided with an eyepiece having a hole 1.5 mm. 
in diameter. At the point M in the tube K is placed 
a diaphragm having an aperture 8 mm. in diameter. All 
parts inside the box / except the mirrors are painted 
black so that no light except that coming thru the tubes 
A and B passes thru the tube K. By having the aper- 



APPARATUS USED. 15 

tures in the eyepiece and the diaphragm properly propor- 
tioned only the images of the bottoms of tubes A and /? 
can be seen, thus preventing the interference of side light 
from the vertical sides of the tubes. 

A person looking thru the eyepiece observes a single 
circular field divided vertically by an almost imper- 
ceptible line when the two solutions are of the same 
intensity. By manipulating the plunger D the level of 
the liquid in B can be easily raised or lowered, thus 
causing the right half to assume a darker or lighter 
shade at will. In matching colors with an ascending 
column in iS, that is, gradually deepening the color of the' 
right half of the field, the usual tendency is to stop a 
little below the true reading, while in comparison with a 
descending column the opposite is the case. At first the 
operator should take a reading in each direction until 
after a little practice this tendency to error has been 
overcome. 

In tests on a large number of titanium solutions by 
oxidation with hydrogen peroxide, using all concentra- 
tions from a very dilute (light yellow) to a fairly con- 
centrated (deep orange) solution, the average percent of 
error was found to be less than one percent and the 
median error less than one half percent. 

In some cases the depth of the column of liquid ob-. 
served cannot be used as a measure of the test substance 
present as the concentration changes not only the depth 
of the color but the color itself. In cases where that is 
so, it will be mentioned in connection with the directions 
for the particular test. 

A modified form of balancing instrument has a mirror 
directly over the tubes A and B reflecting forward the 
two circles of light which come to it. The mirror is 
encased in a lightproof box painted black on the inside 



COLORIMETBTC ANALYSIS. 



PiQ. 5. Lovibond Tintometer. 



APPAEATTIS USED. 



17 




/<^rf^^y. 



K 



2L 







Fig. 6. Duboscq Colorimeter. 



Ig COLORIMETRIC ANALYSIS. 

and about eight inches long. This is fitted to the eyes of 
the operator at the forward extremity so that side light 
cannot enter. The arrangement of the tubes A^ B and O 
and plunger D is the same as in the previous instrument. 
By making the holes under the tubes A and B so small 
that they do not extend to the side walls of the tubes 
side light is largely eliminated and accurate results 
secured. In this case the alteration is made in the color 
of the right hand circle of light as compared with a 
circle of similar size appearing beside it. 

The Lovibond Tintometer is another type of instru- 
ment, one which we shall not describe here at any length 
on account of the limited use that it finds. The instru- 
ment is expensive and so is comparatively rare. The 
comparison is made by graduating a liquid to match a 
glass of a certain color, thus doing away with the use 
of a liquid standard. Since a permanent liquid standard 
may be prepared for this use and so eliminate large 
expense this is not recommended. Glasses may be ob- 
tained for determining carbon in steel and color of water, 
petroleum, etc., in addition to which by combinations of 
glasses the color of any solution may be matched. 

In the Duboscq type of instrument the same result as 
in the Campbell-Hurley type is obtained by a different 
method. The two independent tubes, A and iS, are of the 
same size and are for holding the solutions of the un- 
known and the standard. Each is mounted in a wooden 
holder, M^ N, which slides up and down in a slit cut in 
the backboard of the instrument and which is fastened 
at the chosen place by a thumbscrew. Light is reflected 
upward thru the tubes by a mirror G. Directly over 
tubes A and B which contain the solutions to be tested 
are two glass plungers, (9, P, of a diameter equal to about 
half that of A and B. The bottoms of these plungers 



APPAEATUS USED, 19 

are finely ground and in the best instruments are fused 
on. In the cheaper instruments the bottoms are fastened 
with some sort of adhesive. 

The telescope, K^ for observation of the colors is per- 
pendicular to the base so that the operator looks down- 
ward into the instrument. The light reflected upward 
thru the solutions in A and 5 is so reflected by the 
prisms, /, /, in the box / that two fields appear side by 
side, one from A and one from B, The arrangement of 
the prisms is such that the images observed in the field 
of the telescope are those of the bottoms of the plungers, 
(9, P^ rather than of the entire depth of liquid in A and 
B. By suitable cutting down of the aperture by screens 
all reflection from the sides of the tubes is cut off. 

For use, the instrument is set to face a source of light 
and mirror G is set at the proper angle to reflect skylight 
upward thru A and B, The cylinder having the lighter 
color is then moved upward by sliding its holder in the 
slit in the backboard until the plunger just touches the 
surface of the liquid. The other container is then moved 
upward observing its movement thru the telescope until 
the image of the base of the plunger in that liquid ap- 
pears to be of the same intensity as that observed from 
the other field. The instrument is then said to be bal- 
anced and the depths of liquid underneath the plungers 
have the same relation to each other as the total depths 
of liquid in A and B when the Campbell-Hurley instru- 
ment is balanced. The slits cut in the backboard in 
which the holders of A and B move are graduated so 
that the depths of liquid may be read off directly from 
these and possible errors in reading depths of liquid in 
glass due to the meniscus are eliminated. 

The laws relating to the depths of liquid when the 
instrument is balanced are as follows. The concentration 



20 COLORIMETRIC ANALYSIS. 

per c.c. of the sample is to that of the standard inversely 
as the depths of their liquids. Diameter of the contain- 
ing vessels Iieing the same, the sample solution of the 



Fio. 7. Large Duboaeq Colorimeter. 

depth indicated by A contains identically the same 
amount of test substance as a standard column of the 
depth indicated by B. The figuring of results from this 
data is taken up in the next chapter. The nephelo- 
meter is a modification of this type of instrument. For 
further information regarding that use see Chapter 
Twenty. 



APPABATirs USED. 21 

The Schreiner colorimeter is a simple modification of 
the Diiboscq instrument, eliminating many of the more 
expensive features. The prisms are 
omitted and two round fields near 
each other are observed. The hold- 
erg for A and B are two brass clips 
and the reading is taken by gradua- 
tions on the glass of these tubes. 
The application of this instrument 
is recommended by the manufacturers 
for soil analysis. 

The Stammer colorimeter is also 
a modified form of the Duboscq, and 

is particularly used for determination „ ^ ^ , ^ , 

. , , . , . , . Pig. 8. Small Duboscq 

of the color of sugar solutions and of Colorimeter 

oils. The colors of the fields are 
transmitted to the telescope, K, by prisms, /, /, as in 
the Duboscq type. The alteration is in the character of 
the fields. Instead of two movable containers and two 
fixed pistons as in the other instrument the containers are 
fixed and one movable piston, /*, is provided for the 
variation of the column in B. A false piston, (?, is pro- 
vided so that the light thru A will have to pass thru a 
similar thickness of glass. The column in A remains 
absolutely permanent, no variation in that being jwssible. 
Light is reflected upward by a mirror at G as in previous 
instruments. 

Recently a new type of instrument has been evolved 
from the combination of the Lovibond Tintometer and 
the Stammer Colorimeter, known as the Saybolt Chromo- 
meter. This instrument is intended solely for the evalu- 
ation of the colors of lighting oils. This is similar to 
the Stammer instrument in that two tubes are provided, 
but the left-hand tube, 4, is modified to an empty tube 



22 COLOEIMBTEIC ANALYSIS. 

in which the color is obtained by the addition of one or 
more carefully prepared glass disks. The sample in B 
is drawn off until only 
such an amount is left 
as will duplicate the 
color of the disk used, 
From the graduations on 
the glass of B and the di- 
rections of the makers the 
oil may then be graded as 
so many degrees above or 



fia. 10. StamineT Colorimetei. 



APPAEATUS USED. 23 

below the standard colors for lighting oils (see Chapter 
Nineteen), depending on the disk used. 

One other type, White's colorimeter,^ shows promise of 
great importance because of its ease of operation. In 
this the thickness of two glass prisms observed is the 
variable. The apparatus consists essentially of two 
wedge-shaped hollow glass prisms of exactly equal 
dimensions, open at the large end for the introduction of 
the solutions to be tested. The wedges are held in 
vertical position side by side in a camera and may be 
raised or lowered by rack and pinion actuated by thumb- 
screws. The wedges are screened from view on the side 
of the operator, except for a narrow horizontal slit across 
the middle of the camera thru which the solutions are 
observed when a test is being made. The carriers are 
graduated to correspond to the length of the wedges, the 
zero of the scale being opposite the indicator when the 
sharp edge of the wedge is opposite the narrow opening 
in the screen thru which the color is observed. The 
screens are adjustable so that the opening may be altered 
to suit the operator. The ground glass shutter at the 
forward end of the camera, for diffusing the light, is 
hinged in the manner of a door to facilitate the transfer 
of the wedges to and from the camera. To carry out a 
determination by this method equal quantities of the 
standard and of the material to be tested are diluted to 
equal volumes and convenient amounts of the solutions 
transferred to the wedges. The wedge containing the 
solution of unknown strength is set at the graduation 
representing the percentage, or a multiple of it, of the 
coloring matter in the standard. The wedge containing 
the standard is then adjusted so that the two fields seen 
thru the camera appear identical. The percentage of 

1 J. A. C. S., S4, 659. 



COLORIMBTBIC ANALYSIS. 



coloring matter in the unknown is then indicated by the 
reading of the scale on the carrier containing the stand- 
ard. If the depths of color compared at first are too 
deep for accurate comparison the results may be checked 
without changing the solutions by setting tlie wedge con- 
taining the unknown at a new point on the scale and 



Flo, 11. WliiM'8 Colorimeter. 

again adjusting the other wedge until equality is reached. 
The maximum error reported on a long series of experi- 
ments with this instrument was .6 percent. 

The figuring of results obtained by these balancing 
methods is discussed in the next chapter. 

In some cases it may be impossible to choose a sample 



Al>PABATUa USED. 25 

having as small a content of the test substance as speci- 
fied in the method, particularly if the substance being 
analysed is an ore or alloy containing a fairly large per- 
centage of the test substance. In that case the sample 
may be taken containing somewhat more of the test sub- 
stance than specified for the test and dissolved as 
directed. This solution may then be accurately diluted 
and an aliquot part taken. If, for example, the method 
specifies that a sample must be under .001 gram of the 
test substance and the smallest sample of the substance 
being analysed which can conveniently be weighed out 
will contain nearly one tenth gram of the test substance, 
this may be dissolved as specified and diluted to one liter 
in a volumetric. Ten c.c. of this may then be analysed 
as directed and the weight of test substance found in that 
amount of solution multiplied by 100 will give the 
amount in the original sample. In case it is necessary 
to use such a method the final result cannot be expected 
to be as accurate in regard to the amount of test sub- 
stance by weight as shown in the result but the results 
are identically as accurate so far as the percentage of error 
is concerned. To illustrate, while accuracy to .6 percent 
in the case of a w^eight of substance showing .001 gram 
test substance would allow an error of only .000006 gram 
test substance, if the sample showed .5 gram test sub- 
stance (an amount greater than would ordinarily be 
present in a weight of sample used), the maximum error 
then possible would be .003 gram of the test substance. 

The rapidity with which colorimetric tests may be 
made often admits of the attainment of considerably 
greater accuracy than the theoretical by rapidly doing 
several tests and then averaging the results obtained. 

The intensity of the color of the two solutions is an 
important factor, so results may be expected to be more 

3 



26 COLORIMETRIG ANALYSIS. 

accurate if the standard and sample have nearly the same 
color before a test is carried out by dilution or balancing 
methods than if the colors differ greatly. The intensity 
should also not be too great or too dilute ; in either case 
errors will be greater than if the color produced was 
intermediate between the two extremes. 



CHAPTER THREE. 



FIGURING OF RESULTS. 



In case a series of standards is used the weight of test 
substance in the unknown may be told at once by looking 
at the amount in the standard to which it corresponds. 
The percentage is then readily figured from this by 
dividing by the weight of sample used, thus : 

Weight of sample, 5 grams. 

Weight of test substance present, .0004 gram. 

Sample therefore contains .0004 -^ 5 = .00008 gram test 
substance per gram sample = .008 percent. 

The figuring of results when the dilution method is 
used is somewhat more complex. In that case the darker 
of the two solutions is diluted until the colors of the two 
are identical when observed horizontally thru the tubes, 
at which point the content of test substance per c.c, is 
the same and the content of test substance of one is to 
that of the other directly as their volumes. To illustrate : 

Weight of standard used, .2 gram. 

Weight of unknown used, .2 gram. 

Standard contains .32 percent test substance. 

Readings, standard 38 c.c, unknown 45 c.c. 

The weights of the sample and standard being the 
same in this problem we may eliminate the figuring of 
the weight of test substance in the sample and proceed 
directly to estimate by proportion the percentage. This 
is estimated thus : 

Percent in standard : Percent in unknown = Volume of 
standard : Volume of unknown : 

27 



28 OOLOBIMBTEIG ANALYSIS. 

.32:X=38:45 
Z=.379 

Therefore we find that the sample being tested contains 
.38 percent test substance. 

The figuring of dilution method results does not always 
resolve itself into a mere comparison of percents as 
usually a definite volume of a standard solution whose 
content of test substance per c.c. is known is used as the 
standard. In that case the weight of test substance in 
the standard must be figured and from that the similar 
weight in the sample and then the percentage present in 
the sample. The method follows: 

Weight of sample used, 2 grams. 

Standard used, 20 c.c. of a solution containing .00002 
gram test substance per c.c. 

Eeadings, standard 20 cc, unknown 48 c.c. 

From the readings above it will be seen that in this case 
the color of the sample was darker than that of the 
standard and therefore the sample was diluted. There 
would be no change in the method had the standard been 
diluted. 

Weight of T. S. in standard : Weight of T. S. in 
unknown = Volume of standard : Volume 

of unknown. 

Weight of test substance in standard is 20 X .00002 = 
.0004 gram. 

.0004:X = 20:48; 
X = . 00096. 

Therefore the two grams of sample used contained 
.00096 gram of test substance and the percentage in the 
sample was .048 percent. 

Eesults by the method of duplication are easily ob- 
tained. A standard solution is added to a blank con- 



FTGUEING THE EESULTS. 29 

taining the same reagents as the sample until the color 
of the sample is matched. Water and standard are then 
cautiously added, alternately, until the volume as well as 
the color of the two solutions is identical. The result as to 
how much test substance is present in the sample is given 
by the amount necessary to form an identical standard. 
The percentage in the sample is then figured, thus : 

Weight of sample, 5 grams. 

Standard used contains .0002 gram test substance 
per c.c. 

Standard used for duplication 2.3 c.c. 

Then the total test substance used was 2.3 X .0002 = 
.00046 gram. 

Therefore the sample contained .00046-7-5 gram test 
substance per gram of sample or .000092, which is .0092 
percent. 

Results obtained by the balancing method are some- 
what more complicated to obtain, especially if the ap- 
paratus used is graduated to cm. instead of cc. 

To take, first, a case when the apparatus is graduated 
to c.c. and tubes A and B are identical, whether they be 
the Hehner cylinders or Campbell-Hurley apparatus or 
some other type of instrument, 

Standard used contains .00002 gram test substance 
l^r c.c. 

Volume of standard used to balance colors, 44 c.c. 

Sample used, 4 grams. 

The amounts of test substance in sample and standard 
are the same when the colors are balanced, that is the 
volumes differ but the total content is the same. There- 
fore the amount of test substance in the sample being the 
same as in the standard is 

44 X .00002 = .00088 gram. 



30 COLOBIMETEIG ANALYSIS. 

Since the sample used was 4 grams it was therefore 
.022 percent test substance. 

To take the other case of cylinders graduated to centi- 
meters, the results obtained with a cheaper instrument 
are accurate but more calculation is involved, thus : 

Weight of sample 2 grams. 

Volume of sample after solution before addition to 
tube A = 50 C.C. 

Height of sample in tube A = 7.3 cm. 

Height of standard in B to balance = 8.4 cm. 

Content of standard per c.c. = . 00001 gram test sub- 
stance. 

When the colors of A and B are balanced the concen- 
trations of the contained solutions are inversely propor- 
tional to the volumes used, thus : 

T. S. per c.c. in A: T. S. per c.c. in B = Depth 

of liquid in B : Depth of liquid in A ; 

X:. 00001 = 8.4: 7.3. 

X = . 0000115 gram test substance per c.c. in sample 
solution. 
Then the total test substance in the sample is 

50 X .0000115 = .000575 gram. 

This is .0002875 gram per gram of sample and the 
sample is therefore .02875 percent test substance. 



CHAPTEE FOUR. 

DETERMINATION OF IRON. 

The determination of iron is one of the important ap- 
plications of colorimetry both because of the widespread 
distribution of the metal and because of the inadequacy 
of gravimetric or volumetric methods to measure small 
amounts of the substance, especially in the presence of 
the other third group metals. Lunge recommends that 
some colorimetric method be used for the determination 
of small amounts of iron in preference to other methods. 

There are seven methods for the colorimetric determi- 
nation of iron, details of which follow. 

Method A — Iron by Potassium Sulfocyanate. 

The sulfocyanate method for the determination of iron 
depends upon the oxidation of all iron present to the 
ferric condition after which the reaction between ferric 
iron and potassium sulfocyanate is used to produce a 
red color proportional to the amount of iron present. 
It is a very rapid method and is delicate enuf for the 
detection of one part of iron in fifty million parts of 
water, especially in the presence of free mineral acid.^ 
Use is therefore made of the method for the determina- 
tion of traces of iron in water supplies and of iron in 
small content iron solutions such as solutions of salts and 
minerals known or suspected to contain a small amount 
of iron. It also furnishes a rough and ready method 
for the approximation of the iron content of minerals 

1 0. N., 51, 259. 

31 



32 



OOLOEIMETBIO ANALYSIS. 



A O 






-I 



H ^ 



-i 



which contain considerable amounts of the metal, even 
up to 30 percent, when accuracy within one percent is 

required rather than the greater accuracy 
of the slow analyses. This latter determi- 
nation of course calls for the methods of 
accurate dilution given in Chapter Two. 

Fluorides, phosphates, arsenates, oxal- 
ates, citrates, tartrates and iodates inter- 
fere, also acetates to a lesser degree.* It 
is therefore necessary that these sub- 
stances be albsent from the solution. 
Since the presence of a small amount of 
mineral acid has been found to make the 
test more delicate the determination is 
Fig. 12. Straight carried out in a solution slightly acidified 

Graduated Tubes . ^ '^ 

with sulfuric acid. The amount of the 
iron ionized is dependent upon the concentration of the 
acid solution, so it is advisable to have the acid content 
of sample and standard uniform for the attainment of 
the correct results. 

The amount of the sample must be so chosen that 
the final solution contains between .000001 and .0002 
gram iron per c.c, of which solution ten c.c. is taken for 
the analysis. If the test is being made of a solid dis- 
solved in water the sample may be so chosen as to con- 
form to these limits, if the test is of a water for the iron 
content the water may be concentrated. In some special 
cases where great accuracy is not required it may be 
found possible to use twenty or even forty c.c. of a solu- 
tion somewhat more dilute than the limit as a sample. 
To this solution add ten c.c. of concentrated sulfuric 



2 J. A. C. S., iS9, 409. 



DETEEMINATION OP IBON. 33 

acid and oxidise with potassium permanganate until a 
faint permanent blush appears. Then run in about five 
c.c. of a solution of potassium sulfocyanate, 97 grams to 
the liter, and dilute to fifty c.c. with pure water. The 
color to be tested appears at once and ranges from a faint 
pink to a deep red according to the amount of iron 
present. The test may be made by duplication or by 
balancing but the dilution and series of standards meth- 
ods are not satisfactory in this test. The amount of 
ionization changes with the acid concentration of the 
solution and therefore the color would change with the 
addition of water independent of the amount of iron in 
the solution. A series of standards is not satisfactory 
as they fade out in light so soon. The addition of a 
small amount of potassium persulfate to a solution made 
up will keep the iron oxidised for some time longer if it 
is desired to preserve a sample or standard for several 
days. 

The standards to be used in this test are made up as 
follows, if the result is to be stated as Fe weigh out .1405 
gram of ferrous ammonium sulfate [FeS04(NH4)2S04, 
6H2O], if the result is to be stated as FcjOg the weight 
of salt taken is .1473 gram. The weights given above 
are dissolved in water with the addition of a little dilute 
sulfuric acid. Ten c.c. of concentrated sulfuric acid are 
then added to this and potassium permanaganate run 
in until the iron has all been oxidised to the ferric state. 
Twenty c.c. of the sulfocyanate solution are then added 
and the beaker carefully washed out into a liter volu- 
metric flask and the flask filled to the line. The addition 
of the sulfocyanate to the iron solution before washing 
it into the volumetric enables one to see that the beaker 
in which the oxidation was carried on is thoroughly 
washed out and all the iron transferred to the flask. 



34 COLOBIMETRIG ANALYSIS. 

This standard solution contains .00002 gram Fe or .00003 
gram FegOg according to which weight was used. The 
use of the appropriate weight thus allows for the figuring 
of results more readily. It is well to add one tenth of 
a gram of potassium persulfate to the standard before 
dilution so that the color may remain clear for a longer 
time. The standard should then be kept in the dark 
and renewed at least once a week if used often. 

This standard solution is to be added to thirty c.c. of 
water, and ten c.c. of concentrated sulfuric acid, meas- 
uring the amount of standard used accurately with a 
burette. Not only the color but also the volume of the 
standard must be identical with those of the sample 
before the end point of this method has been reached. 
The amount of iron in the sample is then measured by 
that necessary to make up the duplicate. If the test is 
to be made by the Campbell-Hurley apparatus or other 
type of balancing apparatus the standard mentioned 
above may be diluted to one tenth its present strength 
(ten c.c. diluted to one hundred) unless the color of the 
sample being tested is very deep. 

It is often desirable to make a test of filter papers for 
iron as impurity. Such tests may be made on the original 
filter paper qualitatively by working the paper to a pulp 
in concentrated hydrochloric or nitric acid but the quan- 
titative determination is made by dissolving the ash in 
acid, oxidation of the iron with permanganate, and test- 
ing this sample as a whole by the method already given. 

For the determination of iron in lead and similar sub- 
stances^ soluble in acid a sample of ten to thirty grams, 
dependent on the iron content, is dissolved in nitric acid, 
thirty c.c. of sulfuric acid added, the precipitate allowed 
to settle and filtered. The filter is well washed with warm 

1 J. Ind. Eng. Chem., 7, 1035, 1915. 



DETERMINATION OF IRON. 35 

water, small amounts at a time, and the filtrate rendered 
slightly alkaline by the addition of ammonia. This 
solution is boiled for a minute and the precipitate filtered 
out and well washed with warm water to remove traces of 
lead. Eedissolve this precipitate which contains the iron 
as the hydroxide in dilute hydrochloric acid and test by 
the method given. If the amount of iron is too great 
to test as a single sample dilute to 100 c.c. and use ten or 
twenty c.c. of this solution for testing. For the oxida- 
tion of the iron in this latter test it is possible to add a 
slight excess of hydrogen peroxide to the iron solution 
to be oxidised and thus save the trouble of assuring oxida- 
tion with permanganate. 

Method B. Iron as the CnLORroE in Concentrated 

Acid. 

The yellow color of ferric chloride has been adapted to 
the analysis of samples containing iron for their iron 
content. The sample^ is dissolved in hydrochloric acid or, 
if that is impossible, in aqua regia or nitric acid and the 
nitric acid expelled by heating until all nitric fumes are 
gone. The amount of iron present must be less than .0001 
gram per c.c. and less than half this concentration is 
better. The color concentration of ferric chloride reaches 
its maximum when there is present 28 percent hydro- 
chloric acid, so it is only necessary to make our solution 
more concentrated than that to have the uniformity of 
color desired. The presence of free chlorine will not 
interefere but nitric acid or its oxides will and for that 
reason it is necessary to remove every trace of nitrate if 
nitric acid was used in the dissolving of the sample. If 
copper is present it must be removed by hydrogen 

1 Z. Anorg. Chem., 86, 341. 



36 OOLOBIMETEIG ANALYSIS. 

sulfid as its chloride, when water of crystallization is not 
present, has the same color as iron. Manganese does not 
interfere and the method is therefore applicable to the 
analysis of iron in the presence of that substance and to 
the testing of manganese precipitates for iron. In such 
a dilute solution any small amount of cobalt or nickle 
will not produce sufficient color to cover up the results. 
If other substances with a red or yellow color are present 
the metals which will precipitate with ammonia may be 
separated by the addition of a slight excess of that re- 
agent. The precipitate of hydroxides may then be dis- 
solved in concentrated hydrochloric acid and tested as 
the sample. 

The iron standard is best made up by dissolving one 
gram pure iron wire or such a weight of an impure iron 
that it is known to contain exactly one gram of pure 
metal. This is dissolved in nitric acid, hydrochloric acid 
added, and taken to dryness three times. In that way all 
trace of nitrate will be removed. The residue after these 
ignitions is made up to one hundred c.c. with concen- 
trated hydrochloric acid. Ten c.c. of this solution is 
made up to one liter with concentrated hydrochloric acid 
and furnishes the standard for comparison. This stand- 
ard contains .0001 gram Fe per c.c. The results may be 
obtained either by the dilution method, using concen- 
trated hydrochloric acid for the dilution, or by the bal- 
ancing method. On account of the cost of hydrochloric 
acid if a considerable number of tests are being, run it is 
best to adopt the latter method. 

Method C. Iron by Potassium Ferrocyanide. 

Lunge^ recommends that for some cases the blue color 
produced by the action on ferric iron of potassium ferro- 

1 Technical Methods of Chemical Analysis, Lunge. 



DETERMINATION OF IBON. 37 

cyanide is the most convenient method of analysis for 
iron. This method may be used to determine the ferric 
iron in the presence of ferrous or it may be used to deter- 
mine the entire iron content. In the latter case the 
solution must first be oxidised with permanganate to 
bring all iron present to the ferric condition. Owing to 
the intensity of the color developed this method is only 
applicable to the determination of traces of iron as, for 
example, in water. 

The solution to be tested should be suitably concen- 
trated or diluted so that the iron content will approxi- 
mate .000002 gram iron per c.c. The standards used are 
the same as those for method A. Ferrous ammonium 
sulfate [FeSO^CNHJ^SO^, GH^O], .1473 gram if the 
result is to be determined as FcgOg or .1405 gram if the 
result is to be determined as Fe. In either case the salt 
is dissolved in water, five c.c. concentrated sulfuric acid 
added, and the solution oxidised to the ferric condition by 
permanganate. The solution is then diluted to one liter. 
Ten c.c. of this are pipetted out, ten c.c. potassium ferro- 
cyanide [K4Fe(CN)g] added and the whole diluted to 
100 c.c. This standard now shows the characteristic color 
reaction and contains .000002 gram Fe or .000003 gram 
1^^^20s according to the weight taken. 

For the test five c.c. of potassium ferrocyanide are 
added to fifty c.c. of the solution to be tested and this 
sample then compared with the standard mentioned above 
by dilution or by balancing. These methods are more 
applicable to the particular conditions of this experiment 
than are the other two. 

Method D. Iron by Salicyl.ic Acid. 

Ferric iron in reaction with salicylic aoid produces an 
amethyst coloration, but ferrous iron produces no color. 



38 COLOBIMETEIG ANALYSIS. 

This reaction^ is therefore applicable to the determination 
of ferric iron in the presence of ferrous when the sample 
can be so chosen as to come within the limits of test by 
this method. The color will be too intense for satis- 
factory work if the sample contains more than .0002 gram 
of iron, the color may be detected if as little as .00001 
gram of ferric iron is present. The method is applicable 
to the analysis of salts for impurity except in the presence 
of the following, which interfere with the test: phos- 
phates, fluorides, thiosulphates, sulphites and bisulphites. 
The presence of free mineral acid also interferes with the 
test and it is therefore necessary that such be absent, pre- 
ferably by the solution of the original sample in water, 
if that is not possible then by the neutralization of the 
acid by ammonium hydroxide after solution is complete. 

The sample, if possible, is dissolved in twenty c.c. of 
pure water and any turbidity, the cause of which is known 
to contain no iron, is filtered out. If it is necessary to 
dissolve the sample in acid as dilute hydrochloric acid as 
possible is used, and the free acid after solution is com- 
plete, is neutralized with ammonia, taking care that an 
excess of the ammonia is not added as that would pre- 
cipitate the iron. There should not be more than twenty 
c.c. of the solution after neutralization. If the sample 
to be tested contains organic matter it may be necessary 
to heat to dryness with nitric acid and ignite to remove 
such organic matter. If all the iron content is desired 
the sample is next oxidised with permanganate so that all 
iron is in the ferric condition. For this it may be nec- 
essary to slightly acidify the solution with sulfuric acid 
and later bring it back to neutrality after the oxidation. 
If only the ferric iron is to be determined this step of 
oxidation is omitted. Five c.c. of salicylic acid solution 

1 Adapted from Method of W. S. Allen as given by W. W. Scott. 



DETEEMINATION OF IRON. 39 

ai'e now added and comparison made. Because of 
the fading of the color by exposure to light^ it is not 
advisable to make up a series of standards but rather the 
test should be made by dilution or balancing. The use 
of the duplication method is also inadvisable since for the 
carrying out of it considerable time is required and 
fading may occur during that time. 

A convenient standard is made by dissolving in water 
.0864 gram pure ferric ammonium alum [Fe2( 804)3 
(NH4)2S04, 24H2O] with the addition of two c.c. of con- 
centrated sulfuric acid and diluting accurately to one liter. 
This solution contains .00001 gram F'e per c.c. or .000014 
gram FegO,. For the use of this standard twenty c.c. 
or other suitable quantity may be accurately measured 
out from a burette and five c.c. salicylic acid added just 
before it is used. 

Method E. Iron as the Sulfid. 

The determination of iron as the sulfid in alkaline solu- 
tion is particularly applicable to the testing of drinking 
water but may be adapted to any test for ferrous iron in 
the presence of ferric, or to the analysis for the total iron 
content of a solution by the reduction of the iron content 
to the ferrous condition. If the test is to be made of 
water it is best made immediately after taking the 
sample. If the test^ is not to be made at once it is nec- 
essary to add a few drops of hydrochloric acid and ten 
c.c. hydrogen sulfid water to a sample of not over one 
half liter and cork the bottle tightly at once. By the 
use of these preservatives the oxidation of ferrous to 
ferric iron may be prevented. The sample should have 
an iron content betw^een .0015 and .0003 gram per liter. 

1 Sclireiner and Ferris, J. A. C. S., 26 y 961. 

2 Technical Methods of Chemical Analysis, Lunge. 



40 COLOEIMETEIC ANALYSIS. 

If the amount of ferrous iron exceeds this the solution 
may be diluted, if the iron content is less than the pre- 
scribed limits the results may be attained by the use of a 
larger volume of solution than specified for the test. 

For the determination of the entire iron content of a 
sample of water or of some solid substance dissolved in 
water or acid, the solution is evaporated to dryness with 
a small amount of hydrochloric acid and potassium 
chlorate. The residue is dissolved in warm water with 
the addition of a few c.c. of hydrochloric acid and five 
c.c. of hydrogen sulfide water. This solution is first fil- 
tered to free from sulfur clots and then diluted to 100 cc. 
This is then treated the same as the entire solution would 
have been if analysis was being made for ferrous iron 
only. 

A standard is made up by the solution in a liter of 
water saturated with hydrogen sulfide of .7022 gram of 
ferrous ammonium sulfate [FeS04(NH4)2S04, GH.O] 
with the addition of one c.c. of sulfuric acid. This solu- 
tion must be kept well corked and, if the gas gets free 
to such an extent that the bottle no longer smells of it, 
the liquid must be recharged. This standard contains 
.0001 gram ferrous iron per c.c. 

The comparisons are made in plain glass cylinders by 
the method of duplication. The sample is 100 c.c. of the 
solution to be tested, a blank is run of 95 cc. of distilled 
water. To each, sample and blank, is added five c.c. of 
an aqueous solution of hydrogen sulfid and two drops of 
ammonia. The standard solution is then added to the 
blank drop by drop until the colors of the two solutions 
are approximately equal. The color of the standard will 
show more of a brown hue than that shown by the sample 
and it is therefore necessary to decolorize the standard 
with a few drops of hydrochloric acid and then renew the 



DETEBMINATION OP lEON. 41 

color by the addition of a few drops of ammonia. The 
color which comes back will be the same sort of a brown 
as that shown by the sample. More standard can then be 
added to the blank if necessary in order to bring it to the 
point where the colors of the sample and standard are 
identical. 

If extreme accuracy is desired the standard may again 
be decolorized and should come back by the addition of a 
few drops of ammonia to identically the same color as 
the sample. 

Method F. Iron by Acetylacetone. 

The principal objection to the tests for ferric iron with 
salicylic acid and potassium sulfocyanate already given 
is that they fade in the bright light necessary for the com- 
parison of colors. This test has as its merit the nonfad- 
ing of the results. The color produced^ varies with the 
amount of free acid present and therefore the solution 
should not contain any appreciable amount of free acid. 
The solution to be tested, whether a sample of water or 
some substance in solution, is evaporated to dryness on 
the water bath and a few drops of sulfuric and nitppc 
acids added to the residue to remove any organic matter 
present. If much iron is present the dilution will be 
carried to a considerable degree and the amount of acid 
used will not matter but if only a very slight amount of 
iron is present it may be necessary to heat and drive off 
most of the acid before adding water, in order that the 
acid may not be too strong. The range of greatest accu- 
racy is between .0006 gram and .00005 gram ; the sample 
should therefore be so diluted that fifty c.c. will contain 
an amount of iron somewhere between these limits. The 

1 Pulsif er, J. A. C. S., fS8, 967. 



42 COLORIMETRIC ANALYSIS 

• 

dilution should be to 100 c.c, 500 cc, or 1000 cc, in order 
that the results may then be figured by aliquot parts. 

The sample used is fifty c.c. of the solution above, to 
which there are added two cc. of acetylacetone solution. 
The best method of completing the analysis is by the 
addition of a standard to a blank of 45 c.c. of distilled 
water and two c.c. of acetylacetone solution. The stand- 
ard should consist of .1405 gram ferrous ammonium 
sulfate [FeSO^CNHJ^SO^, GH^O] dissolved in water, 
oxidised with potassium permanganate to the ferric con- 
dition and made up to one liter. This standard contains 
.00002 gram of iron per c.c. The method of balancing 
is not applicable to this test aS the color produced does 
not conform to the necessary laws. If a considerable 
number of tests are to be made it may be well to make up 
a series of standards from the standard solution already 
mentioned, using two to thirty cc. of the solution at two 
c.c. intervals. For the determination of a few tests or 
for experimentation as to its applicability to a given use 
the method of dilution is easiest to arrange. The stand- 
ard previously mentioned is further diluted by the dilu- 
tion of ten c.c. to fifty with the addition of two c.c of 
acetylacetone. This new standard for the dilution method 
contains .000004 gram iron per c.c 

Method G. Iron by Dimethylglyoxime. 

The red coloration produced by this reaction is very 
distinctive and permits of the estimation of the iron 
within an accuracy of one half percent of the metal pres- 
ent in the solution. The iron present^ in the solution for 
this test should be between .00001 and .00006 gram per cc 
The presence of large amounts of aluminum and zinc 
must be avoided. 

1 Z. Anorg. Chem., 89 ^ 401. --^ _- , 



DETERMINATION OF IBON. 43 

To the sample of fifty c.c. of the solution to be 
tested there is added one gram of hydrazine sulfate 
fN2H4H2S04] and five c.c. of a saturated solution of 
dimethylglyoxime in ethyl alcohol and the solution heated 
to boiling. Add ten c.c. of a 25 percent ammonia solution 
and continue the boiling for one half minute. Cool the 
solution rapidly and dilute to 100 c.c. for comparison. 
The comparison is carried out by any of the usual 
methods. 

As a standard the one prepared for methods A and C is 
satisfactory. This consists of .1473 gram of ferrous am- 
monium sulfate dissolved in water, five c.c. concentrated 
sulfuric acid added, oxidised with permanganate and 
made up to one liter. This solution contains .00003 grams 
FCaOg per c.c. For the method of duplication this stand- 
ard is added to seventy-five c.c. of pure water in a tube 
similar to that containing the sample, after the dimethyl- 
glyoxime has been added to the blank in the same propor- 
tions as to the sample. The amount of standard neces- 
sary to bring the blank to the same color as the sample 
when it is of the same volume is the measure of the iron 
present in the sample. For a series of standards varying 
amounts of the standard are used and, after the addition 
of the same reagents as to the sample, each is made up to 
100 c.c. For the method of dilution ten c.c. of the above 
standard are made up to 100 c.c. with the addition of the 
above reagents and tested as usual. As a standard for 
the balancing method it will be necessary to dilute a 
greater quantity of the standard, 25 c.c. diluted to 250 c.c. 
will be ample. The latter two standards after dilution 
contain .000003 gram FcgOg per c.c. 



CHAPTER FIVE. 

DETERMINATION OF COPPER. 

Altho not so widely distributed a metal as iron a greai: 
deal of work has been done on the determination of 
copper to afford methods both rapid and accurate taking 
into account the many interferences apt to occur. The 
seven methods now in use follow. 

Method A. Copper by Ammonia. 

The copper salt of nitric or sulfuric acid dissolves in 
ammonia giving an intense blue color which is propor- 
tional in intensity to the amount of copper present in the 
solution.^ This method is particularly applicable to the 
determination of traces of copper in ores and slags solu- 
ble in nitric acid but is also useful in the analysis of 
salts containing a trace of copper as impurity. Nickel, 
cobalt, and manganese^ interfere with the test so that in 
their presence the solution requires special treatment. 
In the presence of this interference the metals of the 
first two groups are precipitated by hydrogen sulfid in 
acid solution and filtered out. This precipitate of sulfid? 
is washed and redissolved in nitric acid and this solution 
treated as tho it were the original. In the absence of 
fourth group metals the precipitation mentioned above 
is omitted. 

The sample is chosen of from one tenth to ten grams 
according to the approximate copper content, the sample 
being so chosen as to include between .001 and .01 gram 

1 Sutton, Volumetric Anal. 

2 Lunge, Tech. Meth. of Chem. Anal. 

44 ! 



DETERMINATION OP COPPER. 45 

of copper. The sample is dissolved in nitric acid or, if 
it dissolves readily in water, the solution is strongly 
acidified with nitric acid, and placed in a bottle or com- 
parison cylinder. Twenty c.c. of ammonia are then 
added and the solution diluted to 100 c.c. If the solution 
should be so strong that it was not neutralized by the 
twenty c.c. of ammonia mentioned a greater amount 
should be used until the ammonia has redissolved the 
precipitate first formed on its addition. Comparison is 
now made by one of the methods given in Chapter Two. 

A standard solution is prepared by dissolving in nitric 
acid, .5 gram pure copper and diluting the solution to 
500 c.c. If it is desired to prepare a series of standards 
these are made up by using from one to ten c.c. of this 
solution. The standards should each have twenty c.c. of 
ammonia added and be made up to 100 c.c. in bottles 
similar to the one in which the comparison is made. For 
the method of duplication seventy c.c. of water and 
twenty c.c. of ammonia furnishes the blank to which is 
added the standard specified above. This standard con- 
tains .001 gram copper per c.c. If a series of standards 
ai-e used it will be necessary to renew them onoe a month 
as, altho they do not fade in the light, the ammonia pres- 
ent forms a flocculent precipitate with the glass after a 
time. A standard for the dilution or balancing methods 
is made up by diluting ten c.c. of the solution above to 
one liter with fiftv c.c. ammonia and 940 c.c. water. This 
standard contains .00001 gram copper per c.c. 

Experiment has shown that the greatest delicacy of 
test is given when the solution contains .00001274 gram 
copper per c.c. and that the addition of .000000016 gram 
copper can be detected at that concentration. It is with 
this experiment that the greatest amount of study as to 
the delicacy of given reactions has been made. 



46 OOLOBIMETRIC ANALYSIS. 

In some cases it may be found desirable to use the sul- 
fate rather than the nitrate for the foregoing procedure. 
In case sulfuric acid is used for acidifying the sample 
the standard solution used must also have been acidified 
with sulfuric acid since, tho both the nitrate and sulfate 
of copper dissolve in ammonia with a deep blue color, the 
colors of the two for the same amount of copper are not 
the same. 

Method B. Copper as the Chloride in Concentrated 

Acid. 

If no other metals whose chlorides are also deeply 
colored are present the concentration of a copper solution 
may be told, if over thirty percent hydrochloric acid, by 
the intensity of the yellow color produced.^ The actual 
intensity of the color reaches its maximum at an acid 
concentration of 28 percent and after that remains con- 
stant but in order to assure the full depth of color the 
concentrated acid is used. Free chlorine does not inter- 
fere but nitric acid or the presence of any of the nitrous 
oxides does and these must therefore be removed. Cobalt 
or nickel will have very little color in the concentrated 
acid solution unless the amount of them is very great. 
Manganese does not interfere. If nickel, cobalt or iron 
are present in quantities great enuf to color the solution 
the sulfid may be precipitated in acid solution and redis- 
solved, thus getting rid of the disturbing factors quoted. 
Second group sulfids have no effect except those of the 
noble metals. Minute amounts of organic matter may 
produce a yellow color and if present must be burned out 
before dissolving in acid. Iron has a yellow coloration 
which is to that of copper as nine to five. 

1 Huttner, Z. Anorg. Ohem., 86^ 341. 



DETEBMINATION OF COPPER. 47 

For the test a sample is selected such that it contains 
about .1 gram or .01 gram copper. This sample is dis- 
solved in hydrochloric acid directly if possible but such 
a solution is not usually possible. In case the sample 
does not dissolve in hydrochloric acid readily it is dis- 
solved in nitric acid or aqua regia and hydrochloric acid 
added to the solution. This is evaporated to dryness, 
more hydrochloric acid added and the process repeated. 
After repeating the drying a third time the dilution is 
made with concentrated hydrochloric acid. If the sample 
was of one tenth gram of copper the sample is diluted to 
one liter, if the sample was of .01 gram copper it is diluted 
to 100 C.C. A standard is similarly prepared by the solu- 
tion of .5 gram pure copper in nitric acid and subsequent 
expulsion of the nitric acid by the use of hydrochloric 
acid, as directed for the sample. This standard is finally 
taken up with dilute hydrochloric acid and diluted to 500 
c.c. with concentrated hydrochloric acid. This standard 
contains .001 gram copper per c.c. Because of the con- 
centrated acid used it is not advisable to try to keep per- 
manent standards. The test may be made by dilution or 
balancing, if by dilution concentrated hydrochloric acid 
is used for the dilution. 

This method is as accurate as the ammonia method and 
is much easier to apply in some cases. Fifty c.c. of the 
diluted sample is usually convenient for analysis by the 
dilution or balancing methods. 

Method C. Copper by Salicylic Acid. 

The color produced by copper with salicylic acid may 
be used for the quantitative determination of the copper 
present altho the color produced is not permanent. The 
presence^ of free mineral acids, citric and tartaric acids, 

1 Zeitsch. Untersuch. Nahr. Genussm., S2y 727. 



48 COLORIMETBIO ANALYSIS. 

and more than a trace of iron, prevent the use of this 
test. An amount as small as .00001 gram of copper may 
be detected and estimated by this method. 

The comparisons are always made against a series of 
temporary standards made up at the time of the experi- 
ment. A standard copper solution is made up by solu- 
tion of .3928 gram of copper sulfate (CUSO4, SHgO) in 
water and making up to one liter. Of this standard 
amounts varying from one tenth to one c.c. are placed in 
plain test tubes and the sample in solution in about ten 
c.c. of water is placed in a similar plain test tube. The 
standards are each diluted to the volume of the sample, 
ten c.c, or if convenient a five c.c. sample may be used and 
dilution made to five c.c. There is then added to each of 
the tubes five drops of a two percent solution of potas- 
sium nitrate and five drops of ten percent acetic acid, 
then three c.c. of a one half percent solution of salicylic 
acid in dilute alcohol. The tubes are next heated to boil- 
ing in a water bath for three quarters of an hour. The 
reading of the sample is taken to be that of the standard 
to which it most nearly conforms. The standard solution 
for making the series of standards contains .0001 gram 
copper per c.c. 

Method D. Copper by Potassium Ferrocyanide. 

The amount of copper^ in dilute solutions may be deter- 
mined by the intensity of the purple-brown color produced 
in reaction with potassium ferrocyanide. The test is 
rendered more delicate bv the addition of ammonium 
nitrate, ammonium chloride or potassium nitrate to the 
solution. In the presence of these salts the test will show 
one part in 2,500,000. Lead forms a white precipitate 

1 Sutton, Volumetric Analysis. 



DETERMINATION OF COPPER. 49 

with the ferrocyanide but unless the lead is in greater 
excess than it might be expected to be in such a solution 
the small amount of precipitate will be negligible. The 
solution in which the test is to be made must be exactly 
neutral; if it is acid the color produced will turn to an 
earthy brown, if alkaline the color will be partly dis- 
solved and removed by the alkali. In the presence of 
large amounts of iron the solution may be rendered 
alkaline with ammonia and the iron filtered out for use 
in the determination of iron if one is to be made.^ If the 
iron hydroxide is of large volume it may be well to dis- 
solve in acid and reprecipitate to recover copper carried 
down with the precipitate the first time. 

The sample is so chosen as to contain between .005 and 
.00005 gram copper.^ If the sample is a solid, difficulty 
soluble, it is dissolved in nitric acid and the solution 
boiled until the excess of acid has been expelled. P^ive c.c. 
of sulfuric acid are then added and the solution boiled 
again. Filter, wash the precipitate with warm water 
and take the combined filtrate and wash water for the 
test. The sample, if acid, is rendered slightly alkaline 
with ammonia and the solution then boiled until the ex- 
cess has been removed. Five drops of a 1 : 35 solution of 
potassium ferrocyanide are then added to the sample and 
the test made by duplication or dilution. 

The standard is made up by the solution of .3928 gram 
of pure copper sulfate (CUSO4, 5H2O) in water and 
dilution to one liter. This standard contains .0001 gram 
copper per c.c. For use with the method or duplication 
the sample is diluted to 50 or 100 c.c. If the latter is 
the case it may be advisable to divide the solution into 
two parts and use one for testing. The blank is prepared 

1 Jour. Ind. Eng. Chem., 7, 1035. 

2 Mining Sci. Press, 114, 624. 



50 COLORIMETRIO ANALYSIS. 

by adding five drops of ferrocyanide solution to thirty- 
five C.C. of pure water and five c.c. of ammonium chloride 
solution in a tube similar to that containing the sample. 
The standard mentioned above is then run in until the 
color of the sample is duplicated and the volumes cau- 
tiously equalized as described for this method of deter- 
mination. The color produced is not satisfactory for the 
balancing method and fades too soon to make the use of 
a series of standards practicable. For the method of 
dilution ten c.c. of the standard are diluted with twenty- 
five C.C. of water, five c.c. of ammonium chloride and five 
drops of ferrocyanide and made up to fifty c.c. The 
standard if darker in color is now diluted until the colors 
match, and the results figured. 

In some cases for the purpose of concentrating the 
solution the copper may be precipitated with aluminum 
foil and the precipitate gathered and dissolved in nitric 
acid after which the procedure is carried on as before.^ 

Method E. Copper as the Sulfid. 

A trace of copper in solution may be determined by 
the formation of the colloidal sulfid and the estimation 
of the color of the resultant brown solution by the use of 
a standard solution similarly prepared. The amount of 
copper must not be more than .00025 gram and less than 
half that amount is better.^ The method is not applicable 
in the presence of lead but the ferrocyanide method can 
be used in that case and the lead determined from the 
same solution without separating the copper. The ad- 
dition of ammonium chloride to the solution doubles the 
delicacy. 

To the sample in solution are added a few c.c. sodium 

1 Jour. Ind. Eng. Chem., 7, 1035. 

2 Sutton, Volumetric Analysis. 



DETERMINATION OF COPPEB. 51 

acetate, five c.c. saturated ammonium chloride solution 
and three c.c. concentrated nitric acid. The addition of 
two C.C. of hydrogen sulfid water will then produce the 
brown coloration. If the precipitate shows a tendency to 
gather in clots and settle another sample niust be pre- 
pared and to this, in addition to the reagents mentioned, 
add five c.c. of a clear syrup, half sugar and half water. 
This will effectually prevent the settling of the precipitate. 
The standard copper solution consists of .3928 gram 
copper sulfate (CUSO4, SHgO) dissolved in a liter of 
water. This contains .0001 gram copper per c.c. The 
tests are made by any method except that of a series of 
standards. The preparation of such a series even of the 
artificial standards is not satisfactory. For duplication the 
sample is diluted to fifty c.c, 25 c.c. of water are treated 
with the reagents for the sample and the above standard 
added a drop at a time. For dilution or balancing ten 
c.c. of the standard above are treated with the reagents 
specified for the sample with the exception that the 
amount of hydrogen sulfid water is increased to six c.c. 
This is then diluted to 100 c.c. and every c.c. contains 
.00001 gram of copper. 

Method F. Copper by Potassium Ethyl Xanthate. 

Small amounts of copper in solution react with potas- 
sium ethyl xanthate to produce a yellow color suitable for 
colorimetric examination.^ If more than .001 gram cop- 
per is present in the sample the solution will be too strong 
and instead of the desired color there will be formed a 
bright yellow precipitate of copper xanthate. In that 
case the original sample chosen must be smaller and the 
experiment repeated. The presence of small amounts of 

1 Scott, Standard Methods of Chemical Analysis, p. 165. 



52 COLORIMETBIO ANALYSIS. 

iron, lead, nickel, cobalt, zinc or manganese does not 
interfere. The test is one well adapted to the testing of 
salts for copper present as impurity. 

The sample graduated in size according to the copper 
content is dissolved in water and diluted to 90 c.c. Ten 
c.c. of a solution of potassium ethyl xanthate, one gram to 
the liter, are then added. The yellow color appears at 
once. It is not advisable to make up a series of standards 
for the final test as after a time the copper shows a tend- 
ency to precipitate out. A standard copper solution is 
made up by dissolving .3928 gram of pure copper sulfate 
(CUSO4, 5H2O) in water and diluting to one liter. Ten 
c.c. of this solution are then diluted to 100 c.c. so that it 
now contains .00001 gram copper per c.c. and serves for 
the examination by duplication. The blank contains 
sixty c.c. water and ten c.c. potassium ethyl xanthate 
solution into which the above standard is run until the 
colors match after which the colors and volumes are 
cautiously balanced at the same time. For determining 
the results by dilution or balancing another amount of 
standard must be diluted. Ten c.c. of the standard are 
diluted to five hundred c.c. with the addition of ten c.c. 
of the reagent to produce the color. This solution then 
contains .000002 gram copper per c.c. and is suitable for 
the determinations by these other two methods mentioned. 

This method is applicable to the testing of substances 
insoluble in water by treating them with nitric acid to 
render them soluble. If that method is used the nitric 
acid must all be removed before completing the test. 
This is accomplished by adding hydrochloric acid and 
evaporating to dryness. The residue is then extracted 
with water and the copper will be present in the solution. 
If organic matter is to be removed by this ignition a few 
c.c. of sodium hydroxide and a similar amount of sodium 



DETEBMINATION OF COPPBB. 53 

or potassium nitrate are added to the solution before 
evaporation. The residue is then to be ignited until no 
more fumes are given off. Water must now be added, 
then hydrochloric acid and the residue again taken to dry- 
ness to remove all trace of nitrate. Since the examina- 
tion is to be made in neutral or slightly acid solution the 
presence of any considerable amount of acid in the test 
solution must be avoided. 

Method G. Copper as the BROMroE. 

This method^ is suitable for the determination of small 
amounts of copper present as sulfates, nitrates or 
chlorides. If the copper to be tested is present as some 
complex compound it will first be necessary to treat with 
concentrated acid to decompose the complex compound. 
The solution for the test is treated with a reagent made 
up as follows, 100 grams of potassium bromide are added 
to 150 C.C. of boiled water and the solution well shaken. 
This is then transferred to a 200 c.c. volumetric and 
diluted to the line with warm water. After cooling again 
bring up to the mark with cold water. 

Five c.c. of the solution to be tested are used. If the 
sample requires more than that amount of acid or water 
to dissolve it, the above amount is taken from the total 
and the result figured by aliquot parts. For use twenty 
c.c. of the reagent above are precipitated by the careful 
addition with cooling of ten c.c. of concentrated sulfuric 
acid and the resultant potassium sulfate filtered off. Five 
c.c. of this solution are then added to the sample with two 
c.c. of concentrated sulfuric acid. For comparison sev- 
eral standards are made up by the use of a solution of 
.3928 gram of pure copper sulfate (CuSO^, SHgO) in a 
liter of water. The standard contains .0001 gram copper 

1 Bull. Soc. Phar. Bordeaux, Aug.-Dec, 1915. 



54 OOLOEIMETBIC ANALYSIS. 

per C.C. These standards are reasonably permanent and 
so may be kept for some time. The ease with which the 
production of new standards may be carried out renders 
their use, after there is any doubt of their accuracy, un- 
necessary. 

The reagent used is unstable and must be kept in 
brown glass and in a dark place. Renewal not less than 
twice a month is advisable, making up smaller amounts 
at a time if desirable. 



CHAPTER SIX. 

CARBON IN STEEL. 

This is one of the best worked out and most used 
methods of colorimetric analysis, finding use as ordinary 
procedure for rapid determination of the carbon content 
of steels in most works laboratories. The basic principle 
assumed for the test is that a sample of steel when dis- 
solved in nitric acid shows a brown color which is pro- 
portional in intensity to the amount of carbon present 
in the sample. The errors that enter into the determi- 
nation of carbon by this method are these ; only the car- 
bon present in the form of carbide is transformed to this 
brown color, the composition of the brown coloration is 
not constant, presence of many impurities will cause 
colors which interfere. The carbon present in the steel as 
graphite, as it is apt to occur in high carbon steels, is not 
acted on by acids at all. In that case the flakes of carbon 
will be visible in the solution so that one may know that 
there is graphitic carbon present and therefore carry 
out a determination by the combustion method. It is 
sometimes assumed that in steels treated by the same 
identical process the proportion of graphitic carbon to 
carbide carbon will be the same and comparison therefore 
made, even of steels which do contain graphitic carbon. 
These results are not as accurate as those obtained by the 
usual colorimetric tests on steel. 

Manganese, if present, lowers the apparent carbon con- 
tent but may be disregarded if the amount is less than 
one percent. Nickel has a similar but much greater 
eflfect. If much nickel is present a green color is pro- 

55 



56 OOLORIMETRIO ANALYSIS. 

duced which is difficult of comparison. Over one percent 
of silica also produces a green color which is difficult of 
comparison. In cases where the presence of these sub- 
stances is known, a standard should be taken which has 
the same content of the interfering substance as that in 
the sample. Comparison may then usually be made. 
Copper, cobalt and chromium also interfere. 

From the foregoing it follows that the standard used 
should be as nearly as possible of the same composition as 
the sample. The following conditions should be met,^ 

1. Sample and standard should be made by same 
process. 

2. Sample and standard should have same physical 
condition so far as this can be secured by mechanical 
means. 

3. Sample and standard should not differ greatly in 
percentage of carbon. 

4. Solutions of standard and sample should be made 
at the same time under the same identical conditions. 

5. The standard used must be one whose composition 
is accurately known. 

The tints resulting from different forms of steel are 
different, and cannot therefore be matched against each 
other, so as specified above, Bessemer steel must be tested 
against Bessemer steel and open hearth steel against open 
hearth steel. The results obtained are more accurate for 
mild steels than for hard steels. 

Two methods are commonly used in steel works lab- 
oratories, the dilution method and the method by a series 
of standards. In the latter case the standards are arti- 
ficially prepared from the cobalt, iron and copper solu- 
tions described in Chapter Two. In the operation of the 
dilution method the samples and standards are weighed 

1 Steel Works Analysis, Arnold and Ibbotson, p. 50. 



CARBON m STEEL. 



57 



rf^ 



l: 






out alternately into test tubes, using .2 gram of each. To 
the tubes there is then added somewhat more than one 
c.c. of nitric acid (1:1) for every tenth of a percent of 
carbon supposed to be present in the steel. The color 
produced will be darker than it should be if too little of 
the reagent is used but an excess of acid may be added 
without causing any considerable change in the color. 
The test tubes are then placed in a special rack in the 
water bath. If the carbon content of the sample is under 
one percent it should go into solution in less than half 
an hour. A paraffine bath at 120 to 150 degrees may be 
used to hasten solution. Johnson^ sug- 
gests a sand bath at 190 degrees and 
recommends that the tubes be grouped in 
bunches of six and each bunch covered 
with a beaker to prevent too rapid evap- 
oration of the acid. If a temperature 
greater than 100 degrees is to be used 
it is advisable to allow for it by the addi- 
tion of excess acid. Tubes should never 
be immersed in the heated medium below 
the level of the top of the liquid as a 
deposit of iron oxide will form. It is 
always best to add a hollow glass bead to 
each test tube after putting in the acid. 
This will float on the surface of the liquid and prevent the 
formation of a film of iron oxide. 

For convenience in keeping the samples separated it 
is well to have a number scratched on the outside of each 
test tube ordinarily used for this work. The number of 
the test tube used for the sample may then be recorded 
beside the weight when samples are weighed out. This 

1 Chemical Analysis of Special Steels, Steel-making Alloys and 
Graphites, Johnson. 



U» 



1 



Fig. 13. Bent 
Graduated Tubes 



58 



COLORIMETRIO ANALYSIS. 



is somewhat more satisfactory than gummed labels as 
the later are apt to be steamed off. The samples are 
removed from the heating medium as soon as it is seen 
that everything has been dissolved, and cooled by plung- 
ing in cold water. If solution is made by heating on a 

water bath no harm is done by allow- 
ing the sample to stay in the bath two 
or three minutes after the solid matter 
is all dissolved. 

The solution now shows a brown 
coloration w^hich is diluted by the 
addition of two volumes of water to 
the cooled solution. This coloration is 
considered to be due to some variable 
organic compound which at first is to 
be seen as a precipitate in the solution 
and is dissolved by heating.^ Another 
experimenter finds that as much as 
.03 percent of the carbon may be pres- 
ent as CO and COg and therefore not 
show in the results. 

If the sample is to be tested by dilution this is usually 
carried out by use of the camera described under appa- 
ratus. Since the weights of sample and standard used 
are identical the results are easily figured. The solu- 
tions^ must be cooled below fifty degrees for this compari- 
son, otherwise the iron will show color enuf to hide the 
results. The use of ferric nitrate for dilution is some- 
times advised since the solutions are of that salt. Scott 
and Johnson recommend the use of tubes bent at the top, 
for the comparison, so that the w^ater can be added to the 
sample or standard to be diluted and mixed by a gentle 

1 Levy, Analyst, 57, 392. 

2 Ann. Chim. Anal., fS2y 193 and 225. 




Fig. 14. Color 
Camera 



CAEBON IN STEEL. 59 

rocking motion without danger of spilling. The manu- 
facturers furnish both straight and bent graduated tubes 
for this purpose. The sample is carefully washed from 
the original test tube into one of these tubes and the 
standard into another and comparison made. 

Experiments with the balancing instruments seem to 
show that because of the nonpermanenoe of the color ob- 
tained by this method the use of that type of instrument 
is not advisable. The duplication method is of course 
impossible. Standards and samples should be protected 
from light at all times as far as possible, as they fade with 
exposure. Phosphoric acid is sometimes added to the 
sample before dilution. 

For convenience in dilution several little helps have 
been worked out. The standard may be diluted to a 
volume depending upon its carbon content, such as 6 c.c. 
for .6 percent carbon and the sample when diluted to an 
equal color will then give its carbon content by direct 
reading from the scale. When this is not done it is 
common to find an operator adding water, two tenths of 
a c.c. at a time, until the solution being diluted either 
equals the one which was lighter or has become lighter 
than the other solution. If the two colors are identical 
of course that reading is taken but if the solution has 
now been diluted to less than the color of the other the 
reading is taken as one tenth c.c. less than that shown by 
the scale, or what would have been the reading if the last 
addition had been of one tenth instead of two tenths c.c. 

Samples for comparison with a series of standards are 
similarly prepared and washed into containers the same 
as those containing the standards, 



CHAPTER SEVEN. 

LEAD, BISMUTH AND ARSENIC. 

The presence of these substances is important partic- 
ularly in food products because all are poisons. Lead is 
apt to occur as an impurity in many places, notably in 
water supplies not properly taken care of. Bismuth and 
arsenic are not so apt to be found as natural contamina- 
tions but the determination of small amounts of either 
is a problem often apt to arise thru their use as drugs. 

Lead as the Sulfid. 

A small amount of lead in solution on the addition of 
hydrogen sulfid will not precipitate but produces a brown 
coloration due to the presence of the colloidal sulfid in 
the solution.^ If nitric acid is present the color will not 
appear unless sodium acetate is added to the solution. In 
a solution containing hydrochloric acid, sodium acetate 
must be added before the hydrogen sulfid or the solution 
will become turbid on the addition of the latter. The 
amount of lead in the sample must be between .00025 
gram and .000005 gram. If the test is for a contamina- 
tion in water the sample may be suitably concentrated or 
diluted; if the test is of an ore or an alloy for the lead 
content the operator may be governed in his choice of a 
weight of sample by the above conditions. 

The sample, if a solid, is dissolved in twenty c.c. of 
nitric acid, the excess of acid boiled off and the solution 
made up to fifty c.c. If the sample is of water fifty c.c. 

1 J. Ghem. Soc, 97, 841. 

60 



LEAD, BISMUTH AND AESENIC. 61 

are taken. Add a few drops of sodium acetate solution 
and, if the solution is not already strongly acid, acidify 
with two or three c.c. of nitric acid. The color will 
appear upon the addition of two c.c. of a solution of hy- 
drogen sulfid. Comparison is then made with a standard. 
If the color produced shows a tendency to fade due to the 
presence of particles of precipitate the operation must be 
repeated on a new sample adding in addition to the above 
reagents, ten c.c. of a strong sugar syrup, which must be 
fresh and clear. This will prevent the formation and 
deposition of the precipitate. 

Since the color of the sufid fades in the light the use of 
the natural series of standards is not satisfactory. The 
directions already given for artificial standards admit of 
the production of such standards where numerous tests 
are to be made by this method. The dilution and bal- 
ancing methods are also applicable. As a standard dis- 
solve .1598 gram of pure lead nitrate [Pb(N03)2] in 
water and dilute to one liter. This standard contains 
.0001 gram lead per c.c. Amounts from one tenth to ten 
c.c. of this may be used to make up permanent standards 
by mixing the cobalt, iron and copper solutions specified 
so as to duplicate the color produced. If it is desired to 
use the duplication method this solution is run a drop at 
a time into 45 c.c. of water containing the same reagents 
as were specified for the sample. A standard for the dilu- 
tion and balancing methods is made by adding to ten c.c. 
of the above solution, 900 c.c. of water, five c.c. sodium 
acetate solution, ten c.c. concentrated nitric acid and five 
C.C. of hydrogen sulfid water. This is then made up to 
one liter, so that the lead content will be exactly .000001 
gram per c.c. 

This test^ is hindered by the presence of any consider- 

1 J. Soc. Chem. Ind., S8, 636. 



62 OOLORIMETEIC ANALYSIS. 

able amount of iron. This cannot be precipitated by 
rendering the solution alkaline and filtering off the re- 
sultant precipitate since the ferric hydroxide will carry 
down with it the lead present in the solution. The ob- 
struction caused by the iron may be overcome by adding 
sodium thiosulphate to reduce the iron and then a few 
c.c. of sodium or potassium cyanide. A colorless com- 
pound of the ferrous iron with the cyanide is formed and 
the lead may then be treated with sodium acetate and 
ammonium or hydrogen sulfid without rendering the so- 
lution acid. 

It is often desirable to determine lead in the same 
solution in which copper is determined. This is accom- 
plished by determining the copper by the potassium 
ferrocyanide method, then adding a few drops of potas- 
sium cyanide, which will remove the red color leaving the 
solution a light green, and determining the lead as the 
sulfid in the alkaline solution by the method given above. 
In case of the use of the latter procedure it is advisable 
to use the dilution or balancing methods and to prepare 
a special standard for the test, adding to it approximately 
the same amount of copper per c.c. as is present in the 
sample and treating it in the same way. 

Bismuth as the Iodide. 

A solution of bismuth in dilute nitric or sulfuric acid 
when acted on by potassium iodide assumes a yellow color 
due to the formation of bismuth iodide, which color is 
proportional to the bismuth content and therefore suit- 
able for colorimetric examination. The solution must be 
free from large amounts of lead, copper, tin, antimony, 
gold and silver.^ This is accomplished by precipitation. 

1 J. Soc. Chem. Ind., IS7, 102. 



m^ 



■^ 



LEAD, BISMUTH AND ARSENIC. 63 

Arsenic, copper and iron cause the results to be too high 
if they are present. The presence of hydrochloric acid 
and ammonium chloride is also to be avoided for the 
opposite reason. The weight of sample chosen must not 
contain more than .00015 gram bismuth.^ This method is 
suitable for the estimation of bismuth present anywhere 
in small quantities as, for example, in impure copper. 

The sample is dissolved in a small quantity of nitric 
acid and the bismuth precipitated by the addition of an 
excess of sodium carbonate. This will precipitate not 
only the bismuth but also a considerable volume of pre- 
cipitate of other metals. This precipitate is filtered oif 
and a qualitative test made on the filtrate to find out if 
the bismuth has all been removed. If not it must be 
allowed further time to precipitate. In exceptional cases 
it may take as long as five hours for this first precipita- 
tion to be complete. This may be somewhat hastened by 
stirring or shaking. The precipitate is well washed and 
redissolved in nitric acid. Five c.c. of lead nitrate are 
added, the solution nearly neutralized by the addition of 
ammonia and ammonium carbonate added to slight ex- 
cess. The subnitrate of bismuth will be precipitated rea- 
sonably free from other metals. If this seems too impure 
on first precipitation it may be redissolved in nitric acid 
and reprecipitated. This purified precipitate is finally 
carefully dissolved in nitric acid, washing all parts of the 
paper with dilute acid. If the solution from which this 
last precipitation was made shows a high color due to 
copper a few c.c. of potassium cyanide are added before 
precipitation, to decolorize the solution. If the volume 
of the solution of the final precipitate in nitric acid is 
greater than 25 c.c. it is concentrated to approximately 
that volume. Ten drops of sulfurous acid are now added 

1 Eng. Mining J., 104, 1091. 



64 OOIiOEIMETEIC ANALYSIS. 

and five c.c. of a solution of potassium iodide, 17 grams 
to the liter. The color is then tested as desired. 

The color produced is permanent and a series of stand- 
ards may be produced. A standard may be made up by 
dissolving .2327 gram bismuth nitrate, [Bi(N05)8, 5HjO], 
in dilute nitric acid and making up to one liter. A simi- 
lar standard may be made up by dissolving .1 gram of 
pure bismuth in nitric acid and diluting to one liter. 
Sufficient acid must be added to this solution in diluting 
it so that the bismuth does not precipitate out as the sub- 
nitrate. These standards will each contain .0001 gram 
bismuth per c.c. This standard may be used for making 
up the permanent standards or it may be added to 18 c.c. 
water and five c.c. nitric acid, treated with sulfurous acid 
and potassium iodide, and the result obtained by the du- 
plication method. For dilution or balancing it is neces- 
sary that the standard be further diluted. For this pur- 
pose fifty c.c. of the standard are treated with ten c.c. 
concentrated nitric acid, 800 c.c. water, 25 c.c. potassium 
iodide solution and five c.c. sulfurous acid and diluted to 
one liter. This standard then contains .000005 gram bis- 
muth per C.C. 

Method A. Arsenic by the Stain Produced on 
Mercuric Bromide Paper by Arsin. 

In the determination of arsenic by this method the 
arsenic is reduced to arsenious acid and then, by treat- 
ment in a hydrogen generator, to arsin. The arsin is led 
over paper treated with mercuric chloride or better with 
mercuric bromide^ and the stain produced compared with 
stains produced by standard amounts of arsenic. Sulfid, 
phosphoric acid and antimony compounds interfere. The 
first two may be removed by oxidation with nitric 
acid. When antimony is present the arsenic may be 

1 U. S. Dept. of Agriculture, Bureau of Chem., Bull. No. 102. 



LEAD, BISMUTH AND AESENIC. 65 

removed from the solution by precipitation with mag- 
nesium phosphate and filtration, but this method is not 
to be recommended. If results show that antimony is 
present it is advisable to concentrate a large amount of 
substance and carry out the determination by gravimetric 
methods. 

Iron is necessary for the success of the experiment so if 
the sample is iron-free this is added in the ferrous con- 
dition before reduction of the arsenic.^ This reduction 
is accomplished by heating near boiling in a 10 percent 
hydrochloric acid solution with potassium iodide and a 
few c.c. of stannous chloride. The stannous chloride is to 
take up any iodine freed. It is possible to reduce arse- 
nates to arsenious acid by the action of the nascent hydro- 
gen in the hydrogen generator but this is very difficult 
and the reduction previous to their introduction to the 
generator is recommended. 

The generator consists of a one or two ounce wide 
mouthed bottle. An upright tube fifteen centimeters in 
length is attached to this, which tube is divided into two 
chambers. The upper chamber is filled with cotton 
moistened with a five percent solution of lead acetate, the 
lower chamber is filled with strips of lead acetate paper. 
To the top of this tube is attached a capillary tube 3 mm. 
in diameter and 12 cm. long. The lead acetate is for 
removing hydrogen sulfid, which would also stain the 
paper used. The sample chosen for analysis, which must 
contain less than .00004 gram of arsenic is dissolved in 
acid and reduced as previously mentioned. This is then 
introduced into the generator after it has run blank for 
one hour with 40 c.c. one : four sulfuric acid and four or 
five drops of stannous chloride dissolved in hydrochloric 
acid and approximately ten grams of arsenic-free zinc. 

1 Orig*. Com. 8th Inter. Cong. App. ScL, 1, 9. 



66 OOLORIMETRIC ANALYSIS. 

The arsin evolved produces a characteristic stain on the 
prepared paper suspended in the capillary tube, a deep 
orange stain of varying length shading sharply to a light 
yellow and then to the untouched paper. 

For this paper the U. S. Dept. of Agriculture bulletin 
previously quoted recommends that mercuric bromide be 
used, since it does not require that it be developed if the 
results are to be preserved. Other authorities recommend 
the use of mercuric chloride. Strips of heavy drafting 
paper, 2.5 mm. by 12 cm., are used. These are soaked in 
a five percent solution of the salt for an hour, the excess 
of solution squeezed out and then hung up to dry. 

A standard is prepared by dissolving .1320 g. arsenious 
oxide (AsgOg) in sodium hydroxide, partially diluting to 
one liter, acidifying with hydrochloric acid, and com- 
pleting the dilution to one liter. This is too strong and 
ten c.c. of this solution are now to be diluted to 250 c.c. 
This solution will then contain .000004 gram arsenic per 
c.c. • Standards are to be prepared at intervals from 
.000001 to .00005 gram arsenic. These are prepared in 
identically the same manner as that specified for the 
sample and are suitable for keeping for six months if 
the bromide is used but will fade after a week or ten 
days if the chloride is used. The different amounts of 
arsenic present are shown by the differing lengths of the 
orange stain. If the stain produced has been contami- 
nated with antimony it will be longer and lighter in color 
than the standard. This may be confirmed by exposing 
the stain to the fumes of hydrochloric acid. A stain from 
antimony will fade on exposure while an arsenic stain 
will be intensified.^ It has been recommended that in- 
stead of zinc in the generator, arsenic-free iron be used 
for the production of hydrogen. This will produce^ the 

1 Disc. 8th Inter. Con^. App. Sci., ;?7, 4. 

2 Svensk, Farm, Tiskrift, 11, 468, 491. 



LEAD, BISMUTH AND ARSENIC. 67 

desired arsine (AsHg) but not the undesirable stibine 
(SbHg) or phosphine (PH3). The use of a small amount 
of stannous chloride in the generator serves to sensitize 
the zinc. The generator must be so standardized that the 
production of hydrogen is uniform, both while producing 
the sample stain and the standard stains. If the sample 
contains organic materials it must be so treated that these 
will not interfere with the production of hydrogen at the 
desired rate. More than a trace of fluorine must be 
eliminated from the sample before testing. 

Method B. Arsenic by Sil\^r Nitrate. 

In this modification^ of the previous method the arsenic 
is treated as before and passed thru the tubes with lead 
acetate paper but instead of the use of mercuric bromide 
paper the gas is then passed thru a tube containing a few 
crystals of silver nitrate. The black stain produced by 
the arsenic on these crystals is to be compared with stains 
produced by known amounts of arsenic in a similar way. 
The standards after production may be sealed in the 
tubes in which they were produced and kept indefinitely. 

1 Compt. Rend., 158, 869. 



CHAPTER EIGHT. 

ALUMINUM AND CHROMIUM. 

Of the members of the third group of metals, iron was 
worthy of a chapter by itself, leaving us the remaining 
two for discussion here. Aluminum, besides its occur- 
rence in rocks and minerals, is determined in cements 
and plasters. Chromium occurs in minerals to some ex- 
tent but the big field of usefulness of this test for 
chromium is the analysis of chromesteel for the chromium 
content. 

Aluminum by Alizarin- S. 

The sodium salt of alizarin monosulfonic acid is a com- 
mercial indicator which shows purple with alkalies and 
yellow with acids. The coloration produced in acid solu- 
tion^ may be adapted to the determination of aluminum, 
the test being delicate enough for the detection of one 
part of aluminum in ten million parts of water. Small 
amounts of chromium and iron do not interfere, in the 
presence of large amounts of these metals it is necessary 
to add citric acid or a citrate to prevent them from show- 
ing. It is not advisable to render the solution alkaline 
with ammonia for the precipitation of the hydroxides 
from other substances present, as the reagents used are 
almost certain to contain traces of aluminum. The 
presence of phosphates or of moderate amounts of cal- 
cium, magnesium or zinc salts has no effect on the results. 
If cobalt is present the solution must never become alka- 
line. The sample should be chosen to contain between 
.00005 gram and .000005 gram of AI2O3. 

1 J. Soc. Chem. Ind., 84, 936. 

68 



ALUMINUM AND CHROMIUM. 69 

The sample taken is to be dissolved in ten to twenty 
C.C. as dilute acid as possible or, if the sample taken ex- 
ceeds the above limits, in a larger amount and an aliquot 
part taken. If the acid is very concentrated it should be 
partially neutralized so as to leave the final solution acid' 
but not too strongly so. If the solution is made in water 
it is acidified. The use of nitric acid is to be avoided as 
far as possible since it tends to cause the color to fade 
more than do the other acids, altho all cause a slow fading 
of the color produced. Ten c.c. of glycerine are next 
added to the solution to counteract the tendency of the 
aluminum to precipitate and lastly five c.c. of a one per- 
cent solution of the reagent. The solution must stand for 
five minutes after which it is acidified with dilute acetic 
acid until no further change in color occurs, diluted to 
fifty C.C. and compared with a standard. 

Because of the lack of permanency of the color a set of 
natural standards is not satisfactory and, because the 
solution is apt to fade more or less before the comparison 
is made, the use of artificial standards is not advised. 
The method of duplication is out of the question because 
it takes some time for the color to be produced after the 
standard comes into contact with the reagent. It is there- 
fore better to use only the dilution method for this test, 
since to place the standard in the balancing apparatus for 
each determination would be a useless waste of time. A 
standard is prepared at the same time as the sample and 
subjected to the same conditions of temperature, etc. 
Any fading will then be about the same in each. As a 
solution for preparing the standard dissolve .9286 gram 
common alum [K2SO4, Al2( 804)3, 24H2O] in water and 
dilute to one liter. Ten c.c. of this solution are then to be 
pipetted out and again diluted to one liter. The resultant 



70 OOLORIMETRIG ANALYSIS. 

solution contains .000001 gram AI2O3 per c.c. Five to 
twenty-five c.c. of this may be taken as necessary for the 
sample used and prepared as previously directed. 

Method A. Chromium as the Chromate. 

Small amounts of chromium may be determined by the 
oxidation of the chromium to the form of a chromate. 
This method is particularly designed for the analysis of 
minerals but is also adaptable to steel analysis. The 
sample chosen must contain not less than .002 gram 
chromium as experiment shows that the error is too great 
in the oxidation of a lesser amount. The color may later 
be lessened by dilution and test of an aliquot part. 

The sample is dissolved in as dilute sulfuric acid as 
possible and if the acid is very strong it is nearly neutral- 
ized by the addition of sodium carbonate. Concentrated 
sodium thiosulfate is now added to the solution, which 
precipitates chromium and manganese together with some 
other metals which will not interfere with the test. This 
precipitate is filtered off and dissolved in nitric acid. 
A few c.c. of silver nitrate are added as a catlyzer 
and then sodium persulfate which oxidises the chromium 
present to the form of the chromate. If there was man- 
ganese present in the original sample this will have been 
carried thru and wuU now be present as the permanga- 
nate. Since the color of the permanganate will cover up 
the yellow of the chromate it is necessary to remove the 
former, which is accomplished by adding ammonia and 
heating the solution for a short time.^ The permanga- 
nate will be precipitated as the hydroxide and the silver 
as the oxide but the chromate will remain unchanged. 
Hillebrand recommends the addition of methyl or ethyl 

1 Z. Anorg. Chem., 80, 171. 



ALUMINUM AND CHROMIUM. 71 

alcohol to the solution to destroy the color of the per- 
manganate.^ 

The precipitate is filtered off and if iron is present 
sodium phosphate is added to prevent interference from 
that source. The sample is then tested against a stand- 
ard solution of potassium chromate (K2Cr04). This is 
made up by dissolving .2555 gram of the pure salt in 
water and diluting to one liter. The standard then con- 
tains .0001 gram CraOg per c.c. The preparation of a 
series of standards is possible as the standard is per- 
manent, or the sample may be duplicated by running a 
few C.C. of the above solution into a blank of water. The 
use of the above solution for dilution and balancing 
methods is possible or, if the sample is such that it seems 
to require it, 50 c.c. may be diluted to 250 c.c. and will 
then contain .00002 gram Cr^Og per c.c. 

Method B. Chromium by Disodium 1.8 Dihydroxy- 

naphthalene 3.6 dlsulfonate. 

The analysis of steel for chromium has been simplified 
by the adaption of the above i-eagent to the use. The 
resultant color is pink and is so delicate that .0000008 
gram of chromium can be detected. If a considerable 
quantity of vanadium^ is present this produces a brown 
color which is apt to obscure the results. If the vanadium 
present is less than the chromium a correction may be 
introduced. Subtract from the percentage of chromium 
shown, one third of the percentage of vanadium as shown 
by a separate analysis in order to obtain the true per- 
centage of chromium present. If the ratio of chromium 
to vanadium is high the error may be negligible and so 
can be disregarded. A correction may be introduced by 
using a similar amount of vanadium in the standard. 

1 U. S. Geol. Surv., BuU. 176, p. 80. 

2 J. Ind. Eng. Chem., 5, 298. 



72 COLOEIMETEIC ANALYSIS. 

The sample of .2 gram is dissolved in ten c.c. dilute 
sulfuric acid, one half c.c. nitric acid added to oxidise the 
iron and evaporated to fumes. This is then cooled, 
slightly diluted with water, fifty c.c. of a ten percent 
solution of sodium hydroxide and one gram sodium 
peroxide added. After boiling for five minutes to de- 
stroy the excess of hydrogen peroxide the solution is 
cooled, diluted to 200 c.c, and filtered. Use as a sample 
100 c.c. of this to which are added 2 c.c. of 85 percent 
phosphoric acid and two c.c. of Konig's reagent. The 
later is a one percent solution of the reagent for the test, 
disodium 1.8 dihydroxynaphthalene 3.6 disulphonate. 
The color will at once appear. For a steel low in 
chromium (under .15 percent), so that the color does not 
show up as very deep, it may be well to repeat the experi- 
ment using a .4 gram sample. 

The examination may be made by the dilution method 
only, because of the peculiar character of the color devel- 
oped and the danger of interference. The standard 
selected is a steel of known chromium content. The 
standard is treated by the same identical method as the 
sample. 



CHAPTER NINE. 

NICKEL, OOBAI/r, MANGANESE AND ZINC. 

Or the members of this group, two are commonly used 
in the form of alloys with steel and therefore their deter- 
mination assumes vital importance in the steel industry. 
The other two, zinc and cobalt, are of such frequent occur- 
rence in inorganic salts that even the occasions of their 
detection as impurity affords a considerable use of the 
methods here given. 

Method A. Nickel by Potassium Thio-carbonate. 

When nickel in ammoniacal solution^ is acted on by 
potassium thio-carbonate a rose red to brown red is pro- 
duced, suitable for colorimetric examination. Metals of 
the second group interfere but may be removed by pre- 
cipitation with hydrogen sulfid. Manganese and iron are 
precipitated from the solution free from hydrogen sulfid, 
by bromine or hydrogen peroxide and ammonia. At the 
same time any cobalt present as cobaltous compounds is 
oxidised to the cobaltic state in which it does not inter- 
fere. Zinc forms a white precipitate with the reagent 
which is, however, so fine as not apt to obscure the test 
unless a considerable amount is present. If there is any 
doubt as to whether the zinc interferes a similar amount 
may be added to the standard. The sample should be so 
chosen as to contain between .0025 gram and .0004 gram 
of nickel. 

The sample is dissolved in nitric, hydrochloric or sul- 
furic acid as is most convenient, altho nitric is recom- 

1 Z. Anal. Chem., 5S, 165. 
6 73 



74 C50L0RIMETEIG ANALYSIS. 

mended as ordinarily the quickest. If the sample is a 
salt soluble in water the solution is slightly acidified. 
The solution is then almost neutralized with ammonia 
and a solution of potassium thio-carbonate added (five 
c.c. is sufficient) and ammonia in excess. Comparison is 
made by dilution as the color is not permanent. 

The standard is made up by dissolving .6730 gram of 
nickel ammonium sulfate [NiS04, (NH4)2S04, 6H2O] in 
water and diluting to one liter. This solution contains 
.0001 gram nickel per c.c. From one to twenty-five c.c. 
of this solution are treated in the same way as the solution 
of the sample and comparison then made by dilution. A 
series of standards is not permanent enuf. The solution 
may also be added to a blank containing the reagents, 
for the making up of a duplicate. 

Method B. Nickel as the Chloride in Concentrated 

Acid. 

Comparatively large quantities of nickel may be esti- 
mated by the intensity of the yellow color of the pure 
chloride in concentrated hydrochloric acid solution. The 
maximum color^ is reached with the solution about thirty 
percent hydrochloric acid after which the color is con- 
stant regardless of the concentration of the acid above 
that point. The sample, if the ore or alloy to be ex- 
amined contains from .1 percent to 10 percent nickel, is 
.2 gram. If the nickel content is greater than that, .1 
gram will do, if less, the sample may be increased. Iron 
and copper must be removed. Copper is precipitated 
with hydrogen sulfid after which iron is oxidised to the 
ferric condition and precipitated as the basic acetate. 
There must be no nitric acid or free chlorine present, the 
presence of sodium chloride has no effect. 

1 Z. Anorg. Chem., 86y 341. 



NICKEL, COBALT, MANGANESE AND ZINC. 75 

The sample of .2 gram or other weight as specified is 
dissolved in concentrated hydrochloric acid if possible. 
This will not usually be possible and the solution must 
then be made iti nitric acid. In the latter case hydro- 
chloric acid is added to the solution and it is taken to 
dryness after which more hydrochloric acid is added and 
the process repeated, likewise a third time. The residue 
is taken up with concentrated hydrochloric acid, trans- 
ferred to a colorimetric tube and diluted with concen- 
trated hydrochloric acid to fifty or 100 c.c. as is most 
convenient, taking into consideration the depth of color 
present. 

Cobalt produces a color so that the correction for its 
presence is introduced by the use of a standard solution. 
The process also furnishes an estimate of the amount of 
cobalt present. The standards made up are 4.9556 grams 
of pure nickel nitrate [Ni(N03)2, 6H2O] and 4.9361 
grams of pure cobalt nitrate [Co(N03)25 6H2O]. These 
weights of the salts are dissolved in concentrated hydro- 
chloric acid and each evaporated nearly to dryness. 
More concentrated hydrochloric acid is now added to 
each solution and the evaporation repeated, making sure 
that all nitric acid is expelled. The residues are taken 
up with hydrochloric acid and diluted to 500 c.c. each 
with concentrated hydrochloric acid, content .002 gram 
nickel or cobalt per c.c. 

The determination of the nickel present in the sample, 
if it contains no cobalt, may be carried out by the balanc- 
ing method. The making up of a series of standards is 
not advisable as this determination usually has two vari- 
ables, the amount of cobalt solution added to the nickel 
solution and the amount of nickel solution which is used 
with pure acid. The dilution method is carried out by 
diluting the nickel solution with concentrated hydro- 



76 COLORIMETRIO ANALYSIS. 

chloric acid. If the sample solution shows a more or 
less green tinge the same tint is produced in the stand- 
ard by the addition of the standard cobalt solution. The 
amount of this used is then a measure of the cobalt pres- 
ent in the sample. When the colors of the two solutions 
are identical the results may then be figured and by the 
usual proportions the weight and percentages in the 
sample obtained. 

Caution, the nickel must always be in concentrated 
acid or the yellow will turn to green and similarly the 
cobalt must always be in concentrated acid or the blue 
color will turn to pink. 

Method A. Cobalt as the Chloride in Concentrated 

Acid. 

This method for the determination of cobalt is iden- 
tical with the determination of nickel^ which just pre- 
cedes. The sample of .2 gram is dissolved in the same 
way and in the absence of nickel, iron and copper having 
been removed according to instructions given there, the 
test is made by dilution, using the cobalt standard as 
directed and diluting with concentrated acid. If nickel 
is present sufficient is added of the standard solution to 
duplicate the tint in the standard, after which dilution is 
carried out as usual. 

Method B. Cobalt by a-Nitroso-b-Naphthol. 

The red color produced by cobalt in reaction with 
a-nitroso-b-naphthol may be used for the quantitative 
determination^ of small amounts of the metal. If more 
than a trace of copper is present it must be removed by 
precipitation with hydrogen sulfid in acid solution. 

1 Z. Anorg. Chem., 86^ 341. 
2 Analyst, 4S, 317. 



NICKEL, COBALT, MANGANESE AND ZINC. 77 

Nickel, if present, is removed by precipitation with 
dimethylglyoxime, after which the excess of dimethyl- 
glyoxime must be removed by evaporation to dryness and 
treatment of the residue with aqua regia. The aqua regia 
having been driven off by heating, the residue may be dis- 
solved in hydrochloric acid and the process continued. 
If the amount of manganese present is in excess it must 
be removed by the addition, of an equal amount of nitric 
acid to the solution and a small amount of sodium 
bismuthate and boiling. If the manganese is not in ex- 
cess of the cobalt it may be ignored. The manganese 
will be precipitated by the sodium bismuthate and should 
then be filtered off. Organic matter is removed by the 
evaporation of the solution to dryness and taking up 
with aqua regia, evaporating to dryness again and ignit- 
ing, after which it is dissolved in concentrated hydro- 
chloric acid and diluted for analysis. The addition of 
ammonium citrate prevents the interference of metals 
other than those mentioned above. The sample should 
be so chosen as to contain .002 gram to .00005 gram cobalt. 
The sample chosen according to the above limits is dis- 
solved in concentrated hydrochloric acid, the total volume 
of solution being about 25 c.c. Any of the interfering 
metals whose removal is necessary are removed, and the 
solution is then ready for treatment with the reagent. 
The pure solution of the reagent does not keep satis- 
factorily but by conversion to the sodium salt it becames 
permanent. This is accomplished by boiling .1 gram 
a-nitroso-b-naphthol with one c.c. caustic soda and dilut- 
ing to 200 c.c. Before the addition of this to the sample 
the solution is nearly neutralized with ammonia and five 
c.c. of a solution of 500 grams citric acid in 250 c.c. of 
water and 500 c.c. of ammonia, added. This should have 
an excess of ammonia so that on addition of it the solu- 



78 COLOEIMETRIO ANALYSIS. 

tion will become alkaline. If the solution is not slightly 
alkaline a small amount of ammonia is added to render 
it so and then diluted to nearly 100 c.c. Five c.c. of the 
reagent are then added and dilution to 100 c.c. completed. 
A standard is prepared by dissolving .4936 gram of 
cobalt nitrate [Co(N03)25 BHjO] in water and diluting 
to one liter. This solution contains .0001 gram cobalt 
per c.c. One to twenty c.c. according to the estimated 
size of the cobalt content of the sample are taken as a 
standard, treated with the reagents and comparison made 
by the dilution method. A duplicate may be made up 
by running this solution into a blank of 75 c.c. of water 
with the same reagents added to it as to the sample. A 
series of standards cannot well be made up as the cobalt 
comes down after a short time as a red precipitate. For 
the balancing method fifty c.c. of the standard solution 
should be treated with double the quantities of reagents 
specified for the sample and then diluted to 500 c.c. This 
will then show the characteristic color and may be used 
at once but after a time will precipitate out. The latter 
solution contains .00001 gram cobalt per c.c. 

Method A. Manganese as Permanganate OxroATiON 

BY Persulfate. 

The red color of manganese after it has been oxidised 
to the permanganate^ is used as a means of estimation 
of the manganese content of an ore or an alloy. Prac- 
tically no metal interferes except chromium and that not 
unless it is in considerable excess of the amount of per- 
manganate. If a large amount of iron is present as in the 
analysis of a steel for the manganese content, phosphoric 
acid is added to prevent the precipitation of the iron or 
a yellow color from that source. The samples should be 

1 Z. Anorg. Chem., 80, 171. 



NICKEL, COBALT, MANGANESE AND ZINC. 79 

chosen as follows, .02 percent to .25 percent manganese 
take .5 gram, up to 4 percent use .25 gram. Above that 
limit .1 gram is to be used or the sample diluted and an 
aliquot part taken. 

The sample is dissolved in dilute sulfuric or nitric acid 
and any residue of silica filtered off. Ten c.c. of fifth 
normal silver nitrate solution are then added as a catal- 
yzer. If free chlorine or hydrochloric acid have entered 
into the solution in any way it may be necessary to add a 
considerable volume of silver nitrate to precipitate these 
before the catalyzer can be added.^ Potassium persulf ate 
is then added in sufficient quantity to oxidise all the man- 
ganese present to the condition of permanganate. The 
color will at once appear. If there is not too much chro- 
mium present the solution is now ready for testing, other- 
wise a further step must be gone thru. 

In case too much chromium is present ammonia is now 
added and the solution heated. The manganese will be 
precipitated as manganese hydroxide and silver and a 
few other metals present will come down also. The chro- 
mium will, however, remain in the solution as the chro- 
mate unchanged. This is the same method that is used 
to purify the chromate solution for the obtaining of the 
chromium content. The precipitate containing the man- 
ganese is filtered off, dissolved in nitric acid and a further 
quantity of potassium persulfate sufficient to again oxi- 
dise the manganese is added. The permanganate color 
will reappear and the solution is ready for testing. 

An old standard permanganate solution which has been 
used for iron analysis so that the iron factor is known, 
is best used as the standard. In that case the changes 
which occur in the solution when first made up are not 
to be reckoned with. In such a case multiply the iron 

1 C. N., 8S, 76. 



80 COLORIMETRIO ANALYSIS. 

factor of the solution by .2952 to get the manganese 
factor. If a new standard must be made up .1439 gram 
potassium permanganate is dissolved in 100 c.c. of water. 
The water for this solution or for the dilution of the old 
solution, if it is to be diluted, must be boiled and cooled 
before it is ready for use. This 100 c.c. of solution is 
then diluted to one liter with the addition of ten c.c. of 
nitric acid. The latter standard contains .00005 gram 
manganese per c.c. 

The making up of a series of permanent standards is 
not to be relied on. A blank may be prepared of twenty- 
five c.c. of water and the sample diluted to fifty c.c. after 
which the blank is made up as a duplicate by the addi- 
tion of the above standard. The above standard may be 
further diluted, 100 c.c. diluted to 500 c.c. if thought 
desirable for the use of the dilution or balancing methods. 
Water to be used in carrying out the dilution method 
should have been boiled and cooled shortly before using. 

Method B. Manganese as Permanganate, OxroATioN 

BY Periodate. 

Instead of the use of a persulfate which involves the 
use of a catlyzer the oxidation may be carried out by 
the addition of sodium periodate to the solution.^ It is 
necessary that there be an excess of free mineral acid in 
the solution, otherwise the manganese will precipitate out 
as manganese periodate or some oxide of manganese. 
The presence of much iron is taken care of by the ad- 
dition of a few c.c. of strong phosphoric acid, which pre- 
vents the yellow color due to the iron as well as a possible 
precipitation. Other metals do not interfere unless they 
produce a colored solution in which case they must be 
removed. 

1 J. A. 0. S., S9, 2366. 



NICKEL, COBALT, MANGANESE AND ZINC. gl 

The sample is chosen as in the preceding method and 
dissolved in a mineral acid. The presence of hydrochloric 
acid or free chlorine will not interfere with this method. 
The concentration of the acid used is best, 15 c.c. concen- 
trated sulfuric or 20 c.c. nitric acid or 10 c.c. strong phos- 
phoric acid per 100 c.c. of solution. Any reducing sub- 
stances may be removed by adding nitric acid and heating. 
.4 gram sodium periodate is added to the sample and the 
solution boiled for one minute, kept hot for ten minutes, 
cooled and diluted. The comparison is made with stand- 
ards as specified in the previous method. 

Zinc by Resorcinol. 

Zinc is acted on by resorcinol in alkaline solution to 
produce a blue color.^ The character of this alters some- 
what on exposure to the air. The addition of hydro- 
chloric acid turns the solution red and this color is not 
altered. Other metals do not interfere unless they have 
a colored solution before the addition of the reagent. In 
that case the metal causing the color must be removed. 
A sample suitable for examination by this method must 
contain between .003 and .000005 gram zinc. 

The sample is dissolved in acid, diluted to 75 c.c. and 
neutralized by the addition of ammonia. The solution 
is diluted to 100 c.c. Two c.c. of ammonia are then 
added and two c.c. of a five percent solution of resorcinol, 
which will produce the characteristic color. 

A standard is prepared by the solution of .1 gram of 
pure zinc in nitric acid and dilution to one liter. A suit- 
able amount of this is treated the same as the sample. 
The standard so prepared must be very close to the color 
of the sample as the dilution method does not hold thru 
any considerable range in this case. The other methods 

lAnales. Soc. Espan. Fis. Quim., 11, 98. 



g2 COLORIMETRIO ANALYSIS. 

are impracticable except the duplication method. For 
that the blank should contain 70 c.c. of water. After 
results have been obtained by the dilution method in 
alkaline solution a slight excess of hydrochloric acid may 
be added and the test checked by comparing the red color 
produced in acid solution. 



CHAPTER TEN. 

POTASSIUM AND MAGNESIUM. 

The determination of potassium is of considerable im- 
portance in the examination of soils and drainage waters 
as well as the determination of traces of the metal in 
analysis of minerals and substances of all sorts when an 
accurate analysis is made. Magnesium occurs in minerals 
and as one of the dissolved substances in natural waters. 

Method A. Potassium by Determination of the 

Potassium Platino Chloride by Reduction 

WITH Stannous Chloride. 

This method for the determination of potassium de- 
pends on the precipitation and separation of potassium 
as the platino chloride, the solution of this amount of 
platinochloride and reduction with stannous chloride. 
The color produced^ is a yellow proportional to the 
amount of platinum present which is in turn proportional 
to the amount of potassium present. Since the standard 
is a similar salt the proportion may be taken as directly 
resultant from the potassium without consideration of the 
fact that the color is due to the platinum rather than to 
the metal being tested for. This method admits of esti- 
mation of amount so small that they could hardly be 
detected by the usual gravimetric procedure. The amount 
of sample should be so chosen that the potassium content 
is between .00005 and .000005 gram per c.c. in a solution 
of 25 c.c. 

1 J. A. C. S., ^5, 991. 

83 



84 COLORIMETRIG ANALYSIS. 

The sample, if a solid, is dissolved, in water if so solu- 
ble, in hydrochloric or nitric acid if necessary. If the test 
is being made on a solution it may be necessary to con- 
centrate in order to come within the limits. The sample 
solution is then evaporated to dryness with the addition 
of one c.c. of concentrated sulfuric acid. The residue is 
strongly ignited and dissolved in hot water, acidified 
with hydrochloric acid and an excess of chlorplatinic 
acid added. The solution is washed out of the dish onto 
a filter with alcohol (80 percent concentration or more) 
and washed with several successive small portions of 
alcohol. The precipitate is now dissolved in boiling 
water, cooled and made up to a definite volume. This 
should be 50 c.c. if the potassium content is low or 100 c.c. 
if the potassium content is high. In the latter case the 
solution is divided into two parts and one half used for 
the further procedure. 

A stannous chloride solution is made up by boiling 75 
grams of powdered or granulated tin in an erlenmeyer 
flask with 400 c.c. of concentrated hydrochloric acid until 
all the tin has dissolved. Three c.c. of this solution are 
added to the solution of the sample and the yellow color 
will appear. 

A standard is made up by dissolving .0516 gram of 
potassium chlorplatinate (K2PtClo) in water and making 
up to one liter. This solution will contain .00001 gram 
K2O per c.c. If a series of standards is to be prepared 
from one to twenty -five c.c. of this solution are used for 
the original standards, treated with stannous chloride and 
made up to 50 c.c. For the method of duplication the 
above standard is run into a blank of 25 c.c. of water 
treated with three c.c. of stannous chloride. A standard 
solution showing the color so that it may be used for 
comparison by the dilution or balancing methods is made 



POTASSIUM AND MAGNESIUM. 85 

by treatment of 200 c.c. of the solution with five c.c. of 
hydrochloric acid and ten c.c. stannous chloride and mak- 
ing up to 250 C.C. This solution will then contain .000008 
gram potassium per c.c. 

Method B. Potassium as the Chlorplatinate by 

Potassium Iodide. 

This method is for uses similar to those of the preced; 
ing method but is more delicate. The sample should be 
smaller than those recommended for the preceding method. 
The sample is dissolved and treated similarly until the 
solution of potassium chlorplatinate in water is obtained. 
The washing^ of the precipitate should be with stronger 
alcohol than in the preceding case and the last washing 
must be with absolute alcohol. This must be allowed to 
stand for a sufficient time for all of the alcohol to evapo- 
rate before solution is made, otherwise the alcohol in the 
final solution will cause an impediment to correct results. 

The fifty c.c. of sample are then treated with five c.c. of 
a solution of potassium iodide, 86 grams to the liter. A 
red coloration is produced which is darker for the same 
amount of potassium than that produced by reduction of 
the chlorplatinate with stannous chloride. The standard 
used is more dilute, .1032 gram of potassium platino 
chloride (KoPtCle) is dissolved in one liter of water. 
The standard for the series of standards or duplication 
methods is made by dilution of ten c.c. of this solution to 
100 c.c. so that it then contains .000002 gram KgO per c.c. 
For the dilution or balancing methods five c.c. of the 
standard solution are treated with five c.c. of potassium 
iodide solution and made up to 100 c.c. The content per 
c.c. of this is .000001 gram KgO per c.c. 

1 J. A. C. S., fS5, 1063. 



86 COLOBIMETRIO ANALYSIS. 

Magnesium by Determination of the Phosphate 

AS Phosphomolybdate. 

For the determination of magnesium the magnesium is 
changed to the form of the ammonium phosphate.^ This 
is then treated with a standard molybdate solution which 
will give a yellow color proportional to the amount of 
phosphate present, which is in turn proportional to the 
amount of magnesium present. The standard is a phos- 
phate solution and the amount of magnesium represented 
by a gram of PjOg is known so that the percentage is 
obtained by this factor. Calcium precipitates, if present, 
but the addition of a few drops of ammonium oxalate 
solution entirely prevents this interference. The sample 
is so chosen as to contain between .0001 and .00003 gram 
magnesium. 

The sample, if a solid, is dissolved in as small an amount 
of nitric or hydrochloric acid as possible, evaporated to 
dryness to drive off the excess of acid and the residue dis- 
solved in water. If some silica is present this may be 
filtered off before evaporation to dryness. If the sample 
is a liquid at the start such as, for example, a drinking 
water it may be necessary to concentrate. The sample 
should finally be concentrated to about five c.c. of solution 
and cooled. 

The phosphate reagent is made up by dissolving 17.4 
grams potassium hydrogen phosphate (K2HPO4) and 
100 grams ammonium chloride in 900 c.c. water, adding 
fifty c.c. strong ammonia and diluting to one liter. The 
cheaper sodium salt is not satisfactory as it is too diffi- 
cult to wash because of its lesser solubility.^ To the 
sample as specified above is added one drop of ammonia 
and two or three drops of ammonium oxalate and the 

1 J. A. C. S., S6f 961. 

2 J. A. C. S., ^6, 1463. 



POTASSIUM AND MAGNESIUM. 87 

whole evaporated to dryness on the water bath* One c.c. 
of the phosphate solution is added to this residue and the 
whole thoroly stirred with a glass rod and allowed to 
stand for two hours. The precipitate is washed out onto 
a small filter with ammonium hydroxide and washed with 
further quantities of ammonium hydroxide, five c.c. at 
a time, until a total of fifty c.c. has been used. The 
original dish is washed out with each addition before it 
is added to the filter. The dish is washed out with five 
c.c. of pure water and the filter also washed with this. 
A clean beaker is placed under the filter and the dish 
washed out with five c.c. of concentrated nitric acid. 
This same acid is carefully poured over the filter so as 
to dissolve all the precipitate present. The dish and 
filter are washed with hot water, five c.c. at a time, until 
nearly 45 c.c. have been used. The filtrate is cooled, four 
c.c. of ammonium molybdate solution added and diluted 
to fifty c.c. The yellow color now appears and is at its 
greatest intensity after twenty minutes. The solution 
must therefore be allowed to stand until that time before 
comparison. 

A standard is made up by dissolving .5043 gram of 
pure sodium acid phosphate (Na2HP04, I2H2O) in 
water, adding 100 c.c. of nitric acid and making up to one 
liter. This standard contains .0001 gram P2O5 per c.c. 
The test cannot conveniently be made by the production 
of a duplicate since an addition of the standard takes 
twenty minutes to reach its full intensity. The standard 
for use by the balancing or dilution methods is made up 
by adding to ten c.c. of the above solution, 70 c.c. of 
water, 9 c.c. nitric acid and 8 C.c. of ammonium molybdate 
solution and making this up to 100 c.c. accurately. After 
standing, this solution attains the color representing its 
content of .00001 gram PgOg per c.c. or representing 



88 CX)LORIMSTEIC ANALYSIS. 

.00000342 gram magnesium. That is, the amount of 
phosphorus present as PgOg is multiplied by the factor 
.342 to give the amount of magnesium represented by this 
result. The result is in terms of Mg. The factor to rep- 
resent the result in terms of MgO is .568. 



CHAPTER ELEVEN. 

GOLD. 

The quantitative estimation of gold in solution is one 
field left almost exclusively to colorimetry. Tests may 
be performed which are so delicate as to show the pres- 
ence of one cent's worth of gold in a ton of water and by 
concentration of the solution being tested even greater 
delicacy may be secured. It is by these methods that the 
well known fact that every ton of sea water contains a 
few cents' worth of gold is proven. 

Eight brief methods follow. 

Method A. Gold by Method of Dowsett. 

This method for the determination of dilute gold solu- 
tions does not require the making up of a series of stand- 
ards but requires that the worker have sufficient knowl- 
edge of the colors after some use of the methods to be able 
to estimate the gold content from the depth of color pro- 
duced. A solution containing gold up to a value of 8 
cents per ton may be used. If the test is to be made of 
an ore this is dissolved in some manner. The most con- 
venient method is usually by the extraction of the gold by 
a dilute cyanide solution. 

A sample^ of 500 c.c. of the solution to be tested is 
poured into a bottle which has very little shoulder, such 
as a large soft drink bottle, ten to fifteen c.c. of a satu- 
rated solution of sodium cyanide added and two or three 
drops of a saturated solution of lead nitrate. Two 
grams of fine zinc dust are then added and the bottle 

1 Met. and Chem. Eng., IB, 460. 
7 89 



90 COLORIMETRIO ANALYSIS. 

corked and shaken until the precipitate formed settles 
quickly, usually for about two minutes. The bottle is 
then inverted into a casserole and the clear liquid decanted 
off. Nitric acid is added to the precipitate until the 
reaction ceases, a drop or two in excess added and the 
solution evaporated to one or two c.c. This small volume 
of solution is transferred to a small test tube and one c.c. 
of stannous chloride solution added to bring out the 
color. If the quantity of gold is small this may take two 
or three minutes ; in case the color is faint it is more read- 
ily detected by looking down the tube. 
The colors produced are : 

2 cents per ton — Very slight color. 

3 cents per ton — Slight yellow color. 

4 cents per ton — Slight pinkish yellow color. 
6 cents per ton — Strong pink color. 

8 cents per ton — Purple of Cassius. 
For the experimenter to learn this method a few tests 
on known solutions are advisable, to get the colors firmly 
fixed in mind. The smallest amount of lead possible 
should be used so that the precipitate to redissolve will 
be small. Excess nitric acid will cause the colors to differ 
from those specified, therefore the nitric acid should be 
added cautiously. Mercury if present will cause a dark 
color which will cover up the results of the test. 

Method B. Gold by Method of Prister. 

A copper solution^ for carrying out this test is made by 
boiling a solution of two parts salt to one part copper 
sulfate with ten parts water in the presence of scrap 
copper for ten minutes and adding a few drops of acetic 
acid to the solution on cooling. 

A sample of 200 c.c. of solution is acidified with hydro- 

1 Proc. Chem. Met. and Min. Soc. of So. Af ., IV, 235. 



GOLD. 91 

chloric acid and boiled several minutes to decompose 
cyanide if present. A slight excess of the copper solu- 
tion is then added and a few drops of sodium sulfid 
solution. After boiling this solution for five minutes the 
precipitate is allowed to settle and the liquid decanted 
onto a filter. The precipitate in the beaker and on the 
filter is dissolved in three c.c. of a five percent solution of 
potassium cyanide to which has been added a few drops 
of potassium hydroxide. Two grams of zinc dust are 
added to this solution and the mixture then heated to 45 
degrees for a half hour. The liquid is separated by 
decanting thru a filter after which the excess of zinc is 
removed by treatment with hydrochloric acid. The 
residue is now treated several times with ten c.c. of aqua 
regia passing it thru the filter and washing out the beaker 
thoroly. 

The color will be produced by diluting this solution to 
twenty c.c. with stannous chloride. As in the preceding 
case, tests must be made on solutions of known gold con- 
tent to become familiar with the colors produced. 

Method C. Gold by Method of Dorino. 

This method is for the analysis of ores poor in gold. 
One hundred grams of ore are slightly but evenly moist- 
ened in a stoppered bottle with one or two c.c. of equal 
parts of bromine and ether. This is to be shaken at 
frequent intervals for two hours, during which time the 
interior of the bottle must be filled with bromine vapor. 
Fifty C.C. of water are added to the bottle and it is allowed 
to stand for two hours shaking occasionally. The solu- 
tion is then filtered, the filtrate evaporated to one fifth of 
its former volume, a little bromine water added and the 
solution treated with stannous chloride in a test tube. 
Results are as follows: 



92 COLOBIMBTBIO ANALYSIS. 

.1 percent solution, deep brown color, opaque even in 

thin layers. 
.01 percent solution, brown violet immediately, 14 cm. 

column opaque. 
.001 percent solution, pale violet immediately, increases 

after a time. 
.0001 percent solution, evaporated to one fifth former 

volume, and a drop of bromine water added 

shows a distinct rose tint in a 14 cm. column. 
.00005 percent solution shows faint but recognizable pink 

in 14 cm. column, after undergoing the same 

treatment. 
This method will show as little as one half gram per 
ton of ore. From the percent of the solution the gold 
content of the ore may be calculated. 

Method D. Gold by Method of Rose. 

This method is for detection of gold in substances 
which can be dissolved in water, such as a trace of gold 
in a salt, or similar tests. The sample^ is dissolved in 
water, this solution heated to boiling and then poured 
into a solution of stannous chloride. There is formed a 
gelatinous precipitate which has the purple of Cassius if 
considerable gold is present and shows a faint pink for 
amounts as small as one part in one hundred millions. 
The operator is expected to familiarize himself with the 
colorations produced by known quantities of gold. 

Rose suggests the making up of a standard by solution 
of .00003 gram of gold in three liters of water. This 
would be one part in one hundred millions and could be 
concentrated to give a standard of the desired concen- 
tration. 

1 C. N., 66, 271. 



GOLD. 93 

This test used with sea water as a sample gives a very 
distinct reaction. 

Method E. Gold by Method of Cassal. 

This method^ is for treatment of a cyanide solution for 
determination of the gold content. To the sample of 
50 c.c. of solution a small amount of potassium bromate 
is added and then concentrated sulfuric acid until 
effervescence starts. The action will continue without 
further addition of acid until the cyanide is completely 
decomposed. The bromine is then boiled off and stannous 
chloride added in excess which will produce the color 
desired. If the test is made quickly it is not necessary 
to boil off the bromine before addition of the stannous 
chloride but the color will not remain at its true intensity 
for more than a minute unless the bromine is so removed. 
The colors are determined from the memory of those 
caused by standard amounts of gold. This test will show 
a light color from the presence of as little as one cent a 
ton. The purple of Cassius appears when only four dwt. 
of gold per ton are present. 

Method F. Gold by Decomposition of the Cyanide 
BY Potassium Bromide and Sodium Peroxide. 

This method^ is similar to method E except that potas- 
sium bromide is added to the 50 c.c. of sample solution 
and then sodium or potassium peroxide in excess. 
Bromine will be freed and the cyanide will be decom- 
posed. Sulfuric acid is then added to the solution until 
neutrality is reached after which it is acidified with 
hydrochloric acid, stannous chloride added and the deter- 
mination made as in the previous method. 

1 Eng. Min. Jour., 76, 661. 



94 COLOBIMETRIO ANALYSIS. 

Method G. Gold by Decomposition of the 
Cyanide by Ammonia. 

This method is similar to method E except in the 
manner of reduction of the cyanide. To the solution^ to 
be reduced is added one third its bulk of concentrated 
ammonia. Concentrated sulfuric acid is then added until 
neutrality is reached after which the solution is acidified 
with hydrochloric acid and treated with stannous chloride 
for the production of the color. 

Method H. Gold by Metaphenylenediamine. 

The solution of the gold as a cyanide is treated with 
concentrated hydrochloric acid to decompose the cyanide 
and change the gold to the form of the chloride.^ Five 
c.c. of a solution of five grams metaphenylenediamine to 
the liter are then added. Upon the addition of sulfuric 
acid a yellow to dark brown color is formed, dependent 
on the strength of the gold solution. The color shown is 
due tx) the presence of colloidal metallic gold. A solution 
as dilute as .005 percent will show a color by this method. 

The reagent may become pink on standing in the light 
but may be decolorized by animal charcoal. 

1 Eng. Min. Jour., 76, 661. 

2 Chem. Ztg., S6, 934. 



CHAPTEE TWELVE. 

TITANIUM, VANADIUM AND TUNGSTEN. 

These comparatively rare metals are not very widely 
distributed but notwithstanding their scarcity are of 
great importance. Their principal use is in alloys with 
steel where the amount of the metal present would be 
small and well adapted to this type of procedure. 

Method A. Titanium by Hydrogen Peroxide. 

This method of analysis for titanium is ordinarily 
applied to the solution titrated for total iron, after the 
titration has been completed. For use as standard the 
titanium content of a standard ferrotitanium must be 
known. The presence^ of less than half of one percent of 
nickel or chromium may be equalized by the addition of 
a similar amount to the standard, a greater amount than 
this must be removed by methods to be given later. 

The sample, if weighed out, is of .5 gram, the solution 
titrated for iron is just as good and much more con- 
venient. The standard is made up by weighing out an 
amount of ferrotitanium such that its titanium content is 
approximately that of the steel under examination and 
adding to this enuf steel containing no titanium to bring 
the weight up to .5 gram. This standard is then treated 
as specified for the sample. If the sample contains more 
than one half percent nickel or chromium they are re- 
moved as follows : Fuse the .5 gram sample with 5 grams 
sodium carbonate and one gram potassium nitrate. The 

1 Chemical Analysis of Special Steels, Steel Making Alloys and 
Graphites^ Johnson. ^ 

95 



96 COIiOBIMETRIC ANALYSIS. 

melt is heated with 30 c.c. of one-one hydrochloric acid 
in a porcelain casserole for thirty minutes, filtered and 
washed. Thirty c.c. of dilute sulfuric acid are then 
added, the solution evaporated to fumes, transferred to a 
test tube, five c.c. nitric acid added, the red fumes boiled 
off and proceeded with as the solution of the sample. 

The standard, and the sample if it does not have to 
have nickel or chromium removed, is dissolved in ten c.c. 
dilute sulfuric acid, five c.c. concentrated nitric acid 
added and heated until red fumes cease to be given off. 
The standard, and the sample obtained in one of the 
ways mentioned, are then poured into Nessler tubes or 
other tubes suitable for comparison by dilution. Hilde- 
brand^ says that five percent of sulfuric acid is necessary 
in this solution to make sure that any metatitanic acid 
formed is reverted to the metal. If phosphoric acid is 
present a similar amount must be added to the standard.* 
Five C.C. of a solution of hydrogen peroxide are now 
added to each of the tubes and the comparison made by 
dilution. If hydrogen peroxide is not available it may 
be supplied by dissolving 3.5 grams of sodium peroxide 
in 125 C.C. of dilute sulfuric acid and diluting to 500 c.c. 

Since the large amount of iron present is apt to cause 
a yellow color, a correction is sometimes introduced for 
this or the iron removed from the sample before addition 
of the peroxide. The correction for iron present is as 
follows, .1 gram Fe2O3 = .0002 gram TiOg or 10 grams 
Fe2O8 = .02 gram TiOg.^ For the removal of the iron 
the sample is fused with sodium peroxide.* This leaves 
the iron as the insoluble oxide when the melt is extracted. 
The solution is acidified with sulfuric acid, in which case 
no hydrogen peroxide need be added as the excess of 

1 XT. S. Geol. Surv. Bull,, No. 176, p. 80. 

2 J. A. C. S., S9y 481. 



TITANIUM, VANADIUM ANB TUNGSTEN. 97 

sodium peroxide in solution will fonn a sufficient quan- 
tity in reaction with the acid. 

Method B. Titanium by Thymol. 

This method admits of the determination of much 
smaller amounts of titanium, as the color produced by 
titanium in reaction with thymol is about twenty-five 
times as intense as the yellow to orange produced by the 
preceding method. This method produces a red colora- 
tion. The only substance apt to be present which would 
alter the color is tungstic acid. 

The reagent^ is prepared by dissolving one gram thymol 
in 5 c.c. dilute acetic acid and this is then added to 95 c.c. 
of concentrated sulfuric acid. The thvmol if dissolved 
directly in the sulfuric acid would give the solution a yel- 
low color which is not present if solution is made first in 
acetic acid as directed. The solution of the reagent must 
be kept from bright sunlight or it will darken in a few 
hours. The amount of thymol added should be such that 
the ratio of thymol to titanium is less than 100 : 1. 

The sample is placed in solution as in the preceding 
method, using all the same precautions, and then poured 
into Nessler tubes as in that case. Two c.c. of the above 
thymol solution are then added to the sample and stand- 
ard and dilution carried out.^ If desired a series of 
standards may be prepared from a standard ferrotitan- 
ium for this method. For convenience in dissolving the 
original sample it may be desirable to fuse it with potas- 
sium acid sulfate. 

Method A. Vanadium by Hydrogen Peroxide. 

This metal in alloys and ores is determined by solution 
in acid and oxidation with hydrogen i)eroxide. The 

1 J. A. C. S., S5, 138. 

2 Orig. CJom. 8th Inter. Cong. App. Sci., 1, 285. 



98 OOLOBIMETRIG ANALYSIS. " 

resultant color is, in practice, always compared by dilution 
or duplication. If chromium is present an equal amount 
must be introduced into the standard under the same con- 
ditions of temperature, acid concentration, etc. The 
sample is two grams of a steel low in vanadium or one 
gram of a steel with a high vanadium content. If of a 
mineral the weight of sample may be judged by the prob- 
able vanadium content. 

The sample is dissolved in 40 c.c. of concentrated nitric 
acid, one tenth gram permanganate added and the solu- 
tion digested for two minutes. Ammonium bisulfite is 
then added to clarify and the solution boiled to expel 
sulfur dioxide. The solution is then cooled and made up 
to 50C.C.1 (A). 

As an alternate method of solution and treatment of 
the sample it may be dissolved in ten c.c. of sulfuric acid 
(1:3), two c.c. of concentrated nitric acid added, heated 
until the solution clears and fumes disappear, cooled and 
made up to definite volume of 50 c.c.^ (A). 

If the original solution contained chromium the stand- 
ard must also be made up to contain chromium. In that 
case a standard vanadium steel is dissolved in 40 c.c. sul- 
furic acid, an amount of chromium equal to the amount in 
the sample added as potassium dichromate, ten c.c. of 
nitric acid added and the solution heated for ten minutes 
to complete solution and clarify. This standard so pre- 
pared may then be made up to fifty c.c. and is ready for 
comparison by dilution. 

In case chromium is absent and the results are to 
be obtained by dilution an amount of vanadium steel is 
weighed out which contains approximately the same 
weight of vanadium as the sample. This steel is treated 

1 J. Ind. Eng. Chem., 5, 736. 

2 Chem. World, ft, 341. 



TITANIUM, VANADIUM AND TUNGSTEN. 99 

identically the same as the sample and when it reaches 
the point (A) in the process should be diluted to fifty 
C.C. and have the same color as the sample. The color 
present at this point, if any, will be entirely that of some 
metal other than vanadium, if the color of the contami- 
nation of sample and standard differs the determination 
of results will be rendered incorrect. If the results are 
to be obtained by duplication a blank of 45 c.c. of water 
is prepared at this point in the process adding to it as part 
of the total volume the same amount of the same kind of 
acid as is supposed to be present in the sample. Note that 
if nitric acid was used for oxidation the greater part of 
it may have been boiled off as brown fumes before pro- 
ceeding to the next process. 

The solution for determination of the sample by dupli- 
cation consists of .1784 gram of V2O5 dissolved in the 
smallest possible amount of sulfuric acid and diluted to 
one liter. Each c.c. of the solution will then contain 
.0001 gram vanadium. 

To the sample diluted to fifty c.c. and to the standard 
diluted to equal volume or to the blank of 45 c.c. of water 
and acid there is now added one c.c. of the commercial 
3 percent hydrogen peroxide. The color resultant from 
the vanadium present appears and the determination 
is carried out by the method chosen. If commercial hy- 
drogen peroxide is not available it may be made by ad- 
ding two grams of sodium peroxide to 100 c.c. dilute 
sulfuric acid. 

This method of determination of vanadium is accurate 
to about .04 percent.^ 

1 Pickard, Chem. World, £, 341. 



100 OOLOBIMETRIO ANALYSIS. 

Method B. Vanadium by Strychnine. 

By this method the vanadium is separated from the 
solution containing the iron by precipitation with am- 
monium molybdate/ oxidised with potassium chlorate 
and sulfuric acid, treated with a solution of strychnine 
in concentrated sulfuric acid and the results obtained by 
dilution. Ten minutes is necessary for the full color to 
appear after the addition of the last reagent so that deter- 
mination by duplication is impracticable. The prepara- 
tion of the large amount of standard necessary for the 
method by balancing is not satisfactory and the color is 
not permanent enough for the use of a series of standards. 
The sample is chosen as one or two grams as in the pre- 
ceding test. 

The sample is dissolved in 45 c.c. nitric acid (1:2) and 
one c.c. dilute sodium acid phosphate solution added. 
The solution is then cooled and 14 c.c strong ammonia 
added. Boil until all the ferric hydroxide has redis- 
solved in the excess of the solution, remove from the 
flame, add thirty c.c. of ammonium molybdate and shake 
well. Filter and wash free of iron with two percent 
nitric acid and then with water. The precipitate is then 
washed into a large beaker, a small amount of potassium 
chlorate and 20 c.c. sulfuric acid added and evaporated 
until fumes of sulfuric acid begin to pass off. This is 
then cooled and twenty c.c. of a solution of strychnine, 
four grams to the liter, in concentrated sulfuric acid are 
added. After standing for ten minutes the color will 
appear in its full intensity. 

The standard is made by solution of .1784 gram of VgOg 
in as small an amount of sulfuric acid as possible and 
diluted with concentrated sulfuric acid to 100 c.c. Each 
C.C. contains .001 gram vanadium. An amount of this 

1 Chem. World, fS, 341. 



TITANIUM, VANADIUM AND TUNGSTEN. IQl 

standard equal to the amount of vanadium estimated to 
be present in the sample is treated with potassium chlorate 
and sulfuric acid, carried on from that point, and used for 
dilution. 

Tungsten as the Oxide in Colloidal Suspension. 

The sample dissolved in acid is treated with titanium 
trichloride. This frees the blue oxide of tungsten which 
will remain in suspension about one half hour.^ More 
than ten percent of hydrochloric acid in the solution 
weakens the color. The method may not be applied in 
the presence of vanadium, phosphorus and molybdenum 
without first removing these factors. The amount of 
tungsten in the sample must not be over .001 gram, other- 
wise the oxide will precipitate. 

Solution of the sample is made in five c.c. of hydro- 
chloric acid, the solution boiled down to small volume, 
two c.c. are practical, cooled and diluted to forty c.c. 
Five C.C. of TiCla are then added to the solution and dilu- 
tion to 50 c.c. carried out. This is compared with the 
standard by dilution. 

The standard for use by this method is prepared by 
weighing out a standard tungsten steel, of which the 
tungsten content is known, to an amount estimated to be 
near the amount of tungsten contained in the sample and 
adding nontungsten steel until the total weight is equal 
to the weight of sample taken. In case the estimate is 
being made of a mineral rather than of a steel it is 
advisable to use a steel as high in tungsten as possible 
and not to add the additional nontungsten steel as there 
will be no iron in the mineral. This standard is dissolved 
and treated by the same method as the sample. 

1 Compt. Eend., 166, 416. 



CHAPTER THIRTEEN. 

FLUORINE, CHLORINE AND PERCHLORATES. 

Of the members of the halogens these two, fluorine and 
chlorine, may be estimated by colorimetric methods and 
also the compounds of one as perchlorates. The many 
places in which traces of these occur need hardly be 
mentioned, from drinking water to minerals they may be 
found distributed thruout nature, and in synthetic prod- 
ucts the occurrence of chlorine in particular is frequent. 

Fluorine by Estimation of Its Bleaching Action 
ON AN Oxidised Titanium Solution. 

Small amounts of fluorine may be estimated by their 
action on a titanium solution which has been oxidised by 
the addition of hydrogen peroxide.^ The results are ob- 
tained by one of two methods. There may be a similar 
titanium solution containing a known amount of fluorine 
prepared, such that its color is the same, or the bleached 
solution may be examined for its apparent titanium con- 
tent. By subtraction of the apparent content from the 
real content and multiplication by a factor, the fluorine 
which caused the bleaching may be estimated. 

There are a large number of impediments to this 
process. The presence of alkaline sulfates causes a bleach- 
ing action similar to that of fluorine.^ Heating or the 
presence of considerable free acid will serve to bring back 
part of the color thus bleached. The color bleached 
by fluorine as well as that by alkaline sulfates will be 

1 Am. Jour. Sci. (4), S8, 119. 

2 J. A. C. S., 30, 219. 

102 



FLUORINE, CHLORINE AND PERCHLORATES. 103 

restored by a large amount of free acid. If a large 
amount of silica is present, as usually is the case in 
minerals containing fluorine, this must all be- removed 
before the procedure is begun. The color is estimated to 
be increased 5 to 15 percent by raising the temperature 
30 degrees. Phosphoric acid also bleaches the titanium 
solution and so must be absent. Aluminum also bleaches, 
so if present it is removed by precipitation with ammo- 
nium carbonate. Iron prevents the bleaching by fluorine 
and by its color confuses the test, therefore it must be 
removed. A suitable sample contains .01 to .00005 gram 
fluorine. This method is more accurate than gravimetric 
methods for amounts below two percent and will not only 
detect but estimate amounts so small as not to be detected 
by gravimetric methods. 

The sample is fused with three to five grams sodium 
potassium carbonate fusion mixture and the fusion dis- 
solved in hot water as far as possible. Two grams of 
ammonium carbonate are then added and the solution 
evaporated on the water bath to small bulk, the ammo- 
nium carbonate at the same time being decomposed. 
Iron, aluminum and silica are precipitated and may now 
be filtered out. A standard titanium solution is prepared 
by the solution of one gram TiOg in concentrated sulfuric 
acid and diluted to one liter.^ This solution contains 
.001 gram titanium dioxide per c.c. Four c.c. of hy- 

1 Scott gives a somewhat more elaborate method of solution. An 
intimate mixture of one gram of TiOa and three grams of ammonium 
persulfate is heated until the vigorous action has ceased and am- 
monium sulfate is expelled. The residue is treated with twenty c.c. 
strong sulfuric acid, heated to fuming and, when cold, poured into 
about 800 c.c. of cold water. When the suspended salt has dissolved, 
57.5 c.c. of strong sulfuric acid are added, and the solution made up 
to 1000 c.c. Fifty c.c. or more of the solution should be analyzed 
for TiO,. 



104 OOLOBIMETEIC ANALYSIS. 

drogen peroxide are added to the solution to be tested, 
followed by ten c.c. of the above titanium solution and 
five C.C. of sulfuric acid. The addition of the acid should 
render the solution slightly acid. If such is not the case 
a further amount of acid is added until the solution is 
very slightly acid. More acid is now added, dependent 
upon the estimated fluorine content of the solution. For 
the maximum sample, .01 gram, twelve c.c. of acid are 
used, for the minimum sample, .00005 gram, one c.c. will 
be sufficient. The amount used is graded according to the 
amount present between these two extremes. The solu- 
tion is then diluted to 50 to 100 c.c. according to the in- 
tensity of color and the original amount of the solution 
before dilution. 

A standard fluorine solution must be made up if the 
standard is also to be bleached. This consists of 1.2403 
grams potassium zirconium fluoride (potassium fluozir- 
conate, KgZrFe) dissolved in water and diluted to one 
liter. This solution contains .0005 gram fluorine per c.c. 
For determination by this method the fluorine in the 
sample must be estimated, a similar amount added to 
ten C.C. of titanium solution treated with hydrogen perox- 
ide and the agreement with the sample noted. If this 
does not agree it is only a matter of a short procedure to 
make up another which will more nearly approximate the 
sample. 

The desired result may be obtained from the factor 
figured by Steiger^ from a large number of experiments. 
The solution bleached with fluorine may be estimated by 
first determining the apparent titanium present and sub- 
tracting this from the actual weight of titanium known 
to be present. Results show that fluorine will bleach 
slightly more than twice its own weight of titanium, 

1 Steiger, J. A. C. S., SO, 219. 



FLUORINE, CHLORINE AND PERCHLORATES. 105 

accurately 35:75. For general estimation the difference 
between the two titanium readings may therefore be di- 
vided by two and this taken, for accurate work the differ- 
ence between the titanium readings is multiplied by .4667. 
Factors for allowing for the presence of acid and of alkali, 
etc., have been figured but the tables are too long for 
inclusion here. For those records reference should be 
made to Steiger as indicated. The errors due to the 
presence of alkaline sulfates and excess acid are so small 
that they may well be neglected. 

Chlorine by 0-Tolidine. , 

Free chlorine present may be estimated by the green 
color produced by it in reaction with o-tolidine. The 
method^ is especially designed for water analysis but may 
be adapted to other determinations of free chlorine. The 
reagent is so delicate that it will detect .00000005 gram 
of chlorine per c.c. of water, or to phrase it in terms of 
water analysis, .005 part per million. The sample should 
be twenty-five c.c. of a water high in chlorine or fifty c.c. 
of one low in that constituent. 

Five C.C. of a solution of one gram of the reagent in a 
liter of ten percent hydrochloric acid are added. The 
color appears at once. The use of a natural series of 
standards is not advisable as they will not keep from 
day to day but the use of artifical standards made from 
the copper-iron or copper-dichromate series is possible. 
Because of the nature of the test substance the method 
of duplication would be practically impossible to control. 
Dilution or balancing methods may be used, the standard 
should be another solution whose chlorine content has 
been determined, treated with the same proportional 
amount of the reagent as the sample. 

1 J. Ind. Eng. Chem., 5, 915. 
8 



106 COLORIMETBIO ANALYSIS. 

Perchlorates by Methylene Blue. 

The reaction of perchlorates with methylene blue may 
be used for the estimation of the amount of perchlorate 
present.^ This method is used in the test of soluble sub- 
stances such as salts for the presence of these compounds, 
with particular reference to the analysis of Chili salt- 
peter. The sample used is one gram except in rare cases. 
The method has a considerable range. 

The sample is dissolved in water and diluted to 25 c.c. 
Two c.c. of a solution of one gram nitrosodimethylaniline 
in alcohol diluted to one liter with water are then added. 
The preparation of a series of standards for use with this 
is advisable, not permanent standards but temporary 
standards, easily replaced. A standard solution is made 
up of potassium chlorate by the solution of .1468 gram 
in water and dilution to one liter. For use ten c.c. of 
this are diluted to 100 c.c. and will then contain .00001 
gram ClOg per c.c. The non-permanent standards for 
the test may be made up by use of one to ten c.c. of this 
solution. If the tests are all being made on the same 
basic substance as, for example, on Chili saltpeter it is 
well to add that substance in large amount to the standard 
solution before dilution to the line so that the standard 
may contain approximately the same amount as the 
sample solution. It is also possible to add the solution 
of the salt upon which work is being done to the non- 
permanent standards when diluting them to the same 
volume as the sample, before treatment with the reagent. 

It is necessary that the sample and standards be allowed 
to stand for several hours to fully bring out the colors. 
The sample is then compared with the series of standards 
prepared at the same time and an estimation made of its 
perchlorate content. 

1 Arch. Sci. Phys. Nat., i^^ 210. 



CHAPTEB FOURTEEN. 

NITRIC AND NITROUS ACIDS, AND AMMONIA. 

The occurrence of these nitrogen compounds is frequent 
and the determinations made of them are many. The 
test for ammonia with Nessler's reagent is one of the 
oldest of methods for this sort of work. The above sub- 



FlG. 15. Colorimeter with Lamp House Suitable for Night Work. 

stances have to be frequently determined in water sup- 
plies and sewage effluents. The presence of nitric acid 
in sulfuric acid supposed to be chemically pure is also 
often to be detected and estimated by these methods. 
107 



108 OOLOBIMBTEIG ANALYSIS. 

Method A. Nitric Acid by Brucine Eeaction. 

Nitric acid may be determined, colorimetrically, by its 
reaction with brucine in the presence of sulfuric acid, 
making the test by the sulfur yellow that follows the 
initial red coloration rather than by the original colora- 
tion which may not be relied on.^ If nitrous acid is not 
to show in the final results there must be present two 
parts of sulfuric acid for every part of water. This 
method is given on the assumption that only nitric acid 
is to be determined. In case it is desired to determine 
nitrous acid as well as nitric acid, lessening the amount 
of sulfuric acid used, so that there would be present two 
parts of water to one of sulfuric acid, would have the 
effect of causing both to show. In case the solution being 
tested contains much organic matter or ferrous iron these 
must be oxidised by the addition of permanganate until it 
is in slight excess. If this oxidation is carried out it 
must be taken into account that nitrous acid present will 
be oxidised at the same time to nitric acid and that if the 
nitric acid in the presence of nitrous acid is to be deter- 
mined it will be necessary to separately determine the 
nitrous acid, calculate it to the form of nitric acid and 
subtract the result obtained from the total nitric acid 
shown by the test. 

This method is used for determinations on water, 
sewage eflSuents, and sulfuric acid, also for the analysis 
of salts for traces of nitric acid not removed in their 
purification. The sample chosen for the test should con- 
tain .00001 to .0002 gram nitric acid. If a liquid, a quan- 
tity to fall within those limits may be chosen. 

The sample, if a solid, is dissolved in water and diluted 
to not over twenty c.c. The sample, if of a liquid, must 
be under that volume and is diluted to approximately 

1 Chem. Zeit, 23, 454; 25, 586. 



NITEIC AND NITEOUS ACIDS, AMMONIA. 109 

twenty c.c. The determinations are best made in Nessler 
tubes or some type of graduated tubes. One c.c. of a con- 
centrated solution of brucine is added to the sample and 
then 30 c.c. of concentrated sulfuric acid. If the test is 
being made on sulfuric acid for nitric acid as impurity, 
ten C.C. of water and ten c.c. of acid are substituted for 
this addition of acid, leaving the volume somewhat 
smaller than in the other case. Care must be used in the 
addition of this sulfuric acid as it must be carefully 
poured down the side of the tube without causing the 
water to boil. If care is not used the water will boil and 
loss of nitric acid will result. 

The most convenient method for making the test is 
by duplication. The blank of fifteen c.c. of water and 
five c.c. of sulfuric acid is treated as specified for the 
sample. The blank should show no color at this point, 
if it does the solution must be rejected unless it is known 
that the sulfuric acid used contained a small amount of 
impurity in the form of nitrate. In that case the con- 
tamination of sample and blank may be assumed to be the 
same and may be disregarded. The standard solution for 
addition to the blank is made by dissolving .1872 gram 
KNO3 in water and diluting to one liter. This standard 
then contains .0001 gram N2O5 or .000115 gram NO3 or 
.00002594 gram Ng per c.c. The standard is run in and 
the sample duplicated in the usual manner. 

The sample shows first an orange yellow changing to 
light yellow, on the addition of the sulfuric acid. Simi- 
larly the color of the blank will show first orange chang- 
ing to yellow. If the carrying out of the test is delayed 
the solutions may become cooled in which case the change 
of color will be slowed down. In that case it is best to 
heat the solutions over a flame, to have them at sufficient 
heat to show proper color reactions. 



110 COLOEIMETEIG ANALYSIS. 

Method B. Nitric Acid by Diphenylbenzidinb. 

The test for nitrates with diphenylamine is not abso- 
lutely satisfactory as some elements of the test do not 
seem possible to control.^ It has been found by experi- 
ment that diphenylbenzidine is formed in the solution from 
the diphenylamine and that the use of this as the original 
substance gives more satisfactory results. In case di- 
phenylbenzidine is not available diphenylamine may be 
used as a substitute following the same directions as 
those given for the former reagent. The sample chosen 
is to be five c.c. of solution which should contain between 
.000001 and .0000001 gram nitrogen per c.c. The best 
way of handling the test is to make up simultaneously a 
series of nonpermanent standards.^ 

A standard solution may be made up by dissolving 
.3608 gram potassium nitrate in water and making up 
to one liter. Ten c.c. of this solution made up to 100 c.c. 
will then contain .000005 gram nitrogen per c.c. From 
this solution standards are to be made up with .1, .2, .3, 
c.c, etc., of the standard solution diluting each to five c.c, 
similar to the volume of the sample. 

To each of the standards and to the five cc sample 
solution twelve cc of concentrated sulfuric acid are added 
and the solutions set in water to cool. The reagent used 
is .1 gram diphenylbenzidine in 100 cc concentrated sul- 
furic acid. The tubes having been cooled, three c.c. of 
this are added to each of the test tubes and the color of 
the sample compared with that of the standards, after ten 
minutes have been allowed for the color to show. If it is 
desired to obtain the total nitrogen, nitrites may be oxi- 
dised to nitrates by adding permanganate before proceed- 
ing with the determination. For the determination of 

1 Z. Anal. Chem., 56, 28. 

2 Jour. Chem. Soc., 105, 1157. 



NITRIC AND NITROUS ACIDS, AMMONIA. m 

nitrites the total nitrogen and the nitrates may be deter- 
mined and the nitrites present are then represented by the 
difference in the two nitrogen determinations. 

The mixing of the reagent with the solution is accom- 
plished by lightly stirring. Shaking is to be avoided as 
it lessens the color produced. 

Method C. Nitric Acid by Phenolsulfonic Acid. 

The yellow color produced by nitrates is reaction with 
phenolsulfonic acid is found to be a satisfactory method 
of determining the nitrates present in a solution such as, 
for example, drinking water. The phenolsulfonic acid 
for use is prepared^ by heating ten grams of phenol on 
the water bath for two hours with 60 grams concentrated 
sulfuric acid and 30 grams fuming sulfuric acid, or four 
hours with 150 grams concentrated sulfuric acid. The 
sample, if colored, must first be decolorized by the addi- 
tion of aluminum hydroxide, as is usual in water purifica- 
tion. The sample of water or solution selected should 
contain less than .0001 gram nitrogen, preferably less than 
half that amount. 

The sample is evaporated nearly to dryness on the 
water bath and the last few drops of solution then allowed 
to evaporate at room temperature in a place sheltered 
from dust. One c.c. of phenolsulfonic acid is then added 
to the residue and thoroly rubbed over it with a glass rod. 
Ten c.c. of distilled water are now added and stirred until 
all is dissolved. Ammonium or potassium hydroxide is 
then added until the solution is alkaline after which it is 
poured into the tube or bottle in which comparison is 
to be made. 

The standard is made up by dissolving .7217 gram of 
pure potassium nitrate in water and diluting to one liter. 

1 American Committee on Standard Methods of Water Analysis. 



112 COLORIMETEIC ANALYSIS. 

Ten C.C. of this solution are then evaporated the same as 
was specified for the sample, moistened with two e.c. of 
phenolsulfonic acid and diluted to one liter for use as a 
standard. Each c.c. contains .000001 gram nitrogen. 
This standard is suitable for making up a series of stand- 
ards or for the method of duplication. 

For duplication the sample mentioned above is poured 
into a Nessler tube and diluted to 50 c.c. The standard 
is then run into a blank of 25 c.c. and colors and volumes 
equalized. For the preparation of a series of standards 
the following volumes of the dilute standard solution are 
used. The amount of nitrogen represented by each is 
shown opposite. The sample should be diluted to 100 c.c. 
for this method, as should also each of the standards. 
Five c.c. of concentrated ammonia are to be added to the 
solution of standard before it is made up to the final 
volume. 

Vol. of Dil. Stan. Sol. Usecl. Nitrogen Represented in Grama. 

.0 000000 

1.0 000001 

3.0 000003 

5.0 000005 

7.0 000007 

10.0 000010 

15.0 000015 

20.0 000020 

25.0 000025 

30.0 000030 

35.0 000035 

40.0 000040 

In case chlorides are present add two c.c. of the reagent 
to the solution and 1.5 c.c. concentrated sulfuric acid be- 
fore evaporation to small volume.^ The evaporation to 
small volume may then be carried out on the steam bath, 

1 J. Ind. Eng. Chem., 9, 585. 



NITRIC AND NITROUS ACIDS, AMMONIA. 113 

completing evaporation to the final small volume on a 
water bath at not over 70 degrees, after which the method 
is followed out. Chlorides are thus removed without 
serious loss of nitrates. 

Method D. Nitric Acid by Pyrogallol. 

This method is more suited to confirmation of another 
estimate or to quick guesses at the approximate nitrogen 
contents of a sewage effluent or of water than to other use. 

Ten c.c. of the sample are placed in a test tube and ap- 
proximately .2 gram pyrogallol added and thoroly mixed 
with the sample.^ A pipette partially filled with sulfuric 
acid is then inserted, the acid being held in the pipette by 
keeping the finger over the upper end until inserted, and 
two c.c. of sulfuric acid allowed to flow out as a seperate 
layer at the bottom of the tube, the pipette being again 
removed with the finger over the aperture. Upon the 
addition of .1 gram of dry powdered sodium chloride an 
effervescence is set up at the point where the two layers 
join and a purple ring proportional in size and intensity 
to the amount of nitrate present, is formed. 

A standard solution is made up by dissolving .1872 
gram potassium nitrate in water and diluting to one liter. 
This solution contains .0001 gram NgOg per c.c. or .000115 
gram NO3. The eye of the operator using this test soon 
becomes so used to the estimation that he can tell pretty 
accurately from the first glance the amount of nitrate 
present. For confirmation of this he may take that 
amount of standard and another sample and, by carry- 
ing the two tests thru at once, see if they appear to be 
identical. 

1 Report of the Royal Commission on Sewage Disposal, Vol. IV, 
Part V, p. 23. 



114 COLORIMETRIC ANALYSIS. 

Method A. Nitrous acid by Sulfanilic Acid and 

a-Naphthylamine. 

The red coloration which appears after a time when 
acetic acid solutions of sulfanilic acid and a-naphthyla- 
mine are acted on by nitrous acid may be used for the 
determination of the amount of nitrous acid present. In 
the process^ the sulfanilic acid is converted by the nitrous 
acid into the corresponding diazo compound and the 
latter reacts with the a-naphthylamine to form a red azo 
dye. The full color which will result does not appear 
for several hours, often not for days, but if the tempera- 
ture and other conditions of the standard and sample are 
identical the color will appear at the same rate and after 
only a very short time the estimation may be made. 

The standard solution of nitrite is made by preparing 
nitrosyl sulfuric acid. .0493 gram pure sodium nitrite is 
dissolved in 100 c.c. water, and ten c.c. of this solution 
added to 90 c.c. concentrated sulfuric acid. This solution 
contains .00001 gram nitrogen per c.c. in the form of 
nitrite. The solution of sulfanilic acid is prepared by 
warming one gram with 14.7 grams glacial acetic acid 
and 15 c.c. water. A further 270 c.c. of water are then 
added, the solution being kept stirred. Similarly the 
a-naphthylamine is prepared by dissolving .2 gram in 
14.7 grams glacial acetic acid and 25 c.c.^ water, after 
which 300 c.c. of water are added. The solutions are 
mixed for use but only small quantities should be so pre- 
pared at a time. 

For the analysis twenty c.c. of the sample are placed in 
one color comparison tube and eighteen c.c. of pure water 
in the other. Two or three c.c. of a mixture of the above 
reagents are added to each and an amount of the standard 

1 Fowler, Sewage Works Analyses, p. 64. 



NITRIC AND NITROUS AClBS, AMMONIA. US 

added tx) the blank equal to the amount of nitrite esti- 
mated to be present. The colors will be dark enuf after 
five minutes for the operator to see how well they com- 
pare. If they are not identical an experienced operator 
can tell the variation in the amount of nitrous acid be- 
tween the two solutions and may confirm his view by 
adding that amount of standard to a blank and running 
that with a new amount of sample, comparing these after 
five minutes. 

Method B. Nitrous acid by Metaphenylenediamine 

Reaction. 

Metaphenylenediamine reacts with nitrous acid in solu- 
tion to produce triaminoazobenzene (bismark brown), a 
substance of pronounced color.^ Ferric compounds are 
apt to cause trouble but will interfere less if a consider- 
able excess of sulfuric acid is present. The presence of 
organic matter has no effect. The color produced ranges 
from yellow to a yellowish brown. 

The reagent is prepared by dissolving 5 grams of 
metaphenylenediamine in water, adding sulfuric acid at 
once until the liquid is distinctly acid and diluting to one 
liter. If the solution so made up shows color it may be 
decolorized with animal charcoal. The standard solution 
is the same as in the preceding test, .0493 gram pure 
sodium nitrite dissolved in 100 c.c. water and ten c.c. of 
this solution diluted to 100 c.c. with concentrated sul- 
furic acid. The standard contains .00001 gram nitrogen 
per C.C. 

The comparison is made on 100 c.c. of the solution to be 
tested, acidified with two c.c. dilute sulfuric acid, to which 
is added one c.c. of the reagent. A series of standards 
may be prepared, using the nitrosyl sulfuric acid pre- 

1 Ber., 11, 624. 



116 COLORIMETRTG ANALYSIS. 

pared above and diluting the amount used to 100 c.c. 
before the addition of the reagent. The color appears 
quickly so that a duplicate may conveniently be prepared 
from 75 c.c. of water. A standard for dilution and bal- 
ancing methods is prepared by treating ten c.c. of the 
nitrosyl sulfuric acid with two c.c. of the reagent and 
diluting to 100 c.c. The resultant standard contains 
.000001 gram nitrogen per c.c. 

Method C. Nitrous Acid by Zinc Iodide Starch 

Solution. 

The blue color produced by nitrous acid in reaction 
with zinc iodide starch solution is the basis for this 
determination. The test has been estimated to be twenty 
times as delicate^ as the metaphenylenediamine reaction 
(method B). A color will show in seven minutes if 
.00000025 gram nitrous acid is present. 

The starch solution is prepared by boiling five grams 
of starch and 20 grams stannous chloride in 100 c.c. dis- 
tilled water for some hours until the starch has almost 
entirely disappeared. The water is replaced as it evapo- 
rates. To this solution add two grams zinc iodide (Znl2) 
and dilute to one liter. This solution should 'be allowed 
to settle for several weeks, decanting the clear solution as 
needed. For the determination add four c.c. of this solu- 
tion to fifty c.c. of the sample acidified with two c.c. sul- 
furic acid. The blue color develops most rapidly in the 
light. A standard solution of sodium nitrite is made up 
by dissolving .4925 gram pure sodium nitrite in water 
and making it up to one liter. Ten c.c. of this solution 
are then accurately pipetted out and made up to one liter 
giving a standard solution containing .000001 gram nitro- 

1 Analyst, S9, 350. 



NITRIC AND NITROUS ACIDS, AMMONIA. ny 

gen or .000003285 gram NOg or .000004426 gram NO3 per 
c.c. For making up a series of standards this standard 
is used in varying amounts. Because of the fact that the 
color takes some time to appear the use of the duplication 
method is not advisable. The balancing method and di- 
lution method are carried out by using the same standard 
but before the last dilution to one liter fifteen c.c. of the 
reagent are added. This will then show the characteristic 
color and have the same value as the standard above. 

Method D. Nitrous Acid by a-Naphthylamine 

Hydrochloride. 

A very delicate nitrous acid test is obtained by use of 
a-naphthylamine hydrochloride with tartaric acid. The 
reagent may be made up in such a form that it will keep 
indefinitely. The content of nitrous acid of the water 
must be very small, not over .00015 gram per liter or the 
addition of the reagent will cause a precipitate.^ The 
sample may be diluted if the content is too high. 

The reagent for the test consists of one part a-naph- 
thalamine hydrochloride to ten parts sulfuric acid and 
89 parts tartaric acid. A fifty c.c. sample is treated with 
one c.c. of this reagent and the results will show at once. 
If the sample is very dilute it may be well to use 100 c.c. 
A standard nitrite solution is made up by dissolving .4925 
gram pure sodium nitrite in a liter of water. A series 
of standards is not to 'be recommended for as delicate 
a test as this. For duplication the above standard is 
diluted, ten c.c. to 100, and will then contain .00001 gram 
nitrogen per c.c, as nitrite. For factor as nitrate see pre- 
ceding test. As blank for the duplication method use 
48 c.c. water and one c.c. of the reagent. A standard for 

1 Chem. Weekblad, ii, 1115. 



118 COLORIMBTRIO ANALYSIS. 

the use of dilution or balancing methods is made by ad- 
dition of one c.c. of the above standard to 800 c.c. of 
water, adding one c.c. of the reagent and making up to 
one liter. This solution will show the color representing 
.0000001 gram nitrogen as nitrite per c.c. Dihite as this 
is, it is nearly to the upper limit that may be tested by 
this method so from this some idea of the delicacy of the 
test may be gained. 

Method A. Ammonia by Nessler's Beagent. 

In reaction with ammonia Nessler's reagent produces a 
yellow to brown color which is a very accurate indicator 
of the amount of ammonia present in the solution. The 
presence of calcium or magnesium is undesirable as those 
are precipitated with the reagent and the presence of the 
precipitate obscures the results. Rather than precipitate 
and filter it is usual to hold them in solution by adding a 
few c.c. of a solution of 50 grams Rochelle salt in 100 c.c. 
of water. For the preservation of this solution five c.c. 
of Nessler's reagent should be added to it at the time of 
preparation. 

Nessler's reagent, by a recent and highly approved 
formula is prepared as follows,^ dissolve 2.5 grams KI in 
three grams water, add 3.5 grams Hgig and, when this is 
all dissolved, add 100 grams of a 15 percent solution of 
KOH. Allow this to stand and settle and decant. If 
it is necessary to use the reagent at once it should he 
mixed with a little talc, and filtered thru sand if possible. 
The sample used may be of drinking water or sewage 
effluent, or a solution of some substance containing am- 
monia as impurity. 

To 100 c.c. of the sample solution add two or three 
C.C. of the Rochelle salt solution followed by the same 

X Apoth. Ztg., S9, 972. 



NITRIC AND NITROUS ACIDS, AMMONIA. HQ 

quantity of Nessler's reagent. The color appears at once. 
The color is fairly permanent and a series of standards 
may satisfactorily be produced. For this a standard 
solution is prepared by dissolving .3140 gram ammonium 
chloride in water and diluting to one liter. The stand- 
ard then contains .0001 gram ammonia per c.c. For the 
duplication method the blank should be 95 c.c. of water. 
A standard for the dilution or balancing methods is made 
up by adding to 100 c.c. of the above standard, twenty 
c.c. of the Rochelle salt solution and twenty c.c. of Ness- 
ler's reagent and diluting to one liter. This latter stand- 
ard then contains .00001 gram ammonia per c.c. For 
some uses it may be necessary to again dilute this latter 
standard to one tenth of its strength. 

Method B. Ammonia by Phenol. 

The reaction of ammonia with phenol^ gives a method 
of estimation of ammonia more rapid than the test with 
Nessler's reagent and just as delicate but not one admit- 
ting of as grea4; accuracy. For careful work its accu- 
racy is not to be depended on. The time for a complete 
determination should not exceed four minutes. The 
method is applicable to all cases where ammonia is to 
be determined. Calcium is not affected by the test but 
the presence of a considerable amount of free acid spoils 
the test. 

To five c.c. of the solution to be tested add one c.c. of a 
dilute solution of sodium hypochlorite (NaOCl) and one 
c.c. of a 4 percent solution of phenol. This is then 
diluted to ten c.c. The full intensity of the color shows 
on heating in boiling water for two minutes. 

Comparison may be made with a series of standards. 
For their preparation an ammonium chloride solution 

iGaa World (Coking Section), 64, No. 1654, 10. 



120 (X)LORIMBTBia ANALYSIS. 

containing .3140 gram to the liter is prepared. Ten c.c. 
of this solution are diluted to 100 c.c. thus containing 
.00001 gram ammonia per c.c. This standard is also 
suitable for determination of results by the duplication 
method, using a blank of five c.c. of water but diluting to 
only nine c.c. instead of ten before the addition of the 
standard. The application of the duplication method is 
not very satisfactory. Ten c.c. of the original am- 
monium chloride solution treated with five c.c. hypo- 
chlorite solution and five c.c. 4 percent phenol solution are 
diluted to one liter and will then on heating show the 
color. The content of this standard is .000001 gram am- 
monia per c.c, an amount suitable for dilution or balanc- 
ing methods. The method is delicate enuf to detect 
.0000001 gram ammonia in the five c.c. sample. 



CHAPTEE FIFTEEN. 

PHOSPHORUS, SILICA AND BORON. 

These three substances, found occurring naturally in 
the earth, are conveniently determined by colorimetric 
methods in small amounts, somewhat more accurately 
than by gravimetric methods. The first two conflict 
when determined in the same solution so that it has been 
found advisable to work out a new method whereby, by 
introduction of the proper corrections, both may be deter- 
mined in the same solution, each in the presence of the 
other. Boron is to be determined in the form of boric 
acid in which form it occurs to some extent as a preserva- 
tive in food. 

Method A. Phosphorus as the Phosphomolybdate. 

Since the gravimetric method of analysis for phos- 
phorus by precipitation of the phosphomolybdate is very 
inaccurate for small amounts of phosphorus, the solution 
of these small amounts of phosphomolybdate may be 
tested colorimetrically.^ Silica must be removed or the 
results will be too high, for which removal a method is 
given here. The silica forms a silioomolybdate which 
has a color similar to that of the phosphomolybdate. A 
method has also been worked out for the determination 
of the two in the presence of each other. For further 
details of that method see Silica, this same chapter. 

Ten to 100 c.c. of solution which must contain less than 
.0005 gram phosphorus are evaporated to dryness. When 
the solution is partially dry add three c.c. nitric acid. 

1 J. A. C. S., 26, 975. 
9 121 



122 COIiOBIMETBlO ANALYSIS. 

The residue is then heated at 100 degrees for two hours. 
If the substance being tested is such that it contains no 
silica this dehydration may be omitted. The residue 
after dehydration is dissolved in water. It will not be 
necessary to filter as the loss due to filtration would be 
greater than that possible due to the presence of a small 
amount of precipitate of silica in the solution. Two 
c.c. of nitric acid are added and four c.c. of a solution of 
50 grams of ammonium molybdate in a liter of water. 
After three minutes the color is suitable for comparison. 
As a standard dissolve .5043 gram of pure crystallized 
sodium acid phosphate (NajHPO^, I2H2O) in water, 
add 100 C.C. nitric acid and dilute to one liter. This 
solution contains .0001 gram P2O5 per c.c. A series of 
standards prepared from this must be replaced daily as 
the darker solutions precipitate out after a time and the 
lighter solutions fade. The fact that the color is not 
complete for some minutes renders duplication incon- 
venient. For use with the dilution or balancing methods 
ten c.c. of the standard are diluted with 75 c.c. of water, 
five c.c. of the molybdate solution added and the solution 
made up to 100 c.c. This standard will contain .00001 
gram PgOg per c.c. 

Method B. Phosphorus, Separated by Precipitation 

AS Magnesium Phosphate. 

This method is similar to that given for magnesium. 
The phosphate^ is precipitated as magnesium phosphate 
by the addition of an excess of magnesium chloride, the 
precipitate is filtered off, dissolved in nitric acid and 
tested colorimetrically by the addition of a solution of 
ammonium molybdate. The advantages of this method 
are that all silica is removed from the solution being 

1 J. A. C. S., fS6, 1463, 



PHOSPHORUS, SILICA AND BOBON. 123 

tested, thus preventing errors due to the silicomolybdate, 
and that the organic matter is removed from the final 
solution. The interference of calcium is prevented as in 
the magnesium test by the addition of ammonium 
oxalate. Experiments have shown that satisfactory 
samples should range between .004 and .0005 gram phos- 
phorus. Results obtained in the presence of various dis- 
turbing factors were uniformly satisfactory. 

One drop of ammonia and two or three drops of am- 
monium oxalate are added to the sample solution after 
which it is evaporated to dryness on the water bath. A 
solution is made up of 13 grams magnesium chloride and 
twenty grams ammonium chloride in a liter of water 
which includes fifty c.c. strong ammonia. One c.c. of 
this solution is added to the dried precipitate and the 
two well mixed with a glass rod. The precipitate is then 
washed out onto the filter with five c.c. dilute ammonia 
and the dish and filter washed with further successive 
portions until the volume of the filtrate reaches 50 c.c. 
The dish and filter are next washed with five c.c. pure 
water and a new beaker substituted to catch further 
washings. Five c.c. of concentrated nitric acid are used 
to dissolve any precipitate remaining on the dish and the 
precipitate on the filter carefully dissolved with this same 
acid. Dish and filter are washed with successive five c.c. 
portions of hot water until the filtrate amounts to 45 c.c. 
Four c.c. of a solution of 50 grams ammonium molybdate 
in a liter of water are then added and the comparison 
made after twenty minutes. 

The standard is made up by solution of .5043 grams 
sodium acid phosphate (Na2HP045 I2H2O) in water 
with the addition of 100 c.c. nitric acid, and dilution to 
one liter. A series of standards may be prepared from 
this solution, which contains .0001 gram P2O5 per cc.c. A 



124 OOLOBIMBTRIO ANALYSIS. 

duplicate cannot well be made up since the color is so 
long in coming. A standard for dilution or balancing 
methods is made up by adding to ten c.c. of the above 
standard, 70 c.c. water, 9 c.c. nitric acid and 8 c.c. am- 
monium molybdate solution and making up to 100 c.c. 
This standard which contains .00001 gram PgOg (per c.c.) 
must stand for twenty minutes before use in order to bring 
out the color. 

Silica by Ammonium Molybdate in the Presence of 

Phosphorus. 

In reaction with ammonium molybdate silica produces 
a compound, ammonium silicomolybdate, which has a 
color similar to that of ammonium phosphomolybdate, 
a greenish yellow. When silica is to be tested for in the 
solution and it is known that no phosphorus is present 
the test is carried out identically the same as in the test 
for phosphorus. Method A. The color caused by the 
silica diflPers from that caused by the phosphorus in that 
it reaches its full intensity only after some hours, whereas 
phosphorus shows its full color after twenty minutes. 
Another property of the silica coloration, upon which this 
test is based is that its intensity at the end of an hour 
and twenty minutes is twice that at the end of twenty 
minutes.^ 

The samples are treated as in the preceding Method A 
up to the addition of the ammonium molybdate. Two 
samples of the same substance or solution to be tested 
must be so prepared at the same time. Also a standard 
must be available which contains only phosphorus. One 
sample and the standard are acidified and allowed to 
stand for twenty minutes and tested against each other. 
At the time these are acidified and the ammonium molyb- 
date added a similar amount of ammonium molybdate is 

1 J. A. C. S., S6, 975. 



PHOSPHORUS, SILICA AND BORON. 125 

added to the other sample but no nitric acid. After one 
hour has elapsed this second sample is acidified with nitric 
acid and allowed to stand for twenty minutes. It is then 
tested against the same standard as the other sample. 
The reading of the first sample is called the A reading 
and the reading of the second sample the B reading. 
From the law given above the phosphorus can then be 
figured. 

If B reading is twice A reading there is no phosphorus 
present but only silica. If B reading is not twice A 
reading the difference between them is subtracted from A 
reading and the remainder is phosphorus. To repeat, 
the silica reading at the end of one hour and twenty 
minutes from the time the ammonium molybdate as 
added and the end of twenty minutes from the time the 
nitric acid was added to it represents the same amount 
of phosphorus as the reading at the end of twenty 
minutes by the first sample and in addition a reading for 
silica twice that given at the end of twenty minutes by 
the first sample. 

The phosphorus having been determined by this method 
the silica may be determined by the same method taking 
no account of the phosphorus and making the test 
against a standard containing only silica. This silica 
determination may then be corrected by allowing for the 
phosphorus previously determined. This silica standard 
may be prepared at the same time as the original samples 
and the same solution used, as the first sample may be 
kept for four to five hours and tested against a standard 
prepared a similar time before. The test is by dilution 
only. 

If a solution of known silica content is available it is 
best to use that. If such is not available .3657 gram 
potassium silioo fluoride (potassium fluosilicate KaSiFg) 



126 (X)IX>BIMBTRia ANALYSIS. 

is dissolved in hydrochloric acid using as little acid as 
posible, and diluted to one liter. This standard contains 
.0001 gram SiOg per c.c. Such quantities of it may be 
used as are estimated to be necessary to counterbalance 
the contents of the solution being tested. 

Method A. Boric Acid by Curcumin. 

The red coloration produced by boric acid when treated 
with oxalic acid and curcumin is especially adapted to 
the estimation of the quantity of boric acid present in food 
as in adulterant. The advantage is not in the speed but 
in the greater accuracy of the method over the gravimetric 
method. The weight of B2O3 present may be as much as 
.002 gram. 

The sample^ placed in a platinum or porcelain casserol 
is made strongly alkaline and evaporated to dryness. 
The residue is acidified with hydrochloric acid and ex- 
tracted with several successive portions of hot water. 
The extracts are filtered and the filter transferred to a 
dish, made alkaline with barium hydroxide and burned 
to ash. This ash is dissolved in hydrochloric acid and 
added to the original filtrate, and the volume of the filtrate 
made up to 100 c.c. Ten c.c. of this solution are then 
mixed with ten to fifteen grams purified sand in a porce- 
lain dish and made alkaline with barium hydroxide. 
Evaporate to dryness on a paraffine bath, add enough 
dilute hydrochloric acid to make the contents of the dish 
acid, two c.c. saturated solution of oxalic acid and two c.c. 
of an alcoholic solution of curcumin, one gram to the liter. 
Mix the contents of the dish well and cover with a glass 
funnel whose stem is connected to a set of potash bulbs 
containing barium hydroxide. The other end of the set 
of potash bulbs is connected to an aspirator, the dish 
placed on a paraffine bath, the potash bulbs set in water 

1 C. N., 87, 27. 



PHOSPHORUS, SILICA AND BORON. 127 

and air passed thru by suction until the mass is dry. One 
c.c. of the curcumin solution is added to the dish and it is 
again dried. 

Extract the contents of the dish with several portions 
of alcohol and add the contents of the bulbs to the residue 
in the dish. This is dried, made acid with hydrochloric 
acid, treated further with reagents as was the original 
substance, and dried. The alcoholic extract from this is 
added to the previous extract and the color is thus 
obtained. 

A standard solution of boric acid is made up by solu- 
tion of .2886 gram sodium tetraborate (borax NagB^O^, 
lOHgO) and contains .0001 gram BgOg per c.c. Ten c.c. 
of this solution are treated in the same manner as the 
sample and analysis made by dilution with alcohol. It is 
obvious that the series of standards or duplication methods 
would be out of the question. If one tube apparently has 
more of an orange tint than the other that is due to the 
presence of an excess of curcumin and a drop or two 
should be added to the other tube until the colorations are 
similar, after which dilution may be more accurately car- 
ried out. 

Method B. Boric Aero by Turmeric Paper. 

The amount of boric acid in a sample may be esti- 
mated by treatment with methyl alcohol, distillation of 
the resultant methyl borate into alkali and, after further 
treatment, comparing the stain made on a standard tur- 
meric paper with those made by known quantities of 
boric acid. The sample is chosen according to the 
amount of boric acid suspected. If the matter is of 
animal origin ten grams are ordinarily used, if of vege- 
table origin one gram is usually suflScient. 

The sample^ is heated in platinum until the organic 

1 Oompt. Rend., 157^ 1433. 



128 COLOBIMETBia ANALYSIS. 

matter is all removed. Then treat the ash with five to 
ten c.c. of phosphoric acid and rinse into a flask. Wash 
out the dish with twenty c.c. methyl alcohol and add to 
the solution in the flask. Distill on the steam bath re- 
ceiving the distillate in a platinum crucible containing a 
few drops of normal sodium carbonate solution. Add 
ten c.c. more of methyl alcohol to the flask and distill 
this over into the original distillate. This should carry 
over with it the last traces of methyl borate. Evaporate 
the distillate to dryness, cool and add four drops hydro- 
chloric acid and one half c.c. of water. Wash this into a 
small vial and dilute to one and a half c.c. 

A turmeric paper is made by soaking in a solution of 
turmeric and drying, after squeezing out the excess of 
the solution. It is essential that the papers be of uni- 
form thickness, width and saturation with the reagent. A 
good grade of drafting paper is very satisfactory. The 
width may be made such that it readily admits of their 
introduction into the vial used for the determination. A 
strip of this paper is immersed in the vial to a depth of 
15 mm. and the vial exposed to a heat of 35 degrees for 
three hours. If time is not a consideration the vials may 
be allowed to stand at room temperature for 24 hours 
after which the results will be the same as by heating. 
The intensity of the red coloration produced on the 
paper is a measure of the amount of boric acid present in 
the sample. 

The same standard as in the previous test, .2886 gram 
sodium tetraborate (borax, NagB^Oy, lOHgO) to a liter 
of water, is used for the preparation of the standards. 
This standard contains .0001 gram BgOg per c.c. The 
color of the standard papers so prepared is permanent 
and by protecting them from dust they may be kept 
several months. 



CHAPTER SIXTEEN. 

OXYGEN AND HYDEOGEN PEROXIDE. 

The determination of the dissolved oxygen in drink- 
ing water and in sewage effluents is of vital importance 
as it is also in the preliminary examination of streams 
proposed as sources of sewage disposal. The use of hy- 
drogen peroxide as an antiseptic and as a reagent requires 
that its strength be known. 

Method A. Oxygen by Cuprous Chloride. 

The properties of the copper chlorides in ammoniacal 
solution, cuprous chloride being colorless and cupric chlo- 
ride a deep blue, may be made use of for the determination 
of oxygen in a solution.^ The addition of cuprous chlo- 
ride to the water will result in a blue color due to its 
oxidation to cupric chloride which color will be propor- 
tional to the amount of available oxygen in the solution 
tested. If the sample contained much calcium this would 
become turbid on the addition of the ammonia to make 
the solution alkaline, therefore two or three c.c. of hot 
saturated solution of ammonium chloride are added to 
the sample before the addition of the other reagents. If 
the test is being made on a sewage effluent or other solu- 
tion which is slightly colored, a similar yellow may be 
added to the standard in the form of some yellow dye 
such as paranitrophenol. The error which will be intro- 
duced by the original solution being slightly yellow is not 
great. 

1 J. Soc. Chem. Ind., SO, 1071. 

129 



130 COLORIMETRIO ANALYSIS. 

The . sample is gathered in a bottle and the bottle 
tightly corked, taking care that there is not an air bubble 
left below the cork. A special apparatus is used to 
transfer this to the tube in which the test is to be made. 
A double-hole stopper is prepared to fit the sample 
bottle, with two tubes, one of which runs to the bottom 
of the bottle and one of which only enters the cork. A 
rubber tube is attached to the shorter tube. For the trans- 
fer of a sample from the bottle to the tube in which it 
is to be tested the rubber tube, which should be about a 
foot long, is filled half full with kerosene oil. The size 
of tubing used should not be over five to eight mm. 
internal diameter. The double hole stopper is then care- 
fully fitted to the bottle of sample without introducing 
air. The liquids are next allowed to flow out the rubber 
tube. The oil will flow ahead if the tube is carefully 
managed and thus sweep out all air and prevent contact 
of the sample with the air. In the tube in which com- 
parison is to be made the oil will rise to the surface and 
the sample then be run in under the layer of oil. The 
long tube added to the bottle serves as an air inlet, that 
the contents may run out. A small amount of the oil 
will always be introduced into the bottle in the process 
of fitting the stopper, which will rise to the upper sur- 
face of the liquid. The bottle is to be inverted as soon as 
satisfactory connections have been made and the oil will 
then form a layer at the upper, formerly the bottom, 
part of the bottle. The long tube used will let in air 
above this layer so that the sample has no chance to ab- 
sorb oxygen from it. 

The reagent for the test is prepared by warming a 
solution of cupric chloride with some scrap copper and 
pouring into water. The white precipitate which forms 
is filtered and washed, first with hot water, then with 



OXYGEN AND HYDROGEN PEROXIDE. 131 

alcohol and then with ether. The powder is then dried 
and must be pure white. For use a small amount of the 
powder is dissolved in concentrated hydrochloric acid. 
This is added to the sample by placing the solution of 
the reagent in a tap funnel and inserting the tip of the 
tap funnel in the oil layer before the reagent is allowed 
to run into the solution. In that way the introduction 
of air is prevented. 

The test is performed by dilution. A standard water 
for use is made by shaking pure water in a bottle until 
the air bubbles remain in the liquid and then allowing 
this to stand until the air bubbles vanish. A table may 
then be consulted to find the amount of oxygen necessary 
for saturation of a water solution at the given tempera- 
ture.^ 

Temp. Soluble in One Liter. 

5.2*» 8.&56 C.C. 

5.65** 8.744 cc. 

14.78« 7.080 cc. 

24.80'^ 5.762 cc. 

30.00** 2.610 cc 

As a standard one part of this is to be used to three 
parts of boiled water which has shown no oxygen under 
test. The mixing is to be done in the colorimetric tube 
in which the test is to be made and under oil. To each, 
sample and standard, there are added a few cc. of the 
solution of cuprous chloride dissolved in hydrochloric acid 
and immediately after that a few cc. of ammonia. The 
white precipitate which appeared on the addition of the 
cuprous chloride must entirely disappear, otherwise too 
much of the copper solution has been added and a yellow 
interference will result. The anmionia brings out the 
blue colors of the solutions. The boiled water contain- 

iPor further solubilities see Seidell, Solubilities of Organic and 
Inorganic Compounds. 



132 COLOBIMETBIO ANALYSIS. 

ing no oxygen is then used for dilution of the darker 
until results are identical. The readings taken are those 
at the top of the aqueous solution, below the level of the 
oil. Since oil will transmit oxygen after a few hours 
it is necessary that this test be carried along to com- 
pletion in a reasonable length of time. 

If it is desired to make up permanent standards this 
is accomplished by dissolving three grams of copper wire 
in hydrochloric acid and evaporating to dryness. More 
acid is added and this repeated several times. The white 
residue is now dissolved in water, ammonia and hydro- 
chloric acid added until the solution shows a deep blue 
color and the solution smells strongly of ammonia, and 
dilution to one liter carried out. Permanent standards 
-may be made by addition to 100 c.c. of water of the 
following amounts of the standard made up. 

Represents in c.c. 
Standard per 100 c.c. Oxygen per Liter. 

.36 1 

.72 2 

1.08 3 

1.44 4 

1.80 5 

2.16 6 

2.52 7 

2.88 8 

3.24 9 

Method B. Oxygen by Adurol. 

The reagent for this test is adurol with ammonium 
chloride or a mixture of dry powdered borax and adurol. 
The method is not so accurate as the preceding because, 
from the conditions of the method, as great care as to 
accuracy cannot be used. 

The sample of 50 or 100 c.c. water and the standard, 
prepared by saturation of a volume of water, are placed 
in bottles.^ A pinch of adurol is added to each and .5 

1 Z. Angew. Chem., fS4, 341. 



OXYGEN AND HYDROGEN PEROXIDE. 133 

C.C. ammonia containing about 20 percent ammonium 
chloride in solution. This solution sinks to the bottom. 
The standard may be varied by the use of varying 
amounts of boiled water with the saturated aqueous solu- 
tion. To prevent the absorption of oxygen from the air 
in the bottle, before shaking at this point, it is well to 
allow carbon dioxide to flow into the bottles and thus 
displace the air which is present over the liquids. The 
bottles are then corked and shaken. The standard must 
be very nearly the same in color as the sample so that a 
guess may be made as to the content from the experience 
of the operator. Dilution is not satisfactory. 

Borax heated for some hours at 50 degrees and mixed 
six parts of borax, one part adurol and three parts 
Eochelle salt furnishes a convenient reagent.^ This is 
added instead of the adurol at the point in the test. The 
function of the Eochelle salt is to prevent the precipita- 
tion of calcium. 

Hydrogen Peroxide by Its Oxidising Action on 

Ferrous Iron. 

For the analysis of a solution of hydrogen peroxide the 
sample solution is allowed to act on a solution of ferrous 
iron and the amount of the iron oxidised is tested by the 
addition of potassium sulfocyanate as in the iron test. 
The comparison may be made by any method. 

Ten grams of ferrous ammonium sulfate [FeSO^ 
(NH4)2S04, 6H2O] are dissolved in 100 c.c. of water. A 
test should be made on this with potassium sulfocyanate 
to make sure that all iron is in the ferrous condition. If 
not it must be reduced with hydrogen sulfid or by acidi- 
fying with sulfuric acid and adding zinc. This reagent 
must finally be available in a freshly reduced condition 

1 Ihid,, 26y 134. 



134 COLOBIMBTRIG ANALYSIS. 

free from reducing agents. If not already acidified it 
must be made acid for use. The test is made by the addi- 
tion of one c.c. of this solution to a sample of ten c.c. of 
the solution to be tested and then the addition of one c.c. 
of a dilute solution of potassium sulfocyanate. The color 
produced may be compared with a series of standards. 
For these a solution of ferric iron and potassium sulfo- 
cyanate is made up and diluted with water until it is of 
the desired concentration for the varying standards. 
These fade after a time and must be renewed. A blank 
of nine c.c. of water may be used and hydrogen peroxide 
from a known solution added to it. The difficulty with 
that method is the obtaining of a definitely known solu- 
tion. The commercial solution is approximately 3 per- 
cent when fresh. Similarly a standard for the dilution 
or balancing methods is made up by the addition of fer- 
rous iron and potassium sulfocyanate to 25 c.c. of a solu- 
tion of known peroxide content and suitable dilution. 



CHAPTER SEVENTEEN. 

SULFUR, HYDROGEN SULFID AND SELENIOUS ACID. 

The methods for the determination of sulfur given here 
are those for the determination of sulfur in iron and steel 
and in similar substances from which it is possible to 
release it as hydrogen sulfid by the action of an acid. 
The method for hydrogen sulfid is for the analysis of a 
liquid containing dissolved hydrogen sulfid, as for ex- 
ample natural waters in some localities. A method is 
also given for the analysis for selenious acid, an acid of 
the rare element selenium which occurs in the presence 
of sulfur and has properties similar to those of sulfur. 

Method A. Sulfur by Paraphenylenedimethyl- 

DIAMINE. 

The sulfur in iron may be determined by the intense 
blue color produced by its reaction with paraphenylene- 
dimethyldiamine^ (unsymmetrical). The color is so in- 
tense that it is necessary to dilute the sample to a con- 
siderable extent or to use a very small sample. The 
former method is more satisfactory. 

A convenient sample is two grams. This is treated 
with hydrochloric acid and the i-esultant hydrogen sulfid 
absorbed in caustic soda, drawing the air from the flask 
where decomposition is carried on thru a potash bulb, 
by suction. This caustic soda solution is then diluted to 
100 c.c. and ten c.c. pipetted out as a sample. To the 
sample add 1.5 c.c. dilute sulfuric acid and dilute to 50 c.c. 

1 Columbia University, School of Mines Quarterly, fSS, 24. 

135 



136 COLORIMBTRIG ANALYSIS. 

To this 50 c.c. of solution add .1 c.c. of a two percent 
solution of the reagent above and finally a drop of a five 
percent solution of ferric chloride. The color results im- 
mediately on the mixing of these reagents and compari- 
son of the color may then be made by any of the standard 
means. 

As a standard dissolve .2435 gram sodium sulfid (NagS) 
in a liter of water. For making up a series of standards 
five c.c. of this solution are to be diluted to 100 c.c, at 
which concentration the sulfur content will be .000005 
gram per c.c. A series of standards from a solution of 
this concentration would use one to 25 c.c. for the vary- 
ing concentrations. As a basis for duplication use the 
same standard and a blank of twenty-five c.c. of water 
to which have been added the same reagents as to the 
sample. Standards do not keep very well. For dilution 
or balancing treat five c.c. of the standard sohition first 
proposed with double the quantities of reagents specified 
for the sample and dilute to 500 c.c. The standard will 
then contain .000001 gram sulfur per c.c. 

The dilution and aliquot parts of the sample specified 
allow for the testing of one tenth of the sample chosen. 
Simple multiplication of the result obtained, by ten, will 
then give the weight of sulfur in the sample and division 
of the weight of sulfur by the weight of sample will give 
the percent of sulfur. Others of the diamines may be 
used for this reaction as for example p-phenylene- 
diamine. 

Method B. Sltlfur by the Action of Hydrogen SuLrro 

ON Arsenious Oxide Paper. 

The comparison of stains made by known amounts of 
sulfur on arsenious oxide paper with those made by a 
sample of steel or similar substance furnishes a method 



SULFUR, HYDROGEN SUMTD, SELENIOUS ACID. 137 

for the rapid estimation of the sulfur content of a 
sample.^ The standards do not keep well and it is there- 
fore necessary to prepare temporary standards at the 
time the test is made. 

The prepared paper is made up by soaking a good 
grade of drafting paper in a solution of 10 grams AsgOg 
in 30 C.C. hydrochloric acid and 970 c.c. water. The 
paper is soaked, the excess of the reagent removed and 
the paper carefully dried. It is then to be cut into 
squares ten cm. on a side. The sample is placed in a 
glass tumbler and the varying standard amounts of a 
steel of known sulfur content are placed in other tum- 
blers of the same size. Each tumbler is covered with a 
piece of the prepared paper, a piece of felt and a weight 
to hold the paper tightly to the top of the tumbler. To 
each tumbler there are then added, 10 c.c. benzine, sp. gr. 
.710, and 50 c.c. hydrochloric acid sp. gr. 1.1. The stain 
on the prepared paper produced by the sample may then 
be compared with that produced by the standard amounts 
and from that, estim^^tion made. For the analysis of 
several samples at one time it is possible to run them 
simultaneously and compare with the same standards. 

Hydrogen Sulth) by Methyi^ene Blue. 

This method is an adaptation of Method A for sulfur 
in iron and steel to the analysis of natural waters and 
other solutions containing a trace of hydrogen sulfid.* As 
the color produced is very permanent, standards may be 
prepared and kept several weeks before renewal. The ad- 
dition of the reagents in the order specified is imperative 
as the order in which they are added will make a diflfer- 
ence in the tint of the solution produced. Satisfactory 

1 Iron Age, 9S, 1253. 

2 Z. Anorg. Chem., 86, 143. 



138 COLOBIMETRIO ANALYSIS. 

results may be obtained from solutions containing as 
little as .00001 gram hydrogen sulfid per liter. 

The sample is five hundred c.c. of the solution. For 
contents of more than .001 gram hydrogen sulfid per liter 
add ten c.c. concentrated hydrochloric acid, .025 gram 
p-phenylenedimethyldiaminesulfate and 2.5 c.c. tenth 
normal ferric chloride in hydrochloric acid solution. If 
the content of the sample is less than .001 gram per liter 
add ten c.c. hydrochloric acid, .01 gram of the diamine- 
sulfate and one c.c. tenth normal ferric chloride. The 
solutions must stand several hours before comparison is 
made. 

For the preparation of the standard a solution of hy- 
drogen sulfid in water must be determined by the usual 
iodometric method or some other similar way. Known 
quantities of this may then be treated as the samples 
were for their production of color. 

Selenium as Selenious Acid by Potassium Iodide. 

The amount of selenious acid present in a solution may 
be estimated from the intensity of the coloration pro- 
duced in reaction with potassium iodide. The method is 
delicate enuf to show .000001 gram per c.c.^ 

The sample, if fairly concentrated, is five c.c. If the 
sample is very dilute, more may be used. The sample is 
diluted to 70 c.c, one drop gum arabic solution and five 
drops 5 percent hydrochloric acid added, the whole mixed 
and diluted to 99 c.c. Add one c.c. potassium iodide solu- 
tion, allow to stand for five minutes and compare with a 
standard by dilution. The potassium iodide solution 
must be colorless. 

A standard is prepared by the solution of .1632 gram 
crystalline selenious acid (HgSeOg) in water and dilution 

1 Z. Anal. Chem., 55, 29. 



SULFUR, HYDROGEN SULFID, SELENIOUS ACID. 139 

to one liter. This standard contains .0001 gram selenium 
per c.c. as selenious acid.. If it is desired to obtain the 
results in terms of selenious acid the standard should be 
made by solution of .1 gram selenious acid when the same 
weight per c.c. would be present of selenious acid instead 
of selenium. One c.c. of this standard is diluted, treated 
with the same reagents as the sample and results obtained 
by dilution. 



CHAPTER EIGHTEEN. 

SALICYLIC ACID AND CYANIDES. 

These two substances are of very infrequent occur- 
rence. The first, salicylic acid, finds a limited use as an 
antiseptic and is used as a cheap wintergreen flavor. The 
method given for cyanides was developed for the deter- 
mination of the cyanides present in poisonous plants sus- 
pected to derive their poisonous nature from that source 
but is applicable to all determinations of small amounts 
of the simple cyanides. 

Salicylic Acid by Fehling's Solution. 

This method is the opposite application of Method C 
for copper.^ In this method the salicylic acid solution 
is treated with Fehling's solution and compared with a 
series of standards prepared at the same time. The 
determination is prevented by the presence of free min- 
eral acid, citric and tartaric acids and any considerable 
amount of iron. The presence of .000001 gram per c.c. 
may be detected in a ten c.c. sample and correspondingly 
smaller amounts by suitable concentration. The sample 
may contain up to .0001 gram in a ten c.c. sample but if 
its content is over .00001 gram per c.c. the sample solu- 
tion of ten c.c. should be diluted to 100 c.c. and ten c.c. 
taken for analysis. 

A standard salicylic acid solution is made up by solu- 
tion of .1 gram of the substance in water and made up 
to one liter. By the solution of .1160 gram sodium sali- 

1 Zeit. Untersuch Nahr. Gfenussm., 22, 727. 

140 



SALICYLIC ACID AND CYANIDES. 141 

cylate the same strength solution as above may be ob- 
tained. Each contains .0001 gram salicylic acid per 
c.c. Temporary standards are prepared by the use of 
0, .2, .4, .6, .8, and 1.0 c.c. of this solution in plain test 
tubes. The use of graduated tubes for the sample and 
standards is not necessary in this case. The sample is 
concentrated to two or three c.c. and placed in a similar 
tube. To each tube of standard and to the sample there 
are now added two c.c. of a solution of one part Fehling's 
solution to ten parts water, five drops of two percent 
potassium nitrate solution, five drops ten percent acetic 
acid and enough water to bring the volume of each up to 
about five c.c. The volumes must be the same of each. 
The tubes are then heated in boiling water for 45 minutes 
and are ready for comparison. The results must be ob- 
tained by inspection of the temporary standards only, as 
dilution of the samples does not prove satisfactory. 

Fehling's solution is made by mixture of equal volumes 
of the following two reagents just before using, 34.64 
grams pure copper sulfate in 500 c.c. of water, and 60 
grams caustic soda and 173 grams Kochelle salt (sodium 
potassium tartrate, NaKC4H40e, 4H2O) in 500 c.c. 
water. The reagents keep perfectly unmixed but the 
solution must be used immediately after mixing. 

Cyanides by Changing to Sulfocyanates and Coloring 

WITH Iron. 

If the solution is an inorganic one it is only neces- 
sary to render the solution slightly alkaline with caustic 
soda. In the solution of a salt or solid substance to be 
analysed for cyanides the solution should be kept alka- 
line to prevent loss. The preparation of a solution of 
an organic nature is somewhat more complicated.^ The 

1 J. A. C. S., 55, 1624. 



142 coLOHnfETRia analysis. 

substance to be tested is macerated and digested with 100 
C.C. of wat^r for some time. The extract is then filtered 
off and the solid residue washed with another 100 c.c. of 
water. This solution is to be placed in a flask, acidified 
with hydrochloric acid and three fourths of the volume 
distilled over into fifty c.c. of potassium hydroxide. This 
solution is then diluted to a convenient volume and fifty 
c.c. taken for the test. 

To the sample of 50 c.c. of solution is added one c.c. of 
ammonium sulfid and the whole evaporated to dryness on 
the water bath. The residue is taken up with ten to fif- 
teen c.c. hot water and slightly acidified with hydrochloric 
acid. Filter off the sulfur, add one half c.c. hydro- 
chloric acid and boil for five minutes. The sulfur is now 
filtered off and the operation repeated, until the final 
solution comes clear. An alternate method is to use five 
c.c. of a solution of 40 grams potassium sulfid in one liter 
of water, instead of the one c.c. of ammonium sulfid. 
Less trouble will be experienced with sulfur if this is 
used. The clear solution from the previous operation is 
diluted to fifty c.c. and fifteen drops of a five percent 
ferric chloride solution added. 

If the iron is precipitated on addition of this ferric 
solution it indicates that the solution was alkaline and 
must be acidified more. A lemon yellow rather than a 
red color indicates that the solution is too strongly acid. 
The red color produced is a result of the cyanide which 
has been changed to the form of the sulfocyanate. The 
standard is prepared directly from the sulfocyanate since 
the relation between the two compounds is known. A 
standard solution is made up by solution of 1.4923 grams 
potassium sulfocyanate in water and dilution to one liter. 
Each C.C. of this solution contains .001 gram KCN in 
which terms the results of this test will be stated. Per- 



SAXIGYLIC ACID AND CYANIDES. 143 

manent standards may be made up but possess the same 
properties as the results of the iron-sulfocyanate reaction 
always do, that they fade in the light after a time. Du- 
plication is possible by the addition of this standard to a 
blank of 40 c.c. of water containing the ferric solution. 
For dilution ten c.c. of this solution may be diluted to 
fifty c.c, treated with fifteen drops of ferric chloride and 
the sample or standard diluted for obtaining the result. 
The amount used here may be suitably varied from ten 
c.c. if the solution so made is too strong or too weak. As 
a standard for balancing add two c.c. of ferric chloride 
to twenty-five c.c. of the standard and dilute to 250 c.c. 
This latter standard for balancing will then contain .0001 
gram KCN per c.c. 



CHAPTEK NINETEEN. 

COLOR OF WATER, OILS AND DYES. 

The color of water is one of the things determined in 
water analysis in comparison with a standard solution 
officially specified. Oil is graded according to its color. 
The preliminary examination of dyes by colorimetric 
methods is possible in some cases. 

Color of Water. 

The color of water is graded by comparison with a 
standard platinum and cobalt solution made up as fol- 
lows, dissolve .1426 gram platinic chloride and .1 gram 
cobaltous chloride in water, add ten c.c. concentrated 
hydrochloric acid and make up to 500 c.c.^ The usual 
method is to make up a series of standards of 100 c.c. vol- 
ume. These should be at five point intervals up to 40 c.c. 
and at ten point intervals from that point to 70 c.c. 
Grading is never made higher than 70. The number of 
c.c. used, of the solution above specified, is the measure of 
the number of the grade indicated by that standard. 

For the preparation of a standard to be used by the 
balancing method the above 500 c.c. of standard may be 
diluted to one liter. The color value of this solution will 
then be 50 and, after balancing the instrument, the color 
value of the sample will be to that of the standard in- 
versely as the volume of the two used. For rapid work 
in the field permanent glasses are made having the va- 
rious standard colors as given by the official solution 
above specified. 

1 Olsen, Quan. Oliem. Anal. 

144 



COLOB OP WATEB, OILS AND DYES. 145 

Color or Oil8. 

All oil is graded by color, both the refined grades and 
the lubricants and crude oils. The instrument in ordi- 
nary use for this purpose is the Stam- 
mer form of colorimeter. Lovibond's 
Tintometer is also used with a special 
series of glasses for oils. The grad- 
ing of kerosene is into four classes, 
water white, superfine white, prime 
white and standard white. These 
grades all turn yellow on exposure to 
light but that has no effect on their 
burning qualities. The grading of 
oils other than lighting oil is too 
lengthy a subject for discussion here. 

Recently there has appeared on the 
market a new instrument, the Say- 
bolt Chromometer which combines the 
Lovibond Tintometer and the Stam- 
mer Colorimeter. This is treated at 
greater length under Apparatus, 
Chapter Two. 

Color of Dtestutfs. 
The use of the colorimeter for de- 
termination of the strength of solu- 
tions of dyestuffs is very limited in ex- ^^^ ^g^ B^jhoU 
tent, inasmuch as it is not reasonably Chromometer, 

accurate in practice in a majority of cases. It is however 
useful as a method of preliminary classification and may 
then be followed by the longer and more accurate test 
necessary after this preliminary has given a clue to the 
concentration to be expected. 



146 COLOBIMETBIO ANALYSIS. 

Any of the four methods of colorimetric analysis are 
applicable to this use under the proper conditions. Lovi- 
bond's Tintometer finds some use in this field. It is of 
course necessary to dilute the sample to be used as com- 
mercial solutions of dyes are ordinarily so concentrated 
that the eye cannot penetrate the solution. 



CHAPTEK TWENTY. 

NEPHELOMETBY. 

In closing a word about the methods of nephelometric 
measurement may not be amiss. The tests given are 
entirely those depending upon colorimetry, the measure- 
ment of the intensity of light of a given color trans- 
mitted thru a solution. The methods of nephelometry 
measure the amount of a finely divided precipitate sus- 
pended in a solution. This was first tried by a method 
similar to that of colorimetric determination, passing a 
ray of light thru the solution and estimating the amount 
of precipitate present from the interference with light 
caused by its presence. Experiment has shown that the 
measurement of such a precipitate is made many times 
more delicate by measuring the amount of light reflected 
from it rather than the interference of the precipitate 
with light. 

The instrument used for such determinations is a 
modified form of the Duboscq colorimeter with an opaque 
bottom to the cup. Precipitates suitable for such ex- 
amination are those of silver chloride and barium sulfate 
in dilute solutions where the precipitate is described as 
an opalescence. Many applications to bacteriological and 
physiological work have been found. For more complete 
information on this subject see Journal Industrial and 
Engineering Chemistry, Vol. 10, No. 7, P. A. Kober. 



147 



INDEX. 



Acetyl acetone, 41 

Adurol, 132 

Alizarin-S, 68 

Aluminum, 68 

Ammonia, 44, 94, 107, 118, 119 

Ammonium molybdate, 124 

Apparatus, 5 

Arsenic, 60, 64, 67 

Arsenious oxide, 136 

Arsine, 64 

Balancing method, 2, 10 
Benzidene, diphenyl, 110 
Bismuth, 62, 66 
Bleaching of Titanium solution, 

102 
Boron, 121, 126, 127 
Bromide, 53 
Bromide, mercuric, 64 
Brucine, 108 

Calculation of results, 27 
Campbell-Hurley Colorimeter, 12 
Carbon, 55 

Cassal method for gold, 93 
Chlorides, 35, 46, 74, 76 
Chlorine, 102, 105 
Chlorplatinate of potassium, 83, 

85 
Chromate, 70 
Chromium, 68, 70, 71 
Ghromometer, Saybolt, 21 
Cobalt, 73, 76 
Colloidal suspension, 161 
Color camera, 8 
Colorimeters 

Campbell-Hurley, 12 

Duboscq, 18 



Lovibond, 18 

Schreiner, 21 

Stammer, 21 

White's, 23 
Colorimetric apparatus, 5 
Colorimetrie methods, 1 
Colorimetric results, 27 
Copper, 44 
Cuprous chloride, 129 
Curcumin, 126 
Cyanides, 93, 94, 140, 141 

Definition, 1 

Diamine, m phenylene, 94, 115 
Dilution method, 1, 7 
Dimethylglyoxime, 42 
Diphenylbenzidene, 110 
Disodium 1.8 dihydroxynaphtha- 

lene 3.6 disulfonate, 71 
Doring, method for gold, 91 
Dowsett, method for gold, 89 
Duboscq colorimeter, 18 
Duplication method, 2, 9 
Dyes, 145 

Ethyl potassium xanthatc, 51 

Ferrie iron, 31 
Ferrous iron, 31, 133 
Fluorine, 102 

Gold, 89 

Hehner cylinders, 10 
Hydrogen peroxide, 95 

Iodide, 62 

Iodide of Zinc, 116 



148 



INDEX. 



149 



Iron, 31 

Lead, 60 

Lovibond tintometer, 18 

Magnesium, 86 
Magnesium phosphate, 122 
Manganese, 73, 78, 80 
Metaphenylene diamine, 94, 115 
Methods 

Balancing, 2, 10 

Dilution, 1, 7 

Duplication, 2, 9 

Series of standards, 1, 5 
Methylene blue, 106, 137 

Naphthalene derivative, 71 

a Naphthylamine, 114 

a Naphthylamine hydrochloride, 

117 
Nephelometry, 147 
Nessler's reagent, 118 
Nessder tubes, 8 
Nickel, 73, 74 
Nitrate, silver, 67 
Nitric acid, 107 
a Nitroso b naphthol, 76 
Nitrous acid, 107, 114 

Oils, 144, 145 

Oxide of tungsten, 101 

Oxygen, 129, 132 

Paraphenylenedimethyldiamine, 

135 
Perchlorates, 102, 106 
Periodate, 80 
Permanganate, 78, 80 
Peroxide of hydrogen, 95, 97, 

129, 13'3 
Persulfate, 78 
Phenol, 119 

Phenol sulfonic acid, 111 
m Phenylene diamine, 94, 115 



p Phenylene dimethyldiamine, 

135 
Phosphate, 86 
Phosphomolybdate, 8-6, 121 
Phosphorus, 121, 122 
Potassium, 83, 85 
Potassium bromide, 93 
Potassium ethyl xanthate, 51 
Potassium ferrocyanide, 36, 48 
Potassium iodide, 138 
Potassium platino chloride, 88, 

85 
Potassium sulfocyanate, 31 
Potassium thiocarbonate, 73 
Prister, method for gold, 90 
Pyrogallol, 113 

Resorcinol, 81 

Results, 27 

Rose, method for gold, 92 

Salicylic acid, 37, 47, 140 

Saylbolt ohromometer, 21 

Schreiner colorimeter, 21 

Selenious acid, 135, 138 

Series of standards method, 1, 5 

Silica, 121, 124 

Silver nitrate, 67 

Sodium peroxide, 93 

Sodium salt of alizarin, 68 

Stammer colorimeter, 21 

Stannous chloride, 83 

Starch, 116 

Steel, 55 

Strychnine, 100 

Sulfanilic acid, 114 

Sulfides, 39, 60, 135, 136 

Sulfocyanates, 141 

Sulfur, 135, 136 

Thiocyanates, 141 
Thymol, 97 
Titanium, 95, 97, 102 
Tolidine, 105 



150 INDEX, 

Tungsten, 95, 101 White's colorimeter, 23 

Turmeric, 127 

Xanthate, potassium ethyl, 51 
Vanadium, 95, 97, lOQ 

Zino, 73, 81 
Water, color of, 144 Zinc iodide, 116 



LITERATUKE OF THE 
CHEMICAL INDUSTRIES 



On our shelves is the most complete stock of 
technical, industrial, engineering and scientific 
books in the United States. The technical liter- 
ature of every trade is well represented, as is also 
the literature relating to the various sciences, 
both the books useful for reference as well as 
those fitted for students' use as textbooks. 
A large number of these we publish and for an 
ever increasing number we are the sole agents. 

ALL INQUIRIES MADE OF US ARE CHEER- 
FULLY AND CAREFULLY ANSWERED 
AND COMPLETE CATALOGS AS WELL AS 
SPECIAL LISTS SENT FREE ON REQUEST 




D. VAN NOSTRAND COMPANY 
Publishers and Booksellers 

8 WARREN STREET NEW YORK