Wn
M s«
a:j
******
n*
.*A^*M
Jilnatt,
■!MV
SMS
&A*e
^ArfA%*
_ r ^ -
'\<t '«'
%\*:^>sO*
i*>*.
Sw\
^
v 1 '
BpBBt 7 ^
iViTwnHiiBir
s:
^J^^^Sg
^ " " »!
lA-A^n^M**;
H^fl-T
I^a«?V/»7^3Aa .1
rv^A-HS
^^•W'SVSC'v ';
^^&B3^Wm
1\
f>My
■ A* ~ '/ ' 'V pr
WW
^y^^m
' >**'
■ **a?*V^
fflkCbHUi
'-;> ■-■
I
A STUDY OF THE PREPARATION OP PERMANGANIC ACID
BY ELECTROSIS.
A Dissertation presented to the Board of
University Studies of the Johns Hopkins Uni-
versity for the Degree of Doctor of Philoso-
phy.
By
Francis Le Jau Parker, Jr.
1901.
-ooo-
Johns Hopkins University,
Baltimore, Maryland.
Contents.
Page
Acknowledgement
Review of Previous Work 1
Effect of Concentration upon Yield of Acid
in Cups and Jars 14
Effect upon Size of Electrodes upon Yield of Acid -- 18
Effect of Intensity of Current upon Yield of Acid 20
Methods of Analyses 25
Conclusions 27
Electrical Endosmose 29
Biography 36.
ACKNOWLEDGMENT ■
This investigation was undertaken at the suggestion of
Professor H . N. Morse, and carried out under his supervision,
I desire to express my appreciation of the kindly in-
terest and instruction of Professor Horse, and also of Pro-
fessors Ira Remsen and H. C. Jones.
REVIEW OP PREVIOUS WORK
In May, 1900, Professor H. N. Morse and J. C. Olsen pub-
lished an article in the American Chemical Journal entitled
Permanganic Acid by Electrolysis. The results of a number
of experiments were recorded, by which they showed that per-
manganic acid could be prepared in any desired quantity by
the electrolysis of a solution of potassium permanganate.
If a solution of a permanganate is electroMzed in the usual
manner, the acid is quickly reduced by the hydrogen which
appears at the negative pole. In order to prevent this, the
negative electrode was placed in a porous cup filled with
water, end the accumulating alkali drawn off from time to
time by means of a siphon.
The details of the method used a^e. described by them as
follows i /the accompanying figure represents the apparatus
which they employed) "a, a is a galvanized iron tank through
which hydrant water flows, in order to prevent undue rise of
temperature, b, b is a beaker holding 1800 cc, in which are
placed the permanganate solution, the positive electrode e',
and the porous cup c
The cup has a capacity of about 250 cc. and rests upon
the glass tripod d. It contains the negative electrode e,
and one end of the siphon j. The open end of the siphon in
the cup is on a level with the upper edge of the electrode.
The electrodes are each 50 mm. square and are bent to conform
to the sides of the cup; e is of silve'r and e' is of plati-
num, g is a large watch-glas?; with a hole in the center
equal to the outside diameter of the cup. It serves to col-
lect and return the spray from the permanganate solution and
to protect the latter from the dust of the air. h is a
square wooden strip, which is clamped to the edge of the
tank- Into this are screwed the binding posts f, f, etc.
The arrangement by means of which the electrodes are made
adjustable in all directions is more clearly shown in the
supplementary figure, k is a glass tube with stop-cock
through which distilled water is made to flow into up
at any required rate, the rate depending, of course, upon
the frequency with which it is desired to dilute the alkali
by emptying the cup to the upper edge of the electrode, j
and m are siphons connected in an obvious manner in the tube
1. Through this system the alkaline solution in the porous
cup empties into bottle s, whenever its level rises above the
upper bend in m. m empties each time completely, while j
remains always full. The internal diameters of m and j
should be related to each other about as 2 to 3. If they
are equal, m will empty 1 and itself before the required
amount of liquid has passed out of the cup, and v/ill not
again act until the cup has been refilled."
The permanganate solution was made as concentrated as
the temperature of the hydrant water would permit, whi , tir-
ing the winter, was a concentration of about 40 grams per li-
ter. The solutions used were filtered through asbestos in
order to free them of oxides.
8 to 10 of the cells described were arranged in series,
and when necessary, additional resistance was inserted by
means of lamps, so arranged that the current could be passed
through any number of them in parallel, or thrown into series
with each other.
It was found that the resistance of the individual cells
was subject to considerable fluctuations—rising to a maximum
whenever the cup was emptied by the siphon, and falling a £j ain
as the refilling proceeded—and that it was considerably
increased by the deposits of oxide on the walls, and in the
pores of the cup. The resistance amounted to from 5 to 10
ohms on the first day, rising on the second to 25 or 50 ohms,
and declining during the third and fourth to about 10 ohms.
When the average of the observations amounted to less than
13 ohms, the cup was considered a good one.
Ordinary porous battery cups were used. Others of a
more porous nature, and made of various mixtures of clay,
sand, and ground flint were tried, but not found as satisfac-
4
tory. In general the yield increases as the porosity dimin-
ishes .
*
It was found that a 4 per cent, solution of potassium
permanganate would yield a very dilute solution of acid; and,
in order to obtain a more concentrated onc J it was necessary
to add more of the salt from time to time as the alkali is
drawn into the cup and removed by the siphon. Fortunately
for purposes of concentration, the electrolysis is attended
in a very striking manner by the phenomenon of "electrical
endosmose." With a current varying between 1 and 1.5 am-
peres, the v/ater passes out of the permanganate solution in-
to the cup; i.e., in the direction of the current, at the
rate of about 500 cc per day of twenty-four hours. By re-
placing the water thus withdrawn from the beakers by equal
volumes of the 4 per cent, solution of the salt, they were
able to introduce into each cell an additional 20 grams of
permanganate per day; and, accordingly, to increase the con-
centration of the acid to any required extent. They did not
ascertain how far it was profitable or practicable to concen-
trate the acid in this manner. The strongest acid obtained
by them did not exceed a ten per cent, solution. They state
that the limit of concentration was not reached, as the per-
centage yield of the acid showed no tendency to decline as
the strength was increased.
The yield of acid obtained was from 87 to 92 per cent.
of the theoretically possible amount. The loss was due to
various causes: reduction of the acid to oxide and spatter-
ing. The pincipal source of loss, however, was the reten-
tion of the acid by the walls of the cups when they were re-
moved at the close of the experiment. This was extracted and
found to amount to about 3 per cent, of the acid left in the
breakers .
When the acid is made in the manner here described,
several days must elapse before a moderately concentrated
solution can be obtained. They found that the last of the
alkali was extracted very slowly and with a lar^e expendi-
ture of current. These considerations led them to try the
plan of placing both electrodes in porous cups, in order to
determine whether the change would prove economical. They
assumed that such an arrangement would give a fairly concen-
trated acid within a short time, and one which would be
probably free from potassium. It appeared to them doubtful
whether the change would prove economical, inasmuch as they
assumed that the introduction of a second cup would increase
the resistance of the cell. They had tried only one experi-
ment of this kind when the work was discontinued.
Tt was with this object in view,' i.e., the determina-
tion of the best conditions for the preparation of permangan-
ic acid, and the extent to which it could be profitably con-
centrated, that the herein described investigation was under-
taken.
In order to determine approximately the conditions which
seemed most likely to effect the yield of permanganic acid
in the cups and jars a few experiments were tried at random.
In these experiments, and also in all of the experiments
to be described, essentially the same form of apparatus was
employed as that described by Morse and Olsen; with the ex-
ception that in place of beakers ordinary glass battery jars
were used as the receptacles for the permanganate solutions,
and both electrodes were placed in porous battery cups.
Two cells of the kind described were arranged in series,
and placed in a cooling tank through which a continual current
of hydrant water flowed. 2 liters of potassium permanganate-
solution containing 40 grams per liter, which had been prev-
iously filtered through asbestos in order to free it of ox-
ide, were put into each jar and both electrodes placed in
porous battery cups, which were filled with distilled water.
The electrodes were each 50 mm. square and bent to conform to
the sides of the cup. The positive electrode was of plati-
num anci A negative of silver. At the beginning of the experi-
ment 2 or 3 drops of caustic potash were added to the cups
containing the negative electrode, and a few drops of per-
manganic acid to the cups containing the positive, in order
to increase the conductivity of the solutions. An intermit-
tent siphon was placed in the porous cups surrounding the
8
negative electrodes, so that the accumulating alkali was
drawn off automatically at a rate depending upon the amount
of water extracted by the electrical endosmose. About 20
cc. of water were extracted per hour from the permanganate
solution of each cell by this force. By this arrangement
the electrolysis could proceed without much attention.
A current of from 1 to 3 amperes was passed through the
cells for 5 to 6 hours. The fall of potential of each cell
was measured 4 or 5 times during the experiment, and the re-
sistance calculated from Ohms law. To test the relative
merits several kinds of porous cups were tried: the ordinary
porous battery cups; and others of a more porous nature,
which were made of a very pure finely grained clay. The lat-
ter were found useless as acid cups, owing to a very rapid
electrical endosmose in the direction of the current which
sucked the liquid out of the cups into the jars at the rate
of 800 to 1000 cc per hour. The effect of this was to con-
siderably dilute the permanganate solutions in the jars, and
increase the resistance of the cells. These cups were dis-
carded and others made of a similar kind of clay were tried.
The latter were 3 in. longer, and were not supported on tri-
pods, but rested on the bottom of the jars. It was found
that these also were too proous to be of service as acid cups
since the solution was carried by the current into the jari
at the rate of from 100 to 150 cc per hour. These cups were
reburned in order to diminish as much as possible the porosi-
ty; but they were still found to be too porous, and the rate
of flow from the cups into the jars was but little effected.
They served, however, very well as alkali cupSjWhen only one
cup was employed in the electrolysis. When used, in this
manner, singly, they had a very low resistance; and, if re-
moved from the solutions at the close of the experiment^ and
soaked in running water over night, could be used a number
of times without appreciable increase of resistance. But
fvt
one experiment of this kind was tried. The ist day the aver-
age resistance amounted to about 10 ohms. On the 3rd day
the resistance fell to 5 ohms and remained at 5 ohms for 15
days. It was not determined how long these cups could be
profitably employed in this way, as the experiments in this
direction were discontinued.
They did not serve so well as alkali cups when paired
with an ordinary battery cup, as the alkali solution drawn
from them was colored in consequence of the infiltration of
permanganate. When the negative electrode was placed in a
battery cup, the alkali extracted was in general colorless.
In addition to the positive electrical endosmose in the
alkali cups and also in the acid cups, there was a negative,
10
'.'--' ' on
or counter endosmosej in trie opposite^to the current, when
the ordinary battery cups were used in pairs, or paired as
acid cups with the others.
At the same time accompanying the negative endosmose in
the acid cups, there was also a positive endosmose in the
alkali cups, and the solutions rose in both cups at the same
time; i.e., the endosmose went with and against the current
in the same experiment. This appears to be abnormal in the
light of observations heretofore made on electrical endosmose.
From these experiments nothing definite could be ascer-
tained regarding the conditions regulating the yield of acid
in the cups; but it was found that pure acid free from per-
manganate could be concentrated in the cup surrounding the
positive electrode. It was also found that only a portion of
the acid resulting from the electrolysis entered the cups su*»
rounding the anode, since a considerable part remained out-
side in the jars. This could be o btained free from perman-
ganate at any time by continuing the electrolysis until all the
alkali was extracted. It was later found, as in the previous
work, that this required several days for its completion,
and that the last of the alkali was extracted very slowly,
and involved considerable expenditure of current.
A number of systematic experiments were next conducted
with the view of determining the effect of concentration,
11
intensity of current and size of electrodes upon the yield
of acid in the cups and in the jars. It was presumed that a
current of more than two amperes would not prove profitable,
as it was thought that the higher temperatures accompanying
the use of larger currents would increase the tendency of
the acid to decompose.
Two cells, the cathode and anode of each being placed
in porous cups, were arranged in series. All of the cups
were filled with distilled water and allowed to stand until
the walls had become thoroughly saturated. They were then
placed in a 4 per cent, solution of potassium permanganate
and the electrodes introduced. Two or three cups of caustic
potash were added to the cups containing the negative elec-
trodes and a few drops of permanganic acid to those contain-
ing the positive, in order to increase the conductivity of
the solutions in the cups at the beginning of the experiment.
The current at disposal was one of 110 volts, and it was
found necessary to insert addition resistance into the cir-
cuit. For this purpose ordinary incandescent lamps of dif-
ferent candle power were employed, which were so arranged
that the current could be passed through any number of
them in parallel. By this arrangement the current could be
varied as the resistance of the cells increased or diminish-
ed, and was kept approximately constant throughout the ex-
12
perirnent .
Duplicates of these and all experiments described were
made, and the results agreed within the limits of experiment-
al error. The errors are to be attributed to various sour-
ces, some of which could not be regulatedr--absorption of the
acid by the porous cups. It was found by Olsen 'hat this
amounted to from 3 to 3 l/2 per cent, of the acid in the
cups. It also required some time for the current to reach
the maximum that was to be maintained in the experiments, and
the acid formed during this time was not taken into account.
The amount of coursej varied, depending upon the maximum to be
maintained, and the time necessary to reach it. It generally
took from five to fifteen minutes for the current to rise to
1 ampere, and about half an hour to rise to 3 amperes. The
volume of the acid and alkali taken from the cups was measur-
ed with a graduated cylinder, as the accuracy of the method
did not warrant the use of calibrated flasks. All of the re-
sults given are the average of 4 cells. The resistance of
different cups varies considerably and increases with use,
owing to the deposit of oxide on the walls and within the
pores of the cups. The increase is greatly diminished by re-
moving the cups at the end of the experiment and soaking them
in running water over night. This removes the acid from the
13
pores and prevents decomposition. The advantage of this
treatment is clearly indicated by a comparison of the resist-
ance with the resistance of cups which were not treated in
this manner. The average resistance of the cups used by
Morse and Olsen, with a current varying between 1 and 1.5
amperes was 10 ohms per cup, when the current was passed con-
tinually through them for several days and nights. In these
experiments, in which the cups were treated as described, the
average resistance per pair of cups, when the same strength
of current was used, was 7.5 ohms. The resistance of same
cups increases very rapidly with use, while others can be
used successively for 10, or 15 days. When the average of all
observations amounted to less than 10 ohms per pair of cups,
they were regarded as good ones. In some cases the average
was less than 5 ohms for six consecutive days.
To prepare the cups for a second experiment, they were
immersed for five or six days in commercial hydrochloric
acid, and then kept for several days in a bath through which
a constant stream of hydrant water flowed. This treament,
when thorough, put the cup in as good condition as new. It
is doubtful, however, how many times this treatment can be
repeated, as the hydrochloric acid dissolves out the soluble
constituents of the clay and leaves in places small holes,
which, in some cases, extend nearly through the walls of the
14
cup. Several of the cups were cleansed three or four times
in this manner.
EFFECT OF CONCENTRATION UPON YIELD OF ACID
IN CUPS AND JARS.
A current of 1 ampere was passed on successive days
through solutions of different concentrations for six hours,
and the current was kept as constant as possible. The first
day a solution of permanganate containing 40 grams per liter
was electrolized. The^lectrodes used were each 5 mm. square
and bent to conform to the sides of the cups. The convexed
surfaces were turned towards each other and the cups were
placed about 1 l/4 inches apart; care being taken that they
did not touch. The fall of potential was measured four or
five times during the experiment and the resistance calculat-
ed from Ohm's law. During the first hour the resistance was
quite high, but decreased rapidly as the solutions of alkali
and acid became more concentrated in the cups, and continued
to decrease until the end of the experiment. The resistance
of the pair of cups during the experiment is tabulated below:
1st hour. 2nd hr. 4th hr. 5th hr . 6th hr. Average.
17 ohms 8 ohms 7.5 ohms 6.5 ohms 6.5 ohms 7.1 ohms
The average is computed in this and all experiments from
15
the beginning of the 2nd hour, and the number of watts used
calculated on this basis. The resistance during the first
hour could, of course, be reduced by adding larger quanties
of alkali and acid to the cups at the beginning of the ex-
periment. In these experiments the phenomenon of electrical
endosmose was apparent in both cups at the same time, and
the flow proceeded in opposite directions. Water was forced
from the jars into the cups containing the negative elec-
trode, at the rate of about 20 cc per hour. In the acid
cups the flow was against the direction of the current and
less in amount than in the alkali cups. About 20 cc. of wa-
ter were carried into the acid cups by this force in 6 hours.
The average temperature of the cups during the experiment was
20o C. The temperature was subject to very little fluctua-
tion, and did not rise at any time much above 20°.
The yield of acid in the cups and in the jars is given
below:
Yield in grams Yield in grams
per ampere hr. per watt hr.
In cups 0.723 0.050
In jars 1.01 0.0 70
1.733 0.120
In the 2nd experiment the same conditions were maintain-
ed as in the first, with the exception that a less concen-
16
Yield in grams
per
ampere hour.
In cups
0.721
In jars
0.919
Total
1.640
trated solution containing part acid was electrolized. The
solution contained 32 grams of permanganate and 25 grams of
free permanganic acid per liter. The same cups were used,
and the average resistance was a little higher, amounting to
7.5 ohms. The temperature of the cups remained the same.
The yield of acid obtained is given below:
Yield in grams
per watt hour.
0.048
0.061
0.109
In the next experiment a solution containing 16 grams of
permanganate and 20 grams of free permanganic acid was elec-
trolized. To maintain as near as possible the other condi-
tions of the experiments just preceding, the acid cups em-
ployed were cleansed and again used. However, other cups
made of the same material were substituted for the alkali
cups. The resistance was nearly twice as high as the resist-
ance in the above experiments, amounting to 15 ohms. This
increase may be due in part to a greater resistance in the
alkali cups used. But as. the electrical endosmose increases
as the concentration of the solution is diminished, the in-
crease is largely due to the very rapid extraction of water
from the jars, which left a good part of the cups uncovered
17
by the liquid outside. The temperature of the cups was the
same as in the preceding experiments. The yield obtained is
given below:
Yield in grams
per ampere hour.
In cups
0.743
In jars
0.837
Total
1.580
Y
ield in
grams
P
er watt
0.025
0.026
0.051
hour.
For purposes of comparison the yields of acid per ampere
hour and per watt hour in the three experiments is tabulated
below :
TABLE IV.
Concentration Yield in Yield in Yield in Yield in
of solution cups per jars per cups per jars per
in grams per ampere hr. ampere hr. watt hr. watt hr.
liter. in grams. in grams. in grams.
Kmn0 4 40 0.723 1.010 0.050 0.070
(Kmn04 32
) 0.721 0.919 0.048 0.061
(Hmn04 25
(Kmn04 16
) 0.743 * 0.837 0.025 0.026
(Hmn0 4 20
M The slightly greater yield per ampere hour in this experi-
ment is due to the high resistance. A longer time being re-
quired to reach the maximum of 1 ampere, and the acid formed
during that time was not taken into account.
18
These results show clearly that:
The yield per ampere hour in the cups is independent of
the concentration, and does not appear to be Affected by
the resistance of the cups.
The yield per ampere hour in the jars diminishes as the
dilution of the solution is increased.
The yield per watt hour in the cups and in the jars di-
minishes as the dilution of the solution is increased.
EFFECT OF SIZE OF ELECTRODES UPON YIELD OF ACID.
To determine the effect of size of electrodes upon
the yield of acid, two experiments were tried with a current
of 1 ampere, in which the same conditions were maintained as
in the preceding experiments where electrodes 50 mm. square
were used. Later, others were tried, under similar condi-
tions, with currents of 2 and 3 amperes intensity. The size
of the electrodes in these experiments was doubled, and
electrodes 50 mm. wide and 100 mm. long were substituted.
Distilled water was added from time to time to the jars, in
order to replace the water removed by electrical endosmose ,
and to keep the liquid in the cups and jars at the same level,
In the 1st experiment, a 4 per cent, solution of permanganate
was electrolized. The cups used in the other experiments
19
were replaced by new ones of the same material. As is to be
expected, the temperature of the cups was somewhat lower,
amounting to 12° c.; but the average resistance of the cups
was about the same. The size of the electrodes does not ap-
pear to effect the amount of water extracted by electrical
endosmose , as in this experiment the volume of liquid ex-
tracted was the same as in the case where smaller electrodes
were used- The yield of acid is given below:
Yield in grams Yield in grams
per ampere hour. per watt hour.
0.056
0.092
0.148
In the second experiment a solution containing 35
of potassium permanganate and 16 grams of free permanganic
acid was electrolized. The same cups were used and the other
conditions v/ere in general the same as in the preceding ex-
periment. There was no change in the temperature and re-
sistance of the cups. The electrical endosmose in both cups
increased in proportion to the increase in dilution of the
solution electrolized. The yield of acid is given below.
Yield in grams Yield in grams
per ampere hour. per watt hour.
In cups 0.810 0.056
In jars 1.110 0.077
Total 1.920 0.133
In cups
0.758
In jars
1.2 5
Total
2.008
20
A comparison of the results of these experiments with
the results of the experiments in which smaller electrodes
were used shows that there is an increase in the yield of
acid per ampere and per watt hour; and that the use of large
electrodes is to be recommended.
Size of Concentra- Yield in Yield in Yield in Yield
Electrodes, tion of so- cups per jars per cups per in jars
luticn. ampere ampere hr. watt hr. per
hour in in grams. in grams, watt hr.
grams . in grams
50 x 50 mm. A%
50 x.100 mm. 4<
0.723
1.01
0.050
0.070
0.758
1.25
0.056
0.092
Total yield Total yield Percent- Percent- Percent- Percent-
in grams in grams age in age in age in a 6 e in
per ampere per watt hr. cups per jars per cups jars per
hr. amp.hr. ampere per watt hr.
hr. Watt hr.
1.733
0.12
41.72
58.28
41.66
58.33
2.008
0.148
37.75
62.25
37.83
62.16
EFFECT OF INTENSITY OF CURRENT UPON YIELD OF ACID.
To determine the effect of the intensity of the cur-
rent upon the yield of acid, the experiments, in which a
current of 1 ampere was used, were repeated with currents of
2 and 3 amperes intensity. The other conditions of the ex-
21
periments were kept, as neaiyas possible, the same as the
conditions in the experiments referred to- As has been men-
tioned, when a concentrated permanganate solution is electro-
lized with a current of 1 ampere, the resistance decreases
rapidly during the first hour, and continues to decrease as
the acid and alkali solutions in the cups become more con-
centrated. In the experiments where currents of 2 and 3 am-
peres intensity were employed, the resistance of the cells
decreased only during the first three or four hours, and then
began to rise in consequence of the extraction of water, by
endosmose, from the jars. This exposed a part of the cups,
and allowed the acid on the walls to become reduce , hich greatly
increased the resistance of the cell. In some cases, the
increase in resistance towards the end of the experiment was,
in part, reduced by adding distilled water to the jars as
the water was extracted. It can be entirely prevented by
adding, instead of water, a solution of potassium permangan-
ate. This increase was not so noticeable when a concentrat-
ed solution of permanganate was electrolized, but in the ex-
periments in which more dilute solutions were used, on ac-
count of the increased dilution of the solution, the amount
of water extracted, by electrical endosmose, was so great
that a very large portion of the cups was left exposed, and
22
the resistance after the third hour increased very rapidly.
In some cases, the increase was so rapid, that the experi-
ments had to be discontinued before the end of the sixth
hour.
The electrical endosmose in the alkali cups was normal,
and varied in accordance with the laws deduced by Wiedermann;
but in the acid cups, as in the other experiments, it appear-
ed to be abnormal {-liquid being forced from the cathode to
the anode,* i.e., against the current. In other respects, it
appeared to follow the laws referred to, and the amount in-
creased as the intensity of the current became greater. The
resistance of the cells decreased; end the temperature of
the cups increased as the current increased.
A record of t'e resistance and temperature of the cells,
when a solution of permanganate containing 40 grams per liter
was electroljzod, is given below:
Resistance in Ohms.
Cur- Temperature
rent 1st 2nd 3rd 4th 5th 6th Average. of cups,
in hr. hr. hr.hr.hr. hr .
am-
peres.
1 17. 8. 7.5 G.5 6.5 7.1 180 C.
2 10. 5.5 5. 4.4 47 4.9 20
3 5.2 3.6 3.7 3.4 3.7 3.9 3.7 28
23
Although a large number of experiments were tried . r ith
solutions of different concentrations, in which the amperage
was varied from day to day, only in the series in which a
solution of permanganate containing 40 grams per liter was
employed were the conditions at all comparable. In the oth-
ers, in which less concentrated solutions were employed, the
resistance varied so from day to day, that there was no uni-
formity in the yields per watt hour of duplicates. The dis-
crepancies in the results are due to several cause;.', lut
principally to the different character of the cups used, as
regards resistance. In some of the experiments the cups had
to be changed several times, and as they differed very
greatly as to resistance, there was a marked difference in
the yield per watt hour in duplicates. In other experiments
the resistance of the cups changed so rapidly from day to day
that the number of watts consumed in duplicates varied con-
siderably. This difference frequently amounted to from 20 to
40 watts an hour.
In the series in which a concentrated solution of per-
manganate, containing 40 grams per liter, was electrolized,
the conditions remained practically constant. The resist-
ance in the duplicate experiments was the same, and the re-
sults below' are in every respect comparable.
24
Current Concentra- Yield Yield Yield Yield per
in am- tration of per am- per am- per watt watt hr.
peres. solutio* . pere hr. pere hr. hr. in in jars.
in cups, in jars. cups.
1 40 gms.per ltr. 0.723 1.010 0.050 0.070
2 " 0.699 0.922 0.034 0.046
3 " 0.675 0.933 0.029 0.040
Total Total Percent- Percent- Percent- Percent-
yield per yield age in age in age in age in
ampere hr- per cups per jars per cups per jars per
watt hr. ampere ampere hr. watt hr . watt hr.
hr-
1.733 0.120 41.72 58.28 41.66 58.33
1.621 0.080 43.12 56.88 42.50 57.50
1.608 0.069 41.98 58.02 42.03 57.96
A comparison of these results shows that the yield per
ampere is not appreciably affected by the intensity of the
current. The slight differences in the yield per ampere hour
in the results given above are probably due to the sources of
error mentioned in the beginning of the article, as in other
experiments in which the effect of current was tested, the
yield per ampere hour remained practically constant.
The yield per watt hour diminishes in the cups, as well
as in the jars, as the intensity of the current is increased.
It was found that the best results were obtained when a cur-
rent of 1 ampere intensity was employed.
25
When a concentrated solution of permanganate was elec-
trolized with a current of 2 amperes, a two per cent, solu-
tion was obtained in six hours. At the end of six hours the
current was discontinued, the acid transferred to a bottle
and the cups were soaked in running water over night. It was
found that the concentration could be increased from two to
three per cent, each day. A concentration of 17.5 per cent.
was reached, which appears to be the limit that can be ob-
tained by this method. At this concentration the acid solu-
tion has an oily appearance resembling sulphuric acid. It
decomposes very rapidly, and when poured from one vessel
to another the decomposition takes place with an energetic
evolution of gas.
It was found that the acid could not be profitably con-
centrated beyond a ten per cent, solution. Even at this
concentration it is very unstable and deposits oxide on
standing over night. If the acid is to be prepared in large
quantities, it should be used immediately or converted into
some stable salt.
The barium salt is best adapted for this purpose.
METHODS OF ANALYSES.
The amount of acid present in the cups was determined
by titrating a measured volume of the acid against a standard
26
solution of potassium tetroxalate. From the tetroxalate
used the acid was calculated.
To determine the amount of free permanganic acid in the
jars, two equal portions of the solution were measured off.
One was titrated against a standard solution of potassium
tetroxalate, and the amount of sulphuric acid necessary to
convert all of the manganese into manganous sulphate, assum-
ing it to be in the form of free permanganic acid, was cal-
culated. To the other portion a measured quantity of sul-
phuric acid, in excess of the amount calculated, was added.
This was carefully reduced with a neutral solution of hydro-
gen peroxide, and the excess of sulphuric acid determined by
means of an alkali whose relation to the sulphuric acid was
known. The sulphuric acid neutralized in excess of the
amount calculated for the permanganic acid represents the
amount of potassium present in the solution, and from this
the amount of potassium permanganate can be calculated.
As a general rule, however, the quantity of acid in the
jars was calculated from the amount of potassium hydroxide
extracted from the solution. This would give the total per-
manganic acid formed. By subtracting the acid formed in the
cups, the acid in the Jars was ascertained. The results of
the two methods agreed closely.
CONCLUSIONS.
The results of the experiments described leave no doubt
as to the advantage of the use of two porous cups in the
preparation of permanganic acid. When both electrodes are
placed in cups^the resistance of the cell is diminished and
the yield per ampere hour is more than doubled. Morse and
Olsen obtained a yield of 0.3 erf fc gram per ampere hour,
when one cup was employed, and a current of from 1 to 1.5
amperes was passed through a concentrated solution of perman-
ganate. With the same current and concentration a yield of
0.7 of a gram was obtained when two cups were employed.
The advantage of discontinuing the electrolysis each
day, and soaking the cups in water over night is to be empha-
sized. When the electrolysis was conducted in this manner,
the average resistance of the pair of cups did not exceed
7.5 ohms, and some cups were used for 10 or 15 successive
days when a current of one ampere was employed. In experi-
ments in which the electrolysis was continued for several
days and nights without removing the cups, the resistance of
the cell was considerably higher, averaging, after the first
day, from 10 to 15 ohms.
A current of from 1 to 1.5 amperes is the most advantage-
ous. When a larger current is used the yield per watt hour
is diminished, and the level of the liquid in the jars is
28
lowered too much for the electrolysis to be left any lent;
of time unattended. A smaller current will, of cour J&$~
fect decomposition, but prolongs the operation needlessly.
When currents of less than o.5 of am ampere are employed
there is a tendency of the resistance to increase, and the
permanganate solution passes into the alkali cup in small
quantities.
The solution should be as concentrated as possible, and
the water extracted by electrical endosmose should be replac-
ed by some of the same solution, so as to keep the cups as
much covered as practicable. This prevents reduction of the
acid in the walls and undue rise of resistance. The tempera-
ture of the cups should be kept as low as possible.
Care should be exercised in the choice of clays. Some
clays are entirely unsuited for use in this method, since
the endosmose proceeds in the direction of the current, and
the acid is forced from the cups into the cell at a very rap-
id rate. In cups made of these clays, it is not practicable
to concentrate the acid in the cups. The ordinary porous
battery cups were found the most satisfactory. In these the
endosmose was against the current \ but the quantity of li-
quid forced into the cups was not sufficient to injure their
value. It decreased as the acid concentrated in the cups and
became inappreciable when a concentration of 6 per cent, was
29
ELECTRICAL ENDOSMOSE.
Since electrical endosmose plays such an important role
in the preparation of permanganic acid by electrolysis, I
shall describe more fully some apparently abnormal phenomena
which have been observed in the course of these experiments.
Though the fact has been known for some time, having been
observed by Reuss* in 1807, it does not appear to have at-
tracted the general attention of chemists and physicists.
If an electric current is passed through a U-tube con-
taining water, and a porous septum is placed in the bend of
the tube, in addition to the decomposition of the water, a
movement takes place in the direction of the current j i.e.,
the water sinks at the positive electrode, and rises at the
negative. The movement also takes place without the diaphragm
referred to, as determined by Quincke, but to an inappreci-
able extent, and can only be observed with a microscope.
In 1852, Wiedermann 2 published an elaborate investiga-
tion of this subject. He found that the liquid was carried
in the direction of the current when plates of clay or gypsum
were used. The liquids employed were: water, copper, sul-
phate, zinc sulphate, solutions of sodium and potassium hy-
droxide and dilute alcohol. The apparatus which he used, in
Mem de la Societe' des Naturalistes *a Moscow 2, page 327,
1809.
2 Pogg. Ann. 113, page 513, 1861.
30
order to estimate quantitatively the amount of water carried
by the current, consisted of a glass jar in which was placed
a porous clay cell. The cell was surrounded by a cylindri-
cal electrode, and a similar electrode was placed inside.
To the top of the cell a tubular glass bell was cemented,
which connected at right angles with a horizontal side tube.
As the liquid rose to the height of the side tube, it flowed
over and was collected in a tared flask and weighed. Prom
these experiments he deduced the following law: that the
quantity of liquid carried by the current is directly propor-
tional to the intensity of the current, the specific resist-
ance of the liquid and the thickness of the wall; and inverse
ly proportional to the opening of the wall.
A very important observation was made later by Quincke.!
He found that certain fluids moved in the opposite direction
to the current; while others, which generally went in the di-
rection of the current, under certain circumstances were car-
ried against it.
Turpentine in a glass tube showed a positive endosmose,
but if the tube was coated with sulphur the endosmose pro-
ceeded against the current.
Pogg Ann. 87, page 321, 1852.
31
Carbon bisulphide showed in most glass tubes a positive
endosmose, but in special kinds of glass a negative.
In Wiedermann's apparatus with clay cells turpentine was
carried against the current.
Helmholtz proposed the following plausible explanation
of these phenomena:
The fluid and the wall stand in electrical opposition,
just as a metal and a flowing liquid, or as a rubbed and
rubbing body. There exists between them a potential differ-
ence, and an electrical double layer is formed along the sur-
faces of contact. In general the potential of the liquid is
positive, except in the exceptions mentioned by Quincke. The
outer liquid surface is attached to the wall, while the li-
quid otherwise moves with inner friction. The electric cur-
rent detaches the electrically charged parts, which do not
lie directly on the wall. By friction the other parts of the
cross section are put in motion, and thus the Electrical
Endosrnose .
If, on the other hand, the liquid is moved by an exteri-
or hydrostatic pressure, the inner parts at the surface of
contact will also be driven on. These will detach the out-
er portions from the wall or drag them on. In this way
-v positive electricity will be set free.
32
At the beginning of the tube a new layer will come in
contact with the wall and be charged posit ive ^ he remainder
of the liquid at the beginning of the opening will be charged
negative.
In all of the experiments describee thus far, as well
as can be ascertained, only one cell or ; jrous septum was
employed; and the electrical endosmose proceeded either in
one direction, ov the other. In none of the experiments were
both electrodes surrounded by porous septa.
In the experiments described in this investigation, both
electrodes were placed in porous battery cups.
The cell contained three solutions of different charac-
ter and concentration; and, in addition, there was a differ-
ent solution in each cup, containing different ions.
The electrical endosmose in some cases, as noted, pro-
ceeded in both directions at the same time, rising in the
acid cups as well as in the alkali cups when the ordinary
battery cups were employed.
When other cups of a more porous nature, and made of a
different kind of clay, were used, the liquid was carried by
this force, as described, in the same direction in both cups,
viz: in the direction of the current.
In the case of the battery cups the endosmose in the
alkali cups was much greater in volume than in the acid cups,
33
amounting to nearly six times as much; whilst with the more
porous cups the conditions were reversed, and the liquid
flowed out of the acid cups more rapidly than it was carried
into the alkali cups. In both cups, however, the flow pro-
ceeded in the direction of the current.
There appears to be no explanation for these apparently
abnormal phenomena, other than that of Helmholtz. The theory
of Helmholtz accounts for the observations of Quincke; viz.,
that under certain conditions the endosmose proceeds in the
opposite direction to the current. As it assumes that the
liquid in the pores of the septum or cell under these condi-
tions is charged negatively in respect to the walls of the
pores, and, therefore, the particles detached by the current
are attracted to the anode and carried against the current.
The behavior of turpentine in glass tubes and s coated
with sulphur is explained by this assumpti "hen a current
is passed through turpentine contained in glass tubes, the
turpentine at the surface of contact is charged positively in
respect to the glass, and, hence_, attracted to the cathode.
Therefore the endosmose should be positive; i.e., in the di-
rection of the current, as is the case.
When the turpentine is contained in glass tubes coated
with sulphur, the effect of the sulphur coating is such as to
change the character of the charge. The turpentine, in this
case, is charged negatively in respect to the sulphur, and is
attracted to the anode. Hence the endosmose proceeds
against the current.
This assumption is not unwarranted, as it is well known
that very slight differences will effect changes in the char-
acter of the charge of one body in respect to another. Some-
times the mere roughening of the surface of one of the sub-
stances will effect this change.
This explanation might account for the difference in
character of the endosmose in the two kinds of cells referred
to,' i.e., why the endosmose should proceed in opposite direc-
tions in cells made of different clays, as very slight dif-
ferences in the composition of the clays could account for
the change in character of the charges. But it does not ex-
plain why in cells of the same composition the endosmose
should proceed in opposite directions at the same time; es-
pecially as the same conditions are maintained even when the
acid and alkali cups are interchanged.
In order to account for this we must further assume that
the presence of different ions also affects the character of
the charge of the contained liquid in respect to the walls of
the cell. If this be the case, we have an explanation for
this phenomenon, as in the alkali cups we have the ions K and
(OH), and in the acid cups H and Mn04-
35
There is yet one phenomenon to be accounted for: viz.,
why the more porous cups should not be i-f 'ected by the H and
M11O4 ions in the same manner as the battery cups K and OH
ions are present at the cathode, and H and MnC^ itnis at the
anode in both cases. The endosmose at the cathode proceeds
in the same direction in both kinds of cups. At the anode,
however, the endosmose is in the direction of the current
when the more porous cups are employed, and in the opposite
direction when ordinary battery cups are employed. The (OH)
ions appear to affect the charges of the different clays in
the same manner; but the Mn04 ions must effect a change in
the character of the charges of different clays. This is
suggested as an explanation for the different behavior of the
two kinds of cups at the anode.
These speculations are based entirely on the assumption
that the theory of Helmholtz is the correct explanation for
the phenomenon of electrical endosmose.
Biography.
The author of the foregoing thesis was born in
Charleston, South Carolina, January 8, 1876. His early
training was received at the Charleston High School and
University School of Charleston. In 1891 he entered the
South Carolina Military Academy and graduated in June,
1894. The following October he entered the South Caroli-
na College and pursued courses in Chemistry, Geology and
Mineralogy for two years.
In 1896 he was appointed Assistant Professor in
Charge of Chemistry at the South Carolina Military Acad-
emy. During the summers of 1898, 1899 and 1900 he pur-
sued courses in Chemistry and Geology at the University
of Chicago and received the degree of Bachelor of Sci-
ence. In July, 1900, he was granted a leave of absence
for one year from the South Carolina Military Academy and
entered the Johns Hopkins University in October.
-oo-oo-
W?*Vvv v
V w ,v- V
/k'V^'J 1
*V\j*i s
"fT
*^
-'"-" vV;i fe:»» : fe^^:'
- - . - • -
fc*fc
.wVw v ,
* w - • - - - v _, _, - - . ~w
▼fff'
WwhSvyy, V A v \
VVV" v w w -"v-
,;;*>*
.Ml
JL
**W*v%vWtf
'VyW^vrf^^
v^vV
'?M
•«** V..-'
'V'
>WW V
VV^,/
■yv' V
Wff*?'
II.
vv
MRE
fc£
yy-v*y
. JM a.- - v ' ^
J - .
*y*v-
i&fiW-
,^w
sMi
. .v. - - - - • .1."..
, 1 1 Hi
'v$i$$>
VW*^:
^WwW^WWW^^w^^
VvaAa j-^
vV »
^Mwu
u/7 VywwJA/W^^M^Aiiywy^o^.^uwu.
Ai JU . . . V^J r
^WMfc
uw^v r^
S8*L*vfc2 :
