324 W. 3. WENDRIXSON.

TABI,EIII.-EFFEcT OF OZONG AND LIGHTON BLEACHING.

Time, minntes. Reading. Time, minutes. Reading.
0 5 I20
30 8 150
60 IX I 80
90 12 ... ..
In Fig. 3 the above data are plotted together with those for the com-
bined light and peroxide from the earlier table.
Summary.
It has been shown that the oxidizing power of hydrogen peroxide, as
manifested in the bleaching of dyes, is increased by exposing the reac-
tion mixture to light of short wave lengths.
This effect cannot be duplicated by substituting oxygen for hydrogen
peroxide in the reaction mixture, indicating that the result is not due to
formation of ozone.
EVANSTON, ILLINOIS.

[CONTRXBUTXON FaOM THE LABORATORY

OF GRINNELL
COLLEGG. ]
FURTHER WORK ON POTASSIUM HYDROGEN PHTHALATE
AS A STANDARD IN VOLUMETRIC ANALYSIS.
BY W. S. HENDRIXSON.
Received January 28, 1920.
Several years ago the writer showed that potassium hydrogen
phthalate and sodium hydrogen phthalate may be used with advan-
tage as standard in volumetric analysis.' They can be prepared
easily in a state of great purity, they have large molecular weights, 204.14
and 1X8.04, respectively, and O ~ onc Y hydrogen ion. The potassium
salt has the great advantage of absence oi water of crystallization.
little work has been done on these salts and so little is known of their
properties that it seemed desirable to study further the potassium salt,
which is to be preferred as a standard. The objects were: (I) to deter-
mine the practicability of the preparation and use of the salt as a standard,
by others; (2) to determine the necessary degree of purification for ordi-
nary purposes, and the yield on the basis of the phthalic anhydride used;
(3) to determine the hydroscopic character of the salt and the means
necessary to dry it.
The laboratory work on potassium hydrogen phthalate here described
wa.s done by Mr. Sereno Norton, then a senior in Grinnell College, now a
chemist with the Hercules Powder Company.
The potassium hydrogen phthalate was prepared by the method de-
scribed in the paper already cited. Several preparations were made and
titrated. There is one chance of error in its preparation, which might
easily be made by comparatively inexperienced hands, and which should
' %IS JOURNAL+, 37, 2352 ( ~ 9 1 5 ) ~
 ROTASSXUN HYDROGEN PHTHALATE I N VOLUMETRIC ANALYSIS. 725
be mentioned. Phthalic anhydride is very sparingly soluble even in hot
water. If not enough potassium carbonate is used in the neutralization
a portion of the anhydride will be left undissolved, and may easily be
overlooked, since the sparingly soluble acid phthalate is likely to form a
layer of crystals over the hot solution. Since the neutral potassium
phthalate is extremely soluble even in cold water there is no danger of its
contaminating the acid salt. Therefore, a slight excess of the carbonate
sliould be used, and the hot solution should be filtered in the first instance.
A large volume of standard hydrochloric acid was made by the admira-
ble and accurate method of Hulett and Ronner,’ so as to contain o .003647
g. of hydrogen chloride to one cc. This acid was titrated through the
medium of sodium hydroxide against benzoic acid specially prepared for
6wch purposes by the Bureau of Standards, which gave 0.003648 g. to
one cc. This concentration was confirmed by means of silver. The solu-
tion of sodium hydroxide, free from carbonate, was prepared as described
io the first paper, and was used as the intermediary between the acid
phthalate and the hydrochloric acid. ‘The indicator was phenolphthalein,
and as pointed out in the former paper its error in marking the end-point
was cancelled by the method of titration. The same weight burets and
csJibrated flasks described in the former paper were used in this work.
Titrations were made of samples of potassium hydrogen phthalate
which had been 3, 4 and 5 times recrystallized from hot water. ’The ob-
jects were to determine the number of crystallizations necessary to se-
cure a sufficiently pure product, and to determine the constancy of the
acid salt as an individual compound. It is well known that some so-called
acid salts do not show constancy in composition on being repeatedly crys-
tallized. The results are given in the following table in which “three,”
“four” and “five” indicate the number of recrystallizations, and below
each are the values given for the standard hydrochloric acid, assuming
the formula value of the acid phthalate.
&valuation of 0.1 N HC1 with WKC&IpOd through Standard NaOH.
Numbers of titrations. Three. Four. Five.
( I ) ......................... 0.0036489 0.0036470 0.0036519
(2)..~ ...................... 0.0036473 0.0036475 0.0036455
(9) ......................... 0.0036484 0.003644g 0.0036448
(4) ......................... 0.0036480 0.0036450 0.0036410
( 5 )......................... 0.0036471 ..... .....
(6)......................... 0.0036475 ..... . I . * .

(7) ......................... o ,0036478 ..... .....

(8)......................... 0.0036463 . . I . . .....
(9)......................... 0.0036463 ..... .....
(xo) ........................ 0.0036455 ..... .....
Averages: 0.003647 0.003646 0.003646
? h S JOURNAL. 3x9 390 (1909).
726 W. S. HENDRIXSON.

The results seem to show that there is nothing gained by more than 3
recrystallizations. Other results not here given seemed to show that for
ordinary purposes even 2 recrystallizations, that is, 3 crystallizations in all
would suffice. There seems to be no reason to doubt the constancy of
the composition of the acid salt.
Working with no special precautions to secure a large yield our results
show that about 50% of the theoretical yield of the acid salt may be ob-
tained after 4 recrystallizations. 'I'hough expensive in war time, thanks
to recent improvements in its preparation phthalic anhydride of great
purity is again becoming abundant and cheap. Potassium carbonate
of great purity is not difficult to obtain, and therefore, potassium hydrogen
phthalate is one of the standards most easily and cheaply obtained.
We next directed our attention to the hygroscopic character of potas-
sium hydrogen phthalate. It is very desirable that a substance used as a
standard should have a low hygroscopicity. To test this character we
carried out 2 sets of experiments. The first consisted in heating the air-
dried salt in an electric oven at different temperatures, determining the
losses in weight and the effect upon the titration value of the salt. The
second consisted in exposing the substance thus dried in air of known
moisture content, and determining the amount of moisture taken up
from the air,
In carrying out the first set of experiments about 2 0 g. of the salt was
heated in the electric oven in a platinum dish. At the end of each period
the loss was determined and a portion of the substance was taken out
without loss and titrated. The losses and titrations should thus form
checks on each other. There were 11 such titrations, but since only
4 different temperatures were used, the results at the same temperature
are combined in the following table:
Effect of Heat on Potassium Hydrogen Phthalate.
Wt of salt Time Temp Loss Value of €IC1
21 5940 4 hrs. I10 o 0006 o 0036471
20 4956 2 hrs. I20 o 0024 o 0036469
16 0100 8 hrs. 130 0 0034 o 0036490
II 6910 8 hrs. 150 o 0006 0 0036473

It is remarkable that the hygroscopic moisture which should have

been eliminated at 110' is so small, one part in about 36,000. Doubt-
less a part of the loss at higher temperatures is to be ascribed to changing
atmospheric conditions, possibly to undetected losses in taking out por-
tions for analysis. Granting that all the apparent loss was due to water,
its total, calculated on the basis of the original weight of salt taken, i s
0.0088 g. or one part in about 2 5 0 0 , which is outside the limit of accuracy
of titration in ordinary work.
The hygroscopic character was further studied by exposing the salt re-
 T H E E F F E C T OF SOLVENT ON T H E ABSORPTION SPECTRUM. 727

maining, and dried as above, in an atmosphere of known moisture con-

tent. The platinum dish and salt were placed in a desiccator which con-
tained sulfuric acid so diluted as to give to the air above it a humidity
of p"/& which is not likely to be exceeded in the air of a laboratory. From
time to time during 47 days the dish and contents were weighed. The
gain or loss as compared with the previous weight never exceeded 0 . 0 0 2 g.
and a t the end of the period, summing up the gains and losses, there was
found a loss of 0.0001 g.
These 2 sets of experiments seem to show that the hygroscopicity of
potassium hydrogen phthalate is practically nothing, and in using it and
,weighing it attention need be directed only to the containing vessel. By
use of an open platinum crucible in weighing it, it is believed that this
source of error has been almost completely avoided.
In the preparation of potassium hydrogen phthalate it is desirable to
know its solubility in hot and cold water, and this was determined at 2 5
and 35 degrees and a t the boiling point of its saturated solution. At the
lower temperatures the method used was one devised by the writer.'
Por the solubility a t the boiling point the method of Pawlewski2 was used.
Solubility of Potassium Hydrogen Phthalate in Water.
Temp. Wt. solution. W t . acid phthalate. yo in sol.
At 25 degrees.. . . . . . . . . . . . . . . 2 I ,025 2.1531 IO.23
2 1 .a28 2,1553 10.25
A t 35 degrees . . . . . . . . . . . . . . . . 21.198 2.6859 12.67
A t boiling point. . . . . . . . . . . . . 9.18 3.3158 36.12

[CONTRIBUTION FROM THE HAVSMEYERCHEMICAL LABORATORY, NEW YORK

UNIVERSITY.
]
THE EFFECT OF SOLVENT ON THE ULTRA VIOLET ABSORP-
TION SPECTRUM O F A PURE SUBSTANCE.
BY F. 0. R x c ~ .
Received January 28, 1920.
In a previous paper3 some quantitative measurements were described
showing the absorption of acetone and its homologues in the pure liquid
condition. A large quartz spectograph fitted with a spectrophotometric
arrangement was used and by this means the extinction coefficient, e , is
given by e = I / d (log I , / I ) , where I, is the intensity of the incident light, I
the intensity of the emergent light, and d the thickness of the absorbing
layer. E may, therefore, be defined as the reciprocal of that thickness of
medium which reduces the intensity of the light to of its original value.
If G be the normality of the ketone (sp. gr. X (rooo/mol. wt.)) the molecular
extinction is given by M = e/c.
I Proc. Iowa Acad. Sci., 23, 31 (1916).
Rer., 3zY 1040 (1899).
Proc. Roy. SOC.(London),91A, 76 (19r4).