1280 KLOSKY, WOO A N D F L A N I G A N VOl.

49

form is obtained by precipitation of a cold alkaline solution with acetic

acid and is believed to have a lactoid structure. The red form is obtained
by precipitation of a boiling alkaline solution with acids and is believed
to have a p-quinoid structure.
2 . No other forms of fluorescein were obtained than the yellow lactoid
form and the red quinoid form; von Liebig's claims to the existence of
five different yellow forms must be in error.
3. The yellow lactoid fluorescein and the red quinoid fluorescein give
different yellow hydrochlorides.
4. The colorless diacetate, bisphenyl carbamate and hydrazide of fluo-
rescein have been prepared. These are derivatives of lactoid fluorescein.
5 . The red mono-ammonium salt and the red methyl ester and its
orange acetate have been prepared. These are derivatives of quinoid
fluorescein.
6. Dibromofluorescein, its hydrochloride and diammonium salt have
been prepared. These colored compounds are derivatives of quinoid
fluorescein.
7. Dibromofluorescein dibenzoate has been prepared. This is colorless
and a derivative of lactoid fluorescein.
8. Colorless eosin has been prepared. The introduction of bromine
has made the lactoid form stable and colorless.
9. Two colored hydrates of eosin have been prepared.
NEWYORK
ITHACA,

[CONTRIBUTION FROM THE MALONEY

MARTIN CHEMICAL LABORATORY OF THE CATHOLIC
UNIVERSITY OF AMERICA]

THE VAPOR-PRESSURE CURVE OF BENZOIC ACID

LEO P. L. Woo
BY SIMONKLOSKY, AND J. FLANIGAN
ROBERT
N O V E M B E2,
RECEIVED R 1926 M A Y10, 1927
PUBLISHED

Introduction
The determination of the vapor pressure of benzoic acid from 100
t o 180' was suggested by Zanetti.' The authors did not limit themselves
to this range, but have determined values from 50 to 250'. As two differ-
ent methods were used, one for each phase, this paper is divided into two
parts for convenience.
I.2 The Vapor Pressure of Benzoic Acid (Liquid) from 251 to 128.6'
The vapor pressure of liquid benzoic acid has already been determined
by K a h l b a ~ m . ~
He used a static method and his results'are very accurate
Zanetti, Ind. Eng. Chem., 16, 304 (1924).
* Extract from a thesis submitted in partialfulfilment of the requirements for the
degree of Master of Arts, by Robert J. Flanigan.
a Kahlbaum, Z . physik. Chem., 26, 603 (1898).
May, 1927 VAPOR PRESSURE OF BENZOIC ACID 1281

and consistent except a t the lower temperatures fiear the melting point,
121.T0.* I n the redetermination of these values we have used the iso-
teniscope of M e n z i e ~ . ~The only unknown factor in the method, pro-
vided constant temperature is maintained, is the reading of the mano-
meter, which can be done with the aid of a cathetometer.

Materials
The benzoic acid used was Bureau of Standards Sample No. 39b, which
is employed as a standard in calorimetry and acidimetry.
The mercury used in the gage and seals was washed, dried and redis-
tilled under a low pressure.
All thermometers were compared against standards and checked a t
the ice point and boiling point.
The manometer was a Pyrex tube 6 mm. in diameter and 1.5 meters
long, filled with the cleaned mercury and boiled out under an extremely
low pressure, obtained with the aid of a Hyvac pump, and then sealed off.

Results
Several series of determinations were made. The best of these can be
represented by the curve log P, = 30.172 - 4714 (1/T) - 6.720 (log T ) ,
where T represents absolute temperature.
The general agreement of the points with this equation is shown in
Table I.
TABLE
I
LIQUIDBENZOICACID
t , OC. P, obs. P,calcd Deviation, %
128.6 8.4 8.6 -2.3
134.5 11.8 11.6 1.7
148.0 22.2 21.9 1.4
158.2 34.0 34.3 -0.9
172.8 63.8 63.0 1.3
187.6 110.8 110.9 -0.1
199.7 166.8 167.7 -0.5
216.2 286.7 290.4 -1.3
233.0 478.8 482.0 -0.7
247.0 718.1 709.6 1.3

Discussion of Results
The results as a whole are fairly consistent and fall on either side of
the curve. They also agree well with those of Kahlbaum, which only
deviate from our equation on an average of 1%, except below 132', where
this deviation is 5%.
"International Critical Tables," McGraw-Hill Book Co., New York, 1926, p. 208.
Smith and Menzies, THISJOURNAL, 32, 1419 (1910).
1282 KLOSKY, woo AND E’LANIGAN Vol. 49
I1
Vapor Pressure of Benzoic Acid (Solid) from 50 to 121’
The vapor pressure of solid benzoic acid has also been determined from
60 t o 110’ by a dynamic method.6 Although Niederschulte’s results
are fair, they are not consistent and could not be used t o extrapolate to
the melting point, as the slope was uncertain.
The authors have employed the static method of Menzies17 which
turned out to be more difficult to operate, but is capable of giving results
with a high degree of accuracy.
Apparatus
A glass bulb of 733 cc. capacity, containing the solid benzoic acid, was
suspended in an oil-bath whose temperature was controlled by a regu-
lator to 0.1’. This bulb was sealed through a trap to two McLeod gages,
which in turn were connected to a Hyvac pump through a mercury seal.
A vacuum stopco-k was also included to admit nitrogen dried over phos-
phorus pentoxide.
The McLeod gages were made of capillary tubing, calibrated by the
Bureau of Standards. The capillary tubing was 20 cm. long and con-
tained 1 cc. The volumes of the bulbs were measured by filling with
mercury and weighing. They were 139.2 cc. and 10.94 cc., respectively.
The volume of the bulb which held the acid was determined by filling
with water and measuring in a graduate.
The volume of the system was calculated by admitting a known volume
of air and observing the increase in pressure. This method gave from
two determinations results which only differed by one part in 1500.
Manipulation
A measurement was made by evacuating the system and washing i t out
with dry nitrogen. A reading of the pressure (PI) a t room temperature
( T I )was taken; the temperature of the bath was raised to T2 and another
reading of the pressure (Pz) made and all were noted. These second ob-
servations were repeated until they became constant. PI and T I are
called the zero reading.
Calculations
Although Menzies says:’b “, . .an obvious application of the gas laws
enables one to compute the vapor pressure of the substance a t the final
temperature, provided in particular, that Dalton’s law may be assumed,”
no formula is given. With the aid of these assumptions we have derived
the following formula, where V, and V, represent the volumes of the system
and bulb, respectively: P, = P2 - Pl(TZ/Tl) +
V,/Va(P, P1)TXTl.-
6 Niederschulte, Dissertatiolz, Erlangen, 1903.
Menzies, (a) THISJOURNAL, 41, 1783 (1919); (b) 42, 2218 (1920).
May, 1927 VAPOR PRESSURE OF BENZOIC ACID 1283
Results on the Vapor Pressure of the Solid
Several determinations were made on the solid, using various initial
pressures. The vapor pressure was then calculated by the Menzies
formula and found to be on a straight line which is represented by the
equation log P, = 11.956 - 4409 (l/T).
Table I1 shows the deviation of some of the points from this line.
TABLEI1
SOLIDBENZOICACID
1 , T. P. obs. P,calcd. Deviation, %
60.1 0.0527 0.0525 0.13
80.4 ,308 .0303 .15
100.2 1.436 1.400 2.5
121.0 5.830 5.807 0.3
Deviation of Niederschulte's points
60 0.11 0.052 110
75.4 .I9 ,200 65
96 .91 1.014 - 10
110 3.74 2.992 24

Discussion of Results
Several points should be noted about the results on the solid: (1)
the very consistent agreement of all our values; (2) the difference in slopes
of the two curves gives a latent heat of fusion for benzoic acid as 32.8
cal., a value in good agreement with that of Mathews, namely, 34.8 cal./g.s
and that of Garelli and Montanari, 39.65 ~ a l . / g . ~
More than 200 observations were made on the solid, but only those of
final runs whose values are consistent were used t o determine the curve.
Two causes of error may occur: (1) the failure to wait long enough
for the system to reach equilibrium, which yields values that are too low;
(2) adsorption of gases in the bulb, causing a more serious error. The first
run, especially a t low pressure, is never reliable for this reason, and only a
zero reading checked after a run leads t o any consistent values.
As mentioned by Menzies, the ideal condition (for accuracy) is a pres-
sure of nitrogen slightly greater than the vapor pressure t o be measured.
But if i t is too close, rapid sublimation results with the blocking of the tube.
This method should find wider application in the measurement of
vapor pressures of solids, as it is more accurate and faster than the loss-
in-weight air-bubbling method, and makes no assumption as to the molec-
ular weight of the vapor.
Summary
The vapor pressure of benzoic acid has been measured from 250 to
50'.
Mathews, THISJOURNAL, 39, 1125 (1917).
Garelli and Montanari, Gaez. chim. ital., 24 [2], 229 (1894).
1284 W. R. ORNDORFF AND CHEN WANG VOl. 49

The values above 130’ have been found to be in agreement with

those of Kahlbaum. Two of the four values of Niederschulte lie near
the vapor-pressure curve of the solid, as determined by us.
The latent heat of fusion as calculated from the difference in slopes of
the two curves is in agreement with cryoscopic data.
WASHINGTON,
D. C.

[CONTRIBUTION FROM THE BAKERLABORATORY

O F CHEMISTRY, CORNELL UNIVERSITY]

PYROGALLOLBENZEIN AND SOME OF ITS DERIVATIVES

BY W. R. ORNDORFF
AND CHENW A N G ~
RECEIVED
NOVEMBER
13, 1926 PUBLISHED
MAY10, 1927

The method of preparation, purification and the analyses of pyrogallol-

benzein have already been given.2 The red aqueous extracts obtained in
the purification of the benzein2 were concentrated and on cooling gave
red crystals. Yellow crystals were obtained when these were crystallized
from water, after boiling the solution with bone black. Analyses3 showed
that this substance was 2,3,4-trihydroxydiphenylketone, known com-
mercially as Alizarin Yellow A. The triacetate, sodium and lead salts
were also made and analyzed3 and proved that this by-product obtained
in the preparation of the benzein was 2,3,4-trihydroxydiphenylketone.
The pyrogallol probably first reacts with the benzotrichloride to form
trihydroxydiphenyldichloromethane, Cd&(OH)3 +
C6H&C13 = C6HZ-
(OH)3CC12C6Hjf HC1, and this, on boiling with water, gives 2,3,4-tri-
hydroxydiphenyl ketone, C ~ H Z ( ~ H ) ~ C C ~ 4- Z HzO
C ~ H=~ CsH%(OH)3-
+
COC6HB 2HC1. The formation of the pyrogallolbenzein itself prob-
ably takes place as follows.
/O\
CeHz(OH)aCC12Ce,Hs + CaHa(OH)3 = Ce,Hz(OH)zCCIC6Hz(OH)z
I
+ HCl HzO

/O\
C~HZ(OH)ZC=CGHZ(OH)=O.HC1
\
+ H?O = CBHZ(OH)ZC=C~H~(OH)=O
CaH5 + H20 + HCI
\ I
C6H5
Pyrogallolbenzein
1 From a dissertation submitted by Chen Wang to the Faculty of the Graduate

School of Comell University, in partial fulfilment of the requirements for the degree
of Doctor of Philosophy.
2 Omdorff and Wang, THIS JOURNAL, 47, 290 (1925).
3 See Thesis, Cornell University Library. International atomic weights for 1925
were used in all calculations.