1116 T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y Vol. 11, No.

12

SUMMARY
I-A convenient and rapid method was devised for
comparing t h e film-forming properties of possible
anti-dimmers. By this niethod one observer made
56 tests in I j minutes,
11-Organic sulfates as a class are not effective;
the presence of t h e sulfonic acid group does not guar-
antee effectiveness; t h e position of sulfonic acid groups
is of little importance; sodium benzene trisulfonate
is less effective t h a n t h e di- or monosulfonate; naph-
thalene ring sulfonates are no better t h a n benzene
ring sulfonates; t h e amino and hydroxyl groups are
of no material value.
111-High molecular weight is not a vital factor as
shown by poor results with Twitchell’s benzene stearo-
sulfonic acid and similar compounds. Since moisture
should condense in films, rather t h a n in drops, for good
visibility, t h e best anti-dimmers should lower t h e sur- 100 parts sodium salt of 65 per cent sulfonated rape oil, 35
per cent water
face tension of water greatly. They must be soluble, Sulfonated 20 parts sodium hydroxide
but not too soluble or they have no staying powers: Rape Oil 2 5 parts water glass syrup
3 parts glycerol
They should also be deliquescent enough t o prevent 3 parts marine oil
drying out. . Dried to form a solid stick when compressed
IV-Alkali soaps of unsaturated f a t t y acids are
much superior in film-forming power t o soaps of sat- 15 parts sodium hydroxide
urated f a t t y acids. 3 parts water glass syrup
V-Soaps in general are improved b y t h e addition __
3 -Darts alvcerol
I 3 parts marine oil
of a few per cent of glycerol, I O t o I j per cent sodium [ Dried t o form a solid stick when compressed
hydroxide and a few per cent of rnarine oil. Water OBERLINCOLLPIGE
glass in small amount serves well as a binder. OBERLIN,OHIO

ORIGINAL PAPERS -7
PHTHALIC ANHYDRIDE. 11-THE MELTING POINT OF and theoretical considerations led t o t h e present in-
PURE PHTHALIC ANHYDRIDE. THE SYSTEM: vestigation, with t h e object of establishing t h e melt-
PHTHALIC ANHYDRIDE-PHTHALIC ACID ing point of pure phthalic anhydride and also of ob-
By K. P. MONROE serving t h e melting points of known mixtures of phthalic
Reccived September 13, 1919 anhydride with phthalic acid, t h e impurity most
An interesting consequence of t h e recent develop- likely t o be found in t h e technical anhydride as well
ment by Gibbs and co-workers’ of a catalytic process as in t h e samples examined by t h e earlier observers.
for air oxidation of naphthalene t o phthalic anhydride Although numerous titrations of re-sublimed phthalic
has been t h a t t h e consistent melting point ( 1 3 1 ~ ) anhydride derived from t h e air oxidation process
of t h e purified product2 exceeds by three degrees t h a t indicated a high degree of purity, i t was first deter-
previously recorded in t h e literature.3 Both practical mined t o adopt a still more rigorous purification. This
material’ was warmed in a vacuum oven for 5 hrs.
1 See first article of this series, “Phthalic Anhydride, I,” by H. D.
Gibbs, THISJOURNAL, 11 (1919), 1031. a t g o o , fused carefully in a casserole and decanted
* Phthalic anhydride derived from both laboratory and factory opera- into a mortar. For t h e purpose of decomposing possi-
tion of the air oxidation process was tested in the usual capillary tube ble traces of phthalic acid and separating any non-
manner by various workers. In each case the crude product was purified
by re-sublimation; the mateiial was chemically pure so far as could be volatile impurities, t h e pulverized anhydride was
detected by titration. twice re-sublimed in vacuo in t h e presence of phos-
a Reilstein, Handbuch d . org. Chem., 2 (1794) (3rd Ed.), gives 1.78’.
with references t o the following:
phorus pentoxide. The apparatus designed for this
Laurent, Liebig’s Ann. d . Chem., 19 (1836); 38, gives 105’, presumably sublimation is illustrated in Fig. I . A glass tube, A ,
an the Reamur scale (131’ C . ) . Phthalic anhydride was prepared from sealed a t one end and widened t o thrice its diameter
phthalic acid by one sublimation a t atmospheric pressure.
Lossen, I b i d , 144 (1867), 76, gives 128’. Phthalic anhydride was a t t h e other was fitted co-axially inside a longer tube,
prepared from phthalic acid by sublimation a t atmospheric pressure. This B, also sealed a t one end, with a ground glass joint.
melting point was verified by testing phthalic anhydride from phthalic Through a phosphorus pentoxide bulb, C, t h e system
acid ,prepared by various methods, which are not enumerated.
Anschiitz, Ber., 10 (1877), 326, gives 127’. Phthalic acid was de- was evacuated t o a pressure of approximately four
hydrated by acetyl chloride, which was removed by absolute alcohol; millimeters by a vacuum pump. On warming B,
crystals of phthalic anhydride were pressed between filter paper.
Lachowicz, Ibid., 17 (1884), 1283, gives 128O. Phthalyl chloride was
phthalic anhydride contained in t h e lower portion
warmed with lead nitrate; phthalic anhydride was recrystallized from vaporized a n d after filtering through a tight plug
benzene. of glass wool was condensed on t h e lower portion of
Stohmann, J. pvakt. Chem., 40 (1889), 139, gives 128’. Phthalic
anhydride prepared by distillation of commercial phthalic acid was re- A, which was cooled by a n air stream.
crystallized from benzene-ligroin mixture. 1 Thi; crystallized in long, colorless glistening needles.
Dec., 1919 T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 1117

Results of a large number of sublimations indicated in 5 0 cc. of acetone, a few drops of a solution of brom-
t h a t t h e only detectable impurities were occasional phenol blue1 are added, and t h e acetone solution is
minute traces of t a r filtered out by t h e glass wool, titrated with M / I O solution of normal potassium
and a trace of nonvolatile residue. Titration and phthalate in 80 per cent alcohol. If phthalic acid is
combustion' established t h e chemical purity of t h e present the end-point of t h e titration is reached when
final sublimate. N o difference between t h e first sufficient normal phthalate solution has been added
and second vacuum sublimate could be detected just t o produce t h e acid phthalate, and further addi-
either b y analysis or b y melting point test in t h e ac- tion of normal phthalate solution causes a color change
curate manner t o be described later. t o indigo blue. The original color in t h e acetone
solution is pale yellow, while t h e tint a t t h e end-point
is greenish blue. Unfortunately, t h e end-point is
not sharp, since i t is difficult t o reproduce t h e exact
shade of green. Nevertheless, this is a very sensitive
test for t h e absence of traces of phthalic acid in t h e
anhydride. When t h e second sublimate was tested
in this manner, a single drop (0.03 cc.) of normal
phthalate solution caused a color change t o indigo-
blue, showing t h a t t h e end-point must have been ex-
ceeded, and t h a t this sample of anhydride must have
contained less t h a n 0.1 per cent acid.
To determine t h e melting point of t h e second subli-
mate, approximately 15 g. were placed in a wide test
tube, which was closed by a stopper in which were
inserted a Reichsanstalt "Centigrade-Normal" short
stem thermometer2 and a glass guide t u b e for t h e
platinum stirring rod. The test tube was placed in
a snugly fitting stoppered glass air-bath3 clamped
vertically in a liter sulfuric acid bath, which was
stirred b y a turbine. After fusion t h e temperature
was lowered slowly t o t h e point of incipient crystalliza-
tion while t h e melt was stirred constantly b y a me-
chanical device. The equilibrium temperature of
crystals and liquid was 1 3 0 . 8 4 O . ~ A melting point
identical within experimental error was obtained under
similar conditions for t h e original anhydride which was
FIG.1-VACWWM SUBLIMATION APPARATUS t h e source of t h e carefully purified material, confirm-
ing t h e previous conclusion t h a t no more t h a n traces
It will be noted, however, t h a t t h e method of total2 of impurities were contained in this.5
titration leaves much t o be desired as a test for the I n an effort t o search out sources of discrepancy
presence of minute amounts of phthalic acid in t h e between t h e values previously recorded in t h e litera-
anhydride, for t h e difference between t h e molecular ture and t h e melting point of pure phthalic anhydride,
weight of phthalic anhydride (148.07) and t h a t of the which seemed now definitely established, samples of
acid (166.09) is so slight ( 1 8 . 0 2 ) t h a t with usual analy- Kahlbaum's "Phthalsaure Anhydrid"6 were ex-
tical accuracy this method cannot detect with cer- amined in t h e usual capillary tube manner. I t was,
tainty less t h a n approximately I per cent of acid. indeed, observed t h a t t h e material began t o sinter
d number of attempts at digerential titration of acid a t 128-12g0, t h a t a liquid meniscus appeared near
in t h e presence of anhydride, in which mixtures were 1 3 1 O , b u t t h a t t h e solid phase did not completely
dissolved in nonaqueous solvents and titrated with disappear until 178". On examining a larger quantity
N / I O NaOH, indicated t h a t t h e anhydride is rapidly in the manner described for t h e pure anhydride a
hydrated in t h e presence of excess alkali. The fol- well defined freezing point was observed a t 15g.8", but
lowing method was finally devised by E. Q. Adams,3
1 This is the abbreviated name for tetrabromphenolsulfonaphthalein.
of this laboratory, and is a remarkably sensitive test 2 This was graduated t o fifths of degrees, and recently calibrated by the
for traces of phthalic acid in t h e anhydride: Bureau of Standards.
8 For description see Phthalic Anhydride, 111, the accompanying third
A half gram sample of t h e anhydride is dissolved
article of this series (p. 1119).
10.5320 g . and 0.5956 g. sublimate required 7.206 and 8.042 mllli- 4 Under these conditions no correction is necessary, for the exposed
moles "OH. Calculated weights anhydride: 0.5333 g. and 0.5952 g. stem was but 15", and the portion of thermometer not submerged in the
0.3233 g. sub1imat.e gave on combustion 0.6765 g. COz and 0.0905 g. melt was surrounded by air and vapor in the inner tube a t a temperature
1 1 2 0 . Per cent C = 64.8, per cent H = 3.1. Calculated. Per cent differing but slightly from t h a t of the melt.
C = 64.9, per cent H = 2.7. 6 Through the courtesy of Prof, G. A. Hulett the writer obtained a
2 This method involves the addition of excess N / 5 NaOH t o weighed sample of phthalic anhydride prepared by sublimation from pure phthalic
samples of the anhydride in glass stoppered flasks, which are warmed on acid which was dehydrated in vacuo a t 115'. This anhydride melted a t
the steam bath for an hour. T h e excess alkali is then titrated by A7/S 131' when tested in the capillary tube.
HCI. 6 Presumably obtained by sdblimation from phthalic acid derived by
a Dr. Adams is continuing t h e investigation of this and other methods one of the "wet" oxidation methods, which was the universal method of
for differential titration of phthalic acid in the presence of the anhydride. preparation before development of the air oxidation process.
I118 T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY Vol. 11, No. 12

after this lag t h e temperature fell steadily with con- TABLEI-FREEZING POINTS OF MIXTURESOF PHTHALIC ANHYDRIDE
AND
PHTHALICACID
tinual separation of crystals t o 129.6”,where t h e ap- WEIGHT MOL.
PER CENT PER CENT FREEZING
pearance of a distinctly different solid phase was ob- ACID ACID POINT
served and the thermometer halted until solidifica- Solid Phase: Anhydride
0.0 0.0 130.84
tion was complete. Since the impurity most likely 0.29 0.26 130.7
0.98 0.87 130.4
in phthalic anhydride prepared by sublimation from 1.44 1.30 130.1
Eutectic
phthalic acid is t h e acid itself, i t seemed probable 1.96 (interpolated) 1.75 (interpolated) 129.74
t h a t the Hahlbaum material contained a considerable Solid Phase: Acid
2.02 1.81 130.0
admixture of phthalic acid, and t h a t 1 2 9 . 6 ’ would 2.44 2.18 131.9
lie near t h e eutectic temperat cid and anhydride. 2.70 2.41 133.2
3.02 2.70 136.3
Analyses1 confirmed t h e conclusion, and 3.47
3.91
3.10
3.50
137.2
139.7
subsequent determination of t h e eutectic temperature, 4.27 3.83 141.1
4.80 4.30 143.8
using pure materials, gave t h e value 129.74’. 5.57 5.00 146.4
To obtain additional evidence upon the melting
hydride. It may be pointed o u t t h a t t h e capillary
point of pure phthalic anhydride, the Kahlbaum
tube method is a sensitive criterion for presence of
product was twice re-sublimed in the vacuum apparatus.
acid in t h e anhydride if the temperature a t which solid
T h e second sublimate was observed t o have a constant
acid completely disappears is noted, for in t h e capillary
freezing point a t 1 3 0 . & ~ . ~
tube solid acid usually survives much above the true
Since the behavior of Kahlbaum’s “Phthalsaure melting point of t h e mixture.
Anhydrid” confirmed t h e previous conclusion t h a t
presence of phthalic acid is a factor which must be
cons’idered in observations upon anhydride prepared
by sublimation from t h e acid, mixtures of t h e pure
anhydride with very pure phthalic acidg were weighed
into t h e t e s t ’ t u b e and t h e freezing points were de-
te.rmined as for t h e pure a n h ~ d r i d e . ~
unfortunately t h e limitations of this method were
such t h a t only a very restricted region of this system
could be explored, for, as will be observed from t h e
steep slope of the right-hand curve in Fig. 2, the tem-
peratures necessary t o fusion soon became so high t h a t
satisfactory Checks cobld not be obtained. Only those
experiments in which close duplicate freezing points
could be obtained are noted in the graph a n d table.
Although i t was not feasible t o seed t h e melt, very
great care was taken t o lower the temperature ex-
ceedingly slowly5 in t h e neighborhood of t h e freezing
point while t h e walls of t h e tube were continually
scratched by t h e platinum stirring rod, operated by
a mechanical device. 0 2 4 ’ Percent
I n t h e light of these d a t a i t is obvious t h a t any 0 2.24 4.46 6.68 Weiqhf per cent
mixture of pure phthalic anhydride with the pure acid Per cent Acid
will, when tested in t h e capillary tube, begin t o sinter FIG.
2-PREEZING-POINT CURVES FOR S Y S T E M PHTHALIC ACID-PHTHALIC
ANHYDRIDE
a t t h e eutectic temperature (12g.74’), which may
obviously be depressed by other impurities in the an-
It is of interest t o note the bearing of t h e results
10.6165 g. and 0.4423 g. Kahlbaum’s “Phthalsaure Anhydrid” re- in Table I upon t h e melting point of pure phthalic
quired 8.202 and 5.902 millimoles NaOH. calculated weights anhydride:
0.6089 g. and 0.4368 g. Calculated per cent phthalic acid: 10.0 and 10.3. acid, for on account of t h e readiness with which this
0.2891 g. Kahlbaum’s “Phthalsaiire Anhydrid” gave on combustion decomposes near its melting point into the anhydride
0.6729 g. COz and 0.0729 g. Hz0. Per cent C = 63.7, per cent H = 2.8.
Calculated for mixture containing 10 per cent phthalic acid: per cent
and water, the melting point may be approximated only
C = 64.1, per cent H = 2.8 roughly b y direct measurement.’ Direct extrapola-
* 0.5126 g. and 0.6465 g. sublimate required 6.936 and 8.748 milli- tion of t h e right-hand curve in Fig. 2 is not feasible,
moles NaOH. Calculated weights anhydride: 0.5134 g. and 0.6475 g.
a This was prepared by boiling the pure anhydride with distilled water, 1 T h e early impression t h a t phthalic acid melted a t 184’ (Lossen,
subsequent recrystallization from hot water, and thorough drying at room Loc. c&), near which its water vapor tension reaches atmospheric pressure
temperature i n bacwo over phosphorus pentoxide. and the acid decomposes rapidly to a mixture containing anhydride, was
0.7694 g. phthalic acid so prepared required 9.268 millimoles NaOH. first corrected by Ador, A n n . , 163 (1872), 230, who states t h a t pulverized
Calculated weight acid: 0.7697 g. phthalic acid melts a t 203”, while whole crystals melt a t 213O.
4 For discussion of the applicability of this method t o mixtures of I n an attempt t o verify Ador’s results a quantity of very pure phthalic
substances which sublime readily at the melting point see Phthalic An- acid was sealed in an evacuated glass tube, which was attached to a ther-
hydride, 111, the accompanying third article of this series. No ap- mometer and immersed in a paraffin bath which was vigorously turbined.
preciable sublimation was observed during the composition range given. On heating t h e bath very slowly, in order t o prevent thermal lag, the ma-
* This procedure was especially necessary in determining the freezing terial in the tube was observed t o melt a t 199-200°. This must, however,
points on the right-hand curve in Fig. 2, for since each mixture contained be regarded as merely a rough minimum value, for a t this temperature t h e
a relatively minute mol. fraction of acid, separation of this as solid phase water vapor tension exceeds one atmosphere and the melt must contaix
caused a rapid change in composition of the melt. very appreciable amounts of dissolved anhydride and water.
 T B E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C E E M I S T R Y 11’19
Dec., 1919

but if one plots on a large scale t h e reciprocals of t h e near the upper end and SO extended t h a t when t h e
absolute freezing points against t h e logarithms of t h e air-bath was clamped vertically in a liter sulfuric acid
mol. fractions of phthalic acid a straight line results, bath t h e test tube was entirely below t h e acid level.
from which t h e melting point of pure phthalic acid is
indicated t o be 2 0 8 O * 2’. This linear relation is
predicted on t h e assumption t h a t t h e freezing-point
curve conforms t o t h e law deduced for ideal systems,
for it follows from Equation I in the-accompanying
third article of this series t h a t

T
I _--_-
- I
To
log10 x
C
when
c = 0.4343 L R
and T = freezing point of t h e mixture, To = freezing
point of pure phthalic acid, x = mol. fraction of
phthalic acid in t h e mixture, L = molecular heat of
fusion of phthalic acid, and R = t h e gas constant.
SU Y MA R Y
The melting point of pure phthalic anhydride has
been determined ,to be 130.84’, and t h e freezing
curves for t h e system: phthalic anhydride-phthalic
acid have been established within t h e limits of compo-
sition for which t h e experimental method is applicable.
By graphical extrapolation t h e melting point of pure
phthalic acid is indicated t o be 2 0 8 ’ * 2 ” .
COLORLABORATORY
U. 5. BUREAUO F CHEMISVRY
WASHINGTON, D. C .
--
PHTHALIC ANHYDRIDE. 111-THE SYSTEM: NAPH-
THALENE-PHTHALIC ANHYDRIDE
B y K. P. MONROE
Received September 13, 1919
Investigation of t h e freezing-point curves of this
system was undertaken in order t o assist the solution
of certain problems arising in fractional sublimation
of t h e mixture of naphthalene and phthalic anhydride
which is t h e crude product of t h e air oxidation process
developed b y Gibbs and co-w0rkers.l Apart from
their practical value, t h e results are of interest in t h a t --
they offer striking conformation t o t h e familiar thermo-
dynamic law which determines t h e freezing-point FIG.1
curves of ideal systems, while the following experi- Since t h e air-bath was also stoppered, with b u t a
mental method which has been devised would seem narrow opening t o t h e atmosphere through a glassI
particularly well adapted t o t h e investigation of similar guide tube, practically no sublimation was observed
systems of substances which sublime readily a t t h e throughout t h e entire range of mixtures examined;
melting point: for t h e test tube was thus surrounded by air a t a tem-
Mixtures of pure naphthalene2 and phthalic anhy- perature differing b u t slightly from t h a t of the melt,
dride3 were weighed into a wide test tube. After and there was practically no convection current through
closing t h e tube by a stopper in which were inserted the two guide tubes.l
a “TYCOS” short stem thermometer4 and a glass guide
After fusing t o a clear, homogeneous liquid, t h e
tube for t h e platinum stirring rod, the whole was
temperature was permitted t o fall slowly t o t h e point
lowered into a cylindrical glass air-bath which fitted
of incipient crystallization while the melt was stirred
snugly t h e length of the test tube, b u t was widened
constantly by a mechanical device. The freezing
1 THIS
JOURNAL,
11 (1919), 1031.
point was determined in the usual manner from t h e
* The BUredU of Standards’ purest naphthalene was used. Compare
the observed melting point, 80.05’, with 80.09’, the value obtained by time-temperature curve. Although it was not feasible
Washbnrn and Read, PYOC. Nnt. Acad. Sci., 1 (1915), 191. t o seed the melt, care was taken t o lower t h e tempera-
a The fused phthalic anhydride described in the accompanying second
article of this series, Phthalic Anhydride, 11, was used. ture exceedingly slowly in the neighborhood of the
This was graduated t o tenths of degrees, and calibrated by com- 1 The vapor required t o saturate the free space was negligible, for the
parison with a Reichsanstalt “Centigrade-Normal” thermometer. mixtures averaged 20 g. in weight and the free space was less than 60 cc.