906 C. M. CARSON Vol.

48

weight from the total ferrosilicon added. The weight of the lower chlor-
ides, most of which was disilicon hexachloride, was found from the residue
after distilling out the silicon tetrachloride.
An inspection of the table shows that the yield of lower chlorides can be
greatly increased by the addition of silicon tetrachloride to the chlorine up
to a certain point. Beyond that, a further addition reduces the yield which
indicates that the first increase is due to a retarding of the final chlorina-
tion and not to the reduction of silicon tetrachloride by silicon.
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

The authors are indebted to the Chemical Warfare Service for the supply
Downloaded via UNIV NACIONAL DE COLOMBIA on August 30, 2019 at 21:54:28 (UTC).

of silicon tetrachloride used.

Summary
1. The ferric chloride formed by the chlorination of a mixture of iron
and silicon does not act as a catalyst to increase the yield of lower chlorides
of silicon nor to lower the temperature at which silicon is chlorinated.
2. The yield of disilicon hexachloride can be increased by passing sili-
con tetrachloride with chlorine over ferrosilicon heated to 200°.
3. This increase is due to the retarding of the final chlorination of di-
silicon hexachloride to silicon tetrachloride and not to the reduction of
the latter by silicon.
Ames, Iowa

[Contribution from thb Michigan College of Mines]

BASIC STANNOUS SULFATE
By C. M. Carson
Received December 8, 1925 Published April 5, 1926

It is well known that water decomposes stannous salts with the forma-
tion of basic salt precipitates and that sodium hydroxide forms similar
precipitates from solutions containing stannous ion. The composition
of the precipitates for stannous chloride has been determined by Carson1
and it seemed desirable that the basic salts of stannous sulfate should like-
wise be examined.
The stannous sulfate samples obtained from two prominent American
manufacturers contained such a large amount of insoluble material that
they were quite unsuitable either for direct use or as raw material for the
manufacture of a pure product. A good quality of the salt may be made
from tin and sulfuric acid but the following method is better for small
scale operation.
About 25 g. of cupric sulfate is added to 200 cc. of water and 10 cc. of sulfuric acid
in a 500cc. flask, and after the solution has been heated to boiling, small fragments
of tin foil are dropped in, until 20 to 25 g. has been used. With continued boiling, the
1
Carson, This Journal, 41, 1969 (1919).
April, 1926 BASIC STANNOUS SULFATE 907

solution loses its blue color in less than half an hour and while still hot is filtered from the
precipitated copper. The filtrate is evaporated on the water-bath until only about 10
cc. of liquid remains and a deposit of pure white crystals is obtained; but if the crystal-
line material shows a yellow tinge, it should be rejected. The liquid is decanted and
20 to 50 cc. of methanol is added to the residue, the crystalline stannous sulfate is
filtered off, washed with alcohol and then dried in a steam oven. The product is pure
white and dissolves completely in water.
In each experiment of the first series to be recorded here, 5 g. of stannous
sulfate was placed in a 50cc. flask and to it was added 10 cc. of sodium hy-
droxide solution, the concentration of which was varied by taking from 1
to 10 cc. of a 3 IV solution and diluting with water to 10 cc. The corked
flask was suspended in boiling water and shaken for one hour which was
found to be sufficient for the attainment of equilibrium and the resulting
mixture, while still hot, was filtered quickly. The precipitate without
being washed or even thoroughly drained was enclosed in the filter paper
which was then pressed between successive papers until nearly all of the
liquid was removed, and drying was completed in a steam oven. Ob-
viously, the mother liquor cannot be totally removed in this way but, as
will be shown later, the amount remaining and the composition of it may
be determined.
In the analysis of the dried precipitate for tin and the sulfate radical,
the substance was ground, 0.5 g. was dissolved in hydrochloric acid, the
solution was made alkaline with ammonium hydroxide and then acid with
acetic acid. The tin was precipitated as stannous sulfide and weighed as
the dioxide; the determination of the sulfate radical was made in the
filtrate from the stannous sulfide.
The results of the first experimental series are given in Table I.
Table I
Interaction of Five Grams of Stannous Sulfate in Ten cc. of Water and 0.12
1.2 Grams of Sodium Hydroxide
Water, 3 N NaOH, Sn in ppt., SO* in ppt.,
Expt. cc. cc. % %
1 9 1 56.96 41.71
2 8 2 59.93 35.83
3 7 3 63.78 31.99
4 0 4 67.68 27.66
5 5 5 67.82 27.37
e 4 6 68.02 26.63
7 3 7 68.48 23.85
8 2 8 69.39 20.46
9 1 9 69.69 20.21
10 0 10 70.60 18.83

In each experiment of the second series, 5 g. of stannous sulfate was

again used but the volume of 3 N sodium hydroxide solution varied from
11 to 15 cc. and the total volume was made up to 20 cc. with water,
908 C. M. CARSON Vol. 48

Table II
Interaction of Five Grams of Stannous Sulfate in Twenty cc. of Water and
1.32 2.1 Grams of Sodium Hydroxide
Water, 3 N NaOH, Sn in ppt., SO, in ppt.,
Expt. cc. cc. % %
11 9 11 73.52 15.61
12 8 12 76.05 11.80
13 7 13 78.05 9.79
14 6 14 81.48 7.03
15 5 15 83.29 3.56
In Fig. 1, the percentages of sulfate as given in Tables I and II are plotted
as ordinates and the number of cubic centimeters of sodium hydroxide as
abscissas. It will be seen from the horizontal portions of the curve and
also from the percentages in the tables that in each pair of Expts. 4, 5 and
8, 9, the precipitates have nearly the same composition. Since the solid
phases retained unknown amounts of mother liquor, however, it was not

Fig. 1.
certain just how closely alike they really were. Therefore, it was neces-
sary to carry out the experiments in question with the same proportions
of substances as before but by a method that would give not only the
amount of liquid phase left in the solid but also the appropriate corrections,
based on the analyses of the mother liquors, to be applied to the analyses
of the precipitates. Where washing with water, or other solvent, alters
the basic salts to such an extent as to invalidate the results, two different
April, 1926 BASIC STANNOUS SULFATE 909

procedures for analyses have been proposed, either of which would answer
the present purpose.
Kenrick2 adds to the original mixture a substance, called a tell-tale,
that will take no part in the reaction and so will not enter the true solid
phase. By analysis, the amounts of this substance both in the impure
precipitate and in a weighed portion of the liquid phase are determined and
in this way the weight of liquid retained by the solid is calculated.
Schreinemakers8 analyzes both the liquid phases and the partially separated
solid phases for different original mixtures, plots the results with reference
to rectangular coordinates and draws a straight line through the two points
belonging to each mixture. A point where two or more of these lines inter-
sect represents the composition of a compound. Kenrick’s method, with
ammonium sulfate as the tell-tale substance, was first tried but was
abandoned in favor of another plan. A description of this process as ap-
plied to Expt. 4 will now be given.
Five g. of stannous sulfate was heated as before with 4 cc. of 3 N sodium
hydroxide solution and 6 cc. of water. The mixture, while still hot, was
filtered and 1.5 cc. of the filtrate was pipetted into a weighed bottle which
was then weighed again. The bottle was placed in a steam oven and after
evaporation of the mother liquor, the weight was again taken and the
remaining solid was analyzed for tin and the sulfate radical. The main
precipitate was dried as thoroughly as possible between filter papers and
while still slightly moist was transferred to a weighed bottle and
the weight was determined. After being completely dried at 100°, the
precipitate was weighed and analyzed.
The following are the data for the repetition of Expt. 4:
1.5 cc. of Loss on Mother liq. rest
Ppt., moist Dry Loss mother liquor evapn. Residue due in ppt.
Wt.,g. 2.0283 1.8843 0.1440 1.8691 1.3832 0,4859 0.0506

By analysis, 0.4859 g. of mother liquor residue was found to contain

0.1987 g. of tin and 0.2396 g. of sulfate radical. Therefore, 0.0506 g. of
the same residue, which was the amount retained by the solid phase, would
contain 0.0207 g. of tin and 0.0249 g. of sulfate radical. The total dried
precipitate contains 1.2752 g. of tin and 0.5210 g. of sulfate radical, so
that the true basic salt is made up of 1.2545 g. of tin and 0.4961 g. of sul-
fate radical. The ratio, molecules of sulfate to atoms of tin, from the
corrected analysis is 1:2.047 and, therefore, the precipitate has very nearly
the composition SnSCh.SnO.
A similar procedure for Expt. 5 gave the atomic ratio 1 to 2.067 of tin,
while in a new experiment with 5 g. of stannous sulfate, 4.5 cc. of 3 N
2
Kenrick, J. Phys. Chem., 12, 695 (1908).
3
Schreinemakers, Z. physik. Chem., II, 76 (1893).
910 C. M. CARSON Yol. 48

sodium hydroxide solution and 5.5 cc. of water, the ratio 1 of sulfate to
2.047 of tin was obtained. It appears, then, that the precipitate SnS04.-
1.047 SnO remained constant when the amount of sodium hydroxide was
varied by about 12%. A modification of the experimental method will
now be described to support the conclusion that a compound of approxi-

mately that composition is formed.

Five g. of stannous sulfate was thoroughly mixed in a flask with a boiling
solution containing 5 cc. of 3 A sodium hydroxide solution and 45 cc. of
water and the mixture was allowed to settle. After a few minutes, the
clear mother liquor was decanted into a clean, dry flask and was cooled
slowly. A white crystalline deposit formed, which was separated by filtra-
tion and dried between filter papers. Analysis showed the crystals to
have the composition SnSnO4.0.964SnO. A similar preparation that was
washed thrice with water and then dried had the composition SnS04.-
0.982SnO. Similar crystalline precipitates with slightly varying com-
positions were obtained from a mixture of 10 g. of stannous sulfate,
5 cc. of 3 A sodium hydroxide solution, 10 cc. of water and also from 10 g.
of stannous sulfate, 5 cc. of 3 A sodium hydroxide solution, 20 cc. of water.
No corrections were made for the adhering mother liquor as the time for
this was not available, but undoubtedly the true composition in each of
the three cases just described is close to SnSCh.SnO.
Much greater uncertainty exists in applying the corrections for retained
liquid phase in Expts. 8 and 9 than in 4 and 5, for which two causes are
responsible. In the first place, the precipitates in Expts. 8 and 9 were
not so well crystallized as were those in 4 and 5 and, therefore, retained more
liquid; and, in the second place, the liquid in Expts. 8 and 9 differs much
more from the solid than that in Expts. 4 and 5. Thus in Expt. 4 the uncor-
rected and corrected ratios tin to sulfate were 1.980 and 2.047, respectively,
while in Expt. 8 the uncorrected and corrected ratios were 2.700 and 3.017.
It will be seen that the correction in Expt. 8 is more than four times that
of Expt. 4, and consequently the error in calculating the true composition
of the solid is greater.
Expts. 8 and 9 were repeated and corrections such as those introduced
for Expts. 4 and 5, were made. The modified analyses showed that the
precipitates were respectively SnSO4.2.017SnO and SnSO4.2.098SnO with
undetermined amounts of water. While these results do not warrant the
conclusion reached by Ditte4 that a pure compound SnS04.2Sn0.1.5H20
exists, nevertheless, they indicate the probability.
It may appear that the corrections introduced in the results of Expts.
4, 5, 8 and 9 are based on a faulty method since the amount of water
driven off when the moist precipitates are dried exceeds the mere water of
solution in the adhering mother liquor. In the experiments mentioned,
4
Ditte, Ann. chim. phys., [5] 27, 145 (1882).
April, 192G free energy of hydrogen fluoride 911

the mother liquors were weakly acid and therefore during evaporation the
acid interacted with the basic salt with liberation of water. The error
from this cause, however, does not amount to more than 1% of the cor-
rection and may be ignored.
In Expt. 15, where the amount of sodium hydroxide was almost sufficient
to transform the stannous sulfate into stannous oxide, the unwashed
but dried precipitate contained 83.29% of tin and 3.56% of sulfate radical.
With the correction as applied before, the analysis showed only 1.9%
of sulfate. As 1.80 g. of sodium hydroxide was used and 1.86 g. is required,
according to the ordinary equation for the complete transformation into
stannous oxide, it is evident that this reaction takes place very nearly as
the equation indicates.
The volumes of liquid in proportion to the weight of stannous sulfate
were kept small in this work since preliminary experiments showed that
when the volumes were large, the less basic precipitates could not be ob-
tained. This is always true for salts of considerable solubility.

Summary
1. The precipitates formed at 100° by adding sodium hydroxide solu-
tion in various amounts to stannous sulfate have been analyzed.
2. Within two regions of alkali variation, the precipitates remain nearly
constant in composition. The formula SnS04.SnO probably represents a
compound and SnS04.2SnO-unknown-H20 may possibly do so.
3. A method is described for correcting the analyses of incompletely
separated solid phases for the retained liquid phase.
Houghton, Michigan

[Contribution from the Chemical Laboratory

California] of the University of

THE FREE ENERGY OF HYDROGEN FLUORIDE

By R. A. Morgen and J. H. Hildebrand
Received December 9, 1925 Published April 5, 1926

Introduction
The determination of the free energy of fluorine compounds is a problem
of extraordinary interest on account of the unique position of fluorine as the
most electronegative of the elements. Attempts to determine the free
energy of the fluorine electrode made in this Laboratory by Simons and
Hildebrand1 were unsuccessful. Electrodes of various noble metals and
of graphite surrounded by fluorine did not give reproducible potentials
either in liquid hydrogen fluoride or in fused potassium acid fluoride.
These experiments seemed to indicate little prospect of success for the
method and that the only hope would lie in measuring the dissociation
1
Simons and Hildebrand, This Journal, 46, 2223 (1924).