precipitation begins to become incomplete.

I have ascertained, finally, that acetone weakens in no manner
the action of the acetic acid in respect to nickel sulphide.



ISO METHODS OF ANALYTICAL CHEMISTRY

From these different experiments upon nickel acetate, it is evi-
dent that one molecule of sodium acetate prevents the formation of
nickel sulphide in a solution of acetic acid containing as many as
five molecules of the acid. We cannot, moreover, think of explain-
ing the decreasing effect of sodium acetate by the formation of
sodium triacetate, which is produced in a small quantity in dilute
solution as has been shown by Berthelot.

3. I have, therefore, been led to investigate the question as to
whether hydrogen sulphide does not react upon sodium acetate, even
in the presence of free acetic acid, to form sodium sulphide notwith-
standing the great difference in the heats of formation of the acetate
(13.3 Calories) and of the sulphide (7.8 Calories). We have pre-
viously seen, in fact, that the hydrolysis of the acetates of the alkalies
is clearly shown by calorimetric measurements, and from the
moment a little sodium hydroxide is freed by hydrolysis, hydrogen
sulphide in very great excess ought to produce sodium sulphide. As
nickel sulphide once precipitated becomes readily insoluble even in
the strongest acids, one easily understands that, in the equilibrium
which is established between hydrogen sulphide, the acetic acid and
sodium acetate, there may be produced traces of the alkali sulphide
conforming to the equation

2NaC 2 H 3 O 2 +H 2 S = 2HC 2 H 3 O 2 +Na 2 S,

in order that the precipitation of nickel sulphide be accomplished.
I have in fact verified the production of a small quantity of
sodium sulphide in the action of hydrogen sulphide on sodium
acetate, even in the presence of quite large quantities of acetic acid,
by using, to indicate the presence of sodium sulphide, sodium nitro-
prussate in fresh solution. This reaction gives, as we know, no
coloration with free hydrogen sulphide in solution, and produces, on
the contrary, an intense purple coloration with sulphides of the
alkalies, even in minute traces, the purple coloration turning to blue,
as I have shown, in the presence of a very large excess of hydrogen
sulphide in proportion to the sulphide of the alkali. The nitro-
prussate gives immediately a strong blue coloration in the mixture
of aqueous solutions of hydrogen sulphide and sodium acetate. The
formation of the alkali sulphide is decreased and is longer in making
its appearance but not suppressed by the addition of considerable
quantities of acetic acid with the respective concentrations employed
in the preceding experiments of the precipitation of nickel sulphide.



BASED UPOft CHEMICAL REACTIONS 151

The characteristic blue coloration of sodium sulphide is seen at the
end of about one minute with twenty cubic centimeters of acetic
acid, at the end of ten minutes with fifty, and becomes decidedly ap-
parent at the end of 30 minutes when 100 cubic centimeters of acetic
acid are employed. I have in addition verified that, under the same
conditions, no coloration is produced by the nitroprussate with a
mixture of acetic acid and sodium acetate, or with a mixture of
hydrogen sulphide and acetic acid. The blue coloration is therefore
due to the formation of small quantities of sodium sulphide.

It is then the production of the alkali sulphide in equilibrium
between sodium acetate, acetic acid and sodium sulphide that is the
veritable cause of the apparent decrease in the strength of the acetic
acid and of the more or less complete precipitation of the acetate of
the metal, not precipitated in the presence of acetic acid alone under
the same conditions; and it is not at all necessary to employ the
electrolytic dissociation theory to interpret this phenomenon.

If our explanation is correct, this consequence, in appearance
paradoxical, should result, that if we add to the solution of nickel
acetate a neutral salt of a strong acid, like the chlorides or the
sulphates of potassium, sodium or ammonium, the strength of the
acetic acid should appear to increase and consquently the precipi-
tation of nickel by hydrogen sulphide be diminished or completely
prevented by reason of the small quantity of strong acid which
should be liberated by the acetic acid in the partition of the strong
base between the strong acid and acetic acid, as we have seen in
Chapter III (page 56). This is in fact the result we have obtained :
the partial precipitation of nickel sulphide which hydrogen sulphide

gave in the preceding experiments with ten cubic centimeters of -

nickel acetate and five cubic centimeters of acetic acid, is completely
prevented if a few grams of a chloride or sulphate of an alkali are
introduced into the solution.

This apparent increase in the strength of a weak acid produced
by the addition of a neutral salt of a strong acid, which had not yet
been brought to our knowledge, could be used for certain separa-
tions, for example, that of nickel and zinc, whose complete precipi-
tation by hydrogen sulphide is not prevented by traces of free
hydrochloric acid.



152 METHODS OF ANALYTICAL CHEMISTRY

2. Methods for Bringing into Solution a Compound Insoluble
in Water or Acids

This problem, the reverse of precipitation of double decomposi-
tion of salts, is one of those which must be constantly solved in the
chemical analysis of inorganic compounds. It is employed in a
simple form for all the minerals of the metallic sulphides containing
silicious gangue, by treating with strong acids, oxidizing, if neces-
sary; the metallic sulphides are dissolved without changing the
silica, and by simple nitration the dissolved metals are separated
from the insoluble residue.

A more difficult case to solve, which presents itself frequently in
chemical analysis, is that in which there is a mixture of several sub-
stances equally insoluble in acids, such as silica and the sulphate of
barium or of lead. The methods employed in this case are again
derived from the law of equilibrium in double decomposition of
salts.

We can, according to the case, proceed by two different methods :
either transform one of the substances insoluble in acids into another
body insoluble in water, but easily decomposed by acids, by Dulong's
alkali carbonate method of decomposition, or treat the insoluble
substances with a reagent, causing in only one of them a double
decomposition from which results only soluble substances.

Dulong Method by Wet or Dry Way. This method is very gen-
eral and allows the transformation of the acid of an insoluble salt
into a soluble alkaline salt, while the base of the salt passes into the
state of a carbonate insoluble in water but easily dissolved by hydro-
chloric or nitric acids. From these results the possibility of easy
determination, after this transformation of the acid and base of the
salt, each of them being present in combinations of physical states
different and easy to separate by simple nitration.

Let us consider as an insoluble salt, barium sulphate, for example,
and let us treat it with a solution of potassium carbonate. This is
the equilibrium studied by Guldberg and Waage,

BaSO 4 insoluble+K 2 CO 3 = K 2 SO 4 +BaCO 3 (insoluble).

We have seen that the transformation of barium sulphate stops
when the concentration of the potassium sulphate formed reaches
about one fourth of the concentration of the remaining carbonate of
potassium conforming with the equilibrium equation:



BASED UPON CHEMICAL REACTIONS 153



~ 2.26

C K 2 C0 3 _

~~~ T"

C2. 1 1
K 2 S0 4

Then, for a mixture with equal molecules of barium sulphate
and potassium carbonate, only about one fifth of the barium sul-
phate is transformed into carbonate. It is necessary then, in order
that the transformation be complete, to take about five molecules of
potassium carbonate to one molecule of barium sulphate. In prac-
tice, we use sodium carbonate instead of potassium carbonate, be-
cause sodium sulphate is much more soluble than potassium sulphate,
and we use a large excess of alkali carbonate (10 molecules of
Na 2 CO 3 to one molecule of BaSO 4 ). Besides, we heat at the boil-
ing point in order to facilitate the reaction, which is endothermic
and favored consequently by an elevation of the temperature, con-
forming to the principle of opposition of action and reaction. It is
well, for the same reason, to make the filtration while the liquid is
hot. The reaction is, moreover, slow, because it takes place between
substances of different physical states, of which one, the barium sul-
phate, is almost absolutely insoluble.

If one has quartz mixed with barium sulphate, as it is not
attacked by boiling sodium carbonate, it will be sufficient to filter
the residue, and, after washing, treat the mixture with a dilute acid,
hydrochloric or nitric, to dissolve completely the barium carbonate
and leave the insoluble quartz.

The same transformation of an insoluble carbonate, with trans-
formation of the acid of the salt into soluble alkali salt, can be
accomplished by the dry method as well as by the wet method, by
heating to the fusion point at a bright red heat, the mixture of
insoluble salt and of sodium carbonate in excess, or better a mixture
of equal equivalents of potassium carbonate and sodium carbonate,
whose fusion point is lower than the fusion points of the separate
salts. The theory of the dry method is the same as that of the wet
method and it is the respective concentrations of carbonates and
alkali sulphates united in a homogeneous mixture which control the
equilibrium in this case.

But in the dry method, another factor is to be considered, that
of the stability of the carbonates of the metals formed at the fusion
point of the alkali carbonates. Since starting with calcium, all the
carbonates are easily dissociated by heat at the fusion temperature



154 METHODS OF ANALYTICAL CHEMISTRY

of the carbonates of the alkalies, it is generally an oxide which is
produced instead of a carbonate, with the liberation of carbon di-
oxide, and, if this oxide has the function of an acid (A1 2 O 3 for
example), it forms with the alkali a product generally soluble in
water. But the problem is, however, greatly simplified, for it is
sufficient then to produce in the solution obtained, the precipitation
of one of the substances by an appropriate method. This is the case
in the separation of silica from aluminium in clays, by fusing with
an alkali carbonate, dissolving the soluble silicate and aluminate of
the alkalies formed in water, then precipitating the silica by hydro-
chloric acids, etc.

Redissolving an Insoluble Substance by a Suitable Reagent. In
many cases, a salt formed of a weak acid and insoluble in water,
can be dissolved by means of a stronger acid, whenever this latter
can give a soluble salt with the base of the salt. Calcium oxalate
or phosphate, insoluble in water, are immediately transformed by
hydrochloric acid, into soluble calcium chloride and oxalic acid or
phosphoric acid in apparent opposition to the laws of Berthollet.
The reaction is indeed a reaction of equilibrium, but in which, by
reason of the great evolution of heat, the transformation of the sys-
tem is practically complete, as we have seen in Chapter III, Section 3.

In many cases, however, where one fails with a strong acid, he
can frequently succeed in getting into solution an insoluble salt with
a neutral soluble salt, which may present a certain advantage over
the use of acid reagents. Thus, for example, in treating lead sul-
phate which is insoluble not only in water, but also in quite strong
acids, with an alkali salt of a weak acid, as sodium or ammonium
acetate, it is easily redissolved, and one can by this method, separate
easily the lead sulphate obtained from action of oxidizing acids on
galena, from the siliceous gangue frequently accompanying these
minerals.

This action, in appearance paradoxical, is a direct consequence
of the law of equilibrium in the double decompositions of salts. The
reaction

PbSO 4 +2NaC 2 H 3 2 = Na 2 SO 4 +Pb(C 2 H 3 O 2 ) 2

23.8 Cal. 26.6 Cal. 31.4 Cal. 15.5 Cal.

is in fact an equilibrium reaction limited by the reverse reaction.
The direct reaction is produced with the absorption of heat ( 3.5



BASED UPON CHEMICAL REACTIONS 155

'Calories), but the proportion of the first system transformed into the
second is hardly appreciable with an equimolecular mixture. It is
<the predominating stability of the sodium sulphate in the aqueous
solution which is the determining cause of the reaction. It is suffi-
cient then to greatly increase the concentration of the sodium acetate
in order that the concentration of sodium sulphate and of lead
acetate controlled by the equation:

C 2i NaC 2 H 3 2 = k C r Na 2 S0 4 Ci "Pb(C 2 H 3 O 2 ) 2 ,

attain the value corresponding to the complete solution of the lead
sulphate. The following are the quantities of this substance which
are dissolved by the sodium acetate at different concentrations, ac-
cording to H. C. Debbits :*



SODIUM ACETATE IN 100
GRAMS OF SOLUTION


LEAD SULPHATE DISSOLVED


2.05 grams

8.20 "

41.00 "


0.054 gram
o.ooo "
' 1 1. 200 grams



Heat increases, moreover, the solubility, the thermal sign of the
.reaction being negative.

For reasons of the same kind, one must use discretion in allow-
ing the presence of perfectly neutral additional salts, in liquids in
which a definite precipitate is sought, for the preceding example
shows that this precipitate may become incomplete by reason of the
presence of these salts, which might seem a priori to be without
effect.

The study of every method of analysis employing an insoluble
precipitate ought to be made with special attention to the action upon
this precipitate of all reagents which might be present in the solu-
tion at a given moment in the course of the analysis. The failures
that are met at times in the use of a method, come from the fact
that this method was established by its author while using exclu-
sively pure reagents destined to produce the desired reaction, while
in practice the disintegration of the one often necessitates the intro-
duction into the solution of a considerable quantity of foreign sub-
stances (salts of the alkalies, acids, etc.). Hence, while the presence

*H. C. Debbits, Bull, Soc. Chim., Paris (2) xx, 258 (1873).



156 METHODS OF ANALYTICAL, CHEMISTRY

of even a quite considerable excess of hydrochloric acid does not
prevent the precipitation of sulphuric acid by barium chloride, care
must be taken not to cause this precipitation in the presence of free
acetate of the alkalies which produce upon barium sulphate an action
(less, it is true) analogous to that which they have with lead sul-
phate. In this case, it will be sufficient to make the solution decid-
edly acid with hydrochloric acid to destroy the sodium acetate and
to free the acetic acid which is without action upon barium sulphate.
For analogous reasons, the precipitation of sulphuric acid by
barium chloride is incomplete in the presence of copper salts or of
magnesium. On the other hand, it is not perceptibly interfered
with by the presence of zinc salts ; and it is sufficient, for example,
to precipitate copper from its solution by zinc to be able afterwards
to determine quite accurately the sulphuric acid present.*

3. Precipitates of Variable Composition

The principles which we have previously given suffice, in general,
to explain and direct the methods of analysis based upon double
decomposition of salts, each time that the latter causes only a slight
precipitate of invariable composition in equilibrium with the reagents
remaining in solution even while varying within quite wide limits of
temperature and respective concentration. Such are the cases of
the precipitation of hydrochloric acid by silver nitrate, which gives
only silver chloride ; of sulphuric acid by barium chloride, which
gives only barium sulphate, etc., as well hot as cold, and in liquids
more or less strongly acid.

It is not the same, however, when, according to the conditions of
temperature and concentration there may be produced several dif-
ferent precipitates, either pure simple salts or mixtures, in equilib-
rium with the solution.

These variations in the composition may affect either only the
degree of hydration or the respective proportion of the acids and
bases.

Hydrates of Variable Composition. The first case is very fre-
quent in analysis : many metallic hydroxides, sulphides, certain in-
soluble salts, as calcium oxalate, oxides with acid properties :
metastannic, titanic, silicic, etc., have different degrees of hydration
according to the temperature at which they are precipitated. If

*G. Chesneau, C. R., cxxxvii, 653 (1903).



BASED UPON CHEMICAL REACTIONS



157



one were satisfied with weighing them after drying on a tared filter,
he would have quite uncertain results varying from one experiment
to another. In general, however, this uncertainty as to the degree
of hydration is easily removed by transforming the precipitate into
an anhydrous substance of very constant composition, for oxides
and the oxidized salts by simply igniting, for sulphides by heating
in an atmosphere of sulphur or a current of hydrogen sulphide.

Double Ammonium Salts: Quantitative Determination of Arsenic
and of Phosphorus. It is different, however, when the relative
proportions of bases and acids vary in the precipitate with the
factors of equilibrium. The precipitates in the form of insoluble
double ammonium salts belong generally to this category. The
determination of arsenic in the form of ammoniacal arsenate of
cobalt (method of O. Ducru) and of phosphorus in the form either
of ammonium magnesium phosphate or of ammonium phospho^
molybdate offer very striking examples of these variable compounds.

Concerning the precipitation of arsenic in the form of ammoni-
acal arsenate of cobalt produced by the action of cobaltic chloride
upon arsenate of ammonium in the presence of ammonium chloride,
and of free ammonia and prolonged digestion of the precipitate in
the mother-liquor, the author* of the method, has shown that accord-
ing to the concentration of ammonia, which influences only the com-
position of the precipitate, it varies in a continuous manner with the
concentration of ammonium hydroxide from natural erythrine
Co 3 (AsO 4 ) 2 +8H 2 O to Co 3 (AsO 4 ) 2 -3NH 3 +5H 2 O, NH 3 replacing
the water molecularly, as the following diagram shows. The ab-
scissas represent the concentration in free ammonia expressed in

10



3NK



2NH-



1NH 2




10U



200

FIG. VI



300



400



*O. Ducru, These de doctoral, and C. R., cxxxi, 675, 886 (1900).



158 METHODS OF ANALYTICAL CHEMISTRY

cubic centimeters of ammonium hydroxide of a density of 0.921 per
liter of solution and the ordinates the contents of the salt in per
cent of NH 3 . It is indeed a question here, as the author has shown,
of a veritable chemical equilibrium between the solution and a com-
plex precipitate, which appears in addition to be a homogeneous
solid, for the different ammonium arsenates of cobalt are isomor-
phous and consequently form a solid solution. It is then impossible
to obtain a precipitate of an exact fixed composition, but this varia-
tion presents no difficulty because, by reason of a wholly fortuitous
circumstance, NH 3 and H 2 O having almost the same molecular
weight, the content of arsenic in the precipitates does not vary per-
ceptibly whatever may be the proportion of water displaced by am-
monia in the precipitate.

The methods for the quantitative determination of phosphorus
unfortunately do not present the same peculiarities.

In the precipitation of phosphoric acid in the form of ammonium
magnesium phosphate, by magnesium chloride in the presence of
ammonium chloride and ammonium hydroxide, there may be pro-
duced, depending on the concentration of the ammonium hydroxide
either trimagnesium phosphate, Mg 3 (PO 4 ) 2 '5H 2 O, or MgNH 4 PO 4 -
6H 2 O, and we generally have a mixture of the two at the beginning
of the precipitation. Here it is very necessary to have as the final
product the second compound only, to obtain after heating the well-
defined pyrophosphate Mg 2 P 2 O 7 and not a mixture of Mg 2 P 2 O 7 and
Mg 3 (PO 4 ) 2 , which have quite different contents of phosphorus.
From the investigations of H. Lasne* it is the presence of greater
or less proportions of tri-magnesium phosphate in the final product
which explains the variations, often very appreciable, obtained in
the analysis of the same phosphate. The use of strongly ammoniacal
solutions (one-third of the volume concentrated ammonium hydrox-
ide) seems to prevent this error, as H. Lasne indicates, for tri-
magnesium phosphate redissolves little by little in the ammonium
hydroxide and then reprecipitates slowly in the form of ammonium
magnesium phosphate. If one filters too soon there may be even
after two or three hours of digestion, enough tri-magnesium phos-
phate to make the error of the content of phosphorus 0.4 per cent,
and the digestion should be prolonged for sixteen hours at least.

The precipitation of phosphoric acid in the form of ammonium

*H. Lasne, C. R., cxxvii, 62 (1898).



BASED UPON CHEMICAL REACTIONS 159

phosphomolybdate, which is the only known method for the very
accurate quantitative determination of phosphorus in cast irons or
steel, presents sources for error of the same kind. It is known from
the early investigations of Svanberg and Struve* that, if a solution
of phosphoric acid or of a phosphate is poured into a large excess of
nitric acid solution of ammonium molybdate, 7 MoO 3 '3(NH 4 ) 2 O-
4H 2 O, there forms slowly in the cold, rapidly when hot, a yellow
crystalline precipitate containing all the phosphoric acid, with
ammonium hydroxide and an enormous quantity of molybdic acid,
nearly thirty times the weight of the phosphoric acid. Sonnen-
schein,f who first applied this method to the quantitative determina-
tion of phosphorus, fearing that this yellow substance might not be a
definite compound by reason of this very high ratio, recommended its
dissolution in ammonium hydroxide and the reprecipitation of the
phosphorus in the form of ammonium magnesium phosphate. The
advantage of having a precipitate much heavier than the element to
be determined, very important for determining traces of phosphorus,
was thus lost, and a number of workers have sought for the neces-
sary conditions to obtain an exactly definite phosphomolybdate.

H. Debray, who has made a special study of phosphomolybdic
acids,! obtained several hydrates of an acid to which he assigned the
formula, P 2 O 5 -2oMoO 3 , giving in a nitric acid solution containing
ammonium nitrate the familiar yellow precipitate, to which he
assigned the formula 3(NH 4 ) 2 O-P 2 O 5 -2oMoO 3 '3H 2 O, containing
1.918 per cent of phosphorus.

Debray has, moreover, ascertained that this phosphomolybdate is
stable only in the presence of an excess of nitric acid, and that
alkalies transform into ordinary molybdates and phosphomolybdates
of another phosphomolybdic acid, P 2 O 5 -5MoO 3 *3H 2 O+Aq, giving
alkaline phosphomolybdates which crystallize into but slightly soluble
white needles.

Since the work of Debray, the question of the constancy of the
composition of Sonnenschein's precipitate has been frequently
studied by reason of its great importance in the metallurgy of iron.
The following are the results of the principal authorities :

* Svanberg and Struve, Ann. de Millon et Reiset (Annuaire de Chimie)
1849, p. 163.

t Sonnenschein, Journal prakt. Chem., liii, 343 (1851).
t H. Debray, C. R., Ixvi, 702, 732 (1868).



160 METHODS OF ANALYTICAL CHEMISTRY



AUTHOR


FORMULA


PERCENTAGE OF P.


H. Debray
Rammelsberg
W. Gibbs
A. Carnot


3 (NH 4 ) 2 O.P 2 5 .2oMo0 3 . 3 H 2
3 (NH 4 ) 2 OP 2 5 .22Mo0 3 .i2H 2
5 (NH 4 ) 2 0.2P 2 5 . 4 8Mo0 3 .i6H 2
3 (NH 4 ) 2 O.P 2 5 .2 4 Mo0 3 -3H 2


1.918
1.684

1-597
1.628



The differences are, as we can see, very considerable.

Since the work of A. Carnot,* it is assumed in the metallurgical
laboratories that, by following strictly the method employed by the
author, the composition of the precipitate obtained in strong nitric
acid solutions at 40 and then dried at 100, always corresponds to
1.628 per cent of phosphorus, and the numerous verifications made
of the synthetic mixturesf prove this. The precipitates of Debray
and of Rammelsberg probably contain a little phosphomolybdate with