the same ; it is impossible to continue the washings with pure water
without danger of losing in the filtrate all or part of the precipitate,
(case of colloidal ZnS, NiS, etc.).
It is only in quite rare cases in chemical analysis that colloidal
bodies are formed in the absence of all saline matter, and conse-
* In the recent works upon colloidal bodies and colloidal solutions the
former are designated also by the name, hydrogeh, and the latter by the name,
hydro sols.
BASED UPON CHEMICAL REACTIONS *
quently would remain in the original as pseudo-solutions. A well-
known example is that of the cold solution of arsenious acid in pure
water, which, treated with a small quantity of hydrogen sulphide,
gives a yellow solution, passing through the filters and remaining a
long time without change ; a few drops of hydrochloric acid produce
the precipitation of the sulphide of arsenic, which loses at the same
time the property of again forming pseudo-solutions with pure
water. The dilute solutions of stannic and antimonious chlorides
can likewise give with hydrogen sulphide colloidal solutions, but
slightly colored, and existing for some minutes without a precipi-
tate, in spite of the presence of free hydrochloric acid, and this
peculiarity may easily deceive one as to the presence of these metals
in qualitative determinations.
A knowledge, as exact as possible, of the conditions in which are
formed the pseudo-solutions and of the processes which permit their
destruction or prevent their formation, is then necessary in chemical
analysis where colloidal precipitates predominate, on account of the
frequent use of the hydroxides and sulphides of the metals.
From the physical point of view, the colloidal precipitates are
formed of particles or nuclei incomparably finer than crystalline
precipitates, while the dimensions of the finest grains which the
latter give, are of the order of a thousandth of a millimeter, those
of the grains which form the colloidal bodies are of an order a
thousand times smaller, and consequently, invisible under the micro-
scope by the processes of ordinary observation. These particles can,
however, be brought into evidence by a method derived from the
well-known fact that the dust particles of the air, strongly illumin-
ated by a ray of sunlight passing through a dark room, appear
luminous and become perceptible to the naked eye; likewise, any
object, however small it may be, becomes visible under the micro-
scope, if it is lighted strongly by a very powerful ray of light,
perpendicular to the optical axis and incapable -of penetrating the
apparatus. Thanks to this method, due to the German physicists
Siedentopf and Zsigmondy, it has been possible to show the presence
of particles of gold contained in the state of solid colloidal solution
in pink Bohemian glass, as well as the nuclei of the colloidal solu-
tions of copper ferrocyanide in water and bromide of silver in
gelatine (Cotton and Mouton). By counting the particles appear-
ing in a solution of given concentration, it has been possible to
estimate their diameter at a few thousandths of a micron.
20 METHODS OF ANALYTICAL CHEMISTRY
One can understand that, with such small dimensions, the surface
energy can play a preponderant role in the forces involved in the
contact of these particles and the liquids in which they are placed,
and that the colloidal solution resulting from these actions may have
very different properties from those of the ordinary solutions of
crystalloidal bodies.
The desiccation of colloidal solutions or of colloidal precipitates,
at ordinary temperature, produces different effects according to the
bodies in pseudo-solution. Sometimes the residue is more or less
granular with a dull and earthy fracture, and in that case, the
matter, thus desiccated, is no longer capable of again forming a
pseudo-solution by the addition of pure water; such are liquids
made turbid by kaolin, silica, colloidal sulphides of arsenic, copper
and mercury. At other times, on the contrary, the residue after
desiccation furnishes an elastic body with brilliant and vitreous
luster which, placed in contact with pure water, will again give a
colloidal solution identical with the original pseudo-solution, such as
gums, albuminoids, aniline dyes, etc. Colloidal substances are thus
subdivided into unstable and stable colloids ;* these latter allow
indefinite re-solution in water like crystalloidal bodies, but with this
difference, that the stable colloids do not possess maximum coeffi-
cients of solubility at every temperature. There exist between the
matter in the solid phase and its pseudo-solution intermediate gummy
or mucilaginous conditions analogous to the plastic state in the solid
amorphous substances such as glasses.
The stable colloids are most frequently organic substances and
the unstable are inorganic (inorganic colloids, Ag, Pt, Ir, As 2 S 3 ,
SiO 2 , etc.). There does not exist, however, any clearly defined
separation in this respect between the organic and inorganic sub-
stances, certain colloidal hydrates of iron and chromium, even after
prolonged desiccation can again form pseudo-solutions.
By taking the two extremes, the pseudo-solutions of gelatinous
silica, clearly unstable, and the pseudo-solution of mastic, extremely
stable, W. Springf has shown the very clear difference of solidarity
existing between the particles in both pseudo-solutions. By super-
posing a thick, turbid layer of mastic upon pure water in a test
* V. Henri and A. Mayer, I'Etat actuel de nos connaissances sur les
colloides (Revue general des sciences, p. 1015 (igo4.
t W. Spring, Sur la floculation des milieux troubles, Rev. des Trav,
chimiques des Pays-Bas et de la Belgique, xix, 204 (1900).
BASED UPON CHEMICAL REACTIONS 21
tube (the turbid mastic having a specific gravity of 1.0665), W.
Spring has shown that the surface of separation, plane at first,
takes little by little the spherical form, as if the upper layer had
swollen uniformly ; on the contrary, in a layer of silica superimposed
upon pure water, the particles of silica descend parallel, and the
surface of separation remains a plane. It is thus seen that turbid
mixtures of silica finally precipitate at the end of a few weeks,
while, for the mastic, time alone does not produce the clarifica-
tion.
The elevation of temperature favors in certain cases, especially
for the unstable colloids, the decomposition of the pseudo-solutions ;
such is the case of water clouded by clay in suspension, which, if
the clay is very fine, remains turbid almost indefinitely and which
becomes clarified according to C. Barus, about twenty times more
quickly at 100 than at 15. This clarification by elevation of the
temperature can be obtained every time by prolonged digestion ;
heat modifies the precipitate by making it granular instead of gela-
tinous. Such is the case of aluminium precipitated by ammonium
hydroxide, which, digested hot for some hours, becomes granular
and is filtered more easily. Many metallic sulphides (MnS, ZnS,
NiS, etc.), are much less colloidal when precipitated from boiling
solutions ; the flakes are then denser and filter better.
This transformation is still more rapid when the dried substance
is heated, as in the case of insoluble silica obtained by heating the
dried mass for some hours between 100 and 110, which renders , v ,
it not only insoluble in water, but also in very dilute acids. Finally $ v -.
a great number of gelatinous hydroxides of the metals, heated to a ^^\^
high temperature, are transformed into crystalline bodies with liber-
ation of heat (ignited oxides), and lose definitely the faculty of
forming pseudo-solutions; such are the oxides of iron, aluminium,
chromium, etc. Once transformed thus into crystalloids whose
grains are much larger than those of the original colloid, these
oxides in general have no longer any appreciable power of absorp-
tion, and the inorganic salts introduced during the precipitation can
then be easily removed if no chemical combination has been pro-
duced between the oxide and the elements of these salts during the
heating. The hydroxides of nickel and of cobalt, even after igni-
tion still retain strongly, however, the alkali absorbed during pre-
cipitation, and they must be reduced to the metallic state at a high
temperature in a current of hydrogen, in order to be able to free
22 METHODS OF ANALYTICAL CHEMISTRY
them of the alkali, for which the metallic conglomerate of crystalline
granules has no longer any absorbing power.
This transformation by heat of colloidal precipitates, capable of
forming pseudo-solutions, into crystalline bodies insoluble in water
or weak acids, finds an interesting application in the quantitative
determination of fluorine by the Berzelius method. The hydrofluoric
acid is precipitated by calcium chloride in an ammoniacal solution
containing an excess of ammonium carbonate, which gives gelati-
nous CaF 2 , with an excess of CaCO 3 . It is impossible to separate
these two bodies by means of even a dilute acid solution because
gelatinous CaF 2 forms pseudo-solutions with acidulated water, just
as gelatinous silica. But, if we ignite the precipitate at low red heat,
CaF 2 becomes crystalline and ceases to be soluble in acetic acid,
while CaCO 3 is still soluble in it and washing by acetic acid permits
one to obtain finally pure CaF 2 .
Precipitation of Pseudo-Solutions by Salts, All colloidal solu-
tions and turbid liquids, in which the solvent is water, become clari-
fied and allow the matter in pseudo-solution or in suspension to
precipitate whatever may be its chemical nature, by the addition
of a sufficient quantity of a strong acid or one of its salts. The
material first appears in the form of flakes, which deposit at the
bottom of the liquid or come to the surface, depending upon the
density, or whether bubbles of gas are formed during their precipita-
tion. Two phenomena are then distinguished successively and per-
haps independently of each other, that of nocculation and that of
sedimentation*
The proportion of salt or acid necessary to flocculate the pseudo-
solutions must exceed a certain limit, very small, it is true, in order
to be efficacious; for example, with a turbidity formed of pure
kaolin, the limit of action seems reached when the dilution of the
ac ^ * s isooooo (C- Bodlander). In a general way, it is the salts
of the polyvalent metals which are more active, and this fact holds
good so much the more in proportion as the metallic oxide is weaker
as a base, as is shown in the following table of the clarifying powers
of some salts, compiled by H. Schulze, assuming the power of
potassium iodide equal to I :
* W. Spring, loc. cit.
BASED UPON CHEMICAL REACTIONS
SALT
CLARIFYING
POWER
Kl
i.
KC1
2-5
CaCl 2
So.
MgCl 2
182.
Na 2 SO 4
2.3
ZnSO 4
60.
A1 2 (S0 4 ) 3
957-
A1 2 C1 6
1518.
Bodies which yield poorly conducting solutions, on the contrary,
generally retard flocculation ; alcohol, however., accelerates the pre-
cipitation of certain colloidal salts, such as potassium fluosilicate.
When an electric current is passed through a turbid mixture or
a colloidal solution, we notice that a clarification takes place at one
of the two electrodes, according to the chemical nature of the sus-
pended substances, while a flocculation is produced at the opposite
electrode. W. Spring has even shown that the electric current frees
the water from every particle in suspension to the degree of render-
ing the liquid optically empty, that is to say, representing no lumi-
nescence upon the passage of a powerful ray of light. This floccu-
lating property of the electric current joined to the absence of
flocculation by bodies which are poor conductors of electricity, has
caused several writers to attribute the flocculating power of saline
substances, which are electrolytes, to the presence of free electrical
ions from the salts. The following observations of W. Spring*
tend rather to cause this property to be attributed to the presence of
bases and acids freed by the dissociative action of the water upon
the salts in solution.f
This author has first proved that, in the clarification of colloidal
solutions by the electric current, the transportation of the particles
is accomplished in the direction anticipated, or in the opposite direc-
tion by the difference of electrical conductivity of the liquid and of
the colloid; in the second place, that the clarifying power of elec-
trolysis is not in proportion to their respective electrical conductivi-
ties. Then he has observed the method of flocculation of suspended
*Loc. cit.
t See the theory of electrolytic dissociation of salts in Chapter IV. and
that of the dissociation by hydrolysis in Chapter III.
24 METHODS OF ANALYTICAL CHEMISTRY
particles of mastic by different metallic salts, not by mixing directly
the solutions, but by superposing the turbid layer upon the saline
solution and allowing the liquids to diffuse freely. The flocculation
soon begins and the flakes descend as far as the layer, where there
is an equality of density between them and the liquid. With salts
of different metals, it is observed that, at the end of equal periods
of time, the degrees to which the flocculation has progressed are
different, without the possibility of finding a simple relation with the
coefficient of diffusion of the salts ; it is merely observed that salts
derived from polyvalent metals clarify the turbidity to a greater
degree. Colored salts, like sulphate of copper, have permitted the
exact determination of the cause of this difference; it has been
shown with this salt that the flocculation is produced well above the
region where the blue color is manifest, and some tests made with
capillary pipettes at the top of the flocculated zone, have shown, in
fact, the absence of copper, but on the other hand, demonstrated the
presence of sulphuric acid. Likewise, with aluminium chloride, fer-
ric chloride, magnesium chloride, zinc chloride, alum, W. Spring
has always perceived in this part of the solution the presence of the
free acid of the salt without the metal ; on the other hand, the flaky
deposits, collected from the above concentrated saline solutions and
washed with water, have always revealed upon analysis the presence
of the corresponding metallic oxides.
These experiments prove clearly that metallic salts are decom-
posed during their diffusion through the colloidal suspensions. The
metallic hydroxide has inclosed the particles of the colloid and has
precipitated them, while the acid has diffused into the colloidal solu-
tion, flocculating it on its own account; the colloidal matter acts
then upon the metallic salt like the septum of a dialyser and becomes
a sediment while associating itself with the metallic hydroxide. If
the polyvalent inorganic salts are more active than those of the
alkaline metals in producing flocculation, it is because their hydro-
lytic decomposition is much more marked; it is, besides, those salt
solutions which become most strongly illuminated when they are
traversed by a powerful beam of light, consequently constituting
themselves a colloidal solution of metallic hydroxide formed by
hydrolysis which have a very considerable coagulative power.
These experiments show also that the coagulation of colloidal
solutions cannot be produced without an association of the sus-
pended particles with the mineral substance; recent researches,
BASED UPON CHEMICAL REACTIONS 25
notably those of Whitney and Ober* seem even to prove that this
association follows in a certain measure the mathematical laws of
definite compounds. When the colloidal solution of arsenic sulphide
is precipitated by chlorides of potassium, calcium or barium, the
precipitate always contains metallic oxide of the salt, and the liquor
contains on the other hand free hydrochloric acid. If the precipi-
tate is subjected to prolonged washing, a part, only, of the metallic
oxide is eliminated and the rest remains bound to the colloid in an
irreversible manner. If the same colloidal solution of arsenic sul-
phide be precipitated by solutions of different chlorides, but of the
same molecular concentration, one perceives that the weights of
metals carried down, bound in an irreversible manner to the pre-
cipitated colloid, are proportional to the molecular weight of the
metals ; these weights are, in fact, for 200 cubic centimeters of a
five per cent colloidal solution of As 2 S 3 , 0.020 gram of Ca, 0.076
gram Ba, 0.036 gram K. With solutions of the same salt of differ-
ent concentrations, the amount removed by the colloid is in propor-
tion to the concentration. Finally, if the precipitate is washed with
the solution of another salt, the metal of the second salt replaces
quantitatively the metal primarily bound in an irreversible manner
to the colloid. Similar results, obtained by J. Duclauxf upon the
precipitation of a colloidal solution of copper ferrocyanide by various
salts, have led him to conclude that "the materials extracted by
coagulation are only the simple substitution of the radicals of the
precipitant for those which compose the colloid."
Rules for Washing Colloidal Precipitates. Although our knowl-
edge of the formation of colloidal precipitates is yet quite incom-
plete, the preceding results permit us, however, to determine the
rules to be observed for their purification.
Whenever the quantity of mineral salts, bound in an irreversible
manner to the colloidal precipitate, is insufficient by itself to produce
the flocculation of the pure colloidal solution of the same body,
washing upon the filter with pure water reproduces the colloidal
solution of the precipitate, and the precipitate, converted again into
a pseudo-solution, passes through the filter. It is necessary then to
wash, not with pure water, but with a solution of an acid or a salt
which can be easily eliminated by desiccation in a drying oven or by
ignition. The decreasing order of flocculation being the following :
* Whitney and Ober, Zeit. phys. Chew., xxxix, 630 (1902).
tj. Duclaux, C. R., cxxxviii, 571 (1904).
26 METHODS OF ANALYTICAL CHEMISTRY
salts of the polyvalent metals, and acids and salts of the alkaline
earth metals. We are often unfortunately obliged to take the least
efficient. One cannot, in fact, employ any other salts of polyvalent
metals than those of mercury, which can be utilized only in quite
limited cases, on account of the reactions which they can give on
their own account. Acids can be employed only with precipitates
with acid properties (as, for example, in washing titanic acid with
acetic acid), or with salts particularly insoluble in even strong
acids (washing AgCl with dilute HNO 3 ). For salts of the alkaline
metals, we are limited to ammonium salts, easily volatilized by
heat; the chloride, nitrate or acetate.
We are then frequently obliged to introduce into the wash water,
only bodies with a weak flocculating power ; acetic acid or ammon-
ium salts. The latter, having a smaller equivalent weight than the
salts of potassium or sodium, have, moreover, the advantage of
being more active than these latter for the same weights, according
to the observations of Whitney and Ober.
Finally, we can deduce again from these observations, the fol-
lowing general rule, already sanctioned by use in a great many
methods of analysis, "Whenever an hydroxide is to be precipi-
tated by an alkali in a solution containing salts of the heavy metals,
it is necessary, when possible, to redissolve the precipitate in an
acid solution and reprecipitate by ammonium hydroxide; one thus
replaces the heavy metal removed by the precipitate in the first pre-
cipitation, by an ammonium salt, which is easily volatilized during
the ignition of the precipitate."
In the second precipitation, the concentration of the heavy metals
is in fact insignificant in comparison to that of the ammonium salts,
and the irreversible impurity in the precipitate is almost exclusively
formed by the latter. By repeating this treatment, one can finally
obtain precipitates which, when ignited, are pure. This is a rule
which must be followed, for example, in the precipitation of
Fe 2 (OH) 6 or MnO 2 , in solutions containing an appreciable quan-
tity of calcium salts, a very frequent occurrence in the analysis of
minerals, oxides of iron, or manganese.
CHAPTER II
THEORETICAL PRINCIPLES OF THE ANALYTICAL
METHODS BASED UPON IRREVERSIBLE
REACTIONS
i. Principal Types of Irreversible Reactions Employed
in Analysis
THE methods of separation based upon irreversible chemical reac-
tions are extremely varied, by reason of the very different processes
which this category of reactions offers in order to obtain systems
of two distinct phases, separable by mechanical means. We will
indicate by a few examples the principal types to which these
methods belong.
Methods Based upon the Stability of Bodies at Arbitrarily Fixed
Temperatures. The transformation of a great number of precip-
itates into definite stable compounds at high temperature, at which
many of the methods of analysis aim, belongs to this category.
A great number of insoluble bodies obtained by double decomposi-
tion; metallic oxides, oxidized salts, etc., are, in fact, hydrates of a
composition variable with the temperature and cannot serve directly
in the exact determination of the elements which they contain.
Others, like the sulphides precipitated from a salt solution by
hydrogen sulphides, become more or less oxidized in air during
their desiccation. By ignition, one can transform the hydrates of
iron, aluminium, silica, etc., a great number of hydrated salts such
as the oxalate of calcium, ammonium magnesium phosphate, etc.,
into oxides or anhydrous salts, generally crystallized and of a
perfectly definite composition; thanks to the very great stability of
these bodies at high temperatures. Likewise, by heating hydrated
sulphides, more or less oxidized, in an atmosphere of sulphur or
of hydrogen sulphide in a Rose crucible, one obtains the metallic
sulphides, (MnS, Cu 2 S, etc.), crystallized and of definite compo-
sition.
Finally, the unequal stability of different classes of salts, such
27
28 METHODS OF ANALYTICAL CHEMISTRY
as the nitrates, the thiosulphates, etc., heated to definite tempera-
tures, allows the formulation of numerous processes of separation.
Methods Based upon the Employment of Oxidizing and
Reducing Reagents. Most of the methods of attacking mineral
substances by the dry method or the wet method belong to this cate-
gory. They constitute an intermediate stage, a disintegration of a
complex system which is transformed into a more simple system : de-
composition of natural metallic sulphides by aqua regia or potassium
nitrate, chromite by sodium dioxide, of wolfram by aqua regia, etc.
In certain cases, the reaction furnishes at the same time the
determination of the substance : the oxidation at red heat, of alloys
of lead and silver (cupellation), determination of the ash of com-
bustible bodies by roasting in the air, etc.
The numerous oxidizing or reducing reactions used in volu-
metric methods (reduction of potassium permanganate by ferrous
or manganous salts, the oxidation of sodium thiosulphate by iodine,
etc.), or in the methods by precipitation (precipitation of MnO 2 by
ammonium hydroxide and bromine or by hydrogen dioxide, of
mercuric salts by phosphorous acid, of gold by ferrous sulphate or
sulphurous acid, etc.), the ignition in air of bodies easily oxidized
(ZnS, MnS, etc.) in order to transform them into oxides with a
definite composition, belong equally to the same category.
Here may also be incorporated the eudiometric methods for the
analysis of combustible gases, containing hydrogen or carbon, by
combustion with an excess of oxygen. It is to be noted that these
methods permit the analysis of a mixture of combustible gases only
in the case of two gases whose nature is known : it is, in fact, in
this case only that the data from the experiment (contraction of
volume after combustion, volume of carbon dioxide produced) com-
pared with the equation of combustion of the two gases, furnish as
many equations as unknown quantities.
Formation of Soluble or Insoluble Complexes. Another class
of methods, very different from the preceding, consists in bringing
a single one of the bodies of the system to be analyzed into a
complex combination in which the common characteristics of the