C. Remigius Fresenius.

Quantitative chemical analysis online

. (page 59 of 69)
Online LibraryC. Remigius FreseniusQuantitative chemical analysis → online text (page 59 of 69)
Font size
QR-code for this ebook


tion takes place in this after long standing. If no precipitate is
formed the whole of the cobalt has fallen down, otherwise the
small portion must be returned to the principal solution, some
more potassium nitrite added, and after long standing the same
test applied. Thus, and thus alone, can the analyst be sure of
the complete precipitation of the cobalt. Finally filter and
treat the precipitate according to 111, 1, d. Boil the filtrate
with excess of hydrochloric acid, precipitate with potash,
redissolve the precipitate in hydrochloric acid, throw down
the nickel according to 110, 1, 5, 7, as sulphide, and then



* Pogg, Annal , LXXII, 477.

\Annal. d. Chem. u. Pharm., xcvi, 218. \ Ib., civ, 309.

Pogg. Annal., ex, 412.

I Zeitschr.f. analyt. Chem., v, 74.

. t in, 161; Journ. f. prakt. Chem. t xcvn, 387.



656 SEPARATION. [ 160.

convert into metal. In this manner alone can tlie nickel be
obtained pure, as the original filtrate contains so much alkali
salt and also generally alumina and silica.

[When nickel and cobalt are obtained in the form of
sulphides in the process of separation from other metals, the
mixed sulphides may be converted into metals without previous
separation, by the same process that is described for nickel
sulphide 110, 1, J, and 2. Cobalt may then be separated
from a nitric acid solution of the two metals and nickel estimated
by difference.]

10. Methods based upon the different deportment with
Potassium Cyanide.

a. ALUMINIUM FROM ZINC, COBALT, AND NICKEL.

Mix the solution with sodium carbonate, add potassium
cyanide in sufficient quantity, and digest in the cold, until the
precipitated zinc, cobalt, and nickel carbonates are redissolved.
Filter off the undissolved aluminium precipitate, wash, and
remove the alkali which it contains, by resolution in hydro-
chloric acid and reprecipitation by ammonia (FRESENIUS and
HAIDLEN *).

I. COBALT FROM NICKEL.

LIEBIG'S method,-)- which depends upon the conversion of
the cobalt into potassium cobalticyanide, and of the nickel into
double nickel and potassium cyanide, has been carefully studied
in my laboratory by FR. GAUHE.J It has been thus found that
boiling the solution containing potassium cyanide and hydro-
cyanic acid (LiEBio's first method) does not completely convert
the double cobalt and potassium cyanide first formed into
potassium cobalticyanide, but that passing chlorine (LiEBio's
second method) effects a ready and thorough conversion. The
method then gives a very excellent separation, and is more par-
ticularly to be recommended where the quantity of nickel is
small in proportion to the cobalt. We proceed as follows,
taking a hydrochloric solution of the nietals : Remove the
greater part of the free acid by evaporation or neutralize it by
potash, add pure potassium cyanide till the precipitate first

Mil . - . - . i . . -. i . i i

* Annal. d. Chem. u. Pharm., XLIII, 129. \2b. : LXV, 244, and LXXXVII, 1
t Zeitechr. /. analyt. Chem., v. 75.



100.] BASES OF GROUP IV. 667

formed has redissolved ; ilu-ii add more cyanide, dilute, boil for
some rime or not, as von like, pass chlorine through the cold
fluid, adding potash or >oda occasionally, so that the fluid may
remain strongly alkaline to the end. Bromine may bo used
instead of chlorine, and iudood is far more convenient. In the
course of an hour the whole of the nickel will have precipi-
tated as black hydrate of the sesqnioxide. Having taken out a
portion and satisfied yourself of this by addition of a further
quantity of chlorine or bromine, filter, and wash with boiling
water. The precipitate always retains alkali, and must be redis-
solved in hydrochloric acid, and estimated according to. 110,
1, <r, or 'J.

As regards the cobalt, it is most convenient to estimate it
by difference. But if you wish to make a direct estimation, it
will be advisable, in consequence of the large quantity of salts
present in solution, first to evaporate to dryness with excess of
hydrochloric acid, to take up the residue with a little water,
and to heat in a largo platinum dish, with the addition of
excess of pure concentrated sulphuric acid till the greater part
of the sulphuric acid has escaped. The rod mass, consisting
principally of alkali disulphato, is then treated with water, and
the cobalt estimated according to 111, 1, c.

Another method of separating nickel and cobalt by means
of potassium cyanide has been described by FLECK; * it does
not, however, appear to be in any way better. The method is
based on the fact that cobalt monosulphide, as well as nickel
sulphide, dissolves readily in potassium -cyanide solution, but
that this is not the case with the cobalt sulphide precipitated
by ammonium sulphide from a solution of cobalt which has
boon treated with ammonia in excess and exposed to the air
until its color no longer changes.

c. COBALT FROM ZINC.

Add to the solution of the two metals, which must con- 99
tain some free hydrochloric acid, common potassium cyanide
(prepared by LIKING'S method) in sufficient quantity to redis-
solve the precipitate of cobalt cyanide and zinc cyanide which

*Journ.f. prakt. Chem., xcvn, 303; ZeitscJir.f. analyt. Chem., v, 399.



656 SEPARATION. [ 160.

convert into metal. In this manner alone can the nickel be
obtained pure, as the original filtrate contains so much alkali
salt and also generally alumina and silica.

[When nickel and cobalt are obtained in the form of
sulphides in the process of separation from other metals, the
mixed sulphides may be converted into metals without previous
separation, by the same process that is described for nickel
sulphide 110, 1, >, and 2. Cobalt may then be separated
from a nitric acid solution of the two metals and nickel estimated
by difference.]

10. Methods based upon tJie different deportment with
Potassium Cyanide.

a. ALUMINIUM FROM ZINC, COBALT, AND NICKEL.

Mix the solution with sodium carbonate, add potassium
cyanide in sufficient quantity, and digest in the cold, until the
precipitated zinc, cobalt, and nickel carbonates are redissolved.
Filter off the undissolved aluminium precipitate, wash, and
remove the alkali which it contains, by resolution in hydro-
chloric acid and reprecipitation by ammonia (FRESENIUS and
HAIDLEN *).

I. COBALT FROM NICKEL.

LIEBIG'S method,f which depends upon the conversion of
the cobalt into potassium cobalticyanide, and of the nickel into
double nickel and potassium cyanide, has been carefully studied
in my laboratory by FR. GAUIIE.^: It has been thus found that
boiling the solution containing potassium cyanide and hydro-
cyanic acid (LIEBIG'S first method) does not completely convert
the double cobalt and potassium cyanide first formed into
potassium cobalticyanide, but that passing chlorine (LIEBIG'S
second method) effects a ready and thorough conversion. The
method then gives a very excellent separation, and is more par-
ticularly to be recommended where the quantity of nickel is
small in proportion to the cobalt. We proceed as follows,
taking a hydrochloric solution of the metals : Remove the
greater part of the free acid by evaporation or neutralize it by
potash, add pure potassium cyanide till the precipitate first



* Annal. d. Chem. u. Pharm., XLIIT, 129. \lb.. LXV, 244, and LXXXVII, 128,
t Zeiischr. /. analyt. Chem., v. 75.






160.] BASES OF GROUP IV. 657

formed has redissolved ; tlieii add more cyanide, dilute, boil for
some time or not, as you like, pass chlorine through the cold
fluid, adding potash or soda occasionally, so that the fluid may
remain strongly alkaline to the end. Bromine may be used
instead of chlorine, and indeed is far more convenient. In the
course of an hour the whole of the nickel will have precipi-
tated as black hydrate of the sesquioxide. Having taken out a
portion and satisfied yourself of this by addition of a further
quantity of chlorine or bromine, filter, and wash with boiling
water. The precipitate always retains alkali, and must be redis-
solved in hydrochloric acid, and estimated according to. 110,
1, a, or 2.

As regards the cobalt, it is most convenient to estimate it
by difference. But if you wish to make a direct estimation, it
will be advisable, in consequence of the large quantity of salts
present in solution, first to evaporate to dryness with excess of
hydrochloric acid, to take up the residue with a little water,
and to heat in a large platinum dish, w r ith the addition of
excess of pure concentrated sulphuric acid till the greater part
of the sulphuric acid has escaped. The red mass, consisting
principally of alkali disulphate, is then treated with water, and
the cobalt estimated according to 111, 1, c.

Another method of separating nickel and cobalt by means
of potassium cyanide has been described by FLECK ; * it does
not, however, appear to be in any way better. The method is
based on the fact that cobalt monosulphide, as well as nickel
sulphide, dissolves readily in potassium -cyanide solution, but
that this is not the case with the cobalt sulphide precipitated
by ammonium sulphide from a solution of cobalt which has
been treated with ammonia in excess and exposed to the air
until its color no longer changes.

c. COBALT FKOM ZINC.

Add to the solution of the two metals, which must con- 99
tain some free hydrochloric acid, common potassium cyanide
(prepared by LIEBIG'S method) in sufficient quantity to redis-
solve the precipitate of cobalt cyanide and zinc cyanide which

*Journ.f. prakt. Chem., xcvn, 303; ZeitscJir.f. analyt. Chem., y, 399.



558 SEPARATION. [ 160.

forms at first ; then add a little more potassium cyanide and
boil some time, ad ding t occasionally one or two drops of hydro-
chloric acid, but not in sufficient quantity to make the solu-
tion acid. After cooling, add some chlorine or bromine, and
digest for some time to complete the conversion of the cobalt
into potassium cobalticyanide. Mix the solution with hydro-
chloric acid in an obliquely placed flask and boil until the zinc
cobalticyanide which precipitates at first is redissolved, and the
hydrocyanic acid is completely expelled. Add solution of soda
or potassa in excess and boil until the fluid is clear; the solu-
tion may now be assumed to contain all the cobalt as potas-
sium cobalticyanide, and all the zinc as a compound of zinc
oxide and alkali. Precipitate the zinc by hydrogen sulphide
( 108). Filter, and determine the cobalt in the filtrate as in
98. The process is simple and the separation complete (FRE-
SENIUS and HAIDLEN).

d. NICKEL FROM ZINC.

Add to the concentrated solution of the two metals an 100
excess of pure concentrated potassa lye, then sufficient
aqueous hydrocyanic acid to completely redissolve the precipi-
tate, add a solution potassium monosulphide (not ammonium
sulphide), let the precipitated zinc sulphide deposit at a gentle
heat, filter, wash the sulphide with a dilute potassium-sul-
phide solution, treat the precipitate with hydrochloric acid,
and from the solution precipitate the zinc with sodium car-
bonate, as in 108, 1, a. In the filtrate estimate the nickel
by heating for some time with fuming hydrochloric and
nitric acids, or instead of the latter, potassium clilorate, evap-
orating, and finally precipitating with potassa lye (WOHLER *).
KLAYE and DEUs,f who tested the process in my laboratory,
found that instead of potassa lye and hydrocyanic acid,
potassium cyanide could be used if perfectly pure and
recently dissolved. If the solution of the cyanide contains
ammonium carbonate or formate, or potassium cyanate (as
is the case even on short exposure), the complete precipita-
tion of the zinc as sulphide is greatly interfered with. On

* Annul, de Chem. ?/. Pharm., LXXXIX, 37G.
f Zeit8chr. f. analyt. CJu />/ . , x , 1 1)7.



160.] BASES OF GROUP IV. 659

finally washing the precipitated zinc sulphide completely with
water containing hydrogen sulphide, the zinc may be esti-
mated according to 108, 2.

e. COBALT AND NICKEL FKOM MANGANESE AND ZINC
(W. GIBBS*).

Add sodium acetate to the solution of the chlorides and 101
pass in hydrocyanic-acid gas. Zinc cyanide is immediately
more or less completely precipitated as a white powder. Now
add sodium sulphide, which converts the zinc and manganese
into sulphides, while the cobalt and nickel remain in solution
as double cyanides, and may be separated as in 98. The
employment of gaseous hydrocyanic acid renders the method
very unpleasant.

Another method for separating cobalt and nickel from
manganese is as follows: To the acid solution add sodium
carbonate in excess, then acetic acid in liberal excess, then to
the clear fluid, containing say 1 grm. of nickel or cobalt, 30 to
40 c. c. of sodium-acetate solution (1 in 10), and pass hydro-
gen sulphide to saturation, keeping at 70. Filter off the
precipitated nickel or cobalt sulphide, wash and dry it. Con-
centrate the filtrate by evaporation, add ammonium sulphide
and then acetic acid, thus obtaining a second precipitate of
nickel or cobalt sulphide. Test the filtrate again in the same
manner. In the united precipitates determine the nickel or
cobalt according to 110, 1, &, or 111, 1, <?; in the fil-
trate, the manganese according to 109, 2.

11. Methods based on the Volatility of Zinc.
a. COBALT AND NICKEL FROM ZINC.

BEKZELIUS f gives the following method for the absolute 102
separation of cobalt and nickel from zinc : Precipitate the
solution w T ith an excess of potassa lye, boil, and filter the
fluid containing the greater portion of the zinc oxide dis-
solved in the potassa solution from the precipitated nickel and
cobalt hydroxides also containing much zinc, wasli completely

* ZeitscJir. /. anatyl. Chem., in, 332.
f BERZELIUS' Jahresber., xxi, 144.



660 SEPARATION. [ 160.

with boiling water, and determine the zinc in the filtrate (see
108). Dry the precipitate, ignite, and weigh; then mix
with pure sugar (recrjstallized from alcohol) in a porcelain
crucible and heat slowly until all the sugar is completely
carbonized. Then place the crucible, covered with its lid,
in a bath of magnesia within a covered crucible of larger
size, and heat in a wind furnace for one hour at the highest
temperature obtainable. By this treatment the metals are
reduced, the nickel and cobalt mixed with carbon remaining
behind, while the zinc is volatilized. Treat the residue with
nitric acid, and determine the metals by precipitating with
potassa solution and weighing the precipitate. The differ-
ence between this weight and that obtained before gives the
weight of the zinc which had been conjointly precipitated.
KLAYE and DEUS,* who tested the method in my laboratory,
obtained good results with it. They recommend replacing
the sugar by sugar-carbon, as the former causes much intu-
mescence. An attempt to use the gas blowpipe, as given by
BERZELIUS, gave poor results.

J. ZINC FROM IRON IN ALLOYS.

According to BOBIERRE such alloys may be readily and 103
accurately analyzed by ignition in a current of hydrogen.

12. Methods based upon the Volumetric Determi-
nation of one of the Metals, and the finding of the
other from the difference.

a. FERRIC IRON FROM ALUMINIUM.

Precipitate both metals with ammonia ( 105, a, and 104
113, 1). Dissolve the weighed residue, or an aliquot part
of it, by digestion with concentrated hydrochloric acid, or by
fusion with potassium bisulphate and treatment with water
containing sulphuric acid, and determine the iron volumetri-
cally as directed in 113, 3, a or J. The alumina is found
from the difference. This is an excellent method, and to be
recommended more particularly in cases where the relative
amount of iron is small. .If you have enough substance, it is

* Zeitschr. /. analyt. Chem., x, 192.



160.] > BASES Ofc-GROUP IV. 661

of course much more convenient to divide the solution, by
weighing or measuring, into 2 portions, and determine in the
one the sesquioxide of iron + alumina, in the other the
iron. Instead of titrating the iron, this may be precipitated
with ammonium sulphide after the addition of tartaric acid
and ammonia (77).

J. FERRIC IRON FROM FERROUS IRON (ZiNC AND NICKEL).

a. Determine in a portion of the substance the total 105
amount of the iron as sesquioxide, or by the volumetric way.
Dissolve another portion by warming with sulphuric acid in
a flask through which carbonic acid is conducted, to exclude
the air ; dilute the solution and determine the ferrous
iron voluinetrically ( 112, 2, a). The difference gives
the quantity of the ferric iron. Or, dissolve the com-
pound in like manner in hydrochloric acid, and determine
the ferric chloride with sodium thiosulphate according to
113, 3, J. In this case the difference gives the ferrous
iron. If it is desired to determine the ferrous chloride in
the hydrochloric-acid solution directly, it will be well to use
PENNY'S methods ( 112, 2, ). If the compound in which
the ferrous and ferric basic radicals are to be estimated is de-
composed by acids with difficulty, heat it with a mixture of 4
parts sulphuric acid and 1 part water (or with hydrochloric
acid) in a sealed tube for 2 hours at 210 (MITSCHERLICH *)
(see page 521). Or, if this is not enough, fuse it with borax
(1 part mineral, 5 to 6 vitrified borax) in a small retort con-
nected with a flask containing nitrogen (produced by com-
bustion of phosphorus in air) ; an atmosphere of carbonic acid
is less suitable. Triturate the fused mass with the glass,
and dissolve in boiling hydrochloric acid in an atmosphere
of carbonic acid (HERMANN v. KOBELL). Or, as will gener-
ally be the best way, you may dissolve the substance in a
mixture of hydrofluoric and hydrochloric or hydrofluoric and
sulphuric acids with exclusion of air (83). COOKE f dissolves
silicates in a mixture of sulphuric and hydrofluoric acids in

*Journ.f. prakt. Chem., LXXXI, 108, and LXXXIII, 455.
f Amer. Journ. of Science, 2d ser., LXIV, 347.



662 SEPARATION. [ 100.

an atmosphere of steam and carbonic acid, and determines
the ferrous iron by means of potassium permanganate.

Fig. 114 exhibits his apparatus. To the sides of a copper
water-bath are attached three tubes. The tube on the left COD-




Fig. 114.

nects with a MARIOTTE'S flask to maintain the water at a constant
level. The upper tube on the right connects with a carbonic-
acid gas generator, while the third tube carries off any over-
flow of water to the sink.

On the cover of the water-bath close to the rim is a circular
groove which receives the edge of an inverted glass funnel.
When the apparatus is in use this groove is kept full of water
by the spray from the boiling liquid, and thus forms a perfect
water-joint ; but in order to secure this result the bath must be
kept nearly full of water, and holes for the ready escape of the
steam and spray should be provided in the rings, which cover
the bath and adapt it for vessels of various sizes. By this
arrangement the funnel may be kept filled with an atmosphere
of steam or of carbonic acid for an indefinite period. More-
over, we can either pour in fresh quantities of solvent, or we
can stir up the material, in the vessel within, introducing a
tube-funnel or stirrer through the spout of the covering funnel.

The finely pulverized substance (-J to 1 grm.) is placed in a
large platinum crucible. Upon it pour a mixture of dilute
sulphuric acid (sp. gr. 1*5) with as little hydrofluoric acid as
experience may show is required to dissolve or decompose the
substance, stirring up the material with a platinum spatula.



160.] * BASES OF GROUP IV. 663

The crucible is next transferred to the water-bath, the covering
funnel put in place, water poured into the groove, the interior
filled with carbonic acid, and the lamp lighted. As soon as the
water boils, the supply of carbonic acid is stopped ; arid if the
water-level has been properly adjusted, the apparatus will take
care of itself, the groove will be kept full of water, and the
interior of the funnel full of steam. If the materials cake on the
bottom of the crucible, as is not unfrequently the case when a
large amount of insoluble sulphate is formed, the lamp may be
removed, the apparatus again tilled with carbonic acid, and the
contents of the crucible stirred up by aid of a stout platinum
wire about two inches long, fused to the end of a glass tube.
Anything adhering to the rod can easily be washed back into
the crucible by directing the jet from the wash-bottle down the
throat of the covering funnel. The lamp may then be replaced,
the current of carbonic acid interrupted, and the process of
digestion continued. When the decomposition is complete, the
current of carbonic acid gas is re-established, the lamp extin-
guished, and the air-tube of the Mariotte'fi flask raised until its
lower end is above the level of the overflow. A slow current
of water is thus caused to flow through the bath, which soon
cools down the whole apparatus. The crucible may now be
removed, its contents washed into a beaker-glass, and the solu-
tion diluted with pure water until the volume is about 500 c.c.,
when the amount of ferrous iron present can be determined
with a solution of potassium permanganate in the usual way.

Many iron compounds in fine powder are completely decom-
posed by boiling a few minutes only with the mixed acids
above mentioned. If a preliminary experiment shows this to
be the case, a simple and satisfactory way of effecting a solu-
tion is to boil the substance with the solvent acids in a platinum
crucible of 40 to 50 c.c. capacity, provided with a well-fitting
concave cover. By watching the escaping vapor, one can regu-
late the boiling so as to prevent access of air without appreciable
mechanical loss. If on removing the cover the decomposition
is complete, the operation may be considered successful. Put
the crucible and its contents at once into cold water in a beaker
and titrate with permanganate (or thiosulphate if HC1 has been
used).

Iron may also be determined volumetrically in presence of 106



664



SEPARATION.



[ 160.



zinc, nickel, etc. It is, indeed, often the better way, instead
of effecting the actual separation of the oxides, to determine
in one portion of the solution the iron -f- zinc or -f- nickel, in
another portion the iron alone, and to find the quantity of the
other metal by the difference. However, this can be done only
in cases \vhere the quantity of iron is relatively small.

fi. FERRIC IRON FROM FERROUS IRON (BUNSEN).

Fill the flask d, Fig. 89, two-thirds full with fuming 107
hydrochloric acid and replace the air in the flask by carbon
dioxide by throwing a few fragments of sodium carbonate into

the acid. Then immediately in-
troduce the substance previously
weighed off in a small tube and
add a slight excess of potassium
dichromate, also contained in a
similar tube, attach the evolution
tube, and proceed for the rest as
in 130, e, fi. Of course less
free iodine is obtained than had
no potassium dichromate been dis-
solved with the ferrous salt, as a
portion of the chlorine evolved is
used up in converting the ferrous
into ferric chloride. In fact the
difference between the iodine
corresponding to the dichromate
used and that actually obtained
corresponds to the ferrous iron present, hence 1 eq. of iodine
= 1 eq. of ferrous chloride.

To determine the total quantity of iron present, dissolve
another portion of the sample as before in acid in the flask,
and effect the reduction of the ferric to ferrous iron by intro-
ducing a ball of chemically pure zinc cast on the end of a fine
platinum wire. To exclude all access of air, connect the flask
during the boiling with the apparatus lib' shown in Fig. 115.
As soon as the reduction is complete, and which may be
recognized by the colorless appearance of the fluid, cool the
flask by immersing it in cold water, lift the upper stopper,




Fig. 115.



160.] " BASES OF GROUP IV. 665

throw a few fragments of sodium carbonate into the acid,
draw the zinc ball up the tube 5, wash off the fluid adhering
to the ball into the flask, and remove W. Now quickly add
the weighed potassium dichromate and proceed as above
directed.

c. MANGANESE FROM ALUMINIUM AND IRON (KEIEGEE *).

Precipitate with sodium carbonate, digest the precipitate 108
for some time with the fluid, wash first by decantation, then on
the filter, and as thoroughly as possible, dry, ignite, and deter-
mine the manganese in a portion according to 72. Care must
be taken that the precipitate contains the manganese as Mn 3 O 4 ,
and also that, in the case of highly accurate analyses, the small



Online LibraryC. Remigius FreseniusQuantitative chemical analysis → online text (page 59 of 69)