U.S.

Department of Commerce, Bureau of Standards

RESEARCH PAPER RP655

Part of Bureau of Standards Journal of Research, vol. 12, March 1934

METHODS FOR THE SEPARATION OF PLATINUM, PALLA-

DIUM, RHODIUM, AND IRIDIUM FROM ONE ANOTHER
AND FOR THEIR GRAVIMETRIC DETERMINATION
By Raleigh Gilchrist

ABSTRACT
A method is described for the analytical separation of platinum from each or
all ofthe three metals, palladium, rhodium, and iridium. It is based upon the
complete precipitation of the hydrated dioxides of palladium, rhodium, and iridium
in a boiling solution containing sodium bromate and having an acidity of pH 6.
Two precipitations of the hydrated dioxides were found to be sufficient to effect
the quantitative isolation of platinum.
The separation of palladium from rhodium and iridium, by means of dimethyl-
glyoxime, and its subsequent determination are also described. The procedure
for the separation of rhodium from iridium by titanous chloride and for the deter-
mination of these two metals, given in detail in a previous publication from this
laboratory, is discussed only briefly in this paper.

CONTENTS Page
I. Introduction 291
II. Materials used in the experiments 293
III. Separation of platinum from palladium, rhodium, and iridium 293
1. Method recommended for the separation of platinum from
palladium, rhodium, and iridium, either collectively or
singly 294
(a)Preparation of the solution 294
(6) Hydrolytic precipitation of palladium, rhodium, and
iridium 294
IV. Discussion of the hydrolytic procedure 295
V. Determination of platinum 296
VI. Determination of palladium 298
VII. Determination of rhodium and of iridium 299
VIII. Experimental results on the separation of platinum from palladium,
rhodium, and iridium 300
IX. Separation of palladium from rhodium and iridium 302
X. Separation of rhodium from iridium 303

I. INTRODUCTION
In a systematic scheme for the analysis of the platinum group, it is
the usual procedure to remove osmium and ruthenium first as volatile
tetroxides and then to effect the separation of the four
291
292 Bureau of Standards Journal of Research [Vol. i%

platinum metals from one another. It has been the practically uni-
versal custom to effect these latter separations by means of potassium
chloride or ammonium chloride. These reagents produce relatively
insoluble compounds of the type K 2 PtCl 6 (NH 4 ) 2 IrCl 6 etc., with the
, ,

quadrivalent platinum metal chlorides and soluble compounds with

the ter- and bi-valent chlorides. In spite of the general use of
potassium chloride or of ammonium chloride to effect the separation
of iridium from rhodium, or of platinum from rhodium, or of platinum
from iridium, these reagents can hardly be expected to produce
clean-cut separations. Experience in this laboratory with the purifica-
tion of the platinum metals has shown that rhodium contaminates
iridium, when the latter is precipitated as ammonium chloroiridate,
and that it can be completely eliminated only with great difficulty, if
at all, by repeated precipitation of the iridium. Rhodium likewise
contaminates platinum, when the latter is precipitated as ammonium
chloroplatinate. Since the potassium and the ammonium salts of the
quadrivalent platinum metal chlorides are slightly soluble, a small
amount of iridium or of platinum must remain in solution, thereby
contaminating the very metal from which complete separation is
sought. This situation, unfortunately, is often disregarded. Even
though the separations are followed by treatment of the metal residues
with acids or with molten potassium bisulphate, again a sharp separa-
tion cannot be expected. In using such methods of separation, it
often becomes necessary to combine the determinations of a number of
separate fractions in evaluating the determination of a particular
constituent.
The importance of hydrolytic reactions in analytical separations
appears to have been given but little attention by chemists in the past.
The reason for this may possibly be a lack of means for carefully
controlling the pH of the solutions. It is evident that hydroxides of
the platinum metals were known, but their significance in analysis
does not seem to have been realized. The precipitation of hydrated
tervalent rhodium oxide by alkalies and by alkali carbonates, for
instance, was well known, but no use was made of the reaction in the
analytical separation of rhodium. In the purification of platinum,
Doebereiner l did propose a method by which \h&. impurities were
collectively precipitated by hydrolysis, using milk of lime. Claus, 2
however, pointed out that the elimination of impurities by this method
was never complete. About 20 years ago, Mylius and Mazzucchelli 3
proposed a similar collective precipitation of impurities for use in the
approximate analysis of platinum, except that an oxidizing reagent,
sodium hypobromite, was added to a neutral or slightly alkaline
solution to produce hydrated oxides of higher valence. Precipitation
by hydrolysis was later studied by Wichers 4 in the development of an
analytical method for the separation of rhodium from platinum and
in the refining of platinum. 5 In the analytical separation, a definite
alkalinity was established by the addition of an excess of freshly
prepared barium carbonate. In the refining operation, precipitation
was made in a solution slightly alkaline with sodium bicarbonate to
which sodium bromate was added. These researches of Wichers
1
Franz Doebereiner, Ann., vol. 14, pp. 17, 251, 1835.
2 O. Claus, Bull. acad. sci., St. Petersburg, vol. 3, p. 38, 1845.
3 F. Mylius and A. Mazzucchelli, Z. anorg. Chem., vol. 89, p. 1, 1914.
* E. Wichers, Jour. Am. Chem. Soc, vol. 46, p. 1818, 1924.
« E. Wichers, R. Gilchrist, and W. H. Swanger, Trans. Am. Inst. Mining Met. Eng., vol. 76, p. 602, 1928.
Gilchrist] Separation of Pt, Pd, Rh, and Ir 293

clearly demonstrated the value of hydrolytic reactions in the chemis-

try of the platinum metals.
The use of hydrolytic precipitation was further extended by Gil-
christ in studies on the analytical separation and determination of
ruthenium, 6 7 osmium, 8 and iridium. 9 In these studies, it was
found that ruthenium is completely precipitated from a boiling solu-
tion having an acidity of 6;
10
pH
that osmium is quantitatively pre-
cipitated between pH
1.5 and 6; and that iridium is entirely precipi-
tated between pH 4 and 6.
Quite recently, Moser and Hackhofer n have reported on the
separation of iridium from platinum, and Moser and Graber 12 on
the separation of rhodium from platinum, by a hydrolytic procedure
in which the alkalinity is fixed by the reaction between bromate and
bromide ions. This reaction, however, appears to come to equilibrium
so slowly that prolonged boiling and digestion is required to complete
the precipitation.
The purpose of this paper is to present the results of the application
of hydrolytic precipitation to the separation of rhodium from platinum,
of iridium from platinum, of palladium from platinum, and of the
collective separation of rhodium, iridium, and palladium from platinum.
In addition, the paper contains the results of experiments on the
determination of palladium, as well as its separation from rhodium
and iridium, by means of dime thy lglyoxime.

II. MATERIALS USED IN THE EXPERIMENTS

The platinum metals which were used throughout the investigation
were purified by methods previously described 13 and were found to
be free from impurities by spectro chemical analysis. In order to
avoid errors incident to volumetric measurements, weighed portions
of the ammonium salts of the chloro-acids of rhodium and of iridium
were used. The compounds were carefully mixed by grinding in an
agate mortar and their metallic contents determined by direct ignition
of portions of the salts in hydrogen, whenever a series of samples was
weighed. The metallic contents were never calculated from the
theoretical compositions of the salts. With platinum and palladium,
weighed portions of the pure metals in the form of foil were taken.

III. SEPARATION OF PLATINUM FROM PALLADIUM,

RHODIUM, AND IRIDIUM
The simplest and most direct way to separate platinum from palla-
dium, rhodium, and iridium is to precipitate collectively by hydrolysis
the hydrated oxides of palladium, rhodium, and iridium. The method
recommended for this separation is given below and is equally suitable
for the separation of platinum from each of the other three metals
when present singly with platinum.
« Gilchrist, B.S. Jour. Research, vol. 3, p. 993, ]929.
1 Gilchrist, B.S. Jour. Research, vol. 12, p. 283, 1934.
s Gilchrist, B.S. Jour. Research, vol. 6, p. 421, 1931.
» Gilchrist, B.S. Jour. Research, vol. 9, p. 547, 1932.
10 It is to be pH
noted that the values of mentioned in this paper were determined by means of indicator
solutions and were not corrected to correspond to readings of the hydrogen electrode.
" L. Moser and H. Hackhofer, Monatsh., vol. 59, p. 44, 1932.
12 L. Moser and H. Graber, Monatsh., vol. 1932.
59, p. 61,
» See footnote 5, p. 292.
294 Bureau of Standards Journal of Research [Vol. it

1. METHOD RECOMMENDED FOR THE SEPARATION OF PLATINUM

FROM PALLADIUM, RHODIUM, AND IRIDIUM, EITHER COL-
LECTIVELY OR SINGLY
(a) PREPARATION OF THE SOLUTION
Evaporate the solution, containing platinum, palladium, rhodium,
and iridium, to a moist residue on the steam bath in a 600-ml beaker.
Add 5 ml of hydrochloric acid l4 and again evaporate, repeating this
operation until nitroso-compounds are decomposed. Add 2 g of
sodium chloride and 5 ml of hydrochloric acid and evaporate this
time to dryness. Add 2 ml of hydrochloric acid and dilute the solu-
tion to 300 ml with water.
(b) HYDROLYTIC PRECIPITATION OF PALLADIUM, RHODIUM, AND IRIDIUM
Heat the resulting solution to boiling and add to it 20 ml of a filtered
10 percent solution of sodium bromate. Carefully add a filtered 10
percent solution of sodium bicarbonate until the dark green solution
shows evidence of the formation of a permanent precipitate. Test
the acidity of the solution by allowing a drop of brom cresol purple
indicator solution (0.04 percent) to run down the stirring rod into the
drop which clings to it as it is lifted from the hot solution. Sufficient
bicarbonate has been added when the color of the indicator changes
from yellow to blue. At this stage, add 10 ml more of the bromate
reagent and boil the solution gently for 5 minutes. Test the acidity
of the solution again and adjust it to that specified above, if necessary.
The tendency is for the pH of the solution to shift to the acid side
of the indicator as hydrolysis of the platinum metal compounds
progresses. Again add 10 ml of the bromate solution and boil gently
for 15 minutes. During this treatment, the acidity of the hot solution
should stay constant.
As a precautionary measure, it is well to test the acidity again at
the end of the second boiling period. On removing the beaker from
the source of heat, the precipitate which has been formed will settle
quickly, leaving a yellow mother liquor containing the platinum.
Filter 15 the solution. Pour the supernatant liquid through first, then
transfer the precipitate. Rinse the beaker, without attempting to
remove material which clings to the walls, and wash the precipitate
with a hot 1 percent solution of sodium chloride. It is important that
the acidity of the wash solution be tested before it is used because
hydrated palladium dioxide is attacked somewhat by hot slightly acid
solutions. A
solution of sodium chloride whose acidity does not exceed
that at which brom cresol purple indicator contained in it changes
color can be safely used. If the precipitate is caught on filter paper,
return the filter to the beaker and dissolve the hydrated dioxides in
" Throughout this paper, whenever acids are mentioned, hydrochloric acid will mean the concentrated
acid of specific gravity 1.18; nitric acid, the concentrated acid of specific gravity 1.42; and sulphuric acid,
the concentrated acid of specific gravity 1.84. Diluted acids will be designated as follows: For example,
diluted hydrochloric acid (1+99) will mean one volume of the concentrated acid of specific gravity 1.18
diluted with 99 volumes of water.
is The solution may be filtered through paper. It is highly desirable, however, to avoid paper if the
palladium, rhodium, or iridium is to be reprecipitated. Iridium dioxide, which dissolves much less readily
than either palladium or rhodium dioxide, tends to stain the paper pulp and cannot always be removed by
washing. The material of which the paper is composed undoubtedly reacts with acids to some extent and,
as a consequence, small quantities of organic compounds of the platinum metals appear to be formed which
are not easily hydrolyzed. The difficulties encountered with filter paper are avoided by using a porcelain
filtering crucible, having glazed walls and a porous base. Such crucibles are made by the State Porcelain
Works, Berlin, Germany. The ones used in this work were 45 mm mm
in height, 40 in diameter at the top
and had a capacity of 30 ml. While no information could be obtained regarding their degree of porosity,
they bore the mark " Al and were very satisfactory. No trace of the precipitates was ever detected to Lave
'
'

escaped these filters. No doubt, glass crucibles of suitable porosity could also be used.
Gilchrist) Separation of Pt, Pd, Rh, and Ir 295

diluted hydrochloric acid (1 + 1 ) If the filtering crucible is used, place

.

it with the stirring rod in the beaker used for the precipitation. It
may be necessary to remove a small amount of the precipitate which
has crept over the lip of the beaker. This may be done with a very
small piece of filter paper and added to the beaker. Replace the
watch glass and add from 10 to 20 ml of hydrochloric acid, pouring a
portion of it into the crucible and the remainder into the beaker.
Place the covered beaker on the steam bath. The rhodium and pal-
ladium compounds will dissolve quickly. The iridium compound will
dissolve more slowly and may even require a drop of nitric acid to be
added to the solution. Carefully lift the crucible with the stirring
rod, wash it with water and place in in a 250-ml beaker. The porous
bottom will be stained somewhat. Add about 5 ml of hydrochloric
acid, again pouring a portion of the acid into the crucible. Cover the
beaker with a watch glass and set it on the steam bath. This treatment
will usually leach out the small quantity of metal chloride present in
the porous bottom. This operation should be repeated once more with
fresh acid, however, to insure complete removal of metal chloride.
Combine the leachings with the main portion of the solution, add
2 g of sodium chloride, and evaporate to dryness on the steam bath.
Add 2 ml of hydrochloric acid, dilute the solution to 300 ml with water,
and repeat the precipitation of the hydrated dioxides as directed
above. Two such precipitations have been found sufficient to separate
platinum from rhodium, palladium, and iridium. 16

IV. DISCUSSION OF THE HYDROLYTIC PROCEDURE

The normal valence of rhodium in chloride solution is three; that
of iridium, three or four; that of palladium, two; and that of platinum,
owing to the conditions necessary to dissolve it, four.
When a solution of tervalent rhodium as halide or sulphate is
neutralized, the rhodium precipitates as a yellow hydrated tervalent
oxide. This compound does not settle well and has a tendency to
become colloidal when washed. For these reasons, the quantitative
handling of it is somewhat difficult. If, on the other hand, the rhodium
is oxidized to the quadrivalent state, it can be precipitated as an
olive-green hydrated dioxide which does settle rapidly and which
displays no tendency to become colloidal. It was found that complete
precipitation of the olive-green compound occurred at 6 in a pH
boiling solution.
The precipitate of hydrated tervalent iridium oxide is somewhat
flocculent and gelatinous, whereas the deep green hydrated dioxide is
more granular and settles quickly from solution. It was found that
the dioxide is completely precipitated from a boiling solution over a
range of acidity, pH
4 to 6.
19 In the experiments reported in section VIII of this paper, the ratio of platinum to rhodium was approxi-
mately 5 to 1, that of platinum to palladium 2.5 to 1, and that of platinum to iridium 1 to 1. The quantity
of platinum present was usually about 250 milligrams. In the Pt-Rh separation, no platinum was found
in the rhodium solutions after the second hydrolytic precipitation, and in only one of the five experiments
was any rhodium detected in the platinum fractions. The quantity detected in this instance was less than
0. 1 mg. In the Pt-Pd separation, no palladium was detected in the recovered platinum nor was any platinum
found in the nitrates from the precipitation of the palladium with dimethylglyoxime. In the Pt-Ir separa-
tion, only one of the three platinum fractions appeared to contain any iridium. The amount of iridium,
however, was much less than 0.1 mg. For certain work in which a high degree of accuracy in the platinum
determination is not essential, a single hydrolytic precipitation may be sufficient. It must be borne in mind,
however, that a small amount of platinum will contaminate the hydrolytic precipitate. When careful work
is being done, it may be necessary to make a third hydrolytic precipitation, but only in unusual cases.
296 Bureau oj Standards Journal of Research [Vol. 12

Hydra ted palladium dioxide, obtained in the presence of bromate, is

brown in color, is completely precipitated at pH 6, and likewise settles
well from solution.
Owing to the possible interference of nitroso compounds, it is
recommended that the solution first be evaporated several times with
hydrochloric acid. On a number of occasions when solutions contain-
ing rhodium had not been so treated, a small amount of rhodium
failed to precipitate on hydrolysis. When the solutions were eva-
porated with hydrochloric acid, complete precipitation was always
obtained. Unless the solution contains a sufficient amount of sodium
chloride, it should not be evaporated to dryness. A solution containing
only rhodium chloride, for instance, when evaporated to dryness on
the steam bath, forms brown scales which do not redissolve readily in
water. If sodium chloride is present, a double salt is formed which
is easily soluble.
In the precipitation of rhodium and of palladium, sodium bromate
is add 3d to the slightly acidified solution. This is done to prevent the
precipitation of tervalent rhodium and bivalent palladium oxides
when the bicarbonate is added. Some of the bromate is decomposed
by the acid solution so that it is advisable to add this reagent at
intervals. A hot solution of tervalent iridium can be neutralized to
pH 6 without the formation of a precipitate. In the recovery of
iridium from a pure solution, it is often convenient to neutralize to
pH 6 before adding the bromate. As the bicarbonate solution is
added to the acidified bromate solution, an olive-green precipitate
appears with rhodium, a dark green one with iridium, and a brown
one with palladium. These precipitates immediately redissolve,
particularly those of rhodium and palladium, but, as the acidity is
reduced, permanent precipitates form. The appearance of a coagu-
lated precipitate is a convenient indication that the acidity of the
solution is approaching the desired value, namely, pH 6. Owing to
the deep color of the solution, the endpoint cannot be determined in
the usual manner. A convenient method, however, consists in testing
a drop of the hot solution, clinging to the stirring rod, with a drop of
brom cresol purple indicator solution. The actual acidity of the
solution will be slightly greater than that indicated by the drop.
Experiments have repeatedly shown, however, that precipitation is
quantitative at the acidity determined in this manner.
It sometimes happens that the salt used in preparing the wash
solution contains a small amount of acid. Hydrated iridium dioxide
can be safely washed with solutions as acid as pH 4. Hydrated
rhodium dioxide, and more particularly the palladium compound, are
attacked by slightly acid solutions and should be washed only with
solutions which are practically neutral. It is well, therefore, to test
the acidity of the wash solution before using it. A solution more acid
than that which causes a change in color of brom cresol purple indi-
cator contained in it should not be used.

V. DETERMINATION OF PLATINUM
The following procedure is recommended for the recovery and
determination of platinum, after its separation from palladium,
rb odium, and iridium.
Add 20 ml of hydrochloric acid to each of the cooled filtrates from
the hydrolytic precipitation. Warm the solutions carefully until
Gilchrist] Separation of Pt, Pd, Eh, and Ir 297

they become quiescent. Combine the partially concentrated nitrates

and then evaporate to dryness. To make certain that all of the bro-
mate is decomposed, add 5 ml of hydrochloric acid and again evapo-
rate. Dissolve the residue in 100 mt of water and filter the resulting
solution. Wash the filter with diluted hydrochloric acid (1 + 99).
Add 20 ml of hydrochloric acid and dilute the solution to 400 ml.
Precipitate the platinum as sulphide from a gently boding solution
by passing a rapid stream of hydrogen sulphide through it. Allow
the solution to cool somewhat with the hydrogen sulphide still passing
through it.
Filter the solution and wash the sulphide precipitate with diluted
hydrochloric acid (1 + 99). Ignite the dried filter and precipitate in a
porcelain crucible. Leach the metal residue with diluted hydro-
chloric acid, transfer it to a filter, and wash it thoroughly with hot
water. Ignite the filter and metal again strongly in the air. It is
usually customary to weigh the metal so obtained. A slight error
will result, however, because sulphur cannot be entirely eliminated
by ignition either in air or in hydrogen. The amount of sulphur
retained by 0.25-g portions of platinum is usually less than 1 milligram.
In work of the highest accuracy, dissolve the metallic platinum
obtained by ignition of the sulphide in aqua regia. Decompose
nitroso compounds by adding hydrochloric acid and evaporating to a
moist residue. Filter the solution into a clean, unetched beaker.
Wash the filter with diluted hydrochloric acid (1 + 99). Dilute the
so^tion to 100 ml, heat it to boiling, and add to it a solution containing
3 g of sodium acetate and 1 ml of formic acid for each 0.25 g of plati-
uum present. Boil the resulting solution gently until the precipitated
metallic platinum is well coagulated and the supernatant liquid is
colorless. Filter the solution and wash the metallic deposit with a
hot 1 percent solution of ammonium chloride. Place the filter and
the spongy metal in a porcejain crucible and ignite them strongly in
the air. Leach and wash the ignited metai as previously directed.
This precaution is taken to remove traces of soluble salts. Fiually
ignite the platinum again in air. It will now be free from suipnur.
Weigh as metallic platinum.
In the determination of the platinum metals, by ignition of sul-
phide precipitates, a significant error may be caused by the retention
of sulphur. It is well known that when palladium sulphide is ignited
elimination of sulphur is very difficult to attain. In fact, if the sul-
phide is ignited rapidly enough, it will often fuse into a silvery globule.
While the retention of sulphur by platinum is not nearly so marked
as it is by palladium, a small but significant amount does remain
when the sulphide is ignited in air. In 1 7 experiments, the weights of
platinum recovered by ignition of the sulphide in air exceeded those
taken by amounts which ranged from 0.0 to 1.1 mg and averaged
0.5 mg. In 14 of the 17 experiments, the platinum residues were recov-
ered after separation from palladium, rhodium, and iridium. In the
other three, pure platinum was dissolved in aqua regia and simply pre-
cipitated by hydrogen sulphide. These three residues were alternately
ignited in air and in hydrogen. The excesses in weight dropped from
1.1 mg, 0.5 mg, and 0.7 mg to 0.6 mg, 0.2 mg, and 0.4 mg, respectively,
by this treatment. These 3 residues and 4 others were then dissolved
in aqua regia and the resulting solutions evaporated several times with
hydrochloric acid. The solutions were filtered and to each of them
298 Bureau of Standards Journal of Research [vol. it

was added § drops of a filtered 10 percent solution of barium chloride.

A precipitate of barium sulphate was obtained in each solution. No
precipitate of barium sulphate was observed in blank solutions which
contained the same quantities of nitric and hydrochloric acids.
Wichers, 17 a few years ago, found errors of the same magnitude,
when similar quantities of platinum sulphide were ignited in air.
When he charred and ignited the filters and precipitates extremely
slowly, however, no significant errors were observed and the platinum
residues were found to contain only a few hundredths of a milligram of
sulphur. When he recovered the platinum with formic acid, the
results were consistent with those reported in this paper. Wichers also
extended his observations to the ignition of rhodium sulphide. He
found no significant amount of sulphur in the resulting residues nor
did he find that the manner of ignition, whether rapid or slow, affected
the weight of the recovered metal. These conclusions were later
confirmed by the writer.

VI. DETERMINATION OF PALLADIUM

Since the introduction of dimethylglyoxime as a precipitant for palla-
dium by Wunder and Thuringer, 18 this organic compound has prac-
tically, if not entirely, displaced reagents formerly used, such as mer-
curic cyanide and potassium iodide. Wunder and Thuringer, accord-
ing to their original articles, ignited the compound of palladium with
dimethylglyoxime to metal for purposes of weighing. Soon, however,
it became customary to weigh the dried compound as is done in the
case of nickel.
The preparation of palladium in the metallic form suitable for
weighing has always been attended by the difficulty which arises from
the pronounced property of palladium to absorb hydrogen. When
palladium is determined as metal, Swanger 19 recommended the decom-
position of the dried glyoxime precipitate in a Rose crucible under
hydrogen, followed by an ignition in air and another in hydrogen. In
order to eliminate the absorbed hydrogen, he heated the crucible again
in air for a moment, and finally reduced the slightly oxidized metal
with formic acid on the hot plate. The treatment with formic acid
often has to be repeated several times in order to produce a sponge
which does not have a slight purplish color.
According to Wohler, 20 palladium dioxide, Pd0 2 cannot exist above ,

200 C, but decomposes into the monoxide, PdO, and oxygen. The
monoxide, in turn, has a dissociation pressure of 760 at 877 C. mm
Experiments showed that metallic palladium, silvery grey in color,
can be obtained by simply igniting in carbon dioxide for 2 minutes the
palladium which has been reduced in hydrogen. It was also found
that the same result was obtained by igniting oxidized palladium in
carbon dioxide. It was further found that simple strong ignition of
oxidized palladium in the air formed the metal, which, however, took
on a purplish color as it cooled to room temperature. The quantity
of oxide thus formed was found to amount to only 0.1 mgjwith] portions
of palladium weighing 100 mg.
17 Unpublished work by Edward Wichers.
is M. Wunder and V. Thuringer, Z. anal. Chem., vol. 52, pp. 101, 660, 740, 1913.
i» W. H. Swanger, Analysis of Dental Gold Alloys, B.S. Scientific Papers No. 532, Aug. 11, 1926.
so L. Wohler, Z. Elektrochem., vol.
11, p. 836, 1905.
Gilchrist] Separation of Pt, Pd, Rh, and Ir 299

Three experiments were made in which the glyoxime precipitate was

caught on filter paper, washed with diluted hydrochloric acid (1 + 99),
and finally with hot water. The filter containing the precipitate was
wrapped in a second filter and charred carefully in the air. The resi-
due was then ignited strongly in the air and the resulting oxidized
palladium reduced in hydrogen. Finally, the metal was heated in
carbon dioxide for 2 minutes and allowed to cool in an atmosphere of
this gas. The quantities of palladium taken, in the form of foil, were
0.0994 g, 0.0992 g, and 0.1009 g. The metal residues weighed 0.0995
g, 0.0993 g, and 0.1010, respectively, no correction being made for the
filter paper ash.
Experiments were also made in which the glyoxime compound
itself was weighed. One of the precipitates was caught on an asbestos
pad in a Gooch crucible and the other two in porcelain filtering cruci-
bles. The precipitates were washed thoroughly with hot water and
dried for 1 hour at 110 C. The quantities of palladium taken, in the
form of foil, were 0.1061 g, 0.0954 g, and 0.0973 g. The glyoxime pre-
cipitates weighed 0.3348 g, 0.3014 g, and 0.3072 g. Using the theo-
retical factor for the percentage of palladium in the compound, namely,
0.3167, the calculated quantities of palladium recovered were 0.1060 g,
0.0954 g, and 0.0973 g, respectively.
In the preceding experiments, the palladium was precipitated from
solutions of its chloride in diluted hydrochloric acid. Each solution
had a volume of 350 ml and contained 10 ml of hydrochloric acid.
No palladium was detected in the filtrates nor in the wash waters.
In another experiment, the glyoxime precipitate was washed with
200 ml of hot water in order to determine whether it was appreciably
soluble. No palladium was detected in the evaporated washings.
Other experiments showed that palladium was quantitatively pre-
cipitated by dimethylglyoxime not only from solutions of its chloride
in diluted hydrochloric acid but also from solutions of its sulphate in
diluted sulphuric acid. When the yellow precipitate was decomposed
by nitric acid or by aqua regia, as is often done when palladium is to
be reprecipitated, it was found that the metal was not always com-
pletely recovered. A
small but significant amount of it remained in
solution, which could be precipitated only after the filtrate had been
heated until vapors of sulphuric acid were evolved.

VII. DETERMINATION OF RHODIUM AND OF IRIDIUM

Rhodium may be conveniently determined by igniting the hydrated
dioxide to an anhydrous oxide in air and reducing the latter compound
to metal in hydrogen. In order to prevent deflagration, the filter and
precipitate are impregnated with ammonium chloride and carefully
charred before they are strongly ignited in the air.
Rhodium may also be conveniently determined by igniting the
sulphide to an anhydrous oxide and reducing the resulting oxide to
metal in hydrogen. This method is recommended for the recovery
and determination of rhodium after its separation from iridium by
titanous chloride. 21 Rhodium is quantitatively precipitated by hy-
drogen sulphide from a gently boiling solution of its chloride in
diluted hydrochloric acid, but it is only partially precipitated from a
21
See footnote 9, p. 293.

36763—34 3
 —
300 Bureau of Standards Journal of Research [Vol. 12

solution of its sulphate in diluted sulphuric acid. solution of the A

rhodium can be prepared, however, by adding diluted hydrochloric
acid (1 + 2), equal to three times the volume of the concentrated
sulphuric-acid solution, and boiling it. The yellow color of the sul-
phate solution then changes to rose-red, no doubt indicating a con-
version of the sulphate to the chloride. No difficulty attends the
precipitation of rhodium from the solution so treated.
Iridium is best determined by igniting its hydrated dioxide to an
anhydrous oxide and then reducing the oxide in hydrogen. The same
precaution must be used to prevent deflagration as was described for
rhodium. Unfortunately, the precipitation of iridium from acid solu-
tions by hydrogen sulphide is often incomplete and as yet the con-
ditions suitable for quantitative precipitation have not been estab-
lished.

VIII. EXPERIMENTAL RESULTS ON THE SEPARATION OF

PLATINUM FROM PALLADIUM, RHODIUM, AND
IRIDIUM
The method which is outlined in section III was used to effect the
separation of platinum from palladium; the separation of platinum
from rhodium; the separation of platinum from iridium; and the
separation of platinum from these three metals when present
together.
Owing to the difficulty encountered in the direct precipitation of
the platinum from the hydrolytic filtrates by formic acid, the platinum
was first precipitated by hydrogen sulphide. The ignited sulphide
was redissolved and the platinum then precipitated by formic acid
in a solution buffered with sodium acetate, as described in section V.

1. SEPARATION OF PLATINUM FROM PALLADIUM

The results of the determination of platinum and of palladium,
starting with solutions containing both metals, are given in table 1.
The hydrated dioxide of palladium from the second hydrolytic pre-
cipitation was redissolved in hydrochloric acid. The glyoxime pre-
cipitate was ignited in air, then in hydrogen, and finally in carbon
dioxide. The palladium was determined as metal. No palladium
was detected with dimethylglyoxime in the dissolved platinum residues j

nor was any platinum found, by means of ammonium chloride, in the

filtrates from the precipitation of the palladium with dimethyl-
glyoxime.

Table 1. Results of the analysis of solutions containing -platinum and palladium

Palla- Palla-
Platinum Platinum Error dium dium re- Error
Analysis no.
taken recovered
taken covered

9 9 9 9 9 9
25-701 ... 0.2848 0.2849 +0. 0001 0. 1044 0. 1043 -0. 0001
26-702 .2793 .2794 +. 0001 .1243 .1243 .0000

If palladium alone is to be separated from platinum, a single

hydrolytic precipitation, followed by a single precipitation with
dimethylglyoxime, is undoubtedly sufficient to effect complete
 — '

Gilchrist] Separation of Pt, Pd, Rh, and Ir 301

separation. Complete separation of the two metals can also be

accomplished, no doubt, by a double precipitation with dimethyl-
glyoxime alone, provided that certain precautions are observed.
Wunder and Thiiringer, 22 operating in a hot solution, encountered
difficulty in separating palladium from platinum with dimethylgly-
oxime, but Davis 23 later showed that if palladium is precipitated
in a cold solution two precipitations as the glyoxime compound
remove the platinum. Swanger u subsequently followed the pro-
cedure of Davis with excellent results. It should be borne in mind,
however, that, if the palladium is reprecipitated from a solution of
the first precipitate in nitric acid or in aqua regia, a small amount of
palladium may remain in solution.

2. SEPARATION OF PLATINUM FROM RHODIUM

The results of the determination of platinum and of rhodium,
starting with solutions containing both metals, are given in table 2.
The rhodium was recovered by precipitation with hydrogen sulphide
and determined as metal.

Table 2. Results of the analysis of solutions containing platinum and rhodium

Analysis no.
Platinum Platinum
Error
(NEU) 3 RhCl 8 Rhodium Rhodium Error
taken recovered taken present recovered

9 9 9 9 9 9 9
9-646 . 0.2450 2 0.2454 +0.0004 0.1537 0.0469 0.0468 -0. 0001
10-647 .2350 i.2350 .0000 .1682 .0513 .0512 -. 0001
18-652 .2469 .2469 .0000 .0953 .0291 .0291 .0000
19-653 .2479 .2479 .0000 .0800 .0244 .0243 — .0001
20-654 .2617 .2618 +.0001 .0968 .0295 .0296 +. 0001

i Calculated on the basis of an experimentally determined rhodium content of 30.50 percent.

2 Pt determined by ignition of sulphide.
Note.—In analyses nos. 9-646 and 10-647, filter paper was used to catch the hydrated dioxide. In the
other 3 experiments, the porcelain filtering crucible was used. In none of the 5 rhodium solutions, previous
to the recovery of rhodium, was any platinum detected. The test for platinum was made as follows:
The second hydrolytic precipitate of rhodium dioxide was dissolved in hydrochloric acid and the resulting
solution evaporated to a moist residue on the steam bath. 5 ml of hydrochloric acid was added and the
solution diluted somewhat and filtered. The filter was washed with diluted hydrochloric acid (1+99).
The rhodium solution was then evaporated to a small volume and to it was added 10 drops of a saturated
solution of ammonium chloride and 2 or 3 drops of nitric acid. The solution was next carefully evaporated
to dryness on the steam bath. The dry salt mixture was dissolved in the smallest quantity of water pos-
sible and allowed to stand for one half hour. In the solutions tested, no precipitate of ammonium chloro-
platinate appeared, whereas, in solutions of known platinum content 0.1 mg of platinum was readily
detected and a precipitate was formed with 0.05 mg of platinum. A
very faint trace of rhodium was de-
tected in the final platinum residue of analysis no. 10-647. No rhodium was detected in the other 4 plati-
num residues. The test for rhodium was made as follows: The platinum residues were dissolved in aqua
regia, the solutions were evaporated to a sirup with hydrochloric acid, and the bulk of the platinum was
precipitated with ammonium chloride. The mother liquors and the wash waters containing ammonium
chloride were evaporated to dryness. An extremely small amount of rhodium, if present, produces under
these conditions a greenish discoloration of the small amount of ammonium chloroplatinate which has
been carried along with the wash waters.

In determining rhodium, when it alone has been separated from

platinum, it may be more convenient to catch the precipitate from the
second hydrolysis on filter paper. The hydrated dioxide of rhodium
can be ignited to metal in the same manner as iridium dioxide. If
palladium or iridium is present with the rhodium, it is recommended
that the rhodium be recovered, after its separation from these metals,
by precipitation with hydrogen sulphide.
22 See footnote
18, p. 298.
23 C. W. Davis, Separation of Palladium
from Platinum by Means of Dimethylglyoxime, Bur. Mines
Reports of Investigations, serial no. 2351, May 1922.
24 See footnote
19, p. 298.
 —
302 Bureau of Standards Journal of Research [Vol. it

3. SEPARATION OF PLATINUM FROM IRIDIUM

The results of the determination of platinum and of iridium,
starting with solutions containing both metals are given in table 3. ;

The iridium was recovered by hydrolytic precipitation and deter-

mined as metal.
Table 3. Results of the analysis of solutions containing platinum and iridium

Analysis no.
Platinum Platinum (NH4) 2 IrCl 6 Iridium Iridium
Error Error
taken recovered taken present J
recovered

9 9 9 9 9 9 9
21-611 0.2359 0. 2358 -0. 0001 0.4500 0.1944 0.1945 +0. 0001
22-612 .2692 .2694 +. 0002 .4721 .2039 .2040 +. 0001
23-613 .2501 .2501 .0000 .5684 .2455 .2456 +. 0001

1
Calculated on the basis of an experimentally determined iridium content of 43.20 percent.
Note.— The nitrates from the precipitation of platinum by hydrogen sulphide and by formic acid were
heated until vapors of sulphuric acid were evolved and tested for iridium by adding nitric acid. In this
test, with quantities of iridium less than 0.1 mg, the color developed is green, with larger amounts it is
blue. Owing to the incomplete precipitation of iridium by hydrogen sulphide and by formic acid, iridium,
if present with the platinum, will appear in the nitrates from these precipitations, at least partially. The
nitrates from analysis no. 21-611 showed the presence of a quantity of iridium much less than 0.1 mg. No
iridium was detected in the filtrates from the other 2 experiments.

4. SEPARATION OF PLATINUM FROM THE THREE METALS,

PALLADIUM, RHODIUM, AND IRIDIUM
Critical experiments were made in which palladium, rhodium, and
iridium were simultaneously precipitated as hydrated dioxides in the
presence of platinum. The resulting platinum filtrates were carefully
examined for palladium, rhodium, and iridium and the absence of
these metals was established. The third hydrolytic filtrate gave no
evidence of the presence of platinum, when treated with ammonium
chloride.
That no platinum contaminated the second hydrolytic precipitate
of the mixed dioxides of rhodium and iridium was shown by the
following experiment: From a solution containing 0.2642 g of plati-
num, 0.1385 g of iridium, and 0.0465 g of rhodium, 0.0464 g of rhodium
was recovered after its separation from iridium by treatment with
titanous chloride, as described in a previous publication. 25 The ash
of the paper used to filter the solution of rhodium sulphate, produced
by dissolving the metallic rhodium from the second precipitation with
titanous chloride, was found to contain no platinum. If platinum
had contaminated the hydrolytic precipitate, it would have appeared,
at least in part, at this point in the analysis.

IX. SEPARATION OF PALLADIUM FROM RHODIUM AND

IRIDIUM
In one of their papers Wunder and Thuringer 26 reported on the
separation of palladium from rhodium and from iridium, by means
of dimethylglyoxime.
Preliminary experiments in this laboratory showed that palladium,
when precipitated by dimethylglyoxime from a solution of the mixed
precipitate of hydrated dioxides in diluted hydrochloric acid, was not
contaminated by either rhodium or iridium. Apparently, a single
precipitation of the glyoxime compound sufficed to effect a clean
« See footnote 9, p. 293.
28 M. Wunder and V. Thuringer, Z. anal. Chem., vol. 52, p. ), 1913.
Qiichnd] Separation of Pt, Pd, Rh, and Ir 303

separation. To confirm this observation, two solutions were prepared

containing in the one, 0.2300 g of iridium, 0.0177 g of rhodium and
0.0342 g of palladium, and in the other, 0.2154 g of iridium, 0.0112 g
of rhodium and 0.0333 g of palladium.
The palladium precipitates were caught on filter paper, washed
with diluted hydrochloric acid (1+99), and finally with hot water.
The filters and precipitates were placed in 500-ml erlenmeyer flasks
and decomposed by treatment with 5 ml of sulphuric acid to which
portions of nitric acid were added. After destruction of organic
matter, the solutions were heated until heavy vapors of sulphuric
acid were evolved. The solutions were then diluted somewhat and
filtered. The filters were washed with diluted hydrochloric acid (1 + 99)
and with water. The palladium compound was precipitated a second
time from a volume of 350 ml and collected in porcelain filtering cru-
cibles. It was thoroughly washed with hot water, dried at 110 C and
weighed. The weights of the dried precipitates were 0.1080 g and
0.1051 g, respectively. The calculated quantities of palladium pres-
ent, using the theoretical factor 0.3167, were 0.0342 g and 0.0333 g.
The filtrates from the second precipitation of the palladium were
carefully examined for iridium, by heating them until vapors of sul-
phuric acid were evolved and adding nitric acid. No iridium was
detected in them nor did any rhodium appear to be present, since
the solutions remained colorless.

X. SEPARATION OF RHODIUM FROM IRIDIUM

The two metals of the platinum group which remain after the
isolation of osmium, of ruthenium, of platinum, and of palladium
are rhodium and iridium. These two metals can be separated from
each other and determined by the procedure which is described in
detail in a previous paper. 27
Briefly, the method of separation consists in making two precipi-
tations of the rhodium with titanous chloride from a hot solution of
the metal sulphates in diluted sulphuric acid (5 + 95). The iridium is
recovered from the resulting filtrates, after the removal of titanium
with cupferron, by hydrolytic precipitation and determined as metal.
The rhodium is recovered by precipitation with hydrogen sulphide,
in a solution of the metal sulphate which has been boiled with hydro-
chloric acid, and is determined as metal.
The determination of rhodium and of iridium can be greatly sim-
plified and the cupferron precipitations avoided if the solution can
be divided conveniently into aliquot parts. Since the hydrated di-
oxides of rhodium and iridium are quantitatively precipitated under
the same conditions, the two can be precipitated together and finally
recovered as a metallic mixture. In addition, rhodium must be
determined separately in a different portion of the solution.
In a forthcoming paper the various procedures of separation and
determination which have been developed in this laboratory will be
combined into a general scheme of analysis for the platinum group
as a whole.
The writer wishes to express his appreciation to Edward Wichers
for many helpful suggestions.

Washington, December 14, 1933.

27 See footnote 9, p. 293.