‘lktan:a, 1967. Vol. 14. pp. 1133 to 1149. Per@mm Press Ltd.

Printed in Nonhem Irclttnd

TALANTA REVIEW*

A CRITICAL REVIEW OF METHODS OF ISOLATlNG

AND SEPARATING THE NOBLE METALS-III?
CHROMATOG~PHY, F~~CIPITATrON AND FIRE ASSAY

F.E. BEAMISH
Department of Chemistry, University of Toronto, Toronto 5, Ontario, Canada

(Received 23 February 1967. Accepted 2 May 1967)

%mnnary-This review deals with the chromatographic, precipitation

and fire-assay methods of separating the seven noble metals, published
subsequent to the previous reviews and prior to January 1966. The
noble metal species present in solution are discussed in so far as they
relate to separation and determination.

THIS paper continues the review of recent methods of separating and isolating the
noble metals, and the prologue to Part IIt applies equally to Part III.
PAPER CHROMATOGRAPHY
The platinum metal complexes with tin@) chloride and bromide were used by Lederer and
Shukld for a number of separations by paper chromatography: 1 ml of aqueous solution containing
1 mg of each of the six metals was mixed with 1 ml of 1M tin(H) chloride in hydrochloric acid.
Ascending development was used with the metal solution applied as a thin band. The solvent was 1M
tin(R) chloride in 6M hydrochloric acid. Investigation was made of the effects on the R, values when
the hydrochloric acid and tin(n) chloride concentrations in the solvent were varied. The separation
of platinum and palladium was better than that achieved by partition chromatography of the chloro-
complexes. The spots were well deflned, showing for platinum an orange colour and an & value
@21, and for palladium a blue colour with & O-70. For rhodium the spot was elongated with a red
front and a yellow rear and an Rt spread from 0.13 to O-56. Heating the solutions of the metal com-
plexes had little effect on the chromatographic behaviour. Iridium formed two yellow-brown spots
with R, values of @I8 and O-78. Mixtures of Athena-p~adi~, osmium, ~the~i~, platinum-
pahadium and platinum rh~iu~p~dium as the tin@) bromide complexes in 3M hy~ob~~c acid
were successfully separated with the solvent 05M tin(H) bromide in 3M hydrobromic acid. In the
case of potassium chloroiridate(VI) dissolved in water or boiled with hydrobromic acid there were
three spots, one red and two yellow with R, values 0.19, &32,0*87. If iridium(W) was first heated
with tin(n) bromide-hydrobromic acid mixture, four spots developed, one red and three yellow with
R, values 0*23,0*37, @65 and 0.88. No explanation was offered for this phenomenon. The authors
suggested that the complexes were similar in structure to heteropolyacids with the metal as the central
atom surrounded by four or more tin chloro-complex groups. It was noted that the red tin(n)
chloride-rhodium(III) complex could be converted into the yellow form by treatment with hydro-
chloric acid, by extraction with butanol, or by removing tin(H) during eleetromigration.
For the detection of gold, platinum, selenium and tellurium Larsen’ evaporated the chloride
solution with nitric acid to eliminate excess of hydrochloric acid, thus ensuring reduction of selenium
and telhuium to the quadrivalent state. Ascending development was used with a solvent wntaining
n-butanol, methanol and diethyl ether. The R, values were: gold, 0%1.0: platinum, 09: selenium,
OS: tellurium, O=O.Platinum was detected with potassium iodide and gold with tin(H) chloride.
* For reprints of this review see adve~i~ment at the front of this issue.
t Part II-Talanta, 1967,14,991.
1133
1134 F. E. BEAMI~H

A chromatographic separation of gold, palladium, platinum and ruthenium was described by

Vassiliades and Mar~ou~saki~.~ The solvent was isobutanol containing hydrochloric acid to prevent
hydrolysis of the metal salts, and acetic acid to encourage clear separation of the zones. A solution of
dianisidine hydrochloride was used for locating the zones, the colours and R, values being; ruthenium,
brown-yellow, 0.22: palladium, yellow, 0.51: platinum, dark-yellow, 0.62: gold, brown-red, 0.94.
The metal chloride solution contained 150 mg of platinum, 50 mg of palladium, 150 mg of gold and
20 mg of ruthenium per 100 ml. The chrornatogram was run for 21 hr at 26”, dried for 2 hr, and then
sprayed with the reagent. Continued exposure to air or to hydrogen peroxide modified the colours.
For determination of R, values the solvent front was located by use of ultraviolet light.
A useful series of chromatographic separations for the seven noble metals was described by
Blasius and Fischer.4 Ascending and radial development were both applied and three solvents were
used successively, each with some specific advantage. Thiourea was used to locate osmium and ruthe-
nium, and tin(II) chloride and potassium iodide were used for other metals. Iridium was located
with concentrated hydrochloric acid, producing a brown zone at the solvent front. Ammonium
sulphide was also used for ruthenium. The butanol-hydrochloric acid solvent was recommended for
the chloro-complexes of gold, platinum and palladium with either ruthenium or rhodium (the metals
are in order of increasing R,). Sodium nitrite solution was especially useful for the separation of the
chlorocomplexes of iridium and platinum with or without osmium: the R, values increased in the
order given. Sodium nitrite in ethanol and isopropanol was recommended for separation of
the nitrito-complexes of ruthenium and rhodium with either palladium or platinum in nitrite solution.
The solvent ethyl methyl ketone-hydrochloric acid was used for the chloro-complexes of rhodium and
palladium and the nitroso-chloro-complexes of ruthenium in hydrochloric acid. A variety of com-
binations of noble metals was separated successfully, and it appeared that osmium and ruthenium
could he separated with the butanol-hydrochloric acid solvent : the use of thiourea as locating agent
is of special interest. The authors also described a variety of paper ionophoresis separations of the
seven noble metals in various combinations. The electrolytes were hydrochloric acid solutions of
sodium chloride or acetate; the noble metals were present as the chloro-complexes: potentials of
230 or 130 V were applied for 1 or 2 hr. The variation in bands resulting from sample applications
to either the side or centre of the strip and the effect of hydrolysis of platinic acid solution were dis-
cussed. Again, there is particular interest in the clean separation of osmium and ruthenium in sodium
acetate-hydrochloric acid medium, a potential of 230 V being applied for 1 hr.
For the determination of platinum and palladium James6 applied the Kemher and Wells techniques
to silver assay beads obtained by the fire-assay of Merensky Reef ores. The bead was parted with
nitric acid in a specially designed glass bulb with a closed tail calibrated to contain 1 ml, which allowed
control of the evaporation and isolation from the silver chloride formed by treating the parting acid
with hydrochloric acid and sodium chlorate solution, the latter ensuring the oxidation of any bivalent
platinum: @OSml of the resulting solution was used for the chromatogram. The spotted area was
allowed to dry and then inserted into the jar containing the pentanol-hydrochloric acid-hexone
solvent and the chromatogram was allowed to develop overnight. The dried strips were sprayed with
tin(D) chloride solution, platinum appearing as a yellow band and palladium as a dark purple one.
The corresponding areas from a duplicate strip were removed and extracted with 2%%4hydrochloric
acid for platinum and 09t4 hydrochloric acid for palladium. The platinum portion of the stained
strip could also be used for the determination. In the filtered extracts platinum was determined with
tin(I1) chloride and palladium with thioglycollic acid. The results indicate very satisfactory precision.
The author stated that “Results obtained with only slight familiarity with the technique compare very
favourably with those obtained by laboratories practised in and using standard tire assay methods.”
“Capillary chromatography” or, more simply, “spot-testing” has been applied to the separation
and detection of rhodium and iridium. &rn$ and Kubin’ added an acidified solution of the two metals
by capillary to a filter paper impregnated with aluminium hydroxide. Several capillary-loads of water
were added to the spot, which was then dried, and sprayed with a solution of tin(U) chloride and
potassium iodide. Rhodium appeared as a red ring and iridium as a light brown inner ring. The
intensity could be increased by the addition of potassium iodide solution. There was interference
from platinum, palladium and gold.
Filter papers impregnated with cadmium sulphide have been uyd for separation and approximate
determination of platinum, palladium, gold, silver and mercury. Celap ef al.* claimed that the move-
ment of the products of the metathesis was a linear function of the amount of the ion present. Presum-
ably the precision obtained was 2-5%, depending on the amount of ion present. Some additional
evidence for the usefulness of methods such as these might be forthcoming if the technique could be
applied successfully to noble metal alloys.
When investigating the paper chromatographic separation of the platinum metals Noddack and
SanirO were unable to separate each of the six metals on one chromatogram. The amounts of metals
 Methods of isolating and separating the noble metals-Iii 1135

applied were 20-30 ,ug and the ascending technique was used. With the solvent ethyi acetate-
concentrated hydrochloric acid-acetic acid, the Rf values for ruthenium, rhodium, iridium, palladium,
osmium and platinum wererespectively O-10, 0~18,0=19,0-4&O-53 and 058. For rhodium and iridium
Rf vhes of 0.49 and O-88 respective1 were obtained with the solvent methyl ~t~t~on~nt~ted
hydrochloric acid-ethanol. For spot gocation tin(H) chloride was used for rhodium, palladium and
platinum: bromine water for iridium: thiourea for ruthenium: tin(I1) chloride with thiourea for
osmium.
A potentially useful report on ascending cbromatographic separations of the seven noble metals
and copper and silver was recorded by Voicu and Nascutiu,l” who reported the R, values obtained
with eighteen solvents consisting of 1: I mixtures of various alcohols or ketones with various concen-
trations of hy~c~oric acid. No overall pattern of change of R, value with increasing acidity
appeared. With certain solvents the R, values decreased with increase in acidity: in several cases the
R, value increased to a maximum with about 4 or 6M hydr~hloric acid then decreased at higher
acidity: in others the R, valnes increased with acidity. A highly acidic keton~hydr~~o~c acid
solvent separated a large and more varied number of metals. The C&C, alcohols showed maxima in
the R, values. The C,-C, alcohols provided relatively varied separations. Detailed examination of the
tables suggests a variety of chromatographic separations of noble metals which are frequently associated
as groups in analytical separations or as predominant groups in natural deposits. In general, the
solvents allowed good separations of gold and plat~um or gold and ~ladium, On the other hand,
with solutions of platinum and palladium in 1M acid the R, values indicated considerable separational
di@icuhy for all 18 solvents.
For the chromatographic separation of noble metal groups containing rhodium or platinum
with palladium, ruthenium and gold, Kao and Yuan I1 used as solvent oxine dissolved in isobutanol-
ethanol-water-acetic acid (60: 57: 80: 3). The metal solutions were buffered at pH l-2 with dichloro-
acetic acid.
T~~y~ate in hyd~c~oric acid solution was used as locating agent and ethyl methyl ketone
as solvent to separate the platinum metals into the groups: iridium, ~h~u~~th~~, platinum-
palladium and osmium. In a second method Kao and Sun’= used benzene saturated with thiocyanate
as the solvent and hydrochloric acid, potassium chloride and thiocyanate as locating agent. Gold,
palladium and slating-ruthe~um-rh~i~-iridium could be separated. The authors are continu-
mngtheir work with a view to separating iridium from rhodium. _
Filter paper impregnated with butyl phosphate was used by Hu and Shiu+ to separate the groups
~l~pla~um~~adi~-rh~i~ and ~~~p~adiurn-rhodium by revemphase paper
~hromato~phy. Hydrochloric acid was used as solvent, and increasing its concentration had little
effect on the R, values. which for cold, ulatinum, ualladium, and rhodium were 0,0=2. @4-O+ and
0.9 respectively. With.2-fOh4 hy&chl&c acid; gold remained at the origin. With >8M hydro-
chloric acid, silver migrated as AgCl,-. With >2M hydrochloric acid selenium and tellurium could
be separated, In a second paper Hu and Sbii4 added variations to the method. Mixtures of gold or
silver with platinum-palladium-rhodium could be separated with hydrochloric acid that was <6&f.
The R, value of palladium increased with increased acid concentration. With nitric acid as solvent the
R, valuea of rhodium, selenium, tellurium, silver and gold changed very little with increased
~on~ntration of acid while there was a slight increase in the case of ruthenj~, p~adi~, osmium
and platinum. ~~o~unately the ~~o~to~aphi~ behaviour of iridium was not indicated. The
persistently high Rf value of about @93 for rhodium indicates a clear separation from the remaining
noble metals.
Kertes and Beck’” used ~onium thiocyanate and nitric acid solutions of some sixty ions in-
cluding gold and platinum, and extracted with isoamyl alcohol, methyl isobutyl ether and d&isopropyl
ether. The RI values were recorded for all sixty ion species. The use of thiooyanic acid in paper
partition c~o~to~aphy was also discussed by Locke and Martin,‘6 who determined the chromato-
graphic properties of eight solvents. Of the noble metals, only gold was included among the 38 ions
used, The solvents were n-butanol-water-percbloric acid and seven complexing solvents based on an
n-butanol-water-thiocyanic acid mixture to which was added pyridine, aniline or qninoline. The
authors discussed criteria of complexing properties.
For the analysis of platinum metals, Kao and Chang” used two solvents, n-amyl methyl ketone-
hydrochloric acid and acetone-hydrochloric acid-acetophenone. Both solvents provided speed and
stability. The locating reagents were tin(I1) chloride for gold, plat~um-~lla~~, rhodium and
ruthenium: thiourea for osmium: chlorine water for iridium. With the fhst solvent the chromato-
gram was run for f hr at room tem~mture, for the other 3 hr at 95-10” was requhed. With both
solvents silver remained at the start and the osmium spots were scattered. The acetone solvent could
be used to separate rhodium from iridium. Large amounts of sulphate, nitrate, chloride, phosphate
and percblorate did not interfere. Perchlorate concentrated the platinum spots with the acetone sol-
vent. The method was applied to determine gold in electrolytic slimes following the isolation of gold
by hydrogen peroxide.
1136 F. E. BEAMISH

Variations in chromatographic methods are being proposed, some of which offer useful adaptations
to analytical areas of wide interest. Photographic emulsions have been used in many methods to
detect the presence of gold in solid materials. The autora~o~p~c plate was used by Shima18
to trace the losses of radioactive gold to pot wall, cupel etc. during a fire assay. The technique
involved cutting the used pot or cupel into two equal parts. The extent of the penetration of gold,
and thus of the losses, was determined by pressing the thin vinyl membrane against the freshly cut
areas for 3 days and then developing and fixing as usual. The results obtained were discussed in a
recent review.lO
To detect in situ the presence of gold minerals, Rao and Sag+ used gelatine-coated paper
strips impregnated with aqua regia and pressed the paper against a surface of polished mineral sections
for 20-30 sec. The paper was removed and treated with pyridine in hy~obro~c acid. In the presence
of gold, cherry-red spots appeared. Presumably, the associated constituents in minerals do not
interfere. Photographic paper washed free from silver by thiosulphate and water may be used as the
paper base.
A similar method for the detection of gold in ores was reported by Lever” some years earlier.
The ore and paper were pressed between two glass plates and the gold was extracted by nqua regia
onto the paper, treated with ethyl acetate-nitrate acid mixture and finally identified with tin(I1)
chloride. Applied to ores containing O~OsLo~lStroy oz of gold per ton the method gave uncertain
results with some interference from iron. The author reported a more satisfactory set-qu~ti~-
tive chromatographic method for estimating gold in thi range from 25 to 250 (;g per assay ton
of ore. Six samples can be examined in 2 hr and the ranee of annlication can be extended bv
simple modifications. For siliceous ores a lOO-mesh sample weyghing i-5 g was treated with aqua regi&
and the mixture was heated and filtered. Orthophosphoric acid was added to complex iron, and
gold was extracted with ethyl acetate. The aqueous phase was removed and potassium chlorate
added to the organic phase to remove residual water. The organic phase was transferred to chromatog-
raphic strips (pretreated with potassium chlorate solution) over an area previously marked between
1 and 5 in. from one end. The strip was dried and the end nearest the l-in. mark was immersed in the
developing solution (~t~nol-hydrobro~cacid-~ot~siumc~orate). Whenthesolvent had travelled
to the 5-in. mark the strip was removed and air-drted, and the zone at the 5-m mark was sprayed with
tin(n) chloride solution. The intensities of the stains at the S-in. mark were then compared with
standards similarly prepared from a gold solution. The stain could be treated with ammonia fumes
to reduce fading. A method for treating sulphide ores was included, in which case potassium chlorate
was added before aqua regia extraction. The results compared favourably with those of fire assay.
Losses of gold in the aqua regiu and extraction treatments were negligible. The report bears the marks
of a carefully considered and precisely executed research project.
Paper chromatography was also used by Agrinier** to estimate gold in silicate rocks and minerals.
Aqua regiu extracts, made after roasting the minerals, were evaporated and the residue was taken up in
nitric acid. The solvent was ethanol-ethyl acetate-water-nitric acid, and p-dimethylaminobenzyl-
idenerhodanine was used as location reagent. R, values were given for associated base and noble
metals: gold appeared at the solvent front. The results for gold compared favourably with those
obtained by fire assay.
Thin-layer c~o~to~aphy was applied by ~m~avljevi~-J~ovljevi~ er a1.88to separate and then
identify gold, with palladium and platinum present as chlorides and silver as acetate. Dithizone
extracts in benzene were spotted on silica gel plates and chromatographed with benzene-methylene
chloride as solvent. The R, values and colours were; AgHDz, 0, reddish-brown: Au(HDz),, 0.31,
red: At@& 0.20, lemon-yellow: Pd(HDz)*, @43, dark green: Pt(HD&, 0.38, yellow. Silver and
gold dithizonates were selectively extracted at pH 75, palladium at pH 5.5, and platinum at pH 3
after reduction with tin(I1) chloride.
Wu and Wup6 used layers of silica gel mixed with sintered anhydrous calcium sulphate to give
faster separation with better-defined spots. With a~ton~hydr~~oric acid-ethyl acetate or nitric
acid-ethyl acetate as solvent separations into four groups were made; gold-osmium, ruthenium-
palladium, platinum-rhodium, iridium. With increased proportions of concentrated nitric acid
gold-osmium, platinum, ruthenium, palladium, iridium-rhodium could be separated. A potentially
useful technique involved fixing the palladium spot at a suitable R, value by the addition of oxine
solution and then using a solvent such as n-butanol saturated with dilute sulphuric acid to separate
gold, platinum, and palladium.
Paper el~ophoresis was used by Be=en and Hoeye *(Lto determine gold(III) associated with
colloidally dispersed gold. Paper strips, wetted with 0*075M sodium thiosulphate, were used to
provide a clear separation of the ions from the colloidal gold: 5 ,ul of the samples were streaked in the
middle of the strips, and under a constant current of 2.5 mA/cm, the gold(III) ions migrated about
65 mm towards the anode after 1 hr: the gold particles moved 2-5 mm towardtheanode. The potential
dropped from 400 to 200 V. Gold-198 was used for the determination. For l-2% of gold0 the
 Methods of isolating and separating the noble metals-III 1137

method showed a deviation of 30 y0 or less from values obtained by paper chromatography : precision
was about 10%.
Dithiiness has been used for the isolation of platinum and of radium by c~omato~~hy
on an alumina column after nreliminarv extraction of platinum dithizonate from tin(R) &lo&e-
hydrochloric acid solution. &dorofor& was used as ement and 024 ppm of platinum could be
determined. For palladiums’ dithizone separations were made from 1-6N hydrochloric or sulphuric
acid. Chloroform was again used as eluent. The method could be applied to less than 1 ,ug of pallad-
ium. About O-01% of palladium could be determined in 1-mg samples of platinum in 30 min. There
was no interference from copper, mercury and bismuth.
Ash&was8 also separated several new metal dithizonates in a pure state by chromatography
with chloroform as the eluent. Absorption maxima were recorded for the dithizonates of cobalt,
iron, nickel, palladium, gold, copper and platinum.

Ruthenium and osmium

One of the most useful solutions for the collection of the tetroxides of osmium and ruthenium is cold
hydrogen peroxide. Complexing or reducing reagents are sometimes added and determination can
be made without the contaminants which are present in such media as sodium hydroxide or sulphite
solutions. Koch and B~che~~er%* used paper electrophoresis to characterize the ~m~sition of
the peroxide solution of the distillates. The tetroxide was distilled from phosphoric acid solution
intokooled water. Addition of per&lo& acid and hydrogen peroxide gives quantitative formation of
rutheniumc1v). With sulnhuric acid and hydrogen peroxide ruthenium (III) is formed. To denolv-
merize the &henium(IV)ihe solution was addeduto &r exchange-resin, Wofatit KPS in the hyd;og&
form, which was then washed with acidified water and irrigated with perchloric acid.
The effect of aging a solution of [Ru(H,O)C~,]~- in 0*1&fhydrochloric acid was studied by Shukla,sO
by paper electrophoresis and spectraphotometry. It was shown that aging the =compiex resulted in
gradual hydration involving progressive replacement of the chloro- group by an aqu~~oup. The
absorption spectra of electrophoretically pure [Ru~~O~l~]*- and [Ru(&O)#+ were also reported.
Chromatographic and solvent extraction methods have been described for the separation and
determination of the ruthenium constituents in nitric acid solution, specifically in those “aqueous
nrocess solutions from chemical separation plants, designed to separate plutonium and fission
products from irradiated uranium by solvent ixtraction.“- In generac these Constituents are of two
tvnes. nitrato (non-nitro) and nitro comnlexes of nitrosvlruthenium. a1-81 Fletcher et al.*’ suggested
&p&c formulae for eleven nit&o complexes classified according to the presence or abseng of a
nitrato group trans to the NO-group. Paper crouton of ~~e~urn solutions in 3-15M
nitric acid, previously aged to reach equilibrium, indicated a division of the complexes into four groups.
The & values were given for each group with isopropyl methyl ketone (IPMK) and dibutylcellolve
(DBC) as solvent.
Brown and Naylop* dealing with “process solutions” stated that, in contrast to solvent extractive
methods, paper chromatography allowed the characterization and determination of the nitric acid-
ruthenium complexes, which were practically all in the stable nitrosylruthenium form and contained
the nitrato and the nitro groups. The solvents used were IPMK and DBC, equilibrated with 3M
nitric acid. With •kpr~~olu~ions~~ containing mixed fission products, there w& some masking of
ruthenium bv zirconium-95 and niobium-95. which had R, values of about 06 Brown and Navlofl*
determined {he ruthenium distribution on the chromatograms by gamma spectrometry. Wain bt al.-
used the same solvents with OGO5ml or less of ruthenium solution applied as a band at the base of the
paper. The nitrosylruthenium uitro complexes remain at the start while the nitrato complexes, with a
ratio of nitrate to ruthenium greater than 3 to 1 migrate, the RI values being between 0.2 and 0.7.
Points of technique for optimum separation were described and the & values for the various nitro
and nitrato complexes were included. An empirical form&t was given for calculation of the amounts
of nitro and nitrate complexes in nitric acid solutions. Scargill and Lyon- found anomalous results
when the method was applied to uranium solutions. The method indicated “up to 20% of nitro-com-
plexes in solutions known to contain only nitrate-complexes, and also up to 25 % of nitrate-complexes
in solutions where complete conversion to the nitro-complexes had been obtained.” The authors
concluded that the Wain method” would only give accurate results with the use of a carefully delined
technique and in the absence of uranyl nitrate, A revised formula was proposed for general application
to irradiated uranium process solutions to determine the proportions of nitrato- and nitro-complexes.
The ~butyl~ll~lve results were used and showed a higher proportion of nitro-eomplexes in process
solutions than that found by the Wain= method.
ScargilP4 later provided methods for the separation of individual nitrato and dinitrocomplexes of
nitrosylruthenium(III) and a method for the determination of the extractable nitrato and dinitro
complexes in nitric acid solutions. By reverse-phase chromatography at 0”, with nitric acid as the
solvent on paper impregnated with tributyl phosphate, the author was able to obtain 5 groups of
 1138 F. E. &ZAMISH

nitrato-complexes, two major groups of nitro-complexes, and to separate mixtures of nitrato- and
nitro-complexes into the solvent-extractable nitrate group, the dinitro-complex and the cationic
nitro- and ~~to~omplexes. By cation-exchange methods he identifkd the cationic dbitro-coqdex,
r~~~o~~*~~~~~~~~~~and an anionic or neutral complex, ~~~~O~)*~O~~~~U)~~. The existence
of a mononitro-mononitrato complex was also indiated.
The interesting reaction of osn&m(VIII) oxide and potassium iodide, one of the earliest reactions
applied volumetrically, was studied chromatographically by Rallo a”B In hydrochloric acid medium
the constituents OSI,~- and chloroiodo-osmates(IV) were indicated. All of the latter could be con-
verted into the violet OSI,~- by evaporation on a steam-bath with an excess of iodide. Practically
pure OSI,~- could be obtained by performing the reaction in percbloric acid with sodium iodide.

Electrophoresis and ion-exchange chromatography have been used to identify the dissolved
constituents in aqueous acid solutions of rhodium. Shukla and Lederers7 used paper electrophoresis
to determine the conditions for the formation af the yellow and the red rhodium sulphates. In
general, a potential of 25OV was applied for 45 min and tin(H) chloride-potassium iodide was
used for detection. The authors state that the yellow rhodium sulphafe, in which the Rh(II1)
is cationic can be prepared by dissolving freshly precipitated rhod~~{I~) hydroxide in N sulphuric
acid and that in O5N sulplmric acid several, presumably hydroxo-compfex, cations are formed.
Heating or evaporating the N acid solution results in considerable complexation, yielding a mixture
of cationic complexes. Solutions left at room temperature are stable for months. The red rhodium
trisulphate could be obtained by heating to 250” but not by heating on the water-bath In solution it
initially forms a single species but on standing tends to form an equilibrium similar to that for aged
solutions of the yellow form, with some persistence of the anionic form. The author+ also investi-
gated rh~i~(III) perchlorate sofutions, reaching the conclusion, of some analytical signit%cance,
ihat the single ~onsdtuent ~(~~~)~(ClO*)~ couId%e prepared by ~~eeip~tating r~~ium~ydroxid~
from an aoueous solution of Na&hCl, with a slight excess of sodium hvdroxide. dissolvine the
precipitate In perchloric acid and- re-precipitating Until a solution was obtained that had a &gle
absorbance maximum at 396 m,u or gave a single electrophoretic band. Electrophoretic evidence was
obtained of the presence of a second constituent of unknown identity in solutions made by boiling
chlororhodite with perchloric acid. Previous work by the authors ao*puindicated that the removal of
co-ordinated chlorine from rhodium chloro-salts was not complete in one hydroxide precipitation.
Thii wnclusion becomes significant in the analytical separation of rhodium.

Platinum
Paper chromatography has also been applied to separate noble metal complexes in studies of their
structural and behavioural characteristics. Thus, Basolo et aLa1separated cis- and trans-dihalogendi-
ammineplatinum(II) complexes and stated that practically all mixtures of eis- and trans- as well as of
dichloro-, dibromo-, di-iodo- , ehloroiodo- and ~~orobromo-complexes could be thus separated,
Later@ they made a ~hromato~p~i~ study of platinum metals halides on several resin papers_ The
Rt values obtained at various hydrochloric acid concentrations indicated that plat~~um(I~ and
platinum(II) could be readily separated. The RI values for I&PdCl, and H,PtCl, on weak anion-resin
and Amberlite SI32 resin papers were closely similar. A cursory examination of platinum chloride
solutions containing sulphur dioxide indicated the presence of H,PtCl~ and sulphite complexes. Data
were provided to contirm the chromatographic separation of chloroammine, amminepyridine and
chloropyridine platinum(I1) complexes.

Blasius and PreeW used ionophoresis at 3000 V for the separatian of aquo and hydroxychloro
complexes of rhodium(III), iridium(II1) and platinum(IV) in an acetic acid-acetate buffer. Tin(II)
chloride was used as location reagent for the ionophoretograms of rhodium(III) and platinum(N)
and chlorine for iridium(II1). The authors investigated the temperature dependence of the migration
velocity and the effect of aging an the composition of the species. In the case of rhodium, in addition
to the main zone, there were a number of intermediate zones which they interpreted to indicate very
unstable h~droxy~mplexes or higher aggregated ions. The e~oroaqu~~mplexes of rh~i~~III)
and iridium(W) were isolated and their composition determined by reduction to the metal by zinc and
by halogen determination in the filtrate. For rhodium and iridium the cationic zone was characterized
by yellow [MCl,(HeO)J+, the neutral, orange zone by [MCl,(H,O)S] and the anionic zone by
rose [MCl&&,O)J- and red [MCl(HZO)l-“.
 Methods of isolating and separating the noble metals-III 1139

PRECIPITATION
GOId

Procedures for the extraction of noble metals from ores by dissolution in acids ~ve~end~ri~.
A number of these methods have been discussed in a review of neutron activation methodsl* For the
determination of gold in sulphide ores, Assarsson et aLa selectively removed the associated base
metals and silver by treatments in a specially designed apparatus with chlorine-free hydrochloric
acid. The residue was treated with hydrochloric acid and a little nitric acid to remove the gold, which
was finally precipitated by hydrazine sulphate and sodium sulphide in the presence of mercury(B)
chloride. The precipitate was ignited, refined.by cupellation and weighed. The results, compared with
values obtained by fire assay, indicated excellent accuracy.
A method of gold separation from lead and copper concentrates, crude lead, high grade com-
mercial matte and slags was described by Bukharin, b5for concentrations of over 40 g of gold per ton.
A method for silver was included. Initially, the gold and silver were separated by assay melting of the
crude lead which was then treated with nitric acid to produce a gold residue. This was heated at
600”, then dissolved in aqua regia, and the solution was evaporated in the presence of hydrochloric
acid and sodium chloride, then treated with potassium hydrogen fluoride and o-dianisidine, and the
gold determined volumetrically with hydroquinone. The gold losses were O-1-0-4%.
For the separation of gold from copper and lead Ujihirap6 precipitated gold from dilute nitric or
hydrochloric acid withy-dimethyl~no~~ylidenerh~~ne, Z-mer~pto~nzot~~oJe, thionalide
and Methyl Violet. There was interference from iodide, bromide and cyanide.

Palladium
The red-brown precipitate obtained from palladium and Reinecke’s salt was used by Voicu and
Hema to separate palladium from gold and platinum with reasonable accuracy. The data indicated
results for separations from high proportions of rhodium and low values for separations from high
proportions of copper, with some significant lack of precision for separations from high proportions
of copper, iron, cobalt and nickel. The interferences from various other cations were examined.
Recoveries from solutions containing l-460 pg of palladium were very acceptable. The extractant was
ethyl methyl ketone and the medium was 0.1-075M hydrochloric acid. Large excesses of Reinecke’s
salt gave more complete extraction.

As part of an investigation of methods of separating and determining trace amounts of ruthenium,

manganese, t~hneti~, rhenium, moly~enum, Jasim et ai.8LBisolated ruthenium together with
manganese by treating the solution with 10% sodium hydroxide solution and heating in a sealed
capillary. The centrifuged liquid containing manganese and ruthenium was treated with ammonium
sulphide which isolated ruthenium.
The annlication of co-nrecinitated iron(II1) hydroxide as a collecting reagent for a variety of
cations wis” discussed by ~orachevskii et al. ip In t-he case of ruthenium, r%henrum hexachloride-was
completely separated at pH 54-85 whereas in the absence of the iron(II1) hydroxide only 76% of the
ruthenium was precipitated at pH 7-3 and less at higher pH values.

Pa~l~ium, platinum, rhodium and iridium

For the precipitation of palladium and platinum and their collective separation from rhodium and
iridium, Ginzburg et aLso used a suspension of copper(I) chloride in hydrochloric acid that was
<0*15M and contained 05-07 g of sodium chloride per 100 ml. The extent to which rhodium was
co-precipitated depended on the acidity and the chloride concentration, increasing rapidly with a
decrease of sodium chloride concentration below 0,1OM. Selenium was removed by evaporation from
a bromate solution, then tellurium by a cation-exchange (KU-2) separation, The e&rent containing
platinum, palladium, rhodium and iridium was treated with suspended copper(I) chloride. The mixed
precipitate was filtered off and washed with hydrochloric acid at pH 2, then with brine solution to
remove the admixed copper chloride. The platinum and palladium were dissolved in aqua regia and
palladium was determined by dimethylglyoxime and platinum by precipitation with mercury(I)
chloride or snectrophotometrically, The filtrate from the copper(I) chloride together with a solution
of any residuk remahring after dissolution of the platinum andpalladium on the t%er paper was treated
to determine iridium polaro~~~~v and rhodium s~~ophotome~~~v. For relativelv high
rhodium contents, irk&n &d&ern&ed pol~o~aphi~lly : a ~imultaneo~ ~r~ipitationof rhodi<m
and iridium by thiourea was made and rhodium was determined by difference. The data recorded by
the authors, all in the milligram ranges, indicated somewhat indifferent separation efficiency and
. Mitigating against the method are factors such as reagent instability, the sensitivity of
~~~~~-rhodium separations to the sodium chloride content, and the adverse effect of high pro-
portions of rhodium.
f140 F. E. BEAMISH

~d~urn and iridium

For the separation of rhodium from iridium, Tertipis and Beamishsl heated a 1*0&i hydrochloric
acid solution of the metals to 80-85“ in the presence of fresh& reduced copper powder. The nrecinitate
of copper and rhodium was collected and-treated with t&r reg&z, and *%y &sidue was chlorinated
at 700” in the presence of sodium chloride. The solution of canner and rhodium salts was converted
into a chloride solution and the pH adjusted to 1.3-1.5. Copper was selectively removed by the
cation-exchanger Dowex 5OW in the sodium form. Milligram amounts of rhodium were determined
gravimetrically with thiobarbituric acid and microgram amounts spectrophotometrically with tin(H)
chloride. The rhodium-copper filtrate was treated with the cation-exchanger to remove copper, and
iridium was determined ~avime~cally with Z-me~capto~~othiazole or s~trophotome~ca~y with
tin(H) chloride. The method was applied successfuflv to the range of 5 mg to 50 iue of the two metals.
Cross-contamination of the precip&tcs was negligible. - ” .”
For~~mesulph~c acid (thiourea dioxide) was used by Prakwf’eva and Bukanavas’ to pre-
cipitate rhodium in the presence of iridium. The filtrate was treated to determine iridium gravi-
metrically, potentiomet~~y or polaro~aphi~lly. Folio~g d~so~ution of the rho~um complex
a tintID chtoride s~~ro~hotomet~c method was auntied. The method was auulied to amounts
of 5&2&l fig of rhodium knd 50 pg to 23 mg of irid& in ratios from 1: fU to 1 :%O. The accuracy
of recovery was about 14% for rhodium and about iS “f, for iridium: co-precipitation of iridium
with rhodium increased with increasing iridium wncentrations.

Mulricomponent mixtures of the noble metals

A method of separating-and and determining silver, ruthenium, rhodium and palladium, developed
nrimarilv for radiochemical scnaration. was recorded bv Dehmlow and Preiss.“* The nrocedure is
&assical’in approach, with silvir chloride precipitated f&t and purified by dissolution ih ammonia.
The combined filtrates were treated with sodium hydroxide and sodium hypochlorite and the ruthenium
tetroxide was extracted with carbon tetrachloride and snb~q~~y back-extracted into sodium
hydroxide containing sodium hydrogen suIphite, Palladium was removed from the aqueous phase
dimethylglyo~e and then rhodium by reduction with titanium@ii) chloride. The authors wnsidered
the method suitable for sea-~croq~ntitative applications, and 10-12 min sufEced to give reproduc-
ible q~ntitative separations of high chemical yield and high purity. The authors’ conjoint use of the
terms qu~titative and high chemical yield is questionable and the data recorded do not support the
claim for quantitative recovery, though recoveries should be quantitative if the time taken were
extended.
Very few reagents provide a greater separational potential than sulphide salts or acids. On the
other hand the resulting sulphide precipitates am seldom satisfactory weighing forms. Respite this
fact, analytical recipes claiming success with the sulphides of noble metals continue to appear in current
literature. Salarias recorded procedures for the separation and determination of platinum in the
presence of rhenium or zirconium and for gold in the presence of zirconium by a separational precipita-
tion with sodium sulphide followed by adjustment of the hydrochloric acid concentration to 6M for
platinum and to 1M for gold, In each case the noble metal sulphide was used as the weighing form.
An error of 0@-0~15 % was cfaimed for EL-30 mg of platinum and 0@-0~13 % for 20-40 mg of gold.
These methods were discussed in an earlier review.66 It is emphasized again that the accuracy claimed
requires verification by independent investigators.
Thermolysis curves for the sdphides of gold, platinum, rutheni~ and palladium have been
recorded by Taimni and Tandon.~E Oven-dried gold sulphide gave a horizontal up to 220”, being
converted into metal at 280”. In the case of nlatinum. Taimni and Salaria&? had recorded previously
that drying in U(ICIMat room temperature pioduced PtS,5H,O, and that at 130-135” the constant
weight corresponded to PtS,*2H@. The later thermogram by Taimni and Salaria indicated that the
air-dried pentahydrate lost water and sulphur from 50” to 310” at which temperature the metal was
formed. The trihydrate behaved in exactly the same way. In the case of ruthenium the earlier work
indicated the formula Ru&.;ZH,O for the uroduct obtained bv drvina in uacuo to constant weight at
room temperature. The later thermolysis^curve indicated a loss bf weight from 60” to 180” \;ith a
horizontal at 180-240” corresponding to an uncertain composition, then a gain in weight to 340
with a horizontal at 340-410”, again without a stated composition, then a loss of weight until the metal
was obtained at 500”. In the case of palladium the earlier works7 indicated a constant composition of
PdS.2HB0 after drying in vucuo at room temperature. The later thermogram*“ indicated a loss of
weight by the air-dried precipitate from SO” to 420” with subsequent increases and decreases. Prom
800-840’ the horiwntal wrresponded to PdO, the metai being formed at 890”. From the earlier work
of Taimni and Salaria6r the nrecinitate of Ir,Ss.lOHaO could be heated to 85” with the weight remaining
constant. Although these &hors did not&cord thermograms of this precipitate, D&& recorded
his failure to find any level of constant com~sition over the range 44-842”, which suggests that
no iridium compound is formed that is suitable as a weighing form. The thermolysis curves provided
 Methods of isolating and separating the noble metals-111 1141

by Taimni and TandorP do not provide evidence for the acceptability of the weighing forms recom-
mended by Salaria and this failure does not encourage confidence in the sulphide methods described
by Salaria.
A method of isolating platinum metals from partially retined concentrates was dismissed by
Zachariasen and Beamish.6P The technique was designed to provide a method of dissolution of
platinum metal concentrates which could be applied both to large scale operations and to analytical
processes. Methods of isolation were included which would lend themselves to the precipitation
of both large and small amounts of platinum metals. The concentrate used contained some 25 elements
in addition to about 25 ‘A of platinum metals (including traces of osmium). Ths principal base metals
included nickel, tin, copper, lead, chromium and iron, and silica was also present. Standard methods
of dissolution, such as fusion with lead or zinc followed by treatment with acids, fusion with sodium
peroxide, or chlorination with or without a sealed tube, were rejected for various practical reasons.
The process iinally accepted involved an intimate mixing of concentrate and sodium carbonate,
sit&zing at 700” and leaching with water and acid. The dissolution resulted in a residue containng
silica and a very small proportion of platinum metals, ~o~t~g to a total of about 2-3 % of the
original sample. The residue was readily fused with sodium peroxide. This sit&ring method resulted
in very little corrosion of the container and could be made in a silica vessel. The sintered mass was
coherent and could be taken from the vessel as a solid, dry body. In general, subsequent treatments
involved classical separations which would lend themselves to large scale treatments. One distinct
innovation involved a new method of separating rhodium from iridium, developed by Tertipis and
discussed above. The results compared favourably with those obtained by spectrophotometric and
spectrographic methods.
FIRE ASSAY
Although reports of solvent extraction separation of noble metals from aqueous media are numer-
ous, relatively little has been recorded concerning the no less important solvent extraction by metals
or alloys from fused media. Practically no data are available on distribution coefficients, stability
constants of the dissolved constituents etc., which would allow prediction of the effectiveness of extrac-
tion from the associated base metals, even for classical extractants such as lead. The phase diagrams
for these alloys are of very limited assistance because the collecting alloy invariably requires a very
high weight ratio of collector to noble metal. Furthermore, with the present very limited knowledge
of the identity of the noble metal species in aqueous solutions of complex noble metal-base metal alloys,
one must rely to a considerable degree on empirical approaches for the development of improved
methods of separating noble metals from one another and particularly from high proportions of base
metals. Very little progress in investigation of fire assay treatments has been recorded in English-
speaking countries. On the other hand, such assay problems have received incomparably
more attention in Russian laboratories. Recently, Shlenskayaao reviewed some of the Russian
developments in noble metal analysis and stated that “conferences on the analysis of noble metals,
in which workers from research institutes and plant laboratories participate, are regularly held in the
Soviet Union.” The present author has made a considerable effort to secure copies of the original
Russian publications which deal with the separation and determination of the noble metals but with
littie overall success. Wherever possible, however, pertinent material in the various abstracts has
been included.
An excellent general discussion of the problems associated with each stage of the fire assay pro-
cesses was recorded by Britten61 His comment on the future status of 6re assay is of some interest;
“In common with gold assaying elsewhere the cfassical methods of fire assay have continued to
withstand all competition from the new methods of chemical analysis which have been adopted as
standard practice in other f?elds.” He outlined the steps which are being taken to reduce the various
errors caused by pulverizing, cupellation, parting erc. It was considered that for existing processes
of cupellation “The absorbed gold under good conditions of assay varies from about 0.5 to over 2 %
being inversely proportional to the weight of the final bead” and under specific conditions the oven
assay is reasonably in balance with the measured losses of gold in the cupellation stages. The author
concluded that a rise in the price of gold would encourage researches designed to produce increased
accuracy of determination.
In an interesting paper, HanrahaP stated that the results from two sm of assaying showed
that the maxim um diierences in results were 2 parts per 10000. In discussion of the optimum gold-
silver ratio for cupellation he stated that the absorption of gold by the cupel tended to decrease with
increasing proportions of silver. The ratio of silver to gold of 25 represented the highest concentration
of silver which could safely be subjected to normal acid parting.
Among the reports which deal with the losses of gold during the various stages of a fire assay,
Coxon, Verwey and Lock- have shown that routine mine assays give values about 2 % too low, the
losses being made up as follows; 05% of gold to crucibles, slags, and button dressing, 0.7 % to cupel,
1.2 % dilution by iron during pulverizing. These losses are partially compensated for by the 0.4%
 1142 F. E. BEAMISH

of silver retained from the bead after parting. Sinclair64 pointed out that for samples which require
little or no pulverizing the l-2 % loss can be neglected and for those cases where the Giver retention by
the gold is higher there may be a compensation of errors. Sinclair also reported on the amount of
gold lost to slags and cupefs and compared these losses with the silver content retained after parting.
To determine the slag losses he made I32 residue fusions, the slags from which were remelted and re-
assayed without crushing: the loss to slags was 2.20 %. Losses toihe slags and cupels, determined from
44 residue fusions without crushina, were 2.59 and 052 “/, resuectivelv. In the latter case, drillines with
&-in. drill to a depth of # in. were ‘;;sed. The actual loss was;educed by 1.91% because of silver’reten-
tion in the parting acid. The gold content of the litharge averaged OQO36mg for 4-3 g. Experiments
with crushed slags and cupels which showed a nett loss of 12.82 y0 indicated an appreciable contamina-
tion from the manganese-steel jaw-plates, and a direct deter~nation on 8-12 g of sl&nmings from the
latter yielded @041 g of gold. The nett loss on large composite samples of residues averaged l-27%
the slags from the fusions were not crushed. In a similar exueriment in which the slaes were crushed
the apparent nett loss of gold was 9.35 %. Salted composite &mples, for crucibles washgd with litharge,
yielded results 2.23 X low and for unwashed crucibles 1*36X: ,I salted barren silicate samules
I ,vielded
corresponding values low by 2.24 % and 1.25 3:.
For the determination of gold in ores, concentrates and cyanide eluates, Linhart6” rejected the
polarographic determination because of losses of up to 50% caused by adsorption on voluminous
precipitates such as silver chloride and iron hydroxide. He cakined the samples and formed a lead-
gold button by fusion at 800-900” in the presence of borax and sodium chloride The lead was cupelled
to form a silver bead which was parted in nitric acid. The parting acid was drawn off by suction and
the gold dissolved and determined polarographically. To determine gold in biological material the
latter was ignited and treated as described above.
For the determination of gold in ores, Bagdasaryn@ used a wet extraction of 50-100 g with nitric
acid, leaching 7-8 times. The dried residue was treated with aquu regiu; the mixture was evaporated
and treated with hy~ochlori~ acid. The chloride filtrate was evaporated to small volume and neutral-
ized to litmus. Gold was precipitated by excess of standardized iron(B) sulphate, the excess being
back-titrated with potassium dichromate with ~-phenylant~anilic acid as indicator.
Aluminium-gold alloys cannot be cupelled directly because of the oxidation to aluminium oxide.
Scherfaufers7 selectively dissolved the aluminium with 5M sodium hydroxide and filtered the gold.
The paper was ashed in lead foil which was then fused to produce a lead button. Gold was then deter-
mined conventionally by cupellation.
An assay method for the determination of gold in materials containing high proportions of iron
or chromic was recorded by Masl~its~i et al. 68 Iron oxide was removed by dissolution in hydro-
chloric acid and chromium oxide by aqua regia. Gold was then determined by precipitation and
cupellation.
One of the most exhaustive examinations of the Iire assay process for the collection of nanogram
amounts of the noble metals was described by Beyermann .88 The fusion technique was varied some-
what in that a little flour and lead granules were added subsequent to the fusion at 900”. The fusion
mixture contained litharge, borax, silica, sodium carbonate and flour. Re-fusions of the slag were
also made and the lead buttons were then combined. Standard cupels as well as some prepared from
bonemeal and Holland cement were used for ~u~llation after preheating at 900”. In general, the
noble metal content was about 1 ppm. The sampl& analysed weie galena,Teldspar, roast;d ore and a
svnthetic ealena. The author’s results aareed with urevious findings that collection bv silver beads was
heffectivg for nanogram amounts of iyidium, ruthenium and &odium but relati;ely complete for
gold, palladium and platinum.
Losses of radioactive palladium, platinum and gold during the parting process with either con-
centrated sulphuric acid or 50% v/v nitric acid were investigated. The parted residue was filtered
through media with 1-p pores and the radioactivity of the filtrate was measured. For IO-mg beads
and platinum contents ranging from O-02 to 1.8% the relative losses after parting with nitric acid
ranged from 19 to 2% respe&~ely for these extremes. In the case of sulphuri~ acid&e relative losses
were 0*04”/, for @lo/, of ulatinum and 6X for 0.003 “/, of platinum. For O-01“/, of rhodium, with
sulphuric ‘acid the reiatid loss was 81 ‘A;-slightly re&ed*when small amounts of platinum were
associated with the rhodium. Parting gold-silver beads with sulphuric acid resulted in a relative loss
of @4 % for 1.0 % of gold and for WOO5% a loss of 0.8 %.
Investigations were made of losses to the container wall during evaporation of aqueous solutions
of rhodium, rutheni~ and gold and subsequent heating of the residues to 120” and 400”. Solutions
of 50 and of 2 ng of rhodium, gold and ruthenium were evaporated in such a way that the residue
covered an area of l-2 cm*. Foliowing evaporation the residue was heated for 10 min. After cooling
the area was rinsed twice with IM hydrochloric acid and the activity was then measured. With the
residue heated at 120”, the losses of rhodium, ruthenium and gold to glass were respectively @8, 0.7
and 9.5 %. At 300” the losses were appreciably higher. With quartz, the corresponding losses were
 Methods of isolating and separating the noble metals-III 1143

appreciably smaller. The activities retained in the container were not significantly removed by treat-
ment with hydrochloric acid: chlorine and bromine were without effect, but short treatments with
hydrofhoric acid eliminated the activities.
The effects of associated metal constituents on the efficiency of various extractions of each noble
metal were examin ed. The noble metal was present in concentrated hydrochloric acid or ammonia
solution. In each case Z-rig, SO-ng and 2-mg amounts of noble metal were used along with the associ-
ated metals, including silver and lead. The extractants for which there was no appreciable influence
from associated metals were tributyl phosphate in benzene (10 mg of associated metal): n-octylamine
and tri-n-octylamine in chloroform (50 mg of associated metal): butylamine in chloroform (from
ammoniacal media): the extraction of 50 ng of rhodium, ruthenium and iridium with sodium diethyl-
dit~~r~ate in chloroform from sodium carbonate or ammonical media was not affected by
10 mg of platinum, palladium or gold: 2 rug of rhodium could be extracted in the presence of uranium
or thorium. Special precautions were necessary for the separation of rhodium, iridium and ruthenium
from ammoniacal solutions of platinum, palladium and gold. Extraction of 10 ng of gold from hydro-
bromic acid was 99.8 % complete: furthermore, in the presence of 50 mg of platmum, palladium
rhodium and iridium and 50 mg of gold, only traces of the platinum metals were co-extracted. With
dimethylglyoxime in ethanol and chloroform 50 ng to 1 mg of palladium could be extracted from
solutions containing platinum, rhodium, iridium, calcium or ahuninium chloride: only l-2% of the
noble metals were found in the chloroform layer.
The efficiency of co-precipitation of 2 and 50 mg of noble metals with the common precipitants and
collectors was investigated. These were lead and rhenium sulphides, mercury mixed with mercury(I)
chloride, selenium silver and lead chlorides, and various hydrous oxides. In the case of lead sulphide,
additions of 10 g of nickel, ahuninium, calcium and sodium were without adverse influence. Co-
precipitations of the sulphides of rhodium, iridium and ruthenium with rhenium sulphide required
the application of pressure. With selenium, co-precipitation was accomplished by sodium sulphite
in sulphuric acid medium. In the presence of 50 mg of siIver and 2 ng of the noble metals 24% of the
silver was co-precipitated, and if @l mg of silver was present, 80% of it was co-precipitated. Co-
precipitation of the platinum metals with a mixture of mercury and mercury(I) chloride was unaffected
by high proportions of iron, nickel and gold. Co-precipitation of palladium with silver chloride has
been a consistent problem in the parting of noble metal-silver assay beads. Beyermarm found that
dissolution of the silver chloride in ammonia and reprecipitation of the silver chloride removed two-
thirds of the co-precipitated palladium: the results for asolution containing 2 rg of the noble metal
and 10 me of silver are of some si&Iicance-the amounts of iridium(IV) oahadium(I& eold(IIIL
platin~~~ and rh~ium(II1) co@ecipitated were 95, 19, 9, 3 and 2% iespectively. ‘Igthe.case
of 1 g of lead as the chloride in the presence of 2 and 50 ng of rhodium, 2 and 4 ‘A respectively of the
latter were co-precipitated.
He also dealt with co-precipitation of the noble metals with iron hydroxide from various aqueous
media. From 2M ammonia solution containing 50 ng of rhodium(III), gold(III), platinum(IV),
iridium(IV) and nahadium(II) the recoveries were 59. 35, 45, 24 and 3 % resnectivelv. Results were
given for solutions of sodium acetate and concentrated ammonia. In-the-case 01 the potassium
bromate hydrolytic method of Gilchrist and Withers PJ it is noted that in the presence of 5 mg of
nickel, iron, akin or rhodium the amount of platin~(~ co-precipitated from a total of 50 ng
was 2,4,4 and 5 ng respectively. Similarly, small amounts of gold were also co-precipitated, particu-
larly in the case of 50 ng of palladium. About 2 ng of platinum(IV) were co-precipitated in solutions
containing 25 mg of rhodium, palladium or iridium and 50 ng of platinum(IV).
The author’s investigation of the separation by ion-exchange resulted in some unexpected con-
clusions. Dowex 1 was used to adsorb gold, platinum and palladium in 6M hydrochloric acid. only
5-15x of the anions could be eluted with concentrated hydrochloric acid. Continued treatment
removed two-thirds of the metals, and eluting media such as 2M ammonia were less effective.
Recovery was accomplished sucking the resin dry, mixing it with litharge and fusing. Treatment of
the lead regulus resulted in complete recovery of either 10 pg or 50 ng of rhodium. In the case of 50 ng
of rhodium along with 50 mg of palladium, iridium or platinum, the recoveries of rhodium were 92,
100 and 97% respectively, the other three metals remaining adsorbed. In the present author’s
opinion, these results require some coniirmation.
The author’s results dealing with losses by evaporation were significant. Solutions containing 5 ng
to 2 mg of ruthenium could be fumed with hydrochloric or nitric acid or aqua regiu without loss, but
in the presence of 25 mg of osmium and 10 ng of ~thenium 1% of the latter was lost on fuming with
concentrated nitric acid. For the volatilization of 10 ng to 5 mg of ruthenium the author preferred
sulphuric acid-bromate, hypochlorite (chlorine in alkaline solution), and perchloric acid to the sul-
phuric acid-&chromate oxidation. In contrast to other oxidizing media, perchloric acid avoided the
introduction of oxidizing media into the receiving solutions and thus losses of ruthenium. Nitric
acid was an effective oxidant for osmium but the receiving liquid of hydrochloric acid required
1144 F. E. BEAMJSH

saturation with sulphur dioxide, and losses of osmium resulted from boiling off the hydrochloric
acid.
Reyermann proposed three methods of treating silver beads. One applied to beads containing no
more than lO,~g of gold or platinum. A second was intended for the determination of gold and
palladium only. These two procedures involved combinations of standard treatments such as parting
with sulphuric acid. Techniques were included for the recovery of palladium adsorbed by the silver
chloride and for the isolation of palladium by extraction of the dimethylglyoximate by chloroform.
Palladium was finally determined amperometrically. The “insoluble” residue of gold or platinum
from the parting treatment was treated with aqua regia and fumed with perchloric acid, and gold
was extracted as the bromide by diethyl ether and finally determined amperometrically. Platinum
in the aqueous phase was also determined amperometrically.
The third method involved parting the bead with sulphuric acid, then adding selenium dioxide and
sodium sulphite to the mixture. The residue, of gold, platinum, palladium and part of the silver, was
dissolved in aqua regia and fumed to remove the selenium. Gold as the bromide was extracted by ether,
followed by extraction of palladium dimethylglyoximate from the aqueous phase by ether, and then
platinum was determined as in the tirst two methods. The results recorded for salted solutions
indicated good recoveries of gold, palladium and platinum in microgram and submicrogram amounts.
In each case, the solutions contained a second noble metal from which there was no significant
interference. The methods were also applied to a variety of noble metal products such as roasted ore,
iron ore, slags, silver coins and lead, generally with an accuracy acceptable in comparison with other
methods. For the iron ore the values obtained by a spectrochemical method for gold, platinum and
palladium were 0.09,0.24 and 0.4 respectively as compared with the author’s values of 0~13,0*10,0*21.
The present author’s efforts to secure copies of original Russian publications dealing with the
fire assay for noble metals met with little success. Consequently, most of the information included
below has been obtained from various abstracts and thus details of significance may be lacking.
In order to ensure thorough mixing of the pot charge during the assay procedure for noble metals
Anisimov et al.‘” arranged for mechanical mixing during the smelting process: the initial heating
procedure was slow and super-heating was applied at the conclusion of the fusion. An excess of
reducing reagent was required with a limited amount of litharge, part of which was added as a solution
of lead acetate in order to secure uniform distribution and intimate contact with the noble metals.
One may assume that the moistening process involves wetting the whole of the pot charge. Light
and fluid slags were obtained with the ores used and the acidity of the charge varied between 1 and 15.
The authors noted that impurities collected by the lead button assisted in the extraction of the platinum
metals. In the case of noble metal concentrates, the accuracy of recovery was increased by resmelting
the slag and the cupel.
Amsimov and his co-authors made several reports dealing with the fire assay of the noble metals.
The losses of platinum metals at each stage of the assay of concentrates was studied.?’ The techniques
included the preparation of a silver-leaa button, scdrification, reassay of slags and cupels, and-the
analvsis of the silver bead. The ratio of noble metals to silver in the bead was 1: 15 or 1: 20. but in the
case’of high iridium was made 1:40. The fusion was carried out at 1000-1050” for 15 hr: The 40-g
lead buttons were scarified to 18.25 g. The cupellation temperature was 900-950”. The authors
included the losses to slags and cupels for platinum, palladium, rhodium, iridium and ruthenium.
The amounts recovered from two silver beads were respectively lOO*Oand 100.0, 99.9 and lOO*O,
99.5 and 99-7.98-3 and 100-0, 99.2 and 99.9 %. Small amounts of selenium or tellurium were without
effect. A separate lead button was used for ruthenium. The authors found that reassay of the first
pot slag, and of the cupel, was required for good accuracy. To combine two silver beads, fusion was
made in a chamotte crucible under a layer of soda, borax, and wood charcoal. Anisimov et al.‘* also
provided a detailed description of a method for treating silver assay beads. These were parted with
nitric acid and the residues were treated with uqun regiu to dissolve the gold and platinum. The former
was finally precipitated with sulphur dioxide and the latter as ammonium chloroplatinate. The residue
from the uquu regiu treatment was sintered with barium oxide, the sinter was treated with hydrochloric
and sulphuric acids and the barium sulphate was removed. The filtrate was evaporated to fumes and
sodium bromate was added to produce ruthenium tetroxide which was distilled and determined by
potentiometric titration. The pot-liquid was boiled with hydrochloric acid to destroy the bromate,
and sodium nitrate free from ammonium or potassium ion was added. Sodium sulphide was added to
precipitate the rhodium and any residual platinum. The precipitate was dissolved in aqua regiu, and
later evaporated with hydrochloric acid. Platinum was again precipitated as ammonium chloro-
platinate. The filtrate was treated with zinc to precipitate the rhodim, which was purified by fusion
with silver under borax to remove contaminants. Iridium in the filtrate from the sulphide precipitation
was finally determined potentiometrically.
For the determination of gold and silver in ores, Anisimov et ~1.‘~ prepared a second lead button
from the scorikation and crucible slags and scarified the two buttons. Gold was determined by a
potentiometric titration with hydroquinone.
 Methods of isolating and separating the noble metals-III 1145

The techniques and principles involved in the assay methods were. reviewed by Anisimov.?6 The
function cupellation and scarification was outlined and methods for determining osmium and ruthen-
ium were discussed. In particular, the author included a description of the wet methods of parting
and determining the constituents in a noble metal-silver bead.
Methods of preparing ore concentrates for the assay of platinum metals were discussed by Anisimov
et al.76 The concentrates included copper-nickel ores, mattes, cementates, and other products con-
taining large amounts of selenium, tellurium, arsenic, and antimony. The finely ground product was
mixed with sawdust, sulphur and concentrated sulphuric acid, and was heated first at 240-300” and
then at 375’ for 2 hr. The powdered sinter was mixed with aqueous starch, dried, and again heated
at 375” for l-2 hr. The sinter was leached with water and the residue was fused. If appreciable
selenium, tellurium, arsenic, tin and antimony were present the sample was chlorinated after mixing
with sodium chloride and charcoal and wetting with starch to form granules: these. were dried in a
desiccator and heated in a tubular oven for 4 hr at 500-550” in a chlorine current. A single treatment
removed up to 95 ok of the selenium and tellurium. If the sample contained iridium or ruthenium as
well as copper the residue was subjected to two sulphuric acid treatments. The presence of copper in
the lead assay button encouraged dissolution of ruthenium and iridium by the nitric acid parting
solution.
Anisimov et ~1.~7found that a spectral analysis of assay beads obtained from low-grade products
increased the overall sensitivity of determinations relative to classical methods and thus permitted the
use of smaller weights of samples than those normally used for tire assay: 50-mg silver-rhodium
and copper-ruthenium-iridium beads were used. These were obtained from the residue of the lead
parted with dilute nitric acid. The authors stated that when the weight ratio of platinum and palladium
to ruthenium and iridium was less than 100, the latter two metals could be determined spectrochemi-
tally, without a separation of platinum and palladium. When the weight ratio was 200, the residue
from the silver bead could bc treated with aqua regia. It is most unfortunate that the details of the
spectrochemical methods were not available. Normally, one of the most important problems is the
necessity of achieving homogeneity in the beads or residues used by these authors. They stressed
again the intluence of copper and nickel which caused some dissolution of iridium and ruthenium
in the button parting acid. A scheme of analysis of ores and their concentrates for rhodium, ruthenium
and iridium was included.
For almost a century following the discovery of the platinum metals the only effective general
method for the determination of these metals in ores involved a collection by molten lead from fused
media of suitable composition. Furthermore, there can be little doubt that this method suggested
itself through the similarity in properties of the platinum metals and gold, for which the assay-metal-
lurgical method had been an ancient procedure. It is only within the past few years that organized
investigations have been made to determine the effectiveness of lead as a collector for each of the six
platinum metals and indeed even for gold. The results of these investigations, made in various parts
of the world, are in approximate agreement that for such metals as gold, platinum, palladium,
ruthenium, osmium and perhaps for rhodium the lead collection is effective for amounts of the order
of small fractions of a milligram. Thus the classical fire assay will reveal commerically valuable
noble metal values for all ores and concentrates although quantitative success will depend upon the
assayer’s experience and wisdom in the choice of technical conditions. Much of the early work which
was concerned with the examina tion of the efficiency of classical 6re assay has been reviewed.‘*
Results were obtained which indicated some lack of efficiency of the tire assay for iridium. Because of
this and the desirability of having available alternative assay methods for noble metals, various
researchers developed procedures which involved collectors other than lead.
Tin has been used as one alternative. In a series of papers Faye and co-workers’“-8* have developed
a scheme of separation of noble metals incorporating spectrophotometric methods of determination.
The various steps in the method were each carefully proved and the method as a whole was applied
with success to copper-nickel matte, precious metal concentrates, meteorites, and iridosmine.
Although Faye provided a procedure for the continuous determination of the seven noble metals
from a single sample, variations in procedure allowed for simplified determinations of certain noble
metals in the absence of others. If the ore contains much sulphur or arsenic it is roasted at 750-800”
to avoid the formation of a matte or speiss. For samples containing high proportions of copper and
nickel a leaching process is recommended. In general, the precious metals are collected in molten tin
when the sample is fused at 1200-1250” with a flux containing tin(IV) oxide, sodium carbonate, silica,
borax and powdered coke. The resulting tin alloy is placed in a graphite crucible and melted at 600-
700” in an atmosphere of nitrogen and then poured into a gallon of water to disintegrate the button.
The liquid is decanted and the particles of noble metal-tin alloy are dried. These are parted in
concentrated hydrochloric acid and if osmium is present the tetroxide is distilled in the presence of
hydrogen peroxide, collected in hydrobromic acidTo and determined with pyrogallol.80 The pot
residue is treated with hydrochloric and hydrobromic acids to volatilize the tin. In the absence of
1146 F. E. BEAMISH

osmium this treatment is applied directly to the solution of the noble metal residue produced by adding
tin to the hydrochloric parting acid.
The mixed halide solution is treated to remove gold by extractionwi~ diethyl ether and determined
with o-tolidene.81 At this stage, if more than 2 g of base metals are present they are removed by
cation-exchange, and then the acid solution of the remaining noble metals is treated with perchloric
acid to volatilize the ruthenium tetroxide which is collected in concentrated hydrobromic acid and
determined spectrophotometrically .” The perchloric acid in the pot solution is removed by evaporation
with hydrochloric acid* and with ayuu regia, and further evaporations with hydrochloric acid. The
chloride solution is treated with sodium iodide, and rhodium and iridium are extractedwith tri-n-butyl
phosphate (TBP) in hexane .ea--86These metals may be separated by a modification of the method of
Yoe and ~r~and,Bsin which case the parading-~~oso~methyl~iline complex is extracted from the
aqueous phase by chloroform. The aqueous phase is treated with sulphuric and perchloric acids and
platinum is determined with tin(H) chloride. The TBP-hexane phase is evaporated nearly to dryness
and nitric acid is added to destroy the iodides. The metal salts are then converted into chlorides and
hydrogen peroxide is added to ensure oxidation to iridium(N) which is extracted by TBP-hexane,
stripped by hydrochloric acid and finally determined with tin(II) chloride in hydrobromic acid.
Rhodium in the aqueous phase is reconverted into the chloride complex and the solution is evaporated
to dryness. The metal is determined with tin@) bromide in perchloric acid.
The complete procedure, including a number of variations, depending upon the number of metals
present and their proportions, was recorded by Faye .86*87The reports bear the marks of carefully
applied techniques and the results are excellent.
Recent researches, reviewed earlier by the present author, 78have indicated that the composition of
the flux used for the classical assay may affect the recovery of iridium, basic slags producing relatively
high losses. High losses also occur in the presence of much copper and nickel. To explain these losses
it was suggested that the lead collector, which has a minimum crystal size factor of 340 A, is above the
theoretical size required for a good solvent of precious metals. Thus, the collection of iridium by lead
is probably achieved largely through mechanical processes. This and other deficiencies in the quantita-
tive collection of iridium by classical fire assay led to the investigation of other approaches. Because
of the regular occurrence of copper, nickel and iron with the platinum metals it seemed probably that
these metals could serve as collectors in a fire assay. One would guess too that the disintegration of the
base metal silicates and the noble metal constituents would occur simultaneously and that their
juxtaposition would thus encourage the efficiency of collection. The base metal buttons were prepared
by reduction with mechanically mixed graphite, soda ash, borax, and ore with base metals added if
required. The buttons were placed in standard distilling equipment together with hydrochloric acid
(3 + 1) and a stream of nitrogen was passed through .** Heat was applied to accomplish dissolution
Any small amount of osmium tetroxide developed during this treatment was collected in a receiver
trap containing hydrochloric acid into which sulphur dioxide was bubbled. The content of the
distillation flask was heated to yield a moist chloride cake which was moistened with water
and allowed to stand overnight exposed to the atomsphere. The mixture was then diluted with
water and filtered. The filter paper and residue were treated in the distillation flask with sodium
hydroxide so1ution, chlorine was passed through and any osmium and ruthenium tetroxide were
collected in hy~ogen peroxide. This solution was added to the pot-liquid as indicated below.
The filtrate from the pot-liquid was diluted with water, the pH adjusted to @9-l-5 and the base metals
selectively removed by a cation-exchanger which was then washed with hydrochloric acid at pH 1 (see
also reference 90). The effluent was treated with a stream of sulphur dioxide for 1 hr followed by a
6-hr standing period. The solution was evaporated to 500 ml on a hot-plate and subsequently to 15 ml
in the distillation flask under an atmosphere of sulphur dioxide.
This solution was combined with the osmium and ruthenium from the filter paper and residue as
indicated above, and then treated as described bv Westland and Reamishsl to remove osmium, and
the pot-liquid cantoning the remaining platinum~met~s was also treated by Westland’s procedure to
remove ruthenium.
The pot-liquid now contained all of the platinum, palladium, iridium and rhodium. Rhodium and
iridium may be separated and determined as described by Tertipis and BeamishSIJ’a or by Sant and
Bear&has and also by the various other methods recorded above. Platinum and palladium may be
separated and determined as described by Coburn et aLs4and by Plummer.*6
The physical characteristics of the iron-copper-nickel-noble metal alloys were described by
Plummer et ~1.~~ It was concluded that palladium and probably some of the other platinum metals
form solid solutions with the alloy.
+ Recent researches by the author have shown that though perchloric acid is removed by evapora-
tions with hydrochloric acid and/or treatment with ayua regia, the removal is not sufficiently
complete to allow the subsequent use of a cation-exchanger for the separation of decagram
amounts of base metals.
 Methods of isolating and separating the noble metals-III 1147

The iron-copper-nickel collection was applied to determine the composition of iridosmines.

Sen Gupta and Beamishg7 found that the fusion allowed the dissolution of even such resistant alloys
as Tasmanian iridosmines. The results obtained by the method agreed with those. from a chlorination
of the iridosmines.
The ironapper-nickel collection was used by Kavanagh and Beamisho8 to collect osmium and
ruthenium quantitatively from roasted sulphide ores. The method involved parting of the alloy
and distillation of the tetroxides of osmium and ruthenium from perchloric acid. Thiouma in 6M
hydrochloric acid was used as the collecting liquid when both metals were distilled simultaneously,
and hydrobromic acidwhen ruthenium onlywas distilled. The metals were precipitated as the thionalide
complexes. Perchloric acid was also used for the dissolution of siderite meteorites.0e The platinum
metals were separated and determined as in the case of the iron-copper-nickel assay button.
This flm assay for osmium was extended to apply to sulphide concentrates. This adaption avoided
the oxidation previously used by Kavanagh to eliminate sulphur, a process which could cause the loss
of osmium. Van Loon and Beamishroo used hydrogen to reduce the sulphur content to a level which
allowed a clear fusion. For flotation concentrates, salted with 16-805 pg of osmium per assay ton,
only microgram losses were found.
Because. of the high temperature required for the fusion of iron-nickelapper buttons an effort
was made to reduce the required temperature by the addition of copper to the ore. Banbury and
Beamishrol collected platinum and palladium quantitatively in this way. Buttons of 25-40 g in weight
were prepared from a roasted artificial sulphide containing 815 g of copper sulphide. The palladium
content varied from 40 pg to 10 mg: the platinum content from 50 pg to 10 mg. Recoveries were
acceptable in all cases. Initial furnace temperatures were 1100” and were raised to 1200” over a period
of 15 hr. The chemical treatment involved the removal of copper by cation-exchange from the parting
solution adjusted to pH 1.0 f O-1. By a similar process, Banburyl”* collected rhodium by means of
copper. As in the case of platinum and palladium, losses to siliceous residue, slag, pot-wall, and
ion-exchange resin were found to be negligible. Copper was also used as a collector or iridium.
Agrawal and Beamish*o8 considered copper superior to lead for the collection of iridium in that it can
be applied successfully in the presence of the commonly associated transition metals. It was superior
to copper-iron-nickel in that comparable collections could be made at lower temperatures. Maximum
recovery was achieved with acid slags of 3-O silicate degree.
Despite the availability of new collectors for the noble metals, the classical lead collector remains
clearly superior for the determination of platinum, palladium, gold and perhaps rhodium. The
device of cupellation to remove the lead selectively requires very little attention and it is a reasonably
rapid process compared to the methods required for the removal of copper, nickel, iron or tin.
Furthermore, the ease of forming a silver-noble metal alloy by cupellation allows use of a variety of
potentially useful methods of separation and determination. It should be noted that though osmium
is collected quantitatively by lead, cupellation is not acceptable because of losses of osmium tetroxide,
and any attempt to form a silver bead will increase the osmium losses. On the other hand, the removal
of both osmium and ruthenium from lead by a parting distillation with perchloric acid is accomplished
with ease.
For those analytical consultants who frequently must deal with the claim that the classical lead
collection will sometimes fail to reveal noble metal values of commercial importance, the present
author emphasizes the fact that no valid support has ever been recorded for this claim and one
cannot doubt that the classical method used by competent assayers will invariably indicate
whether a noble metal content is of commercial value.
Zusammenfaasrmt-Diese uhersicht handelt von den Trennungen der
sieben Edehnetalle durch Chromatogra hie, Fallung oder Schmelzver-
fahren, die zwischen den frtlheren &Je rsichten und Januar 1966
publiziert wurden. Die in Lbsung vorliegenden Edelmetallspezies
werden diskutiert, sofem sie zur Trennung und Bestimmung in
Beziehung stehen.

R&annUette revue traite des methodes de separation des sept

metaux nobles par chromatographie, precipitation et voie s&he,
publiQs ult&ieurement aux ptidentes revues et anterieurement a
janvier 1966. On discute des esp&ces chimiques sous lesquelles les
metaux nobles sont presents en solution dans la mesure oit elles sont
rel&s B la separation et au dosage.
REFERENCES
1. M. Lederer and S. K. Shukla, J. Chromamg., 1961, 6, 353.
2. F. Larsen, J. Chem. E&x., 1964,4,435.
2
1148 F. E. BEAMLQX

3. C. Vassiiiades and G. Manoussakis, Chim. Anal., 1961,43,225.

4. E. Blasius and M. Fischer, Z. Anal. Chem., 1960,177,412.
5. G. S. James, S. African Ind. Chemist, 1961,15, No. 4, 62.
6. N. F. Kember and R. A. Wells, Analyst, 1955,80,735.
7. P. Cexn~ and J. Kubin, Acta Fat. Rerum Nat. Univ. Comenianae, Chfmmia,1960,4,667.
8. M. B. &lap T. J. JanjiC and i. F. Spanovic, Glasnik Hem. Drustva Beograd, 1960-1961,
25-26,527: ibid., 1960-61,25-26,125, 133.
9. W. Noddack and S. Sanir, Chemiker Ztg., 1964,88,19,738.
10. V. G. Voicu and T. Nascutiu, Acad. Rep. Populare Romine Studii Cercetari Chim., 1961,9,699.
11. Chia-Lung Kao and Chin-Yao Yuan, Hua Hsueh Hsueh Pao, 1964,30,418.
12. Chia-Lung and Ming Sun, K’o Hsueh T’ung Pao, 1964,5,434.
13. Chih-Te Hu and Hsiu-Chun Shi, ibid., 1963, 8, 54.
14. Z. T. Hu and S. C. Shi, Hua Hsueh Hsueh Pao, 1964,30,352.
15. A. S. Kertes and A. Beck, J. Chromatog. 1959,2,362.
16. L. C. Locke and E. C. Martin, ibid., 1962,7,120.
17. Chia-Lung Kao and Hsiu-Hua Chang, Huu Hsueh Hsueh Pao, 1963,29,421.
18. M. Shima, Radioisotopes (Tokyo), 1957,5, No. 2, 1.
19. F. E. Beamish, K. S. Chung and A. Chow, Tahantu, 1967,14, 1.
20. A. N. Rao and T. V. S. R. Sagar, Current Sci (India), 1963, 32, No. 2, 79.
21. R. R. Lever, Australasian Inst. Mining &Met. Proc. 1964, No. 210, 139.
22. H. Agrinier, Collpt. Rend., 1%2,255,2801.
23. H. HranisavljeviGJakovljevic, I. Pejkovic-TadiC and J. MiljkoviCStajanoviC, Mikrochim.
Ichnoanal. Acta, 1965, 141.
24. Meng-Yen Wu and Liang-Fang Wu, K’o Hsueh T’ung Pao 1964 (lo), 939: c$ Lan Chou Ta,
Hsueh Hsueh Pao, Tzu Jan K’o Hsueh 1964 (l), 57.
25. V. Been and A. Hoeye, J. Chromatog., 1965,17, 631.
26. T. Ashizawa, Bunseki Kagaku, 1961,10,555.
27. Zdem, ibid., 1961, 10,449.
28. Zdem, Japan Analyst, 1961, 10, 350.
29. H. Koch and H. Bruchertseifer, Radiochimica Acta, 1965,4,82.
30. S. K. Shukla, .7. Chromatogr., 1962, 8,96.
31. J. M. Fletcher, and P. G. M. Brown, E. R. Gardner, C. J. Hardy, A. G. Wain and J. L. Wood-
head, J. Znorg. Nucl. Chem., 1959,12,154.
32. P. G. M. Brown and A. Naylor, J. Appl. Chem., 1960,10,422.
33. A. G. Wain, P. G. M. Brown and J. M. Fletcher, J. Itwrg. Nucl. Chem., 1960,12,346.
34. D. Scargill and C. E. Lyon, At. Energy Res. Stab. (Gr. Brit.) Rept., 1963, R4292, 14pp.
35. D. Scargill, ibid., 1964, R4583,22 pp.
36. F. RalIo, J. Chromatog., 1962,8,132.
37. S. K. Shukla and M. Lederer, J. Less-Common Metals, 1959, 1,255.
38. Zdem, ibid., 1959,1,202.
39. S. K. Shukla, J. Chromatog., 1958, 1,457.
40. M. Lederer, ibid., 1958,1,279.
41. F. Basolo, M. Lederer, L. Ossicini and K. H. Stephen, ibid., 1963, 10,262.
42. I&m, Ric. Sci. Rend. Ser. 2, 1962,2,485.
43. E. Blasius and W. Preetz, Z. Anorg. Allgem. Chem., 1965,335,1.
44. G. V. Assamson, K. Peterson and A. M. Asklund, Z. Anal. Chem., 1960,174,194.
45. A. F. Bukharin, Sbk. Tr. Vses. Nauchn.-Issled. Inst. Gorno-Met. Inst. Tsveinoi Met., 1959, No. 5,
209.
46. Y. Yjihira, Ni pon Kagaku Zasshi, 1963,84,642.
47. V. Voicu and c . Dema, Acad. Rep. Populare Romine, Studii Cercetari Chim., 1959, 7, 431.
48. F. Jasim, R. J. Magee and C. L. Wilson, Talanta, 1960,4,17.
49. Yu. V. Morachevskii and A. I. Novikov, Khim. Nauk, 1950, No. 18, 129.
50. S. I. Ginzburg, M. I. Yuz’ko and N. N. Chalisova, Zh. Analit. Khim., 1963, 18,222.
51. G. Tertipis and F. E. Beamish, Anal. Chem., 1960,32,486.
52. I. V. Prokof’eva and A. E. Bukanova, Zh. Analit. Khim., 1965,20, 598.
53. S. S. Dehmlov and I. L. Preiss, Z. Anal. Chem., 1963,194, 183.
54. G. B. S. Salaria, ibid., 1961,180,358.
55. F. E. Beamish, Tulanta, 1958, 1,3.
56. I. K. Taimni and S. N. Tandon, Anal. Chim. Acta, 1960,22,553.
57. I. K. Taimni and G. P. S. Salaria, ibid., 1954, 11, 329.
58. C. Duval, ibid., 1959,20, 152.
59. H. Zachariasen and F. E. Beamish, Talantu, 1959,4,44.
 Methods of isolating and separating the noble metals-III 1149

60. V. I. Shlenskaya, Zh. Analit. Khim., 1965,20,136.

61. H. Britten, Trans. Commonwealth Minilap Met. Congr., 1961,3,1005.
62. J. F. Hamahan, J. S. African Inst. Mining Met., 1962, 62,712.
63. C. H. Coxon, C. J. Verwey and D. N. Lock, ibid., 1962,62,546.
64. W. A. Sinclair, ibid., 1964,64,333.
65. F. Linhart, Chem. L&y, 1961,55,552.
66. A. M. Bugdasaryan, Nauchn. Tr., Nauchn.-Issled. Gorn.-Met. Inst., Sov. Nar. Khoz. Arm. SSR,
1962, No. 3,349.
67. A. M. Scherhaufer, Z. Anal. Chem., 1960,174,352.
68. I. M. Maslenitskii and L. N. Polievskii, Izo. Vysshikh. Uchebn. Zavedenii, Tsvetn. Met., 1961,
4, No. 5, 139.
69. K. Beyerrmum, Z. Anal. Chem., 1964,200,183.
70. R. Gilchrist and E. Withers, J. Am. Chem. Sot., 1935,57,2565.
71. V. I. Anisiiov, I. E. Karbovskaya, V. I. Gorlanov and Z. F. Oberenko, Met&y Analiza Platin.
Metallov, Zolota i Serebra (Moscow; Gos. Izd. Lit. Chemoi i Tsvetn. Met.) Sb. 1960,33.
72. S. M. Anisimov, N. Ya Semenova, A. Z. San’ko and V. I. Gerlanov, ibid., 1960,70.
73. S. M. Anisiiov, I. N. Firsova, P. G. Shulakov and Kh. I. Tsybulevskii, f&f., 1960,70.
74. S. M. Anisiiov, I. E. Karbovskaya and V. M. Alyanchikova, ibid., 1960, 132.
75. S. M. Anisimov, ibid., 1960,7.
76. S. M. Anisimov, E. I. Nikitina and A. P. Rodnova, ibid., 1960,151.
77. S. M. Anisimov, A. P. Osipov, A. I. Chemova and M. A. Orekhov, Vopr. Analiza Blagorodn.
Metal. (Novosibirsk; Sibirsk. Otd. Akad. Nauk SSSR) Sb. 1%3, 3.
78. F. E. Beamish, Takznta, 1960,5,1.
79. C. H. Faye, Anal. Chem., 1965,37,696.
80. Idem, ibid., 1965, 37,259.
81. G. H. Faye and W. R. Inrnan, ibid., 1961,33, 1914.
82. Idem, ibid., 1961,33,278.
83. I&m, ibid., 1962, 34,972.
84. G. H. Faye, W. R. Inman and P. E. Moloughney, ibid., 1964,36,366.
85. G. H. Faye and W. R. Inman, Talanta, 1960,3,277.
86. G. H. Faye, Can. Dept. Mines Tech. Surv. Mines Branch Res. Rept. R., 1965,154,27 pp.
87. Idem, Anal. Chem., 1963,35,985.
88. J. H. Yoe and J. J. Kirkland, ibid., 1954,26,1335.
89. J. C. Van Loon and F. E. Beamish, ibid., 1964,36,1771.
90. H. Zachariasen and F. E. Beamish, ibid., 1962,34,964.
91. A. Westland and F. E. Beamish, Micro&m. Acta, 1957,5,40.
92. G. G. Tertipis and F. E. Beamish, Anal. Chem., 1962,34,108.
93. B. R. Sant and F. E. Beamish, ibid., 1961,33,304.
94. H. G. Cobum, F. E. Beamish and C. L. Lewis, Anal. Chem., 1956,28,1297.
95. M. E. V. Phunmer, C. L. Lewis and F. E. Beamish, ibid., 1959,31,254.
96. M. E. V. Plummer, J. M. Kavanagh, J. C. Hole and F. E. Beamish, Trcuzr.AIME, 1961,221,145.
97. J. G. Sen Gupta and F. E. Beamish, Anal. C&m., 1962,34,1761.
98. J. M. Kavanagh and F. E. Beamish, ibid., 1960,32,490.
99. J. G. Sen Gupta and F. E. Beamish, Am. Mineralogist, 1963,18,379.
100. J. C. Van Loon and F. E. Beamish, Anal. Chem., 1964,36,872.
101. L. M. Banbury and F. E. Beamish, Z. Anal. Chem., 1965,211, 178.
102. Idem, ibid., 1966, 218,263.
103. K. C. AgrawaI and F. E. Beamish, ibid., 1965,211,265.