platinum metals review · platinum metals rev., 2006, 50, (3) 114 [pt(nh 2 eh) 4][ptcl 4] with a...

E-ISSN 1471–0676

PLATINUM METALS REVIEWA Quarterly Survey of Research on the Platinum Metals and

of Developments in their Application in Industrywww.platinummetalsreview.com

VOL. 50 JULY 2006 NO. 3

ContentsElectro-Spun, Semiconducting, Oriented Fibres of 112

Supramolecular Quasi-Linear Platinum CompoundsBy Margherita Fontana, Walter Caseri and Paul Smith

Crystallographic Properties of Platinum 118By J. W. Arblaster

Launch of the Low Carbon and Fuel Cell 119Knowledge Transfer Network

By M. Hugh

The Minting of Platinum Roubles: Part IV 120By Thilo Rehren

Centenary of the Discovery of Platinum 130in the Bushveld Complex

By R. Grant Cawthorn

Phoscorite-Carbonatite Pipe Complexes 134By Juarez Fontana

“Platinum 2006” 143

Thermophysical Properties of Palladium 144By Claus Cagran and Gernot Pottlacher

Abstracts 150

New Patents 154

Final Analysis: Mercury as a Catalyst Poison 156By J. K. Dunleavy

Communications should be addressed to: The Editor, Barry W. Copping, Platinum Metals Review, [email protected]; Johnson Matthey Public Limited Company, Orchard Road, Royston, Hertfordshire SG8 5HE

In their solid state, some metal complexesassume quasi-one-dimensional structures compris-ing linear arrays of metal atoms (mostly present ascations) (1–3). The anisotropic nature of suchmaterials renders them attractive for their opticaland electrical properties (1–3). Among these com-pounds, the most widely investigated have beenMagnus’ green salt (4, 5), [Pt(NH3)4][PtCl4], and itsderivatives of the type [Pt(L1)4][Pt(L2)4] (3, 6–8),

where L1 denotes a neutral ligand and L2 denotesan anionic ligand or a corresponding part of a mul-tidentate ligand, respectively. These speciescomprise a supramolecular arrangement of the pla-nar [Pt(L1)4]2+and [Pt(L2)4]2– units which arestacked alternately (Figure 1). Notably, the align-ment of the platinum atoms in these compounds isa consequence of the electrostatic forces betweenthe alternately charged coordination units (7, 9).

112Platinum Metals Rev., 2006, 50, (3), 112–117

DOI: 10.1595/147106706X128412

Electro-Spun, Semiconducting,Oriented Fibres of SupramolecularQuasi-Linear Platinum CompoundsOPTICAL AND ELECTRICAL PROPERTIES OF MAGNUS’ GREEN SALT DERIVATIVES

By Margherita Fontana, Walter Caseri* and Paul SmithDepartment of Materials, ETH Zürich, CH-8093 Zürich, Switzerland; *E-mail: [email protected]

The semiconducting Magnus’salt derivatives [Pt(NH2eh)4][PtCl4] and [Pt(NH2dmoc)4][PtCl4],with eh = (R)-2-ethylhexyl and dmoc = (S)-3,7-dimethyloctyl, are compounds that exhibit asupramolecular structure comprising a backbone of linear arrays of platinum atoms. Thesecompounds behave essentially like ordinary polymers. In this work they were processed intofibres by electrospinning from organic solvents such as toluene. X-ray diffraction patternsindicated that the platinum arrays in the fibres were oriented parallel to the axis of thefibres. Accordingly, the fibres show anisotropic optical and electrical properties. The electricalconductivities observed along the fibre axis were 2 × 10–5 S cm–1 for [Pt(NH2dmoc)4][PtCl4]and 7 × 10–7 S cm–1 for [Pt(NH2eh)4][PtCl4]. These exceeded the values for the respective bulkcompounds by 2–3 orders of magnitude, in agreement with comparable observations in orientedsemiconducting organic polymers.

Fig. 1 Chemical structure ofMagnus’ green salt and some of itsderivatives (eh = (R)-2-ethylhexyl;dmoc = (S)-3,7-dimethyloctyl)

These electrostatic forces, as well as crystal packingeffects, determine the interplatinum distances thattypically amount to 3.1–4.0 Å (7, 10–12). In com-pounds with Pt–Pt distances shorter than ca. 3.5 Å,the orbitals of adjacent platinum atoms overlapsignificantly. As a consequence these compoundsbecome semiconductive.

Processing of compounds derived fromMagnus’ salt has hitherto been cumbersome, if notimpossible, because these complexes are largelyinsoluble and do not melt prior to decomposition.Only recently have solution-processible Magnus’salt derivatives of the type [Pt(NH2R)4][PtCl4], withR being an alkyl group, been synthesised (7, 9, 13,14). The side groups R not only favour solubility inorganic solvents, but also influence the interplat-inum distances, and therefore dictate the physicalproperties of the corresponding bulk materials.For example, the complex with R = octyl, having arelatively large Pt–Pt distance (ca. 3.6 Å), wasfound to be an electrical insulator, whereas thosewith R = (R)-2-ethylhexyl (eh) and R = (S)-3,7-dimethyloctyl (dmoc), by contrast, weresemiconductors as a consequence of their relative-ly short Pt–Pt spacings (< 3.3 Å). Importantly, thesoluble [Pt(NH2R)4][PtCl4] derivatives could beprocessed like common polymers (14–16),enabling, for instance, the manufacture of fibres bystandard technologies. In the present work, wehave explored the use of a process known as elec-trospinning to produce filaments of selectedMagnus’ salt derivatives.

Conducting Fibres of Magnus’ SaltDerivatives

Electrospinning (17–20) is a relatively oldprocess (dating from the late 1910s), which hasprovoked renewed interest for the preparation ofpolymer fibres in recent years. In this technique,fibres are produced from polymer solutions underthe action of electrostatic forces. Upon charging apolymer solution, the electric forces at the liquidsurface can overcome the surface tension, resultingin a jet of electrically charged solution that is eject-ed towards an oppositely charged collector. If thesolvent concomitantly evaporates, then polymerfibres accumulate at the collector. We have already

reported the preparation of fibres by electrospin-ning of the insulating complex [Pt(NH2R)4][PtCl4]with R = octyl (14). The production of these fibreswas supported by the particular phase behaviour ofthese Magnus’ salt derivatives. These compoundsformed thermoreversible gels even at low concen-trations of the Pt-complex, which stronglyfacilitated the spinning process. However, theresulting fibres are of only modest interest due totheir insulating nature. Hence, in the following,emphasis is placed on the preparation and charac-terisation of fibres of semiconducting[Pt(NH2R)4][PtCl4] compounds with R = dmocand eh. This treatment significantly expands on aprevious brief sketch on R = dmoc filaments (16).

Electrospinning of fibres of previously synthe-sised [Pt(NH2R)4][PtCl4] (R = eh or dmoc) wasperformed from solutions of the compound intoluene, using a simple laboratory setup (Figure 2).In principle, it is possible to make very long fibreswith this method by using a rotating metal cylinderinstead of a metal plate as the negative electrode.Process parameters for the experiment (includingoptimal conditions) are given in Table I.

Uniform fibres of [Pt(NH2dmoc)4][PtCl4] wereproduced, with lengths of 1–5 mm and diametersranging from 0.1 μm to 2 μm. The large differencebetween the lengths of fibres obtained for the twocompounds is attributed to the shorter Pt chainlength of the [Pt(NH2dmoc)4][PtCl4] compound intoluene solution, compared with that for[Pt(NH2eh)4][PtCl4] (9, 13). Fibres of

Platinum Metals Rev., 2006, 50, (3) 113

Fig. 2 Schematic illustration of the simple laboratorysetup for the creation of filaments by electrospinning


[Pt(NH2eh)4][PtCl4] with a diameter of about 1 μmcould be formed into a loop of 20 mm cross-sec-tion (Figure 3). This implies that the fibrespossessed a considerable flexibility and resistanceto bending stresses. These particular fibres broke

as the loop was further tightened. Unfortunately,the [Pt(NH2dmoc)4][PtCl4] fibres were too short tocarry out this experiment on them.

Both types of fibres displayed notable birefrin-gence in the polarising optical microscope. Forfibres observed between crossed polarisers, lightwas transmitted at maximum intensity at an angleof 45º between the polarisers and the axis of thefibres. The transmitted light intensity was drasti-cally reduced when the fibre’s axis was parallel orperpendicular to one of the polarisers, indicatingthat the quasi-one-dimensional structures wereindeed oriented within the filaments along the fil-ament axis. Molecular orientation of thePt-structures within [Pt(NH2eh)4][PtCl4] fibres wasconfirmed by strong equatorial arcs in wide-angleX-ray diffraction pattern from a bundle of fibres(Figure 4). The average orientation angle derivedfrom the half-width at half-maximum intensity ofthe major equatorial reflections (21) was found tobe ~ 13º. This value indicates a low degree of ori-entation of the Pt-fibres as compared with that of,for instance, liquid crystalline polymer fibres suchas aramids, high-performance polyethylene andpoly(hexyl isocyanate) (PHIC) (21). Nevertheless,the orientation of the Pt-fibres is still significantconsidering that the organisation of the Pt-struc-tures within the fibres is governed by electrostaticforces.

A tilt compensator (MgCl2 crystal) revealed apositive sign for the birefringence of

Table I

Process Parameters for Electrospinning of [Pt(NH2R)4][PtCl4] Fibres

Parameter Range Optimal processing

Solution temperature, ºC Room temperature to 70 Room temperature

Compound concentration, 30–50 (R = dmoc) 45 (R = dmoc)% w/w 15–35 (R = eh) 30 (R = eh)

Applied voltage, kV 1–20 10 (R = dmoc)7–10* (R = eh)

Distance between tip of glass 3–15 12vessel and ground plate, cm

Angle of inclination between 0–45 30glass vessel and ground plate, º

Fig. 3 Optical micrograph of a [Pt(NH2eh)4][PtCl4]fibre taken between crossed polarisers. The doublearrows indicate the configuration of the polarisers. Theintensity minima of the fibre parts oriented parallel toone of the polarisers and the intensity maxima of thefibre parts oriented at a 45º angle to the polarisers indi-cate that the platinum compound is oriented within thefilament

* For R = eh, long, thin fibres were obtained when the voltage was gradually increased from 7 to 10kV during processing


[Pt(NH2dmoc)4][PtCl4] and [Pt(NH2eh)4][PtCl4],demonstrating that the refractive index was greaterin the direction of the fibre’s axis than perpendic-ular to it. Interestingly, at higher magnifications,alternating bright and dark parallel bands appearedwhen [Pt(NH2eh)4][PtCl4] fibres were examinedbetween crossed polarisers (Figure 5). These bandswere separated by 10–15 μm and were perpendic-ular to the fibre’s axis. These rather strikingfeatures strongly resemble those found in aramidfibres, for which their occurrence has beenexplained by a pleated sheet structure (22).

For reasons as yet unknown to us, such bandswere not observed in [Pt(NH2dmoc)4][PtCl4]fibres. Scanning electron microscopic studiesrevealed that the surfaces of the[Pt(NH2dmoc)4][PtCl4] and [Pt(NH2eh)4][PtCl4]fibres were relatively smooth (Figure 6).

The bulk electrical conductivity along the axisof [Pt(NH2dmoc)4][PtCl4] fibres amounted to 2 ×10–5 S cm–1, and for [Pt(NH2eh)4][PtCl4] to 7 × 10–7

S cm–1. These values exceeded those for therespective bulk compounds by approximately 2 or3 orders of magnitude (1.6 × 10–7 S cm–1 and 7 ×10–10 S cm–1, respectively). Similar enhancementswith orientation have also been observed for(semi-)conducting organic polymers (23). This is

consistent with the assertion that the platinumchain structures in the filaments are oriented pref-erentially, with the quasi-one-dimensional axisparallel to the fibre’s axis, since the platinum arraysare the principal conduction path.

Finally, in order to explore whether metallicplatinum fibres could be produced by degradationof the solution-spun Pt-compounds, a number offibres were subjected to plasma etching under anoxygen atmosphere for 20 min and 1 h periods,conducted in a plasma chamber. Interestingly, the

Fig. 4 Wide-angle X-ray diffraction pattern of a bundleof oriented [Pt(NH2eh)4][PtCl4] fibres. The double arrowindicates the direction of the fibres

Fig. 6 SEM image of [Pt(NH2eh)4][PtCl4] fibres pro-duced by electrostatic spinning from solution in toluene

Fig. 5 Optical micrograph of a [Pt(NH2eh)4][PtCl4]fibre between crossed polarisers displaying bands whichare indicative of a pleated sheet structure. The doublearrows indicate the configuration of the polarisers


macroscopic shape of the fibres was preserved andthe samples could still be handled after this plasmatreatment. SEM analysis of the morphology of[Pt(NH2eh)4][PtCl4] fibre surfaces after 20 minplasma treatment revealed a surface structure of acharacteristic length of 100–200 nm (Figure 7(a)).Upon exposure of 1 h, the structure coarsened andappeared to increase in density (Figure 7(b)). Thisseries of samples showed a weight loss of 25–30%

after 20 min plasma treatment and of ca. 50% after1 h. After 5 h plasma treatment the fibres had bro-ken into several parts and could not be handledfurther. Considering that the initial platinum con-tent of [Pt(NH2eh)4][PtCl4] is 37.2% w/w, it wasunfortunate that plasma exposure for 1 h did notyield pure elemental platinum, which would haveenabled the fabrication of platinum fibres by sim-ple processing of a precursor. Consistently, theplasma-treated fibres did not exhibit electrical con-ductivities in the metallic range.

ConclusionsLike common organic polymers, the quasi-one-

dimensional platinum compounds[Pt(NH2eh)4][PtCl4] and [Pt(NH2dmoc)4][PtCl4]can be processed to fibres by electrospinning fromsolution. The thickness and length of the resultingthin crystalline fibres depend on various processparameters such as solution concentration of theplatinum compounds, applied voltage, and the dis-tance between the polymer solution and thecollector. The longest fibres (up to 30 cm) wereobtained with [Pt(NH2eh)4][PtCl4]. The fibres wererather flexible and allowed the formation of loops.X-ray diffraction patterns indicated that the linearsupramolecularly arranged coordination units wereoriented along the axis of the fibres, which,accordingly, showed highly anisotropic opticalproperties. The fibres showed electric conductivi-ties in the semiconductor range. Gratifyingly, theconductivities along the axis of the fibres exceed-ed the values for the non-oriented bulk materialsby 2–3 orders of magnitude.

Given that films of [Pt(NH2dmoc)4][PtCl4] withoriented fibres are capable of functioning as theactive semiconducting layer in, for example, FETs,Magnus’ salt derivatives might possibly pave theway for mass-produced “plastic electronics” (16).Other potential applications include environmen-tally stable semiconducting fibres (8, 16).

AcknowledgementsThe authors are grateful to A. P. H. J.

Schenning (Technical University Eindhoven, TheNetherlands) for supplying the dmoc and to M.Müller (ETH Zürich) for the SEM studies.

Fig. 7 SEM image of a [Pt(NH2eh)4][PtCl4] fibre after(a) 20 min and (b) 1 h etching with an oxygen plasma

The Authors

Margherita Fontana is presently workingat Ciba Speciality Chemicals in Basel asa Laboratory Head. She is leadingprojects in the development of colourmaterials for E-paper displayapplications based on electrophoreticparticles as well as electrochromicmaterials. She is also involved in thecharacterisation of semiconductingmaterials and related devices.

Walter Caseri has been active as aSenior Scientist in the Institut fürPolymere at the ETH Zürich,Switzerland, since 1996. He is involvedin research and teaching. His interestsare in polymers containing both organicand inorganic components (polymericstructures with inorganic backbones,nanocomposites and polymers atinorganic interfaces).

Paul Smith has been Professor ofPolymer Technology at the ETH Zürichsince 1995. His interests lie in thedevelopment of advanced polymermaterials, polymer phase behaviour,mechanical properties of polymersystems, electrically and optically activepolymers, and polymer/metal systems.


1 K. Carneiro, in “Electronic Properties of InorganicQuasi-One-Dimensional Compounds”, Part II(Experimental), ed. P. Monceau, D. ReidelPublishing Co., Dordrecht, 1985, p. 1

2 S. Kagoshima, H. Nagasawa and T. Sambongi,“One-Dimensional Conductors”, Springer,Heidelberg, 1988

3 L. V. Interrante, ‘Electrical Property Studies ofPlanar Metal Complex Systems’, in “InorganicCompounds with Unusual Properties”, ed. R. B.King, Advances in Chemistry Series, No. 150,American Chemical Society, Washington, DC, 1976,p. 1

4 G. Magnus, Pogg. Ann., 1828, 14, 2395 G. Magnus, Ann. Chim. Phys. Sér. 2, 1829, 40, 1106 J. S. Miller and A. J. Epstein, Prog. Inorg. Chem., 1976,

20, 17 J. Bremi, V. Gramlich, W. Caseri and P. Smith, Inorg.

Chim. Acta, 2001, 322, (1–2), 238 W. Caseri, Platinum Metals Rev., 2004, 48, (3), 919 M. Fontana, H. Chanzy, W. R. Caseri, P. Smith, A.

P. H. J. Schenning, E. W. Meijer and F. Gröhn,Chem. Mater., 2002, 14, (4), 1730

10 L. V. Interrante and R. P. Messmer, Inorg. Chem.,1971, 10, (6), 1174

11 J. R. Miller, J. Chem. Soc., 1965, 71312 M. L. Rodgers and D. S. Martin, Polyhedron, 1987, 6,

(2), 225

13 J. Bremi, W. Caseri and P. Smith, J. Mater. Chem.,2001, 11, (10), 2593

14 J. Bremi, D. Brovelli, W. Caseri, G. Hähner, P. Smithand T. Tervoort, Chem. Mater., 1999, 11, (4), 977

15 M. G. Debije, M. P. de Haas, T. J. Savenije, J. M.Warman, M. Fontana, N. Stutzmann, W. R. Caseriand P. Smith, Adv. Mater., 2003, 15, (11), 896

16 W. R. Caseri, H. D. Chanzy, K. Feldman, M.Fontana, P. Smith, T. A. Tervoort, J. G. P.Goossens, E. W. Meijer, A. P. H. J. Schenning, I. P.Dolbnya, M. G. Debije, M. P. De Haas, J. M.Warman, A. M. Van De Craats, R. H. Friend, H.Sirringhaus and N. Stutzmann, Adv. Mater., 2003, 15,(2), 125

17 A. Formhals, US Patent 1,975,504; 193418 J. Zeleny, Phys. Rev., 1917, 10, (1), 119 J. Doshi and D. H. Reneker, J. Electrostat., 1995, 35,

(2–3), 15120 R. Dersch, T. Liu, A. K. Schaper, A. Greiner and J.

H. Wendorff, J. Polym. Sci. A: Polym. Chem., 2003, 41,(4), 545

21 A. R. Postema, K. Liou, F. Wudl and P. Smith,Macromolecules, 1990, 23, (6), 1842

22 H. H. Yang, “Kevlar Aramid Fiber”, Wiley,Chichester, 1993

23 K. Kaeriyama, in “Handbook of OrganicConductive Molecules and Polymers”, ed. H. S.Nalwa, Wiley, Chichester, 1997, Vol. 2, p. 271

References


DOI: 10.1595/147106706X129088

Crystallographic Properties of PlatinumNEW METHODOLOGY AND ERRATUM

By J. W. ArblasterColeshill Laboratories, Gorsey Lane, Coleshill, West Midlands B46 1JU, U.K.; E-mail: [email protected]

Equations are given to represent the lattice parameter thermal expansion of platinum from293.15 K to the melting point at 2041.3 K. This treatment is intended to supersede a combinationof dilatometric equations with corrections for thermal vacancy effects.

In the review of the crystallographic propertiesof platinum by the present author (1), the high-temperature data were represented by expressionsderived from precision dilatometric thermal expan-sion measurements (Equations (i) and (ii)). Above1000 K temperature, not only did length changemeasurements derived from lattice parameter mea-surements fail to agree with one another, they alsoshowed marked scatter around the dilatometricresults. The length change measurements weretherefore unsuitable for calculating the lattice para-meter thermal expansion. This problem wasaddressed by correcting the dilatometric data forthermal vacancy effects (Equations (iii) and (iv)),based on the consistent set of thermal vacancyparameters given in Table I and explained in theoriginal review (1).

On reflection, this procedure is cumbersome

and might be considered unsatisfactory. It hastherefore been replaced here by Equations (v) and(vi) which are based on a combination ofEquations (i) and (iii), and which represent the lat-tice parameter thermal expansion from 293.15 Kto the melting point at 2041.3 K. Equation (v)agrees with a combination of Equations (i) and (iii)

High Temperature Dilatometric Thermal Expansion (293.15–2041.3 K)

α* = 7.08788 × 10–6 + 1.04970 × 10–8 T – 2.00846 × 10–11 T2 + 2.28200 × 10–14 T3

– 1.18453 × 10–17 T4 + 2.37348 × 10–21 T5 K–1 (i)

δL/L293.15 K = 7.08788 × 10–6 T + 5.24850 × 10–9 T2 – 6.69487 × 10–12 T3

+ 5.70500 × 10–15 T4 – 2.36906 × 10–18 T5 + 3.95580 × 10–22 T6 – 2.39745 × 10–3 (ii)

Thermal Vacancy Corrections (1300–2041.3 K)

α*(lattice) = α*(dilatometric) – (5841/T2) e(1.32 – 17523/T) K–1 (iii)

δa/a293.15 K = δL/L293.15 K – (1/3) e(1.32 – 17532/T) (iv)Note: In Equation (x) of Ref. (1), to which Equation (iv) corresponds, the second δ was incorrectly given as d.

High Temperature Lattice Parameter Thermal Expansion (293.15–2041.3 K)

α* = 7.03139 × 10–6 + 1.08937 × 10–8 T – 2.10071 × 10–11 T2 + 2.36623 × 10–14 T3

– 1.20728 × 10–17 T4 + 2.34219 × 10–21 T5 K–1 (v)

δa/a293.15 K = 7.03139 × 10–6 T + 5.44686 × 10–9 T2 – 7.00236 × 10–12 T3

+ 5.91557 × 10–15 T4 – 2.41456 × 10–18 T5 + 3.90366 × 10–22 T6 – 2.39164 × 10–3 (vi)

Table I

Thermal Vacancy Parameters for Platinum

Parameter Symbol Value

Thermal vacancy concentration cv 7 × 10–4

at melting point

Enthalpy of monovacancy formation Hvf 1.51 eV

Entropy of monovacancy formation Svf 1.32k

Note: k is the Boltzmann constant, given at the time ofpublication of Reference (1) as 8.617385 × 10–5 eV K–1

to within 4 × 10–9 K–1 and to within ± 2 × 10–9 K–1

overall, well within the accuracy of Equation (i) of± 2 × 10–8 K–1.

In Equations (i), (iii) and (v), α* is the thermalexpansion coefficient relative to 293.15 K.

ErratumIn the review (1), equations were given repre-

senting a precision relationship between thermalexpansion and specific heat. However, the third

equation on page 19 of (1) (at the top of the right-hand column) was incorrectly given. It should haveread:

n

α = Cp( A + BT + Σ Cj T – j )j = 1


Reference1 J. W. Arblaster, Platinum Metals Rev., 1997, 41,

(1), 12

On the 25th May, 2006, Fuel Cell Today(www.fuelcelltoday.com), along with its partnersCENEX (the U.K.’s newly formed Centre ofExcellence for Low Carbon and Fuel CellTechnologies), Fuel Cells UK and ForesightVehicle, announced the launch of the LowCarbon and Fuel Cell KnowledgeTransfer Network (LCFC-KTN).

This new development, designedto enhance the U.K.’s competitiveposition in emerging clean energytechnologies, was instigated by theDepartment of Trade and Industry.The Network was launched simulta-neously in Yokohama, Japan, at the Japan Societyof Automotive Engineers congress.

Fuel Cell Today and the other KTN partnershave combined their specialist knowledge tocover broad aspects of sustainable transportation(www.low-carbon-ktn.org.uk) and the full com-plement of commercial opportunities for fuelcells, from portable battery replacement throughto power generation and transport applications.A principal aim is to accelerate the developmentand deployment of fuel cells in the U.K.

The KTN will provide a range of services tothe U.K. low carbon and fuel cell communityincluding a dedicated website, Business toBusiness facilities, networking opportunities,online conferencing, briefing notes, and expertopinions on technology and policy. The launch

of the KTN is timely. As the commercial phaseof fuel cell development gets underway, the U.K.fuel cell community now has a real opportunityto influence domestic and even worldwide mar-kets.

The advent of the Low Carbon and Fuel CellTechnology KTN is evidence that theU.K. Government is reshaping itsapproach to boosting U.K. fuel cellindustry capabilities and competitive-ness in line with broader internationalindustry trends. Currently, the U.K.does not sit with the United States,Japan, Canada and Germany in the

first tier of international fuel cell development,but it has an undeniable depth of expertise whichbodes well for the future. The U.K. also hassome of the most innovative companies in thebusiness. With effort, and continued Govern-ment support, the U.K. might yet take a place atthe top table.

The Fuel Cell KTN website can be viewed at:www.fuelcellktn.com. For further informationon this KTN and the services which it offers,contact: [email protected].

M. HUGH

Mike Hugh is the Moderator of the Fuel Cell TechnologyKnowledge Transfer Network website. He is interested in fuelcell paths to market and the corresponding policy process.He is on the staff of Fuel Cell Today, and his Ph.D. thesisfocused on drivers and barriers for stationary fuel cellmarkets in the U.K.

Launch of the Low Carbon and Fuel CellKnowledge Transfer Network

DOI: 10.1595/147106706X129853

The first half of the 19th century was a crucialperiod in the discovery and metallurgical study ofplatinum and its allied metals, iridium, osmium,palladium, rhodium and ruthenium. Only platinumwas known in 1800, but all six were known by 1844(4–7). The subsequent development of their refin-ing and production processes is not well known,probably due to commercial secrecy.Contemporary ‘best practice’ reports, for instanceby Sobolevsky’s Russian Royal Mint in St.Petersburg (8), are therefore not necessarily com-prehensive or reliable in their technical detail.

When significant deposits of platinum were dis-covered in the Ural mountains, the Russianauthorities, and in particular the then Minister ofFinance, Count Egor F. Kankrin wanted to use itfor coinage along with gold and silver denomina-tions. The value ratio between the three metals wasset at 15.6:5.2:1 for gold, platinum and silver,

respectively. Large-scale platinum ore processingbegan following the decision in April 1828 to issueplatinum roubles. This was done at the Royal Mintin St. Petersburg, supervised by GeneralSobolevsky. A technically successful process usedabout 20 tonnes of platinum ore from 1828 to1845, striking more than 1.3 million 3 roublepieces (Figure 1), almost 15 thousand 6 roublepieces and 3474 12 rouble pieces, with a total plat-inum weight of 485,505 troy ounces(approximately 15.1 tonnes) (Reference (4), page247). The monetary side, however, was less suc-cessful. In 1845 the Russian governmentdemonetised the entire platinum coinage, whichwas sold to various European platinum refineriesfor reworking.

There was something of an ‘afterlife’ for theRussian platinum roubles when the Russian RoyalMint produced fresh coins (‘Novodels’) for collec-tors in the late 19th century. Officially struck, usingthe original dies, these are numismatically identicalto the original series. It may be difficult to distin-guish authentic early to mid-19th century coinsfrom the ‘Novodels’ by established numismatic cri-teria, particularly since the latter are typically inmint condition, and more likely to be found inmajor reference collections. A written provenanceis often missing, so a scientific protocol is requiredto distinguish ‘Novodel’ issues from monetary

120

The Minting of Platinum RoublesPART IV: PLATINUM ROUBLES AS AN ARCHIVE FOR THE HISTORY OF PLATINUM PRODUCTION

By Thilo RehrenInstitute of Archaeology, University College London, London WC1H 0PY; E-mail: [email protected]

This paper augments a series of articles on Russian roubles in this Journal (1–3) with asummary of recent research into the manufacturing history and materials characterisationof these coins. The results are not only significant for the identification of genuine roublesissued between 1828 and 1845, ‘Novodel’issues produced in the late 19th century, and outrightforgeries of the 20th century, but offer a fascinating insight into the difficulties encounteredat the time in the large-scale refining and processing of platinum metal. A range of instrumentalmethods have been used to elucidate the magnetic properties, chemical composition andlow density of genuine roubles, and to reveal their complex internal structure. The resultingnew insights into the historical practice of platinum metallurgy are unbiased by concernsabout industrial espionage, state secrets, and professional rivalry.

Platinum Metals Rev., 2006, 50, (3), 120–129

DOI: 10.1595/147106706X128890

Fig. 1 3 roublepiece, dated 1831,struck by the RoyalMint in St.Petersburg underthe supervision ofGeneral Sobolevsky(diameter 23 mm)

coins. A further complication is the existence of20th century forgeries, allegedly produced in theLebanon, and perhaps also elsewhere.

Recent literature (1–3, 9) has drawn attention toa valuable body of technical information. It is theaim of this paper to give an up-to-date, first-handaccount of key results from the ongoing researchof the author’s group into these coins, and to out-line the potential of early coins and medals forproviding direct and precise information regardingthe development of platinum metallurgy overmore than a century.

Early 19th Century PlatinumRefining

The early metallurgy of platinum was consider-ably hampered by its chemically and thermallyrefractory nature, and by the presence in the ore ofother, not easily separable elements (typicallyabout 25% in total of varying amounts of the otherplatinum group elements, plus iron and copper).Most proposed refining methods relied on dissolv-ing the ore in aqua regia (mixed concentratedhydrochloric and nitric acids), followed by selec-tive precipitation of platinum as ammoniumhexachloroplatinate. Careful washing was neededto remove as much of any coprecipitated iridiumand iron salts as possible without excessive loss ofplatinum. The dried precipitate was brought to redheat, driving off the ammonia and chlorine, andyielding a metallic platinum sponge. The spongewas then ground, forged and hammered, withrepeated annealing cycles. The result was a solidmetal which was forged into bars and sheets. Thedensity of the metal sponge increased progressive-ly during hammering to a maximum of around 21g cm–3, close to that of pure platinum at 21.45 gcm–3.

Over twenty years, Wollaston perfected therefining and working of platinum at a laboratoryscale to economic success (6). The Royal Mint inSt. Petersburg, on the other hand, pioneered indus-trial use of the early powder metallurgy, reportedlywith several refining variants along the way (4, 7, 9,10). The relative merits and efficacy of the variantsin terms of finished metal quality cannot be judgedfrom these publications alone.

Investigation of Russian PlatinumCoins

Bachmann and Renner’s (11) were the first ana-lytical results, based on scanning electronmicroscopy and X-ray fluorescence analysis on an1829 3 rouble piece. There was a significant degreeof porosity at the surface, as expected for materialproduced by powder metallurgy, and with 0.5 wt.%iron and about 0.1 wt.% each of palladium, rhodi-um and chromium present. The density of thematerial was 20.7 g cm–3. The authors record a vis-ible improvement of the surface quality of thecoins over the production period. There has beenno metallographic study, or discussion of the rangeand origins of impurities, until recently.

The present study used a series of 3 roublecoins from 1828 to 1842. It was prompted by theobservation of a magnetic moment and substan-dard density for most of them. Only the 1828 coinwas in mint condition; the other eight showed clearsigns of wear and circulation. A Russian Olympiccommemorative platinum coin struck in 1977 wasincluded to represent more recent metallurgicalstandards. Analysis was largely non-sampling andnon-destructive. Only one of the coins was sam-pled for metallographic study. Full analytical detailsand results have been published elsewhere (9,12–14 and literature cited therein). This papersummarises the results and addresses the coins’potential significance for the history of platinummetallurgy.

Material CharacterisationThe 3 rouble coins are inscribed with their

nominal weight of 2 zolotnik (zol.) 41 dolya (dol.),or 10.35 g, pure Ural platinum; the measuredweights vary from 10.35 g for the 1828 issue to lessthan 10.1 g for the 1837 coin (Table I). Densityvalues were scattered in the range 20.0 to 20.7 gcm–3; the 1828 issue had a density of 21.3 g cm–3.This agrees with contemporary values from theearly 19th century literature of about 20 to 21 gcm–3, while placing the 1828 issue close to the the-oretical value for pure platinum. Three of the coinsshowed a considerable response to an ordinaryhand-held magnet; a fourth could be lifted bodily;the other three showed no perceptible response,


among them the 1828 coin. It became evident thatthe uncirculated coin dated 1828 was probably alate-19th century ‘Novodel’ issue. This was usedthereafter as an internal standard for technicallypure platinum; its chemical characteristics aregiven below, together with those of the 1977 com-memorative issue.

Where coins showed a magnetic moment, thiswas quantified using a Förster deflectometer. Theinstrument was calibrated on the 1828 coin ratherthan on a sheet of pure platinum, so that the geom-etry of the reference piece would be identical tothat of the samples. The six other coins gave val-ues between 8 and 35 units – results of littlesignificance in themselves, but clearly not randomin the light of other observations (Table I).

From an initial qualitative chemical analysis byscanning electron microscopy with microanalysisby energy-dispersive X-ray spectroscopy (SEM-EDS), the only readily detectable impurities(chlorine and calcium among others) were proba-bly surface contaminants. Of likely primarycontaminants from the ore, only iron was detected.However, due to the extremely dense matrix, thedetection limits for this and other elements wererather high, precluding reliable quantification andinterpretation of the results. The peaks for otherelements such as gold and iridium were too closeto the dominant platinum peaks to be properlyresolved at low concentrations. Two series of X-ray fluorescence (XRF) analyses were performed,

both measuring approximately two thirds of thecoins’ surfaces. Iridium, gold and iron were detect-ed in most coins, followed by minor signals forcopper, nickel and occasionally zinc. The secondseries of analyses by energy-dispersive spec-troscopy identified iron and iridium as the maincontaminants, both at around 1% by weight, fol-lowed by copper and gold in the range 0.1–1%.Rhodium, palladium and nickel were typically pre-sent at around 0.1% or less. Elements such astitanium, zinc and tin rarely exceeded a few hun-dred parts per million (Table II).

The results for the obverse and reverse of eachcoin are typically very similar, but several coinsshowed much higher readings for some elementson one side only – for instance, copper and gold ataround 1% each on the reverse of the 1838 issue.This was consistent with macroscopically visiblegold specks on the coin’s surface. One of theseparticles proved to be a high-copper gold alloywith a low silver content, unlike natural goldnuggets which have a rather lower copper contentand a higher silver content. The 1844 3 roublepiece analysed by Lupton (2) gave a similar result.The 1837 coin studied here shows abnormally highlevels of nickel, silver and tin on its reverse, togeth-er with an elevated copper level.

Both the magnetic and chemical analyses indi-cated a significant presence of iron in these coins,but did not distinguish between mechanicallyincorporated iron-rich particles (either oxide or

Table I

Physical Properties of Seven 3 Rouble Platinum Coins

Year Weight, g Density, g cm–3 *Magnetic value Theoretical Fe content Impurities(XRD), wt.% (XRF), wt.%

1828** 10.351 21.32 0 0 0.71832 10.281 20.25 35 2.5 3.51835 10.165 20.15 8 0.4 1.81836 10.251 20.42 22 1.9 3.51837 10.076 20.03 21 1.2 4.11838 10.279 20.12 13 0.4 4.11842 10.311 20.69 17 1.2 3.1

Density determined by immersion in water. *Magnetic value is the dimensionless reading from the Förster deflectometer calibrated tozero on the 1828 issue. Weight per cent impurities are taken from Table II. **‘Novodel’ issue


metal) and iron alloyed to the platinum matrix. X-ray diffraction (XRD) was used to determine this.The XRD pattern for the 1828 ‘Novodel’ issueexactly matched the indexed peak positions ofpure platinum, demonstrating that the surfacemorphology of the coins had little or no influenceon the XRD pattern. The only peaks found werethose of metallic platinum, with the peak positionsfor all coins other than the ‘Novodel’ issue beingshifted to slightly higher 2θ values, that is, smallerlattice spacings. The atomic radius of iron in themetallic bond is about 10% smaller than that ofplatinum; hence, substituting iron for platinumwill decrease the cell parameters of the resultingalloy, as observed in the coins.

After Cabri and Feather (15), the deviation ofthe measured peak from the ideal was used to cal-culate a theoretical value for the iron content in

the alloy. Cabri and Feather’s Figure 3 gives therelationship between iron content and peak shiftfor chemically analysed binary platinum-ironalloys. The coins, however, are more complexmultielement systems with significant amounts ofiridium, rhodium, gold, copper, and possibly otherelements, as well as iron. All these elements willaffect lattice parameters, but the whole effect ishere ascribed to iron, that is, a binary system isassumed. Thus the calculated theoretical iron con-centrations represent the total effect of all alloyingelements rather than that of iron alone. However,one may be confident of the dominance of iron inthe peak shift since the atomic radii of other majorcontaminants are very similar to that of platinum,or they are present in concentrations much lowerthan that of iron. In particular, the atomic radii foriridium and gold are only slightly lower (iridium)


Table II

XRF Analyses of 3 Rouble Platinum Coins

Coin Ti Mn Fe Ni Cu Rh Pd Ag Sn Ir Au Sum, %

1829 110 < 20 12,300 60 2800 1100 1350 680 10 3500 1100 2.3

100 < 20 12,500 80 3000 1150 1270 900 < 5 3700 1050

1831 40 < 20 6700 250 1600 790 540 220 < 5 9400 1100 2.1

300 < 20 6900 230 1850 770 530 160 10 9200 850

1832 60 < 20 15,800 680 4200 1100 760 220 10 13,200 510 3.5

180 180 14,300 600 5100 1050 710 140 10 12,100 280

1835 140 < 20 5700 125 1200 520 210 240 30 9200 630 1.8

240 620 5200 140 1050 610 260 200 30 8400 1000

1836 140 100 13,800 470 3100 1400 610 215 60 14,100 220 3.5

280 480 14,500 410 4200 1400 590 475 20 13,300 190

1837 50 190 10,600 440 5900 870 570 195 15 14,100 550 3.6

25 180 13,700 1540 7300 890 550 1180 850 10,800 430

1838* 530 < 20 11,800 340 2850 760 290 100 60 12,200 1100 4.1

280 < 20 11,300 440 10,300 760 280 825 80 15,100 13,200

1842 50 < 20 7000 340 2900 1500 1380 220 30 16,700 950 3.1

160 70 5100 230 2250 1650 2000 190 75 18,700 1050

1828** 520 90 6100 50 290 90 140 50 10 2250 < 300 0.7

< 20 < 20 130 < 20 260 85 140 50 5 2300 < 300

1977 < 20 < 20 480 < 20 110 60 140 < 5 < 5 < 300 < 300 < 0.2

< 20 < 20 650 < 20 170 50 90 < 5 < 5 < 300 < 300

Measured using a “Spectro X-LAB 2000” XRF spectrometer. All data in ppm. The upper row gives values for the obverse, the lowerrow for the reverse. Other metals were also found in most coins, up to hundreds of ppm, such as zinc. *The 1838 coin has gold specksvisible on its reverse. **The 1828 coin is a ‘Novodel’ issue, probably produced in the late 19th or early 20th century

or higher (gold) than that of platinum; thus, themain contribution to the peak shift is probably dueto iron and copper, which are both significantlysmaller in radius than platinum.

The copper concentration is known to be aboutone quarter of the iron concentration. Most of thecrystal lattice deformation is therefore due to theiron component, with possibly up to one quarterdue to the copper, and much less due to iridium.The maximum theoretical iron content calculatedfrom the XRD pattern is about 2 to 2.5 wt.% Fe,and can be as little as 0.5 wt.% (13). This agreesvery well with the iron content measured by XRFof around 0.5 to 1.5 wt.% (see Table I), with up toone quarter of the calculated shift ascribed to cop-per. In summary, the XRD pattern confirms asignificant presence of iron in the platinum lattice,as an alloying component in solid solution ratherthan as a mechanical impurity.

It is important to recall that, while eddy currentand density measurements test the entire coins,both XRF and XRD analyses only characterise thenear-surface parts of the coins, as the extremelydense matrix will have prevented any penetrationof the X-rays beyond a few tens of μm.

Metallographic InvestigationThe 1837 coin was chosen for metallographic

investigation of its interior, having shown thestrongest magnetic response to the hand-held mag-net and among the highest impurity content in theXRF analysis. However, its calculated iron contentaccording to XRD and eddy current readings wasonly moderate. Sampling was done with a slow-moving diamond-impregnated cutting wheel,removing a triangular cross-section from the rim.The sample was mounted in cold-setting resin, per-pendicular to the coin’s flat surfaces; then groundand polished by standard procedures down to aquarter micron diamond finish for optical andscanning electron microscopy.

Most striking under the optical microscope wasthe high density of tiny oxide inclusions through-out the body of the coin, but notably absentimmediately beneath the two main surfaces(Figures 2(a) and 2(b)). The inclusions consistedmainly of oxides of iron and nickel. Two distinct,adjacent oxide phases were observed, haematiteand magnetite, which are described in Table III.

The marked absence of metal oxide inclusionsnear the surfaces, and within certain layers in the


Table III

Oxide Inclusions Throughout Body of 1837 Coin

Oxide inclusion, identified by Colour Composition, determined byRaman spectroscopy (12) electron micro-probe analysis

Haematite Bluish, with intense Pure iron oxidered internal reflections

Magnetite Greyish 5 wt.% nickel oxide(iron-deficient)

Fig. 2(a) (Below) Optical micrograph of cross-section of1837 coin, showing oxide inclusions present throughoutits body, but notably absent beneath the two principalsurfacesFig. 2(b) (Right) As Fig. 2(a), but at higher magnifica-tion

body, was interpreted as a result of pickling of thehammered sheet metal during the process, beforeits being folded over for repeated hammering, andthen of pickling of the coin blanks after the finalannealing. This would have removed any oxidescale from the surface and the immediate sub-sur-face layer. Notably, there were no voids or poresin the body, nor were any grain boundaries appar-ent. However, in cross-section the outer surfacesdid show some imperfections and irregularities(e.g., Figure 2(a), top part of machined rim), com-parable to those observed earlier by Bachmannand Renner (11) using non-destructive SEM imag-ing

The clean near-surface layers and the inclusion-rich body were analysed separately by a scanningelectron microscope fitted with a wavelength-dis-persive spectrometer. Results are summarised inTable IV. The iron results were the most interest-ing, showing a clear tendency to higherconcentrations in the body than near the surface.This further corroborates the hypothesis of partialiron depletion of the surface metal by oxidationand leaching.

To elucidate the metallographic structure of theplatinum matrix, and to better understand the rela-tionship of the inclusions to the matrix, we turnedto the Johnson Matthey Technology Centre forhelp with etching. This was done in hydrochloricacid saturated with sodium chloride, and applyingan alternating current to the sample. Initially, theetching attacked the oxide inclusions only. Themetal grain structure which eventually became vis-ible was relatively coarse, with a clear elongation ofthe individual grains parallel to the flat surfaces ofthe coin (Figures 3(a) and 3(b)). The cycles of forg-ing and annealing, the latter often at hightemperatures and over many hours (4), prior to the

striking of the coins, had clearly obliterated the ini-tial structure of the metal sponge. Grains were onaverage about two orders of magnitude larger thanthe inclusions. No systematic spatial relationshipbetween the inclusions and the metal grains wasapparent.

Compositional Characterisation ofthe 1828 and 1977 Coins

Together with the six coins dated between 1829and 1842, two later coins not intended for circula-tion were analysed by XRF: the ‘Novodel’ issuelabelled 1828, but probably made in the late 19thor early 20th century, and the 1977 commemora-tive issue. Their density and XRD pattern weremuch closer to the theoretical values for pure plat-inum (see above). They gave no unusual magneticresponse. Their chemical (surface) composition isgiven in the bottom two rows of Table II. It isobvious that they are made from more highlyrefined platinum, with much lower levels of allcontaminants. Not unexpectedly, the 1977 coinshows the least contaminations, with about 550ppm iron as the major impurity, whereas the‘Novodel’ issue (Figure 4) still has more than 2000ppm iridium and iron, and a slightly higher copper


Table IV

Average Contents of 1837 Coin Regions

Region Average content, wt.%*Fe Ni Ir Rh

Near-surface layers 0.73 0.03 0.85 0.57Body of the coin 1.40 0.05 1.06 0.51

Fig. 3(a) Elongation of individual metal grains parallelto flat surfaces of 1837 coin, apparent after etching

Fig. 3(b) As Fig. 3(a), but at higher magnification

*From Ref. (9), p. 85

content than the 1977 issue. Notable is the almostcomplete separation of gold, first witnessed here(Figure 5).

The occurrence of anomalously high levels ofcertain elements on one surface only is not restrict-ed to the early coins; the ‘Novodel’ issue has on itsobverse much higher readings for titanium, ironand zinc. The origins of these and other similarhigh readings are not yet known, requiring moreresearch on both manufacture and subsequenttreatment.

Interpretation of the ResultsThe physical, chemical and metallurgical char-

acteristics of the coins relate directly to their

manufacture. The considerable contamination,both as discrete inclusion of iron oxide, and asalloying elements in the platinum, such as iron,iridium, copper, rhodium and gold, reflects thelimitations of the refining operation based onselective precipitation from aqua regia. The practi-tioners of the time were well aware of the necessityto rinse the precipitate sufficiently to remove asmuch of the iron as possible, but not so as to losetoo much platinum.

The methods available to separate iridium andplatinum were all laborious and far from quantita-tive, so it is unsurprising that iron and iridium arethe main contaminants in the coins, with concen-trations of around 1 wt.% each. Other elements,such as rhodium, palladium, copper and gold,accompany platinum into the precipitate to someextent. Iron was probably the most deleteriouscontaminant, at higher concentrations renderingthe metal too hard and brittle for successful forg-ing and striking. The upper limit of tolerability ofiron contamination obviously depended stronglyon the intended use of the metal, while the purityattainable depended on the skills and experience ofthe practitioners as much as on the quality ofreagents and tools available.

The research and development which eventual-ly made platinum workable centred on mechanicaltreatment as much as on the refining procedure(4). Even Wollaston, the undisputed authority onrefining and working platinum in the first decadesof the 19th century, was not able to obtain plat-inum metal free of iron; after each forging andannealing cycle he found iron scales which had tobe removed by pickling before resuming the treat-ment (16). Analysis of platinum wire made byWollaston, and now held in the Science Museum,London, found about 0.35 wt.% iron and 0.2 wt.%iridium in the metal (10).

Metallographic investigation of one of the coinsdemonstrated that the iron is partly present as ironoxide particles within the body of the coin, andpartly as an alloying element in the platinummatrix. For wire drawing, both oxide inclusionsand the hardening effect of the iron and iridiumalloy component would have been detrimental,requiring purer platinum than for coin minting.


Fig. 5 The1838 coinhas a macro-scopicallyvisible goldspeck on itssurface(lower right;adjacent todate)

Fig. 4 X-ray fluorescence spectra for (top) ‘Novodel’issue and (bottom) 1837 coin

Fe Cu

Pt

Pt

Pt

Pt Pt

Pt

Pt

Fe

Fe NiCu

Pt Zn

Pt Pt Pt

Pt

Pt

IrPt

Zn

6 7 8 9 10 11 12Energy, keV

50.040.532.024.518.012.5

8.04.52.00.5

0

Inte

nsity

, kC

PS

6 7 8 9 10 11 12Energy, keV

50.040.532.024.518.012.5

8.04.52.00.5

0

Inte

nsity

, kC

PS

Ir

Cu Ir

Cu

This observation is borne out in the different lev-els of residual contamination in the metal ofWollaston’s wire and the Russian coins.

The degree of contamination in the eight gen-uine coins analysed here falls in the range 2 to 4wt.% of combined impurities near the surface;there is a further impurity contribution from theoxide inclusions within the coin matrix. The threecoins with the lowest impurity levels are thoseminted in 1829, 1831 and 1835, that is, during thefirst half of the period of coin production. Iron,the element most critical to the malleability ofrefined platinum, shows no decrease during theseyears. Iridium levels apparently increase slightly.This absence of a consistent trend suggests thatrefining practice remained basically largelyunchanged, resulting in a platinum content ofprobably only around 95 to 97 wt.% for the gen-uine coins.

Looking at the data reveals a positive correla-tion in the concentrations of the three platinumgroup metals (pgms): iridium, rhodium and palla-dium, with the highest value for all three elementsfound in the 1842 coin, and generally low values inthe 1835 issue. Iron, on the other hand, follows adifferent pattern, reasonably well correlated withcopper. This probably reflects the different behav-iour of these elements during refining andmanufacture; we may assume that they all co-var-ied throughout the precipitation of the platinumsponge and the subsequent washing steps. Onlycopper and iron are likely to burn out as oxidesduring the hot forging and pickling. The nature ofthe oxide inclusions – haematite and oxidisedmagnetite – clearly indicates aggressively oxidisingconditions during metal processing. This is con-ducive to the further removal of metallic iron fromthe alloy during hot forging.

It is known that the platinum ore wasprocessed in batches of around 10 to 15 kg per day(8). One may assume that the observed variabilityin coin composition reflects variability betweenbatches rather than systematic changes in practice.On the other hand, Sobolevsky (8) mentionsimprovements in the refining procedure followedat the Royal Mint in St. Petersburg, so a reductionof overall impurity levels among the coins over the

production period would thus not be unexpected.Schneider (17) reports that at some point the pro-cedure to separate iridium from platinum changedfrom using an initial excess of hydrochloric acid inthe solution, with selective precipitation of theplatinum, to selective precipitation of all pgmsother than platinum by adding limewater in dark-ness. It is plausible that the generally higher andmore constant levels of iridium in the second halfof the production period reflect this change. Thequality achieved early on was evidently fit for thepurpose, while the slight increase in impuritiesover time might even indicate that the mint mas-ters learned to cope with them. However, tofurther explore the issue of variability and trendsin composition, a much more comprehensiveseries of analyses is required, covering multiplecoins for each year, as well as more archivalresearch in Russia.

A clearer distinction emerges between the gen-uine and the later coins, which have total impuritylevels of less than one per cent (Table II, bottomtwo rows). Not unexpectedly, the 1977 coin showsthe least contaminations, with about 550 ppm ironas the major impurity. In contrast, the 1828‘Novodel’ issue still has more than 2000 ppm eachof iridium and iron, although most contaminantsare present in significantly lower concentrationsthan previously. The platinum content is betterthan 995/1000, if one ignores the possibly errati-cally high iron content on the coin’s obverse. The‘Novodel’ and 1977 issues demonstrate majorprogress in refining and manufacturing practicesover more than a century, first following the intro-duction of hydrogen-oxygen burners in 1847 andthen, in the 20th century, the introduction of elec-trochemical refining. The almost completeseparation of platinum and iridium, evident fromthe 1977 coin, leaves iron as the last major impuri-ty – present at more than just a couple of hundredppm – thus qualifying the platinum for “999” fine-ness.

The content of most impurities is below thedetection limit for typical EDS systems attached toscanning electron microscopes, and iron levelsalone are no reliable discriminator for genuinecoins versus ‘Novodel’ issues. Willey and Pratt (3),


for instance, report very low levels of iron in their1834 coin and use this to interpret this coin as aforgery. However, the EDS spectrum of that coinshows in addition to the iron peak a significantrhodium peak (cf. their figure on page 137 of (3)),unknown in other ‘Novodel’ issues but appearingin other genuine coins; also, its weight is only 10.16g, very much in line with the genuine coins studiedhere. This coin may have been made from a partic-ularly iron-poor batch of metal, but still within theperiod of genuine coin production. Lupton (2, p.78) found even lower iron levels in the two 3 rou-ble pieces analysed, of 0.3 and 0.4 wt.%,respectively; these coins had densities of around20.8 and 20.4 g cm–3 and sufficient other impuritiesto suggest that they were indeed genuine.

ConclusionsRussian platinum coins have been analysed to

characterise their physical and chemical properties.In conjunction with published information on therefining and working of platinum in general, and ofthese coins in particular, criteria have been estab-lished to distinguish authentic coins, issuedbetween 1828 and 1845, from later official reis-sues, known as ‘Novodels’. The density of theauthentic coins generally falls between 20 and 21 gcm–3, while later coins seem to have a density com-fortably above 21 g cm–3. Similarly, weights belowthe nominal 10.35 g seem to be common amonggenuine coins (Table I). A recognisable magnetismappears to be an indicator, though not a prerequi-site, for authenticity, as is a complex pattern ofspecific chemical impurities in the metal (14).

A full understanding of most of the observedphysical and chemical characteristics of the ninecoins studied here was only possible through themetallographic investigation of one of them (the1837 issue). Even allowing that one sample may bemisleading, we feel that the metallography ren-dered much more reliable the interpretation ofresults obtained by non-sampling and non-destruc-tive methods for all these coins.

Platinum refining during the 19th century reliedprimarily on a complex and then only partiallyunderstood sequence of dissolution and precipita-tion. The main criteria for the purposes of the

Royal Mint in St. Petersburg were the malleabilityof the resulting metal, to be balanced against theoverall costs of the operation, and manageability atan industrial scale. The analysis of the genuinecoinage of the first half of the century suggests thatimpurity levels were tolerable, particularly asregards iron and iridium. For both elements, typi-cal concentrations were found to be in the one percent range, clearly worse than in Wollaston’s con-temporary metal, refined at a laboratory scale (10).The relatively wide scatter in impurity concentra-tions found among the coins analysed so farindicates a degree of flexibility in refining practiceat St. Petersburg. The ‘Novodel’ issues are of aconsiderably higher purity than even the best gen-uine coins in terms of several critical elements,including gold, iridium, copper, nickel and iron.The present analysis uses only a single ‘Novodel’issue, so quantitative characterisation must be cau-tious. A marked increase in refining quality is,however, to be expected over the fifty yearsbetween the production of the original and the‘Novodel’ issues, and apparent in all four majorcontaminants, iron, iridium, copper and gold. TheRussian 1977 issue, analysed as an example of amodern use of platinum for commemorative coinsand medals, is almost pure platinum, with onlyminute transition metal concentrations. It is veryobviously different from the 19th century metal.

Future work should concentrate on characteris-ing the ‘Novodel’ issues more fully, both in theirchemical composition and physical properties suchas magnetic response, density and possiblymicrostructure. This would greatly improve ourdiscrimination between the two series, which areotherwise almost indistinguishable. It would be ofinterest to study the homogeneity within and vari-ability between individual metal batches of thegenuine coinage on a year-to-year basis. This couldshow whether any of the indicated changes in St.Petersburg’s refining procedures resulted in sys-tematic shifts in composition, if not in improve-ments in platinum fineness, or whether theobserved variability of the composition simplyreflects the variability of ore batches or individualbatch preparations, without a specific chronologi-cal trend in quality.


AcknowledgementsThe work presented here (begun while the

author was a research scientist at the GermanMining Museum in Bochum) could not have beencarried out without the cooperation, help andadvice of many friends and colleagues. Particularthanks are due, in chronological order, toEberhard Auer, from Wertheim; Alex von Bohlenand Reinhold Klockenkämper, Institut fürSpektrochemie und Angewandte Spektroskopie inDortmund; Andreas Ludwig and Dirk Kirchner of

the Deutsches Bergbau-Museum in Bochum;Hans-Gerd Bachmann, from Hanau; EckhardPappert, then Spectro Analytical, Wesel; AllinPratt of the Johnson Matthey Technology Centre;Kevin Reeves from the Wolfson ArchaeologicalScience Laboratories at the Institute of Archae-ology, University College London (UCL), andmost recently to Jaap van der Weerd and RobinClark from the Christopher Ingold Laboratories,Department of Chemistry, UCL, for all their prac-tical, analytical and academic contributions.


1 C. Raub, Platinum Metals Rev., 2004, 48, (2), 662 D. F. Lupton, Platinum Metals Rev., 2004, 48, (2), 723 D. B. Willey and A. S. Pratt, Platinum Metals Rev.,

2004, 48, (3), 1344 D. McDonald and L. B. Hunt, “A History of

Platinum and its Allied Metals”, Johnson Matthey,London, 1982, 450 pp

5 L. F. C. Vallvey, Platinum Metals Rev., 1999, 43, (1), 316 W. P. Griffith, Platinum Metals Rev., 2003, 47, (4), 1757 W. P. Griffith, Platinum Metals Rev., 2004, 48, (4), 1828 P. G. Sobolewsky, ‘Über das Ausbringen des Platins

in Rußland’, Ann. Pharm., 1835, 13, 429 E. Auer, Th. Rehren, A. von Bohlen, D. Kirchner

and R. Klockenkämper, 1998, Metalla, 5, (2), 7110 B. Kronberg, L. Coatsworth and M. Usselman,

‘Mass spectrometry as a historical probe:Quantitative answers to historical questions in met-allurgy’, in “Archaeological Chemistry III”, ed. J. B.Lambert, Advances in Chemistry Series, No. 205,American Chemical Society, Washington, DC, 1984,pp. 295–310

11 H.-G. Bachmann and H. Renner, Platinum MetalsRev., 1984, 28, (3), 126

12 J. van der Weerd, Th. Rehren, S. Firth and R. J. H.Clark, Mater. Charact., 2004, 53, (1), 63

13 Th. Rehren, D. Kirchner and E. Auer, ‘The metallur-gy of 19th century Russian platinum coins’, in“Founders, Smiths and Platers”, ed. P. Northover, inprint

14 Th. Rehren, E. Pappert and A. von Bohlen, ‘Thechemical composition of early Russian platinumcoins as an indicator of authenticity’, in “Metallurgyin Numismatics V”, ed. M. Cowell, Spink, London,in print

15 L. J. Cabri and C. E. Feather, Can. Mineral., 1975, 13,(2), 117

16 W. H. Wollaston, ‘Ein neues Verfahren schmied-bares Platin darzustellen’, Dinglers Polytech. J., 1829,34, 1

17 W. v. Schneider, ‘Über die technische Darstellungvon chemisch reinem Platin’, Dinglers Polytech. J.,1868, 190, 359

References

The Author

Dr-Ing habil. Thilo Rehren FSA is Professor ofArchaeological Materials and Technologies at theInstitute of Archaeology, University College London.His interest focuses on the study of the primaryproduction of metals and glass in prehistory andhistory, using materials science methods within anarchaeological and historical framework.

The story of the discoveries by Dr HansMerensky (Figure 1) of the platinum-rich pipes andthe Merensky Reef itself in 1924 has been well doc-umented (1). However, the events preceding thediscoveries have not been summarised. In theprobable centenary year of the first report of plat-inum in the Bushveld, it is appropriate to reviewthose events from 1906 to 1924.

Bushveld Platinum Reported on10th November 1906

In geology it is risky to claim a date for the“first” documentation of any event. However, it issuggested that for the occurrence of platinum inthe Bushveld Complex, this can reasonably be con-sidered to have been a report (2) by William Bettelon 10th November 1906 in an article in SouthAfrican Mines, Commerce and Industries, a weekly jour-nal then published in Johannesburg.

Platinum in South AfricaTo cover all possibilities concerning first dates,

a reference to “platina” (the old name for plat-inum) should be mentioned. A specimen, togetherwith assorted other geological samples, was dis-played on Church Square, Pretoria, by aprospector, Dick Hart. It was collected from anarea of ~ 130 km by 75 km around Pretoria. Theevent was recorded in the Pretoria newspaper DieVolkstem on 27th July 1885 (cited in (3), p. 52).There is no reason to doubt the prospector’s iden-tification (“platina” had little value then), or theprobability that it came from the BushveldComplex, but the display had no impact on themining community.

To return to Bettel: he was the chief chemist atthe Robertson gold mine in Johannesburg at thetime. His story begins in 1890 when he analysed a“black sand” concentrate from a stamp battery(used for crushing gold ore) from a gold mine inKlerksdorp, 100 km southwest of Johannesburg (amere four years after the first discovery of the goldreef in Johannesburg). Bettel found the concen-trate to contain “silver, gold, platinum and iridium(with osmium)”. Hence, the presence of the plat-inum group elements in South Africa in minoramounts was well established by the end of thenineteenth century.

In Situ PlatinumBettel stated in his article that he “recently” (i.e.

before November 1906) analysed half-a-dozen


DOI: 10.1595/147106706X119746

Centenary of the Discovery ofPlatinum in the Bushveld ComplexBy R. Grant CawthornSchool of Geosciences, University of the Witwatersrand, PO Wits, 2050, South Africa; E-mail: [email protected]

The earliest authenticated scientific report of the occurrence of platinum in rocks from theBushveld Complex, South Africa, appears to be that of William Bettel on 10th November 1906.Thereafter, prospecting of the chromite-rich rocks for platinum proved frustrating. It is arguedthat the resurgence of interest by Dr Hans Merensky in 1924 resulted from his realisation thatnewly panned platinum had a grain size different from that in the chromite layers and indicateda different source rock, which he promptly located as the Merensky Reef.

Fig. 1 Dr Hans Merensky,taken in 1917 atPietermaritzburg (7)

samples of chromite-bearing rock, which hedescribed as “olivine gabbro”, and had foundthem to contain platinum. He regarded this docu-mentation as marking the first instance ofplatinum in situ in South Africa. Bettel referred tothe samples as being from the Transvaal, but didnot have permission to divulge exact details of thelocality. His description is sufficiently precise thatthese samples can safely be considered to be fromthe Bushveld Complex. This report therefore rep-resents the first published documentation ofplatinum in the Bushveld Complex.

Russian AnalogiesBettel commented on analogies with the

Russian occurrences of platinum, which were themajor source of platinum at that time. Thus begana mistake or digression by South African geolo-gists to which Percy Wagner referred. Wagnerwrote (4): “The professional geologist made onlyone mistake. He followed too closely the experi-ence gained in the Urals, where platinum is alwaysassociated with chromite”.

The Russian deposits were all alluvial, but thesource rock was known to be chromitite, occur-ring in peridotite (an olivine-rich rock). The rockswere all uneconomic to mine. It was only thedecomposition of the peridotite and chromitite,and upgrading of the dense minerals by riveraction, that made the alluvial Russian occurrencespayable. Indeed, so closely was the Russian ana-logue followed, that once Merensky found the firstoutcrops of dunite pipes and the Merensky Reef inthe eastern Bushveld in 1924, he focused a greatdeal of his attention on exploring alluvium in theconfluence of two perennial rivers downstreamfrom the outcrop. He incorrectly thought thatthere might be major concentrations of easilyworked alluvial platinum derived from these out-crops.

Chromitite in the BushveldThe South African geologists followed this

Russian model closely and began investigationsinto the chromite-rich rocks of the BushveldComplex. By contrast, “the rocks associated withthe chromite were neglected” (4). Geologists of

the Geological Survey of South Africa, Wagner’semployer at the time, made a study of thechromite-rich rocks of the Bushveld Complex.Hall and Humphrey reported the occurrence ofplatinum in these rocks in 1908 (5), a publicationthat is often quoted as the first reference to plat-inum in the Bushveld Complex. Fifteen years later,Wagner (6) reviewed all the information availableon platinum in chromite and concluded “that itwould never pay to work the chromite rock forthat metal [platinum] alone”. The highest gradequoted was about 2 g t–1.

During the period 1906 to 1923, it can beassumed that it was not only the Geological Surveythat was actively evaluating the platinum potentialin chromite. It would appear that considerableexploration was also being undertaken. The extentof this can only be guessed, but the biography ofDr Hans Merensky by Olga Lehmann, “LookBeyond the Wind” (7), contains an interesting fewsentences. Referring to the period before 1924,Lehmann wrote “Many prospectors, includingMerensky, found copious chromite…”, but therehad been “four or five disastrous platinum discov-eries of former years that had not covered theirfinders in glory”. If Merensky had been involvedin previous unsuccessful exploration projects inthe eastern Bushveld, why should he try to raisemoney again in 1924 for yet another prospectingcampaign?

“Look Beyond the Wind”The above review is based entirely on pub-

lished documents, but I now speculate on whyMerensky would contemplate a subsequent explo-ration project when the previous attempts hadbeen unsuccessful. Admittedly, the first platinummine in South Africa had just opened in 1923, nearNaboomspruit, 150 km away, but Merensky knewwell that the host to the platinum there was inquartz veins, geologically apparently totally unre-lated to the Bushveld Complex. That wastherefore not the incentive.

To get inside Merensky’s mind, I must refer toanother incident related by Olga Lehmann (7).Merensky had been contracted by a major mininghouse in Johannesburg to evaluate a reported gold


discovery in Madagascar. Several consultantsjoined ship in July 1905 en route to Madagascar,and were shown an area in which gold had beenfound. Merensky and others retraced this gold withtheir pans through several streams and small pits.Then Merensky “looked beyond the wind”. Heturned in the opposite direction from his hotel andbegan panning other streams. He again found goldand realised that the area which the consultantswere meant to investigate had been salted (illegalenrichment of an ore in an area or sample to beassayed). How? Merensky recognised that the goldgrains panned in the area being promoted and thegold grains he found from elsewhere were of dif-ferent shapes, and that the associated denseminerals in his pan were different in the two local-ities. Sadly, salting was not an unknown activity inthose days, but Merensky had looked beyond theobvious, used his mineralogical acumen and recog-nised the fallacy!

Merensky’s 1924 ExplorationThe next question is how the Madagascar salt-

ing incident relates to Merensky’s 1924 visit to theeastern Bushveld. Previous exploration projectshad focused on the chromitite layers. A great dealis now known about the platinum group mineralsand their sizes in the chromitite layers, especiallythe Upper Group 2 chromitite layer. Their typicalsize is from 2 to (rarely) 30 μm (8). Merenskywould have known that panning in the field fromcrushed chromitite yielded very little platinum,because it was so fine grained that it was washedout of the pan. Had he ever found any in his pan,it would have been almost submicroscopic.Presumably, platinum grades based on panningwould not have agreed with chemical analyses ofchromite ore samples. Lest modern mineralogistsquestion the accuracy of such comparative tests, itshould be noted that Merensky stated in his earlyreports that panning and chemical analysis of sam-ples from the Merensky Reef gave remarkablysimilar grades.

In 1924, Andries Lombaard, a farmer in theeastern Bushveld, sent Merensky an “aspirin bot-tle” containing a white concentrate, panned from astream on his farm, Maandagshoek. Merensky had

it chemically analysed to confirm that it was plat-inum. Merensky evidently used his experience inMadagascar to good effect. He looked at the parti-cle size of the platinum group minerals in theconcentrate, and realised that they were enormousby comparison with everything that had beenfound in the chromitite layers. Merkle andMcKenzie (8) reported typical grain sizes from theMerensky Reef as 10–200 μm, and Wagner (9)reported a grain of 0.9 cm from the dunite pipes.

In 1998 the present author revisited the area onMaandagshoek from where Lombaard pannedplatinum. Some soil samples were analysed byAnglo Platinum (10). Subsequently some Germancolleagues undertook a mineralogical study of thesame area and found grains of various platinumgroup minerals in excess of 0.2 mm or 200 μm size(11) (see Figure 2). Merensky performed bothstudies in a matter of a day in 1924, and came tothe right conclusion. Merensky realised that theplatinum grains on Maandagshoek were totally dif-ferent from those found in the chromitite layers,and indicated a different source rock. The materialwas also coarse enough to be separable mechani-cally (the main extraction process in those days),with very good recovery of up to 85% (7). Hecommenced his often-documented explorationwith Lombaard. This ultimately had enormousconsequences for the world platinum industry.

Sceptics may claim that my suggestion cannotbe verified. None of the reports written byMerensky himself contain any interpretation orrationale to his prospecting, merely very factualstatements. However, the many and varied discov-eries made by Merensky and documented by OlgaLehmann (7) demonstrate his remarkably astutegeological sense. His appreciation of the signifi-cance of grain size would have been an obviousparameter in his prospecting skills.


References1 R. G. Cawthorn, Platinum Metals Rev., 1999, 43, (4),

1462 W. Bettel, ‘Occurrence of platinum in the

Transvaal’, S. Afr. Mines Comm. Ind., 10th Nov., 1906,206

3 J. Gray, “Payable Gold: An Intimate Record of theHistory of the Discovery of the PayableWitwatersrand Goldfields and of Johannesburg in

133

1886–1887”, Central News Agency, Johannesburg,1937, 286 pp

4 P. A. Wagner, S. Afr. J. Ind., 1925, 8, 905 A. L. Hall and W. A. Humphrey, Trans. Geol. Soc. S.

Afr., 1908, 11, (1), 696 P. A. Wagner, S. Afr. J. Sci., 1923, 20, 2237 O. Lehmann, “Look Beyond the Wind”, 3rd

Impression, Howard Timmins, Cape Town, 1959,232 pp

8 R. K. W. Merkle and A. D. McKenzie, ‘The Miningand Beneficiation of South African PGE Ores – An

Overview’, in “The Geology, Geochemistry,Mineralogy and Mineral Beneficiation of thePlatinum-Group Elements”, ed. L. J. Cabri,Canadian Institute of Mining, Metallurgy andPetroleum, Montreal, 2002, Special Vol. 54, pp.793–809

9 P. A. Wagner, “Platinum Deposits and Mines ofSouth Africa”, Oliver and Boyd, Edinburgh, 1929,326 pp

10 R. G. Cawthorn, J. Geochem. Explor., 2001, 72, (1), 5911 T. Oberthür, F. Melcher, L. Gast, C. Wöhrl and J.

Lodziak, Can. Mineral., 2004, 42, (2), 563

Platinum Metals Rev., 2006, 50, (3)

The Author

Grant Cawthorn comes from England. He has degrees in geology from Durham and Edinburgh Universities.After a post-doctoral fellowship in Newfoundland, he now teaches igneous petrology in the Department ofGeology at the University of the Witwatersrand, South Africa. His main interest is in the formation of theBushveld Complex with its vast reserves of platinum, chromium and vanadium. His post at the University issupported by the mining industry, and he holds the title of the Platinum Industry’s Professor of IgneousPetrology.

Fig. 2(c) Grain (c) has been cutthrough its centre and is photographedthrough a microscope. Its overallrounded shape suggests long trans-portation. It is made of several discreteminerals. The platinum alloy islabelled Pt-Fe, and other minerals arelabelled as follows: 1: laurite (RuS2);2: an unnamed mineral (Pd11Te2As2); 3:palladoarsenide (Pd2As); 4: sperrylite(PtAs2); 5: irarsite (IrAsS)

Fig. 2(b) Grain (b) is a near-perfectcube, also of platinum-iron alloy. Thecorners of grain (b) are still sharp,suggesting a very local derivation

Fig. 2(a) Grain (a) is a well-roundedgrain of platinum-iron alloy; its shapesuggests that it has been transportedover a long distance and rolled aboutin a river system

Grains of platinum group minerals panned from the farm Maandagshoek in theeastern Bushveld Complex, where the Merensky Reef was found in 1924. Thesephotographs were reported by Oberthür et al. (11), and are reproduced here bypermission of the Editor of The Canadian Mineralogist. Note the scale, indicat-ing that these are large grains (by platinum group mineral standards), and arevery different from anything found in chromitite layers

Layered mafic-ultramafic massifs remain themain focus of mineral exploration for the platinumgroup elements (PGEs). The large PGE reserves inthe Bushveld, Stillwater, Sudbury and Great Dykecomplexes provide the major share of platinummetals production, particularly the Bushveld, inSouth Africa. However, looking back to the begin-ning of the 20th century, it is surprising howimportant PGE lode mining was from unique geo-logical structures like Onverwacht and Mooihoek(South Africa), and Nizhny Tagil (Urals, Russia).The iron-rich platinum pipe bodies found through-out the eastern lobe of the Bushveld Complexwere probably discovered during early explorationof this igneous complex. In fact, the first discover-ies of economic levels of PGEs were in theOnverwacht and Mooihoek platinum pipes. Theterm IRUP, iron-rich ultramafic pegmatite (1, 2),describes the crosscutting pipes, veins and sub-concordant sheets of very coarse-grainedultramafic pegmatite that disrupt the BushveldCritical Zone in South Africa (3).

Ural-Alaskan Type Complexes (U-ATCs) are agroup of zoned mafic-ultramafic intrusive bodies,generally of sub-circular/stock-like shape, rarely

exceeding 10 km in diameter. They are well knownas a source of platinum placers in the Urals, in thefar east of Russia, in south-east Alaska, Colombia,Australia and other regions (4–6). Some of thebiggest platinum placers in the world were discov-ered in the Urals over 150 years ago, and duringthe first hundred years of operation yielded about400 tons of platinum. Despite this placer produc-tion, there has been very limited lode mining fromthe source rocks. The disparity in productionbetween placer and lode deposits is due partly tothe disseminated nature of the PGE mineralisationand partly to the difficulty of identifying andexploring narrow diameter pipes. Typical PGEmineralisation, found in Nizhny Tagil and theIRUPs of South Africa, is concentrated in discretecore zones (of 10 to 24 m diameter); it comprisesfine to very fine disseminated platinum alloys witha large range of compositions. Platinum is the mainelement. There are also bonanzas with rich con-centrations of PGEs up to 2000 g tonne–1 (7–9).

The occurrence of these platiniferous pipes isclosely linked to geodynamic environments thathave a long history of multistage magmatic-meta-somatic-hydrothermal evolution.

Platinum Metals Rev., 2006, 50, (3), 134–142 134

DOI:10.1595/147106706X128791

Phoscorite-Carbonatite Pipe ComplexesA PROMISING NEW PLATINUM GROUP ELEMENT TARGET IN BRAZIL

By Juarez FontanaPolytechnic School of University of São Paulo, Mining and Petroleum Engineering Department, LCT – Technological

Characterization Laboratory, São Paulo, Brazil; E-mail: [email protected]

The background to a project in Brazil is described that has found promising concentrationsof platinum group elements (PGEs) in phoscorite-carbonatite complexes. Further geochemicaland mineralogical research is underway to determine their potential as ore deposits. The well-established industrial demand and current level of prices for the platinum group metalshave encouraged the exploration of geological environments other than the layered mafic-ultramafic intrusions that provide the bulk of platinum metals. Environments, such as the Ural-Alaskan Type Complexes (U-ATCs) and the associated placer deposits were for many yearsthe only known sources of the PGEs. This paper attempts to show a connection betweenplatiniferous dunite-pyroxenite pipes in the Ural Platinum Belt and those on the eastern marginof the Bushveld Complex, both being significant PGE producers in the past, and phoscorite-carbonatite pipe (PCP) complexes. PCP complexes may be a promising source of PGEs. FourBrazilian PCP complexes are sampled (Salitre, Tapira, Ipanema and Catalão) as well as thePhalaborwa PCP complex in South Africa.


Phoscorite-Carbonatite Pipe (PCP)Complexes

Multistage, phoscorite-carbonatite zoned com-plexes are found worldwide, but the mostinteresting are in the Kola Peninsula, atMaymeicha-Kotui in northern Siberia, at the AldanShield in the far east of Russia, in Alaska, in theeast coast countries of Africa, and in the central-southern region of Brazil.

Phoscorite-carbonatite complexes are small,pipe-like bodies. They may take the form of dikes,sills, small plugs or irregular masses. A typical pipe-like body is of subcircular or elliptical cross-sectionand 3 to 4 km in diameter. Pipes extend to depthsof 3 to 13 km or more (10).

The magmatic mineralisation in pipe-like car-bonatite is commonly found in crescent-shaped,steeply dipping zones. Metasomatic mineralisationoccurs as irregular forms or veins (11). PCP com-

plexes possess a deep-step structure, and haveexperienced a prolonged, multistage metasomatic-hydrothermal evolution. Their complexcomposition and internal structure is clearlyindicative of multiple intrusions/injections ofmolten partially crystallised magmatic masses.Tectonic discontinuities, layering, banding andsteep inward concentric dips all indicate episodicdiapiric emplacement mechanisms. Phoscoritesand carbonatites are present as paired rocks; theyoccur near the nuclear area of the PCP complexes.Phoscorites are practically always associated withcarbonatites, but many carbonatite complexes,especially those of linear structure, do not containphoscorites.

Phoscorites, by definition, are medium andcoarse-grained igneous rocks of magnetite-forsterite-apatite composition (12). Besides typicalphoscorites, there are a variety of rock types of

Fig. 1 Map of southern Brazilshowing alkaline and alkaline-carbonatite occurrences (afterUlbrich and Gomes (26)).Catalão, Salitre, Tapira andIpanema PCP complexes areindicated. Some occurrences arein the sea; the dots near theshore and in the sea representintrusive alkaline volcanicrocks, including some volcanicsea islands. Those alkalineintrusions and volcanic islandsare related to the post-creta-ceous mechanism that resultedin the formation of the SouthAtlantic Ocean


similar mineral composition in the same geologicalenvironment. Some of these contain diopsideinstead of forsterite, while others contain actino-lite, with or without relicts of diopside orphlogopite and tetraferriphlogopite (13). In thePCP complexes calcite-carbonatite is the dominantrock type, followed by dolomite-carbonatite.

Carbonatites occur in close spatial and tempo-ral association with phoscorites and often formmultiphase phoscorite-carbonatite series. Thephoscorite-carbonatite series are mantle-derivedrocks, and much discussion centres on whethercarbonatite and phoscorites separate by fractionalcrystallisation and accumulation, by liquid immisci-bility, or by both mechanisms. The existence ofmultistage phoscorite-carbonatite complexes, thesimilar mineral association and their geochemicalpeculiarities indicate, without doubt, that bothrocks were formed from a parental magma bysome method of differentiation. The mantle originof phoscorite has been confirmed by both radi-ogenic and stable isotope studies (13, 14).Recently, a growing body of evidence has suggest-ed that such magmas could be derived directlyfrom depths in excess of 70 km by partial meltingof the carbonated mantle peridotite (15, 16).

Multiple intrusions are revealed by the litholog-ical and structural arrangement of successive stagesof paired phoscorite-carbonatite; there are six atKovdor (Russia) (14), two at Solki (Finland) (17),at least two in Phalaborwa (South Africa) (18, 19)and five in Tapira (Brazil) (20, 21).

The economic relevance of the PCP complexesis demonstrated by the presence of large hosted

mining operations, such as phosphate mining atPhalaborwa, Jacupiranga, Tapira and Catalão(Brazil); Kovdor (Russia); baddeleyite mining atPhalaborwa and Kovdor; copper mining atPhalaborwa; iron mining at Kovdor; niobium min-ing at Araxá and Catalão (Brazil), see Figures 1 and2, and uranium mining at Phalaborwa and Araxá.

The evolution of carbonatite and alkaline com-plexes could be related to crustal evolutiondynamics and to continental rifting; carbonatiteand alkaline complexes were formed shortly beforethe continental break-up and the opening of theocean basins. Continental-type U-ATCs are close-ly related to PCP complexes; for instance, the largecylindrical platiniferous dunite-pyroxenite/alkalineintrusive complexes at Inagly, Kondyor and Guliplatinum province in the Aldan Shield, easternSiberia. These alkali-ultrabasic complexes are wide-spread in tectonically stable areas of eastern Russia.These complexes usually exhibit well-formed, con-centric zoned structures, and are mainly composedof platinum-bearing chrome-spinel dunites andpyroxenites, which are very similar to those occur-ring in the Ural Platinum Belt (UPB) (22).However, in contrast to the Uralian ultramaficmassifs, the Aldanian massifs occur as isolated,pipe-like bodies intruding the Archean crystallinebasement.

The strong similarities between the Uralian andAldanian platinum-bearing ultramafic rocks isthought to reflect similar melt-rock interactionmechanisms during their individualisation in theshallow mantle, before their emplacement in dif-ferent geodynamic contexts. The geodynamic

Fig. 2 A panoramic view of the Catalão PCP minetaken in October 2004. The sides show the benchesof typical mines. This mine is still in operation andbesides phosphate, a significant amount of niobi-um ore is being mined here


setting was probably subduction-related for theUralian-Alaskan zoned complexes and it is clearlyintra-continental, possibly related to a rift zone forthe Aldanian massifs (23).

Could PCP Complexes GeneratePGE Ore Deposits?

PCP complexes are assumed to be associatedwith the continental-type U-ATCs series, similarto the Aldanian massifs, and have been examinedas a potential source for research on minerals ofthe PGEs. The question is, could some highlysiderophile mantelic magma, by recurrent andmultiphase magmatic-metasomatic-hydrothermalevolution, result in PGE ore deposits? Indicationsof such a process are isolated, for instance:• Platinum group metals have been recoveredfrom electrolytic refinery sludge at Phalaborwa formany years (24).• PGE mineral phases have been reported insome ore concentrates from Phalaborwa andKovdor (25).

Despite the suggestion that the PGE and gold-silver mineralisation of the Kovdor andPhalaborwa complexes has a close spatial andprobably genetic relationship with the multistagemagmatic and post-magmatic processes, there isno consistent empirical evidence to support this.

The strategy of the current research was to usegeological knowledge accumulated at PCP com-plexes that have been mined over a number ofyears. A field and laboratory research programmewas drawn up to survey four typical PCP intru-sions located in central southern Brazil: Ipanema,Tapira, Salitre I and Catalão II, and also atPhalaborwa in South Africa.

It was assumed that a selective survey of rockgeochemistry could define any potential PGEenrichment in the PCP complexes and could pro-vide sufficient data to establish a connectionbetween the evolution of the PCP phases and thePGE concentration. The survey used a mineralexploration approach to look for PGE concentra-tions at ore standard. Current geological conceptsand exploration methodology, known to be effec-tive in this area, were used. However, explorationis not driven by highly specialised models or by

using a range of techniques, so a follow-up explo-ration is necessary in order to establish appropriatetechniques that can be adapted for effective explo-ration.

The Brazilian PCP Complexes Voluminous flood basalt magmatism occurred

in central and southern Brazil from the EarlyCretaceous period to the Eocene time. Theseregions include the extensive Early CretaceousParaná continental flood basalt province and anumber of Early Cretaceous to Eocene alkalineigneous provinces that surround the Paraná Basin,see Figure 1.

Some quite large (up to 65 km2) intrusive, car-bonatite-bearing ultramafic complexes are locatedin these provinces (21). The main PCP complexesare: Catalão I, Catalão II, in southern Goias State;Serra Negra, Tapira, Salitre I, Salitre II and Araxáin south-west Minas Gerais State; Ipanema,Jacupiranga and Juquiá in east São Paulo State, andLajes and Anitapolis, the southern intrusionsfound in Sta Catarina State.

With kamafugites, lamproites and kimberlites,these complexes are multistage intrusionsemplaced into Late-Proterozoic metamorphic ter-rains (20). The PCP complexes are formed by theamalgamation of multiphase intrusions comprisingmainly ultramafic rocks (dunite, wehrlite andclinopyroxenite), see Figure 3, with subordinatecalcite-carbonatite, phoscorite and syenite.

The phoscorite-carbonatite rock pair in theBrazilian PCP complexes is represented mainly byapatite, magnetite, diopside (minor phlogopite)

Fig. 3 A specimen of magnetite-biotite-perovskite-pyroxenite from the Tapira phosphate mine in Brazil


phoscorite rock associated with calcite-carbonatite,see Figure 4.

The surface shape expression of all PCP com-plexes is almost rounded or oval, internallyconcentrically zoned, with diameters varying from1 to 7 km, depending on the level of erosion.However, the size of the actual phoscorite-carbon-atite bodies is even smaller (27). Rockidentification at the surface is quite difficult, aschemical weathering results in deep soil formationand a poor fresh rock exposure, see Figure 5.

The Phalaborwa PCP Complex The Phalaborwa complex intruded the Archean

basement at the edge of the Kapvaal Craton inEarly Proterozoic times (2.06 million years ago)

and consists of concentrically zoned, multipleintrusions, which decrease in age from the marginto the core. The core is an elliptically-shaped verti-cal intrusion known as the Loolekop pipe (18, 19).The host Phalaborwa complex is mainly composedof ultramafic rocks (dunite and pyroxenite) with acore of carbonatite and phoscorite. Minor rocktypes include glimmerite, syenite and fenite. Thecore of the composite intrusion shows a concentricarrangement of phoscorite around the margin anda core of banded carbonatite. Both these rocktypes were intruded by the central transgressivecarbonatite, see Figure 6.

The phoscorite is composed of olivine, mag-netite, apatite and phlogopite. The mineralcomposition and grain size present a wide rangevariation. The banded carbonatite consists largelyof magnetite-rich calcite-carbonatite, with minoramounts of apatite, olivine, phlogopite and biotite.The transgressive carbonatite is mineralogicallysimilar to the banded carbonatite, but lacks thebanding and represents a younger crosscuttingintrusive rock (24).

Field and Laboratory Surveys The present research was oriented to identify

possible PGE concentrations (at ore standard) andto establish if there was a relationship between thePGE concentration and individual rock types fromthe selected PCP complexes. To avoid the weath-ered upper zone, sampling was done on fresh rock

Fig. 4 A rock specimen from a Brazilian PCP complexof a phoscorite-carbonatite rock pair showing whitestreaks of calcite-carbonatite in phoscorite rock

Fig. 5 A Fosfertil mine geologist standing among deeplyweathered magnetite rich phoscorite in a phosphate minein Catalão, Brazil

Fig. 6 A specimen of banded carbonatite (white rock)and phlogopite magnetite phoscorite (dark rock) from theFoskor phosphate mine in Phalaborwa, South Africa


exposures from the lower benches of the open pitphosphate mines.

In order to avoid the “nugget effect”, large vol-ume chip rock samples (15 kg on average) werecollected from every kind of rock type at theIpanema, Tapira, Salitre I, Catalão II andPhalaborwa PCP intrusions. In all 70 samples werecollected, amounting to 1500 kg (Table I).

Rock sampling was planned after considerationof all geological data available from the selectedPCP complexes. Representative samples were col-lected from the main rock types, represented by:

Tapira: pyroxenite (Figure 3) and phoscorite;Salitre: pyroxenite, phoscorite and carbonatite; Catalão: phoscorite and carbonatite (Figure 4);Ipanema: phoscorite and magnetitite; andPhalaborwa: pyroxenite, phoscorite, bandedcarbonatite (Figure 6) and transgressive car-bonatite.Each rock sample was crushed, homogenised,

cross-split and finally pulverised in an oscillatingmill to less than 0.044 mm particle size, beforebeing taken for chemical assay. The first round ofchemical analyses (54 samples), was performed bythe Analytical Science Division of Mintek (SouthAfrica), using the fire assay technique with NiScollection and ICP-OES for the final PGE nota-tion. The laboratory procedures were certified;SARM7 (Standard South Africa) was the referencesample and every sample was assayed twice.

The results from Mintek at the beginning of2005 were conclusive in indicating that PCP com-

plexes are, in fact, very suitable for the enrichmentof PGEs at ore level standards. Three PCP intru-sions (Ipanema, Catalão and Phalaborwa) gavevery consistent data indicating a high concentra-tion of PGE:Ipanema: up to 5.15 ppm in total of Pt, Pd,

Rh, Ru, Os;Catalão: 4.47 ppm in total of Pt, Pd, Rh, Os;Phalaborwa: 16.67 ppm in total of Ir, Os.The other two (Salitre and Tapira) show no PGEenrichment at all.

It should be noted that the rock samples fromSalitre and Tapira were taken from available drillcores, so were not as representative as the samplesfrom Catalão, Ipanema and Phalaborwa collectedfrom the mining pit benches. The negative resultobtained for these areas must take the samplingconstraints into account, and consequently shouldnot be taken as a definitive statement.

On comparing the data, it can be seen that therelative abundance of PGE in Brazilian PCP com-plexes shows a distinct enrichment in rhodium (upto 0.60 ppm), with no indication of iridium con-centration, whereas the rocks from Phalaborwa donot register rhodium enrichment, but show a verysignificant iridium concentration (up to 13.5 ppm).

A general statement of PGE concentration is:Catalão: Pt > Pd > Ru > Rh > OsIpanema: Pt > Pd > Os > Rh > RuPhalaborwa: Pt > Pd > Ru > Ir > Os

All the rock samples were graded for particlesize distribution by wet screening, followed by

Table I

Main Characteristics of Samples from the Brazilian and Phalaborwa PCP Complexes

Main Tapira* Salitre I* Catalão II* Ipanema* Phalaborwa†characteristics

Age 70 Ma. 80 Ma. 83 Ma. 123 Ma. 2060 Ma.

Surface, km2 33.0 15.0 14.5 9.0 20.0

Shape Oval Oval Circular Circular Kidney-shaped

Main rocks Peridotite Peridotite Peridotite — —Clinopyroxenite Clinopyroxenite Clinopyroxenite Clinopyroxenite ClinopyroxeniteCalcite-carbonatite Calcite-carbonatite Calcite-carbonatite Calcite-carbonatite Calcite-carbonatitePhoscorite Phoscorite Phoscorite Phoscorite PhoscoriteSyenite Alkalic-syenite — Syenite —

* Ref. (27) CBMM †Ref. (18) Vielreicher et al.


chemical and mineralogical analyses as well as mag-netic separation studies.

The research, which is still in progress, will befollowed by systematic chemical and mineralogicalstudies. At present, total rock geochemical analysis,including major and minor trace elements, includ-ing the rare earths, is being determined by AASand/or ICP spectroscopy. The rock samples willbe surveyed for process mineralogy studies, andthin-polished samples will be prepared for opticaland scanning electron microscopy analyses.Mineralogical analyses will be performed by X-raydiffraction, optical mineralogy and scanning elec-tron microscopy with energy dispersive X-rayspectrometer (SEM-EDS). Gravitational concen-tration and heavy minerals extraction for asensitive mineralogical (microprobe) study will alsobe undertaken.

The author is aware that the following miner-alogical study will be much more difficult and time

consuming, and is of higher risk. The Brazilian lab-oratory expertise and technical apparatus may notbe sufficient. So it is further intended to ask for thecollaboration of Russian mineralogists, especiallythose linked with St. Petersburg University or withNATI Research JSC, St. Petersburg.

The main aim will be to identify the PGE min-eralogical phases and to establish their relationshipwith associated silicate and oxide minerals, lookingfor a probable genetic connection between thePGE concentration process and the evolution ofthe phoscorite-carbonatite complexes. The labora-tory procedures will be performed at theUniversity of São Paulo, and other internationalacademic and commercial research facilities.

Discussion and ConclusionsThe PGEs, rhenium and gold comprise the so-

called highly siderophile elements (HSE), theabundances of which in the upper mantle are often

Table II

PGE Crust and Mantle Abundance Compared with PGE Concentration in Combined Rock Types fromPCP Complexes

Material Platinum, Paladium, Rhodium, Osmium, Iridium, Ruthenium,ppt* ppt ppt ppt ppt ppt

Crustal average† 400 400 60 50 50 100PGE abundance

Upper mantle (peridotite)‡Average abundance 6500 5700 n.a. 3500 3500 5800

Catalão PCP 300,000 300,000 140,000 120,000 n.a. 120,000up to up to up to up to up to

2,860,000 810,000 600,000 200,000 200,000

Ipanema PCP 300,000 200,000 400,000 400,000 n.a. 150,000up to up to up to up to

3,200,000 1,330,000 540,000 490,000

Phalaborwa PCP 150,000 260,000 n.a. 300,000 1,100,000 120,000up to up to up to up to

580,000 3,280,000 13,500,000 160,000

Ore concentration, 0.15 0.26 0.14 0.12 1.10 0.12g t–1 up to up to up to up to up to up to

3.20 1.33 0.60 3.28 13.50 0.49

Concentration rate: 89 45 n.a. 57 315 27PCP composition/ up to up to up to up to up tomantle abundance 492 times 233 times 937 times 3857 times 85 times

* ppt = parts per trillion †Crustal average, Ref. (29) Wedepohl (1995) ‡Upper Mantle, Ref. (28) Morgan et al.


in the parts per trillion (ppt) range, and, which inthe undepleted primitive upper mantle, are inapproximately chondritic proportions (28, 29).

Phoscorite-carbonatite pipes (PCP) are gener-ated in a geodynamic environment which providesappropriate structural control to enable the verti-cal migration of dense iron-rich melts, overthousands of metres, and, equally important, theirconcentration into relatively small core zones.There is a limited understanding of the role of themagma mixing, assimilation, crustal metasomatismand other subsolidus processes in the origin andevolution of PCP complexes. However, the role ofmetasomatism, including late-stage alteration isalso important in the ultimate understanding ofthe PGE enrichment processes.

PGE mineralisation has a close spatial andprobably genetic relationship with the multistagemagmatic and post-magmatic evolution of PCPcomplexes. The PGE enriched zones, in the corezones of PCP complexes, represent the final prod-uct of a series of superimposed events like theprogressive PGE fractionation during the evolu-tion of mantle magma and the recurrence ofmagmatic pulse events.

The parental sulfur-poor/oxygen-rich melt sys-tem tends to result in a PGE ore mineralogyassemblage similar to those from classical ultra-mafic platiniferous pipes. These pipes arecommercially very attractive because of the lowcapital cost of establishing ore dressing facilities,that is, the bulk of the PGEs could be recoveredby conventional gravity and magnetic separationtechnologies.

The PGE concentrations in Catalão, Ipanemaand Phalaborwa provide solid evidence of thePGE potential of PCP complexes, particularly inregare to their ore concentration level. The PGEoccurrence in Ipanema shows the need to payattention to the detection of Fe-Cr and PGE-richvein-type varieties of spinels. The wide develop-ment of such Fe-spinel veins is indicative of a lowerosion level at the PCP complex cupola, and con-sequently, a better PGE potential (22).

The theoretical and factual data encourage theassumption that PCP intrusions are a very promis-ing target for PGE mineralisation. It is proposed

that the platiniferous pipe conceptual modelshould be extended to take into account PCPcomplexes as a promising new member. For sys-tematic mineral exploration, the PGE explorationstrategy for PCP complexes must be directedmainly to a selective structural and geochemistrysurvey, instead of to conventional saturation rocksampling geochemistry.

AcknowledgementsFoskor Limited, Jan H. van der Merwe and the

mining staff from Fosfertil SA, Tapira and Catalãoare gratefully thanked for permission, technicalassistance, hospitality and discussions during fieldwork and sampling at the Phalaborwa, Tapira andCatalão mines, respectively. This study has beenpartly supported by the São Paulo Research andDevelopment Agency (FAPESP – Grant03/09481-0) and hosted by the LCT –Technological Characterization Laboratory,Polytechnic School of the University of São Paulo(USP, Mining and Petroleum EngineeringDepartment).

References1 R. N. Scoon and A. A. Mitchel, ‘Discordant iron-

rich ultramafic pegmatites in the Bushveld Complexand their relationship to iron-rich intercumulus andresidual liquids’, J. Petrol., 1994, 35, 881

2 M. J. Viljoen and M. J. Scoon, ‘The distribution andthe main geologic features of discordant bodies oniron-rich Ultramafic pegmatite in the BushveldComplex’, Econ. Geol., 1985, 90, 1109

3 M. J. Viljoen and L. W. Shürman, ‘Platinum-GroupMetals’, in “The Mineral Resources of SouthAfrica”, 6th Edn., eds. M. C. G. Wilson and C. R.Anhaueusser, Council for Geoscience, Handbook16, Pretoria, 1998, pp. 532–630

4 H. P. Taylor, ‘The Zoned Ultramafic Complexes ofSoutheastern Alaska’, in “Ultramafic and RelatedRocks”, ed. P. J. Wyllie, John Wiley and Sons, Inc.,New York, 1967, pp. 97–121

5 P. Laznika, ‘Zoned Mafic-Ultramafic Complexes inPhanerozoic Orogenic Belts (Alaska or Ural Types)’,in “Empirical Metallogeny”, Elsevier, Amsterdam,1985, Vol. 1, Pt. A, pp. 243–250

6 A. J. Teluk, ‘Platinum Metallogenesis, Ural-AlaskanType Complexes’, Fifield Platinum Project, NSW,Australia, Geodyne Pty Ltd., Technical Report,2001, seehttp://www.rimfire.com.au/PDF/Geodyne%20report%20complete%20with%20figs.pdf

7 V. I. Smirnov, 1977, ‘Deposits of Platinum Metals,Late Magmatic Deposits’, in “Ore Deposits


of USSR”, ed. V. I. Smirnov, Pitman, London, pp.112–121

8 ‘Platinum group minerals and associated chrome-spinels of the Alaskan-type Nizhny Tagil Massif,middle Urals’, in “Mineral Deposits: Processes toProcessing”, eds. C. J. Stanley et al., Balkema,Rotterdam, 1999, pp. 787–790

9 E. F. Stumpfl and J. C. Rucklidge, ‘The platiniferousdunite pipes of the Eastern Bushveld’, Econ. Geol.,1982, 77, 1419

10 “Phoscorites and Carbonatites from Mantle to Mine:the Key Example of the Kola Alkaline Province”,eds. F. Wall and A. N. Zaitev, The MineralogicalSociety of Great Britain and Ireland, London, 2004,498 pp

11 A. G. Bulakh et al., ‘Overview of carbonatite-phoscorite complexes of the Kola Alkaline Provincein the context of a Scandinavian North AtlanticAlkaline Province’, in “Phoscorites and Carbonatitesfrom Mantle to Mine: the Key Example of the KolaAlkaline Province”, eds. F. Wall and A. N. Zaitev,The Mineralogical Society of Great Britain andIreland, London, 2004, pp. 1–36

12 “Igneous rocks. A Classification and Glossary ofTerms”, 2nd Edn., ed. R. W. Le Maitre, CambridgeUniv. Press, Cambridge, U.K., 2002, 236 pp

13 N. I. Krasnova et al., ‘Introduction to phoscorites:occurrence, composition, nomenclature and petro-genesis’, in “Phoscorites and Carbonatites fromMantle to Mine: the Key Example of the KolaAlkaline Province”, eds. F. Wall and A. N. Zaitev,The Mineralogical Society of Great Britain andIreland, London, 2004, pp. 45–74

14 N. I. Krasnova et al., ‘Kovdor – classic phoscoritesand carbonatites’, in “Phoscorites and Carbonatitesfrom Mantle to Mine: the Key Example of the KolaAlkaline Province”, eds. F. Wall and A. N. Zaitev,The Mineralogical Society of Great Britain andIreland, London, 2004, pp. 100–132

15 M. E. Wallace and D. H. Green, ‘An experimentaldetermination of primary carbonatite magma com-position’, Nature, 1988, 335, (6188), 343

16 P. J. Willie and W. J. Lee, ‘Model system controls onconditions for formation of manesiocarbonatite andcalciocarbonatite magmas from the mantle’, J.Petrology, 1988, 39, (11/12), 1885

17 M. J. Lee et al., ‘Carbonatites and phoscorites fromthe Sokli complex, Finland’, in “Phoscorites andCarbonatites from Mantle to Mine: the KeyExample of the Kola Alkaline Province”, eds. F.Wall and A. N. Zaitev, The Mineralogical Society ofGreat Britain and Ireland, London, 2004, pp.133–162

18 N. M. Vielreicher, D. I. Groves and R. M.

Vielreicher, ‘The Phalaborwa (Palabora) deposit andits potential connection to iron-oxide copper-golddeposits of Olympic Dam Type’, in “HydrothermalIron-Oxide Copper-Gold and Related Deposits. AGlobal Perspective”, ed. T. M. Porter, PGCPublishing, Adelaide, Australia, 2000, Vol. 1, pp.321–329

19 Palabora Mining Corp. Ltd., ‘The geology and eco-nomic deposits of copper, iron and vermiculite inthe Palabora igneous complex: a brief review’, Econ.Geol., 1976, 71, 177

20 J. F. Brod, ‘Petrology and Geochemistry of theTapira Alkaline Complex, Minas Gerais State,Brazil’, Ph.D. Thesis, Dept. of Geological Sciences,University of Durham, U.K., 1999

21 J. F. Brod et al., ‘The kamafugite-carbonatite associ-ation in the Alto Paranaíba Igneous Province (APIP)Southeastern Brazil’, Rev. Bras. Geoc., 2000, 30, (3),408

22 K. N. Malitch, ‘Assessment of the platinum poten-tial of clinopyroxenite-dunite massifs’, Trans. Russ.Acad. Sci., Earth Sci. Sect., V, 1994, 347A, (3), 400

23 A. A. Efimov et al., ‘Platiniferous dunites in the Uralsand the Aldan Shield, Russia: structural, mineralogi-cal and geochemical evidence for a similar origin’,Abstracts 9th Int. Platinum Symposium, Billings,Montana, U.S.A., 2002

24 M. G. C. Wilson, ‘Copper’, in “The MineralResources of South Africa”, 6th Edn., eds. M. C. G.Wilson and C. R. Anhaueusser, Council forGeoscience, Handbook 16, Pretoria, 1998, pp.209–225

25 N. S. Rudaschevsky et al., ‘A review and comparisonof PGE, noble-metal and sulphide mineralization inphoscorites and carbonatites from Kovdor andPhalaborwa’, in “Phoscorites and Carbonatites fromMantle to Mine: the Key Example of the KolaAlkaline Province”, eds. F. Wall and A. N. Zaitev,The Mineralogical Society of Great Britain andIreland, London, 2004, pp. 375–405

26 H. H. G. J. Ulbrich and C. B. Gomes, ‘Alkaline rocksfrom continental Brazil’, Earth Sci. Rev., 1981, 17, 135

27 C. S. Rodrigues et al., “Complexos Carbonatiticos doBrasil: Geologia”, Companhia Brasileira deMineracao e Metalurgia, Sao Paulo, Brasil, 1984, 44pp

28 J. W. Morgan et al., ‘Sideophile elements in Earth’supper mantle and lunar breccias: data synthesis sug-gests manifestations of some late influx’, Meteorit.Planet. Sci., 2001, 36, 1257

29 K. H. Wedepohl, ‘The composition of the continen-tal crust’, Geochim. Cosmochim. Acta, 1995, 59, (7),1217

The AuthorDr Juarez Fontana is professor and a mineralexploration expert at the University of São Paulo,Brazil. He is interested in both academic (geologicalmodels) and commercial projects, connected withPGE mineralisation and metallogenic processes,especially those related to the alkalic phoscorite-carbonatite intrusive complexes.

Johnson Matthey’s annual surveys of supplyand demand of the platinum group metals con-tinue with “Platinum 2006”, published in May2006 and reporting on the calendar year 2005.

Johnson Matthey records a world demandfor platinum of 6.7 million oz in 2005, an annu-al rise of 160,000 oz (2 per cent). Purchases bythe autocatalyst sector again grew strongly, withdemand increasing by 330,000 oz to a new highof 3.82 million oz. Europe accounted for mostof this growth, attributable to continued tight-ening of emissions rules and greater use ofcatalysed soot filters in light-duty diesel vehicleapplications.

Purchases of platinum for jewellery manu-facture fell by 200,000 oz (9 per cent) to 1.96million oz. A strong platinum price promptedstock reductions across the trade, and encour-aged the recycling of old jewellery. Chinesejewellery demand for platinum fell to its lowestfor seven years. Demand in Japan and NorthAmerica also contracted.

Industrial demand for platinum climbed by 9per cent to 1.675 million oz in 2005, cited as anall-time high. In the electrical sector there wasfurther growth in the production of data storagedisks using a platinum alloy layer. Continuingexpansion of liquid crystal display glass manu-facturing in Asia drove demand for platinum inglass applications to a record level.Consumption of platinum for making catalystsfor petroleum refining and chemical manufac-ture also increased.

World supplies of platinum increased by 2per cent in 2005, rising to 6.63 million oz,primarily due to greater output from SouthAfrica, which increased by 2 per cent to 5.11million oz. This increase was less than anticipat-ed, since efforts to expand output werehampered by a number of operational prob-lems. Supplies from North America and Russiafell slightly.

Demand for palladium increased by 7 percent to 7.04 million oz in 2005, due almostentirely to substantially greater use of the metalin jewellery. Palladium purchases for jewellerymanufacture, driven by rapid market develop-ment in China, rose by 54 per cent to 1.43million oz. Autocatalyst demand for palladiumincreased marginally to 3.81 million oz.Although automotive manufacturers madegreater use of palladium catalyst systems than in2004, average loadings of palladium on catalystscontinued to decline.

Palladium supplies fell by 2 per cent to 8.39million oz; growth in South African output didnot offset lower production in North Americaand a drop in sales of Russian metal.

Purchases of rhodium expanded by 11 percent to 812,000 oz in 2005, equalling the previ-ous high recorded in 2000. Use of the metal inautocatalyst, glass and chemical applicationsincreased.

A special feature, ‘Other Applications forPlatinum’, highlights a wide range of furtheruses of platinum. These vary from medium-scale automotive and medical applications tomany small end uses such as stationary sourcepollution control, gas safety sensors andcathodic protection. Each of the latter usesrequires just a few thousand ounces. In theautomotive sector, spark plugs and oxygen sen-sors account for a combined platinumconsumption of more than 130,000 oz in 2005.Biomedical uses of platinum (with an estimatedconsumption of a little over 100,000 oz in 2005)range from anticancer drugs to devices associat-ed with innovative treatments for heart andbrain disease.

“Platinum 2006”, Johnson Matthey PLC,Precious Metals Marketing, Orchard Road,Royston, Hertfordshire SG8 5HE, U.K.; E-mail: [email protected]; website:http://www.platinum.matthey.com/.

Platinum Metals Rev., 2006, 50, (3), 143 143

DOI: 10.1595/147106706X123930

Platinum 2006

IntroductionThe Subsecond Thermophysics Group at Graz

University of Technology has been working todetermine the thermophysical properties of liquidmetals for about 25 years. The work remains rele-vant and of current interest for scientificapplications as well as for the metalworking indus-try. Accurate data for the melting transition andthe liquid state are often sparse, but are essentialinputs to computer simulations, for instance thoseof solidification or die-casting.

Palladium is used in dentistry, jewellery, watch-making, spark plugs, the production of electricalcontacts, and metallising ceramics (1). Finely divid-ed palladium makes a good catalyst, used toaccelerate hydrogenation and dehydrogenationreactions, and for petroleum cracking. The palladi-um-hydrogen electrode is used in electrochemicalstudies. Palladium has recently attracted muchinterest as a potential replacement for higher-priced platinum in catalytic converters forcontrolling emissions from diesel vehicles.

The present experiments on palladium formpart of a systematic investigation of the thermo-physical properties of the platinum group metals.Measurements on rhodium are scheduled in thepresent programme; osmium and ruthenium arenot available in wire shape. Platinum and iridiumhave already been investigated (2–4), and show aslight increase of normal spectral emissivity atwavelength 684.5 nm in the liquid phase, similar to

the trend in emissivity values for palladium report-ed in the present work.

Experimental and Data ReductionProcedures

Using a pulse-heating apparatus (Figure 1), thethermophysical properties of conducting materialsare accessible from the solid state up to the end ofthe stable liquid phase. For the present investiga-tions, palladium samples in the form of wire (0.5mm diameter, 60 mm length, purity 99.9%, pur-chased from Alfa Aesar, Stock 10279, lot F28J28)were incorporated in a capacitor-driven dischargecircuit and resistively pulse-heated.

The following parameters were determineddirectly:• electric current• voltage drop• surface radiance• thermal expansion• normal spectral emissivity.From these, the following thermophysical proper-ties were derived:• sample temperature• enthalpy of fusion• isobaric heat capacity• electrical resistivity at initial geometry• electrical resistivity under thermal expansion• thermal conductivity• thermal diffusivity.

The accessible range of measurement extends


DOI: 10.1595/147106706X129079

Thermophysical Properties ofPalladiumDETERMINATIONS (INCLUDING SPECTRAL EMISSIVITY AT 684.5 nm) AT THE MELTING TRANSITIONAND IN THE LIQUID STATE

By Claus Cagran and Gernot PottlacherInstitut für Experimentalphysik, Technische Universität Graz, Petersgasse 16, A-8010 Graz, Austria; E-mail: [email protected]

The results from fast-pulse heating experiments (of duration 60 μs) performed on pure palladiumare presented. Thermophysical properties derived include specific enthalpy, enthalpy of fusion,electrical resistivity, isobaric heat capacity, thermal conductivity and thermal diffusivity, overa range of temperatures from the melting transition up to some hundred degrees higher inthe liquid state. Additionally, normal spectral emissivity at wavelength 684.5 nm is presented,

from room temperature up to superheated liquidstates. Experimental details are described exten-sively elsewhere (2, 5).

To enable accurate and unambiguous tempera-ture determination over such a large range,pyrometric detection based on Planck’s law ofblack-body radiation (6) was used. Normal spectralemissivity data were determined by an ellipsomet-ric method (division of amplitude polarimeter;μs-DOAP) (7, 8) to avoid uncertainties arisingfrom the unknown emissivity and its behaviour

over the temperature range of the measurement.

ResultsIn Figure 2, the normal spectral emissivity, ε, of

palladium at wavelength 684.5 nm is plottedagainst radiance temperature, Trad, and comparedwith literature results. The melting temperature ofpalladium, Tm, is 1828 K (9), whereas the radiancetemperature at melting is 1680 K for wavelength650 nm. At the latter temperature, the value of theemissivity is 0.36. An average of seven measure-


1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 32000,34

0,36

0,38

0,40

0,42

0,44

0,46

0,48

0,50

norm

al s

pect

ral e

mis

sivi

ty a

t 684

.5 n

m

radiance temperature at 650 nm, K

Fig. 1 Sketch of pulse-heating circuit

Fig. 2 Normal spectral emis-sivity for palladium atwavelength 684.5 nm versusradiance temperature at650 nm for palladium

0.50

0.48

0.46

0.44

0.42

0.40

0.38

0.36

0.34

Nor

mal

spe

ctra

l em

issi

vity

, ε, a

t 684

.5 n

m

Radiance temperature, Trad, at 650 nm, K

Key:Average of seven measure-ments from this work

—— Value from least squares fitRef. (10)

•


ments in the liquid phase produced the followinglinear fit for normal spectral emissivity in the radi-ance temperature range 1680 K < Trad < 3200 K:

ε = 0.3063 + 3.2258 × 10–5Trad (i)

At the end of the solid phase, emissivity valueswere around 0.49 (following surface preparationwith abrasive paper of grade 1200 or 4000). As thesurface smooths during liquefaction, a strongdecrease can be observed; the emissivity is 0.360 atthe end of melting. An emissivity of 0.3602 for liq-uid palladium at the melting temperature isreported (10); this was interpolated for wavelength684.5 nm. A slight increase of normal spectralemissivity is observed up to 3200 K; this is similarto the behaviour reported for platinum (11).

Figure 3 is a plot of specific enthalpy, H, versustemperature, T. Since this work focuses mainly onmelting and the beginning of the liquid phase,temperature dependences are shown in all plotsfrom around 1500 K upwards.

For the solid and liquid phases in the tempera-ture ranges: 1550 K < T < 1828 K and 1828 K <T < 2900 K respectively, averages of seven pulse-heating measurements give:

Hs(T) = –95.8103 + 0.2854T (iia)

Hl(T) = –55.5552 + 0.3507T (iib)

where H is in kJ kg–1 and T in K. The slope of

Equation (iib) gives a constant value of the isobar-ic heat capacity, cp, of (351 ± 36) J kg–1 K–1;Arblaster (11) recommends for the liquid a valueof 387 J kg–1 K–1. (Our conversion from the molarvalue uses an atomic weight of 106.42 (11)).

At melting, which is indicated in Figure 3 by avertical broken line, the specific enthalpy changesfrom Hs = 425.9 kJ kg–1 to Hl = 585.5 kJ kg–1 (thesubscripts s and l denoting solid and liquid respec-tively.) These results yield ΔH = (159.6 ± 16) kJkg–1 for the enthalpy of fusion. Arblaster (11) rec-ommends specific enthalpy values of 442.3 kJ kg–1

for the onset of melting and 593.4 kJ kg–1 for theend of melting, yielding ΔH = (151.1 ± 6.9) kJ kg–1

for the enthalpy of fusion. Dinsdale (12) reports avalue of ΔH = 157.3 kJ kg–1for the enthalpy offusion. Seydel et al. (13, 14) report specific enthalpyvalues of 445 kJ kg–1 for the onset of melting and609 kJ kg–1 for the end of melting, giving ΔH =164 kJ kg–1 for the enthalpy of fusion.

Figure 4 shows electrical resistivity, ρ, as afunction of temperature, T. At the onset of melt-ing, indicated by a vertical broken line, a resistivityvalue of 0.461 μΩ m is obtained for the initialgeometry (i.e. with no correction for thermalexpansion). The corresponding value at the end ofmelting is 0.724 μΩ m. Thus an increase Δρ =0.263 μΩ m is observed at melting.

The linear fit to the present values for the liq-

1600 1800 2000 2200 2400 2600 2800300

400

500

600

700

800

900

1000

spec

ific

enth

alpy

, kJ·

kg-1

temperature, K

Fig. 3 Specific enthalpyversus temperature forpalladium

Spec

ific

enth

alpy

, H, k

J kg

–1

Temperature, T, K

Key:Average of seven measure-ments from this workLinear least squares fit to meanvalues of measured dataRecommended values fromRef. (11)Melting temperature (1828 K)

——

- - -


uid in the temperature range 1828 < T < 2900 Kis:

ρ = 0.7777 – 2.9226 × 10–5T (iii)

where ρ is in μΩm and T in K.For electrical resistivity that is compensated for

thermal expansion, ρv, Seydel and Kitzel’s (13)thermal expansivity values for palladium wereadopted. Several other authors also report densityvalues for palladium (15–17). The change in diam-eter (and hence cross-section) of the sample withheating results in a shift to higher resistivity values.At the onset of melting, a volume-adjusted resistiv-ity of 0.495 μΩm was obtained, and at the end ofmelting 0.844 μΩm. Thus an increase Δρ = 0.349μΩm at melting is observed. Matula (18) recom-mends a resistivity at the onset of melting of 0.46μΩm and for the end of melting 0.83 μΩm, givingan increase Δρ = 0.37 μΩm at melting.

The polynomial fit to the present volume-adjusted values for liquid palladium in thetemperature range 1828 K < T < 2900 K is:

ρv = 0.8372 + 3.5413 × 10–6T (iv)

where ρv is in μΩm and T in K.The present resistivity values, compensated for

thermal expansion, agree excellently with Matula’srecommendations (18) for the liquid phase, andwell for the solid.

Figure 5 is a plot of thermal conductivity, λ,against temperature, T. To estimate thermal con-ductivity via the Wiedeman–Franz law (6), Seydeland Kitzel’s density data (13) were again used tocorrect electrical resistivity for actual thermalexpansion. For liquid palladium in the temperaturerange 1828 K < T < 2900 K:

λ = 0.8131 + 0.0286T (v)

where λ is in W m–1 K–1 and T in K.At the onset of melting a thermal conductivity

value of 85.3 W m–1 K–1 was obtained, and at theend of melting, for the beginning of the liquidphase, 54 W m–1 K–1. Zinovyev (19) reports a valueof 86 W m–1 K–1 at 1600 K. Vlasov et al. (20) ascited by Mills et al. (21) report thermal conductivi-ty values for the end of the solid phase and thebeginning of the liquid phase of 99 and 87 W m–1

K–1, respectively.Thermal diffusivity, a, can be estimated from

thermal conductivity (2). Thermal diffusivity is notplotted against temperature here; this gives noadditional relevant information since λ and cp areused for the calculation. The corresponding fit forliquid palladium in the temperature range1828 K < T < 2900 K yields:

a = –1.7636 × 10–6 + 8.9500 × 10–9T (vi)

where a is in m2 s–1 and T in K. At the onset of

1600 1800 2000 2200 2400 2600 28000,35

0,40

0,45

0,50

0,55

0,60

0,65

0,70

0,75

0,80

0,85

0,90

elec

trica

l res

istiv

ity,

μΩ·m

temperature, K

Fig. 4 Electrical resistivityversus temperature for palla-dium

Key:Average of seven measurements from thiswork, without correction for volume expansion,and least squares fitCalculated with volume data from Ref. (13),with correction for volume expansionMelting temperatureRecommended values from Ref. (18)

——

- - - E

lect

rical

resi

stiv

ity, ρ

, μΩ

m

Temperature, T, K

0.90

0.85

0.80

0.75

0.70

0.65

0.60

0.55

0.50

0.45

0.40

0.35


melting a value for a of 2.75 × 10–5 m2 s–1 wasobtained, and at the end of melting, for the begin-ning of the liquid phase, a was 1.46 × 10–5 m2 s–1.

DiscussionThermophysical data for liquid palladium are

quite sparse in the literature. Seydel and Kitzel (13)report only enthalpy dependences and no temper-ature dependences, since they did not perform thelatter measurements. The values found here forthe normal spectral emissivity ε for liquid palladi-um at wavelength 684.5 nm at the end of meltinggive an excellent match to that reported byMcClure et al. (10). The enthalpy of fusionobtained here compares very satisfactorily with theresults reported by Arblaster (11), Dinsdale (12)and Seydel et al. (13, 14), within the range of uncer-tainty of the present work. Further, selected valuesfrom (22), considered the best of their time, mustbe considered outdated for comparison purposes.

The temperature-dependent resistivity valuesreported here agree well with the recommendedvalues of Matula (18), within the present experi-mental uncertainty. Only a comparison of thepresent thermal conductivity values at the onsetand end of melting with corresponding data fromVlasov et al. (20) as reported by Mills et al. (21)shows a significant discrepancy, but Zinovyev’sdata (19) for the end of the solid phase appear toconfirm the thermal conductivity values obtained

in the present work.

UncertaintiesWithin the terms of (23), the uncertainties

reported here are expanded relative uncertaintieswith a coverage factor of k = 2. The uncertaintiesgiven in Table I have been derived for the thermo-physical properties calculated here.

ConclusionFor liquid palladium, a set of thermophysical

data is reported here: enthalpy, isobaric heatcapacity, electrical resistivity, thermal conductivityand thermal diffusivity as a function of tempera-ture. Since temperature measurement is combinedwith simultaneous emissivity measurements, thereis no ambiguity in the present temperature-depen-dent data.

AcknowledgementThis research was supported by the Austrian

Fonds zur Förderung der WissenschaftlichenForschung (FWF), Grant No. P15055.

1400 1600 1800 2000 2200 2400 2600 2800 3000

50

60

70

80

90

100

110th

erm

al c

ondu

ctiv

ity, W

·m-1·K

-1

temperature, K

Fig. 5 Thermal conductivityof palladium versus tempera-ture

Key:Average of seven measurementsfrom this workValues for melting transition fromRef. (20) as reported in Ref. (21)Value from Ref. (19)

•

Ther

mal

con

duct

ivity

, λ, W

m–1

K–1

Temperature, T, K

References1 “Platinum 2006”, Johnson Matthey PLC, Royston,

Hertfordshire, U.K. , 2006, pp. 32–392 B. Wilthan, C. Cagran, C. Brunner and G.

Pottlacher, Thermochim. Acta, 2004, 415, (1–2), 473 B. Wilthan, C. Cagran and G. Pottlacher, Int. J.

Thermophys., 2004, 25, (5), 1519


4 C. Cagran and G. Pottlacher, submitted to Int. J.Thermophys.

5 R. Gallob, H. Jäger and G. Pottlacher, Int. J.Thermophys., 1986, 7, (1), 139

6 M. Boivineau and G. Pottlacher, Int. J. Mater. ProductTechnol., 2006, 26, (3/4), 217

7 A. Seifter, F. Sachsenhofer and G. Pottlacher, Int. J.Thermophys., 2002, 23, (5), 1267

8 C. Cagran, B. Wilthan and G. Pottlacher, Int. J.Thermophys., 2004, 25, (5), 1551

9 R. E. Bedford, G. Bonnier, H. Maas and F. Pavese,Metrologia, 1996, 33, (2), 133

10 J. L. McClure, A. Cezairliyan and E. Kaschnitz, Int. J.Thermophys., 1999, 20, (4), 1149

11 J. W. Arblaster, Calphad, 1995, 19, (3), 36512 A. T. Dinsdale, Calphad, 1991, 15, (4), 31713 U. Seydel and W. Kitzel, J. Phys. F: Metal Phys., 1979,

9, (9), L15314 U. Seydel, H. Bauhof, W. Fucke and H. Wadle, High

Temp.–High Press., 1979, 11, (6), 63515 P. S. Martsenyuk and Yu. V. Ivanschenko, Ukr. Khim.

Zhur (SU), 1974, 40, 43116 L. D. Lucas, C. R. Acad. Sci. (Fr.), 1961, 253, 2526

17 T. Iida and R. I. L. Guthrie, “The Physical Propertiesof Liquid Metals”, Oxford University Press, Oxford,1988, p. 71

18 R. A. Matula, J. Phys. Chem. Ref. Data, 1979, 8, (4),1147

19 V. E. Zinovyev, “Metals at High Temperatures –Standard Handbook of Properties”, HemispherePublishing Corporation, New York, 1990

20 B. V. Vlasov, S. G. Taluts, V. E. Zinov’ev, N. A.Korenovskii and V. P. Polykova, Phys. Met. Metall.,1992, 74, (4), 371

21 K. C. Mills, B. J. Monaghan and B. J. Keene,‘Thermal Conductivities of Liquid Metals – Part 1:Pure Metals’, NPL Report CMMT(A), Teddington,U.K., 1997, p. 53

22 R. Hultgren, P. D. Desai, D. T. Hawkins, M. Gleiser,K. K. Kelley and D. D. Wagman, “Selected Values ofthe Thermodynamic Properties of the Elements”,American Society for Metals, Metals Park, OH, 1990

23 GUM, “Expression of the Uncertainty ofMeasurement in Calibration”, EA-4/02, EuropeanCo-operation for Accreditation, 1999,http://www.european-accreditation.org/n1/doc/EA-4-02.pdf

Table IUncertainties in Thermophysical Properties Determined for Liquid Palladium

Thermophysical parameter Symbol Uncertainty, %

Temperature T 4

Normal spectral emissivity ε 6

Enthalpy H 4

Enthalpy of fusion ΔH 10

Specific heat capacity at constant pressure cp 8

Electrical resistivity with initial geometry ρ 4

Expansion-corrected electrical resistivity ρV 6

Thermal conductivity λ 12

The Authors

Claus Cagran studied physics atGraz University of Technology,Austria, from which he received hismaster’s (Dipl.-Ing.) and doctoral(Dr. Techn.) degrees. He currentlyworks for the Optical TechnologyDivision at NIST, Gaithersburg,MD, dealing with opticalreflectance and emittancemeasurements of metals,ceramics, and surface coatings.

Gernot Pottlacher studied physicsat the Technical University of Graz,Austria, and holds its Dipl.-Ing. andDr. Techn. degrees. He is aprofessor at the Institute forExperimental Physics. His mainfields of activity are thethermophysical properties of pulse-heated liquid metals and alloys, aswell as didactic courses in physicsfor teachers in training.

PROPERTIESStudy of an Internally-Oxidized Pd0.97Ce0.03 AlloyV. M. AZAMBUJA, D. S. DOS SANTOS, L. PONTONNIER, M.MORALES and D. FRUCHART, Scr. Mater., 2006, 54, (10),1779–1783

Cold-worked foils of Pd0.97Ce0.03 underwent an inter-nal oxidisation heat treatment at 1073 K for 72 h.TEM showed the precipitation of needle-shapedCeO2 (1) with a cubic lattice parameter of 5.4 Å. (1)exhibited preferential growth directions relative to thePd matrix which correspond to the diagonal of the Pdcube. (1) were ~ 20–40 nm wide and 1–2 μm long, incoherence with the Pd matrix.

X-ray Photoelectron Spectroscopy and Magnetismof Mn–Pd AlloysM. COLDEA, M. NEUMANN, S. G. CHIUZBAIAN, V. POP, L. G.PASCUT, O. ISNARD, A. F. TAKÁCS and R. PACURARIU, J. AlloysCompd., 2006, 417, (1–2), 7–12

MnxPd1–x alloys and compounds (1) were preparedby Ar arc melting. The samples were melted repeat-edly (four times) in the same atmosphere to ensurehomogeneity. The electronic structures of (1) werestudied using XPS. Both valence band and core levelspectra were analysed. The magnetic properties of (1)are strongly correlated with their crystallographicproperties and can be explained considering only thenear-neighbour antiferromagnetic interactionsbetween both Mn and Pd atoms and Mn–Mn pairs.

Dramatic Evolution of Magnetic PropertiesInduced by Electronic Change in Ce(Pd1–xAgx)2Al3

P. SUN, Q. LU, T. IKENO, T. KUWAI, T. MIZUSHIMA and Y.ISIKAWA, J. Phys.: Condens. Matter, 2006, 18, (24), 5715–5723

Measurements of lattice parameters (a, c), magneticsusceptibility χ(T) and magnetisation M(H), specificheat C(T), and electrical resistivity ρ(T) were made forCe(Pd1–xAgx)2Al3. It was found that with increasing xthe system varies from antiferromagnetism to ferro-magnetism at x ~ 0.05, then back at x ~ 0.45. Themagnetic evolution resembles that of Ce(Pd1–xCux)2Al3.

CHEMICAL COMPOUNDSProtonation of Platinum(II) Dialkyl ComplexesContaining Ligands with Proximate H-BondingSubstituentsG. J. P. BRITOVSEK, R. A. TAYLOR, G. J. SUNLEY, D. J. LAW andA. J. P. WHITE, Organometallics, 2006, 25, (8), 2074–2079

Pt(II) dimethyl complexes [Pt(L)Me2], L = unsym-metrically substituted bipyridine, were prepared.Reactions in MeCN with 1 equiv. of a strong acidgave [Pt(L)Me(CH3CN)]+. The selectivity of the pro-tonation reactions is reported to be governed by stericeffects rather than H-bonding effects.

Synthesis and Structure of NbPdSiM. VALLDOR and R. PÖTTGEN, Z. Naturforsch., 2006, 61b, (3),339–341

NbPdSi (1) was prepared by melting the elements inan arc furnace. Well-shaped single crystals of (1) wereobtained by annealing in an induction furnace. ThePd and Si atoms were shown by powder and singlecrystal XRD analysis to build up a 3D [PdSi] networkwhere each Pd atom has a strongly distorted tetrahe-dral Si coordination at Pd–Si of 242–250 pm. The Nbatoms fill channels left in the [PdSi] network.

N-Heterocyclic Carbenes: Synthesis, Structures,and Electronic Ligand PropertiesW. A. HERRMANN, J. SCHÜTZ, G. D. FREY and E. HERDTWECK,Organometallics, 2006, 25, (10), 2437–2448

Rh(COD)X(NHC) complexes were synthesised.The relative σ-donor/π-acceptor quality of variousNHC ligands was classified by means of IR spec-troscopy at the corresponding Rh(CO)2I(NHC).Single crystal XRD studies of Rh pyrazolin- and tetra-zolinylidene complexes are reported. Differentazolium salts were applied to obtain Rh and Ir com-plexes with two and four carbene ligands.

Bis[iridium(I)] Complex of Inverted N-ConfusedPorphyrinM. TOGANOH, J. KONAGAWA and H. FURUTA, Inorg. Chem.,2006, 45, (10), 3852–3854

When a N-confused tetraphenylporphyrin wastreated with 2.0 equiv. of IrCl(CO)2( p-toluidine) and10 equiv. of NaOAc in toluene/THF = 20/1 (v/v) at100ºC for 3.5 h, a novel bis[iridium(I)] complex (1),wherein the confused pyrrole ring took an invertedconformation, was obtained in 17% yield. The reac-tions were significantly accelerated by THF. (1) canbe handled in air without special care. No decompo-sition was observed by heating in 1,2-Cl2C6H4. Nodemetallation occurred on CF3COOH addition.

Fullerene Polypyridine Ligands: Synthesis,Ruthenium Complexes, and Electrochemical andPhotophysical PropertiesZ. ZHOU, G. H. SAROVA, S. ZHANG, Z. OU, F. T. TAT, K. M.KADISH, L. ECHEGOYEN, D. M. GULDI, D. I. SCHUSTER and S.R. WILSON, Chem. Eur. J., 2006, 12, (16), 4241–4248

Fullerene coordination ligands (1) with a single bpyor tpy unit were synthesised. Coordination of (1) toRu(II) gave linear rod-like donor-acceptor systems.Steady-state fluorescence of [Ru(bpy)2(bpy-C60)]2+

showed a rapid solvent-dependent, intramolecularquenching of the Ru(II) MLCT excited state.Electrochemical studies on [Ru(bpy)2(bpy-C60)]2+ and[Ru(tpy)(tpy-C60)]2+ indicated electronic couplingbetween the Ru centre and the fullerene core.

Platinum Metals Rev., 2006, 50, (3), 150–153 150

ABSTRACTSof current literature on the platinum metals and their alloys

DOI: 10.1595/147106706X129097


ELECTROCHEMISTRYChemical and Electrochemical Synthesis ofPolyaniline/Platinum CompositesJ. M. KINYANJUI, N. R. WIJERATNE, J. HANKS and D. W.HATCHETT, Electrochim. Acta, 2006, 51, (14), 2825–2835

The direct chemical synthesis of Pt-polyaniline (1)composites was achieved by the oxidation of anilineby PtCl62–. The Pt particles were ~ 1 μm in diameter.Electrochemical synthesis of (1) was initiated by theuptake and reduction of PtCl62– into an a priori elec-trochemically deposited polyaniline film. This methodproduced a uniform dispersion of Pt particles withdiameters of 200 nm–1 μm.

Electrocatalytic Activity for Hydrogen Evolution ofPolypyrrole Films Modified with Noble MetalParticlesM. TRUEBA, S. P. TRASATTI and S. TRASATTI, Mater. Chem. Phys.,2006, 98, (1), 165–171

Polypyrrole (Ppy) films with Pt, Ru and Ir particleswere electrosynthesised on the surface of austeniticstainless steel by: (a) electrodeposition of a polymerfilm from a solution already containing an anionicmetal complex, followed by potentiodynamic or gal-vanostatic reduction; or (b) presynthesised Ppy filmsmodified by galvanostatic electrodeposition of themetals from solutions of their metal complexes. Theelectrocatalytic activity of the modified electrodes forthe H2 evolution reaction was tested in H2SO4 (0.05M) by potentiodynamic techniques (0.5 mV s–1).

PHOTOCONVERSIONPlatinum–Acetylide Polymer Based Solar Cells:Involvement of the Triplet State for EnergyConversionF. GUO, Y.-G. KIM, J. R. REYNOLDS and K. S. SCHANZE, Chem.Commun., 2006, (17), 1887–1889

Blends of a blue-violet absorbing Pt-acetylidepolymer (1) with 1-(3-(methoxycarbonyl)propyl)-1-phenyl[6.6]C61 (PCBM), can be used as the activematerial in a photovoltaic device. (1) acts as the chro-mophore and electron donor blended with PCBM asan electron acceptor. Photoinduced charge separationin the blends is believed to occur via the triplet excit-ed state of the organometallic polymer.

Structurally Integrated Organic Light EmittingDevice-Based Sensors for Gas Phase andDissolved OxygenR. SHINAR, Z. ZHOU, B. CHOUDHURY and J. SHINA, Anal. Chim.Acta, 2006, 568, (1–2), 190–199

The O2-sensitive dyes Pt- or Pd-octaethylporphyrin(1), were embedded in polystyrene, or dissolved insolution. Their performance was compared to that ofRu(dpp)3

2+. A green OLED, based on Alq3, was usedto excite (1). The O2 level was monitored in the gasphase and in H2O, EtOH and toluene by measuringchanges in the PL lifetime τ of (1).

Photophysical Properties of the Photosensitizer[Ru(bpy)2(5-CNphen)]2+ and IntramolecularQuenching by Complexation of Cu(II)M. G. MELLACE, F. FAGALDE, N. E. KATZ, H. R. HESTER and R.SCHMEHL, J. Photochem. Photobiol. A: Chem., 2006, 181, (1),28–32

The lifetime of the 3MLCT emitting state of[Ru(bpy)2(5-CNphen)]2+ has been determined inMeCN by flash photolysis and time correlated singlephoton counting techniques. The value obtained, τ =2.2 μs, suggests its potential use as a photosensitiserin molecular devices. Static and dynamic quenchingof the complex luminescence by Cu2+ ions was seen.

ELECTRODEPOSITION AND SURFACECOATINGSAdhesion and Bonding of Pt/Ni and Pt/CoOverlayers: Density Functional CalculationsG. F. CABEZA, N. J. CASTELLANI and P. LÉGARÉ, J. Phys. Chem.Solids, 2006, 67, (4), 690–697

The electronic and energetic properties of Pt/Niand Pt/Co surfaces are examined using the full-potential linearised augmented plane wave method.The results of the shifts in the d-band centers whenone metal (Pt) is pseudomorfically deposited onanother with smaller lattice constant (Ni, Co) are pre-sented, together with those corresponding to thesurface and adhesion energies. The results for pureNi, Co and Pt surfaces are given to compare with datain the literature.

Self-Assembled Palladium Nanowires byElectroless DepositionZ. SHI, S. WU and J. A. SZPUNAR, Nanotechnology, 2006, 17,(9), 2161–2166

The self-assembly production of Pd nanowires (1)has been carried out by electroless deposition on aporous stainless steel template. Various arrays of self-assembled (1) in the form of single wire, parallel andcurved wires, intersections and network structures areobtainable. (1) can be built in a self-assembled man-ner by the assembly of nanoparticles generated in theinitial stages of the deposition without any externalfield except the chemical reaction.

Selective Growth of IrO2 Nanorods UsingMetalorganic Chemical Vapor DepositionG. WANG, D.-S. TSAI, Y.-S. HUANG, A. KOROTCOV, W.-C. YEHand D. SUSANTI, J. Mater. Chem., 2006, 16, (8), 780–786

Area-selective growth of IrO2 nanorods (1) wasachieved via MOCVD using (MeCp)Ir(COD) on asapphire (012) or (100) substrate which consisted ofpatterned SiO2 as the nongrowth surface. Orientationof (1) was controlled by the in-plane epitaxial relationbetween the IrO2 crystal and sapphire, along with theIrO2 growth habit in the [001] direction. The pho-tolithography method gave better resolution inpreserving rod orientation of (1) at the growth andnongrowth boundary zone.

151


APPARATUS AND TECHNIQUEHigh-Purity COx-Free H2 Generation from NH3 viathe Ultra Permeable and Highly Selective PdMembranesJ. ZHANG, H. XU and W. LI, J. Membrane Sci., 2006, 277, (1–2),85–93

A compact H2 generation system combining NH3

decomposition with separation by a series of Pdmembranes (3 μm) has been developed to providehigh-purity, COx-free H2 for fuel cell applications.Removal of H2 product in a Pd membrane reactorwas shown to promote NH3 conversion over a Ni-based catalyst. However, ex situ integration, in whichan NH3 cracker was followed by a Pd membrane puri-fier, was deemed more suitable for practical uses dueto its high productivity of pure H2.

Nanocomposite of Pd-Polyaniline as a SelectiveMethanol SensorA. A. ATHAWALE, S. V. BHAGWAT and P. P. KATRE, Sens.Actuators B: Chem., 2006, 114, (1), 263–267

A Pd-polyaniline nanocomposite (1) was synthe-sised by oxidative polymerisation of an anilinesolution containing Pd nanoparticles. (1) was highlyselective and sensitive to MeOH vapours. The selec-tivity of (1) was further investigated by exposing it toMeOH-EtOH and MeOH-isopropanol. Here (1)exhibited a response identical to that for pure MeOH,except for the response time.

Hydrogen Permeation Characteristics of Thin PdMembrane Prepared by MicrofabricationTechnologyY. ZHANG, J. GWAK, Y. MURAKOSHI, T. IKEHARA, R. MAEDAand C. NISHIMURA, J. Membrane Sci., 2006, 277, (1–2),203–209

A Pd membrane (1), ~ 2.5 μm thick, on Si waferwas successfully prepared using microfabricationtechnology. H2 permeability of (1) was investigatedwithin 473–673 K, and found to be ~ 50–65% that ofa 0.70 mm thick Pd membrane. Grain growth wasfound in (1) after permeation, and the presence ofCO2 reduced H2 permeability significantly.

HETEROGENEOUS CATALYSISEffect of Pt Precursors on Catalytic Activity ofPt/TiO2 (Rutile) for Water Gas Shift Reaction atLow-TemperatureH. IIDA, K. KONDO and A. IGARASHI, Catal. Commun., 2006,7, (4), 240–244

Pt/TiO2 (rutile) catalysts for the low temperature-WGSR were prepared from various Pt precursors.The catalytic activity decreases for the precursorsused: H2PtCl6·6H2O, Pt(C5H7O2)2 > [Pt(NH3)4]Cl2 >[Pt(NH3)4](NO3)2 > cis-[Pt(NO2)2(NH3)2]. There was alinear relationship between catalytic activity and Ptdispersion. The TOF for the LT-WGSR was almostconstant regardless of Pt dispersion.

Enhancement of Naphthalene Hydrogenation overPtPd/SiO2-Al2O3 Catalyst Modified by GoldB. PAWELEC, V. LA PAROLA, S. THOMAS and J. L. G. FIERRO, J.Mol. Catal. A: Chem., 2006, 253, (1–2), 30–43

The effect of the support (amorphous silica-alumi-na (ASA) and C multiwall nanotubes (MWNTs)) onthe activity of PtPd catalysts in naphthalene hydro-genation is described. Also, the effect of Auincorporation on PtPd/ASA was studied. AuPtPd/ASAshowed the highest naphthalene conversion and low-est deactivation. The less acidic PtPd/C MWNTs didnot show S-resistance. The contribution of the acidsites of the support to S-resistance and their deactiva-tion by coke are discussed.

Improved CO Oxidation in the Presence andAbsence of Hydrogen over Cluster-DerivedPtFe/SiO2 CatalystsA. SIANI, B. CAPTAIN, O. S. ALEXEEV, E. STAFYLA, A. B. HUNGRIA,P. A. MIDGLEY, J. M. THOMAS, R. D. ADAMS and M. D. AMIRIDIS,Langmuir, 2006, 22, (11), 5160–5167

Pt5Fe2/SiO2 and PtFe2/SiO2 samples (1), preparedfrom organometallic cluster precursors decarbonylat-ed in H2 at 350ºC, were found to be highly active forthe oxidation of CO in the presence or absence of H2.Pt-Fe nanoparticles were formed with sizes of 1–2nm. A higher degree of dispersion and more homo-geneous mixing of the metals were observed in (1) ascompared to a conventionally impregnation preparedPtFe/SiO2 (2). (1) were also more active than Pt/SiO2

or (2) for the oxidation of CO in air.

Hydrogenation of Sunflower Oil on Pd Catalysts inSupercritical Conditions: Effect of the ParticleSizeC. M. PIQUERAS, M. B. FERNÁNDEZ, G. M. TONETTO, S. BOTTINIand D. E. DAMIANI, Catal. Commun., 2006, 7, (6), 344–347

Sunflower oil hydrogenation was carried out usingsupercritical propane and Pd/γ-Al2O3. The selectivityto cis-isomers and the production of saturated fattyacids was favoured by a small Pd particle size (< 2nm). There was no significant variation in the reactionrate nor in the TOF. Despite the fact that during thereaction a phase separation occurred, propane was insupercritical state in both phases.

Effects of Natural Water Ions and Humic Acid onCatalytic Nitrate Reduction Kinetics Using anAlumina Supported Pd-Cu CatalystB. P. CHAPLIN, E. ROUNDY, K. A. GUY, J. R. SHAPLEY and C. J.WERTH, Environ. Sci. Technol., 2006, 40, (9), 3075–3081

The NO3– reduction rate of a H2O sample using Pd-

Cu/γ-Al2O3 was 2.4 × 10–01 l/min g cat. The additionof SO4

2–, SO32–, HS–, Cl–, HCO3

–, OH– and humic aciddecreased the NO3

– reduction rate. Preferentialadsorption of Cl– inhibited NO3

– reduction to agreater extent than NO2

– reduction. Dissolved con-stituents in groundwater decreased the NO3

–

reduction rate. Removal of dissolved organic matterusing activated C increased the NO3

– reduction rate.

152

HOMOGENEOUS CATALYSISRecovery and Reuse of Ionic Liquids andPalladium Catalyst for Suzuki Reactions UsingOrganic Solvent NanofiltrationH. WONG, C. J. PINK, F. C. FERREIRA and A. G. LIVINGSTON,Green Chem., 2006, 8, (4), 373–379

Organic solvent nanofiltration was used for separat-ing ionic liquids (1) and the catalyst Pd2(dba)3-CHCl3from Suzuki cross-couplings. The reactions were car-ried out in 50:50 wt.% ethyl acetate and (1). The postreaction mixture was diluted further with ethyl acetateand then separated by nanofiltration. The productwas recovered in the nanofiltration permeate, while(1) and Pd catalyst were retained by the membrane.

Unexpected Roles of Molecular Sieves inPalladium-Catalyzed Aerobic Alcohol OxidationB. A. STEINHOFF, A. E. KING and S. S. STAHL, J. Org. Chem.,2006, 71, (5), 1861–1868

The effect of molecular sieves (MS3A) onPd(OAc)2/pyridine (1) and Pd(OAc)2/DMSO (2) wasinvestigated by performing kinetic studies of alcoholoxidation. MS3A enhanced the rate of (1)-catalysedoxidation of alcohols. This was attributed to the abil-ity of MS3A to serve as a Brønsted base. In contrast,no rate enhancement was observed with (2). Both (1)and (2) exhibit improved catalyst stability in the pres-ence of MS3A, resulting in higher catalytic TONs.The MS3A provided a heterogeneous surface thathinders bulk aggregation of Pd metal.

Unsymmetric-1,3-Disubstituted Imidazolium Saltfor Palladium-Catalyzed Suzuki-Miyaura Cross-Coupling Reactions of Aryl BromidesH.-W. YU, J.-C. SHI, H. ZHANG, P.-Y. YANG, X.-P. WANG and Z.-L. JIN, J. Mol. Catal. A: Chem., 2006, 250, (1–2), 15–19

Unsymmetric 1,3-disubstituted-imidazolium salts(1) derived from ferrocene were prepared, and theirpreliminary activities as precursors of N-heterocycliccarbene ligands for Pd-catalysed cross-coupling ofaryl bromides with phenylboronic acid were studied.A combination of Pd(OAc)2 and (1) was an excellentcatalyst system for the Suzuki-Miyaura cross-couplingof aryl bromides with phenylboronic acid in the pres-ence of Cs2CO3.

Rh(0) Nanoparticles as Catalyst Precursors for theSolventless Hydroformylation of OlefinsA. J. BRUSS, M. A. GELESKY, G. MACHADO and J. DUPONT, J.Mol. Catal. A: Chem., 2006, 252, (1–2), 212–218

The hydroformylation of 1-alkenes can be performedin solventless conditions, using ligand-modified orunmodified Rh(0) nanoparticles (1) prepared in imi-dazolium ionic liquids as catalyst precursors.Aldehydes were generated when 5.0 nm (1) are used.With smaller nanoparticles, chemoselectivity isdecreased; large sized nanoparticles (15 nm) produceonly small amounts of aldehydes, similarly to a classi-cal heterogeneous Rh/C catalyst precursor.

FUEL CELLSThermal Stability in Air of Pt/C Catalysts and PEMFuel Cell Catalyst LayersO. A. BATURINA, S. R. AUBUCHON and K. J. WYNNE, Chem.Mater., 2006, 18, (6), 1498–1504

The thermal stability of Pt/Vulcan XC 72 and a 46wt.% Pt/Vulcan XC 72/Nafion layer was studied.Low temperature (100–200ºC) C combustion occuredin the presence of Pt. In PEMFC catalyst layers, thethermal decomposition temperature of Nafion is low-ered by ~ 100ºC to 300ºC in the presence of Pt/C.

High Performance PtRuIr Catalysts Supported onCarbon Nanotubes for the Anodic Oxidation ofMethanolS. LIAO, K.-A. HOLMES, H. TSAPRAILIS and V. I. BIRSS, J. Am.Chem. Soc., 2006, 128, (11), 3504–3505

PtRuIr/C MWNTs system (1) was prepared usingan organic colloid synthesis method. (1) has a veryhigh real surface area and is highly active toward theoxidation of MeOH. The Ir component acts as a pro-moter. The splitting of the Pt(111) XRD feature intofour peaks and the shift to larger d spacing reflect thehigh dispersion of the metallic components.

ELECTRICAL AND ELECTRONICENGINEERINGInterface Effect on Ferroelectricity at theNanoscaleC.-G. DUAN, R. F. SABIRIANOV, W.-N. MEI, S. S. JASWAL and E. Y.TSYMBAL, Nano Lett., 2006, 6, (3), 483–487

A first-principles study of ultrathin KNbO3 ferro-electric films (1) placed between two metal electrodes,either Pt or SrRuO3, was carried out. The strength ofbonding and intrinsic dipole moments at the inter-faces was shown to control the ferroelectricity. Thepolarisation profile was inhomogeneous across thefilm thickness. The critical thickness for the net polar-isation of (1) was predicted to be ~ 1 nm for Pt and1.8 nm for SrRuO3 electrodes.

Calculations and Measurements of ContactResistance of Semi-Transparent Ni/Pd Contacts top-GaNK. H. A. BOGART and J. CROFTON, J. Electron. Mater., 2006, 35,(4), 605–612

Calculations of specific contact resistance (1) as afunction of doping and barrier height were performedfor p-GaN. (1) were measured for oxidised Ni/Au,Pd, and oxidised Ni/Pd ohmic contact metal schemesto p-GaN. The Ni/Pd contact had the lowest (1).Some Ni had diffused away from the GaN surface tothe contact surface, with the bulk of the Pd located inbetween two areas of Ni. Both Ni and Pd interdif-fused with the GaN at the semiconductor surface.The majority of the O was as NiO. PredominantlyNiO and PdO species were formed, with higher Niand Pd oxides at the contact surface.



METALS AND ALLOYSNickel-Based SuperalloyHITACHI LTD British Appl. 2,418,207

A nickel-based superalloy (1) includes (in wt.%):3–7 Cr, 3–15 Co, 4.5–8 W, 3.3–6 Re, 4–8 Ta, 0.8–2Ti, 4.5–6.5 Al, 0.1–6 Ru, 0.01–0.2 Hf, < 0.5 Mo, ≤0.06 C, ≤ 0.01 B, ≤ 0.01 Zr, ≤ 0.005 O, ≤ 0.005 N, andoptionally a rare earth element at 0.1–100 ppm, withthe balance Ni. Single crystal turbine blades can bemade from (1). (1) has excellent mechanical strengthand resistance to corrosion and oxidation.

ELECTRODEPOSITION AND SURFACECOATINGSPd-Containing CoatingELTECH SYST. CORP World Appl. 2006/028,443

An electrocatalytic coating (1) of mixed metal oxide,preferably containing Pt group metal oxides, and anelectrode using (1) are used for the electrolysis of ahalogen-containing solution. (1) includes a topcoatinglayer of oxides of Pd, Rh or Co. The Pd oxide com-ponent reduces the operating potential of theelectrode and removes the necessity of a ‘break-in’period to reach the lowest electrode potential.

Low-Pressure Deposition of Ru and Re LayersTOKYO ELECTRON LTD U.S. Appl. 2006/068,588

A low pressure method for depositing Ru and Relayers at high deposition rates, with low particulatecontamination and good step coverage is described. ARu- or Re-carbonyl precursor is processed with a car-rier gas in a process chamber at a pressure of < 20mTorr. The metal is deposited onto a surface by ther-mal chemical vapour deposition.

Ru Film-Forming InkJAPAN SCI. TECHNOL. AG Japanese Appl. 2005-336,263

An ink can be used to form films of metallic Ru orRu oxide on a resin film substrate such as polyimides.The ink contains a compound obtained by preheatinga β-diketone, β-ketoester or β-diester complex of Ruin an alcohol solution. This ink can be applied to asurface and heated to give the films.

APPARATUS AND TECHNIQUEMembrane for Diffusion Limited Gas SensorsGENERAL ELECTRIC British Appl. 2,417,561

A micro fuel cell sensor for measuring selectedgases in fluid streams is claimed. The sensing elementhas identical first and second gas diffusing electrodes,made from at least one of Pt/C, Au, Pd, Pd-Pt, Ru,Ir, Os, Rh or Ta. The electrodes are separated fromgas-containing media by gas-permeable membranes.A spacer having an acidic electrolyte is placedbetween the electrodes, facilitating electrochemicaloxidation and reduction of gases at the electrodes.

Graphitic Nanotubes in Luminescence AssaysMESO SCALE TECHNOL. LLC U.S. Patent 7,052,861

Electrochemiluminescent complexes of Ru, Os orRe, in particular of Ru (1), are attached to C nanotubesupports together with an enzyme cofactor (2), andare used in luminescence assays. The analyte of inter-est can be detected by bringing the sample intocontact with the assay composition, causing oxidationor reduction of (2) and electrochemiluminescence of(1). The latter can be correlated to the presence oramount of analyte.

HETEROGENEOUS CATALYSISMultiple Layer Exhaust Gas CatalystCATALER CORP European Appl. 1,640,575

An exhaust gas catalyst system contains multiplelayers of catalysts on solid support. The first layercontains a noble metal active component includingRh, and optionally Pt, with a barely soluble Ba com-pound. The second layer contains another noblemetal which may include Pt or Pd. The system isstructured so that the two layers come into contactwith the exhaust gas sequentially.

Manufacture of Noble Metal Alloy CatalystsUMICORE AG CO KG U.S. Appl. 2006/094,597

Supported noble metal catalysts (1) are manufac-tured with a high degree of alloying and smallcrystallite size, < 3 nm, using polyol solvents in a twostep process. The first component is a transitionmetal such as Co, Cr, Ru, preferably Ru; the second isa noble metal such as Pt, Au, Ag, Pd, Rh, Os, Ir or amixture. (1) can be used as electrocatalysts in fuel cellsor as gas-phase catalysts for CO oxidation or exhaustgas purification.

Low-Emissions Diesel FuelCLEAN DIESEL TECHNOL. INC U.S. Patent 7,063,729

A low-emissions diesel fuel is composed of aviationkerosene, detergent, lubricity additive and a bimetal-lic, fuel-soluble Pt and Ce fuel-borne catalyst.Retarding engine timing can further reduce NOx andthe use of a diesel particulate filter and/or diesel oxi-dation catalyst can further reduce CO, unburnedhydrocarbons and particulates.

Exhaust Gas Cleaning CatalystTOYOTA CENTRAL RES. DEV. LAB. INC

Japanese Appl. 2005-305,217An exhaust gas cleaning catalyst exhibits high NOx

removal activity even after exposure to a high-tem-perature atmosphere, and is prevented from beingpoisoned by sulfur. There are two catalysts, preparedwith a porous oxide powder carrier and a NOx occlu-sion material, the first carrying Pt and the secondcontaining θ-alumina and carrying Rh. The catalystsintermingle but remain separated by their carriers.

NEW PATENTS

DOI: 10.1595/147106706X124669

HOMOGENEOUS CATALYSISRing-Closing Metathesis ProcessBOEHRINGER ING. INT. GmbH U.S. Appl. 2006/063,915

A process is described for preparing compounds (1)which are active for the treatment of hepatitis C viral(HCV) infections, or are intermediates for the prepa-ration of active compounds. (1) are formed bycyclising diene compounds (2) in a suitable organicsolvent, in the presence of a Ru catalyst (3). Theprocess is performed in a gas such as CO2 or a gasmixture at supercritical or near-supercritical condi-tions, with (3) at ~ 25–50 mol% relative to (2).

Process for Oxidation of AlkanesCSIR U.S. Appl. 2006/142,620

A Pd complex (1) catalysed process for the oxida-tion of linear alkanes is claimed which employsmolecular O2 as the oxidant, to produce secondaryalcohols and ketones in high selectivity. (1) mayinclude monodentate, bidentate or polydentate lig-ands up to a maximum Pd coordination number of 4.The process may be carried out under a continuousfeed of pure or diluted O2 or in air, in the presence orabsence of a solvent, and does not require the use ofa co-catalyst.

Acetic Acid Production MethodsCELANESE INT. CORP U.S. Patent 7,053,241

A method for production of acetic acid by carbony-lation of MeOH involves a Rh-based catalyst system,with at least one catalyst stabiliser (1) selected fromRu- or Sn-salts or a mixture. Precipitation of Rh dur-ing recovery of acetic acid is minimised by (1), even inlow H2O content reaction mixtures and in the pres-ence of an iodide salt copromoter at > ~ 3 wt.% ofthe reaction mixture. (1) may be present at molar con-centrations of metal to Rh from ~ 0.1:1–20:1.

Curable Silicone Releasing Agent CompositionSHIN-ETSU CHEM. CO LTD Japanese Appl. 2005-314,510

The composition of a curable silicone film withsmall peel resistance at low-speed/high-speed peel-ing, with slipperiness and good resistance to airexposure is described. The film includes: (A) adiorganopolysiloxane containing 0.5–5 mol% alkenylgroups bonded to Si; (B) a diorganopolysiloxane withalkenyl groups bonded to Si at the terminals of themolecular chain; (C) an organohydrogenpolysiloxane;and (D) a catalytic amount of Pt-based catalyst.

FUEL CELLSCarbon Nanotube Pastes and Methods of UseO. MATARREDONA et al. U.S. Appl. 2006/039,848

Dispersible pastes consisting of C single-wallednanotubes (SWNTs) in H2O or an organic solvent areprepared. These pastes can be impregnated withnoble metal precursors including compounds of Pt.The SWNT-Pt composites can have small Pt clustersdistributed evenly over the surface and can be used ascatalysts or as electrodes for fuel cells.

Carbon-Supported Alloy Nanoparticle CatalystsRES. FOUND. STATE UNIV. NEW YORK

U.S. Patent 7,053,021C-supported core-shell PtVFe nanoparticle electro-

catalysts (1) are formed from a reaction solution (2)including precursors containing metals or salts of Pt,V and Fe plus an organic compound. The processproduces nanoparticles of controlled size within therange 1.0–10.0 nm, the size being determined by thecomposition of (2). (1) are particularly useful for O2

reduction reactions (ORR), exhibiting ORR catalyticactivities in the range ~ 2–4 times that of a standardPt/C catalyst.

Reduced Cost Catalyst for Fuel CellNISSAN MOTOR CO LTD Japanese Appl. 2005-332,662

Manufacturing costs for a fuel cell catalyst contain-ing Pt plus Ir oxide (1) are reduced using thedescribed method. First an Ir complex is formedfrom a mixture of Ir chloride solution and a hydrox-ide of an alkali metal or alkaline earth metal. Thiscomplex is deposited onto a C support, then baked tohigh temperature to form (1), without burning thesupport. Finally Pt is added to the catalyst supportwhich contains (1).

ELECTRICAL AND ELECTRONICENGINEERINGSelf-Aligned Silicide ContactIBM CORP U.S. Appl. 2006/051,961

A self-aligned Ni alloy silicide contact is described.A conductive Ni-Pt alloy is first deposited onto a Si-containing semiconductor structure. An O2 diffusionbarrier is deposited to prevent metal oxidation, thenan annealing step causes formation of a Ni-Si, Pt-Sicontact in regions in contact with Si. Finally a selec-tive etching step removes unreacted Ni-Pt fromregions not in contact with Si.

Iridium Oxide NanowiresSHARP LAB. INC U.S. Appl. 2006/086,314

Ir oxide nanowires (1) are grown from an Ir-con-taining precursor, using a MOCVD process from thesurfaces of a growth promotion film which has non-continuous surfaces. (1) have a diameter in the rangeof 100–1000 Å, a length in the range of 1000 Å–2 μmand an aspect ratio (length:width) of > 50:1. (1)include single-crystal cores covered with an amor-phous layer of < 10 Å thickness.

Magnetic Recording MediaSEAGATE TECHNOL. LLC U.S. Patent 7,041,394

A magnetic recording disc includes a disc substrateof glass, quartz, Si, SiO2, ceramic or AlMg, with anetched locking pattern of multiple pits, completelyfilled with chemically synthesised nanoparticles (1) ofa single magnetic species. (1) may consist of Fe-Pt,Co-Pt, Fe-Pd or Mn-Al, and have a grain size of 3–10nm. (1) exhibit short-range order characteristics,forming self organised magnetic arrays.


The effects of permanent poisons on platinumgroup metal (pgm) catalysts are irreversible (1).The poisons are so strongly absorbed that theycannot be adequately removed, even with aggres-sive remedial actions such as steaming or thermalregeneration.

Typical examples of permanent poisonsencountered when processing hydrocarbon feed-stocks would be contaminant metals such asnickel, copper, vanadium and, in particular, ‘heavymetals’ such as lead, arsenic and mercury (Hg). Allthese metals have a significant deactivating effecton precious metal catalysts, but the ‘heavy metals’have a particularly dramatic impact. This is so evenat very low concentrations, either present in thefeedstock or accumulated on the catalyst.

The metals content of crude oils may vary froma few ppm to several thousand ppm, dependingupon the geographical and geological origin of thecrude. The primary refining process involves dis-tilling and fractionating the crude into separatehydrocarbon ‘cuts’. Distillation tends to concen-trate any metallic components in the oil residue.However, some organometallic compounds arevolatilised at distillation temperatures; they there-fore distribute in some of the higher-boilingdistillate fractions which are then processed fur-ther for refinery and petrochemical applications.

Mercury is an increasingly common contami-nant, detected in natural gas deposits, coal depositsand in some crude oils from a variety of geo-graphical regions such as South East Asia, parts ofthe Middle East and South America. Hg may bepresent in a range of chemical forms, includingelemental, ionic and organic, and as suspendedsolids, in concentrations ranging from a few partsper billion (ppb) to several thousand ppb.

Catalysts consisting of palladium supported onalumina are widely used for the selective hydro-genation of acetylenes in petrochemical processing(2). These catalysts are adversely affected by heavymetals that accumulate on the surface, and effec-tively alloy with the precious metal through dπ-dπ

bonding. The result is to effectively remove activecentres from the desired reaction scheme. Catalystpoisoning reduces yields and shortens catalyst life.It is therefore essential to reduce inlet Hg concen-trations to < 5 ppb to achieve an economicallyacceptable service life. Hg poisoning, if leftunchecked, may require an unplanned and prema-ture catalyst change-out, with all its associatedcosts, including downtime. Unfortunately, it is noteasy to detect and measure Hg at these very lowconcentrations as the metal is lost on sample linesand container walls. Exacting analytical methodsand sampling techniques are necessary if theresults are to be reliable and consistent. Neutronactivation analysis and cold vapour atomic fluores-cence spectroscopy have both proven successfulfor determining Hg concentrations.

The effects of poisons cannot be avoided com-pletely, but they can be mitigated by a carefullydesigned upstream purification system to protectthe pgm catalysts. Although the Hg compoundsare not the most reactive, several products havebeen developed to remove Hg from both gaseousand liquid hydrocarbon streams. Elemental Hgand inorganic Hg species can be very effectivelyremoved (to levels of less than 1 ppb) through sur-face adsorption on non-regenerable metal sulfidepellets in fixed-bed reactors at ambient tempera-ture. Hg has a high affinity for sulfur, and themetal sulfide reacts chemically with the Hg impu-rity to form a stable mercuric sulfide.

Several approved disposal and treatmentoptions are available for the Hg-laden spentabsorbent. J. K. DUNLEAVY

References1 D. E. Grove, Platinum Metals Rev., 2003, 47, (1), 442 ‘Selective hydrogenation of acetylenes and dienes’,

Johnson Matthey Catalysts, http://www.jmcata-lysts.com/pct/marketshome.asp?marketid=10&id=374

Platinum Metals Rev., 2006, 50, (3), 156 156

FINAL ANALYSIS

Mercury as a Catalyst Poison

DOI: 10.1595/147106706X128403

The AuthorDr John Dunleavy is Business Director – Refinery, Oil & GasSection, Johnson Matthey PCT, PO Box 1, Belasis Avenue,Billingham TS23 1LB, U.K. He has over 20 years’ experience in thecatalyst industry.

platinum metals review · platinum metals rev., 2006, 50, (3) 114 [pt(nh 2 eh) 4][ptcl 4] with a...

Documents