View Article Online / Journal Homepage / Table of Contents for this issue Analyst, May 1995, Vol. 120 1391 Determination of Platinum, Palladium, Ruthenium and Iridium in Geological Samples by Isotope Dilution Inductively Coupled Plasma Mass Spectrometry Using a Sodium Peroxide Fusion and Tellurium Coprecipitat ion* Jacinta Enzweilert and Philip J. Potts* Department of Earth Sciences, The Open University, Walton Hall, Milton Keynes, UK MK7 6AA Kym E. Jarvis NERC ICP-MS Facility, Centre for Analytical Research in the Environment, Imperial College at Silwood Park, Buckhurst Road, Ascot, Berkshire, UK SL5 7TE A method was developed for the determination of Ru, Pd, Ir determination of PGEs a challenge. Appropriate analytical and Pt in geological samples by isotope dilution inductively procedures generally have three steps: chemical decomposi- coupled plasma mass spectrometry. After fusion of the tion of the sample, separation of the PGEs from the matrix sample with sodium peroxide, the platinum group elements (i.e. , preconcentration) and finally elemental determination. were preconcentrated by Te coprecipitation. Results One of the most successful methods for both decomposing the obtained for the reference materials WGB-1, TDB-1, UMT- sample and concentrating the PGEs is the nickel sulfide fire 1, WPR-1, WMG-1 and SARM-7 are in excellent agreement assay procedure coupled to the high sensitivity of inductively with the recommended values for elements above the coupled plasma mass spectrometry (ICP-MS) instrumentation detection limit level of 0.3-2.0 ng g-1 (whole rock). Although for the determination ~tage~7.8later improved by an inter- the method used only 0.5 g of sample, no errors were found mediate Te coprecipitation step.9-12 The nickel sulfide fire that could be associated with sample inhomogeneity effects in assay method offers the advantage of accepting large sample the analysis of the above reference materials. Further aliquots (routinely 20-50 g), which is essential in the analysis measurements indicated that the technique could be of any sample that might contain discrete PGE minerals. In extended to the determination of Rh and Au by external such circumstances, serious sampling errors are likely to occur calibration. (caused by the 'nugget effect') if techniques that employ Keywords: Platinum group element determination; geological smaller sample masses (e.g., <1 g) are used. However, the sample; sodium peroxide fusion; isotope dilution; inductively nickel sulfide procedure has some disadvantages. One is the Published on 01 January 1995. Downloaded by UNIVERSIDAD ESTADUAL DE CAMPINAS 19/06/2015 16:35:23. coupled plasma mass spectrometry relatively large amounts of reagents used and the contribution they may make to the analytical blank unless very high purity reagents are used. Another potential difficulty is that judge- ment is required to optimize flux composition in the analysis of Introduction some sample types (e.g., chromite and sulfide-bearing samples The development of methods for the determination of the or kimberlites).13-15 A third potential difficulty is that some platinum group elements (PGEs; Ru, Rh, Pd, Os, Ir and Pt) in investigations have indicated that PGE recoveries by nickel geological samples has attracted considerable interest in sulfide fire assay may not always be quantitative, with losses recent years for several reasons. The economic importance of occurring during the fusion procedure itself16 and during these elements has stimulated the development of analytical subsequent crushing17 and dissolution of the nickel sulfide techniques appropriate for geochemical exploration and ore button.9 Alternative pi-ocedures are often used for the analysis processing. 1 However, there is also increasing academic of mantle rocks, in Os/Re isotope studies for example, where interest in the geochemistry of these elements in studies of, for geochemical and thermodynamic constraints would indicate example, mantle and metallogenetic processes273 and the that discrete PGE grains are unlikely to be present. In these Cretaceous-Tertiary boundary layer with its associated Ir circumstances, a low blank decomposition procedure is more anomaly.4 This resurgence of interest has been further important than the capability to analyse 50 g sample aliquots stimulated by the recent deve!opment of isotopic techniques and indeed, many studies have been undertaken successfully for OdRe geochronometry studies .5,6 using techniques requiring less than 1 g of sample.18-2" The low concentration of PGEs in geological samples (with In the analysis of mantle rocks, several approaches have typical background levels of a few ng g-l) makes the been used to dissolve the sample. Some of the most common are those based on acid attack (total or partia1)21322 and fusion * Presented at Geoanalysis 94: An International Symposium on the Analysis of procedures.23J4 The digestion with aqua regia in sealed Geological and Environmental Materials, Ambleside. England, September 18- tubes25926 is also being used because of its effectiveness in 22, 1994. + Present address: Universidade Estadual de Campinas, lnstituto de Geo- dissolving PGE-bearing phases and the low blank levels, the cisncias, Campinas, SP, P.O. Box 6152, CEP 13081, Brazil. latter being crucial in Re/Os geochemistry. The fusion of * To whom correspondence should be addressed. powdered samples with oxidizing fluxes such as sodium View Article Online 1392 Analyst, May 1995, Vol. 120 peroxide has been widely used both to decompose resistant by heating gently in 6 ml of 6 rnol I-' HCI. Deionized water minerals after a preliminary acid attack27 and to decompose was added to dilute the solution to a final volume of 20 ml. the entire sample in radiochemical neutron activation analysis methods.20 Indeed, fusions with oxidizing fluxes have the advantage that they are very effective in dissolving both resistant and sulfide mineral phases? However, such pro- Sample Decomposition cedures must normally be restricted to small samples, usually 1 Samples were dried at 105°C overnight and cooled in a g or less, in applications such as those referred to above where desiccator bcfore weighing. Conical crucibles made of glassy sampling errors are not likely to occur. carbon with a capacity of 60 ml (Sigri, Meitingen, Germany) In this work, a method has been developed for the were used in the sample fusion procedure. In order to protect determination of Ru, Pd, Ir and Pt using a fusion with sodium the crucible from attack by the sodium peroxide flux, a layer of peroxide to decompose the sample in glassy carbon crucibles. sodium carbonate (3.5 g) was used to line the bottom and walls After dissolution of the fused cake, the sample solution was of the crucible bcfore each fusion. Sodium peroxide (0.5 g), spiked and the PGEs were separated by Te coprecipitation sample (0.5000 g) ar *:en more sodium peroxide (1.5 g) were prior to determination of the PGEs by isotope dilution ICP- weighed into the crucible onto the sodium carbonate lining. MS. The method was applied to five Canada Centre for The sample and sodium peroxide were carefully mixed using a Mineral and Energy Technology (CANMET) PGE reference nickel spatula. The crucible was covered with a lid and materials: WGB-1 (gabbro rock), TDB-1 (diabase rock), transferred into a cold furnace. 1hc temperature of the UMT-1 (ultramafic ore tailings), WPR-1 (altered peridotite) furnace was raised to 200 "C and left for 15 min and then raised and WMG-1 (mineralized gabbro), and to the mineralized ore to 490 "C and left for 1 h. The crucible was then removed from of the South African Bureau of Standards (SARM)-7. the furnace and allowed to cool. With the crucible transferred Additional studies were carried out to evaluate the determina- into a 400 ml beaker (covered with a watch-glass) the cake was tion of the mono-isotopic elements Rh and Au by conven- dissolved in 50 ml of warm de-ionized water, swirling the tional external calibration. The fusion and Te coprecipitation contents periodically. After dissolution, the spike solution procedure described here is similar to that used in the containing Pt, Pd, Ir and Ru was added to the alkaline sample radiochemical neutron activation method of Stone and solution. The crucible was removed from the beaker using a Crocket.20 The over-all method is similar to the isotope PTFE tweezer and rinsed repeatedly with small volumes of de- dilution ICP-MS procedure described by Sen Gupta and ionized water and 6 rnol I-' HCl to ensure quantitative Gregoire.29 However, instead of using a two-stage sample transfer of the contents into the beaker. More acid was added dissolution procedure based on an initial acid attack, followed (total 45 ml) to dissolve completely the hydroxides and to by fusion of any residue with sodium peroxide, we have acidify the solution. The solution was then gently boiled for a adopted a single-stage digestion based on a sodium peroxide few minutes to destroy peroxides. Any precipitated silica was fusion and have included Pt in the list of analysed elements. removed by filtration using a Millipore filtration apparatus (with Whatman No. 4 filter-paper) and the filtrate collected in a 250 ml Pyrex filter-flask. The original beaker and the precipitate were washed with hot 1 rnol I-' HCl(20 ml), which Experimental was added to the filtrate. The filter-flask containing the solution was then heated to Reagents boiling and 2 ml of Te solution were added. The solution was Quartz-distilled HCI and sub-boiling distilled HN03 were brought back to boiling and tin(I1) chloride solution (10 ml) used in the analytical procedure. De-ionized water was used was slowly added. The solution was boiled for a few minutes and all other reagents were of analytical-reagent grade. until the black precipitate was thoroughly coagulated. At this Poly(tetrafluoroethy1ene) (PTFE), glassware and glassy car- point, more Te solution (1 ml) was added.