ANALYST, AUGUST 1991, VOL. 116 773 Evaluation of a Rapid Technique for the Determination of Precious Metals in Geological Samples Based on a Selective Aqua Regia Leach Charles J. B. Gowing and Philip J. Potts" Department of Earth Sciences, The Open University, Walton Hall, Milton Keynes MK7 6AA, UK Results are described for optimizing the rapid determination of precious metals in geological samples (principally of ophiolitic origin, including chromitites) using an aqua regia [HCI + HN03 (3 + I)] leach followed by inductively coupled plasma mass spectrometry. Little improvement in extraction efficiency is afforded by aqua regia attack for longer than 30 min and heating the extraction mixture appears t o have a detrimental effect on the extraction efficiency of the precious metals.The procedure was evaluated by analysis of a suite of standard reference materials and some independently analysed ophiolitic rock samples. Generally, Au and Pd are quantitatively extracted, Pt, Rh, Ru and 0 s are extracted t o a lower but significant extent (2040% recovery) and Ir is poorly extracted (typically I-10% recovery). The insolubility of selected platinum group element minerals in aqua regia is considered t o be the dominant effect for the non-quantitative recoveries. Keywords: Aqua regia leach; geological sample; platinum group elements and gold determination; inductively coupled plasma mass spectrometry; mineral insolubility In recent years, considerable interest has been shown in the development of techniques for the determination of Au and the platinum group elements (PGEs) in geological samples.This has been stimulated not only by. the extensive exploration interest in these metals but also by the development of instrumentation of high sensitivity [particularly inductively coupled plasma mass spectrometry (ICP-MS)] capable of extending applications to samples containing parts per billion (ppb) levels of these elements. Despite these advances in instrumentation, most analytical techniques require a preconcentration step so that these ppb detection limits, essential in geological applications, can be achieved. Many current procedures are based on fire assay. The traditional lead fire assay procedure'-3 is normally restricted to the quantitative recovery of Au, Pt and Pd.435 Over the last decade, there has also been considerable interest in the nickel sulphide fire assay procedure,6,7 not the least because this procedure is capable of recovering quantitatively all the precious metals from a range of sample types.Indeed, the nickel sulphide procedure is effectively the analytical 'bench mark' against which other techniques are normally evaluated and has been subjected to useful further refine- ments in reducing the mass of the button necessary to recover the PGE from 20 g to 1 g.8 Despite the widespread use of fire assay techniques, these procedures have some drawbacks. One reservation is the low recoveries that can be encountered from some sample types, a second fusion then being necessary to recover precious metal residues from the silicate glass.3.4 Another reservation is that the success of the fire assay procedure depends to some extent on the skill and judgment of the operator in optimizing both the flux composition and fusion conditions.With this back- ground, it was decided to evaluate an alternative, very simple, extraction procedure based on a selective aqua regia [HCl + HN03 (3 + l)] leach. This procedure requires only standard laboratory equipment and can be undertaken without diffi- culty by non-specialized laboratory personnel. Furthermore, the technique has the potential of being adapted for operation in a field laboratory. The aim of this paper is, therefore, to report preliminary data on investigations of the aqua regia leach procedure for recovering the PGEs. Aqua regia leach is currently used widely for the determina- tion of Au in geological samples.9-12 The success of the * To whom correspondence should be addressed.technique depends on the ability of aqua regia to dissolve native Au and associated alloys with Cu and Ag. This property arises from the reactivity of nitrosyl chloride (NOCl) and/or free chlorine formed in freshly prepared aqua regia solution (see for example Latimer and Hildebrand13). In some procedures, hydrofluoric acid is added to the reaction mixture to attack silicate phases and facilitate the liberation of Au grains that might not otherwise be wetted by aqua regia alone .14-16 Although there are some reports in the literature concern- ing the use of aqua regia in the determination of the PGEs,17-*4 this preconcentration procedure is not extensively used in practical applications, in part because of uncertainties in the quantitative recovery efficiencies and how this is influenced by sample matrix and PGE mineralogy. One of the aims of the present study is to investigate further the influence of PGE mineralogy on the recovery efficiency by using aqua regia leach.However, as little information is available on the detailed PGE mineralogy of the geological reference materials that are currently available, these materials are not ideal candidates for this study. Following an extensive project to characterize the PGE mineralogy of the Unst ophiolite (Shetland Isles, N. Scotland), detailed minera- logical descriptions of several samples are available .25-*7 One of the samples described as part of this project was a serpentinized chromitite (CHR-C) from Cliff, Unst , contain- ing exceptionally high abundances of the PGEs (in the ppm range).This sample was an obvious candidate for the present investigation, partly because of the availability of a detailed PGE mineralogical description, and partly because one of the secondary aims of the present work was to develop an analytical method suitable for chromitites, recognizing that such samples are not always easy to analyse by conventional fire assay techniques.2.28 Complementary measurements were undertaken on other chromitite samples from Unst (more detailed descriptions of which can be found in the references cited above) and on a reference sample as follows: CHR-A, a chromitite from Harold's Grave, Unst, containing enhanced PGE abundances, particularly Os, Ru and Ir; CHR-B, a 'barren' chromitite from Unst ; CHR-C, an exceptional PGE mineralized chromitite from Cliff, Unst (referred to above); and SARM 7 (previously PTO-l), a platinum ore Certified Reference Material (CRM), available from the South African Bureau of Standards (SABS) for which certified values of all the PGEs are available.29 This sample consists of a gabbroic rock and is therefore representative of silicate materials.The samples CHR-B and CHR-C are sub-samples of the774 ANALYST, AUGUST 1991, VOL. 116 proposed chromitite reference samples CHR-Bkg and CHR- Pt+ , respectively, which are, at the time of writing, the subject of an international cooperative study.30 These sub-samples were separated after jaw-crushing the bulk material used to prepare the reference materials.Although never fully homogenized at the milled powder stage, the mineralogies (though not necessarily the abundances) of CHR-B and CHR-C are expected to be virtually identical to those of CHR-Bkg and CHR-Pt+ , respectively. Experimental Aqua regia was freshly prepared before each experiment by mixing AnalaR or Aristar grade concentrated nitric and hydrochloric acids (BDH Chemicals). Although experimental conditions were varied in evaluating the optimum extraction conditions, the standard procedure was the same. Standard Procedure for the Aqua Regia Leach Freshly prepared aqua regia (20 ml) was added to the rock powder (10.0 f 0.1 g) in borosilicate glass beakers (150 ml) and mixed using a polytetrafluoroethylene rod until com- pletely wetted. The beaker was covered with a watch-glass and the mixture stirred on a magnetic stirring table (together with other samples in the batch) for 1 h at room temperature.The contents of the beaker were then poured into a filter funnel fitted with a Whatman 40 filter-paper (55 mm) together with washings from both beaker and watch-glass cover, rinsed with the minimum volume of de-ionized water. After filtration, the filter-paper and residue were transferred into a polythene bag and sealed for subsequent analysis, as appropriate. The filtrate was transferred into a graduated flask (100 ml) and made up to volume with de-ionized water. Beakers and storage flasks were cleaned before use by soaking in laboratory cleaner overnight and then rinsing with de-ionized water.Between runs, other glassware was rinsed with 50% nitric acid (AnalaR) and then de-ionized water. Each batch of samples included a blank aqua regia control (processed in the same way as the samples) and all leach experiments were performed in duplicate. Analysis of Solutions The aqua regia leach solutions were analysed by ICP-MS (VG Elemental PlasmaQuad) using the Natural Environmental Research Council (NERC) facility then based at the Univer- sity of Surrey or the British Geological Survey (BGS) instrument then based in London. The operating conditions broadly followed those described by Jarvis31 and Gray and Williams32 with some minor amendments to take into account the aggressive nature of the sample solutions (20% aqua regia, and >O.1% total dissolved solids). The auxiliary gas flow rate was set at 0.5 1 min-l, the sample uptake rate being about 0.5 ml min-1. The instrument was set up using a nine element standard solution and the ion lens settings were optimized in order to maximize the 115In+ signal and minimize the doubly charged and oxide ion interferences as monitored by '40Ce2+ and 140CeO+. In order to match the matrix of the sample solutions, the standard solution was prepared with elemental concentrations of 100 pg 1-1 in 20% aqua regia (Aristar grade). This standard solution was run after a calibration blank (20% aqua regia) at the beginning of each batch and then repeated after at least every fifth sample. These data from the standard solution were used to apply a correction for drift in instrument sensitivity by interpolation of count data from adjacent standard solutions.In the course of this work, it was realized that additional interference effects were occurring owing to the suppression of signals in sample solutions containing particularly high contents of dissolved salt. An additional correction was then applied by adding Re and In (both at 100 pg I-* in the analysed solution) as internal standards to all sample and standard solutions. Data for Rh, Ru and Pd were normalized against the In signal and Os, Ir, Pt and Au against the Re internal standard signal, thus minimizing any mass-dependent effects. A blank solution (20% aqua regia) was nebulized for 70 s between sample solutions and 90 s between a drift correction standard (containing higher concentrations of the PGE) and a sample solution in order to avoid memory effects.Analysis of Residues A further series of measurements was carried out on the residue after the aqua regia leach, in order to determine the proportion of concomitant elements co-extracted by aqua regia and therefore present in the analyte solutions as potential interferences. After drying, aliquots of the residue were prepared as powder pellets and analysed by energy dispersive X-ray fluorescence (ED-XRF) , generally following the procedures described by Potts et aZ.33 The proportion of the element leached by aqua regia was calculated from the ratio of measured intensity data between the leached and unleached sample. No formal correction to these data was applied for matrix effects because of the very small fraction of matrix elements leached by the aqua regia.However, some compensation for drift and any residual matrix effects was applied by normalizing the intensity data against Cr counts for the current sample. Other aliquots of the residue were analysed by instrumen- tal neutron activation analysis (INAA) following the proce- dures described by Potts et ~ 1 . 3 ~ A calibration was undertaken relative to a sample of Ailsa Craig microgranite that had been spiked with standard solutions, mixed as a slurry, dried and lightly ground in order to disaggregate the resultant cake. The spiked concentrations in this standard were 1.02 pg g-1 Ir, 10.4 pg 8-1 Pt and Sb, 4.17 pg 8-1 Au and 104 pg g-l As.Neutron flux corrections were made by interpolating the response from two samples of chromitite CHR-C placed at either end of the irradiation can, each of which contained a total of 11 samples. Results and Discussion Optimizing the Aqua Regiu Leach A series of experiments was carried out to test the influence of various parameters on the aqua regia extraction efficiency. The parameters tested were contact time, heating and aqua regia composition. In order to simplify procedures, all experiments were carried out on 10 g of rock powder and 20 ml of aqua regia. Results were obtained both by the analysis of filtrate solutions (ICP-MS) and of residues (ED-XRF and INAA). Experiments were performed on chromitites CHR- A, CHR-B, CHR-C and the platinum ore SARM 7, represent- ing a silicate matrix. Varying the aqua regia composition between the limits of 1 + 3 and 3 + 1 HCl + HN03 appeared to have little influence on the extraction data and all subsequent measurements were made with a standard aqua regia mixture (3 + 1, HCl + HN03).The effect of contact time and heating on the aqua regia-sample reaction mixture are shown in Figs. 1-3. In Fig. 1, data are plotted for the proportion of PGE extracted from the mineralized chromitite CHR-C and the platimum ore SARM 7 at room temperature for a period from 15 to 120 min and also for a separate experiment in which the extraction mixture was heated for 120 min. The results are calculated as the proportion (%) of the element extracted, arbitrarily normalized to the maximum value in order to simplify the visual interpretation of the diagrams.It must be emphasized,ANALYST, AUGUST 1991, VOL. 116 cn 775 50 therefore, that the results do not represent the absolute extraction efficiency, which for several elements is substan- tially less than 100%. The data in Fig. 1 indicate that, for the samples invest- igated, the reaction with aqua regia proceeds rapidly at room temperature and that prolonging the leach time beyond about 30 min does not appear to improve significantly the proportion recovered. The data also indicate that significant proportions of 0 s and Ru are recovered by an aqua regia leach at room temperature. However, for these elements, there is some indication that the corresponding ICP-MS signal is reduced as the contact time is extended.This trend may be due to losses of 0 s and Ru from the analyte or an increased ICP-MS suppression effect if the dissolved salt content of the leachate solutions increases with aqua regia contact time. None of the elements measured showed an improved recovery efficiency, based on the ICP-MS signal obtained in experiments where the extraction mixture was heated for 120 min. The reason for this was not established unambiguously. However, one cause - 50s 0 150 100 50 0 I 0 s I I I L 30 60 90 120 Ti mehi n z 20 60 Tern peratu rePC Fig. 1 Proportion of ( a ) Pd, (b) Pt and (c) Rh extracted from 1, CHR-C and 2, SARM 7 and (d) proportion of 0 s and Ru extracted from CHR-C by aqua regia extraction at room temperature for specified times and for 120 min at specified temperatures.Data have been arbitrarilv normalized to the maximum measured value in order to simplify visial interpretation of the graphs. Analyses were obtained by ICP-MS 0 ' I I I I /+ 3 1 30 60 90 120 240 Time of aqua regia attacwmin Fig. 2 Proportion of ( a ) Ni, (b) Cu, (c) As, (d) Mn, (e) Fe, and (j) Zn remaining in the residue after aqua regia leach for specified extraction times at room temperature of chromitites 1, CHR-A; 2, CHR-B; and 3, CHR-C. Cu in CHR-A and As in CHR-A and CHR-B were below the detection limits. Extraction data are normalized to intensity data obtained from unleached samples. Analyses were made by XRF776 100 4 ANALYST, AUGUST 1991, VOL. 116 1 4 3 could be the reduced activity of aqua regia on heating due to the loss of volatile nitrosyl chloride or chlorine components or, alternatively, the suppression effect due to dissolution of other elements mentioned. Results from the analysis of the residue after aqua regia attack for the elements Ni, Cu, As, Mn, Fe and Zn (by ED-XRF) and Au, As, Sb, Ir, Fe and Co (by INAA) are plotted in Figs.2 and 3, respectively. The true proportion of the element remaining in the residue is plotted as a function of contact time. The data indicate that a high (85-95%) proportion of Au, As and Sb in CHR-C is leached by aqua regia. For Cu, Ni and Co significant proportions are extracted that vary in magnitude when data from different chromitites are compared. Only a very small (if not negligible) proportion of Mn, Fe, Zn and Ir is extracted. As can be concluded from the data in Fig.1, the extraction reaction appears to be completed within 15-30 min. loo[ 50 50 t 1 2 50 t 0 4 u 30 60 90 120 240 Time of aqua regia attacklmin Fig. 3 Proportion of (a) Au, As and Sb in CHR-C; ( b ) Ir in 1, CHR-A and 3, CHR-C; (c) Fe in 1, CHRA, 2, CHR-B and 3, CHR-C; and (d) Co in 1, CHR-A, 2, CHR-B and 3, CHR-C remaining in the residue after aqua regia leach for specified extraction times at room temperature. Au, As and Sb in CHR-A and CHR-B and Ir in CHR-B were below the detection limits. Extraction data represent the true proportion of element extracted. Analyses were made by INAA Extraction Efficiency An estimate of the true proportion of PGE and Au extracted by aqua regia was made by comparing the concentrations of elements leached from chromitite OU-CX (a sub-sample of CHR-C), with the best estimates of results obtained by fire assay from five commercial laboratories.The results listed in Table 1 show the average fire assay concentration and the corresponding proportion measured by aqua regia leach-ICP- MS. They indicate that the aqua regia leach of Pd and Au is about 100%; Pt and Rh, 18%; Ir, 1%; and Ru and Os, uncertain. This uncertainty in the 0 s and Ru data arises from the variability and paucity of results for these elements obtained from commercial laboratories. Application to Other Samples In order to evaluate the more general application of this procedure, the aqua regia leach technique was applied to a suite of samples that had been independently analysed by nickel sulphide fire assay with ICP-MS detection by a commercial laboratory.This suite comprised ophiolitic rocks, including 14 chromitites, three pyroxenites and two wehrlites; the results for which are listed in Table 2. Further comparisons are made in Table 3 in which aqua regia leach data are compared with certifiedrecommended values for a range of PGE reference materials including: PTA-1 (Platiniferous Black Sand), PTC-1 (Sulphide Concentrate), PTM-1 (Ni-Cu Matte), SARM 7 (Platinum Ore), GXR-1 (Jasperoid Soil), and GXR-4 (Copper Mill-head) and SU-la (Ni-Cu-Co Ore). Reviewing these data reveals significant variability for a particular element over a range of different samples. However, some general observations on trends in these data can be made. (i) Although a few samples show high recoveries of Pt by aqua regia leach, the proportion extracted varies significantly from sample to sample and is usually less than 2040%.(ii) The proportion of acid-leachable Pd appears to be high, often near-quantitative, based on an over-all evalua- tion of the data listed in Table 2 (ophiolitic samples). The results of Pd determinations on reference samples are more variable, particularly low recoveries being obtained for PTC-1 and PTM-1. Very high recoveries (>> 100%) were measured from GXR-1 and GXR-4, although there is some uncertainty as to the reliability of the compiled values with which the measured data are compared. (iii) Significant proportions of Rh and Ru were recovered from samples, although in general the proportion is less than 30% of the expected composition.(iv) Although some 0 s is extracted, it is difficult to judge the over-all trend for this element owing in part to determinations falling below the detection limit of the technique used here and in part to the absence of 0 s data for reference samples. (v) The proportion of Ir extracted by aqua regia is low, usually significantly less than 10%. (vi) In general recoveries of Au are high and often near-quantitative, a notable exception being data for PTM-1. (vii) Over-all recoveries are very low for sulphur-rich samples and clearly the extraction technique used here is not suitable for such materials without extensive modification. It may be that this problem is associated with the oxidation of sulphide sulphur by aqua regia leading to a loss in the potency of the acid andor high base metal concentrations in solutions, leading to interferences in the ICP-MS measure- ments.Interpretation of Results Excluding analytical discrepancies, two factors, both related to sample composition, might influence the proportion of PGE extracted by aqua regia leach. The first is whether a proportion of the elements is distributed in the solid solution or is present as discrete mineral grains embedded within the resistant matrix phases (such as chromite) and are notANALYST, AUGUST 1991, VOL. 116 777 ~ ~ ~ ~ ~ ~ ~~~~~~ ~~~~~ ~ Table 1 Composition of mineralized chromitite (OU-CX) in ppm and % recovery of PGE by aqua regia leach of this sample. Bulk composition data are the best estimates derived from five commercial laboratories, obtained from a round robin analytical test organized by Rio Tinto Zinc.Aqua regia recovery is the proportion (%) of the element recovered in the filtrate after a 60 min aqua regia extraction on a 10 g sample at room temperature with measurements made by ICP-MS Sample Pt Pd Rh Ru 0 s Ir Au Bulk composition 40.0 53.0 3.7* - - 6.0* 2.7 Aqua regia recovery 18 104 18 uncertain? uncertain? 1 110 * Additional analytical uncertainty due to the variability of reported results. I Recovery efficiencies of Ru and 0 s are uncertain owing to the lack of data reported by commercial laboratories. Table 2 Comparison of the extraction efficiency of the aqua regiu leach (this work) with samples independently analysed by nickel sulphide fire assay-ICP-MS; all data listed in ppb (whole rock) Pt Pd Rh Ru 0 s Ir Au Sample Chromitite 1" Chromitite 2* Chromitite 39 Chromitite 4§ Chromitite 54 Chromitite 64 Chromitite 7§ Chromitite 8§ Chromitite 94 Chromitite 107 Chromitite 117 Chromitite 127 Chromitite 134 Chromitite 149 Pyroxenite 19 Pyroxenite 29 Pyroxenite 34 Wehrlite 19 Wehrlite 28 This Fire work assay 200 800 110 110 30 83 200 510 17 52 1390 4050 --$ 16 8 51 130 500 230 300 7290 40000 210 280 1140 870 7 9 120 110 29 44 65 100 --$ 48 22 22 This Fire work assay 730 1300 490 1670 36 66 300 340 50 83 3600 3150 --$ 11 --$ 18 450 520 51000 53000 226 220 2200 2100 8 10 --$ 4 40 62 180 220 22 18 37 48 --t - This Fire work assay 216 470 191 610 6 20 79 120 3 13 159 605 1 9 4 46 6 55 --t 662 3700 -1- 87 110 --$ 6 --$ 3 3 9.5 --$ 4 --$ 2.5 - - 1 5.5 This Fire work assay 85 960 76 1200 29 110 460 390 --$ 69 250 1150 --$ 45 --$ 310 35 190 -1- --t --t 290 190 74 44 --$ 26 --$ 4 12 13 --$ 19 --$ 5.5 - - - This Fire work assay --t -t --$ 42 9 88 --$ 34 --$ 675 --$ 24 --$ 160 6 50 -? --t --t 45 98 --$ 12 --$ 4 --$ 4 --4 6 --$ 2 --f 4 - - - - - This Fire work assay 30 380 16 --$ 9 57 18 250 5 50 29 1550 --$ 25 2 180 3 97 -? 70 12000 18 3000 17 120 --$ 14 --$ 4.5 --$ 2.5 4 5.5 --$ 5 --$ 4 - * Samples and analyses for fire assay from R.V. D. Robert, MINTEK, Randberg, South Africa. 7 Not analysed by NiS fire assay. -$ Below the detection limit. 4 Fire assay analyses by Sheen Analytical, Perth, Australia. fl in-house reference samples (Open University); best estimate data derived from five commercial laboratory analyses.This Fire work assay 30 70 13 50 -t --t --t --t -? --t 2 6 2 30 2960 2700 --t 155 130 --$ 2 16 8 --f 6 34 28 -? -4 2 - - - - - - - - Table 3 Comparison of aqua regia leach-ICP-MS data (this work) with expected compositions (given value) of selected reference samples; all data listed in ppb (whole rock) Sample Pt Pd Rh Ru 0 s Ir Au This Given This Given This Given This Given This Given This Given This Given work value work value work value work value work value work value work value Platinum Ore, Platiniferous Black Sulphide Concentrate, Ni-Cu Matte, Ni-Cu-Co Ore, Jasperoid Soil, Copper Mill-head, SARM7(PTO-l)* 1400 3740 1180 1530 174 240 115 430 9 63 19 74 302 310 Sand, PTA-1-t 3320 3050 16 --$ 5 --$ -3 --$ -$ --$ 15 -$ 363 --$ PTC- 17 210 3000 6010 12700 71 620 171 65011 49 2401) -4 17011 136 650 PTM- 1 * * 45 5800 5 8100 132 900 294 59011 31 14011 44 47011 -4 1800 SU-lat-t 59 410 265 370 16 8011 29 5611 -§ 1111 -4 2511 105 160(( GXR-l-$-$ -9 <lo 181 <0.1 12 --$ -§ --$ -5 --$ -§ --$ 2830 3300 GXR-4$$ -9 <10 55 0.2 2 --$ -5 --$ -9 --$ -9 --$ 618 470 * SARM 7: SABS certified values from Steele et al.29 t PTA-1: Canadian Certified Reference Material Project recommended values from McAdam et af.35 $ No reported data.9 Below the detection limit. 7 PTC-1: Canadian Certified Reference Material Project recommended values from McAdam et uf.36 11 Data are additional non-certified values. ** PTM-1: Canadian Certified Reference Material Project recommended values from McAdam et al.37 I-? SU-la: Canadian Certified Reference Material Project recommended values from Steger and B0wman.3~ $$ GXR-1 and GXR-4: compiled values from Gladney and R ~ e l a n d t s .~ ~778 ANALYST, AUGUST 1991, VOL. 116 Table 4 PGE mineralogy of a chromitite from the Cliff area, Unst, Shetland, summarizing aqua regia solubility data Conclusion The results obtained in this work indicate that high, and for ophiolitic rocks, near-quantitative recoveries of Pd and Au are achieved by a room temperature aqua regia leach of l o g samples using 20 ml of acid. Heating the aqua regia leach mixtures led to a reduction in the PGE signal in ICP-MS measurements from sample solutions. Significant but lower recoveries (usually 2040%) of Pt, Rh, Ru and 0 s were also observed but the recovery of Ir was much lower (generally 1-10%).Low recoveries of all the elements were observed from sulphur-rich samples. In assessing the analytical perfor- mance, account must be taken of other elements that are co-extracted (including As, Sb, Ni and Cu from chromitites), which might cause interference effects in analyte solutions. Comparison of the recovery efficiency data with a detailed PGE mineralogical description available for one mineralized chromitite sample indicated that the chemical solubility of individual PGE mineral types could be the dominant factor in dictating recovery efficiencies from ophiolitic samples. Dominant element Mineral* Pt Sperrylite Genkinite Hongshiite Alloy Alloy Stibiopalladinite Potarite Alloy Rh Hollingworthite Unidentified Unidentified Unidentified Rut henian Pd Mertieite Ir Irarsite Ru Laurite pentlandite 0 s Native metal Iridosmine Formula PtAs, (Pt,Pd),Sb3 PtCuAs P t-Pd-Cu Pt-Pd-Au-Cu (Pd,Cu)g(Sb,As)3 PdHg AuPd (Rh,Pt,Pd)AsS Rh(Sb,S) (Rh ,Ni)Sb (Ir,Ru,Rh,Pt)AsS I r (S b , S) (Ru,0s,Ir)S2 (Ni ,Fe, Ru)$& (Pd,Cu)dSb,Ash 0 s Os,Ir Solubility in aqua regia No - - - - - Yes - - - - - - - No - - No * Other minerals reported to be soluble in aqua regia: native gold, platinum and palladium. Minerals reported to be insoluble in aqua regia: cooperite (Pt,Pd,Ni)S; braggite (Pt,Pd,Ni)S; osmiridium (Ir,Os); and platiniridium (Ir,Pt).therefore effectively wetted by aqua regia. The second factor relates to the chemical solubility of individual PGE mineral species in aqua regia. In respect of chromitite CHR-C, the detailed mineralogical descriptions of Prichard et aZ.25 have characterized a variety of PGE mineral types found in this sample.These mineral types are listed within each element category in decreasing fre- quency of observation in Table 4. These data indicate, for example, that the dominant Pt minerals in CHR-C are sperrylite and genkinite, with hongshiite and alloy phases being observed less frequently. A literature survey has revealed some data on the solubility of specified mineral types in aqua regia40.41 and these are also listed in Table 4. Evaluation of the data indicates that there is some correla- tion between the recovery efficiencies listed in Table 1 and the aqua regia solubility data in Table 4. Thus, high recovery efficiencies of Pd appear to correlate directly with the dominant presence of this element in the mineral stibiopalladi- nite and mertieite.The former mineral is reported to be soluble in aqua regia, as might be the latter in view of the similarity in chemical composition. Conversely, the dominant mineral containing Pt in chromitite CHR-C is sperrylite, a mineral that is insoluble in aqua regia, an observation that is consistent with the much lower recovery efficiency of this element (18% in OU-CX). Finally, although information is not available for irarsite, it appears significant that other Ir bearing minerals including osmiridium, iridosmine, platiniri- dium and laurite are all insoluble in aqua regia, so accounting for the very low recovery (1%) of Ir by aqua regia leach of this chromitite. Although these observations relate only to one sample, and discrepancies due to non-wetting by aqua regia of PGE source minerals cannot be discounted, the results described here suggest that the solubility of individual PGE mineral types may be the limiting factor in the more general application of the aqua regia leach method for the determination of the PGEs.The authors are grateful to NERC, Alan Gray, Kym Jarvis (Surrey) and Yuk Cheung (BGS) for the use of the ICP-MS facilities and John Williams and Ed McCurdy for assistance in running the instrumentation, and to Hazel Prichard, Richard Lord, John Bridges of the Open University (OU), Chris Morrissey (Rio Tinto Zinc) and Rob RobCrt (MINTEK) for the loan of samples and comparative data. Support from the Minerals Industry Research Organisation (MIRO) is also gratefully acknowledged.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 References Smith, E. A., The Sampling and Assay of the Precious Metals, Revised 2nd Ed., Met-Chem Research, Boulder, Colorado, 1987. Haffty, J., Riley, L. B., and Goss, W. D., U.S. Geol. Surv. Bull., 1977, No. 1445. Van Loon, J. C., Pure Appl. Chem., 1977,49, 1495. Bowditch, D. C., Aust. Min. Dev. Lab. Bull. (AMDEL), No. 15, 1973, p. 71. Hall, G. E. M., and Bonham-Carter, G. F., J . Geochem. Explor., 1988,30, 255. RobCrt, R. V. D., Van Wyk, E., and Palmer, R., Nut. Znst. Metall. Repub. S. Afr. Rep., 1971, No. 1371. Hoffman, E. L., Naldrett, A. J., VanLoon, J. C., Hancock, R. G. V., and Manson, A., Anal. Chim.Acta, 1978, 102, 157. Asif, M., and Parry, S. J., Analyst, 1989, 114, 1057. Beevers, J. R., Econ. Geol., 1967, 62, 426. Strong, M. B., and Murray-Smith, R., Talanta, 1974,21, 1253. Terashima, S., Geostand. Newsl., 1988, 12,57. Rivoldini, A., and Haile, T., At. Spectrosc., 1989, 10, 89. Latimer, W. M., and Hildebrand, J. H., Reference Book of Inorganic Chemistry, Revised edn., The Macmillan Company, New York, 1940, p. 206. Sen Gupta, J. G., Talanta, 1989, 36, 651. Trancoso, M. A., and Barros, J. S., Analyst, 1989, 114, 1053. Alvarado, J., and Petrola, A,, J. Anal. At. Spectrom., 1989, 4, 411. Grimaldi, F. S., and Schnepfe, M. M., U.S. Geol. Surv. prof. pap. 1967,575-C, C141. Grimaldi, F. S., and Schnepfe, M. M., U.S. Geol. Surv. prof. pap. 1968,600-B, B99. Palmer, I., Streichert, G., and Wilson, A., Nut.Znst. Metall. Repub. S. Afr. Rep., 1971, No. 1218. Palmer, I., Palmer, R., and Steele, T. W., J. S. Afr. Chem. Znst., 1972, 25, 190. Fryer, B. J., and Kerrich. R., At. Absorpt. Newsl., 1978, 17,4. Sighinolfi, G. P., Gorgoni, C., and Mohamed, A. H., Geostand. Newsl., 1984, 8, 25. Sen Gupta, J. G., and Gregoire, D. C., Geostand. Newsl., 1989, 13, 197. Kontas, E., Niskavaara, H., and Virtasalo, J., Geostand. Newsl., 1990, 14, 477.ANALYST, AUGUST 1991, VOL. 116 779 2.5 26 27 28 29 30 31 32 33 34 35 Prichard, H. M., Potts, P. J., Neary, C. R., Lord, R. A., and Ward, G. R., European Communities Commission, Luxem- bourg Rep., 1989, No. CD-NA-11631-EN-C. Tarkian, M., and Prichard, H. M., Miner. Deposita, 1987, 22, 178. Prichard, H. M., and Tarkian, M., Can. Mineral., 1988,26,979. Lenahan, W. C., and Murray-Smith, R. de L., Assay and Analytical Practice in the South African Mining Industry, The South African Institute of Mining and Metallurgy, Johannes- burg, 1986, p. 228. Steele. T. W.. Levin, J . , and Copelowitz, I . , Nat. Inst. Metall., Repub. S. Afr. Rep., 1975, No. 1696. Potts, P. J . , and Govindaraju, K., Geostand. Newsl., 1989, 13, 193. Jarvis, K. E., Chem. Geol., 1988, 68, 31. Gray, A. L., and Williams, J . G., J. Anal. At. Spectrom., 1987, 2, 599. Potts. P. J., Webb, P. C., and Watson, J . S . , X-ray Spectrom., 1984, 13, 2. Potts, P. J . . Thorpe, 0. W., and Watson, J . S . , Chem. Geol., 1981,34, 331. McAdam, R. C., Sutarno, and Moloughney, P. E., Dept. of Energy, Mines and Resources, Mines Branch, Ottawa Tech. Bull., 1971, No. TB138. 36 McAdam, R. C., Sutarno, and Moloughney, P. E.. Dept. of Energy, Mines and Resources, Mines Branch, Ottawa Tech. Bull., 1973, No. TB176. 37 McAdam, R. C., Sutarno, and Moloughney, P. E., Dept. of Energy, Mines and Resources, Mines Branch, Ottawa Tech. Bull., 1973, No. TB182. 38 Steger, H. F., and Bowman, W. S., Canadian Certified Reference Materials Project, CANMET Rep. , 1980, No. 80-9E. 39 Gladney, E. S . , and Roelandts, I . , Geostand. Newsl., 1990, 14, 21. 40 Schoeller, W. R., and Powell, A. R., The Analysis of Minerals and Ores of the Rarer Elements, 3rd edn., Charles Griffin, London, 1955, p. 32.5. Powell, A. R., in Comprehensive Analytical Chemistry, Volume I c, Classical Analysis: Gravimetric and Titrimetric Determina- tion of the Elements, eds. Wilson, C. L., and Wilson, D. W., Elsevier, Amsterdam, 1962, p. 702. 41 Paper 1 l00989C Received March 4th, 1991 Accepted April 5th, 1991