1050 Analyst, November, 1979, Vol. 104, pp. 1050-1054 X-ray Fluorescence Determination of Platinum and Palladium in Platinum Concentrates Using a Solution Technique Z. Cruickshank and H. C. Munro Johannesburg Consolidated Investment Company Limited, Minerals Processing Research Laboratory, P.O. Box 13017, Knights, Transvaal 1413, South Africa An X-ray fluorescence solution technique for the determination of platinum and palladium in platinum-bearing material is described. Ruthenium has to be removed prior to the measurement of the platinum and palladium. Mercury and thorium are used as internal standards. The method is precise and is more rapid than the gravimetric method normally used. Keywords : X-ray fluorescence spectrometry ; platinum determination ; palladium determination ; mercury and thorium internal standards ; platinum concentrates An accurate X-ray fluorescence spectrometric technique was required for the determination of platinum and palladium in a variety of platinum-bearing materials. A literature survey yielded few instrumental methods.lS2 Owing to the complexity and heterogeneity of these materials, it was decided to use a solution technique to obviate possible mineralogical and particle size effects.Absorption effects would be countered by the use of internal standards. It is possible by correct planning and scheduling to analyse concurrently large numbers of samples for platinum and palladium, thus halving the analysis time normally required for the determination of these elements by the gravimetric procedure.This method does not require pre-analysed samples to be used as standards, but uses pure platinum and palladium solutions covering the calibration range required. Experimental Dissolution The silica content of the samples varied widely, in some instances being as high as 20%, which made treatment with hydrofluoric acid advisable. This was carried out in PTFE beakers either on a steam-bath or on a low-temperature hot-plate. The hydrofluoric acid treatment was followed by leaching of the residue with aqua regia and subsequent filtration through a Millipore filter. The resultant residue was scanned on an X-ray fluorescence spectrometer and found to contain appreciable amounts of platinum group metals, which made the dissolution of the residue necessary. It was also found that the mass of residue varied considerably from sample to sample.The residue was fused with sodium peroxide, leached and added to the aqua regia solution. Choice of Internal Standards The above solution was scanned on an X-ray fluorescence spectrometer so that suitable internal standards could be chosen. The main criteria for choosing an internal standard are that it must not be present in the sample, it must have an X-ray fluorescence line whose energy is similar to that of the line being analysed, and no absorption edges may fall between these two lines except absorption edges generated by the two elements concerned. According to the above criteria, mercury was chosen as the internal standard for platinum. A few trial solutions were prepared and the optimum mercury concentration was deter- mined.The choice of an internal standard for palladium was more difficult. On the short wave- length side, the spectrum was very crowded and full of absorption edges, the ruthenium KP line overlapped the palladium Ka line, and no suitable internal standard or background position could be found. On the long wavelength side, there were the rhodium Ka and ruthenium K a lines. If the ruthenium was removed, both the ruthenium Ka line, and theCRUICKSHANK AND MUNRO 1051 ruthenium KP line interference on the palladium Ka line, would disappear. This would have made niobium a good internal standard for palladium if the niobium K/3 line was used. However, it was found that the preparation of a 40 g 1-1 niobium solution was difficult and, as an alternative, thorium was chosen.Although thorium dissolved easily, it was found that on mixing with the sample solution a crystalline precipitate formed. This was overcome by vigorously boiling the thorium solution for 1-2 min during the original dissolu- tion process. This stabilised the thorium standard solution and precipitation no longer occurred on mixing with the sample solution, even when this mixture was allowed to stand for over 2 weeks. The usual method of removing ruthenium is by the addition of sodium bromate in the presence of sulphuric acid. Although this works well, the sulphuric acid causes precipitation of thorium and this method is therefore unsuitable. Fuming perchloric acid was also tried, but proved to be time consuming. The most suitable method was found to be the stepwise addition of sodium bromate and hydrochloric acid,3 which resulted in the complete removal of ruthenium, provided that no nitric acid was present in the sample.The analytical line used was the thorium Ly,. Proposed Method Samples and 11-19yo of palladium were analysed. Samples of platinum-bearing concentrate containing approximately 16-35y0 of platinum Reagents All reagents used were of analytical-reagent grade. Hydrojuoric acid, sp. gr. 1.16. Hydrochloric acid, sp. gr. 1.19. Nitric acid, sp. gr. 1.40. Sodium peroxide. Granular. Sodium bromate solution. Mercury internal standard solution. Tlzoriztm internal standard solution. Platinum standard solution. Palladium standard solution. A 10% solution in de-ionised water.Dissolve 30 g of mercury(I1) chloride in 1 1 of 3 M Dissolve 200 g of Th(N03),.(4-6)H20 in 1 1 of 3 M Prepared by dissolving Specpure platinum sponge in aqua Prepared by dissolving Specpure palladium wire in aqua hydrochloric acid. hydrochloric acid. regia to give solutions containing 6, 9, 13, 16 and 19 g 1-l. regia to give solutions containing 4, 5, . . ., 10 g 1-1. The solution was boiled vigorously for 1-2 min after preparation. Apparatus Philips PW 1450/20 sequential X-ray jhorescence spectrometer. sample changer and related equipment were used. Millipore filter apparatus with a 0.45-pm membrane. Vitreous carbon crucibles with 30-cm3 capacity. A 60-position automatic Any similar instrument could be used. Procedure Samples (5 g) were weighed in triplicate into PTFE beakers and 15 cm3 of hydrofluoric acid were added.The PTFE beakers were placed on a low-temperature hot-plate and the samples dried; if large amounts of silica were known to be present the dried sample was re-dissolved in another 15 cm3 of hydrofluoric acid and the drying step repeated. Aqua regia (80 cm3) was then added, the samples were digested on a steam-bath for approximately 2 h and, after cooling, filtered through a Millipore filter apparatus. The filtrate was evaporated to dryness on a steam-bath, 15 cm3 of hydrochloric acid were added and the beakers returned to the steam-bath to dissolve the residue from the evaporated filtrate. The Millipore membrane was transferred into a carbon crucible, dried and ignited at 800 "C. Sodium peroxide (4 g) was added to the cool crucible containing the residue and the sample was carefully fused.The fusion mixture was leached with approximately 80 cm3 of distilled water in a 400-cm3 squat-form beaker and the carbon crucible removed. The hydrochloric1052 CRUICKSHANK AND MUNRO: X-RAY FLUORESCENCE DETERMINATION OF Analyst, Vd. 104 acid filtrate was also transferred into the same beaker, the resultant solution tested with pH paper to ensure that it was acidic and if necessary a few more drops of hydrochloric acid were added. The solution was brought to the boil, and 60 cm3 of 10% sodium bromate solution were slowly added to the boiling solution, in 5-cm3 aliquots, followed by 50 om3 of concentrated hydrochloric acid added in the same manner. When the addition of the sodium bromate and the hydrochloric acid was complete, the solution was boiled to reduce the volume, cooled and diluted to volume in a 100-cm3 cali- brated flask.Two 10-cm3 aliquots of the solution (one for the platinum determination and the other for the palladium determination) were pipetted into 100-cm3 beakers. Mercury(I1) chloride solution (10cm3) was pipetted into the beaker containing the aliquot for the platinum measurement and thorium nitrate solution (10 cm3) was pipetted into the beaker containing the aliquot for the palladium determination. The platinum and palladium were measured by X-ray fluorescence spectrometry using the Feather and Willis* background-correction method. According to this method, background intensities at differing wavelength positions are linearly related provided that they fall between adjacent major element absorption edges.This allows for the determination of the background by a single measurement at an interference-free background position. Background-correction calibrations may be set up at the beginning of a run by using blanks. The background-correction calibration will be of the form y = mx + c , where y is the blank reading at the peak position, x the blank reading at the background position and c and m are constants. The choice of blanks for the platinum and mercury background corrections is relatively easy. At least two solutions must be used, one with a light matrix and the other with a heavy matrix such that the background intensity of the sample falls between those of the two blanks.De-ionised water and a 100 g 1-1 nickel solution would be suitable blanks for the platinum and mercury background corrections. The choice of blanks for the palladium and thorium background corrections is, however, more difficult, in that a thorium absorption edge falls between the background position and the palladium and thorium lines. For the palladium background correction, this is over- come by adding thorium to the blanks in the same ratio as to the samples. The background correction is thus always made across the thorium absorption edge, and as the amount of thorium is always constant, no matrix variations due to thorium can occur. A 1 + 1 mixture of de-ionised water and the thorium standard solution, plus a 1 + 1 mixture of a 9Og1-1 platinum solution and the thorium standard solution, would be suitable blanks for the palladium background correction.For the thorium background correction de-ionised water and a 90 g 1-1 platinum solution could be used provided that the water is “infinitely thick” to thorium radiation. Alternatively, instead of water, a slightly heavier matrix can be used, e.g., a nickel solution of sufficient concentration to be “infinitely thick” to thorium radiation. This will depend on the geometry and capacity of the sample holders used. Tube* .. . . .. Counter . . .. .. Collimator . . . . . . Discriminator setting . . Counting time, peak and background/s . . .. Peak angle, “28 . . .. Background angle, “219 . . Voltage/kV . . .. .. CurrentlmA .. .. Vacuum . . .. .. Counter H.T./V .. .. Spinner . . .. . . Crystal . . .. .. TABLE I INSTRUMENTAL PARAMETERS Platinum determination Palladium determination Mo LiF (220) Scintillation Fine (160 pm) LL = 200, window = 200 2 x 20 Hg La = 51.69; Pt La = 54.92 67.70 80 36 Off -1040 On Au LiF (220) Scintillation Fine (160 pm) LL = 200, window = 200 2 x 20 Pd Ka = 23.77; Th Lyl = 26.56 28.70 80 35 Off -1040 On * It was not possible to use the same tube for both platinum and palladium determinations owing to line interferences.November, 1979 Pt and Pd IN Pt CONCENTRATES USING A SOLUTION TECHNIQUE 1053 As the thorium content is always constant, the thorium absorption edge between the back- ground position and the thorium line can be disregarded. The instrumental parameters used in the determinations are shown in Table I.Results Values obtained by the X-ray fluorescence spectrometry solution technique compared well with values obtained by the two other laboratories using wet-chemical procedures (see Table 11). TABLE I1 X-RAY FLUORESCENCE VALUES AND VALUES OBTAINED FROM TWO OTHER LABORATORIES USING WET-CHEMICAL PROCEDURES Sample No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16 16 17 18 19 20 21 22 23 24 26 Average . . Platinum determined, yo Laboratory Laborator; X-ray value No. 1 No. 2 17.70 17.88 17.61 16.64 16.84 16.40 23.77 24.31 24.17 26.49 27.44 27.63 28.97 29.23 29.04 30.32 30.50 30.37 27.80 27.25 27.05 22.14 22.62 22.26 24.93 24.98 25.12 24.62 24.96 24.90 24.18 24.25 24.24 28.16 28.52 28.28 28.34 28.52 28.07 24.97 25.19 24.89 31.32 28.11 31.51 28.86 28.54 28.37 26.12 26.37 26.20 28.07 27.47 27.96 25.05 25.05 25.17 27.86 27.96 28.17 29.30 29.49 29.21 35.74 35.21 35.11 29.97 29.87 30.13 32.32 32.17 32.77 32.32 32.23 32.38 Palladium determined, yo r A > Laboratory Laboratory X-ray value No.1 No. 2 13.40 13.90 13.68 14.11 14.67 14.34 12.58 13.20 12.93 13.89 14.64 14.40 13.92 14.28 13.89 11.86 12.14 11.76 12.91 13.50 13.23 15.40 15.81 15.62 14.39 14.57 14.70 14.52 14.78 14.81 14.33 14.10 14.20 14.43 14.75 14.50 15.67 16.12 15.48 15.23 15.59 16.39 14.87 15.35 14.92 15.69 16.08 15.99 16.08 16.20 16.20 16.22 16.10 16.25 16.37 16.52 16.32 15.86 15.87 16.12 15.22 15.31 15.26 13.62 13.87 13.63 17.82 18.36 18.11 18.81 19.25 18.67 19.09 19.78 19.33 . . 27.04 27.00 27.08 15.05 15.40 16.17 Statistical comparisons of the results indicate that while there is no significant difference at the 0.05 level for platinum, there is, however, a significant difference for palladium between the three laboratories (see Table 111).The sum of squares (for X-ray fluorescence spectrometric results and results for labora- tories 1 and 2) was decomposed into two contrasts, one comparing laboratory 1 and laboratory 2, and the other comparing the X-ray fluorescence spectrometry technique with the mean values obtained from laboratory 1 and laboratory 2.5 As expected, the t values obtained from the contrasts for platinum were not significant, but for palladium the t values obtained from the contrasts were significant. It can be seen that the reproducibility of the palladium results obtained by laboratory 1 and laboratory 2, is of the same order as the reproducibility of results obtained by the X-ray fluorescence spectrometry technique and the mean of the results of laboratory 1 and laboratory 2. The removal of the ruthenium is time consuming, but with proper planning large numbers of samples can be treated simultaneously, whereas the wet-chemical procedure as used by this laboratory requires, after the dissolution of the sample, the separation of the platinum group metals and there- fore only a limited qumber of samples can be conveniently handled.The method, as described, is reasonably fast and accurate.1054 (1) Platinum Analysis of variance- Source of variation CRUICKSHANK AND MUNRO TABLE I11 STATISTICAL COMPARISON OF RESULTS Sum of squares Laboratories 0.084 Samples .. .. . . .. . . 1274.136 Residuals . . .. .. .. .. 9.971 Total . . . . .. .. .. .. 1284.191 (X-ray method and laboratories 1 and 2) Contrasts- Laboratory 1 versus laboratory 2 . . .. . . . . laboratory 2 . . . . . . .. . . .. X-ray method versus mean of laboratory 1 and (2) Palladium Analysis of variance- Sum of Source of variation squares Laboratories 1.553 Samples .. .. .. . . . . 222.673 Residuals . . .. .. .. . . 1.044 Total . . .. .. .. .. .. 226.271 (X-ray method and laboratories 1 and 2) Contrasts- Laboratory 1 versus laboratory 2 . . .. . . laboratory 2 . . .. .. .. .. .. X-ray method versus mean of laboratory 1 and Degrees of freedom 2 24 48 74 t value - 0.636 - 0.008 Degrees of freedom 2 24 48 74 t value 5.387 - 6.508 Mean square F value 0.042 0.202 53.089 255.563 0.207 - - - Significance at 0.05 level Not significant Not significant Mean square F value 0.776 35.696 9.278 426.342 0.021 - - - Significance a t 0.05 level Significant Significant The authors thank Mr. M. Laws, Matthey Rustenburg Refiners, Wadeville, Mrs. B. Fourie and Mr. D. Nicolas for their assistance in the development of the method, Dr. D. Hawkins for his assistance with the statistical evaluation of results and the Management of Johannes- burg Consolidated Investment Company (Pty) Ltd. for permission to publish this paper. References 1. Austen, C. E., and Steele, T. W., “The Determination by X-Ray Fluorescence Spectrometry of Noble and Base Metals in Matte-Leach Residues,” Report No. 1912, Laboratory Method No. 78/26, National Institute for Metallurgy, Johannesburg, September 1977. Shestakov, V. A., Arkhipov, N. A., Makarov. D. F., and Kukushkin, Yu. N., J. Analyt. Chem. USSR, 1974, 29, 1872. Schoeller, W. R., and Powell, A. R., “Analysis of Minerals and Ores of the Rarer Elements,” Second Edition, Griffin, London, 1940, pp. 240-296. Feather, C. E., and Willis, T. P., X-Ray Spectrom., 1976, 5, 41. Chatfield, C. C . , “Statistics for Technology,” Chapman and Hall, London, 1970. 2. 3. 4. 5. Received March l6th, 1979 Accepted June Sth, 1979