ANALYST, JANUARY 1986, VOL. 111 15 Simultaneous Determination, by Hydride Generation and Inductively Coupled Plasma Atomic Emission Spectrometry, of Arsenic, Antimony, Selenium and Tellurium in Silicate Rocks Containing the Noble Metals and in Sulphide Ores L. Halicz" Geological Survey of Israel, 30 Malkhe Yisrael St., Jerusalem, Israel and G. M. Russell Council for Mineral Technology, Private Bag X30 15, Randburg 2 125, Republic of South Africa A method is described for the continuous generation of the hydrides of arsenic, antimony, selenium and tellurium and the simultaneous determination of these elements by an atomic emission spectrometer with a 5-kW nitrogen - argon inductively coupled plasma source. After digestion of the sample, the analytes are separated from the matrix by coprecipitation with iron(ll1) hydroxide at pH 2.40.A standard additions technique is used for quantification. This method is relatively free from interferences and is applied to the determination of arsenic, antimony, selenium and tellurium in silicate rocks containing gold and the platinum-group metals (noble metals), and also in sulphide ores rich in transition metals and lead. Keywords: Hydride generation; inductively coupled plasma atomic emission spectrometry; arsenic, antimony, selenium and tellurium determination; silicate rocks; sulphide ores The purpose of this investigation was the development of a method for the simultaneous determination of trace amounts of arsenic, antimony, selenium and tellurium in complex geochemical and industrial matrices. The simultaneous determination by inductively coupled plasma atomic emission spectrometry (ICP-AES) of these elements after the generation of their hydrides was first described by Thompson and co-workers,1-3 who established the best compromise conditions for the generation of the hydrides from 5 M hydrochloric acid using a 1% solution of sodium tetrahydroborate(II1).Wolnik et aZ.4 and Nygaard and Lowry5 also described methods for the simultaneous and sequential determination of the volatile hydride-forming elements. It is well known that arsenic, antimony, selenium and tellurium commonly exist in solution in two oxidation states. Unfortunately, reduction with tetrahydroborate(II1) is not efficient for these metals in high oxidation states. Potassium iodide has been used as a pre-reducing agent before final production of the hydrides with sodium tetrahydrobor- ate(III).s7 In the determination of total selenium and tellurium,l the addition of potassium bromide377 and boiling of the sample in 4 M hydrochloric acids10 for pre-reduction has been found to be beneficial.Almost all publications on hydride generation mention possible inter-elemental interferences. Smith12 carried out the first systematic study of the effects of 48 elements on the determination of hydride-forming elements. Smith's findings12 and those of Thompson et a1.2 indicate that interference effects are less severe in the determination of arsenic and antimony but that, in the determination of selenium and tellurium, the analytes must usually be sepa- rated from interfering elements such as the transition metals, gold and the platinum-group metals. This can be effected partly by the coprecipitation of the analytes on lanthanum hydroxide (a method devised by Bkdard and Kerbyson13) or by the coprecipitation of the analytes with iron(II1) hydrox- ide.14915 A variety of procedures for the digestion and * To whom correspondence should be addressed.decomposition of the sample are described in the literature as being satisfactory. Dissolution techniques are considered extremely important because losses of the analytes may occur.1G18 Experimental Apparatus Continuous hydride generation was accomplished by the use of three channels of a four-channel peristaltic pump (Gilson Instrument Co., Minipuls 11). The sample solution, sodium tetrahydroborate(II1) reagent and potassium iodide solution were delivered to a modified separator based on the design of Thompson et aZ.1 and similar to that of De Oliveira et aZ.19 A schematic representation of the hydride-flow manifold, which is similar to that of Nygaard and Lowry,S is shown in Fig.1. The gaseous hydrides and the hydrogen are swept from the separator into the plasma by a continuous flow of argon. I 1 7 Fig. 1. H dnde-generation manifold. (1) Flow-rate of argon, 2.1 I min-1; (2; flow-rate of sam le, 9.5 ml min-1; (3) flow-rate of 1% NaBH4 in 0.1 M NaOH, 4.7 mfmin-1; (4) flow-rate of 10% KI in 20% HCI, 9.5 ml min-1; 5 , gas exit to plasma torch; (6), phase separator; and (7), drain16 ANALYST, JANUARY 1986, VOL. 111 A Hilger Analytical Polyvac El000 spectrometer, with a Radyne R50P r.f.generator (frequency 27.12 MHz) was used for the detection of the hydrides of the elements. An Orion pH meter (Research 501) with an Orion glass electrode was used for the adjustment of the pH of the solution. Table 1. Experimental parameters Torch . . . . . . . . . . . . Incident power . . . . . . . . Coolant outer gas (nitrogen) . . . . Intermediate gas (argon) . . . . . . Carrier gas (argon) . . . . . . . . Observation height . . . . . . . . Wavelength: As1 . . . . . . . . . . . . Sb I . . . . . . . . . . . . Se I . . . . . . . . . . . . TeI . . . . . . . . . . . . Integration time . . . . . . . . Sample matrix . . . . . . . . . . Sample flow-rate . . . . . . . . Sodium tetrahydroborate(II1) flow-rate Potassium iodide flow-rate .. . . Greenfield type 4.75 kW 23 1 min-1 16.8 1 min-1 2.1 1 min-1 9.0 mm 197.197 nm 206.833 nm 196.026 nm 214.281 nm 20 s 1 + 1 hydrochloric acid 9.5 ml min-1 4.7 ml min-1 9.5 ml min-1 Table 2. Effect of selected ions on hydride production Concentration/ Suppression of analyte signal, YO Interferent pg ml-1 As* Sb* Se* Te* 500 2 0 40 0 Pb(I1) . . . . 400 10 30 40 90 100 0 6 10 65 Au(II1) . . . . 1 0 10 85 >95 Pt(1V) . . . . 1 20 20 30 >95 Pd(I1) . . . . 1 20 20 70 >95 Fe(II1) . . . . 5000 4 3 65 5 Ir(1V) . . . . 1 0 0 0 0 Rh(II1) . . . . 1 0 0 10 0 Ru(1V) . . . . 5 0 0 3 8 Au+PGMt . . 0.2each 5 0 48 89 * Analyte concentration 0.1 pg ml-1. t Platinum-group metals. Table 3. Efficiency of coprecipitation with Fe(OH), at pH 2.40 in the presence of 10% NH&l Concentration before precipitation/ Element pg ml-1 As .. . . . . . . . . . . 20 Sb . . . . . . . . . . . . 20 Se . . . . . . . . . . . . 20 Te . . . . . . . . . . . . 20 Pb . . . . . . . . . . . . 200 c u . . . . . . . . . . . . 200 Au, Pt, Pd, Ir, Rh . . . . . . 10 Ru . . . . . . . . . . . . 10 0.1 0.1 0.1 0.1 Recovery, % 100 100 100 99 100 99 100 94 c0.2 <0.2 <2 15 Reagents De-ionised water and analytical-grade reagents were used throughout the experimental work. Hydrochloric acid, relative density 1.18. Nitric acid, relative density 1.40. Hydrofluoric acid, 40% mlm. Perchloric acid, 70% mlm. Sodium peroxide, 95% mlm. Merck. Sodium hydroxide, 99% mlm. Merck. Ammonium chloride, 99.5% mlm. Saarchem UnivAR grade. Sodium tetrahydroborate (III), solution, 1 YO.A mass of 10 g of sodium tetrahydroborate(II1) was dissolved in 1 1 of 0.1 M sodium hydroxide solution. Potassium iodide solution, 100 g 1-1 in 20% VIV hydro- chloric acid. Arsenic standard solution, 1.00 g 1-1. Arsenic trioxide was dissolved in aqua regia and diluted with 3 M hydrochloric acid. Antimony standard solution, 1.00 g 1-1. Antimony (gran- ules) was dissolved in aqua regia and diluted with 3 M hydrochloric acid. Selenium standard solution, 1 .OO g 1-1. Selenium (granules) was dissolved in aqua regia and diluted with 3 M hydrochloric acid. Tellurium standard solution, 1.00 g 1-1. Tellurium (gran- ules) was dissolved in hydrochloric acid - nitric acid - water (1 + 1 + 1) without heating and diluted with 3 M hydrochloric acid.The optimum conditions for the formation of the hydrides of arsenic, antimony, selenium and tellurium and the com- promise instrumental parameters are given Table 1. Decomposition of Silicate Rocks A mass of 1.000 g of finely ground sample was treated with 10 ml of nitric acid and 10 ml of perchloric acid in a Teflon beaker. This solution was evaporated to 2-3 ml on a hot-plate, 10 ml of hydrofluoric acid and 5 ml of perchloric acid were added to the residue and the solution was evaporated nearly to dryness. After decomposition, the residue was dissolved in 30 ml of 1 + 1 hydrochloric acid and heated on a water-bath for 1 h. Decomposition of Sulphide Ores A mass of 1 .OOO g of finely ground sample was fused with 3.0 g of sodium peroxide in a zirconium crucible.The melt was cooled and then dissolved in 40 ml of 1 + 1 hydrochloric acid, after which the solution was heated on a water-bath for 1 h. Separation of Arsenic, Antimony, Selenium and Tellurium from the Matrix The acid solution was diluted to 125 ml with water and 15 g of ammonium chloride were added. The solution was heated to approximately 50 "C, neutralised with pellets of sodium hydroxide and finally adjusted to a pH of 2.40 with a 0.2 M solution of sodium hydroxide. The solution was heated on a water-bath for 2 h and the precipitate obtained was filtered through a Whatman No. 542 filter-paper. The precipitate of Table 4. Effect of ammonium chloride on the efficiency of the coprecipitation with Fe(OH), at pH 2.40 Recovery, % I 11 14-1, YO As* Sb* Se * Te * Aul- Ptt Pdt Irt Rht Rut 0 25 25 33 17 5 12 6 17 25 73 2.5 97 98 94 60 <2 3 <2 9 10 55 5 99 99 98 75 <2 (2 <2 <2 <2 40 10 100 99 98 94 <2 (2 <2 <2 <2 15 * Analyte concentration 0.1 pg ml-1. t Analyte concentration 10 pg ml-1.ANALYST, JANUARY 1986, VOL.111 17 iron(II1) hydroxide (only a partial precipitation of iron is effected), including arsenic, antimony, selenium and tel- lurium, was dissolved in 50 ml of hot, concentrated hydro- chloric acid, transferred into a 100-ml calibrated flask and diluted to volume with water. Determination of Arsenic, Antimony, Selenium and Tellurium The compromise conditions for reduction and measurement (Table 1) were used for the determination of the elements. A portion of the sample solution was spiked with a mixed standard solution of arsenic, antimony, sellenium aqd tel- lurium to give an additional concentration of 100 ng ml-1 of each of the elements.The concentration of the elements was calculated from the net signal for each element and compared with the net signal of the addition. Results and Discussion Silicate material is usually decomposed by fusion with sodium hydroxide, sodium peroxide or lithium metaborate. Decom- position by fusion can be used only as a preliminary step, as silicic acid must be removed. If the latter is not removed, the silica separates as a gel when the solution is acidified. This gel contains the hydride-forming elements. A more commonly used method for the decomposition of silicates is to heat the samples with nitric and perchloric acids followed by a further treatment with perchloric and hydrofluoric acids.14 This combination of acids has the advantage that no arsmic,16 antimony,l6 selenium,1618 or tellurium20 is lost. Sulphide ores containing low levels of silica are decomposed by fusion with sodium peroxide at 600 "C for 5 min to prevent the loss of the hydride-forming elements. The advantage of a fusion over the acid digestion previously discussed is that the sulphides are completely converted into Table 5. Effect of residual gold and platinum-group metals on hydride production after coprecipitation with Fe( OH), Initial concentrations of individual Suppression of analytical signal, % metals/ No. of pgml-* precipitations As* Sb* Se* Te* 2 1 4 5 46 50 10 1 6 5 55 90 10 2 7 6 40 35 * Analyte concentration 0.1 pg ml-1.sulphates, whereas acid digestion can result in the formation of a considerable amount of elemental sulphur. Each digestion procedure is followed by heating of the strong hydrochloric acid. This results in the reduction of selenium(V1) and tellurium(VI)+11 to selenium(1V) and tellurium(IV), respectively, the +4 oxidation state being preferred for quantitative hydride generation.1 However, arsenic(V) and antimony(V) remain unchanged and must be reduced to arsenic(II1) and antimony( 111) , respectively, with an auxiliary reducing agent such as potassium iodide before their hydrides can be produced. Unfortunately, pre-reduction with potassium iodide also reduces the oxidation state of selenium to zero, from which no hydride can form. It was found that the addition of the potassium iodide to the sample after the sodium tetrahydroborate(II1) had been added allowed the hydride of selenium to form before the oxidation state of the selenium could be reduced to zero, while allowing arsenic(V) and antimony(V) to be reduced successfully to arsenic(II1) and antimony(III), respectively.For the equi- valent production of the arsenic and antimony hydrides to be obtained in the absence of a pre-reduction step with potassium iodide, a potassium iodide solution of 100 g 1-1 in a 20% V/V hydrochloric acid medium must be used at a flow-rate of 9.5 ml min-1. The flow manifold is detailed in Fig. 1. The detection limits obtained with this system were calculated according to the equation used by Winge et aZ.21 and were 1.1 ng ml-1 for arsenic, 0.8 ng ml-1 for antimony, 0.7 ng ml-1 for selenium and 2.2 ng ml-1 for tellurium.The calibration graphs were linear up to 5 pg ml-1 for arsenic, antimony and selenium, and up to 3 pg ml-1 for tellurium. The coprecipitation of the hydride-forming elements with lanthanum hydroxide at pH 9.0 greatly reduced the interfer- ence effects caused by copper, nickel, cobalt, zinc, cadmium and related matrix elements. 13-21 Unfortunately, this method involves the coprecipitation of iron, lead, gold and the platinum-group metals with the lanthanum hydroxide. The interference due to iron is minimal (Table 2), and quantitative analysis is possible by the use of a technique in which the analyte solutions are spiked. However, the interference effects caused by lead, gold and the plantinum-group metals are extremely high (Table 2), especially in the production of the selenium and tellurium hydrides, and a more efficient separation of the interfering elements is necessary.The coprecipitation of arsenic, antimony, selenium and tellurium Table 8. Concentration of the major and minor components in anode slime (Mintek 1/77)22 Table 6. Concentrations of major elements in in-house reference materials Concentration, % Mintek No. c u Fe Pb S Zn 3 1/74 0.35 9.06 1.95 27.43 44.77 32/74 1.07 3.38 71.06 NR* 3.15 33/74 20.25 20.80 16.93 26.96 6.30 * NR = not reported. Concentration, Concentration, Component YO Component Y O Pt . . . . . . 0.04* Ag . . . . 22.5 Sb . . . . . . 7.60 As . . . .1.0* Se . . . . . . 5.03 Au . . . . 1* c u . . . . 6.15 SiOz . . . . 4.45 Ni 2.06 Sn 4.32 Te . . . . 1.36 Pb . . . . 17.82 Pd . . . . 0.2* . . . . . . . . . . . . * Tentative value. Table 7. Analytical results for arsenic, antimony, selenium and tellurium in selected standards As/pg gg1 Sblpg g-1 Se/pg g-1 Telpg g-1 Material This work Reported This work Reported This work Reported This work Reported Mintek31/7422 . , . . 3.6 k 0.3 NR* 1 4 f 3 1 9 f 5 1 k 0.3 NR 1 k 0.4 1.5 k 0.7 Mintek3217422 . . , , 2.5 f 0.5 4.0 k 0.8 NDt NR 3.2 k 0.3 4.4 k 1 1 f 0.4 NR Mintek3317422 . . . . 3.640.5 2.6$ k 0.8 ND NR 1.6 4 0.4 1.5$ k 0.5 4 4 0.5 NR NBS1633 . . . . . . 5 5 f 3 61 6 f 0.5 7$ 10 & 0.5 9.4 0.5 f 0.2 NR * NR = Not reported. t ND = not determined. $ Tentative value.18 ANALYST, JANUARY 1986, VOL.111 Table 9. Analytical results for arsenic, antimony, selenium and tellurium in mixtures of standards Aslpg g-1 Sblpg g-’ Selpg g-’ Te1p.g g-1 Calculated* Calculated Calculated Calculated Material This work value This work value This work value This work value SARM 4 - 1/177 (2500+ 1) . . . . 4.5 kO.3 4.0 21 + 2 30 f 6 2 0 f 2 20+1 6.2 + 0.5 5.4 k 0.5 (5OOO+ 1) . . . . . . 2.4k0.3 2.0 1 2 k 1 1 5 k 3 1 0 k 1 10 k 0.5 3.2 +. 0.4 2.7 k 0.3 SARM 4 - 1/77 * Calculated from tentative value. with iron(II1) hydroxide at pH 2.40 in the presence of 100 g 1-1 ammonium chloride solution has the advantage that almost 100% of the analyte elements are recovered with minimum coprecipitation of the interfering elements (Table 3). Therefore, this technique is recommended.The high concentration of ammonium chloride is necessary to ensure good separation of the gold and the platinum-group metals and complete coprecipitation of the hydride-forming elements (Table 4). In the analysis of samples with a total platinum-group metals content of less than 200 pg g-1, only one coprecipita- tion step is necessary. However, for high-grade ores and concentrates, it may be necessary for the original precipitate to be dissolved and coprecipitated a second time. This reduces the suppression caused by the interfering elements sufficiently for an analysis to be carried out (Table 5). It is also theoretically possible for losses of the analyte elements to occur during the coprecipitation stage, which would also reduce the analytical signal.This hypothesis was tested by the analysis of a synthetic solution containing a known concentration of the hydride-forming elements and iron, gold and platinum-group elements. After separation of the interfering elements, an aliquot of the solution was spiked with the hydride-forming elements to give a final concentra- tion equivalent to twice their concentration prior to separa- tion. A comparison between the net analytical signal of the unspiked sample and the relative increase due to the spike addition indicated that the recovery of the analytes was excellent. These experiments also confirmed the need for additions of the analytes to be used for quantification instead of calibration with synthetic solutions. Analyte addition compensates for the suppression caused by residual gold and platinum-group metals in the final solution (Table 5).Results were also obtained for arsenic, antimony, selenium and tellurium on the “in-house” reference materials of the Council for Mineral Technology (Mintek),22 viz., Mintek 31/74, 32/74 and 33/74, and in the standard reference material NBS 1633 (coal fly ash) (Tables 6 and 7). Owing to the lack of suitable international reference materials containing gold and the platinum-group metals, an in-house reference material was prepared for arsenic, antimony, selenium and tellurium by mixing of the South African reference material SARM 4 (norite) with Mintek 1/77 (anode sludge)22 in the proportions 2500 + 1 and 5000 + 1 (Table 8). The analytical results indicate good agreement with the calculated values (Table 9).Conclusions An analytical method has been developed for the simul- taneous determination of arsenic, antimony, selenium and tellurium in silicate rocks containing gold and the platinum- group metals and in sulphide ores containing lead and transition elements. The method utilises a separation tech- nique in which the hydride-forming elements are coprecipi- tated with iron(II1) hydroxide in the presence of ammonium chloride at pH 2.40. Interference effects are reduced to a minimum. This paper is published by permission of the Council for Mineral Technology (Mintek). The authors extend their thanks to Mr. R. V. D. Robert for many useful discussions on hydride-generation chemistry. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.13. 14. 15. 16. 17. 18. 19. 20. 21. 22. References Thompson, M., Pahlavanpour, B., Walton, S. J., and Kirk- bright, G. F., Analyst, 1978, 103, 568. Thompson, M., Pahlavanpour, B., Walton, S . J., and Kirk- bright, G. F . , Analyst, 1978, 103, 705. Pahlavanpour, B., Pullen, J . H., and Thompson, M., Analyst, 1980, 105, 274. Wolnik, K. A., Fricke, F. L., Hahn, M. H., andcaruso, J. A., Anal. Chem., 1981, 53, 1030. Nygaard, D. D., and Lowry, J. H., Anal. Chem., 1982,54,803. Nakahara, T., Anal. Chim. Acta, 1981, 131, 73. Azad, J., Kirkbright, G. F., and Snook R. D., Analyst, 1980, 105, 79. Thompson, M., Pahlavanpour, B., and Thorne, L. T., Water Res., 1981, 15, 407. Nadkarin, R. A., Anal. Chim. Acta, 1982, 135,363. Verlinden, M., Talanta, 1982, 29, 875. Nazarenko, V. A., “Analytical Chemistry of the Elements: Selenium and Tellurium,” Israel Program for Scientific Trans- lations, Wiley, New York, 1972, p. 190. Smith, A. E., Analyst, 1975, 100,300. BCdard, M., and Kerbyson, J. D., Anal. Chem., 1975,47,1441. Nakashima, S . , Analyst, 1978, 103, 1031. Nakashima, S . , Anal. Chem., 1979, 51, 654. Bajo, S . , Anal. Chem., 1978, 50, 649. Chan, F. L., and Riley, J. P., Anal. Chim. Acta, 1965,33, 36. Bock, R., and Jacob, B., Fresenius 2. Anal. Chem., 1964,200, 81. De Oliveira, E., McLaren, J. W., and Berman, S. S . , Anal. Chem., 1983,55, 2047. Nazarenko, V. A., “Analytical Chemistry of the Elements: Selenium and Tellurium ,” Israel Program for Scientific Trans- lations, Wiley, New York, 1972, p. 190. pp. 204-205. Winge, R. K., Peterson, V. J., and Fassel, V. A., Appl. Spectrosc., 1979, 33, 206. Stoch, H., and Ring, E. J., Report M104, Council for Mineral Technology, Randburg, 1983. Paper A51 70 Received February 20th, 1985 Accepted August 8th, 1985