ANALYST SEPTEMBER 1986 VOL. 111 1023 Application of Electrothermal Atomic Absorption Spectrometry to the Determination of Trace Amounts of Indium in Metallic Zinc and Krystyna Brajter and Ewa Olbrych-Sleszynska Department of Chemistry University of Warsaw Pasteura 1 02-093 Warsaw Poland Trace amounts of indium in metallic zinc and lead have been determined by graphite furnace atomic absorption spectrometry. The detection limits for three operated furnace systems were evaluated for SP9-01 (1 ) 6.0 x 10-13 g; for HGA-500 (2) 7.0 x 10-13 g; and for GRM-1268 (3) 1.1 x 10-11 g each using 20 pl of sample. Linear calibration graphs were obtained between 0.01 and 0.1 p.p.m. (I) 0.05 and 0.6 p.p.m. (2) and 0.05 and 0.2 p.p.m. (3). Ion-exchange separation employing Xylenol Orange modified Amberlyst A-26 anion exchanger was used as the preliminary step for the determination of indium in Zn - In and Pb - In alloys as interferences from other metals were observed in the indium absorbance.In order to compare the results two other separation methods for the lead matrix and several different graphite furnace atomisers were used. Molar excesses of Ni Co Fe Cu and Zn of less than 1000 and between a 100 and 1000 molar excess of Al cause a decrease in the indium absorbance. With molar ratios greater than 1000 the suppression caused by Al Co and Zn disappears and that of Fe becomes less pronounced. Ga causes an enhancement of the indium atomic absorption signal. Equivalent amounts of As Sb Bi do not interfere. Keywords Indium determination; electrothermal atomisation; atomic absorption spectrometry; ion-exchange separation; Xylenol Orange modified Amberlyst A-26 Indium is frequently found in trace amounts in metallic zinc and lead.Its determination in metals and ores (in nature indium occurs in some zinc and lead ores) is important from an analytical point of view. The determination of indium, especially in the presence of the other metal matrices presents many problems for the analytical chemist. The aim of the work reported in this paper was to determine trace amounts of indium in metallic zinc and lead using atomic absorption spectrometry (AAS) with electrothermal atomisation (ETA). AAS was chosen for indium determination as the most convenient means of comparing the results with other methods of separation.In preliminary experiments we found that there were interferences from matrix metals. Preliminary separation of indium from matrix metals was adopted as the most practical procedure to avoid matrix interferences. Three independent methods of separation were investigated co-precipitation of hydroxides precipitation of lead sulphate and ion-exchange separation. We found ion-exchange separation on Xylenol Orange modified anion exchanger to be the most convenient method. Experimental Instrumentation A Pye Unicam SP9-01 atomic absorption spectrometer a Perkin-Elmer Model 2380 atomic absorption spectrometer equipped with an HGA-500 graphite furnace atomiser and a Beckman Model 1272D spectrophotometer equipped with an Unilam 1288 burner and a Pye Unicam GRM-1268 graphite furnace atomiser were used.The three instrument systems were equipped with deuterium background correction. The deuterium lamp background compensator could not eliminate all the matrix interferences observed. The operating con-ditions are given in Table 1. The SP9-01 atomic absorption spectrometer was used for the determination of indium in all the metallic lead and zinc samples. For the examination of the indium content in the weighed samples and for comparison of results indium was also determined using the Beckman 1272D spectrophotometer and the GRM-1268 graphite furnace atomiser. Table 1. Optimum operating conditions for the determination of indium by ETA atomic absorption spectrometry Parameter Wavelengthhm . . . . . . Dryingtemperature/"C .. . . Drying time/s . . . . . . . . Charring temperaturePC . . . . Charringtimels . . . . . . Atomisation temperature/'C . . Atomisationtimek . . . . . . CleaningtemperaturePC . . . . Cleaningtimek . . . . . . Band-passhm . . . . . . . . Injectionvolumelpl . . . . . . Argon flowratell rnin- I . . . . HGA-500 Pye Unicam Pye Unicam (Perkin-Elmer GRM- 1268 SP9-01 2380) (Beckman 1272A) 303.9 110 15 500* 15 2400 15 3000 5 0.7 20 7 L. 303.9 100 20 1200 20 2500 10 2500 3 0.7 20 Standard (0.3) 303.9 100 20 1100-1350 30 2700-2900 10 3000 3 0.7 20 Standard ( 1 3) * Explanation in text 1024 Reagents A standard indium solution was prepared by dissolving indium (spectral grade 1.000 g) in 2 M HCl and diluting to 500 ml.This solution was further diluted as required. Metal ion solutions were generally prepared by dissolving the appropriate masses of the sulphates and nitrates of the metals in doubly distilled water and were standardised by EDTA titration. The stock solutions were diluted as required. ANALYST SEPTEMBER 1986 VOL. 111 Procedure B Ion-exchange Columns Columns 25 cm long and 5 mm i.d. with a stopcock at the end, were used. They were packed with the macroporous strong base anion-exchange resin Amberlyst A-26 with a bead size of 0.1-0.2 mm (Rohm and Haas) modified by the use of Xylenol Orange (XO) tetrasodium salt.1 In procedures A and B in order to obtain XO-loaded resin the Amberlyst A-26 chloride form with a bead size of 0.1-0.2 mm was shaken with a 4 x 10-5 M aqueous solution of XO until the supernatant became colourless.The resin was then filtered off washed with water and ethanol dried and stored in a refrigerator.' The modified resin had a capacity of 0.4 mmol of XO per gram of Amberlyst A-26 in the primary chloride form. A 4 cm bed height of 200 mg of modified resin was used in both procedures. Flow-rates of 40 ml h-1 were used. Ion-exchange Separation of Indium from Zinc and Lead Preliminary experiments confirmed the usefulness of XO-loaded resin for the separation of indium from zinc and lead. To optimise the conditions for the separation of trace amounts of indium from the great excess of lead and zinc some investigations were performed with the use of synthetic solutions simulating the composition of the real metallic zinc and lead samples being analysed.The effect of bed height and the size of the resin beads was investigated. The following procedures for the rapid separation of indium from zinc (A) and from lead (B) were deduced. Procedure A A 13-ml volume of solution containing 0.005 mg of In and 669 mg of Zn adjusted to pH 2.0 was introduced into the column. Zn was eluted with 25 ml of water acidified to pH 4.0 with H2S04. Indium was eluted with 20 ml of 0 . 2 ~ HN03 into a 25-ml Calibrated flask. After dilution to volume indium was determined according to the conditions given in Table 1. The indium concentration was obtained from a calibration graph. The mean obtained from six separate determinations was 0.0049 L- 1.3 X 10-4 mg (95% confidence limit).Each result was the mean of four AAS measurements by the GRM-1268. The mean obtained using the Pye Unicam SP9-01 spectropho-tomer was 0.0048 k 1.2 x mg (95% confidence limit). A 13-ml volume of solution containing 0.005 mg of In 669 mg of Pb and 2 mg of XO adjusted to pH 3.0-3.8 was introduced into the column. Pb was not retained on the column and passed into the eluent. To wash the Pb from the void volume, 25 ml of water acidified to pH 4.0 with HN03 were used. Indium was eluted with 20 ml of 0 . 2 ~ HN03 into a 25-ml calibrated flask. The volume was adjusted to the mark and indium was determined by AAS. The indium concentration was obtained from the calibration graph. The mean obtained from six separate determinations was 0.0049 _+ 1.3 x 10-4 mg (95% confidence limit).Each result was the mean of four AAS measurements by the GRM-1268. The mean obtained using the Pye Unicam SP9-01 spectrometer was 0.0049 k 1.2 x 10-4 mg (95% confidence limit). Analysis of Metallic Zinc Procedure 1 and Metallic Lead, Procedures 2 3 and 4 The sample of metallic zinc was dissolved by heating it with 5 ml of HN03 (1 + 1). The solution was evaporated twice and the residue was dissolved in 0 . 2 ~ HN03 transferred into a 25-ml calibrated flask and diluted to volume. Five different samples of metallic lead were dissolved in the same way and transferred into five 25-ml calibrated flasks. Procedure 1 A 5-ml aliquot was diluted to 10 ml adjusted to pH 2.0 and introduced into the ion-exchange column.Zn was not retained on the resin. Zinc remaining in the void volume was washed out with 25 ml of water acidified to pH 4.0 with H2S04. Indium was then eluted with 20 ml of 0.2 M HN03 into a 25-ml calibrated flask and was diluted to volume. The indium was determined by AAS according to the conditions given in Table 1. The concentration was obtained from the calibration graph, and the results of the analysis are presented in Table 2. Procedure 2 A 5-ml aliquot was diluted to 10 ml and adjusted to pH 3.0-3.8; 2 mg of XO were then added and the aliquot was introduced into the column. Pb was not retained on the resin bed and passed into the eluent. Lead was washed out with 25 ml of water acidified to pH 4.0 with HN03. Indium was eluted with 20 ml of 0 .2 ~ HN03 into a 25-ml calibrated flask. The solution was diluted to volume and indium determined by AAS; the concentration was obtained from the calibration graph. The results are presented in Table 3. Procedure 3 Five samples of metallic lead containing different amounts of indium were analysed. They were dissolved by heating with 5 ml of nitric acid (1 + 1) and then 20 mg of metallic iron, dissolved in 1 + 1 nitric acid were added to all the samples and metal hydroxides were coprecipitated with ammonia solution. Table 2. Determination o f indium in metallic zinc by ETA - AAS De t t' rm i na t ion after ion-exchange separation+ (Procedure 1 ) D i rec t de t t' r ni i n ;i t ion Instrument Sam ple/mg 111. '%I Sa ni plt'img In '%, GRM-1268 .. 1020 0.0009 I125 0.0005 * 0.0002 SP9-01 . . . I005 0.0009 1015 0.0000 5 0.0003 * Obtained from 10 AAS measurements without zinc separation after dissolving the sample as described before Procedure 1. t Mean and range (95% confidence limit) of four separate determinations ANALYST SEPTEMBER 1986 VOL. 111 Q 0.030 0.020 0.010 1025 (a) - /x-x-x ,x-x -4 X -I I I Table 3. Determination of indium by ETA - AAS in five different samples of metallic lead after the separation step obtained for the GRM-1268 ( a ) and SP9-01 (b). All samples analysed contained a mixture of metal ions (Cu. Ag. Bi. As Sh. Sn. TI) in the range 0.045-0.0006% for each metal ion Separation by coprecipitation of hydroxides (Procedure 3) Saniple/mg 544.7 512.0 1126 1 108 933.8 921.4 1010 1013 1103 1105 In," %" 0.028 0.025 0.0 12 0.01 1 0.002 0.003 0.0007 0.0006 0.008 0.008 Separation by precipitation of lead sulphate (Procedure 4) Sample/mg In.* O/" 512.1 0.020 543.2 0.020 1140 0.013 1114 0.01 1 1002 0.00 1 101 1 0.002 1115 0.0006 1112 0.0005 1232 (1.007 1114 0.007 Ion-exchange separation (Procedure 2) Sample/mg In,? YO 532.0 0.025 k 0.001 508.1 0.024 k 0.002 1012 0.015 k 0.004 1007 0.012 2 0.003 1 104 0.0018 k 0.0003 0.002 ? 0.0004 1103 120s 0.0007 ? 0.0003 120 1 0.0006 ? 0.0003 1125 0.0070 ? 0.00 1 8 1112 0.0070 k 0.0021 * Obtained from four AAS measurements for one sample of metallic lead.after one separation step. i- Mean and range (95% confidence limits) for four AAS measurements in each of four separate determinations.This was repeated twice. The hydroxide precipitate was dissolved in 5 ml of HN03 (1 + l ) transferred into a 25-ml calibrated flask and diluted to volume. Indium was deter-mined by AAS according to the conditions given in Table 1. The results were corrected for blanks determined under the same conditions as the lead samples. Procedure 4 For the comparison of results the lead matrix in all alloy samples was separated as lead sulphate after the dissolution step described above and indium was determined in solution after filtering off the precipitate. For this 5 ml of concentrated H2S04 were added to the aliquot and the sample was heated until white fumes were observed.The sample was then cooled and after adding 50 ml of water heated to boiling. In this instance the indium standards were mixed with lead nitrate and then analysed in exactly the same way. The results obtained are presented in Table 3. Results and Discussion Results were obtained by three different ETA - AAS systems. The Beckman 1272D system belongs to an earlier generation of spectrophotometers but is still used by many laboratories. The results obtained using it are comparable with those from a newer type of spectrophotometer the Pye Unicam SP9-01. This implies that even those laboratories that possess older types of spectrophotometers may use our method with good results. To optimise the charring temperature the dependence of absorbance on charring temperature was investigated (Fig.1). A charring temperature of 1200°C applied for 20 s was the maximum temperature at which no loss of analyte occurred when the HGA-500 was used. For the SP9-01 spectrometer a charring temperature of 500 "C applied for 15 s was advised by Pye Unicam for indium determination. There was no differ-ence in the absorption peak at charring temperatures in the range 500-1000°C in the determination of indium. For the GRM-1268 an optimum charring temperature of 1100 "C applied for 20 s was chosen. Atomisation temperature versus absorbance is presented in Fig. 2. The maximum temperatures available using the HGA-500 2400°C for the SP9-01 and 2700-2900°C for the GRM-1268 were chosen. These maximum temperatures were also recommended by Dittrich.2~3 As expected the best analytical results were obtained using the HGA-500 and SP9-01 systems.For the GRM-1268 L'vov's pyrolytic graph-ite platform was used but no improvement was obtained. The 3 c I 500 1000 1500 Tc ha rl0c Fig. 1. Height of indium absorption peak as a function of charring temperature. GRM-1268 Tat,, = 2800 "C cln = 0.1 p.p.m tchar. = 30 s 1500 2000 2500 3000 T,t"lll/"C Fig. 2. (a) Absorbance of indium; and (b) peak height as a function of atomisation temperature. ( a ) HGA-500 Tchar. = 1200 "c cIn = 0.4 p.p.m. ( b ) GRM-1268 Tchar. = llOO°C cIn = 0.1 p.p.m. detection limit with the GRM-1268 is similar to2,4 or better than3 those in the literature. The presence of NaOH HCl or HN03 has a strong influence on the determination of indium, as has also been observed by others.2,4,5 According to Dittrich,zJ the suppression of the indium absorbance in the presence of HC1 is due to InCl formation.We think that a more probable explanation for the decrease in the indium signal is the diffusion of volatile InC13 from the furnace in the period of time before the atomisation temperature has been reached. The presence of HCl and NaOH causes a propor-tional suppression of absorbance whereas the addition of HN03 (>O. 18 M) stabilises the signal. Hence 0.2 M HN03 was chosen as the best medium for the determination of indium. Linear calibration graphs were observed between 0 and 0.6 p.p.m. for the HGA-500 between 0 and 0.2 p.p.m. for the GRM-1268 and between 0 and 0.1 p.p.m. for the SP9-01.In all instances a 20+1 aliquot was used 1026 ANALYST SEPTEMBER 1986 VOL. 111 Table 4. Detection limit and sensitivity (for 0.0044 absorbance) for the determination of indium with different spectrophotometers. In all instances indium was determined under the optimum conditions given in Table 1; 20-yl volumes were used Perkin-Elmer Beckman 1272D Pye Unicam 2380 (Pye Unicam Parameter SP9-0 1 (HGA-500) GRM- 1268) Detection limithg ml-l . . 0.030 0.035 0.55 Detection limit/pg . . . . 0.60 0.70 11 Sensitivityhgml-1 . . . . 0.050 0.050 2.2 Sensitivitylpg . . . . . . 1 .0 1 .0 44 Reproduci bilityhg ml- 1 . . k 0.0052 k 0.0052 +0. 10 Fig. 3. Effect of Al Fe Co Ni Cu Zn Ga and Pb on the AAS signal of In expressed as relative change in signal for the HGA-500 and GRM-1268 systems.A Ga; B Co; C Fe; D Al; E Zn; F Ni; G Pb; and H Cu Zinc lead and other matrix metals have a marked effect on the indium absorbance; metals were taken as nitrates through-out to avoid the influence of chlorides. Molar excesses of Ni, Co Fe Cu and Zn of less than 1000 and of A1 between 100 and 1000 cause a decrease in the indium absorbance. With molar ratios greater than 1000 the suppression caused by Al Co and Zn disappears and that of Fe becomes less pronounced. Gallium causes an enhacement of the indium atomic absorp-tion signal. These results were obtained by the HGA-500 system and by the GRM-1268 system for comparison and are presented in Fig. 3. Equivalent amounts of As Sb and Bi do not interfere with the indium atomic absorption signal.The complex nature of the chemical process taking place in the graphite furnace makes the correct explanation of the interference effects observed extremely difficult. Campbell and Ottaway6 and Aggett and Sprott7 have postulated that atoms are produced by direct reduction of the metal oxides by graphite. If this is true the suppression of the indium absorbance in the presence of an excess of the other metals, especially of those forming more stable carbides or refractory oxides may be explained simply by the occlusion of the indium or the graphite surface. Under such conditions the absorption peak of indium was observed to be delayed and was of a different shape. We observed the suppression of the indium absorption signal in the presence of Fe Co and Ni, which form carbides with high melting-points and in the presence of A1 (refractory oxide) Zn (high sublimation point) Pb and Cu.A strong influence of the presence of zinc and lead on the absorption signal of indium was observed; this was the reason for developing Procedures 1 2 3 and 4. Some modification of the charring step by an extension of time could not remove the interference of lead and zinc on the indium signal. Magnesium nitrate was investigated as a matrix modifier but it did not remove the interferences. Mg(N03)2 causes an enhacement of the indium atomic absorption signal related to its concentration in the sample. If the excess of interfering metal is sufficiently large a new effect is observed an increase of the indium signal. This also suggests that the mechanism of atomisation may have been changed.This effect is very similar to that observed if a modified graphite furnace (modified by means of some metal ions forming stable carbides or other compounds) is used for the atomisation process.sl0 No presence of indium in interfering reagents even at high interfering metal concen-trations was reported. Owing to the complicated effects on the indium signal resulting from the presence of other metal ions we decided to eliminate matrix interferences by a preliminary separation of indium from the sample of metallic zinc and lead by ion exchange. The methods were applied to the determination of indium in metallic zinc and lead. The results are presented in Tables 2 and 3. In order to compare the determination in metallic lead, we also used other separation methods namely coprecipi-tation and precipitation of the matrix by the use of H2S04.In the coprecipitation method indium as In( OH)3 coprecipi-tates on Fe(OH)3. This method was also adopted for the separation of indium. The results are presented in Table 3. The XO-loaded resin permits the simple fast separation of indium from an excess of lead and zinc using a very short resin bed. No influence from other metal ions present in up to a 10-fold molar excess was observed ANALYST SEPTEMBER 1986 VOL. 111 References 1. 2. 3. 4. 5. Brajter K. and Olbrych-Sleszynska E. Talanta 1983 30, 355. Dittrich K. Talanta 1977 24 735. Dittrich K. Talanta 1977 24 725. Yudelevich I. G. Burynova L. M. Bakhturova N. F. and Korda T. M. Zh. Anal. Khim. 1977 32 28. Martinsen I . and Langmyhr F. J. Anal. Chim. Acta 1982, 135 137. 1027 6. 7. 8. 9. 10. Campbell W. C . and Ottaway J. M. Talanta 1974 21 837. Aggett J . and Sprott A. J. Anal. Chim. Acta 1974 72 49. Brajter K. and Kleyny K. Talanta 1985 7 521. Thomson K. C . Godden R. G . and Thomerson D. K., Anal. Chim. Acta 1975 74 289. Lagas P. Anal. Chim. Acta 1978,98 261. Paper A51349 Received September 30th) 1985 Accepted April 14th) 198