Analyst, January 1996, Vol. 121 (7-11) 7 Determination of Trace Amounts of Cadmium in a Hydrometallurgical Zinc Refining Process Stream by a Flow-injection Method With On-line Preconcentration and Spectrophotometric Detection Yutaka Hayashibe and Yasumasa Sayama Materials Characterization Center, Central Research Institute, Mitsubishi Materials Co., 1-297, Kitabukuro-cho, Omiya, Saitama 330, Japan A flow-injection (FI) method was developed for the spectrophotometric determination of cadmium in a hydrometallurgical zinc refining process stream using l-(4-nitrophenyl)-3-(4-phenylazophenyl)triazene (Cadion) as the chromogenic reagent. The sample solution was injected into a carrier containing potassium iodide. The sample was then passed through an anion-exchange mini-column on which the analyte was concentrated as a cadmium-iodo complex.In order to extend the detectable range of cadmium, a multiple sample injection method, in which the sample solution was repeatedly injected into the carrier at regular intervals of 30 s, was applied. Cadmium on the column was eluted with 1 moll-1 nitric acid and merged with a stream of a mixture of masking agent (trisodium citrate-potassium sodium tartrate-potassium hydroxide) and Cadion. Finally, the absorbance of the cadmium-Cadion complex was measured at 480 nm. The proposed FI was fully controlled by a personal computer. The proposed system permitted throughputs of 6 samples h-l for single injection, and 2 samples h-l for single injection, and 2 samples h-1 for multiple sample injection (50 injections). The reproducibility was satisfactory with a relative standard deviation of less than 5.0% (0.14 pg ml-1 Cd level, n = 5) for the single injection method and 10% (2.0 ng ml-1 Cd level, n = 5) for the multiple sample injection method (50 injections).The detection limits were 0.028 pg ml-1 of cadmium for the single injection method and 0.83 ng ml-1 of cadmium for the multiple sample injection method (50 injections). The absolute amount of cadmium detectable, defined as the analytical signal equal to twice the uncertainty in the background, was 10 ng. Keywords: Flow injection; cadmium determination; on-line preconcerztration; anion exchange; spectrophotometric detection Introduction For the preparation of the zinc electrolyte in hydrometallurgical zinc refining, it is necessary to eliminate impurities which originate from zinc concentrate (sphalerite, calamine and smithsonite).In the preparation of the electrolyte, zinc powder is commonly utilized to eliminate impurities by cementation. The amount of zinc powder added to the process stream is controlled on the basis of the cadmium content, which must be monitored continuously. Flow injection (FI) is a rapid and precise technique, and a number of automated FI systems are commercially available. Although several methods for the determination of cadmium with FI have been published,2-5 their applications to on-line process analysis have scarcely been reported. Sensitive spectrophotometric methods for the determination of cadmium have been reported and several sensitive chromo- genic reagents for the spectrophotometric detection of cadmium are commercially available. Among these, 1 -(4-nitrophenyl)- 3-(4-phenylazophenyl)triazene (Cadion) forms a water-soluble complex with cadmium in basic media containing Triton X-100.6 In order to improve the selectivity and sensitivity of the Cadion method, however, the separation of cadmium from matrix zinc and preconcentration of cadmium are necessary before spectrophotometric detection.C~precipitation,~ solvent extraction,*-lO ion exchange, etc., have usually been used for separation of cadmium from the zinc matrix. Because of the applicability to aqueous media, we selected an ion-exchange technique. Various ion-exchange procedures have been de- scribed for the separation of cadmium from other elements.Although cadmium can be separated from zinc by cation- exchange or anion-exchange methods,' '-I4 their selectivity for cadmium is poor and a large amount of resin is required for zinc to be adsorbed. The anion-exchange adsorption behaviour of many elements in hydriodic acid media has been reported by Marsh et al.15 We have also investigated the anion-exchange behaviour of 12 elements in potassium iodide media and found that cadmium is selectively adsorbed on the strongly basic anion-exchange resin Bio-Rad AG1 -X8 from dilute solutions of potassium iodide, and completely separated from matrix zinc.16 In this work, an FI system for the determination of trace amounts of cadmium in a hydrometallurgical zinc refining process stream was developed. In order to determine cadmium accurately and rapidly, the sample solutions were taken directly from a continuous-flow process stream.The sample solution was injected into a carrier (potassium iodide solution) and cadmium was concentrated on an anion-exchange mini-column as the iodo complex. After elution with nitric acid, cadmium was detected spectrophotometrically with Cadion as the chromogenic reagent. In general, the sensitivity of an FI system is defined and controlled easily by varying the volume of the injection loop. However, replacement of the injection loop for the adjustment of sensitivity to detect various amounts of analyte (from nanograms to micrograms) is tedious and time- consuming. Hence, a multiple sample injection method was studied in order to control the sensitivity of the proposed FI system.Experimental Reagents All reagents used were of analytical-reagent grade and all solutions were prepared with distilled water.8 Analyst, January 1996, Vol. 121 w2 A cadmium stock standard solution was prepared by dissolving 1.00 g of cadmium (99.999% purity) in 30 ml of nitric acid, expelling the nitrogen oxides and diluting to 1000 ml with 1 moll-1 nitric acid to yield a 1 mg 1-1 Cd-1.4 mol 1-1 nitric acid solution. Working standard solutions were prepared by appropriate dilution of the stock standard solution. A stock solution of Cadion was prepared by dissolving 200 mg of the commercial reagent (Aldrich, Milwaukee, WI, USA) in 1000 ml of 0.1 moll-1 potassium hydroxide-O.l% v/v Triton X-100 solution. A working solution of Cadion was prepared by diluting the stock solution appropriately with 0.1 mol 1-1 potassium hydroxide-O.l% v/v Triton X- 100.For the prepara- tion of 0.1 mol 1-l potassium iodide solution as the carrier solution, 16.8 g of potassium iodide were dissolved in the appropriate amount of water and diluted to 1000 ml with water. About 5.0 g of trisodium citrate dihydrate, 2.5 g of potassium sodium tartrate tetrahydrate and 112 g of potassium hydroxide were dissolved in 1000 ml of water to yield 1.7 X 10-2 moll-' trisodium citrate-8.8 X 10-3 mol 1-l potassium sodium tartrate-2.0 mol 1-1 potassium hydroxide mixture as the masking solution. The anion-exchange resin Bio-Rad AGl-X8 (100-200 mesh, C1- form, Bio-Rad Labs.) was used as received for preparing the anion-exchange mini-column.The capacity of the resin used was 1.2 mequiv. ml-1. The mini-column was prepared in the following manner: the resin was slurry-packed in a PTFE tube (100 X 1.0 mm id) and each end was plugged with the appropriate amount of cotton wool. - p2 6 - FI Manifold and General Procedure Fig. 1 shows a schematic diagram of the manifolds used. A Hitachi U- 1000 ratio-beam spectrophotometer equipped with a 60 pl flow cell (20 mm pathlength) was used as the detector. Sanuki-kogyo DMX-2400T double-plunger pumps were used for delivering the carrier, the eluent, the chromogenic reagent solution and the masking solution. A peristaltic pump, ATTO AC-2120, was used to pump sample solutions. Sanuki-kogyo SVA-6M2H automated six-way valves, made of ceramic material, were used for sample introduction and line switching for the ion-exchange column.PTFE tubing (1.0 mm id) was used to construct the analytical manifold. The dimensions are specified in the discussion. All of the pumps and valves and the spectrophotometer were controlled by an NEC-PC980 1 perso- nal computer, for which a control program (MS-DOSm88- BASIC ver.6.1) written in this laboratory was used. - C D P3 - Carrier solution was pumped into the analytical line at a flow rate of 1 .O ml min-l by the pump P1. The sample, taken directly from the zinc refining process with an on-line sampler, was injected into the carrier stream with the six-way valve V1 (350 pl), and passed through the mini-column connected to the six- way valve V2 to adsorb the cadmium-iodo complex.For the multiple sample injection method, 350 pl portions of sample solution were repeatedly injected into the carrier at regular intervals of 30 s. At 220 s after the sample injection, the eluent (1.0 moll-1 nitric acid) was introduced into the mini-column at a flow rate of 1.0 ml min-1 by switching V2. The effluent was merged with the chromogenic reagent solution and the masking solution at a flow rate of 0.75 ml min-1, and the absorbance of the cadmium-Cadion complex was monitored in the flow- through cell at 480 nm. Results and Discussion Anion-exchange Adsorption Behaviour of Cadmium and Zinc in Acidic Potassium Iodide Media Table 1 shows the compositions of typical zinc electrolytes used for hydrometallurgical zinc refining. Because of the high matrix concentration and interferences caused by the co-existing elements, it is difficult to determine cadmium directly with a spectrophotometric method.The adsorbabilities of many ele- ments on an anion-exchange resin in hydriodic acid media have already been published. 1 1915 However, details of the distribution coefficients in lower concentrations of hydriodic acid have not been reported. We have measured the anion-exchange distibu- tion coefficients of 12 elements in acidic potassium iodide media.I6 It is well known that cadmium forms stable anionic iodo complexes ([CdI3]- and [CdI,]*-). We believe that the form of the cadmium-iodide complex is [CdI#-- in 0.1 moll-' potassium iodide solution from an estimate of the stability constants of the two species.17 The results obtained suggest that cadmium is almost completely sorbed on the anion-exchange resin column at potassium iodide concentrations above 0.05 moll-1, whereas zinc is weakly adsorbed at such concentration levels, and can be washed out from the column with 0.05 moll-' potassium iodide.Although iodide is unstable in acidic media, potassium iodide in neutral solution can be stored for at least 7 d. Furthermore, because of the short mixing time of the iodide solution with the acidic solution, and hence little liberation of iodine, the anion-exchange method in iodide media can be employed for the on-line preconcentration of cadmium. The cadmium-iodo complex can be rapidly and completely eluted from the column with dilute nitric acid. Spectrophotometric Detection of Cadmium The conventional methods used for the determination of trace amounts of cadmium are AAS and extraction photometry with dithizone.The dithizone method, which requires extraction and back-extraction, is tedious and time-consuming. Other conven- ~ ~~~~ Table 1 Composition of typical zinc electrolyte used for hydrometallurgical zinc refining Species Ca2+ Co*+ Fez+ Mn2+ Ni*+ SW+ s o p * mg ml-1. Content/ pg ml-' 200 0.1 15 5* <0.1 <0.1 = 180* Species Cd*+ c u2+ K+ Mg2+ Pb2+ Bi'+ Zn'+ Content/ pg ml-* 0.2 0.1 5* 8* <0.1 <0.1 > 150*Analyst, January 1996, Vol. 121 9 ient methods, which are more selective and sensitive than the dithizone method, have scarcely been reported. Hence, the Cadion6 and 5,10,15,20-tetraphenyl-2 lH,23H-porphyrinetetra- sulfonic acid, disulfuric acid, tetrahydrate (TPPS)18 methods were considered for the spectrophotometric detection of cadmium.These reagents react with cadmium and form water- soluble complexes with high molar absorptivity (E > 105 m2 mol-I). Although Cadion is less sensitive than TPPS, 0.1 pg ml-1 levels of cadmium can be detected with the former in alkaline media in combination with Triton X- 100 as solubilizing agent. Taking into account the high reaction rate and low cost, Cadion was chosen as the chromogenic reagent for the detection of cadmium. Optimization of the FI System The single-line manifold shown in Fig. 1 was constructed to introduce the sample into the FI system. It would seem to be more appropriate to use a multiline arrangement and merge the iodide stream with the sample.However, the total time for sample introduction using a multiline manifold would probably be longer than that using a single-line manifold. Hence, the single-line manifold was used. The dispersion of the sample injected is sufficient to ensure mixing with the carrier and complex formation with iodide in the proposed system. The influence of the concentration of the potassium iodide solution (carrier) was examined in the range 0.01-0.5 rnol 1-1 and it was found that a potassium iodide concentration above 0.05 mol 1-l was necessary for cadmium to be adsorbed strongly on the anion-exchange mini-column. However, a concentration of potassium iodide of more than 0.2 mol 1-l cannot be employed because matrix zinc is also strongly adsorbed on the mini-column and cadmium cannot be com- pletely separated from matrix zinc.A potassium iodide concentration of 0.1 mol 1-1 was selected. The effect of the concentration of nitric acid as the eluent was investigated in the range 0.1-2.0 mol 1-1. The peak height increased steadily with increasing concentration of nitric acid, and hence the cadmium-iodo complex adsorbed on the column was rapidly eluted with increasing concentration of nitric acid. A concentration below 0.5 moll-1 is not recommended because the system peak caused by the zinc weakly adsorbed on the column appears in close proximity to the cadmium peak. Hence, 1 .O moll-1 nitric acid was employed to elute the cadmium-iodo complex from the mini-column. Cadmium can be separated from most of the co-existing elements in the zinc refining process stream by the anion- exchange mini-column.However, small amounts of bismuth, copper and lead accompany cadmium. These elements react with Cadion and interfere with the spectrophotometric detection of cadmium. Hence, 1.7 X 10-2 mol 1-1 trisodium citrate-8.8 X 10-3 moll-1 potassium sodium tartrate was utilized to mask these interferents.5 Cadion reacts with cadmium to form a stable complex in alkaline media; therefore, the effluent from the column must be made alkaline. Hence, a mixture of 1.7 X 10-2 moll-' trisodium citrate-8.8 X mol I-' potassium sodium tartrate-2 mol 1-1 potassium hydroxide was used as the masking stream. The effect of the Cadion concentration in the range 0.00005-0.02% m/v was tested, using 350 pl of sample.A maximum and constant response for 1.0 pg ml-1 of cadmium was obtained at a reagent concentration above 0.002% m/v. In order to dissolve Cadion in water, potassium hydroxide and Triton X- 100 were necessary as solubilizing agents.5 Hence, a mixture of 0.002% m/v Cadion solution-0.1% v/v Triton X-100-0.1 rnol 1-1 potassium hydroxide was used as the chromogenic reagent. The effect of the length of the reaction coils (Fig. 1) was examined at various flow rates of the carrier and reagent solutions. Coils of 0,0.5 and 1 .O m were tested for C1. The coil length was found to have no effect on the response of cadmium. This result suggested that cadmium reacts rapidly with iodide ion to form a stable [CdLJ2+ complex. The length of C2 was then varied from 1 to 5 m.A 3 m coil gave the most sensitive and precise results. Hence, the optimum lengths were 0.5 m for C1 and 3 m for C2. The effect of the flow rate of the carrier (Pl) on the peak height was studied in the range 0.5-2.0 ml min-l, by injecting 350 p1 of cadmium standard solution and a real sample solution which was spiked with cadmium standard solution to contain 0.4 pg ml-1 of cadmium. A constant response was obtained in the range tested. This result demonstrated that the anion- exchange reaction between the cadmium-iodo complex and iodide is rapid and that the complex is strongly adsorbed on the anion-exchange resin. A flow rate of PI of 1.0 ml min-l was selected. The influence of the flow rate of P2 for delivering the eluent was examined in the range 0.6-2.0 ml min-l.The peak height decreased steadily (Fig. 2) and the peak width broadened with increasing flow rate. This suggests that [CdIJ*- is strongly adsorbed on the ion-exchange resin and that the ion- exchange reaction between [Cd14]2- and Nos- is slow. The peak separation between the system peak caused by zinc adsorbed weakly on the mini-column and the cadmium peak was improved by decreasing the flow rate. The optimum flow rate of P2 was 1.0 ml min-1 to ensure sufficient sensitivity for the determination of 0.1 pg ml-1 levels of cadmium. The effect of the flow rate of P3 was investigated in the range 0.6-2.0 ml min-1. An optimum flow rate of 1.0 ml min-1 was employed to reduce the consumption of expensive chromogenic reagent and also to avoid dilution of the effluent.The influence of the sample volume on the absorbance was investigated by injecting various volumes ( I 65-870 pl) of cadmium standard and sample solutions into the carrier stream at the recommended flow rate and coil length. The peak height increased steadily with increasing injection volume. It was found that sufficient sensitivity to detect 0.1 pg ml-1 levels of cadmium was obtained by injecting volumes above 350 p1. Therefore, an injection volume of 350 pl was used subse- quently. Switching Sequence of Valves The influence of the adsorption time, which is the time from sample injection to the start of elution of cadmium by switching 0 0.5 1.0 1.5 2.0 Flow rate of P2 / ml min-1 Fig. 2 Flow rate of P2 versus maximum peak height absorbance. Filled circle, standard solution (0.2 yg ml-I Cd); open circle, sample solution (0.4 yg ml-I Cd-150 mg ml-1 Zn).The flow rate of pump P1 was kept at 1.0 ml min-1. The ratio of the flow rate of pump P2 to that of pump P3 was kept at 1.3.10 Analyst, January 1996, Vol. 121 V2, was investigated in the range 140-480 s by injecting 350 pl of cadmium standard solution or sample solution into the carrier stream at the recommended flow rate. The adsorption time had no effect on the response of cadmium and the maximum peak height absorbance was consistently obtained 80-90 s after switching V2. Furthermore, the system peak decreased with increasing adsorption time. This suggested that the cadmium- iodo complex is strongly adsorbed on the anion-exchange mini- column, whereas zinc is weakly adsorbed and washed out by the carrier.Hence, 220 s was chosen as the adsorption time. The elution time, which is the period from the start of the elution to the next sample injection, was also examined. An elution time of 200 s was found to be necessary for the quantitative recovery of cadmium, for which the column wash volume by the eluent was about 3.3 ml. The signal profiles obtained by using the proposed system are shown in Fig. 3. Concentration of Cadmium on the Anion-exchange Mini-column by the Multiple Sample Injection Method As mentioned above, cadmium is strongly adsorbed on the anion-exchange mini-column from potassium iodide media. The concentration of cadmium as the cadmium-iodo complex on the mini-column was attempted. In general, the sensitivity of an FI system is defined and controlled easily by varying the volume of the injection loop, However, replacement of the injection loop for the adjustment of sensitivity to detect various amounts of analyte (from nanograms to micrograms) is tedious and time-consuming.The multiple sample injection method was studied to vary the total injection volume of the sample by using the proposed FI system. Portions (350 pl) of the standard solution (20 ng ml-1 Cd) were repeatedly injected at regular intervals of 30 s. After the last sample injection, the anion- exchange mini-column was washed with the carrier for 220 s, and then the adsorbed cadmium was eluted with 1 moll-’ nitric acid. As can be seen in Fig. 4, the relationship between the number of sample injections and the peak height absorbance was linear from 1 to 50 injections, and the maximum peak height absorbance was consistently observed 80-90 s after switching V2.A ‘carryover effect’ can be observed in Fig. 4 when more than 50 multiple injections are performed. This suggests that a portion of the cadmium-iodo complex adsorbed on the resin was flowing out from the column. For multiple injections at regular intervals of 30 s using a 350 pl loop at a carrier flow rate of 1.0 ml min-1, the total time for sample introduction is 1500 s and the total sample volume injected is 4 min A - 17.5 ml. On the other hand, for a single injection using a 17.5 ml loop at the same carrier flow rate, the total time required for sample introduction is more than 1050 s, because of the dispersion of the injected sample zone in the injection loop.It is suggested that the multiple injection method would allow the preconcentration of a large volume as rapidly as a single injection of the same volume of sample, and the injection of the total volume of sample without changing the injection loop. The reproducibility of the multiple injection method for 50 injec- tions was poorer than that of the single injection method at the same absolute amount of cadmium loaded, because of pulsa- tions caused in the carrier flow. Furthermore, in the multiple injection method, the adsorption band of the cadmium-iodo complex is extended in the mini-column by the large volume of carrier delivered. Consequently, an equivalent amount of the complex is eluted in a larger volume, bringing the peak down closer to the detection limit.Hence, the number of injections affects the reproducibility. If necessary, levels of cadmium of several ng ml-l can be determined by the multiple sample injection method. It should be noted, however, that co-existing elements in the sample solution affect the adsorptivity of the cadmium-iodo complex and that it is necessary to optimize the number of injections for each sample composition. Calibration Graphs A calibration graph was obtained by the procedure described under Experimental. The calibration graph was linear over the range 0.05-2.0 pg ml-1 of cadmium, using the single injection method, and from 0.002 to 0.04 pg ml-1 of cadmium, using the multiple sample injection method (50 injections).Equations of the calibration graphs obtained by the least-squares method are: single sample injection: y = (0.1645 k 0.0082) x + (0.0079 k 0.0021) (n = 3); multiple-sample injection: y = (6.1 f 0.0005) x + (0.0086 f 0.0025) (50 injections, n = 3), where y is the maximum peak height absorbance and x is the concentration of cadmium in pg ml-1. The data regarding each calibration were evaluated: the multiple injection calibration had a sensitivity that was almost 37 times larger than that of the single injection calibration. The peak height of a 0.04 pg ml-1 solution injected 50 times using a 350 p1 injection loop was 15% lower than that of a 2.0 pg ml-1 solution injected in one lot using the same injection loop. It is suggested that the adsorption band of the cadmium-iodo complex is extended in the mini-column by the large volume of carrier delivered.The responses that were obtained by using solutions prepared by the addition of various increments of the standard cadmium solution to the real sample solution were equal to those obtained with the standard 0.4 pg Cd rnl-1 h 0.1 pg Cd rnl-’ t- Scan Fig. 3 Sample solution contained 150 g 1-1 Zn. Typical analytical signals obtained with the proposed FI system. The total volume injected / ul 0 3500 7000 10500 14000 17500 21000 0.2 ! I 5: / I The number of sample injections Fig. 4 Number of sample injections versus maximum peak height absorbance. Sample solution contained 20 ng ml-1 Cd. Aliquots of 350 pl were injected at constant intervals of 30 s.Analyst, January 1996, Vol.121 11 cadmium solution. Therefore, it was decided to quantify cadmium by a simple calibration method, and by using the same multiple injection method for the sample solution. Effect of Foreign Ions The influence of foreign metal ions was studied. In the determination of 0.5 pg ml-1 of cadmium by using the single injection method, the following ions when present in the amounts (pg ml-1) shown in parentheses do not interfere: Cu2+ (lo), Fe2+ (lOOO), Fe3+ (loo), Sn2+ (lOOO), Bi3+ (50), and Pb2+ (30). These elements, therefore, do not affect the determination of cadmium in a hydrometallurgical zinc refining process stream. Other elements such as alkali metals and alkaline-earth metals, and species such as nitrate and sulfate do not interfere at all.Analysis of Zinc Electrolyte and River Water The proposed FI method was applied to the determination of cadmium in several samples. The results obtained for the high- salt concentration solutions of zinc refining process streams by the single sample injection method and for the river water collected from Shibakawa river (Omiya, Saitama, Japan) are given in Table 2. The values obtained with modified JIS HlllOl9 and JIS KO10220 methods are also listed. In order to compare the validity of the results obtained in Table 2, the I t I values for the determination of cadmium in the sample solutions were calculated. For sample No. 6, It1 was higher than the critical value (1.96, degrees of freedom = 00 for a 0.05 significance level), which can be interpreted as the probable existence of a systematic error for this determination. For the other determinations the presence of a systematic error was not proven. The reproducibility was satisfactory with a relative standard deviation of less than 5.0% (0.14 pg ml-1 Cd level, n ~~ Table 2 Results of the determination of cadmium in hydrometallurgical zinc electrolyte and river water (pg ml-l) Proposed FI Sample No.method JIS-method* I t I Zinc electrolyte 1 0.12, 0.13 0.14, 0.14 7.36 2 0.24, 0.24 0.24, 0.25 -t 3 0.26, 0.26 0.24, 0.25 -t 4 0.09, 0.08 0.08, 0.10 1.41 5 0.22, 0.24 0.21, 0.22 2.12 6 0.12 f 0.012$ 0.12, 0.13 15.4 River water <0.01§ < 0.01 -t * Zinc electrolyte: JIS-H11 lO;*9 river water: JIS-K0102.20 t Not calculated. * Average of 152 determinations k standard deviation.Sample analysed 5 Sample injected 50 times at regular intervals of 30 s. at regular intervals of 30 min. = 5 ) for the single injection method and 10% (2.0 ng ml-1 Cd level, n = 5) for the multiple sample injection method (50 injections). The detection limits were 0.028 pg ml-1 of cadmium for the single injection method and 0.83 ng ml-1 of cadmium for the multiple sample injection method (50 injections). The absolute amount of cadmium detectable, defined as the analytical signal equal to twice the uncertainty in the background, was 10 ng. The FI system permits a throughput of 6 samples h-1 for the single injection method and 2 samples h-1 for the multiple sample injection method (50 injections). These results indicate that the proposed FI method is suitable for the on-line routine determination of trace amounts of cadmium in the hydrometallurgical zinc refining process stream.The authors gratefully acknowledge Professor K. Oguma of Chiba University for his useful suggestions. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Hughes, D. V., and Nyberg, J. R., in Process Control and Automation in Extractive Metallurgy, ed. Partelpoeg, E. H., and Himmesoete, D. C., The Minerals, Metals & Materials Society, 1988, p. 157. Gomes Neto, J. A., Bergamin Filho, H., Sartini, R. P., and Zagatto, E. A. G., Anal. Chim. Acta, 1995, 306, 343. Xu, S., Sperling, M., and Welz, B., Fresenius’ J. Anal. Chem., 1992, 344, 535. Purohit, R., and Devi, S., Analyst, 1991, 116, 825. Hirata, S., Honda, K., and Kumamaru, T., Anal. Chim. Acta, 1989, 221, 65. Chung-Gin, H., Chao-Sheng, H., and Ji-Hong, J., Talanta, 1980, 27, 676. Fischer, H., and Leopoldi, G., Mikrochim. Acta, 1937. 1, 37. Saltzman, B. E., Anal. Chem., 1953, 25, 493. Bode, H., and Wulff, K., 2. Anal. Chem., 1966, 32, 219. Escriche, J. M., Estelles, M. L., and Reig, F. B., Talanta, 1983, 30, 915. Kallman, S., Oberthin, H., and Lin, R., Anal. Chem., 1958, 30, 1846. Kallman, S., Oberthin, H., and Lin, R., Anal. Chem., 1960, 32, 58. Korkisch, J., and Klakl, E., Talanta, 1969, 16, 377. Strelow, F. W. E., Anal. Chim. Acta, 1978, 97, 87. Marsh, S. F., Alarid, J. E., Hammond, C. F., Meleod, M. J., Roensch, F. R., and Rein, J. E., Los Alamos Scientific Laboratory report LA- 7084, 1978, February. Sayama, Y., Tokuda, M., and Hayashibe, Y., Anal. Sci., 1995, 11, 849. Stability Constants of Metal-Ion Complexes, ed. Sillen, L. G., and Martell, A. E., Chemical Society Special Publication No. 17, Suppl. 1, The Chemical Society, London, 1971. Igarshi, S., Itoh, J., Yotsuyanagi, T., and Aomura, K., Nippon- Kagakukai-shi, 1978, 2, 212. JIS H1110, Method for Determination of Cadmium in Zinc Metal, Japanese Industrial Standard Committee, Tokyo, 1989. JIS K0102, Testing Methods for Industrial Wastewater, Japanese Industrial Standard Committee, Tokyo, 1989. Paper 5/03134F Received 16 May, 1995 Accepted I 8 September, I995