Comparison of Systems for Eliminating Interferences in the Determination of Arsenic and Antimony by Hydride Generation Inductively Coupled Plasma Atomic Emission Spectrometry ANNA RISNES AND WALTER LUND" Department of Chemistry University of Oslo P.O. Box 1033 N-031 i Oslo Norway The determination of arsenic and antimony by hydride generation inductively coupled plasma atomic emission spectrometry is discussed. A comparison is made of different procedures for eliminating redox (HNO,) and transition metal interferences; two pre-reduction agents three continuous-flow systems and two gas-liquid separators were studied. The pxe- reduction agents were KI and thiourea which were added in a continuous-flow mode either before or after the introduction of tetrahydroborate. The spray chamber served as a gas-liquid separator; alternatively a U-tube separator was used.In the recommended system 10% m/v thiourea is added to the sample stream before 1.0% m/v tetrahydroborate is introduced using the spray chamber as a gas-liquid separator. Thiourea eliminated the oxidation interference from HNO I and also the interference from 50 mg 1-' of Ag+ Cu2+ hi2+ and Fe3+. Keywords Hydride generation; inductively coupled plasma atomic emission spectrometry; arsenic and antimony determination; on-line pre-reduction; thiourea; potassium iodide Sample introduction by hydride generation is used frequently when As Sb and Se are determined by atomic spectrom:try; the technique provides very low detection limits and in add1 tion removal of interferences. Hydride generation is particu! arly well established in the field of AAS,' but it is even riore suitable for ICP-AES because this technique eliminates the gas-phase interferences which may be encountered in AA ?i.2-4 For ICP-AES a continuous-flow hydride generation system is more suitable than a batch system because it ensures plasma stability and multi-element capability In the determination of As and Sb a pre-reduction agent is normally used to reduce any pentavalent As and Sb ta the trivalent state because a low sensitivity is obtained when these elements are present in the higher oxidation state.For samples containing HNO As and Sb may be present in the pentavalent state since trivalent As and Sb are easily oxidized by HNO,. The pre-reduction of the elements to the trivalent state can be achieved by adding K15-8 or t h i ~ u r e a ~ .~ to the sample solution prior to the hydride generation process. However the pre- reduction will interfere with the determination of Se becmse Se" is partly reduced to the elemental state.'-'' Schramel and Xu" found that this interference was eliminated when tht. KI addition was carried out in a flow system. The addition of KI or thiourea also serves to minimize the interference froin a number of transition metals.' The hydrides can be separated from the liquid phase by different devices; the most common is probably a U-tube Journal of Analytical Atomic Spectrometry separator." From the U-tube the hydrides can be introduced directly at the bottom of the torch without disconnecting the spray chamber using a special interface.' However the normal spray chamber can also be used as a gas-liquid separator.In this work different continuous-flow procedures for elimin- ating redox (HNO,) and transition metal interferences are compared. The pre-reduction agents KI and thiourea were added either before or after the introduction of the tetrahydro- borate. The spray chamber or a U-tube served as a gas-liquid separator. EXPERIMENTAL Apparatus and Operating Conditions A Perkin-Elmer (Norwalk CT USA) ICP 5500B sequential inductively coupled argon plasma atomic emission spec- trometer was used. The plasma generator (27.12 MHz) and torch box were identical with those of the Perkin-Elmer ICP 6500 whereas the data system was a Model 3600 computer with a PR-100 printer.An all-quartz torch was used; the outer diameter was 20 mm and the inner diameter of the injector tip was 1.4 mm. For the hydride generation Gilson ( Villiers-le-Bel France) Minipuls 3 (4 channels) peristaltic pumps PVC/PTFE tubing and poly (propylene) T-junctions were used. The acidified sample sodium tetrahydroborate solution and pre-reduction agent were continuously pumped in separate streams to the T- junctions where the reagents were mixed and the hydrides generated. The three different flow systems used are shown in Fig. 1; the flow rates are given in Table 1. In Flow System 1 the pre-reduction agent is added ofS-line to the tetrahydrobo- rate solution; in Flow System 2 the pre-reduction agent is Flow System 1 Sample1 blank Sample/ Flow System 2 Flow System 3 * To whom correspondence should be addressed.Fig. 1 Flow diagrams for the three flow systems Journal of Andytical Atomic Spectrometry October 1996 Vol. 11 (943-948) 943Table 1 Flow rates of sample and reagents Flow rate/ml min-' Gas-liquid Pre-reduction separator Sample NaBH agent Waste U-type 9.2 4.2 9.2* Spray chamber 3.5 1.5 3.5* 13.5t * Only in Flow Systems 2 and 3. 7 6.9 ml min-' for Flow System 1. introduced into the sample stream after the hydride generation whereas in Flow System 3 the pre-reduction agent is introduced before the hydride generation takes place. When the spray chamber served as a gas-liquid separator the reaction mixture passed through the solution inlet of the cross-flow nebulizer while argon at a rate of 1.0 1 min-' passed through the gas inlet.When a U-type gas-liquid phase separ- ator12 was used the hydrides were transported by an argon flow of 1.0 1 min-' to the bottom of the quartz torch through a specially designed ball-joint adapter with a separate hydride inlet. The adaptor which is shown in Fig. 2 constitutes the interface between the spray chamber and the quartz torch with its ball-joint. The adaptor allowed the introduction of hydrides without disconnecting the spray chamber. The operating conditions for the ICP-AES measurements and the hydride generation are given in Table 2. The emission signals were corrected for the blank which was determined using 5 moll-' HCl as the sample. The blanks were only a few per cent. of the analyte signals. 1 cm Fig. 2 Ball-joint adaptor with separate hydride inlet.The adaptor constitutes the interface between the quartz torch and the spray chamber Table 2 Operating conditions for the ICP-AES instrument and the hydride generation ICP-AES- Rf-effect Flow rate of plasma gas Flow rate of auxiliary gas Observation height Integration time Wavelength As Sb Se Background correction Hydride generation- Reducing agent Acid concentration in sample Flow rate of carrier gas (Ar) 1.05 kW 14.0 1 min-' 0.6 1 min- 18 mm 1.0 s 193.70 nm 217.58 nm 196.03 nm kO.08 nm 1.0% m/v NaBH in 0.1 moll-' NaOH 2.0 2.4 or 5.0 mol 1- ' HCl 1.0 mi min-' Reagents and Samples All reagents were of analytical-reagent grade and de-ionized water was used throughout. The 1% m/v sodium tetrahydro- borate solution was prepared by dissolving NaBH (Fluka Buchs Switzerland) in 0.1 moll-' NaOH. The solution was filtered before use and stored in a polyethylene flask at 4°C.Solutions containing 1-10% m/v of KI and thiourea respect- ively were prepared daily by dissolving the reagents in 5 moll-' HCl. Single-element 1 g 1-' stock standard solutions were pre- pared from the salts listed in Table 3. Working standard solu- tions of concentrations lower than 1 g 1-' were prepared fresh daily by dilution. The sample solutions contained 2.0 2.4 or 5.0moll-' HCl and in some experiments also l.lrnoll-' HNO,. The analyte concentration was usually 50 pg 1-' . Recommended System and Procedure The recommended system for hydride generation is shown in Fig. 3. Flow System 3 is used; the sample (in 2.0 moll-' HCl) is first mixed with the pre-reduction agent 10% m/v thiourea (in 5.0mol1-' HC1).The flow rate of both streams is 3.5 ml min- '. The hydrides are then generated by introducing 1.0% m/v NaHB4 (in 0.1 moll-' NaOH) at a rate of 1.5 mlmin-'. The spray chamber is used as a gas-liquid separator. RESULTS AND DISCUSSION Different pre-reduction agents can be used for the selective reduction of As and Sb from the pentavalent to the trivalent state. Potassium iodide is probably the most commonly used reagent but KBr,I3 ~-cysteine'~*'~ and thiourea6*' have also been recommended. Preliminary experiments showed that L-cysteine worked well for Sb but the intensity of the As signal decreased with increasing concentration of L-cysteine. For 4-6% m/v L-cysteine the emission intensity for As was equal to the blank value.In this work the effects of KI and thiourea were studied. Flow Systems The three flow systems shown in Fig. 1 were tested by analysing 50pgl-I solutions of pentavalent and trivalent As and Sb. For each system the effect of pre-reduction agent and gas- liquid separation was studied by calculating the signal intensity ratios AsV :As"' and SbV Sb"'; the ratios should be 1.0 when the pre-reduction is complete. The following concentrations of KI and thiourea were used 0 2 4 6 8 and 10% m/v. In Flow System 1 the pre-reduction agent is added ofl-line to the tetrahydroborate solution; Schramel and Xu" used the same approach. In Flow System 2 the pre-reduction agent is introduced into the flowing stream after the introduction of tetrahydroborate; this approach was also used by Nygaard Table 3 Salts used for preparation of 1 g 1-' stock standard solutions of analytes and interferents; the solutions also contained 0.05 mol 1-' HCl Analyte As"' AsV Sb"' SbV Se" Interfering metal ions Fe3 + cu2 + Ni2 - Ag + Salt Na AsO Na,HAsO * 7H20 Spectrascan (Teknolab Norway) Na2Se03 5H20 KSbOC4H404 '0.5H20 FeC13 -6H,O CuCl 2H20 NiSO 6H,O Ag2S04 944 Journal of Analytical Atomic Spectrometry October 1996 Vol.1 1I I Argon 1.0 I min-1 Sample 3.5 ml min-1 in 2.0 mol 1-1 HCI r I u 13.5 ml min-1 10 9% thiourea 3.5 ml min-1 in 5.0 mol F1 BCI Waste Fig. 3 The Recommended System for hydride generation based on Flow System 3 with 10% m/v thiourea as pre-reduction agent and the spray chamber as gas-liquid separator. All experimental parameters ai.e given in the figure and Lowry' and Pretorius et In Flow System 3 the pre- reduction agent is mixed with the sample before tetrahydrobo- rate is added. The results obtained for the three flow systems are shown in Table 4. For all three systems the signals obtained for trivalent As and Sb were unaffected by the presence of the pre-reduction agent. From Table4 it can be seen that low values are obtained for the As" As"' and Sb" Sb"' ratios in the absence of a pre- reduction agent; the As ratio is in the range 0.2-0.5 whereas the Sb ratio is 0.05-0.2. The As and Sb ratios are lower for Flow Systems 2 and 3 than for Flow System 1. The use of an extra channel in Flow Systems 2 and 3 (in the absence of pre- reduction agent 2.0moll-' HCl was introduced) leads to a dilution which affects the two oxidation states differently. When the pre-reduction agent was introduced by means of Flow Systems 1 and 2 complete pre-reduction was obtained for SbV with 10% m/v KI and the spray chamber as a gas- liquid separator as shown in Table4.However the pre- reduction of AsV was far from complete. Experiments were also carried out with reagent concentrations higher Lhan 2.0 moll-' HCl 1 % tetrahydroborate and 10% pre-reduction agent and with a higher flow rate of the pre-reduction agent but the As ratio remained below 1.0. The best results were obtained when Flow System 3 was used as shown in Table 4. Here complete pre-reduction was achieved for both Asv and SbV by 10% thiourea (and for Sb" also by 10% KI) although only when the spray chamber was used as a gas-liquid separator.The results indicate that it is preferable to carry out the pre-reduction before the hydrides are generated by NaBH4. For all three flow systems the pre-reduction by KI was more effective (faster) for SbV than for As" as has also been observed by other worker^.^*'^ From Table4 it can be seen that it is best to use the spray chamber as a gas-liquid separator. This may partly be due to the low flow rates which were used when the reaction mixture was introduced through the solution inlet of the cross-flow nebulizer because this resulted in a prolonged reaction time. From the experiments with the three flow systems the following conditions are recommended to achieve complete pre-reduction of both As" and Sb" (hereafter called the Recommended System) 10% m/v thiourea is used as pre- reduction agent Flow System 3 is used for mixing the different reagent streams and the spray chamber is used as a gas-liquid separator.The Recommended System is shown in Fig. 3. The concentration of thiourea used is much higher than that required when thiourea is added directly to the sample solution prior to the hydride generationg The detection limits obtained with the Recommended System were 0.3 pg 1-' As and 0.5 pg 1-' Sb; the values correspond to twice the standard deviation for a solution of 1/5 the back- ground equivalent concentration of each element.'8w19 The detection limit was also calculated for Se; a value of 0.5 pg I-' Se was obtained. The detection limits are higher than those obtained using the U-type separator; in the latter instance the detection limits were 0.1 pg 1-' As and 0.2 pg 1-' Sb.The Table 4 Effect of pre-reduction agent and gas-liquid separator on the signal intensity ratios Asv As"' and Sbv Sb"' for Flow Systems 1 2 and 3. The concentration of each species was 50 pg 1-' in 2.0mol1-' HCl. The standard deviations were 0.01-0.05 (n=3) Pre-reduction agent Flow system 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 3 KI KI Thiourea Thiourea KI KI Thiourea Thiourea KI KI Thiourea Thiourea Thiourea Concentration (YO m/v) 0 2 10 2 10 0 2 10 2 10 0 2 10 2 6 10 Gas-liquid separator U-type Spray chamber AsV:As"' 0.46 0.54 0.60 0.60 0.64 0.29 0.32 0.45 0.33 0.64 0.22 0.26 0.5 1 0.34 0.70 0.89 SbV Sb'" 0.20 0.58 0.99 0.28 0.36 0.05 0.3 1 0.58 0.2 1 0.46 0.05 0.37 0.76 0.21 0.41 0.54 AsV:Asl" 0.54 0.57 0.55 0.59 0.65 0.43 0.42 0.46 0.47 0.77 0.32 0.33 0.43 0.38 0.82 0.99 SbV Sb"' 0.19 0.67 0.97 0.25 0.47 0.17 0.69 0.99 0.39 0.77 0.08 0.93 1.03 0.50 0.86 0.98 Journal of Analytical Atomic Spectrometry October 1996 Vol.1 1 945inferior detection limits for the Recommended System are due to the low sample introduction rate and a high background caused by a less complete gas-liquid separation with droplets entering the p l a ~ m a . ~ Unfortunately the presence of thiourea depresses the Se" signal,' because SeIV is partly reduced to the elemental state. The effect of the Recommended System on this process was studied; it was found that the Se" signal was depressed by 14%. Effect of HNO in Sample Solution Nitric acid is often present in solutions of real samples after the sample decomposition step.The effect of HNO is of particular interest because the acid can oxidize trivalent As and Sb to the pentavalent state and in addition it will lower the effective concentration of the pre-reduction agent. The effect of 1.1 mol I-' HN03 on the signal intensities of As"' Sb"' and Se" is shown in Table 5 using the Recommended System. As can be seen the signal intensities for As and Sb in the presence of HNO were only 0.33 and 0.07 respectively relative to the intensities obtained in the absence of HNO,. These values are almost the same as those given in Table 4 for the AsV As"' and SbV Sb"' ratios using Flow System 3/spray chamber without a pre-reduction agent.Thus the results in Table 5 show that HNO oxidizes trivalent As and Sb to the pentavalent state even at room temperature. For our study this explanation is more likely than inhibition of hydride evolution by nitrogen oxides.20 The positive effect of thiourea is also demonstrated in Table 5; 10% m/v thiourea eliminates the depression of the As and Sb signals caused by 1.1 moll-' HNO,. For Se the presence of HNO has no effect since Se" is not oxidized by HNO,. However thiourea depresses the Se" signal because SeIV is partly reduced to the elemental state. Interference From Transition Metals In high concentrations metals such as Fe Cu Ni Co Ag Au Pt Pd and W can interfere with the generation of the h y d r i d e ~ . ' - ~ - ~ ' - ~ ~ The predominating reaction is probably the reduction of the interfering metal ion by tetrahydroborate which results in the formation of metal particles or metal b o r i d e ~ ; ~ ~ ~ ' the finely dispersed precipitate may adsorb and decompose the hydrides.The interference can be minimized in different ways.26 A simple approach is to increase the concen- tration of HCl.27*28 A decrease in the tetrahydroborate concen- tration may also have a beneficial effect.29 Alternatively a masking agent can be used; a variety of such agents have been evaluated.,' More recently the use of EDTA diethylenetriami- nepentaacetic acid (DTPA) and tartrate,2-4*31 ~ - c y s t e i n e ' ~ * ' ~ ' ~ ~ and thiourea' was described. Thiourea has been used to mask the interference from transition metals in the determination of As9* and Sb.6,9734 Thiourea was found to be a better masking agent than KI.6 Thiourea is one of a few ligands that forms complexes with metal ions in strongly acidic solution.In this work the interference from Ag' Cu2+ Ni2+ and Fe3+ was studied; these are among the elements that often interfere the most.' The interfering metals were studied at concentrations of 0 0.05 0.5 5.0 and 50 mg 1-'; the results for 5.0 and 50 mg 1-' concentrations are shown in Tables 6 7 and 8 for As Sb and Se respectively. When the concentration of Ag and Cu was above 5 mg 1 - ' a precipitate was formed when NaBH was introduced in the absence of thiourea. The precipi- tation was prevented by thiourea and thus a time-consuming acid cleaning of the system was avoided. From Table 6 it can be seen that the As signal is depressed by all of the transition metals at a concentration of 50 mg 1-'.However in the presence of thiourea the interference from the transition metals is eliminated. It can also be seen that HNO alone depresses the As signal more than the transition metals do owing to the oxidation of As"' to AsV; the signal is not further depressed by 50 mg 1-' of Ag Ni Cu or Fe. Thiourea eliminates the combined interference from HNO and trans- ition metals. The effects observed for Sb are similar to those described for As but the signal depression by the transition metals and HNO is more severe as shown in Table 7. The depression by HNO is ascribed to the oxidation of Sb"' to SbV; the signal is not further depressed by 50mg 1-1 of Ag Ni Cu or Fe.Again thiourea eliminates the interference from the transition metals and the combined effect of HNO and these metals. Table 5 signal intensities; 1.00 means no depression effect. The Recommended System was used for hydride generation; n = 8 Effect of HN03 and thiourea on the signal intensity of 50 pg l-.' of AS"' Sb"' and Se" respectively. The results are given as relative 2.4 mol I-' HCl 1.1 mol 1-' HN03+2.4mol I-' HCl Without thiourea 10% thiourea Without thiourea 10% thiourea Analyte Mean S Mean S S Mean S Mean As"' 1 .oo 0.022 1.01 0.013 0.33 0.01 1 0.97 0.008 Sb"' 1 .oo 0.0 12 1 .oo 0.013 0.07 0.007 0.97 0.012 Se" 1 .oo 0.022 0.84 0.019 0.99 0.01 5 0.72 0.065 Table 6 Effect of transition metals HN03 and thiourea on the signal intensity of 50 pg 1-' As"'.The results are given as relative signal intensities; 1.00 means no depression effect. The Recommended System was used for hydride generation. The standard deviations were 0.01-0.06 (n = 2) Interferent 2.4 mol 1 - ' HCI 1.1 mol I-' HN03+2.4mol 1-' HCI Metal ion Concentration/mg 1- 0 5.0 50 Ni2 + 5.0 50 c u 2 + 5.0 50 Fe3 + 5.0 50 Ag+ Without thiourea 1 .00 1.01 0.59 0.98 0.85 0.95 0.67 0.64 0.60 10% thiourea 1.01 1.02 0.98 0.98 1.04 0.99 1.03 0.96 1.03 Without thiourea 0.33 0.33 0.32 0.32 0.30 0.3 1 0.30 0.3 1 0.32 10% thiourea 0.97 0.96 0.95 0.93 0.96 0.97 0.97 0.94 0.95 946 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11Table7 Effect of transition metals HNO and thiourea on the signal intensity of 50pg I-' Sb"'. The results are given as relative signal intensities; 1.00 means no depression effect.The Recommended System was used for hydride generation. The standard deviations were 0.01-0.04 (n=2) In te rfe ren t 2.4 rnol I - ' HCI 1.1 mol 1-' HN03+2.4 mol 1-' HCI Metal ion Concentration/mg I-' Ag+ Ni2 + c u 2 + Fe3 + 0 5.0 50 5.0 50 5.0 50 5.0 50 Without thi1:urea 1 .oo 0.80 0.20 1 .oo 0.81 1.03 0.07 0.37 0.08 10 % thiourea 1 .oo 1.01 0.97 0.99 1 .oo 1 .oo 0.98 0.96 0.95 Without thiourea 0.07 0.06 0.06 0.10 0.05 0.07 0.07 0.10 0.1 1 10% thiourea 0.97 0.98 0.98 0.94 0.99 0.95 0.97 0.99 0.91 Table 8 Effect of transition metals HNO and thiourea on i.he signal intensity of 50 pg I-' SetV. The results are given as relative signal intensities; 1.00 means no depression effect. The Recommended System was used for hydride generation.The standard deviations were 0.01 -0.07 (n=2) Interferent 2.4 rnol 1-' HCI 1.1 rnol I - ' HN03+2.4mol I - ' HCI Metal ion Concentration/mg 1- ' 0 5.0 50 5.0 50 5.0 50 5.0 50 Without thiourea 1 .oo 0.13 0.98 1.01 1.04 0.72 0.97 0.93 - 10% thiourea 0.84 0.85 0.82 0.8 1 0.83 0.77 0.15 0.82 0.57 Without thiourea 0.99 0.1 1 0.96 0.94 0.96 0.41 0.97 0.98 - 10% thiourea 0.72 0.80 0.62 0.67 0.67 0.59 0.09 0.70 0.49 Table9 Effect of a mixture of Ag+ Ni2+ Cu2+ and Fe3+ in equal amounts HN03 and thiourea on the signal intensity of As"' Sb"' and Se" for a mixture of the analytes (50pg I-' of each). The results are given as relative signal intensities; 1.00 means no depression effect. The Recommended System was used for hydride generation.The standard deviations were 0.01-0.06 (n = 2 ) Analyte As'" As"' A s'" Sb"' Sb"' Sb"' SeiV SetV Interferent concentration/ mg 1-'* 0 5 .O 50 0 5.0 50 0 5.0 2.4 m01 1- ' HCl Without thiourea 10% thiourea 1 .oo 0.95 0.78 1.00 0.60 0.40 1.00 0.18 1.03 0.98 0.99 1.03 0.97 0.96 0.80 0.62 1.1 rnol 1-1 HNO,+2.4 rnol 1-' HCI Without thiourea 10% thiourea 0.37 0.35 0.35 0.07 0.08 0.05 0.94 0.25 0.96 0.95 0.96 0.95 0.93 0.92 0.73 0.5 1 * Concentration of each of the interfering metals Ag+ Ni2+ a h 2 + and Fe3+. For Se the signal depression by the transition metals is very marked for Ag and less so for Cu while Ni and Fe do not interfere significantly at 50 mg l-' as shown in Table 8. Nitric acid does not depress the Se signal because HNO cannot oxidize Set" to Se".However thiourea depresses the Se" signal particularly when 50 mg 1 - l Cu is present and has a positive effect only on the interference from Ag. In Table 9 the relative signal intensities of As"' Sb"' and Se" are given for solutions containing a mixture of As Stl and Se the transition metals Ag Ni Cu and Fe and HNO For As and Sb it can be seen that thiourea eliminates the inter- ference from HN03 and a mixture of 50 mg 1-' of each of the transition metals. Even for Se thiourea partly eliminate#*; the interference from the transition metals probably becaus.;e of the positive effect on the interference from Ag (Table 8). CONCLUSION Trivalent As and Sb are oxidized to the less sensitive ptmta- valent state when HNO is present in the sample solution. Pentavalent As and Sb are fully reduced to the trivalent state by 10% m/v thiourea in a continuous-flow system using the spray chamber as a gas-liquid separator.Thiourea is more effective than KI for the pre-reduction of AsV. Thiourea also eliminates the interference from 50mg1-' of each of the transition metal ions Ag' Cu2+ Ni2+ and Fe3+. However the use of a continuous-flow system for the addition of the pre-reduction agent does not eliminate the interference from thiourea on the SeIV signal. REFERENCES 1 Dedina J. and Tsalev D. L. Hydride Generation Atomic Absorption Spectrometry Wiley Chichester 1995. 2 Wickstrram T. Lund W. and Bye R. J . And. At. Spectrom. 1995 10 809. 3 Wickstr~m T. Lund W. and Bye R. Analyst 1995 120 2695. 4 Wickstrarm T. Lund W. and Bye R. Analyst 1996 121 201.5 Nakahara T. Anal. Chim. Acta 1981 131 73. 6 Nakahara T. and Kikui N. Anal. Chim. A d a 1985 172 127. Journal of Analytical Atomic Spectrometry October 1996 Vol. 11 9477 Sinemus H. W. Melcher M. and Welz B. At. Spectrosc. 1981 2 81. 8 Nygaard D. D. and Lowry J. H. Anal. Chem. 1982 54 803. 9 Uggerud H. and Lund W. J. Anal. At. Spectrom. 1995 10 405. 10 Thompson M. Pahlavanpour B. Walton S. J. and Kirkbright G. F. Analyst 1978 103 705. 11 Schramel P. and Xu L.-Q. Fresenius’ J. Anal. Chem. 1991 340,41. 12 Thompson M. Pahlavanpour B. Walton S. J. and Kirkbright G. F. Analyst 1978 103 568. 13 Thompson M. Pahlavanpour B. and Thorne L. T. Water Res. 1981 15 407. 14 Chen H. Brindle I. D. and Le X.-C. Anal. Chem. 1992 64 667. 15 Chen H. Brindle I. D.and Zheng S. Analyst 1992 117 1603. 16 Pretorius L. Kempster P. L. van Vliet H. R. and van Staden J. F. Fresenius’ J. Anal. Chem. 1992 342 391. 17 Nadkarni R. A Anal. Chim. Acta. 1982 135 363. 18 Inductively Coupled Plasmas in Analytical Atomic Spectrometry eds. Montaser A. and Golightly D. W. VCH New York 2nd edn. 1992 p. 262. 19 Method Development for ZCP Spectrometry User Manual Perkin- Elmer Norwalk CT. 20 Brown R. M. Fry R. C. Moyers J. L. Northway S. J. Denton M. B. and Wilson G. S. Anal. Chem. 1981 53 1560. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Nakahara T. and Kikui N. Spectrochim. Acta Part B 1985 40,21. Smith A. E. Analyst 1975 100 300. Henden E. Analyst 1982 107 872. Bye R. Talanta 1986 33 705. Bax D. Agterdenbos J. Worrell E. and Kolmer J. B. Spectrochim. Acta Part B 1988 43 1349. Nakahara T. Spectrochim. Acta Rev. 1991 14 95. Welz B. and Melcher M. Spectrochim. Acta Part B 1981,36,439. Hershey J. W. and Keliher P. N. Spectrochim. Acta Part B !986 41 713. Astrsm O. Anal. Chem. 1982 54 190. Nakahara T. Prog. Anal. At. Spectrosc. 1983 6 163. Wickstrsm T. Lund W. and Bye R. J. Anal. At. Spectrom. 1995 10 803. Le X.-C. Cullen W. R. Reimer K. J. and Brindle I. D. Anal. Chim. Acta 1992 258 307. Peacock C. J. and Singh S. C. Analyst 1981 106 931. Narsito Agterdenbos J. and Bax D. Anal. Chim. Acta 1991 244 129. Paper 61033066 Received May 13 1996 Accepted July 5 1996 948 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11