JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 23 1 Sensitive Inductively Coupled Plasma Atomic Emission Spectrometric Determination of Cadmium by Continuous With Sodium Tetraethylborate* M. C. Valdes-Hevia y Temprano M. I?. Fernandez de la Campa and A. Sanz-Medelt Department of Physical and Analytical Chemistry Faculty of Chemistry University of Oviedo clJulian Claveria 8 33006-Oviedo Spain Al ky I at i on Continuous flow generation of volatile Cd species using NaBEt as a means of gaseous sample introduction into an inductively coupled plasma atomic emission spectrometry (ICP-AES) instrument has been investigated in detail. The assumed generation of CdEt2 is discussed and critically evaluated in terms of the sensitivity selectivity and accuracy of the corresponding ICP-AES determination of low levels of Cd.The proposed method for the determination of Cd by ICP-AES using NaBEt provided detection limits of 0.4 ng ml-’ which are ten times better than those of conventional nebulization ICP-AES. A precision of & 1.4% at the 50 ng ml-’ level of Cd was observed. Interference studies have been carried out which demonstrated high selectivity of the proposed method. This method has been applied to the determination of low levels of Cd by ICP- AES in samples of sea-water and tea infusions. Satisfactory validation of the results obtained has been provided by electrothermal atomic absorption spectrometric analysis of the same samples. Keywords Cadmium; inductively coupled plasma atomic emission spectrometry; continuous vapour generation; sodium tetraethylborate(iti) The most important feature that distinguishes heavy metals from other toxic pollutants is that metals as such are not biodegradable even though their potential toxicity in the environment as with other toxic metals is controlled largely by their physico-chemical form.Cadmium is therefore charac- terized by a long persistence in the environment and biological ‘half-life’ which accounts for its bioaccumulation in individuals.’ In recent years several cases of Cd poisoning have been reported2 and experimental results have confirmed its high toxicity. In fact Cd is considered together with Hg by national and international legislation to be the most toxic of metals. Consequently the maximum permissible concentrations of Cd in environmental food and freshwater samples3 are always extremely low.Therefore reliable control of this element in those samples requires very sensitive analytical techniques and the detection power of inductively coupled plasma atomic emission spectrometry (ICP-AES) with conventional nebuliz- ation is insufficient to determine the concentrations of Cd usually found in such samples. Although determinations of Cd using the generation of the volatile chloride have been reported the generation of the volatile hydride from aqueous solutions at room temperature which is so useful to increase the sensitivity for other elements in atomic absorption spectrometry (AAS) and ICP-AES is very difficult with Cd2+ owing to the instability of its hydride at temperatures above those of liquid nitrogen.’ In spite of such instability CdHz has been proposed as the volatile species of Cd formed when treating Cd2+ solutions with NaBH in an organic medium of dimethylformamide6 or in a vesicular organized medium of didodecyldimethylammonium bromide ( DDAB).7 A survey of the organometallic literature’ shows that the alkylboranes are able to alkylate several cationic compounds of metals such as Pb Hg Sn T1 Cd Zn and Cu to give fairly volatile organometallic species. Thus a possible alternative route to the more stable volatile Cd species is alkylation of the metal using alkylboranes.Honeycutt and Riddle,’ first reported the use of sodium tetraethylborate (NaBEt,) in aqueous solutions to produce volatile organometallic species * Presented at the XXVIIl Colloquium Spectroscopicurn Internationale (CSI) York UK June 29-July 4 1993.t To whom correspondence should be addressed. of Pb and Hg from their inorganic salts and this approach has already been applied to the determination and speciation and their alkyl ions. D’Ulivo and Chen25 have reported the use of such alkylation reactions for the atomic fluorescence spectrometric (AFS) determination of very low amounts of Cd even if no conclusive identification of the exact Cd species formed could be achieved. In the present paper the formation of such volatile Cd species with NaBEt and its application to increase the sensi- tivity of Cd in determinations by ICP-AES have been investi- gated thoroughly. The effect of the nature of organized media and temperature on the efficiency of generation of the Cd volatile species have been studied.A new highly sensitive vapour generation (VG) ICP-AES method for the determi- nation of Cd is proposed which has been succesfully applied to the determination of low levels of this toxic metal in samples of sea-water and tea infusions. of pb,1&13 ~~,10,14-18 Sn 10.19-22 Sb and Ge,10,23 Se 24 7 Experimental Instrumentation An inductively coupled plasma atomic emission spectrometer Philips Model PU7000 was used for detection by ICP-AES. An atomic absorption spectrometer Perkin-Elmer Model 2280 equipped with a hollow cathode lamp (Perkin-Elmer) and a recorder (Perkin-Elmer Model 56) was used for AAS detection in connection with a laboratory-made batch hydride generator. The gas-liquid separation interface used was a grid- type nebulizer and spray chamber provided with the ICP instrument (see below). The Gilson Minipuls 2 peristaltic pump and the experimental flow system used are shown diagramat- ically in Fig.1. Reagents A 1000 pg ml-’ Cd” stock standard solution was stabilized in 0.5 mol 1-1 HN03 (Merck Darmstadt Germany). Working solutions were freshly prepared daily by diluting appropriate aliquots of this stock solution with ultrapure water. Sodium tetraethylborate(rI1) solutions were prepared just before use by dissolving the solid reagent (Alfa Ventron Danvers MA USA) in ultrapure water (Milli-Q Millipore Milford MA USA) stabilized in a 1% m/v NaOH solution. The reagent was stored in the dark in poly(tetrafluoroethy1ene)232 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 n ICP-AES r- 1 - 1 I I +111 A Per is t a I t i c Pump I I Ar nebulizer Fig. 1 Continuous vapour generation flow system (PTFE) vessels. Aqueous NaBEt solutions were stable for about 1 week when stored at 4 "C in darkened capped bottles. Cetyltrimethylammonium bromide (CTAB) solution (1 x lop2 moll-') was prepared by dissolving the surfactant powder (Fluka Buchs Switzerland) in water by gentle warm- ing. Other surfactants such as sodium lauryl sulfate (SLS) solution (1 x lop2 mol l-l Sigma St. Louis MO USA) Triton X-100 (TX-100) solution (2% v/v Merck) Zwittergent-3.16 (ZW-3.16) solution (1 x mol I-' Carbiochem-Behring La Jolla CA USA) were prepared in a similar way. The didodecyl- dimethylammonium bromide (DDAB) and the dihexadecyl phosphate (DHDP) vesicles solutions (1 x lo-' moll-') were prepared by dissolving the surfactant powders (Kodak Rochester NY USA and Aldrich Milwaukee WI USA respectively) in water and sonicating at room temperature (DDAB) or at 90 "C (DHDP) at a power of 60 W for about 12min with the tip of a high-intensity ultrasonic processor (Sonics & Materials Danbury CT USA).All mineral acids and metal salts used were of analytical- reagent grade and ultrapure water (Milli-Q) was used throughout. General VG-ICP-AES Procedure Continuous generation of volatile diethylcadmium In the flow system shown schematically in Fig. 1 the Cd sample dissolved in 0.2 mol I-' HCl was pumped continuously through one of the channels of the peristaltic pump at a rate of 0.75mlmin-' and merged with a 1.2% m/v solution of NaBEt at the same flow rate.This mixed solution feeds the grid nebulizer of the ICP instrument detuned in order to allow for separation of the volatile species that would eventually reach the torch of the plasma while the liquid phase goes to drain. Cadmium is measured at the 214.440 nm emission line and the instrumental conditions are detailed in Table 1. Background correction using two off-the-line points at 214.408 and 214.468 nm was used throughout. Samples Samples of sea-water from Villaviciosa (Asturias Spain) and infusions of commercial tea were analysed without any pre- treatment simply by following the continuous flow generation of volatile Cd species and the general ICP-AES procedure described above.The infusions of tea were prepared by extrac- tion of 2 g of tea in 250ml of Milli-Q water with gentle warming. Results and Discussion Optimization of Instrumental and Chemical Parameters Using the procedure outlined above for continuous gas-liquid separation of volatile Cd and ICP detection the effect on the Table 1 Optimum conditions for generation of CdEt (a) Generation by continuous VG-ICP-AES- Optimum plasma experimental conditions 'Wavelength 214.440 nm 1h.f. forward power 1.0 kW Aerosol gas pressure Outer gas flow htermediate gas flow :Final sample introduction [ntegration time 3 s 40 psi* 13 1 min-' 0 1 min-' flow rate 1.5 ml min Optimum chemical parameters HCl NaBEt (b) Generation by batch VG-AAS- 0.2 mol 1-' (flow rate 0.75 ml min-') 1.2% m/v in 1% m/v NaOH (flow rate 0.75 ml min-') Optimum AAS experimental conditions Wavelength 228.8 nm Lamp intensity 7 mA Slit 0.7 nm Gain 75 v Air flow rate C2H flow rate Argon flow rate Total volume 5 ml 15.5 1 min-' 2 1 min-' 1.5 1 min-' Optimum chemical parameters HCl 0.2 rnol I-' NaBEt Injection of 2 ml of 1.2 O/O m/v in 1% m/v NaOH * 1 psi = 6894.8 Pa.Cd signals of plasma instrumental variables such as nebulizer gas pressure forward r.f. power and coolant gas flow were studied. The variables in the chemical generation of the volatile species such as concentration of reagents and flow rates were investigated by following a univariant-type experimen- tal search. Maximum signal-to-background ratio at the 214.440 nm line of Cd was always the optimization criterion. The optimum ICP instrumental values observed in such experi- ments are summarized in Table 1 (a).For optimum instrumental settings chemical generation parameters were then investigated (concentration flow and final pH). The results observed have been plotted in Fig. 2. As can be seen in Fig. 2 A the effect of HCl concentration is fairly critical with an optimum pH of 2.1 for generation. Fig. 2 B shows the effect of the concentration of NaBEt,. This was not as critical as pH and a concentration PH .- - +d v) 1.0 1;2 1.; 2;OA c S i- e m- 200 2 250 *.I .- 0 [NaBEt,] (YO m/v) Fig. 2 in the Cd signal by VG-ICP-AES Effect of A sample acidity (HCl) and B NaBEt concentrationJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 233 of 1.2% m/v was eventually selected to save on this expensive reagent without important losses in sensitivity.The influence of sample flow rate was also studied between 0.25 to 5 ml min-'. As expected the signal increased with the sample flow rate then a plateau was reached for stationary vapour generation-excitation conditions after a flow of 1.5 ml min-' which was the final value selected for subsequent experiments. Optimum values selected for ICP instrumental plasma con- ditions and for continuous chemical generation of CdEt are summarized separately in Table 1 (a). Studies of the influence of temperature and efficiency of the generation-volatilization of Cd were carried out by batch AAS because of the impossibility of obtaining these data by continu- ous VG-ICP-AES. Optimum parameters for CdEt generation by batch AAS are given in Table l(b) the acidity of the solution and the flow rate of the carrier gas being the most critical parameters affecting the Cd AAS signal.Analytical Performance Characteristics Using all the experimental conditions given in Table l(a) the analytical performance characteristics of the corresponding VG-ICP-AES method were evaluated. The observed slope of the calibration lines detection limits and precision of the determinations of Cd by the proposed VG-ICP-AES method are summarized in Table 2. Characteristics observed for conventional nebulization are also given for comparison. The ICP-AES signals observed for conventional nebulization ( 100 ng ml- ' of Cd) CdH generation from DDAB7 and CdEt generation with NaBEt are shown for comparison in Fig. 3.The calibration graph in the last case was linear up to 1 pg ml-' of Cd and the detection limit (3sb IUPAC criterion) was 0.4 ng ml-I of Cd (10-fold better than the value observed Table 2 Analytical performance characteristics for determination of Cd by ICP-AES Method Detection limit*/ Slope/ Precision ng ml-' ml ng-' (%)? Conventional Vapour generation nebulization 5 2.66 x 105 1.3 with NaBEt 0.4 2.92 x lo6 1.4 *Detection limit = 3sb/slope. ?Precision (RSD %)=relative standard deviation at the 50 ng ml-' level. Precision values are within-run results (ten replicates). 350 A .g 300 3 > 2 250 L c .- + 2 200 g 100 .- fn .- 50 I I 214.41 214.44 214.47 Wavele ng t h/nm Fig.3 Emission spectra of 100ngmlF' of Cd. A Generation of CdEt from water; B generation of CdH from DDAB vesicles; and C conventional nebulization using conventional nebulization and twice better than CdHz generation from DDAB).The within-run precision evaluated by analysing ten independent replicates of a solution containing 50 ng ml-' of Cd was & 1.4%. Effect of 'Organized Media' on the Generation of CdEt It has been shown previously by this group that 'organized media' such as micelles and vesicles can improve the sensitivity of hydride generation because of the new microenvironment created in the solution. The kinetics of the generation of volatile species from the desired analyte can be improved,26 and this effect has been applied to improve the determination of Cd As and Pb by ICP-AES via 'cadmium hydride' arsine and plumbane generation in organized media.7-26.27 Several types of organized assemblies were assayed including cationic (CTAB) anionic (SLS) zwitter-ionic (ZW-3.16) and non-ionic (Triton X-100) micelles and also anionic (DHDP) and cationic (DDAB) vesicles.Using the experimental con- ditions given in Table l(a) the analytical parameters of the corresponding VG method for the different reaction media were evaluated. The observed influence of different organized media on the slope detection limits and precision of the determination of Cd by continuous VG-AAS and by VG-ICP- AES (in this latter case for those two surfactants which had shown the best behaviour in AAS i.e. TX-100 and DDAB) is summarized in Table 3. As can be seen unexpectedly the effect of organized media on the generation of volatile CdEt proved not to be as benefitial as in the case of volatile CdH (poss- ible specie^).^ Interference Studies Using the selected optimum conditions given in Table 1 (a) the effect of the presence of foreign elements on the VG-ICP-AES signal of Cd was investigated.All of the potentially interfering elements tested and the levels of tolerance observed in the corresponding determi- nations of Cd are summarized in Table4. Hydride forming elements and high levels of alkali alkaline earth metals or common anions were found not to affect generation of CdEt,. Only Zn and Ni could be a problem but only if present at relatively high excesses ( 1 500 = Cd interferent). In other words the proposed method is fairly selective for Cd. Effect of Mineral Acids and Their Concentration on the Cd Signal The different mineral acids presently used in many sample digestions including HC1 HN03 H2S04 HClO and HF were tested in the alkylation of Cd with NaBEt,. All these mineral acids (except HF which was tested in connection with H3BO3) showed similar behaviour to that illustrated for HCl and HNO in Fig.4 showing a maximum signal at a concen- tration of 0.2 mol I-' (pH 2.1). The use of HCl provided the highest signal for Cd and was selected for use in further experiments. In order to clarify this behaviour of HCl in the generation of volatile Cd species the effect of chloride as NaCI was also studied using HNO for sample acidification. Concentrations of NaCl of between 0.01 and 1 moll-' were found not to affect the observed Cd signals.Thus it seems that it is HCl that favours the metal alkylation that is observed. Influence of Generation Temperature The effect of the reaction vessel temperature where alkylation takes place was studied between 0 and 65°C using AAS measurements of Cd. Results showed that temperature greatly affects the kinetics and also the observed efficiency of the reaction for generation of the volatile species. During these234 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Table 3 AES detection Comparison of different organized media for volatile species of Cd using NaBEt with continuous VG-AAS and continuous VG-ICP- (a) AAS detection- Media Water DDAB SLS TX-100 CTAB DHDP ZW-3.16 (b) ICP-AES detection- Media Water DDAB TX-100 Surfactant concentration/ mol I-' 1 x 1 x - 1 x 10-3 1 x 10-3 0.15 1 x lo- Surfact ant concentration/moll-' - 10-3 Slope/ ml ng-' 1.668 1.990 0.359 1.888 1.449 1.751 1.749 Detection limit?/ ng ml-' 0.4 0.62 0.36 A -k sb* 0.028 f 0.002 0.03 1 & 0.002 0.020 & 0.004 0.039 k 0.002 0.028 & 0.003 0.041 f 0.005 0.028 k 0.003 Slope/ ml ng-' 2.92 x lo6 3.30 x lo6 3.48 x lo6 DLt/ ng ml-' 3.6 2.4 3.8 6.2 8.6 5.1 35 Precision 1.4 2.3 1.6 W)$ Precision W)$ 2.3 2.5 3.9 2.8 2.4 3.7 2.5 * Ab (mean value of absorbance for ten replicates of the blank); and sb standard deviation for ten replicates of the blank. ?Detection limit = 3sb/slope.1 Precision =relative standard deviation (RSD %) on 50 ng ml-I (n = 10). Table 4 Interference studies on 50 ng ml-' of Cd by generation of CdEt and ICP-AES ~~ Interferent As"' Zn" CU" M n" Ni" Fe"' Hg" Pb" M g" Ca" Na' K' Nitrate Chloride Cd interferent mass ratio 1:lO 1 100 1 500 1 10 1 100 1 500 1 10 1 100 1:500 1:lO 1:100 1:500 1 10 1 100 1 500 1 10 1:100 1 500 1 10 1 100 1:500 1 10 1 100 1 500 1:lO 1 100 1:500 1 10 1 100 1 500 1 500 1 1000 1 500 1 1000 1 1000 1 10000 1 1000 1 10000 Amount of interferent Recovery pg ml-' (%)* 0.5 99 5 103 25 96 0.5 99 5 101 25 93 0.5 100 5 99 25 96 0.5 102 5 99 25 100 0.5 99 5 93 25 91 0.5 98 5 99 25 98 0.5 99 5 99 25 99 0.5 100 5 99 25 99 0.5 99 5 102 25 101 0.5 99 5 100 25 98 25 98 50 99 25 98 50 99 50 99 500 100 50 99 500 101 *The precision observed in each case (expressed as RSD) were within +3%.0 0.05 0.10 0.15 0.20 0.25 0.30 Acid concentratiordmol r' Fig.4 Effect of A HCI B HNO and C HF on the ICP-AES signal of Cd by VG-ICP-AES temperature experiments the generation was carried out using a Cd concentration of 25 ng ml-' in a final volume of 5 ml. An adequately thermostated typical batch AAS hydride gener- ation system and the conditions specified in Table l(b) were used. The results obtained have been plotted in Fig. 5 for peak height and peak area of the observed AAS transient signals. They show that with increasing reaction temperatures forma- tion of Cd and transport to the atomizer increases. At tempera- tures around 60°C the reaction for generation of the volatile species of Cd and transport is very fast. In other words at temperatures higher than room temperature for the vapour generation a further decrease in the detection limit for Cd as compared with the values given in Table 2 could be obtained.Efficiency of the Generation-Volatilization of the Cd Species The efficiency of the generation of volatile Cd using NaBEt was evaluated. To do so the amount of Cd in the residual aqueous solutions after the corresponding tetraethylborate reaction in batch was determined by electrothermal (ET) AAS. Six independent standard solutions of 100 ng ml-' of Cd were treated with tetraethylborate at room temperature under the optimized experimental conditions [Table 1 (b)] and theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 235 Oa7* 25 Table 5 Determination of Cd in real samples by generation of CdEt and ICP-AES 0 10 20 30 40 50 60 U Temperatu rePC Fig.5 Temperature effect on Cd signal by batch-VG-AAS remaining Cd was determined in the residual aqueous solu- tions. This was repeated at a reaction temperature of 60°C. The results obtained showed that the efficiency of analyte volatilization was (79.5 f 1.9)’?40 at room temperature and (98.8 +_ 1.2)% at 60 “C. Preliminary efforts to characterize the exact nature of the observed volatile species of Cd measured in the ICP (formed when the analyte and NaBEt solutions merge in the continu- ous VG system of Fig. 1) have failed so far [attempts to generate the possible ‘CdEti2’ in the batch mode and analyse it by gas chromatography-mass spectrometry (GC-MS) for element identification have been unsuccessful so far but studies of alternative techniques are in progress]. However it is worth noting that the VG (volatile species generation) technique proposed here offers ten times better detection limits for Cd than conventional nebulization ICP- AES.Analysis of Real Samples Once the best conditions for the generation of diethylcadmium had been established the recommended VG-ICP-AES pro- cedure was applied to the direct determination of low levels of Cd in sea-water and tea infusions. Background correction (at 214.408 and 214.468nm) was employed and the other con- ditions used were as detailed in Table l(a). For sea-water samples known amounts of Cd were spiked into the real samples as their Cd contents were undetectable and the samples were then analysed using the recommended procedure without any pre-treatment.Tea samples were ana- lysed directly and the results obtained by the proposed method were compared with those by ETAAS for the same samples obtained in our laboratory. The values observed in all cases are summarized in Table 5. As can be seen very good agreement between the expected and the obtained results was observed. Therefore the validity of the new VG-ICP-AES method proposed for the determi- nation of low levels of Cd has been shown in the analysis of sea-water and tea infusions. Conclusions It has been demonstrated that the generation of alkylated volatile species of Cd (probably CdEt2) using NaBEt as the reducing agent can be applied to the determination of low levels of Cd. This method has proved to be much more sensitive than conventional nebulization.The selectivity is also Sea-water- Cd concentration*/ng ml-’ Sample 1 2 3 4 5 6 Tea infusion- ~ ~~~ ~~~ VG-ICP Cd added 4.1 f0.9 4 7.3 k 0.8 7 9.8 & 0.5 10 5.2 & 0.8 5 10.6 f 0.6 11 3.3 f 0.9 3 Cd concentration*/ng ml-’ Sample VG-ICP ETAAS Normal tea 1 5.3 f0.8 5.1 k0.3 Normal tea 2 22.0 f 0.5 22.8 f 0.9 Jasmine tea 6.0 f 0.9 6.1 rfi 0.2 Grey tea 1 17.4f0.6 17.1 k0.8 Grey tea 2 10.9f0.7 10.7k0.2 Orange tea 8.2 k 1.0 7.9 0.4 *Mean f SD (n = 3). Analysis using calibration line. high and so it is more appropriate for the determination of the metal in environmental and food samples as demonstrated here for sea-water and tea infusion samples. The sensitivity of the proposed method (measured by the slope of the calibration graphs) is about ten times higher than that obtained with conventional nebulization ICP-AES.Moreover ten times lower detection limits (DL=0.4 ng ml-1 at room temperature versus 5 ng ml-I) as compared with conventional nebulization can be achieved via continuous VG with NaBEt,. Using AFS measurements and similar VG D’Ulivo and Chen2’ have reported a DL of 0.2 ng ml-I for Cd and using AAS measure- ments the DL was 1 ng ml-’. As the volatile species formed probably has two alkyl groups2’ the presence of organized media such as micelles of vesicles could improve the efficiency of the generation at room temperature. The hydrocarbon layers of the micelles and could exert ‘concentration effects’ on the reagents and so a higher efficiency of alkylation (VG) could be expected. Unexpectedly the organized media tested proved not to be so beneficial for alkylation at Cd and generation of CdEt probably because the efficiency of the volatile alkylated species formed seems to be very high in aqueous media (80-100%).The temperature has a great effect on the kinetics (the peaks are narrower) and the efficiency (80% at room temperature while at 65°C it was 99Y0) of the volatile Cd compound evolved. The definitive nature of the volatile compound formed reaching the atomizer most probably CdEtz by analogy with Hg,” has yet to be positively confirmed. However a new ICP- AES method for the determination of Cd in real samples has been established based on the generation of such volatile species. Financial support from FICYT (Fundacion para el Foment0 en Asturias de la Investigacibn Cientifica Aplicada y la Tecnologia) and CICYT (Comision Interministerial de Ciencia y Tecnologia) and also the FICYT grant to M.C. Valdes- Hevia y Temprano is acknowledged. The loan of the ICP PU7000 instrument by Unicam Analytical Systems (Cam- bridge UK) is also gratefully acknowledged. References 1 Page A. L. Bingham F. L. and Chang A. C. in E’ect of Heavy Metal Pollution on Plants ed. Lepp N. W. Applied Science Publishers London 1981 1 pp. 77-109.236 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Yamagata N. and Shigematsu I. Bull. Inst. Public Health (Tokyo) 1970,19 1. Ravera O. Experientia 1984 40 2. Skogerboe R. K. Dick D. L. Pavlica D. A. and Lichte F. E. Anal. Chem. 1975 47 568. Barbaras G.D. Dillard C. Finholt A. E. Wartik T. Wilzbach K. E. and Schlesinger H. I. J. Am. Chem. SOC. 1951 73,4585. Cacho J. Beltran I. and Nerin C. J. Anal. At. Spectrom. 1989 4 661. Valdes-Hevia y Temprano M. C. Fernandez de la Campa M. R. and Sanz-Medel A J. Anal. At. Spectrom. 1993 8 847. Negishi E. Comprehensive Organometallic Chemistry Pergamon Press Oxford 1982 vol. 7 pp. 276-277. Honeycutt J. B. and Riddle J. M. J. Am. Chem. SOC. 1961,83,369. Ashby J. Clark S. and Craig P. J. J. Anal. At. Spectrom. 1988 3 735. Rapsomanikis S. Donard 0. F. X. and Weber J. H. Anal. Chem. 1986 58 35. Blais J. S. and Marshall W. D. J. Anal. At. Spectrom. 1989,4,641. Sturgeon R. E. Willie S. N. and Berman S. S. Anal. Chem. 1989 61 1867. Rapsomanikis S. and Craig P. J. Anal. Chim. Acta 1991,248 563. Bloom N. Can. J. Fish. Aquat. Sci. 1989 46 1131. AL-Rashdan A. Vela N. P. and Caruso J. A. J. Anal. At. Spectrom. 1992 7 551. 17 18 19 20 21 22 23 24 25 26 27 Fischer R. Rapsomanikis S. and Andreae M. O. Anal. Chem. 1993 65 763. Craig P. J. Mennie D. Osteah N. Donard 0. F. X. Martin F. Analyst 1992 117 823. Cai Y. Rapsomanikis S. and Andreae M. O. J. Anal. At. Spectrom. 1993 119. Cai Y. Rapsomanikis S. Andreae M. O. Anal. Chim. Acta 1993 274 243. Ashby J. R. and Craig P. J. Appl. Organomet. Chem. 1991,5 173. Ashby J. R. and Craig P. J. Sci. Total Environ. 1989 78 219. Yan D. Yan Z. Cheng G. Li A. Talanta 1982 29 519. Clark S. and Craig P. J. Mikrochim. Acta 1992 109 141. DUlivo A. and Chen Y. J. Anal. At. Spectrom. 1989 4 319. Valdes-Hevia y Temprano M. C. Aizpun Fernandez B. Fernandez de la Campa M. R. Sanz-Medel A. Anal. Chim. Acta. 1993 283 175. Aizpun Fernandez B. Valdes-Hevia y Temprano M. C. Fernandez de la Campa M. R. Sanz-Medel A. and Neil P. Talanta 1992 39 1517. Paper 3/03897A Received July 6 1993 Accepted October 5 1993