JOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY JANUARY 1994 VOL. 9 33 Determination of Arsenic Species in Fish by Directly Coupled High- performance Liquid Chromatography-Inductively Coupled Plasma Mass Spectrometry* Simon Branch Les Ebdont and Peter O'Neill Plymouth Analytical Chemistry Research Unit Department of Environmental Sciences University of Plymouth Drake Circus Plymouth UK PL4 8AA Using directly coupled high performance liquid chromatography-inductively coupled plasma mass spec- trometry non-toxic arsenobetaine was identified as the major arsenic species in cod dab haddock mackerel plaice and whiting. The fish was caught in coastal waters around Plymouth UK and purchased from local markets. Arsenic levels ranged between 1.0 mg kg-' dry mass in the mackerel to 187 mg kg-' dry mass in the plaice.Mackerel also contained dimethylarsinic acid (DMAA) and possibly a lipid bound arsenic species. Levels of the toxic inorganic species were low in all cases. No monomethylarsonic acid was recorded in any of the fish. No degradation of arsenobetaine to more toxic species was observed when an enzymic digestion procedure based on the action of trypsin was applied to the fish with the possible exception of one of the plaice samples for which DMAA was characterized in the digest at the mg kg-' level. For total arsenic determinations nitrogen addition ICP-MS was used to overcome the potential interference from 40Ar35CI +. The results obtained compared well with certified values for the dogfish reference material DORM-1. Keywords Inductively coupled plasma mass spectrometry; high-performance liquid chromatography; arsenic speciation; fish; arsenobetaine Fish and marine-based products are the major source of arsenic in the human diet.'*2 Since the turn of the ~ e n t u r y ~ .~ fish and shellfish have been known to contain relatively high levels of arsenic being in excess of 1 mg kg-'. In 1926 Chapman4 proposed that the arsenic compound in lobster was a low toxicity organic molecule. However it was not until 1977 that Edmonds et d5 identified the molecule as arsenob- etaine a compound that is now known to be widely distributed in marine organisms.G12 Although the total arsenic concen- tration in fish and shellfish from UK waters has been deter- mined,2 little attempt has been made to identify the nature of the arsenic species present.Furthermore in the many studies of arsenic species in fish few have addressed the problem of quantifying the species. The analytical approach in the majority of arsenic speciation studies involves the coupling of the separatory powers of liquid chromatography with the inherent selectivity and sensitivity of atomic spectrometry. Gas chromatography cannot be used without derivatization as a number of species such as arsenob- etaine are non-volatile. The most common couplings are high- performance liquid chromatography (HPLC) with hydride generation atomic absorption spectrometry (HGAAS),'3-15 HPLC-inductively coupled plasma atomic emission spec- trometry ( ICP-AES),16-'8 and HPLC-electrothermal atomic absorption spectrometry ( ETAAS).lS2' Although these methods are satisfactory they all suffer certain disadvantages.22 Hydride generation AAS can only be applied to a limited number of applications since some of the major arsenic species eg.arsenobetaine are non-reducible. The principal problem with ICP-AES is that it does not have the required sensitivity for environmental analysis particularly as the arsenic wave- lengths are in the low UV region. Electrothermal AAS offers the required sensitivity for the determination of arsenic and responds to all arsenic species. However an elaborate interface is usually required to introduce the sample into the cuvette. Real time analysis is impossible owing to the furnace cycle hence discontinuous chromatograms are obtained. Ebdon * Presented in part at the 1993 Winter Conference on Plasma t To whom correspondence should be addressed.Spectrochemistry Granada Spain January 10-15 1993. et have reviewed coupled techniques for the speciation of arsenic. Recently the coupling of HPLC with inductively coupled plasma mass spectrometry (ICP-MS) has received interest in speciation No complicated interfacing is required HPLC flow and ICP-MS uptake rates are compatible and for arsenic high sensitivity to all species is attainable. Continuous chromatograms are also produced. The application of HPLC-ICP-MS to the determination of arsenic species in marine fish caught from the waters around Plymouth Devon UK is described in this paper. Experimental Materials and Methods Duplicate samples of fish were purchased from Plymouth Fish Market.The vendor prepared the fish as for a conventional customer. In the laboratory each fish was cut into small pieces and freeze dried (Edwards Super Modulyo Edwards High Vacuum Crawley Sussex UK). The dried fish was then ground in a pestle and mortar until a fine powder was produced and the powder was stored in a desiccator until required for analysis. Total arsenic was determined by accurately weighing 0.1-0.2 g of the fish sample into a poly(tetrafluoroethy1ene) digestion bomb (Savillex Corporation Minnetonka MN USA) and adding 2 ml of nitric acid (Aristar grade Merck Poole Dorset UK). The samples were left to pre-digest overnight and were then digested in a microwave oven. The digests were diluted to 10 or 100 ml with doubly de-ionized water (Millipore Bedford MA USA) which was used throughout.The samples and appropriate matrix matched standards were spiked with a solution (100 pg 1-') of indium (Aldrich AAS standard Aldrich Milwaukee USA) and ana- lysed by ICP-MS with addition of nitrogen to the injector gas (N,-ICP-MS) using the operating conditions shown in Table 1. The technique of N,-ICP-MS has recently been shown to overcome the interference of chloride on arsenic determinations by ICP-MS.29 A dogfish reference material (DORM-1 Canadian Research Council Ottawa Canada) was simul- taneously analysed to maintain quality control. Arsenic species34 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JANUARY 1994 VOL. 9 Table 1 Instrumental parameters for the HPLC-ICP-MS coupling HPLC parameters - Mobile phase Flow rate Injection volume Run time ICP-MS parameters - Nebulizer gas flow rate Coolant gas flow rate Intermediate gas flow rate Forward power Reflected power Data acquisition mode survey scan Equilibration on 1 mmol 1-' K,SO pH 10.5.Switched on injection or after 3 mins to 50 mmol 1-' K2S04 pH 10.5 1.5 ml min-' 175 pl 671 s 0.85 1 min-' 14 1 min-' 0.8 1 min-' 1.5 kW <low Local mass set to 75 u in the fish were determined following two different extraction procedures. Method 1 was a variation on the method pre- viously described by Beauchemin et and Method 2 was based on the principle of enzymic digestion reported by Crews et aL31 Method 1 A 1.0 g-portion of the dry sample was accurately weighed into a glass beaker and 20 ml of methanol and 10 ml of chloroform were added (both Aristar grade Merck).After covering the sample was sonicated for 1 h. The solution was transferred into a centrifuge tube and centrifuged for 10min at 2500 rev min-'. The supernatant was transferred into a 100 ml separating funnel and the process was repeated on the remain- ing pellet. Chloroform (20 ml) and water (20 ml) were then added to the combined supernatants in the separating funnel which was shaken vigorously. The chloroform layer was run off and the water-methanol layer was filtered (Whatman 541 paper Whatman Kent UK) into a 100ml round-bottom flask The solvent was rotary evaporated and the residue was re-dissolved in 10 ml of water. The samples were further filtered using a 0.45 pm pore size filter (Millipore) into plastic storage bottles prior to analysis by HPLC-ICP-MS.The samples were refrigerated at 4°C if there was a delay in the final analysis. Recovery experiments and determinations of arsenic in the solid residue and the chloroform layer were also performed on representative samples. Method 2 A 1.Og portion of sample was accurately weighted into a plastic centrifuge tube along with 100mg of trypsin (Sigma Dorset UK). Ammonium carbonate solution (20 ml 0.1 mol l-' AnalaR Merck) was added the tube was sealed and placed in a shaking water-bath for 4 h at 37 "C. The sample was then ultracentrifuged at 11 000 rev min- for 20 min (MSE Highspeed 18 MSE Scientific Instruments Crawley Sussex UK). As with Method 1 the samples were filtered and refriger- ated prior to analysis by HPLC-ICP-MS.Recovery tests were performed and representative residues analysed for total arsenic concentration. Arsenic species were determined in the extracts using HPLC-ICP-MS. The HPLC separation has previously been describedz7 and involves separation of five species on a column (12.5 x0.4cm id.) packed with a strong anion- exchange resin (7-10 pm SAX Benson Reno NV USA). The column was immersed in a 60°C water-bath to improve peak shape. The HPLC system was interfaced to the ICP-MS by connecting the column outlet to the ICP-MS nebulizer by the shortest possible length of capillary tubing. Dilute solutions of alkaline potassium sulfate (Aristar Merck) were used as the mobile phase. Arsenic standards were freshly prepared from 1000 mg 1-1 As stock solutions of sodium arsenite (AnalaR) disodium hydrogen arsenate ( AnalaR Merck) dimethylarsinic acid (DMAA) (Sigma) monomethylarsonic acid (MMAA) (donated as disodium methanearsonate by Dr.A. Howard Southampton University Southampton UK) and arsenobetaine (AB) (donated by Professor K. J. Irgolic Texas A and M University TX USA). Instrumentation The ICP-MS measurements were performed using a VG Plasmaquad 2 ICP mass spectrometer (VG Elemental Winsford Cheshire UK) fitted with a high solids nebulizer (Ebdon nebulizer PSA Sevenoaks Kent UK). This arrange- ment overcame slight problems of salt deposition seen around the orifice of a glass concentric nebulizer supplied with the instrument. Chromotographic separations utilized a Waters 6000A pump (Waters Milford MA USA) with a Rheodyne 7125 (Rheodyne Cotati CA USA) injection valve used for sample introduction. Results and Discussion The moisture content of the various fish are shown in Table 2.In most cases the results for two separate fish are presented and these are represented by (1) and (2) in the table. All the fish with the exception of the mackerel contained greater than 70% water. The lower result for the mackerel may be related to the oily nature of the fish. Arsenic species in the various fish obtained using the two extraction procedures are shown in detail in Tables 3 and 4. No arsenite or MMAA were identified in any of the samples. The total arsenic concentration in pairs of fish of the same species were similar to each other the one exception being lemon sole. It is interesting that the two cod specimens one coming from the Atlantic the other from the local waters around Plymouth had markedly different arsenic concen- trations.Other workers6.' have also noted that the arsenic content of the same species of fish varies with the location of the catch. This is tentative evidence to support the theory of Edmonds and France~coni~~ that arsenobetaine does not arise within higher marine animals endogenously but is derived from microbial action upon algae within sediments releasing arsenobetaine to the water column. Therefore fish from differ- ent areas may contain differing concentrations of arsenobetaine as the products of microbial action will vary from site to site. The flatfish surveyed in the study i.e. dab plaice and sole were found to contain the highest levels of arsenic.All of these fish are carnivorous bottom feeders existing on diets of mol- luscs and bivalves which filter feed on the sea bed. Thus they are more likely to ingest arsenobetaine before it is dissipated in the body of the ocean. It is also possible that the variation Table 2 Moisture content of fish in % m/m Fish Atlantic Cod 'Local' cod Dab Haddock Lemon Sole Mackerel Plaice Whiting Moisture content 86 86 74 73 77 76 77 60 63 81 74 70 75JOURNAL OF ANALYTICAL ATOMIC SPECTKOMETRY JANUARY 1994 VOL. 9 35 Table 3 Arsenic species extracted by Method 1 (methanoi-chloroform extraction). All results in mg kg ~ ' of dry mass Sample Atlantic cod 'Local' cod Dab (1) Dab ( 2 ) Haddock Lemon Sole ( 1 ) Lemon Sole ( 2 ) Mackerel (1 1 Mackerel ( 2 ) Plaice (1) Plaice (2) Whiting ( 1 ) Whiting (2) DORM-] * Total arsenic 17.8 k0.7 13.7 F 0.7 27.1 & 0.7 28.3 5 0.4 30.7 & 0.2 62.2 & 2.8 172.952.6 149.65 15.1 4.1 50.5 3.5 k 0.3 196.1 5 3.5 183.1 5 1.6 15.9 0.5 16.4 k 0.2 AB 15.1 k0.6 13.2 rfr 0.6 3 1.2 & 0.8 26.9 k 1.9 27.5 5 1.8 59.3 k 2.1 145.6 rt 12.8 102.6f 19.4 1.1 k0.3 1.OkO.1 187.3 k 20.3 102.9 k 13.6 10.45 1.8 9.8 -+_ 0.7 DMAA N P t < 0.3 < 0.3 < 0.3 < 0.3 < 0.5 < 0.5 < 0.5 0.550.1 0.3 f 0.i < 0.5 < 0.5 < 0.3 < 0.3 Arsenate N P < 0.5 < 0.5 < 0.5 < 0.5 < 1.0 < 0.5 < 0.5 < 0.5 1.1 k0.S < 1.0 < 1.0 < 0.5 < 0.5 Total As in residue N P 0.6 f 0.01 1.3 50.01 N P NP 2.5 5 0.04 N P N P 0.5 F 0.03 0.5 * 0.02 NP NP NP N P Total As in methanol water phase NP 12.0 k0.1 22.5 & 0.8 NP NP 53.4f 1.1 N P NP 1.5 f 0.8 2.0 k 0.2 NP NP NP NP Total As in chloroform phase NP 0.09 & 0.0 1 0.3 5 0.03 NP NP < 0.05 NP NP 0.08 * 0.0 1 : 0.08 f 0.01 : N P N P NP NP * Certificate values for DORM-1 AB = 15.7 k 0.8 mg kg-'; total arsenic = 17.7 k 2.1 mg kg -' .i.NP = Not performed. : These results should be regarded as 'water-leachable' arsenic. Table 4 Arsenic species extracted by Method 2 (trypsin digestion). All results in mg kg-.' of As dry mass Sample Atlantic cod Local cod Dab (1) Dab ( 2 ) Haddock Lemon Sole ( 1 ) Lemon Sole (2) Mackerel ( 1 ) Mackerel ( 2 ) Plaice ( 1 ) Plaice ( 2 ) Whiting ( 1) Whiting ( 2 ) DORM- 1 * Total arsenic 17.8 k 0.7 13.7 k 0.7 27.1 f 0.7 28.3 & 0.4 30.7 * 0.2 61.2 k 2.8 172.9 2.6 149.6 k 15.1 4.1 k0.5 3.5 -t- 0.3 196.1 53.5 183.1 5 1.6 15.9 k 0.5 16.4 k 0.2 AB 16.1 k0.4 13.5 k0.9 26.9 & 1.2 40.0 5.7 41.4 k0.4 50.0 k 2.1 183.2 k 14.1 77.0 k0.2 0.8kO.l 0.2 k 0.01 106.0 k 19.0 77.0 k 5.0 12.lf0.2 15.8 f 1.3 DMAA NPt < 0.5 t 0 .5 ~ 0 . 5 < 0.5 < 0.5 < 0.5 < 0.5 0.4 k 0.1 0.4 k 0.1 <0.5 21.6 k 2.3 < 0.5 < 0.5 Arsenate NP < 1.0 1.0k0.4 2.4 k 0.3 < 1.0 < 1.0 < 1.0 < 1.0 0250.1 0.3 k 0.1 < 1.0 < 1.0 I .3 k0.2 < 1.0 Total As in residue NP 2.3k0.1 3.7 -t 0. I NP NP 15.7 k 0.2 NP NP 2.8 & 0.2 1.4 & 0.2 NP NP NP NP Total As in extract NP 13.2 * 0.6 29.6 k 2.0 NP NP 8.0 i 3.0 N P NP 2. I * 0.2 2.3 &- 0.1 NP NP NP NP * Certificate values for DORM-I AB= 15.7+0.8 mg kg-'; total arsenic= 17.7F2.1 mg kg-' t NP = Not performed.in arsenic concentrations might be related to other factors such as age or health status. The concentrations of arsenic in some of the fish were relatively high e.g. in plaice over 180 mg kg-' dry mass. These concentrations are not unprecedented. Luten et found the arsenic concentration in plaice from various regions of the North Sea to be in the range 3-166 mg kg-l. These values are higher than reported in the UK Ministry of Agriculture Fisheries and Food (MAFF) arsenic surveillance paper,2 which listed results in the range 1-20 mg kg-' fresh mass. The results for cod dab and haddock in this study also compare well with those detailed in the MAFF report. The mackerel results in this study 3.5-3.4mg kg-' dry mass compared well with results of surveys of fish from Scottish waters,2 the results for which lay in the range 0.2-1.6 mg kg-' fresh mass.The results of this study for cod haddock mackerel and sole are similar to results for the same fish caught in the Atlantic6 Luten et reported arsenic concentrations in lemon sole of 24.9-31.4 mg kg-' fresh mass. which compared well with the values from this study of 150-173 mg kg-' dry mass. Using both extraction methods arsenobetaine was found to be the major species present in all of the fish. In a number of specimens eg. the cod samples. arsenobetaine comprised approximately 100% of the arsenic. In a few cases the total arsenic remaining in the residue was determined by N2- ICP-MS. The results confirmed that the extraction procedures had recovered the majority of the arsenic.Total arsenic was also determined in the extract solutions and found to offer only moderate agreement with the sum of the species. There was some evidence to suggest that this was due to the methanol in the extract causing perturbation in the ICP-MS. Attempts to matrix-match samples and standards proved fruitless. Similarly arsenic was determined in the chloroform phase of selected specimens and low results were obtained. The results for the mackerel specimens were of interest. Following extraction by Method 1 (Table4) the sum of the individual species the arsenic in the extracted residue and the arsenic in the chloroform phase was less than the total arsenic determined in the original fish.This was almost certainly because the mackerel an oily fish contains a significant proportion of arsenic bound to lipids. This arsenic would be extracted into the chloroform layer. The chloroform was evaporated to leave a viscous orange phase. The solid was insoluble in water. After shaking the water was itself analysed by ICP-MS to give the figure in the 'total arsenic in CHCI phase' column in Table 4 which in this case would be more accurately described as 'water-leachable' arsenic. Attempts to digest the solid with nitric acid proved unsuccessful owing to sample charring. If a proportion of the arsenic in mackerel were lipid bound it might explain why low recoveries were recorded for mackerel using Method 2. Trypsin a protease would be unable to break down such a compound.As the sum of the individual species was less than the total arsenic in the extract e.g. 1.4 and 2.1 mg kg-' respectively for mackerel ( l ) this might indicate that some of the lipid- bound arsenic was extracted but remained bound to the36 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JANUARY 1994 VOL. 9 column. During the analyses the column had to be periodically reversed and flushed in order to overcome pressure build-up due to the accumulation of material on the column. The discrepancy between the sum of the arsenic species and the total arsenic determined by ICP-MS for lemon sole plaice and whiting may be indicative of the fish containing arsenic in a lipid-bound form. In planned future studies the determination of arsenic in all residues and chloroform layers will assist in elucidating the nature of the speciation. Using extraction Method ( l ) DMAA was identified only in the mackerel and at toxicologically insignificant concen- trations.The level of arsenate in mackerel (2) is of more significance although it would still be less than 1 mg kg-' in the fresh fish. The results for the trypsin extractions produced a number of peculiarities. It was predicted prior to the study that trypsin would yield greater recoveries than the methanol-chloroform approach as the protease would disrupt the lipid-protein membrane and release the cell contents. This hypothesis was supported by the results for DORM-1 and the whiting which gave higher results for Method 2. However both plaice both mackerel and one of the lemon sole samples gave lower recoveries in the extracts.This may have been related to the composition of the fish tissues. The haddock results indicate only about 80% of the arsenic was extracted from the fish tissue the predominant species again being arsenobetaine. The results for the dab with extraction Method 2 were unsatisfactory. Arsenobetaine recoveries were in excess of 130% of the total arsenic figure. This was a reproducible phenomenon observed in both fish. Why this occurred is unclear. The dab was unique in that it was the only fish in which an unidentified species appeared on the chromatogram (see Fig. 1). Co-injection showed it not to be any of the five species in the study. The species was not identified in the extracts obtained using Method 2 and it could be that the species co-eluted with arsenobetaine giving an erroneously high result.Further work to identify this arsenic compound is planned. A high concentration of DMAA was recorded in one of the plaice. This was possibly owing to partial enzymic degradation of the arsenobetaine by trypsin. The sum of the two species correlates well with the speciation in the same fish using methanol-chloroform extraction. The results of the determinations in no case yielded inorganic arsenic results with implications for human health. The highest arsenate level recorded was 2.4 mg kg-' but as these results 7450 3725 c I v) v) c r O ; 1000 1. C c)) m .- 500 0 I I I Ti rne/s 335 67 1 Fig. 1 (a) HPLC-ICP-MS chromatogram of a mixed arsenic stan- dard 1 arsenobetaine; 2 DMAA; 3 arsenite; 4 MMAA; and 5 arsen- ate.(b) HPLC-ICP-MS chromatogram of a methanol-chloroform extract from Dab (1) showing the presence of unidentified species 1 arsenobetaine; 2 unknown species; and 3 arsenate were for dry mass the exposure would be less than 1 mg kg-' in the fresh fish. Quality control experiments were performed throughout the study. Experimental results for DORM-1 were in good agree- ment with certified values for total arsenic and arsenobetaine concentrations. The slightly higher value for arsenobetaine by Method 2 in comparison to the certificate value probably arose as the certificate value is quoted for arsenobetaine extracted by Method 1. The recoveries of individual species taken through extraction Method 1 using plaice as the sample matrix varied from 80% for arsenobetaine to 108% for arsenate.The recovery of spikes taken through extraction Method 2 and using whiting as the sample matrix varied between 85% for arsenobetaine to 115% for arsenate. Determination of arsenic in plaice (1) using standard additions gave a recovery of 190 mg kg-' dry mass or 97%. For all of the samples peak identity was confirmed by co-injection of individual standards. Conclusions The technique of directly coupled HPLC-ICP-MS has been shown to be useful in the investigation of arsenic species in fish. In the fish studied the predominant arsenic compound present was non-toxic arsenobetaine. The concentrations of the toxic inorganic species arsenite and arsenate were neglible or below the limits of detection in all of the fish analysed.Since these are amongst the most commonly eaten fish these are significant results when assessing dietary risk. We gratefully acknowledge the financial support of the UK Ministry of Agriculture Fisheries and Food for this research. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 References Penrose W. R. CRC Crit. Rev. Environ. Control 1974 4 465. Survey of Arsenic in Food Food Surveillance Paper Number 8 Her Majesty's Stationery Office London 1982. Cox H. E. Analyst 1925 50 3. Chapman A. Analyst 1926 51 548. Edmonds J. S. Francesconi K. A. Cannon J. R. Raston C . L. Skelton B. W. and White A. H. Tetrahedron Lett. 1977,18 1543. Lawrence J. F. Michalik P. Tam G. and Conacher B. S. J. Agric. Food Chem. 1986 34 315. Francesconi K.A. Micks P. Stockton R. A. and Irgolic K. J. Chemosphere 1985 14 1443. Edmonds J. S. and Francesconi K. A. Chemosphere 1981 10 1041. Luten J. B. Riekwel-Booy G. and Rauchbaar A. Environ. Health Perspect. 1982 45 165. Hanoaka K. and Tagawa S. Bull. Jap. SOC. Sci. Fish. 1985 51 681. Morita M. and Shibata Y. Anal. Sci. 1987 3 575. Hanaoka K. Yamamoto H. Kawashima K. Tagawa S. and Kaise T. Appl. Organomet. Chem. 1988 2 371. Ricci G. R. Shepard L. S. Colovos G. and Hester N. E. Anal. Chem. 1981 53 610. Tye C. T. Haswell S. J. O'Neill P. and Bancroft K. C. C. Anal. Chim. Acta 1985 169 195. Chana B. S. and Smith N. J. Anal. Chim. Acta 1987 197 177. Morita M. Uehiro T. and Fuwa K. Anal. Chem. 1981,53 1806. LaFreniere K. E. Fassel V. A. and Eckels D. E. Anal. Chem. 1987 59 879.Low G. K.-C. Batley G. E. and Buchanan S . J. Chromatographia 1986 22 292. Brinckman F. E. Blair W. R. Jewett K. L. and Iverson W. P. J. Chromatogr. Sci. 1977 15 493. Brinkman F. E. Jewett K. L. Iverson W. P. Irgolic K. J. Ehrhardt K. C. and Stockton R. A. J. Chrornatogr. 1980,191,31. Haswell S . J. Stockton R. A. Bancroft K. C. C. O'Neill P. Rahman A. and Irgolic K. J. J. Autom. Chem. 1987 9 6. Ebdon L. Hill S. Walton A. P. and Ward R. W. Analyst 1988 113 1159.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JANUARY 1994 VOL. 9 37 23 Dean J. R. Munro S. Ebdon L. Crews H. M. and Massey R. J. Anal. At. Spectrom. 1987 2 607. 24 Thompson J. J. and Houk R. S. Anal. Chem. 1986 58 2541. 25 Shibata Y. and Morita M. Anal. Sci. 1989 5 107. 26 Beauchemin D. Siu K. W. M. McLaren J. W. and Berman S. S. J. Anal. At. Spectrom. 1989 4 285. 27 Branch S. Bancroft K. C. C. Ebdon L. and O’Neill P. Anal. Proc. 1989 26 73. 28 Hansen S. H. Larsen E. H. Pritzl G. and Cornett C. J. Anal. At. Spectrom. 1992 7 629. 29 Branch S. Ebdon L. Ford M. Foulkes M. E. and O’Neill P. J. Anal. At. Spectrom. 1991 6 151. 30 Beauchemin D. Bednas M. E. Berman S. S. McLaren J. W. Siu K. W. M. and Sturgeon R. E. Anal. Chem. 1988 60 2209. 31 Crews H. M. Burrell J. A. and McWeeney D. J. Z. Lebensm. Unters Forsch. 1985 180 221. 32 Edmonds J. S. and Francesconi K. A. Appl. Organomet. Chem. 1988 2 297. 33 Luten J. B. Riekwel-Booy G. Greef J. v. d. and ten Noever de Brauw M. C. Chemosphere 1983 12 131. Paper 3102642 F Received May 10 1993 Accepted September 6 1993