JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 297 Quality Control of a Recently Developed Analytical Method for the Simultaneous Determination of Methylmercury and Inorganic Mercury in Environmental and Biological Samples* Hgkan Emteborg Negassi Hadgu and Douglas C. Baxter Department of Analytical Chemistry University of Umea S-90 7 8 7 Umea Sweden A recently developed method for the simultaneous determination of methylmercury and inorganic mercury based on extraction butylation capillary gas chromatographic separation and atomic emission detection has been evaluated with respect to analytical quality. A number of reference and candidate reference materials have been analysed within international intercalibration exercises showing good agreement with respect to methylmercury and total mercury.Other small scale laboratory intercomparisons have also been made in order to assess the analytical performance. The materials analysed have a wide range of mercury concentrations from low ng g-' to pg g -'. Finally important characteristics such as artifact formation detector selectivity chromatographic performance and stability of mercury compounds during acid leaching extract ion and der ivatization are discussed. Keywords Mercury speciation; capillary gas chromatography microwave-induced plasma atomic emission spectrometry; environmental fish and biological samples; reference materials Trace metals and persistent organometallic species pose a greater risk to man and the environment than the combined total toxicity of radioactive and organic wastes each year through their accumulation in the food chain.' It has also been recognized for a considerable period of time that organometal- lic compounds are generally much more toxic than their inorganic counterparts owing to biological compatibility where these species penetrate vital barriers such as the blood- brain barrier and cell membranes.This has led to a desire to distinguish between the different forms of specific elements an analytical discipline that is termed speciation and some methods have been developed for the determination of different mercury compounds. The classical methods for mercury speciation are the gas- chromatographic procedure proposed by Westoo' for the determination of methylmercury chloride (MeHgCl) with elec- Iron-capture detection of the halide moiety (GC-ECD) and the method from Magos3 based on a selective reduction of inorganic mercury and total mercury yielding cold mercury vapour (CV) which is detected with atomic absorption spec- trometry3 (AAS). Selective reduction is achieved with SnC1 and SnC1 + CdC1 for inorganic and total mercury respect- ively.The difference between total mercury and inorganic mercury corresponds to the content of methylmercury in the sample. Both methods have been improved by several workers but some drawbacks still persist. These limitations are dis- cussed further below. Newer variants that are now in use include gas chromatography cold vapour atomic fluorescence spectrometry (GC-CVAFS) following ethylation of both Hg2+ and MeHg' with sodium tetraethylb~rate~ and a method analogous to that of Magos involving selective reduction with BrCl+SnCl where bromine monochloride is used as an oxidizing agent for methylmercury prior to addition of the reducing agent SnCl giving results for total mercury.Addition of SnCl alone will yield results for inorganic mercury after which the reduced mercury is amalgamated and thermally desorbed and detected by AFS4 or AAS.' The method discussed here is based on butylation of both Hg2+ and MeHg' with a Grignard reagent after extraction into toluene as diethyldithiocarbamate complexes.6 The butylated species are separated on an open tubular gas- chromatographic column then atomized and excited in an atmospheric pressure microwave induced helium plasma and * Presented at the XXVIII Colloquium Spectroscopicurn Inter- nationale (CSI) York UK June 29-July 4 1993.detected using atomic emission spectrometry at 253.7 nm In the above mentioned methods it is important to dis- tinguish between species-specific and operationally defined methods. The first category is represented by the GC-MIP- AES based procedure and the latter by the method of M a g o ~ . ~ A species-specific method gives a positive identification of a species e.g MeHg through high separation efficiency and high detector selectivity whereas the operationally defined method gives indirect answers since the response in the case of the method proposed by Magos is dependant on which reagent is added to the sample. All operationally defined methods should be compared with species-specific methods as otherwise the analyst is 'blind' in terms of speciation.In order to ensure results of high analytical quality from either category it is essential that certified reference materials (CRMs) or reference materials (RMs) are used. Until recently it was impossible to purchase RMs for methyl- mercury' and there is a need for more materials covering a broad range of concentrations to assess an analytical method in terms of ruggedness accuracy and applicability to various types of biological and environmental samples. It is however impossible to produce a CRM for mercury speciation for every conceivable sample matrix and thus when there are no CRMs or RMs available it is the responsibility of the analyst or laboratory to verify accuracy by means of small scale labora- tory intercomparisons on the matrix of interest or by other measures to assess the analytical quality of the results.It is also important to report on the accuracy of new methods in this field which offer ways to circumvent difficulties by provid- ing either improved chromatography or detector selectivity. In the work described here attempts have been made to verify the analytical quality of mercury speciation results obtained using the GC-MIP-AES method outlined re~ently.~.*>~ Results for speciation of mercury in two Northern pikes (Esox lucius) are also included to demonstrate the applicability of this method to fresh samples. Total mercury in pike samples was determined by CVAAS as described el~ewhere.~ (GC-MIP-AES). Experimental Instrumentation The instrumentation has been discussed in detail pre- v i o ~ s l y ~ ~ * ~ ~ slight alterations being summarized in Table 1.The operating conditions for GC can be found in Table 2.298 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Table 1 Components of the GC-MIP-AES system Chromatograph Varian 3300 (Solna Sweden) with 15 m fused silica non-polar bonded phase wide-bore capillary column (0.53 mm i.d. 1.5 pm DB-1 J & W Scientific Rancho Cordova CA USA); equipped with a pneumatic automatically controlled four-way valve from Valco TX USA and a Varian 1093 SPI injector with on-column insert Beenakker TM, configuration; torch assembly AHF-Ingenieurburo H. Feuerbacher Tubingen Germany. Details given in refs. 6 and 8 AHF-Ingenieurburo operated with 150 W forward power tuned to give reflected power of <0.5 W polychromator part of 'organic analyser' system MPD 850 AG (Luton Bedfordshire UK).Varian Star integration software for PC use. 20 Hz sampling rate and 1 V max output from photomultiplier tube amplifier read-out Cavity Generator GMW 24-1000 DRL microwave generator Spectrometer 0.75 m Rowland circle direct reading Integrator Table 2 Operating conditions for the gas chromatograph; carrier gas helium flow rate 18 cm3 min-' Column oven Initial column temperature/"C Initial hold time/min Final column temperature/"C Ramp rate/"C min-l Isothermal hold time/min Injector temperature/"C Initial relay Switch delay time/min Final relay Injector Relay 50 1 180 40 1 180 - 1* 2.1 + I t * - 1 =Column effluent vented.t + 1 =Column effluent directed to plasma. Sample Materials Dogfish liver tissue (DOLT-2) and lobster hepatopancreas tissue (TORT-2 presently not certified) were supplied in the form of lyophilized homogeneous powders by the National Research Council of Canada (NRCC Ottawa Canada). Two similarly prepared tuna fish tissue samples (BCR-463 and BCR-464 certification report in preparation) were obtained from the Community Bureau of Reference (BCR Brussels Belgium)." A human whole blood sample (Seronorm 904 Nycomed A/S Oslo Norway) which had previously been analysed,' is also included in the ensuing discussion. Two Northern pikes (wet masses 0.8 and 1.3 kg) were caught in Rickleiin (a stream 50 km north of Umei) and Tavelsjon (a lake 30 km west of Umei).The fish were stored frozen and thawed before analysis. Reference material IAEA MA B3/TM was obtained from the International Atomic Energy Agency (Vienna Austria) and is certified for total mercury only. Sample Preparation The sample preparation procedure has been modified slightly and hence requires description. Samples other than blood and the tissue from pike were dried in a vacuum dessicator at 100mm Hg for 48 h over anhydrous Mg(C104) (Merck Darmstadt Germany)," to determine the moisture content. Separate samples 50-100 mg were accurately weighed into 10 ml glass centrifuge tubes used as supplied. Although sample masses of 200-500mg are generally recommended to avoid problems with inhomogeneity recent results suggest that many RMs exhibit fairly homogeneous trace element distributions even at the milligram mass level.' Thus samples of 50-100 mg were considered to be sufficiently representative of the bulk material as demonstrated by the results given below.Larger amounts of sample would have led to higher reagent blanks and the need for numerous extractions and such practical considerations also dictated the sample masses used. The lyophilized and powdery samples were wetted with 0.5-1.0 ml of saturated NaCl solution (Merck Darmstadt Germany pro analysi quality). Acid leaching to liberate the mercury species was achieved with an addition of 0.1-0.2ml of concentrated sub-boiling distilled hydrochloric acid (orig- inally pro analysi quality from Merck) and shaking for 15 rnin on an automatic shaker (IKA-VIBRAX-VXR Labassco Partille Sweden).Then 2.1-2.2 ml of 1 mol I-' NaOH solution (Eka Nobel Bohus Sweden) was added to neutralize the solution. It had previously been found that the greatest contri- bution to the reagent blank came from the NaOH solution and so purification procedures as reported in ref. 9 were applied. Next 1.5 ml of a pH 9 borate buffer (Merck) were added followed by 1.0 ml of 0.5 moll-' sodium diethyldithio- carbamate (DDTC) solution supplied by Aldrich (Milwaukee WI USA). Subsequently the samples were shaken for 5 min and 1.0 ml of 'distilled in glass quality' toluene (Burdick and Jackson Muskegon MI USA) was added. The sample was shaken for 5 rnin and centrifuged for 5 rnin at 3200g (Wifug centrifuge Bradford UK) after which 0.8 ml of the toluene phase was withdrawn with a Gilson pipette (Villiers-le Bel France) and transferred into another centrifuge tube standing in an ice-water bath.Another 1 ml of toluene was added to the sample and the above procedure was repeated except that 1.0ml was withdrawn the second time. To the combined toluene phases (0.8 + 1.0 ml) 0.3 ml of 2.0 mol dmP3 of n-butylmagnesium chloride in tetrahydrofuran (Aldrich) was added. The reaction was completed within 5 rnin and the excess of Grignard reagent was quenched with 0.5 ml of 0.6 moll-' HC1. Finally the samples were centrifuged and the organic layer containing the butylated mercury species was transferred to 2 ml screw capped glass vials. Sample preparation of the pike samples was performed with an alkaline digestion procedure similar to that described by Cappon and Smith.I3 A sample of 0.5-5.Og of tissue was weighed accurately into a 50ml polycarbonate tube and 10-25 ml of 15% m/v of NaOH (E Nobel) in 0.2% m/v NaCl (Merck) solution were added with agitation.The tube was loosely stoppered and warmed at 50-60°C in a water-bath until the fish tissue was completely dissolved (10-25 min). After cooling 2 ml of the alkaline digest were placed in a 10 ml screw-capped glass centrifuge tube and neutralized with 5-20 drops of concentrated sub-boiling distilled HC1. Next the sample was shaken for 5 rnin on an automatic shaker followed by addition of 1 ml of pH 9 borate buffer (Merck) and 1 ml of 0.5 moll NaDDTC (Aldrich). Extraction and derivatization were then performed as described above.Details of the blood sample preparation procedure can be found in ref. 8. Solutions All solutions were prepared using Milli-Q quality water (Millipore Milford MA USA). Aqueous standards of methyl- mercury chloride (MeHgCl) were diluted from a 158 mg I-' stock solution prepared by dissolving an appropriate amount of the salt (Merck >98% pure) in Milli-Q water. This stock solution was calibratedg against a commercial 1000 mg 1-' Hg standard (as HgC1,) supplied by Analytical Standards AB (Kungsbacka Sweden). Inorganic mercury standards were diluted from the 1000mgl-' standard. The content of inor- ganic mercury was quantified in the MeHgCl standards at a level of 2% using the GC-MIP-AES procedure.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 299 Calibration For the purposes of calibration the standard additions method was employed to account for non-quantitative extraction and derivatization of the mercury species. Additions at two concen- tration levels were made to each type of sample material on at least two different occasions prior to the acid leaching step allowing 15 min with shaking for eq~i1ibration.l~ The added concentrations were selected to increase the total levels of mercury by factors of two and three following preliminary screening analyses to determine the approximate concen- trations in the samples. The additions were found to give a linear response with correlation coefficients always above 0.995. Calibration of the pike samples was achieved with external aqueous Hg2+ standards which were extracted and derivatized as described above.Results reported in Table 7 are corrected for recovery. Chromatograms were evaluated using the Varian Star software (Table 1 ) employing peak-area measurements. Reported concentrations in Table 4 have been corrected for determined moisture levels in the lyophilized materials. Results Recoveries of added methylmercury and inorganic mercury from various sample matrices are reported in Table 3. The recoveries were assessed relative to aqueous standard solutions that were processed in the same fashion as the samples i.e by complexometric extraction and butylation. It should be noted that at least for simple aqueous standards essentially complete recovery of mercury species is achieved following a double extraction into tol~ene,'~ and that the derivatization reaction appears to be quantitative." Results for the determination of methylmercury and inor- ganic mercury in several candidate RMs and RMs using the present GC-MIP-AES method are summarized in Table 4 and some representative chromatograms are displayed in Figs.1-4. Total mercury concentrations determined were obtained by summing the results for the two individual species and compared with certified values. Extraction efficiencies of mer- cury from pike tissue are reported for different sample work- up procedures in Table 5. It should be noted that materials other than lyophilized finely dispersed and homogenous samples should be digested with the alkaline digestion pro- cedure prior to extraction and derivatization.To demonstrate the possibility of using acid leaching for sample work-up in fine ground lyophilized materials results are reported in Table6 which indicate that there are no differences between alkaline digestion or acid leaching sample work-up for these ma trices. In Table 7 results for speciation of mercury in Northern pike (Esox lucius) can be found. Note that pikes with a mass of approximately 1 kg are routinely used to monitor mercury levels in aquatic ecosystems. Lakes where the content of total mercury is higher than 1 mg kg-I in pike are deemed unsuit- able for commercial fishing and some 10000 lakes are presently 'blacklisted' in Sweden. Discussion Recoveries In scrutinizing the data given in Tables 3 and 4 it can be seen that better recoveries are generally obtained at lower concen- trations of mercury species in the samples in agreement with previous work.14 Significant differences in recoveries between matrices is also apparent in Table3 even in the case of the two BCR tuna fish samples an effect which was observed by several of the laboratories participating in the certification campaign.1° Unless the extraction efficiency for the work-up procedure is established standard additions to the matrix of interest are still an absolute requirement for calibration pur- Table 3 Recoveries for methylmercury and inorganic mercury values given % & one standard deviation for n = 5 Addition Sample BCR-463 BCR-464 TORT-2 DOLT-2 MeHg 72.5 f 4.1 84.9 f 2.5 100.3 f 12.3 72.1 f 10.8 Hgz+ 76.3 2.2 69.7 f 3.0 90.4 & 6.6 73.3 & 9.0 Table 4 Results for the speciation of mercury in various environmen- tal and biological materials by the GC-MIP-AES method uncertain- ties given as & one standard deviation Material n Certificate value (Certificate value)? (Certificate value)? (Certificate value)? DOLT-2 8 TORT-2 7 BCR-463 5 BCR-464 5 SE-904 5 SE-9047 2 Reference value MeHq Hg2+/1 g- Pg g- 0.740 & 0.03 1 0.693 f0.053 1.1 58 f 0.085 - 0.154+0.009 0.086 fO.010 - -t 3.25 k 0.169 3.04 f 0.244 - (11 labs.) 5.75 f 0.374 5.46 f 0.339 - (12 labs) 3.6k0.4 ng g-' 3.4 ng g-' 0.280 f 0.020 0.403 f 0.003 1.8 k0.5 ng g-' 1.0 ng g-' - - Hgto( P!2 g- 1.90 f 0.09* 1.99 4 0.10 0.240f0.013* -$ 3.24 & 0.16* 2.88 & 0.17 (7 labs.) 5.75 40.37* 5.26k0.21 (7 labs) 5.4f0.6* ng g-' 4.0k0.49 ng g-' *Sum of inorganic and methylmercury.t(Certificate value) = certification report in preparation. $=Data from GC-MIP-AES method will be included in the 9= Expressed as pg g- of MeHg in BCR-463 and BCR-464. 7SE-904 = Data from Brunmark et ~ 1 . ' ~ certification; certified values will be submitted as soon as possible. T 1 50 mV J 1 1 1 1 1 0 1 2 3 4 5 Time/m i n Fig. 1 Chromatogram for 2 pl injection of BCR-464 (49.9 mg) A methylbutylmercury (2.6 min); and B dibutylmercury (4.1 min). Concentrations of different species are presented in Table 4300 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 t a t 0 a v) U 0 1 2 3 4 5 Time/m in Fig. 2 Chromatogram for 2 p1 injection of TORT-2 (103.1 mg) A methylbutylmercury (2.6 min); and B dibutylmercury (4.1 min).Concentrations of different species are presented in Table 4 t a v) C 0 a v) a U B 120 m" A 0 1 2 3 4 5 Time/min Fig. 3 Chromatogram for 2 pl injection of DOLT-2 (102.2 mg) A methylbutylmercury (2.6 min); and B dibutylmercury (4.1 min). Concentrations of different species are presented in Table 4 1 I 1 I 1 1 0 1 2 3 4 5 Time/m in Fig. 4 (4.1 min) peak arising from reagent contamination Chromatogram for 2 pl injection of blank A dibutylmercury Table 5 Extraction efficiencies of mercury from mixed pike tissues for different sample work-up procedures Extraction Sample Leaching digestion efficiency (YO) Pike homogenate Acid leaching HCI + NaCl, 12.2 (mixed tissues) Precipitate* Alkaline digestion (15% NaOH+0.2% NaCl) 22.5 (15% NaOH+0.2% NaCl) 88.0 * Re-digestion of precipitate from acid leaching.Table 6 Comparison of alkali digestion and acid leaching for hom- ogenous finely dispersed samples typical for reference materials Concentration one stan- dard deviation/ng g-l Leaching/ Sample Method digestion MeHg Hg2+ €lgtot [AEA MA Certificate - B3/TM IAEA MA GC-MIP* Acid leaching 420 150 5701 B3/TM IAEA MA GC-MIP Alkaline 411 +9 141 +9 552f401 B3/TM digestion - - 510+70 *Ref. 6. ?Sum of MeHg and Hg2+ (n=3). poses and recovery assessment even when reference materials are available to check the accuracy of the analytical procedure. As indicated in Table 5 very large differences in extraction efficiencies can be found for different sample work-up pro- cedures i.e alkali digestion or acid leaching of pike tissues.To study the efficiency of an acid-leaching procedure it is common to fortify the homogenized fish tissues with a known amount of mercury species and determine the recovery after storage for a few hours. The high percentage recovery values obtained are used as evidence of the acid-leaching efficiency of the method. This type of recovery experiment does not however provide unambiguous evidence for the digestion efficiency because the added mercury species probably adhere to the surface protein thiol groups of the homogenized sample and are easily liberated when the acid is added. However mercury bound within sample particles on a-helix coil proteins are not accessible to acid liberation. Characteristics of the Method A potential disadvantage of the sample preparation method used here is that ethylmercury is converted into inorganic mercury during the complexometric extraction step.This has been discussed previously in a study of mercury speciation in human whole blood' and was observed in the present work when ethylmercury was added to several of the matrices shown in Tables 3 and 4. Although not produced naturally in the environment ethylmercury has been detected by GC-ECD in several reference materials (certified for total mercury only) distributed by the IAEA.I7 Horvat17 speculated that the ethyl- mercury in the IAEA samples could have been produced during the ethanol extraction stage of the analytical procedure analogous to the artifact formation reported by Panaro et a1." These latter workers noted that signals corresponding to the retention time of methylmercury in a GC-ECD system using a packed column 'passivated' with inorganic mercury,lg were obtained when acetone or diethyl ether were injected. This problem became particularly pronounced at higher injector temperatures. Indeed Panaro et al." observed significantly higher methylmercury levels in a variety of marine samples when using GC-ECD than total mercury concentrations deter-JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 30 1 Table 7 Speciation of mercury in Northern pike (Esox lucius) Concentration & one standard deviation (n = 3)/ng g- GC-MIP CVAAS Sample Location of catchment Mass/kg MeHg Hg2+ Hgtot Hgtot Pike Pike 236 f 18 1400 f 126* 1363 f 85 Tavelsjon Sweden 1.3 1164&55 Ricklein Sweden 0.8 850 f 40 55 & 10 905 f 170* 858 f 65 *Sum of MeHg and Hg2+ mined by CVAAS.Note that a standard analytical method for methylmercury determinations involves the use of acetone in the sample work-up and GC-ECD.,' The GC-MIP-AES method described here presents a number of features that should preclude any such artifact formation. Firstly no passivation of the chromatographic column is required eliminating the source of inorganic mercury that could participate in unexpected on-column reactions. Secondly toluene is the only organic solvent used in the initial extraction steps for which Panaro et d . l s observed no artifact formation. Thirdly the detection system used is element specific and thus any co-eluting species will in general yield no response. (Only one exception has so far been observed with the present system probably resulting from a spectral inter- ference from phosphorus containing species.') The lack of interfering peaks and excellent chromatographic behaviour of the butylated mercury species is readily apparent in Figs.1-3. As to the inability of the method to differentiate between ethylmercury and inorganic mercury in samples of biological origin the same is true of the aqueous-phase ethylation cryogenic GC-CVAFS technique proposed by Bloom2' and the headspace technique devised by Lansens and Bayens.22 Ethylmercury is thus not amenable to detection in biological samples by any of the modern GC based techniques. The absolute detection limit for the GC-MIP-AES system in its present configuration is 0.4pg of Hg (using the 3s criterion) as either methylmercury or inorganic mercury.Con- centration detection limits depend on the mass of the initial sample volume of toluene used for extraction injection volume and reagent blanks. The long-term stability of the system is excellent during one day the baseline drift is about 5 mV and additionally the response from a single sample remains largely the same from day to day. The stability of the chromatographic system is also very high largely owing to the high capacity of the wide-bore column. Nevertheless following frequent use of the column for the analysis of samples of biological origin involatile organic residues are deposited close to the inlet leading to peak broadening (after > 500 injections) although this problem can easily be overcome by removing the first 40 cm or so of the column at the injector side.Accuracy of the Method As noted above the objective of this work was to verify the accuracy of the proposed GC-MIP-AES based method for the determination of mercury species in environmental and biologi- cal samples. To this end this laboratory has participated in a number of certification programmes (for BCR-463 BCR 464 and TORT-2 yet to be certified) with the objective of producing new RMs. Thus the concentrations of mercury species in these materials were unknown at the start of the exercises. This also applies to DOLT-2 although the results obtained were submit- ted too late to be included in the certification stage. As seen from Table 4 the GC-MIP-AES method has in these intercali- bration exercises given accurate and precise results for marine biological samples.Also included in Table4 is a comparison of results reported previously' with data obtained by Brunmark et for methylmercury and inorganic mercury in human whole blood reference material Seronorm 904. (This material is provided with a reference value for total mercury only.) Brunmark et aZ.I6 determined methylmercury following extrac- tion according to a modified version of the procedure proposed by Westoo2 (as described by Cappon and SmithI3) derivatiz- ation using diazomethane separation by capillary GC and detection by mass spectrometry in the single ion monitoring mode. Inorganic mercury was determined by CVAAS using a version of the method of Magos3 proposed by Velghe et ~ 1 .~ This comparison demonstrates the viability of the GC-MIP- AES method for the determination of mercury species at the low levels typical of human whole blood. Further analytical quality assurance for the determination of mercury species in blood by GC-MIP-AES was obtained in an interlaboratory comparison the results of which were recently reported by Lind et Spiked blood samples were speciated in three laboratories (i) total and inorganic mercury by CVAAS following selective reduction using CdCl + SnCl and SnCl respectively according to a modified version of the method by Magos3 described by Lind et ~ 1 . ; ~ ' ( i i ) methylmer- cury was determined after alkaline digestion aqueous-phase ethylation and cryogenic GC-CVAFS,21 and the total mercury following acid digestion BrCl + SnCl reduction amalga- mation and detection by CVAFS;4 and (iii) methylmercury and inorganic mercury by the present GC-MIP-AES pro- cedure.As shown by the results presented by Lind et all methods performed well and the recoveries of the spiked methylmercury inorganic mercury and total mercury deter- mined by GC-MIP-AES were 96+1 84k2 and 94+8% respectively. Note that for these samples calibration was carried out by the external standard method which could explain the slightly lower recoveries for inorganic mercury. Although these last results represent elevated mercury levels in combination with the data given in Table 4 it seems fairly obvious that the GC-MIP-AES method and associated sample work-up procedure can provide accurate analytical results for methylmercury and inorganic mercury in a variety of matrices over a reasonably extensive concentration range.Conclusions The obtainment of precise and accurate measurements is of fundamental importance in modern society. Quality assurance rules and guidelines (IS0 9000 and EAN 45000) acreditation authorities as well as CRMs have been established to ensure high-quality measurements in vital sectors in society e.g food agriculture environment industrial products and consumer protection. The GC-MIP-AES procedure described provides an accu- rate and precise (Table 4) method for the simultaneous determi- nation of methylmercury and inorganic mercury in environmental and biological samples. This paper demon- strates the viability and accuracy of the proposed method employing only simple wet chemistry and provides highly efficient chromatography (Figs.1-3) coupled with a very sensitive and specific detection system. It would be interesting to perform laboratory intercomparisons between GC-ECD based standardized methods and element specific methods to elucidate whether there is a significant difference in the results as pointed out by other workers."302 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 This work was supported in part by the Swedish Environmental Protection Board and the Centre for Environmental Research in UmeA. References 1 Donard O. and Quevauviller P. Mikrochim. Acta 1992 109 1. 2 Westoo G. Acta Chem. Scand. 1967 21 1790. 3 Magos L. AnaZyst 1971 96 847. 4 Bloom N.and Fitzgerald W. F. Anal. Chim. Acta 1988,208 151. 5 Bloom N. S. and Crecelius E. A. Mar. Chem. 1983 14 49. 6 Bulska E. Baxter D. C. and Frech W. Anal. Chim. Acta 1991 249 545. 7 Byrne A. R. Analyst 1992 117 251. 8 Bulska E. Emteborg H. Baxter D. C. Frech W. Ellingsen D. and Thomassen Y. Analyst 1992 117 657. 9 Emteborg H. Baxter D. C. and Frech W. Analyst 1993 118 1007. 10 Quevaviller Ph. Drabaek I. Muntau H. and Griepink B. Certification Report EUR. Report. CEC Brussels Belgium in preparation. NBS Certijicate of Analysis Standard Reference Material 1566 Oyster Tissue National Bureau of Standards (now National Institute for Standards and Technology) Washington DC 1979. 12 Kurfurst U. Pauwels J. Grobecker K.-H. Stoeppler M. and Muntau H. Fresenius’ J. Anal Chem. 1993 345 112. 11 13 14 15 16 17 18 19 20 21 :22 :23 124 :25 Cappon C. J. and Smith J. C. Anal. Chem. 1977 49 365. Emteborg H. Bulska E. Frech W. and Baxter D. C. J. Anal. At. Spectrom. 1992 7 405. Donard 0. F. X. and Pinel R. in Environmental Analysis using Chromatography Interfaced with Atomic Spectroscopy ed. Harrison R. M. and Rapsomanikis S. Ellis Horwood Chichester Brunmark P. Skarping G. and Schutz A. J. Chrornatgr. 1992 573 35. Horvat M. Water Air Soil Pollut. 1991 56 95. Panaro K. W. Erickson D. and Krull I. S. Analyst 1987 112 1097. O’Reilly J. E. J. Chromatogr. 1982 238 433. Hight S. C. and Corcoran M. T. J. Assoc. Of. Anal. Chem. 1987 70 24. Bloom N. Can. J. Fish. Aquat. Sci. 1989 46 1131. Lansens P. and Bayens W. Anal. Chim. Acta. 1990 228 93. Velghe N. Campe A. and Claeys A. At. Absorpt. Newsl. 1978 17 139. Lind B. Body R. and Friberg L. Fresenius’ J. Anal. Chem. 1993 345 314. Lind B. Friberg L. and Nylander M. J. Trace. Elem. Exp. Med. 1988 1 49. 1989 ch. 7 pp. 189-222. Paper 3/04427K Received July 29 1993 Accepted October 5 1993