JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL. 6 Comparison of Dry Mineralization and Microwave-oven Digestion for Determination of Arsenic in Mussel Products by Platform in Furnace Zeeman-effect Atomic Absorption Spectrometry N. Ybbiiez M. L. Cewera and R. Montoro 379 the lnstituto de Agroquimica y Tecnologia de Alimentos (CSIC) Jaime Roig 7 1 46070 Valencia Spain Miguel de la Guardia Depafiamento de Quimica Analitica Universidad de Valencia Or. Moliner 50 46 100 Burjassot Valencia Spain Two digestion procedures were compared in order to obtain an accurate method for the determination of As in mussel products using platform in furnace Zeeman-effect atomic absorption spectrometry. One procedure was based on dry mineralization of the samples and the other on microwave-oven sample digestion in closed polytetrafluoroethylene reactors.Microwave-oven digestion with HN03 and H202 allows the accurate determination of As in certified oyster and mussel tissue samples and provides results comparable to those found by dry mineralization of real mussel product samples with a sensitivity of 3.9 absorbance units per pg ml-’ and a relative standard deviation of 5% in the analysis of samples containing 8 pg g-‘ dry mass of As. Experimental conditions for the determination of As in real samples were optimized and a series of real samples analysed in order to determine the analytical characteristics of the proposed procedure. Sample treatment time was reduced from 2-3 d using dry mineralization to 20 min per sample using microwave-oven digestion. Keywords Arsenic determination; mussel product sample; dry mineralization; microwave-oven digestion; Zeeman-effect atomic absorption spectrometry The microwave decomposition technique seems to be the method currently preferred when preparing samples for the determination of the elemental composition of organic materials.At present marine biological samples are usually prepared by wet digestion with concentrated acids includ- ing perchloric acid or by dry mineralization with ashing as an aid prior to instrumental analysis using techniques such as flame and electrothermal atomic absorption spectrome- try (AAS). The use of HC104 can be dangerous in the presence of organic material and so the dry mineralization procedure is becoming more widely used because of its simplicity and safety.14 Nevertheless this sample treat- ment requires long heating periods which becomes the limiting factor determining the speed of analysis. Recently very fast safe and efficient acid decomposition methods based on the use of microwave ovens have been proposed for the determination of As in a variety of samples but few reports have been published on the determination of As in marine samples using microwave heating.Commercially available polytetrafluoroethylene (PTFE) pressure relief vessels or completely closed vessels have been used in microwave ovens with monitoring of the pressure during d i g e s t i ~ n . ~ ~ ~ The accuracy of the methodol- ogies has been tested by analysis of a National Research Council of Canada (NRCC) biological standard (lobster hepatopancreas TORT- 1 ).However until now there have been no publications in which real samples of marine foods have been analysed by electrothermal AAS following microwave-oven pressure digestion. It is evident that incomplete destruction of organic matter in the microwave oven could affect the accuracy of the analysis. The determination of As in marine biological samples by atomic absorption using the stabilized temperature plat- form furnace (STPF)’ concept in combination with Zee- man-effect background correction provides a dramatic reduction of non-spectral interference in fish tissues,* and the results are not affected by spectral interferen~e.~ In the present paper complete destruction of mussel products carried out by dry mineralization and acid extraction inside a microwave oven are compared as sample preparation methods for the platform in furnace Zeeman-effect atomic absorption spectrometric determina- tion of As.Experimental Equipment A Heraeus Model 1100/3 mume furnace equipped with a Jumo DPG-4411 digital microprocessor was used for mineralizing the samples. A domestic micro-wave oven Bauknecht Model MWT-732 programmable for time and microwave power in four discrete steps with eight power settings (ranging from 120 to 750 W) and a maximum time of 99 min 59 s was used without further modification. It is important to point out that different positions in the microwave cavity are not identical from the microwave irradiation point of view and so a prior study of the irradiation efficiency in each position is necessary. In order to carry out this study 35 Pyrex beakers (55 mm in diameter and with a volume of 250 ml) each containing 100 ml of distilled water were placed inside the cavity of the microwave oven covering the whole of the bottom surface of the oven.The oven was operated at maximum power for a total time of 2 h. After the beakers had cooled the loss of water from each beaker was measured and related to the level of microwave power absorbed taking the maximum loss of water as equivalent to a power absorption level of 1OOOh. Fig. 1 shows the microwave distribution in the oven from which it is concluded that the central position must be used in order to obtain effective sample decomposition. The polytetrafluoroethylene (PTFE) vessels used for the solutions were laboratory made and had a volume of approximately 120 ml with a 10 mm wall thickness and tight-fitting screw-cap lids.For pressurized acid digestion of the samples in thermal ovens hermetically sealed borosilicate glass tubes (40 ml) with Bakelite screw caps and Parr digesters with a volume of 55 ml were employed. A Perkin-Elmer Model Zee- man/3030 atomic absorption spectrometer equipped with an HGA-600 graphite furnace and an AS-60 autosampler was also used. This instrument includes a graphics display380 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 Fig. 1 Distribution of microwave radiation within the oven cavity unit and highly time-resolved signals were plotted with a Perkin-Elmer PR- 100 printer. Pyrolytic graphite coated graphite tubes with a L'vov platform inserted (Perkin- Elmer Part No.112660) were used exclusively. Reagents Analytical-reagent grade water with a metered resistivity of 18 MR cm was used to prepare all samples and standards. All reagents used were of the highest purity available and at least of analytical-reagent grade. An aqueous stock solution of As"' was prepared by dissolving arsenic(rI1) oxide. The ashing aid suspension was prepared by stirring 20 g of Mg(NOJ2-6H20 and 2 g of MgO in 100 ml of water until a homogeneous mixture was obtained. A nickel chemical modifier solution (0.1% m/v Ni) was prepared by dissolving the Ni(N03)2.6H20 in 1% v/v HN03. General Procedures Real and certified samples of mussel products were treated by both dry mineralization and microwave-oven decompo- sition and the digested solutions analysed by platform in furnace Zeeman-effect AAS.The experimental conditions of the microwave-oven digestion were optimized and a study of the atomization of samples and standards was carried out in order to obtain the best analytical character- istics and results from the two digestion procedures. 43.4 Samples Several cans of mussel products in brine and in pickled sauce were purchased at a local retail market. The brine or sauce was removed from the mussels by the method for determining drained mass in canned foods. The total contents of the can were emptied onto a sieve with a 5 mm stainless-steel mesh made of 1 mm gauge wire the sieve being tilted slightly to facilitate drainage. To ensure complete drainage the mussels were allowed to drain for 5 min.The drained mussels were then crushed and homogen- ized to a fine paste in a domestic Moulinex blender stored in previously de-contaminated glass flasks and kept in the freezer until the analysis could be carried out. A reference mussel product sample was prepared so that the different sample treatments could be compared. The paste obtained by crushing and homogenizing the mussels was dried at 100 k 3 "C to constant mass and then ground to pass through a 250 pm sieve and stored at -4 "C until required. Reference Samples Validation of the methods presented in this study was performed by using two reference materials one from the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 1556a Oyster Tissue and the other from the Community Reference Bureau (BCR) Certified Reference Material (CRM) 278 Mussel Tissue. Results and Discussion Determination of Moisture in Mussel Products A microwave oven can also be employed as a drying system.Io Therefore in the present study the same appara- tus as was employed for sample digestion was used to determine moisture level.The results obtained using conventional drying in a thermal oven for 24 hours at 110 "C and those obtained in the microwave oven after 3 min at 385 W and 5 min at 555 W are summarized in Table 1. It can be seen that the differences between the two series of results are of the order of f 0.7%. Dry Mineralization The various experimental parameters for the mussel pro- ducts such as sample size amount of HN03 and ashing agents added mineralization steps in the muffle furnace and solubilization of the ash obtained were optimized.ll The different steps involved in dry mineralization of the mussel products are summarized in Table 2.The recom- mended procedure is as follows 5 ml of 5Ooh v/v HN03 and 1 ml of ashing aid suspension containing 20% m/v Mg(N03)2 and 2Oh m/v MgO were added to 0.250 g (dry mass) of mussel product or certified sample or 1.00 g (wet mass) of mussel product and the solution was mixed well. After evaporation to total dryness in a sand-bath the dry residue was subjected to a careful mineralization process (programme 1) in a muffle furnace at a temperature lower Table 1 Determination of moisture in mussel products Moisture content* (%) Sample Conventional Microwave E n o e (%I) 1 69.9 67.3 0.4 2 67.0 67.6 0.6 3 59.5 58.8 -0.7 * Average of three determinations.t The percentage ratio of the results obtained using a microwave oven and a thermal oven. Table 2 Steps involved in the dry mineralization of mussel products. The time taken to increase the temperature by 100 "C was 15 min and by 150 "C was 30 min Programme Step Temperature/"C Time/h I 1 150 1 2 200 0.5 3 250 1 4 300 3 5 350 0.5 6 450 12 I1 1 150 1 2 300 0.5 3 450 12JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 38 1 than 450 "C until a white ash was obtained. In general it was necessary to wet the ash with 50% v/v HN03 evaporate in a sand-bath and perform a second short mineralization process (programme 11) (once or twice) until the ash was completely white.The ash was dissolved with 2 ml of 50% v/v HN03 washed with water and filtered through What- man No. 1 filter-paper into a 50 ml flask. The solution was stored in a polyethylene bottle. Microwave-oven Digestion The dry mineralization procedure involves a very long (2 or 3 d) sample treatment time. To ensure complete dissolution of the marine samples for the determination of As by AAS the conventional sample preparation procedure is wet digestion. In order to release As from organoarsenic compounds mixtures of corrosive and sometimes explosive acids (including HClO,) and long heating periodsl1J2 or pressure decomposition9 are necessary. Moreover other workers have stated that it is necessary to maintain oxidizing conditions at all times particularly if chloride is present in order to prevent volatilization of As as the trichloride.The vital point seems to be that the sample must be heated in the presence of excess of nitric acid first and only allowed to char when the chloride has been removed as nitrosyl chloride.13 In order to avoid the dangerous use of HClO in the presence of organic material and to test whether the sample can be decomposed by acid digestion pressure decomposi- tion with nitric acid in a conventional oven was used initially since it has been reported that if a sample material cannot be decomposed by a conventional method prob- lems using the microwave heating technique will occur.14 Simple rapid pressure extraction with 0.5 ml of HN03 in a stoppered borosilicate glass container which was adequate for the determination of Cd Cu Fe Pb and Zn in mussels,15 provided unsatisfactory precision (1 4%) and low recoveries (95*39%) for the determination of As by platform in furnace AAS.For this reason acid digestion in a PTFE bomb heated to 140 "C for 1 h in a conventional oven with HN03 (1 2 or 5 ml) and HN03-HCl was tried prior to digestion in the microwave oven. The results obtained using 2 ml of acid are in agreement with the contents found in the reference mussel sample digested by dry mineralization. Finally the efficiency of the acid digestion procedure was enhanced by using a microwave oven. Different volumes of HN03 and H202 and different times and power settings of the microwave oven employing a single-stage or two-stage power and time setting were tested in order to ensure total recovery of As after sample treatment.Heating for 5 min at 555 W resulted in an excess of vapour pressure inside the vessel and sample charring. Therefore sample treatment was reduced to a two-stage power and time setting. The oven was operated at 555 W for 1 min and then the power was reduced to 300 W for 4 min. A 0.5 ml volume of HNOJ provided unsatisfactory recovery of As in the reference mussel sample similar to that obtained using 0.5 ml of HN03 in a stoppered borosilicate glass container. A 2 ml volume of HN03 digests 0.250 g (dry mass) of mussel sample yielding a solution with fat residues which gives promising results for the determination of As by platform in furnace AAS as indicated in Table 3.It is generally accepted that organic materials are not totally decomposed to C02 and water by HN03. Perchloric acid is used for the oxidation of resistant fat tissue and organoarsenic compound^.^ However using H202 avoids the use of HClO, eliminating difficulties associated with corrosive fumes and the potentially explosive nature of some nitric-perchloric acid mixture^.^ The addition of 2 ml of H202 with 2 ml of HN03 improves the efficiency of the digestion providing a clearer solution after sample diges- tion and greater accuracy in the analysis of real samples according to the results obtained by Matusiewicz et aL6 from an NRCC reference sample (lobster hepatopancreas The recommended procedure for microwave-oven diges- tion of samples was as follows 0.250 g of dry sample (mussel product or certified sample) or 1.00 g of wet sample (mussel product) was placed in a high pressure PTFE vessel and 2 ml of 65Oh m/v HN03 and 2 ml of 30% m/v H202 were added.The vessel was sealed with the screw cap and placed inside the microwave oven. Samples were irradiated at a 555 W power setting for 1 min and then the power was reduced to 300 W for 4 min. The vessel was removed from the oven and placed in a tray of ice-water for 15-20 min. The cap was removed the contents of the vessel washed with water and filtered through Whatman No. 1 filter-paper into a 50 ml flask and the solution was stored in a polyethylene bottle until required for analysis. TORT- 1 ). Table 3 Analytical characteristics of the methods Analytical procedure Microwave digestion Recovery of As Dry mineralization HN03 HN03-H202 Sensitivity/absorbance unit per pg ml-l Detection limit*/pg g-I Precision RSDt (4s) Mussel reference samplefl/pg g-' NIST SRM 1566a Oyster Tissuell/pg g-I (certified value 14.0 f 1.2/pg g-l) BCR CRM 278 Mussel TissueJI/pg g-' (certified value 5.9 f 0.2/pg g-l) 4.6 4.0 3.9 2.4 1.5 1.8 2 3% 69 5 8.3 +- 0.2 7.5 f 0.2 8.2 f0.3 13.1 k0.8 14.0k 0.6 13.1 & 1.2 6.0 k 0.2 - 6.2 f 0.3 * Eight reagent blanks were employed.$ Sample mass=0.250 g (dry mass). 9 Sample mass = 1 .OO g (wet mass). fl Mean and confidence semi-interval at the 95% confidence level based on analyses of eight replicate 0.250 g samples. 11 Mean and confidence semi-interval at the 95% confidence level based on analyses of four replicate 0.250 g samples.RSD = relative standard deviation from eight independent analyses.382 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 Table 4 Graphite furnace operating conditions for the determination of As in mussel products Operating conditions- Electrodeless discharge lamp Wavelength Spectral bandwidth Background correction Measuring mode Graphite furnace Furnace tube Injection mode Sample volume Chemical modifier 8.5 W 193.7 nm 0.7 nm (low) Zeeman Integrated absorbance calculated by HGA software Pyrolytic graphite coated graphite tubes with a L'vov platform Autosampler AS40 10 pl of Ni (0.1 Oh mlv) HGA-600 20 pl Furnace programme- Timels Internal Step "C Ramp Hold ml min-l Temperature1 Ar flow rate1 Drying Drying Mineralization Mineralization Atomization Cleaning Cooling 90 10 20 300 120 10 20 300 800 10 10 300 1100 10 10 300 2300 0 5 0 2650 1 5 300 20 10 10 300 Optimization of the Graphite Platform in Furnace Pro- gramme For the graphite platform in furnace Zeeman-effect atomic absorption spectrometric determination of As 1 ml of H20 and 20 pl of standard solutions containing 0 2 4 and 6 pg ml-l of As were added to digested sample aliquots of 1 ml and the resulting solutions were analysed using the standard additions method under the experimental condi- tions indicated in Table 4.The recommended analytical conditions for pyrolytic graphite coated platform in furnace Zeeman-effect determi- nation of As with STPF conditions16 are the addition of 10-20 pg of Ni (as nitrate) as a chemical modifier a thermal pre-treatment temperature of 1300 "C and an atomization temperature of 2300 "C.The atomization and mineraliza- tion temperature-time programme was optimized in order to provide maximum matrix decomposition without loss of As compatible with minimum background and maximum As absorbance in the atomization stage. Experimentally it has been confirmed using As standards that the Zeeman- effect signal increases between 1800 and 2300 "C thereafter decreasing. In order to enhance the lifetime of the graphite tube the atomization temperature for real samples can be decreased from 2300 to 2100 "C without any significant differences. Between 1000 and 1500 "C the mineralization temperature has no influence on the absorbance values obtained for an acidic standard solution of As and real samples.On the other hand a mineralization time of 10 s gives similar sample mineralization to a time of 60 s but the corresponding background signal is slightly increased. Matrix Effects In order to study the effect of the matrix on absorbance and background signals solutions were prepared containing 40 ng ml-l of As and various percentages of digested sample (0 10 25 50 75 and 100%). The matrix effect was evaluated in the presence of 10 and 20 pg of Ni as chemical modifier. Fig. 2 shows the results obtained. Significant differences were not detected between the use of 10 and 20 pg of Ni as chemical modifier. The variations in absorbance and background using solutions with 10 and 25% of digested 0.25 1 0 20 40 60 [AsVng mi-' Fig.3 Analytical curves obtained using the standard additions method A 0; B 25; C 50; D 75; and E 100% of the digested sample in the volume injected on the platformJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL. 6 383 0.4 0.3 9) C n U 5 0.2 9" 0.1 0 0 20 40 60 80 [AsYng ml-' Fig. 4 Calibration graphs obtained using A aqueous acid stan- dards and B the standard additions method for 10 pg of Ni added I A I -2 8 0 fG1-O s 11 ,... ....._ AA 0.5 I BG 1.0 I C 0 0 5 Timeis Fig. 5 Absorbance versus time profiles. for As determination @g g-I) in mussel products A dry mineralized mussel product; B HN03 microwave digested mussel product; and C HN03-H202 microwave digested mussel product. Solid lines correspond to absorbance readings and broken lines correspond to background readings slight increase in back ground.The influence of the matrix effect on the slope of the calibration graph was also studied. Fig. 3 shows the decrease in the slope of the standard additions graph when the percentage amount of digested sample in the solution injected on the platform increases and indicates that solutions with > 50% of digested sample strongly reduce the slope of the graph. In view of these results solutions with 50% of digested sample were selected to be injected onto the platform. In order to express the results in pg g-l from solutions containing a lower percentage of digested sample the instrumental readings must be multiplied by a very high factor (1 000 for 10% of digested sample and 800 for 25% of digested sample) which makes the analytical character- istics detection limit precision and accuracy unacceptable.The solution containing 50% of digested sample allows the determination of the As content of the spiked sample in the concentration range between the detection limit and 11.6 pg g-l for the microwave digested sample working within the linear range up to 0.900 absorbance units. The matrix effect was also observed in dry mineralized samples as can be seen in Fig. 4. It was verified that the effect is not due to the use of Mg as an ashing aid by comparing calibration graphs for As without Mg and those containing Mg in amounts 5 and 10 times higher than are present in dry mineralization (0 3 1 and 62 pg of Mg). The slopes of the calibration graphs show no significant differ- ences (3.26 3.48 and 3.24 absorbance units per pg ml-').A 50% solution of the ashed sample was selected to be injected onto the platform because the linear working range for analysis between the detection limit and 12.8 pg g-l (dry mass) was similar to that employed in microwave diges- tion. Moreover maintaining the same dilution factor in both methodologies allows a more accurate comparison of them. Non-spectral interference in the determination of As in mussel products caused by sample matrix cannot be totally corrected for in spite of the use of a Ni chemical modifier and STPF conditions and the standard additions method should be employed. Fig. 5 indicates the shape and size of the atomic absorption peaks for As and the corresponding background values when A dry mineralization; €3 microwave digestion with HN03; and C microwave digestion with HN03-H202 were used.The use of magnesium salts in dry mineraliza- tion increases the background signal. A typical background value for the aqueous As standards and HN03 microwave digests is of the order of 0.075 A and for HN03-H202 microwave digests 0.122 A. However atomization of dry mineralized samples gives background values of 0.190 A. Comparison of Analytical Characteristics of the Method In order to check the validity of the proposed methods for the determination of As in mussel products the analytical characteristics such as sensitivity detection limit precision and accuracy were evaluated in accordance with IUPAC recommendations. Sensitivity was established from the mean value of the slopes of the standard additions curves and expressed in absorbance units per pg ml-l.The detection limit established as the As concentration in pg g-l (dry mass) of mussel product which provides an absorbance reading statistically different from that of the blank was calculated by dividing three times the standard deviation of the absorbance readings of the reagent blanks (which were always of a very low magnitude) by the sensitivity taking into account the sample mass and dilution employed. Precision is expressed as the relative standard deviation (RSD) of eight independent analyses of the same sample of mussel product (reference sample). In order to evaluate the accuracy of the methods two certified samples SRM 1566a Oyster Tissue and CRM 278 Mussel Tissue were analysed. The results are given in Table 3 and demonstrate that both the proposed methods dry mineralization and HN03-H202 microwave digestion are appropriate for the determination of As in real samples.The higher matrix effect in the microwave digests slightly decreases the analytical sensitivity. The detection limits obtained by the three methods (including HN03 microwave digestion) are comparable because as the concentrations of As in the reagent blanks are almost indetectable in all instances their reproducibil- ity mainly depends on instrumental noise. The three detection limits are adequate for the determination of As in the mussel products analysed. The precisions found for the three digestion methods are comparable.The greater efficiency of HN03-H202 com- pared with HN03 digestion of the mussel reference sample in the microwave oven does not provide greater analytical precision. On the other hand the precision is slightly less when a wet mussel product sample instead of the dried sieved reference mussel product sample is analysed. The high level of accuracy of the methods is demon- strated by the good agreement of the results obtained in the analysis of the reference materials with the certified results.384 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL. 6 Table 5 Analysis of real samples Concentration of Aslpg g-* Sample Dry mineralization A B C D E Averagelpg g-' Standard deviationlpg g-I RSD (%) xk c10.05*/pg g-1 5.4 f 0.2 6.9 k 0.4 8.8 f0.5 6.1 f 0.3 6.0 f 0.6 6.6 1.3 20 6.6 +- 1.6 * X+ - CI0.05 = mean f confidence semi-interval at the 95% confidence level.Microwave digestion 7.2 f 1 .O 5.9 f 0.8 9.3 k 0.8 5.5 ? 0.3 6.3 f 0.3 6.8 1.1 22 6-82 1.9 Nevertheless the microwave-oven digestion with HN03-H202 provides the most accurate results for the analysis of Oyster Tissue. In the analysis of the reference mussel product sample by microwave digestion the additional use of H202 gives higher As contents than that obtained employing only nitric acid and this is in good agreement with the data obtained by dry mineralization. Analysis of Real Samples The determination of As in real mussel product samples was carried out by dry mineralization and microwave-oven digestion procedures. The results obtained in three replicate analyses of five real samples by the two methods used are comparable as can be seen in Table 5 Applying a Student's t-test to the mean concentrations obtained by the two methods the experimental t value obtained falls within the acceptance area of the null hypothesis at a significance level of a=0.05 (te,,=0.24<t8=2.306).Table 5 also shows that the confidence limits for the mean concentrations obtained by the two methods almost completely coincide. Conclusions Mussel product digestion with HN03-H202 in a pressur- ized PTFE vessel using microwave heating for a graphite platform in furnace AAS determination of As provides an efficient alternative to conventional dry mineralization. The detection limit precision and accuracy are similar for the two methodologies and appropriate for the determina- tion of As in real samples.Sample preparation times are reduced from 2-3 d using dry mineralization to about 20 min including subsequent cooling time employing the microwave heating source. Funds to carry out this work were provided by the Comisi6n Interministerial de Ciencia y Tecnologia (CI- CyT) Projects ALI89-052 1 and PB88-035 1 for which we are deeply indebted. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Sturgeon R. E. Willie S. N. and Berman S. S. J. Anal. At. Spectrom. 1986 1 11 5 . Dabeka R. W. and Lacroix G. M. A. J. Assoc. Off Anal. Chem. 1987,70 866. Krynitsky A. J. Anal. Chem. 1987 59 1884. Brumbaugh W. G. and Walther M. J. J. Assoc. Off Anal. Chem. 1989,72,484. Nakashima S. Sturgeon R. E. Willie S. N. and Berman S. S. Analyst 1988 113 159. Matusiewicz H. Sturgeon R. E. and Berman S. S. J. Anal. At. Spectrom. 1989 4 323. Slavin W. Manning D. C. and Carnrick G. R. At. Spectrosc. 1981 2 137. May T. W. and Brumbaugh W. G. Anal. Chem. 1982 54 1032. Welz B. and Schlemmer G. J. Anal. At. Spectrom. 1986 1 119. de la Guardia M. Empleo de 10s Hornos Microondas en Quimica Universidad de Valencia 1990. Agemian H. and Thomson R. Analyst 1980 105 902. Maher W. A. Talanta 1983 30 534. Gorsuch T. T. The Destruction of Organic Matter eds. Belcher R. and Frieser H. Pergamon Press 1st edn. vol. 39 1970. Ohls K. ICPZnJ Newsl. 1990 15 784. Solchaga M. Montoro R. and de La Guardia M. J. Assoc. Ofl Anal. Chem. 1986,69 874. Perkin-Elmer Analytical Methods for-Furnace Atomic Absorp- tion Spectrometry Perkin-Elmer Uberlingen Publication B332 1984. Irving H. M. N. H. Freiser H. and West T. S. International Union of Pure and Applied Chemistry Compendium of Analytical Nomenclature Pergamon Press Oxford 1978. Paper Om485 5 K Received October 29th I990 Accepted February 8th 1991