JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 29 1 On-line Microwave Oxidation for the Determination of Organoarsenic Compounds by High-performance Liquid Chromatography-Hydride Generation Atomic Absorption Spectrometry* M. Angeles Lopez-Gonzalvez M. Milagros Gomez Carmen Camara and M. Antonia Palaciost Departamento de Quimica Analitica Facultad de Ciencias Quimicas Universidad Complutense de Madrid 28040 Madrid Spain An on-line high-performance liquid chromatography (HPLC)-microwave oxidation-hydride generation atomic absorption spectrometry coupled system has been developed for the determination of arsenite arsenate dimethylarsinate (DMA) monomethylarsonate (MMA) arsenobetaine and arsenocholine in environmental samples. An anionic cartridge placed before the HPLC anionic column (Hamilton PRP-X1 00) quantitatively retains anionic species such as arsenite arsenate MMA and DMA but not cationic species which are separated and quantitatively determined after microwave-K,S,O decomposition. The anionic species can be separated and determined quantitatively by removing the anionic cartridge from the system and introducing water instead of K,S,O solution.Detection limits of between 0.3 and 0.9 ng were achieved for all species. A conversion efficiency close to 100% was achieved for the species tested working with a microwave oven power of 700 W and 5% K2S208 oxidizing solution. Keywords Arsenic speciation; high-performance liquid chromatography; microwave oxidation; hydride generation; atomic absorption spectrometry Arsenic is widely distributed in the biosphere as a result of its widespread use in industry and agriculture.' Since arsenic species differ greatly in toxicity there is increasing interest in developing analytical procedures to determine quantitatively arsenic compounds in the environment and in biota.Arsenite and arsenate are very toxic whereas dimethylarsinic acid (DMA) monomethylarsonic acid (MMA) and arseno- choline (AsC) are generally less toxic and arsenobetaine (AsB) seems to be n o n - t ~ x i c . ~ ~ ~ The combination of high-performance liquid chromatography (HPLC) and hydride generation atomic absorption spectrometry (HG-AAS) is an important tool for the speciation of these six arsenic species in environmental samples. Nevertheless two problems have not yet been satisfact- orily resolved.Firstly it is difficult to separate quantitatively arsenic species by HPLC and thus chromatographic peak overlapping occurs when certain columns are used. For example arsenite and AsB drift together in the most common anionic and reversed-phase c o l ~ m n s . ~ ~ The coupling of differ- ent columns to separate species more effectively has been reported,&' but the detection limits are unacceptably high. Secondly hydride generation from DMA is less efficient than from inorganic arsenic species and furthermore AsB and AsC do not generate arsine.'-1° Organoarsenic compounds are usually decomposed by photo-oxidation using a combination of UV radiation and K2S208-NaOH when AsB and AsC are to be determined by on-line HG-AAS." Recently papers dealing with on-line analy- sis of liquid samples by microwave digestion with flow injection (FI) AAS and by HG or cold vapour AAS for the determi- nation of Hg As Bi Pb and Sn in urine and waters have been p~blished'~.'~ In a previous paper14 a more efficient HG from organoarsenicals by on-line HPLC-thermo- oxidation-HG-AAS using a powdered graphite oven at 140 "C instead of a UV radiation lamp was reported.The efficiency of conversion of the organoarsenic compounds into arsenate by microwave oven using an FI system has been recently reported by Le et a1." The present work investigates the suitability of an anionic cartridge for separating AsB and arsenite and the feasibility of an on-line HPLC system coupled ~~ ~ * Presented at the XXVIII Colloquium Spectroscopicum Inter- t To whom correspondence should be addressed.nationale (CSI) York UK June 29-July 4 1993. to a microwave oven for the decomposition of organoarsenicals before HG. Experimental Apparatus The hyphenated H PLC-micro wave oxida t ion-HG- AAS apparatus is shown in Fig. 1. Chromatographic module Separation was by HPLC using a Hamilton PRP-X 100 column (25 cm x 4.1 mm i.d.) and a high pressure solvent pump (Waters Model 590). A Waters IC-H anionic cartridge located upstream of the ionic column was used when AsB and AsC species were determined. The anionic cartridge was regenerated after 40 injections for multi-standard solutions and after about 20 injections for samples. For anionic cartridge regeneration 10 ml of 10% NaOH or 10% KOH were forced through the cartridge using a syringe. The solution was removed from the cartridge with 20 ml of Milli-Q water (Millipore).Samples and standard solutions were injected through a six-port Rheodyne type 50 low-pressure sample injection valve fitted with a 100 p1 loop. Microwave oxidation reactor A domestic microwave oven (Balay Model BAHM-111) with a maximum power output of 700 W (variable in nine steps) and an operating frequency of 2450 MHz was used. A loop of poly(tetrafluoroethy1ene) (PTFE) tubing (1.5 m 0.5 mm id.) was placed inside the microwave oven through the ventilation holes. A 250ml beaker filled with water was used to prevent overheating. Hydride generation module The thermo-oxidized effluent was dipped in an ice-bath (Teflon tubing 0.5 m x 0.5 mm i.d.) before HG in order to lower its high temperature and to avoid over-pressure and decompo- sition of sodium tetrahydroborate. The continuous manifold used to generate arsine consisted of PTFE tubing a four-channel peristaltic pump (Gilson HP4),292 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 1 o r 2 Sample I-i Cartridge Microwave oven NaH2P0 Na,HPO ml min-' - U HPLC Pump I K2S208 0.6 ml m i K ' t Ice-bath Gas-I i q u i d TI HCI NaBH separator 1.9 ml min-' 1.9 ml min-' mixing and reaction joints (Teflon tubing 0.5 mm i.d.) and a V-tube gas-liquid separator (Philips). Hydrides were trans- ported by intermediate argon flow to a quartz atomization cell heated by an acetylene-air flame. Atomic absorption spectrometer A Model 2380 Perkin-Elmer atomic absorption spectrometer equipped with an electrodeless discharge lamp operated at 10 W from an external power supply was used.A spectral bandwidth of 0.7nm was selected to isolate the 193.7 nm wavelength. The signals were relayed to a printer and peak height was recorded. Ar Fig. 1 HPLC-microwave oxidation-HG-AAS manifold for arsenic speciation Reagents All reagents used were of analytical-reagent grade. De-ionized water from a Milli-Q system was used throughout. Stock solutions of arsenic compounds (1000rngl-' as As) were prepared by dissolving appropriate amounts of NaAsO (Carlo Erbaj Na,HAsO4.7H2O (Merck) CH,As0,Na26H20 (Carlo Erbaj and (CH3),As0,Na-3H20 (Sigma). The AsB and AsC were reference standards from the Community Bureau of Reference (BCR). The standard stock solutions were stored in glass bottles kept at 4 "C in darkness.Dilute arsenic solutions for analysis were prepared daily. A 5% m/v potassium persulfate (Merck) solution stabilized in 2.5% m/v NaOH (Merck) prepared daily was used as oxidizing solution. A 3% m/v NaBH (Aldrich) solution stabilized in 1.5% m/v NaOH was used as reducing solution. The concentration of HCl was 3 moll-'. An HPLC mobile phase of 17mmoll-' was prepared by mixing solution A [ 1.17 g of NaH,PO,.H,O (Panreac) in 500ml of water] with an appropriate volume of solution B [ 1.20 g of Na,HPO (Panreac) in 500 ml water] to give pH 6.0. A 5 mmol 1-1 mobile phase was prepared similarly. The resulting solutions were filtered through a 0.45 pm membrane filter and de-gassed before use. Sample Pre-treatment Samples were collected and stored in acid-washed polyethylene bottles.Samples not used on the day of collection were stored at 4 "C in darkness. Carbonated waters were de-gassed by sonication for 30 min. Analytical Procedure On-line determination of AsB and AsC Samples and multi-standard solutions (six arsenic species) were injected through the anionic cartridge leading into a 100 pl sample loop. Arsenite arsenate MMA and DMA were retained in the anionic cartridge and AsB and AsC passed through it into the 100 p1 sample loop and entered the chromatographic column. The mobile phase was 5 mmol 1-' phosphate buffer at pH 6.0 and at a flow rate of 1 ml min-'. The column effluent mixed with the oxidizing solution (5% K2S20s in 2.5% NaOH) flowed to the digestion coil in the microwave oven for decomposition.The solution from the microwave oven was cooled in an ice-bath and a T-junction was used to acidify the sample with 3 mol I-' HCl. In a second T-junction the acidified solution was mixed with the 3% NaBH solution. The resulting solution containing volatile arsine flowed to the gas-liquid separator where the liquid phase was drained off and the gas phase entered the quartz atomization cell. The chromatogram peak height signals were recorded. In all subsequent experiments the peak height was the average of at least three injections of each solution. On-line determination of arsenite arsenate M M A and DMA Arsenite arsenate MMA and DMA were determined as described above for AsB and AsC except that the anionic cartridge was removed a mobile phase of 17mmoll-' phos- phate buffer at a flow rate of 2 ml rnin-l was employed and water was introduced instead of persulfate. Under these con- ditions no AsB or AsC signal was observed.Results and Discussion The optimization of the experimental HG conditions is described in a previous paper.' Chromatographic Parameters It has been shown in previous work14 that the arsenite and AsB peaks overlap under all the chromatographic conditions tested. Since AsB and arsenite exhibit different ionic behaviour the arsenite and AsB overlap was prevented by placing anionic cartridges up-stream of the anion-exchange column. Initial experiments in which 100 pl of each arsenic species at 12Opg1-' were injected into the system through anionic cartridges showed that the anionic species arsenite arsenate MMA and DMA are quantitatively retained in the cartridges while the cationic species AsC and AsB (or neutral depending on pHj break through.To optimize the chromatographic AsB and AsC separation in the isocratic mode the phosphate buffer was varied in the ranges 1-20 mmol l-l pH 4-8 and flow rates 0.5-2 ml min-'. Optimum chromatographic resolution was obtained at 5 mmol 1-' pH 6.0 and a flow rate of 1 ml min-'. Fig. 2 shows a typical chromatogram when the six standard species are injected through the anionic cartridge. Only the AsB and AsC broke through the cartridge giving the corresponding signals. Fig. 3 shows the chromatogram for the injection of the six arsenic species onto the HPLC anionic column after removing the anionic cartridge.Only the arsenite DMA MMA and arsenate peaks were obtained because AsB and AsC do not generate arsine in the absence of an oxidizing agent.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 293 0 2.0 4.0 6.0 Tim e/rn i n Fig.2 Chromatographic separation of AsB and AsC using anionic cartridge-HPLC-microwave oxidation-HG-AAS ( 120 ng ml- ' of each six arsenic species were injected injection volume = 100 pl) 0.15 1 0 2.0 4.0 6.0 8.0 Time/m in Fig. 3 of As injection volume 100 pl) using HPLC-heat-HG-AAS Chromatographic separation of arsenic species (120 ng ml- ' Microwave-oxidation Parameters The experimental conditions14 used in on-line AsB and AsC thermo-decomposition were adapted to microwave oxidation which was studied separately for each species using an FI system.The effect of residence time on microwave conversion efficiency at maximum microwave power (700 W) is shown in Fig. 4. Arsenobetane (100 ng ml-') and 100 ng ml-' of AsC 150 A s g 120 - 2. a 0 .- c 9c n $ 60 .- +- a (0 $ 30 E 0 .- 0 C / / I / / I / B 'v 7' 6.0 8.0 10.0 12.0 14.0 16.0 Residence ti rn e/s Fig. 4 species ( 100 ng ml- of As) A arsenate; B AsB; and C AsC Microwave oxidation efficiency uersus residence time of arsenic were injected separately into the chromatographic carrier of phosphate buffer (17 mol l-' pH 6.0 and flow rate 1 ml min- ') and run with 3% persulfate into the microwave oven. Fig. 4 shows that AsC behaves like arsenate i.e. AsC is oxidized to an arsine generator species in a few seconds while AsB is fully converted after 11 s.At a constant microwave power of 700 W and 1.5 m of reaction coil the percentage conversion of both species varies with the concentration of potassium persulfate as shown in Fig. 5. At 1 YO K2S208 no AsB signal was obtained while conversion of AsC was 60% as arsenate. Above 3% of potassium persulfate HG efficiency for both species was 100% as arsenate. In order to identify the decomposition products of the six arsenic species after microwave oxidation at three concen- trations of K2S20s the effluent solution from the microwave oven was collected and injected into the chromatographic column for separation and determination by HPLC-HG-AAS. The results are shown in Table 1. At 1% persulfate solution all organoarsenic species remain in the original form except that AsC is transformed mostly into arsenite and partially into DMA. These results agree with Fig.5 which indicates that AsB does not generate arsine and that the efficiency of AsC conversion is less than 100%. A 3% m/v concentration of K2S208 transforms all compounds into arsine generator species. The transformation of AsB was about 65% to arsenate and 30% to MMA and AsC was converted quantitatively into arsenate. Treatment with 5% K2S208 converts all organic species into arsenate. This concentration was chosen for further experiments. On-line microwave oxidation is as feasible as the previously proposed on-line thermo-oxidation system12 for decomposing organoarsenic compounds into arsine generator species. However the coil length necessary for on-line organoarsenical microwave oxidation is much shorter (1.5 m) than for thermo- oxidation (4.5 m) and microwave power has the advantages of high efficiency fast decomposition and ease of operation.Analytical Performance The retention times detection limits relative standard devi- ations (RSD) and calibration parameters of each species are listed in Table2. At the 120ngml-' level the RSDs were 3-5% (n = 5). These analytical characteristics make the method very promising for application to most environmental samples. Under the conditions given under Analytical Procedure each peak is completely separate. The chromatogram run time is 3.5 min for the determination of AsB and AsC and 8 min for arsenite arsenate MMA and DMA. The different slopes are due to the different peak widths of all species except DMA - 2.100 - C e 0 .- $ 7 5 - 8 0 .- 4- f 50- c a m Fig. 5 Hydride generation efficiency versus concentration of K,S,O (100 ng ml-' of As) A arsenate; B AsB; and C AsC294 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Table 1 coil length. Products identified by HPLC-microwave oxidation-HG-AAS; results given in YO On-line microwave oxidation products of 100 ng of organoarsenic compounds at three different concentrations of persulfate and 1.5 m Species obtained* Species injected AsC As"' AsB DMA MMA AsV Persulfate concentration 1 3 5 1 3 5 1 3 5 1 3 5 1 3 5 1 3 5 AsV - 98 98 100 100 100 63 95 59 91 71 97 100 100 100 - - - ~ *Note AsC not found. Table 2 Analytical characteristics and calibration parameters for arsenic species; n = 10 ~ ~ _ _ _ _ _ _ _ Species Retention time/s DL*/ng RSD (%) Slope Intercept Correlation coefficient AsC 2.20 0.3 4 0.0090 0.002 0.9994 AsB 3.00 0.4 5 0.0071 0.010 0.998 DMA 2.45 0.9 5 0.0035 0.006 0.992 MMA 3.70 0.4 4 0.0069 0.003 0.9999 AsV 7.20 0.6 4 0.0047 0.001 0.9998 As"' 1.85 0.3 3 0.0099 0.0 13 0.9999 *Detection limit =& f 30.which is only 64% as efficient as an equivalent amount of the other arsenic species. Application The proposed HPLC-microwave oxidation-HG-AAS method was used to determine six arsenic species in mineral water sewage water harbour sea-water synthetic fish extract and sediment extract (Table 3) by calculating their concentration using the respective calibration curve of each species.The spike recovery data for 100ngml-' of each arsenic species in synthetic fish samples are acceptable and confirm the reliability of the proposed method. Total content of As in sewage was determined by ICP-AES by another research institute [Centro de Investigaciones EnergCticas Medioambientales y Tecnologicas (CIEMAT)] and the As found (36+4 mg I-') was similar to the sum of the different species determined by the proposed method. The concentration of AsB found in the harbour sea-water and in the sewage sea-water are unusual. The concentration of AsB in other sea-water samples was lower than the detec- tion limit. Conclusions The proposed on-line HPLC-microwave oxidation-HG-AAS system has been successfully used to determine arsenite arsen- ate MMA DMA AsC and AsB in drinking water sewage and harbour sea-water synthetic fish extract and sediment extract.The novel coupling of anionic cartridges to the HPLC column is of great interest and makes it possible to determine the six arsenic species without the chromatographic overlap- ping of arsenite and AsB. The microwave oxidation system developed is a good alternative to photo-oxidation and thermo-oxidation. The feasibility of thermo-oxidation and microwave oxidation for the decomposition of the organoarsenicals AsB and AsC is Table 3 Speciation of As in waters synthetic fish extract and synthetic sediment extract. Results are expressed as As; nks (n=5) Fish extract*/ Species pg 1-' As"' n.d. 7 AsV 347 f 20 MMA 11Of8 DMA 140+10 AsC 106 + 7 AsB 450 + 30 Recovery in fish (spiked samples %) 110 96 105 108 95 108 Sediment extract*/ n.d.34f 1 18.2f0.8 8.3 k0.3 n.d. n.d. I % I-' Mineral watert/ I-' n.d. 51 + 6 n.d. n.d. n.d. n.d. Sewage water$/ mg 1-' 7.4 f 0.2 n.d. n.d. n.d. n.d. 30f2 Sea-water (harbour&/ n.d. n.d. n.d. n.d. n.d. 47f1 I-' *Provided by the Bureau of Community Reference Material of the European Communities. tCommercially available. $From CIEMAT (Spanish research centre). §From Santander harbour Spain. 1n.d. =Not detected.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 295 similar and the conversion efficiency was close to 100% for the species tested when 5% K2S20 was used. The proposed on-line system is efficient reproducible simple fast cheap and free from cross-contamination. The authors thank Isabel Martin for her collaboration Direccion General de Investigacion Cientifica y Tkcnica (Project PB91-0376) for financial support Max Gormann for revision of the manuscript and BCR for providing reference materials.References 1 Buchet J. P. and Lauwerys R. in Analytical Techniques for Heavy Metals in Biological Fluids ed. Facchetti S. Elsevier Amsterdam 1981. p.75. 2 Lewis R. J. and Tatken R. L. Registry of Toxic Efects of Chemical Substances US Department of Health Education and Welfare Cincinnati OH USA 1978. 3 Marafante E. Vahter M. and Dencker L. Sci. Total Enuiron. 1984 34 223. 4 5 6 7 8 9 10 11 12 13 14 15 Beauchemin D. Siu K. W. M. McLaren J. W. and Berman S . S. J. Anal. At. Spectrom. 1989 4 285. Low G. K. C. Batley G. E. and Buchanan S . J. J. Chromatogr. 1986,368,423. Cullen W. R. and Dodd M. Appl. Organornet. Chem. 1989,3,401. Murer A. J. L. Abildtrup A. Poulsen 0. M. and Christense J. M. Analyst 1992 117 677. Branch S. Bancroft K. C . C. Ebdon L. and O’Neill P. Anal. Proc. 1989 6 73. Shibata Y. and Morita M. Anal. Chem. 1989 61 2116. Rauret G. Rubio R. and Padro A. Fresenius’ J. Anal. Chem. 1991,340 157. Atallah R. M. and Kalman D. A. Talanta 1991 38 167. Tsalev D. L. Sperling M. and Welz B. Analyst 1992 117 1729. Tsalev D. L. Sperling M. and Welz B. Analyst 1992 117 1735. Lopez M. A. Gomez M. M. Palacios M. A. and Camara C. Fresenius’ J . Anal. Chem. 1993 346 643. Le X.-C. Cullen W. R. and Reimer K. J. Appl. Organomet. Chem. 1992 6 161. Paper 3/05883 B Received July 13 1993 Accepted September 29 1993