Analyst, December, 1983, Vol. 108, $p. 1495-1499 1495 Anion-exchange Method for Speciation of Arsenic and Its Application to Some Environmental Analyses John Aggett and Rabiya Kadwani Chemistry Department, University of Auckland, Auckland, New Zealand A relatively straightforward two-stage anion exchange method has been developed for the speciation of inorganic arsenic(V) and arsenic(III), mono- methylarsonic acid and dimethylarsinic acid. Arsenic(II1) and dimethyl- arsinic acid are eluted in succession with a carbon dioxide - hydrogen carbonate buffer a t pH 5.5. Separation of monomethylarsonic acid and inorganic arsenic(V) is then obtained by elution with carbon dioxide - sodium chloride solution a t pH 4.0-4.2. The method has been applied to the analysis of sediment interstitial waters, and three species of aquatic plants.Keywords : Anion-exchange method ; arsenic speciation ; environmental analysis In recent years several methods have been published for the separation and analysis of environ- mentally important arsenic species. In these methods separation has been based on selective generation of arsines,l92 sequential volatilisation of generated ar~ines,~-~ conventional ion- exchange chrornatography,6-11 high-performance liquid chromatography12 and ion chromato- graphy.12913 Analyses have been completed through the use of hydride gene~ation~y~s~s~~ and graphite furnace6,8,10-12 atomic-absorption spectroscopy, diff erential-pulse polar~graphy,~ d.c. discharge3 and microwave emi~sion.~ Although speciation by methods based on hydride generation is attractive for those systems to which it is applicable it is of course limited to the determination of those species which can be converted into volatile hydrides.Fortunately, the four species most commonly considered to be of environmental importance at the present time, i e . , arsenate, arsenite, monomethylarsonic acid, and dimethylarsinic acid are amenable to this form of analysis. However, conventional ion-exchange and ion-chromatographic methods appear to possess the potential advantage that it should be possible to extend or modify them to include the analysis of additional environmentally important arsenic species should that become necessary. The purpose of this paper is to report the development and application of a relatively simple anion-exchange method for the speciation of arsenate, arsenite, monomethylarsonic acid and dimethylarsinic acid.As these four arsenic species are weak acids the dissociation constants of which are quite different (Table I) it seemed that separation by anion-exchange chromato- graphy was both logical and possible, However, the first two published ion-exchange methods for speciation of arsenic697 both used cation-exchange chromatography. The mechanism for these separations has not been established and attempts in this laboratory to use them to separate inorganic arsenic( 111) and inorganic arsenic(V) were not successful. Subsequently, Henry and Thorpeg and other workers10911 have published methods that require the use of both cation-and anion-exchange columns.TABLE I DISSOCIATION CONSTANTS OF ARSENIC SPECIES Acid Pkai Pka2 pka, Arsenic acid .. .. . . 2.25 7.25 12.30 Arsenious acid . . .. . . 9.23 Monomethylarsonic acid . . . . 4.26 8.25 Dimethylarsinic acid . . . . 6.25 The method presented here is a two-stage single column anion-exchange method using hydrogen carbonate and chloride as eluate anions. These species appear to have no adverse effects in subsequent analytical procedures. Its successful application is dependent on careful control of pH. It has been applied to the speciation of arsenic in samples obtained during studies on the fate of arsenic released into the Waikato River from geothermal sources in its catchment. For developmental purposes analyses were performed by hydride generation atomic-absorption spectroscopy, although any extension to more general speciation would1496 AGGETT AND KADWANI : ANION EXCHANGE FOR SPECIATION Analyst, VoZ.108 require the use of a more general technique for analysis, such as graphite furnace atomic- absorption spectroscopy or inductively coupled plasma atomic-emission spectroscopy. Experimental Anion-exchange Resins and Procedures Four anion-exchange materials were chosen for the initial investigation, viz. Zerolit FF(ip) SRA 62 (BDH Chemicals Ltd.), Zerolit FF(ip) SRA 70 (BDH Chemicals Ltd.), Amberlyst A-26 (Aldrich) and Whatman DE52 (Whatman). Columns (25 cm long) were prepared in 25-ml burettes using small glass-wool pads to support the resin. In some of these experiments the length of the Whatman DE52 column was reduced to 12 cm.The columns were converted into the hydrogen carbonate form and then prepared for use by elution with the appropriate buffer until the effluent reached the required pH. They were then loaded with 10-ml aliquots of the solutions under study. These solutions normally contained 1 pg ml-l of arsenic. During the elution procedure 5-ml aliquots were collected with an ISCO, Model 273, fraction collector. These aliquots were subsequently analysed by a hydride generation atomic- absorption method1 in which generation was carried out in 1 moll-1 hydrochloric acid. Separ- ate standards were used for the different compounds. Ekates As theoretical considerations suggested that separations were most likely to be achieved where the pH of the eluate was between 4 and 7 three buffer media were investigated as possible eluates, viz., citrate, acetate and carbon dioxide - hydrogen carbonate. Carbon dioxide - hydrogen carbonate buffers were prepared by saturating de-ionised distilled water with carbon dioxide.The solution was left overnight to ensure that equilibrium solubility of carbon dioxide was achieved and then solid sodium hydrogen carbonate was added until the desired pH was reached. The citrate and acetate buffers were prepared by dissolving the appropriate amounts of analytical-reagent grade materials in de-ionised distilled water. The carbon dioxide - ammonium chloride eluate was prepared by dissolving ammonium chloride (10 g 1-l) in saturated carbon dioxide solution. The pH of this was in the range 4.04.2. Speciation in interstitial water from sediments Sediment samples were packed into 50-ml centrifuge tubes immediately after collection and the tubes sealed to prevent intrusion of air.Interstitial water was obtained by centrifuging the samples in a refrigerated centrifuge for 15 min at 10000 rev min-l. This was normally done within 4 h of sample collection. The interstitial water was preserved at pH 2 with hydrochloric acid until speciation was carried out. Immediately prior to speciation the pH of the samples was adjusted to 5.0-5.5 and trace amounts of iron(II1) were removed by extrac- tion into chloroform with acetylacetone; 10-ml aliquots were speciated. Iron(II1) was removed prior to speciation to avoid the possibility that at the pH of speciation it might precipitate as hydrous iron(II1) oxide and in doing so coprecipitate arsenic.Speciation in lakeweeds For speciation of arsenic in lakeweeds, fresh samples of lakeweed (100 g) were cut into small pieces and homogenised with 500 ml of de-ionised water. They were then centrifuged in a refrigerated centrifuge for 1 h at 15000 rev min-l. The supernatant liquid and residue were separated and stored at 4 "C until required for analysis. The supernatant liquid was passed through a 0.45-pm membrane filter immediately before speciation. The solid material obtained by centrifuging the lakeweeds was refluxed with methanol and ethanol in separate procedures. The extracts were evaporated to dryness and the residues re-dissolved in de-ionised water and speciated. Results and Discussion Preliminary experiments were conducted with various combinations of resins and eluates in the pH range 5-7.In these the only promising separations were obtained with carbon dioxide-hydrogen carbonate as the eluate and this was used as the basis for subsequent development .December, 1983 OF As AND APPLICATION TO ENVIRONMENTAL ANALYSES 1497 The behaviour of arsenic(II1) and dimethylarsinic acid on SRA 70 as a function of the pH of the carbon dioxide - hydrogen carbonate buffer is shown in Fig. 1. Neither monomethyl- arsonic acid nor arsenic(V) was eluted by 250 ml of eluate in the pH range 4.8-6.4. These observations are understandable in terms of the dissociation constants of the respective acids and the nature of the buffer solution used for elution. Arsenic(II1) exists as an undissociated molecule over the pH range studied and as a conse- quence is eluted rapidly in a manner independent of pH.At lower pH the elution of dimethyl- arsinic acid overlaps that of arsenic(II1) but as the pH is raised there is greater tendency for dimethylarsinic acid to be retained by the resin, presumably a consequence of dissociation. This is reversed somewhat in the region of pH 6.0-6.4. The cause of this is believed to be the increase in hydrogen carbonate concentration in the eluent in this pH region. In practice, the hydrogen carbonate concentration increased from about 6 x mol 1-1 at pH 5.4 to about 6 x mol 1-1 at pH 6.4. T 2 pH 6.4 pH 4.8 10 20 30 Volume of aliquotlml Fig. 1. Separation of arsenic(II1) and dimethylarsinic acid on SRA70 by elution with carbon dioxide - hydrogen carbonate as a function of pH.1, Arsenic(II1) ; 2, dimethylarsinic acid. The results showed that separation of arsenic(II1) and dimethylars.inic acid by elution with carbon dioxide - hydrogen carbonate was satisfactory in the pH range 5.2-6.0. Elution of monomethylarsonic acid was accelerated and satisfactory separation from arsenic(V) achieved using saturated aqueous carbon dioxide (pH 4.04.2) containing 10 g 1-1 of ammonium chloride. Thus complete separation of the four species was obtained by eluting first with 150 ml (30 x 5-ml aliquots) of carbon dioxide - hydrogen carbonate buffer a t pH 5.5 & 0.3 followed by elution with a further 150 ml(30 x 5-ml aliquots) of carbon dioxide - ammonium chloride buffer solution at pH 4.04.2.Although the pH range for successful separation of arsenic( 111) from dimethylarsinic acid appears small, no problems have been encountered in meeting that specification during subsequent speciation procedures. Similar results were obtained with the other resins although the optimum pH for separation of arsenic( 111) from dimethylarsinic acid with carbon dioxide - hydrogen carbonate eluate varied: SRA 62, 6.0; SRA 70,5.5; A-26,5.2; and DE52, 6.0. The SRA 70 was preferred as it appeared to provide slightly better separation between arsenic(II1) and dimethylarsinic acid1498 AGGETT. AND KADWANI: ANION EXCHANGE FOR SPECIATION Analyst, V d . 108 TABLE I1 RECOVERIES OF ARSENIC SPECIES ON TREATED SRA 70 Species Mean recovery, yo* Arsenic(II1) .. .. .. 97.0 Arsenic(V) . . .. .. 98.1 Monomethylarsonate . . .. 98.8 Dimethylarsinate . . .. 94.8 * Four samples analysed. than did the others. In addition, peaks obtained with SRA 62, which has a lower percentage of cross-linking than SRA 70, were more inclined to tail. Elution was normally carried out at a flow-rate of 1 ml min-l. Increasing this to 4 ml min-1 did not appear to affect the separations. Initial investigation into quantitative aspects revealed a significant problem, i.e., that arsenic(II1) was oxidised to arsenic(V) during elution. This was indicated by the 70-80% recoveries for arsenic(II1) associated with the elution of low concentrations of arsenic in aliquots 35-45 when arsenic(II1) was eluted on its own. Although the cause of this problem was not positively identified the problem was removed when resins were treated with nitric acid (1 moll-l) and ethylenediaminetetraacetic acid (0.1 mol l-l, pH 5) before use. Recover- ies obtained with resin treated in this way are shown in Table I1 and the chromatogram of a synthetic mixture of the four species in Fig.2. The method has been applied to the analysis of the interstitial waters of sediments in Lake Ohakuri, and also to three species of lakeweed, Egeria densa, Largarosiphon major and Cerato- phyllum demersum, from the same lake as part of an investigation into the fate of arsenic released into the aquatic system from geothermal sources in the Waikato River catchment. No significant concentrations of methylated arsenic species were found in the interstitial waters examined and a typical sample (collected 7th July 1982) contained 0.52 pg ml-l of arsenic(II1) and 0.50 pg ml-l of arsenic(V).Chromatograms of the supernatant liquids obtained from centrifuging the lakeweed homo- genates showed that the two major arsenic species were inorganic arsenic(II1) and arsenic(V). No dimethylarsinic acid was detected in any of the supernatant samples analysed. It is possible that small amounts of monomethylarsonate (less than 3% of total arsenic) were present as very small concentrations of arsenic were detected in aliquots 3545 in most samples. However, no maxima were observed in the vicinity of aliquot 36, which suggests that these As (Ill 1 u 10 DMA 20 30 II I MMA I 40 50 L 60 Volume of a I iq u ot/m I Separation of a synthetic mixture of arsenic(II1) (0.5 pg), dime'hylarsinic acid (0.5 pg), monomethylarsonic acid (0.5 pg) and arsenic(V) (1 pg) on SRA7O.1499 December, 1983 OF As AND APPLICATION TO ENVIRONMENTAL ANALYSES 10(1 L .10 20 30 40 50 60 Volume of aliquotlml Fig. 3. Speciation of arsenic extracted by methanol from the solid phase of Largarosiphon major produced by centrifugation of the homo- genate. small peaks could be the result of oxidation of arsenic(II1) during elution. Attempts to clarify the situation by chromatographing samples at 4 "C were not successful. Only one of a number of solid phase lakeweed samples speciated appeared to contain a significant fraction of monomethylarsonic acid and none gave any indication of the presence of dimethylarsinic acid.The chromatogram of the Largarosiphon major sample containing monomethylarsonic acid is shown in Fig. 3. These applications indicate the usefulness of the method. By comparison with other ion- exchange methods for speciation of arsenic it is relatively simple and requires the use of eluates that are unlikely to interfere with subsequent quantitative analyses. Recoveries of the individual compounds are adequate and the only significant limitation appears to be associated with the problem of distinguishing between monomethylarsonic acid present in very small amounts and arsenic(V) formed by oxidation of arsenic(II1) on the ion-exchange column. At a flow-rate of 4 ml min-1 the ion-exchange procedure takes 75 min and it should be possible to make the procedure semi-automatic by coupling the ion-exchange system to the hydride generation system or other alternative analytical systems.Although dimethylarsonic acid was found in lakeweeds, dimethylarsenic acid might be found by other methods. Different digestion methods are at present being investigated for their effectiveness. The authors are grateful for the assistance of NZ Electricity Department personnel in collection of samples and Dr. E. White and Dr. M. H. Timperley, DSIR, for access to the centrifuge for separating interstitial waters from sediment solid phases. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Aggett, J . , and Aspell, A. C., Analyst, 1976, 101, 341. Howard, A. G., and Arbab-Zavar, M. H., Analyst, 1980, 105, 338. Braman, R. S., Johnson, D. L., Foreback, C. C., Ammons, J . M., and Bricker, J. L., Anal. Chem., Andreae, M. O., Anal. Chem., 1977, 49, 820. Talmi, Y., and Bostick, D. T., Anal. Chem., 1975, 47, 2145. Yamamoto, M., Soil Sci. SOC. Am. Proc., 1975, 39, 859. Dietz, E. A., and Perez, M. E., Anal. Chem., 1976, 48, 1088. Iverson, D. G., Anderson, M. A., Holm, T. R., and Stanford, R. R., Environ. Sci. Technol., 1979,13, Henry, F. T., and Thorpe, T. M., Anal. Chem., 1980, 52, 80. Pacey, G. E., and Ford, J. A., Talanta, 1981, 28, 935. Grabinski, A. A., Anal. Chem., 1981, 53, 966. Brinckman, F. E., Jewett, K. L., Iverson, W. P., Irgolic, K. J., Ehrhardt, K. C., and Stockton, Ricci, G. R., Shepard, L. S., Colovos, G., and Hester, N. E., Anal. Chem., 1981, 53, 610. 1977, 49, 621. 1491. R. A., J . Chromatogr., 1980, 191, 31. Received July 4th, 1983 Accepted August lst, 1983