Membrane solubilization with tetramethylammonium hydroxide for the preconcentration and electrothermal atomic absorption spectrometric determination of trace amounts of arsenic in water Noriko Hata,* Hiromi Yamada, Issei Kasahara and Shigeru Taguchi Faculty of Science, Toyama University, Toyama 930-8555, Japan. E-mail: noriko@sci.toyama-u.ac.jp Received 2nd September 1998, Accepted 11th November 1998 Solubilization of a mixed cellulose ester membrane filter (MF) with tetramethylammonium hydroxide (TMAH) is proposed for the preconcentration and electrothermal atomic absorption spectrometric (ETAAS) determination of trace amounts of arsenic in water.Arsenic at not more than 0.4 mg in 100 ml of sample solution was retained on the MF by filtration as an ion associate of arsenomolybdate and tetraphenylphosphonium ions.The ion associate was dissolved in a small volume of TMAH together with the MF. After being made up to 2 ml with water, the arsenic in the concentrate was determined by ETAAS in the presence of zirconyl nitrate as a chemical modifier.This method is very simple and rapid. The detection limit, defined as three times the standard deviation of the blank, was 0.04 mg l21. Inorganic matrix components in river waters, Na+, K+, Ca2+, Mg2+, SO4 22, NO3 2 and silicate at high concentrations did not interfere with the determination. Trace components, phosphate at 0.3 mg l21, dodecyl sulfate at 14 mg l21 and aluminium at 5 mg l21, also did not interfere with the determination.The proposed method was applied to the analysis of river water samples. Electrothermal atomic absorption spectrometry (ETAAS) is widely applied for trace analysis because of its high sensitivity and the requirement for only small amounts of samples. In environmental trace analysis, however, sometimes its sensitivity is insufficient and interference from matrix components also causes serious problems. In most such cases preconcentration techniques such as solid-phase extraction and solvent extraction are employed to concentrate and separate the analyte from the matrices before instrumental analyses. We proposed a soluble membrane filter (MF) technique for solid-phase extraction of trace elements in water.1 In this technique, the analyte was converted into hydrophobic species and the species was retained on an MF by filtration, then the collected material was dissolved in a small volume of organic solvent or sulfuric acid together with the MF.This technique is simple, rapid and versatile and has been applied to the spectrophotometric determination of trace amounts of phosphate,1,2 ETAAS determination of cadmium,3 copper4 and chromium5 and inductively coupled plasma atomic emission spectrometric (ICP-AES) determination of arsenic.6 Most of the MF solubilization methods are associated with sulfuric acid3,5,6 or organic solvents such as N,N-dimethylformamide (DMF),4 dimethyl sulfoxide (Me2SO)1 and 2-methoxyethanol.2 In our preliminary experiments on the application of the soluble MF technique to the ETAAS determination of arsenic, solubilizers such as DMF, Me2SO and 2-methoxyethanol, which have been successfully applied in previous studies, were not suitable for the ETAAS determination of arsenic, because white curds of the material of the mixed cellulose ester MF were precipitated when a chemical modifier aqueous solution was added to the concentrate.Probably in these solvents the material of the MF is dissolved as the original polymer and is not decomposed to small molecules. Therefore, the addition of an aqueous solution of a modifier to the polymer solution decreased the solubility of the material of the MF and the material was precipitated. Concentrated sulfuric acid, which dissolves the mixed cellulose ester MF, can be diluted with an aqueous solution to prevent the precipitation, but its viscosity is extremely high.In our preliminary experiments it was found that a small volume of tetramethylammonium hydroxide (TMAH) dissolves a mixed cellulose ester MF quickly and does not give viscous solution. TMAH has been used as a ‘tissue solubilizer’ for various zoological samples prior to analysis for minor inorganic elements by ETAAS.7 In this work, TMAH was studied as a solubilizer for MF and the solubilization technique was applied to preconcentration and ETAAS determination of trace amounts of arsenic in water.Experimental Apparatus A Hitachi (Tokyo, Japan) Model Z-8000 flame and graphite furnace atomic absorption spectrometer equipped with a Zeeman-effect background corrector and an optical temperature control system (Hitachi Model 180-0342) was used. Sample solution (20 ml) was injected by an autosampler. The analytical wavelength and slit width were 193.7 and 1.3 nm, respectively. The electric current of the hollow cathode lamp was 17.5 mA. The lamp current was set according to the manufacturer’s instrumental manual.A graphite tube cuvette was used as the furnace. The argon gas flow rate was 200 ml min21 except for atomization, where it was 30 ml min21. The optimum furnace operating conditions for the determination of arsenic used in this study are given in Table 1. Reagents All the chemicals were analytical-reagent grade or of the highest purity available. They were used as received. Analyst, 1999, 124, 23–26 23Molybdate reagent solution.Dissolve 18 g of sodium molybdate dihydrate (Na2MoO4·2H2O) (Wako, Osaka, Japan) in water, add 100 ml of concentrated sulfuric acid, then dilute to 400 ml with water. Tetraphenylphosphonium bromide solution, 0.02 mol l21. Dissolve 0.84 g of tetraphenylphosphonium bromide [(C6H5)4PBr] (Tokyo Chemical Industry, Tokyo, Japan) in 100 ml of water. TMAH solution (25%, TAMAPURE AA grade). This was purchased from Tama Chemicals (Tokyo, Japan) and was diluted to 12.5% with water before use.Zirconyl nitrate solution, 0.02 mol l21. Dissolve 0.53 g of zirconyl nitrate dihydrate [ZrO(NO3)2·2H2O] (Wako, extra pure grade) in 100 ml of water. Membrane filter An Advantec Toyo (Tokyo, Japan) A045A025A MF (25 mm in diameter, 0.45 mm pore size, mixed cellulose ester) was used. Mixed cellulose ester MFs of different brands may also be used. An Advantec Toyo KG-25 filter support (effective filtration area 2.0 cm2) was used as a filter support.Recommended procedure 1. Determination of inorganic AsV (arsenate) (Procedure A). Place 100 ml of a sample solution containing not more than 0.4 mg of AsV in an Erlenmeyer flask. Poor linearity for ETAAS determination of arsenic was obtained with more than 0.4 mg of arsenic. Add 8 ml of the molybdate reagent solution and set aside for 5 min, then add 2 ml of tetraphenylphosphonium bromide solution, set aside for at least 2 min and pass the solution through an MF by filtration under suction, collecting the arsenomolybdate as its ion associate with a tetraphenylphosphonium cation.Wash the MF twice with about 5 ml portions of water. Remove the MF from the holder and place it in a 10 ml beaker containing 0.4 ml of 12.5% TMAH solution. Heat the beaker at about 100 °C for 1 min on an electric heating plate with swirling to dissolve the MF completely. Alternatively, without heating, allow the beaker to stand for at least 6 h to dissolve the MF in 0.4 ml of 12.5% TMAH solution with swirling at intervals.Add 1.2 ml of 0.02 mol l21 zirconyl nitrate solution and dilute to 2 ml with water. Inject 20 ml of the solution into the cuvette with an autosampler and measure the absorption at 193.7 nm. 2. Determination of total inorganic arsenic [arsenite (AsIII) plus arsenate (AsV)] (Procedure B). Place 100 ml of a sample solution containing not more than 0.4 mg of arsenic in an Erlenmeyer flask. Before addition of 8 ml of molybdate solution as in Procedure A, add and dissolve 1 g of potassium peroxodisulfate to oxidize arsenite to arsenate. Arsenite is oxidized to arsenate immediately after dissolving potassium peroxodisulfate.Heating should be avoided in this procedure. Follow Procedure A from ‘Add 8 ml of the molybdate reagent solution . . .’. 3. Determination of total arsenic (Procedure C). Place 100 ml of a sample solution containing not more than 0.4 mg of arsenic in an Erlenmeyer flask. Prior oxidation and decomposition of organoarsenic species to arsenate can be accomplished by heating for 20 min after addition of 1 g of potassium peroxodisulfate before addition of 8 ml of molybdate solution as in Procedure A.Follow Procedure A from ‘Add 8 ml of the molybdate reagent solution . . .’. Results and discussion Formation and collection of arsenomolybdate The conditions for the formation of arsenomolybdate6 are similar to those reported by Wadelin and Mellon,8 except that sulfuric acid is used in place of hydrochloric acid.The choice of a counter ion is very important for the quantitative collection of the ionic species. In our previous study we found that both tetrapentylammonium and tetraphenylphosphonium are effective and successful for the collection of arsenomolybdate.6 In this work, tetraphenylphosphonium cation was applied. Material of MF and solubilizer A solubilizer applicable to the MF solubilization and ETAAS method for the determination of trace amounts of arsenic was investigated.Organic solvents, in which the materials of MF are dissolved as an original polymer, can hardly be diluted with water. Both hydrochloric acid and nitric acid, which dissolve polyamide MF, can also hardly be diluted with water for a similar reason. As stated above, concentrated sulfuric acid was not suitable because of its extremely high viscosity. Although both TMAH solution and sodium hydroxide solution in which mixed cellulose ester MF dissolves can be diluted with water, the ETAAS signal for arsenic in TMAH solution was sharper and more reproducible than that in sodium hydroxide solution.Eventually, among the materials of MF and the solubilizer tested, the combination of mixed cellulose ester and TMAH solution was selected as the best. A piece of mixed cellulose ester MF (25 mm in diameter) was dissolved and decomposed in 0.3 ml of 12.5% or 0.2 ml of 25% TMAH solution by heating for about 1 min.In the range 2.5–6.3% of TMAH in the concentrate, a constant absorbance in ETAAS determination was obtained. Chemical modifier Chemical modifiers have been recommended to improve the AAS sensitivity.9 Twenty metal salts were investigated as chemical modifiers for the preconcentration and ETAAS determination of arsenic. Cobalt nitrate, nickel nitrate, zinc nitrate and potassium permanganate were not suitable because they gave precipitates when their solutions were added to the TMAH solution for dissolving the mixed cellulose ester MF.Zirconyl nitrate, chromium(iii) nitrate, iron(iii) nitrate, palladium chloride, ammonium tungstate, ammonium vanadate(v), copper(ii) sulfate and lead nitrate solutions did not give precipitates in the TMAH solution for dissolving the MF and Table 1 Temperature–time programme optimized for the determination of arsenic Temperature/°C Stage No. Stage Start End Ramp or hold time/s 1 Drying 80 120 30.0 2 Ashing 400 800 10.0 800 800 20.0 3 Atomization 2800 2800 10.0 4 Cleaning 2900 2900 3.0 24 Analyst, 1999, 124, 23–26were effective in enhancing the sensitivity.Although some metals such as iron(iii) produced precipitates with pure TMAH solution they did not produce precipitates in the TMAH solution for dissolving the MF. Among the metal salts tested, zirconyl nitrate, chromium(iii) nitrate and iron(iii) nitrate gave comparable and excellent results. The effects of their concentrations on the arsenic signal are shown in Fig. 1. All of these three nitrates could be used successfully and their effects were almost the same. Zirconyl nitrate was adopted in this study as a chemical modifier. Ashing Temperature The influence of the ashing temperature on the signal was studied in the range 400–1200 °C. When zirconyl nitrate was used as a chemical modifier the maximum absorbance was independent of ashing temperature over the range 400–900 °C. The ashing temperature adopted in the ETAAS determination was 800 °C.Effect of foreign substances Interference from aluminium, sodium, potassium and sulfate in the direct ETAAS determination of arsenic has been reported.10 However, as previously reported in the ICP-AES determination of arsenic,6 most of spectrally interfering ions are eliminated by the MF retention procedure. Table 2 shows the effects of foreign substances on the determination of arsenic. Concentrations of 5 mg l21 of aluminium, 2.0 mol l21 of sodium, 0.1 mol l21 of potassium and 1.0 mol l21 of sulfate did not interfere.Interference from phosphate in the determination of arsenic without separation has been reported.11,12 In this study, phosphate up to 0.3 mg PO4 l21 did not interfere. In our experience, in most river waters not strongly polluted, the concentrations of the phosphate were below this limit. At phosphate levels higher than the limit, this method was not suitable unless the sample solution was diluted.Phosphate also formed a heteropolymolybdate and was collected on the mixed cellulose ester MF under the same conditions as arsenate. A sample solution containing 1 mg l21 of phosphate took longer than 8 min to be filtered and the mixed cellulose ester MF with collected heteropolymolybdates did not dissolve completely in 0.4 ml of 12.5% TMAH solution. Reproducibility The correlation of the calibration graphs based on peak height (correlation coefficient = 0.999) was better than that based on peak area (correlation coefficient = 0.994), and the peak height was more reproducible than the peak area.Peak height was therefore adopted. At a concentration of 0.5 mg l21 As, the relative standard deviation (RSD) was 8.8%, at 1.0 mg l21 the RSD was 6.3%, at 2.0 mg l21 the RSD was 4.9% and at 4.0 mg l21 the RSD was 2.9% (n = 6). These are mainly due to the instrumental errors in measurement. The blank was 0.016 absorbance units (peak height) and the RSD was 12.4%.The detection limit, defined as three times the standard deviation of the blank, was 0.04 mg l21 (n = 7). Determination of arsenic in natural water samples The acute toxicity of arsenic species decreases in the order arsenite (AsIII) > arsenate (AsV) > > dimethylarsinic acid (DMAA) > monomethylarsonic acid (MAA).13 Trimethylarsine oxide (TMAO) and arsenobetaine (AB) are considered to be relatively non-toxic. The acute toxicities of organoarsenic species are several orders of magnitude less than those of inorganic arsenic (arsenite and arsenate).With respect to the acute toxicity, at least arsenite and arsenate should be determined. Table 3 gives the analytical results of arsenic in natural water samples. Using Procedure A, only arsenate (AsV) formed arsenomolybdate and was collected on the MF and was determined by ETAAS. Therefore, arsenate is identical with the value obtained from Procedure A. Arsenite (AsIII) was oxidized to arsenate by addition of peroxodisulfate.AsIII is given by subtraction of the value obtained by Procedure A from that obtained by Procedure B. The decomposition of methylated arsenic, DMAA, with peroxodisulfate was investigated. Although the addition of 1 g of potassium peroxodisulfate did not decompose DMAA to arsenate, heating for 10 min after addition of the reagent gave arsenate. In natural water, arsenic species are mainly present as inorganic arsenic and most of the rest is as methylated arsenic.The total As is given by the value obtained by Procedure C. However, in sea-water samples the value from Procedure C was slightly lower than that from Procedure B, perhaps because some of the chloride may be oxidized to hypochlorite and be converted into chlorine by acidification. When hypochlorite solution was added to tetraphenylphosphonium after acidification, yellowish white curds precipitated and were collected Fig. 1 Effect of metal salts as chemical modifiers on arsenic signal intensity (arbitrary units) at 193.7 nm.Arsenic, 2 mg l21; sample volume, 100 ml; concentration factor, 50. Metal salts: 5, ZrO(NO3)2·2H2O; :, Fe(NO3)3·9H2O; 8, Cr(NO3)3·9H2O. Table 2 Effect of foreign substances on the determination of arsenic in water Substance added Added as Concentration/ mg l21 As founda (%) Sodium Na2SO4 46 000 102 Magnesium MgCl2·6H2O 2 400 88 1 200 96 Potassium KCl 3 900 99 Calcium Ca(NO3)2·4H2O 4 000 99 Aluminium Al(NO3)3·9H2O 5 95 Chloride NaCl 35 000 100 Sulfate Na2SO4 96 000 102 Phosphate KH2PO4 0.3 (as PO4) 96 0.5 (as PO4) 90 Silicate SiO2 + Na2CO3 200 (as SiO2) 102 Dodecyl sulfate SDSb 28.8 85 14.4 96 a Arsenic concentration, 2 mg l21; sample volume, 100 ml; concentration factor, 50.b Sodium dodecyl sulfate. Analyst, 1999, 124, 23–26 25on the MF. Arsenic is known to form gaseous molecules with chlorine.14,15 Therefore, this pre-treatment was not adopted for higher salinity samples. Table 3 gives the results of recovery tests at different arsenic levels. The recoveries of the added arsenic were quantitative, which indicates that the major salts in the river water did not interfere.Conclusion TMAH solution was successfully applied as the solubilizer in MF solubilization methods for the ETAAS determination of trace amounts of arsenic. TMAH solution dissolves the mixed cellulose ester MF to give a solution of low viscosity. The method will be applied to other trace elements and/or other instrumental analyses, especially atomic spectroscopic determinations.Acknowledgement The authors thank Dr. K. Goto, Professor Emeritus at Toyama University, for his valuable suggestions. References 1 S. Taguchi, E. Ito-oka and K. Goto, Bunseki Kagaku, 1984, 33, 453. 2 C. Matsubara, M. Takahashi and K. Takamura, Yakugaku Zasshi, 1985, 105, 1155. 3 S. Taguchi, S. Yamazaki, A. Yamamoto, Y. Urayama, N. Hata, I. Kasahara and K. Goto, Analyst, 1988, 113, 1695. 4 M. Kan, T. Nasu and M. Taga, Anal. Sci., 1991, 7(Suppl.), 1115. 5 Z. Q. Li, Y. Z. Shi, P. Y. Gao, X. X. Gu and T. Z. Zhou, Fresenius’ J. Anal. Chem., 1997, 358, 519. 6 N. Hata, I. Kasahara, S. Taguchi and K. Goto, Analyst, 1989, 114, 1255. 7 S. B. Gross and E. S. Parkinson, At. Absorpt. Newsl., 1974, 13, 107. 8 C. Wadelin and M. G. Mellon, Analyst, 1952, 77, 708. 9 D. L. Tsalev, V. I. Slaveykova and P. B. Mandjukov, Spectrochim. Acta Rev., 1990, 13, 225. 10 D. Chakraborti, W. D. Jonghe and F. Adams, Anal. Chim. Acta, 1980, 119, 331. 11 K. Saeed and Y. Thomassen, Anal. Chim. Acta, 1981, 130, 281. 12 Z-M. Ni, Z. Rao and M. Li, Anal. Chim. Acta , 1996, 334, 177. 13 M. Andou and Y. Magara, Shigen Kankyo Taisaku, 1997, 33, 113. 14 K. Fujiwara, J. N. Bower, J. D. Bradshow and J. D. Winefordner, Anal. Chim. Acta, 1979, 109, 229. 15 J. Koreckov, W. Frech, E. Lundberg, J. A. Persson and A. Cedergren, Anal. Chim. Acta, 1981, 130, 267. Paper 8/06856I Table 3 Analyses of natural water samples and recovery of arsenic added Inorganic Recovery of added As Samplea AsV added/mg l21 AsV found/mg l21 AsIII foundb/mg l21 Total As foundc/mg l21 mg l21 % River water A 0 0.97 ± 0.14(5)d 0.29 1.95 ± 0.07(4)d — — River water B 0 0.62 — — — — 1.00e 1.67 — — 1.05 105 2.00e 2.67 — — 2.05 103 River water C 0 0.09 ± 0.03(5)d NDf 0.32 ± 0.15(4)d — — Sea-water 0 0.50 ± 0.27(3)d 0.19 NAg — — a River waters A and B were taken from Jinzu River on different days. Concentration factor, 50. River water C was taken from Oyabe River. Concentration factor, 50. Sea-water was taken from the seashore at Toyama Bay on the Japan Sea. Concentration factor, 50. b AsIII is given by subtraction of the value obtained by Procedure A from that obtained by Procedure B. c Total As is given by the value obtained by Procedure C. d Mean ± confidence limits (No. of runs). e Arsenic added as arsenic(v). f Not detected. g Not available (see text). 26 Analyst, 1999, 124, 23–26