ANALYTICAL SCIENCES DECEMBER 2009, VOL. 25 1491 2009 © The Japan Society for Analytical Chemistry

Notes

determination of Chloroacetic acids in drinking water using suppressed Chromatography with solid-Phase Extraction

kenji Yoshikawa,† Yuko soda, and akio sakuragawa

Department of Materials and Applied Chemistry, Collage of Science and Technology, Nihon University, 1-5-1 Kanda-Surugadai, Chiyoda, Tokyo 101–8308, Japan

Suppressed ion chromatography with a conductivity detector was developed for the determination of trace amounts of underivatized chloroacetic acids (CAAs). When carbonate and methanol were used as a mobile phase, the simultaneous determination of each CAA took approximately 25 min. The linearity, reproducibility and detection limits were determined for the proposed method. For the solid-phase extraction step, the effects of the pH of the sample solution, sample volume and the eluting agent were tested. Under the optimized extracting conditions, the average recoveries for CAAs spiked in tap water were 83 – 107%, with an optimal preconcentration factor of 20. The reproducibility of recovery rate for CAAs was 1.2 – 3.8%, based upon 6 repetitions of the recovery experiments.

(received september 14, 2009; accepted November 11, 2009; Published december 10, 2009)

However, the required instruments are very expensive. Although introduction HPLC and CE do not require a derivatization step, the detection limits are unsatisfactory compared with MS-based methods. Chloroacetic acids (CAAs) are disinfection by-products,1–3 and Among the wide variety of methods for the micro-determination are generated by the chlorination of natural organic matter in of , ion chromatography (IC) with ion-exchangers and raw water. CAAs are well-known carcinogens, and may have conductivity detection has been extensively studied.21–26 some influences on heredity. Analyzing CAAs by IC coupled with conductivity detection The United States Environmental Protection Agency (EPA) does not generally require a derivatization step. has promulgated regulations to control haloacetic acids (HAAs). In addition, a poly(vinyl alcohol) column (Shodex IC SI-90 The HAAs include monochloroacetic acid (MCA), 4E) in this work is more durable. It can be used for pH 3 – 12 monobromoacetic acid (MBA), (DCA), and a wide temperature range. Controlling the position of the dibromoacetic acid (DBA) and (TCA). system peak, resulting from carbonic acid in the mobile phase, According to the EPA drinking water regulations,4 the total is possible. The authors investigated the determination of maximum concentration limit for all should not exceed inorganic anions in several water samples using IC with the 0.06 mg/l; they further state that DCA should never be present, column, and with suppressed conductivity detection.27 and the TCA concentrations should not be more than 0.03 mg/l. This study demonstrates the simultaneous determination of In Japan, a radical review of water quality standards for tap CAAs by suppressed IC, using this column with conductivity water was performed, and was enacted in April, 2004.5 The detection. Moreover, the separation behaviors of inorganic regulations for the presence of MCA, DCA and TCA were listed anions to be presented in drinking water were examined. in the standard for tap water. The standard values are below Because the standard values of CAAs in the water quality 0.02 mg/l for MCA, 0.04 mg/l for DCA and 0.2 mg/l for TCA. standard for tap water are under the determination limit, only Currently, CAAs are typically analyzed using solvent ion chromatographic determination is difficult. Therefore, extraction–gas chromatography mass spectrometry (SE-GC/MS) solid-phase extraction (SPE)22,25,26 was used to achieve the after chemical derivatization. This is the official analysis preconcentration of CAAs. SPE has been an extensive method method used in Japan. to concentrate the constituents present in sample solutions. Other analytical methods are gas chromatography with However, the use of anion exchange cartridges was deemed to electron-capture detection (GC-ECD),6–8 gas chromatography– be unsuitable, since matrix anions would be preconcentrated. In mass spectrometry (GC-MS),9–11 high-performance liquid addition, matrix anions would inevitably affect the recovery of chromatography (HPLC),12,13 liquid chromatography–mass CAAs.28,29 Hence, the preconcentration of protonated CAAs on spectrometry (LC-MS),14–16 capillary electrophoresis (CE),17 etc. polymeric reversed-phase materials was examined. For GC, CAAs must be extracted from drinking water, which The objective of this study is to establish optimal conditions requires a derivatization step. However, the extraction operation for simultaneously determining and attaining the preconcentration is very complicated. The MS detection technique14–16,18–20 is a of underivatized CAAs, and to apply them to the determination sensitive and selective method for analyzing compounds. of CAAs in drinking water.

† To whom correspondence should be addressed. E-mail: [email protected] 1492 ANALYTICAL SCIENCES DECEMBER 2009, VOL. 25

Fig. 1 Effect of the sodium carbonate concentration on the retention Fig. 2 Chromatogram of CAAs and several inorganic anions. Ions –1 – time (methanol concentration fixed 5.0 (v/v)%). Ions (mg l ): ○, F –1 – – – (mg l ): 1, F (5); 2, MCA (10); 3, Cl (10); 4, DCA (10); 5, NO3 (20); – – (5); ●, MCA (10); △, Cl (10); ▲, DCA (10); □, NO3 (20); ■, TCA 6, TCA (10). Column, Shodex IC SI-90 4E; mobile phase, 0.5 mM (10). Column, Shodex IC SI-90 4E; mobile phase, 0.25 – 2.0 mM Na2CO3–5.0 (v/v)% CH3OH; flow rate of mobile phase, 1.0 ml/min; Na2CO3–5.0 (v/v)% CH3OH; flow rate of mobile phase, 1.0 ml/min; column temperature, 30°C. column temperature, 30°C.

and 3.0 ml of 0.5 M . After 100 ml of sample Experimental solution pH was adjusted to 1.0 with sulfuric acid, a solution containing CAAs was passed through the cartridge at a rate of Apparatus and operating conditions 1.0 ml/min. The cartridge was washed with 0.5 ml of deionized The equipment consists of a DGU-12 degasser, an LC-10ADvp water to remove any excess sulfate and other inorganic anions, pump, a CTO-10Avp column oven, a CDD-6Avp conductivity and the retained CAAs were eluted by 5.0 ml of 100 (v/v)% detector and a C-R6A Chromatopac (Shimadzu, Kyoto, Japan). methanol at 1.0 ml/min. In this case, CAAs resulted in a 20-fold A manual injection valve, Model 7725 (Rheodyne, Cotati, CA, enrichment. Finally, the eluted solution was injected into a USA) with a 25-μl sample loop was used. All separations with separation column. the IC system were performed using a Shodex IC SI-90 4E (4.0 mm i.d. × 250 mm, Showa Denko, Tokyo, Japan). An anion-exchange micro-membrane suppressor (AMMS III, results and discussion Dionex, Sunnyvale, CA, USA) was used. We used 0.5 mM sodium carbonate and 5 (v/v)% methanol Effect of sodium carbonate and methanol concentration in the (flow rate, 1.0 ml/min) as the mobile phase. The column mobile phase temperature was maintained at 30°C. The suppressor was To achieve the separation of CAAs, sodium carbonate and continuously regenerated with 12.5 mM sulfuric acid at methanol were applied as a mobile phase. The effect of the 1.0 ml/min. sodium carbonate concentration on the retention behavior was examined at 0.3 – 2.0 mM, with the concentration of methanol Reagents fixed at 5 (v/v)%. The relationship between the retention time Sodium carbonate and methanol (Wako Pure Chemical, of the analyte and the concentration of sodium carbonate is Osaka, Japan) were used as a mobile phase. A standard solution shown in Fig. 1. The retention time of all analytes decreased of CAAs was provided by MCA, DCA and TCA (Kanto with an increase in the sodium carbonate concentration. At a Chemical, Tokyo, Japan). Standard solutions of anions (F–, Cl– concentration of sodium carbonate greater than 0.7 mM, the – – and NO3 ) were prepared by dissolving sodium (Wako Pure separation of MCA and Cl was difficult, and the optimum Chemical, Japan). Each standard solution was prepared by sodium carbonate concentration was determined to be 0.5 mM. diluting 1000 mg/l certified standards of CAAs and inorganic Furthermore, the effect of the methanol concentration on the anions. Sulfuric acid (Wako Pure Chemical, Japan) was used retention behavior was examined at 1.0 – 6.0 (v/v)%, with the for a scavenger of any suppressor and acidification of the sample concentration of sodium carbonate fixed at 0.5 mM. Regardless solutions. All chemicals were of analytical grade purity. of the methanol concentration, the retention time of all analytes Deionized water was further purified with a Simplicity UV was approximately constant. When the concentration is low, the system (Millipore, Bedford, MA, USA). baseline tends to be unstable. Because the maximum allowable concentration of organic solvent with the separation column is Enrichment of CAAs by a solid phase 10 (v/v)%, 5.0 (v/v)% methanol was adopted. The purpose of this study is to develop a reliable preconcentration procedure for CAAs, prior to detection by IC. Effect of the column temperature and the flow rate of the mobile For the SPE, LiChrolut EN cartridges (Merck, Darmstadt, phase Germany) with 200 mg of sorbent serving as a reversed-phase Under the composition of the mobile phase, the effect of the were employed. A cartridge was conditioned by 6.0 ml of column temperature on the retention behavior was examined at 100 (v/v)% methanol, followed by 5.0 ml of deionized water 20 – 40°C. As the temperature increased, the separation of ANALYTICAL SCIENCES DECEMBER 2009, VOL. 25 1493

Table 1 Analytical precision of chloroacetic acids Detection Linearity Chloroacetic –1 %RSD limit/mg l acid Range/mg l–1 r2 (n = 6) (S/N = 3)

MCA 0.1 – 0.3 0.9999 2.3 0.05 DCA 0.1 – 0.3 0.9999 1.6 0.09 TCA 0.1 – 0.3 0.9985 1.6 0.36

RSD: relative standard deviation at 3.0 mg/l (n = 7).

MCA and Cl– became difficult. In particular, DCA and TCA showed a tendency to delay the retention time. Considering the resolution and retention time for each analyte, 30°C was chosen as the optimal column temperature. Fig. 3 Effect of the pH of the sample solution on the recovery rate –1 Because the retention time of each species was not significantly with a SPE cartridge. Ions (0.1 mg l ): ●, MCA; ▲, DCA; ■, TCA. SPE column, LiChrolut EN cartridge; pH of sample, 0.5 – 3.0; sample affected by the flow rate of the mobile phase, its optimum flow volume, 100 ml; eluting agent, 5 ml of methanol. rate was 1.0 ml/min.

Chromatogram of CAAs and several inorganic anions A typical chromatogram of CAAs and several inorganic – – – anions (F , Cl and NO3 ) is shown in Fig. 2. The separation of Considering the recovery rate, 100 ml was adopted as the CAAs and inorganic anions present in drinking water was sample volume and 1.0 ml/min as the load rate. satisfactory. Incidentally, it was found that a large excess of Finally, the effects of several eluting agents on the recovery sulfate ion from the sample acidification step was eluted in rate were examined. Methanol, acetonitorile, acetone and ethyl nearly 70 min. Therefore, it does not interfere with the other acetate were tested as eluting agents. The load volume and the ions. load rate of the eluting agent were fixed at 5.0 ml and 1.0 ml/min, respectively. Ethyl acetate could not elute a considerable Analytical precision amount of any of the CAAs. When using an acetone as an The linearity of the calibration curves, reproducibility and eluting agent, MCA and DCA were eluted, but TCA was not. In detection limits are shown in Table 1. The calibration curves contrast, methanol and acetonitorile successfully eluted CAAs. obtained from the peak height for CAAs were linear, with a The recovery rate of CAAs was found to range from good correlation coefficient of 0.999. The relative standard approximately 60 to 90% in methanol, but was less than 50% in deviations (RSD) at 3.0 mg/l from seven repeated measurements acetonitrile. From the point of view of a high recovery rate, were between 1.5 and 2.3%. The detection limits were methanol was found to be suitable for the eluting agent. calculated at 3 for the S/N ratio. Each detection limit was 0.05 mg/l for MCA, 0.09 mg/l for DCA and 0.36 mg/l for TCA. Analysis of CAAs in drinking water The proposed method was applied to the determination of Extraction and elution of CAAs CAAs in several drinking-water samples. The analytes were Because the standard values of CAAs in water-quality extracted immediately upon collection by the optimized SPE standards for tap water are under the determination limit, SPE procedure. After elution, 25 μl of the eluates were injected into was used to achieve preconcentration of the protonated CAAs. the IC. As shown in Table 2, although MCA and TCA were not In SPE, the pH of a sample solution is important. However, it found, DCA was detected in drinking-water samples. According is hard to retain the ionic state into a reversed-phase cartridge. to previous studies, a comparable level of DCA present in 24,25 Considering the acid dissociation constant (pKMCA = 2.86, drinking-water samples has been quantified. 29 pKDCA = 1.30, pKTCA = 0.70), the effect of the pH on the To confirm the reliability of this method, the CAAs of each recovery rate was examined at pH 0.5 – 3.0. In addition, 100 ml concentration were added to tap water, spiked with 0.02 mg/l of of sample solutions each containing 0.1 mg/l of CAAs were MCA, 0.04 mg/l of DCA and 0.2 mg/l of TCA, and the recovery introduced into the cartridge at 1.0 ml/min, and 5.0 ml of test was demonstrated. Figure 4 shows a chromatogram of the 100 (v/v)% methanol was used as an eluting reagent at a flow standard addition of CAAs in tap water. Because of the rate of 1.0 ml/min. The relationship between the pH of the satisfactory recovery rate and reproducibility, this method can sample solution and the recovery rate of each CAA is shown in be applied for the determination of CAAs in drinking water, Fig. 3. Except for TCA, the recovery rate decreased as the pH while avoiding the effect of coexisting ions. increased. Although the recovery rate was satisfactory at pH 0.5, this was beyond the usable pH of the cartridge, and the reproducibility of the recovery rate was extremely poor. Conclusions Moreover, the baseline was unsuitable after eluting the sulfate ion. In view of these facts, the optimum pH was determined to In this study, an ion-chromatographic determination of three be 1.0. underivatized CAAs and inorganic anions was demonstrated. In Moreover, the effect of the sample volume on the recovery addition, a 20-fold preconcentration of CAAs is possible in the rate was examined. The sample solution was fixed at pH 1.0. optimal SPE procedure using a LiChrolut EN cartridge. This At higher sample volumes, it exceeded the overall capacity of method can be applied for the determination of CAAs in the cartridge. Moreover, an analyte breakthrough is occurring. drinking water. Moreover, in the determination of CAAs, 1494 ANALYTICAL SCIENCES DECEMBER 2009, VOL. 25

Table 2 Analytical results for chloroacetic acids in tap water

Sample Chroloacetic acid Founda/mg l–1 Added/mg l–1 Calculated value/mg l–1 Foundb/mg l–1 Recovery, % %RSD (n = 6)

Tap water A MCA n.d. 0.02 0.40 0.43 108 3.8 DCA 0.03 0.04 1.40 1.28 91 2.3 TCA n.d. 0.2 4.00 3.34 84 1.2 Tap water B MCA n.d. 0.02 0.40 0.39 98 3.8 DCA 0.04 0.04 1.60 1.47 92 2.7 TCA n.d. 0.2 4.00 3.28 82 1.6

a. Original water samples. b. Water samples preconcentrated 20-fold by a LiChrolut EN cartridge. n.d.: not detected. RSD: relative standard deviation.

Fig. 4 Chromatograms of tap water extracts (samples preconcentrated 20-fold on LiChrolut EN cartridge). Ions: 2, MCA; 3, Cl–; 4, DCA; 6, TCA. Column, Shodex IC SI-90 4E; mobile phase, 0.5 mM Na2CO3–5.0 (v/v)% CH3OH; flow rate of mobile phase, 1.0 ml/min; column temperature, 30°C; SPE column, LiChrolut EN cartridge; pH of sample, 1.0; sample volume, 100 ml; eluting agent, 5 ml of methanol.

the same value as that in the water quality standard for tap water 1055, 63. in Japan is possible. 14. M. Takino, S. Daishima, and K. Yamaguchi, Analyst, 2000, 125, 1097. 15. B. Ells, D. A. Barnett, R. W. Purves, and R. Guevremont, references Anal. Chem., 2000, 72, 4555. 16. R. Loos and D. Barcero, J. Chromatogr., A, 2001, 938, 45. 1. K. L. Simpson and K. P. Hayes, Water Res., 1998, 32, 1522. 17. D. Martinez, F. Borrull, and M. Calull, J. Chromatogr., A, 2. L. Liang and P. C. Singer, Environ. Sci. Technol., 2003, 37, 1998, 827, 105. 2920. 18. R. Roehl, R. Slingsby, N. Avdalovic, and P. E. Jackson, J. 3. S. K. Golfinopoulos and A. D. Nikolaou, Desalination, Chromatogr., A, 2002, 956, 245. 2005, 176, 13. 19. Y. Liu, S. Mou, and D. Chen, J. Chromatogr., A, 2004, 4. US EPA, 2006 Edition of the Drinking Water Standards and 1039, 89. Health Advisories, 2006, 1 – 7. 20. L. Barron and B. Paull, Talanta, 2006, 69, 621. 5. A. Kobayashi and Y. Fuchu, Industrial Water (in Japanese), 21. L. M. Nair, R. Saari-Nordhaus, and J. M. Anderson, Jr., J. 2003, 537, 37. Chromatogr., A, 1994, 671, 309. 6. D. Benanou, F. Acobas, and P. Sztajnbok, Water Res., 1998, 22. C. Sarzanini, M. C. Bruzzoniti, and E. Mentasti, J. 32, 2798. Chromatogr., A, 1999, 850, 197. 7. A. A. Kampioti and E. G. Stephanou, J. Chromatogr., A, 23. Y. Liu and S. Mou, J. Chromatogr., A, 2003, 997, 225. 1999, 857, 217. 24. Y. Liu and S. Mou, Chemosphere, 2004, 55, 1253. 8. A. D. Nikolaou, S. K. Golfinopoulos, M. N. Kostopoulou, 25. L. Barron and B. Paull, Anal. Chim. Acta, 2004, 522, 153. and T. D. Lekkas, Water Res., 2002, 36, 1089. 26. L. Barron, P. N. Nesterenko, and B. Paull, J. Chromatogr., 9. M. N. Sarrion, F. J. Santos, and M. T. Galceran, J. A, 2005, 1072, 207. Chromatogr., A, 1999, 859, 159. 27. A. Isozaki, K. Yoshikawa, and H. Nagashima, Industrial 10. Y. Xie, Water Res., 2001, 35, 1599. Water (in Japanese), 2003, 541, 11. 11. M. Jia, W. W. Wu, R. A. Yost, P. A. Chadik, P. W. Stacpoole, 28. D. Martínez, F. Borrull, and M. Calull, J. Chromatogr., A, and G. N. Henderson, Anal. Chem., 2003, 75, 4065. 1998, 827, 105. 12. H. Carrero and J. F. Rusling, Talanta, 1999, 48, 711. 29. B. Paull and L. Barron, J. Chromatogr., A, 2004, 1046, 1. 13. D. Kou, X. Wang, and S. Mitra, J. Chromatogr., A, 2004,