Detection of Nine Chlorinated and Brominated Haloacetic Acids at

Anal. Chem. 2000, 72, 4555-4559

Detection of Nine Chlorinated and Brominated Haloacetic Acids at Part-per-Trillion Levels Using ESI-FAIMS-MS Barbara Ells,† David A. Barnett,‡ Randy W. Purves,§ and Roger Guevremont*,‡

University of Alberta, Edmonton, Alberta, Canada, T6G 2G3, National Research Council of Canada, Ottawa, Ontario, Canada K1A 0R6, and PE-Sciex, Concord, Ontario, Canada, L4K 4V8

A combination of electrospray ionization, high-field asymmetric waveform ion mobility spectrometry, and mass spectrometry (ESI-FAIMS-MS) was used for the analysis of a solution containing a mixture of the nine chlorinated and brominated haloacetic acids. For a carrier gas of nitrogen in the FAIMS analyzer, haloacetate anions of the mono- and dihalogenated acids and the decarboxylated anions of three of the trihalogenated acids were detected. No signal was observed for bromodichloroacetic acid (BDCAA) at a dispersion voltage of -3400 V. The addition of a small amount of carbon dioxide to the nitrogen carrier gas resulted in the detection of the pseudomolecular trihaloacetate anions, including BDCA-, and significant increases in sensitivities for the trihalogenated species. The addition of carbon dioxide to the nitrogen carrier gas had little effect on the mono- and dihalogenated anions. Quantitative analysis of the nine haloacetic acids, using flow injection, gave detection limits between 5 and 36 parts-per-trillion in 9/1 methanol/water (v/v) containing 0.2 mM ammonium acetate. The number of deaths occurring annually from the outbreak of waterborne diseases has been dramatically reduced by the chlorination of drinking water.1 However, naturally occurring humic and fulvic acids also react with chlorine, forming organohalogen compounds that are referred to as disinfection byproducts (DBPs). Many of these DBPs are suspected or known carcinogens,2-8 and as such, maximum contaminant levels for these compounds in water have been introduced by the United States Environmental Protection Agency (U.S. EPA). The second most prevalent class of known DBPs is composed of the haloacetic acids * Corresponding author. e-mail: [email protected]. † University of Alberta. ‡ National Research Council of Canada. § PE-Sciex. (1) Akin, E. W.; Hoff, J. C.; Lippy, E. C. Environ. Health Perspect. 1982, 46, 7-12. (2) DeAngelo, A. B.; Daniel, F. B.; Stober, J. A.; Olsen, G. R. Fundam. Appl. Toxicol. 1991, 16, 337-47. (3) Jorgenson, T. A.; Meierhenry, E. F.; Rushbrook, C. J.; Bull, R. J.; Robinson, M. Fundam. Appl. Toxicol. 1985, 5, 760-9. (4) Herren-Freund, S. L.; Pereira, M. A.; Khoury, M. D.; Olsen, G. Toxicol. Appl. Pharmacol. 1987, 90, 183-9. (5) Bull, R. J.; Sanchez, I. M.; Nelson, M. A.; Larson, J. L.; Lansing, A. J. Toxicology 1990, 63, 341-59. (6) Mather, G. G.; Exon, J. H.; Koller, L. D. Toxicology 1990, 64, 71-80. 10.1021/ac000341v CCC: $19.00 Published on Web 08/24/2000

© 2000 American Chemical Society

(HAAs). The current maximum contaminant level for the sum of five of the chlorinated and brominated haloacetic acids (monochloro- (MCAA), dichloro- (DCAA), trichloro- (TCAA), monobromo(MBAA), and dibromoacetic acid (DBAA)) is 60 µg/L, as set out in the stage 1 DBP rule. This level may be lowered to 30 µg/L in upcoming stage 2 DBP negotiations. Currently, the U.S. EPA methods for HAA analysis (EPA methods 552 9 and 552.2 10) involve liquid-liquid extraction of the acids from water, followed by derivatization prior to gas chromatographic analysis. Derivatization of the nonvolatile acids is labor-intensive, is time-consuming, and uses reagents that are toxic and carcinogenic. Methods for the direct determination of the haloacetic acids (i.e., no derivatization) are therefore desirable. Recent studies have focused on the application of ion chromatography11,12 and capillary electrophoresis13-16 to HAA analysis. Both of these methods require preconcentration of the sample and still may not meet detection limits required for routine water monitoring. Liquid chromatography/electrospray ionization-mass spectrometry (LC/ESI-MS) has been reported17 to give detection limits comparable with those obtained by GC analysis with electron capture detection. However, the 2000-fold preconcentration of the sample and the chromatographic separation of the acids prior to mass spectrometry results in long analysis times. Recently, high-field asymmetric waveform ion mobility spectrometry (FAIMS), was applied to HAA analysis.18 The separation of ions in FAIMS is the result of compound-dependent changes (7) Toxicity and carcinogenesis studies of bromodichloromethane (CAS No. 75-27-4) in F- 344/N rats and B6C3F1 mice (gavage studies), U.S. Department of Human Health and Human Services; National Institutes of Health, Research Triangle Park, NC, 1987. (8) Report on carcinogenesis bioassay of chloroform, Carcinogenesis Program; National Cancer Institute, Bethesda, MD, 1976. (9) United States Environmental Protection Agency, EPA Method 552, Environmental Monitoring Systems Laboratory, Cincinnati, OH, 1990. (10) United States Environmental Protection Agency, EPA Method 552.2, National Exposure Research Laboratory, Cincinnati, OH, 1995. (11) Sarzanini, C.; Bruzzoniti, M. C.; Mentasti, E. J. Chromatogr., A 1999, 850, 197-211. (12) Vichot, R.; Furton, K. G. J. Liq. Chromatogr. 1994, 17, 4405-29. (13) Ahrer, W.; Buchberger, W. Fresnius J. Anal. Chem. 1999, 365, 604-9. (14) Martinez, D.; Farre, J.; Borrull, F.; Calull, M.; Ruana, J.; Colom, A. J. Chromatogr., A 1998, 808, 229-36. (15) Martinez, D.; Borrull, F.; Calull, M. J. Chromatogr., A 1998, 827, 105-12. (16) Martinez, D.; Borrull, F.; Colull, M. J. Chromatogr., A 1999, 835, 187-96. (17) Hashimoto, S.; Otsuki, A. J. High Resolut. Chromatogr. 1998, 21, 55-8. (18) Ells, B.; Barnett, D. A.; Froese, K.; Purves, R. W.; Hrudey, S.; Guevremont, R. Anal. Chem. 1999, 71, 4747-52.

Analytical Chemistry, Vol. 72, No. 19, October 1, 2000 4555

Figure 1. Schematic of the ESI-FAIMS-MS system.

in ion mobility that occur at high electric fields.19 The principles of separation in FAIMS have been described elsewhere.20-22 Experimentally, the change in ion mobility between high and low electric fields is reflected in the compensation voltage (CV) at which an ion is transmitted through the FAIMS analyzer. For lowmass, negatively charged ions, such as the HAAs, the mobility at high field is greater than at low field, and the CV required for ion transmission becomes more positive with increasing electric field. ESI-FAIMS-MS studies of a mixture of six haloacetic acids (MCAA, DCAA, TCAA, MBAA, DBAA, BCAA) in 9/1 methanol/ tap water (v/v) showed FAIMS to be capable of separating the haloacetic acids in the gas phase.18 Mass spectral background was dramatically reduced as voltages were tuned to transmit selected ions. Detection limits ranged from 1 to 4 ppb for the six acids in the 9/1 methanol/tap water solutions (v/v) with no extraction, preconcentration, derivatization, or chromatographic separation of the acids required. These detection limits, however, were insufficient to allow direct monitoring of water samples. Recent advances in the design of the FAIMS analyzer,23 along with observed effects of carrier gas composition on ion intensities and peak separation,24,25 have led to further investigation of the HAAs by ESI-FAIMS-MS. In this study, the nine chlorinated and brominated haloacetic acids are selectively transmitted through FAIMS and detected by mass spectrometry at levels suitable for their direct monitoring in source and treated drinking water. EXPERIMENTAL SECTION Instrumentation. A schematic of the ESI-FAIMS-MS instrument is shown in Figure 1. A similar version of this FAIMS device has been described previously.23 Briefly, the FAIMS device consists of two concentric cylinders. The end of the inner cylinder (16 mm o.d.) facing the mass spectrometer was machined to a (19) Mason, E. A.; McDaniel, E. W. Transport properties of ions in gases; John Wiley & Sons: New York, 1988. (20) Buryakov, I. A.; Krylov, E. V.; Nazarov, E. G.; Rasulev, U. K. Int. J. Mass Spectrom. Ion Processes 1993, 128, 143-8. (21) Purves, R. W.; Guevremont, R.; Day, S.; Pipich, C. W.; Matyjaszczyk, M. S. Rev. Sci. Instrum. 1998, 69, 4094-105. (22) Purves, R. W.; Guevremont, R. Anal. Chem. 1999, 71, 2346-57. (23) Guevremont, R.; Barnett, D. A.; Purves, R. W.; Vandermey, J. Anal. Chem., in press. (24) Barnett, D. A.; Purves, R. W.; Ells, B.; Guevremont, R. J. Mass Spectrom., in press. (25) Barnett, D. A.; Ells, B.; Guevremont, R.; Purves, R. W. J. Am. Soc. Mass Spectrom., in press.

4556

Analytical Chemistry, Vol. 72, No. 19, October 1, 2000

spherical surface. The inner surface (nearest the mass spectrometer) of the outer cylinder (20 mm i.d.) was machined to a concave spherical shape so that the analyzer region was kept constant at a width of 2 mm. The outer surface (nearest the mass spectrometer) of the outer cylinder was machined flat with a 1-mm aperture in the center. The outer cylinder made electrical contact with the orifice plate (-8 V) of the PE Sciex API 300 triple-quadrupole mass spectrometer. The entire FAIMS device was held by a PEEK insulating sleeve (not shown) that was attached to the orifice plate. This design holds several advantages23 over an earlier FAIMS prototype in which ions were extracted at 45 deg21,22,26 including the following: implementation of an external electrospray source (thereby decreasing the danger of contaminating the gas in the FAIMS analyzer region); more efficient sampling of ions by the mass spectrometer; and elimination of the need for a custom MS interface. The electrospray needle and sample delivery system have been described previously.27 The electrospray needle (-3800 V, -45 nA) was positioned at an angle of ∼45 deg, and ∼1 cm from a 2-mm opening in the curtain plate, which was electrically insulated from the outer cylinder of the FAIMS device. The curtain gas (4.5 L/min) was passed through a charcoal/molecular sieve filter and introduced into the gap (∼1.5 mm) between the curtain plate and the outer FAIMS cylinder. The gas split into two flows with a larger portion of the gas flowing out of the opening in the curtain plate as a countercurrent flow against the arriving electrospray ions, thus facilitating desolvation. A smaller portion of the gas flow carried the ions inward through a 1-mm opening in the outer FAIMS cylinder and along the analyzer region of the FAIMS device. This gas is referred to as the carrier gas. An asymmetric waveform, which provided a dispersion voltage (DV) of up to -3600 V with a frequency of 750 kHz, was applied to the inner cylinder of the FAIMS analyzer. The relative amplitude of the sinusoidal wave to its harmonic28 was ∼3:1. Under the appropriate electrical conditions of DV and CV, ions were transmitted through FAIMS and focused to a region in front of the spherical tip of the inner cylinder where they were sampled by the mass spectrometer. Note that differences in tuning the asymmetric waveform might cause slight day-to-day variations in CV. For acquisition of CV spectra, sample solutions were delivered by a Harvard Apparatus model 22 syringe pump at a flow rate of 1 µL/min. Flow injection analysis, used for calibration of the HAAs, employed an HPLC pump (Waters 515) to deliver the 9/1 methanol/deionized water (v/v) running buffer containing 0.2 mM ammonium acetate at a flow rate of 140 µL/min to a Rheodyne injection valve (model 7725). The solvent stream was split using a Valco T-union after a 20-µL injection loop, providing a flow rate to the electrospray needle of ∼1 µL/min. Comparisons of ESI-MS and ESI-FAIMS-MS experiments were facilitated by the relative ease of conversion between these modes of operation. ESI mass spectra were collected by removing the FAIMS device from the orifice plate and returning the API 300 curtain plate to its original configuration. The time needed for (26) Guevremont, R.; Purves, R. W. J. Am. Soc. Mass Spectrom. 1999, 10, 492501. (27) Ells, B.; Froese, K.; Hrudey, S. E.; Purves, R. W.; Guevremont, R.; Barnett, D. A. Rapid Commun. Mass Spectrom. 2000, 14, 1538-42. (28) Viehland, L. A.; Guevremont, R.; Purves, R. W.; Barnett, D. A. Int. J. Mass Spectrom. 2000, 197, 123-30.

Table 1. Calibration Figures of Merit for Nine Haloacetic Acidsa HAA

CV (V)

(M - H)- (m/z)

R2

detection limit (ppt)

MCAA DCAA TCAA MBAA DBAA TBAA BCAA BDCAA CDBAA

38.4 21.4 13.1 30.0 15.0 32.8 17.8 12.5 16.7

93 127 161 137 217 295 173 207 251

0.9995 0.9930 0.9936 0.9983 0.9965 0.9900 0.9986 0.9958 0.9991

28 10 36 18 14 6 27 13 5

a

Figure 2. IS-CV spectra (CV ) 6 to 30 V, DV ) -3400 V) of a solution containing a mixture of 1 ppm of each of the nine chlorinated and brominated haloacetic acids in 9/1 methanol/water (v/v) containing 0.2 mM ammonium acetate with pure nitrogen as the carrier gas. (a) Trihalogenated species, detected as their decarboxylated anions. (b) Mono- and dihalogenated species, detected as their haloacetate anions. No signal is observed for BDCAA.

the conversion of the instrument was less than 10 min and did not require breaking vacuum. The ESI and MS operating conditions were identical for both setups. A nebulizer gas was not used for any of the ESI-MS or ESI-FAIMS-MS experiments. Reagents. The individual haloacetic acids used in these studies had purities of at least 97%. MCAA, TCAA, MBAA, and DBAA were purchased from Fluka. Tribromoacetic acid (TBAA), DCAA, and BCAA were supplied by Aldrich Chemical Co., while bromodichloroacetic acid (BDCAA) and chlorodibromoacetic acid (CDBAA) were obtained from Supelco. Individual stock solutions of ∼1000 ppm were prepared in deionized water and stored at 4 °C. Solutions for analysis were prepared by dilution of the stocks with an ESI buffer of 9/1 methanol/deionized water (v/v) containing 0.2 mM ammonium acetate. Note, for trihaloacetic acid analysis, dilute solutions were prepared immediately before use. RESULTS AND DISCUSSION To date, most of the ESI-FAIMS-MS studies have employed either nitrogen or compressed air as the carrier gas in the FAIMS analyzer. Shown in Figure 2 are ion-selective compensation voltage (IS-CV) spectra for a solution containing a mixture of 1 ppm of each of the nine chlorinated and brominated haloacetic acids, using a carrier gas of nitrogen. The spectra were acquired by setting the DV to -3400 V and scanning the CV from 6 to 30 V while monitoring the m/z values of the nine halogenated species. For clarity, the individual ion traces for the trihalogenated species are presented in Figure 2a, while the mono- and dihalogenated

DV ) -3600 V, 95/5 nitrogen/CO2 carrier gas.

species are shown in Figure 2b. Note, the haloacetic acids were detected in their anionic forms (the masses for the haloacetate anions are given in Table 1) and have been labeled as acetate anions in figures (i.e., MCA-, MBA-, etc). The mono- and dihaloacetic acids were detected as their acetate anions; however, the trihaloacetic acids were detected only as their decarboxylated anions, the masses of which are given in the figure. Note that, at this DV, with nitrogen as the carrier gas, no signal for BDCA-, or its corresponding decarboxylated anion, was observed. Decarboxylation of the trihaloacetic acids complicates the assignment of the peaks in the CV spectrum to the individual HAA species because of significant isobaric overlap. The isotope pattern of decarboxylated TBA- overlaps with the isotopomers of CDBA-; CDBA- can lose carbon dioxide to give an overlap with the isotopomers of BDCA-; and decarboxylation of BDCA- results in an isobaric overlap with TCA-. Furthermore, the separation of isotopes by FAIMS29 skews the expected isotope distribution across the peak in the CV spectrum. As a result, mass spectra acquired at peak maximums within the CV spectrum were insufficient to determine to which trihaloacetic acid the peaks corresponded. The identities of the peaks in the CV spectrum were determined through analysis of individual component solutions. Recent studies have shown that the carrier gas composition affects the transmission of ions through the FAIMS analyzer.25 For some analytes, binary mixtures of carrier gases have been shown to provide better resolution and ion intensities than either of the pure gases alone.24 The effect of using a binary gas mixture of nitrogen and carbon dioxide on the analysis of the haloacetic acids is shown in Figure 3. Figure 3 shows IS-CV spectra (CV ) 6 to 30 V) for a solution containing a mixture of 1 ppm of each of the nine chlorinated and brominated haloacetic acids collected at DV ) -3400 V with a carrier gas of nitrogen containing 3% CO2. For clarity, the individual ion traces for the trihalogenated species are shown in Figure 3a, while traces for the mono- and dihalogenated species are shown in Figure 3b. Addition of 3% CO2 to the carrier gas caused the mono- and dihaloacetic acids to be transmitted through the FAIMS analyzer at higher CVs with slightly greater signal intensities than observed in pure nitrogen. More importantly, the addition of a small amount of CO2 to the carrier gas resulted in the detection of the pseudomolecular trihaloacetate anions and dramatic increases in sensitivity for the trihalogenated species. The change in species detected for TCA(29) Barnett, D. A.; Purves, R. W.; Guevremont, R. Nucl. Inst. Methods A. 2000, 450, 179-85.

Analytical Chemistry, Vol. 72, No. 19, October 1, 2000

4557

Figure 3. IS-CV spectra (CV ) 6 to 30 V, DV ) -3400 V) of a solution containing a mixture of 1 ppm of each of the nine chlorinated and brominated haloacetic acids in 9/1 methanol/water (v/v) containing 0.2 mM ammonium acetate with a carrier gas of nitrogen containing 3% CO2. (a) Trihalogenated species. (b) Mono- and dihalogenated species. All of the haloacetic acids were detected as their haloacetate anions.

and TBA- was also accompanied by a large shift in CV. The anion BDCA-, which was not observed at DV ) -3400 V using a nitrogen carrier gas, now appears at a CV of 9.2 V and has the second highest signal intensity of the haloacetate anions. Note that the voltages within the mass spectrometer interface, which were optimized for MCAA sensitivity, resulted in some fragmentation of CDBA- and TBA-, shown by the presence of the decarboxylated anion appearing at the same CVs in the spectrum. The effect of the percentage of CO2 in the carrier gas on the signal for the trihaloacetate anions is shown in Figure 4. IS-CV spectra (CV ) 4 to 26 V) from a solution containing a mixture of 2.5 ppm of each of the four trihaloacetic acids using a carrier gas of pure nitrogen at DV ) -3400 V is shown in Figure 4a. The decarboxylated anions of TBA- (m/z -251), CDBA- (m/z -207), and TCA- (m/z -117) were detected, but no signal for BDCAwas observed. With the addition of 3% CO2 to the nitrogen carrier gas (Figure 4b), the haloacetate anions of the four trihalogenated species were detected, and large CV shifts, compared with Figure 4a, were observed. The decarboxylated TBA- anion transmitted at CV ) 9.2 V in a carrier gas of nitrogen was transmitted as TBAat CV ) 21.3 V when the carrier gas included 3% CO2. The decarboxylated TCA- anion transmitted at CV ) 15.9 V in nitrogen was transmitted as TCA- at CV ) 11.2 V with the addition of 3% CO2. The BDCA- anion, missing from the CV spectrum collected when a carrier gas of nitrogen was used, was now an intense peak at CV ) 11.1 V. With the addition of 8% CO2, the TCA- peak shifted to higher CV (CV ) 12.7 V), but little change was observed 4558 Analytical Chemistry, Vol. 72, No. 19, October 1, 2000

Figure 4. IS-CV spectra (CV ) 4 to 26 V, DV ) -3400 V) of a solution containing a mixture of 2.5 ppm of TBAA, TCAA, CDBAA, and BDCAA in 9/1 methanol/water (v/v) containing 0.2 mM ammonium acetate. (a) Carrier gas of pure nitrogen. Three of the four trihaloacetic acids are detected as the decarboxylated ion, while no signal is observed for BDCAA. (b) Carrier gas of nitrogen containing 3% CO2. The four trihaloacetic acids are detected as their haloacetate anions. (c) Carrier gas of nitrogen containing 8% CO2. (d) Carrier gas of 60% nitrogen/40% CO2.

in the CV spectra of the other ions. With the addition of 40% CO2, TCA- was transmitted at the highest CV of the four species (CV ) 17.5 V). The TBA- ion, which was transmitted at CV ) 21.2 V at both 3 and 8% CO2 in the nitrogen carrier gas, was transmitted at CV ) 13.1 V and decreased in sensitivity by a factor of 20. The CDBA- anion was now transmitted at a lower CV than BDCAand decreased in sensitivity by 40% over that observed at 3% CO2. Therefore, the addition of CO2 to the nitrogen carrier gas can be used to optimize separation of the haloacetate anions and their sensitivities. However, the composition of the carrier gas may not optimize both separation and sensitivity simultaneously. For quantitative analysis, a carrier gas composed of 95% nitrogen and 5% CO2, which optimized HAA ion sensitivity, was used. Solutions containing mixtures of all nine acids were analyzed by flow injection (DV ) -3600 V and the CV of optimal transmission for each ion) to determine the detection limits of the haloacetate anions. Representative flow injection peaks,

Figure 5. Flow injection peaks for TBA- (DV ) -3600 V, CV ) 32.8 V) for concentrations ranging from 20 to 1000 ppt. Inset is an expansion of the 40 and 20 ppt signals.

Figure 6. Mass spectra of a 100 ppb solution of BDCAA in 9/1 methanol/water (v/v) containing 0.2 mM ammonium acetate. (a) ESI-MS, (b) ESI-FAIMS-MS at CV ) 12.5 V, DV ) -3600 V. Inset is an expansion of the baseline.

acquired at CV ) 32.8 V, showing the signal reproducibility and the response near the detection limit for TBA- (m/z -295) are shown in Figure 5 for a concentration range of 20-1000 partsper-trillion (ppt). Based on 3 times the standard deviation of the background, the detection limit for TBA- is 6 ppt. Table 1 gives the detection limits for the nine haloacetic acids, which range from

5 ppt for CDBA- to 36 ppt for TCA-, along with the isotopomer monitored and their CVs of optimal transmission. These detection limits are approximately 100-fold lower than those previously reported using ESI-FAIMS-MS.18 To demonstrate the advantage of ESI-FAIMS-MS over conventional ESI-MS, mass spectra of a 100 ppb solution of BDCAA were collected using each method. Each mass spectrum shown in Figure 6 represents an average of 10 scans. The ESI mass spectrum, Figure 6a, has a relatively high background, with a signal-to-background ratio (S/B) for m/z -207 of 12, where the background was averaged between m/z -185 and -195. The detection limit for BDCA- by ESI-MS is approximately 100 ppb. The ESI-FAIMS-MS mass spectrum, Figure 6b, acquired at DV ) -3600 V and CV ) 12.5 V, shows a small increase in absolute signal over conventional ESI-MS, and a virtual elimination of the background, resulting in a S/B of 25 000, again the background being taken as the average signal between m/z -185 and -195. The virtual elimination of the background by FAIMS lowers the detection limit of BDCA- to 13 ppt, an improvement of roughly 4 orders of magnitude over conventional ESI-MS. Note, peaks appearing between m/z -161 and -167 in Figure 6b are the result of decarboxylation of BDCA- within the mass spectrometer interface, and m/z -169 is an unknown background ion that was transmitted through the FAIMS device at a similar CV value. CONCLUSIONS Transmission of ions through the FAIMS analyzer is affected by the carrier gas. Analysis of haloacetic acids using a pure nitrogen carrier gas resulted in the detection of haloacetate anions for the mono- and dihaloacetic acids and decarboxylated anions for the trihaloacetic acids. The addition of a small percentage of CO2 to the nitrogen carrier gas had little effect on the transmission of the mono- and dihaloacetate anions, but produced dramatic changes in the transmission of the trihalogenated species. A small amount of CO2 in the nitrogen carrier gas resulted in the detection of trihaloacetate anions with a high degree of sensitivity. The high sensitivity, combined with the selectivity of ions transmitted through FAIMS, improves the detection limit by 3-4 orders of magnitude over conventional ESI-MS. The detection limits for the nine haloacetic acids in 9/1 methanol/water (v/v) containing 0.2 mM ammonium acetate ranged between 5 and 36 ppt with no sample preconcentration prior to analysis. The elimination of chromatographic separation prior to analysis along with the low detection limits makes this technique a fast, simple, and sensitive method for monitoring HAA concentrations in water. ACKNOWLEDGMENT The authors thank the American Water Works Association Research Foundation (AWWARF) for their support in this project. Received for review March 22, 2000. Accepted June 27, 2000. AC000341V

Analytical Chemistry, Vol. 72, No. 19, October 1, 2000

4559