Anal. Chem. 1999, 71, 4747-4752
Detection of Chlorinated and Brominated Byproducts of Drinking Water Disinfection Using Electrospray Ionization-High-Field Asymmetric Waveform Ion Mobility Spectrometry-Mass Spectrometry Barbara Ells,† David A. Barnett, Kenneth Froese,† Randy W. Purves,‡ Steve Hrudey,† and Roger Guevremont*
National Research Council of Canada, Ottawa, Canada K1A 0R6, Department of Public Health Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2G3, and PE-SCIEX, Concord, Ontario, Canada L4K 4V8
The lower limit of detection for low molecular weight polar and ionic analytes using electrospray ionization-mass spectrometry (ESI-MS) is often severely compromised by an intense background that obscures ions of trace components in solution. Recently, a new technique, referred to as high-field asymmetric waveform ion mobility spectrometry (FAIMS), has been shown to separate gas-phase ions at atmospheric pressure and room temperature. A FAIMS instrument is an ion filter that may be tuned, by control of electrical voltages, to continuously transmit selected ions from a complex mixture. This capability offers significant advantages when FAIMS is coupled with ESI, a source that generates a wide variety of ions, including solvent clusters and salt adducts. In this report, the tandem arrangement of ESI-FAIMS-MS is used for the analysis of haloacetic acids, a class of disinfection byproducts regulated by the US EPA. FAIMS is shown to effectively discriminate against background ions resulting from the electrospray of tap water solutions containing the haloacetic acids. Consequently, mass spectra are simplified, the selectivity of the method is improved, and the limits of detection are lowered compared with conventional ESI-MS. The detection limits of ESI-FAIMS-MS for six haloacetic acids ranged between 0.5 and 4 ng/mL in 9:1 methanol/tap water (5 and 40 ng/mL in the original tap water samples) with no preconcentration, derivatization, or chromatographic separation prior to analysis. The introduction of chlorination as a means to disinfect drinking water has dramatically reduced the number of deaths occurring annually from outbreaks of waterborne diseases.1 However, chlorination also leads to the formation of organohalogen species from the reaction of hypochlorous acid with the humic and fulvic acids found naturally in water. These organohalogens * Corresponding author. † University of Alberta. ‡ PE-SCIEX. (1) Akin, E. W.; Hoff, J. C.; Lippy, E. C. Environ. Health Perspect. 1982, 46, 7-12. 10.1021/ac990343j CCC: $18.00 Published on Web 09/18/1999
© 1999 American Chemical Society
are classified as disinfection byproducts (DBPs) and, among others, include trihalomethanes (THMs) and haloacetic acids (HAAs). Toxicological studies on laboratory animals have found that several of the DBPs are carcinogenic2-8 and may have adverse reproductive consequences.9-13 The United States Environmental Protection Agency (US EPA) has introduced guidelines for the maximum contaminant levels of these organohalogen compounds. The Stage 1 DBP rule for total THMs is set at 80 µg/L, and the limit for the sum of five haloacetic acids (monobromo- (MBAA), monochloro- (MCAA), dibromo- (DBAA), dichloro- (DCAA), and trichloroacetic acid (TCAA)) is set at 60 µg/L. It is proposed that these levels be lowered to 40 µg/L and 30 µg/L, respectively, during upcoming Stage 2 DBP negotiations. Presently, practical methods are available for the analysis of volatile THMs (i.e., bromoform, chloroform, bromodichloromethane, (2) Jorgenson, T. A.; Meierhenry, E. F.; Rushbrook, C. J.; Bull, R. J.; Robinson, M. Fundam. Appl. Toxicol. 1985, 5, 760-9. (3) Report on carcinogenesis bioassay of chloroform, Carcinogenesis Program; National Cancer Institute: Bethesda, MD, 1976 (4) Toxicity and carcinogenesis studies of bromodichloromethane (Cas No. 7527-4) 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, N. C, 1987. (5) DeAngelo, A. B.; Daniel, F. B.; Stober, J. A.; Olsen, G. R. Fundam. Appl. Toxicol. 1991, 16, 337-47. (6) Herren-Freund, S. L.; Pereira, M. A.; Khoury, M. D.; Olsen, G. Toxicol. Appl. Pharmacol. 1987, 90, 183-9. (7) Bull, R. J.; Sanchez, I. M.; Nelson, M. A.; Larson, J. L.; Lansing, A. J. Toxicology 1990, 63, 341-59. (8) Mather, G. G.; Exon, J. H.; Koller, L. D. Toxicology 1990, 64, 71-80. (9) Katz, R.; Tai, C. N.; Diener, R. M.; McConnell, R. F.; Semonick, D. E. Toxicol. Appl. Pharmacol. 1981, 57, 273-87. (10) Bhat, H. K.; Kanz, M. F.; Campbell, G. A.; Ansari, G. A. S. Fundam. Appl. Toxicol. 1991, 17, 240-53. (11) Cicmanec, J. L.; Condie, L. W.; Olsen, G. R.; Wang, S.-R. Fundam. Appl. Toxicol. 1991, 17, 376-89. (12) Kramer, M. D.; Lynch, C. F.; Isacson, P.; Hanson, J. W. Epidemiology 1992, 3, 407-13. (13) Bove, F.; Fulcomer, M. C.; Klotz, J. K.; Esmart, J.; Dufficy, E. M.; Savrin, J. E. Am. J. Epidemiol. 1995, 141, 850-62. (14) Munch, D. J. Determination of chlorination disinfection byproducts, chlorinated solvents and halogenated pesticides/herbicides in drinking water by liquidliquid extraction and gas chromatography with electron-capture detection. Revision 1.0; US EPA Method 551.1; U. S. Environmental Protection Agency: Cincinnati, OH, 1995
Analytical Chemistry, Vol. 71, No. 20, October 15, 1999 4747
chlorodibromomethane) by gas chromatography (GC) using both electron capture14,15 and mass spectrometric detection.16,17 However, the polar and ionic compounds, including the HAAs, are not volatile, which makes their analysis more complex. The current US EPA method for the analysis of HAAs (EPA Method 552.2)18 involves liquid-liquid extraction followed by derivatization of the acids prior to GC analysis. The extraction, derivatization, and separation steps are labor-intensive, time-consuming, and involve reagents that are toxic and carcinogenic. Methods for the direct determination (i.e., no derivatization) of the haloacetic acids using chromatography19,20 and capillary electrophoresis21,22 have been reported using amperometric20 or UV absorbance detection.19,21,22 However, detection by these methods is not sufficiently specific or sensitive for the determination of HAAs in complicated matrixes. The applicability of electrospray ionization-mass spectrometry (ESI-MS) for the direct determination of HAAs in both water and biological samples has been reported in recent literature.23,24 While analysis by ESI-MS is considered feasible, ESI produces an abundance of solvent- and salt-related ions that give an intense mass spectral background that severely limits the detection of trace levels of HAAs. In an effort to overcome the chemical background, Hashimoto and Otsuki23 evaluated an LC/ESI-MS method for the determination of chloro- and bromo-substituted haloacetic acids in environmental waters following a 2000-fold sample preconcentration. This direct detection of the haloacetate anions resulted in comparable detection limits with those obtained by GC-ECD.25 However, LC separation of the HAAs prior to ESI-MS resulted in an analysis time of 20 minutes per sample. Tandem MS was used to eliminate the effect of the high ESI background by Brashear et al.,24 who reported detection limits at low ppb levels for the chlorosubstituted HAAs in plasma. A new continuous flow technique for the separation of gasphase ions at atmospheric pressure (760 Torr) and room temperature (298 K), referred to as high-field asymmetric waveform ion mobility spectrometry (FAIMS), has recently been described.26,27 The application of FAIMS, using an ESI source and MS detection (ESI-FAIMS-MS), has considerable potential for improving the detection of low m/z ions that are obscured in conventional ESI-MS by solvent- and salt-related ions.28 The (15) Greenberg, A. E.; Eaton, A. D. In Standard Methods for the Examination of Water and Wastewater; American Public Health Association, American Water Works Association, Water Environment Federation: New York, 1998. (16) Weisel, C. P.; Jo, W.-K. Environ. Health. Perspect. 1996, 104, 48-51. (17) Lebel, G. L.; Williams, D. T. Int. J. Environ. Anal. Chem. 1995, 60, 21320. (18) USEPA. Determination of haloacetic acids and dalapon in drinking water by liquid-liquid extraction, derivatization and gas chromatography with electron capture detection; Method 552.2; US Environmental Protection Agency: Cincinnati, OH, 1995. (19) Vichot, R.; Furton, K. G. J. Liq. Chromatogr. 1994, 17, 4405-29. (20) Bachmann, K.; Blaskowitz, K.-H.; Bukatsch, H.; Pohl, S.; Sprenger, U. J. Chromatogr. 1989, 382, 307. (21) Martinez, D.; Farre, J.; Borrull, F.; Calull, M.; Ruana, J.; Colom, A. J. Chromatogr., A 1998, 808, 229-36. (22) Martinez, D.; Borrull, F.; Calull, M. J. Chromatogr., A 1998, 827, 105-12. (23) Hashimoto, S.; Otsuki, A. J. High Resolut. Chromatogr. 1998, 21, 55-8. (24) Brashear, W. T.; Bishop, C. T.; Abbas, R. J. Anal. Toxicol. 1997, 21, 3304. (25) Barth, R. C.; Fair, P. S. J. AWWA 1992, 84(11), 94-8. (26) Purves, R. W.; Guevremont, R.; Day, S.; Pipich, C. W.; Matyjaszczyk, M. S. Rev. Sci. Instrum. 1998, 69, 4094-105. (27) Guevremont, R.; Purves, R. W. Rev. Sci. Instrum. 1999, 70, 1370-83. (28) Purves, R. W.; Guevremont, R. Anal. Chem. 1999, 71, 2346-57.
4748 Analytical Chemistry, Vol. 71, No. 20, October 15, 1999
Figure 1. Illustration of the dependence of ion mobility on electric field strength for three different types of ions.
Figure 2. An illustration of the ion motion between two parallel plates during application of an electric potential shown as V(t); the ion is transported horizontally by a gas flow (distance not to scale).
FAIMS technique takes advantage of the dependence of gas-phase ion mobility on the applied electric field, E. At low electric fields, ion mobility, K, is independent of field strength, while at high electric fields (e.g., E > 104 V/cm) ion mobility is a function of E and may be denoted as Kh. An illustration of the dependence of ion mobility, plotted as Kh/K, on E is shown in Figure 1 for three types of ions that have been observed using FAIMS. In the figure, as the electric field strength increases, the mobility of a type A ion increases, the mobility of a type C ion decreases, and the mobility of a type B ion increases initially before decreasing. A mathematical description of the operation of a parallel plate FAIMS has been presented elsewhere.26-29 For illustrative purposes, consider an ion (e.g., type A, Figure 1) that is transported by a gas stream between the two parallel plates, as shown in Figure 2. The lower plate is kept at ground potential while an asymmetric waveform is applied to the other. The waveform, V(t), a simplified version shown in Figure 2, is composed of a brief high-voltage component, thigh, and a longer lasting, oppositepolarity, low-voltage component, tlow. The integrated voltage-time product of one complete cycle of the waveform is zero. If the highfield portion of V(t) is sufficiently large, such that Kh > K, the distance traveled by the type A ion during thigh is greater than the distance traveled during tlow. The ion will experience a net displacement from its original position during each cycle of V(t), and will begin to move toward the lower plate, as illustrated by the dashed line in Figure 2. For an ion experiencing a net migration away from the upper plate, a constant dc voltage may be applied to this plate to reverse, or “compensate” for the drift, such that the ion does not travel toward either plate. This dc voltage is referred to as the compensation voltage (CV). If two types of ions respond differently to the applied electric field (i.e., their ratios of Kh to K are not (29) Buryakov, I. A.; Krylov, E. V.; Nazarov, E. G.; Rasulev, U. K. Int. J. Mass Spectrom. Ion Processes 1993, 128, 143-8.
Table 1. Concentrations of HAAs in Standard Solution formula conc. in weight standard HAA (g/mol) (µg/mL) MCAA DCAA TCAA MBAA DBAA BCAA a
Figure 3. (a) Three-dimensional schematic of the ESI-FAIMS-MS instrument (the FAIMS and MS interface plate (OR) were connected at a 45° angle) and (b) the actual asymmetric waveform used in this study. The maximum value of the waveform is called the dispersion voltage (DV).
identical), the CV values necessary to prevent net drift will also differ, thereby allowing selective transport of one ion over the other. To analyze a mixture of ions, the CV may be scanned to transmit each of the components in a mixture, generating a CV spectrum. The FAIMS analyzer used in this study consisted of concentric cylinders instead of parallel plates, following the design modification of Carnahan and Tarassov.30 This modification in geometry has improved sensitivity compared with that of parallel plates because of an ion-focusing mechanism that was described by Guevremont and Purves.27 This focusing of ions under specific instrumental conditions results in four modes of operation of FAIMS, namely, modes P1, P2, N1, and N2. The P and N refer to positive and negative ions respectively, and modes 1 and 2 refer to the polarity of the applied asymmetric waveform.26-28 Low massto-charge (m/z) ions tend to be transmitted in mode 1, while larger m/z ions are generally transmitted in mode 2. Detailed principles of operation of the FAIMS analyzer have been described elsewhere.26,27 ESI-FAIMS-MS is implemented in this study to selectively transmit haloacetate anions. The five regulated HAAs, along with bromochloroacetic acid (BCAA), in solutions of 9:1 methanol/ tap water (v/v), are selectively transmitted through FAIMS and detected by mass spectrometry without preconcentration, derivatization, or chromatographic separation prior to analysis. EXPERIMENTAL SECTION FAIMS Instrumentation. Figure 3a shows a three-dimensional view of the ESI-FAIMS-MS instrument used in this study. A detailed description of the instrument has been given previously.28,31 For the generation of negative ions, the electrospray needle was held at -1950 V, giving an electrospray current of approximately -40 nA. The asymmetric waveform was applied (30) Carnahan, B. L.; Tarassov, A. S. U.S. Patent 5,420,424, 1995. (31) Guevremont, R.; Purves, R. W. J. Am. Soc. Mass Spectrom. 1999, 10, 492501.
94.5 128.9 163.4 138.9 217.8 173.4
300 300 100 200 100 200
(M-H)-, -m/z
(M-CO2H)-, -m/z
CV (V)
93,a 95 127, 129, 131 161, 163, 165, 167 137, 139 215, 217, 219 171, 173, 175
n/a 83, 85, 87 117, 119, 121, 123 n/a 171, 173, 175 127, 129, 131
21.7 14.5 18.0 17.9 9.9 11.6
Isotope m/z ratios used for monitoring are given in bold.
to the long inner cylinder of the FAIMS analyzer. The waveform (N1 mode) that was used to generate all spectra in this study is shown in Figure 3b. The maximum voltage of this waveform, referred to as the dispersion voltage (DV), was varied between 0 and -3300 V. The frequency of the asymmetric waveform was constant at 210 kHz. The CV, which was also applied to the long inner cylinder of the FAIMS analyzer, was scanned from -3 to 27 V when collecting ion-selected (IS) CV spectra (i.e., specific m/z values monitored as a function of CV) or set to a specific value when acquiring a mass spectrum. Compressed air was introduced into the carrier gas inlet (Cin) of the FAIMS analyzer at a flow rate of 5 L/min. Gas exited through the carrier gas outlet (Cout) at 4 L/min and through the sample gas outlet (Sout) at 1 L/min. The pressure inside the FAIMS analyzer was maintained at approximately 770 Torr. If the combination of DV and CV was appropriate, ions were transferred to the vacuum chamber of a mass spectrometer (PE SCIEX API 300 triple quadrupole) through a sampler cone (260µm orifice) placed at the end of the FAIMS analyzer at a 45° angle relative to the axis of the FAIMS cylinders. For clarity, the FAIMSMS interface illustrated in Figure 3a shows the ions exiting the FAIMS at a 90° angle. The sampler cone was electrically insulated from the FAIMS and a separate voltage (OR), which varied between -24 and -29 V depending on the compound being studied, was applied to it. An offset voltage ranging between -34 and -37 V was also applied to the entire FAIMS unit (VFAIMS) to enhance the sensitivity of the FAIMS-MS. The skimmer cone was held at ground potential and the small ring electrode normally located behind the orifice of the API 300 was not incorporated into the present interface. Flow injection analysis, used for calibration of the HAAs, employed an HPLC pump (Waters 6000A) to deliver the solvent at a flow rate of 100 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 approximately 1 µL/min. Reagents. A standard acid mixture (EPA 552.1 Acids Calibration Mix ICR, 100-300 µg/mL, Supelco, USA), containing the nine chloro- and bromo-substituted haloacetic acids along with 2,2dichloropropionic acid and 2-bromopropionic acid in methyl tertbutyl ether (MTBE), was diluted in an ESI buffer of 9:1 methanol/ tap water (v/v) containing 0.2 mM ammonium acetate (Sigma Aldrich) prior to mass spectrometric analysis. The concentrations of the haloacetic acids of interest in the standard are given in Table 1. The individual DCAA solutions used for the determination of the upper limit of quantitation were prepared from a 1000 µg/mL Analytical Chemistry, Vol. 71, No. 20, October 15, 1999
4749
Figure 4. Effect of DV on the IS-CV spectra of the five US EPA regulated HAAs and BCAA in a solution of 9:1 methanol/tap water (v/v) containing 0.2 mM ammonium acetate. (a) DV ) 0 V, (b) DV ) -1300 V, (c) DV ) -1700 V, (d) DV ) -2500 V, (e) DV ) -3300 V.
Figure 5. Mass spectra collected at CV values corresponding to the peak maxima in the IS-CV spectrum presented in Figure 4e: (a) CV ) 9.9 V, (b) CV ) 11.6 V, (c) CV ) 14.5 V, (d) 18.0 V, and (e) 21.7 V. DV ) -3300 V.
standard in MTBE (Supelco), while the TCAA solutions were prepared from reagent-grade TCAA (Fisher). Glass-distilled HPLC grade methanol (Anachemia) was used as received. RESULTS AND DISCUSSION As described earlier, ions are transmitted through FAIMS at combinations of CV and DV that are determined by the iondependent mobilities at high electric field. The effect of DV on the CV spectra of the five regulated HAAs and BCAA is illustrated in Figure 4 for a 500-fold dilution of the EPA 552.1 standard. In each trace, the IS-CV spectrum was collected using the m/z values indicated in Table 1. The dwell time and number of scans were the same for each spectrum. At DV ) 0 V, Figure 4a, ions are transmitted with no change in mobility and appear at CV ≈ 0 V. At DV ) -1300 V, Figure 4b, the ions have experienced small, field-induced, changes in their mobility and appear at CV values greater than 0 V. As the DV becomes more negative, Figure 4ce, the effect of the electric field on ion mobility increases and the peaks shift toward more positive CV values. Furthermore, in these CV spectra, ion separation is improved and the peaks become more intense. The latter result is a consequence of ion focusing within the FAIMS analyzer, the mechanism for which has been described in detail elsewhere.27 Mass spectra were acquired at CV values corresponding to the peak maxima within the IS-CV spectrum shown in Figure 4e and are presented in Figure 5a-e. Each mass spectrum represents the sum of 10 scans using a dwell time of 5 ms per 0.1 amu. Note that even though it was the haloacetate anions (or in the case of TCAA, the fragment ion CCl3- that was formed in the mass spectrometer interface) that were detected, the spectra are labeled with the parent compound abbreviations for simplicity. The peak at 9.9 V in the CV spectrum is the result of the transmission of DBAA, as shown in Figure 5a. In addition to DBAA, other species are transmitted through the FAIMS at this CV value. For example, the distribution of peaks around m/z -141 in this mass spectrum corresponds to 2,2-dichloropropionic acid, which is present within the EPA 552.1 standard mixture. The mass spectrum collected at CV ) 11.6 V, Figure 5b, shows BCAA along with bromodichloroacetic acid (m/z -217), which has optimal transmission through FAIMS at a similar CV. The mass spectrum of the peak at 14.5 V, Figure 5c, shows the transmission of DCAA, as well as acetate 4750 Analytical Chemistry, Vol. 71, No. 20, October 15, 1999
Figure 6. (a) ESI-MS of a 500-fold dilution of EPA 552.1 standard solution with 9:1 methanol/tap water (v/v) containing 0.2 mM ammonium acetate. (b) ESI-FAIMS-MS of the same solution at DV ) -3300 V and CV ) 18.0 V.
(m/z -59) and the acetate dimer, H(CH3COO)2- (m/z -119). Both TCAA and MBAA are present in the mass spectrum shown in Figure 5d, which was acquired at CV ) 18.0 V. The optimal transmission of MBAA and TCAA through FAIMS occurs at CV ) 17.9 and 18.0 V, respectively. As a result of this similarity in CV values for optimal transmission, MBAA and TCAA do not appear as separate peaks in the CV spectrum. Finally, Figure 5e shows that MCAA is transmitted through FAIMS at CV ) 21.7 V. To evaluate the improvement in the signal-to-background ratio (S/B), mass spectra collected with conventional ESI-MS were compared with those collected using ESI-FAIMS-MS. Mass spectra of a 500-fold dilution of the EPA 552.1 standard are shown in Figure 6. The upper trace, Figure 6a, was collected using ESIMS, and has been plotted such that the MBAA peak at m/z -137 is full-scale. In this spectrum, peaks are observed at essentially every m/z value. These peaks are attributable to several of the HAAs in the solution as well as to solvent-related background ions. Fragmentation of HAAs via the neutral loss of CO2 also contributes to the complexity of this mass spectrum and will hamper the accurate identification and quantitation of some species. For
Table 2. Calibration Figures of Merit for Six Haloacetic Acids HAA
background (cps)
slope (cps/ppb)
R2
D. L. (ppb)
MBAA MCAA DBAA DCAA TCAA BCAA
8(3 18 ( 5 1(1 56 ( 10 13 ( 4 2(2
11.9 22.6 5.66 21.6 3.45 6.33
0.9996 0.9995 0.9975 0.9986 0.9965 0.9970
0.8 0.7 0.5 1.4 4 1
Figure 7. Flow-injection peaks for increasing concentrations of TCAA in 9:1 methanol/tap water (v/v) solutions containing 0.2 mM ammonium acetate. DV ) -3300 V, CV ) 18.0 V.
Figure 8. Calibration curve for TCAA in 9:1 methanol/tap water (v/ v) containing 0.2 mM ammonium acetate. DV ) -3300 V, CV ) 18.0 V.
example, the loss of CO2 (m/z 44) from BCAA (m/z -171) results in an ion that will overlap with DCAA (m/z -127). The lower trace, Figure 6b, was acquired using ESI-FAIMS-MS at DV ) -3300 V and CV ) 18.0 V. Under these conditions, TCAA and MBAA are both transmitted by FAIMS. The spectrum is dramatically simplified over that acquired using ESI-MS. This combination of DV and CV effectively eliminated background ions, other HAAs, and their neutral loss fragments from the mass spectrum. The major peaks in the mass spectrum correspond to TCAA, MBAA, and Br-. Bromide (m/z -79 and -81) is observed at this CV because of MBAA fragmentation within the mass spectrometer interface. The TCAA signal in Figure 6b is also no longer subject to overlap from acetate dimer ions (m/z -119), as observed in Figure 6a. As a result, the TCAA signal in Figure 6b shows the expected isotope pattern. Also, the overall abundance of the ions has increased with ESI-FAIMS-MS because of ion-focusing. This focusing, along with the decreased background signal, provides improvement in the lower level of detection of these compounds. With conventional ESI-MS, the lower limit of detection of MBAA is about 40 ng/mL (ppb), whereas the detection limit of MBAA using FAIMS is improved to approximately 1 ng/mL.
Figure 9. Effect of total HAA concentration on analyte signal using solutions of DCAA and TCAA in 9:1 methanol/tap water (v/v) containing 0.2 mM ammonium acetate. (a) [TCAA] ) 800 ng/mL and [DCAA] ) (i) 130 ng/mL, (ii) 650 ng/mL, (iii) 1300 ng/mL, and (iv) 2600 ng/mL. (b) [DCAA] ) 1000 ng/mL and [TCAA] ) (i) 160 ng/mL, (ii) 800 ng/mL, (iii) 1600 ng/mL, and (iv) 3200 ng/mL.
Quantitative Analysis with ESI-FAIMS-MS. Flow-injection analysis was used for generating standard addition curves for the five regulated HAAs and BCAA. Calibration curves consisted of a minimum of five points, with concentrations spanning at least one and a half orders of magnitude. Each point was determined on the basis of average peak intensities from four injections. The data were collected at DV ) -3300 V, and the CV was tuned to the optimal voltage for transmission of the desired ion. Representative flow-injection peaks for TCAA (CV ) 18.0 V), illustrating the signal near the lower detection limit as well as reproducibility, are shown in Figure 7. The calibration curve generated for TCAA is shown in Figure 8. On the basis of three times the standard deviation of the zero addition signal, the detection limit of TCAA was calculated to be 4 ppb. The detection limits for all six haloacetic acids are given in Table 2. These detection limits range between 0.5 and 4 ng/mL in a 9:1 methanol/tap water solution (v/v) with no preconcentration of the sample. Also, because no chromatographic separation of the HAAs was required before detection, sample analysis time is greatly reduced. Analytical Chemistry, Vol. 71, No. 20, October 15, 1999
4751
The upper limit of the linear calibration region is affected by the total HAA concentration in the sample. At high concentrations of HAAs in solution, signal suppression becomes apparent. Shown in Figure 9 are IS-CV spectra for solutions of DCAA and TCAA in 9:1 methanol/tap water (v/v) containing 0.2 mM ammonium acetate. In Figure 9a, the concentration of TCAA is maintained at 800 ng/mL, whereas the concentration of DCAA was 130 ng/mL in trace (i) and 650, 1300, and 2600 ng/mL in traces (ii) through (iv), respectively. The signal for DCAA increased linearly to a concentration of 1300 ng/mL, but deviated negatively from linearity at higher concentration. The signal for TCAA was constant for all of these solutions. In Figure 9b, the concentration of DCAA was maintained at 1000 ng/mL, whereas the concentration of TCAA was 160 ng/mL in trace (i) and 800, 1600 and 3200 ng/mL in traces (ii) through (iv). The increase in signal for TCAA was linear over the concentration range studied; however, the signal for DCAA decreased at high concentrations of TCAA. Analyte-dependent signal suppression is a common observation when using ESI-MS at high concentration.32-34 However, this suppression will only be a concern in the analysis of treated water when samples have been very highly preconcentrated. CONCLUSIONS FAIMS is a relatively new technique for the separation of gasphase ions at room temperature and atmospheric pressure. ESI(32) Enke, C. G. Anal. Chem. 1997, 69, 4885-93. (33) Tang, L.; Kebarle, P. Anal. Chem. 1993, 65, 3654-68. (34) Agnes, G. R.; Horlick, G. Appl. Spectrosc. 1994, 48, 649-54.
4752
Analytical Chemistry, Vol. 71, No. 20, October 15, 1999
FAIMS-MS was used to detect and quantify haloacetic acids from complex solutions. The ability to tune the voltages applied to the FAIMS analyzer, such that only selected ions were continuously transmitted through the device, greatly simplified mass spectra over that acquired using conventional ESI-MS. This selectivity, coupled with the ion-focusing provided by FAIMS, also significantly improved the detection capabilities relative to conventional ESI-MS. The detection limits for six haloacetic acids in 9:1 methanol/tap water ranged from 0.5 to 4 ng/mL using ESI-FAIMSMS with no preconcentration or derivatization prior to analysis, resulting in detection limits ranging between 5 and 40 ng/mL in the original tap-water sample. The application of existing preconcentration methods for the haloacetic acids in water samples prior to ESI-FAIMS-MS analysis is expected to lower detection limits to the mid-to-low pg/mL range. The elimination of chromatographic separation prior to detection, along with the low detection limits, make this technique a fast, simple, and sensitive method for monitoring HAA concentrations in source and treated water. ACKNOWLEDGMENT The authors thank Mine Safety Appliances Co., Pittsburgh, PA, and the Natural Science and Engineering Research Council (NSERC) for their support in this project.
Received for review April 5, 1999. Accepted August 11, 1999. AC990343J
