Anal. Chem. 2000, 72, 2308-2312
Microextraction of Nine Haloacetic Acids in Drinking Water at Microgram per Liter Levels with Electrospray-Mass Spectrometry of Stable Association Complexes Matthew L. Magnuson* and Catherine A. Kelty
National Risk Management Research Laboratory, Water Supply and Water Resources Division, Treatment Technology Evaluation Branch, U.S. Environmental Protection Agency, 26 West Martin Luther King Drive, Cincinnati, Ohio 45268
Haloacetic acids are disinfection byproducts of the chlorination of drinking water. This paper presents the analysis of all nine chloro and bromohaloacetic acids (HAA9) at the submicrogram per liter level by microextraction with detection by electrospray-mass spectrometry. The haloacetic acids are extracted from acidified water through a microscale liquid-liquid extraction. Perfluoroheptanoic acid is added to the extracts, and the haloacetic acids are detected with electrospray-mass spectrometry. Confidence in the selective quantification of the haloacetic acids is achieved by observing the stable association complexes that are formed between the haloacetic acids and perfluoroheptanoic acid. The method detection limits for the haloacetic acids are less than 1 µg L-1, depending on the haloacetic acid. Standard addition is used to quantify the haloacetic acids in several water matrixes. The disinfection of drinking water to control microbial contaminants results in the formation of chemical disinfection byproducts (DBPs). One class of disinfection byproducts which has been the subject of much investigation comprises the haloacetic acids (HAAs). Dichloroacetic acid and trichloroacetic acid are animal carcinogens.1 Due to the potential adverse human health effects of these compounds, five haloacetic acids are listed in the Stage 1 Disinfectants and Disinfection Byproducts Rule.2 Under this proposal, the sum of the concentrations of chloro-, dichloro-, trichloro-, bromo-, and dibromoacetic acids (HAA5) is regulated at a maximum contaminant level of 60 µg/L. “HAA6”, composed of HAA5 + bromochloroacetic acid, are readily determined through Method 552.03 by extraction from drinking water, methylation by diazomethane, and analysis by gas chromatography with electron capture detection (GC-ECD). Because the remainder * Corresponding author. Phone: 513-569-7321. Fax: 513-569-7658. E-mail:
[email protected]. (1) Richard, A. M.; Hunter, E. S. Teratology 1996, 53, 352-365. (2) U.S. Environmental Protection Agency Stage 1 Disinfectants and Disinfection Byproducts Rule; EPA Document No. 815-F-98-010; GPO: Washington, DC, 1998. (3) Hodgeson, J. W.; Collins, J.; Barth, R. E. Determination of Haloacetic Acids in Drinking Water by Liquid-Liquid Extraction, Derivization and Gas Chromatography with Electron Capture Detection. EPA Method 552.0, Revision 1. In Methods for the Determination of Organic Compounds, Supplement II; EPA Document No. 600-4-88-039; GPO: Washington, DC, 1990.
2308 Analytical Chemistry, Vol. 72, No. 10, May 15, 2000
of the HAAs may be subject to future regulation, EPA Method 552.24 adds bromodichloro-, dibromochloro-, and tribromoacetic acids (to give HAA9) and uses acidic methanol for methylation. The decarboxylation of these last three trihaloacetic acids (HAA9) can occur during methylation. The degree of decomposition is matrix and operator dependent. Therefore, there is a need for alternative means to quantify haloacetic acids in drinking water, particularly for fundamental studies involving the formation, decay, and transformation of these compounds. These types of studies may help to reveal the significance of HAA9 in the overall haloacetic acid contamination of a disinfected water. Understanding the behavior of the brominated HAAs may be particularly important because bromide-containing waters can produce another carcinogenic disinfection byproduct, bromate, when ozone is used as a pretreatment process. Quantifying brominated HAAs is necessary to balancing risks of producing brominated HAAs with the risks of producing bromate. Most alternative HAA analyses have focused on HAA5. For instance, liquid chromatography techniques with ion-exchange5 or reversed-phase6 columns have been used with a conductivity detector,5 an ultraviolet detector,5 and an electrochemical detector coated with a surfactant-Nafion film.6 Capillary electrophoresis combined with solid-phase extraction7 has been used for the haloacetic acids. Solid-phase microextraction of methylated haloacetic acids has been investigated,8 as has derivatization of HAAs using 2,4-difluoroaniline, with GC separation followed by detection with a mass spectrometer or an atomic emission detector.9 Other reports have addressed the analysis of HAA9. Modifications to EPA Method 552.2 were recently investigated.10 Simultaneous (4) Munch, D. J.; Munch, J. W.; Pawlecki, A. M. Determination of Haloacetic Acids and Dalapon in Drinking Water by Liquid-Liquid Extraction, Derivization and Gas Chromatography with Electron Capture Detections, EPA Method 552.2, Revision 1. In Methods for the Determination of Organic Compounds, Supplement III; EPA Document No. 600-R-95-131; GPO: Washington, DC, 1995. (5) Sarzanini, C.; Bruzzoniti, M. C.; Mentasti, E. J. Chrom. A. 1999, 850, 197211. (6) Carrero, H.; Rusling, J. F. Talanta 1999, 48, 711-718. (7) Martinez, D.; Borrull, F.; Calull, M. J. Chromatogr. A 1998, 827, 105-112. (8) Aikawa, B.; Burk, R. C. Int. J. Environ. Anal. Chem. 1997, 66, 215-224. (9) Scott, B. F.; Alaee, M. Water Qual. Res. J. Can. 1998, 33, 279-293. (10) Brophy, K. S.; Weinberg, H. S.; Singer, P. C. Preprints of Extended Abstracts, 217th National Meeting of the American Chemical Society, Anaheim, CA, March 21-25, 1999, pp 256-259. 10.1021/ac991469j Not subject to U.S. Copyright. Publ. 2000 Am. Chem. Soc.
Published on Web 04/12/2000
extraction-derivatization using solid-phase extraction was investigated with GC-ECD detection.11 Electrospray mass spectrometry was also used recently to investigate HAAs. Tandem mass spectrometry has been used with an electrospray interface to quantitate trichloroacetic acid, dichloroacetic acid, and monochloroacetic acid in biological samples.12 The direct electrospray introduction of aqueous haloacetic acids has been investigated, but the presence of other ionic species in the water caused matrix effects.13 To reduce matrix effects, solvent extraction was combined with liquid chromatography-mass spectrometry (LC-ESIMS), achieving submicrogram per liter detection limits.14 Due to limitations in the choice of solvents for LC-ESI-MS,15 the resolution of all nine haloacetic acids was difficult in a reasonable run time and selectivity for the identification relied on mass spectrometry. Flow injection (FI) analysis avoids long chromatographic run times. FI-ESI-MS analysis is complicated by the presence of interfering ions, and considerations for experimental protocol have been discussed in some detail.16 Recently, in this laboratory, we investigated using the selective complexation of an analyte with another substance to increase the selectivity of the mass spectrometric identification. The perchlorate ion was complexed with cationic surfactants and organic bases to increase the selectivity of the FI-ESI-MS detection.17,18 In this manner, submicrogram per liter detection limits were achieved with analysis times on the flow injection time scale (minutes). The advantage in the mass spectrometric identification of observing the analyte as a complex is that the mass of the analyte complex can be increased to be outside the region of “chemical” noise in the mass spectrum, which is that region of the mass spectrum exhibiting interference due to naturally occurring ions, as well as ions formed by the electrospray process. In this paper, we report the determination of haloacetic acids as complexes with a suitable association-complexation agent. The optimization of the extraction procedure, the choice of complexing agent, and the optimization of the electrospray-mass spectrometry experiment are described. Method detection limits are calculated, and the technique is used to measure concentrations of HAA9 in different water matrixes. EXPERIMENTAL SECTION Reagents. HAA9 were obtained as a standard solution in methyl tert-butyl ether from Supelco (Bellafonte, PA; part 47630U). Perfluorobutyric acid [375-22-4], perfluoroheptanoic acid [37585-9], and perfluorotetradecanoic acid [376-06-7] were obtained from Oakwood Chemical Co. (West Columbia, SC). Optima methyl tert-butyl ether (MTBE) was obtained from Fisher Scientific Co. (Fairlawn, NJ), and sodium sulfate and cupric sulfate hydrate were (11) Benanou, D.; Acobas, F.; Sztajnbok, P. Water Res. 1998, 32, 2798-2806. (12) Brashear, W. T.; Bishop, C. T.; Abbas, R. J. Anal. Toxicol. 1997, 21, 300334. (13) Alaee, M.; Sergeant, D. B.; Ikonomou, M. G.; Wilkinson, R. J.; Luross, J.; Proceedings of the 46th ASMS Conference on Mass Spectrometry and Allied Topics, Orlando, FL, May 31-June 4, 1998; p 738. (14) Hashimoto, S.; Otsuki, A. J. High Resolut. Chromatogr. 1998, 21, 55-58. (15) Niessen, W. M. A. Liquid Chromatography-Mass Spectrometry; Marcel Dekker: New York, 1999. (16) Sharp, B. L.; Sulaiman, A. B.; Taylor, K. A.; Green, G. N. J. Anal. At. Spectrom. 1997, 12, 603-609. (17) Urbansky, E. T.; Magnuson, M. L.; Freeman, D.; Jelks, C. J. Anal. At. Spectrom. 1999, 14, 1861-1866. (18) Magnuson, M. L.; Urbansky, E. T.; Kelty, C. A. Anal. Chem. 2000, 72, 2529.
Table 1. Summary of Experimental Conditions for the Determination of Haloacetic Acids by Electrospray Ionization Mass Spectrometry acquisition mode selected ion-monitoring scan window scan time applied ESI spray potential interface capillary temperature (optimized) sheath gas pressure injection mode/injection volume carrier liquid/flow rate extraction solvent extraction preconcentration ratio (nominal)
negative ESI-MS on Q3 0.5 amu 0.25 s 4.0 kV 130 °C 70 psi (480 kPa) flow injection/30.0 mL methanol/0.3 mL min-1 methyl tert-butyl ether 63:1
obtained from GFS Chemical Co. (Columbus, OH). The sodium sulfate was heated in a muffle oven at 400 °C for at least 4 h prior to use. Sulfuric acid was from J.T. Baker (Phillipsburg, NJ). Deionized water was prepared by a system combining ionexchange cartridges and UV irradiation. The toxicity or carcinogenicity of each reagent used has not precisely been defined; however, each chemical compound must be treated as a potential health hazard.4 See sections 5, 14, and 15 of ref 14 for more complete safety, pollution prevention, and waste management information regarding the analysis of haloacetic acids and the handling of MTBE. Water Samples. Water samples were collected headspacefree into class 3000 cleaned 1 L bottles with PTFE-faced septa, as described in EPA Method 552.2.4 Prior to collection, 100 mg of ammonium chloride was added to each bottle. This amount of ammonium chloride serves to dechlorinate the water. Apparatus. Injections were made with a Rheodyne (Rohnert Park, CA) model 7725 injector with a 200 mL loop. A Waters (Milford, MA) 600 pump was used for the carrier liquid. The mass spectrometer was a Finnigan (San Jose, CA) MAT TSQ-700 equipped with a Finnigan electrospray interface. Mass spectra were acquired in the negative-ion mode by scanning Q3 over appropriate mass ranges. Other experimental conditions are listed in Table 1. Procedure. Solutions of perfluoro acids were prepared in MTBE and were refrigerated when not in use. Aqueous solutions were extracted according to the following procedure. Initially, a separatory funnel was used to separate the reaction mixture. The separatory funnel tended to allow a small quantity of water into the solvent phase. The large amounts of sulfuric acid and sodium sulfate in this water interfered with the electrospray process in a nonreproducible manner. In the final procedure, 188 mL (graduated cylinder) of the water sample was added to a 250 mL cleaned and muffled clear Boston round screw cap bottle. The water was acidified with 20 mL (repeating dispenser) of 50% sulfuric acid. A 58 g quantity of sodium sulfate and 7 g of copper sulfate pentahydrate were added. The bottle was capped with a PTFE-lined septum and inverted for 1 min. This allowed the sodium and copper sulfates to dissolve sufficiently so that an additional 20 mL of 50% sulfuric acid could be added. The bottle was recapped and shaken for an additional 4 min. Three milliliters of MTBE (repeating dispenser) was added, and the capped bottle was shaken vigorously for 2 min. Five minutes was allowed for separation. This procedure permitted the MTBE layer to be drawn with a Pasteur pipet from the neck of Analytical Chemistry, Vol. 72, No. 10, May 15, 2000
2309
Figure 1. Peak area of tribromoacetic acid versus the concentration of perfluoroheptanoic acid (13F) added to methyl tert-butyl ether.
the Boston round bottle without any contamination from the aqueous solution. The extract was transferred to a 1.8 mL glass vial. A syringe was used to precisely transfer 500 µL of the extract to a second 1.8 mL glass vial. A precisely measured (10 µL syringe) quantity of solution containing perfluoroheptanoic acid (13F) was added to the MTBE extract to produce the desired 13F concentration. A 30 µL portion of the resulting solution was then injected into the electrospray mass spectrometer. RESULTS AND DISCUSSION Complexation Agent. The ESI-MS analysis involves the formation of a stable association complex of the haloacetic acid during the electrospray ionization. Undesirable complexes of analytes with other species are commonly formed in the electrospray process, sometimes from solvents or contaminants. This phenomenon has been put to constructive use in the analysis of chromium species19 and the perchlorate ion17,18 through the correct choice of complexing agent. Carboxylic acids were investigated as complexation agents because it was reported that, in the electrospray of haloacetic acids using an acetic acid eluent, complexes of the haloacetic acids with acetate were observed.15 Acetic acid was initially investigated as the complexation agent, but the complexes had poor sensitivity. Difluoroacetic acid produced complexes with better sensitivity than acetic acid, and it was surmised that fluorinated acids might produce superior results. Since one of the goals of forming the haloacetic acid complexes is to increase the mass of the ion used to quantitate the haloacetic acid, large perfluoro carboxylic acids were investigated: perfluorobutanoic acid (7F), perfluoroheptanoic acid (13F), and perfluorotetradecanoic acid (23F). Haloacetic acid complexes of 23F had low sensitivity, compared to those of 7F and 13F. The complexes of 7F, however, had insufficient mass to avoid the noisy region of the mass spectrum. Therefore, 13F was selected as the complexing agent. The amount of 13F to be added to the solutions was investigated. The peak area of a constant concentration of tribromoacetic acid as a function of the concentration of 13F in solution is shown in Figure 1. The analyte response decreases 6-fold as the 13F concentration is increased from 1 to 200 mg/L. The decrease in peak area is probably due to suppression of the electrospray signal15 when the ionic strength increases with increasing 13F concentration. Higher concentrations were not investigated, to (19) Gwizdala, A. B.; Johnson, S. K.; Mollah, S.; Houk, R. S. J. Anal. At. Spectrom. 1997, 12, 504-506.
2310 Analytical Chemistry, Vol. 72, No. 10, May 15, 2000
Figure 2. Peak area of tribromoacetic acid versus the temperature of the interface capillary of the electrospray mass spectrometer.
avoid further suppression of the analyte response. A concentration of 50 mg/L was selected because this is expected to be ∼1000 times greater than typical haloacetic acid concentrations (
