Hydrogen Abstraction and Decomposition of Bromopicrin and Other

Figure 1 GC/MS chromatograms of bromopicrin analyzed at a GC injection port ... GC/MS analyses were performed on a VG 70-SEQ mass spectrometer equippe...
0 downloads 0 Views 166KB Size
Environ. Sci. Technol. 2002, 36, 3362-3371

Hydrogen Abstraction and Decomposition of Bromopicrin and Other Trihalogenated Disinfection Byproducts by GC/MS PAUL H. CHEN AND SUSAN D. RICHARDSON* U.S. Environmental Protection Agency, National Exposure Research Laboratory, Athens, Georgia 30605 STUART W. KRASNER Metropolitan Water District of Southern California, LaVerne, California 91750 GEORGE MAJETICH Department of Chemistry, University of Georgia, Athens, Georgia 30602 GARY L. GLISH Department of Chemistry, University of North Carolina, Chapel Hill, North Carolina 27599

Tribromonitromethane (bromopicrin), dibromochloronitromethane, bromodichloronitromethane, and trichloronitromethane (chloropicrin) have been identified as drinking water disinfection byproducts (DBPs). They are thermally unstable and decompose under commonly used injection port temperatures (200-250 °C) during gas chromatography (GC) or GC/mass spectrometry (GC/MS) analysis. The major decomposition products are haloforms (such as bromoform), which result from the abstraction of a hydrogen atom from the solvent by thermally generated trihalomethyl radicals. A number of other products formed by radical reactions with the solvent and other radicals were also detected. The trihalonitromethanes also decompose in the hot GC/MS transfer line, and the mass spectra obtained are mixed spectra of the undecomposed parent compound and decomposition products. This can complicate the identification of these compounds by GC/MS. Trihalomethyl compounds that do not have a nitro group, such as tribromoacetonitrile, carbon tetrabromide, methyl tribromoacetate, and tribromoacetaldehyde, do not decompose or only slightly decompose in the GC injection port and GC/MS transfer line. The brominated trihalomethyl compounds studied also showed H/Br exchange by some of their fragment ions. This H/Br exchange also makes the identification of these compounds in drinking water more difficult. The extent of H/Br exchange was found to depend on the mass spectrometer ion source temperature, and it is proposed that the internal surface of the ion source is involved in this process.

Introduction Disinfection byproducts (DBPs) are formed during drinking water treatment when a disinfectant, such as chlorine or * Corresponding author phone: (706)355-8304; fax: (706)355-8302; e-mail: [email protected]. 3362

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002

ozone, reacts with organic matter and/or bromide that is naturally present in the water. Rook identified the first DBP from the chlorination of drinking water (chloroform) in 1974 (1), and since then, a number of DBPs have been reported for chlorine and other disinfectants (2). Many DBPs are trihalomethyl compounds, which include compounds such as chloroform, bromoform, bromodichloroacetic acid, trichloroacetaldehyde (chloral hydrate), bromodichloroacetonitrile, 1,1,1-trichloropropanone, and trichloronitromethane (2). Gas chromatography/mass spectrometry (GC/MS) has been the primary analytical tool used to identify DBPs in drinking water, and it continues to be widely used for identifying and quantifying DBPs in drinking water (2, 3). Recently, we discovered that the GC/mass spectra of some of the trihalomethyl compounds (particularly those containing tribromomethyl groups) show unusual [fragment + 1]+ ions in their mass spectra, which distort their mass spectra, causing the spectra to be different from those in the mass spectral library databases (NIST and Wiley). A few of these trihalomethyl compounds also partially decompose in the GC/ MS transfer line (which is generally heated to the same or higher temperature as the highest temperature of the GC program), and the mass spectra become a mixture of the undecomposed parent compound and the decomposition products. These effects can complicate or prevent the identification of these compounds in drinking water by GC/ MS. The following trihalomethyl compounds were investigated: tribromonitromethane (bromopicrin), dibromochloronitromethane, bromodichloronitromethane, trichloronitromethane (chloropicrin), tribromoacetonitrile, carbon tetrabromide, methyl tribromoacetate, and tribromoacetaldehyde. All of these compounds are either DBPs or contaminants that have been identified in drinking water (methyl tribromoacetate is the methylation product of tribromoacetic acid, which requires methylation for successful analysis by GC/MS). Bromopicrin can be formed in chlorinated (4) and ozonated (5) drinking water as well as in waters treated with a combination of ozone and chlorine (6, 7). Recently conducted toxicity studies are revealing that bromopicrin and other bromonitromethanes are extremely cytotoxic and genotoxic to mammalian cells (8, 9). Because of their potential adverse health effects, bromopicrin and other halonitromethanes are included in a current nationwide occurrence study of DBPs (10, 11). As part of this study, bromopicrin, dibromochloronitromethane, and bromodichloronitromethane were found at 2.5, 1.5, and 0.7 µg/L, respectively, in finished water from a California drinking water treatment plant that used pre-ozone and post-chlorine/chloramine disinfection on a high-bromide source water (0.14 mg/L) (11). Bromopicrin is thermally labile and decomposes at commonly used GC injection port temperatures (200-250 °C) (5), largely due to a low bond dissociation energy for the CBr3-NO2 bond (approximately 40 kcal/mol) (12). Thermolysis of bromopicrin generates free radicals (12), such as •CBr , •NO , and •Br, which can undergo reactions with the 3 2 solvent or combine with other radicals to form reaction products in the GC injection port (e.g., bromoform, carbon tetrabromide). Three other trihalonitromethanessdibromochloronitromethane, bromodichloronitromethane, and chloropicrinsalso decompose readily in the GC injection port at commonly used temperatures. The objectives of this investigation were as follows: (i) identify and study the decomposition/reaction products formed in the GC injection port; (ii) determine the impact 10.1021/es0205582 CCC: $22.00

 2002 American Chemical Society Published on Web 06/26/2002

TABLE 1. Formulas and Molecular Weights for Halogenated Compounds compound

formula

mol wt (monoisotopic)

bromopicrin dibromochloronitromethane bromodichloronitromethane chloropicrin tribromoacetonitrile carbon tetrabromide methyltribromoacetate tribromoacetaldehyde

CBr3NO2 CBr2ClNO2 CBrCl2NO2 CCl3NO2 CBr3CN CBr4 CBr3CO2CH3 CBr3CHO

295 251 207 163 275 328 308 278

of decomposition in the GC/MS transfer line on the mass spectrum for each compound; (iii) study the unusual [fragment + 1]+ behavior in the mass spectrum of each compound and determine the mechanism; and (iv) study the impact of these processes on real drinking water analyses.

Experimental Section Standard and Sample Preparation. Bromopicrin standard solutions were prepared in ethyl acetate, acetone, and methylene chloride at 50 µg/mL for the study of its decomposition products in the GC injection port. In addition to bromopicrin, standard solutions for other trihalomethyl compounds (50 µg/mL) were prepared in ethyl acetate and carbon tetrachloride for the study of H/Br exchange in the ion source and decomposition in the transfer line. These standards included dibromochloronitromethane, bromodichloronitromethane, chloropicrin, tribromoacetonitrile, carbon tetrabromide, methyl tribromoacetate, and tribromoacetaldehyde. Molecular formulas and weights are provided in Table 1 for reference. Bromopicrin solutions were also prepared in methanol, hexane, and benzene (at 50 µg/ mL) to determine whether H/Br exchange occurs in these solvents. Water samples for GC/MS analysis were treated

with ozone and chlorine and concentrated on XAD resins, followed by elution with ethyl acetate as described in detail elsewhere (6). Bromopicrin Measurements for Treated Drinking Water Samples. Water samples for GC analysis and quantitation involved the extraction of 30 mL of water with 3 mL of methyl tert-butyl ether (MTBE) in the presence of 11 g of sodium sulfate and 1 g of copper sulfate (11). Salt was added to improve the partitioning of bromopicrin from the aqueous phase to the solvent. The copper sulfate was needed to facilitate decanting a small amount of extract from the water layer. Common dechlorination agents (e.g., ascorbic acid) destroy bromopicrin (11). Therefore, ammonium chloride (100 mg/L) was used to quench free chlorine residuals for bromopicrin samples. A pH of 3.5 or 4 was required to minimize base-catalyzed hydrolysis of the bromopicrin (5, 11). A few drops of a 0.25 or 0.5 M sulfuric acid solution were used to adjust the pH of the sample (10). Chemical Standards. Synthetic standards of bromodichloronitromethane, dibromochloronitromethane, and bromopicrin were purchased from CanSyn (Toronto, Ontario, Canada; currently available from Helix Biotech, New Westminster, BC, Canada). The bromopicrin used for the GC analysis was synthesized by Columbia Organic Chemical Co., Inc. (Camden, SC). A chemical standard of tribromoacetonitrile was prepared synthetically in two steps by reacting commercially available tribromoacetaldehyde (Aldrich) with hydroxylamine and dehydrating the resulting oxime using Mexican Bentonite and microwave irradiation (13). All other chemicals were purchased at the highest level of purity from Aldrich, Fisher, Chem Service, or Supelco. GC/MS Analysis. GC/MS with electron ionization (EI) was used for most of the experiments presented in this paper: study of decompositions in the GC injection port, decompositions in the GC/MS transfer line, and H/Br exchange in the mass spectrometer source. GC with electron capture detection (ECD) was used for the quantitative experiments

FIGURE 1. GC/MS chromatograms of bromopicrin analyzed at a GC injection port temperature of 250 °C in different solvents. Peak A, artifact products from the aldol condensation of acetone [the second peak represents the initial acetone condensation product, CH3COCH2C(OH)(CH3)2; the first peak represents its dehydration product, CH3COCHdC(CH3)2]. VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3363

SCHEME 1

involving actual drinking water samples (experimental details in the next section). GC/MS analyses were performed on a VG 70-SEQ mass spectrometer equipped with a HP 5890A gas chromatograph. Samples were analyzed at GC injection port temperatures of 250 °C for normal routine analyses and 170 or 140 °C for lower temperature analyses. A GC/MS transfer line temperature of 290 °C was employed for normal routine analyses, while 225 °C was used for lower temperature analyses (where compounds were not expected to decompose). The mass spectrometer ion source temperature was maintained at 200 °C, except when the effect of source temperature on H/Br exchange was studied. In this case, the source temperature was varied from 130 to 290 °C, and the transfer line was maintained at the lower temperature (225 °C). A J&W Scientific (Agilent) DB-5 column (30 m, 0.25 mm i.d., 0.25 µm film thickness) was used, with the following GC temperature program: initial temperature of 35 °C held for 4 min, followed by a rate increase of 9 °C/min to 285 °C. When comparing the differences in GC injection port decompositions between different mass spectrometers, the mass spectrometer ion sources were maintained at the same temperature (200 °C). GC Analysis. A GC (model 3600; Varian Analytical Instruments) equipped with a split/splitless injector, two electron capture detectors, and an autosampler (model 8100) was used for the quantitative analyses. A 30 m × 0.25 mm i.d. DB-1 (1-µm film thickness) and a 30 m × 0.25 mm i.d. DB-5 (1-µm film thickness) capillary column (J&W ScientificAgilent) were used as the primary (analytical) and secondary (confirmation) columns, respectively. The temperature program was as follows: initial temperature of 35 °C for 23 min; increase to 139 °C at 4 °C/min and hold at 139 °C for 0 min; increase to 301 °C at 27 °C/min and hold at 301 °C for 5 min. Bromopicrin eluted at ∼130 °C with this temperature program. The injector temperature was set at 117 or 87 °C 3364

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002

to prevent the degradation of thermally labile compounds such as bromopicrin (5, 11).

Results and Discussion Decomposition in the GC Injection Port. When bromopicrin was analyzed at a GC injection port temperature of 140 °C, no significant decomposition was observed in any of the three solvents studied (ethyl acetate, acetone, and methylene chloride). At 250 °C, extensive decompositions occurred (Figure 1). The major decomposition product was always bromoform, which was likely formed by hydrogen abstraction from solvents by the tribromomethyl radical (14). An analysis of bromopicrin in deuterated ethyl acetate (ethyl-d5 acetated3) supported this hydrogen abstraction mechanism, as deuterium was incorporated into the bromoform formed under these conditions. A number of other products formed by radical reactions with solvents and by radical recombinations were also detected. For ethyl acetate, these included ethyl nitroacetate, an unidentified compound, carbon tetrabromide, hexabromoethane, and tetrabromoethene (Scheme 1). In acetone, the reaction products included bromoacetone, nitroacetone, tetrabromoethene, hexabromoethane, carbon tetrabromide, and 2,5-hexanedione; while in methylene chloride they included dichloronitromethane, 3-bromocyclohexene, 1,1,2,2-tetrachloroethane, tetrabromoethene, hexabromoethane, and carbon tetrabromide. 3-Bromocyclohexene is likely formed by the reaction of a bromine radical with cyclohexene, which is a preservative added to methylene chloride (15, 16). It should be noted that the dibromonitromethane evident in Figure 1 was mainly due to its impurity present in the bromopicrin standard, and it was not formed substantially by decomposition of bromopicrin (dibromonitromethane was still observed at the lower injection port temperatures where bromopicrin does not decompose). As evident in Figure 1, bromopicrin

extract would be small unless a large amount of bromopicrin was present in the water.

FIGURE 2. Effect of GC/MS transfer line temperature on the decomposition of bromopicrin; m/z 252 (CHBr3•+) is the molecular ion (isotopic peak) for bromoform; m/z 251 (CBr3+) is the base peak for bromopicrin. undergoes greater hydrogen abstraction decomposition in methylene chloride (forming higher quantities of bromoform) than in ethyl acetate and acetone. This is most likely due to the greater polarity of methylene chloride, which would generate a more stable secondary free radical, as compared to a primary free radical with ethyl acetate or acetone. Methylene chloride is a widely used solvent for liquid-liquid extraction, and if this solvent is used for drinking water analysis by GC/MS, the chance of seeing bromopicrin in the

We also tested two additional solventssMTBE and pentanesbecause they are used in EPA Method 551 (for the determination of chlorinated DBPs and other analytes) (17). As with the other solvents, significant decomposition of bromopicrin was observed with these solvents at a GC injection port temperature of 250 °C, whereas minimal decomposition was observed at 140 °C. In each case, the major decomposition product was bromoform (as with the other solvents tested), and radical reaction products were also observed. For MTBE, the reaction products included bromoform, carbon tetrabromide, tetrabromoethene, hexabromoethane, and a compound tentatively identified as bromomethyl tert-butyl ether. For pentane, the reaction products included bromoform, 1-bromopentane, carbon tetrabromide, tetrabromoethene, and hexabromoethane. The extent of decomposition was also found to vary with the type of mass spectrometer. When these bromopicrin solutions were analyzed using different mass spectrometers (two different magnetic sector mass spectrometers and a quadrupole mass spectrometer), significant variations were observed in the decomposition of bromopicrin at 250 °C. For example, for the bromopicrin samples dissolved in MTBE, one magnetic sector mass spectrometer (a VG 70 SEQ) gave a ratio of approximately 5:1 bromopicrin to bromoform, and another magnetic sector instrument (a Micromass Autospec) gave a ratio of approximately 2:1 bromoform to bromopicrin (showing much greater decomposition). Thus, the extent of decomposition of bromopicrin using different instruments can vary. We noticed that small air leaks in the GC and the condition of the injection port liner (buildup of charred deposits) can significantly affect the extent of decomposition. Therefore, we suspect surface reactions may be involved in

FIGURE 3. EI mass spectrum of bromopicrin analyzed at high (a) (290 °C) and low (b) (225 °C) GC/MS transfer line temperatures producing a mixed and pure spectrum, respectively. Pure EI mass spectrum of bromoform (c) provided for comparison. VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3365

FIGURE 4. EI mass spectrum of chloropicrin analyzed at high (a) (290 °C) and low (b) (225 °C) GC/MS transfer line temperatures producing a mixed and pure spectrum, respectively. Pure EI mass spectra of hexachloroethane (c) and chloroform (d) provided for comparison. these injection port reactions, and it is likely that differences observed between GC/mass spectrometers may be due to differences in the GC injection ports (including their cleanliness and the amount of air present). Decomposition of Bromopicrin at Different GC Injection Port Temperatures. The extent of decomposition of bromopicrin in ethyl acetate at different GC injection port temperatures (140-280 °C) was studied by monitoring the formation of bromoform using GC/MS. The results showed that decomposition of bromopicrin below 170 °C was negligible. At 180 °C, a small amount of bromoform was detected, while at 250 and 280 °C there was significant formation of bromoform. At 250 °C, bromoform was approximately 70% the level of bromopicrin (as judged by chromatographic peak areas); at 280 °C, bromoform became the dominant peak at a ratio of 6:1 to bromopicrin. GC injection port temperatures of 250 and 280 °C were included because the former is a commonly applied injection port temperature for the analysis of semivolatile organic compounds, while the latter is usually the highest injection port temperature employed in the routine analysis of the semivolatile compounds. Decomposition of Other Trihalonitro Compounds. Trihalonitro compounds other than bromopicrin, such as dibromochloronitromethane, bromodichloronitromethane, and chloropicrin, also decompose extensively (about 50%) at an injection port temperature of 250 °C. The major decomposition products are as follows: For dibromochloronitromethane, the products are dibromochloromethane and a small amount of tribromochloromethane; for bromodichloronitromethane, the major products are bromodichloromethane, dibromodichloromethane, and two unidentified products from the radical reaction with solvent 3366

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002

(ethyl acetate); for chloropicrin, the major reaction products are chloroform and hexachloroethane. Most of these compounds are formed by thermal cleavage of the trihalomethyl and nitro bond to generate the trihalomethyl radical, which undergoes hydrogen abstraction (18, 19), bromine abstraction (20), or radical recombination (21) to form the final products. Decomposition of Other Tribromo Compounds. Tribromomethyl compounds that do not have a nitro group, such as tribromoacetonitrile, methyl tribromoacetate, tribromoacetaldehyde, and carbon tetrabromide, do not decompose or only decompose slightly (less than 5%) at a GC injection port temperature of 250 °C. The major products are bromoform and carbon tetrabromide. Bromoform is likely formed by the abstraction of a hydrogen from the solvent (ethyl acetate) by the CBr3 radical, while carbon tetrabromide is likely formed by either combination of CBr3 and Br radicals (21) or by the abstraction of Br by the CBr3 radical (20). Decomposition in the GC/MS Transfer Line. Thermal decompositions in the GC/MS transfer line were also observed for trihalogenated compounds. Bromopicrin decomposed in the GC/MS transfer line as temperatures were varied from 225 to 290 °C. Because bromopicrin decomposes only within a fraction of a second before it enters the ion source, bromopicrin and its decomposed products enter the ion source at roughly the same time (there is no separation in time). In other words, a mixed spectrum of the parent and the decomposed compound will be obtained. The degree of decomposition was measured by monitoring the appearance of bromoformsa major decomposition productsin the mass spectrum of bromopicrin. The results are shown in Figure 2. The degree of decomposition to bromoform is expressed by monitoring the area ratio of m/z 252 (molecular ion isotope peak for bromoform) and m/z 251 (base peak for bromo-

FIGURE 5. EI mass spectrum of bromopicrin analyzed in (a) ethyl acetate and (b) carbon tetrachloride. Extensive H/Br exchange is evident in panel a. picrin). At 225 and 240 °C, the ratios are 0.0107 and 0.0113, respectively. These ratios are very close to the 13C/12C isotopic ratio of 0.011 for an ion such as CBr3+, which contains only one carbon. This means that at 225 and 240 °C, the decomposition of bromopicrin to bromoform is negligible. As the temperature increased from 250 to 290 °C, the decomposition to bromoform becomes more significant. The mass spectrum obtained at a transfer line temperature of 225 °C yielded a spectrum totally due to bromopicrin (Figure 3b). When the sample was analyzed at a transfer line temperature of 290 °C, the mass spectrum obtained (Figure 3a) is a mixed spectrum of undecomposed bromopicrin and bromoform. This mixed mass spectrum could make it more difficult to identify bromopicrin in drinking water if one is not aware of this decomposition phenomenon. When comparing the decomposition in the GC/MS transfer line to the decompositions occurring in the GC injection port, those in the transfer line are not as great (bromoform was the only decomposition product noted for bromopicrin in the GC/MS transfer line, whereas several decomposition products were observed in the GC injection port, Figure 1). This is likely due to the fact that bromopicrin has less time to react in the transfer line (smaller volume and high flow rate) and to the lack of reactive species (such as the solvent) present with bromopicrin in the transfer line (the solvent and other potential reactive species have previously eluted from the column). Mass spectra of other trihalonitromethanes also showed a mixed spectrum of their parent compound and their decomposed products (Figure 4 and Supporting Information). Chloropicrin demonstrated the most pronounced effect by forming two decomposition productsshexachloroethane and chloroform (Figure 4). Its resulting mass spectrum exhibited ions at m/z 83, 85, and 87 that are due to the CHCl2+ ion of

chloroform and ions at m/z 94/96/98, 164/166/168/170, and 199/201/203/205/207 that are due to the C2Cl2•+, C2Cl4•+, and C2Cl5+ ions, respectively, from the overlapping hexachloroethane spectrum (Figure 4). Similarly, the mass spectrum of dibromochloronitromethane showed a mixed spectrum of dibromochloronitromethane and dibromochloromethane, and the mass spectrum of bromodichloronitromethane showed a mixed spectrum of bromodichloronitromethane and bromodichloromethane (Supporting Information). It should be noted that the extent of decomposition depends not only on the temperature but also on the length of the transfer line, which varies from instrument to instrument. The mass spectra of tribromoacetonitrile, methyl tribromoacetate, carbon tetrabromide, and tribromoacetaldehyde obtained at a transfer line temperature of 290 °C were very similar to those obtained at 225 °C, indicating that decomposition of these compounds at the transfer line temperature of 290 °C is negligible for these compounds that do not contain a nitro group. Mass Spectrometry of Bromopicrin and Other Trihalo Compounds. Bromopicrin contains no hydrogen, but its mass spectrum shows peculiar [fragment + 1]+ ions for the fragment ions CBr2•+, Br+, and CBr+, and to a lesser extent, for CNOBr•+ (Figure 5a). In addition to a small (1.1%) contribution from the 13C isotopes, these [fragment + 1]+ ions are most likely due to a H/Br exchange, generically shown in eq 1:

CClxBry+ + H• a CClxBry-1H+ + Br•

(1)

where x ) 0-2, y ) 1-3, and x + y e 3. A hydrogen abstraction mechanism, like the one occurring in the GC injection port (before ionization in the mass spectrometer), is unlikely in this situation because there were cases where the same VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3367

FIGURE 6. EI mass spectrum of bromodichloronitromethane analyzed in (a) ethyl acetate and (b) carbon tetrachloride. fragment ion (e.g., CClBr•+) in the spectra of two different compounds (e.g., dibromochloronitromethane and bromodichloronitromethane) exhibited different behavior (one compound showed the [fragment + 1]+ ion and the other did not). Thus, the behavior cannot be explained by a hydrogen abstraction by the CClBr•+ ion. However, an exchange of a H• from the solvent (or another source of H•) with a Br• from CClBr2+ could explain this behavior. Thermochemically, the reaction shown would be 41 kcal/mol exothermic, and the same reaction with Cl instead of Br would be 10 kcal/mol exothermic (22). If reactions between the ion and the solvent are considered, covalent bonds are involved, and the Cl-based reaction would become endothermic and thus not seen (as with CCl2BrNO2, Figure 6), while the Br reaction would remain exothermic and observed (as with CBr2ClNO2, Figure 7). This H/Br exchange was observed in the bromopicrin standard solutions prepared in ethyl acetate, methylene chloride, acetone, methanol, hexane, and benzene. Available thermochemical values support the proposition that the Br exchange is exothermic, while Cl exchange is endothermic (22). Carbon tetrachloride was the only solvent tested that did not show this behavior (Figure 5b). This was not unexpected as carbon tetrachloride has no hydrogens in its structure that could be exchanged. However, the injection of carbon tetrachloride had a profound effect on the H/Br exchange of subsequent analyses of bromopicrin in ethyl acetate. After carbon tetrachloride was introduced into the GC/mass spectrometer (for the analysis of bromopicrin in carbon tetrachloride), the next few injections of bromopicrin prepared in ethyl acetate showed little or no H/Br exchange. This suggests that carbon tetrachloride replaced surfacebound or adsorbed species in the ion source that provided hydrogen for the H/Br exchange and that carbon tetrachloride remained in the ion source for a considerable length of time 3368

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002

(few hours). This provides further support that the internal surfaces of the ion source are probably involved in this process. Others have previously observed similar surfaceinduced reactions in the ion source of mass spectrometers (23-26). The extent of H/Br exchange by the CBr3+ ion as expressed by the ratio of CHBr2+/CBr3+ was studied at different ion source temperatures. As shown in Figure 8, the lower the source temperature, the greater the H/Br exchange. This is consistent with the suggestion that the internal surface of the source is involved (23). In addition to the source temperature, the extent of H/Br exchange also depends on the other source conditions. For example, if the GC injection port liner was changed or if other GC column maintenance was performed (such as clipping off part of a guard column or replacing the GC column), the extent of H/Br exchange increases. This can be explained by the fact that as the column head is exposed to atmosphere (during such GC maintenance), air enters the GC column and the mass spectrometer ion source, and it can oxidize the metal surface of the hot ion source. It is likely that this oxidized ion source surface facilitates the surface-induced H/Br exchange. This behavior was observed for other halogenated compounds, including tribromoacetonitrile, carbon tetrabromide, methyl tribromoacetate, and tribromoacetaldehyde (Supporting Information). For example, extensive H/Br exchanges were observed when ethyl acetate was used as the solvent, but no H/Br exchanges occurred when carbon tetrachloride was used as the solvent. It should be noted that tribromoacetaldehyde itself already has a hydrogen in its structure, so the [fragment + 1]+ ions could also have a contribution from intramolecular hydrogen transfers in addition to H/Br exchange.

FIGURE 7. EI mass spectrum of dibromochloronitromethane analyzed in (a) ethyl acetate and (b) carbon tetrachloride.

FIGURE 8. Effect of source temperature on H/Br exchange by CBr3+.

Impact on Drinking Water Analyses Impact of GC Injection Port Temperature. To study the GC thermal decomposition of bromopicrin in a real drinking water extract, ethyl acetate extracts of ozone-chlorinetreated drinking water were spiked with known amounts of bromopicrin and dibromonitromethane and analyzed at GC injection port temperatures of 140 and 250 °C. When the injection port temperature was increased from 140 to 250 °C, the quantity of bromopicrin decreased 45% (Figure 9), accompanied by an increase in bromoform. The small, insignificant change in the dibromonitromethane chromatographic peak is likely due to a slight formation of dibromonitromethane as a very minor decomposition product of bromopicrin. Radical recombination products of bromopicrin were also detected in the drinking water extracts, including carbon tetrabromide, hexabromoethane, and tet-

rabromoethene, in samples analyzed at 250 °C. Ethyl nitroacetate, one of the decomposition products of bromopicrin in ethyl acetate, was not detected in the drinking water extracts. This compound is polar, and its detection by GC/ MS likely depends on the injection port and column conditions and possibly on the sample matrix. Impact of the GC/MS Transfer Line. For trihalonitromethanes, decompositions in the GC/MS transfer line have obvious implications for their analysis in real drinking water samples. As discussed earlier, if the transfer line is maintained at temperatures of 250 °C or higher, significant decompositions occur and result in mixed mass spectra of the parent compound and its decomposition product(s). These mixed mass spectra could result in a poor library match when analyzing a real drinking water extract and could result in a misidentification or a lack of identification. Therefore, for drinking water analyses, we would recommend lowering the GC/MS transfer line to 225 °C to prevent these decompositions. As mentioned earlier, this effect is not as pronounced for other trihalogenated compounds that do not contain a nitro group. Impact of H/Br Exchange in the Mass Spectrometer. The H/Br exchange behavior in the mass spectrometer has implications for the analysis of real drinking water sampless both for the identification and for the quantification of DBPs and other drinking water contaminants. As discussed, the solvent can have an effect on the occurrence or extent of H/Br exchange, and the introduction of a different solvent (such as carbon tetrachloride) can change the extent of exchange for subsequent analyses. Ion source conditions can affect the H/Br exchange, and even GC maintenance can affect the extent of exchange (through the introduction of air into the mass spectrometer). As a result, the extent of H/Br exchange will be highly variablesit can vary from instrument to instrument and from day to day on the same VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3369

FIGURE 9. GC/MS chromatograms of a real drinking water extract spiked with bromopicrin and dibromonitromethane. Top: 140 °C GC injection port; bottom: 250 °C GC injection port. instrumentsand the resulting mass spectra could pose a problem for an inexperienced analyst in identifying trihalogenated compounds using library database matching. In conclusion, instrumental and analytical conditions can have a significant effect on analysis of bromopicrin and other trihalogenated compounds. Typically used GC injection port temperatures (200-250 °C) can thermally decompose trihalonitro compounds and produce a suite of decomposition products. Typically used GC/MS transfer line temperatures (250-290 °C) can also thermally decompose these compounds and produce mass spectra that are mixtures of the parent compound and decomposition products. Finally, H/Br exchange can occur in the mass spectra of trihalonitro and other trihalogenated compounds, and this phenomenon is variable depending on the mass spectrometer ion source conditions, solvent, and GC maintenance. These phenomena can impact the identification and quantification of DBPs and other contaminants in drinking water samples. For example, the formation of bromoform as a dominant decomposition product of bromopicrin (with GC injection port temperatures higher than 170 °C) could contribute to overestimations of bromoform concentrations in drinking water samples, and the presence of bromopicrin may go undetected, even when it is there at appreciable levels (µg/ L) in the original drinking water sample. With a low GC injection port temperature, both bromoform (3 µg/L) and bromopicrin (2.5 µg/L) were detected and quantified in finished water from a California drinking water treatment plant. Bromopicrin and other brominated nitromethanes are becoming important DBPs, as early screening studies are showing them to be more genotoxic (in mammalian and bacterial cells) than many of the DBPs that are currently regulated. The analyst needs to be aware of these potential 3370

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002

problems for drinking water analyses and alter the analytical and instrumental conditions when necessary.

Acknowledgments The authors gratefully acknowledge the assistance of Terry Floyd and Al Thruston of the U.S. EPA with GC/MS analyses and Russell Chinn of Metropolitan Water District for developing the GC method for the analysis of bromopicrin. This paper has been reviewed in accordance with the U.S. Environmental Protection Agency’s peer and administrative review policies and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use by the U.S. EPA.

Supporting Information Available Table and six figures. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Rook, J. J. Water Treat. Exam. 1974, 23 (2), 234. (2) Richardson, S. D. Drinking Water Disinfection Byproducts. In The Encyclopedia of Environmental Analysis and Remediation; John Wiley & Sons: New York, 1998; Vol. 3, p 1398. (3) Richardson, S. D. J. Environ. Monit. 2001, 3, 1. (4) Thibaud, H.; De Laat, J.; Dore´, M. Water Res. 1988, 22, 381. (5) Krasner, S. W.; Chinn, R.; Hwang, C. J.; Barrett, S. E. Proceedings of the 1990 American Water Works Association Water Quality Technology Conference; American Water Works Association: Denver, CO, 1991. (6) Richardson, S. D.; Thruston, A. D., Jr.; Caughran, T. V.; Chen, P. H.; Collette, T. W.; Floyd, T. L.; Schenck, K. M.; Lykins, B. W., Jr.; Sun, G.-R.; Majetich, G. Environ. Sci. Technol. 1999, 33, 3368. (7) Richardson, S. D.; Thruston, A. D., Jr.; Caughran, T. V.; Chen, P. H.; Collette, T. W.; Floyd, T. L.; Schenck, K. M.; Lykins, B. W., Jr.; Sun, G.-R.; Majetich, G. Environ. Sci. Technol. 1999, 33, 3378.

(8) Plewa, M. J.; Kargalioglu, Y.; Vankerk, D.; Minear, R. A.; Wagner, E. D. Water Sci. Technol. 2000, 42, 109. (9) Kargalioglu, Y.; Wagner, E. D.; Richardson, S. D.; Minear, R. A.; Plewa, M. J. Environ. Mol. Mutagen. 2000, 35 (35), Suppl. 31. (10) Gonzalez, A. C.; Krasner, S. W.; Weinberg, H.; Richardson, S. D. Proceedings of the 2000 American Water Works Association Water Quality Technology Conference; American Water Works Association: Denver, CO, 2000. (11) Krasner, S. E.; Pastor, S.; Chinn, R.; Sclimenti, M. J.; Weinberg, H. S.; Richardson, S. D.; Thruston, A. D., Jr. Proceedings of the 2001 American Water Works Association Water Quality Technology Conference; American Water Works Association: Denver, CO, 2001. (12) Kumskov, M. I.; Peshkova, S. E.; Ponomareva, L. A.; Rezchikova, K. I. Russ. Chem. Bull. 1996, 45 (8), 1840. (13) Delgado, F.; Cano, A. C.; Garcfa, O.; Alvarado, J.; Velasco, L.; Alvarez, C.; Rudler, H. Synth. Commun. 1992, 22, 2125. (14) Wampler, F. B.; Kuntz, R. R. Int. J. Chem. Kinet. 1971, 3, 283. (15) Campbell, J. A.; Lapack, M. A.; Peters, T. L.; Smock, T. A. Environ. Sci. Technol. 1987, 21, 110. (16) Chen, P. H.; VanAusdale, W. A.; Roberts, D. F. Environ. Sci. Technol. 1991, 25, 540. (17) EPA Method 551.1. Determination of Chlorination Disinfection Byproducts, Chlorinated Solvents, and Halogenated Pesticides/ Herbicides in Drinking Water by Liquid-Liquid Extraction and

(18) (19) (20) (21) (22) (23) (24) (25) (26)

Gas Chromatography with Electron-Capture Detection; U.S. EPA, National Exposure Research Laboratory, Office of Research and Development: Cincinnati, OH, 1995. Gershuni, S.; Itzhak, N.; Rabani, J. Langmuir 1999, 15 (4), 1141. Hunter, W. H.; Edgar, D. E. J. Am. Chem. Soc. 1932, 54, 2025. Nelson, E. D.; Tichy, S. E.; Kentta¨maa, H. I. J. Chem. Soc., Perkin Trans. 2, 1999, 2267. Fossey, J.; Lefort, D.; Sorba, J. Free Radicals in Organic Chemistry; John Wiley & Sons: New York, 1995; Chapter 10. Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17 (Suppl. 1), 1-861. Van Berkel, G. J.; Glish, G. L.; McLuckey, S. A.; Tuinman, A. A. J. Am. Chem. Soc. 1989, 111, 6027. Budzikiewicz, H. Org. Mass Spectrom. 1988, 23, 561. Madhusudanan, K. P.; Murthy, V. S.; Fraisse, D. Org. Mass Spectrom. 1987, 22, 665. Sears, L. J.; Campbell, J. A.; Grimsrud, E. P. Biomed. Environ. Mass Spectrom. 1987, 14, 401.

Received for review January 25, 2002. Revised manuscript received May 8, 2002. Accepted May 14, 2002. ES0205582

VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3371