Fast Selective Detection of Polar Brominated Disinfection Byproducts

Aug 5, 2008 - Brominated disinfection byproducts (DBPs), formed from the reactionofdisinfectant(s)withnaturalorganicmatterandbromide in raw water, are...
0 downloads 0 Views 285KB Size
Environ. Sci. Technol. 2008, 42, 6598–6603

Fast Selective Detection of Polar Brominated Disinfection Byproducts in Drinking Water Using Precursor Ion Scans X I A N G R U Z H A N G , * ,† JEFFREY W. TALLEY,‡ BILL BOGGESS,§ GUOYU DING,† AND DENNIS BIRDSELL| Department of Civil Engineering, Hong Kong University of Science and Technology, Hong Kong, China, Department of Civil and Environmental Engineering and Geological Sciences, Department of Chemistry and Biochemistry, and Center for Environmental Science and Technology, University of Notre Dame, Notre Dame, Indiana 46556

Received November 8, 2007. Revised manuscript received June 9, 2008. Accepted June 30, 2008.

Brominated disinfection byproducts (DBPs), formed from the reaction of disinfectant(s) with natural organic matter and bromide in raw water, are generally more cytotoxic and genotoxic than their chlorinated analogues. Brominated DBPs have been intensively studied over the past 35 years, yet only a fraction of the total organic bromine formed during disinfection has been identified. A significant portion of the unaccounted total organic bromine may be attributed to polar/highly polar brominated DBPs. In this work, a method for fast selective detection of polar/ highly polar brominated DBPs in drinking water was developed using negative ion electrospray ionization-triple quadrupole mass spectrometry (ESI-tqMS) by setting precursor ion scans of m/z 79 and 81. This method was conducted withoutliquidchromatographyseparation.Theresultsdemonstrate that the ESI-tqMS precursor ion scan is an effective tool for the selective detection of electrospray ionizable brominecontaining compounds in a complex mixture. Many polar/ highly polar bromine-containing DBPs were tentatively found in two drinking water samples, and some of them may be new brominated DBPs that have not been previously reported. This method was also extended for the selective detection of polar bromine-containing compounds/contaminants in groundwater, surface water and wastewater.

Introduction Bromide is naturally present in many source waters across the world (1). Upon chlorine, chloramine, chlorine dioxide, or ozone disinfection, bromide can be oxidized to hypobromous acid which then reacts with natural organic matter (NOM) to form brominated disinfection byproducts (DBPs) (1-6). Evidence has shown that brominated DBPs generally * Corresponding author phone: +852-2358-8479; fax: +852-23581534; e-mail: [email protected]. † Hong Kong University of Science and Technology. ‡ Department of Civil Engineering and Geological Sciences, University of Notre Dame. § Department of Chemistry and Biochemistry, University of Notre Dame. | Center for Environmental Science and Technology, University of Notre Dame. 6598

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008

are significantly more cytotoxic and genotoxic than their chlorinated analogues (7), e.g., bromoacetic acid is 18.4× more cytotoxic in Salmonella typhimurium and 89.8× more cytotoxic in Chinese hamster ovary cells than chloroacetic acid; bromoacetic acid is 201.3× more mutagenic in Salmonella typhimurium stain TA100 and 23.6× more genotoxic in Chinese hamster ovary cells than chloroacetic acid (8, 9). Accordingly, brominated DBPs likely warrant more attention than their chlorinated analogues. Like chlorinated DBPs, brominated DBPs have been intensively studied over the past 35 years, yet only a fraction of the total organic bromine (TOBr) formed during disinfection has been chemically identified or even well characterized. By simulating a typical drinking water disinfection condition, Zhang et al. (4) demonstrated that the percentages of TOBr that could be represented by the commonly known brominated DBPs in the chlorine, chlorine dioxide, chloramine, and ozone treated samples were only 60, 26, 14, and 8.2%, respectively. Hua et al. (5) reported that at higher levels of bromide, the ratio of “unknown TOBr” to TOBr during chlorination increased from 17 to 33% with an increase in the added bromide level from 2 to 30 µmol/L. Richardson (2) compiled a review on drinking water DBPs in which she listed approximately 90 brominated DBPs, and almost all of them were identified by gas chromatography/mass spectrometry (GC/MS) or derivatization-GC/MS (10, 11). In recent years, some new brominated DBPs have been reported and surveyed including tribromopyrrole, brominated acids, bromonitromethanes, etc (1, 6), and all of them were identified by GC/MS. Even with derivatization, however, GC/MS generally is not amenable to the detection of highly polar compounds, which contain more than two hydrophilic functional groups. Thus, liquid chromatography/MS or liquid direct infusion-MS using electrospray ionization (ESI) would be a logical choice for detecting polar/highly polar brominated DBPs. A raw water sample or a chlorinated humic substance sample could produce an ESI full scan spectrum with hundreds to thousands of ions distributed on almost every mass unit (12). In order to differentiate the brominecontaining DBPs from the bromine-free DBPs, Minear (13) proposed that radioisotopic bromine be introduced into NOM to label the bromine-containing DBPs, as radioisotopic chlorine (36Cl) has been introduced into NOM successfully to label the chlorine-containing DBPs (14). However, the most stable bromine radioisotope, 82Br, has a half-life of 35 h only, which hinders its use as a specific probe for brominecontaining DBPs. Recently, Zhang et al. (12) reported the characterization of high molecular weight chlorinated DBPs using an ESI-triple quadrupole mass spectrometer (ESItqMS). The mass spectrometer has a useful MS/MS mode, precursor ion scan, which can be used to look for the precursor of a particular fragment ion. Assuming that the fragment ion is selected as chloride ion m/z 35 or 37, it should be possible to find all the chlorine-containing DBPs in a mixture by the ESI-tqMS precursor ion scan. The results have demonstrated that the ESI-tqMS precursor ion scan is an effective tool for fast selective detection of polar chlorinecontaining DBPs in a complex mixture. It is anticipated that a method based on the ESI-tqMS precursor ion scan could be used for fast selective detection of polar brominecontaining DBPs. Although the primary purpose of the method to be developed was to uncover the “mysterious” unknown part of TOBr that have eluded from researchers over the past 35 years, the significance of the method may be even greater. 10.1021/es800855b CCC: $40.75

 2008 American Chemical Society

Published on Web 08/05/2008

TABLE 1. Isotopic Abundances of Haloacids in Full Scan haloacids

molecular ions

monochloroacetic acid dichloroacetic acid monobromoacetic acid 2,2-dichloropropanoic acid 2-bromopropanoic acid trichloroacetic acid chlorobromoacetic acid dichlorobromoacetic acid dibromoacetic acid chlorodibromoacetic acid tribromoacetic acid

COO-

ClCH2 Cl2CHCOOBrCH2COOCH3CCl2COOCH3CHBrCOOCl3CCOOClBrCHCOOCl2BrCCOOBr2CHCOOClBr2CCOOBr3CCOO-

M-

(M+2)-

93 127 137 141 151 161 171 205 215 249 293

95 129 139 143 153 163 173 207 217 251 295

First, this method could be extended from fast selective detection of brominated DBPs to fast selective detection of bromine-containing contaminants in drinking water, which could be an important issue with regard to drinking water security. Second, this method could be extended from fast selective detection of brominated DBPs in drinking water to fast selective detection of brominated DBPs in wastewater, and to fast selective detection of bromine-containing compounds/contaminants in surface water and groundwater. The objectives of this work were to develop a method for fast selective detection of polar bromine-containing DBPs in drinking water by using ESI-tqMS precursor ion scan, to apply the method for the selective detection of bromine-containing DBPs in drinking water, and to extend the method for the selective detection of bromine-containing compounds/ contaminants in other waters. Theoretical Calculation of the Isotopic Abundances for Precursor Ion Scans. The information about isotopic abundances is useful in predicting the number of atoms of a given element in a formula. The prediction of the isotopic abundances for full scans is well-known (15), but there is no literature available for the prediction of the isotopic abundances for precursor ion scans. Accordingly, the theoretical isotopic abundances for precursor ion scans were developed. For a compound containing m bromine atoms and n chlorine atoms, it should have a total of m + n + 1 isotopic peaks in full scan. The abundance of the Nth isotopic peak in precursor ion scan of m/z 79 (AN,pre79) can be expressed by eq 1: min(N-1,m)



AN,pre79 )

[CmkCnN-k-13-(N-k-1)(m - k)]

(1)

k)max(N-n-1,0)

where k is any integer between max (N - n - 1,0) and k and CN-k-1 are defined to the natural min (N - 1,m); Cm n numbers k ) Cm

m! n! and CnN-k-1 ) k ! (m - k)! (N - k - 1) ! (n - N + k + 1)!

where ! denotes the factorial. The abundance of the Nth isotopic peak in precursor ion scan of m/z 81 (AN,pre81) can be expressed by eq 2: min(N-1,m)

AN,pre81 )



[CmkCnN-k-13-(N-k-1)k]

(2)

k)max(N-n-1,0)

For a compound containing m bromine atoms only, eqs 1 and 2 can be simplified to eqs 3 and 4, respectively. N-1 (m - N + 1) AN,pre79 ) Cm

(3)

N-1 AN,pre81 ) Cm (N - 1)

(4)

(M+4)-

(M+8)-

ratio

(M-CO2)-

167

3:1 9:6:1 1:1 9:6:1 1:1 3:3:1:0.1 3:4:1 9:15:7:1 1:2:1 3:7:5:1 1:3:3:1

49 83 93 97 107 117 127 161 171 205 249

131 145 165 175 209 219 253 297

211 255 299

The derivation of eqs 1-4 is detailed in the Supporting Information. Once the abundances of the Nth isotopic peak in precursor ion scans of m/z 79 and 81 are obtained, the ratio of product ions m/z 79 and 81 of the Nth isotopic peak in full scan (R79/81,prodN) can be determined: R79⁄81,prodN ) AN,pre79⁄AN,pre81

(5)

Experimental Methods Instrument Optimization. The mass spectra were acquired on a Micromass Quattro LC tqMS fitted with a Z-Spray interface (Waters). To obtain the optimal detection, all the parameters related to the ESI-tqMS were optimized with a haloacid standard consisting of 10-30 mg/L each of nine haloacetic acids and two halopropanoic acids in 1:1 water/ acetonitrile. The haloacid standard was prepared by dilution of the U.S. Environmental Protection Agency (EPA) 552.2 acids calibration mix (Supelco) that contains 11 components (Table 1). Haloacids were chosen as the standard because the majority of polar/highly polar brominated DBPs in drinking water are believed to be brominated acids, which could obtain their carboxyl groups either by transfer from humic substances or by oxidation of humic substances during disinfection. First, the ESI full scan was optimized with the haloacid standard. The goal was to maximize the molecular ions of all the components in the sample. The appropriate operation parameters were set as follows: Sample flow rate via an infusion pump (without using liquid chromatography), 5 µL/ min; ESI source, atmospheric pressure; ESI mode, negative; ESI needle potential, -3.0 kV; Ion source temperature, 65 °C; Nebulizer gas flow rate, 26 L/h; Desolvation gas flow rate, 512 L/h. Quadrupoles 1 and 3 (Q1 and Q3) were set to unit resolution, which corresponds to a numerical value of 15, in all experiments unless otherwise specified. To examine the possible fragmentation of analytes in the ion source, a series of source cone potentials were tested: 10, 15, 20, 25, 30, 35, 40, and 45 V. Second, the product ion scan was optimized with the haloacid standard. The goal was to maximize the production of fragment ion Br- (m/z 79 or 81). Collision gas (Argon) pressure in the collision induced decomposition (CID) chamber was set at 1.55 × 10-3 mb, and collision energy was optimized by varying it from 1 to 200 eV. Third, the precursor ion scan was optimized with the haloacid standard. The goal was to maximize the bromine-containing precursor ions by setting the fragment bromide ion m/z 79 or 81. For precursor ion scans, a Q1 resolution of 12 was also used. The Q1 resolution of 12 increases the mass width of ions admitted into the collision cell beyond 1 unit, and thus amplifies the intensities of precursor ions dramatically, but it may possibly allow unintended ions to be fragmented and detected at Q3. Sample Pretreatment. Two drinking water samples were collected. One was the finished drinking water taken at a U.S. drinking water treatment plant, and the other was the VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6599

drinking water taken at a U.S. consumer’s tap fed by the drinking water treatment plant. The nominal linear distance between the consumer’s tap and the drinking water treatment plant was 3.1 km. Prior to shipment to the laboratory, the samples were spiked with 100 mg/L of granular ammonium sulfate to quench free chlorine. The source groundwater that supplied the drinking water treatment plant was collected as a control sample. The characteristics of the groundwater were as follows: Br-, 84 µg/L; dissolved organic carbon, 2.3 mg/L as C; specific UV absorbance, 2.4 L/mg-m; alkalinity, 285 mg/L as CaCO3; hardness, 402 mg/L as CaCO3; total iron, 0.31 mg/L; pH, 7.4. The treatment processes of the drinking water treatment plant consisted of prechlorination, filtration, and chlorination (with a dose of 3.6 mg/L as Cl2). One lake water sample was collected at Lake Michigan. One wastewater effluent sample was collected at a U.S. wastewater treatment plant, which was a conventional activated sludge plant with treatment processes including bar screens, grit removal, primary settling, activated sludge, anaerobic digesters, secondary clarification, chlorination (with a dose of 5.0 mg/L as Cl2), and dechlorination. All the samples were collected in amber glass containers with Teflon-lined screw caps. The collected samples were immediately shipped on ice to the laboratory where they were stored in a cold room at 4 °C. A high concentration of inorganic ions in water samples could complicate MS spectra and lower the electrospray ionization efficiency for organics, and thus such ions need to be eliminated or minimized prior to the mass spectrometry analysis. Also, the concentrations of polar brominated DBPs in original water samples are generally below the detection limits of mass spectrometry. Accordingly, an appropriate pretreatment is required for reducing inorganic ions and for concentrating brominated DBPs. Since liquid-liquid extraction (LLE) with methyl tert-butyl ether (MtBE) can effectively extract haloacetic acids in water (16), the LLE with MtBE was adopted in this study. A 1.7 L portion of each water sample was acidified to pH e0.5 with sulfuric acid so that most polar/highly polar DBPs, mainly including phenolic and carboxyl anions, can be neutralized. The water sample was then saturated with sodium sulfate for salting out during extraction, followed by the addition of 170 mL MtBE. The sample was shaken vigorously and then allowed to settle in a separatory funnel. After the water layer was drained off, the organic layer was transferred to a rotary evaporator and concentrated to 0.5 mL. The 0.5 mL solution in MtBE was mixed with 20 mL acetonitrile, and the mixture was rotoevaporated back to 0.5 mL, leading to the desired solvent exchange from MtBE to acetonitrile due to the lower boiling point of MtBE. The 0.5 mL solution in acetonitrile was diluted with Milli-Q water to 1 mL and transferred to a 1 mL vial for analysis. To determine whether there were any impurities in the solvents or any artifacts in the extraction and concentration, a control sample was generated by repeating the same procedure with 1.7 L Milli-Q water. Sample Analysis. Each pretreated sample was then analyzed with the optimized ESI-tqMS system. First, the pretreated sample was analyzed by the precursor ion scans of m/z 79 and 81. Second, for those brominated molecular ions detected by the precursor ion scans with relatively strong intensities, they were analyzed by product ion scans to obtain some structural information. All the samples were performed with the LLE pretreatment and analysis of duplicate portions.

Results and Discussion Method Development. To selectively detect brominated DBPs, the pivotal step is to generate the true molecular ions of all the electrospray ionizable compounds by ESI full scan. One possible problem related to an ESI full scan spectrum is its low molecular weight bias for higher molecular weight 6600

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008

FIGURE 1. Variation of ESI full scan spectra of the haloacid standard with source cone potential (V): (a) 10, (b) 15, (c) 20, (d) 25, (e) 30, (f) 35, (g) 40, and (h) 45. The y-axes are on the same scale with a maximum intensity of 2.06 × 109. DBPs, most likely attributed to fragmentation. McIntyre et al. (17) studied humic substances by negative ESI full scan with a Micromass ESI-tqMS, and suggested that a source cone potential of 35 V should be used to acquire the full scan spectra of humic substances because no fragmentation seemed to occur at that potential. Accordingly, the effect of source cone potential on the full scan spectra of DBPs was examined in detail. Figure 1 shows the ESI full scan spectra of the haloacid standard at different source cone potentials. As the cone potential increased from 10 to 15 V, the intensities of all ions increased; By increasing the cone potential from 20 to 45 V, the intensity of higher m/z ions decreased, accompanied by an increase of lower m/z ions. When the cone potential was less than 25 V, the measured isotopic ratios of all the molecular ion clusters matched the expected isotopic ratios (Table 1). Relative intensities of molecular ions ClBr2CHCOOand Br3CCOO- were very weak because these two components are unstable in water (18) and might partially decompose to the corresponding trihalomethanes. When the cone potential was above 30 V, an interesting phenomenon occurred: the measured isotopic ratios of several molecular ion clusters began to deviate from those in Table 1, e.g., as the cone potential increased from 30 to 45 V, the abundance ratio of m/z 127/129/131 gradually shifted from 9:6:1 to 3:4: 1, which means that Cl2CHCOO- (m/z 127/129/131 with a ratio of 9:6:1) gradually became Cl2CH-, and ClBrCHCOOgradually became ClBrCH- (m/z 127/129/131 with a ratio of 3:4:1); the abundance ratio of m/z 171/173/175 gradually shifted from 3:4:1 to 1:2:1, which indicates that ClBrCHCOO(m/z 171/173/175 with a ratio of 3:4:1) gradually became ClBrCH-, and Br2CHCOO- gradually became Br2CH- (m/z 171/173/175 with a ratio of 1:2:1). These shifts correspond to the neutral loss of CO2, as shown in Table 1. The effect of source cone potential on the molecular ions can be exemplified with a specific ion, m/z 127, in the haloacid standard. The appearance of product ion spectra of m/z 127 varied with the cone potential (Figure S1 in the Supporting Information (SI)). When the cone potential was less than 15 V, only product ions 35Cl- at m/z 35 and 35Cl2CH- at m/z 83

FIGURE 2. Variation of fragment Br- intensity with CID collision energy for different precursor ions of the haloacid standard. were observed from the expected molecular ion Cl2CHCOO(m/z 127). At a cone potential of 20 V, another product ion 79Br- at m/z 79 was detected besides the product ions m/z 35 and 83, which suggests that another molecular ion also with m/z 127, ClBrCH-, showed up at this potential. Figure 1 and SI Figure S1 demonstrate that the appropriate cone potential for haloacids was around 15 V; when the cone potential was above 20 V, decarboxylation would occur. The cone potential of 15 V was selected for all analyses. It has been reported in previous work that maximizing the abundance of Cl- during CID required high collision energy (19), sometimes up to 200 eV (12). It turned out that the formation of Br- during CID could be maximized at a much lower collision energy. The variation of product ion spectra of m/z 137 (BrCH2COO-) of the haloacid standard with the collision energy is shown as an example (SI Figure S2): Br- intensity reached the highest value of 3.49 × 106 at a collision energy of 10 eV; while at the collision energy of 200 eV, Br- intensity decreased by 99% to 3.49 × 104, which was just ∼2 times higher than background intensity. The effect of collision energy on the formation of Brfrom different precursor ions, including BrCH2COO-, ClBrCHCOO-, and Br2CHCOO-, was also examined with the haloacid standard (Figure 2). As the collision energy increased from 1 to 200 eV, the Br- intensity rose rapidly to a maximum, and then fell rapidly until reaching a collision energy of 80 eV. Above a collision energy of 80 eV, the Br- intensity slowly decreased and gradually reached the background level at 200 eV collision energy. The optimum collision energy that corresponds to the maximum Br- intensity had a slight variation from compound to compound, but generally fell in the range of 10-25 eV. This is completely different from the formation of Cl- during CID (SI Figure S3). It indicates that the precursor ion scan for bromine-containing compounds is significantly more effective at lower collision energies than the precursor ion scan for chlorine-containing compounds. Under the optimized conditions (source cone potential 15 V and collision energy 20 eV), the ESI-tqMS precursor ion scan was performed for the haloacid standard. Figure 3a shows the precursor ion scan spectrum of m/z 79, which consisted of eight peaks/clusters: m/z 127/129, 137, 151, 171/ 173, 205/207/209, 215/217, 249/251/253, and 293/295/297. Figure 3b shows the precursor ion scan spectrum of m/z 81, which also consisted of eight peaks/clusters: m/z 129/131, 139, 153, 173/175, 207/209/211, 217/219, 251/253/255, and 295/297/299. The eight peaks/clusters in the precursor ion scan of m/z 79 or 81 corresponded to ClBrCH-, BrCH2COO-, CH3CHBrCOO-, ClBrCHCOO-, Cl2BrCCOO-, Br2CHCOO-, ClBr2CCOO-, and Br3CCOO-, respectively. ClBrCH- was not an original component in the haloacid standard, and may come from the decarboxylation of ClBrCHCOO-. The ions ClBr2CCOO- and Br3CCOO- were relatively weak mainly

FIGURE 3. Spectra of precursor ion scans of the haloacid standard obtained under the optimized instrument conditions (source cone potential 15 V and CID collision energy 20 eV): (a) precursor ion scan of m/z 79, and (b) precursor ion scan of m/z 81. The y-axes are on the same scale with a maximum intensity of 4.28 × 106. because they are quite unstable in water (18), and also because their initial concentrations were relatively low. As expected, ClCH2COO-, Cl2CHCOO-, CH3CCl2COO- and Cl3CCOO- were not observed in the precursor ion scan of either m/z 79 or m/z 81. These spectra indicate that the precursor ion scan can selectively detect bromine-containing compounds in a mixture. For the same bromine-containing compound, the peak/cluster detected by the precursor ion scan of m/z 79 and the peak/cluster detected by the precursor ion scan of m/z 81 have a 2-unit difference, and have the same isotopic peak abundances due to the fact that the natural abundance ratio of 79Br/81Br is ∼1:1. Here, a peak/ cluster detected by the precursor ion scan of m/z 79 and the counterpart peak/cluster detected by the precursor ion scan of m/z 81 are defined as a pair. The characteristics of the pair is its approximate “symmetry”, i.e., the same isotopic peak abundances. Each pair generally corresponds to one brominated compound, but may possibly correspond to two or more isobaric brominated compound homologues. By seeking such pairs in the precursor ion scan spectra of m/z 79 and 81, it can be assured that bromine-containing compounds are selectively detected. It is of note that the ESItqMS precursor ion scan can only detect those brominecontaining compounds that are ionized by negative ion ESI. Humic substances/brominated humic substances are generally rich in carboxylic and phenolic groups, which make polar/ highly polar brominated DBPs suitable for analysis with negative ion ESI. More importantly, the isotopic abundances in the precursor ion scan spectra of the haloacid standard match the theoretical calculations (Supporting Information): e.g., BrCH2COO- with 1Br generated one isotopic peak m/z 137 in the precursor ion scan of m/z 79; ClBrCHCOO- with 1Br + 1Cl generated two isotopic peaks m/z 171/173 with an abundance ratio of 3:1 in the precursor ion scan of m/z 79; Br2CHCOO- with 2Br generated two isotopic peaks m/z 215/ 217 with an abundance ratio of 1:1 in the precursor ion scan of m/z 79. The same isotopic peak abundances for the same compound were also observed in the precursor ion scan of m/z 81. VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6601

FIGURE 4. Spectra of precursor ion scans of the finished drinking water sample: (a) precursor ion scan of m/z 79 (Q1 resolution 12, Y-axis maximum intensity 1.39 × 106), (b) precursor ion scan of m/z 81 (Q1 resolution 12, Y-axis maximum intensity 1.39 × 106), (c) a tentative structure for pair (395/397/ 399, 397/399/401), and (d) a tentative structure for pair (410/412/ 414, 412/414/416).

FIGURE 5. Spectra of precursor ion scans of the tap drinking water sample: (a) precursor ion scan of m/z 79 (Q1 resolution 12, Y-axis maximum intensity 2.61 × 106), (b) precursor ion scan of m/z 81 (Q1 resolution 12, Y-axis maximum intensity 2.61 × 106), and (c) structure of chlorobromopropanedioic acid, one of the isobaric brominated DBPs corresponding to pair (215/217, 217/219).

Method Application. Several pretreated water samples were analyzed with the ESI-tqMS precursor ion scans by setting m/z 79 and 81, respectively. By comparing the spectra of precursor ion scans of m/z 79 and 81, the pairs corresponding to the bromine-containing DBPs in each sample could be determined. By comparing the isotopic peak abundances in the precursor ion scan spectrum of m/z 79 or 81 to the theoretical ones, the exact numbers of Br and Cl atoms in a compound can be determined. Since Br and Cl atoms are significantly heavier than H, C, O, and N atoms, a large portion of the molecular weight of a formula is determined. Then, the pairs with relatively strong intensities were selected, and were further analyzed with product ion scans to gain additional information that could potentially be used to propose structures for the brominated DBPs. SI Figure S4 shows the precursor ion scan spectra of a groundwater sample, which was collected at the point of the groundwater entering the drinking water treatment plant. As expected, except for one pair m/z (456, 458) with a low intensity of 1.20 × 105, there were no other pairs seen by the precursor ion scans of m/z 79 and 81. Figure 4 shows the precursor ion scan spectra of the finished drinking water sample collected at the drinking water treatment plant. There were 10 major pairs seen by the precursor ion scans of m/z 79 and 81: m/z (171/173, 173/175), (193, 195), (205/207, 207/ 209), (215/217, 217/219), (376, 378), (395/397/399, 397/399/ 401), (410/412/414, 412/414/416), (428, 430), (456, 458), and (512, 514). The pairs (171/173, 173/175), (205/207, 207/209), and (215/217, 217/219) can be ascribed to the commonly known DBPs ClBrCHCOO-, Cl2BrCCOO-, and Br2CHCOO-, respectively. The other pairs might represent some new brominated DBPs that have not been previously reported in drinking water. The pair m/z (456, 458) in the finished drinking water sample had an intensity of 1.39 × 106, which was about 12 times higher than that in the groundwater sample. With the aid of the precursor ion scans in finding brominecontaining molecular ions, product ion scans were performed to confirm whether those ions could produce product bromide ions, and to gain some structural information. As illustrated in the Supporting Information, and SI Figures S6 and S7, pair (395/397/399, 397/399/401) should be assigned as 1,1,2-tribromo-1,2,2-tricarboxylethane, and pair (410/412/ 414, 412/414/416) could correspond to 1-bromoamino-1,2dibromo-1,2,2-tricarboxylethane (or its isomers). Their structures are shown in Figures 4c and 4d. To our knowledge, it is the first time that these highly polar tricarboxyl brominated

DBPs have been tentatively identified and observed in drinking water. The product ion scan spectra confirmed that all the other pairs including (376, 378), (428, 430), (456, 458), and (512, 514) also corresponded to bromine-containing compounds in SI Figures S5, S8, S9 and S10. Figure 5 shows the precursor ion scan spectra of a tap drinking water sample. This sample was collected at a consumer’s tap into which the drinking water treatment plant distributed its water. There were nine major pairs seen by the precursor ion scans of m/z 79 and 81: m/z (127/129, 129/131), (137, 139), (149, 151), (171/173, 173/175), (193, 195), (205/207, 207/209), (215/217, 217/219), (249/251, 251/253), and (456, 458). Pairs (127/129, 129/131), (137, 139), (149, 151), (171/173, 173/175), (205/207, 207/209), (215/217, 217/ 219), and (249/251, 251/253) can readily be attributed to ClBrCH-, BrCH2COO-, BrCHdCHCOO-, ClBrCHCOO-, Cl2BrCCOO-, Br2CHCOO-, and ClBr2CCOO-, respectively. As illustrated in SI Figure S11, pair (193, 195) can be determined as bromobutenedionic acid, which has been identified as a DBP produced by disinfection of drinking water rich in bromide (1). As illustrated in SI Figure S12, pair (215/217, 217/219) consisted mainly of dibromoacetic acid and a small portion of chlorobromopropanedioic acid, whose structure is shown in Figure 5c. It is the first time that this polar brominated DBP has been tentatively identified and observed in drinking water. Compared to the finished drinking water collected at the drinking water treatment plant, some higher molecular weight DBPs did not show up, such as pairs (376, 378), (395/397/399, 397/399/401), (410/412/ 414, 412/414/416), (428, 430), and (512, 514); Some higher molecular weight DBPs were significantly reduced, e.g., the intensity of pair (456, 458) decreased from 1.39 × 106 in the finished drinking water sample to 1.33 × 105 in the tap drinking water sample; Meantime, some lower molecular weight DBPs emerged, such as pairs (127/129, 129/131), (137, 139), (149, 151), (205/207, 207/209), and (249/251, 251/253). This might suggest that some higher molecular weight brominated DBPs were transformed into lower molecular weight ones when they moved along the distribution system: some NOM/chlorinated NOM may continue to react with the residual chlorine, and some may undergo hydrolysis or other decomposition reactions (20, 21). To demonstrate that this method could be extended from fast selective detection of brominated DBPs in drinking water to fast selective detection of brominated compounds/ contaminants in other waters, it was also conducted for the analysis of numerous surface water and wastewater samples.

6602

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008

The application of the method to the analysis of a lake water sample and a wastewater effluent sample is exemplified as shown in SI Figures S13-S15. In conclusion, a method for fast selective detection of polar/highly polar brominated DBPs in drinking water was developed by using an ESI-tqMS following a simple LLE pretreatment. This method was successfully applied to fast selective detection of brominated DBPs in two drinking water samples: by seeking pairs in the spectra of precursor ion scans of m/z 79 and 81, many polar brominated DBPs were found; by comparing the pairs with DBPs in the literature, it was readily asserted whether the pairs are new brominated DBPs; in addition, by further analyzing the pairs with product ion scans, some structural data of the detected brominated DBPs were obtained. Therefore, this work is a pivotal step toward resolving the unknown portion of TOBr formed during disinfection. This method was also successfully applied to the selective detection of polar brominated compounds/ contaminants in groundwater, lake water, and wastewater.

Acknowledgments This work was supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (project no. HKUST622607). We are grateful to the three anonymous referees for constructive critique and useful suggestions.

(6)

(7)

(8)

(9)

(10) (11) (12)

(13) (14)

Supporting Information Available

(15)

Additional details and Figures S1-S15. This material is available free of charge via the Internet at http://pubs.acs.org.

(16)

Literature Cited (1) Richardson, S. D., Jr.; Rav-Acha, C.; Groisman, L.; Popilevsky, I.; Juraev, O.; Glezer, V.; Mckague, A. B.; Plewa, M. J.; Wagner, E. D. Tribromopyrrole, brominated acids, and other disinfection byproducts produced by disinfection of drinking water rich in bromide. Environ. Sci. Technol. 2003, 37, 3782–3793. (2) Richardson, S. D. Drinking water disinfection by-products. In Encyclopedia of Environmental Analysis and Remediation; John Wiley & Sons: New York, 1998; pp 1398-1421. (3) Cowman, G. A.; Singer, P. C. Effect of bromide ion on haloacetic acid speciation resulting from chlorination and chloramination of aquatic humic substances. Environ. Sci. Technol. 1996, 30, 16–24. (4) Zhang, X.; Echigo, S.; Minear, R. A.; Plewa, M. J. Characterization and comparison of disinfection by-products of four major disinfectants. In Natural Organic Matter and Disinfection ByProducts: Characterization and Control in Drinking Water; Barrett, S. E., Krasner, S. W., Amy, G. L., Eds.; American Chemical Society: Washington, DC, 2000; pp 299-314. (5) Hua, G.; Reckhow, D. A.; Kim, J. Effect of bromide and iodide ions on the formation and speciation of disinfection by-

(17) (18) (19)

(20) (21)

products during chlorination. Environ. Sci. Technol. 2006, 40, 3050–3056. Krasner, S. W.; Weinberg, H. S.; Richardson, S. D.; Pastor, S. J.; Chinn, R.; Sclimenti, M. J.; Onstad, G. D., Jr. Occurrence of a new generation of disinfection byproducts. Environ. Sci. Technol. 2006, 40, 7175–7185. Bull, R. J. Carcinogenic Properties of Brominated Haloacetates. Workshop Report on Disinfection By-Products in Drinking Water; Critical Issues in Health Effects Research: Chapel Hill, NC, 1995, p 29. Plewa, M. J.; Kargalioglu, Y.; Vankerk, D.; Minear, R. A.; Wagner, E. D. Mammalian cell cytotoxicity and genotoxicity analysis of drinking water disinfection by-products. Environ. Mol. Mutagen. 2002, 40, 134. Plewa, M. J.; Wagner, E. D.; Richardson, S. D., Jr.; Woo, Y. T.; McKague, A. B. Chemical and biological characterization of newly discovered iodoacid drinking water disinfection byproducts. Environ. Sci. Technol. 2004, 38, 4713–4722. Richardson, S. D. The role of GC/MS and LC/MS in the discovery of drinking water disinfection by-products. J. Environ. Monit. 2002, 4, 1–9. Xie, Y. Analyzing haloacetic acids using gas chromatography/ mass spectrometry. Water Res. 2001, 35, 1599–1602. Zhang, X.; Minear, R. A.; Barrett, S. E. Characterization of high molecular weight disinfection byproducts from chlorination of humic substances with/without coagulation pretreatment using UF-SEC-ESI/MS/MS. Environ. Sci. Technol. 2005, 39, 963–972. Minear, R. A. Personal communications 2002. Zhang, X.; Minear, R. A. Characterization of high molecular weight DBPs resulting from chlorination of aquatic humic substances. Environ. Sci. Technol. 2002, 36, 4033–4038. McLafferty, F. W.; Turecek, F. Interpretation of Mass Spectra; 4th ed.; University Science Books: Mill Valley, CA, 1993. USEPA. Method 552.3 for Determination of Haloacetic Acids and Dalapon in Drinking Water by Liquid-Liquid Microextraction, Derivatization, And Gas Chromatography with Electron Capture Detection, EPA 815B03002; U.S. Environmental Protection Agency: Washington, DC, 2003. McIntyre, C.; Batts, B. D.; Jardine, D. R. Electrospray mass spectrometry of groundwater organic acids. J. Mass Spectrom. 1997, 32, 328–330. Zhang, X.; Minear, R. A. Decomposition of trihaloacetic acids and formation of the corresponding trihalomethanes in drinking water. Water Res. 2002, 36, 3665–3673. Zhang, X.; Minear, R. A.; Guo, Y.; Hwang, C. J.; Barrett, S. E.; Ikeda, K.; Shimizu, Y.; Matsui, S. An electrospray ionizationtandem mass spectrometry method for identifying chlorinated drinking water disinfection byproducts. Water Res. 2004, 38, 3920–3930. Koch, B.; Krasner, S. W.; Sclimenti, M. J.; Schimpff, W. K. Predicting the formation of DBPs by the simulated distribution system. J. Am. Water Works Assoc. 1991, 83 (10), 62–70. Zhang, X.; Minear, R. A. Formation, adsorption and separation of high molecular weight disinfection byproducts resulting from chlorination of aquatic humic substances. Water Res. 2006, 40, 221–230.

ES800855B

VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6603