Environ. Sci. Technol. 2008, 42, 543–549
Identification of the Novel Cycloaliphatic Brominated Flame Retardant 1,2Dibromo-4-(1,2-dibromoethyl)cyclohexane in Canadian Arctic Beluga (Delphinapterus leucas) G R E G G T . T O M Y , * ,†,‡ K E R R I P L E S K A C H , † GILLES ARSENAULT,§ DAVE POTTER,§ ROBERT MCCRINDLE,4 CHRIS H. MARVIN,⊥ ED SVERKO,⊥ AND SHERYL TITTLEMIER# Department of Fisheries and Oceans, Arctic Aquatic Research Division, Winnipeg, MB, R3T 2N6 Canada, Department of Environment and Geography, University of Manitoba, Winnipeg, MB, R3T 2N2 Canada, Wellington Laboratories Incorporated, Research Division, Guelph, ON, N1G 3M5, Canada, Department of Chemistry, University of Guelph, Ontario, ON, N1G 2W1, Canada, Environment Canada, Burlington, ON, L7R 4A6 Canada, and Food Research Division, Health Canada, Ottawa, ON, K1A 0L2 Canada
Received August 15, 2007. Revised manuscript received October 16, 2007. Accepted November 7, 2007.
1,2-Dibromo-4-(1,2-dibromoethyl)cyclohexane (TBECH) is used primarily as an additive flame retardant. Technical grade TBECH consists of near equimolar amounts of two (of a possible four) diastereoisomers: rac-(1R,2R)-1,2-dibromo-(4S)-4-((1S)-1,2dibromoethyl)cyclohexane (R-TBECH) and rac-(1R,2R)-1,2dibromo-(4S)-4-((1R)-1,2-dibromoethyl)cyclohexane (β-TBECH). The two other possible isomers, γ- and δ-TBECH, appear in the technical mixture when heated at temperatures above 120 °C. Careful selection of GC-capillary column length was critical in resolution of the two main diastereoisomers. Column lengths of 60 or 30 m (0.25 µm film thickness) resulted in incomplete separation of the R- and β-isomers, while on a 10 m column, the isomers were baseline separated. The γ- and δ-isomers could not be resolved on any column length in this study. Increased injector port temperature induced thermal conversion of the R- and β-isomers to γ- and δ-TBECH. Electron impact ionization (EI) was used to provide specificity because no characteristic ions in the electron capture negative ionization (ECNI) mass spectrum of TBECH were evident. In EI, the dominant ions in the mass spectrum corresponded to a concomitant loss of HBr and Br from the molecular ion; the biggest peak in this ion cluster (m/z 266.9208) was used for quantitation and the second biggest peak (m/z 264.9227) was used for confirmation. Beluga (Delphinapterus * Corresponding author phone: (204)-983-5167; fax:(204)-984-2403; e-mail:
[email protected]. † Department of Fisheries and Oceans, Arctic Aquatic Research Division. ‡ Department of Environment and Geography, University of Manitoba. § Wellington Laboratories Incorporated, Research Division. 4 Department of Chemistry, University of Guelph. ⊥ Environment Canada. # Food Research Division, Health Canada. 10.1021/es072043m CCC: $40.75
Published on Web 12/13/2007
2008 American Chemical Society
leucas) blubber extracts of animals from the Canadian Arctic (n ) 29) were analyzed using low resolution (LR) MS and high resolution (HR) MS run at a resolving power of 10 000. β-TBECH was the only isomer observed in the samples and was detected in 17 samples. The LRMS technique appeared to overestimate β-TBECH concentrations compared to HRMS, suggesting a small interference arose at the nominal mass monitored. This potential interference also led to some false positive and negative values (n ) 7) based on the expected ion ratio of the quantitation and confirmation ions. Observed concentrations of the β-isomer as measured by HRMS ranged from 1.1 to 9.3 ng/g (lipid weight).
Introduction 1,2-Dibromo-4-(1,2-dibromoethyl)cyclohexane [also known as tetrabromoethylcyclohexane (TBECH)], is a polybrominated cycloaliphatic flame retardant. It is produced by Albemarle Corporation and marketed as Saytex BCL-462. This material is primarily used as an additive flame retardant in expandable polystyrene beads, which are largely used to produce thermal insulation for housing (1). Concentrations of TBECH in these insulation products are at about 1% w/w (1). TBECH is also incorporated as a flame retardant into extruded polystyrene foam, adhesive in fabric and vinyl lamination, electrical cable coatings, high-impact plastic parts of appliances, and some construction materials (1, 2). Information on recent production amounts of TBECH is limited; one report lists production for the years 1986, 1990, 1994, 1998, and 2002 between 4.5 and 226 t per year (3). Older information reports production at less than 4535 t in 1977, 272 t in 1982, and 453 t in 1983 (1, 2). TBECH can exist as four pairs of enantiomers, the stereochemistries of which are represented in Figure 1. The nomenclature adopted relates to the elution order from a DB-5 capillary column (4). Technical TBECH consists of near equimolar amounts of two diastereomers, R and β (4). At temperatures above 120 °C, a small amount of thermal conversion of the R- and β-isomers occurs, resulting in the formation of γ- and δ-isomers. A similar thermal conversion has been observed to occur with another cycloaliphatic polybrominated flame retardant, hexabromocyclododecane (HBCD) (5). While the technical TBECH formulation consists primarily of R and β, like HBCD, we might expect to detect the thermally formed isomers, γ and δ, in the environment because TBECH is also incorporated into products at high temperatures. To date, no information or data have been published about the identity or proportions of the diastereoisomers in the environment. Larsson et al. have recently demonstrated that TBECH binds to and activates the human androgen receptor in vitro (6). It has also been shown that TBECH bioaccumulates in captive zebrafish (7). This is not surprising considering that the estimated log Kow value of TBECH (based on nonisomeric structural information) is 5.2 (KowWin, Syracuse Research Corporation), which is indicative of a compound with a potential to bioaccumulate (8). Estimated ultimate degradation and atmospheric oxidation half-lives of TBECH also indicate that this compound has long-range atmospheric transport potential (8). Little is known about the environmental fate and behavior of TBECH; however, the initial studies discussed above suggest that it may be persistent, bioaccumulative, and bioactive. The objectives of this study were to develop an analytical method for the detection of the isomers of TBECH VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
543
FIGURE 1. Structures and nomenclature of the four possible TBECH diastereomers. in environmental samples based on mass spectrometry and to investigate the possibility that TBECH is present in beluga sampled from the Canadian Arctic.
Materials and Methods Chemicals. Technical TBECH was obtained from SigmaAldrich (Oakville, ON Canada). Individual solutions of brominated diphenyl ether congeners (71, 126, 156, 197 and 207), a mixture containing equimolar amounts of R- and β-isomers, pure β-isomer (>99.5% purity) and a mixture containing equimolar amounts of the γ- and δ-isomers were from Wellington Laboratories (Guelph, ON Canada). Distilled in glass hexane, dichloromethane (DCM), and acetone were from Caledon Laboratories (Georgetown, ON). Deuterochloroform was obtained from (CDN Isotopes, Quebec, Canada). Sources of organohalogen test mixtures were as follows: toxaphene and technical chlordane (Cambridge Isotope Laboratories, Andover, MA), a mixture of 87 PCB congeners (Ultra Scientific, North Kingston, RI), SRM 2261 (a chlorinated pesticide mixture, NIST, Gaithersburg, MD), and MMQA (marine mammals quality assurance, an “in-house” test mixture of persistent organochlorine compounds). Arctic Samples. Beluga blubber samples were obtained through the Department of Fisheries and Oceans (DFO) ongoing collection programs as part of subsistence hunts of Northern communities in the Canadian Arctic. Animals selected for study were from Hendrickson Island (69°32′N, 133°36′W, Northwest Terrorities, 2006, n ) 8, all males), Pangnirtung, (66°08′N, 65°43′W, SE Baffin Island, 2005, n ) 10, 1F, 9M), Igloolik (69°22′N, 81°48′W Nunavut, 2005, n ) 5, all females), and Resolute (74°41′N, 94°49′W, Nunavut, 2003, n ) 6, all males). Every effort was made to select male animals in our study to avoid contaminant losses that can occur during lactational transfer in females. Blubber was shipped to DFO frozen in plastic bags and stored at -30 °C until analysis. Sample Extraction. To avoid contamination from the sampling bags, inner portions of blubber, not in contact with the inner surface of the plastic bag, were chosen for analysis. Blubber (1 g) were spiked with a suite of recovery internal standards (RIS, 10 µL of standard solution of BDE-71, 126, 197, 207 each at concentration of 2 ng/µL) and extracted by addition of 10 mL of hexane/DCM/acetone (45:45:10) and homogenization with a Polytron hand blender. Tubes were centrifuged at 4 000 g for 10 min. The supernatant was removed, and the extraction process was repeated a second time. The combined extract (20 mL) was then reduced in volume; lipids were removed using an automated gel permeation chromatograph and extracts were further processed by adsorption chromatography using Florisil. Further details of the Florisil step can be found in Law et al. (9). Fraction 1 off the Florisil column was reduced in volume and 544
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
an instrument performance internal standard, BDE-156, was added prior to injection on the GC. GC-MS Analysis. Extracts were injected onto an Agilient 5890 gas chromatograph (Mississauga, ON, Canada) coupled either to a Kratos triple-sector (EBE geometry) high-resolution mass spectrometer (HRMS) or to an Agilent 5973 quadrupole low-resolution mass spectrometer (LRMS). A DB-5 capillary column (0.25 µm film thickness × 0.25 mm i.d.; J&W Scientific, Folsom, CA) of initial length of 60 m was employed in our study; the column was subsequently cut into lengths of 30 or 10 m. Splitless injections of 2 µL on the HRMS were made by an autosampler into the injector port set isothermally at 250 °C. An on-column injector was used for the LRMS analyses in oven-tracking mode. The same oven program was used for both LRMS and HRMS analyses: an initial oven temperature of 90 °C with no hold time, ramped at 20 °C/min to 310 °C, and held for 5 min. The carrier gas, UHP helium, was set at constant flow, 0.75 mL/min. The MS analysis was performed in the electron impact ionization (EI) mode. For the LRMS, source and quadrupole temperatures were both set to 250 °C. A full-scan LRMS EI mass spectrum [m/z 50–450 amu, see Figure 1, Supporting Information (SI)] of an external standard solution of the β-isomer (1 ng/µL, 2 µL injection) gave a base peak cluster starting at m/z 265 corresponding to the [M - HBr - Br]+ ion. The EI-mass spectrum of the R-isomer was identical to that of the β-isomer. For monitoring purposes and under LRMS conditions, the two most dominant ions in the [M-HBr-Br]+ cluster, m/z 267 and 265, were selected. The most abundant was used for quantitation (m/z 267) and the other for confirmation (m/z 265). For HRMS analysis, the resolving power was set at 10 000, the electron energy was 70 eV, the ion source temperature was set at 250 °C, and perfluorokerosene (PFK) was used as the reference gas. One selected ion monitoring (SIM) window was used and monitored the m/z 264.9227 and 266.9208 ions of TBECH. The Kratos Mach 3 software arbitrarily selected a start mass of m/z 251.1608 and a PFK lock mass of m/z 268.9824. Dwell times on the two ions of TBECH were 555.6 and 50 ms for the PFK lock mass. Two important criteria were used for confirmation of TBECH isomers in our beluga samples. First, the elution time of the isomers was within (2 s in the sample and external standard. Second, the ratio of the quantitation to confirmation ion was within (20% of the mean observed value in our β-isomer external standard solution. The BDE congeners 71, 126, 197, and 207 (used as RIS) were measured using LRMS under electron-capture negative ionization using methane as the buffer gas and monitoring the [Br]- ions (m/z 79 and 81). Source and quadrupole temperatures were both set to 150 °C. Quality Control. The extracts employed in this study were previously used for our work on characterizing the spatial
TABLE 1. Average ± 1 Standard Error (n = 6) Percent Recoveries of Brominated Diphenyl Ether (BDE) Internal Recovery Standards and TBECH Isomers fortification level R-TBECH β-TBECH BDEs
low
medium
high
70 ( 3% 59 ( 3% 71 ( 3%
69 ( 5% 94 ( 9% 83 ( 4%
92 ( 13% 98 ( 9% 100 ( 5%
distribution of BFRs in the Canadian Arctic (10). In that study, environmentally nonrelevant brominated diphenyl ethers (71, 126, 197 and 207) were used as RIS. To test if these BDEisomers are extracted in a similar manner to the TBECH isomers, a controlled experiment was conducted whereby the suite of BDE isomers, along with the R- and β-isomers (using a technical TBECH mixture) were spiked into a 15 mL polypropylene tube containing precleaned Ottawa sand (baked for 6 h at 600 °C) at low (0.5 ng, n ) 6), medium (1 ng, n ) 6), and elevated (4 ng, n ) 6) levels. The spiked tubes were extracted, following the same method used for the beluga blubber extracts. Mean recoveries of the four BDE recovery standards and R- and β-TBECH are listed in Table 1. One-way ANOVA testing indicated that there was no difference between the recoveries of the BDEs or TBECH-isomers at either fortification level. Therefore, these results suggest that it would be suitable to use the recoveries of the BDE internal standards analyzed previously to monitor TBECH concentrations in the analytical method. Because recoveries of BDEs in the blubber extracts processed previously were consistent in the sample extracts and greater than 80%, no recovery correction was applied to the TBECH. One duplicate sample of blubber from Pangnirtung Island, extracted separately and analyzed to check for repeatability, was within 75% of each other for the β-isomer suggesting good method repeatability. Method detection limits (MDLs) were estimated from the procedural blanks (n ) 5) which consisted of Ottawa sand. TBECH was not detected in the blanks, thus blank extracts were intentionally fortified with a small amount of the β-isomer (2 ng in a final volume of 100 µL), and five separate injections of the spiked extracts were then made. The ion signals obtained for the β-isomer were used to estimate concentrations that would give a signal-to-noise ratio of 3:1. In this manner and using an average sample mass of 1.2 g, MDLs of the β-isomer were determined to be 0.8 pg/g (wet weight) by HRMS and 4 pg/g (wet weight) by LRMS. The linear dynamic ranges of the instruments were LRMS, 2-500 pg (r2 > 0.99, on-column); HRMS 0.8-500 pg (r2 ) 0.99, splitless injection). Statistical Analysis. Statistical treatment of the data was done using SigmaStat (version 9.01, Systat Software Inc.).
Results and Discussion Selection of MS Ionization Mode and Elution Behavior of the Isomers. ECNI is now well established for the analysis of brominated flame retardants (BFRs). Because of the labile nature of the carbon-bromine bond (bond energy ) 276 kJ/mol), BFRs amenable to GC analysis produce an intense m/z 79/81 ion signal when ionized. The use of shorter analytical columns (10–15 m) can often provide sufficient separation of environmentally relevant BFRs and also allow more thermally susceptible compounds like BDE-209 and decabromodiphenylethane (DBDPE) to be analyzed in one GC-injection (9, 11). However, because this approach offers little specificity, care is often required for positive identification of BFRs. Our initial attempt was to develop a method that was based on ECNI using the m/z 79 and 81 ions so that it could
be incorporated easily into our existing protocols. Even under the fast oven conditions used (20 °C/min), we were able to get near baseline separation of the two isomers [retention time (rt) of R, 7.01, and β, 7.04 min] when the technical formulation was injected (see Figure 2). Unfortunately, because of the short analytical column, there was coelution of the β-isomer with the BDE-15 congener. Despite our best efforts, we were unable to improve the chromatography any further on the short column. Because there were no diagnostic ions in the ECNI fullscan mass spectrum of the β-isomer that would allow us to mass discriminate it from BDE-15, injections on analytical columns of longer lengths were attempted to try to resolve BDE-15 from the β-isomer. While injections on both columns resulted in separation of the β-isomer from BDE-15 congener, we found that when the technical formulation was injected onto the longer columns, both the R and β peaks coalesced (see Figure 2). On the basis of our observations with the behavior of hexabromocyclododecane (HBCD), it was hypothesized that this broad single peak (∼25 s wide) arose from thermally induced interconversion on the longer columns because of the greater on-column residence times of TBECH. Because of the lack of TBECH-specific ions in ECNI and the issues of thermal conversion on columns of longer lengths, our next step then was to use EI monitoring of the abundant [M - 2HBr - Br]+ fragment ion using the short 10 m column. The β-isomer was first injected (Figure 3). Similar to what was observed by injection of the technical mixture, a small peak appeared downfield at rt ≈ 7.25 min; increasing the amount of material injected onto the column further increased the intensity of this peak (data not shown). Increases in injector port temperatures also reduced the overall intensity of the β-isomer (see Figure SI2) and increased the intensity of the peak at rt ≈ 7.25 min. Because the additional peak also appeared in the same ion channel (m/z 266.9208) to that of the β-isomer suggests that the molecule is likely an isomer of TBECH. It was hypothesized that a small amount of thermal conversion on the column was occurring and was dependent on both the injector port temperature and the amount of material injected onto the column. To verify the identity of the peaks, a solution containing equimolar amounts of the γ- and δ- isomers was injected separately but under identical conditions (see Figure SI3). Although the γ- and δ- isomers could not be resolved, their retention time matched that of the additional peak appearing in the ion chromatogram of the β-isomer. Interestingly, the ion chromatogram of the γ- and δ-isomers also resulted in the formation of R- and β-isomers. Taken together, these results further support the hypothesis of thermally mediated isomerization. The fact that these rearrangements were also observed using a cool on-column injector suggests that minor amounts of rearrangement are taking place in the injector and that the majority is occurring on the column itself. Further details on the mechanism of these reactions can be found in Arsenault et al. (4). It could be argued that TBECH isomers might be ideal candidates for analyses by liquid chromatorgraphy mass spectrometry. However, neither electrospray or atmospheric pressure ionization produced ions other than the nondiagnostic m/z 79/81 ions. Furthermore, when the m/z 79 ion was selected, all 4 isomers coeluted on a C18 reverse-phase analytical column. Detection of TBECH in Canadian Arctic Beluga. Because the isomers of TBECH are not discriminated against in our extraction and workup procedures for other BFRs, extracts that were analyzed previously for other BFRs were used for this study (10). We chose to perform our analyses using both LRMS and HRMS which enabled a natural comparison of VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
545
FIGURE 2. HRGC/EI-HRMS selected ion chromatogram (m/z 266.9208) of technical TBECH (100pg/µL, 1µL injection) on a 60 (top panel), 30 (middle panel), and 10 m (bottom panel) DB-5 capillary column. the two methods. Even so, considerable care was taken to ensure that the ratio of the quantitation and confirmation ions for the TBECH isomers were within (20% of their theoretical value of 1.96. The ion ratios observed for the β-isomer from repeated injections of the external standard was 1.97 ( 0.01 (n ) 30). Figure 4 shows a HRGC/EI-HRMS ion chromatogram of a beluga blubber extract. The chromatogram shows a number of peaks that arise in the m/z 266.9208 ion channel, some of which are quite intense. Of the 29 samples examined in our study, 17 (59%) were found to contain the β-isomer based 546
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
on matching retention times and correct ion ratios of m/z 266.9208 to m/z 264.9227 under HRMS. The EI mass spectrum of TBECH (see Figure SI1) centered at nominal mass m/z 267 clearly shows an ion cluster pattern consistent with 2 Br atoms. The theoretical ratio for the three most abundant peaks in the isotopic cluster for fragment ion C8H11Br2+ is 0.51:1:0.49 (for m/z 264.9227, 266.9208, and 268.9189, respectively). Exact mass measurements could not be taken on the 17 samples that met the two criteria of matching retention time and correct quantitation/confirmation ion ratios because of inadequate concentrations of
FIGURE 3. HRGC/EI-HRMS selected ion chromatogram (m/z 266.9208) of β-isomer (100pg/µL, 1µL injection) on a 10m DB-5 capillary column.
FIGURE 4. HRGC/EI-MS selected ion chromatogram (m/z 266.9208) of a beluga blubber extract (Hendrickson Island) showing the β-peak. TBECH. Thus the ratio of m/z 268.9189 to 266.9208 was also monitored to provide further confirmation of the detected compound as β-TBECH. Figure SI4 shows a sample ion chromatogram of a blubber extract for an animal from Pangnirtung. The abundances of the three ions, that is, electronically integrated areas, are in very good agreement with the theoretical ratios; m/z 268.9189 to m/z 264.9228 is 0.91 (expected ) 0.96), and m/z 266.9208 to m/z 268.9198 is 2.19 (expected ) 2.05). Overall, the ion ratios of m/z 268.9189 to m/z 264.9227 and m/z 266.9208 to m/z 268.9198 were
