Environ. Sci. Technol. 2010, 44, 8615–8621
High-Sensitivity Method for Determination of Tetrabromobisphenol-S and Tetrabromobisphenol-A Derivative Flame Retardants in Great Lakes Herring Gull Eggs by Liquid Chromatography-Atmospheric Pressure Photoionization-Tandem Mass Spectrometry ROBERT J. LETCHER* AND SHAOGANG CHU Ecotoxicology and Wildlife Health Division, Wildlife and Landscape Science Directorate, Science and Technology Branch, Environment Canada, National Wildlife Research Centre, Carleton University, Ottawa, ON, K1A 0H3, Canada
Received August 3, 2010. Revised manuscript received September 25, 2010. Accepted October 10, 2010.
Tetrabromobisphenol-A-bis(2,3-dibromopylether) (TBBP-Adbpe), tetrabromobisphenol-A-bis(allyl ether) (TBBP-A-ae), and tetrabromobisphenol-S-bis(2,3-dibromopropyl ether) (TBBP-Sdbpe) are derivatives of tetrabromobisphenol-A (TBBP-A), and are all used as brominated flame retardants (BFRs). Using highperformanceliquidchromatography-quadrupole-time-of-flightmass spectrometry with atmospheric pressure photoionization (APPI) in the negative mode (LC-APPI(-)-Q-TOF-MS) and the novel use of pure acetone as dopant and LC mobile phase, full scan mass spectra showed that for these BFRs the dominant isotopic ion cluster was [M + O2]-, and with other lesser abundant [M + O2 - HBr]-, and [M - H]- fragment ions. Subsequently, highly sensitive quantification of TBBP-Adbpe, TBBP-A-ae, and TBBP-S-dbpe was accomplished via LCtriple quadrupole mass spectrometry with APPI(-) (LC-APPI(-)MS/MS) via multiple ion monitoring based on the [M + O2]- > [Br]- transition. Low to sub ng/g (wet weight (w.w.)) method limits of detection (LODs) were achieved, i.e., 0.07, 0.03, and 1.28 ng/g w.w. for TBBP-A-dbpe, TBBP-A-ae, and TBBPS-dbpe, respectively. A variety of herring gull eggs were screened for these BFRs. The eggs were collected during 2008-2009 from several colony sites in the eastern Laurentian Great Lakes (Ontario) and in the St. Lawrence River (Que´bec). All egg samples had TBBP-S-dbpe concentrations below the LOD, and TBBP-A-ae and TBBP-A-dbpe were quantifiable in 67%-83% of the samples at concentrations up to 0.56 ng/g wet weight. Thus, TBBP-A-ae and TBBP-A-dbpe are present in herring gull eggs from these populations, bioaccumulate in the herring gull food chain, and are transferred from gull to egg.
* Corresponding author phone: 613-998-6696; fax 613-998-0458; e-mail:
[email protected]. 10.1021/es102135n
Published 2010 by the American Chemical Society
Published on Web 10/21/2010
Introduction Brominated flame retardants (BFRs) are a broad class of additive and reactive substances and technical mixtures used in commercial polymeric materials. Tetraromobisphenol A (TBBP-A) is the most extensively used BFR (1) with a global market of about 170,000 tonnes in 2004 and representing more than 50% of the production of all BFRs (2), and has come under increasing environmental scrutiny (3). TBBP-A is used in the manufacture of several derivatives (1, 4-6). Major TBBP-A derivatives that are produced include tetrabromobisphenol-A-bis(allyl ether) (TBBP-A-ae), tetrabromobisphenol-A-bis(2,3-dibromopropyl ether) (TBBP-Adbpe), TBBP-A bis(2-hydroxyethyl ether), TBBP-A brominated epoxy oligomers, and TBBP-A carbonate oligomers. TBBP-A derivatives utilized as flame retardants and TBBP-A oligomers represent 18% of the total sales of TBBP-A (2), are relatively new as BFRs, and the proportions in use are expected to increase (7). TBBP-A and oligomers and ether derivatives of TBBP-A are used mainly in specialized applications as either reactive or additive intermediates in polymer manufacture. They are incorporated into a broad range of plastic products and epoxy resins that are used in electrical and electronic equipment and installations worldwide (2, 6). TBBP-A-dbpe is an additive for polyolefins and polymers, including polypropylene, high-density polyethylene, and low-density polyethylene, whereas TBBP-A-ae is used as a reactive BFR in polystyrene foams (4, 8). Another TBBP-A-dbpe-like BFR, which is widely used in Asia, is tetrabromobisphenol-S-bis-(2,3-dibromopropyl ether) (TBBPS-dbpe). TBBP-S-dbpe is used in electronic devices that are distributed worldwide. TBBP-S-dbpe was largely produced and marketed in Japan, and there are indications that it continues to be produced and marketed in China in high production volumes (9). Regulation of TBBP-A is recent or underway in several global countries or jurisdictions such as the European Union (EU) (6). In North America, to date, legislative focus on TBBP-A has received little attention, although Canada is in the process of assessing the human and environmental risks of TBBP-A and its diglycidyl and allyl ether derivatives including TBBP-A-dbpe and TBBPA-ae (10). Covaci et al. (4) recently reviewed analytical methods and the determination of TBBP-A and its derivatives in environmental samples. TBBP-A, and likely its derivatives, are thermally unstable and thus direct determination by GC-MS based methods has been generally unsuccessful. A GC-MS based method has been reported for TBBP-A analysis but only after chemical derivatization by, e.g., methylation using methyl chloroformate (11). TBBP-A is a phenolic compound with acid-base pKa1 and pKa2 values estimated at 7.5 and 8.5, respectively. LC-MS(/MS) based methods have been reported as being ideal for TBBP-A determination in samples (4). Analysis of TBBP-A derivatives, such as TBBP-A-dbpe, using LC-MS with electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) was recently reported but with poor quantitative sensitivity (7). Other methods for TBBP-A determination based on approaches incorporating LC-diode array detection (LC-DAD) and laser desorption/ionization-mass spectrometry (LDI-MS) have also been reported (7, 9, 12, 13). Information is limited for TBBP-A in environmental samples, and for biological samples reports are restricted to plasma and serum and exceedingly rare examples of animal tissue and eggs (14). Even more rare are reports of TBBP-A derivatives in environmental matrices, which is due in part VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8615
to the current lack of sensitive and validated analytical methods for their determination (4, 7, 10). We recently reported on an increasingly complex array of BFRs including polybrominated diphenyl ether (PBDE) and non-PBDE BFRs in herring gulls (Larus argentatus), specifically in eggs, from colony sites spanning the Laurentian Great Lakes of North America. Quantifiable non-PBDE BFRs included HBCD (and isomers), 1,2-bis(2,4,6-tribromophenoxy)ethane (BTBPE), and decabromodiphenyl ethane (DBDPE) (15, 16). Levels of these non-PBDE BFRs in herring gull eggs ranged from sub ng/g wet weight to several hundreds of ng/g wet weight depending on the location of the gull nest and the year of collection. The objectives of the present study were to (1) develop a sensitive method capable of quantifying TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe at the low to sub ng/g wet weight levels in avian eggs, and (2) screen for these TBBP-A derivatives in representative herring gull eggs collected (in 2008) from selected sites in the lower Great Lakes basin and St. Lawrence River areas.
Experimental Methods Standards and Chemicals. Tetrabromobisphenol-A-bis(2,3dibromopropyl ether) (TBBP-A-dbpe) and tetrabromobisphenol-A-bis(allyl ether) (TBBP-A-ae) were purchased from Wellington Laboratories (Guelph, ON, Canada). Tetrabromobisphenol-S-bis(2,3-dibomopropyl ether) (TBBP-S-dbpe) was purchased from Hainan Zhongxin Chemical Co. (Haikou, China). Their simplified molecular structures are shown in Figure 1. All these chemicals were used without further purification. Individual stock standard solutions of analytes were prepared by dissolution in 2-propanol (100-200 mg/ L), and the mixtures or individual working standard solutions were prepared by dilution of stock solution with the appropriate volume of methanol. All the stock and working standard solutions were stored at 4 °C in the dark to prevent any possible photodegradation. Diatomaceous earth (DE) was purchased from Varian Inc. (Mississauga, ON, Canada). DE was treated in a muffle furnace at 600 °C overnight (>12 h) prior to use. Bakerbond SPE silica gel (SiOH) disposable, solid phase extraction (SPE) columns (6 mL, 500 mg, 47-60 µm) were purchased from VWR (Mississauga, ON, Canada). Anisole, chlorobenzene, and methyl tert-butyl ether (MTBE) were purchased from SigmaAldrich (Oakville, ON, Canada). All the solvents (2-propanol, dichloromethane, methanol, acetone, toluene, and water) were HPLC grade from Fisher Scientific (Ottawa, ON, Canada) or Caledon Laboratories Ltd. (Georgetown, ON, Canada). Samples. Herring gull egg samples were collected during 2008-2009 from breeding colony sites in the St. Lawrence River at Ile Deslauriers, Montreal, Canada, and in Lakes Ontario, Huron, and Erie in the Great Lakes basin (Figure S1 in Supporting Information). The Great Lakes eggs samples were collected as part of Environment Canada’s Great Lakes Herring Gull Monitoring Program (GLHGMP) (17). Individual eggs were homogenized and stored at -20 °C at Environment Canada’s National Wildlife Specimen Bank (EC-NWSB) prior to chemical analysis. Chicken eggs purchased from local markets in Ottawa (Canada) were used to assess recovery efficiency and method limits of detection. Sample Extraction and Cleanup. About 3 g of egg sample was weighed and transferred into a glass mortar, and then ground with about 6 g of DE until a free-flowing homogeneous mixture was obtained. Extraction was performed by accelerated solvent extraction (ASE200, Dionex, Oakville, ON, Canada) using DCM/hexane (50:50 v/v). The optimized operating parameters were as follows: heat, 5 min; static, 5 min; flush, 100%; purge, 30 s; cycles, 3; pressure, 1500 psi; temperature, 100 °C. The resulting extract was dried using a sodium sulfate column, concentrated and cleaned up by gel permeation chromatography (GPC) and SPE. The GPC 8616
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
FIGURE 1. APPI(-)-ToF full scan (m/z 430-1060 amu) mass spectra of (A) TBBP-A-ae, (B) TBBP-A-dbpe, and (C) TBBP-S-dbpe. The chemical structures of the TBBP-A derivatives are also shown (hydrogen atoms omitted for clarity). was performed on an Autoprep 2000 (O.I. Analytical, College Station, TX) with a S-X3 column (50 cm ×2.5 cm i.d.). DCM/ hexane (50:50 v/v) was used as the mobile phase and the flow rate was 5 mL/min. The fraction containing the target compounds (150-300 mL) was collected, concentrated, and solvent exchanged into hexane. The sample was then loaded on a silica gel SPE column, which was preconditioned by elution with 4 mL of DCM and 4 mL of hexane. The first fraction was eluted with 6 mL of DCM/hexane (10:90 v/v) and discarded. The target compounds were then eluted from the column with 6 mL of DCM/hexane (50:50 v/v). This fraction was collected and solvent evaporated to dryness, and reconstituted in 200 µL of methanol for analysis by liquid chromatography-mass spectrometry (LC-MS). Mass Spectral Characterization By LC-APPI(-)-Q-ToFMS. LC-APPI-quadrupole-time-of-flight-MS (LC-APPI-QToF-MS) was used to generate the mass spectra of TBBPA-ae, TBBP-A-dbpe, and TBBP-S-dbpe by injection of individual standard solutions. The analyses were performed on an Agilent 1200 LC coupled to an Agilent 6520 Q-ToF with APPI source operating in the negative mode. The LC column
TABLE 1. Retention Times and Operating Parameters for the Quantitative LC-APPI(-)-MS/MS Analysis of TBBP-A-dbpe, TBBP-A-ae, and TBBP-S-dbpe compound
RT (min)
transition (m/z)
cone (V)
coll. (eV)
TBBP-S-dbpe TBBP-A-ae TBBP-A-dbpe
6.41 6.49 6.78
997.4 > 79 655.8 > 79 975.5 > 79
35 50 35
70 60 70
was a ZOBRAX SB-C18 column (2.1 mm × 30 mm, with 3.5 µm particle size). Acetone was used as mobile phase and the flow rate was 0.3 mL/min with no additional dopant being necessary to add after LC separation. The MS source parameters were set with a capillary voltage of 5.8 kV. The fragmentor and skimmer voltage were 150 and 80 V, respectively. The gas temperature was 200 °C with a dry gas flow rate of 10 L/min and nebulizer set at 30 psi. Mass spectra were acquired by full scan analysis over a range of m/z 50 to 1700 and a scan rate of 1.4 spectra/sec. The ToF-MS demonstrated high resolution as the resolution was over 18,000 at m/z 601.978977 within 3 ppm mass error. Quantitative Sample Analysis By LC-APPI(-)-MS/MS. To increase the sensitivity of quantitative analysis of the target compounds in egg sample extraction fractions, LC-APPI(-)tandem quadrupole MS (LC-APPI(-)-MS/MS) was used. The separation of target compounds was carried out using a Waters 2695 liquid chromatograph equipped with an ACE 3 C18 analytical column (2.1 mm × 50 mm, 3 µm particle size) and an ACE 3 C18 guard column (2.1 mm × 10 mm, 3 µm particle size; Advanced Chromatography Technologies, Aberdeen, UK). Temperature of the LC column was kept constant at 40 °C. The mobile phases consisted of water (A) and acetone (B), and the LC flow rate was 0.2 mL/min. The following gradient was employed: 0% B ramped to 100% B in 0.01 min after injection and held for 15 min, followed by a final change to 100% A in 1 min, and then held for 14 min. A 50-µL aliquot of the sample was injected into the LC system. The mass spectrometer was a Waters QuattroUltima triple quadrupole MS (Waters, Milford, MA), with PhotoMate APPI (krypton UV lamp) source operated in the negative mode (Syagen Technology, Tustin, CA). Nitrogen was used as nebulizing gas and dissolvent gas. Argon was used as collision gas when multiple reaction monitoring (MRM) mode was used. The source temperature and probe temperature were 150 and 300 °C, respectively. Cone and desolvation gas flow rates were 100 and 700 L/h, respectively. The compounddependent operation parameters and retention times are listed in Table 1. The data analysis was performed using QuanLynx software, Version 4.0. Quantification was performed using external standard method with a five-point calibration curve spanning the range of anticipated analyte concentrations in the egg samples. The matrix effect (ME) value was measured and calculated as described previously (18). The linearity was tested with concentrations of target compounds ranging from 0 to 1000 ng/mL. The instrument detective limit (IDL) was measured by the lowest amount of target compounds introduced into the LC-MS/MS system and resulting in a response peak with S/N of 3 (peak to peak). The limit of detection (LOD) and the limit of quantitation (LOQ) were defined as the concentration of target compounds in spiked egg sample (3 g) producing a peak in chromatogram with a S/N ratio of 3 and 10 (peak to peak), respectively. A series of standard solutions ranging from 0 to 1000 ng/mL for each of TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe were prepared and analyzed to determine the dynamic linear APPI(-) response ranges. All the target compounds showed very good linearity and coefficients of determination of >0.996. For each batch of n ) 5 gull egg samples, one procedure blank sample was run. There were
no detectable (>LOD) concentrations of target compounds found in the procedure blank samples.
Results and Discussion LC-APPI(-)-Q-ToF Mass Spectral Characterization. Recently, an array of analytes, but especially for several neutral compounds, has been shown to be highly responsive to APPI (19). In the present study, individual solutions of TBBP-Aae, TBBP-A-dbpe, and TBBP-S-dbpe were analyzed by LCAPPI(-)-Q-ToF-MS. The ToF full scan mass spectrum of TBBP-A-ae showed a very weak [M - H]- isotope ion cluster (e.g., m/z 622.808) and two dominating isotope ion clusters of [M + O2]- (e.g., m/z 655.805) and [M + O2 - HBr]- (e.g., m/z 575.880), as well as an isotope ion cluster identified as [M + O4]- (e.g., m/z 687.784) (Figure 1A). The ion clusters centered at m/z 655.805 and m/z 687.784 also had isotopic distribution patterns expected for the presence of four bromine atoms. The mass spectrum of TBBP-A-dbpe was dominated by a [M + O2]- ion cluster (e.g., m/z 975.473), while [M - H]- (e.g., m/z 942.475) was very low in abundance (Figure 1B), and their isotopic distribution pattern was in accordance with the presence of eight bromine atoms. In addition to the ion clusters of [M + O2 - HBr]- (e.g., m/z 895.548) and [M - H - HBr]- (e.g., m/z 864.564) for TBBPA-dbpe, there was another ion cluster centered at m/z 1018.485 identified as [M + O2 + C2H3O]-. The mass spectrum of TBBP-S-dbpe (Figure 1C) showed fragment ions similar to TBBP-A-dbpe (Figure 1B), with the exception that along with the dominant ion cluster of [M + O2]- (e.g., m/z 997.388) there were ion clusters of [M + O2 + C2H3O]- (e.g., m/z 1040.402), [M - H]- (e.g., m/z 964.391), [M + O2 - HBr](e.g., m/z 917.471) and [M - H - HBr]- (e.g., m/z 884.481). When the Q-ToF was operated in the MS/MS mode, for all three analytes the [Br]- (e.g., m/z 78.918 and 80.917) fragment ion dominated in abundance relative to the precursor ion of [M + O2]-. To our knowledge, definitive mass spectra of any kind have not as yet been reported for TBBP-A-ae, TBBP-A-dbpe, or TBBP-S-dbpe. Koppen et al. (7) unsuccessfully attempted to analyze TBBP-A-dbpe in environmental samples by LCMS. They found that neither molecular ions nor any typical fragment ions of TBBP-A-dbpe could be detected when ESI was used. APCI could be used to obtain its mass spectrum by direct infusion, but the low sensitivity rendered it impractical for analysis at trace levels. Shi et al. (20) reported on the analysis of TBBP-A-dbpe by low-resolution GCquadrupole-MS but the obtained mass spectrum could not be explained by the molecular structure of TBBP-A-dbpe. For example, an isotopic ion fragment cluster centered near m/z 493 was typical of the presence of three bromine atoms rather than four bromine atoms, which they concluded to be present. Some experiments in other laboratories (7), as well as in our laboratory also showed that common GC-MS method was unsuitable for analysis of these compounds, because they are not thermally stable. Quantitative Analysis By LC-APPI(-)-MS/MS and Method Validation. Using high resolution APPI(-)-Q-ToF-MS we were able to obtain detailed mass spectral information for TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe. However, the sensitivity was found to be much higher using low-resolution LC-APPI(-)-MS/MS rather than high-resolution LC-APPI(-)Q-ToF-MS, because ToF can only be operated in full scan mode. Therefore, depending on the mass spectra obtained from Q-ToF, specific MRM transitions were used to make quantitative analysis for these target compounds by LCAPPI(-)-MS/MS (Table 1). Water, methanol, and acetonitrile are all common mobile phases used in reversed phase HPLC. It is well-known that the use of acetonitrile results in lower sensitivity using an APPI source (Kr UV lamp) due to its high ionization energy VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8617
(IE ) 12.20 ev). Therefore, methanol/water was chosen as the mobile phase for further assessment. For LC-APPI(-)MS/MS analysis, different dopants were initially tested. In initial tests methanol was used as mobile phase and the dopant was introduced into the APPI source by a “T” connector and a syringe pump after LC separation under different flow rates ranging from 0 to 10% v/v of the mobile phase. In this range, toluene as dopant resulted in the greatest APPI(-) response for TBBP-A-ae, TBBP-A-dbpe, and TBBPS-dbpe, whereas acetone or MTBE as dopants resulted in about 30% lower responsiveness relative to toluene. Anisole and chlorobenzene as dopants gave even lower relative responsiveness accompanied by higher noise and thus lower S/N ratios. However, when the LC separation was operated in gradient mode (methanol/water (5:95 v/v) to 100% methanol) the resulting mass chromatogram had increased noise with toluene as dopant. This is because toluene has low water miscibility and this resulted in high noise in the chromatogram when the mobile phase was not 100% methanol. Therefore, acetone was deemed the most ideal dopant. A common dopant flow rate is 1-10% of the flow rate of the mobile phase. However, in this study the APPI(-) response of TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe showed increases even after the acetone flow rate was increased to 10% of the mobile phase flow rate. Therefore, different mobile phases, which contained different proportions of acetone in methanol, were prepared. For TBBP-Sdbpe and TBBP-A-dbpe, the APPI(-) responses were not increased when the acetone proportion was increased to around 30%. In contrast, for TBBP-A-ae the APPI(-) responsiveness generally continued to increase until the proportion of acetone was 90% of the total mobile phase volume (Figure S2). Since APPI(-) responsiveness could be maximized for TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe using pure acetone as the mobile phase, acetone was used for LC-APPI(-)-MS/MS analysis of these BFRs. In most cases acetone is not considered as a suitable mobile phase for LC when a UV detector is used, since acetone has a high absorption wavelength cutoff (λcutoff ) 330 nm). However, given the low viscosity (0.32 cP) of acetone and that UV absorption issues are not a problem in LC-MS, it was a very suitable mobile phase for the analysis of TBBP-A-ae, TBBPA-dbpe, and TBBP-S-dbpe when APPI source was used. To further improve the limits of detection (LODs) of TBBPA-ae, TBBP-A-dbpe, and TBBP-S-dbpe, a proven on-column concentration approach was used. It has been shown that this approach utilizes a relatively large volume (50 µL) of sample injection into the LC system with 100% water for the mobile phase (21). In the present study, the BFRs were first concentrated on the head of column post-injection, with subsequent separation with increasing proportions of acetone in the LC mobile phase. The mass chromatographic peak shapes of TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe were found to be improved (e.g., narrowing) as well as lowering the background noise and thus increasing the S/N ratio. It should be mentioned that the solvent used for preparing standard solutions and samples was methanol even though the LC mobile phase was acetone. We found that if acetone was used for preparing standard solutions and sample fractions, the chromatographic peaks were more broad for these BFRs. Using the optimized LC-APPI(-)-MS/MS quantitative method, low IDLs, LODs, and LOQs were achieved (Table 2) for TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe. Although other studies on environmental matrices for comparative purposes are extremely limited, the present LOD concentrations for these BFRs of low and sub ng/g wet weight were 1-2 orders of magnitude lower than what has been previously reported. Koppen et al. (7) used LC-DAD/MS to analyze TBBPA-dbpe and reported an LOD of 10 ng/g dry weight (dw) for 8618
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
TABLE 2. Quantitative LC-APPI(-)-MS/MS Analysis of TBBP-A-ae, TBBP-A-dbpe, and TBBP-S-dbpe and Their Matrix Effects (ME) in Spiked Chicken Egg Samples, Instrument Detection Limits (IDLs), Limits of Detection (LODs), and Limits of Quantitation (LOQs)a IDL LOD LOQ (pg) (ng/g wet weight) (ng/g wet weight) TBBP-A-ae 12 TBBP-A-dbpe 32 TBBP-S-dbpe 112
0.03 0.07 1.28
0.08 0.25 4.27
ME (%) (n ) 7) 108.3 ( 15.5 99.6 ( 8.2 97.3 ( 9.7
a See the Experimental Methods section for details on the approaches for IDL, LOD, and LOQ determination.
sediment samples and 22 ng/g dw for sewage sludge samples. Knudsen et al. (2007) recently reported using LC-DAD/MS that the LOD for TBBP-A-dbpe was 0.99 µg/mL (or µg/g wet weight) for dosing solutions and blood and bile samples from dosed rats. Furthermore, in the same rat dosing study with [14C]-labeled TBBP-A-dbpe, it was reported that LC-scintillation analysis of blood and bile resulted in a LOD of 19 ng/mL (13). The recovery efficiencies of target compounds by the present LC-APPI(-)-MS/MS based method were determined by spiking 20 ng of the target compounds to 3 g of chicken egg samples (n ) 7 replicate spike analysis). No quantifiable concentrations of the target compounds were found in the original chicken egg sample. Using an external standard method, the recovery efficiencies were 89 ( 14%, 55 ( 16%, and 63 ( 17% for TBBP-S-dbpe, TBBP-A-ae, and TBBP-Adbpe, respectively. Based on calculations reported in Chu and Letcher (18), the percent matrix effects (%MEs) of LCAPPI(-)-MS/MS analysis of TBBP-S-dbpe, TBBP-A-ae, and TBBP-A-dbpe were about 100% (Table 2), and clearly showed that the ionization suppression or enhancement was negligible, which attested to the minimal presence of ion suppressing components in real elution time. It is more accurate to use exact isotope-enriched target compounds as internal standard to properly compensate for ionization effects by matrix and the loss during sample preparation. However, at present, isotope (e.g., 13C) labeled internal standards of TBBP-S-dbpe, TBBP-A-ae, and TBBP-A-dbpe are not currently available. Target BFRs in Herring Gull Eggs. Out of a list of 22,263 substances, Howard and Muir (22) very recently identified a list of 610 commercially produced substances of probable persistence and bioaccumulation relevance to the environment, although only 101 of these 610 chemicals have been measured in environmental media. Of these 610 substances, 13% were brominated. TBBP-A was included in the top 10 brominated substances of environmental priority, and also on the list of 610 were TBBP-A-ae and TBBP-A-dbpe. This sublist of 610 environmental priority substances was determined using criteria such as log Kow values as a measure of bioaccumulation potential, and for TBBP-A-ae and TBBPA-dbpe were quoted as 10.02 and 11.52 and thus with log Kow values >3 can be considered bioaccumulative (22). These log Kow values are roughly consistent with those of Jonsson and Ho¨rsing (23) who recently calculated (using ALOGPS 2.1 modeling software) the log Kow values for TBBP-A-ae and TBBP-A-dbpe as being 8.63 and 9.99, respectively. Substances are also considered as being persistent in the atmosphere if they have an atmospheric half-life (AOt1/2) of >2 days, although the AOt1/2 for TBBP-A-dbpe and TBBP-A-ae were reported to be 0.42 and 1.02 days, respectively (22). When considering the reported log Kow and AOt1/2 values combined, the present finding of low levels of TBBP-A-dbpe and TBBP-
FIGURE 2. LC-APPI(-)-MS/MS analysis and MRM mass chromatograms (top to bottom) of (A) TBBP-S-dbpe, TBBP-A-dbpe, and TBBP-A-ae and TIC (m/z 50-1700) of all MRM transitions in standard solution (25 ng/mL each), and (B) TBBP-S-dbpe, TBBP-A-dbpe, and TBBP-A-ae and TIC (m/z 50-1700) of all MRM channels in a representative herring gull egg sample (Location No. 7, Big Chicken Is. (Lake Ontario); see Table 3 and Figure S1). For each mass chromatogram, on the far right are the MRM or TIC ion details as well as the maximum mass chromatographic response intensities. A-ae in herring gull eggs (Table 3) is consistent with their low atmospheric persistence and bioaccumulative potential in biota. In the Laurentian Great Lakes basin of North America, the herring gull is an opportunistic, fish-eating bird, and nests in key colonies spanning this region. The herring gull
has been used for several decades to monitor bioaccumulative contaminants in this watershed ecosystem (17). Recently, we have reports on numerous non-PBDE BFRs in the eggs of this biomonitoring species (15, 16). In BFR-containing fractions from a subset of these herring gulls eggs, GC-ECNI/ MS revealed rather strong [Br]- mass chromatographic peaks suspected to be TBBP-A-dbpe and TBBP-S-dbpe (16). However, these [Br]- responsive peaks were found not to be TBBP-A-dbpe and TBBP-S-dbpe, as both compounds were found to be not amenable to GC as they were found to undergo total thermal degradation in the injection port with only slightly less degradation when using on-column GC injection. Using the present LC-APPI(-)-MS/MS based method, various individual herring gull eggs recently collected from several colonial sites in the Great Lakes-St. Lawrence River system (Figure S1) were analyzed for TBBP-A-ae, TBBP-Adbpe, and TBBP-S-dbpe. In comparison to a standard solution containing TBBP-A-ae, TBBP-A-dbpe, and TBBPS-dbpe (Figure 2A), the mass chromatograms from the optimized MRM transitions for a representative herring gull egg fractions revealed the presence of quantifiable TBBPA-ae and TBBP-A-dbpe, but no detectable (
