- Email: [email protected]
Determination of hexabromocyclododecane diastereoisomers in air and soil by liquid chromatography–electrospray tandem mass spectrometry
Journal of Chromatography A, 1190 (2008) 74–79
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Determination of hexabromocyclododecane diastereoisomers in air and soil by liquid chromatography–electrospray tandem mass spectrometry Zhiqiang Yu, Ping’an Peng, Guoying Sheng, Jiamo Fu ∗ State Key laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China
a r t i c l e
i n f o
Article history: Received 20 April 2007 Received in revised form 14 February 2008 Accepted 22 February 2008 Available online 2 March 2008 Keywords: Hexabromocyclododecane Brominated flame retardants LC–MS/MS Diastereoisomers Matrix effects
a b s t r a c t Hexabromocyclododecanes (HBCDs), used as additive brominated flame retardants, are of high concern due to their widespread use and increasing levels in various environmental systems. High-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry (LC–MS/MS) was developed for the determination of HBCD diastereoisomers. A detailed study was carried out to optimize the composition of the mobile phase involving methanol/acetonitrile/water, and the values of MS/MS parameters. It was found that the mobile phase could simultaneously affect the chromatographic separation and sensitivity. The instrumental limits of detection (LODs) on column in this study were 0.5, 0.3 and 0.3 pg for ␣-HBCD, -HBCD and ␥-HBCD, respectively. The effects of extracted matrix components on HBCD determination were investigated by spiking air and soil sample extracts with three 13 C-labelled individual stereoisomers. The results indicated that the responses of the HBCD analysis in air and soils were not significantly affected by matrix effects. The method reported here was further applied to the air and soil samples. Three HBCD diastereoisomers were detected in all the air and soil samples, with levels ranging from 1.2 to 1.8 pg/m3 and 1.7 to 5.6 ng/g dry weight, respectively. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Hexabromocyclododecanes (HBCDs) are additive brominated flame retardants (BFRs) and are mainly used as thermal insulation in buildings, in upholstery textiles and in electronics [1]. Since the production and use of penta- and octabromodiphenyl ethers has been banned in Europe, HBCDs might become the replacement for polybrominated diphenyl ethers (PBDEs) in some applications [2] and slight increases in their concentrations have been found in various environmental systems [3]. Recent studies indicated that HBCDs are ubiquitous organic contaminants and share the major characteristics of persistent organic pollutants (POPs): persistency, bioaccumulation, long-range transport, and toxicity [3]. They have been included on the OSPAR (The Convention for the Protection of the Marine Environment of the North-East Atlantic) list of chemicals for priority action [4]. Technical grade HBCD mixtures are produced via bromination of cyclododecane-1,5,9-triene isomers. The commercial HBCDs on the market mainly consist of ␣-HBCD, -HBCD and ␥-HBCD. Although HBCD concentration can be determined as the total concentration of three diastereoisomers by gas chromatography (GC) with electron-capture detection (ECD) or mass spectrometry (MS), these techniques cannot separately measure different stereoiso-
∗ Corresponding author. Tel.: +86 20 85290196; fax: +86 20 85290706. E-mail address: [email protected] (J. Fu). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.02.082
mers due to thermal rearrangement at temperatures above 160 ◦ C and thermal decomposition at temperatures above 240 ◦ C [5–7]. Further details regarding GC and GC–MS analysis can be found in recent reviews [8,9]. Liquid chromatography coupled to mass spectrometry (LC–MS) or tandem mass spectrometry (LC–MS/MS) is currently the preferred method for determining diastereoisomers in environmental samples and biota [10]. Signal intensity improvements gained using quadrupole and ion-trap mass spectrometry in different ionization modes such as atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) have been discussed previously [10,11]. Using LC–MS/MS and multiple reaction monitoring for [M−H]− (m/z 640.6) → [Br− ] (m/z 79 and 81), Budakowski and Tomy [11] developed a sensitive method with the instrument limit of detection (LOD) of 4–6 pg for on column for a standard solution of ␥-HBCD. Comparison of the results of different ionization modes indicated that ESI has a greater signal intensity than the APCI mode. Janak et al. [12] further reported a baseline separation method using an RP-C18 column and methanol/acetonitrile/water as the mobile phase. The LOD, defined as three times the noise level, was 0.5, 1 and 5 pg for ␥-HBCD, ␣HBCD and -HBCD, respectively. Although most studies have used the ESI ionization mode of HBCDs, Suzuki and Hasegawa [13] first reported the HBCDs diastereoisomers and tetrabromobisphenol A (TBBPA) in water and sediment samples by using LC–APCI–MS. Their results showed that ionization of HBCDs by APCI led to a 2–5 times higher S/N ratio compared with ESI. They also concluded that ESI was more sensitive to matrix components. In addition,
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
atmospheric pressure photoionization (APPI) was also compared with the ESI and APCI modes for HBCD analysis [14]. Although APPI can better resolve common problems in ESI and APCI such as matrix effects, the APPI signal intensities were approximately 10 times lower compared with ESI. Many researchers [15–17] have reported that matrix effects might adversely affect the reproducibility and quantification of target analyte when developing an LC–MS/MS method, especially when using ESI as the ionization mode. For HBCD diastereoisomer analysis, different studies have presented contrary results. Budakowski and Tomy [11] confirmed the existence of the “ion suppression phenomenon” when determining HBCD diastereoisomers in both biotic and sediment samples using LC–ESI–MS/MS. Dodder et al. [18] developed a baseline separation method for HBCD diastereoisomer determination in the biological tissues. Investigation of the matrix effect showed that potential matrix interference did not significantly influence the LC–MS/MS analyses of the diastereoisomers, whereas the response of the HBCD enantiomers in tissue samples were greatly influenced by matrix effects and other changes to the ionization conditions. As we know, HBCD can be analyzed using the same protocols as those reported in the PBDE study [12,19]. After finishing the PBDE measurements, the solvent was exchanged to methanol or acetonitrile for further HBCD analysis. Therefore, it is necessary to investigate whether the sample preparation and clean-up procedure can fulfill the determination requirements of diastereoisomers. In this study, a widely used clean-up procedure for soil and air samples was investigated for matrix effects by spiking 13 C-labelled HBCDs before injection. Also the optimization of MS/MS parameters for improving signal responses and the effects of mobile phase on the chromatographic separation and sensitivity are discussed in detail. The method developed in this study obtained completely baseline chromatographic separation and lower LODs. 2. Experimental 2.1. Chemicals and solvents Unlabelled and 13 C12 -labelled ␣-, -, ␥-HBCD were purchased from Cambridge Isotope Labs. (Andover, MA, USA) and used as received. The HPLC-grade methanol and acetonitrile were obtained from Merck (Darmstadt, Germany). Ammonium acetate and acetic acid were acquired from J.T. Baker (Phillipsburg, NJ, USA). Analytical grade hexane and methylene chloride were redistilled by a glass system. 2.2. Liquid chromatography An Agilent 1100 series HPLC system (Agilent Technologies, Palo Alto, CA, USA) equipped with a vacuum degasser, quaternary pump and autosampler was used. The separation was performed on a Zorbax SB-C18 reversed-phase column (250 mm × 4.6 mm I.D., 5 m, Agilent). The gradient mobile phase consists of methanol (A)/acetonitrile (B)/water with 10 mM ammonium acetate (C). The flow rate was set at 0.5 mL/min. The gradient program started at an initial composition of 80:10:10 A/B/C (v/v) and was ramped to 50:40:10 A/B/C in 18 min, followed by 30:70 A/B at 23 min, and was held for 7 min, then returned to 80:10:10 A/B/C in 8 min. The column was equilibrated for a further 6 min. The detailed program is listed in Table 1. 2.3. Mass spectrometry An Applied Biosystems-Sciex API 4000 (Applied Biosystems, Foster City, CA, USA) triple quadrupole mass spectrometry
75
Table 1 Gradient mobile phase program for the separation of ␣-, - and ␥-HBCD Time (min)
Methanol
Acetonitrile
Water (10 mM ammonium acetate)
0.0 18.0 23.0 30.0 37.0 43.0
80 50 30 30 80 80
10 40 70 70 10 10
10 10 0 0 10 10
equipped with a TurboIonSpray ionization interface was used. The Q1 scan range was m/z 630–660, operated with unit resolution for a scan time of 0.5 s. The quantities of the three diastereoisometric HBCDs were detected in electrospray ionization negative ion mode using multiple reaction monitoring (MRM) for [M−H]−1 → Br− (m/z 640.6 → 79 and 652.6 → 79 for native and 13 C-labelled HBCD, respectively), utilizing unit resolution and a 200 ms dwell time per transition. Details of the optimized MS/MS parameters are listed in Table 2. 2.4. Sample collection Four air samples were collected during 11–15 November 2006 from the Tianhe district, Guangzhou. The sampling procedure was described elsewhere [20,21]. Before sampling, glass fiber filters (GFFs, Whatman, Maidstone, UK) were baked at 450 ◦ C for 4 h to remove any organic contaminant, and polyurethane foam (PUF) plugs were Soxhlet extracted for 48 h with each of following solvents:methanol and 1:1 acetone–hexane. Air volumes of 800–860 m3 were drawn at 0.3–0.5 m3 /min for 32–35 h using a high-volume air sampler. After sampling, loaded GFFs were wrapped with prebaked aluminum foil and sealed with double layers of polyethylene bags, and PUFs were placed in solvent rinsed glass jars with aluminum foil-lined lids and stored at −20 ◦ C until extraction. Three surface soils were sampled from the Tianhe, Huangpu and Liwan districts in the Guangzhou urban areas in December 2006. Each of the surface soil samples was collected using a pre-cleaned stainless steel scoop, and mixed with 5 cores from 0 to 10 cm depth. The soil samples were kept at −18 ◦ C until analysis. Before extraction, soil samples were freeze-dried for 24 h and sieved through a 10-mesh stainless steel mesh. 2.5. Sample extraction and preparation The sample extraction and preparation procedures have been described elsewhere [20,21] and modified in this paper. In brief, soil and air samples (including PUF plugs and GFF glass fiber filters) were extracted with a mixture of acetone and hexane (1:1) for 72 h with a Soxhlet extractor. Activated copper granules were added into the extractor flask to remove elemental sulfur. After extraction, the concentrated extracts from the soil and air samples were cleaned Table 2 Optimised MS/MS parameters for the determination of ␣-, - and ␥-HBCD Parameter
Optimised value
Source temperature, TEM ( ◦ C) Ionization voltage (V) Ion source (GS1) setting Ion source (GS2) setting Curtain gas settings CAD gas setting Declustering potential (V) Entrance potential (V) Collision energy (V) Collision cell exit potential (V)
300 4500 50 55 10 25 −100 −8 −50 −6
76
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
and fractionated in a 10-mm I.D. silica/alumina column packed, from the bottom to top, with neutral alumina (6 cm, 3% deactivated), neutral silica gel (4 cm, 3% deactivated), silica with 50% conc. sulfuric acid (8 cm), and anhydrous sodium sulfate (1 cm). The PBDE and HBCD mixture was eluted with 30 mL of hexane and 60 mL of hexane:methylene chloride (1:1). The final extracts containing PBDEs of both soils and air samples were concentrated to 200 L in hexane for GC–MS analysis. Before HBCD analysis, the extracts were further cleaned using 1 g acidified silica with 50% conc. sulfuric acid and eluted with 10 mL of hexane:methylene chloride (1:1), which is helpful for decreasing matrix effects. For determination of HBCD diastereometers, the solvent was exchanged to methanol for LC–MS/MS analysis. 2.6. Evaluation of matrix effects The matrix effects were investigated following the procedures reported elsewhere [15]. In brief, MS/MS areas with known amounts of three 13 C-labelled diastereoisomer standards (A) were compared with those measured in soil and air extracts spiked, after extraction, with the same analyte amount (B). The ratio (B/A) is defined as the absolute matrix effect (ME). If the value of ME is 1, it means that there is no matrix effect. There is signal enhancement if ME >1 and signal suppression if ME <1. 2.7. Calibration curves Quantification was performed by an isotope-dilution method using 13 C12 -labelled HBCD as an internal standard, which is described elsewhere [22]. The calibration curve was derived with a series of standards ranging from 1 to 100 g/L and fixed concentrations (20 g/L) of the internal standards (r2 > 0.998). The response data were adjusted according to the relative ratios of the responses to the stable isotope internal standards in the sample and the standard. 3. Results and discussion 3.1. Optimization of chromatographic resolution and MS/MS parameters 3.1.1. Effects of mobile phase on the chromatographic separation and sensitivity For improving HBCD chromatographic separation, different columns (such as C8 and C18 ) and mobile phases were investigated in this study. The results indicated that the C18 column was the optimal one. Fig. 1a–c shows the chromatograms of HBCD diastereoisomer analysis using the C18 column and three mobile phase compositions. It was found that (1) the three diastereoisomers keep the same elution order when eluting by three mobile phases; (2) methanol and acetonitrile present different elution ability for the three individual stereoisomers, and -HBCD was retained in the column for a longer time when acetonitrile was replaced by methanol as the mobile phase. This result was also confirmed by Morris et al. [10] and Hoh and Hites [19]. The final gradient composition (Table 1) is composed of methanol, acetonitrile and water. As shown in Fig. 1c, the three diastereoisomers were adequately separated from each other. The results also suggest that the sufficient baseline chromatographic separation is benefit to reduce the matrix effect. Fig. 2 shows the effects of the mobile phase on sensitivity. Methanol and acetonitrile were chosen as the organic modifiers and the MS/MS was operated in a negative ESI using MRM for [M−H]− (m/z 640.6) → Br− (m/z 79.0 and 81.0). A series of mobile phase compositions were tested with the ratio of organic modifier to water in the range 10:0 to 6:4. The results indicated that: (1) the signal
Fig. 1. Chromatographic separation of three HBCD diastereoisomers by different mobile phases. (a) 9:1 methanol/water, (b) 9:1 acetonitrile/water and (c) gradient compositions described in Table 1.
responses of the three diastereoisomers follow the order  > ␥ > ␣, and (2) ␣- and -diastereoisomers have a stronger response when using methanol as the organic modifier in comparison with that using acetonitrile, whereas ␥-HBCD shows the reverse response. In order to enhance the signal response, mobile phase modifiers such as ammonium acetate and formic acid are also discussed. 10 mM ammonium acetate was found to provide the maximum response and was used in the following method development. 3.1.2. MS/MS optimization For quantification purposes, the MRM signal from the m/z 640.6 → 79 transition was used. The MS/MS optimization was performed through flow injection analysis (FIA) using the ␣-, -, ␥-diastereoisomers. In FIA, the flow rate was set at 0.5 mL/min and the mobile phase represented the composition present when the analyte was entering the source from the column. The compositions (methanol/acetonitrile/water) were 58/32/10, 55/35/10, and 51/39/10 for ␣-, -, ␥-HBCD, respectively. Each parameter set was run six times to ensure response stability. The final optimized MS/MS parameters are listed in Table 2 and are chosen
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
77
Fig. 3. Effects of different turbo-gas temperatures on the sensitivity of three diastereoisomers using MRM for [M−H]− (m/z 640.6) → Br− (m/z 79.0 and 81.0).
3.2. Matrix effects
Fig. 2. Effects of different mobile phases on the sensitivity of three diastereoisomers using MRM for [M−H]− (m/z 640.6) → Br− (m/z 79.0 and 81.0).
to optimize maximum signal response and stability. Compared with the results in previous papers, it was found that optimization of MS or MS/MS parameters on different instruments might be inconsistent. For example, the optimum temperature of the turbo gas was about 300 ◦ C (Fig. 3) in this study. It was shown in Fig. 3 that the three diastereoisomers showed different sensitivities when increasing the temperature from 150 to 450 ◦ C. -HBCD has the highest signal response when the temperature was increased to 400 ◦ C, whereas the maximum sensitivity of ␣- and ␥-HBCD was observed at 300 ◦ C. The compromise temperature was set at 300 ◦ C because most environmental research focuses on the determination of ␣- and ␥-HBCDs. Budakowski and Tomy [11] and Dodder et al. [18] reported that the optimal temperatures of the turbo gas were 500 and 450 ◦ C, respectively, whereas temperatures of 160 and 250 ◦ C have been used elsewhere [10,13].
When developing an LC–MS/MS method, the matrix effect is a major issue. When co-eluting compounds and analytes enter the ion source together, the ionization efficiency of the analyte might be affected, resulting in signal enhancement or suppression, which are named matrix effects. In this study, the matrix effects on HBCD determination were evaluated by spiking three individual 13 C-labelled stereoisomers into the air and soils matrix after completing sample preparation and clean-up. The detailed evaluation method of the matrix effects was described in Section 2.4. The three diastereoisomers were spiked with concentration levels of 5, 20 and 100 g/L, respectively, and the results are listed in Table 3. As Table 3 shown, ME values range from 0.95 to 0.99, clearly indicated that the extracted matrix components in the air and soil samples will not interfere with the accuracy of the HBCD diastereomer analysis in this study. This result might arise from (1) improved sample preparation and clean-up; or (2) improved chromatographic separation. The gradient mobile phase program not only allows clear baseline separation for all three diastereoisomers, but also sufficiently removed the majority of potential interferences in the sample extracts. It is also necessary to continue the gradient program after the HBCDs are eluted from the column in order to remove the apolar components in the sample matrix. The extent of matrix effects is dependent on the removal of interference in sample extracts and chromatographic separation. Budakowski and Tomy [11] confirmed the “ion suppression phenomenon” when measuring HBCD diastereoisomers for both biotic and soil samples using LC–ESI-MS/MS. This might be explained by simple sample preparation procedures and poor chromatographic Table 3 Matrix effects in air and soil extracts ME (matrix effect) 5 g/L
20 g/L
100 g/L
Soil ␣-HBCD -HBCD ␥-HBCD
0.93 0.95 0.95
0.98 0.98 0.99
0.99 1.00 0.99
Air–gas ␣-HBCD -HBCD ␥-HBCD
0.95 0.95 0.96
0.99 0.99 0.99
0.99 1.00 1.00
Air–particulate matter ␣-HBCD -HBCD ␥-HBCD
0.95 0.94 0.94
0.99 0.99 0.98
0.99 0.99 0.99
78
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
Fig. 4. LC–MS/MS chromatograms of three HBCD diastereomers in the technical mixture, air and soil samples.
separation between - and ␥-HBCD. Dodder et al. [18] developed a baseline separation method for HBCD diastereoisomer determination in the biological tissues. Investigation of the matrix effect showed that potential matrix interferences did not significantly influence the LC–MS/MS analyses of the diastereoisomers. 3.3. Method performance and its application to environmental samples For the validation of the method performance, within- and between-day reproducibility was assessed at two concentrations levels (10 and 50 g/L) in triplicate. The relative standard deviations range from 1.7 to 3.1% and 2.7 to 4.3%, respectively. The LOD on column, defined as a signal-to-noise (S/N) ratio of 3:1, was 0.5, 0.3 and 0.3 pg for ␣-HBCD, -HBCD and ␥-HBCD, respectively. Recovery experiments were performed at concentration levels of 5, 25 and 50 ng by spiking three 13 C-labelled diastereoisomers into air and soil samples. The mean recoveries were 84 ± 11%, 71 ± 7% and 87 ± 8% (n = 6) for ␣-HBCD, -HBCD and ␥-HBCD, respectively. Four air samples (including gas phase and particle phase) and three surface soil samples from Guangzhou city in South China were analyzed using the method developed here. The results are shown in Fig. 4. Levels of total HBCDs found in air samples were in the range of 1.2–1.8 pg/m3 , and they were detected in both the gas phase and particle phase. The mean values of total HBCDs were 1.4 pg/m3 . These data show comparable levels with the results reported in USA background air [19], whereas they are somewhat lower than the observations from background air in Sweden [23]. All the air samples in this study had similar ratios of the three diastereoisomers.
The mean percentages of ␣-HBCD, -HBCD, and ␥-HBCD were 58%, 15%, and 27%, respectively, which is different from commercial mixtures. This may be explained by leaching from the treated products, which has been experienced in high temperature processes [19]. The concentration levels of total HBCDs in soils from urban areas in Guangzhou city ranged from 1.7 to 5.6 ng/g dry weight. To date, few data have been reported on concentration levels in background sites, especially in soils. The only data found by the authors are from soil samples collected near point-sources in Sweden and Belgium/Germany [3], which contained high concentrations of HBCDs, ranging from 111 to 23,200 ng/g dry weight. However, our data reported here are comparable with those in non-point source sediments where concentration levels were less than 10 ng/g dry weight [3]. The ratios of the three diastereoisomers in three soils presented different patterns. Two samples were similar to the commercial product, with ␥-HBCD being the most abundant stereoisomer, whereas the other was dominated by ␣-HBCD. At present, it is still unclear whether this difference in the composition of the HBCD stereoisomers between some soils and technical HBCDs is caused by thermal isomerization during the processing of HBCDs or by stereoisomer-specific processes in the environment. More research is needed to improve our understanding of the source, distribution and fate of individual diastereomers. 4. Conclusion A high-sensitivity and baseline chromatographic separation method for analysis of HBCD diastereoisomers was developed. Chromatographic conditions (including mobile phase, column and
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
modifier) and MS/MS parameters were optimized for a combination of sensitivity and stability. The LODs on column with this method were 0.5, 0.3 and 0.3 pg for ␣-HBCD, -HBCD and ␥-HBCD, respectively. Using the analytical method described in this study, matrix effects were evaluated and show that matrix interferences in soil and air samples did not significantly influence the accuracy and precision of the diastereomer analysis. It should be noted that there still is an important gap in our understanding regarding the determination of HBCDs in environmental matrixes. The optimization of MS or MS/MS parameters in the literature might vary between studies and matrix effects were discussed only in a few papers. A comparison of different methods and international intercalibration studies will be needed for improving the quality of analytical data and comparability of laboratory results. Acknowledgements This study was supported by National Natural Scientific Foundation of China (No. 40590392), and Chinese Academy of Sciences (No. KZCX2-YW-403). Partial funding was also provided by Natural Scientific Foundation of Guangdong Province (No. 5006280) and the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry to Z.Y. References ˚ Bergman, Environ. Int. 29 (2003) 683. ˝ [1] M. Alaee, P. Arias, A. Sjodin, A.
79
[2] U. Sellstrom, A. Kierkegaard, C. de Wit, B. Jansson, Environ. Toxicol. Chem. 17 (1998) 1065. [3] A. Covaci, A.C. Gerecke, R.J. Law, S. Voorspoels, M. Kohler, N.V. Heeb, H. Leslie, C.R. Allchin, J.D. Boer, Environ. Sci. Technol. 40 (2006) 3679. [4] Draft of the EU Risk Assessment Report on Hexabromocyclododecane, National Chemicals Inspectorate (KEMI), Sundbyberg, Sweden, 2005. [5] E. Eljarrat, A. De La Cal, C. Duran, D. Barcelo, Environ. Sci. Technol. 38 (2004) 2603. [6] F. Barontini, V. Cozzani, L. Petarca, Ind. Eng. Chem. Res. 40 (2001) 3270. [7] E.R. Larsen, E.L. Ecker, J. Fire Sci. 4 (1986) 261. [8] R.J. Law, M. Kohler, N.V. Heeb, A.C. Gerecke, P. Schmid, S. Voorspoels, A. Covaci, G. Becher, K. Janak, C. Thomsen, Environ. Sci. Technol. 39 (2005) 281A. [9] A. Covaci, S. Voorspoels, L. Ramosc, H. Neels, R. Blust, J. Chromatogr. A 1153 (2007) 145. [10] S. Morris, P. Bersuder, C.R. Allchin, B. Zegers, J.P. Boon, P.E.G. Leonards, J. de Boer, Trends Anal. Chem. 25 (2006) 343. [11] W. Budakowski, G. Tomy, Rapid Commun. Mass Spectrom. 17 (2003) 1399. [12] K. Janak, A. Covaci, S. Voorspoels, G. Becher, Environ. Sci. Technol. 39 (2005) 1987. [13] S. Suzuki, A. Hasegawa, Anal. Sci. 22 (2006) 469. [14] S. Backus, S. Batchelor, M. Alaee, D. Ueno, L.M. Hewitt, Organohalogen Compd. 67 (2005) 240. [15] J.C. Van De Steene, K.A. Mortier, W.E. Lambert, J. Chromatogr. A 1123 (2005) 71. [16] T. Benijts, R. Dams, W. Lambert, A. De Leenheer, J. Chromatogr. A 1029 (2004) 153. [17] H.B. Lee, K. Sarafin, T.E. Peart, J. Chromatogr. A 1148 (2007) 158. [18] N.G. Dodder, A.M. Peck, J.R. Kucklick, L.C. Sander, J. Chromatogr. A 1135 (2006) 36. [19] E. Hoh, R.A. Hites, Environ. Sci. Technol. 39 (2005) 7794. [20] B.X. Mai, S.J. Chen, X.J. Luo, L.G. Chen, Q.S. Yang, G.Y. Sheng, P.A. Peng, J.M. Fu, E.Y. Zeng, Environ. Sci. Technol. 39 (2005) 3521. [21] L.G. Chen, B.X. Mai, X.H. Bi, S.J. Chen, X.M. Wang, Y. Ran, X.J. Luo, G.Y. Sheng, J.M. Fu, E.Y. Zeng, Environ. Sci. Technol. 40 (2006) 1190. [22] S.G. Chu, C.D. Metcalfe, J. Chromatogr. A 1164 (2007) 212. [23] M. Remberger, J. Sternbeck, A. Palm, L. Kaj, K. Stromberg, Chemosphere 54 (2004) 9.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Determination of hexabromocyclododecane diastereoisomers in air and soil by liquid chromatography–electrospray tandem mass spectrometry Zhiqiang Yu, Ping’an Peng, Guoying Sheng, Jiamo Fu ∗ State Key laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China
a r t i c l e
i n f o
Article history: Received 20 April 2007 Received in revised form 14 February 2008 Accepted 22 February 2008 Available online 2 March 2008 Keywords: Hexabromocyclododecane Brominated flame retardants LC–MS/MS Diastereoisomers Matrix effects
a b s t r a c t Hexabromocyclododecanes (HBCDs), used as additive brominated flame retardants, are of high concern due to their widespread use and increasing levels in various environmental systems. High-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry (LC–MS/MS) was developed for the determination of HBCD diastereoisomers. A detailed study was carried out to optimize the composition of the mobile phase involving methanol/acetonitrile/water, and the values of MS/MS parameters. It was found that the mobile phase could simultaneously affect the chromatographic separation and sensitivity. The instrumental limits of detection (LODs) on column in this study were 0.5, 0.3 and 0.3 pg for ␣-HBCD, -HBCD and ␥-HBCD, respectively. The effects of extracted matrix components on HBCD determination were investigated by spiking air and soil sample extracts with three 13 C-labelled individual stereoisomers. The results indicated that the responses of the HBCD analysis in air and soils were not significantly affected by matrix effects. The method reported here was further applied to the air and soil samples. Three HBCD diastereoisomers were detected in all the air and soil samples, with levels ranging from 1.2 to 1.8 pg/m3 and 1.7 to 5.6 ng/g dry weight, respectively. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Hexabromocyclododecanes (HBCDs) are additive brominated flame retardants (BFRs) and are mainly used as thermal insulation in buildings, in upholstery textiles and in electronics [1]. Since the production and use of penta- and octabromodiphenyl ethers has been banned in Europe, HBCDs might become the replacement for polybrominated diphenyl ethers (PBDEs) in some applications [2] and slight increases in their concentrations have been found in various environmental systems [3]. Recent studies indicated that HBCDs are ubiquitous organic contaminants and share the major characteristics of persistent organic pollutants (POPs): persistency, bioaccumulation, long-range transport, and toxicity [3]. They have been included on the OSPAR (The Convention for the Protection of the Marine Environment of the North-East Atlantic) list of chemicals for priority action [4]. Technical grade HBCD mixtures are produced via bromination of cyclododecane-1,5,9-triene isomers. The commercial HBCDs on the market mainly consist of ␣-HBCD, -HBCD and ␥-HBCD. Although HBCD concentration can be determined as the total concentration of three diastereoisomers by gas chromatography (GC) with electron-capture detection (ECD) or mass spectrometry (MS), these techniques cannot separately measure different stereoiso-
∗ Corresponding author. Tel.: +86 20 85290196; fax: +86 20 85290706. E-mail address: [email protected] (J. Fu). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.02.082
mers due to thermal rearrangement at temperatures above 160 ◦ C and thermal decomposition at temperatures above 240 ◦ C [5–7]. Further details regarding GC and GC–MS analysis can be found in recent reviews [8,9]. Liquid chromatography coupled to mass spectrometry (LC–MS) or tandem mass spectrometry (LC–MS/MS) is currently the preferred method for determining diastereoisomers in environmental samples and biota [10]. Signal intensity improvements gained using quadrupole and ion-trap mass spectrometry in different ionization modes such as atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) have been discussed previously [10,11]. Using LC–MS/MS and multiple reaction monitoring for [M−H]− (m/z 640.6) → [Br− ] (m/z 79 and 81), Budakowski and Tomy [11] developed a sensitive method with the instrument limit of detection (LOD) of 4–6 pg for on column for a standard solution of ␥-HBCD. Comparison of the results of different ionization modes indicated that ESI has a greater signal intensity than the APCI mode. Janak et al. [12] further reported a baseline separation method using an RP-C18 column and methanol/acetonitrile/water as the mobile phase. The LOD, defined as three times the noise level, was 0.5, 1 and 5 pg for ␥-HBCD, ␣HBCD and -HBCD, respectively. Although most studies have used the ESI ionization mode of HBCDs, Suzuki and Hasegawa [13] first reported the HBCDs diastereoisomers and tetrabromobisphenol A (TBBPA) in water and sediment samples by using LC–APCI–MS. Their results showed that ionization of HBCDs by APCI led to a 2–5 times higher S/N ratio compared with ESI. They also concluded that ESI was more sensitive to matrix components. In addition,
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
atmospheric pressure photoionization (APPI) was also compared with the ESI and APCI modes for HBCD analysis [14]. Although APPI can better resolve common problems in ESI and APCI such as matrix effects, the APPI signal intensities were approximately 10 times lower compared with ESI. Many researchers [15–17] have reported that matrix effects might adversely affect the reproducibility and quantification of target analyte when developing an LC–MS/MS method, especially when using ESI as the ionization mode. For HBCD diastereoisomer analysis, different studies have presented contrary results. Budakowski and Tomy [11] confirmed the existence of the “ion suppression phenomenon” when determining HBCD diastereoisomers in both biotic and sediment samples using LC–ESI–MS/MS. Dodder et al. [18] developed a baseline separation method for HBCD diastereoisomer determination in the biological tissues. Investigation of the matrix effect showed that potential matrix interference did not significantly influence the LC–MS/MS analyses of the diastereoisomers, whereas the response of the HBCD enantiomers in tissue samples were greatly influenced by matrix effects and other changes to the ionization conditions. As we know, HBCD can be analyzed using the same protocols as those reported in the PBDE study [12,19]. After finishing the PBDE measurements, the solvent was exchanged to methanol or acetonitrile for further HBCD analysis. Therefore, it is necessary to investigate whether the sample preparation and clean-up procedure can fulfill the determination requirements of diastereoisomers. In this study, a widely used clean-up procedure for soil and air samples was investigated for matrix effects by spiking 13 C-labelled HBCDs before injection. Also the optimization of MS/MS parameters for improving signal responses and the effects of mobile phase on the chromatographic separation and sensitivity are discussed in detail. The method developed in this study obtained completely baseline chromatographic separation and lower LODs. 2. Experimental 2.1. Chemicals and solvents Unlabelled and 13 C12 -labelled ␣-, -, ␥-HBCD were purchased from Cambridge Isotope Labs. (Andover, MA, USA) and used as received. The HPLC-grade methanol and acetonitrile were obtained from Merck (Darmstadt, Germany). Ammonium acetate and acetic acid were acquired from J.T. Baker (Phillipsburg, NJ, USA). Analytical grade hexane and methylene chloride were redistilled by a glass system. 2.2. Liquid chromatography An Agilent 1100 series HPLC system (Agilent Technologies, Palo Alto, CA, USA) equipped with a vacuum degasser, quaternary pump and autosampler was used. The separation was performed on a Zorbax SB-C18 reversed-phase column (250 mm × 4.6 mm I.D., 5 m, Agilent). The gradient mobile phase consists of methanol (A)/acetonitrile (B)/water with 10 mM ammonium acetate (C). The flow rate was set at 0.5 mL/min. The gradient program started at an initial composition of 80:10:10 A/B/C (v/v) and was ramped to 50:40:10 A/B/C in 18 min, followed by 30:70 A/B at 23 min, and was held for 7 min, then returned to 80:10:10 A/B/C in 8 min. The column was equilibrated for a further 6 min. The detailed program is listed in Table 1. 2.3. Mass spectrometry An Applied Biosystems-Sciex API 4000 (Applied Biosystems, Foster City, CA, USA) triple quadrupole mass spectrometry
75
Table 1 Gradient mobile phase program for the separation of ␣-, - and ␥-HBCD Time (min)
Methanol
Acetonitrile
Water (10 mM ammonium acetate)
0.0 18.0 23.0 30.0 37.0 43.0
80 50 30 30 80 80
10 40 70 70 10 10
10 10 0 0 10 10
equipped with a TurboIonSpray ionization interface was used. The Q1 scan range was m/z 630–660, operated with unit resolution for a scan time of 0.5 s. The quantities of the three diastereoisometric HBCDs were detected in electrospray ionization negative ion mode using multiple reaction monitoring (MRM) for [M−H]−1 → Br− (m/z 640.6 → 79 and 652.6 → 79 for native and 13 C-labelled HBCD, respectively), utilizing unit resolution and a 200 ms dwell time per transition. Details of the optimized MS/MS parameters are listed in Table 2. 2.4. Sample collection Four air samples were collected during 11–15 November 2006 from the Tianhe district, Guangzhou. The sampling procedure was described elsewhere [20,21]. Before sampling, glass fiber filters (GFFs, Whatman, Maidstone, UK) were baked at 450 ◦ C for 4 h to remove any organic contaminant, and polyurethane foam (PUF) plugs were Soxhlet extracted for 48 h with each of following solvents:methanol and 1:1 acetone–hexane. Air volumes of 800–860 m3 were drawn at 0.3–0.5 m3 /min for 32–35 h using a high-volume air sampler. After sampling, loaded GFFs were wrapped with prebaked aluminum foil and sealed with double layers of polyethylene bags, and PUFs were placed in solvent rinsed glass jars with aluminum foil-lined lids and stored at −20 ◦ C until extraction. Three surface soils were sampled from the Tianhe, Huangpu and Liwan districts in the Guangzhou urban areas in December 2006. Each of the surface soil samples was collected using a pre-cleaned stainless steel scoop, and mixed with 5 cores from 0 to 10 cm depth. The soil samples were kept at −18 ◦ C until analysis. Before extraction, soil samples were freeze-dried for 24 h and sieved through a 10-mesh stainless steel mesh. 2.5. Sample extraction and preparation The sample extraction and preparation procedures have been described elsewhere [20,21] and modified in this paper. In brief, soil and air samples (including PUF plugs and GFF glass fiber filters) were extracted with a mixture of acetone and hexane (1:1) for 72 h with a Soxhlet extractor. Activated copper granules were added into the extractor flask to remove elemental sulfur. After extraction, the concentrated extracts from the soil and air samples were cleaned Table 2 Optimised MS/MS parameters for the determination of ␣-, - and ␥-HBCD Parameter
Optimised value
Source temperature, TEM ( ◦ C) Ionization voltage (V) Ion source (GS1) setting Ion source (GS2) setting Curtain gas settings CAD gas setting Declustering potential (V) Entrance potential (V) Collision energy (V) Collision cell exit potential (V)
300 4500 50 55 10 25 −100 −8 −50 −6
76
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
and fractionated in a 10-mm I.D. silica/alumina column packed, from the bottom to top, with neutral alumina (6 cm, 3% deactivated), neutral silica gel (4 cm, 3% deactivated), silica with 50% conc. sulfuric acid (8 cm), and anhydrous sodium sulfate (1 cm). The PBDE and HBCD mixture was eluted with 30 mL of hexane and 60 mL of hexane:methylene chloride (1:1). The final extracts containing PBDEs of both soils and air samples were concentrated to 200 L in hexane for GC–MS analysis. Before HBCD analysis, the extracts were further cleaned using 1 g acidified silica with 50% conc. sulfuric acid and eluted with 10 mL of hexane:methylene chloride (1:1), which is helpful for decreasing matrix effects. For determination of HBCD diastereometers, the solvent was exchanged to methanol for LC–MS/MS analysis. 2.6. Evaluation of matrix effects The matrix effects were investigated following the procedures reported elsewhere [15]. In brief, MS/MS areas with known amounts of three 13 C-labelled diastereoisomer standards (A) were compared with those measured in soil and air extracts spiked, after extraction, with the same analyte amount (B). The ratio (B/A) is defined as the absolute matrix effect (ME). If the value of ME is 1, it means that there is no matrix effect. There is signal enhancement if ME >1 and signal suppression if ME <1. 2.7. Calibration curves Quantification was performed by an isotope-dilution method using 13 C12 -labelled HBCD as an internal standard, which is described elsewhere [22]. The calibration curve was derived with a series of standards ranging from 1 to 100 g/L and fixed concentrations (20 g/L) of the internal standards (r2 > 0.998). The response data were adjusted according to the relative ratios of the responses to the stable isotope internal standards in the sample and the standard. 3. Results and discussion 3.1. Optimization of chromatographic resolution and MS/MS parameters 3.1.1. Effects of mobile phase on the chromatographic separation and sensitivity For improving HBCD chromatographic separation, different columns (such as C8 and C18 ) and mobile phases were investigated in this study. The results indicated that the C18 column was the optimal one. Fig. 1a–c shows the chromatograms of HBCD diastereoisomer analysis using the C18 column and three mobile phase compositions. It was found that (1) the three diastereoisomers keep the same elution order when eluting by three mobile phases; (2) methanol and acetonitrile present different elution ability for the three individual stereoisomers, and -HBCD was retained in the column for a longer time when acetonitrile was replaced by methanol as the mobile phase. This result was also confirmed by Morris et al. [10] and Hoh and Hites [19]. The final gradient composition (Table 1) is composed of methanol, acetonitrile and water. As shown in Fig. 1c, the three diastereoisomers were adequately separated from each other. The results also suggest that the sufficient baseline chromatographic separation is benefit to reduce the matrix effect. Fig. 2 shows the effects of the mobile phase on sensitivity. Methanol and acetonitrile were chosen as the organic modifiers and the MS/MS was operated in a negative ESI using MRM for [M−H]− (m/z 640.6) → Br− (m/z 79.0 and 81.0). A series of mobile phase compositions were tested with the ratio of organic modifier to water in the range 10:0 to 6:4. The results indicated that: (1) the signal
Fig. 1. Chromatographic separation of three HBCD diastereoisomers by different mobile phases. (a) 9:1 methanol/water, (b) 9:1 acetonitrile/water and (c) gradient compositions described in Table 1.
responses of the three diastereoisomers follow the order  > ␥ > ␣, and (2) ␣- and -diastereoisomers have a stronger response when using methanol as the organic modifier in comparison with that using acetonitrile, whereas ␥-HBCD shows the reverse response. In order to enhance the signal response, mobile phase modifiers such as ammonium acetate and formic acid are also discussed. 10 mM ammonium acetate was found to provide the maximum response and was used in the following method development. 3.1.2. MS/MS optimization For quantification purposes, the MRM signal from the m/z 640.6 → 79 transition was used. The MS/MS optimization was performed through flow injection analysis (FIA) using the ␣-, -, ␥-diastereoisomers. In FIA, the flow rate was set at 0.5 mL/min and the mobile phase represented the composition present when the analyte was entering the source from the column. The compositions (methanol/acetonitrile/water) were 58/32/10, 55/35/10, and 51/39/10 for ␣-, -, ␥-HBCD, respectively. Each parameter set was run six times to ensure response stability. The final optimized MS/MS parameters are listed in Table 2 and are chosen
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
77
Fig. 3. Effects of different turbo-gas temperatures on the sensitivity of three diastereoisomers using MRM for [M−H]− (m/z 640.6) → Br− (m/z 79.0 and 81.0).
3.2. Matrix effects
Fig. 2. Effects of different mobile phases on the sensitivity of three diastereoisomers using MRM for [M−H]− (m/z 640.6) → Br− (m/z 79.0 and 81.0).
to optimize maximum signal response and stability. Compared with the results in previous papers, it was found that optimization of MS or MS/MS parameters on different instruments might be inconsistent. For example, the optimum temperature of the turbo gas was about 300 ◦ C (Fig. 3) in this study. It was shown in Fig. 3 that the three diastereoisomers showed different sensitivities when increasing the temperature from 150 to 450 ◦ C. -HBCD has the highest signal response when the temperature was increased to 400 ◦ C, whereas the maximum sensitivity of ␣- and ␥-HBCD was observed at 300 ◦ C. The compromise temperature was set at 300 ◦ C because most environmental research focuses on the determination of ␣- and ␥-HBCDs. Budakowski and Tomy [11] and Dodder et al. [18] reported that the optimal temperatures of the turbo gas were 500 and 450 ◦ C, respectively, whereas temperatures of 160 and 250 ◦ C have been used elsewhere [10,13].
When developing an LC–MS/MS method, the matrix effect is a major issue. When co-eluting compounds and analytes enter the ion source together, the ionization efficiency of the analyte might be affected, resulting in signal enhancement or suppression, which are named matrix effects. In this study, the matrix effects on HBCD determination were evaluated by spiking three individual 13 C-labelled stereoisomers into the air and soils matrix after completing sample preparation and clean-up. The detailed evaluation method of the matrix effects was described in Section 2.4. The three diastereoisomers were spiked with concentration levels of 5, 20 and 100 g/L, respectively, and the results are listed in Table 3. As Table 3 shown, ME values range from 0.95 to 0.99, clearly indicated that the extracted matrix components in the air and soil samples will not interfere with the accuracy of the HBCD diastereomer analysis in this study. This result might arise from (1) improved sample preparation and clean-up; or (2) improved chromatographic separation. The gradient mobile phase program not only allows clear baseline separation for all three diastereoisomers, but also sufficiently removed the majority of potential interferences in the sample extracts. It is also necessary to continue the gradient program after the HBCDs are eluted from the column in order to remove the apolar components in the sample matrix. The extent of matrix effects is dependent on the removal of interference in sample extracts and chromatographic separation. Budakowski and Tomy [11] confirmed the “ion suppression phenomenon” when measuring HBCD diastereoisomers for both biotic and soil samples using LC–ESI-MS/MS. This might be explained by simple sample preparation procedures and poor chromatographic Table 3 Matrix effects in air and soil extracts ME (matrix effect) 5 g/L
20 g/L
100 g/L
Soil ␣-HBCD -HBCD ␥-HBCD
0.93 0.95 0.95
0.98 0.98 0.99
0.99 1.00 0.99
Air–gas ␣-HBCD -HBCD ␥-HBCD
0.95 0.95 0.96
0.99 0.99 0.99
0.99 1.00 1.00
Air–particulate matter ␣-HBCD -HBCD ␥-HBCD
0.95 0.94 0.94
0.99 0.99 0.98
0.99 0.99 0.99
78
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
Fig. 4. LC–MS/MS chromatograms of three HBCD diastereomers in the technical mixture, air and soil samples.
separation between - and ␥-HBCD. Dodder et al. [18] developed a baseline separation method for HBCD diastereoisomer determination in the biological tissues. Investigation of the matrix effect showed that potential matrix interferences did not significantly influence the LC–MS/MS analyses of the diastereoisomers. 3.3. Method performance and its application to environmental samples For the validation of the method performance, within- and between-day reproducibility was assessed at two concentrations levels (10 and 50 g/L) in triplicate. The relative standard deviations range from 1.7 to 3.1% and 2.7 to 4.3%, respectively. The LOD on column, defined as a signal-to-noise (S/N) ratio of 3:1, was 0.5, 0.3 and 0.3 pg for ␣-HBCD, -HBCD and ␥-HBCD, respectively. Recovery experiments were performed at concentration levels of 5, 25 and 50 ng by spiking three 13 C-labelled diastereoisomers into air and soil samples. The mean recoveries were 84 ± 11%, 71 ± 7% and 87 ± 8% (n = 6) for ␣-HBCD, -HBCD and ␥-HBCD, respectively. Four air samples (including gas phase and particle phase) and three surface soil samples from Guangzhou city in South China were analyzed using the method developed here. The results are shown in Fig. 4. Levels of total HBCDs found in air samples were in the range of 1.2–1.8 pg/m3 , and they were detected in both the gas phase and particle phase. The mean values of total HBCDs were 1.4 pg/m3 . These data show comparable levels with the results reported in USA background air [19], whereas they are somewhat lower than the observations from background air in Sweden [23]. All the air samples in this study had similar ratios of the three diastereoisomers.
The mean percentages of ␣-HBCD, -HBCD, and ␥-HBCD were 58%, 15%, and 27%, respectively, which is different from commercial mixtures. This may be explained by leaching from the treated products, which has been experienced in high temperature processes [19]. The concentration levels of total HBCDs in soils from urban areas in Guangzhou city ranged from 1.7 to 5.6 ng/g dry weight. To date, few data have been reported on concentration levels in background sites, especially in soils. The only data found by the authors are from soil samples collected near point-sources in Sweden and Belgium/Germany [3], which contained high concentrations of HBCDs, ranging from 111 to 23,200 ng/g dry weight. However, our data reported here are comparable with those in non-point source sediments where concentration levels were less than 10 ng/g dry weight [3]. The ratios of the three diastereoisomers in three soils presented different patterns. Two samples were similar to the commercial product, with ␥-HBCD being the most abundant stereoisomer, whereas the other was dominated by ␣-HBCD. At present, it is still unclear whether this difference in the composition of the HBCD stereoisomers between some soils and technical HBCDs is caused by thermal isomerization during the processing of HBCDs or by stereoisomer-specific processes in the environment. More research is needed to improve our understanding of the source, distribution and fate of individual diastereomers. 4. Conclusion A high-sensitivity and baseline chromatographic separation method for analysis of HBCD diastereoisomers was developed. Chromatographic conditions (including mobile phase, column and
Z. Yu et al. / J. Chromatogr. A 1190 (2008) 74–79
modifier) and MS/MS parameters were optimized for a combination of sensitivity and stability. The LODs on column with this method were 0.5, 0.3 and 0.3 pg for ␣-HBCD, -HBCD and ␥-HBCD, respectively. Using the analytical method described in this study, matrix effects were evaluated and show that matrix interferences in soil and air samples did not significantly influence the accuracy and precision of the diastereomer analysis. It should be noted that there still is an important gap in our understanding regarding the determination of HBCDs in environmental matrixes. The optimization of MS or MS/MS parameters in the literature might vary between studies and matrix effects were discussed only in a few papers. A comparison of different methods and international intercalibration studies will be needed for improving the quality of analytical data and comparability of laboratory results. Acknowledgements This study was supported by National Natural Scientific Foundation of China (No. 40590392), and Chinese Academy of Sciences (No. KZCX2-YW-403). Partial funding was also provided by Natural Scientific Foundation of Guangdong Province (No. 5006280) and the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry to Z.Y. References ˚ Bergman, Environ. Int. 29 (2003) 683. ˝ [1] M. Alaee, P. Arias, A. Sjodin, A.
79
[2] U. Sellstrom, A. Kierkegaard, C. de Wit, B. Jansson, Environ. Toxicol. Chem. 17 (1998) 1065. [3] A. Covaci, A.C. Gerecke, R.J. Law, S. Voorspoels, M. Kohler, N.V. Heeb, H. Leslie, C.R. Allchin, J.D. Boer, Environ. Sci. Technol. 40 (2006) 3679. [4] Draft of the EU Risk Assessment Report on Hexabromocyclododecane, National Chemicals Inspectorate (KEMI), Sundbyberg, Sweden, 2005. [5] E. Eljarrat, A. De La Cal, C. Duran, D. Barcelo, Environ. Sci. Technol. 38 (2004) 2603. [6] F. Barontini, V. Cozzani, L. Petarca, Ind. Eng. Chem. Res. 40 (2001) 3270. [7] E.R. Larsen, E.L. Ecker, J. Fire Sci. 4 (1986) 261. [8] R.J. Law, M. Kohler, N.V. Heeb, A.C. Gerecke, P. Schmid, S. Voorspoels, A. Covaci, G. Becher, K. Janak, C. Thomsen, Environ. Sci. Technol. 39 (2005) 281A. [9] A. Covaci, S. Voorspoels, L. Ramosc, H. Neels, R. Blust, J. Chromatogr. A 1153 (2007) 145. [10] S. Morris, P. Bersuder, C.R. Allchin, B. Zegers, J.P. Boon, P.E.G. Leonards, J. de Boer, Trends Anal. Chem. 25 (2006) 343. [11] W. Budakowski, G. Tomy, Rapid Commun. Mass Spectrom. 17 (2003) 1399. [12] K. Janak, A. Covaci, S. Voorspoels, G. Becher, Environ. Sci. Technol. 39 (2005) 1987. [13] S. Suzuki, A. Hasegawa, Anal. Sci. 22 (2006) 469. [14] S. Backus, S. Batchelor, M. Alaee, D. Ueno, L.M. Hewitt, Organohalogen Compd. 67 (2005) 240. [15] J.C. Van De Steene, K.A. Mortier, W.E. Lambert, J. Chromatogr. A 1123 (2005) 71. [16] T. Benijts, R. Dams, W. Lambert, A. De Leenheer, J. Chromatogr. A 1029 (2004) 153. [17] H.B. Lee, K. Sarafin, T.E. Peart, J. Chromatogr. A 1148 (2007) 158. [18] N.G. Dodder, A.M. Peck, J.R. Kucklick, L.C. Sander, J. Chromatogr. A 1135 (2006) 36. [19] E. Hoh, R.A. Hites, Environ. Sci. Technol. 39 (2005) 7794. [20] B.X. Mai, S.J. Chen, X.J. Luo, L.G. Chen, Q.S. Yang, G.Y. Sheng, P.A. Peng, J.M. Fu, E.Y. Zeng, Environ. Sci. Technol. 39 (2005) 3521. [21] L.G. Chen, B.X. Mai, X.H. Bi, S.J. Chen, X.M. Wang, Y. Ran, X.J. Luo, G.Y. Sheng, J.M. Fu, E.Y. Zeng, Environ. Sci. Technol. 40 (2006) 1190. [22] S.G. Chu, C.D. Metcalfe, J. Chromatogr. A 1164 (2007) 212. [23] M. Remberger, J. Sternbeck, A. Palm, L. Kaj, K. Stromberg, Chemosphere 54 (2004) 9.