Analysis of halogenated flame retardants in Canadian wastewater treatment plants using gas chromatography–tandem mass spectrometry (GC-MS/MS)


 An investigation of brominated flame retardants (BFRs), polybrominated diphenylethers (PBDEs), and chlorinated flame retardants was conducted in samples from Canadian wastewater treatment plants (WWTPs). The penta-BDE mixture and deca-BDE were most commonly detected. Median removal efficiencies of PBDEs by WWTPs were 71% to 99% and higher concentrations of PBDEs were found in influent samples in summer, compared to winter. Three novel brominated flame retardants (nBFRs) including 1,2-bis(2,4,6-tribromophenoxy)ethane (BTBPE), 2-ethylhexyl-2,3,4,5-tetrabromobenzoate (EHTeBB or TBB), and bis(2-ethyl-1-hexyl)tetrabromophthalate (BEHTBP or TBPH) were detected in greater than 80% of influents and biosolids. Median removal efficiencies were 45% to 99% for BTBPE, 63% to 99% for TBB, and 88% to 99% for TBPH. Detection of total dechlorane plus ((DP) syn and anti) was greater than 90% in both wastewater and biosolids samples. However, median removal efficiencies were between 51% and 66% for total DP. Overall, PBDEs were effectively removed via different treatment processes, while BTBPE, TBB, and DP exhibited lower removal rates, which could be due to their physical chemical properties.


INTRODUCTION
int/TheConvention/ThePOPs/TheNewPOPs/tabid/2511/ Default.aspx). As additive FRs, PBDEs are not covalently bound and can be released into the environment during production, application, and waste disposal processes.
Although hexabromocyclododecane was at one point the second-most heavily produced BFR globally and is included in monitoring programs for a wide range of environmental compartments, most laboratories globally have transitioned to methods based on liquid chromatography/mass spectrometry (LC/MS) that enables diastereomer-specific analysis (Marvin et al. ). for dechlorane (i.e., Mirex) (Sverko et al. ). In the DP technical mixture, two stereoisomers (syn and anti) are present at an approximate ratio of 1:2, respectively. DP is used mainly in cable coatings, plastic roofing materials, and hard connectors in computers and televisions (Qiu et al. ).
The growing human population and the expansion of industrial activities generate large volumes of wastewater which, in developed countries, is regularly treated by wastewater treatment plants (WWTPs) prior to discharge to the environment (Rocha-Gutierrez & Lee , ).
Although WWTPs are effective at removing many contaminants, such as nutrients and biodegradable organics, from residential, industrial, commercial, and institutional wastewater, WWTPs are not designed to remove POPs such as PBDEs (Kim et al. a). Therefore, WWTP effluents and biosolids are considered to be pathways where organic pollutants, including PBDEs, are discharged to river ecosystems, to land and to lakes and oceans. Previous studies showed that lower concentrations of PBDEs were detected in treated effluent samples than in raw influents; this observation could be mainly due to the PBDEs' high partitioning affinity (octanol-water partition coefficient known as Log K ow , log K ow 5-10) to be adsorbed on suspended particles in wastewater (Song et al. ; Kim et al. a). Despite the fact that PBDEs appear to be effectively removed in the liquid stream of WWTPs, the overall fate and occurrence of PBDEs during wastewater treatment requires further investigation. WWTP performance can be affected by environmental and operational parameters, including season, treatment time, and process complexity; however, only a handful of studies have investigated the relationship between the removal efficiency for PBDEs and varying environmental and/or operational conditions (Kim et al. a, b, ).
Eight Canadian WWTPs, encompassing lagoon, primary, secondary, and advanced liquid treatment processes, were selected for the analysis of flame retardants in this study. An analytical method based on gas chromatography (GC) with tandem mass spectrometry (GC-MS/MS) was used for identification and quantification of 15 PBDEs and 16 alternative flame retardants in a total of 186 liquid and 58 solid samples, which were collected over a three-year period. In addition, the removal efficiencies of the various types of liquid treatment processes were investigated. This study also examined trends related to seasonal, annual, and input (industrial versus domestic) variations. To the best of our knowledge, this is the first study to report the concentration ranges of the alternative flame retardants TBB and TBPH in influent, effluent, and biosolids samples from WWTPs.

Sample information
All detailed WWTP characteristics and sampling pro-

Sample pre-treatment, wastewater
Each wastewater sample (approximately 0.6 L and 1.6 L of sample for influent and effluent, respectively) was spiked with 100 ng of BDE-71 as a recovery standard or surrogate, and extracted sequentially with dichloromethane (DCM, 100 mL, 50 mL, and 50 mL) by liquid-liquid extraction. Following each extraction, the organic layers were dried over anhydrous Na 2 SO 4 (granular certified ACS, fired in a 600 C muffle furnace overnight) and combined. If emulsion was present, the emulsion layer was removed and centrifuged for 10 min at 2,500 rpm and 24 C to isolate the organic layer, which was subsequently added to the existing extracts. The combined extract was solvent-exchanged to iso-octane and concentrated to approximately 1 mL using a rotary evaporator followed by evaporation with nitrogen. One laboratory blank (deionized water, no standards spiked) and spiked blank (target standards spiked in influent sample) was included for every six-sample batch. Individual samples, laboratory blanks and spiked blanks were spiked with BDE-71. Laboratory blanks reflected extraneous FRs present in the laboratory environment while spiked blanks accounted for matrix effects. It is acknowledged that labeled surrogate standards are now available for many of the analytes targeted in our schema, which is an approach we fully advocate; however, use of BDE-71 was standard practice during the period of time we began our method development.

Sample pre-treatment, biosolids
Approximately 1 g of biosolids sample was ground with 15 g of anhydrous Na 2 SO 4 to a homogenous mixture.
This solid mixture was transferred to a centrifuge tube, spiked with 100 ng of BDE-71 as a recovery standard, and then 25 mL of hexane:acetone (1:1) was added as the extraction solvent. The mixture was vortexed for 30 min at 2,500 rpm then centrifuged for 10 min at 2,000 rpm at 24 C. The supernatant was collected and the remaining solid mixture was extracted twice more using the same vortex/centrifuge procedure. The combined extract was solvent-exchanged to iso-octane and reduced in volume to about 1 mL using a rotary evaporator followed by evaporation with nitrogen.

Column clean-up
Liquid and solid extracts were further pre-treated using an open column cleanup. A chromatography column was prepared containing 10 g of 5% deactivated silica gel (dried at 160 C overnight prior to the deactivation), which was rinsed and saturated with hexane and topped with 2 g of anhydrous Na 2 SO 4 . The concentrated extract was loaded onto the column and eluted with an additional 45 mL of a 1:1 mixture of hexane:ethyl acetate. The extract was solvent exchanged into iso-octane and concentrated to 1 mL using a rotary evaporator followed by evaporation with nitrogen; the extracts were analyzed using GC with tandem mass spectrometry (MS/MS).

Instrumental parameters
All sample extracts were analyzed using an Agilent 7890A  Table 2.
Quality assurance and quality control Three main quality control criteria were used to confirm the correct identification of the target flame retardants: (i) the GC retention times of FRs must be matched within ±0.1 min of standards, (ii) the ratio between the quantifier (primary) and qualifier (secondary) ions must be   in influent and from not detected to 34 ng L À1 in effluent   -47, 99, 100, 153, 154, and 209, in the biosolids were 90%, 100%, 100%, 84%, 83%, and 100%, respectively as shown in Table 3. The total concentrations of major PBDE congeners in biosolids samples were in the range of 70 to 6,390 ng g À1 dry weight (dw), falling in the range previously reported for North America as shown in

).
In addition, deca-BDE (BDE 209) was the predominant congener found in the wastewater and biosolids samples as shown in Table 6. To the best of our knowledge, this study is the first to report concentrations of TBB and TBPH in influent, effluent, and biosolids samples, as well as their removal efficiency via various types of treatments in Canadian WWTPs. The concentrations of TBB ranged from 11 to 477 ng L À1 , nd to 29 ng L À1 , and 5 to 1,227 ng g À1 dw, while the concentrations of TBPH ranged from 0.4 to 326 ng L À1 , nd to 44 ng L À1 , and 56 to 1,820 ng g À1 dw in influent, effluent, and biosolids samples, respectively, as shown in Table 7. Greater than 96% median removal efficiency of TBB was observed for AL, FL, ST, and AT, while the PT rendered a median of 63% removal, presented in Figure 2. In contrast, a median of over 88% removal efficiency of TBPH was   Peng et al. ). In this study, the concentrations of BTBPE ranged from 0.6 to 25 ng L À1 , nd to 7 ng L À1 , and 2-54 ng g À1 dw in influent, effluent, and biosolids samples, respectively (Table 7). It was recently reported that the concentrations of BTBPE found in biosolids samples from 20 different Canadian WWTPs ranged from 0.005 to 140 ng g À1 , which was approximately three times higher than the concentrations found in this study. In contrast, the concentrations of BTBPE found in surface sediment from Lake Ontario (2.7 to 3.8 ng g À1 ) were about an order of magnitude lower than BTBPE concentrations found in biosolids from this study. Furthermore, it was reported that median removal efficiencies of BTBPE across 20 Canadian WWTPs ranged from 61% to 97%, whereas in this study a greater than 85% median removal efficiency of BTPBE was observed in AL, FL, ST, and AT, but the PT removed a median of only 45% of BTPBE as shown in Figure 2.

Dechlorane plus (DP)
Two stereoisomers, syn and anti, are present at an approximate ratio of 1:2, respectively, in the dechlorane plus (DP) technical mixture (De la Torre et al. ). It was previously shown that the total DP concentration in suspended sediment samples collected from the Niagara River had been declining, from 89 ng g À1 to 7.0 ng g À1 , for over two decades biosolids samples were stored in iso-octane at À20 C. The percent detection of total DP was greater than 90% in both wastewater and biosolids samples, ranged from 17 to 247 ng L À1 , 2 to 139 ng L À1 , and 96 to 740 ng g À1 dw in influent, effluent, and biosolids, respectively. Similar to PBDEs, DP also has a relatively high affinity for lipids and solid particles due to its high molecular weight and high log Kow (9.3). Unlike the PBDEs, DP median removal efficiencies were lower and ranged from 51% to 70% for AL, FL, ST, PT, and AT treatment processes as shown in Figure 2.
In addition, higher concentrations of total DP were found in influent samples collected in winter relative to summer, whereas no statistically significant correlations were found between total DP concentrations and annual or input variations (p < 0.05, data not shown). Syn-DP not being detected in any of the wastewater or biosolids samples could be due to conversion of syn-DP to anti-DP under extended storage conditions. This hypothesis was tested by analyzing pre-treated influent samples that were only stored for a few days at À20 C, and syn-DP was detected in those samples. However, the conversion of syn-DP to anti-DP under extended storage conditions and low removal efficiency of DP within different liquid treatment processes in WWTPs will be investigated further. All WWTPs reduced levels of TBB and BTBPE in wastewater; however, the PT process had relatively low removal efficiency for these compounds. In addition, WWTPs did not effectively remove high levels of the DP found in influent and biosolids samples; only 66% median removal of DP was observed in AT. Despite the effective removal of PBDEs in WWTPs, some nFRs, including BTBPE, TBB, and DP, have lower removal rate in WWTPs, which could be due to their physical/chemical properties. Further study is required to investigate conditions that would improve the removal efficiency of these nFRs both in biosolids and in effluent samples prior to discharge to the environment.