- Email: [email protected]
Discrimination of hexabromocyclododecane from new polymeric brominated flame retardant in polystyrene foam by nuclear magnetic resonance
Chemosphere 144 (2016) 1391e1397
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Discrimination of hexabromocyclododecane from new polymeric brominated flame retardant in polystyrene foam by nuclear magnetic resonance bastien Schweizer b, Yavor Nikolaev Mitrev a, c, Damien Jeannerat a, *, Marion Pupier a, Se b b Philippe Favreau , Marcel Kohler a b c
D epartement de chimie organique, Universit e de Gen eve, 30 quai Ernest Ansermet, 1211 Gen eve 4, Switzerland ^ti, 23 avenue de Sainte Clotilde, Case postale 78, 1211 Gen Service de Toxicologie de l'Environnement Ba eve 8, Switzerland Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Acad. G. Bonchev str., bl 9, 1113 Sofia, Bulgaria
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
During a 2013/2014 survey, the HBCDD content in polystyrene ranged from 0.2 to 2.4% by weight in a total of 98 samples. A trend towards the use of copolymerized brominated flame retardants in polystyrene foam was observed between 2013 and 2014. Analyses indicated that expandable and extruded polystyrene foams significantly differed in the a/g HBCDD isomer ratio. Investigation by NMR identified polystyrene samples containing HBCDD or a brominated butadiene styrene (BBS) as copolymer. NMR constitutes a simple method to discriminate flame-retardants in insulating polystyrene foam materials.
a r t i c l e i n f o Article history: Received 27 July 2015 Received in revised form 2 October 2015 Accepted 3 October 2015 Available online xxx
a b s t r a c t Hexabromocyclododecane (HBCDD) is a brominated flame retardant (BFR) and major additive to polystyrene foam thermal insulation that has recently been listed as a persistent organic pollutant by the Stockholm Convention. During a 2013/2014 field analytical survey, we measured HBCDD content ranging from 0.2 to 2.4% by weight in 98 polystyrene samples. Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) analyses indicated that expandable (EPS) and extruded (XPS) polystyrene foams significantly differed in the a/g HBCDD isomer ratio, with a majority of a and g isomers in XPS and EPS, respectively. Interestingly, this technique indicated that some recent materials did not contain HBCDD, but demonstrated bromine content when analysed with X-ray fluorescence (XRF). Further
* Corresponding author. E-mail address: [email protected] (D. Jeannerat). http://dx.doi.org/10.1016/j.chemosphere.2015.10.021 0045-6535/© 2015 Elsevier Ltd. All rights reserved.
1392
Keywords: HBCDD Brominated butadiene styrene Polystyrene insulation foam XRF LC-MS/MS NMR
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
investigation by Nuclear Magnetic Resonance (NMR) was able to discriminate between the BFRs present. In addition to confirming the absence or presence of HBCDD in polystyrene samples, high-field NMR spectroscopy provided evidence of the use of brominated butadiene styrene (BBS) as copolymer in the production of polystyrene. Use of this alternative flame retardant is expected to cause fewer health and environmental concerns. Our results highlight a trend towards the use of copolymerized BFRs as an alternative to HBCDD in polystyrene foam boards. In addition to providing a rapid NMR method to identify polymeric BFR, our analytical approach is a simple method to discriminate between flameretardants in polystyrene foam insulating materials. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction The cycloaliphatic hexabromocyclododecane (HBCDD) is a common brominated flame retardant (BFR) with applications in the polymer and textile industries (Fig. 1) (Alaee et al., 2003). One of its main uses is in thermal insulation for buildings and constructions, where it is added to expanded (EPS) and extruded (XPS) polystyrene foam to reduce their flammability. The increasing use of flame retardant in EPS and XPS has been mainly driven by national regulations, construction safety rules and insurance policies (Lassen et al., 2011), with the aim of significantly reducing large building fires. In the early 2000s, the European market for EPS was approximately 470,000 tons, with increasing demand in Eastern Europe (ECHA, 2008). Recently, the total world market for EPS was estimated at 1.8 million tons per year (Lassen et al., 2011). The use of BFRs has led to some serious concerns about their toxicological and environmental impact (De Wit, 2002; Darnerud, 2003; De Boer, 2004; Covaci et al., 2006; Messer, 2010; Eljarrat and Barcelo, 2011). Environmental data unambiguously demonstrate persistence of HBCDD in various environmental compartments, as well as bioaccumulation in numerous living organisms (Law et al., 2008; Tanabe et al., 2008). Repeat dose studies in animals have shown that HBCDD has significant effects on the liver and thyroid gland (Van der Ven et al., 2006; ECHA, 2008) and on reproductive toxicity (Ema et al., 2008). In addition, HBCDD has been identified as a disruptor of endocrine function, with long-term exposure to low doses shown to have profound effects on hormonal pathways and neurodevelopment (Legler, 2008). In view of its environmental and toxicological impact, HBCDD was recently added to the list of persistent organic pollutants in annex A of the Stockholm Convention. Its use is therefore expected to decline, with a complete commercial ban in mid-2015. Recycling of HBCDD-containing materials will be prohibited in order to avoid uncontrolled dissemination in other materials. However, there is concern that landfills with construction waste containing EPS or XPS materials constitute emission sources of HBCDD to soil, air and water (Remberger et al., 2004). Over the past three decades EPS and
XPS materials have been widely installed in buildings where their presence is raising concerns over disposal during renovation and demolition work. With the exception of information from commercial sources, there is very little published data in the scientific literature on the HBCDD content of insulation materials. In view of these concerns, an efficient methodology is needed to detect and quantify HBCDD in existing and replacement building products. The primary objective of this study was therefore to evaluate the total bromine content of samples using a hand-held Xray fluorescence instrument in EPS and XPS materials, then to submit samples to liquid chromatography and tandem mass spectrometry (LC-MS/MS) analyses to enable precise quantification and profiling of the three major HBCDD isomers. In addition, we used Nuclear Magnetic Resonance (NMR) to further investigate the significant bromine presence, but lack of HBCDD, in some recently collected samples. 2. Experimental 2.1. Chemicals Acetonitrile, methanol and toluene were all of HPLC-grade (Merck Millipore, Darmstadt, Germany) and milliQ water was used throughout the analyses. a-, b-, g- and 13C-g-HBCDD analytical standards were purchased from Cambridge Isotope Laboratories (CIL, USA) at 50 mg/mL in toluene. 2.2. Samples Samples were collected from construction sites in the Geneva region of Switzerland during April 2013 (n ¼ 53) and April 2014 (n ¼ 45). In total, 45 collection sites were sampled, including 30 residential, 14 public (commercial, medical), and one industrial location. Samples were taken from brand new expanded (EPS) and extruded (XPS) polystyrene foams being installed on new building constructions for a broad range of thermal insulation applications (roof, floors, walls, foundations). The material type was identified at the field site by the product name and by visual inspection. Samples originated from various manufacturers, distributors and importers, and corresponded to 50 different trade names. No replicate samples were collected at any one single site in order to avoid sampling boards from the same production batch. However, some board references were found in duplicates or replicates at different collection sites. In these cases, all samples were kept to reflect the true situation encountered in the field. 2.3. X-ray fluorescence analysis
Fig. 1. Molecular structure of HBCDD (left), and a new BFR (right) with bromo butadiene styrene (BBS) blocks inserted in the polymer.
A 40 keV tube-emitting Niton Xlt 700 Series, model Xlt792W (Thermo Fisher Scientific Inc., USA) was used for all XRF analyses. The Niton XLt is a small, field-portable instrument designed for the
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
chemical characterization of soils, sediment, and other thick homogeneous samples, such as plastics and metals. Use of the instrument for very low-density materials (polystyrene) may be considered a specific application. Acquisitions were performed using the ‘Plastic analysis’ mode with reading times (typically 10e20 s acquisition) in duplicate measurements. The instrument was factory-calibrated with polyethylene samples containing 1.04 and 0.1% bromine. An instrument calibration check was carried out prior to each round of analyses using three multi-element reference samples distributed by TechLab (Metz, France) and provided by Modern Analytical Techniques, LLC (Hillsborough, NJ, USA). The three bromine-containing reference materials were in polyvinyl chloride. ‘PVC-01A’, ‘PVC-L-20A’ and ‘PVC-H-19A’ contained 0, 497 and 1094 ppm of bromine respectively, with an overall error of 4% relative to a 95% confidence level, except below 100 ppm, where the error was estimated to be 5 ppm at 95% confidence level (Supplementary Data 3, Table S1). The polyvinyl chloride reference samples were used to qualify the XRF instrument for measuring bromine content under standard conditions. 2.4. Sample preparation Raw samples were analysed directly by XRF without any specific treatment. For each sample, the material thickness under X-ray beam exposure was measured precisely. Prior to LC-MS/MS analysis, 20 mg of sample was collected from the inner part of the material in order to avoid any external contamination, and directly dissolved in 20 mL of toluene. After thorough mixing to obtain complete dissolution (approx. 10 sec), the solution was filtered using a 0.45 mm nylon disposable filter from Merck Millipore (Darmstadt, Germany). The 13C-g-HBCDD internal standard (50 mL at 1 mg/mL) was then added to the filtrate (10 mL) and the mixture was completed with 940 mL of methanol. This solution was injected directly into the LC-MS/MS apparatus. 2.5. LC-MS/MS analysis An Agilent 1200 series HPLC instrument equipped with a degasser, autosampler and oven was used together with a Synergy polar RP column (50 2 mm, 2.5 mm, 100 Å). The oven temperature was set to 25 C. A constant flow of 250 mL/min was applied using a gradient of water (solvent A) and methanol (solvent B). Initial conditions were 50/50 (A:B) and were maintained for 0.3 min after injection of 5 mL of sample. At 3.2 min, a linear gradient of 50e90% B was applied and left for 3.5 min. Then a final step was made to 100% B for 0.1 min and left for 5.9 min before returning to the initial conditions. Mass spectrometry analyses were performed using a QTRAP 3200 from Applied Biosystems (ABSciex, USA) in multi-reaction monitoring (MRM) and electrospray negative ionization modes. Instrumental parameters such as declustering potential, entrance potential, and collision energy were all optimized using the ‘compound optimisation’ mode. Data acquisition and treatment were performed using the Analyst software (version 1.6.1) from Applied Biosystems. Quantification was based on the transitions 640.6/78.9 and 652.6/78.9 for HBCDD isomers and the internal standard, respectively, while the transitions 640.6/80.9 and 652.6/80.9 were used as qualifiers for HBCDD isomers and the internal standard, respectively. Calibration curves covering the quantification range 1e1000 ng/mL were generated for each analyte using 13C-g-HBCDD as an internal standard. Accuracy and relative standard deviation (RSD) were determined by performing addition of HBCDD analytical standards to HBCDD-free EPS material at three concentrations (low: 5 ng/mL, medium: 50 ng/mL and high: 500 ng/mL).
1393
Reproducibility was tested by performing six sample preparations and analyses with one EPS and one XPS sample. Relative recovery (accuracy) was calculated as the ratio (%) between the experimental and the theoretical values. Relative standard deviation was calculated as the standard deviation of all values divided by the mean value. Reproducibility was estimated as the relative standard deviation of all values divided by the mean value. 2.6. NMR experiments NMR experiments were carried out using a Bruker Avance III 500 MHz spectrometer with a 5-mm DCH 13Ce1H/D helium-cooled cryogenic probe equipped with a z-gradient coil. Analyses were performed on fresh samples prepared by dissolving 75 mg (c.a. 2e3 cm3) of raw foam in 750 mL of CDCl3. Samples were filtered using a 0.45 mm nylon disposable filter (EXAPURE SY13TF, Alys Technologies), and 600 mL of the filtrate were then transferred into 5 mm NMR tubes. The measurement temperature was 25 C for all analyses, except those aimed at identifying g-HBCDD (Supplementary Data 1, Fig. S1), which were recorded at 10 C. 1D 1 H spectra were recorded with four scans, a relaxation delay of 2 s, and a total acquisition time of 30 s 1D T2-filtered experiments (Bruker pulse program “cpmg1d”) were acquired with 32 scans and a 205 ms CPMG (Car-Purcell-Meiboom-Gill) period with an interpulse delay of 400 ms. 2D HSQC experiments (Bruker pulse program “hsqcetgpsisp2.2”) were recorded over 2.5 h with a 10 or 15 ppm width in the 1H dimension and a 200 ppm spectral width in the 13C dimension (Supplementary Data 1, Fig. S1). Each of the 512 increments in t1 was acquired with 8 scans, a 1.7 s recycle delay between scans and 4096 points in the direct dimension. Multiplicity-edited experiments (Supplementary Data 1, Fig. S1) were performed using the “hsqcedetgp” Bruker pulse program (Willker et al., 1993). Each of the 256 t1 increments was recorded with 16 scans and 4096 points in the direct dimension. The T2edited HSQC included a CPMG element (see above for details) inserted after the first pulse of the HSQC sequence. Spectra were processed using the standard Fourier transform after application of a squared, shifted sine-bell window function in both dimensions. 3. Results and discussion 3.1. Measurement of the bromine content of EPS and XPS using XRF X-ray fluorescence spectra acquired from the various insulation foam materials typically showed two intense peaks at 11.9 and 13.3 keV that were both characteristic of the elemental bromine emission rays Ka and Kb, respectively (Supplementary Data 2, Fig. S2). We detected no interfering peak or other major emission ray specific to a particular element. Results showed a relative standard deviation of less than 10% for all samples (Supplementary Data 3, Table S2); it should be noted that most EPS and XPS measurements were in the 300 to 20,000 ppm range, therefore higher than the highest calibration sample of 1094 ppm. The XRF instrument detected the presence of bromine in all samples collected in 2013 and 2014, with values ranging from 0.2 to 1.5% atomic bromine (Fig. 2). XFR only indicates the presence of elemental bromine, and cannot distinguish HBCDD from other brominecontaining compounds. Therefore, we subsequently analysed all samples by LC-MS/MS and NMR in order to characterize bromine compounds present in more detail. 3.2. XRF quantification versus LC-MS/MS data We devised the LC-MS/MS quantification method to quickly and
1394
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
clearly separate the three HBCDD isomers in the ng/mL range. First, we validated the instrument and methodology for precision, accuracy and reproducibility. The methodological validation was performed on BFR-free polystyrene material that was spiked with HBCDD in the range 1e1000 ng/mL (Table 1 and Supplementary data 4, Table S3). We devised a quick and simple preparation method based on the favourable dissolution properties of polystyrene in toluene. Filtration, dilution in methanol and addition of the internal standard were carried out prior to sample injection. Typical extracted ion chromatograms for the three HBCDD isomers are shown in Fig. 3. The a-HBCDD eluted first, followed by the g- and b-HBCDD isomers. Surprisingly, the isomer elution profile was different from that observed in previous studies with different reverse-phase columns and gradients (Heeb et al., 2010a). However, the injection of each individual isomer confirmed the observed profile under our analytical conditions. One should thus be careful when changing experimental parameters that could affect the elution order of the HBCDD isomers, thus potentially giving rise to errors in peak identification. Overall, we observed a good correlation between XRF and LC-MS/MS data (Fig. 2). The bromine content of EPS samples was sometimes underestimated by XRF because the measurements were dependent on the sample thickness (Supplementary Data 5, Fig. S3). We obtained the best match with LC-MS/MS data using a minimum thickness of 12 cm. In addition to HBCDD, two new families of polybrominated compounds have recently been found in HBCDD-treated polystyrene materials (Heeb et al., 2010a, 2012); these are likely byproducts of HBCDD chemical synthesis. Isobutoxypentabro mocyclododecanes (iBPBCDs) and pentabromocyclododecanols (PBCDOHs) were previously detected in EPS and XPS insulation foam boards as well as in HBCDD technical mixtures; however, quantification of these compounds was prevented by the unavailability of pure standards, and these were not included in the targeted LC-MS/MS analytical method. When HBCDD was detected, the total bromine content correlated well with HBCDD quantification, suggesting that iBPBCDs and PBCDOHs did not represent a significant portion of the total brominated content of EPS and XPS. 3.3. HBCDD quantification in polystyrene insulation foams All samples were analysed by LC-MS/MS to quantify the three
1.60% 1.40% 1.20% 1.00%
0.80% 0.60% 0.40% 0.20% 0.00% 0.00%
0.50%
1.00%
1.50%
2.00%
Fig. 2. Correlation of bromine content based on HBCDD quantified by LC-MS/MS and XRF analysis. Note that the lower limit of quantification of LC-MS/MS was 0.03%; lower limit of detection for XRF measurements was estimated to be less than 50 ppm (0.005%).
Table 1 Validation of relative recovery (accuracy) and precision from measurements of HBCDD and its a, b and g-isomers using LC-MS/MS. Sample preparation
a-HBCD
QC low, 5 ng/mL, (n ¼ 3) Relative recovery (%) 100.9 RSD (%) 6.6 QC medium, 50 ng/mL, (n ¼ 3) Relative recovery (%) 102 RSD (%) 5.5 QC high, 500 ng/mL, (n ¼ 3) Relative recovery (%) 100.2 RSD (%) 10.6
b-HBCD
g-HBCD
Total HBCD
102 1.2
109.9 5.1
103.6 4.0
104.2 5.7
102.8 7.7
102.4 2.0
103 7.9
102.3 4.0
101.1 1.8
HBCDD isomers. We were therefore able to precisely measure the HBCDD content of EPS and XPS based on the detection of the major a-, b- and g-isomers (Supplementary data 6, Table S4). Using LCMS/MS, we detected HBCDD in all samples collected in 2013, but in only 70% of those collected in 2014. EPS materials had a mean HBCDD total content of 0.67%, with values ranging from 0.25 to 2.4%; for XPS insulation foams, values ranged from 0.2 to 2.29%, with a mean of 1.1% (Table 2). We found the mean HBCDD content of the EPS boards we sampled was slightly higher than the usual value (0.5%) reported by the industry (Morose, 2006). The maximum concentration of HBCDD in EPS beads was previously reported to be 0.7% (ECHA, 2008), which is much lower than the maximum of the 2.4% reported here. The overall range observed in the present study (0.2e2.4%) is also wider than the previously reported range of 0.5e1% (Lassen et al., 2011). This difference may be attributed to the lack of previous representative and robust data, and could also reflect variations in HBCDD used in polystyrene beads at different production sites. We also observed a significant variation in total HBCDD content of foam insulation boards displaying an identical trade name and reference, which probably arises from different production sites or batches of origin (Supplementary Data 6, Table S5). The mean HBCDD content of XPS was higher than for EPS, with a similar range of variation. Our observations therefore do not entirely support previously published data on the HBCDD content range of EPS and XPS materials (0.5e1% for EPS and 1e2% for XPS). In contrast, we found that both materials contained a similar mean amount of HBCDD (0.7 and 1.1 for EPS and XPS, respectively), with high variation in the amounts of both materials 0.2e2.4%. In addition to the variation possibly due to the diversity in production batches or sites, the use of synergists (i.e. dicumyl peroxide), which were not investigated in this study, could affect the HBCDD content. Such additives would enhance the flame retardant activity, allowing for varying amounts of HBCDD to be used to obtain an equivalent efficacy. The relative ratio of HBCDD isomers in our samples was tightly linked to the type of polystyrene foam in all cases. It has been reported that the usual HBCDD technical mixture contains mainly the g-HBCDD isomer (Heeb et al., 2005). We found that the relative presence of this isomer in EPS appeared to be in line with the technical HBCDD mixture used during polystyrene bead formation. But for XPS, the major isomer present was a-HBCDD, which is probably due to the thermal transformation of g-to a-HBCDD (Heeb et al., 2010b). Indeed, the industrial manufacturing of XPS uses a higher temperature (200 C) than EPS (120 C) during the foaming process. We found the mean proportion of g-HBCDD in EPS was 75%, which corresponded to the proportion of a-HBCDD content in XPS. The proportion of the least abundant b-HBCDD isomer in both EPS and XPS was always in the range of 1e19%. Interestingly, 30% of the 45 samples collected in 2014 did not show a correlation between the bromine content measured by XRF
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
1395
Fig. 3. LC-MS/MS ion chromatogram traces of a-, g- and b-HBCD (respectively at 7.7, 8.4 and 8.8 min) in EPS and XPS insulation foam materials.
Table 2 HBCDD content and isomer ratios of EPS and XPS samples. HBCDD content (%)
EPS samples (n ¼ 32) Min. value Max. value Average RSD (%) XPS samples (n ¼ 54) Min. value Max. value Average RSD (%)
HBCDD isomer relative ratio/total HBCDD (%)
a-HBCDD
b-HBCDD
g-HBCDD
Total HBCDD
a-HBCDD
b-HBCDD
g-HBCDD
0.02 0.45 0.10 88
0.00 0.33 0.06 104
0.17 1.88 0.51 79
0.25 2.40 0.67 73
5 41 16 47
1 19 9 57
40 93 75 17
0.15 1.71 0.83 44
0.03 0.38 0.17 47
0.02 0.24 0.11 48
0.20 2.29 1.11 44
64 83 75 4
11 19 15 10
6 19 10 21
and the HBCDD content measured by LC-MS/MS (Fig. 2), demonstrating that the bromine content was not originating from HBCDD. Using LC-MS/MS, we did not detect HBCDD in around 60% of EPS samples (12 out of 21). We therefore performed investigations using NMR in order to elucidate the nature of the brominated material present in EPS foam boards.
HeCeBr signals in the 3.5e4.5/47e60 1H/13C ppm (Arsenault et al., 2007) region; see Fig. 4a and a more complete set of spectra in the Supplementary Data 1). The methylene signals of HBCDDs (see the dotted box in Fig. 4b) were located in a crowded region of the spectrum near the very intense polymer signals from CH2-Ar groups (1.5e2.5 ppm region), making them much less attractive for reliable identification of HBCDD.
3.4. Discrimination of HBCDD-containing polystyrene by NMR The three most abundant HBCDD isomers (Supplementary data 7, Fig. S4 for the full structures) showed well-characterised fluxional behaviour (Arsenault et al., 2007) enlarging NMR signals. The a-isomer was the least affected and presented resolved signals in HSQC spectra recorded at 25 C. As predicted from the dynamic properties of HBCDDs, decreasing the temperature to 10 C reduced the dynamic of the g-isomer sufficiently to present resolved signals in the HSQC spectrum (Supplementary data 1 and 8). To observe signals of the least-abundant b-isomer, it would be necessary to further reduce the temperature; however, these would still be difficult to observe due to the absence of a C2 symmetry element increasing the sensitivity to the two major isomers (Supplementary data 7, Fig. S4). As the three stereoisomers of HBCDD are always observed together, we discriminated HBCDDcontaining samples at room temperature based on the fingerprint signals of the a-isomer (three characteristic and well resolved
3.5. Characterisation of a polymeric brominated polystyrene by NMR In an attempt to confirm if HBCDDs have been replaced by new BFR structures in recent polystyrene materials, we combined standard HSQC and multiplicity-edited HSQC experiments (Willker et al., 1993) on HBCDD-free samples (Fig. 4b and Supplementary Data 1). The spectra showed broad signals at 3.86e4.02/33.0 1 H/13C ppm, 4.4/55.4 ppm and 4.6/56.4 ppm that were in perfect accordance with the expected position of brominated carbons in a new BFR (Khoee and Sorkhi, 2007) alternative to HBCDD (Fig. 1) (USEPA, 2014). These signals probably correspond to the eCH2eBr and >CHeBr of the polymer (the type of carbon was confirmed with edited HSQC). This assignment was further supported by T2 relaxation-edited experiments (Tang et al., 2004) (Supplementary data 8 and 9) that confirmed that these signals had short T2 relaxation times, which would be expected if these brominated
1396
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
Fig. 4. Typical HSQC spectra of polystyrene dissolved in CDCl3. (a) The three characteristic CH signals of a-HBCD indicate the presence of prohibited-HBCDD flame retardants. The dotted square indicates the location of the methylene signals of a- and g-HBCDD. (b) In samples containing no HBCDD, broad signals were observed (dotted circles), which probably correspond to eCH2eBr and two types of HeCeBr (possibly corresponding to backbone and side-chain locations) on a bromine-containing polymer (see Supplementary data 1 for more details).
carbons were indeed part of the polymer. Such T2-edited experiments improve the quality of the spectra of small molecules, as the free HBCDD present in samples containing large amount of slowrelaxing signals, such as those of the polymer. However, we found that ‘standard’ HSQC experiments were sufficient for clear discrimination, provided that high-sensitivity probes (helium or liquid nitrogen cooled) were used in order to obtain clear 2D spectra within a reasonable acquisition time (Supplementary data 10). We analysed the three main types of polystyrene foams among collected samples using NMR, and found EPS and XPS foams containing HBCDD, and EPS foams containing bromo butadiene styrene copolymer (Supplementary data 1).
4. Conclusion This study provides precise quantitative data on HBCDD incorporated in expanded and extruded polystyrene foams used in building thermal insulation. Our results confirm that the use of these construction materials over the past decades now constitutes a huge reservoir of a toxic, bioaccumulative and persistent organic pollutant. The waste generated by demolition and renovation works will undoubtedly result in long-term environmental dissemination of HBCDD (Morf et al., 2008; Managaki et al., 2009) if this pollution source is not controlled by an appropriate waste management strategy (UNEP, 2011). Standard 2D HSQC NMR spectra clearly distinguished
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
polystyrene foams containing conventional HBCDD from new foams containing a bromo butadiene styrene copolymer, providing a qualitative method to discriminate BFRs in polystyrene. This technique will enable the evaluation of BFRs in materials and in the environment, and the monitoring of the human and environmental fate and impact of these substances, contributing to the global effort to reduce the environmental release of an important worldwide pollutant. Acknowledgements We would like to thank Jean-Luc Bailly for his help with field sampling, Natacha Duran for her dedication to experimental work, and Ron Hogg of OmniScience SA. for carefully reading and correcting the manuscript. We also acknowledge Marc Dupayrat for his expertise in XRF technology. The authors thank the State of Geneva and the Swiss NSF (grant no. 200021_147069 and 206021_128746) for funding. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2015.10.021. References Alaee, M., Arias, P., Sjodin, A., Bergman, A., 2003. An overview of commercially used brominated flame retardants, their applications, their use patterns in different countries/regions and possible modes of release. Environ. Int. 29, 683e689. Arsenault, G., Chittim, B., McAlees, A., McCrindle, R., 2007. Nuclear magnetic resonance spectral characterization and semi-empirical calculations of conformations of a- and g-1,2,5,6,9,10-hexabromocyclododecane. Chemosphere 67, 1684e1694. Covaci, A., Gerecke, A.C., Law, R.J., Voorspoels, S., Kohler, M., Heeb, N.V., Leslie, H., Allchin, C.R., de Boer, J., 2006. Hexabromocyclododecanes (HBCDs) in the environment and humans: a review. Environ. Sci. Technol. 40, 3679e3688. Darnerud, P.O., 2003. Toxic effects of brominated flame retardants in man and in wildlife. Environ. Int. 29, 841e853. De Boer, J., 2004. Brominated flame retardants in the environment-the Price for our convenience? Environ. Chem. 1, 81e85. De Wit, C.A., 2002. An overview of brominated flame retardants in the environment. Chemosphere 46, 583e624. ECHA, 2008. Risk Assessment Hexabromocyclododecane. CAS-No.: 25637-99-4, EINECS-No.: 247-148-4, Final report May 2008. Eljarrat, E., Barcelo, D., 2011. Brominated Flame Retardants. Springer-Verlag Berlin Heidelberg. Ema, M., Fujii, S., Hirata-Koizumi, M., Matsumoto, M., 2008. Two-generation reproductive toxicity study of the flame retardant hexabromocyclododecane in rats. Reprod. Toxicol. 25, 335e351.
1397
Heeb, N.V., Schweizer, W.B., Kohler, M., Gerecke, A.C., 2005. Structure elucidation of hexabromocyclododecanesea class of compounds with a complex stereochemistry. Chemosphere 61, 65e73. Heeb, N.V., Graf, H., Schweizer, W.B., Lienemann, P., 2010a. Isobutoxypentabromocyclododecanes (iBPBCDs): a new class of polybrominated compounds. Chemosphere 78, 950e957. Heeb, N.V., Graf, H., Schweizer, W.B., Lienemann, P., 2010b. Thermally-induced transformation of hexabromocyclo dodecanes and isobutoxypenta bromocyclododecanes in flame-proofed polystyrene materials. Chemosphere 80, 701e708. Heeb, N.V., Zindel, D., Schweizer, W.B., Lienemann, P., 2012. 2,5,6,9,10Pentabromocyclododecanols (PBCDOHs): a new class of HBCD transformation products. Chemosphere 88, 655e662. Khoee, S., Sorkhi, M., 2007. Microstructure analysis of brominated styrenebutadiene rubber. Polym. Eng. Sci. 47, 87e94. Lassen, C., Maag, J., Hoibye, L., Vesterlykke, M., Lundegaard, T., 2011. Alternatives to the Use of Flame Retarded EPS in Buildings. COWI A/S Denmark for the Norwegian Climate and Pollution Agency. Law, R.J., Herzke, D., Harrad, S., Morris, S., Bersuder, P., Allchin, C.R., 2008. Levels and trends of HBCD and BDEs in the European and Asian environments, with some information for other BFRs. Chemosphere 73, 223e241. Legler, J., 2008. New insights into the endocrine disrupting effects of brominated flame retardants. Chemosphere 73, 216e222. Managaki, S., Miyake, Y., Yokoyama, Y., Hondo, H., Masunaga, S., Nakai, S., Kobayashi, T., Kameya, T., Kimura, A., Nakarai, T., Oka, Y., Otani, H., Miyake, A., 2009. Emission load of hexabromocyclododecane in Japan based on the substance flow analysis. Organohalogen Compd. 71, 2471e2476. Messer, A., 2010. Mini-review: polybrominated diphenyl ether (PBDE) flame retardants as potential autism risk factors. Physiol. Behav. 100, 245e249. Morf, L., Buser, A., Taverna, R., Bader, H.P., Scheidegger, R., 2008. Dynamic substance flow analysis as a valuable tool e a case study for brominated flame retardants as an example of potential endocrine disruptors. Chimia 62, 424e431. Morose, G., 2006. An Overview of Alternatives to Tetrabromobisphenol A (TBBPA) and Hexabromocyclododecane (HBCD). Lowell Center for Sustainable Production. University of Massachusetts, Lowell. Remberger, M., Sternbeck, J., Palm, A., Kaj, L., Stromberg, K., Brorstrom-Lunden, E., 2004. The environmental occurrence of hexabromocyclododecane in Sweden. Chemosphere 54, 9e21. Tanabe, S., Ramu, K., Isobe, T., Takahashi, S., 2008. Brominated flame retardants in the environment of Asia-Pacific: an overview of spatial and temporal trends. J. Environ. Monit. 10, 188e197. Tang, H.R., Wang, Y.L., Nicholson, J.K., Lindon, J.C., 2004. Use of relaxation-edited one-dimensional and two dimensional nuclear magnetic resonance spectroscopy to improve detection of small metabolites in blood plasma. Anal. Biochem. 325, 260e272. UNEP, 2011. The Persistent Organic Pollutants Review Comitee. United Nations Environment Programme. POPS/POPRC.7/19/Add.1. USEPA, 2014. Flame retardant alternatives for hexabromocyclododecane (HBCD). EPA publication 740R14001. Van der Ven, L.T.M., Verhoef, A., Van de Kuil, T., Slob, W., Leonards, P.E.G., Visser, T.J., Hamers, T., Herlin, M., Hakansson, H., Olausson, H., Piersma, A.H., Vos, J.G., 2006. A 28-day oral dose toxicity study enhanced to detect endocrine effects of hexabromocyclododecane in wistar rats. Toxicol. Sci. 94, 281e292. Willker, W., Leibfritz, D., Kerssebaum, R., Bermel, W., 1993. Gradient selection in inverse heteronuclear correlation spectroscopy. Magn. Reson. Chem. 31, 287e292.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Discrimination of hexabromocyclododecane from new polymeric brominated flame retardant in polystyrene foam by nuclear magnetic resonance bastien Schweizer b, Yavor Nikolaev Mitrev a, c, Damien Jeannerat a, *, Marion Pupier a, Se b b Philippe Favreau , Marcel Kohler a b c
D epartement de chimie organique, Universit e de Gen eve, 30 quai Ernest Ansermet, 1211 Gen eve 4, Switzerland ^ti, 23 avenue de Sainte Clotilde, Case postale 78, 1211 Gen Service de Toxicologie de l'Environnement Ba eve 8, Switzerland Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Acad. G. Bonchev str., bl 9, 1113 Sofia, Bulgaria
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
During a 2013/2014 survey, the HBCDD content in polystyrene ranged from 0.2 to 2.4% by weight in a total of 98 samples. A trend towards the use of copolymerized brominated flame retardants in polystyrene foam was observed between 2013 and 2014. Analyses indicated that expandable and extruded polystyrene foams significantly differed in the a/g HBCDD isomer ratio. Investigation by NMR identified polystyrene samples containing HBCDD or a brominated butadiene styrene (BBS) as copolymer. NMR constitutes a simple method to discriminate flame-retardants in insulating polystyrene foam materials.
a r t i c l e i n f o Article history: Received 27 July 2015 Received in revised form 2 October 2015 Accepted 3 October 2015 Available online xxx
a b s t r a c t Hexabromocyclododecane (HBCDD) is a brominated flame retardant (BFR) and major additive to polystyrene foam thermal insulation that has recently been listed as a persistent organic pollutant by the Stockholm Convention. During a 2013/2014 field analytical survey, we measured HBCDD content ranging from 0.2 to 2.4% by weight in 98 polystyrene samples. Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) analyses indicated that expandable (EPS) and extruded (XPS) polystyrene foams significantly differed in the a/g HBCDD isomer ratio, with a majority of a and g isomers in XPS and EPS, respectively. Interestingly, this technique indicated that some recent materials did not contain HBCDD, but demonstrated bromine content when analysed with X-ray fluorescence (XRF). Further
* Corresponding author. E-mail address: [email protected] (D. Jeannerat). http://dx.doi.org/10.1016/j.chemosphere.2015.10.021 0045-6535/© 2015 Elsevier Ltd. All rights reserved.
1392
Keywords: HBCDD Brominated butadiene styrene Polystyrene insulation foam XRF LC-MS/MS NMR
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
investigation by Nuclear Magnetic Resonance (NMR) was able to discriminate between the BFRs present. In addition to confirming the absence or presence of HBCDD in polystyrene samples, high-field NMR spectroscopy provided evidence of the use of brominated butadiene styrene (BBS) as copolymer in the production of polystyrene. Use of this alternative flame retardant is expected to cause fewer health and environmental concerns. Our results highlight a trend towards the use of copolymerized BFRs as an alternative to HBCDD in polystyrene foam boards. In addition to providing a rapid NMR method to identify polymeric BFR, our analytical approach is a simple method to discriminate between flameretardants in polystyrene foam insulating materials. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction The cycloaliphatic hexabromocyclododecane (HBCDD) is a common brominated flame retardant (BFR) with applications in the polymer and textile industries (Fig. 1) (Alaee et al., 2003). One of its main uses is in thermal insulation for buildings and constructions, where it is added to expanded (EPS) and extruded (XPS) polystyrene foam to reduce their flammability. The increasing use of flame retardant in EPS and XPS has been mainly driven by national regulations, construction safety rules and insurance policies (Lassen et al., 2011), with the aim of significantly reducing large building fires. In the early 2000s, the European market for EPS was approximately 470,000 tons, with increasing demand in Eastern Europe (ECHA, 2008). Recently, the total world market for EPS was estimated at 1.8 million tons per year (Lassen et al., 2011). The use of BFRs has led to some serious concerns about their toxicological and environmental impact (De Wit, 2002; Darnerud, 2003; De Boer, 2004; Covaci et al., 2006; Messer, 2010; Eljarrat and Barcelo, 2011). Environmental data unambiguously demonstrate persistence of HBCDD in various environmental compartments, as well as bioaccumulation in numerous living organisms (Law et al., 2008; Tanabe et al., 2008). Repeat dose studies in animals have shown that HBCDD has significant effects on the liver and thyroid gland (Van der Ven et al., 2006; ECHA, 2008) and on reproductive toxicity (Ema et al., 2008). In addition, HBCDD has been identified as a disruptor of endocrine function, with long-term exposure to low doses shown to have profound effects on hormonal pathways and neurodevelopment (Legler, 2008). In view of its environmental and toxicological impact, HBCDD was recently added to the list of persistent organic pollutants in annex A of the Stockholm Convention. Its use is therefore expected to decline, with a complete commercial ban in mid-2015. Recycling of HBCDD-containing materials will be prohibited in order to avoid uncontrolled dissemination in other materials. However, there is concern that landfills with construction waste containing EPS or XPS materials constitute emission sources of HBCDD to soil, air and water (Remberger et al., 2004). Over the past three decades EPS and
XPS materials have been widely installed in buildings where their presence is raising concerns over disposal during renovation and demolition work. With the exception of information from commercial sources, there is very little published data in the scientific literature on the HBCDD content of insulation materials. In view of these concerns, an efficient methodology is needed to detect and quantify HBCDD in existing and replacement building products. The primary objective of this study was therefore to evaluate the total bromine content of samples using a hand-held Xray fluorescence instrument in EPS and XPS materials, then to submit samples to liquid chromatography and tandem mass spectrometry (LC-MS/MS) analyses to enable precise quantification and profiling of the three major HBCDD isomers. In addition, we used Nuclear Magnetic Resonance (NMR) to further investigate the significant bromine presence, but lack of HBCDD, in some recently collected samples. 2. Experimental 2.1. Chemicals Acetonitrile, methanol and toluene were all of HPLC-grade (Merck Millipore, Darmstadt, Germany) and milliQ water was used throughout the analyses. a-, b-, g- and 13C-g-HBCDD analytical standards were purchased from Cambridge Isotope Laboratories (CIL, USA) at 50 mg/mL in toluene. 2.2. Samples Samples were collected from construction sites in the Geneva region of Switzerland during April 2013 (n ¼ 53) and April 2014 (n ¼ 45). In total, 45 collection sites were sampled, including 30 residential, 14 public (commercial, medical), and one industrial location. Samples were taken from brand new expanded (EPS) and extruded (XPS) polystyrene foams being installed on new building constructions for a broad range of thermal insulation applications (roof, floors, walls, foundations). The material type was identified at the field site by the product name and by visual inspection. Samples originated from various manufacturers, distributors and importers, and corresponded to 50 different trade names. No replicate samples were collected at any one single site in order to avoid sampling boards from the same production batch. However, some board references were found in duplicates or replicates at different collection sites. In these cases, all samples were kept to reflect the true situation encountered in the field. 2.3. X-ray fluorescence analysis
Fig. 1. Molecular structure of HBCDD (left), and a new BFR (right) with bromo butadiene styrene (BBS) blocks inserted in the polymer.
A 40 keV tube-emitting Niton Xlt 700 Series, model Xlt792W (Thermo Fisher Scientific Inc., USA) was used for all XRF analyses. The Niton XLt is a small, field-portable instrument designed for the
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
chemical characterization of soils, sediment, and other thick homogeneous samples, such as plastics and metals. Use of the instrument for very low-density materials (polystyrene) may be considered a specific application. Acquisitions were performed using the ‘Plastic analysis’ mode with reading times (typically 10e20 s acquisition) in duplicate measurements. The instrument was factory-calibrated with polyethylene samples containing 1.04 and 0.1% bromine. An instrument calibration check was carried out prior to each round of analyses using three multi-element reference samples distributed by TechLab (Metz, France) and provided by Modern Analytical Techniques, LLC (Hillsborough, NJ, USA). The three bromine-containing reference materials were in polyvinyl chloride. ‘PVC-01A’, ‘PVC-L-20A’ and ‘PVC-H-19A’ contained 0, 497 and 1094 ppm of bromine respectively, with an overall error of 4% relative to a 95% confidence level, except below 100 ppm, where the error was estimated to be 5 ppm at 95% confidence level (Supplementary Data 3, Table S1). The polyvinyl chloride reference samples were used to qualify the XRF instrument for measuring bromine content under standard conditions. 2.4. Sample preparation Raw samples were analysed directly by XRF without any specific treatment. For each sample, the material thickness under X-ray beam exposure was measured precisely. Prior to LC-MS/MS analysis, 20 mg of sample was collected from the inner part of the material in order to avoid any external contamination, and directly dissolved in 20 mL of toluene. After thorough mixing to obtain complete dissolution (approx. 10 sec), the solution was filtered using a 0.45 mm nylon disposable filter from Merck Millipore (Darmstadt, Germany). The 13C-g-HBCDD internal standard (50 mL at 1 mg/mL) was then added to the filtrate (10 mL) and the mixture was completed with 940 mL of methanol. This solution was injected directly into the LC-MS/MS apparatus. 2.5. LC-MS/MS analysis An Agilent 1200 series HPLC instrument equipped with a degasser, autosampler and oven was used together with a Synergy polar RP column (50 2 mm, 2.5 mm, 100 Å). The oven temperature was set to 25 C. A constant flow of 250 mL/min was applied using a gradient of water (solvent A) and methanol (solvent B). Initial conditions were 50/50 (A:B) and were maintained for 0.3 min after injection of 5 mL of sample. At 3.2 min, a linear gradient of 50e90% B was applied and left for 3.5 min. Then a final step was made to 100% B for 0.1 min and left for 5.9 min before returning to the initial conditions. Mass spectrometry analyses were performed using a QTRAP 3200 from Applied Biosystems (ABSciex, USA) in multi-reaction monitoring (MRM) and electrospray negative ionization modes. Instrumental parameters such as declustering potential, entrance potential, and collision energy were all optimized using the ‘compound optimisation’ mode. Data acquisition and treatment were performed using the Analyst software (version 1.6.1) from Applied Biosystems. Quantification was based on the transitions 640.6/78.9 and 652.6/78.9 for HBCDD isomers and the internal standard, respectively, while the transitions 640.6/80.9 and 652.6/80.9 were used as qualifiers for HBCDD isomers and the internal standard, respectively. Calibration curves covering the quantification range 1e1000 ng/mL were generated for each analyte using 13C-g-HBCDD as an internal standard. Accuracy and relative standard deviation (RSD) were determined by performing addition of HBCDD analytical standards to HBCDD-free EPS material at three concentrations (low: 5 ng/mL, medium: 50 ng/mL and high: 500 ng/mL).
1393
Reproducibility was tested by performing six sample preparations and analyses with one EPS and one XPS sample. Relative recovery (accuracy) was calculated as the ratio (%) between the experimental and the theoretical values. Relative standard deviation was calculated as the standard deviation of all values divided by the mean value. Reproducibility was estimated as the relative standard deviation of all values divided by the mean value. 2.6. NMR experiments NMR experiments were carried out using a Bruker Avance III 500 MHz spectrometer with a 5-mm DCH 13Ce1H/D helium-cooled cryogenic probe equipped with a z-gradient coil. Analyses were performed on fresh samples prepared by dissolving 75 mg (c.a. 2e3 cm3) of raw foam in 750 mL of CDCl3. Samples were filtered using a 0.45 mm nylon disposable filter (EXAPURE SY13TF, Alys Technologies), and 600 mL of the filtrate were then transferred into 5 mm NMR tubes. The measurement temperature was 25 C for all analyses, except those aimed at identifying g-HBCDD (Supplementary Data 1, Fig. S1), which were recorded at 10 C. 1D 1 H spectra were recorded with four scans, a relaxation delay of 2 s, and a total acquisition time of 30 s 1D T2-filtered experiments (Bruker pulse program “cpmg1d”) were acquired with 32 scans and a 205 ms CPMG (Car-Purcell-Meiboom-Gill) period with an interpulse delay of 400 ms. 2D HSQC experiments (Bruker pulse program “hsqcetgpsisp2.2”) were recorded over 2.5 h with a 10 or 15 ppm width in the 1H dimension and a 200 ppm spectral width in the 13C dimension (Supplementary Data 1, Fig. S1). Each of the 512 increments in t1 was acquired with 8 scans, a 1.7 s recycle delay between scans and 4096 points in the direct dimension. Multiplicity-edited experiments (Supplementary Data 1, Fig. S1) were performed using the “hsqcedetgp” Bruker pulse program (Willker et al., 1993). Each of the 256 t1 increments was recorded with 16 scans and 4096 points in the direct dimension. The T2edited HSQC included a CPMG element (see above for details) inserted after the first pulse of the HSQC sequence. Spectra were processed using the standard Fourier transform after application of a squared, shifted sine-bell window function in both dimensions. 3. Results and discussion 3.1. Measurement of the bromine content of EPS and XPS using XRF X-ray fluorescence spectra acquired from the various insulation foam materials typically showed two intense peaks at 11.9 and 13.3 keV that were both characteristic of the elemental bromine emission rays Ka and Kb, respectively (Supplementary Data 2, Fig. S2). We detected no interfering peak or other major emission ray specific to a particular element. Results showed a relative standard deviation of less than 10% for all samples (Supplementary Data 3, Table S2); it should be noted that most EPS and XPS measurements were in the 300 to 20,000 ppm range, therefore higher than the highest calibration sample of 1094 ppm. The XRF instrument detected the presence of bromine in all samples collected in 2013 and 2014, with values ranging from 0.2 to 1.5% atomic bromine (Fig. 2). XFR only indicates the presence of elemental bromine, and cannot distinguish HBCDD from other brominecontaining compounds. Therefore, we subsequently analysed all samples by LC-MS/MS and NMR in order to characterize bromine compounds present in more detail. 3.2. XRF quantification versus LC-MS/MS data We devised the LC-MS/MS quantification method to quickly and
1394
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
clearly separate the three HBCDD isomers in the ng/mL range. First, we validated the instrument and methodology for precision, accuracy and reproducibility. The methodological validation was performed on BFR-free polystyrene material that was spiked with HBCDD in the range 1e1000 ng/mL (Table 1 and Supplementary data 4, Table S3). We devised a quick and simple preparation method based on the favourable dissolution properties of polystyrene in toluene. Filtration, dilution in methanol and addition of the internal standard were carried out prior to sample injection. Typical extracted ion chromatograms for the three HBCDD isomers are shown in Fig. 3. The a-HBCDD eluted first, followed by the g- and b-HBCDD isomers. Surprisingly, the isomer elution profile was different from that observed in previous studies with different reverse-phase columns and gradients (Heeb et al., 2010a). However, the injection of each individual isomer confirmed the observed profile under our analytical conditions. One should thus be careful when changing experimental parameters that could affect the elution order of the HBCDD isomers, thus potentially giving rise to errors in peak identification. Overall, we observed a good correlation between XRF and LC-MS/MS data (Fig. 2). The bromine content of EPS samples was sometimes underestimated by XRF because the measurements were dependent on the sample thickness (Supplementary Data 5, Fig. S3). We obtained the best match with LC-MS/MS data using a minimum thickness of 12 cm. In addition to HBCDD, two new families of polybrominated compounds have recently been found in HBCDD-treated polystyrene materials (Heeb et al., 2010a, 2012); these are likely byproducts of HBCDD chemical synthesis. Isobutoxypentabro mocyclododecanes (iBPBCDs) and pentabromocyclododecanols (PBCDOHs) were previously detected in EPS and XPS insulation foam boards as well as in HBCDD technical mixtures; however, quantification of these compounds was prevented by the unavailability of pure standards, and these were not included in the targeted LC-MS/MS analytical method. When HBCDD was detected, the total bromine content correlated well with HBCDD quantification, suggesting that iBPBCDs and PBCDOHs did not represent a significant portion of the total brominated content of EPS and XPS. 3.3. HBCDD quantification in polystyrene insulation foams All samples were analysed by LC-MS/MS to quantify the three
1.60% 1.40% 1.20% 1.00%
0.80% 0.60% 0.40% 0.20% 0.00% 0.00%
0.50%
1.00%
1.50%
2.00%
Fig. 2. Correlation of bromine content based on HBCDD quantified by LC-MS/MS and XRF analysis. Note that the lower limit of quantification of LC-MS/MS was 0.03%; lower limit of detection for XRF measurements was estimated to be less than 50 ppm (0.005%).
Table 1 Validation of relative recovery (accuracy) and precision from measurements of HBCDD and its a, b and g-isomers using LC-MS/MS. Sample preparation
a-HBCD
QC low, 5 ng/mL, (n ¼ 3) Relative recovery (%) 100.9 RSD (%) 6.6 QC medium, 50 ng/mL, (n ¼ 3) Relative recovery (%) 102 RSD (%) 5.5 QC high, 500 ng/mL, (n ¼ 3) Relative recovery (%) 100.2 RSD (%) 10.6
b-HBCD
g-HBCD
Total HBCD
102 1.2
109.9 5.1
103.6 4.0
104.2 5.7
102.8 7.7
102.4 2.0
103 7.9
102.3 4.0
101.1 1.8
HBCDD isomers. We were therefore able to precisely measure the HBCDD content of EPS and XPS based on the detection of the major a-, b- and g-isomers (Supplementary data 6, Table S4). Using LCMS/MS, we detected HBCDD in all samples collected in 2013, but in only 70% of those collected in 2014. EPS materials had a mean HBCDD total content of 0.67%, with values ranging from 0.25 to 2.4%; for XPS insulation foams, values ranged from 0.2 to 2.29%, with a mean of 1.1% (Table 2). We found the mean HBCDD content of the EPS boards we sampled was slightly higher than the usual value (0.5%) reported by the industry (Morose, 2006). The maximum concentration of HBCDD in EPS beads was previously reported to be 0.7% (ECHA, 2008), which is much lower than the maximum of the 2.4% reported here. The overall range observed in the present study (0.2e2.4%) is also wider than the previously reported range of 0.5e1% (Lassen et al., 2011). This difference may be attributed to the lack of previous representative and robust data, and could also reflect variations in HBCDD used in polystyrene beads at different production sites. We also observed a significant variation in total HBCDD content of foam insulation boards displaying an identical trade name and reference, which probably arises from different production sites or batches of origin (Supplementary Data 6, Table S5). The mean HBCDD content of XPS was higher than for EPS, with a similar range of variation. Our observations therefore do not entirely support previously published data on the HBCDD content range of EPS and XPS materials (0.5e1% for EPS and 1e2% for XPS). In contrast, we found that both materials contained a similar mean amount of HBCDD (0.7 and 1.1 for EPS and XPS, respectively), with high variation in the amounts of both materials 0.2e2.4%. In addition to the variation possibly due to the diversity in production batches or sites, the use of synergists (i.e. dicumyl peroxide), which were not investigated in this study, could affect the HBCDD content. Such additives would enhance the flame retardant activity, allowing for varying amounts of HBCDD to be used to obtain an equivalent efficacy. The relative ratio of HBCDD isomers in our samples was tightly linked to the type of polystyrene foam in all cases. It has been reported that the usual HBCDD technical mixture contains mainly the g-HBCDD isomer (Heeb et al., 2005). We found that the relative presence of this isomer in EPS appeared to be in line with the technical HBCDD mixture used during polystyrene bead formation. But for XPS, the major isomer present was a-HBCDD, which is probably due to the thermal transformation of g-to a-HBCDD (Heeb et al., 2010b). Indeed, the industrial manufacturing of XPS uses a higher temperature (200 C) than EPS (120 C) during the foaming process. We found the mean proportion of g-HBCDD in EPS was 75%, which corresponded to the proportion of a-HBCDD content in XPS. The proportion of the least abundant b-HBCDD isomer in both EPS and XPS was always in the range of 1e19%. Interestingly, 30% of the 45 samples collected in 2014 did not show a correlation between the bromine content measured by XRF
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
1395
Fig. 3. LC-MS/MS ion chromatogram traces of a-, g- and b-HBCD (respectively at 7.7, 8.4 and 8.8 min) in EPS and XPS insulation foam materials.
Table 2 HBCDD content and isomer ratios of EPS and XPS samples. HBCDD content (%)
EPS samples (n ¼ 32) Min. value Max. value Average RSD (%) XPS samples (n ¼ 54) Min. value Max. value Average RSD (%)
HBCDD isomer relative ratio/total HBCDD (%)
a-HBCDD
b-HBCDD
g-HBCDD
Total HBCDD
a-HBCDD
b-HBCDD
g-HBCDD
0.02 0.45 0.10 88
0.00 0.33 0.06 104
0.17 1.88 0.51 79
0.25 2.40 0.67 73
5 41 16 47
1 19 9 57
40 93 75 17
0.15 1.71 0.83 44
0.03 0.38 0.17 47
0.02 0.24 0.11 48
0.20 2.29 1.11 44
64 83 75 4
11 19 15 10
6 19 10 21
and the HBCDD content measured by LC-MS/MS (Fig. 2), demonstrating that the bromine content was not originating from HBCDD. Using LC-MS/MS, we did not detect HBCDD in around 60% of EPS samples (12 out of 21). We therefore performed investigations using NMR in order to elucidate the nature of the brominated material present in EPS foam boards.
HeCeBr signals in the 3.5e4.5/47e60 1H/13C ppm (Arsenault et al., 2007) region; see Fig. 4a and a more complete set of spectra in the Supplementary Data 1). The methylene signals of HBCDDs (see the dotted box in Fig. 4b) were located in a crowded region of the spectrum near the very intense polymer signals from CH2-Ar groups (1.5e2.5 ppm region), making them much less attractive for reliable identification of HBCDD.
3.4. Discrimination of HBCDD-containing polystyrene by NMR The three most abundant HBCDD isomers (Supplementary data 7, Fig. S4 for the full structures) showed well-characterised fluxional behaviour (Arsenault et al., 2007) enlarging NMR signals. The a-isomer was the least affected and presented resolved signals in HSQC spectra recorded at 25 C. As predicted from the dynamic properties of HBCDDs, decreasing the temperature to 10 C reduced the dynamic of the g-isomer sufficiently to present resolved signals in the HSQC spectrum (Supplementary data 1 and 8). To observe signals of the least-abundant b-isomer, it would be necessary to further reduce the temperature; however, these would still be difficult to observe due to the absence of a C2 symmetry element increasing the sensitivity to the two major isomers (Supplementary data 7, Fig. S4). As the three stereoisomers of HBCDD are always observed together, we discriminated HBCDDcontaining samples at room temperature based on the fingerprint signals of the a-isomer (three characteristic and well resolved
3.5. Characterisation of a polymeric brominated polystyrene by NMR In an attempt to confirm if HBCDDs have been replaced by new BFR structures in recent polystyrene materials, we combined standard HSQC and multiplicity-edited HSQC experiments (Willker et al., 1993) on HBCDD-free samples (Fig. 4b and Supplementary Data 1). The spectra showed broad signals at 3.86e4.02/33.0 1 H/13C ppm, 4.4/55.4 ppm and 4.6/56.4 ppm that were in perfect accordance with the expected position of brominated carbons in a new BFR (Khoee and Sorkhi, 2007) alternative to HBCDD (Fig. 1) (USEPA, 2014). These signals probably correspond to the eCH2eBr and >CHeBr of the polymer (the type of carbon was confirmed with edited HSQC). This assignment was further supported by T2 relaxation-edited experiments (Tang et al., 2004) (Supplementary data 8 and 9) that confirmed that these signals had short T2 relaxation times, which would be expected if these brominated
1396
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
Fig. 4. Typical HSQC spectra of polystyrene dissolved in CDCl3. (a) The three characteristic CH signals of a-HBCD indicate the presence of prohibited-HBCDD flame retardants. The dotted square indicates the location of the methylene signals of a- and g-HBCDD. (b) In samples containing no HBCDD, broad signals were observed (dotted circles), which probably correspond to eCH2eBr and two types of HeCeBr (possibly corresponding to backbone and side-chain locations) on a bromine-containing polymer (see Supplementary data 1 for more details).
carbons were indeed part of the polymer. Such T2-edited experiments improve the quality of the spectra of small molecules, as the free HBCDD present in samples containing large amount of slowrelaxing signals, such as those of the polymer. However, we found that ‘standard’ HSQC experiments were sufficient for clear discrimination, provided that high-sensitivity probes (helium or liquid nitrogen cooled) were used in order to obtain clear 2D spectra within a reasonable acquisition time (Supplementary data 10). We analysed the three main types of polystyrene foams among collected samples using NMR, and found EPS and XPS foams containing HBCDD, and EPS foams containing bromo butadiene styrene copolymer (Supplementary data 1).
4. Conclusion This study provides precise quantitative data on HBCDD incorporated in expanded and extruded polystyrene foams used in building thermal insulation. Our results confirm that the use of these construction materials over the past decades now constitutes a huge reservoir of a toxic, bioaccumulative and persistent organic pollutant. The waste generated by demolition and renovation works will undoubtedly result in long-term environmental dissemination of HBCDD (Morf et al., 2008; Managaki et al., 2009) if this pollution source is not controlled by an appropriate waste management strategy (UNEP, 2011). Standard 2D HSQC NMR spectra clearly distinguished
D. Jeannerat et al. / Chemosphere 144 (2016) 1391e1397
polystyrene foams containing conventional HBCDD from new foams containing a bromo butadiene styrene copolymer, providing a qualitative method to discriminate BFRs in polystyrene. This technique will enable the evaluation of BFRs in materials and in the environment, and the monitoring of the human and environmental fate and impact of these substances, contributing to the global effort to reduce the environmental release of an important worldwide pollutant. Acknowledgements We would like to thank Jean-Luc Bailly for his help with field sampling, Natacha Duran for her dedication to experimental work, and Ron Hogg of OmniScience SA. for carefully reading and correcting the manuscript. We also acknowledge Marc Dupayrat for his expertise in XRF technology. The authors thank the State of Geneva and the Swiss NSF (grant no. 200021_147069 and 206021_128746) for funding. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2015.10.021. References Alaee, M., Arias, P., Sjodin, A., Bergman, A., 2003. An overview of commercially used brominated flame retardants, their applications, their use patterns in different countries/regions and possible modes of release. Environ. Int. 29, 683e689. Arsenault, G., Chittim, B., McAlees, A., McCrindle, R., 2007. Nuclear magnetic resonance spectral characterization and semi-empirical calculations of conformations of a- and g-1,2,5,6,9,10-hexabromocyclododecane. Chemosphere 67, 1684e1694. Covaci, A., Gerecke, A.C., Law, R.J., Voorspoels, S., Kohler, M., Heeb, N.V., Leslie, H., Allchin, C.R., de Boer, J., 2006. Hexabromocyclododecanes (HBCDs) in the environment and humans: a review. Environ. Sci. Technol. 40, 3679e3688. Darnerud, P.O., 2003. Toxic effects of brominated flame retardants in man and in wildlife. Environ. Int. 29, 841e853. De Boer, J., 2004. Brominated flame retardants in the environment-the Price for our convenience? Environ. Chem. 1, 81e85. De Wit, C.A., 2002. An overview of brominated flame retardants in the environment. Chemosphere 46, 583e624. ECHA, 2008. Risk Assessment Hexabromocyclododecane. CAS-No.: 25637-99-4, EINECS-No.: 247-148-4, Final report May 2008. Eljarrat, E., Barcelo, D., 2011. Brominated Flame Retardants. Springer-Verlag Berlin Heidelberg. Ema, M., Fujii, S., Hirata-Koizumi, M., Matsumoto, M., 2008. Two-generation reproductive toxicity study of the flame retardant hexabromocyclododecane in rats. Reprod. Toxicol. 25, 335e351.
1397
Heeb, N.V., Schweizer, W.B., Kohler, M., Gerecke, A.C., 2005. Structure elucidation of hexabromocyclododecanesea class of compounds with a complex stereochemistry. Chemosphere 61, 65e73. Heeb, N.V., Graf, H., Schweizer, W.B., Lienemann, P., 2010a. Isobutoxypentabromocyclododecanes (iBPBCDs): a new class of polybrominated compounds. Chemosphere 78, 950e957. Heeb, N.V., Graf, H., Schweizer, W.B., Lienemann, P., 2010b. Thermally-induced transformation of hexabromocyclo dodecanes and isobutoxypenta bromocyclododecanes in flame-proofed polystyrene materials. Chemosphere 80, 701e708. Heeb, N.V., Zindel, D., Schweizer, W.B., Lienemann, P., 2012. 2,5,6,9,10Pentabromocyclododecanols (PBCDOHs): a new class of HBCD transformation products. Chemosphere 88, 655e662. Khoee, S., Sorkhi, M., 2007. Microstructure analysis of brominated styrenebutadiene rubber. Polym. Eng. Sci. 47, 87e94. Lassen, C., Maag, J., Hoibye, L., Vesterlykke, M., Lundegaard, T., 2011. Alternatives to the Use of Flame Retarded EPS in Buildings. COWI A/S Denmark for the Norwegian Climate and Pollution Agency. Law, R.J., Herzke, D., Harrad, S., Morris, S., Bersuder, P., Allchin, C.R., 2008. Levels and trends of HBCD and BDEs in the European and Asian environments, with some information for other BFRs. Chemosphere 73, 223e241. Legler, J., 2008. New insights into the endocrine disrupting effects of brominated flame retardants. Chemosphere 73, 216e222. Managaki, S., Miyake, Y., Yokoyama, Y., Hondo, H., Masunaga, S., Nakai, S., Kobayashi, T., Kameya, T., Kimura, A., Nakarai, T., Oka, Y., Otani, H., Miyake, A., 2009. Emission load of hexabromocyclododecane in Japan based on the substance flow analysis. Organohalogen Compd. 71, 2471e2476. Messer, A., 2010. Mini-review: polybrominated diphenyl ether (PBDE) flame retardants as potential autism risk factors. Physiol. Behav. 100, 245e249. Morf, L., Buser, A., Taverna, R., Bader, H.P., Scheidegger, R., 2008. Dynamic substance flow analysis as a valuable tool e a case study for brominated flame retardants as an example of potential endocrine disruptors. Chimia 62, 424e431. Morose, G., 2006. An Overview of Alternatives to Tetrabromobisphenol A (TBBPA) and Hexabromocyclododecane (HBCD). Lowell Center for Sustainable Production. University of Massachusetts, Lowell. Remberger, M., Sternbeck, J., Palm, A., Kaj, L., Stromberg, K., Brorstrom-Lunden, E., 2004. The environmental occurrence of hexabromocyclododecane in Sweden. Chemosphere 54, 9e21. Tanabe, S., Ramu, K., Isobe, T., Takahashi, S., 2008. Brominated flame retardants in the environment of Asia-Pacific: an overview of spatial and temporal trends. J. Environ. Monit. 10, 188e197. Tang, H.R., Wang, Y.L., Nicholson, J.K., Lindon, J.C., 2004. Use of relaxation-edited one-dimensional and two dimensional nuclear magnetic resonance spectroscopy to improve detection of small metabolites in blood plasma. Anal. Biochem. 325, 260e272. UNEP, 2011. The Persistent Organic Pollutants Review Comitee. United Nations Environment Programme. POPS/POPRC.7/19/Add.1. USEPA, 2014. Flame retardant alternatives for hexabromocyclododecane (HBCD). EPA publication 740R14001. Van der Ven, L.T.M., Verhoef, A., Van de Kuil, T., Slob, W., Leonards, P.E.G., Visser, T.J., Hamers, T., Herlin, M., Hakansson, H., Olausson, H., Piersma, A.H., Vos, J.G., 2006. A 28-day oral dose toxicity study enhanced to detect endocrine effects of hexabromocyclododecane in wistar rats. Toxicol. Sci. 94, 281e292. Willker, W., Leibfritz, D., Kerssebaum, R., Bermel, W., 1993. Gradient selection in inverse heteronuclear correlation spectroscopy. Magn. Reson. Chem. 31, 287e292.