Tribromophenoxy Flame Retardants in the Great Lakes Atmosphere

an atmospheric concentration of ∼180 ng/m3 in Arkansas, where a flame retardant manufacturer, Great Lakes Chemical (El Dorado, Arkansas), was lo...
1 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Tribromophenoxy Flame Retardants in the Great Lakes Atmosphere Yuning Ma, Marta Venier, and Ronald A. Hites* School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405, United States S Supporting Information *

ABSTRACT: The 2,4,6-tribromophenoxy moiety is a common structural feature of several brominated flame retardants, and we have previously reported on the environmental concentrations of one such compound, 1,2bis(2,4,6-tribromophenoxy) ethane (TBE). Here we report the atmospheric concentrations of TBE and three other tribromophenoxy compounds: allyl 2,4,6-tribromophenyl ether (ATE), 2-bromoallyl 2,4,6-tribromophenyl ether (BATE), and 2,3-dibromopropyl 2,4,6-tribromophenyl ether (DPTE). The samples were collected at five sites near the shores of the Great Lakes during the period 2008−2009, inclusive. Of these four compounds, TBE and ATE are currently used as flame retardants, and DPTE was formerly used as a flame retardant until its production ceased in the mid-1980s. The total concentrations of ATE, BATE, and DPTE were ∼2 pg/m3 in the cities of Chicago and Cleveland and 0.1−0.4 pg/m3 at the rural and remote sites. The concentrations of TBE were ∼1 pg/m3 in these cities and 0.2−0.8 pg/m3 at the rural and remote sites. In both cases, this was a very significant urban effect. The concentrations of ATE, BATE, and DPTE did not change significantly over the two-year study, but the concentrations of TBE decreased by about a factor of 2 during this time. This temporal change was statistically significant but not strong compared to the urban effect.

■

(TBP), which is itself a flame retardant, commercially known as PH-73 FF and currently marketed by Chemtura.5

INTRODUCTION Flame retardants, which are added to many materials to give people extra time to escape from structural fires,1 have become ubiquitous and persistent in the environment. The most wellknown of these compounds are the polybrominated compounds, such as the polybrominated diphenyl ethers (PBDEs), the polybrominated biphenyls (PBBs), and the hexabromocyclododecanes (HBCDs). With the exception of decabromodiphenyl ether, PBDEs and PBBs are no longer produced, but other polybrominated compounds are. These include 2ethylhexyl-2,3,4,5-tetrabromobenzoate (TBB) and di(2-ethylhexyl) tetrabromophthalate (TBPH), which are the major components of an “alternative” flame retardant called Firemaster 550. We have recently shown that the atmospheric concentrations of these two compounds have been increasing over time at five locations around the Great Lakes and that their concentrations are now 0.31−14 pg/m3, depending on the site.3 Other less well-known flame retardants include hexabromobenzene (HBB), pentabromobenzene (PBBz), pentabromotoluene (PBT), pentabromobenzylacrylate (PBBA), pentabromobenzyl bromide (PBBB), tetrabromo-pxylene (pTBX), and pentabromoethyl benzene (PBEB), and their atmospheric concentrations near the Great Lakes are now 0.21−4.6 pg/m3, depending on location.4 This paper focuses on four structurally related tribromophenoxy flame retardants: allyl 2,4,6-tribromophenyl ether (ATE), 2-bromoallyl 2,4,6-tribromophenyl ether (BATE), 2,3-dibromopropyl 2,4,6-tribromophenyl ether (DPTE), and 1,2bis(2,4,6-tribromophenoxy)ethane (TBE or BTBPE). These four compounds are all synthesized from 2,4,6-tribromophenol © 2012 American Chemical Society

TBE is synthesized by the condensation of two moles of TBP with 1,2-dibromoethane.6 TBE was introduced into the market in the mid-1970s by the Great Lakes Chemical Corporation, which is now part of Chemtura.7 It is marketed under the trade name FF-680 and used in acrylonitrile-butadiene-styrene polymers, high impact polystyrene, polycarbonate, thermoplastic elastomers, unsaturated polyesters, adhesives, coatings, and textiles.5 The total aggregated production of TBE in the United States’ Environmental Protection Agency’s Inventory Update Report in 2006 was between 500 and 5000 t.8 TBE has Received: Revised: Accepted: Published: 13112

August 20, 2012 October 16, 2012 November 7, 2012 November 26, 2012 dx.doi.org/10.1021/es3033814 | Environ. Sci. Technol. 2012, 46, 13112−13117

Environmental Science & Technology

Article

Table 1. Summary of Detection Frequencies and Total Concentrations (in pg/m3) Measured in the Atmosphere near the Great Lakes Regiona detection frequency %

mean conc. ± std err

median conc.

conc. range

ATE BATE DPTE ABDb,c TBE ATE BATE DPTE ABDb TBE ATE BATE DPTE ABDb TBE ATE BATE DPTE ABDb TBE

Chicago 76 18 60 91 93 1.6 ± 0.4 0.41 ± 0.11 0.49 ± 0.10 1.79 ± 0.36 0.86 ± 0.11 1.2 0.35 0.24 1.0 0.63 0.012−15 0.012−1.1 0.046−2.7 0.076−16 0.11−3.4

Cleveland 80 32 22 86 98 1.9 ± 0.3 0.12 ± 0.03 0.33 ± 0.07 1.88 ± 0.30 0.79 ± 0.11 1.2 0.092 0.21 1.3 0.51 0.046−7.0 0.012−0.44 0.11−0.78 0.046−7.7 0.025−3.1

Sturgeon Point 58 34 31 80 92 0.35 ± 0.09 0.29 ± 0.19 0.059 ± 0.009 0.40 ± 0.11 0.44 ± 0.14 0.12 0.062 0.049 0.13 0.14 0.012−2.2 0.036−3.9 0.024−0.15 0.012−4.0 0.012−7.2

Sleeping Bear Dunes 41 44 5 61 98 0.31 ± 0.12 0.076 ± 0.018 0.069 ± 0.051 0.27 ± 0.09 0.42 ± 0.06 0.052 0.045 0.024 0.081 0.26 0.012−2.5 0.012−0.46 0.012−0.17 0.012−3.0 0.037−2.7

Eagle Harbor 27 43 8 57 50 0.076 ± 0.014 0.043 ± 0.008 0.11 ± 0.04 0.084 ± 0.015 0.17 ± 0.04 0.064 0.031 0.11 0.044 0.086 0.018−0.19 0.012−0.20 0.024−0.20 0.012−0.39 0.025−1.0

all sites inclusive 55 34 24 73 85 1.06 ± 0.15 0.15 ± 0.04 0.31 ± 0.05 0.96 ± 0.12 0.56 ± 0.05 0.22 0.05 0.17 0.22 0.32 0.012−15 0.012−3.9 0.012−2.7 0.012−16 0.012−7.2

The concentrations observed in the vapor and particle phases are added together; all means are arithmetic averages ± their standard errors. bABD represents the sum of ATE, BATE, and DPTE concentrations. cThe detection frequency of at least one of the three compounds.

a

propyl 2,4,6-tribromophenyl ether (DPTE), 1,2-bis(2,4,6tribromophenoxy)ethane (TBE), and the isotopically labeled PBDE congener 13C12−BDE-209 were purchased from Wellington Laboratories, Guelph, ON. Native PBDE congeners BDE-118, BDE-77, and BDE-166 were purchased from AccuStandard, New Haven, CT. Sample Information. Air samples were collected for 24 h every 12 days from January 2008 to December 2009 using modified Anderson high-volume air samplers (General Metal Works, model GS2310) at five IADN stations. The average air flow rate was ∼0.6 m3/min, and the total sampling volume was ∼850 m3. The sampling sites included two urban sites, one in Chicago (Chi), IL (41.8344 N, 87.6247 W) and one in Cleveland (Clev), OH (41.4921 N, 81.6785 W); a rural site in Sturgeon Point (StPt), NY (42.6931 N, 79.0550 W); and two remote sites, one in Eagle Harbor (EH), MI (47.4631 N, 88.1497 W) and one in Sleeping Bear Dunes (SBD), MI (44.7611 N, 86.0586 W). The high-volume air samplers were equipped with 20.3 × 25.4 cm Whatman QM-A quartz fiber filters to first collect the particle phase compounds and with a cartridge containing ∼40 g of Amberlite XAD-2 resin (20−60 mesh) to collect the vapor phase compounds. Details are available on the IADN Web site.20 All the samples were spiked with known amounts of the three surrogate recovery standards (BDE-77, BDE-166, and 13 C12−BDE-209). The first two recovery standards were selected because they do not occur in the commercial PBDE mixtures, and the third was selected because it allows the determination of BDE-209 by isotopic dilution. The samples were Soxhlet extracted with 25 mL of 50% acetone in n-hexane (OmniSolv, Inc.; Gibbstown, NJ) for 24 h, and then fractionated through a column of 3.5% water deactivated silica (Fisher Scientific Inc.; Hampton, NH). Two fractions were collected: One was eluted with n-hexane, and the other was eluted with 50% dichloromethane (OmniSolv, Inc.) in nhexane. The quantitation internal standard (BDE-118) was

been widely detected in the environment. It has been found in sediments,9 air,10 indoor dust,11 fish,12 and baby products.13 ATE is formed by the condensation of one mole of TBP with 3-bromo-1-propene (allyl bromide).14 ATE is currently produced by Chemtura Corporation and marketed under the trade name PHE-65. ATE is commonly added to expandable or foamed polystyrene.5 The total production of ATE given in the United States’ Environmental Protection Agency’s Inventory Update Report in 2006 was less than 230 t.8 DPTE was manufactured in Germany, by bromination of ATE, until the mid-1980s by Chemische Fabrik Kalk (Köln, North Rhine Westphalia) under the trade name Bromkal 73−5PE, but it does not seem to be a commercial product anymore. BATE does not seem to have been used as a flame retardant,15 but it is potentially a transformation product of DPTE. Publications on the environmental presence of ATE, DPTE, and BATE are rare. These compounds have been found in blubber and brain samples from hooded and harp seals at levels of 2.0−470 μg/kg wet weight.16 DPTE was also found in air samples from Asia to Antarctica.17 In addition, ATE and TBE have been listed in the San Antonio Statement.18 ATE, BATE, and DPTE can penetrate the blood-brain barrier,16 and inhalation of TBE may cause behavioral and gastrointestinal changes and dermatitis.19 In this study, we report the atmospheric concentrations of these four tribromophenoxy compounds in about 300 atmospheric samples collected near the Great Lakes by the Integrated Atmospheric Deposition Network (IADN) in 2008 and 2009, inclusive. We also investigate the spatiotemporal factors controlling their concentrations. This is the first systematic study on the spatiotemporal trends of these compounds in North American air.

■

EXPERIMENTAL SECTION Chemicals. Allyl 2,4,6-tribromophenyl ether (ATE), 2bromoallyl 2,4,6-tribromophenyl ether (BATE), 2,3-dibromo13113

dx.doi.org/10.1021/es3033814 | Environ. Sci. Technol. 2012, 46, 13112−13117

Environmental Science & Technology

Article

spiked into the final extracts prior to instrumental analysis. Details of the sampling and extraction procedures are available at the IADN Web site.20 To check the efficiency of these extraction procedures for ATE, BATE, DPTE, and TBE, a matrix-spike study was performed. The recoveries for these four spiked compounds were between 84% and 92% (N = 6), which was acceptable. Analytical Methods. The extracts were analyzed for ATE, BATE, DPTE, and TBE by an Agilent (Palo Alto, California) 7890 series gas chromatograph coupled to an Agilent 5975C mass spectrometer (GC-MS) operating in the electron capture negative ionization (ECNI) mode. Chromatographic separation was achieved on an Rtx-1614 fused silica capillary column (15 m × 250 μm i.d.; 0.1 μm film thickness; Restek Corporation, Bellefonte, CA). One μL of sample was injected in the pulsed splitless mode at an injection port temperature of 240 °C. The GC oven temperature was programed as follows: 100 °C (for 2 min), 25 °C/min to 250 °C, 3 °C/min to 270 °C, then 25 °C/ min to 300 °C, 300 °C (for 6 min). The helium carrier gas (99.999%; Liquid Carbonic, Chicago) was at a constant flow rate of 1.5 mL/min. The temperatures of the GC/MS transfer tube and the ion source were maintained at 280 and 200 °C, respectively. Selected ion monitoring of the bromide ions at m/ z 79 and 81 was used for the quantitation of ATE, BATE, DPTE, TBE, and BDE-118. Quality Control. All solvents were high resolution gas chromatography grade and were checked to show that there were no interferences with the project’s other analytes. Data were reported only if the recoveries of at least two of the three surrogate standards were in the range of 70−130%. A procedural blank was run with every other batch of ∼8 samples. Field blanks were collected regularly at each sampling site. ATE, BATE, DPTE were not detected in any procedural or field blank sample. The annual field blank average of TBE was less than 0.01 pg/m3. Concentrations below this level were reported as nondetects and replaced by empty cells in the data spreadsheets. We did not correct any concentrations in the samples for blank levels or recoveries.

tetrabrominated benzoate and phthalate compounds measured at the same sites in 2008−2009 (0.31−14 pg/m3, filters only);3 and about one-twelfth of the previously reported concentrations for the no longer used polybrominated diphenyl ether flame retardants (e.g., BDE-47 at 1.8−23 pg/m3) measured at the same sites in 2005−2009.21 TBE was the most abundant of these four compounds at the rural and remote sites, with median concentrations ranging from 0.09 to 0.26 pg/m3. At the urban sites, the concentrations of ATE surpassed those of TBE, with median concentrations of 1.2 pg/m 3 . In fact, ATE showed the single highest concentration of any compound measured in this study at 15 pg/m3 on 26 November 2008 in Chicago. We have verified that this high value was not the result of sample contamination. In fact, we observed unusually high concentrations of PAHs and, to some extent, PBDEs on this same date. Unfortunately, we do not have an explanation for these relatively high concentrations. DPTE and BATE were found in only 24−34% of the samples, and their levels were the lowest among these four compounds with average concentrations of 0.31 ± 0.05 and 0.15 ± 0.04 pg/ m3, respectively. ATE and DPTE were more frequently detected at the urban sites compared to the remote sites. On the other hand, BATE had its lowest detection frequency at Chicago for reasons that are not yet clear. Atmospheric concentrations of ATE and BATE have not been reported previously. Lee et al. reported only the detection frequencies of ATE and BATE in the Global Atmospheric Passive Sampling (GAPS) network.22 DPTE has been found in the marine atmosphere (over the Atlantic, Pacific, Indian, Arctic, and Southern Oceans) at concentrations up to 5.9 pg/ m3.17 Interestingly, TBE was identified as early as 1977 at an atmospheric concentration of ∼180 ng/m3 in Arkansas, where a flame retardant manufacturer, Great Lakes Chemical (El Dorado, Arkansas), was located.23 Years later, lower concentrations of TBE were detected in the east-central United States in samples collected during 2003−2004 with an average of 3.4 pg/m3 in Arkansas, 1.6 pg/m3 in Chicago, and 0.16 pg/m3 in Sleeping Bear Dunes.7 Since that time, TBE has been routinely measured by the IADN project, which reported averages of 0.5−1.2 pg/m3 (2005−2006) and 0.2−1.0 pg/m3 (2005− 2009).21,24 Outside the U.S., TBE has been found in the atmosphere at levels less than 0.06 pg/m3 over European Arctic,25 less than 4 pg/m3 near Taihu Lake in China,26 and 0.16−20 pg/m3 around the Canadian High Arctic and Tibetan plateau,27 indicating the potential for long-range atmospheric transport of this compound. Spatiotemporal Trends. There was a consistent spatial variability to the concentrations, with the urban sites in Chicago and Cleveland always showing very much higher concentrations for TBE and ABD compared to the other three sites. Figure 1 shows box plots for these concentration distributions, along with the site-by-site ANOVA results. For both compound sets, the concentrations were always the highest at the two urban sites, Chicago and Cleveland, but the concentrations were not distinguishable from one another at these sites. These concentrations were generally not distinguishable from one another at rural and remote sites with the exceptions of relatively high level of TBE at Sleeping Bear Dunes and relatively low level of ABD at Eagle Harbor. This trend seems to follow that of the local human population, with the highest concentrations observed near densely populated areas. A similar trend was previously observed for PBDEs, PCBs, PAHs, organochlorine pesti-

■

RESULTS AND DISCUSSION Atmospheric Concentrations. ATE, BATE, DPTE, and TBE were present in both the particle and vapor phases, and the Supporting Information gives the full data set. For simplicity, the concentrations of the compounds in the particle and vapor phases have been added together, and a summary of these concentrations (means, medians, and ranges) and the detection frequencies are given in Table 1. Because the detection frequencies of ATE, BATE, and DPTE were relatively low and because these three compounds are closely related to each other structurally, we have summed their concentrations, and these values are also reported in Table 1. This approach gave similar numbers of measurements for the two compound groups (ATE + BATE + DPTE, on the one hand, and TBE, on the other), which in turn provided for more balanced statistical analyses. We will refer to the sum of the ATE, BATE, and DPTE concentrations as ABD. In the atmosphere over the Great Lakes region, the average concentrations of these four tribromophenoxy flame retardants are 0.043−1.9 pg/m3 at the five sampling locations. These concentrations were about one-third of the previously reported concentrations for the bromobenzene flame retardants (0.21− 4.6 pg/m3) measured at the same sites in 2008−2010;4 about one-sixth of the previously reported concentrations for the 13114

dx.doi.org/10.1021/es3033814 | Environ. Sci. Technol. 2012, 46, 13112−13117

Environmental Science & Technology

Article

Figure 1. Concentrations (given as the natural logarithms of the pg/ m3 values) of TBE and the sum of ATE + BATE + DPTE (ABD) segregated by sampling site. Distributions that share a letter are not significantly different (P < 0.05). The red lines represent the arithmetic mean, and the black lines inside each box represent the median. The boxes represent the 25th and 75th percentiles. All the outliers beyond the 10th and 90th percentiles are shown individually.

Figure 2. Simplified illustration of the relationship between the human population near a sampling site (pop) and the concentration of a pollutant measured at that site (C) using eq 1 with β′ = 0.06 and arbitrary population data.

cides.24,28 These overall spatial trends indicate that these tribromophenoxy compounds, like other flame retardants, have been widely used in commercial products that have reached the general public and have been gradually released back into the environment. We have previously parametrized the effect of local human population on the atmospheric concentrations (typically given as C in units of pg/m3) of various persistent organic pollutants using the functional relationship:

ln(C) = β′log 2(pop)

(1)

where pop is the number of people living and working within a 25-km radius of the sampling site29 and β′ is a proportionality constant. The squared logarithm of the population is used to give the relationship some curvature such that the concentrations at very low populations will be more similar to one another than at high populations.28 This effect is illustrated in Figure 2 for arbitrary data. Note that the concentrations at low population densities approach an asymptotic concentration, which is the result of long-range atmospheric transport. In other words, this minimum atmospheric concentration represents a more or less global background level. So that we can use linear regression statistical tools, we usually plot ln(C) versus log2(pop), and Figure 3 shows this relationship for the TBE and ABD data as presented in this paper. Both of these regressions are statistically significant, and based on the General Linear Model in Minitab 16, the slopes are significantly different from one another (P < 0.0001). The regression lines for TBE and for ABD intersect at a population of about 250 000 people. The slope of 0.047 for TBE indicates that an increase in population from 100 000 to 1 000 000 people would cause an increase in its atmospheric concentration by a factor of 1.7. The slope of 0.088 for ABD indicates that the same increase in population would cause an increase in its atmospheric concentration by a factor of 2.6. Both of these fractional increases are significant. We have also used a simple multiple regression technique for teasing apart spatial and temporal trends in atmospheric

Figure 3. Natural logarithmically scaled average concentrations of TBE and ABD (the sum of ATE + BATE + DPTE) at the five IADN sampling sites as a function of the squared logarithm of the human population living or working within a 25 km radius of each sampling site. The error bars show standard errors. The slopes of the two lines are: TBE, 0.0475 ± 0.0054; and ABD, 0.0878 ± 0.0068. These regressions are both statistically significant (P < 0.015), and the slopes are significantly different from one another (P < 0.0001).

concentration data measured as a function of time.28,30 The approach is to fit the coefficients of ln(C) = a0 + a1log 2(pop) + a 2t

(2)

where t is time in Julian days relative to the reference date of 1 January 2008, a0 is an intercept that rectifies the units, a1 is a unitless parameter indicating the importance of population on the concentrations, and a2 is a first-order rate constant (in days−1). Previously, we have also included seasonal parameters in this regression, but for the data reported here, no such variations were observed. The regression parameters for TBE and ABD are given in Table 2 along with their errors. As suggested by Figure 3, the urbanization regression coefficients, a1, are positive and statistically significant both for TBE and ABD; these coefficients are also significantly different from one another. This result indicates that the 13115

dx.doi.org/10.1021/es3033814 | Environ. Sci. Technol. 2012, 46, 13112−13117

Environmental Science & Technology

■

Table 2. Spatiotemporal Regression Coefficients for TBE and ABD (= ATE + BATE + DPTE) along with Statistical Errors and P-Valuesa variable

intercept population time

a0 a1 a2

value

−2.27 ± 0.20 0.0480 ± 0.0054 −0.00072 ± 0.00031 ABD = ATE + BATE + DPTE

P

SOS %