Environ. Sci. Technol. 2002, 36, 146-151
Concentrations and Spatial Variations of Polybrominated Diphenyl Ethers and Several Organochlorine Compounds in Fishes from the Northeastern United States NATHAN G. DODDER, BO STRANDBERG, AND RONALD A. HITES* Environmental Science Research Center, School of Public and Environmental Affairs, and Department of Chemistry, Indiana University, Indiana 47405
Fish from four lakes, two small lakes in the northeastern United States and two of the Great Lakes, were analyzed to determine the concentrations and spatial variations of polybrominated diphenyl ethers (PBDE) in this region. Three of the lakes were considered to have background levels of PBDEs; one lake was close to a suspected source. The PBDE concentrations were compared to organochlorine pesticide and PCB concentrations. Age and trophic position did not influence the organohalogen concentrations in the fish collected from the four locations. At the three background locations, the ΣPBDE concentrations ranged from 6.9 ( 1.4 to 18 ( 1 ng/g wet weight, or 150 ( 9 to 300 ( 80 ng/g lipid, and these values were similar to those of some of the organochlorine pesticides, such as total chlordane, but lower than ΣPCB concentrations. At the lake near the suspected source, the ΣPBDE concentration was 65 ( 8 ng/g wet weight, or 2400 ( 600 ng/g lipid, and it exceeded the ΣPCB concentration. The hexabrominated congeners made up 43% of the total mass of PBDEs at this lake, and 7% at the background locations. Fish to sediment concentration ratios indicated that the tetra- through hexasubstituted congeners have a similar bioavailability, while the deca-substituted congener does not seem to be bioavailable at all.
Introduction Polybrominated diphenyl ethers (PBDE) are flame-retardants that are added to a variety of materials, including synthetic polymers and textiles. There are three technical PBDE products, each with varying degrees of bromination and each with different physical-chemical properties. The technical products are commonly referred to as penta-BDE, octa-BDE, and deca-BDE, and each consists of a relatively limited number of congeners. [The IUPAC numbering scheme used for polychlorinated biphenyl (PCB) congeners is also used for PBDEs.] One penta-BDE technical product (DE-71) consists of approximately 43% 2,2′,4,4′,5-pentabromodiphenyl ether (BDE-99) and 27% 2,2′,4,4′-tetrabromodiphenyl ether (BDE-47) (1). The remaining five congeners are penta* Corresponding author e-mail:
[email protected]. 146
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 2, 2002
and hexa-substituted and include 2,2′,4,4′,5,5′-hexabromodiphenyl ether (BDE-153) and 2,2′,4,4′5,6′-hexabromodiphenyl ether (BDE-154). Octa-BDE (DE-79) contains nine hexa- through nona-substituted congeners, primarily BDE183 at 44% and an unidentified octa-BDE at 21%. Deca-BDE (DE-83) contains almost exclusively 2,2′,3,3′,4,4′,5,5′,6,6′decabromodiphenyl ether (BDE-209). PBDEs are often added at 5-30 wt % to a polymer; for example, penta-BDE is added to polyurethane foam used in furniture. These compounds are noncovalently incorporated into the polymer, and they can migrate out of the treated product. In 1999, deca-BDE represented 72% of the U.S. sales of PBDE products, octaBDE represented 4%, and penta-BDE represented 24% (2). Total U.S. sales of these three products were estimated at 34 000 metric tones (2). Tetra- and penta-substituted PBDEs have been shown to be persistent, lipophilic, bioaccumulating, and ubiquitous in the environment (3-5), and they have been found in a variety of media, including humans (6). The congener profile commonly found in organisms and in air often resembles that of the penta-BDE technical product, except for elevated amounts of BDE-47 relative to BDE-99. Studies of trout from Lake Ontario (7), beluga blubber from the Cumberland Sound (8), sediment off the coast of Sweden (9), and Swedish human milk (10) have demonstrated that the environmental concentrations of PBDE have been increasing since the mid 1970s. The toxicology of PBDEs has been reviewed (3, 11, 12), but these studies are not extensive. Current data suggest that the less brominated congeners are more toxic than the higher brominated ones (12); this may be a result of the limited bioavailability of the more highly brominated congeners. The European Parliament has voted to ban the pentaBDE technical product (13), but there are currently no regulations on PBDE production or use in the United States. Relatively little data exist on the occurrence and spatial distribution of PBDEs in the United States. PBDEs have been found in carp from the Buffalo River (14), trout from Lakes Ontario, Huron, and Superior (15), fish from waters in Virginia and sewage sludge (16), in steelhead trout from Lake Michigan (17), and in salmonids from Lake Michigan (18). PBDEs have also been found in air along the coast of the Great Lakes (19). The aim of this paper is to determine the concentrations and spatial distributions of PBDEs (relative to organochlorine pesticides and PCBs) in fish collected near a suspected source and from more remote locations.
Experimental Section Sample Collection and Preparation. Twelve white crappie (Pomoxis annularis) and 12 bluegill (Lepomis macrochirus) were collected from Hadley Lake (40.47° N; 86.93° W) in 1999 (see Figure 1). Hadley Lake was chosen because of its proximity to a suspected PBDE manufacturing facility about 1.3 km away. Air deposition is the only known method for PBDEs to enter Hadley Lake. Three white crappie and three bluegill were collected from the Lake of the Ozarks (38.20° N; 92.62° W) the same year. The Lake of the Ozarks was chosen because it was relatively remote from any known point sources of PBDE. Two carp (Cyprinus carpio) were collected from Hadley Lake only. Fish from Hadley Lake were collected using electrofishing equipment (Coffelt Manufacturing, Flagstaff, AZ). The traditional rod and reel method was used to collect the six fish from the Lake of the Ozarks. The collection, homogenization, and transport of smelt (Osmerus mordax) taken in 1994 from the Great Lakes have been described previously (20). The Great Lakes sites, the Apostle Islands (46.93° N; 90.65° W) at Lake Superior and Oswego 10.1021/es010947g CCC: $22.00
2002 American Chemical Society Published on Web 12/12/2001
FIGURE 1. Fish sampling sites in the Great Lakes and in north central United States. (43.45° N; 76.50° W) at Lake Ontario (see Figure 1), are located at offshore fishing grounds that are not impacted by local sources (20). Three composite samples were used from each location. All fishes were collected in the months of July through October. Following collection, the fish from Hadley Lake and from the Lake of the Ozarks were wrapped in aluminum foil, stored at 0 °C, and shipped to Bloomington, IN. The fish were then washed with water, individually wrapped in aluminum foil, sealed in plastic bags, and stored at -30 °C until analysis. Prior to extraction, the fish were weighed and measured for body length. The age of each fish was determined by counting its scale annuli (21). The white crappie and bluegill from Hadley Lake were composited (four whole fish per sample) to give three composite samples of each species. To reduce the effect of age and size variation on the composite samples, the fish were selected so that the average length of fish in each composite was similar. In the case of the carp from Hadley Lake, muscle tissue was sampled instead of the whole fish. The two carp from Hadley Lake and the six white crappie and bluegill from the Lake of the Ozarks were individually homogenized because of the small number of samples. All homogenized samples were transferred to solvent washed glass jars with foil-lined screw caps and stored at -30 °C until analysis. Extraction Procedure. A 20 g portion of the homogenized fish sample was ground with 130 g of sodium sulfate; loaded into a glass wool plugged glass thimble; placed in a Soxhlet extractor; spiked with the internal standards [13C12-2,3,3′,4,4′,5hexachlorodiphenyl ether (CDE-156), 13C12-2,2′,3,3′,4,4′,5,5′octachlorodiphenyl ether (CDE-194), 13C6-γ-hexachlorocyclohexane (HCH), and 13C12-2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB-153)]; and extracted for 24 h with 50% acetone in hexane. All of the internal standards and authentic compounds were from Cambridge Isotope Laboratories (Cambridge, MA) or from Dr. Ehrenstorfer (Augsburg, Germany). Throughout the extraction and analysis procedure, the analytes were protected from degradation by light by using amber glassware or by wrapping the containers in aluminum foil. The extracts were reduced in volume and centrifuged to remove any water that may have been coextracted with the lipids. The lipid concentration was determined by gravimetric measurement. The majority of the lipids were removed using a gel permeation chromatography system consisting of a 100-cm × 2.5-cm i.d. glass column with SX-8 “Bio-Beads” (Bio-Rad Laboratories, Her-
cules, CA). The eluent was 60% cyclohexane in dichloromethane. Two fractions were collected: the first of 220 mL and the second of 380 mL. The PBDEs were in the second fraction. This sample was then evaporated to about 2 mL and cleaned up on a 20-cm × 1.9-cm i.d. 3% water deactivated silica column. One fraction, eluted with 150 mL of dichloromethane,wascollected.Theperformancestandard,2,2′,3,4,4′,5,6,6′octachlorobiphenyl (PCB-204), was added, and the samples were further reduced in volume using a stream of nitrogen to about 50 µL and analyzed by gas chromatographic mass spectrometry. Instrumental Analysis and Parameters. The fish were analyzed for PBDE congener numbers 47, 99, 100 (2,2′,4,4′,6penta BDE), 153, 154, 190 (2,3,3′,4,4′,5,6-hepta BDE), and 209. These compounds, except for BDE-190 and BDE-209, are the most abundant congeners in the penta-BDE technical mixture. The internal standard 13C12-CDE-156 was used to quantitate BDE-47, and the internal standard 13C12-CDE-194 was used to quantitate the other PBDEs. The standard for BDE-209 was made using the deca-BDE technical product, DE-83. This technical product is more than 98% BDE-209. The fish were also analyzed for total polychlorinated biphenyls (ΣPCB), total hexachlorocyclohexanes (ΣHCH), total chlordanes (ΣCHL), hexachlorobenzene (HCB), and dieldrin. The ΣPCB concentrations are the sum of the concentrations of 85 tri- through decachlorinated PCB congeners. The ΣHCH concentrations are the sum of the concentrations of R- and γ-HCH. The ΣCHL concentrations are the sum of the concentrations of cis- and trans-chlordane, trans-nonachlor, cis-heptachlorepoxide, and oxychlordane. The samples were analyzed for PBDEs on a HewlettPackard 6890 series gas chromatograph coupled to a HP 5973 mass spectrometer with helium as the carrier gas. The 1 µL injections were made in the pulse splitless mode, with a purge time of 2.0 min. The injection port was kept at 285 °C. This instrument was operated in the electron impact (EI) or in the electron capture negative ionization (ECNI) mode, with selected-ion monitoring (SIM). ECNI was used to quantitate decabromodiphenyl ether (BDE-209); all other PBDEs were analyzed by EI. The GC column used for the tetra- through heptabrominated PBDEs was a 30-m × 250-µm i.d. fused silica capillary tube coated with DB-5-MS (0.25-µm film thickness; J&W Scientific, Folsom, CA). The temperature of the GC oven was programmed as follows: isothermal at 110 °C for 1 min, 15 °C/min to 180 °C, 2 °C/min to 270 °C, 0.50 °C/min to 275 °C, and held at 275 °C for 10 min. BDE-209 was analyzed on a shorter DB-5-MS column (15-m × 250µm i.d.; 0.25-µm film thickness; J&W Scientific) with the following temperature program: 110 °C for 1 min, 15 °C/ min to 300 °C, and held at 300 °C for 17 min. The following ions (m/z) were used to detect PBDE compounds in the EI mode: tetra-Br, 325.9 (M - Br2 + 2) and 485.7 (M + 4); pentaBr, 403.8 (M - Br2 + 2), and 405.8 (M - Br2 + 4); hexa-Br, 481.7 (M - Br2 + 2) and 483.7 (M - Br2 + 4); and hepta-Br, 561.6 (M - Br2 + 4) and 563.6 (M - Br2 + 6). In the ECNI mode, the two bromide ions at m/z 79 and 81 were used for maximum sensitivity. In each case, the first ion was used for quantitation and the second for confirmation. The PCBs and pesticides were analyzed on a DB-5-MS 30-m or 60-m column (both 250-µm i.d., 0.25-µm film thickness; J&W Scientific), with the mass spectrometer operated in ECNI or EI mode, depending on which mode gave the highest sensitivity for a given compound class. The splitless injection port was held at 285 °C, and helium was the carrier gas. The temperature program for the 30-m column was as follows: isothermal at 40 °C for 1 min, 30 °C/min to 130 °C, 3 °C/min to 241 °C, 30 °C/min to 285 °C, and held at 285 °C for 25 min. The temperature program for the 60-m column was the same as for the 30-m except it was held at 285 °C for 20 min, ramped at 30 °C/min to 300 °C, and held VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
147
at 300 °C for 20 min. The two most abundant ions of the molecular ion isotopic clusters were used as the SIM descriptors for the 13C-labeled and native PCBs and pesticides (22). The compounds, except for the PCBs, were quantitated using an external quantitation standard with known amounts of all the target compounds, internal standards, and recovery standards. The individual PCB congeners were quantitated using the relative response factor from a native PCB congener from the same or similar homologue group as described previously (22). Four quality control criteria were used to ensure the correct identification of the target compounds in the tissue samples: (a) The signal-to-noise ratio was greater than three. (b) The GC retention times matched (( 0.05 min) those of the standard compounds. (c) The internal standard recoveries were between 70% and 120%. (d) The isotopic ratio between the quantitative and confirmation ions was within ( 15% of the theoretical value. Sediment Collection and Analysis. Four grab samples of sediment were collected from Hadley Lake in November 2000. At this time, one field blank was collected by exposing an empty jar to the atmosphere for 15 min. The top five centimeters of sediment were stored in solvent washed glass jars, shipped to Bloomington, IN, and stored at -30 °C until analysis. The dry weight was determined gravimetrically after heating an aliquot of the sample at 75 °C for 48 h. The loss on ignition was determined after heating the dried sample at 525 °C for 24 h. A 15-g portion of each wet sediment sample was ground with 100 g of Na2SO4, spiked with the internal standards, and Soxhlet extracted with 50% hexane in acetone for 24 h. The samples were then exchanged into hexane, reduced in volume, and fractionated on a 20-cm × 1.9-cm i.d. 1% water deactivated silica column. Three 75-mL fractions were collected with the following solvents: hexane, 30% hexane in dichloromethane, and dichloromethane. The second fraction was reduced in volume to 50 µL under a steady stream of nitrogen and analyzed for PBDEs by gas chromatographic mass spectrometry as described above. These sediment samples were not analyzed for PCBs or pesticides. Quality Control. A procedural blank was run in parallel with every batch of four samples. The procedural blank consisted of the appropriate amount of Na2SO4 spiked with the internal standards. The amounts of PBDEs and organochlorine pesticides in the laboratory and field blanks were well under 10% of the levels found in the fish tissue and sediment. The background levels of PCBs in our building are high (23). Therefore, if a PCB congener level found in the blank was under 10% of that in the tissue, no correction was made. If the blank PCB level was 10-30% of the tissue sample, the blank level was subtracted from the tissue level. If the blank PCB level was over 30%, that congener was considered unquantifiable and was excluded from the data set. Our laboratory participated in an interlaboratory study that analyzed biological and sediment samples for PBDEs. Our results were within ( 14% of the mean values obtained by the other laboratories for BDE-47, -99, -100, -153, and -154. The agreement was also good for BDE-209 analyses in sediment, but all laboratories showed wide variability for BDE-209 analyses in biota. This may have been caused by the degradation of BDE-209 by light (24), or BDE-209 may not have been efficiently separated from the lipids during gel permeation chromatography. Experiments in our laboratory showed that only about 20% of the BDE-209 could be recovered from spiked fish samples. Because of this low recovery and the lack of suitable internal standards for the analysis of this compound, the results for BDE-209 in fish in this study should be considered as estimates. Nitrogen Isotope Measurements. The trophic position of each fish sample (except the carp) was found by measuring 148
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 2, 2002
the nitrogen isotope ratio of the biomass. A 1-g aliquot of each of the homogenized whole fish samples was freezedried and finely ground in a Wiley-mill. Approximately 11 mg of each sample was placed in a Vycor tube, along with 1 g of copper oxide, 0.5 g of copper, and a 1-cm length of silver wire. The tubes were evacuated and heated at 750 °C for 12 h. Passing the gas through liquid nitrogen cooled U-tubes cryogenically purified the elemental nitrogen that resulted from the combustion. Nitrogen gas also passed through a zone filled with elemental copper turnings held at 350 °C to remove traces of oxygen and nitrogen oxides. The stable nitrogen isotope ratios were analyzed on a magnetic sector, dual inlet, isotope ratio mass spectrometer (Finnigan MAT 252, Bremen, Germany) at the Indiana University Biogeochemistry Laboratories. The isotope ratio was standardized against air according to
δ15N )
[
]
Rsample - 1 × 1000 Rair
(1)
where R is the isotopic ratio of 15N to 14N. The δ15N values (in ‰) were calibrated against two NIST standards. The precision of the analytical method and instrumentation was determined by running three samples taken from the same ground fish sample. The relative standard deviation of these three samples was 2.3%.
Results and Discussion BDE-47, -99, -100, -153, and -154 were found in all fish samples. BDE-190 was not found in any of the samples, and it is not known to be in any of the technical mixtures. BDE209 was also not observed in the fish samples, except for a trace amount in one smelt sample, even though this congener is produced in the greatest amount. An analysis of variance between fish species did not show a statistically significant difference in ΣPBDE concentrations (p ) 0.095) at Hadley Lake or the Lake of the Ozarks; therefore, the concentration averages discussed below for these locations include both species. Table 1 lists the average concentrations of each PBDE congener in the fish at the four sampling locations. The ΣPBDE concentration in the crappies and bluegills from Hadley Lake (65 ( 8 ng/g wet weight or 2400 ( 600 ng/g lipid) was higher than at the other three locations (6.9 ( 1.4 to 18 ( 1 ng/g wet weight or 150 ( 9 to 300 ( 80 ng/g lipid). Figure 2 shows the organohalogen compound profiles in the fish (except the carp). ΣPBDE was the third most abundant pollutant class, after ΣPCB and ΣCHL, for fish from the Lake of the Ozarks, Lake Superior, and Lake Ontario. In these three cases, the ΣPBDE concentration was similar to that of dieldrin and greater than that of HCB and ΣHCH. The ΣPBDE concentration was within a factor of 2 of the ΣCHL concentration in fish from these three lakes, and the ΣPCB concentration was 5-13 times higher than the ΣPBDE concentration. In Hadley Lake, located near a suspected source, the concentration of ΣPBDE exceeded that of ΣPCB (65 ( 8 ng/g wet weight vs 45 ( 5 ng/g wet weight). The concentrations of ΣPCB, ΣCHL, and dieldrin at Hadley Lake were similar to the concentrations of these compounds at the other three lakes, suggesting that these compounds have become relatively evenly distributed throughout the environment since they were banned. The concentrations of ΣHCH and HCB varied dramatically among the lakes, with the Great Lakes having the highest concentrations of these compounds. The relatively high ΣPBDE concentration in fish from Hadley Lake suggests a local source of PBDEs; however, even in lakes distant from any source, the ΣPBDE concentration is as high as those of some other banned, environmental contaminants. There are a few reports of PBDEs in fish from the Great Lakes. Trout were reported to have a ΣPBDE (sum of di-
TABLE 1. Species, Age, Average Mass, Length, δ15N, % Lipid, and Concentrations (ng/g Wet Weight), of PBDE in Fish from Hadley Lake, the Lake of the Ozarks, Lake Superior, and Lake Ontarioe Hadley Lake speciesa num. samples age (yr) fish weight (g) fish length (cm) δ15N (‰) % lipid BDE-47 BDE-99 BDE-100 BDE-153 BDE-154 BDE-190b BDE-209b ΣPBDE (ng/g wet weight)d ΣPBDE (ng/g lipid)d
WC, B 6 3-4 69.6 ( 4.0 13.0 ( 0.3 15.2 ( 0.1 3.2 ( 0.4 13 ( 2 16 ( 2 7.4 ( 1.0 15 ( 2 13 ( 1
