Spatial Distribution of Polybrominated Diphenyl Ethers in Southern

Organic Composition, Chemistry, and Photochemistry of Urban Film in Leipzig, ...... ethers-47 in male adult rat is modulated by intracellular glutamat...
0 downloads 0 Views 139KB Size
Environ. Sci. Technol. 2004, 38, 724-731

Spatial Distribution of Polybrominated Diphenyl Ethers in Southern Ontario As Measured in Indoor and Outdoor Window Organic Films CRAIG M. BUTT, MIRIAM L. DIAMOND,* AND JENNIFER TRUONG Department of Geography, University of Toronto, Toronto, Ontario, Canada M5S 3G3 MICHAEL G. IKONOMOU Contaminants Science Section, Institute of Ocean Sciences, Department of Fisheries and Oceans Canada, Sidney, British Columbia, Canada V8L 4B2 ARNOUT F. H. TER SCHURE Chemical Ecology and Ecotoxicology, Department of Ecology, Lund University, Ecology Building, S-223 62 Lund, Sweden

Organic films were collected from indoor and outdoor window surfaces, along an urban-rural transect extending northward from Toronto, Ontario, Canada, and analyzed for 41 polybrominated diphenyl ether congeners (PBDE). For exterior films, urban ΣPBDE concentrations were ∼10× greater than rural concentrations, indicating an urbanrural gradient and greater PBDE sources in urban areas. Urban films ranged from 2.5 to 14.5 ng/m2 (mean ) 9.0 ng/ m2), excluding the regional “hotspot” Electronics Recycling Facility, compared to 1.1 and 0.56 ng/m2 at the Suburban and Rural sites. Interior urban films (mean ) 34.4 ng/m2) were 3 times greater than rural films (10.3 ng/m2) and were representative of variations in building characteristics. Indoor films were 1.5-20 times greater than outdoor films, consistent with indoor sources of PBDEs and enhanced degradation in outdoor films. Congener profiles were dominated by BDE-209 (51.1%), consistent with deca-BDE as the main source mixture, followed by congeners from the penta-BDE mixture (BDE-99:13.6% and -47:9.4%) and some octa-BDE (BDE-183:1.5%). Congener patterns suggest a degradative loss of lower brominated compounds in outdoor films versus indoor films. Gas-phase air concentrations were back-calculated from film concentrations using the film-air partition coefficient (KFA). Mean calculated air concentrations were 4.8 pg/m3 for outdoor and 42.1 pg/m3 for indoor urban sites, indicating that urban indoor air is a source of PBDEs to urban outdoor air and the outdoor regional environment.

Introduction Polybrominated diphenyl ethers (PBDEs) are a class of halogenated flame-retardants commonly added to plastics, polyurethane foams, and textiles. End products for PBDEs * Corresponding author phone: 416-978-1586; fax: 416-946-5992; e-mail: [email protected]. 724

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 3, 2004

include electronics, computers, and foam padding in furniture. PBDEs are incorporated into materials using an “additive” process and thus may slowly volatilize into the environment during the product’s lifetime (1). Recent studies have shown that PBDEs are globally distributed and detected in most environmental media including air (2-4), water (5), soil (6), sediment (7, 8), vegetation (3), and biota (5, 9, 10). The presence of PBDEs in higher-order Arctic marine mammals (9) and the congener fractionation observed in frogs along a latitudinal gradient in Sweden (10) indicates that these compounds undergo long-range transport (LRT) and bioaccumulation. Temporal studies have shown that PBDEs are increasing in the environment (9, 11) and human breast milk at an exponential rate (12, 13), which is consistent with the increase in global demand since the early 1980s. While the European Union has restricted the use of some PBDEs (14), demand in North America continues to increase. Global demand for PBDEs in 1999 reached 67 125 metric tons of which the deca-BDE mixture comprised 81.6%, followed by 12.7% and 5.7% for penta-BDE and octa-BDE mixtures, respectively (15). Urban environments are characterized by numerous PBDE sources relative to rural areas. Consequently, elevated concentrations of PBDEs are measured near urban areas (2, 11). Urban areas are likely sources of PBDEs to surrounding areas as seen with other semivolatile organic compounds (SOCs) such as PAH and PCBs (16, 17). While separate measurements have been made in both urban and rural sites, few studies have examined PBDE concentrations along an urban-rural transect (4). It is now appreciated that indoor air is a significant route of exposure for environmental contaminants because humans spend over 80% of time indoors and air concentrations of SOCs are frequently several times higher indoors than outdoor (18, 19). Also, contaminants in the indoor environment are less prone to degradation and atmospheric dilution, thereby increasing their persistence. The indoor environment contains many PBDE sources, resulting in elevated concentrations (1). However, few ambient measurements of PBDEs in indoor air have been reported (1). As a result of higher concentrations, indoor air likely represents a significant source of SOCs to outdoor air (20, 21). Organic films have been shown to form on impervious building surfaces in both urban and rural areas (22, 23). The composition of the organic film is representative of ambient air quality since the film is hypothesized to form through the condensation of primary gas-phase species and secondary organic aerosols. This condensation forms a “sticky” layer and is followed by the wet and dry deposition of particulateassociated compounds (24). Further, gas-phase compounds are capable of partitioning into the film dependent upon the relative fugacity between the air and film. With knowledge of the sampling rate and film-air partition coefficient (KFA), it is possible to quantitatively determine ambient air concentrations (25). Thus, the organic carbon reservoir in window films can be used as a time-integrated passive sampler for gas-phase air concentrations. The octanol-air partition coefficient (KOA) has often been used to describe partitioning between atmospheric gas-phase species and environmental surfaces (26). KFA has been observed to be approximately equal to the KOA for compounds whose log KOA < 11 (Butt et al., manuscript in preparation), as described by the following equation where foc is equal to the fraction of organic carbon in the film

KFA ) focKOA 10.1021/es034670r CCC: $27.50

(1)

 2004 American Chemical Society Published on Web 12/17/2003

FIGURE 1. Regional southern Ontario and downtown Toronto window film sampling locations. Site abbreviations are as follows: “RU”, Rural; “SU”, Suburban; “U/LT”, Urban/Lt. Industrial; “UW”, Urban West; “RS”, Residential; “UT”, University-Laboratory and UniversityBookstore; “UE”, Urban East. Electronics Recycling Facility not labeled for purposes of anonymity. Average gas-phase air concentrations may be back-calculated using derived film-air (KFA) partition coefficients, and thus the organic film can be used to assess contaminant deposition and can act as a surrogate of ambient air quality. This paper presents PBDE concentrations and congener profiles in organic films from interior and exterior window surfaces, collected along an urban-rural transect. Five urban sites were chosen to assess spatial variability within the urban core. In addition, samples were collected at an electronics recycling facility, a site hypothesized to exhibit elevated PBDE levels. Finally, average gas-phase air concentrations were calculated using derived film-air partition coefficients (KFA).

Experimental Section Collection of Organic Films. Organic films were collected from exterior and interior window surfaces using previously described methods (22). Briefly, films were sampled by scrubbing window surfaces with precleaned laboratory Kimwipes, initially wetted with HPLC-grade organic solvent. Exterior films were collected with dichloromethane (DCM), whereas isopropyl alcohol (IPA) was used for interior films. Previous work demonstrated no difference in film collection efficiency between the two solvents (Tieu, unpublished data). Contamination from building materials was avoided by leaving a 10 cm border around the outside of the window area. The time of last window cleaning was unknown but was longer than 4 months at all sites. Samples were collected from late July to early August, 2001, at 9 sites (14 samples in total, 9 exterior, 5 interior)

situated along an urban-rural transect extending from downtown Toronto, Ontario, approximately 80 km northward (Figure 1). Exterior window films were collected at the following seven sites in downtown Toronto: two buildings at the University of Toronto campus (“University-Bookstore”, “University-Laboratory”), a community health center (“Urban East”), an automobile sales office (“Urban West”), a composite of four homes near downtown (“Residential”), the Meteorological Service of Canada’s Downsview headquarters (“Urban/Lt. Industrial”), and a computer/electronic recycling facility (“Electronics Recycling Facility”). In addition, exterior films were collected at two sites outside downtown Toronto: a community center in Richmond Hill (“Suburban”) and Environment Canada’s Centre for Atmospheric Research Experiments (CARE) in Egbert (“Rural”). Interior window films were collected from the following sites: University-Bookstore, Residential, Urban/Lt. Industrial, Electronics Recycling Facility, and Rural. The mean outside air temperature during the 2-week sampling period was ∼23 °C. Indoor air temperatures were not recorded but are expected to be slightly cooler as all buildings were air conditioned, with the exception of the Residential site. Sample Preparation and Analysis. Sampled Kimwipes were kept frozen until extraction with Soxhlet apparatus for approximately 18 h using a 80:20 mixture of toluene:acetone. Extracts were acid-base washed with H2SO4 and KOH and returned to neutral pH by rinsing with HPLC-grade water, reduced to dryness, and reconstituted in 1:1 DCM:hexane. Sample cleanup took place sequentially using three columns. VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

725

First, the sample was passed through a multilayered acidic/ basic silica column, eluted with 1:1 DCM:hexane, blown down to dryness, and reconstituted in hexane. The sample was then passed through a copper column and eluted with hexane for the removal of sulfur impurities, reduced in volume, and applied directly to an alumina column for fractionation. The fraction containing the PBDEs was eluted with 1:1 DCM: hexane, blown down to dryness, and reconstituted in toluene. All sample extracts were analyzed using GC-HRMS operated in the positive electron impact ionization mode and acquiring data under selected ion monitoring (SIM) conditions at 10 000 resolution. Complete details of the sample extraction, cleanup, surrogate standards used, the instrumental analysis conditions, and the criteria used for congener identification and quantification are presented elsewhere (27). A total of 41 PBDE congeners were quantified along with several unidentified congeners. Total PBDE concentrations presented below refers to the sum of identified and unidentified congeners. Unidentified congeners comprised 1-8.1% of the total sample concentration. Quality Control and Assurance. Prior to extraction, samples were spiked with internal recovery standards (13C-BDE28, 13C-BDE47, 13C-BDE99, 13C-BDE154, 13C-BDE183, 13 C-BDE209). Recoveries ranged from 35% to 119%, within the allowable limits (34). Reported concentrations presented below were corrected for recovery of the internal standards. Laboratory blanks (precleaned Kimwipes exposed only to indoor laboratory air) were extracted and analyzed prior to field blank and sample extraction. Laboratory blank levels were low and within the range of field blank values. Field blanks were collected at 3 sites by waving 10 solvent-wetted Kimwipes in the air until dry, about 15 s per Kimwipe. Method detection limits (MDL) were calculated as the mean of three field blanks plus three times the standard deviation. With the exception of the outdoor samples from the Rural and Suburban sites, MDLs were generally less than 5% of the sample concentration. Reported concentrations have been adjusted by subtracting the MDL from the gross congener value. Data Analysis and Treatment. In calculating congener profiles only, values less than the MDL were replaced with one-half the detection limit to facilitate the computation of geometric means. Specific differences in congener profiles were examined using the statistical technique of principal component analysis (PCA) using SPSS version 11.0 (Chicago, IL). PCA has been widely used to identify differences and similarities between environmental samples (5, 28, 29) and is particularly useful when a large number of variables are present in the data set. PCA reduces the data set into a new set of uncorrelated variables (principal components) such that each successive component contains decreasing proportions of the total variance. Differences between samples can be ascertained through visual observation when two principal components are projected on a two-dimensional plot. Compounds were removed from the data set, for the purposes of PCA only, if more than 30% of the samples had concentrations below the MDL. Congener profiles were selectively normalized as described by Johansson et al. (30). For comparison, a particulate deposition profile was calculated from 16 out of the 27 samples originally reported by ter Schure and Larsson (31), with only those samples that had detectable quantities of BDE-209 included in the analysis. It was noted that samples with undetectable amounts of BDE209 also had low ΣPBDE concentrations; in many cases ΣPBDE concentrations were near the relatively high detection limit of BDE-209. Thus, these sample profiles were suspected to be biased and removed from the data set for profile calculation and statistical analysis. The indoor and outdoor organic window film profiles that were statistically compared to the particulate deposition profile were calculated as the 726

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 3, 2004

arithmetic mean of the nine congeners analyzed in the particulate deposition (BDE-28, -47, -66, -99, -100, -153, -154, -183, -209). Statistical comparison was done by a three-way multiple analysis of variance (MANOVA) using SPSS. Using the measured PBDE film concentrations, the average gas-phase air concentrations were calculated. Film concentrations were converted to mass per unit volume assuming a film thickness of 60 nm for exterior films and 30 nm for interior (Bahavar, unpublished data). The fraction of organic carbon was assumed to be 10% (Bahavar, unpublished data). Literature KOA values (32) were temperature corrected using the mean air temperature (23 °C) during the sampling period. Air concentrations were calculated for the 12 congeners in which both film concentrations and KOA values were available.

Results and Discussion Spatial Trends in PBDE Concentrations. A spatial gradient in total PBDE concentration (ΣPBDE) was observed with urban film concentrations greater than rural film concentrations in both exterior and interior films (Figure 2). Considering exterior window films, ΣPBDE ranged from 2.5 to 14.5 ng/m2 (n ) 6, geometric mean ) 9.0 ng/m2) at urban sites, excluding the Electronics Recycling Facility, compared to 1.1 and 0.56 ng/m2 at the Suburban and Rural sites, respectively. Expressed on an aerial basis, ΣPBDE concentrations varied by ∼10-fold between urban and rural films. An order of magnitude difference between the same urban and rural sites has also been observed in PCBs and PAHs in surface organic films (22) and soil (33). The spatial variation observed is probably due to the increased sources within the urban environment that results in increased air concentrations. Several studies have demonstrated elevated urban air concentrations, for example, total PBDE air concentrations from Chicago were ∼50 pg/m3 compared to 5-15 pg/m3 for background sites from around the Great Lakes (2). Further, during summer 2000, Harner et al. (4) observed a 2-fold gradient in gas-phase air concentrations determined from passive samplers deployed along the same urban-rural transect as the present study. Elevated PBDE levels have also been found in biota collected near urban areas, which can be related to elevated emissions that result in increased dietary concentrations. Norstrom et al. (11) found the greatest PBDE levels in herring gull eggs near major urban areas in the Great Lakes. Frogs from the city center of Umeå, Sweden, were also found to contain elevated PBDE levels as compared to more pristine areas of Sweden (10). ΣPBDE in interior films ranged from 19.4 to 75.9 ng/m2 (n ) 3, geometric mean ) 34.4 ng/m2) at the urban sites, again excluding the Electronics Recycling Facility, compared to 10.3 ng/m2 at the rural site. Thus, interior urban films were about 3 times greater than that from the rural location, a lesser gradient than exterior films. Variations in interior films were consistent with variations in building characteristics, such as the number of potential PBDE sources and building ventilation. For example, the Residential site had few PBDE sources compared to the Urban/Lt. Industrial site, a location that is a large government research facility containing numerous computers and foam furniture. Since the building ventilation systems may be an effective mechanism of contaminant dispersion in the indoor environment, indoor concentrations are likely representative of the entire building rather than the individual rooms in which the films were collected. Detailed comparison of the sites is not possible given the limited number of samples and the fact that no replicates were collected. Total PBDE concentrations measured at the Electronics Recycling Facility site were 38.7 and 755 ng/m2 for exterior and interior films, respectively. These concentrations were ∼4.4 and ∼22 times above ambient Toronto films, confirming

FIGURE 2. Total PBDE concentrations (ng/m2) from interior and exterior window films in southern Ontario.

TABLE 1. Congener Profiles of Indoor and Outdoor Window Organic Films As Compared to Atmospheric Particulate Deposition and Major PBDE Commercial Mixturesa window filmb

BDE-15 BDE-28/33 BDE-75 BDE-49 BDE-47 BDE-66 BDE-100 BDE-119 BDE-99 BDE-155 BDE-154 BDE-153 BDE-183 BDE-209

outdoor

indoor

0.0 0.3 0.0 0.6 9.2 0.5 2.3 0.1 13.1 0.1 0.8 1.2 0.9 58.2

0.0 0.3 0.0 0.7 11.5 0.6 2.7 0.0 17.7 0.1 1.4 3.9 4.2 47.9

particulate depositionc

n/a 1.4 n/a n/a 8.8 0.5 2.4 n/a 15.3 n/a 1.5 3.2 5.6 61.4

penta-BDEd

octa-BDEd

DE-71

70-5DE

DE-7 9

79-8DE

0.0 0.2 0.0 0.5 30.8 0.4 8.8 0.0 48.1 0.2 4.4 6.6 n/a n/a

0.0 0.3 0.0 0.3 33.6 0.2 8.4 0.0 51.1 0.0 2.3 3.8 n/a n/a

0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.3 0.8 0.0 13.9 84.5 n/a n/a

0.0 0.0 0.0 0.0 0.0 0.0 0.0 11.8 21.8 0.0 37.1 29.3 n/a n/a

deca-BDEe

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 97.0

a n/a denotes congener not analyzed. b Window film calculated as geometric mean of indoor (n ) 5) and outdoor (n ) 9) film samples, normalized to sum of 14 congeners presented here. Note: outdoor and indoor profiles do not sum to 100% since geometric means were used in the calculation. c Particulate deposition calculated as arithmetic mean of 16 out of 27 samples originally reported by ter Schure and Larsson (31). d Bromkal/Great Lakes Chemical commercial mixture profiles from Rayne and Ikonomou (42). e “DecaBDE” constructed assuming 97% composition of BDE-209 (40).

that the Electronics Recycling Facility was a PBDE “hotspot” and thus a likely source of PBDEs to the surrounding environment. These results support those of Sjo¨din et al. (1), who reported concentrations in indoor air from an electronics recycling plant that were several orders of magnitude greater than reference indoor sites. Further, elevated levels of PBDEs have been reported in the blood serum of workers with high occupational exposure to computers and at an electronicsdismantling plant (34, 35). Indoor and Outdoor Film Comparisons. Direct comparison between interior and exterior films at five sites was possible and on a site-by-site basis; interior films were 1.520 times greater than exterior films. This is consistent with the source of PBDEs being within the indoor environment. As the contaminants are transported to the outside, their concentrations are expected to decrease through dilution from atmospheric mixing and loss processes. Further, unlike exterior films, interior window films are not subject to the degradation process of wash-off during rain events. The indoor air concentrations are significantly higher than those measured outdoors (1). The trend of higher indoor

versus outdoor air concentrations has also been observed with other SOCs, such as PCBs, PAH, and organohalogen pesticides, and particularly those contaminants that may originate indoors (18, 19, 36). Considering that humans spend 80% of their time indoors, indoor air likely represents a significant exposure route for PBDEs. In addition, our results support the hypothesis that indoor air acts as a significant source of PBDEs to outdoor air (31), in a manner similar to other SOCs (20, 21). PBDE Congener Profiles. General. Congener profiles were similar among all samples with domination by congeners that comprise the commercial PBDE mixtures (Table 1). Six congeners (BDE-209, -99, -47, -100, -153, -183) accounted for 86.5-93.3% of the total PBDE measured (Figure 3). Geometric means (and ranges) of the six main congeners, calculated as a percent of the total mass, were as follows: BDE-209, 51.1% (17.0-86.0%); BDE-99, 13.6% (4.8-27.0%); BDE-47, 9.4% (2.6-27.7%); BDE-100, 2.3% (0.8-5.0%); BDE153, 1.7% (0.2-5.9%); BDE-183, 1.5% (0.4-11.0%). While the overall PBDE pattern in window films was similar between samples, the predominant congeners were VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

727

FIGURE 3. Compositional profiles of the six main PBDE congeners from window films collected in southern Ontario. Mass fractions are normalized to the total of the six congeners presented. representative of those found in the commercial mixtures; subtle differences among samples were examined using principal component analysis (PCA). Pretreatment of the data set resulted in 18 congeners being retained for the analysis. Data pretreatment did not remove a significant proportion of the sample variability as after adjustment 99.2-100% of the original sample mass remained. Indoor versus outdoor samples were separated on the score plot by the second principal component (PC2), which accounted for 22% of the total variability (Figure 4). As indicated by the loading plot and confirmed by examination of the film profiles (Figure 3), samples in the lower right quadrant (indoor samples) have greater proportions of octaBDE-associated congeners BDE-183 and -153. In contrast, samples toward the upper left quadrant (most outdoor samples) exhibit higher proportions of deca-BDE-associated congeners and less of the lower brominated congeners (e.g., less BDE-47 and -99). The Urban/Lt. Industrial and Rural sites are government research buildings that contain many computers and electronic products, which are potential sources of octa-BDE. The Electronic Recycling Facility, by virtue of its building use, also contains many octa-BDE sources. Therefore, the differences between the indoor and outdoor film profiles, as suggested by the PCA for most sites, should be interpreted as a depletion of octa-BDE in the outdoor films rather than enrichment of deca-BDE. This analysis suggests a loss of the less brominated congeners as one moves from indoors to outdoors. As mentioned above, outdoor window films are subject to losses through washoff, volatilization, and chemical degradation (e.g., photolysis). Preliminary wash-off studies suggest that SOCs are removed independently of compound solubility (37) facilitated by the high concentration of polar compounds (38). Thus, wash-off 728

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 3, 2004

is not expected to preferentially remove the lower brominated compounds from outdoor films. Neither is volatilization expected to result in the preferential loss of lower brominated compounds in outdoor versus indoor window films, assuming that compounds in the gas- and film-phases are in equilibrium. However, rates of chemical degradation are expected to be higher outdoors than indoors, which would, in particular, deplete the lower brominated compounds in outdoor films. Reported photolytic half-lives of deca-BDE in sediment and soil are ∼100 and ∼200 h, respectively, in natural sunlight (39). Another point raised by this analysis is the relative enrichment of the more bioavailable, lower brominated congeners in indoor air which could be an important route of human exposure. It was expected that the octa-BDE and deca-BDE congeners would have clustered together on the loadings plot considering that these formulations are used for similar industrial applications, primarily electronic products. PentaBDE is used almost entirely on polyurethane foam in upholstered furniture, consistent with higher proportions of BDE-47 and -99 measured in the Residential films. OctaBDE is used exclusively in acyrlonitrile-butadiene-styrene (ABS), a resin used in electronics housings and is consistent with the positioning of the Electronics Recycling Facility on the PCA score plot and relatively high percentage of BDE183 and -153 in these films. The primary use of deca-BDE is in thermoplastics, specifically high-impact polystyrene, for electronic equipment housings; however, smaller quantities are used to flame retard upholstery textiles (40, 41). Paired indoor and outdoor samples were similar at the Electronics Recycling Facility and Residential sites, suggesting that the outdoor films were strongly influenced by indoor air. This is consistent with the elevated outdoor film concentrations

FIGURE 4. PCA score and loading plots for PBDE window film profiles collected from southern Ontario. “I” and “O” denote indoor and outdoor samples. measured at the Electronics Recycling Facility as compared to background. Technical Mixtures. Table 1 compares the mean congener profile of all samples to four Bromkal/Great Lakes Chemical commercial PBDE mixtures analyzed previously under the same conditions as the samples of our study (42). DE-71 and 70-5DE are representative penta-mixtures, whereas, DE-79 and 79-8DE are representative octa-mixtures. Although BDE209 was not analyzed in the penta- and octa-mixtures, it is generally accepted that these mixtures do not contain quantities of this congener (40). Deca-BDE was constructed assuming 97% BDE-209 (40). Comparisons between the film and technical mixtures revealed that, in general, the decaBDE mixture was the major commercial mixture contributing to the film. Penta-BDE also contributed through the abundance of moderate proportions (∼3-28%) of BDE-47 and -99. Octa-BDE may have also been a source to the film as small proportions of BDE-153 and -183 were detected, although BDE-153 is also found in trace amounts in penta-

BDE. The film profiles observed were not indicative of “pure” PBDE mixtures but were likely altered somewhat by the various degradation processes described above, such as photolysis, which may debrominate BDE-209 into lower brominated congeners (39), in addition to partitioning from the gas phase, which favors deposition of higher brominated congeners into films. The high proportion of BDE-209 in the organic film was expected since this congener comprises nearly all of the decacommercial mixture, which constituted ∼82% of the global PBDE demand in 1999 (15). In general, BDE-209 has only been measured in significant quantities adjacent to industrial sources or regional “hotspots” in air (1) and sediment samples (8). Away from “hotspots”, Strandberg et al. (2) measured low concentrations, ∼0.40 pg/m3 compared to total (gas and particulate-associated) PBDE of ∼50 pg/m3. However, high proportions of BDE-209 (arithmetic mean ) 77%) have been reported in house dust (43). The high proportions of BDE209 observed in the organic film raises the issue of environmental significance. Once captured in the film, BDE-209 is not expected to volatilize back into the atmosphere (2) but will be efficiently conveyed to urban surface waters (44) due to the high concentration of polar compounds in the film (38). Atmospheric Deposition. The indoor and outdoor film profiles were also similar to particulate atmospheric deposition from an urban area in Sweden (Table 1). All three profiles were dominated by BDE-209, followed by BDE-99 and BDE47. A three group multiple analysis of variance (MANOVA) revealed no significant difference (p > 0.05) between the indoor (n ) 5) and outdoor (n ) 9) film profiles and particulate deposition profiles (n ) 16). These results indicate that the film profiles were similar to condensed environmental media such as atmospheric particles. This is consistent with the hypothesized film formation processes that includes enhanced particulate deposition (24). Comparison of PBDE versus PCB Concentration. PBDEs are structurally similar to PCBs and thus have similar physical-chemical properties. Consequently, PBDEs and PCBs are expected to behave analogously in the environment. Because of this, environmental levels of PBDEs are frequently compared to PCBs in the literature. In the present study, ΣPBDE concentrations were greater than ΣPCB in all samples; ΣPBDE concentrations were 1.1-63 times greater than ΣPCB on a site-by-site basis (Table S1 in Supporting Information). Although some of the reported PCB levels were from films collected 1 year earlier (July 2000), this variation is not thought to be significant. With the exception of this work and that of Gouin et al. (3), reported PCB levels in the environment are greater than PBDEs and in several cases by as much as an order of magnitude. This trend has been observed in biota (9, 10, 45), human blood and adipose tissue (12, 46), and air (2, 31). Our findings confirm that trends in PBDE versus PCB levels are largely system dependent (47). The organic film trends may be explained by its relatively low residence time of hours to days, resulting from the high surface area-to-volume ratio (44). Thus, the film is postulated to respond more rapidly to contaminant releases and thus reflect present day nonpoint emissions, such as those from PBDEs, from urban and regional areas. In contrast, biota and humans have longer residence times and thus take longer to respond to changes in contaminant emissions. Our results indicate that greater quantities of PBDEs than PCBs are present in the more transient environmental compartments, such as organic film and air. This reasoning implies that PBDE concentrations in other media will gradually rise over time. Thus, while PCBs are declining or remaining steady in the environment, PBDEs are increasing at an exponential rate and may soon overtake PCB concentrations (9, 11, 47). Recently, Rayne et al. (47) VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

729

TABLE 2. Back-calculated Gas-Phase Air Concentrations (pg/m3). Geometric Mean of “Ambient” Urban Toronto Sites Excludes Electronic Recycling Facilitya outdoor

indoor

passive sampler, summer 2000b

0.67 1.30 4.20 0.09 0.00 0.22 0.90 0.02 0.00 0.01 0.03 0.01 4.8

2.50 6.25 25.21 0.70 0.04 1.30 6.45 0.15 0.00 0.13 0.45 0.40 42.1

0.18 0.45 7.56 0.22 0.01 0.94 4.29 0.15 n/a 0.22 0.32 0.26 14.9

BDE-17 BDE-28/33 BDE-47 BDE-66 BDE-77 BDE-100 BDE-99 BDE-85 BDE-126 BDE-154 BDE-153 BDE-183 ΣPBDE

a Note: Since geometric means were calculated for both mean congener concentrations and mean ΣPBDE, sum of congener concentrations does not equal ΣPBDE concentration. Thus, values cannot be used to calculate percent congener contributions. b Data obtained using passive samplers situated at locations near the outdoor film sampling sites. ΣPBDE originally reported in Harner et al. (4), congener data provided by T. Harner (personal communication).

estimated that PBDE levels may surpass those of PCB as early as 2003, with a doubling time of 1.6 years. Calculation of Average Gas-Phase Air Concentrations Using KFA. As mentioned in the Introduction, we can use organic films as convenient passive samplers of ambient air and can estimate gas-phase air concentrations assuming that compounds in film and the gas-phase in air are at equilibrium and quantified by KOA (eq 1). Table 2 presents the calculated gas-phase air concentrations for the representative Toronto sites (calculated as the mean of the urban sites, excluding the PBDE “hotspot”, the Electronics Recycling Facility). Despite comprising the bulk of the film mass, BDE-209 was not calculated since reliable KOA values are not available and this compound is not expected to attain significant concentrations in the gas-phase. ΣPBDE concentrations ranged from 0.76 to 15.8 (mean ) 4.8) and 31.2 to 59.2 pg/m3 (mean ) 42.1) for the outdoor and indoor sites, respectively. Calculated outdoor air concentrations are within the range reported for downtown Toronto in summer 2000 (4), indicating that the organic film is a suitable passive sampler for PBDEs. Elevated indoor air concentrations relative to outdoor concentrations is consistent with the source of PBDEs being within the indoor environment. As mentioned above, these results have implications for human exposure to PBDEs considering that humans spend most of their time indoors. The major congener comprising the air profiles was BDE-47 (approximately 60% of the total mass in both indoor and outdoor air), followed by smaller contributions of BDE-99 (15%). The predominance of BDE-47 is the result of its high film concentration and relatively low KOA. Similar congener patterns have been reported in ambient air from the Great Lakes, rural southern Ontario, and downtown Toronto (2-4). These results confirm that organic films on windows can be used as a simple tool to assess atmospheric concentrations of SOCs such as PBDEs. The measured film and calculated gas-phase air concentrations indicate, not surprisingly, that urban indoor air is a source to urban outdoor air, which in turn is a source to the regional environment.

Acknowledgments We thank members of the Diamond research lab for their assistance with field sampling and the analysts of the Regional 730

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 3, 2004

Dioxin Laboratory at the Institute of Ocean Sciences for their assistance with sample analysis. Funding was provided by the Toxic Substances Research Initiative of Health Canada (Project 227, to M.L.D.), the Natural Sciences and Engineering Research Council of Canada (NSERC to M.L.D.), and the Canadian Foundation for Climate and Atmospheric Sciences (CFCAS).

Supporting Information Available Table containing comparison of PBDE and PCB concentrations in window organic films. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Sjo¨din, A.; Carlsson, H.; Thursesson, K.; Sjolin, S.; Bergman, A.; Ostman, C. Environ. Sci. Technol. 2001, 35, 448-454. (2) Strandberg, B.; Dodder, N. G.; Basu, I.; Hites, R. A. Environ. Sci. Technol. 2001, 35, 1078-1083. (3) Gouin, T.; Thomas, G. O.; Cousins, I.; Barber, J.; Mackay, D.; Jones, K. C. Environ. Sci. Technol. 2002, 36, 1426-1434. (4) Harner, T.; Ikonomou, M.; Shoeib, M.; Stern, G.; Diamond, M. Organohalogen Compd. 2002, 57, 33-36. (5) Ikonomou, M. G.; Rayne, S.; Fischer, M.; Fernandez, M. P.; Cretney, W. Chemosphere 2002, 46, 649-663. (6) Environmental Health Criteria 162: Brominated dipheynyl ethers. World Health Organization, 1994. (7) Sellstro¨m, U.; Kierkegaard, A.; de Wit, C.; Jansson, B. Environ. Toxicol. Chem. 1998, 17, 1065-1072. (8) Allchin, C. R.; Law, R. J.; Morris, S. Environ. Pollut. 1999, 105, 197-207. (9) Ikonomou, M. G.; Rayne, S.; Addison, R. F. Environ. Sci. Technol. 2002, 36, 1886-1892. (10) ter Schure, A. F. H.; Larsson, P.; Merila¨, J.; Jo¨nsson, K. I. Environ. Sci. Technol. 2002, 36, 5057-5061. (11) Norstrom, R. J.; Simon, M.; Moisey, J.; Wakeford, B.; Weseloh, D. V. C. Environ. Sci. Technol. 2002, 36, 4783-4789. (12) Ryan, J. J.; Patry, B.; Mills, P.; Beaudoin, N. G. Organohalogen Compd. 2002, 58, 173-176. (13) Nore´n, K.; Meironyte´, D. Chemosphere 2000, 40, 1111-1123. (14) Renner, R. Environ. Sci. Technol. 2000, 34, 223A-226A. (15) Renner, R. Environ. Sci. Technol. 2000, 34, 452A-453A. (16) Gigliotti, C. L.; Dachs, J.; Nelson, E. D.; Brunciak, P. A.; Eisenreich, S. J. Environ. Sci. Technol. 2000, 34, 3547-3554. (17) Simcik, M. F.; Zhang, H.; Eisenreich, S. J.; Franz, T. P. Environ. Sci. Technol. 1997, 31, 2141-2147. (18) Currado, G. M.; Harrad, S. Environ. Sci. Technol. 1998, 32, 30433047. (19) Naumova, Y. Y.; Eisenreich, S. J.; Turpin, B. J.; Weisel, C. P.; Morandi, M. T.; Colome, S. D.; Totten, L. A.; Stock, T. H.; Winer, A. M.; Alimokhtari, S.; Kwon, J.; Shendell, D.; Jones, J.; Maberti, S.; Wall, S. J. Environ. Sci. Technol. 2002, 36, 2552-2559. (20) Currado, G. M.; Harrad, S. Environ. Sci. Technol. 2000, 34, 7882. (21) Halsall, C. J.; Lee, R. G. M.; Coleman, P. J.; Burnett, V.; HardingJones, P.; Jones, K. C. Environ. Sci. Technol. 1995, 29, 23682376. (22) Diamond, M. L.; Gingrich, S. E.; Fertuck, K.; McCarry, B. E.; Stern, G. A.; Billeck, B.; Grift, B.; Brooker, D.; Yager, T. D. Environ. Sci. Technol. 2000, 34, 2900-2908. (23) Gingrich, S. E.; Diamond, M. L.; Stern, G. A.; McCarry, B. E. Environ. Sci. Technol. 2001, 35, 4031-4037. (24) Liu, Q.-T.; Diamond, M. L.; Ondov, J. M.; Maciejczyk, P.; Stern, G. A. Environ. Pollut. 2003, 122, 51-61. (25) Shoeib, M.; Harner, T. Environ. Sci. Technol. 2002, 36, 41424151. (26) Wania, F.; Lei, Y. D.; Harner, T. Anal. Chem. 2002, 74, 34763483. (27) Ikonomou, M. G.; Fraser, T. L.; Crewe, N. F.; Fischer, M. B.; Rogers, I. H.; He, T.; Sather, P. J.; Lamb, R. F. Can. Data Rep. Fish Aquat. Sci. 2001, 2389, 1-95. (28) Stern, G. A.; Halsall, C. J.; Barrie, L. A.; Muir, D. C. G.; Fellin, P.; Rosenberg, B.; Rovinsky, F. Y.; Kononov, E. Y.; Pastuhov, B. Environ. Sci. Technol. 1997, 31, 3619-3628. (29) Shine, J. P.; Ika, R. V.; Ford, T. E. Environ. Sci. Technol. 1995, 29, 1781-1788. (30) Johansson, E.; Wold, S.; Sjodin, K. Anal. Chem. 1984, 56, 16851688. (31) ter Schure, A. F. H.; Larsson, P. Atmos. Environ. 2002, 36, 40154022.

(32) Harner, T.; Shoeib, M. J. Chem. Eng. Data 2002, 47, 228-232. (33) Wong, F.; Diamond, M. L.; Harner, T.; Liu, Q.-T. Environ. Pollut., in press. (34) Sjo¨din, A.; Hagmar, L.; Klasson-Wehler, E.; Kronhom-Diab, K.; Jakobsson, E.; Bergman, A. Environ. Health Prespect. 1999, 107, 643-648. (35) Jakobsson, K.; Thuresson, K.; Rylander, L.; Sjo¨din, A.; Hagmar, L.; Bergman, A. Chemosphere 2002, 46, 709-716. (36) Vorhees, D. J.; Cullen, A. C.; Altshul, L. M. Environ. Sci. Technol. 1997, 31, 3616-3618. (37) Diamond, M. L.; Gingrich, S. E.; Stern, G. A.; McCarry, B. E. Organohalogen Compd. 2000, 45, 272-275. (38) Liu, Q.-T.; Chen, R.; McCarry, B. E.; Diamond, M. L.; Bahavar, B. Environ. Sci. Technol. 2003, 37, 2340-2349. (39) Tysklind, M.; Sellstro¨m, U.; So¨derstro¨m, G.; de Wit, C. In The Second International Workshop on Brominated Flame Retardants; Asplund, L., Bergman, A., de Wit, C., Jansson, B., Lund, B.-O., Marsh, G., Sellstro¨m, U., Wijk, M., Eds.; World Health Organization, The Swedish Minsitry of the Environment, The Royal Swedish Academy of Sciences: Stockholm, Sweden, 2001.

(40) Hardy, M. L. Chemosphere 2002, 46, 717-727. (41) Wenning, R. J. Chemosphere 2002, 46, 779-796. (42) Rayne, S.; Ikonomou, M. G. Environ. Toxicol. Chem. 2002, 21, 2292-2300. (43) Knoth, W.; Mann, W.; Meyer, R.; Nebhuth, J. Organohalogen Compd. 2002, 58, 213-216. (44) Priemer, D. A.; Diamond, M. L. Environ. Sci. Technol. 2002, 36, 1004-1013. (45) Manchester-Neesvig, J. B.; Valters, K.; Sonzogni, W. C. Environ. Sci. Technol. 2001, 35, 1072-1077. (46) Covaci, A.; de Boer, J.; Ryan, J. J.; Voorspoels, S.; Schepens, P. Environ. Res. 2002, 88, 210-218. (47) Rayne, S.; Ikonomou, M. G.; Antcliffe, B. Environ. Sci. Technol. 2003, 37, 2847-2854.

Received for review June 27, 2003. Revised manuscript received October 21, 2003. Accepted October 30, 2003. ES034670R

VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

731