Polybrominated Diphenyl Ethers in Indoor Dust in Ottawa, Canada

Air Quality Research Branch, Meteorological Service of Canada, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada, M3H 5T4, Environmen...
0 downloads 0 Views 214KB Size
Environ. Sci. Technol. 2005, 39, 7027-7035

Polybrominated Diphenyl Ethers in Indoor Dust in Ottawa, Canada: Implications for Sources and Exposure B R Y O N Y H . W I L F O R D , †,‡ M A H I B A S H O E I B , † T O M H A R N E R , * ,† JIPING ZHU,§ AND KEVIN C. JONES‡ Air Quality Research Branch, Meteorological Service of Canada, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada, M3H 5T4, Environmental Science Department, Institute of Environmental and Natural Sciences, Lancaster University, Lancaster, LA1 4YQ, U.K., and Chemistry Research Division, Health Canada, Ottawa, Ontario, Canada, K1L 0L2

Polybrominated diphenyl ethers (PBDEs) are widely used as additive flame retardants in plastics, soft furnishings, electrical and electronic equipment, and insulation in the indoor environment, and may be released indoors via volatilization or as dusts. The penta-and octa-brominated mixes are now banned in most parts of Europe, and phasing out of their use has recently begun in North America. This study follows a previous investigation into indoor air levels of PBDEs. House dust was analyzed from the family vacuum cleaners of 68 of the same 74 randomly selected homes, in Ottawa, Canada during the winter of 20022003. PBDEs, comprising on average 42% BDE-209, were found in all samples. The levels were log-normally distributed with a geometric mean ΣPBDE of 2000 ng g-1, and a median of 1800 ng g-1 dust. The levels in dust did not correlate with questionnaire information on house characteristics. Correlations were found between pentamix congener levels in dust and in air from the same homes, but not for congeners of the more highly brominated mixes. Exposure scenarios are presented for mean and high dust ingestion rates, and compared against exposures from other pathways, for both adults and toddlers (6 months-2 years). Assuming a mean dust ingestion rate and median dust and air concentrations, adults would be exposed to ca. 7.5 ng ΣPBDE d-1 via the dust ingestion pathway, which represents ∼14% of total daily exposure when compared to diet (82%) and inhalation (4%). However, for toddlers the equivalent intakes would be 99 ng d-1, representing 80% of their daily PBDE exposure. At high dust ingestion rates these values increase to 180 ng d-1 (80% daily intake) for adults and 360 ng d-1 (89% daily intake) for toddlers. The data give a clearer picture of sources of PBDE exposure in the home environment and suggest that dust could be a significant exposure pathway for some individuals, particularly children.

* Corresponding author tel.: +1 416 739 4837; fax: +1 416 739 5708; e-mail: [email protected]. † Environment Canada. ‡ Lancaster University. § Health Canada. 10.1021/es050759g CCC: $30.25 Published on Web 08/10/2005

 2005 American Chemical Society

Introduction Polybrominated diphenyl ethers (PBDEs) are used as flame retardants in plastic polymers for many household materials. The global demand for PBDEs has been substantialsan estimated 67 000 tonnes in 2001, for example, manufactured and used as three main technical mixtures, the penta-, octa-, and deca- products, accounting for 11%, 6%, and 83% of the global market, respectively (1). The penta-mix has been used mainly in polyurethane foam for soft furnishings, mattresses, and car seating and insulation; the octa-mix was used mainly in acrylonitrile-butadiene-styrene (ABS) polymer for equipment casings; the deca-mix is used in hard plastics for electronic equipment components and casings, plus flameretarded rubbers and textile back coatings (2). Despite decaBDE having the major market share, BDE-47 and -99, the principal congeners of the penta-mix, are generally those detected at the highest levels in biota and the environment (3). The European Union has recently prohibited the marketing and use of the penta- and octa- commercial mixes due to concerns for secondary poisoning (4). Rapidly increasing levels have been found in humans, other biota, and environmental media (3), and concerns have been raised over some congeners’ potential long-range atmospheric transport (LRAT), tendency to bioaccumulate, and potential toxicity (5, 6). European manufacturers also avoided use of the penta-mix in their products leading up to this ban, and as a result, 95% of recent global penta-mix use has been in North America (7). PBDE levels in the North American population appear to be ∼15-20 times higher than those found in European studies (3). The penta-mix is currently being considered as a potential “persistent organic pollutant” (POP) under the 1998 United Nations Economic Commission for Europe (UNECE) “POPs convention”. Regulatory action has been slower in North America, although California has announced a plan to ban both the penta- and the octaformulations by 2008 (8). However, the main U.S. producer of penta-BDE and octa-BDE recently announced it would discontinue these products by the end of 2004 as part of a voluntary agreement with the United States Environmental Protection Agency (USEPA) (9) which may preclude the need for regulation. There is considerable interest in how PBDEs escape from sources, enter environmental compartments, and reach biota, including humans. Possible indoor pathways of release include volatilization for lower-brominated congeners (10) or formation of dusts by the shearing off of fragments (11) during the use of treated products. Other entry routes into the environment include emissions during PBDE manufacture, incorporation into products, and the recycling and disposal of flame-retarded products (12-15). Treated products represent significant reservoirs of PBDEs (16), so they are the focus of attention for emission estimates (10, 16-19). Expected exposure scenarios for PBDEs have generally been based on PCB exposure pathways (and those of other older-generation POPs from industrial or agricultural applications, such as those now covered under international agreements e.g. the Stockholm Convention of the United Nations Environment Program). For most traditional POPs, dietary intake dominates human exposure (20). However, in contrast to these older legacy POPs, the PBDEs have been added at 5-30% (1) to products used predominantly indoorssin homes and workplacessand in vehicles. This leads to PBDE indoor air levels being much higher than outdoor levels (21-23), and significant indoor dust PBDE concentrations have been found in homes (24-28) and offices (29) in various countries. It has been found in various studies VOL. 39, NO. 18, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7027

that PBDE levels in 5-10% of any population studied are significantly higher than those of the rest, which could not be completely accounted for by dietary intake or occupational exposure (30). This suggests that the population as a whole is subject to an erratic/variable source, and it is possible that PBDE-loaded house dust could be one such “missing” source of variability not currently taken into account in the intake pathways considered. On average, North Americans spend approximately 90% of their time indoors, with the majority of this being at home (31, 32). In Canada, this figure increases to 96-98% during winter months (32). Accidental ingestion of house dust containing elevated levels of PBDEs, e.g., on food and by hand-to-mouth transfer from other surfaces, could therefore represent an important pathway of human exposure (24, 26, 30). Dermal uptake from dust might also be conceivable. Dust ingestion may be particularly significant for young children (24) who are estimated by the USEPA to consume at least double the amount typically ingested by adults (33), due to their proximity to the floor and their tendency to put their hands and other objects into their mouths. This scenario is also of concern given that indoor sourcessboth primary, i.e., products in use, and secondary, i.e., dustswill continue to emit flame retardants after any phase-out or restrictions on PBDE use, and time scales for such releases are unclear or unknown (10, 18, 19). Such compounds may also persist indoors if dust is allowed to accumulate, since they are unlikely to experience the same wind dispersal and degradation processes (e.g., solar, microbial if any) as those outdoors. Any PBDE-laden particulates which are released to the outdoors via ventilation or household cleaningseither by direct sweeping or via landfill sitesscan then be distributed regionally by wind (34, 35). House dusts are also likely to end up in the sewerage system, from where they may be spread on crop fields as sludge, or released into rivers (and deposited to sediments) (11). House dust is a heterogeneous mixture of biologically derived material, including human and pet skin and hair, mites, and fungus spores; particulate matter from indoor aerosols, e.g., fibers from clothing and furnishings, cooking and heating emissions, and cigarette smoke; and soil brought in by foot traffic (33). Its composition depends on a large number of factors, including the location and construction of the building, the room’s use, type of decorating and furnishing materials, heating and ventilation systems, how well and often the area is cleaned, and the time of year (36, 37). It is therefore likely to vary between rooms of a house. Previous studies have suggested that the congener profile of PBDEs in indoor air appears to be mainly a result of volatilization of PBDEs from products in use (21, 22), dominated by the more volatile lower-brominated congeners of the commercial penta-mix. Dusts are more likely to contain the more highly brominated, less volatile congeners, including BDE-209 which makes up 97% of the deca-BDE commercial mix (1). This could be due to sheared-off particles of flameretarded plastics or textile fibers, similar to the pathway suggested for the penta-mix (11). Deca-BDE has increasingly been the subject of attention, since it is the only remaining PBDE mix not subject to regulatory action; 56 100 tonnes of it was used globally in 2001 (1), mainly in hard plastics such as electrical components and equipment casings. Despite assurances from industry that BDE-209 has low environmental mobility (38) due to its low water solubility and vapor pressure, its fate and transport in the particulate phase is little understood (35). It is poorly bioavailable with low absorption from the gastrointestinal tract and rapid elimination, possibly due to its large molecular size, yet it has been found in Peregrine falcon eggs (39) and also in human blood (40, 41), serum and adipose tissue (42), and breast milk (43). BDE-209 can be broken down photolytically (44-46) to products including lower-brominated BDEs and PBDD/Fs 7028

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 18, 2005

of greater toxicological concern, and may be metabolically degraded in the gut of some fish (47). In this study, part of a wider investigation into factors affecting indoor levels of various pollutants, dust samples were analyzed from the family vacuum cleaners in 68 of the 74 randomly selected homes in Ottawa, Canada, which were concurrently sampled for PBDEs in air (21). Briefly, passive air samplers were placed in homes for a period of 3 weeks, and a calibration study was conducted such that an actual air concentration for each home could be calculated. The objectives of this dust analysis were to (1) establish the range of indoor dust concentrations; (2) compare the range of indoor dust concentrations to indoor air levels found in the same houses; (3) consider the implications for human exposure via indoor dust, in comparison to diet and indoor air exposures.

Materials and Methods Study Location. The study was conducted in the city of Ottawa, Canada (population 780 000). A detailed description of selection of the residences sampled is given elsewhere (21). Briefly, a two-stage stratified sampling process was used to select a representative sample of residences, which should provide a good estimate of levels. First, the city was divided into three geographical areas, based on level of urbanization (urban core, urban fringe, and rural fringe) using 2001 Canada Census data provided by Statistics Canada. Stratified random sampling was used to select the number of zones within each area-type to be sampled, and simple random sampling was used to select specific zones from each geographical area. Homes to be sampled in each zone were selected by random sampling. A questionnaire was completed in each home covering house characteristics (e.g., dwelling type, size, age, carpeting, heating type, ventilation and air purification, renovations, new furniture), residents’ lifestyle (e.g., number of adults and children resident, smoking, hobbies, products stored in the home or garage, occupational exposures to dusts/fumes/ gases) to attempt to identify factors affecting chemical loadings. Sample Collection. Dust samples were collected directly from the vacuum cleaner in regular use by the householder, hence the vacuum cleaner used and vacuuming practices of the household could not be standardized. The typical frequency between vacuuming was noted in the questionnaire for each household. The whole vacuum cleaner bag was removed from the family vacuum cleaner, placed in a clean polyethylene zip-seal plastic bag (Fisher Scientific, Nepean, Ontario), from which the dust contents were transferred to another clean polyethylene zip-seal plastic bag for transport to the Health Canada laboratory. The bags were then cut open with clean scissors and the whole contents sieved according to the following procedure. A portion of the bag contents was transferred to an all stainless steel testing sieve VWR ∼100 mesh for 150-µm cutoff (VWR International, Montreal, Quebec), covered with the lid, and shaken vigorously for 3-5 min. The sieve was allowed to sit for 30 s, then the top two sections were set aside and the dust in the collection pan was transferred to a 125-mL preweighed jar (VWR International, Montreal, Quebec) using a clean paintbrush that was discarded after each use. Clean tweezers were used to remove collected hair. This process was repeated until all dust in the collection bag had been sieved. Dust collected in the jar was mixed by shaking, and a 3-g portion was transferred to a 20-mL wide-mouth vial with aluminumlined screw-cap and stored at -20 °C before despatch to Environment Canada laboratories. Samples were then frozen at -20 °C until analysis. Sample Preparation and Analysis. Ceramic thimbles were baked overnight at 450 °C, then pre-extracted in DCM for >3

h. A 70-ng aliquot of BDE-35 was added as a surrogate standard to the samples prior to extraction. Samples were dry and friable so no drying agent was used. Approximately 0.25 g of each sample was Soxhlet extracted for 15 h in dichloromethane (DCM), concentrated to ∼4 mL, and transferred to petroleum ether/isooctane by rotary evaporation. Extracts were then made up to 10 mL, homogenized by stirring, and volumetrically divided into 2 aliquots. One 5-mL aliquot was blown down and frozen at -20 °C for later analysis for other compounds. The remaining 5-mL aliquot was blown-down and solvent exchanged to 3 mL of petroleum ether/isooctane then treated by shaking with 1-2 mL of concentrated sulfuric acid for at least 1 min using a vortex mixer. The sample was centrifuged and the acid layer was removed by pipet. Samples were acid-washed a second time, rinsed with 3 mL of deionized water, and then placed in the freezer (-20 °C) overnight; any remaining water was removed by pipet. Samples were concentrated to 1 mL in isooctane for analysis by GC/MS. Mirex was added as an internal standard (10 µL of 10 ng µL-1) for volume correction and to adjust for variations in instrument response. BDE-209 may experience photodegradation, hence sample exposure to light was minimized during extraction and cleanup, and amber GC vials were used for storage. Extracts were analyzed for lower-brominated BDE congeners (BDEs 17, 28, 71, 47, 66, 100, 99, 85, 154, 153, 138, 183, and 190) by GC/MS on a Hewlett-Packard 6890 GC fitted with a DB-5 60-m column (J&W Scientific, Rancho Cordova, CA), i.d. 0.25 mm, film thickness 0.25 µm, and operated with helium as the carrier gas. This was coupled to a HewlettPackard 5973 mass selective detector MS in negative chemical ionization (NCI) mode with methane as reagent gas. Temperatures were as follows: injector 250 °C; transfer line 300 °C; source 230 °C; quadrupole 106 °C. Sample injection volume was 2 µL in splitless mode. Masses 79 (PBDEs) and 404 (Mirex) were monitored for quantification, plus 81 for PBDE confirmation. Compounds were identified and quantified by comparison with a series of standard solutions of known concentration (from Cambridge Isotope Laboratories, Andover, MA). Peaks were quantified only if the signal to noise ratio was >3, and the ratio of the ion to its qualifier ion was within 15% of the standard value. Samples were also analyzed for BDE-209 using the same GC/MS system as above in NCI mode, on a 15-m DB-5 column (J&W Scientific, Rancho Cordova, CA), i.d. 0.25 mm, film thickness 0.25 µm, operated with helium as the carrier gas. Sample injection volume was 2 µL in splitless mode. Temperatures were as follows: injector 250 °C; transfer line 290 °C; source 250 °C; quadrupole 106 °C. The most abundant ion, mass 81, was used for quantification, and ions 79, 487, and 486 were used for confirmation. Subsets of all dust samples were sent for total organic and inorganic carbon analysis at the University of Guelph (Laboratory Services, Guelph, Ontario). QA/QC. Since an equivalent matrix with sorptive properties similar to dust is not readily available for performing field blanks, sodium sulfate granules were placed in a 20-mL vial after the sieving step (since these were larger than the sieve mesh size) and shipped/stored and processed in the same manner as the samples, as a storage blank. Solvent blanks were also analyzed to check for contamination from the laboratory and equipment. One blank (solvent or sodium sulfate storage blank alternately) was extracted in each batch of 10-12 samples. For analytes found in blanks, the detection limit was raised to the mean of the blank levels plus 3 standard deviations. The instrument detection limit for each congener was calculated by linear extrapolation from the lowest standard in the calibration, to the point where the signal to noise ratio was equal to 3. Detection limits typically increased with level of bromination and average detection limits ranged between 0.5 and 8 ng g-1 (0.1-14 ng g-1 for individual sample

TABLE 1. Summary of PBDE Levels Found in House Dust in ng g-1 Dry Weight (Only Congeners Found above the Detection Limit Are Shown)a median mean BDE 17 BDE 28 BDE 47 BDE 66 BDE 100 BDE 99 BDE 85 BDE 154 BDE 153 BDE 138 BDE 183 BDE 190 BDE 209 ΣBDE excl. -209 ΣBDE incl. -209

range