Environ. Sci. Technol. 2007, 41, 6370-6377
Atmospheric Concentrations of Polybrominated Diphenyl Ethers at Near-Source Sites T H O M A S M . C A H I L L , * ,† DANKA GROSKOVA,‡ M . J U D I T H C H A R L E S , ‡,⊥ JAMES R. SANBORN,§ MICHAEL S. DENISON,‡ AND LYNTON BAKER| Department of Integrated Natural Sciences, Arizona State University West Campus, P.O. Box 37100, Phoenix, Arizona 85069, Department of Environmental Toxicology, University of California, Davis, Davis, California 95616, Office of Environmental Health Hazard Assessment, P.O. Box 4010, Sacramento, California 95812-4010, and California Air Resources Board, P.O. Box 2815, Sacramento, California 95812
Concentrations of polybrominated diphenyl ethers (PBDEs) were determined in air samples from near suspected sources, namely an indoors computer laboratory, indoors and outdoors at an electronics recycling facility, and outdoors at an automotive shredding and metal recycling facility. The results showed that (1) PBDE concentrations in the computer laboratory were higher with computers on compared with the computers off, (2) indoor concentrations at an electronics recycling facility were as high as 650 000 pg/ m3 for decabromodiphenyl ether (PBDE 209), and (3) PBDE 209 concentrations were up to 1900 pg/m3 at the downwind fenceline at an automotive shredding/metal recycling facility. The inhalation exposure estimates for all the sites were typically below 110 pg/kg/day with the exception of the indoor air samples adjacent to the electronics shredding equipment, which gave exposure estimates upward of 40 000 pg/kg/day. Although there were elevated inhalation exposures at the three source sites, the exposure was not expected to cause adverse health effects based on the lowest reference dose (RfD) currently in the Integrated Risk Information System (IRIS), although these RfD values are currently being re-evaluated by the U.S. Environmental Protection Agency. More research is needed on the potential health effects of PBDEs.
Introduction Polybrominated diphenyl ethers (PBDEs) are fire retardants that were used in foams and electronics applications, but there has been increasing concern about the safety of these chemicals for several reasons. First, the PBDEs are hydrophobic chemicals (log Kow values 4-10) due to the presence of a large number of bromines attached to the molecule, which predisposes them to bioaccumulate in both humans * Corresponding author phone: (602) 543-6021; e-mail: tmcahill@ asu.edu. † Arizona State University. ‡ University of California. § Office of Environmental Health Hazard Assessment. | California Air Resources Board. ⊥ Deceased Oct 2004. 6370
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
and wildlife. Second, PBDEs are detected in human milk (1), blood serum (2, 3), and adipose tissue (4-6), which demonstrates human exposure to these chemicals. Last, these chemicals may be endocrine disruptors since the hydroxy metabolites of some PBDEs are structurally similar to thyroid hormones, so prolonged exposure to PBDEs may cause thyroidogenic, estrogenic, hepatic, and/or neurodevelopmental effects (7). Because of the increasing concern about the possible health effects from PBDEs, both the European Union and the State of California have prohibited the use of the Penta- and Octa-formulations of PBDEs although the Deca- formulation has no such restrictions. Considerable research has determined PBDE concentrations in environmental media, such as indoor air (8-10), outdoor air (11-17), dust (18-20), freshwater (21-23), and marine organisms (6, 24). PBDEs are now ubiquitous in the environment, but higher concentrations of PBDEs are found in indoor and occupational situations (10, 19, 20, 25, 26). Since high human exposure to these chemicals is more likely through occupational exposure rather than environmental or dietary exposure, these sites need to be monitored more carefully to assess human exposure to PBDEs. The objective of this research was to determine the air concentrations of PBDEs at three sites expected to have elevated PBDE concentrations to both assess the PBDE emissions from these sites and the potential human exposure to these chemicals. The exposure sites chosen for study were (1) a computer laboratory in a government office building, (2) an electronics shredding and recycling facility, and (3) an automotive shredding and metal recycling operation. In each case, the use or handling of products that contained PBDEs was expected to generate significant PBDE emissions. The air concentrations from these sites were used to estimate human exposure in these situations which were then compared to toxicological data to assess if the exposures posed a health risk.
Experimental Section Sampling Site Locations. Atmospheric concentrations of PBDEs were determined at three near-source sites and a control site. The first site was a computer training laboratory with a volume of 114 m3 in a government office building in California. The room was sealed and the ventilation turned off to prevent dilution of the computer emissions. The laboratory contained 13 personal computers and monitors of sufficient age that they likely contained PBDEs. Indoor air samples of 24h duration, both particulate and gas phases, were collected on 8 days in January and February 2004. The computers were turned on for six sampling days and off for two sampling days to determine contributions from operating computers. The second near-source site was an electronics recycling facility where electronic equipment was dismantled and shredded. To represent the worst case scenario, a recycling facility was selected that lacked any emission control or dust suppression measures. Air samples of 8h duration were collected inside the dismantling hall with two samplers located 1.7 m from the circuit board shredder as well as from outside the building by four samplers located around the building (See Figure S-1, Supporting Information). Indoor samples (n ) 2 per day) and outdoor samples (n ) 4 per day) were collected for 3 working days (June 2-4, 2004). There was no shredding activity on June 3, 2004, which provided a contrast to the other sampling days. The third near-source site investigated was an automotive shredding and metal recycling facility. PBDEs were expected 10.1021/es070844j CCC: $37.00
2007 American Chemical Society Published on Web 08/17/2007
to originate from the foams and plastic components of the vehicles. Since the entire operation was conducted outdoors, only outdoor samples of 24h duration were collected with three samplers located around the site for 3 days in September 2004. (See Figure S-2, Supporting Information). The site was not in operation for 1 day, so the samples from that day represent a “no activity” control condition. The last site was outdoors at the University of California, Davis, to determine the PBDE background in a semiurban site. Duplicate 24-hour outdoor samples were collected on March 17 and 18, 2004. These samples represented the control site for the other outdoor samples. Sample Collection and Extraction Procedures. The air samples were collected by two types of samplers. For the indoor air samples, 13 to 18 L/min of air was pulled through a 47 mm Tissuquartz filter (PALL Life Sciences, Ann Arbor, MI) to collect the particulate phase and then 6 g of Amberlite XAD-2 resin (Supelco Inc., Bellefone, PA) to collect the vaporphase chemicals and blow-off from the filter. The sample substrates were precleaned before use by baking the quartz filters at 550 °C for 8 h and Soxhlet extracting the XAD resin for 24 h with dichloromethane. No size-selective inlets were used so the sampler could collect large particles. The sampling apparatus was constructed from stainless steel to reduce PBDE adsorption or contamination. The outdoor air samples were collected by Andersen Highvolume total suspended particulate samplers (model GBM2000H, Andersen Instruments Inc.). The samplers collected particulate matter onto 20.3 × 25.4 cm Tissuquartz filters and then collected gas-phase and filter blow-off onto 90 g of Amberlite XAD-2 resin mounted in an aluminum holder downstream of the filter. The samplers were operated at a flow rate of 30 m3/hour. After sampling, the filters and XAD-2 adsorbents were individually sealed in amber glass jars and stored in the freezer at -20 °C until extraction, which was generally less than two weeks. Several quality control measures were taken to ensure the accuracy of the results. For each field sampling campaign, three filters and three batches of XAD-2 were mounted in the samplers and then removed for field blanks. In addition, three filters and three XAD-2 samples were spiked with six PBDEs congeners in the field to verify sample stability and extraction efficiencies. All samples were extracted and concentrated in a similar fashion. The samples (filters or XAD-2) were Soxhlet extracted for 24 h with dichloromethane (B&J CG2 grade, Honeywell International, Inc., Muskegon, MI) in a darkened fume hood. The extracts were then concentrated by rotoevaporation, transferred to a silica gel cleanup column (5 g of 70-230 mesh gel, Aldrich, Milwaukee, WI), and eluted with 50 mL dichloromethane. The extracts were nitrogen evaporated to 300-500 µL and solvent exchanged into isooctane (nanograde, Mallinckrodt Baker Inc., Paris, KY). PCB-65, PCB-209 and 13C12-deca-PBDE (Cambridge Isotope Laboratories, Andover, MA.) were then added as injection standards to account for any potential instrument drift. PCB-65 was used for the mono- to tri- congeners; PCB-209 was used for the tetra- to hepta-congeners and 13C12-deca-PBDE was used for the octato deca-congeners. Analyte quantification was conducted by gas chromatography-negative chemical ionization mass spectrometry, which is detailed in the Supporting Information. The filter and XAD-2 resin PBDE masses were totaled to give a total atmospheric PBDE concentration. Exposure Assessment. Female and male human inhalation exposures were compiled from the PBDE air concentrations and physiological factors from the US EPA Exposure Factors Handbook (27). An exposure duration of 8 h was assumed for the occupational settings. Inhalation uptake and respiratory retention of 100% was used in the calcula-
tions, but this may overestimate actual exposure since some coarse particles may not be inhaled into the lung. A Hazard Quotient (HQ) approach was used to estimate potential for adverse health outcome from the PBDE inhalation exposures estimates. The HQ is the ratio of exposure (mg/kg/day) to the reference dose (RfD, mg/kg/day) for the PBDEs in the Integrated Risk Information System (IRIS) (28). HQs less than or equal to 1 are not expected to result in adverse health effects.
Results The large number of sites, activity conditions, and PBDE congeners detected resulted in large raw data tables. Accordingly, the results will be summarized herein while the full data tables are presented in the Supporting Information as well as a report prepared for the California Air Resources Board (29). The mono- and di-PBDEs had numerous interfering compounds in the chromatogram and lacked sufficiently unique ions in the mass spectra, so they were not quantified. In addition, octa- and nona-PBDEs appeared to have interferences arising from the degradation of PBDE209 in the injection system as evidenced by degradation of the PBDE-209 standard analyzed after the field samples. The loss of PBDE-209 was minor (
