Environ. Sci. Technol. 2004, 38, 1313-1320
Indoor and Outdoor Air Concentrations and Phase Partitioning of Perfluoroalkyl Sulfonamides and Polybrominated Diphenyl Ethers M A H I B A S H O E I B , † T O M H A R N E R , * ,† MICHAEL IKONOMOU,‡ AND K U R U N T H A C H A L A M K A N N A N §,| Meteorological Service of Canada, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4, Institute for Ocean Sciences, Fisheries and Oceans Canada, Sidney, British Columbia, Canada V8L 4B2, and Wadsworth Center, New York State Department of Health, Empire State Plaza, P.O. Box 509, and Department of Environmental Toxicology and Health, State University of New York, Albany, New York 12202-0509
Perfluoroalkyls (PFAs) and polybrominated diphenyl ethers (PBDEs) are two classes of emerging persistent organic pollutants (POPs) that are widely used in domestic and workplace products. These compounds also occur in remote locations such as the Arctic where they are accumulated in the food chain. This study makes connections between indoor sources of these chemicals and the potential and mode for their transport in air. In the case of the PFAs, three perfluoralkyl sulfonamides (PFASs) were investigatedsN-methyl perfluorooctane sulfonamidoethanol (MeFOSE), N-ethyl perfluorooctane sulfonamidoethanol (EtFOSE), and N-methyl perfluorooctane sulfonamidethylacrylate (MeFOSEA). These are believed to act as precursors that eventually degrade to perfluorooctane sulfonate (PFOS), which is detected in samples from remote regions. High-volume samples were collected for indoor and outdoor air to investigate the source signature and strength. Mean indoor air concentrations (pg/m3) were 2590 (MeFOSE), 770 (EtFOSE), and 630 (∑PBDE). The ratios of concentration between indoor and outdoor air were 110 for MeFOSE, 85 for EtFOSE, and 15 for ∑PBDE. The gas and particle phases were collected separately to investigate the partitioning characteristics of these chemicals. Measured particulate percentages were compared to predicted values determined using models based on the octanol-air partition coefficient (KOA) and supercooled liquid vapor pressure (p°L); these models were previously developed for nonpolar, hydrophobic chemicals. To make this comparison for the three PFASs, it was necessary to measure their KOA and vapor pressure. KOA values were measured as a function of temperature (0 to +20 °C). Values of log KOA at 20 °C were 7.70, 7.78, and 7.87 for MeFOSE, EtFOSE, and MeFOSEA, respectively. Partitioning * Corresponding author phone: (416)739-4837; fax: (416)739-5708; e-mail:
[email protected]. † Environment Canada. ‡ Institute for Ocean Sciences, Fisheries and Oceans Canada. § New York State Department of Health. | State University of New York. 10.1021/es0305555 CCC: $27.50 Published on Web 01/15/2004
2004 American Chemical Society
to octanol increased at colder temperatures, and the enthalpies associated with octanol-air transfer (∆HOA, kJ/ mol) were 68-73 and consistent with previous measurements for nonpolar hydrophobic chemicals. Solid-phase vapor pressures (p°S) were measured at room temperature (23 °C) by the gas saturation method. Values of p°S (Pa) were 4.0 × 10-4, 1.7 × 10-3, and 4.1 × 10-4, respectively. These were converted to p°L for describing particle-gas exchange. Both the p°L-based model and the KOA model worked well for the PBDEs but were not valid for the PFASs, greatly underpredicting particulate percentages. These results suggest that existing KOA- and p°L-based models of partitioning will need to be recalibrated for PFASs.
Introduction Until recently, research into the environmental fate and toxicological effects of persistent halogenated compounds has focused largely on chlorinated compounds. Now, other classes of halogenated chemicals have come to the forefront due to recent studies (1-6) showing their widespread use, distribution in humans and wildlife, and concern regarding their environmental persistence. Two such classes include perfluorinated surfactants and brominated flame retardants. Perfluoroalkyl compounds (PFAs) have been manufactured since the 1950s for use as refrigerants; surfactants; fire retardants; stain-resistant coatings for fabrics, carpets, and paper; and insecticides. Production of perfluorinated (fully fluorinated) chemicals increased by 220% over the period 1988-1997 (7). The total production of fluorinated surfactants (anionics, cationic, and neutral) was 200 t in 1979 (8), whereas in 2000, the total production of one PFA, perfluorooctane sulfonate (PFOS), was nearly 3000 t (9). This represents only a small fraction (∼0.01%) of total surfactant production (810). PFOS is one of the main perfluorinated products. It is a ubiquitous pollutant and has been detected in wildlife samples from remote geographic locations such as the Arctic and the Mid-North Pacific Ocean (1-3). In addition to its presence in various perfluorinated products, PFOS could also be a stable degradation product/metabolite of several other PFAs. These are referred to as PFOS precursors and include N-methyl perfluorooctane sulfonamidoethanol (MeFOSE, C8F17SO2N(CH3)CH2CH2OH), N-ethyl perfluorooctane sulfonamidoethanol (EtFOSE; C8F17SO2N(CH2CH3)CH2CH2OH), and N-methyl perfluorooctane sulfonamidethylacrylate (MeFOSEA) that are incorporated into surface treatment formulation for textile and paper products to impart oil and water resistance. Because PFOS itself is relatively less volatile, it is hypothesized that its occurrence in remote regions is the result of atmospheric transport of relatively more volatile precursors that are later transformed into PFOS (11). Several other PFAs compounds were found to degrade to PFOS using the CATABOL model stimulation (12). A detailed assessment of this pathway is difficult due to the lack of information on the key physical-chemical properties for these chemicals. It is also not clear how basic physical-chemical property data (vapor pressure; Henry’s law constant; octanol-water partition coefficient, KOW; and octanol-air partition coefficient, KOA) relate to the environmental partitioning behavior of these chemicals. This complexity arises because PFAs are different from typical persistent pollutants, having both hydrophobic and hydrophilic ends. Although the toxicity of PFOS and related compounds is still relatively unknown, some PFAs have been linked to VOL. 38, NO. 5, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1313
carcinogenicity and disruption of gap junction intracellular communication (13). As a result of some of these concerns, 3M has recently phased out the production of POSF (perfluorooctanesulfonylfluoride), the basic chemical building block for many sulfonyl-based fluorochemicals (7). Polybrominated diphenyl ethers (PBDE) are a class of brominated flame retardants (BFRs) that are added to a variety of materials, including computers, textiles, and polyurethane foam used in domestic and construction materials. Global production of PBDEs was ∼40 000 t in 1992 (14) and ∼67 000 t in 1999 (BSEF, 2000; http://05.232.112.21/ bsef/docs/major_Brominated.doc, Brussels, Belgium). PBDEs are manufactured as three commercial products having different degrees of bromination (14). The deca-product, which contains about 97% of the fully brominated diphenyl ether (BD-209, using the IUPAC naming scheme), comprised about 80% of the world PBDE demand in 1999 (15). The other two PBDE formulations include the octa-product, consisting of 70-80% hepta and octa congeners, and the penta-product, which consists mainly of tetra and penta congeners. The penta-product was heavily used between 1992 and 1999 (15). PBDEs are ubiquitous environmental pollutants and are detected in air (16-20); in sediments and sewage sludge (2124); in biological samples including fish, birds, and marine mammals (5, 25-27); and in human plasma and milk (4, 28-30). Dramatic increases in human milk concentrations have been observed in Sweden (29) and in North America (4), although North American values were higher by more than an order of magnitude (30). A recent study in the Canadian Arctic showed that levels of PBDEs in ringed seal have increased exponentially over the period from 1981 to 2000, suggesting that they are efficiently transported long distances via the atmosphere (5). Air concentrations of PBDEs have been reported at background locations in the Great Lakes region (16) and the Canadian Arctic (31). Higher concentrations have been measured in source areas such as electronic recycling plant (18, 19), offices containing computers (20), domestic and workplace environments (32), and urban air (16). Recent measurements of the octanol-air partition coefficient (KOA) (33) and vapor pressure (34) have allowed estimates of the particle-gas partitioning of PBDEs that agree with measured data (16). KOA and vapor pressure are also useful for predicting partitioning of gas-phase chemicals to other environmental organic surfaces such as soil and vegetation. Such basic physical-chemical property data are essential for understanding the transport and partitioning of chemicals in the environment. Toxicological studies on PBDEs and other BFRs are limited but suggest that the less brominated congeners may compromise endocrine and hepatic system function as well as neurodevelopment (35, 36). The thyroid hormone system is also sensitive to PBDE exposure (37). As a result of their predominant usage in domestic and workplace materials, both PFAs and PBDEs are expected to be present in indoor air. This has important consequences for human exposure and associated health risk since people spend much of their time indoors. Furthermore, indoor air will serve as a source to the outdoor environment. Despite these concerns, few studies have measured PFAs and PBDEs in indoor air or attempted to assess their partitioning status in air. This study addresses these data gaps by measuring indoor and outdoor concentrations of PBDEs and PFAs including their partitioning to aerosols. The octanol-air partition coefficient (KOA) and solid-phase vapor pressure are measured directly for three sulfonamides and used to test KOA-based and vapor pressure-based models of particlegas partitioning that have been useful for other classes of POPs. 1314
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 5, 2004
Experimental Section Chemicals. The perfluoroalkyl sulfonamides (PFASs) MeFOSE, EtFOSE, and MeFOSEA were obtained from the 3M Company, with purity >90%. Polybrominated diphenyl ether (PBDE) and 13C-labeled PBDE standards with purity of 98% were purchased from Cambridge Isotope Laboratories, Inc. (Andover, MA). Mirex was also used as an internal standard for volume correction. Octanol-Air Partition Coefficient (KOA) Measurements. KOA values were measured as a function of temperature (0, +10, and +20 °C) for MeFOSE, EtFOSE, and MeFOSEA using an established method (38). Briefly, octanol-saturated nitrogen was passed through a generator column consisting of glass beads coated with octanol solution of the target analytes at concentration CO. Equilibrated chemicals in the monitored gas stream exiting the generator column were then collected on an adsorbent trap consisting of C18-bonded silica (IST Limited, Mid Glamorgan, U.K.). The trap was extracted with 15 mL of acetone/ethyl acetate (1:1) to yield an air concentration CA. Extracts were blown down to approximately 1 mL with a gentle stream of nitrogen at low temperature to prevent evaporative losses. The analysis of the extracts is described later. Several measurements were conducted at each temperature, and KOA values were calculated as the ratio CO/CA. Solid Phase Vapor Pressure (p°S) Determinations. The solid-phase vapor pressure of MeFOSE, EtFOSE, and MeFOSEA was determined using the generator column method described by Wania et al. (39). Briefly, glass beads were soaked in a hexane solution consisting of dissolved sulfonamide crystals (approximately 5-20 mg of each component) and two reference compoundssp,p′-DDE and p,p′-DDT. The reference compounds were added for validation of the method by comparing with published values. After evaporating the hexane to dryness, the coated glass beads were loaded into a generator column (same dimensions as used in the KOA measurements described above). Nitrogen was passed through the column at a flow rate of approximately 13-320 mL/min, and the exiting airstream was sampled using traps consisting of C18-bonded silica (same as for the KOA determinations). Traps were extracted with 20 mL of acetone/ ethyl acetate (1:1) and concentrated into ethyl acetate. A range of flow rates was used to confirm that the airstream was in fact saturated (in equilibrium) with the solid crystals. Vapor pressures were calculated assuming ideal gas behavior [i.e., P ) nRT/V, where P is vapor pressure in Pa, n is moles (determined from the amount of analyte collected on C18), T is temperature in Kelvin, V is air volume (m3, measured at the exit of the generator column), and R (J/mol‚K) is the ideal gas constant. Four determinations of p°S were made at room temperature (∼23 °C). Air Sampling. Concentrations of the PFASs and PBDEs in indoor and outdoor air were determined using a conventional high-volume air sampler (PS-1, Tisch Environmental, Cleves, OH; sampling rate ∼0.4 m3/min). The sampling train consisted of one or two glass fiber filters (GFFs, 20 cm × 25 cm, Whatman International Ltd, Maidstone, England) for trapping particles, followed by two polyurethane foam (PUF) plugs (80 mm diameter, 75 mm long, PacWill Environmental, Stoney Creek, ON) for collecting the gas phase. The purpose of the second GFF was to assess adsorptive partitioning of gas-phase chemical that has penetrated the first filter. The second PUF plug was included to assess breakthrough of gas-phase chemical. Air sampling times and volumes were typically 12-24 h (300-600 m3) for outdoor samples and 3-6 h (100-200 m3) for indoor samples. Teflon gaskets were not used on the high-volume sampler to eliminate potential source of contamination. A total of 10 indoor and 3 outdoor samples was collected. Field blanks were also collected by placing PUF and GFF in the sampling
TABLE 1. Sampling Details for Indoor and outdoor Air Samples location outdoor 1 outdoor 2 old laboratory new laboratory house 1 house 2 house 3 house 4 a
sampling date Nov 2001 Mar 2003 Nov 2001 Dec 2001 Jan 2003 Nov 2001 Dec 2001 Dec 2001 Mar 2003 Nov 2001 Mar 2003 Nov 2001 Feb 2003
temperature (°C)
sample volume (m3)
-2 -5 1 23 23 23 23 24 24 22 22 20 20
647 686 798 199 162 702 224 679 180.8 629.5 165 356 212
compound analyzed PBDE FOSE PBDE & FOSEa PBDE FOSE PBDE & FOSEa PBDE & FOSEa PBDE & FOSEa FOSE PBDE FOSE PBDE FOSE
details semi-urban semi-urban built 1970 built 1994 built 1940 built 1980; basement 4th floor built 1991 built 1995
Only GFF was analyzed.
TABLE 2. Analysis Information for Perfluoroalkyl Sulfonamides (PFASs) ions monitored (m/z) compound
acronym
molecular formula
N-methyl perfluorooctane sulfonamidoethanol MeFOSE C8F17SO2N(CH3)CH2CH2OH N-ethyl perfluorooctane sulfonamidoethanol EtFOSE C8F17SO2N(CH2CH3)CH2CH2OH N-methyl perfluorooctane sulfonamido MeFOSEA C8F17SO2N(CH3)CH2CH2OCOCHdCH2 ethylacrylate
head, turning the pump on and off, and then removing the media and treating as a collected sample. Some locations were sampled twice but at different times and using slightly different sampling approaches. Sampling details are summarized in Table 1. PUF plugs were precleaned by first washing separately with water and then acetone. PUFs were then extracted using a Soxhlet apparatus (18-24 h for each extraction): first with acetone and second with petroleum ether. GFFs were prepared by baking for 18-24 h at 450 °C in individually wrapped aluminum foil sleeves. After baking, the sleeves were sealed and placed in polyethylene bags that were used for storing GFFs. Samples (PUF and GFF) were extracted for 18-24 h using a Soxhlet apparatus. GFFs were extracted separately using dichloromethane (DCM), and PUF plugs were extracted with 1:1 petroleum ether/acetone. For some samples, front and back PUF plugs were extracted separately to assess breakthrough. All extracts were concentrated by rotary evaporation, then further concentrated to about 1 mL using a gentle stream of nitrogen. Samples were split; half in isooctane keeper for PBDE analysis and the other half in ethyl acetate keeper for sulfonamide analysis. The choice of ethyl acetate as keeper was based on previous development work that showed good instrument response for the sulfonamides. Analysis. Analysis of sulfonamides was by gas-chromatography electron impact mass spectrometry (GC-EIMS) using a Hewlett-Packard 6890 GC-5973 mass selective detector MSD in selective ion monitoring (SIM) mode. Sample analysis in negative chemical ionization (NCI) SIM mode was also investigated, where methane was used as reagent gas with flow of 2.2 mL/min. Although this exhibited good sensitivity for standard injections, results were unfavorable for real sample extracts due to interferences. Samples were not taken through cleanup steps. Analytes were separated on a 60 m DB5 column with 0.25 mm i.d. and 0.25 µm film thickness, and helium was used as the carrier gas. The GC oven temperature was 60 °C, 0.5 min, 3 °C/min to 160 °C, then 20 °C/min to 260 °C. Splitless injections were 2 µL with split opened after 0.5 min, and the injector was at 200 °C.
MW (g/mol) 557 571 577
EI (pg)
NCI (pg)
526, 462 483, 400 540, 448 483, 400 526, 462 483, 400
instrumental detection limit EI (pg)
NCI (pg)
1.2 1.2 5.4
3.6 3.7 1.2
The ion source and quadrupole were 230 and 150 °C for EI and 150 and 106 °C, respectively, for NCI analysis. Analysis details are given in Table 2. A total of 10 µL of mirex (10 ng/µL) was used as the internal standard. Response linearity was obtained for over 3 orders of magnitude for sulfonamide compounds; however, standards were included every eight samples to monitor changes in sensitivity. This was especially necessary for MeFOSE and EtFOSE that experienced reduced sensitivity after numerous successive injections; however, no change in response was observed for MeFOSEA. This problem was controlled by frequent maintenance of the injector port. For analysis of PBDEs samples, extracts were first spiked withy [13C12] bromodiphenyl ether surrogate internal standards - congeners 28, 47, 99, 100, 126, 154 and 183 - loaded onto a 30-cm silica column packed with (from bottom to top) four layers of silica gel (2 g of basic, 1 g of neutral, 4 g of acidic, and 1 g of neutral) and Na2SO4 (1 g). The column was eluted with 60 mL of 1:1 DCM/hexane. The eluant was then loaded onto a 30-cm alumina column dry packed with 10 g of neutral alumina and 1 g of Na2SO4 and eluted with 60 mL of 1:1 DCM/hexane. The eluants from the alumina column were concentrated to less than 10 mL and spiked with the 13C12-labeled, method-performance standard ([13C12]BDE 77)prior to congener-specific PBDEs analyses by highresolution gas chromatography/high-resolution mass spectrometry (GC/HRMS). The GC/HRMS system used was a VGAutospec high-resolution mass spectrometer (Micromass, Manchester, U.K.) equipped with a Hewlett-Packard model 5890 series II gas chromatograph and a CTC A200S autosampler (CTC Analytics, Zurich, Switzerland). A 15-m DB5HT column (0.25 mm i.d. × 0.1 µm film thickness) coupled with 1.2 m of precolumn of the same properties was used with UHP-He at 42 kPa and the following temperature program: hold at 100 °C for 1 min; 2 °C/min to 140 °C; 4 °C/min to 220 °C; 8 °C/min to 330 °C; and hold 1.2 min. The splitless injector port, direct GC/HRMS interface, and HRMS ion source were maintained at 300, 275, and 315 °C, respectively. Detailed analytical conditions, the criteria for identification and quantification, and the quality control VOL. 38, NO. 5, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1315
TABLE 3. Physical-Chemical Properties for Perfluoroalkyl Sulfonamides and PBDE Congeners compound
log KOA (20 °C)
A
B
r2
∆Hoa (kJ/mol)
P°S(Pa)a (23 °C)
P°L(Pa)b,c 23 °C)
MeFOSE EtFOSE MeFOSEA PBDE 17 PBDE 28/31 PBDE 47 PBDE 85 PBDE 99 PBDE 100 PBDE 153 PBDE 154 PBDE 183
7.7 7.78 7.87 9.53d 9.73d 10.8d 12.0d 11.6d 11.5d 12.1d 12.2d 12.2d
-4.534 -5.267 -4.88 -3.45 -3.54 -6.47 -6.22 -4.64 -7.18 -5.39 -4.62 -3.71
3570 3811 3724 3803 3889 5068 5331 4757 5459 5131 4931 4672
0.97 0.98 0.98
68.3 72.9 71.2 72.8d 74.5d 973 102d 91.1d 105d 98.2d 94.4d 89.5d
(4.0 ( 0.7) × 10-4 (1.7 ( 0.6) × 10-3 (4.1 ( 0.9) × 10-4
2.0 × 10-3 8.6 × 10-3 2.1 × 10-3 2.2 × 10-3 2.0 × 10-3 1.7 × 10-4 2.8 × 10-5 3.3 × 10-5 9.1 × 10-5 4.0 × 10-6 1.8 × 10-5 4.0× 10-7
a Mean ( one standard deviation of four determinations. b p° values for sulfonamides calculated from p° using eq 2. c p° data for PBDEs L S L determined by GC-retention method from Wong et al. (34). KOA at any temperature calculated using eq 2, log KOA ) A + B/T. d From Harner and Shoeib (33).
measures undertaken for the GC/HRMS analysis of all the analytes of interest are discussed elsewhere (40).
Results and Discussion Quality Control/Quality Assurance. Three to five determinations of KOA were made at each temperature with standard deviations in the range 6-23% (Table 3), which is consistent with previous KOA data (38). The possibility of chemical breakthrough from the C18 absorbent trap used for KOA and vapor pressure determinations was checked by connecting a second trap in series. No breakthrough was observed. Complete extraction of compounds from the trap was tested by spiking with known quantities of target chemical. Two 15-mL extracts of 1:1 acetone/ethyl acetate were collected with none of the target compounds detected in the second extract. Recovery of target compounds in the first extract was >90%. A third extraction was performed with 15 mL of methanol with none of the compounds detected in this third extraction, indicating complete elution in the first 15 mL of extract. Three field blanks were taken to assess method detection limits (MDLs) of target chemicals in PUF and GFF media. Method detection limits (MDL, ng) were calculated as mean blank + 3 SD. MDL values for PUF were 8.2, 3.4, and
