Environ. Sci. Technol. 2007, 41, 3529-3536
Perfluoroalkyl Contaminants in the Canadian Arctic: Evidence of Atmospheric Transport and Local Contamination NAOMI L. STOCK,† VASILE I. FURDUI,† DEREK C. G. MUIR,‡ AND S C O T T A . M A B U R Y * ,† Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada, and Water Science and Technology Directorate, Environment Canada, 867 Lakeshore Drive, Burlington, Ontario L7R 4A6, Canada
Perfluorosulfonates (PFSAs) and perfluorocarboxylates (PFCAs) have been hypothesized to reach remote locations such as the Canadian Arctic either indirectly as volatile precursor chemicals that undergo atmospheric transport and subsequent degradation, or directly via oceanic and atmospheric transport of the PFSAs and PFCAs themselves. Water, sediment, and air samples were collected from three Arctic lakes (Amituk, Char, and Resolute) on Cornwallis Island, Nunavut, Canada. Samples were analyzed for PFSAs and PFCAs, precursor chemicals including the fluorotelomer alcohols (FTOHs) and polyfluorinated sulfonamides (FSAs), and precursor degradation products such as the fluorotelomer unsaturated carboxylates (FTUCAs). PFSAs and PFCAs were detected in water and sediment of all three Arctic lakes (concentrations ranged from nondetect to 69 ng/L and nondetect to 85 ng/g dry weight, respectively). FTOHs and FSAs were observed in air samples (mean concentrations ranged from 2.8 to 29 pg/ m3), and confirm that volatile precursors are reaching Arctic latitudes. The observation of degradation products, including FTUCAs observed in sediment and atmospheric particles, and N-ethyl perfluorooctanesulfonamide (NEtFOSA) and perfluorooctanesulfonamide (PFOSA) in air samples, indicate that degradation of the FTOHs and FSAs is occurring in the Arctic environment. PFSAs and PFCAs were also observed on atmospheric particles (mean concentrations ranged from LOD in 80% of the samples. In terms of frequency, the dominant polyfluorinated sulfonamide was perfluorooctanesulfonamide (F(CF2)8SO2NH2, PFOSA). The mean value of total concentrations of PFOSA was 20 pg/m3, with concentrations as high as 64 pg/m3 on an individual sample. N-methyl perfluorooctanesulfonamidoethanol (NMeFOSE, F(CF2)8SO2N(CH3)CH2CH2OH) was the FSA with the largest mean concentration (29 pg/m3), observed >LOD in 60% of samples. NEtFOSE, NEtFOSA, N-methyl perfluorobutanesulfonamidoethanol (NMeFBSE, F(CF2)4SO2N(CH3)CH2CH2OH), and N-ethyl perfluorobutanesulfonamidoethanol (NEtFBSE, F(CF2)4SO2N(CH2CH3)CH2CH2OH) were observed at similar concentrations (mean values of total concentrations ranged from 11 to 23 pg/m3). Mean concentration of ΣFSAs was 112 pg/m3. Concentrations of FSAs and FTOHs determined in this study are similar to those recently reported by Shoeib et al. (42) who investigated the occurrence of contaminants during a crossing of the North Atlantic and Canadian Archipelago,
and observed the 6:2, 8:2, and 10:2 FTOHs, NMeFOSE, and NEtFOSE at concentrations ranging from below detection limits to 31 pg/m3. An investigation of precursor compounds at six lower latitude locations throughout North America in 2001 (7) reported mean values of total concentrations of ΣFTOHs (6:2, 8:2, and 10:2) ranging from 11 to 165 pg/m3 and ΣFSAs (NEtFOSA, NMeFOSE, and NEtFOSE) ranging from 22 to 403 pg/m3, which are elevated compared to those observed in the current study. To investigate possible sources, 3-day back trajectories were calculated for the air sampling site, using the Canadian Meteorological Centre (CMC) Trajectory Model (43). These trajectories were used to produce a back trajectory probability map (44; Figure S1), useful in identifying those regions from which the parcels of air arriving at the sampling site were most frequently associated. As illustrated in Figure S1, 3-day back trajectories were generally associated with air masses originating from south and southwest of the sampling location and did not have trajectories which overlapped with urban areas in mid-latitude North American or Eurasian sources. PFSAs and PFCAs were also observed on air filter samples (Figure 3a). The dominant contaminant observed was PFOS (mean concentration of 5.9 pg/m3), observed >LOQ in 90% of samples. PFOS concentrations were 1-2 orders of magnitude greater than other detected PFSAs and PFCAs. PFHxS and PFDS were also observed in filter samples (mean concentrations of both 0.2 pg/m3). The dominant PFCA observed was PFOA (mean concentration of 1.4 pg/m3), observed >LOQ in 60% of samples. Longer-chain PFCAs were VOL. 41, NO. 10, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3533
also observed. PFNA and PFDA were both observed at mean concentrations of 0.4 pg/m3, while PFUA, perfluorotridecanoate (PFTriA), and perfluorotetradecanoate (PFTetA) were observed at mean concentrations ranging from 0.02 to 0.06 pg/m3. The frequency of observation of the long-chain PFCAs was typically less than that of PFOA. Interestingly, PFHpA and PFDoA were not observed in any samples, although this may be explained by poor extraction recoveries (Table S2). The 8:2 and 10:2 FTUCA were observed at mean concentrations similar to those observed for the longer-chain PFCAs (0.06 and 0.07 pg/m3, respectively). Concentrations of PFOS detected in the particulate phase above Japan (ranging from nondetect values to 21.8 pg/m3; 27) and Lake Ontario (concentrations up to 8.1 pg/m3; 28) are similar to the mean values reported in this study for Cornwallis Island. It should also be noted that concentrations of PFOA observed in samples from Cornwallis Island are many orders of magnitude less than the concentrations of PFOA (0.1-3.84 µg/m3) observed on filter samples collected outside of a manufacturing facility (29). Implications of Observations. Air sampling results confirm that volatile precursors, such as the FTOHs and FSAs, reach Arctic latitudes. These results are consistent with the PAART hypothesis and recent observations of Shoeib et al. (42). The observation of degradation products in air samples, including FTUCAs, NEtFOSA, and PFOSA, confirm that degradation of FSAs and FTOHs is occurring in the Arctic atmosphere. The presence of PFSAs and PFCAs on atmospheric particles in the Arctic was also confirmed in this investigation and could be due to contributions from both transport hypotheses. The observation of PFDA, PFUA, PFTriA, and PFTetA on arctic atmospheric particles is indicative of atmospheric oxidation. These compounds do not have any significant commercial production (23), and observation presumably is a result of the degradation of precursor compounds such as longer-chain FTOHs (PAART hypothesis). If the presence of these longer-chain PFCAs is due to the atmospheric oxidation of precursors, some proportion of the other PFCAs observed on particles is likely from the same source, as a homologous series of PFCAs have been shown to be formed during the atmospheric oxidation of FTOHs and FSAs (6, 12). It is also possible that marine aerosols moving contamination from the Arctic Ocean may contribute to the presence of PFSAs and PFCAs in Arctic atmospheric particles (direct hypothesis). Although filters were not analyzed for sodium or chloride, the presence of these elements in Amituk, Char, and Resolute Lakes has been established (Table S6), and indicates that sea salt aerosols are likely deposited onto Cornwallis Island. However, concentrations of sodium and chloride vary widely among the lakes on Cornwallis Island, and in addition to sea salt aerosol inputs, concentrations may also be influenced by many factors including local geology and catchment area. Thus, the actual contribution of sea salt is uncertain. A smog chamber study investigating the atmospheric oxidation of FTOHs reported yields of FTCAs and PFCAs of approximately 25 and 1%, respectively (6). A more detailed modeling study on the conversion of 8:2 FTOH to PFOA indicated yields of 1-10% based on location and season (13). No such modeling effort has yet been completed for the fluorotelomer carboxylates (FTCAs). In the current investigation, lower levels of FTCAs (as the FTUCAs) were measured on particles than the corresponding PFCAs. This observation may be due to differential yields of FTCAs and PFCAs under realistic environmental conditions, or that FTCAs/FTUCAs themselves are reactive on atmospheric particles. As noted previously, the extraction methodology used was not optimized for FTUCAs or PFCAs. 3534
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007
In addition to the observation of FTUCAs on atmospheric particles, these degradation products were also observed in water and sediment samples of all three arctic lakes. The observed concentrations of FTUCAs were similar within all three lakes and further confirm the degradation of FTOHs in the arctic environment. Observations made in this investigation also indicate another route by which perfluoroalkyl contaminants reach the Canadian Arctic, and that is local contamination. Concentrations of PFHxS, PFOS, PFHpA, and PFOA in the water and sediment of Resolute Lake are much greater, up to 60-fold higher, than those observed in Amituk or Char Lakes. In addition, concentrations of these contaminants in Resolute Lake are greater than those typically observed in low-latitude environments, such as Lake Ontario. Given the close proximity of the three Arctic lakes, it is assumed that the air samples collected on Cornwallis Island are representative of the ambient air above each lake (Figure S1). As such, the contamination observed in Resolute Lake is likely from a non-atmospheric source. One possibility for the contamination observed in Resolute Lake is the outflow from nearby Meretta Lake (Figure 1). It has been documented that Meretta Lake received raw sewage, and wastewater discharge from 1949 until the early 1990s, from the “North Base” of the Canadian Department of Transport Airport Base, located just south of the current Resolute Airport (45). As a result, Meretta Lake has undergone eutrophication and has historically had significantly higher nutrient levels than any other high arctic lake (45). However, current concentrations of Chlorophyll A in Meretta Lake are not elevated relative to other lakes on Cornwallis Island (Table S6). To investigate this possibility, samples from Meretta Lake and from the outflow of Meretta Lake were collected in the summer of 2005 and analyzed as previously described. Concentrations of perfluoroalkyl contaminants in Meretta Lake and its outflow were similar to those observed for Resolute Lake (Table 1, Table S3). Elevated concentrations of PFHxS and PFOS (ranging from 10.9 to 56.1 ng/L), in addition to PFHpA and PFOA (ranging from 12.9 to 25.5 ng/ L) were observed. As such, it is also possible that perfluoroalkyl contaminants, for example those used in consumer products such as food packaging and textile protectors, could have entered Meretta Lake via the wastewater discharge and subsequently contaminated Resolute Lake. Another possibility for the contamination observed in both Resolute and Meretta Lakes, was use of aqueous fire-fighting foams (AFFF) in the vicinity of these lakes or their watersheds. Following a spill of AFFF from the L.B. Pearson International Airport into Etobicoke Creek, a tributary of Lake Ontario, concentrations of ΣPFSAs (PFHS, PFOS, and PFDS) in the creek ranged from
