Document MJdYpLmwJM5Xq429X83d3MEVy

ELSEVIER Chemosphere 129 (2015) 33-38 Contents l ists available at ScienceDirect Chemosphere journal homepage: www.elsevier.com/locate/chemosphere Chemosphere Stockholm Arlanda Airport as a source of per- and polyfluoroalkyl substances to water, sediment and fish Lutz Ahrens a, Karin Norstrom b'*, Tomas Viktor b, Anna Palm Cousins b, Sarah Josefsson a Dept. of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences (SLU), Box 7050, SE-750 07 Uppsala, Sweden b II/L Swedish Environmental Research Institute, BOX 21060, SE-100 31 Stockholm, Sweden 11) CrossMark ARTICLE INFO Article history: Received 19 December 2013 Received in revised form 18 March 2014 Accepted 22 March 2014 Available online 10 May 2014 Handling Editor: Klaus Kfimmerer Keywords: Per- and polyfluoroalkyl substances (PFASs) Perfluorooctane sulfonate (PFOS) Aqueous fire-fighting foams (AFFFs) Bioconcentration factor (BCF) Sediment/water partition coefficient (Kd) ABSTRACT Fire training facilities are potential sources of per- and polyfluoroalkyl substances (PFASs) to the nearby environment due to the usage of PFAS-containing aqueous fire-fighting foams (AFFFs). The multimedia distribution of perfluoroalkyl carboxylates (PFCAs), perfluoroalkyl sulfonates (PFSAs), perfluorooctanesulfonamide (PFOSA) and 6:2 fluorotelomer sulfonate (FTSA) was investigated near a fire training facility at Stockholm Arlanda Airport in Sweden. The whole body burden of PFASs in European perch (Perca fluviatilis) was 334 80 g absolute and was distributed as follows: Gonad > liver muscle > blood > gill. The bioconcentration factor (BCF) and sediment/water partition coefficient (Ici) increased by O.6-1.7 and 0.2-0.5 log units, respectively, for each additional CF2 moiety for PFCAs and PFSAs. PFAS concentrations in water showed no significant decreasing trend between 2009 and 2O13 (p > 0.05), which indicates that Stockholm Arlanda Airport may be an important source for long-term contamination of the nearby environment with PFASs. 2O14 Elsevier Ltd. All rights reserved. 1. Introduction Aqueous fire-fighting foams (AFFFs) are concentrated liquids designed to extinguish hydrocarbon fuel fires and are used by commercial, public, and military fire-fighting organizations (Paul et al., 2009). Since the 196O5, per- and polyfluoroalkyl substances (PFASs) have been used in AFFFs due to their surface active characteristics to enable the formation of an aqueous film and to resist heat, oil, and water. PFASs have received increasing public attention due to their persistence, bioaccumulative potential, and possible adverse effects on humans and wildlife. As a consequence, perfluorooctane sulfonate (PFOS) and its precursors have been added to the persistent organic pollutants (POPs) list of the Stockholm Convention in May 2009, resulting in global restrictions on its use and production (Paul et al., 2009). After the voluntary phase-out or ban of the C8-based PFASs, such as PFOS, the production shifted to PFAS precursors, due to their higher degradation potential, and to short-chain PFASs (perfluoroalkyl chain length of C < 8), due to their lower bioaccumulation potential (Martin et al., 2003; Moller et al., 2010; Ahrens, 2011). However, PFAS precursors can be (bio)degraded under aerobic and anaerobic conditions to long- and short-chain perfluoroalkyl carboxylates (PFCAs) and perfluoroalkane sulfonates (PFSAs) which are the final * Corresponding author. Tel.: E-mail address: 20. . . orstr6m). http://dx.doi.org/10.1016/j.chemosphere.2014.03.136 0045-6535/ 2014 Elsevier Ltd. All rights reserved. degradation products and extremely persistent in the environment (Rhoads et al., 2008; Wang et al., 2011). Previous studies have linked the usage of AFFFs at fire training facilities to the contamination of the environment with PFASs (Moody et al., 2002, 2003; Awad et al., 2011; Karrman et al., 2011; De Solla et al., 2012). PFOS and 6:2 fluorotelomer sulfonate (FTSA) were the dominant PFASs detected at AFFF contaminated sites with maximum concentrations in the mg L-1 and g g-1 range for water and fish, respectively (Moody et al., 2003; Awad et al., 2011; Karrman et al., 2011; De Solla et al., 2012). Once released into the environment, PFASs are subject to various partitioning, degradation, and transport processes depending on their physicochemical properties and environmental conditions (Ahrens, 2011). The short-chain PFSAs and PFCAs are potentially more water soluble, whereas long-chain PFSAs and PFCAs seem to bind more strongly to particles and accumulate in the food chain (Higgins and Luthy, 2006; Ahrens et al., 2010). However, there is a lack of knowledge of the multimedia distribution of PFASs in the environment. AFFFs are a particularly problematic source of PFASs since high amounts are used in liquid form during a relatively short time period, which increases the potential for the PFASs to be released into the aqueous environment. The aim of this study was to examine the transport and fate of PFASs in water, sediment and fish near a fire training facility at Stockholm Arlanda Airport. The specific objectives were to (i) determine the concentrations and composition profile of various 34 L. Ahrens et al. / Chemosphere 129 (2015) 33-38 PFASs in the multimedia environment (i.e., water, sediment and fish), (ii) to calculate fish bioconcentration factors (BCF) and the sediment/water partition coefficient (Kd) of individual PFASs, and (iii) to investigate the spatial distribution and temporal trends of PFASs in a lake near the fire training facility. PFHxS, PFOS, PFDS), perfluorooctanesulfonamide (PFOSA, C8F17SO2NH2) and 6:2 FTSA (C6F13CH2CH2SO3-) plus [13C4]-PFOS and [13C4]-PFOA as mass-labeled internal standards (IS). 2.4. Analysis 2. Materials and methods 2.1. Sampling sites The study was carried out in an area surrounding Stockholm Arlanda Airport, which is the largest airport in Sweden. Stockholm Arlanda Airport is an international airport located 37 km north- northwest of Stockholm and is used by approximately 19 million people annually (2012). There is a fire training facility located in the north part of the airport area (Fig. 1) where PFOS-containing AFFFs (STHMEX-AFFF 3%, Dr. Richard Sthamer GmbH & Co. KG, Hamburg, Germany) were used freguently from the 1980s until 2003. From 2003, remaining stocks of STHMEX-AFFF 3% were used and then replaced by PFOS-free AFFFs (Presto AFFF and Moussol APS-P, Dr. Richard Sthamer GmbH & Co. KG, Hamburg, Germany); however, these PFOS-free AFFFs still contained other PFASs. According to the manufacturer, the PFOS-free AFFFs contain <10% of PFASs, but further details about the AFFF products are not available. In 2011, Stockholm Arlanda Airport stopped purchasing PFAS-containing AFFFs and a fluorine-free (FF)-AFFF (Moussol FF 3/6, Dr. Richard Sthamer GmbH & Co. KG, Hamburg, Germany) is used instead. 2.2. Sampling Surface water (n = 33), sediment (n = 3) and fish (n = 21, European perch (Perca fluviatilis)) samples were collected at Lake Halmsjn nearby Stockholm Arlanda Airport between 2009 and 2013 (Fig. 1, and Supplementary Tables S1, S2, S3). Perch tissue samples included muscle tissue (n = 18), and blood, liver, muscle, gill and gonad tissues from three individual fish. In addition to the sampling at Lake Halmsjn (sampling site 13, n = 13, 2009- 2013), surface water samples were collected at two reference lakes (sampling sites 1 and 2, n = 2, 2009), artificial ditches downstream of the fire training facility (sampling sites 5-10, n = 6, 2011), and various streams in the area surrounding Stockholm Arlanda Airport (sampling sites 3, 4, 11, 12, 14-18, n = 12, 2011) (for details see Supplementary information). 2.3. Chemicals The target analytes included C6-C11 PFCAs (CnF2n+1COO-, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA), C6, C8, C10 PFSAs (CnF2n+1SO3-, The water samples were extracted by solid phase extraction (SPE) according to the ISO/DIS 25101:2009 method using Oasis WAX cartridges (Waters, 150 mg, 6 mL, 30 lm) (ISO/DIS25101, 2008). The fish tissue and unsieved sediment samples were extracted based on the solid-liquid extraction method described by Powley et al. (2005). The separation and detection of PFASs were performed by high-performance liquid chromatography (HPLC, UFLC Shimadzu, model CBN-20A, Japan) coupled to a mass spectrometer interfaced with an electrospray ionization source in a negative-ion mode (ESI-MS/MS, API 4000, AB Sciex, Foster City, CA, USA). The isotope dilution method was used for quantification. As standard procedure, laboratory blanks, method detection limits (MDLs), and recoveries were examined. Blank concentrations were <1% of the concentrations measured in the samples. MDLs were in the range of tens to hundreds of picograms per liter and grams, respectively. The recoveries in the sample were typically in the range of 70-110% (for details see Supplementary information). 3. Results and discussion 3.1. PFAS composition profiles in water, sediment and fish All 11 target PFASs were detected and quantified in water, sediment and perch tissues (i.e. blood, liver, muscle, gill, gonad). The composition profiles PFASs varied between compartments in Lake Halmsjn (Supplementary Fig. S1). In water, PFOS (36%), PFHxS (29%), and PFOA (20%) were the dominant PFASs, followed by PFHxA (8%). PFOS was even more predominant in sediment and perch tissues with a contribution of 86% and 99%, respectively. This indicates a compound-specific distribution of PFASs in the multi-compartment environment (Martin et al., 2003; Higgins and Luthy, 2006; Ahrens, 2011). 3.2. Tissue distribution of PFASs in perch PFASs were determined in blood, liver, muscle, gill and gonad tissues in perch from Lake Halmsjn. No significant decreasing trend was observed for the PFAS concentrations in muscle tissues between 2009 and 2012 (p > 0.05, ANOVA), and therefore they are discussed as average values in the following. PFOS was the predominant PFAS in all tissue samples. The highest average RPFAS concentrations were detected in blood with 5900 900 ng g-1 ww (n = 3), followed by liver (3900 500 ng g-1 ww, n = 3), gill Fig. 1. Sampling sites for water sampling (1-18) and sampling of water, sediment and European perch (Perca fluviatilis) (H) at Stockholm Arlanda Airport, Sweden. L. Ahrens et al. / Chemosphere 129 (2015) 33-38 35 (2100 700 ng g-1 ww, n = 3), gonad (1600 800 ng g-1 ww, n = 3) and muscle (330 190 ng g-1 ww, n = 21) (Supplementary Fig. S2). Similar to our finding, the PFAS concentrations were highest in blood, followed by kidney > liver > gall bladder > gonads > adipose > muscle tissue in rainbow trouts (Oncorhynchus mykiss) exposed in a flow-through system (Martin et al., 2003). The tissue distribution of individual PFASs in perch varied depending on the functional group and fluorocarbon chain length (Supplementary Fig. S3). Most PFASs were mainly distributed in blood (PFOSA, PFDS and C8-C11 PFCAs with a contribution of 77%, 58%, 46%, 49%, 54% and 54%, respectively). In contrast, 6:2 FTSA, PFOS, and PFHxS were more distributed in gonad tissue (76%, 54%, and 36% respectively). The compoundspecific distribution of PFASs in perch can be influenced by differences in accumulation and elimination of individual compounds (Martin et al., 2003). 3.3. Whole body burden of PFASs in perch The whole body burden distribution was calculated by multiplying the PFAS concentration in the sub-sample with the individual tissue weight. The mean whole body burden in perch of all detected PFASs was estimated to be 334 80 lg absolute (n = 3). PFOS was the predominant compound in all tissue samples with a contribution of 99% (330 80 lg absolute). PFAS whole body burden distribution for the individual tissues decreased in the following order: gonad (54%) > liver (19%) muscle (17%) > blood (7%) > gill (2%) (Fig. 2). More studies are required to investigate the whole body burden in fish and other species (Ahrens et al., 2009) for a better understanding of the bioaccumulation process of PFASs. The relatively high body burden in perch in relation to the total body weight (in average, 1.3 0.3 kg) may have an effect on the physiological functions of the fish. It has been shown that PFOS can be transferred to developing embryos in zebrafish (Danio rerio), where a chronic exposure to PFOS can alter the sex ratio, induce histopathological alterations and have a negative effect on reproduction, embryonic growth, and subsequent offspring development (Du et al., 2009). More ecotoxicological studies are needed to investigate the effects of individual PFASs and their mixtures on individual and community/ecosystem level. Recent studies have found a positive correlation of the consumption of PFAS contaminated freshwater fish and PFAS levels in human blood which indicate a risk for human health (Hlzer et al., 2011; Zhao et al., 2011; Zhang et al., 2011). In 2008, Swedavia Swedish Airports restricted fishing in Lake Halmsjn due to the high PFOS contamination in the lake. 3.4. Fish bioconcentration factors (BCF) and sediment/water partition coefficients (Kd) for PFASs BCF were estimated in various tissues of perch in Lake Halmsjn, based on the concentrations in water and fish tissue. BCFtissue ctissue=cwater 1 where BCFissue is the bioconcentration factor (L kg-1) based on the tissue concentration, ctissue is the concentration of chemical in the tissue of perch (ng kg-1 ww), and cwater is the concentration of the target analyte in water (ng L-1). Since the concentrations differed between tissues for perch, the BCFs also varied. Generally, the highest BCF was found for BCFblood (maximum BCFblood = 65 000), followed by BCFliver, BCFgill and BCFgonad, whereas the BCF was lower in muscle and whole body (Supplementary Table S3). Different BCFs were found for individual PFASs depending on the perfluoroalkyl chain length and functional group (Table 1). Highest average BCFwhole body value was found for PFOS (6400 3600), whereas the C6 chained PFHxS had an average BCFwhole body of only 26 10. For the PFCAs, the BCFwhole body decreased from an average of 181 58 for PFDA to an average of 1.0 0.2 for PFOA. The influence of the functional group on the BCF can be illustrated for three different PFAS classes (i.e., PFCAs, FOSAs, and PFSAs) by comparing the BCFwhole body between PFNA, PFOSA and PFOS, which all have a perfluoroalkyl chain length of C8. The highest average BCFwhole body was found for PFOS (6400 3600), whereas the BCFwhole body for PFOSA was 5 times lower (1200 900) and the BCFwhole body for PFNA was more than two orders of magnitude lower (48 8.8). In general, the log BCFwhole body values increased by 0.6-1.7 log units for each additional CF2 moiety and the bioaccumulation of PFOS was 0.7 log units and 2.1 log units higher compared to PFOSA and PFNA, respectively. Overall, the BCF calculated in this study were in the same range as previously reported in rainbow trout (Martin et al., 2003). However, it has to be considered that perch are carnivorous fish and it is not known whether the water phase or the food is the primary route of PFAS uptake. The distribution of PFASs between water and sediment can be described by the Kd (cm3 g-1). Kd csediment=cwater 2 where csediment is the concentration of chemical in sediment (ng g-1 dry weight (dw)) and cwater is the concentration of the target analyte in water (ng cm-3). Similar as for the BCF, the perfluoroalkyl chain length and functional group had an influence on the Kd values (Table 1). The log Kd for PFOS was 2.3 0.3, whereas the C6 chained PFHxS had an average log Kd of only 1.3 0.4. For the PFCAs, the log Kd decreased from an average of 1.2 0.3 for PFOA to an average of 0.8 0.3 for PFHxA. In general, the log Kd values increase by 0.2-0.5 log units for each additional CF2 moiety. The highest log Kd was calculated for PFOSA with an average of 3.1 0.1, whereas the log Kd of the C8 chained PFOS was 0.8 log units lower (2.3 0.3). These findings are in agreement with previous reported Kd values (Higgins and Luthy, 2006; Ahrens et al., 2010), but it is important to bear in mind that the Kd values can be influenced by environmental factors such as the organic carbon content of sediment, pH and salinity of water (Higgins and Luthy, 2006). 3.5. Spatial distribution of PFASs in water around Stockholm Arlanda Airport Fig. 2. PFAS whole body burden distribution in percent and lg per tissue in brackets for European perch (Perca fluviatilis) collected from Lake Halmsjn. The spatial distribution of PFASs was investigated in water around Stockholm Arlanda Airport (Fig. 3). RPFAS concentrations 36 L. Ahrens et al. / Chemosphere 129 (2015) 33-38 Table 1 Whole body bioconcentration factors (BCF, L kg-1) of European perch (Perca fluviatilis) and sediment/water partition coefficients (Kd, cm3 g-1) for PFASs in Lake Halmsjn near the Stockholm Arlanda Airport (average values and ranges are given in parentheses; N/A = not available). PFHxA PFOA PFNA PFDA PFHxS PFOS PFOSA BCFwhole body N/A 1.0 (0.9-1.1) 48 (42-54) 181 (140-220) 26 (14-33) 6400 (3200-10 200) 1200 (170-1800) Log BCFwhole body N/A 0.0 (0.07-0.05) 1.7 (1.6-1.7) 2.3 (2.2-2.4) 1.4 (1.2-1.5) 3.8 (3.5-4.0) 3.1 (2.2-3.3) Kd 8.3 (4.2-16) 17 (9.7-30) N/A N/A 23 (10-48) 240 (120-450) 1300 (1000-1700) Log Kd 0.8 (0.6-1.2) 1.2 (1.0-1.5) N/A N/A 1.3 (1.0-1.7) 2.3 (2.1-2.7) 3.1 (3.0-3.2) showed a high variability ranging from a few nanograms to several thousands nanograms per liter depending on the location. The highest PFAS concentrations were found in a ditch close to the fire training facility with 4000 ng L-1 for RPFASs (sampling site 5, see Figs. 1 and 4). PFOS (2340 ng L-1) was the predominant compound at sampling site 5, followed by PFHxS (980 ng L-1), PFHxA (290 ng L-1), and PFOA (210 ng L-1). The PFAS levels around Stockholm Arlanda Airport were comparable with those found in seepage and surface water contaminated with AFFFs, which ranged from hundreds to a few thousand nanograms per liter (Krrman et al., 2011; De Solla et al., 2012). The PFOA concentration typically exceed the PFOS concentration in the aqueous environment (Ahrens, 2011). The PFOS concentration in this study, however, was consistently higher than PFOA with an average PFOS/PFOA ratio of 6.5 4.0, which confirms the presence of a local contamination source of PFOS in this area (Naile et al., 2010). RPFAS concentration in the two reference lakes (sampling sites 1 and 2) and two reference streams (sampling sites 3 and 14) were much lower (<10 ng L-1) than at the contaminated sampling site 5 (4000 ng L-1) indicating that the Stockholm Arlanda Airport is the dominating source of PFASs in this region. With increasing distance from the fire training facility, the RPFAS Fig. 4. Temporal trends of PFASs (with a contribution >5%) in water in Lake Halmsjn south of Stockholm Arlanda Airport between 2009 and 2013. concentrations decreased along the stream towards the outlet into the Lake Mlaren (sampling site 18) by a factor of 10-50, i.e. from 4000 to a few hundred nanograms per liter. The same decrease was observed for individual PFASs (Fig. 3). The decreasing water concentrations with increasing distance from the source can be explained by the dilution effect and the partitioning as described in the previous chapter 3.4. Fig. 3. Spatial distribution of PFASs in water around Stockholm Arlanda Airport, Sweden (note: nd = not detected; na = not available). L. Ahrens et al. / Chemosphere 129 (2015) 33-38 37 PFOS water concentrations at sampling sites 5, 9, and 10 were above the Estimated No-Effects Value (ENEV) of 491 ng L-1 for aquatic organisms (used by Environment Canada and calculated from the NOEC of 49 100 ng L-1 for Chironomus tentans (MacDonald et al., 2004) divided by the safety factor of 100). In addition, for almost all sampling sites (i.e., sampling sites H, 1, 3-13, and 15-18) PFOS concentrations were above the proposed Annual Average Environmental Quality Standard (AA-EQS) of 0.65 ng L-1, but always below the proposed Maximum Acceptable Concentration (MAC-EQS) of 36 000 ng L-1 under the Water Framework Directive (WFD) of the European Union (EU) (EU, 2011). To estimate the yearly PFAS fluxes (kg year-1) from the catch- ment to Lake Mlaren, the PFAS concentration at the river outlet to Lake Mlaren (sampling site 18) was multiplied by the yearly water flow of 18 million m3 year-1 (Water Information System Sweden (VISS), http://www.viss.lansstyrelsen.se/). It should be noted that the estimated yearly PFAS fluxes are subject to uncer- tainties due to the possible variations in PFAS concentrations in the river throughout the year, which was not covered in our sam- pling. 5.3 kg The yearly year-1 for PPFAPSFAflSusx,esdiwviedreedesitnimtoat2e.d6 to be approximately kg year-1 for PFOS, 1.7 kg year-1 for PFHxS, 0.6 kg year-1 for PFHxA, and 0.4 kg year-1 for PFOA. The PFAS fluxes estimated in this study were consider- ably lower than the PFAS fluxes estimated for the major European rivers (McLachlan et al., 2007; Mller et al., 2010), which is not suprising considering the relatively low water flow of the catchment. 3.6. Temporal trends of PFASs in water Temporal and seasonal trends were investigated in Lake H(2a0l0m9s-j2n01(3s)a. mPplPiFnAgSssitrean1g3e)d over a from time period of 146-344 ng L-1 five with years PFOS (59-137 ng L-1), PFHxS (54-104 ng L-1), PFOA (13-80 ng L-1), and PFHxA (8.2-34 ng L-1) as the dominant compounds (Fig. 4). PFAS concentrations in the two reference lakes (sampling sites 1 and 2) north of Stockholm Arlanda Airport were two orders of magni- tude lower, clearly indicating a contamination of Lake Halmsjn from the airport. However, no direct stream connects the fire train- ing facility with Lake Halmsjn, indicating a sub-surface transport of PFASs into the lake. It is interesting to note that the PFAS com- position profile close to the point source at the fire training facility (sampling site 5) was different compared to Lake Halmsjn. The contribution of PFOS decreased (60% vs. 40%), whereas the contri- bution of PFHxS (24% vs. 32%) and PFOA (5% vs. 19%) increased in Lake Halmsjn compared to sampling site 5. The changed compo- sition profile can be explained by the higher sorption potential of PFOS to particles compared to PFHxS and PFOA (Higgins and Luthy, 2006). 6:2 FTSA displayed highest concentrations near the fire training facility (176 ng L-1, sampling site 5), but was not detected in Lake Halmsjn. It is likely that 6:2 FTSA is biotrans- formed to primarily 5:3 acid (F(CF2)5CH2CH2COOH), PFPA, and PFHxA during the sub-surface transport into Lake Halmsjn (Wang et al., 2011). Although no PFOS- or PFAS-containing AFFFs were purchased after 2003 and 2011, respectively, the PFAS concentrations in Lake Halmsjn showed no significant decreasing trend between 2009 and 2013 (p > 0.05, ANOVA) (Fig. 4). It is interesting to note that not even PFOS showed a significant decreasing temporal trend (frpo>m02.0050,3A. NItOisViAn)tedreessptiintegtthoenpohteasteh-aotuttheofaPvFeOraSg-ecoPntPaFiAnSincgoAncFeFnFstrations were slightly higher in summer (296 ng L-1) compared to spring, autumn and winter periods (254 ng L-1, 209 ng L-1 and 272 ng L-1, respectively) (Fig. 4). The low water flow conditions during the dry summer season (0.3 m3 s-1) may diminish the dilution capacity of the river compared to the spring, autumn and winter periods (0.7-0.9 m3 s-1). However, the PFAS concentrations were relatively constant over time and no significant seasonal trend was observed (p > 0.05, ANOVA). This indicates that PFASs are still leaching out from the contaminated soil or are slowly transported from contaminated groundwater at the fire training facility. Previous studies have reported elevated PFAS concentrations even after several years of decommissioning of the fire training facility (Moody et al., 2002, 2003). Ultimately, the continued presence of PFASs in Lake Halmsjn may reflect a long-term impact of the AFFF contaminated fire training facility on the nearby environment. 3.7. Implications PFASs are transported from the contaminated area at Arlanda airport through the watercourse towards Lake Mlaren. The con- tamination of Lake Mlaren is of concern since the water is used by 2 million people in the Stockholm area as drinking water. TPhPeFAouStflgouixnegs fluxes into Lake are estimated to Mlaren be 5.3 from Arlanda airport of kg year-1, of which PFOS represents approximately 50% (2.6 kg year-1). This can be com- pared to the estimated atmospheric deposition of PFOS to Lake Mlaren (1.5-5.3 kg year-1), based on the measured concentration in precipitation on the Swedish west coast (2.2-7.7 ng L-1) (Woldegiorgis et al., 2006), the average annual precipitation rate at Arlanda (0.6 m3 m-2, average 1979-2012, SMHI database, http://luftwebb.smhi.se/) and the total surface area of Lake Mlaren (1140 km2). PFASs may also enter the lake through other pathways, such as riverine inflow and via stormwater and effluent water from municipal and private treatment plants, but the magni- tude of such releases is not known. PFAS concentrations in the water phase of Lake Halmsjn did not show a decreasing trend over a time period of five years (2009-2013) which indicates that Stockholm Arlanda Airport may be an important source for long-term contamination of the nearby environment with PFASs. PFASs can sediment or bioaccu- mulate in the food chain depending on their chain length and func- tional group (see field-based Kd and BCF in Table 1). Generally, shorter-chained PFASs (perfluoroalkyl chain length of C < 8) are mainly transported in the water phase, whereas longer-chained PFASs (perfluoroalkyl chain length of C P 8) have a higher poten- tial to partition to particles and accumulate in the food chain. More studies on the whole body burden of PFASs in different species are needed for a better toxicology and risk assessment. In addition, the ecotoxicological and human impact of other PFASs (e.g., shorter chain PFASs) needs to be investigated (Mller et al., 2010) and more research on the multimedia distribution of PFASs is needed for a better understanding of the fate of PFASs in the environment. PFAS-containing AFFFs are used worldwide by commercial, public, and military fire-fighting organizations. Although most airports are now using PFOS-free AFFFs, many of these AFFF products still con- tain high amounts of PFASs. Alternative fluorine-free FF-AFFFs are available to be used against hydrocarbon-fuel fires (Schaefer et al., 2008), but due to the lack of legal restrictions, PFAS-containing AFFFs are still being used or stockpiled and pose a potential threat for the environment, for example for human drinking water reservoirs. Acknowledgements The authors gratefully acknowledge Swedavia Swedish Airports and the Foundation for IVL Swedish Environmental Research Institute (SIVL) for funding this research. SIVL is funded by grants from the Swedish Research Council FORMAS and the Swedish Environmental Protection Agency. We also thank Swedavia Swedish Airports for access to the sampling sites. 38 L. Ahrens et al. / Chemosphere 129 (2015) 33-38 Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.03.136. References Ahrens, L., 2011. Polyfluoroalkyl compounds in the aquatic environment: a review of their occurrence and fate. J. Environ. Monitor. 13, 20-31. Ahrens, L., Siebert, U., Ebinghaus, R., 2009. Temporal trends of polyfluoroalkyl compounds in harbor seals (Phoca vitulina) from the German Bight, 1999-2008. Chemosphere 76, 151-158. Ahrens, L., Taniyasu, S., Yeung, L.W.Y., Yamashita, N., Lam, P.K.S., Ebinghaus, R., 2010. Distribution of polyfluoroalkyl compounds in water, suspended particulate matter and sediment from Tokyo Bay, Japan. Chemosphere 79, 266-272. Awad, E., Zhang, X., Bhavsar, S.P., Petro, S., Crozier, P.W., Reiner, E.J., Fletcher, R., Tittlemier, S.A., Braekevelt, E., 2011. Long-term environmental fate of perfluorinated compounds after accidental release at Toronto airport. Environ. Sci. Technol. 45, 8081-8089. De Solla, S.R., De Silva, A.O., Letcher, R.J., 2012. Highly elevated levels of perfluorooctane sulfonate and other perfluorinated acids found in biota and surface water downstream of an international airport, Hamilton, Ontario, Canada. Environ. Int. 39, 19-26. Du, Y., Shi, X., Liu, C., Yu, K., Zhou, B., 2009. Chronic effects of water-borne PFOS exposure on growth, survival and hepatotoxicity in zebrafish: a partial life-cycle test. Chemosphere 74, 723-729. EU, 2011. Perfluorooctanat sulphonate (PFOS). PFOS EQS dossier 2011. Higgins, C.P., Luthy, R.G., 2006. Sorption of perfluorinated surfactants on sediment. Environ. Sci. Technol. 40, 7251-7256. Hlzer, J., Gen, T., Just, P., Reupert, R., Rauchfuss, K., Kraft, M., Mller, J., Wilhelm, M., 2011. Perfluorinated compounds in fish and blood of anglers at Lake Mhne, Sauerland area, Germany. Environ. Sci. Technol. 45, 8046-8052. ISO, 2008. Water quality - determination of perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) - method for unfiltered samples using solid phase extraction and liquid chromatography/mass spectrometry, ISO/DIS25101 (ISO/ TC147/SC2/WG56). Krrman, A., Elgh-Dalgren, K., Lafossas, C., Mskeland, T., 2011. Environmental levels and distribution of structural isomers of perfluoroalkyl acids after aqueous firefighting foam (AFFF) contamination. Environ. Chem. 8, 372-380. MacDonald, M.M., Warne, A.L., Stock, N.L., Mabury, S.A., Solomon, K.R., Sibley, P.K., 2004. Toxicity of perfluorooctane sulfonic acid and perfluorooctanoic acid to Chironomus tentans. Environ. Toxicol. Chem. 23, 2116-2123. Martin, J.W., Mabury, S.A., Solomon, K.R., Muir, D.C.G., 2003. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 22, 196-204. McLachlan, M.S., Holmstrom, K.E., Reth, M., Berger, U., 2007. Riverine discharge of perfluorinated carboxylates from the European continent. Environ. Sci. Technol. 41, 7260-7265. Mller, A., Ahrens, L., Surm, R., Westerveld, J., Van Der Wielen, F., Ebinghaus, R., De Voogt, P., 2010. Distribution and sources of polyfluoroalkyl substances (PFAS) in the River Rhine watershed. Environ. Pollut. 158, 3243-3250. Moody, C.A., Martin, J.W., Kwan, W.C., Muir, D.C.G., Mabury, S.A., 2002. Monitoring perfluorinated surfactants in biota and surface water samples following an accidental release of fire-fighting foam into Etobicoke Creek. Environ. Sci. Technol. 36, 545-551. Moody, C.A., Hebert, G.N., Strauss, S.H., Field, J.A., 2003. Occurrence and persistence of perfluorooctanesulfonate and other perfluorinated surfactants in groundwater at a fire-training area at Wurtsmith Air Force Base, Michigan, USA. J. Environ. Monitor. 5, 341-345. Naile, J.E., Khim, J.S., Wang, T., Chen, C., Luo, W., Kwon, B.-O., Park, J., Koh, C.-H., Jones, P.D., Lu, Y., Giesy, J.P., 2010. Perfluorinated compounds in water, sediment, soil and biota from estuarine and coastal areas of Korea. Environ. Pollut. 158, 1237-1244. Paul, A.G., Jones, K.C., Sweetman, A.J., 2009. A first global production, emission, and environmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol. 43, 386-392. Powley, C.R., George, S.W., Ryan, T.W., Buck, R.C., 2005. Matrix effect-free analytical methods for determination of perfluorinated carboxylic acids in environmental matrixes. Anal. Chem. 77, 6353-6358. Rhoads, K.R., Janssen, E.M.-L., Luthy, R.G., Criddle, C.S., 2008. Aerobic biotransformation and fate of n-ethyl perfluorooctane sulfonamidoethanol (n-EtFOSE) in activated sludge. Environ. Sci. Technol. 42, 2873-2878. Schaefer, T.H., Dlugogorski, B.Z., Kennedy, E.M., 2008. Sealability properties of fluorine-free fire-fighting foams (FfreeF). Fire Technol. 44, 297-309. Sveriges Meteorologiska och Hydrologiska Institut (SMHI). <http:// luftwebb.smhi.se/> (26.10.13). Wang, N., Liu, J., Buck, R.C., Korzeniowski, S.H., Wolstenholme, B.W., Folsom, P.W., Sulecki, L.M., 2011. 6:2 Fluorotelomer sulfonate aerobic biotransformation in activated sludge of waste water treatment plants. Chemosphere 82, 853-858. Water Information System Sweden (VISS). <http://www.viss.lansstyrelsen.se/> (26.10.13). Woldegiorgis, A., Andersson, J., Remberger, M., Kaj, L., Ekheden, Y., Blom, L., Brorstm-Lundn, E., Borgen, A., Dye, C., Schlbach, M., 2006. Results from the Swedish National Screening Programme 2005. Subreport 3: Perfluorinated alkylated substances (PFAS). IVL Report B1698, 1-46. Zhang, T., Sun, H., Lin, Y., Wang, L., Zhang, X., Liu, Y., Geng, X., Zhao, L., Li, F., Kannan, K., 2011. Perfluorinated compounds in human blood, water, edible freshwater fish, and seafood in China: daily intake and regional differences in human exposures. J. Agric. Food Chem. 59, 11168-11176. Zhao, Y.G., Wan, H.T., Law, A.Y.S., Wei, X., Huang, Y.Q., Giesy, J.P., Wong, M.H., Wong, C.K.C., 2011. Risk assessment for human consumption of perfluorinated compound-contaminated freshwater and marine fish from Hong Kong and Xiamen. Chemosphere 85, 277-283. ) ARTICLE pubs.acs.org/est Wastewater Treatment Plant and Landfills as Sources of Polyfluoroalkyl Compounds to the Atmosphere Lutz Ahrens,*, Mahiba Shoeib,*, Tom Harner, Sum Chi Lee, Rui Guo,,|| and Eric J. Reiner,|| Environment Canada, Science and Technology Branch, Toronto, ON, Canada, M3H 5T4 University of Toronto, Department of Chemistry, Toronto, ON, Canada, M5S 3H6 Ontario Ministry of the Environment, Toronto, ON, Canada, M9P 3V6 bS Supporting Information ABSTRACT: Polyfluoroalkyl compounds (PFCs) were determined in air around a wastewater treatment plant (WWTP) and two landfill sites using sorbent-impregnated polyurethane foam (SIP) disk passive air samplers in summer 2009. The samples were analyzed for five PFC classes (i.e., fluorotelomer alcohols (FTOHs), perfluorooctane sulfonamides (FOSAs), sulfonamidoethanols (FOSEs), perfluoroalkyl sulfonic acids (PFSAs), and perfluoroalkyl carboxylic acids (PFCAs)) to investigate their concentration in air, composition and emissions to the atmosphere. PFC concentrations in air were 315 times higher within the WWTP (228024 040 pg/m3) and 530 times higher at the landfill sites (278026 430 pg/m3) compared to the reference sites (5971600 pg/m3). Variations in the PFC pattern were observed between the WWTP and landfill sites and even within the WWTP site. For example, FTOHs were the predominant PFC class in air for all WWTP and landfill sites, with 6:2 FTOH as the dominant compound at the WWTP (89512 290 pg/m3) and 8:2 FTOH dominating at the landfill sites (129017 380 pg/m3). Furthermore, perfluorooctane sulfonic acid (PFOS) was dominant within the WWTP (43171 pg/m3), followed by perfluorobutanoic acid (PFBA) (55116 pg/m3), while PFBA was dominant at the landfill sites (101102 pg/m3). It is also noteworthy that the PFCA concentrations decreased with increasing chain length and that the emissions for the even chain length PFCAs outweighed emissions for the odd chain length compounds. Furthermore, highly elevated PFC concentrations were found near the aeration tanks compared to the other tanks (i.e., primary and secondary clarifier) and likely associated with increased volatilization during aeration that may be further enhanced through aqueous aerosol-mediated transport. PFC yearly emissions estimated using a simplified dispersion model were 2560 g/year for the WWTP, 99 g/year for landfill site 1, and 1000 g/year for landfill site 2. These results highlight the important role of WWTPs and landfills as emission sources of PFCs to the atmosphere. ' INTRODUCTION In recent years, the input of polyfluoroalkyl compounds (PFCs) to the atmosphere and their fate has been recognized as one of the emerging issues in environmental chemistry. PFCs comprise a diverse group of chemicals that have been widely used as processing additives during fluoropolymer production and as surfactants in consumer applications, including stain repellents in textile, furniture and paper products for over 50 years.1 During production and usage, PFCs can be released into the environment,2,3 where they have been found to be ubiquitous in water,4 air,5,6 sediment,7 and organisms.8 Once released in the environment, fluorotelomer alcohols (FTOHs), perfluorooctane sulfonamides (FOSAs) and perfluorooctane sulfonamidoethanols (FOSEs) can be degraded in the atmosphere or under aerobic conditions to perfluoroalkyl carboxylic acids (PFCAs) and perfluoroalkyl sulfonic acids (PFSAs).911 PFCAs and PFSAs are resistant to typical environmental degradation processes and PFSAs and longer chain PFCAs are known Part of the Perfluoroalkyl Acid Special Issue to be bioaccumulative12 and have possible adverse effects on humans and wildlife.13,14 They can be transported by ocean currents4 and atmosphere15 and their volatile precursors (i.e., FTOHs, FOSAs, FOSEs) can undergo long-range transport via the atmosphere.5,6,16 As a result, perfluorooctane sulfonic acid (PFOS) has been added to the persistent organic pollutants (POPs) list of the Stockholm Convention in May 2009 resulting in global restrictions on its uses and production.17 Based on the production volume, historic emissions are estimated to be 680045 300 t for perfluorooctylsulfonyl fluoride (POSF) (19722002).3 The majority of the emissions were estimated to have been released to the aqueous environment (45 000 t) and Special Issue: Perfluoroalkyl Acid Received: Accepted: Revised: Published: October 26, 2010 March 28, 2011 March 8, 2011 April 05, 2011 r 2011 American Chemical Society 8098 dx.doi.org/10.1021/es1036173 | Environ. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology only a small amount into the air (235 t).3 The emissions for PFCAs range between 3200 and 7300 t (19512004), from which the indirect emissions (PFCA impurities and/or precursors) were estimated to be 15% of the total source emissions.2 A recent study indicates that the aqueous phase can act as a net source for perfluorooctanoic acid (PFOA) to the atmosphere;18 however, field measurements of PFOA in the atmosphere are rare. Industrial or municipal wastewater treatment plants (WWTPs) are mostly discussed as point sources for PFCs into the aqueous environment.19,20 WWTP mass flow studies found similar or higher PFC concentrations in the effluent in comparison to the influent, indicating that conventional WWTPs are not effective for removing PFCs. In fact, biodegradation of precursor compounds within the WWTP could actually increase concentrations of PFCAs and PFSAs.19,20 Recently, landfill leachate was also identified as a potential point source to the aquatic environment.21 However, these studies focused on the discharge into the aquatic environment, whereas to our knowledge, there are no data available on the emissions of PFCs to the atmosphere from this potential source. The aim of this study was to examine atmospheric PFC concentrations at numerous sites on and around one WWTP, and upwind and on-site at two solid waste landfills in Ontario, Canada, using passive air samplers (PAS). The specific objectives of this study include (i) to determine if WWTPs and landfills are important emission sources of PFCs to air by measuring five PFC classes (i.e., FTOHs, FOSAs, FOSEs, PFSAs, PFCAs) in the atmosphere at the WWTP and landfill sites, (ii) to investigate the composition of PFCs in these samples in order to identify sectorspecific differences, and (iii) to model and quantify for the first time, emissions of PFCs to air from a WWTP and landfills. ' EXPERIMENTAL SECTION Sampling. A new type of PAS comprising sorbent-impregnated polyurethane foam (SIP) disks was deployed in two independent field campaigns to assess PFCs in air from a WWTP and two landfill sites. Details on the preparation of SIP disk samplers have been previously reported.22 Briefly, precleaned polyurethane foam (PUF) disks (14 cm diameter 1.35 cm thick; surface area 365 cm2, mass 4.40 g, volume 207 cm3, Tisch Environmental, Cleves, OH) were impregnated with finely ground XAD-4 resin (Supelco, Bellefonte, PA) (0.5 g per PUF disk). During field deployment, the SIP disks were individually housed inside precleaned stainless steel chambers (model TE-200-PAS, Tisch Environmental) and deployed 2 m above the ground/ water surface. This new PAS type improves the sorptive capacity of the conventional PUF disk for more volatile and polar chemicals such as PFCs.22,23 To assess emissions of PFCs to air from the WWTP, SIP disks PAS (n = 12) were deployed for 63 days around a municipal WWTP in Ontario, Canada, between July and September 2009. The samplers were deployed at the primary clarifier (sites 4 and 5), aeration tank (sites 69), secondary clarifier (sites 10 and 11), which were all open to the atmosphere, and at four nearby reference sites (sites 13 and 12) (Figure 1 and Figure S1 in the Supporting Information (SI)). In terms of process flow, the wastewater which is first screened to remove large debris begins at the primary clarifiers, where solids separate from the wastewater through gravity sedimentation. From there, the liquid effluent goes through an aeration stage where activated sludge (microorganisms/bacteria) is used to degrade the organic ARTICLE Figure 1. Sampling sites on and around the wastewater treatment plant. Prevailing wind direction is indicated by the red arrow. matter. Air is injected through the bottom of the aeration tanks to enhance microbial activity. This introduces turbulence to the water surface and bubbling/aqueous aerosol emission. The aeration tank effluent then goes through a secondary clarifier, where the bacteria and remaining particles undergo further settling from the wastewater. The secondary clarifier effluent is then passed through a closed chlorination process (disinfection) prior to discharge to the lake. Four field blanks were collected by placing the SIP disks in the sampler housing and then removing them after 1 min. To check reproducibility, duplicate samples were collected at sites 8 and 9, between two aerations tanks (5 m from each tank). To assess emissions to air from landfills, SIP disks PAS were deployed for 55 days at two municipal solid waste landfill sites in Ontario, Canada, between June and August 2009. The samples were deployed upwind and on-site of the active zone of each landfill site (n = 4) and one field blank was collected at each site as described earlier (Figure S2 in the SI). It should be noted that both sites collected landfill gas and the active area of the landfill was kept to a minimum by covering the waste with soil and a plastic film. Details of the meteorological data during the sampling period of the WWTP and landfill sites can be found in Table S1 and Figure S3 and S4 in the SI. All SIP disk samples and field blanks were stored in amber glass jars having aluminum-lined lids at 20 C until analysis. Deriving Concentrations in Air. The individual air concentrations derived from the SIP disks were calculated as previously described.22,23 The air sample volume is chemical specific and based on the passive samplerair partition coefficients (KSIPA) for each chemical and the kinetic sampling rate. The sample volumes for the 5563 day deployment at the WWTP were calculated to be 106114 m3 for FTOHs and 188236 m3 for FOSAs and FOSEs, respectively. The lower value for the FTOHs reflect their lower KSIPA and the fact that the SIP disk reaches saturation (maximum capacity) for the FTOHs. For the PFCAs and PFSAs, an average sampling rate of 4 m3/ day was used.23 It is important to note that the calculation of the air sample volume is only an estimation based on the previous calibration of the samplers. The sample air volume is a function of deployment time, average temperature and KSIPA. Recent field intercomparison studies have shown a good agreement between SIP disk PAS and high-volume samplers, with the exception of PFOS which showed that concentration was underestimated (i.e., sample air volume overestimated) using SIP disk PAS.24,25 The existing passive sampler configuration results in the collection of mainly gas-phase contaminants.22 For the particulate 8099 dx.doi.org/10.1021/es1036173 |Environ. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology ARTICLE Figure 2. (A) FTOHs, (B) FOSAs, and FOSEs in pg/m3 measured in air at the wastewater treatment plant and the two landfill sites using SIP disk air samplers. phase sampling rate, we referred to the study by Klanova et al.26 which showed this to be 10% of the gaseous phase rate. One major advantage of the PAS is that it performs time-integrated sampling which provides the "average" air concentration over the deployment period. This is especially useful in cases of air concentrations that fluctuate greatly with time, since short-term high volume air samplers that are collected typically for several hours to a day may not adequately represent the average scenario. Chemicals. The target analytes include 22 PFCs (i.e., C4C12, C14 PFCAs, C4, C6, C8, C10 PFSAs, 6:2, 8:2, 10:2 FTOH, perfluorooctane sulfonamide (PFOSA), methyl and ethyl FOSA and methyl and ethyl FOSE) plus 16 mass-labeled internal standards (IS) and N,N-dimethylperfluoro-1-octanesulfonamide (N,N-Me2FOSA), 13C8 PFOS, and 13C8 PFOA as injection standards (InjS) (for details see Table S2 in the SI). Sample Extraction and Instrumental Analysis. SIP disks were extracted as described previously.23 Briefly, prior the extraction the SIP disks were spiked with 5 ng and 50 ng absolute of an IS mixture containing mass-labeled PFSAs, PFCAs and FTOHs, FOSAs, FOSEs, respectively. SIP disks were Soxhlet extracted with petroleum ether/acetone (50/50, v/v) for 18 h, followed by a 4 h extraction with methanol. The petroleum ether/acetone and methanol extracts were concentrated by rotary evaporation followed by gentle nitrogen blow down to 0.5 and 1 mL, respectively. The methanol extract was further purified using the dispersive clean up with ENVI-Carb (100 mg, 1 mL, 100400 mesh, Supelco, St. Louis, MO) and glacial acetic acid.27 After centrifugation 0.5 mL was transferred into a polypropylene vial. Prior to injection, 10 ng of N,N-Me2FOSA and 4 ng of 13C8 PFOS and 13C8 PFOA were added to the petroleum ether/acetone and methanol extract as InjS, respectively. Analysis was performed using gas chromatography (Agilent 7890A; Agilent Technologies, Palo Alto, CA)mass spectrometry (Agilent 5975C; Agilent Technologies, Palo Alto, CA) (GC/MS) in selective ion monitoring (SIM) mode using positive chemical ionization (PCI) for the FTOHs, FOSAs and FOSEs.5 Aliquots of 2 L were injected on a DB-WAX column (30 m, 0.25 mm inner diameter, 0.25 m film, J&W Scientific, Folsom, CA). The separation and detection of the PFCAs, PFSAs, and PFOSA was performed by liquid chromatography (Agilent 1100; Agilent Technologies, Palo Alto, CA) with triple quadrupole mass spectrometer interfaced with an electrospray ionization source in negative-ion mode (LC()ESIMS/MS; API 4000, Applied Biosystems/MDS SCIEX, Foster City, CA). Aliquots of 25 L were injected on a Luna C8(2) 100A column (50 2 mm, 3 m particle size; Phenomenex, Torrance, CA) using a gradient of 200 L/min methanol and water (both with 10 mM aqueous ammonium acetate solution (NH4OAc)). QA/QC. As part of the QA/QC, duplicate measurements (sites 8 and 9) and field blanks were evaluated. Field blank concentrations (n = 6) were <1% of the concentrations measured in the samples. Method detection limits (MDLs) were calculated as 3 times signal-to-noise for each compound in the actual air samples and ranged from 4.043.0 pg/m3 for the FTOHs, FOSAs and FOSEs using GC/PCIMS and from 0.040.87 pg/m3 for PFCAs, PFSAs, and PFOSA using LC()ESIMS/MS. Method recovery values calculated from mass-labeled IS spiked before injection and InjS spiked before injection ranged from 55 to 163% for FTOHs, FOSAs, and FOSEs and from 41100% for the PFCAs, PFSAs, and PFOSA. All results were recovery corrected for PFCs. Further details on the surrogates chosen for each native compound have been previously reported.4,22 Duplicate measurements showed excellent agreement and very good reproducibility with a mean standard deviation of 5% for the PFCs. In addition to our QA/QC protocol, SIP disks PAS were evaluated in an international field intercomparison against active (high-volume) samplers, which showed generally good agreement between both types of samplers.25 ' RESULTS AND DISCUSSION WWTP Concentrations. Concentrations of PFCs in air around a WWTP using SIP disk PAS are shown in Figure 2 and 3 (for details see Table S3 and S4 in the SI). 8100 dx.doi.org/10.1021/es1036173 |Environ. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology ARTICLE Figure 3. PFOS and PFCAs in pg/m3 measured in air at the wastewater treatment plant and the two landfill sites using SIP disk air samplers. Note: The sample from WWTP site 7 was lost during sample treatment. Based on other work,24 the derived air concentration for PFOS is likely to be underestimated using the SIP disk sampler. Air concentrations at reference sites (sites 13 and 12) of FTOHs, FOSAs, and FOSEs were usually in the range as reported for urban areas.5,28 However, FTOH concentrations were 210 times higher at the sites 1 and 2 (8561190 pg/m3) compared to previous measurements in Toronto (50123 pg/ m3)5 and Hamburg, Germany (179531 pg/m3).28 This may reflect proximity to the WWTP itself (Figure 1). FTOH concentrations were lower at the more distant reference sites (sites 3 and 12) (304504 pg/m3) in comparison to the sites 1 and 2. Overall, FTOHs, FOSAs, and FOSEs concentrations at site 12 were similar to reported values at urban sites (see Table S5 in the SI). Site 12 was approximately 600 m from the perimeter of the WWTP and not downwind and therefore unlikely to be impacted by the emissions from the WWTP. In contrast, the other reference sites (i.e., sites 13), which were much closer (within 200 m of the treatment tanks) were able to detect the emissions from the WWTP. Air concentrations of FTOH were 11 times higher within the plant (8080 ( 7250 pg/m3, sites 411) compared to the reference locations (sites 13, 12); whereas air concentrations of FOSA/FOSE were 4 times higher within the plant (66 ( 37 pg/ m3). The differences in the magnitude of FTOHs and FOSAs/ FOSEs air concentrations may reflect differences in wastewater concentrations combined with differences in their volatilization potentials from water. FTOHs have higher liquidvapor pressures (PL) (531670 Pa at 25 C) compared to the FOSAs and FOSEs (0.357.0 Pa at 25 C).29 The most abundant PFC class within the WWTP was the FTOHs with 6:2 FTOH as the dominant compound (58% of the FTOHs), followed by 8:2 FTOH (36%). This is contrary to results for ambient urban air where 8:2 FTOH typically has higher concentrations than 6:2 FTOH.5,28,30 However, the PFC pattern varied within the plant indicating different emissions for individual compounds and processes. For example, the N-ethyl perfluorooctane sulfonamidoethanol (EtFOSE) was detected at the primary clarifier and aeration tank (sites 49) but not at the secondary clarifier. The FTOH concentrations at the two primary clarifier sites varied by up to a factor of about 2. Furthermore, the FTOH concentrations at the primary clarifier (sites 4 and 5) and above the aeration tank (sites 6 and 7) were 59 times higher in comparison to the secondary clarifier (sites 10 and 11; Figure 2). The higher FTOH concentration at site 4 is likely due to the fact that this sampler was mounted 1.5 m above a metal grate that sat directly above the outflow from the clarifier. The outflow exhibited some turbulence that may have contributed to emission of PFCs that impacted the sample. It should be noted that the primary clarifiers are downwind (prevailing wind direction) of the aeration tanks and therefore samples 4 and 5 may receive a substantial portion of PFCs emitted from the aeration process (see Figure 1). In comparison, the FOSA/ FOSE concentrations above the aeration tank (sites 6 and 7) were 1.5 and 5 times higher in comparison to the primary clarifier (sites 4 and 5) and secondary clarifier (sites 10 and 11), respectively (Figure 2). This likely reflects enhanced volatilization associated with the turbulence/bubbling of the aeration process, whereas the primary and secondary clarifiers have a relatively calm water surface.31 Interestingly, the air concentrations over the primary clarifier (which also has a calm water surface) were similar to that for the aeration tank for FTOHs, FOSAs, and FOSEs although the primary clarifier has no aeration. This may reflect higher wastewater concentrations of the FTOHs, FOSAs, and FOSEs at the upstream stages of the process prior to their degradation.10 This is consistent with EtFOSE not being detected in samples collected above downstream stages in the secondary clarifier. It is also possible that emissions to air from the aeration tanks impact the samplers at the primary clarifiers that are less than 100 m downwind of the aeration tanks (Figure 1). Of the PFSAs and PFCAs targeted in this study PFOS and C4C12, C14 PFCAs were detected in air at the WWTP, whereas perfluorobutane sulfonic acid (PFBS), perfluorohexane sulfonic acid (PFHxS), perfluorodecane sulfonic acid (PFDS) and PFOSA were below the MDL. Within the WWTP, PFOS was dominant with air concentrations of 43171 pg/m3 (41% of the PFOS/PFCAs), followed by perfluorobutanoic acid (PFBA) with 55116 pg/m3 (32%) and perfluorohexanoic acid (PFHxA) with 1063 pg/m3 (11%). Mean PFOS (113 ( 39 pg/ m3) and PFCA (170 ( 60 pg/m3) concentrations in air within the WWTP (sites 411) were 7 and 2 times higher, respectively, compared to the reference sites (sites 13 and 12; 16 ( 15 pg/ m3 and 89 ( 33 pg/m3, respectively). This indicates that the wastewater is a source of PFOS and PFCAs to the atmosphere. Notably, the reference sites 2 and 3, which were within a few hundred meters of Lake Ontario (Figure 1), had 7.5 times higher PFOS concentrations (24 pg/m3 and 34 pg/m3, respectively) compared to the reference sites 1 and 12 (4.7 pg/m3 and 3.0 pg/ m3, respectively). This might be due to the proximity of reference sites 2 and 3 to the lakeshore where PFOS emissions may be 8101 dx.doi.org/10.1021/es1036173 |Environ. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology ARTICLE elevated because of evaporation and generation of aqueous aerosols (e.g., wave breaking). A multimedia multispecies environmental fate model for the Lake Ontario indicates that the direct volatilization of PFOA accounts for over 99% of the water to air flux, whereas the aqueous aerosol production of PFOA is negligible (<1%).18 The volatilization of PFOA from the Lake Ontario can explain the increased PFOA concentration at the reference site 3 that was closest to the lake. Unlike the FTOHs, FOSAs, and FOSEs, which showed little difference in air concentrations between the primary clarifier and aeration tank, the PFOS and shorter chain PFCA (C4C8) concentrations above the aeration tanks were elevated (351 ( 39 pg/ m3) and were 1.6 times higher than at the primary and secondary clarifier (214 ( 12 pg/m3) (Figure 3). This indicates that the aeration process is especially relevant for emitting PFOS and shorter chain PFCAs (including PFOA) and not as important in emitting the longer chain PFCAs (C9C14). It has been shown that PFOA can be emitted directly by volatilization from the wastewater and/or associated with the aqueous aerosol and then evaporated to the gas phase in its neutral form.32 These pathways may contribute to the observed elevated air concentrations of PFOA and the shorter chain PFCAs and PFOS. PFOS and PFCAs are mainly present in their ionized form in the wastewater (pH 68),33 however, the anion can be protonated and the protonated form can volatilize to the atmosphere.15 The other possibility is the aqueous aerosol-mediated transport pathway,32 which is supported by highest air concentrations of PFOS, PFOA, and the shorter chain PFCA at the aeration tanks, where aqueous aerosols are generated. This can be further facilitated by the surface-active characteristics of PFCs which result in an enrichment of PFCs at the surface microlayer.34 The elevated levels of PFOS and shorter chain PFCAs in air may also be reflecting their dominance in wastewater in comparison to the longer chain PFCAs.19,35 Wastewater concentrations of the longer chain PFCAs may be decreased as they become associated with particles36 that are removed as sludge from the wastewater. It would be useful in the future to conduct coupled air and water sampling at the different treatment stages to investigate the waterair exchange dynamics and how it is affected by meteorology, water-side concentrations, and operating conditions. Landfill Concentrations. Air concentrations of PFCs at the two landfills determined using SIP disks deployed upwind and on-site are shown in Figure 2 and 3 (for details see Table S3 and S4 in the SI). The FTOH/FOSA/FOSE upwind concentrations (658 ( 150 pg/m3) are in the same range as for the WWTP reference sites. Substantially higher FTOH concentrations (536 times) were found on-site at the landfills compared to the upwind samples, while FOSA/FOSE concentrations were only 23 times higher (on-site vs upwind) reflecting lower landfill emission strength for the FOSA and FOSE classes. Some key differences were observed between the two landfills. For example, the FTOH concentration at landfill 2 were about 1 order of magnitude higher (26 000 pg/m3) than for landfill 1 (2590 pg/m3), whereas the FOSA/FOSE concentrations were about two times higher at landfill 2 (114 pg/m3) compared to landfill 1 (63 pg/m3). The type of waste deposited at the two landfills was similar, containing mostly residential waste. However, the active dumping area of landfill 2 was larger than for landfill 1 which may explain the higher air concentrations at landfill 2. Air concentrations of PFCs at the landfills were dominated by FTOHs (9398% of the PFCs), but, in contrast to the WWTP where 6:2 FTOH was dominant, 8:2 FTOH was the dominant compound (58% of the FTOHs), followed by 6:2 FTOH (32%). This relative abundance (i.e., 8:2 FTOH greater than 6:2 FTOH) is typical for urban areas.5,28,30 Emissions from the landfills are likely associated with waste products containing polymeric and surfactant materials. Theses polymeric materials can contain small percentages of unbound residual FTOHs and N-methyl perfluorooctane sulfonamidoethanol (MeFOSE) (0.043.8%).37 PFC emissions to air can be also affected by the degradation of contaminated materials during waste aging, by waterair transfer of water-soluble PFCs in the aqueous phase (i.e., landfill leachate) and by landfill gas production and utilization. The mean concentrations of PFCAs (226 ( 136 pg/m3) and PFOS (3.9 ( 1.0 pg/m3) were about 3 times higher at the landfill sites compared to the upwind sites. The distributions of the individual PFCAs were consistent, with PFBA (54%) > PFHxA (12%) > PFOA (8.2%), while at the upwind sites PFBA was predominant (80%). The PFCA pattern was similar to the pattern at the WWTP, with dominance of the even chain length PFCAs. This is consistent with the findings that the PFCAs originated from the telomer-derived products which contain only even chain length PFCs,9 whereas biodegradation of FTOHs to PFCAs produces even and odd chain length PFCAs.38 Similar PFCA patterns, with dominance of even chain length PFCAs and shorter chain PFCAs, were observed in landfill leachates from 22 sites in Germany,21 indicating that this pattern is typical for landfill emissions. It is interesting to note however, that PFOS exhibited very low air concentrations at the landfill sites (3.9 ( 1.0 pg/m3) contributing only 2% to the PFOS/PFCAs, whereas PFOS was predominant at the WWTP with a 41% contribution to the PFOS/PFCAs (113 ( 39 pg/m3). The low emissions of PFOS at the landfills may be due to several factors including: strong sorption of PFOS to landfill solids; efficient trapping of PFOS in landfill gas collection; and partitioning of PFOS to landfill leachate. FTOH Concentration Ratios. Individual FTOH air concentrations at the WWTP and landfill sites are positively correlated with each other (p < 0.0001, R = 0.870.98, Pearson Correlation, SPSS version 17.0 for Windows). The ratio of 6:2 FTOH to 8:2 FTOH to 10:2 FTOH on a concentration basis may help to distinguish sources of FTOHs to the atmosphere.30 The 6:2 FTOH to 8:2 FTOH to 10:2 FTOH ratio at the reference sites was 2.7 to 2.7 to 1.0 which is in the range of values reported for urban areas.28 It is interesting to note that 8:2 FTOH is enhanced at the landfill sites (i.e., ratio of 3.1 to 6.1 to 1.0) while 6:2 FTOH and 8:2 FTOH are elevated relative to 10:2 FTOH at the WWTP (i.e., 10.6 to 6.5 to 1.0). It should be noted that the generation of PFCAs from precursor FTOHs in air at the WWTP and landill sites is unlikely to occur due to the abundance of nitric oxide (NO) in these predominantly urban/suburban areas. Conversion of FTOH to PFCAs is expected to occur following atmospheric transport to background and remote regions where NO concentrations are low.9,11 However, biodegradation of FTOHs is possible in the wastewater and lodged landfill material.19,38 WWTP and Landfill Inputs to the Atmosphere. A simplified Gaussian dispersion model was applied to the study results to estimate emissions of PFCs. For a point source with instantaneous emissions, the simplified Gaussian formula is defined by:39 e cuyz 1 where c is air concentration (pg/m3), e is the emission rate (g/day), y and z are a horizontal fetch (m distance away 8102 dx.doi.org/10.1021/es1036173 |Environ. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology 1500 ?,1200 2 900 0 7) E 600 O 300 0 < w w Co < O U- O u_ O U- Ou_ Ou_ Ou_ Ocun_ aou- amu- co cc.os! 0 ID a) 2 WWTP 50 landfill site 1 CI landfill site 2 40 LL ^;-3 w 30u_22 g 20 (OL1L) - - 10mO1 LL 0 0. O u_ z u_ O O O 0_ a a 0 u_ u_ Figure 4. Estimated PFC emissions from a WWTP and two landfill sites in g/year. ARTICLE from the sampling site) and the height of sampler above a ground surface (m), and u is the mean wind speed (m/s) at the sampling height (note: air concentrations from the reference sites were not subtracted from the air concentration c, for details see text in the SI). The calculations are based on the air concentrations at the landfills determined for the on-site sample and the average for sites 6 and 7 at the WWTP, which were situated above the aeration tank. The calculated emission rates for the WWTP were multiplied by 8 to account for emissions from all eight identical aeration tanks. This calculation assumes that the aeration tanks were the predominant point source for PFCs at the WWTP. This is supported by the data and knowledge of the importance of aeration in enhancing water--air exchange of PFCs. However, if the other tanks have nonnegligible contributions to air emissions, then this calculation underestimates the emission rate from the WWTP. The calculated residence time over the WWTP due to advection was only 77 s (for details see the SI). Ultimately, the yearly PFC emission rates (g/year) were estimated based on the average daily emission rates (g/day) and are summarized in Figure 4 (for details see Table S6 in the SI). The estimated average yearly PFC emissions from the WWTP and landfills are subject to additional uncertainties due to the variations in PFC concentrations in wastewater and air and variable wind speeds during different seasons of the sampling year. The influence of rain events could not be determined due to the lack of temporal resolution of PAS. However, substantial rain events did occur during the sampling period (see Figure S3 and S4 in the SI) and it would be interesting to examine this influence in future studies. Further studies should also include the sampling of wastewater from all treatment stages in order to follow changes in PFCs throughout the process; investigation of which types of landfill materials are contributing most to air emissions is another question that should be addressed. EPFC yearly emissions were estimated to be approximately 2560 g/year for the WWTP, 99 g/year for landfill 1 and 1000 g/ year for landfill 2 in 2009. Emissions of FTOHs predominated and were about 2 orders of magnitude higher than the other PFC classes investigated in this study. Emissions of 6:2 FTOH at the WWTP were dominant, whereas 8:2 FTOH was dominant at landfill site 1 and 2. Among the PFSAs and PFCAs, PFOS and PFBA had the highest emissions to the atmosphere from the WWTP and PFBA emissions were highest at the landfill sites. These results are consistent with lake studies where upward fluxes were also found.'5'4 It should be noted that both landfills investigated here had processes for capturing landfill gas and further reduced volatilization to air by keeping the active, exposed area to a minimum. The per capita emissions for the WWTP were estimated using the emissions rates in yg per year and the population served by the WWTP (--1 000 000 people) (see Table S6 in the SI). This per capita estimation must be considered with caution because it is based on one WWTP and does not account for possible commercial and industrial wastewater contributions. The estimated per capita emissions of the EPFCs from WWTPs to air were 2560 jug/year/person. In comparison, the reported WWTP discharge of PFCs to the aqueous environment in effluents is ^-2-10 times higher.2O'35 For example, the mean EPFC discharge in effluents from 7 WWTPs in Switzerland was ^-34 000 jug/year/person (85-2670 g/year) and dominated by PFOS,20 and the mean EPFC discharge in effluents from 9 WWTP in Germany was ^4400 jug/year/person (18-20 300 g/year), dominated by PFOA.35 In the case of landfill discharges to the aqueous environment, a study of 22 landfill sites in Germany estimated that the mean EPFC discharge was "49 g/year, much lower compared to efluents from WWTPs21 Aqueous discharges of FTOHs were not identified in any effluents from the WWTP and landfill studies discussed above. Thus, for the FTOHs in particular, emissions to air from WWTPs and landfills are important in contributing to atmospheric burdens and for their role as precursors in the long-range atmospheric transport of PFCAs. The emission data for PFCs presented here provide a basis and justification for future waste sector studies. ASSOCIATED CONTENT Supporting Information. Additional details on sampling sites, meteorological data and PFC concentrations. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION Corresponding Author *Phone: (L. A.); gc.ca; ec.gc.ca. (M. S.); fax: M. S.); e-mail: lutz.ahrens@ec. ACKNOWLEDGMENT We kindly acknowledge Dr. Jianmin Ma (Environment Canada, Downsview) for his help in calculating the emission rates and Shirley-Anne Smyth (Environment Canada, Burlington) for 8103 dx.doi.org/10.1021/es1036173 lEnviron. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology ARTICLE assistance with initiating the field campaigns. We thank the Chemicals Management Plan (Government of Canada) for partial funding. We also thank the operators of the landfill sites and WWTP for access to the sampling sites. ' REFERENCES (1) Kissa, E. Fluorinated Surfactants and Repellents; Marcel Dekker: New York, 2001. (2) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44. (3) Paul, A. G.; Jones, K. C.; Sweetman, A. J. A first global production, emission, and environmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol. 2009, 43, 386-392. (4) Ahrens, L.; Barber, J. L.; Xie, Z.; Ebinghaus, R. Longitudinal and latitudinal distribution of perfluoroalkyl compounds in the surface water of the Atlantic Ocean. Environ. Sci. Technol. 2009, 43, 3122-3127. (5) Shoeib, M.; Harner, T.; Vlahos, P. Perfluorinated chemicals in the Arctic atmosphere. Environ. Sci. Technol. 2006, 40, 7577-7583. (6) Stock, N. L.; Furdui, V. I.; Muir, D. C. G.; Mabury, S. A. Perfluoroalkyl contaminants in the Canadian Arctic: Evidence of atmospheric transport and local contamination. Environ. Sci. Technol. 2007, 41, 3529-3536. (7) Ahrens, L.; Yamashita, N.; Yeung, L. W. Y.; Taniyasu, S.; Horii, Y.; Lam, P. K. S.; Ebinghaus, R. Partitioning behaviour of per- and polyfluoroalkyl compounds between pore water and sediment in two sediment cores from Tokyo Bay, Japan. Environ. Sci. Technol. 2009, 43, 6969-6975. (8) Giesy, J. P.; Kannan, K. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342. (9) Ellis, D. A.; Martin, J. W.; Mabury, S. A.; De Silva, A. O.; Hurley, M. D.; Sulbaek Anderson, M. D.; Wallington, T. J. Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321. (10) Rhoads, K. R.; Janssen, E. M.-L.; Luthy, R. G.; Criddle, C. S. Aerobic biotransformation and fate of n-ethyl perfluorooctane sulfonamidoethanol (n-EtFOSE) in activated sludge. Environ. Sci. Technol. 2008, 42, 2873-2878. (11) Wallington, T. J.; Hurley, M. D.; Xia, J.; Wuebbles, D. J.; Sillman, S.; Ito, A.; Penner, J. E.; Ellis, D. A.; Martin, J.; Mabury, S. A.; Nielsen, O. J.; Sulbaek Anderson, M. P. Formation of C7F15COOH (PFOA) and other perfluorocarboxylic acids during the atmospheric oxidation of 8:2 fluorotelomer alcohol. Environ. Sci. Technol. 2006, 40, 924-930. (12) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204. (13) Giesy, J. P.; Naile, J. E.; Khim, J. S.; Jones, K. C.; Newsted, J. L. Aquatic toxicology of perfluorinated chemicals. Rev. Environ. Contam. T. 2010, 202, 1-52. (14) Joensen, U. N.; Bossi, R.; Leffers, H.; Jensen, A. A.; Skakkebk, N. E.; Jrgensen, N. Do perfluoroalkyl compounds impair human semen quality?. Environ. Health Perspec. 2009, 117, 923-927. (15) Webster, E.; Ellis, D. A.; Reid, L. K. Modeling the environmental fate of perfluorooctanoic acid and perfluorooctanoate: An investigation of the role of individual species partitioning. Environ. Toxicol. Chem. 2010, 29, 1466-1475. (16) Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Hurley, M. D.; Sulbaek Anderson, M. D.; Wallington, T. J. Atmospheric lifetime of fluorotelomer alcohols. Environ. Sci. Technol. 2003, 37, 3816-3820. (17) UNEP. United Nations Environment Programme (UNEP), Geneva hosts Stockholm Convention on Persistent Organic Pollutants from 4 to 8 May, from http://www.unep.org/newscentre/ (accessed September 22, 2010). (18) Webster, E.; Ellis, D. A.; Reid, L. K. Modeling the environmental fate of perfluorooctanoic acid and perfluorooctanoate: An investigation of the role of individual species partitioning. Environ. Toxicol. Chem. 2010, 29, 1466-1475. (19) Schultz, M. M.; Higgins, C. P.; Huset, C. A.; Luthy, R. G.; Barofsky, D. F.; Field, J. A. Fluorochemical mass flows in a municipal wastewater treatment facility. Environ. Sci. Technol. 2006, 40, 7350-7357. (20) Huset, C. A.; Chiaia, A. C.; Barfosky, D. F.; Jonkers, N.; Kohler, H.-P. E.; Ort, C.; Giger, W.; Field, J. A. Occurrence and mass flows of fluorochemicals in the Glatt Valley watershed, Switzerland. Environ. Sci. Technol. 2008, 42, 6369-6377. (21) Busch, J.; Ahrens, L.; Sturm, R.; Ebinghaus, R. Polyfluoroalkyl compounds in landfill leachates. Environ. Pollut. 2010, 158, 1467-1471. (22) Shoeib, M.; Harner, T.; Lee, S. C.; Lane, D.; Zhu, J. Sorbentimpregnated polyurethane foam disk for passive air sampling of volatile fluorinated chemicals. Anal. Chem. 2008, 80, 675-682. (23) Genualdi, S.; Lee, S. C.; Shoeib, M.; Gawor, A.; Ahrens, L.; Harner, T. Global pilot study of legacy and emerging persistent organic pollutants using sorbent-impregnated polyurethane foam disk passive air samplers. Environ. Sci. Technol. 2010, 44, 5534-5539. (24) Vierke, L.; Ahrens, L.; Shoeib, M.; Reiner, E. J.; Guo, R.; Palm, W.-U.; Ebinghaus, R.; Harner, T. Air concentrations and particle-gas partitioning of polyfluoroalkyl compounds at a wastewater treatment plant. Environ. Chem. 2011, accepted. (25) Dreyer, A.; Shoeib, M.; Fiedler, S.; Barber, J. L.; Harner, T.; Schramm, K.-W.; Jones, K. C.; Ebinghaus, R. Field intercomparison on the determination of volatile and semivolatile polyfluorinated compounds in air. Environ. Chem. 2010, 7, 350-358. (26) Klanova, J.; upr, P.; Kohoutek, J.; Harner, T. Assessing the influence of meteorological parameters on the performance of polyurethane foam-based passive air samplers. Environ. Sci. Technol. 2008, 42, 550-555. (27) Powley, C. R.; George, S. W.; Ryan, T. W.; Buck, R. C. Matrix effect-free analytical methods for determination of perfluorinated carboxylic acids in environmental matrixes. Anal. Chem. 2005, 77, 6353-6358. (28) Jahnke, A.; Ahrens, L.; Ebinghaus, R.; Temme, C. Urban versus remote air concentrations of fluorotelomer alcohols and other polyfluorinated alkyl substances in Germany. Environ. Sci. Technol. 2007, 41, 745-752. (29) Lei, Y. D.; Wania, F.; Mathers, D.; A., M. S. Determination of vapor pressures, octanol-air, and water-air partition coefficients for polyfluorinated sulfonamide, sulfonamidoethanols, and telomer alcohols. J. Chem. Eng. Data 2004, 49, 1013-1022. (30) Piekarz, A. M.; Primbs, T.; Field, J. A.; Barofsky, D. F.; Simonich, S. Semivolatile fluorinated organic compounds in Asian and western U.S. air masses. Environ. Sci. Technol. 2007, 41, 8248-8255. (31) Tansel, B.; Eyma, R. R. Volatile organic contaminant emissions from wastewater treatment plants during secondary treatment. Water, Air, Soil, Pollut. 1999, 112, 315-325. (32) McMurdo, C. J.; Ellis, D. A.; Webster, E.; Butler, J.; Christensen, R. D.; Reid, L. K. Aerosol enrichment of the surfactant PFO and mediation of the water-air transport of gaseous PFOA. Environ. Sci. Technol. 2008, 42, 3969-3974. (33) Barton, C. A.; Kaiser, M. A.; Russell, M. H. Partitioning and removal of perfluorooctanoate during rain events: The importance of physical-chemical properties. J. Environ. Monit. 2007, 9, 839-846. (34) Ju, X.; Jin, Y.; Sasaki, K.; Saito, N. Perfluorinated surfactants in surface, subsurface water and microlayer from Dalian coastal waters in China. Environ. Sci. Technol. 2008, 42, 3538-3542. (35) Ahrens, L.; Felizeter, S.; Sturm, R.; Xie, Z.; Ebinghaus, R. Polyfluorinated compounds in effluents of waste water treatment plants and surface water along the River Elbe, Germany. Mar. Pollut. Bull. 2009, 58, 1326-1333. (36) Ahrens, L.; Taniyasu, S.; Yeung, L. W. Y.; Yamashita, N.; Lam, P. K. S.; Ebinghaus, R. Distribution of polyfluoroalkyl compounds in water, suspended particulate matter and sediment from Tokyo Bay, Japan. Chemosphere 2010, 79, 266-272. (37) Dinglasan-Panlilio, M. J. A.; Mabury, S. A. Significant residual fluorinated alcohols present in various fluorinated materials. Environ. Sci. Technol. 2006, 40, 1447-1453. 8104 dx.doi.org/10.1021/es1036173 |Environ. Sci. Technol. 2011, 45, 8098-8105 Environmental Science & Technology (38) Liu, J.; Wang, N.; Szostek, B.; Buck, R. C.; Panciroli, P. K.; Folsom, P. W.; Sulecki, L. M.; Bellin, C. A. 62 Fluorotelomer alcohol aerobic biodegradation in soil and mixed bacterial culture. Chemosphere 2010, 78, 437-444. (39) Wei-Mei, J.; Xiao-Ming, W. A linked three-dimensional PBL and dispersion model in coastal regions. Boundary-Layer Meteorol. 1990, 53, 43-62. (40) Kim, S.-K.; Kannan, K. Perfluorinated acids in air, rain, snow, surface runoff, and lakes: relative importance of pathways to contamination of urban lakes. Environ. Sci. Technol. 2007, 41, 8328-8334. ARTICLE 8105 dx.doi.org/10.1021/es1036173 |Environ. Sci. Technol. 2011, 45, 8098-8105 Environ. Sci. Technol. 2009, 43, 3122-3127 Longitudinal and Latitudinal Distribution of Perfluoroalkyl Compounds in the Surface Water of the Atlantic Ocean LUTZ AHRENS,*' t ' # JONATHAN L. BARBER," ZHIYONG XIE,t AND RALF EBINGHAUS t Department for Environmental Chemistry, Institute for Coastal Research, GKSS Research Centre Geesthacht, Max-Planck-Strasse 1, DE-21502 Geesthacht, Germany, Institute for Ecology and Environmental Chemistry, Faculty of Environment and Technique, Leuphana University of Luneburg, DE-21335 Liineburg, Germany, and Centre for Chemicals Management, Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, United Kingdom Received December 10, 2008. Revised manuscript received February 13, 2009. Accepted March 2, 2009. Perfluoroalkyl compounds (PFCs) were determined in 2 L surface water samples collected in the Atlantic Ocean onboard the research vessels Maria S. Merian along the longitudinal gradient from Las Palmas (Spain) to St. Johns (Canada) (15 W to 52 W) and Polarstern along the latitudinal gradient from the Bay of Biscay to the South Atlantic Ocean (46 N to 26 S) in spring and fall 2007, respectively. After filtration the dissolved and particulate phases were extracted separately, and PFC concentrations were determined using high-performance liquid chromatography interfaced to tandem mass spectrometry. No PFCs were detected in the particulate phase. This study provides the first concentration data of perfluorooctanesulfonamide (FOSA), perfluorohexanoic acid, and perfluoroheptanoic acid from the Atlantic Ocean. Results indicate that trans-Atlantic Ocean currents caused the decreasing concentration gradient from the Bay of Biscay to the South Atlantic Ocean and the concentration drop-off close to the Labrador Sea. Maximum concentrations were found for FOSA, perfluorooctanesulfonate, and perfluorooctanoic acid at 302, 291, and 229 pg L--', respectively. However, the concentration of each single compound was usually in the tens of picograms per liter range. South of the equator only FOSA and below 4 S no PFCs could be detected. Introduction Perfluoroalkyl compounds (PFCs) are persistent against the typical environmental degradation processes and have been found in water, wildlife, and human tissues around the globe (1-3). Because of their unique physicochemical properties due to their combination of lipophilic and * Correspondin author hone: fax: e-mail: @gkss.de. GKSS Research Centre Geesthacht. Leuphana University of Luneburg. Lancaster University. " Current address: Cefas Lowestoft Laboratory, Centre for Envi- ronment, Fisheries and Aquaculture Science, Pakefield Road, Low- estoft, Suffolk NR33 OHT, U.K. 3122 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009 hydrophilic characteristics, PFCs have been widely used in a lot of consumer products such as polymerization aids and stain repellents on carpets, textiles, leather, and paper products for over 50 years (4). From the production and use of these products, PFCs can be released into the environment. PFCs could be bioaccumulative (5) and have toxic effects in biota (6, 7). The transportation pathways of PFCs to remote regions have not been conclusively characterized to date. Two main hypotheses were proposed for the global transportation of PFCs. First, neutral, volatile precursor compounds could undergo long-range atmospheric transport and be degraded in remote regions (8), or second, ionic PFCs could be transported directly by oceanic currents or by means of sea spray (9, 10). The first hypothesis is supported by the determination of precursor compounds, such as fluorotelomer alcohols (FTOHs), perfluoalkanesulfonamidoethanols, in the Arctic atmosphere (11). The second hypothesis is supported by the fact that ionic PFCs such as the perfluorinated carboxylic acids (PFCAs) and perfluorinated sulfonates (PFCAs) have high water solubilities and low pia values and are therefore dissociated at environmentally relevant pH values (4). The ocean currents were calculated by Prevedouros and coworkers to be the major transportation pathway for PFCAs in comparison to atmospheric transportation (12). However, irrespective of the transportation pathway involved, high concentrations of PFCs have been found in biota from the Canadian Arctic, especially in marine mammals which are top predators in the marine ecosystem (13). The global occurrence of PFCs in open-ocean water was described first by Yamashita et al. (1). Further investigations of PFCs in the Indian Ocean and close to Antarctica were described subsequently (14). Detected concentrations are usually around some tens to hundreds of picograms per liter, depending on the location and the compound. It was hypothesized that PFCs could be transported globally with the thermohaline circulation system (15), with the openocean water acting as a final sink for perfluorooctanesulfonate (PFOS) and perfluorooctanoic acid (PFOA) (1). However, ocean measurements are very limited and essential for the validation of models as well as quantifying inputs of PFCs to remote environments such as the Arctic. Further, seawater measurements of PFCs are very useful for determining the dominant transportation pathway, either oceanic currents or atmospheric transport of precursors. The aim of this study was to investigate the longitudinal and latitudinal gradient of PFCs in surface water in the Atlantic Ocean. Sixty water samples were collected during cruises on the research vessels Maria S. Merian, from Las Palmas (Spain) to St. Johns (Canada) (15 W to 52 W), and Polarstern, from the Bay of Biscay to the South Atlantic Ocean (46 N to 26 S), in spring and fall 2007, respectively. Concentrations of various PFCs, including perfluorobutanesulfonate (PFBS), PFOS, perfluorooctanesulfonamide (FOSA), perfluorohexanoic acid (PFFIxA), perfluoroheptanoic acid (PFHpA), PFOA, and perfluorononanoic acid (PFNA), were quantified in openocean water samples. The observed distribution, characterized by increasing and decreasing concentration gradients of PFCs, can be explained by the pattern of ocean water currents. A comparison with open-ocean water PFC data from Yamashita et al. (1, 15) and Wei et al. (14) is given. This study provides the first evidence for the presence of FOSA, PFHpA, and PFHpA in the Atlantic Ocean. 10.1021/es803507p CCC: $40.75 2009 American Chemical Society Published on Web 03/30/2009 FIGURE 1. Map showing the sampling stations of the Maria S. Merian (15 W to 52 W, A-R) and Polarstern (46 N to 26 S, 1-42) cruises in 2007. The yellow arrows display the main surface currents in the Atlantic Ocean. Experimental Section Chemicals. The target analytes include 33 ionic PFCs (PFCAs, PFSAs, perfluorinated sulfinates (PFSiAs), fluorotelomer carboxylic acids (FTCAs), and unsaturated fluorotelomer carboxylic acids (FTUCAs)) as well as 7 neutral PFC precursor compounds (perfluoroalkanesulfonamides, perfluoroalkanesulfonamidoethanols) (for details, see Table S2 in the Supporting Information). Methanol (SupraSolv), acetonitrile (LiChrosolv), ammonium hydroxide (25% for analysis), formic acid (98-100% suprapure), and ammonium acetate were purchased from Merck (Darmstadt, Germany). Sampling Campaign. Surface water samples were collected with the research vessels Maria S. Merian (Leibniz Institute for Baltic Sea Research (IOW), Warnemu nde) and Polarstern (Alfred-Wegener-Institut (AWI), Bremerhaven) from April 14 to April 30 (cruise "MSM05") and Oct 29 to Nov 22 (cruise "ANT XXIV-1"), 2007, respectively. The first cruise with the R/V Maria S. Merian was performed along the longitudinal gradient from 15 W to 52 W, and the second cruise with the R/V Polarstern was performed along the latitudinal gradient from 46 N to 26 S (Figure 1, Table S1 in the Supporting Information). Water samples (62 L) were collected in brown glass bottles via the ships' intake systems at approximately 11 m below the surface at sampling stations 1-42 (46 N to 26 S) and A-R (15 W to 52 W). In addition, at sampling stations J, K, L, O, and R water samples were collected at 2 m depth and directly at the water surface by an external sampler in a brown glass bottle, and at sampling station L two deep-water samples were taken at depths of 200 and 3800 m with a rosette-type sampler (Seabird SBE-32 carousel water sampler equipped with 24 10 L HydrobiosFreeflow bottles) to examine concentration differences between different water layers. The different sampling techniques, ship intake systems, rosette-type sampler, and outboard sampler using 2 L brown glass bottles were tested to evaluate for possible background contamination during sampling. The samples were filtered directly after sampling onboard using glass fiber filters (GFF, GC/C, Whatman, L 47 mm, >1.2 m). The dissolved-phase samples were stored at 4 C prior to solid-phase extraction (SPE) onboard ship on the same or following days, whereas the GFFs were stored in sealed test tubes in a freezer at -20 C and extracted after the end of the sampling cruise in a clean laboratory (class 10 000) within a few days of arrival at the laboratory. Field blanks (FBs) were taken every 10th sample for the filtrate and GFF to test for possible blank contamination. For the dissolved-phase FB, 100 mL of Millipore water (Millipore Elix 5 and Millipore Milli Q Plus) was added to a 2 L brown glass bottle and then put through SPE extraction. The sources of blank contamination are mostly caused by sampling, the extraction process, and instrument analysis (1), which indicates that the amount of contamination is independent of the volume of Millipore water extracted. FB GFFs were prepared by placing them on the filtration equipment for 1 min. Both types of field blanks were then stored and extracted in the same manner as "real" samples. Sample Extraction and Instrumental Analysis. The filtrate and the GFF were separately spiked with 10 ng absolute of an internal standard (IS) mix (i.e., [13C2]PFHxA, [13C4]PFOA, [13C4]PFNA, [13C4]PFDA, [13C2]PFUnDA, [13C2]PFDoDA, [18O2]PFHxS, [13C4]PFOS, [13C4]PFOSi, d3-MeFOSA, d5-EtFOSA, d7-MeFOSE, d9-EtFOSE, [13C2]FHEA, [13C2]FOEA, [13C2]FDEA, [13C2]FHUEA, [13C2]FOUEA, [13C2]FDUEA, 100 L of a 0.1 g mL-1 solution; see Table S1 in the Supporting Information). The filtrate was spiked with the IS mix and extracted by SPE with Oasis WAX cartridges (Waters, 150 mg, 6 cm3, 30 m), as described elsewhere (16) with some modifications. Briefly, after being preconditioned with 5 mL of methanol and Millipore water, the cartridge was loaded with the 2 L sample at approximately 4 drops s-1 (0.1 mL min-1). The cartridge was then washed with 5 mL of 0.1% formic acid in Millipore water and dried for 30 min under vacuum. After the loading and drying steps onboard the ships, the cartridges were stored in a freezer at -20 C. The cartridges were eluted after the end of the sampling cruise in a clean laboratory (class 10 000) within a few days of arrival at the laboratory. The elution was divided into two parts: The sulfonamides were eluted with 14 mL of acetonitrile; thereafter the acids were eluted with 5 mL of 0.1% ammonium hydroxide in methanol. The combined extract was reduced to 150 L under a nitrogen stream and spiked with 20 ng absolute of the injection standard d5-EtFOSAA (InjS, 50 L of a 0.4 g mL-1 solution; see Table S1). The particulate matter (>1.2 m) was analyzed by sonication as described elsewhere (17) with some modifications. The GFF was spiked with the same IS mix as the filtrate and sonicated with 100 mL of methanol for 1 h. This extraction was performed twice, and the two fractions were combined, evaporated by rotary evaporation, and filtered. The extract was reduced to 150 L under a nitrogen stream and spiked with 20 ng of the InjS (see above). Finally, the extracts from the dissolved- and particulate-phase samples were analyzed using high-performance liquid chromatography coupled with tandem mass spectrometry (HPLC-MS/MS). An HP 1100 HPLC system (Agilent Technologies) was used with a Synergi Hydro RP 80A column (150 2 mm, 4 m) by Phenomenex, combined with a suitable guard column: Synergi 2 Hydro RP Mercury (20 2 mm, 2 m). The triple-quadrupole mass spectrometer, supplied by Applied Biosystems/MDS SCIEX (API 3000), used an electrospray ionization (ESI) interface in negative ionization mode (for details see ref 18). Results and Discussion Quality Assurance. The analytical quality of the laboratory has been approved in interlaboratory studies (19). As standard procedure, FBs, method detection limits (MDLs), method quantification limits (MQLs) (see Table 1), and recoveries of spiked samples were examined (see Table S3 in the Supporting Information). Matrix spike recoveries of the target VOL. 43, NO. 9, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3123 TABLE 1. MDLs and MQLs for the Dissolved Phase and Field Blank Concentrations (Dissolved and Particulate Phases) for Cruises Onboard the Research Vessels Maria S. Merian (FB 15 W to 52 W, n = 4 + 4) and Polarstern (FB 46 N to 26 S, n = 6 + 6) in the Atlantic Ocean (pg L-1)a dissolved-phase FB particulate-phase FB analyte MDLb MQLb R/V Maria S. Merian R/V Polarstern R/V Maria S. Merian R/V Polarstern PFBS 0.49 1.6 <1.6 <1.6 nd nd PFOS 3.1 10 <10 <10-25 nd nd FOSA 5.1 17 <17 <17 nd nd PFHxA 1.7 5.7 <5.7 <5.7-9.2 nd nd PFHpA 1.8 5.9 <5.9 <5.9-9.7 nd nd PFOA 1.2 4.0 <4.0-15 <4.0-28 nd nd PFNA 1.5 5.1 <5.1 <5.1 nd nd a Blank levels were calculated from a sample volume of 100 mL of Millipore water. One FB sample from the R/V Maria S. Merian cruise and two FB samples from the Polarstern cruise showed a blank contamination. Details are given in the text. nd ) not detected. <x ) below the respective MQL. b MDL and MQL (ng L-1) at 3 and 10 times the signal-to-noise ratio in natural samples (n ) 4), respectively. analytes ranged from 77% to 131% for the dissolved phase and from 72% to 113% for the particulate phase. A variety of laboratory products contain fluoropolymers such as polytetrafluoroethylene (PTFE) (20). All fluorinated materials which could come in contact with the sample during the sampling (including SPE block), sample preparation, and instrumental analysis were removed (for details see ref 1). After the removal of all PTFE parts from the HPLC system, no instrument blank was detected. All procedure blanks, using 1 L of Millipore water, which were extracted in the same manner as the samples, were below the MQL. No background contamination was detected in the FB for the particulate phase. For the dissolved phase, the FBs from both sampling cruises were usually below the MQL, but in three FBs contamination levels of a few picograms per liter up to 28 pg L-1 (PFOA) were quantified (see Figure S1 in the Supporting Information). For control of the repeatability and blank contamination of the ship inlet system at stations J, K, L, O, and R, five samples were taken in parallel using the outboard sampler and the ship inlet system; for all detected PFCs no significant differences were observed (Mann-Whitney U-test [p < 0.01]). MDLs and MQLs were calculated for substances that were found in real samples using signal-to-noise ratios of 3 and 10, respectively. The MDLs were in the low picogram per liter range for the 2 L water samples. Matrix spike recoveries of the IS at two different spike levels (5 and 20 ng L-1) ranged from 23% (d3-MeFOSA) to 90% ([13C2]FHUEA) for the dissolved phase and from 50% (d3-MeFOSA) to 124% ([13C2]PFHxA) for the particulate phase. Concentrations of PFCs in the Atlantic Ocean. In this study, 40 PFCs (see Table S2 in the Supporting Information) were measured in the water samples. PFBS, PFOS, FOSA, PFHxA, PFHpA, PFOA, and PFNA could be quantified in the dissolved phase of the marine water samples in a concentration range of <MQL to 1115 pg L-1; all other PFCs were below the corresponding MDLs. PFCs were not detectable in the particulate phase. The low particle mass in the 2 L water samples could be responsible for the not detectable PFC concentration in the particulate phase in the Atlantic Ocean; however, the partitioning behavior of PFCs is an important future research field to evaluate their physical state and bioavailability. To the authors' knowledge, this is the first report of quantifiable concentrations of PFHxA, PFHpA, and FOSA in surface water in the Atlantic Ocean. At five sampling locations, water samples were taken in parallel at 11 and 2 m 3124 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009 FIGURE 2. Individual PFC concentrations (pg L-1) and ambient temperature from the Polarstern cruise (46 N to 26 S, 1-42) in 2007. FIGURE 3. Individual PFC concentrations (pg L-1) and ambient temperature from the Maria S. Merian cruise (15 W to 52 W, A-R) in 2007. depths and directly at the surface. No correlation between sampling depth and concentration levels was observed, which indicates that there is a well-mixed zone between the surface and 11 m water depth in open-ocean waters. Ju et al. found at three near-shore sites that the surface microlayer (50 L thickness) had by a factor of 24-109 higher PFOS concentration than the subsurface water (>30 cm depth) (21); however, the results are not comparable because of the different sampling techniques. Concentrations of PFCs at location L in the two deep-water samples at 200 and 3800 m were below the MDL. Yamashita et al. investigated vertical profiles in the Labrador Sea, Middle Atlantic Ocean, South Pacific Ocean, and Japan Sea, where they detected relatively constant concentration levels of PFOS, PFOA, and PFBS down to 2000 m in the Labrador Sea and a decreasing concentration gradient with water depth in the other areas. They suggested that the global circulation system has an influence on the occurrence of PFCs in deep water (15). Our deep-water samples were taken far away from downwelling currents, which explains why no PFCs were detected. The longitudinal and latitudinal distribution of PFC concentrations in the Atlantic Ocean is shown in Figures 2 and 3 and Table S4 in the Supporting Information. The highest PFC concentration (1115 pg L-1) was found in the Bay of Biscay close to the European source area. A decreasing north to south latitudinal gradient was observed toward the Canary Islands where the mean PFC concentration declined by a factor of 6 to 142-191 pg L-1 (sampling stations 16 and 17). The PFC concentrations remained relatively constant toward the south from the Canary Islands down to 10 N. In the equator area, the PFC concentrations decreased by a factor of 4, relative to those of the Canary Islands down to 10 N, with only FOSA quantifiable at 37-53 pg L-1 (sampling stations 31-34). South of 4 S, no PFCs were detected. The west to east transect in the North Atlantic showed a different pattern of concentrations. The PFC concentrations east of 25 W (sampling stations G-R) and at sampling points A and B at the coast from Canada ranged from 52 to 117 pg L-1 and were on average a factor of 2 higher than those of samples C-F, clustered north of the main transect close to the Labrador Sea (nd to 40 pg L-1). The increasing concentrations, observed from sampling station L to sampling station R, could be influenced by latitudinal as well as longitudinal trends. Two-thirds of the water samples contained quantifiable concentrations of PFOA, which was the most abundant compound in the water samples from the Atlantic Ocean, with a mean contribution of 37% to the PFCs and a concentration range from <4.0 to 229 pg L-1. PFNA was detected in 52% of the water samples at concentrations greater than the LOQ, but the concentrations from <5.1 to 107 pg L-1 were less than those of PFOA. Concentrations of the other PFCs ranged from <17 to 307 pg L-1 for FOSA, from <10 to 291 pg L-1 for PFOS, from <5.7 to 127 pg L-1 for PFHxA, from <5.9 to 104 pg L-1 for PFHpA, and from <1.6 to 60 pg L-1 for PFBS; however, the concentrations were below the MQL in more than half of the samples. Comparison with Other Ocean Water PFC Measurements. Minimum and maximum PFC concentrations in surface open-ocean water from the Atlantic, Pacific, and Indian Oceans are shown in Table 2. In previous studies, the highest concentrations of PFOA and PFOS were found in the North and Mid Atlantic Ocean and the Western Pacific Ocean, while the lowest concentrations were observed in the Central and Southern Pacific and the Indian Ocean (1, 14, 15, 22). The results presented here can be compared with open ocean water samples presented from Yamashita et al. (1). They collected samples from the North and Middle Atlantic Ocean in 2002-2004 and found concentration levels of several tens of picograms per liter for perfluorohexanesulfonate (PFHxS), PFOS, and PFNA to a few hundreds of picograms per liter for PFOA. A similar study from Theobald et al. (22) was carried out from 53 N to 30 S in the Atlantic Ocean in 2005. The concentrations of PFOA and PFOS were in a range of a few tens of picograms per liter with a maximum concentration of 170 pg L-1 for PFOS. Both studies (1, 22) reported concentrations in the same range as in this study, except for PFHxS, which was found in the first study, but could not be detected by us. Concentrations of PFOS and PFOA reported in the West Pacific Ocean are in the same range as found in the Atlantic in this study, with concentrations of PFOS and PFOA in the Central and South Pacific and Indian Oceans about 1 magnitude lower than in the North Atlantic Ocean. PFBS and PFNA were also found in the Central to East Pacific Ocean and the Antarctic region, respectively, but the concentrations were less than those found in the Atlantic Ocean. The global distribution of PFBS might originate from the increasing production of n-methylperfluorobutanesulfonamidoethanol (MeFBSE) and related products with four perfluorinated carbons, which was introduced after the voluntary phaseout of perfluorooctanesulfonyl fluoride (POSF) by the 3M Co. in 2000 (23, 24). In comparison to this study, FOSA, PFHxA, and PFHpA could not be detected or were not analyzed in the Pacific and Indian Oceans, respectively. Conversely, PFHxS and PFDoDA were detected in the Indian Ocean and TABLE 2. Comparison of Minimum and Maximum PFC Concentrations in Surface Open-Ocean Water with Literature Data (pg L-1)a location PFBS Yamashita et al.(1) Theobald et al.(22) Yamashita et al.(15) Wei et al.(14) this study North Atlantic Ocean (n ) 9) Middle Atlantic Ocean (n ) 7) West Pacific Ocean (n ) 2) Central to East Pacific Ocean (n ) 12) North to South Atlantic Ocean (n ) 22) South Pacific Ocean (n ) 5) Central and South Pacific Ocean (n ) 9) Indian Ocean (n ) 7) Antarctic region (n ) 5) North Atlantic Ocean (n ) 40) Middle Atlantic Ocean (n ) 10) South Atlantic Ocean (n ) 10) na na na na na na <25 <5 <1(5)-2.9 <1.6-60 <1.6 <1.6 PFHxS 4.1-6.1 2.6-12 2.2-2.8 0.1-1.6 na na <5 <5 <1(5) nd nd nd PFOS 8.6-36 37-73 54-78 1.1-20 <14-170 <5-11 <5-21 <5-8.6 5.1-22.6 <10-291 <10-60 <10 FOSA na na na na na na <5 na na <17-307 <17-60 <17-53 PFHxA na na na na na na <5 <5 <5 <5.7-127 <5.7 <5.7 PFHpA na na na na na na <5 <5 <5 <5.9-104 <5.9-9.7 <5.9 PFOA 160-338 100-439 136-142 15-62 <17-90 <5-11 <5-7.0 <5-11 <5 <4.0-229 <4.0-87 <4.0 PFNA 15-36 na na 1.0-16 na na <5 <5 <5 <5.1-107 <5.1-35 <5.1 PFDoDA na na na na na na <1 <1-1.4 <1-1.1 nd nd nd VOL. 43, NO. 9, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3125 a nd ) not detected. na ) not analyzed. <x ) below the respective MQL. Arctic region in the few picograms per liter range, whereas both compounds could not be detected in this study. Yamashita et al. have studied vertical profiles of several PFCs in the Labrador Sea, Middle Atlantic Ocean, South Pacific Ocean, and Japan Sea in 2004 and 2005 (15). The surface water concentrations of PFOA and PFOS in the Northwest Atlantic Ocean were comparable with those from this study, but in addition in this study, PFHpA and PFNA were detected and PFBS was not detected in this area. Impact of the Ocean Currents on the PFC Pattern and Concentration Level. The occurrence of elevated PFC levels in the Arctic ecosystem (25) raises the question about the global transportation and fate of PFCs. In addition to the atmosphere, the ocean currents could be an important global transport pathway for transport of PFCs from industrial to remote areas (1). This study examined the impact of environmental factors, such as ocean currents, on the distribution of PFCs in the Atlantic Ocean. The PFC concentration distribution along the latitudinal gradient will be influenced by the Canary, Equatorial Counter, and Benguela Currents (see Figures 1 and 2). The Canary Current comes from the north from the European Continent source region and crosses the Equatorial Counter Current and the Benguela Current in the equator area. The Benguela Current has for its origin Antarctic water with low PFC loading (14), and the influence of this water body resulted in a rapid decrease of PFC concentrations to below the MDLs. The presence of cold surface water from the Benguela Current was confirmed by the drop-off of the water temperature from the equator region to the south. Furthermore, the West African coast seemed to have no impact on the concentration level, which suggests lack of sources (e.g., river discharge) from this area in contrast to the industrial European area. It is probable that PFC-laden water from the Canary Current was transported to the west and northwest by the North and South Equatorial Currents and possibly further south along the coast of Brazil by the Brazil Current. The decreasing latitudinal gradient is consistent with the decreasing gradient of PFOA and PFOS on airborne particles described by Jahnke et al. (26). What it is not known, however, is whether the airborne particle-bound fraction originated from sea spray or atmospheric degradation of volatile precursor compounds, i.e., whether the ocean was the source or sink for this airborne contamination. It is noteworthy that the pattern of importance of individual PFCs changed depending on the sampling area. In the northeast of the Atlantic Ocean in the Bay of Biscay, all PFCs, except for PFHpA, were detected, with the latter detected for the first time at 37 N. The concentration ranged from several tens of picograms per liter (PFBS, PFHxA, and PFNA) to a few hundreds of picograms per liter (PFOS, PFOA, and FOSA). The concentration of PFSAs, PFOS, and PFBS dropped below the MDLs south of 32 N and 25 N, respectively. The occurrence of PFCAs toward the south depended on their chain length, with the longer the chain length of the PFCAs (C6 to C9), the further southward they were detected. The reason for this behavior could be different physicochemical characteristics (e.g., vapor pressure, partition coefficient) and/or input from atmospheric sources (17, 8). The increasing PFC concentration at sampling stations 24-28, attended by an increasing water temperature, could be caused by higher rainfall in this area, leading to increased deposition of PFCAs from the atmosphere (27). Of all detected PFCs, FOSA was found furthest south, down to 4 S, possibly as a result of its higher vapor pressure increasing the importance of atmospheric transport. The location of the Labrador, North Atlantic, and Canary Currents will similarly affect the distribution of PFCs along the longitudinal transect (see Figures 1 and 3). PFC concentrations dropped off in sampling stations C-F. These results imply that the northeastern samples were influenced 3126 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009 FIGURE 4. Correlations between PFNA and PFOA concentrations in surface water in the Atlantic Ocean. by the Labrador Current, whose origin is the Arctic Ocean and is relatively "clean" (15), and the low ocean temperatures in this area support this hypothesis. In contrast, sampling stations A and B had elevated concentrations and were probably influenced by inputs from the Canadian coast and North Atlantic Current, respectively. In the area of the North Atlantic Current, PFOS, PFHpA, and PFOA dominated, with sum concentrations of 52-117 pg L-1 (sampling stations G-O). This is twice the concentration found at sampling stations C-F influenced by the Labrador Current. The warm temperature of the North Atlantic Current became noticeable at sampling stations L and M, which were at latitudes similar to those of stations C-F, but with surface water temperatures much higher than those close to the Labrador Sea. The highest concentrations where found at sampling stations P-R, which could be induced by the Canary Current carrying PFCs from the European Continent source region. The concentrations of PFNA and PFOA were positively correlated (r2 ) 0.52; see Figure 4), which indicates that the sources of both compounds are related (28). Young et al. found a positive correlation with a gradient of 1 in snow on remote ice caps that were contaminated atmospherically by precursors, in this case 8:2 FTOH (28). The gradient in the Atlantic Ocean is 0.4, and it is possible that the degradation of perfluoalkanesulfonamides could lead to a higher amount of PFOA relative to PFNA (29). Plots of PFHxA, PFOA, and PFNA concentrations versus PFOS concentrations are also correlated for the Atlantic Ocean, but the significance is lower because the calculations are based on only a few data points (Figure S2 in the Supporting Information). Young et al. found no correlation among PFHxA, PFNA, and PFOS on the ice cap, which further supported the source there being the atmospheric pathway (28). Concentrations of PFHpA and PFOA were negatively correlated in the Atlantic Ocean (r2 ) 0.32, see Figure S2), probably as a consequence of the patchy distribution of PFHpA. Simcik et al. found that the ratio of PFHpA to PFOA increased with increasing distance from nonatmospheric sources and suggested that a high ratio would be a good tracer of atmospheric deposition (30). In this study, only a few stations (15, 20, M, and N) had PFHpA/ PFOA ratios greater than 1, which suggests that direct releases are important determinants of open Atlantic Ocean surface water concentrations. Several reasons for the distribution pattern of PFCs in the Atlantic Ocean have been suggested, but for each location several factors could be responsible for the occurrence of the PFCs. Ocean currents and related dilution effects have a crucial influence on PFC distribution (15). The spatial distribution data obtained in this study are useful for global transportation models (31), in which industrial areas are considered as source of PFCs, and ocean waters and the atmosphere are important as sinks and for transportation of these compounds. This transportation to remote regions could have adverse effects in top predators here, because of the high bioaccumulation potential for PFOS and longer chained PFCAs in the marine food web (32). Further investigations of the biochemical cycle of PFCs in ocean waters are necessary for understanding the transportation and fate of PFCs in the marine environment. Acknowledgments We kindly acknowledge the German Federal Environmental Foundation for sponsoring the project. We kindly acknowledge the Alfred Wegener Institute for Polar and Marine Research (AWI), Bremerhaven, for the possibility of taking part in expedition ANTXXIV-1. We thank Armando Caba for performing sampling onboard the R/V Maria S. Merian and Clare Benskin, Claudia Moeckel, and Jasmin Shuster for help with sampling onboard the R/V Polarstern. Note Added after ASAP Publication Reference 18 was modified in the version of this paper published March 30, 2009; the corrected version published ASAP April 6, 2009. Supporting Information Available Additional information about the location of the sampling sites, target analytes, and method recovery rates, overview of PFC concentrations in the Atlantic Ocean, chromatograms of a real sample and a blank sample, and correlations between PFC concentrations. This information is available free of charge via the Internet at http://pubs.acs.org. Literature Cited (1) Yamashita, N.; Kannan, K.; Taniyasu, S.; Horii, Y.; Petrick, G.; Gamo, T. A Global Survey of Perfluorinated Acids in Oceans. Mar. Pollut. Bull. 2005, 51, 658-668. (2) Giesy, J. P.; Kannan, K. Global Distribution of Perfluorooctane Sulfonate in Wildlife. Environ. Sci. Technol. 2001, 35, 1339- 1342. (3) Hansen, K. J.; Clemen, L. A.; Ellefson, M. E.; Johnson, H. O. Compound-Specific, Quantitative Characterization of Organic Fluorochemicals in Biological Matrices. Environ. Sci. Technol. 2001, 35, 766-770. (4) Kissa, E. Fluorinated Surfactants and Repellents; Marcel Dekker: New York, 2001; Vol. 97. (5) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Bioconcentration and Tissue Distribution of Perfluorinated Acids in Rainbow Trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204. (6) Austin, M. E.; Kasturi, B. S.; Barber, M.; Kannan, K.; Mohan Kumar, P. S.; Mohan Kumar, S. M. J. Neuroendocrine Effects of Perfluorooctane Sulfonate in Rats. Environ. Health Perspect. 2003, 111, 1485-1489. (7) Oakes, K. D.; Sibley, P. K.; Solomon, K. R.; Mabury, S. A.; Van Der Kraak, G. J. Impact of Perfluorooctanoic Acid on Fathead Minnow (Pimephales Promelas) Fatty Acyl-CoA Oxidase Activity, Circulating Steroids, and Reproduction in Outdoor Microcosms. Environ. Toxicol. Chem. 2004, 23, 1912-1919. (8) Ellis, D. A.; Martin, J. W.; Mabury, S. A.; De Silva, A. O.; Hurley, M. D.; Sulbaek Anderson, M. D.; Wallington, T. J. Degradation of Fluorotelomer Alcohols: A Likely Atmospheric Source of Perfluorinated Carboxylic Acids. Environ. Sci. Technol. 2004, 38, 3316-3321. (9) Armitage, J.; Cousins, I. T.; Buck, R. C.; Prevedouros, K.; Russell, M. H.; Macleod, M.; Korzeniowski, S. H. Modeling Global-Scale Fate and Transport of Perfluorooctanoate Emitted from Direct Sources. Environ. Sci. Technol. 2006, 40, 6969-6975. (10) McMurdo, C. J.; Ellis, D. A.; Webster, E.; Butler, J.; Christensen, R. D.; Reid, L. K. Aerosol Enrichment of the Surfactant PFO and Mediation of the Water-Air Transport of Gaseous PFOA. Environ. Sci. Technol. 2008, 42, 3969-3974. (11) Shoeib, M.; Harner, T.; Vlahos, P. Perfluorinated Chemicals in the Arctic Atmosphere. Environ. Sci. Technol. 2006, 40, 7577- 7583. (12) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, Fate and Transport of Perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44. (13) Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A. Identification of Long-Chain Perfluorinated Acids in Biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380. (14) Wei, S.; Chen, L. Q.; Taniyasu, S.; So, M. K.; Murphy, M. B.; Yamashita, N.; Yeung, L. W. Y.; Lam, P. K. S.; Wei, S.; Chen, L. Q. Distribution of Perfluorinated Compounds in Surface Seawaters between Asia and Antarctica. Mar. Pollut. Bull. 2008, 54, 1813- 1838. (15) Yamashita, N.; Taniyasu, S.; Petrick, G.; Wei, S.; Gamo, T.; Lam, P. K. S.; Kannan, K. Perfluorinated Acids as Novel Chemical Tracers of Global Circulation of Ocean Waters. Chemosphere 2008, 70, 1247-1255. (16) Taniyasu, S.; Kannan, K.; So, M. K.; Gulkowska, A.; Sinclair, E.; Okazawa, T.; Yamashita, N. Analysis of Fluorotelomer Alchols, Fluorotelomer Acids, and Short- and Long-Chain Perfluorinated Acids in Water and Biota. J. Chromatogr., A 2005, 1093, 89-97. (17) Higgins, C. P.; Luthy, R. G. Sorption of Perfluorinated Surfactants on Sediment. Environ. Sci. Technol. 2006, 40, 7251-7256. (18) Ahrens, L.; Siebert, U.; Ebinghaus, R. Total Body Burden and Tissue Distribution of Polyfluorinated Compounds in Harbor Seals (Phoca vitulina) from the German Bight. Mar. Pollut. Bull. 2009, 58, 520-525. (19) van Leeuwen, S. P. J.; de Boer, J. Extraction and Clean-up Strategies for the Analysis of Poly- and Perfluoroalkyl Substances in Environmental and Human Matrices. J. Chromatogr., A 2007, 1153, 172-185. (20) Yamashita, N.; Kannan, K.; Taniyasu, S.; Horii, Y.; Okazawa, T.; Petrick, G.; Gamo, T. Analysis of Perfluorinated Acids at PartsPer-Quadrillion Levels in Seawater Using Liquid Chromatography-Tandem Mass Spectrometry. Environ. Sci. Technol. 2004, 38, 5522-5528. (21) Ju, X.; Jin, Y.; Sasaki, K.; Saito, N. Perfluorinated Surfactants in Surface, Subsurface Water and Microlayer from Dalian Coastal Waters in China. Environ. Sci. Technol. 2008, 42, 3538-3542. (22) Theobald, N.; Gerwinski, W.; Jahnke, A. Occurrence of Perfluorinated Organic Acids in Surface Sea-Water of the East Atlantic Ocean between 53 North and 30 South. Poster presentation at the SETAC Europe Annual Meeting, Porto, Portugal, May 20-24, 2007. (23) D'Eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A. Atmospheric Chemistry of N-Methyl Perfluorobutane Sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and Mechanism of Reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868. (24) 3M Phase-Out Plan for POSF-Based Products; USEPA Docket ID OPPT-2002-0043; U.S. EPA: Washington, DC, 2000. (25) Dietz, R.; Bossi, R.; Riget, F. F.; Sonne, C.; Born, E. W. Increasing Perfluoroalkyl Contaminants in East Greenland Polar Bears (Ursus maritimus): A New Toxic Threat to the Arctic Bears. Environ. Sci. Technol. 2008, 42, 2701-2707. (26) Jahnke, A.; Berger, U.; Ebinghaus, R.; Temme, C. Latitudinal Gradient of Airborne Polyfluorinated Alkyl Substances in the Marine Atmosphere between Germany and South Africa (53 N-33 S). Environ. Sci. Technol. 2007, 41, 3055-3061. (27) Scott, B. F.; Spencer, C.; Mabury, S. A.; Muir, D. C. G. Poly and Perfluorinated Carboxylates in North American Precipitation. Environ. Sci. Technol. 2006, 40, 7167-7174. (28) Young, C. J.; Furdui, V. I.; Franklin, J.; Koerner, R. M.; Muir, D. C.; Mabury, S. A. Perfluorinated Acids in Arctic Snow: New Evidence for Atmospheric Formation. Environ. Sci. Technol. 2007, 41, 3455-3461. (29) Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J. Atmospheric Chemistry of Perfluoroalkanesulfonamides: Kinetic and Product Studies of the OH Radical and Cl Atom Initiated Oxidation of N-ethyl Perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 864-872. (30) Simcik, M. F.; Dorweiler, K. J. Ratio of Perfluorochemical Concentrations as a Tracer of Atmospheric Deposition to Surface Waters. Environ. Sci. Technol. 2005, 39, 8678-8683. (31) Wania, F.; Mackay, D. A Global Distribution Model for Persistent Organic Chemicals. Sci. Total Environ. 1995, 160/161, 211-232. (32) Houde, M.; Martin, J. W.; Lethcher, R. J.; Solomon, K. R.; Muir, D. C. G. Biological Monitoring of Polyfluoroalkyl Substances: A Review. Environ. Sci. Technol. 2006, 40, 3463-3473. ES803507P VOL. 43, NO. 9, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3127 ELSEVIER Chemosphere 76 (2009) 179-184 Contents lists available at ScienceDirect Chemosphere journal homepage: www.elsevier.com/locate/chemosphere Spatial distribution of polyfluoroalkyl compounds in seawater of the German Bight Lutz Ahrens a'b'*, Sebastian Felizeter a'c, Ralf Ebinghaus 'GKSS Research Centre Geesthacht, Institute for Coastal Research, Max-Planck-Str. 1, D-21502 Geesthacht, Germany b Leuphana University of Luneburg, Institute for Ecology and Environmental Chemistry, D-21335 Luneburg, Germany Carl-von-Ossietzky-University of Oldenburg, Institute for Chemistry and Biology of the Marine Environment, D-26129 Oldenburg, Germany ARTICLE INFO Article history: Received 30 January 2009 Received in revised form 24 March 2009 Accepted 24 March 2009 Available online 25 April 2009 Keywords: PFCs Spatial distribution PFOS PFOA German Bight ABSTRACT The spatial distribution of polyfluoroalkyl compounds (PFCs) and their composition profile was investigated in 48 water samples collected from the German Bight. All samples were prepared by solid-phase extraction with Strata XAW cartridges and analysed using high performance liquid chromatography/negative electrospray ionisation-tandem mass spectrometry (HPLC/(-)ESI-MS/MS). Concentrations of various PFCs, including perfluorinated sulfonates (PFSAs), perfluorinated carboxylic acids (PFCAs), unsaturated fluorotelomercarboxylic acids, perfluoralkyl sulfonamide and sulfonamidoethanol, were quantified. The EPFC concentration ranges from 9.36 ng L-1 to 31.2 ng L-1, while perfluorobutane sulfonate (PFBS, 3.38-17.7 ng L-1) and perfluorooctanoic acid (PFOA, 2.67-7.83 ng L-1) dominated. The rivers Elbe, Weser and Ems had a high influence on the distribution of most PFCs in the German Bight, with maximum PFC concentrations found in their estuaries, and concentrations decreasing with increasing distance from the coast. Conversely, PFBS had its maximum concentration not in the estuaries but in the western German Bight, which suggest an additional source, where PFBS was transported into the German Bight with the westerly current. 2009 Elsevier Ltd. All rights reserved. 1. Introduction Polyfluoroalkyl compounds (PFCs) are persistent against typical environmental degradation processes and are ubiquitous in the environment, having been found in water (Yamashita et al., 2005), air (Jahnke et al., 2007) and organisms (Giesy and Kannan, 2001) around the globe. Because of their unique physicochemical properties due to their combination of lipophilic and hydrophilic characteristics, PFCs have been widely used in many consumer products, such as polymerisation aids, stain repellents on carpets, textiles, leather, and paper products for over 50 years (Kissa, 2001). From the production and use of these products, PFCs can be released into the environment. In general, neutral PFCs like perfluoroalkyl sulfonamides and fluorotelomer alcohols are less water-soluble and more volatile than perfluorinated acids. In the atmosphere as well as under aerobic conditions, e.g. in activated sludge, they can be degraded to perfluorinated carboxylic acids (PFCAs) and perfluorinated sulfonates (PFSA5) (Ellis et al., 2004; Martin et al., 2006; Rhoads et al., 2008). Previous studies examined the release of PFCs into the aqueous environment by runoff from contaminated soil or waste * Corresponding author. Address: GKSS Research Centre Geesthacht, Institute for Coastal Research, Max-Planck-Str. 1, D-21502 Geesthacht, Germany. Tel.: fax: E-mail address: @gkss.de (L. Ahrens). 0045-6535/$ - see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2009.03.052 water treatment plants (WWTPs) (Schultz et al., 2006; Skutlarek et al., 2006) and their riverine transportation (McLachlan et al., 2007). The longer-chained PFCs are known to be bioaccumulative (Martin et al., 2003) and have toxic effects in biota (Austin et al., 2003; Oakes et al., 2004). As a result, the 3M Company, major producer of perfluorooctyl sulfonyl fluoride (POSF, which is a major precursor for several PFCs), voluntarily phased out the production in 2002, but a variety of related PFCs are still being produced by other manufacturers (Prevedouros et al., 2006). In addition the European Union (EU) formed a directive in October 2006, which prohibits the general use of perfluorooctane sulfonate (PFOS) and their derivates from June 2008 (European Parliament and Council, 2006). The former POSF-based products are now substituted by perfluorobutyl sulfonyl fluoride (PBSF)-based products. Highest PFC concentrations are found in top predators (Kannan et al., 2001). Only a little is known about how PFCs reach the marine environment and about their spatial distribution in the coastal area. In this study 48 surface water samples were collected in coastal water from the German Bight for the determination of PFCs in the water phase. We investigated the spatial distribution of 18 PFCs in coastal water and their composition profiles to identify sources in the urbanized/industrial sampling area. Furthermore, the relationship between concentrations of PFCs, and dissolved organic carbon (DOC) as well as suspended particulate matter (SPM) were examined. 180 L. Ahrens et al. / Chemosphere 76 (2009) 179-184 2. Materials and methods 2.1. Chemicals and standards The standards used in this study are described elsewhere (Ahrens et al., 2009) and are listed in Table S1 in the Supplementary material. Methanol (SupraSolv), acetonitrile (LiChrosolv), ammonium hydroxide (25% for analysis), formic acid (98-100% suprapure) and ammonium acetate were purchased from Merck (Darmstadt, Germany). 2.2. Sample collection and sample pretreatment Surface water samples were taken onboard the research vessel Ludwig Prandtl at 48 sampling stations in the German Bight in August 2007 (see Fig. 1). Details of the sampling and the physicochemical parameters of the water samples are presented in Table S2 in the Supplementary material. Sampling sites were chosen to show the influence of the rivers Elbe, Weser and Ems in comparison to the westerly current along the coast of the German Bight. 5 L water samples were collected in brown glass bottles via a metal ships' intake system at approximately 1 m below the surface. In addition, at sampling stations 7, 29, 35 and 43 duplicate samples were collected for quality control. The samples were filtered directly after sampling onboard using glass fibre filters (GFF, GC/C, Whatman, 47 mm, >1.2 lm). The filtrated samples were stored at 4 C prior to solid-phase extraction (SPE) on board ship on the same or following days. Five field blanks (FB) were taken to test for possible blank contamination. For the FB, 100 mL Millipore water (Millipore, Elix 5 and Millipore Milli Q Plus) was added to a 5 L brown glass bottle and then put through SPE extraction. The FB were then stored and extracted in the same manner as ``real" samples. 2.3. Solid-phase extraction The filtrate was extracted by SPE with Strata XAW cartridges (Phenomenex, 500 mg, 12 cc, 33 lm), similar as described else- where (Taniyasu et al., 2005). Prior to the extraction, the samples were spiked with 10 ng absolute of an internal standard (IS) mix (i.e., [13C2]-PFHxA, [13C4]-PFOA, [13C4]-PFNA, [13C4]-PFDA, [13C2]-PFUnDA, [13C2]-PFDoDA, [18O2]-PFHxS, [13C4]-PFOS, [13C4]-PFOSi, d3-MeFOSA, d5-EtFOSA, d7-MeFOSE, d9-EtFOSE, [13C2]-FHEA, [13C2]-FOEA, [13C2]-FDEA, [13C2]-FHUEA, [13C2]-FOU- EA, [13C2]-FDUEA, 100 lL of a 0.1 lg mL1 solution, see Table S1 in the Supplementary material). Briefly, after preconditioning with 10 mL methanol and Millipore water, the cartridge was loaded with the 5 L sample at approximately 5 drops per second. The cartridge was then washed with 10 mL 0.1% formic acid in Millipore water and dried for 30 min under vacuum. The elution was divided into two parts: the sulfonamides were eluted with 20 mL acetonitrile; thereafter the acids were eluted with 15 mL 0.1% ammonium hydroxide in methanol. Both extracts were collected separately in brown glass vials and closed with a phenolic resin/aluminium caps. The samples were stored in a freezer at 20 C after the elution steps on board ship. After the end of the sampling cruise, both extracts were concentrated in a clean lab (class 10 000) to 2 mL using rotary evaporator within a few days of arrival at the lab. Fi- nally, both extracts were combined and reduced to 150 lL under a nitrogen stream and spiked with 20 ng absolute of the injection standard d5-EtFOSAA (InjS, 50 lL of a 0.4 lg mL1 solution, see Ta- ble S1 in the Supplementary material). 2.4. Instrument analysis An HP 1100 HPLC-system (Agilent Technologies) was used with a Synergi Hydro RP 80A column (150 2 mm, 4 lm) by Phenomenex, combined with a suitable guard column: Synergi 2 l Hydro RP Mercury (20 2 mm, 2 lm). Modifications of the HPLC system were made as described elsewhere (Yamashita et al., 2004) to eliminate instrumental blank contamination. The triple-quadrupole mass spectrometer, supplied by Applied Biosystems/MDS SCIEX (API 3000), used an electrospray ionisation (ESI) interface in negative ionisation mode (for details see (Ahrens et al., 2009)). 2.5. Data analysis Quantification was performed by the internal standard method with an external calibration. A 10-point calibration curve (1, 5, 10, 25, 50, 100, 500, 1000, 2000 and 3000 pg injected) was used for calculation. For the compounds PFPS, PFNS, PFPeDA and PFHpDA no standards were available, thus these PFCs were calculated from the calibrations of corresponding substances with plus and minus one carbon atom in the carbon chain. For peak integration only the main peak of a compound was used. The isomers were not included in the peak integration, because of the lack of standards. 2.6. Quality assurance The analytical quality of the laboratory has been approved in interlaboratory studies (van Leeuwen et al., 2009). As standard procedure, instrument detection limits (IDLs), method quantification limits (MQLs), FB (see Table 1), recoveries and duplicate samples were examined. After the removal of all Teflon parts from the HPLC system, no instrument blank was detected. In some FB contamination levels of PFOA, PFUnDA and PFDoDA were found, but all the concentrations were below the MQL. IDLs and MQLs were calculated for sub- Fig. 1. Map showing the sampling locations in the German Bight. L. Ahrens et al. / Chemosphere 76 (2009) 179-184 181 Table 1 Instrument detection limits (IDLs), method quantification limits (MQLs) and field blank concentrations for the German Bight survey in August 2007.a IDLb (pg absolute) MQLc (ng L1) Field blanks (ng L1) PFBS PFPSd PFHxS PFOS PFNS 6:2 FTS PFPA PFHxA PFHpA PFOA PFNA PFDA PFUnDA PFDoDA FOSA MeFBSA MeFBSE FDUEA 0.50 0.31 0.48 0.17 1.36 0.51 0.27 0.36 0.36 0.35 0.40 0.29 0.37 0.33 1.25 1.09 0.58 0.367 0.080 0.097 0.120 0.072 0.120 0.158 0.084 0.077 0.067 0.039 0.047 0.019 0.008 0.004 0.180 0.241 0.016 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.-(0.022) n.d. n.d. n.d.-(0.006) n.d.-(0.004) n.d. n.d. n.d. n.d. a n.d. = not detected. Blank levels were calculated from a sample volume of 100 mL Millipore water. Details are given in the text. Values in brackets were below the respective method quantification limit (MQL). b IDL (ng absolute) at 3 times of the signal to noise in the calibration standards (n = 4). c MQL (ng L1) at 10 times of the signal to noise in natural samples (n = 4). d Have to be considered as estimates, because no standards were available for this compound. stances that were found in real samples using the signal to noise ratios of 3 and 10, respectively. The IDLs were usually lower than 1 pg absolute, while the MQLs were in low ppq level for the 5 L water samples. These MQLs were approximately three times lower in comparison of using only 1 L sample volume. Conversely to a previous study (Yamashita et al., 2004), for this method the background noise increase was negligible in comparison to the increasing target peak response. The mean recoveries of the IS ranged from 23% (d3-MeFOSA) to 102% ([13C2]-FHUEA). The low recoveries of the perfluoroalkyl sulfonamides could possibly due to a breakthrough because of the high water volume of 5 L. However 19 IS were used to correct matrix effects as well as losses during sampling, sample extraction, concentration, and analysis. Duplicate samples showed a good agreement with a relative standard deviation of lower than 20% for each compound. 3. Results and discussion 3.1. Concentrations of PFCs in the German Bight Overall 18 of the 39 examined analytes were found at the 48 sampling stations. The PFCs quantified included C4-C6, C8 and C9 PFSAs, 6:2 FTS, C5-C12 PFCAs, MeFBSA, FOSA, MeFBSE and FDUEA (see Tables SP3-S5 in the Supplementary material). The spatial distribution of PFC and individual PFCPconcentrations in the German Bight is shown in Fig. 2. The PFC concentration ranges from 9.36 ng L1 (sampliPng station 26) to 31.2 ng L1 (sampling station 13). The highest PFC concentrations were found in the wElebset,eWrnessearmapnldinEgmstsa. tTihoen P24PFaCndcotnhceeensttruaatiroynms doeuctrhesasoefdthbey raivfaecrstor of 2-3 towards the offshore stations. The dominant compound of the PFSAs was PFBS with concentrations ranging from 3.38 to 17.7 ng L1, while PFOS was detected in concentrations ranging from 0.69 to 3.95 ng L1. The PFCA concentrations were dominated by PFOA (2.67-7.83 ng L1) and PFHxA (0.47-9.56 ng L1), whereas the longer-chaPined PFCAs (C9-C12) had usually a contribution of under 3% of the PFCs. Four precursor compounds of the PFCAs and PFSAs were detected (i.e., MeFBSA, FOSA, MeFBSE and FDUEA), with a concentration level of lower than 0.75 ng L1. Overall, PFBS was the predominated PFC in the German Bight with a composition of 40%, followed by PFOA (26%), PFOS (9%) and PFHxA (8%). Interestingly, the compounds PFOS and FOSA contained a contribution of 60% branched isomers, while the PFCAs had a contribution of <10% and for the other PFCs no branched isomers were observed. It was previously hypothesized that the presence of branched isomers may indicate exposure from historical releases of electrochemical fluorination (ECF) manufacturing process (De Silva and Mabury, 2006) or may be a sign of local ECF production. However, the specific distribution and composition profile of PFCs indicates an input from the rivers and western current and additionally an atmSoigsnpihfiecraicntdceoprorseiltaiotinonisbpeotwsseibelne.DOC and PPFC, PFOS and PFOA concentration (p < 0.0001, see Fig. 3) and PFHxS, C6-C10 PFCA and FOSA concentration (p < 0.0001, see Fig. S1 in the Supplementary material), respectively, was found, while no correlation had the DOC with PFBS, PFPA and PFUnDA. This correlation corresponds with the positive relationship between sorption of PFCs to sediment and total organic carbon (TOC) amount (Higgins and Luthy, 2006). Furthermore, significant correlation between SPM and PFPS, PFHxS, PFHpA, PFOA, PFNA (p < 0.0001), PFOS (p < 0.001), PFHxA (p < 0.012), PFDA (p < 0.007) and FOSA (p < 0.024) was observed (see Fig. S2 in the Supplementary material). This indicates that sedimentation could be an effective removal mechanism for PFCs in the water phase. However, the DOC and SPM concentrations decreased with increasing distance from the coast, which suggests that the distance from the coast has a high influence on the distribution and concentration level of PFCs in the aquatic environment. Further investigations of their physical state are necessary. 3.2. Identification of sources of individual PFCs Most PFCs had their highest concentrations in the river estuary and decreased with increasing distance from the coast. But the composition pattern in the three estuaries Ems, Weser and Elbe were not the same (Fig. 1). PFOA and PFNA had a similar distribution in all estuaries, whereas PFHpA and PFDA were mainly distributed in the rivers Weser and Elbe. Interestingly, PFHxA was mainly found in the river Weser and not in the other both rivers. PFHxS and PFOS were distributed in all river estuaries with highest concentrations in the estuary of the river Ems. Investigations by Caliebe et al. (2004) determined mean concentrations of PFCs in the river Elbe in 2003 of about 20 ng L1 for PFOA and PFOS, and 1- 3 ng L1 for other PFCs like PFHxA, PFNA, PFDA, PFHxS and FOSA. Another study in the river Elbe reported PFHxA, PFHpA, PFOA and PFNA concentrations of 15.4 ng L1, 2.7 ng L1, 7.6 ng L1 and 0.27 ng L1, respectively, in 2005 (McLachlan et al., 2007). These concentrations were 2-5 times higher than measured in the estuaries in this study. Potential sources for PFCs into the rivers Ems, Weser and Elbe could be caused by effluents from domestic and industrial WWTPs and/or diffuse sources. The two big cities Bremen and Hamburg, located at the rivers Weser and Elbe, respectively, could be an additionally source for PFCs because of their several industries like petroleum, textile, paper and polymer industries. Furthermore, the rivers transported the PFCs from the source regions into the German Bight, while their discharge (the mean discharge for the rivers Ems, Weser and Elbe was 105 m3 s1, 433 m3 s1 and 697 m3 s1, respectively, for the year 2007) could have a high influence on the distribution pattern of PFCs in the German Bight. Finally, McLachlan et al. (2007) estimated a total flux of 0.26 (PFNA) to 14.3 (PFOA) tonnes per year for 14 major rivers in Europe indicating the rivers as the major input pathway into the marine environment. 182 L. Ahrens et al. / Chemosphere 76 (2009) 179-184 Fig. 2. Spatial distribution of PPFCs and individual PFC concentrations in the German Bight in ng L1. Note: the different circle sizes at the sampling stations are in proportion to the concentration which is shown on the right side of each map. Fig. 3. Relationship between concentrations of PPFCs, PFOS, PFOA and dissolved organic carbon (DOC) in surface water in the German Bight. The occurrence of PFPS and PFNS at low concentration levels (0.06-0.86 ng L1 and 0.11-0.28 ng L1, respectively) could be due to the POSF-based and PBSF-based production, where uneven carbon chained PFSAs can be produced as by-products (Giesy and Kannan, 2002). In contrast to the other PFCs, PFBS composition decreased from 40-60% in the offshore area to 15-20% in the river estuary and its highest composition was found at the western sampling stations 20 (58%) and 24 (61%). This is probably the result of an additional source, where PFBS was transported into the German Bight with the westerly current. It is possible that this contamination of PFBS was originating from the river Rhine, where concentrations of up to 46 ng L1 were found (Skutlarek et al., 2006). Positive correlations between C6-C10 PFCAs and C5, C6 and C8 PFSAs (see Table S6 in the Supplementary material) suggest a common pollution source of these compounds into the marine environment. Possible sources could be the effluents of WWTPs (Schultz et al., 2006) and rain or surface runoff (Kim and Kannan, 2007). In addition, the usage of aqueous film-forming foams (AFFF) could be a source of PFCs, which correspond with the detection of 6:2 FTS in this study (Schultz et al., 2004). On the other hand, the detection of the perfluoralkyl sulfonamide (i.e., MeFBSA, FOSA, MeFBSE) and FDUEA indicates atmospheric deposition and/or incomplete biodegradation (Loewen et al., 2005; Martin et al., 2006; Rhoads et al., 2008). So et al. (2004) found differences in the distribution patterns due to a seasonal shift of the water currents. The circulation in the North Sea is stable throughout the year in contrast to the situation at the Pearl River Delta, China, and thus it is unlikely that the distribution pattern in the North Sea will change dramatically. Nevertheless, unusual weather conditions like east wind might change the pattern on a smaller scale. 3.3. Comparison of PFC concentrations in the German Bight with other coastal water studies Minimum and maximum PFOS and PFOA concentrations in coastal water are shown in Table 2. The concentration of PFOS and PFOA in previous studies in the German Bight, Pearl River Delta (China), coastal area of Hong Kong, coastal area of Korea, coastal area of Dalian (China) were in the same range as in this study (Caliebe et al., 2004; So et al., 2004; Yamashita et al., 2005; Theobald L. Ahrens et al. / Chemosphere 76 (2009) 179-184 Table 2 Global comparison of PFOA and PFOS concentrations in seawater from the German Bight with coastal water studies from other areas. Location Coastal area of Japan German Bight, Germany Coastal area of South Korea Perl River Delta, China Hong Kong, China Tokyo Bay, Japan Coastal area of Hong Kong Coastal area of Korea South China Sea German Bight, Germany West Baltic Sea Coastal area of Dalian, China German Bight, Germany Concentration (ng L1) PFOS <2.5-59 0.25-7.0 0.04-730 0.02-12 0.09-3.1 0.34-58 0.07-2.6 0.04-2.5 0.008-0.11 0.28-3.1 0.33-0.90 <0.10-2.3 0.69-3.95 PFOA n.a. 3-13 0.24-320 0.24-16 0.73-5.5 1.8-192 0.67-5.5 0.24-11 0.16-0.42 0.54-5.9 0.47-1.1 0.17-38 2.67-7.83 n.a. = not available. 183 References Taniyasu et al. (2003) Caliebe et al. (2004) So et al. (2004) So et al. (2004) So et al. (2004) Yamashita et al. (2005) Yamashita et al. (2005) Yamashita et al. (2005) Yamashita et al. (2005) Theobald et al. (2007) Theobald et al. (2007) Ju et al. (2008) This study et al., 2007; Ju et al., 2008). Lower concentrations of PFOS and PFOA were only found in the South China Sea and West Baltic Sea (Yamashita et al., 2005; Theobald et al., 2007). Highest concentrations of PFOA (320 ng L1) and PFOS (720 ng L1) were found at a contaminated coastal area of South Korea (So et al., 2004). In general, except for the coastal area of South Korea, the concentration of PFOA was higher than of PFOS, which suggests that similar sources exist in the urbanized/industrial coastal areas at the different locations. This corresponds with decreased contamination levels of PFOS and PFOA with increasing distance from the coast. 4. Conclusions Only a few studies exist about the spatial distribution of PFCs in coastal waters. It is very important to identify the sources for individual PFCs and their distribution mechanism. High concentrations (>1 ng L1) of shorter-chained PFCAs (C5-C8), PFBS, PFHxS and PFOS in coastal water of the German Bight suggests that these compounds are entering the marine environment from rivers and can potentially undergo long-range transportation via the ocean currents (Yamashita et al., 2005). Dilution processes and/or adsorption to suspended particle matter could be responsible for the decreasing concentrations, however, PFCs are very persistent and only the longer-chained PFCs (PC8) have a high affinity to the sediment (Higgins and Luthy, 2006; Theobald et al., 2007). In addition, long-range transportation could also be possible by the atmosphere, which is in agreement with the detection of the precursor compounds MeFBSA, FOSA, MeFBSE and FDUEA. Most PFCs had their highest concentrations in the river estuaries, but in contrast, PFBS had its maximum concentration in the western German Bight. This study suggests that PFBS has a significant source outside the study region, which makes research on the short-chained PFCs even more important. However, most studies have usually focused on PFOA and PFOS and future studies should be expanded to include the shorter-chained PFCAs and PFSAs. The occurrence of high concentrations of PFCs in coastal water could possibly be problematic, because they are bioavailable and can accumulate in the marine food chain. Chemical `fingerprints' may help to identify specific sources of PFC contamination into the aqueous environment. This research, the spatial distribution of PFCs in coastal area, is very important for the understanding of the transportation and fate of PFCs in the marine environment. Acknowledgement We kindly acknowledge the German Federal Environmental Foundation for sponsoring this project. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.chemosphere.2009.03.052. References Ahrens, L., Siebert, U., Ebinghaus, R., 2009. Total body burden and tissue distribution of polyfluorinated compounds in harbor seals (Phoca vitulina) from the German Bight. Mar. Poll. Bull. 58, 520-525. Austin, M.E., Kasturi, B.S., Barber, M., Kannan, K., Mohan Kumar, P.S., Mohan Kumar, S.M.J., 2003. Neuroendocrine effects of perfluorooctane sulfonate in rats. Environ. Health Persp. 111, 1485-1489. Caliebe, C., Gerwinski, W., Hhnerfuss, H., Theobald, N., 2004. Occurrence of perfluorinates organic acids in the water of the North Sea. Organohalogen Compd. 66, 4074-4078. De Silva, A.O., Mabury, S.A., 2006. Isomer distribution of perfluorocarboxylates in human blood: potential correlation to source. Environ. Sci. Technol. 40, 2903- 2909. Ellis, D.A., Martin, J.W., Mabury, S.A., De Silva, A.O., Hurley, M.D., Sulbaek Anderson, M.D., Wallington, T.J., 2004. Degradation of fluorotelomer alcohols: a likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 38, 3316-3321. Giesy, J.P., Kannan, K., 2001. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 35, 1339-1342. Giesy, J.P., Kannan, K., 2002. Perfluorochemical surfactants in the environment. These bioaccumulative compounds occur globally, warranting further study. Environ. Sci. Technol. 36, 146A-152A. Higgins, C.P., Luthy, R.G., 2006. Sorption of perfluorinated surfactants on sediment. Environ. Sci. Technol. 40, 7251-7256. Jahnke, A., Berger, U., Ebinghaus, R., Temme, C., 2007. Latitudinal gradient of airborne polyfluorinated alkyl substances in the marine atmosphere between Germany and South Africa (53 N-33 S). Environ. Sci. Technol. 41, 3055-3061. Ju, X., Jin, Y., Sasaki, K., Saito, N., 2008. Perfluorinated surfactants in surface, subsurface water and microlayer from Dalian coastal waters in China. Environ. Sci. Technol. 42, 3538-3542. Kannan, K., Koistinen, J., Beckmen, K., Evans, T., Gorzelany, J.F., Hansen, K.J., Jones, P.D., Helle, E., Nyman, M., Giesy, J.P., 2001. Accumulation of perfluorooctane sulfonate in marine mammals. Environ. Sci. Technol. 35, 1593-1598. Kim, S.-K., Kannan, K., 2007. Perfluorinated acids in air, rain, snow, surface runoff, and lakes: relative importance of pathways to contamination of urban lakes. Environ. Sci. Technol. 41, 8328-8334. Kissa, E., 2001. Fluorinated Surfactants and Repellents, vol. 97. Marcel Dekker. Loewen, M., Halldorson, T., Wang, F., Tomy, G., 2005. Fluorotelomer carboxylic acids and PFOS in rainwater from an urban center in Canada. Environ. Sci. Technol. 39, 2944-2951. Martin, J.W., Ellis, D.A., Mabury, S.A., Hurley, M.D., Wallington, T.J., 2006. Atmospheric chemistry of perfluoroalkanesulfonamides: kinetic and product studies of the OH radical and Cl atom initiated oxidation of n-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 40, 864-872. Martin, J.W., Mabury, S.A., Solomon, K.R., Muir, D.C.G., 2003. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 22, 196-204. McLachlan, M.S., Holmstrm, K.E., Reth, M., Berger, U., 2007. Riverine discharge of perfluorinated carboxylates from the European continent. Environ. Sci. Technol. 41, 7260-7265. Oakes, K.D., Sibley, P.K., Solomon, K.R., Mabury, S.A., Van Der Kraak, G.J., 2004. Impact of perfluorooctanoic acid on fathead minnow (Pimephales promelas) fatty acyl-CoA oxidase activity, circulating steroids, and reproduction in outdoor microcosms. Environ. Toxicol. Chem. 23, 1912-1919. 184 L. Ahrens et al. / Chemosphere 76 (2009) 179-184 Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 40, 32-44. Rhoads, K.R., Janssen, E.M.-L., Luthy, R.G., Criddle, C.S., 2008. Aerobic biotransformation and fate of n-ethyl perfluorooctane sulfonamidoethanol (nEtFOSE) in activated sludge. Environ. Sci. Technol. 42, 2873-2878. Schultz, M.M., Barofsky, D.F., Field, J.A., 2004. Quantitative determination of fluorotelomer sulfonates in groundwater by LC MS/MS. Environ. Sci. Technol. 38, 1828-1835. Schultz, M.M., Higgins, C.P., Huset, C.A., Luthy, R.G., Barofsky, D.F., Field, J.A., 2006. Fluorochemical mass flows in a municipal wastewater treatment facility. Environ. Sci. Technol. 40, 7350-7357. Skutlarek, D., Exner, M., Frber, H., 2006. Perfluorinated surfactants in surface and drinking waters. Environ. Sci. Pollut. Res. 13, 299-307. So, M.K., Taniyasu, S., Yamashita, N., Giesy, J.P., Zheng, J., Fang, Z., Im, S.H., Lam, P.K.S., 2004. Perfluorinated compounds in coastal waters of Hong Kong, South China, and Korea. Environ. Sci. Technol. 38, 4056-4063. Taniyasu, S., Kannan, K., Horii, Y., Hanari, N., Yamashita, N., 2003. A survey of perfluorooctane sulfonate and related perfluorinated organic compounds in water, fish, birds, and humans from Japan. Environ. Sci. Technol. 37, 2634-2639. Taniyasu, S., Kannan, K., So, M.K., Gulkowska, A., Sinclair, E., Okazawa, T., Yamashita, N., 2005. Analysis of fluorotelomer alcohols, fluorotelomer acids, and short- and long-chain perfluorinated acids in water and biota. J. Chromatogr. A 1093, 89- 97. Theobald, N., Gerwinski, W., Caliebe, C., Haarich, M., 2007. Development and validation of a method for the determination of polyfluorinated organic substances in sea water, sediments and biota. Occurrence of these compounds in the North and Baltic Seas. Report No. UBA-FB: 001049 (in German). van Leeuwen, S.P.J., Swart, C.P., de Boer, J., 2009. Significant improvements in the analysis of perfluorinated compounds in water and fish: results from an interlaboratory method evaluation study. J. Chromatogr. A 1216, 401-409. Yamashita, N., Kannan, K., Taniyasu, S., Horii, Y., Okazawa, T., Petrick, G., Gamo, T., 2004. Analysis of perfluorinated acids at parts-per-quadrillion levels in seawater using liquid chromatography-tandem mass spectrometry. Environ. Sci. Technol. 38, 5522-5528. Yamashita, N., Kannan, K., Taniyasu, S., Horii, Y., Petrick, G., Gamo, T., 2005. A global survey of perfluorinated acids in oceans. Mar. Pollut. Bull. 51, 658-668. Environ. Sci. Technol. 2009, 43, 6969-6975 Partitioning Behavior of Per- and Polyfluoroalkyl Compounds between Pore Water and Sediment in Two Sediment Cores from Tokyo Bay, Japan LUTZ AHRENS,*'** NOBUYOSHI YAMASHITA,* , u LEO W. Y. YEUNG, II SACHI TANIYASU, YUICHI HORII, PAUL K. S. LAM," AND RALF EBINGHAUS t Institute for Coastal Research, GKSS Research Centre Geesthacht, DE-21502 Geesthacht, Germany, Institute for Ecology and Environmental Chemistry, Leuphana University of Luneburg, DE-21335 Luneburg, Germany, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan, and Department of Biology and Chemistry, City University of Hong Kong, HKSAR, China Received December 2, 2008. Revised manuscript received July 6, 2009. Accepted August 10, 2009. The partitioning behavior of per- and polyfluoroalkyl compounds (PFCs) between pore water and sediment in two sediment cores collected from Tokyo Bay, Japan, was investigated. In addition, the fluxes and temporal trends in one dated sediment core were studied. Short-chain perfluoroallwl carboxylic acids (PFCAs) (C 7) were found exclusively in pore water, while long-chain PFCAs (C 11) were found only in sediment. The perfluoroalkyl sulfonates (PFSAs), n-ethylperfluoro-1octanesulfonamidoacetic acid (N-EtFOSAA), and perfluorooctane sulfonamide (PFOSA) seemed to bind more strongly to sediment than PFCAs. The enrichment of PFCs on sediment increased with increasing organic matter and decreasing pH. The perfluorocarbon chain length and functional group were identified as the dominating parameters that had an influence on the partitioning behavior of the PFCs in sediment. The maximum I PFC contamination in sediment was observed in 2001-2002 to be a flux of 197 pg cm-2 yr-1. Statistically significant increased concentrations in Tokyo Bay were found for perfluorooctanesulfonate (PFOS) (1956-2008), perfluorononanoic acid (PFNA) (1990-2008), and perfluoroundecanoic acid (PFUnDA) (1990-2008). Concentrations of PFOSA and N-EtFOSAA increased between 1985 and 2001, but after 2001, the concentration decreased significantly, which corresponded with the phase out of perfluorooctyl sulfonyl fluoride-based compounds by the 3M Company in 2000. * Corresponding author hone: L.A (N.Y.); fax: (L.A.), (N.Y.); e-mail: @gkss.de. (L.A.), nob.yamashita@ aist.go.jp. (N.Y.). * GKSS Research Centre Geesthacht. * Leuphana University of Luneburg. National Institute of Advanced Industrial Science and Tech- nology (AIST). " City University of Hong Kong. 10.1021/es901213s CCC: $40.75 Published on Web 08/19/2009 2009 American Chemical Society Introduction In recent years, per- and polyfluoroallcyl compounds (PFCs) such as perfluorooctanoic acid (PFOA) and perfluorooctanesulfonate (PFOS), have received increasing public attention because of their persistence, bioaccumulative potential (1), and possible adverse effects on living organisms (2, 3). Over the past 50 years, PFCs have been widely used as processing additives during fluoropolymer production and as surfactants in consumer applications, including surface coatings for carpets, furniture, and paper products (4). The production of C8--PFC homologous has been restricted in many countries, e.g., the 3M Company, the major producer of perfluorooctyl sulfonyl fluoride (POSF), voluntarily phased out its production in 2000; however, they are still used by other manufacturers (5). Through the production and use of these products, PFCs can be released into the environment, and they have been found ubiquitous in water (6), sediment (7), wildlife (8), and humans (9), with the highest concentrations found in top marine predators (10). Sediment is an important sink and reservoir of persistent organic pollutants and has a large impact on their distribution, transport, and fate in the aquatic environment. Sediment--water distribution depends on such solution parameters as pH (11, 12) and on the organic carbon fraction (foc) (13). Higgins et al. found that the compound-specific adsorption of PFCs depended on their lipophilic and hydrophilic characteristics, and perfluorooctanesulfonamidoacetic acids (FOSAAs) were found to adsorb more strongly to sediment than PFOS (11). Volatile, neutral precursors can undergo long-range atmospheric transportation (14), whereas ionic PFCs can be transported by ocean currents (6). Sediment possibly acts as a sink for long-chain PFCs (11), but data on the partitioning of PFCs between sediment and pore water are scarce. In the study reported here, two sediment cores were collected from Tokyo Bay, which is situated near one of the most populous areas in the world, for a determination of the PFCs in pore water and sediment. We investigated the vertical concentration profile and distribution of the PFCs in both. Furthermore, the influence of organic carbon and pH on sediment concentrations was examined. This study provides information on the current and historical contamination status of PFCs in Tokyo Bay and reconstructs the pollution history of these contaminants in the waters. To the best of our knowledge, this is the first field study to examine the partitioning behavior of PFCs between pore water and sediment. Experimental Section Sampling Campaign. Two sediment cores were collected from Tokyo Bay using an acrylic tube (120 cm long and 12 cm i.d.) in May 2008 (cores A and B, Figure 1). These cores were sliced at 3 cm intervals for the first 9 cm for core A (6 cm for core B) and then at 2 cm intervals for up to 79 and 70 cm, respectively, using a clean stainless steel slicer and then stored in polypropylene (PP) tubes. Only the inner part of the sediment core was analyzed for PFCs to avoid contaminations from sampling. All of the samples were transported in an ice-cooled box to the lab and stored in a refrigerator at 4 C. The sampling conditions, including total organic carbon (TOC), total nitrogen (TN), pH, oxygen reaction potential (ORP), moisture and dry density, are shown in Tables 51 and S2 of the Supporting Information. Sediment Dating. The sedimentation rate was estimated from the excess 210Pb (dpm) in each layer and the cumulative VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 6969 Downloaded by GKSS FZ GEESTHACHT GMBH on October 8, 2009 | http://pubs.acs.org Publication Date (Web): August 19, 2009 | doi: 10.1021/es901213s FIGURE 1. Map showing sampling locations A (353460 N/ 1395501 E) and B (352918 N/1395424 E) in Tokyo Bay, Japan. weight (g cm-2) in core A. The average sedimentation rates of the dry matter were calculated to be 0.76 g cm-2 yr-1 for the 1958-2008 time period, which is approximately 1.5 cm yr-1 for the core sections. Separation of Pore Water and Sediment. Pore water was extracted from wet sediment within 72 h by centrifugation at 10000 rpm for 10 min at a constant temperature of 10 C (AvantiTM J-25 Centrifuge, Beckman, U.S.A.). The sediment was weighed before and after centrifugation to calculate the efficiency of the pore water extraction and then stored in a PP bottle (for details, see Table S3 of the Supporting Information). The pore water was then filtered through 0.45 m nylon syringe filters (Iwaki, Fukushima, Japan) into a PP bottle. All of the bottles were stored in a refrigerator at 4 C until analysis. Extraction and Analysis. The standards and reagents used have been previously reported (15), and a full list of the PFCs analyzed is shown in Table S4 of the Supporting Information. The 20 to 96 mL pore water samples were extracted by solidphase extraction with Oasis WAX cartridges (Waters, 150 mg, 6 cm3, 30 m) as described by Taniyasu et al. (16). Briefly, after preconditioning with 4 mL ammonium hydroxide in methanol, 4 mL methanol, and then 4 mL Millipore water, the cartridges were loaded with 20-96 mL samples at approximately 1 drop sec-1. Before loading, the pore water samples were spiked with 1 ng absolute of an internal standard (IS) mix (i.e., [13C4]-PFBA, [13C4]-PFOA, [13C5]-PFNA, [13C2]-PFDA, [13C2]-PFUnDA, [13C2]-PFDoA, [13C4]-PFOS, and 100 L of a 10 ng mL-1 solution; Table S4 of the Supporting Information). The cartridges were then washed with 4 mL of 25 mM ammonium acetate buffer (pH 4) in Millipore water and dried by centrifugation at 3000 rpm for 2 min. The elution was then divided into two fractions. The first fraction was carried out with 4 mL methanol and the second with 4 mL 0.1% ammonium hydroxide in methanol. Both fractions were reduced to 0.5 mL under a nitrogen stream and analyzed separately. The half-dry sediment was extracted using the method described by Powley et al. (17) with a few modifications. Briefly, a 5 g sediment sample was weighed into a PP tube, and 2 mL of 100 mM sodium hydroxide in 20% Millipore water and 80% methanol was added and then soaked for 30 min. The extraction was carried out with 20 mL methanol, and 1 ng absolute IS mix was spiked. The sample was then shaken in a wrist-action shaker at 250 rpm for 30 min. After shaking, the tube was centrifuged at 3000 rpm for 15 min, and the supernatant was decanted into another PP tube. The extraction was repeated with 1 mL of 100 mM sodium hydroxide in 20% Millipore water and 80% methanol, soaked for 30 min, and then with 10 mL of methanol, shaken at 250 rpm for 30 min, and centrifuged at 3000 rpm for 15 min. Both extracts were combined and acidulated with 0.1 mL of 4 M hydrochloric acid. This extract was centrifuged again at 3000 rpm for 5 min, and an aliquot of one-eighth (4.15 mL) of the 6970 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009 supernatant was used for the cleanup with Supelclean ENVICarb cartridges (100 mg, 1 mL, 100-400 mesh, Supelco, U.S.A.). The conditioning of the cartridges was carried out three times with 1 mL of methanol. Afterward, the sample extract and then three times 1 mL of methanol were added to the cartridge and directly collected in another vial. Finally, the extract was reduced to 1 mL under a nitrogen stream. All of the extracts were analyzed for PFCs in their ionic form using high-performance liquid chromatography coupled with tandem mass spectrometry (HPLC-MS/MS). An HP 1100 HPLC-system (Agilent Technologies, Palo Alto, CA) was equipped with a Betasil C18 column (2.1 mm i.d. 50 mm in length, 5 m; Thermo Hypersil-Keystone, Bellefonte, PA), with XDB-C8 (12.5 mm 2.1 mm, 5 m; Agilent Technologies, Foster City, CA) as a guard column and a RSpak JJ-50 2D column (2.0 mm i.d. 150 mm in length, 5 m; Shodex, Showa Denko K.K., Kawasaki, Japan). A triple-quadrupole mass spectrometer, supplied by Micromass (Quattro Ultima Pt, Beverly, MA), was used an electrospray ionization (ESI) interface in negative ionization mode. The flow rate was set to 300 L min-1, and 10 L of the sample was injected. Details of the extraction, instrumental conditions, and quantification for PFC analysis are shown in the Supporting Information and have been described elsewhere (15). Results and Discussion Blank Contamination and Method Quantification Limits and Recoveries. All fluorinated materials that could come into contact with the sample during sampling, sample preparation, or instrumental analysis were removed to avoid contaminations (for details, see ref 18). All procedural blanks, which were extracted in the same manner as the samples, using 100 mL of Millipore water for pore water and 1 mL of methanol for sediment analysis, were below the method quantification limits (MQLs). The MQLs were determined at a signal-to-noise ratio (S/N) of 10 and ranged between 0.1 and 0.5 pg cm-3 for pore water (50 mL of the pore water sample used) and between 0.37 and 1.86 pg cm-3 for sediment (5 g of the sediment sample used), depending on the sample size and relative response. The recoveries for water analysis ranged between 81% (N-EtFOSAA) and 128% (PFHxA), with a mean standard deviation (SD) of 3.4% (n ) 3), and for sediment between 78% (N-EtFOSAA) and 112% (PFHxA), with a mean SD of 5.3% (n ) 3) (Table S4 of the Supporting Information). The analysis of the sediment samples without cleanup using the Supelclean ENVI-Carb cartridges resulted in the degradation of the recoveries (Figure S1 of the Supporting Information). This could have been caused by the matrix effects, which led to signal enhancement for PFBS, PFHxS, and PFOS and to signal suppression for all other PFCs. Vertical Profiles. The PFC concentrations were corrected for dry density by multiplying the concentrations measured in pore water and sediment by their observed density. The concentration of the remaining pore water in the sediment was subtracted from the sediment concentration. In the following, the concentrations in pore water and sediment (solid phase) are expressed in pg cm-3 to enable a direct comparison of their concentration levels. Perfluoroalkyl sulfonate (PFSA), PFOSA/N-EtFOSAA, and perfluoroalkyl carboxylic acid (PFCA) concentrations were observed in pore water and sediment at different depths in two sediment cores (Figure 2 and Tables S5-S8 of the Supporting Information). In this study, 11 PFCs were found in pore water with concentration levels of <0.1-0.54 pg cm-3 for PFHxS, <0.1-3.49 pg cm-3 for PFOS, <0.1-0.84 pg cm-3 for TH-PFOS, <0.1-6.26 pg cm-3 for N-EtFOSAA, <0.5-26.1 pg cm-3 for PFBA, <0.1-3.15 pg cm-3 for PFPeA, 1.17-4.74 pg cm-3 for PFHxA, <0.1-0.45 pg cm-3 for PFHpA, 0.44-11.0 pg cm-3 for PFOA, <0.5-6.29 pg cm-3 for PFNA, and <0.1-0.47 pg cm-3 Downloaded by GKSS FZ GEESTHACHT GMBH on October 8, 2009 | http://pubs.acs.org Publication Date (Web): August 19, 2009 | doi: 10.1021/es901213s FIGURE 2. Vertical profile of PFSAs, PFOSA/N-EtFOSAA, and PFCAs in pore water and sediment in pg cm-3 from two sediment cores (A and B) collected from Tokyo Bay, Japan. for PFDA. To the best of the authors' knowledge, this is the first report of quantifiable concentrations of PFCs in pore water. The maximum surface water concentrations of PFHxS, PFOS, PFOA, and PFNA observed in Tokyo Bay were approximately 1 order of magnitude higher than those in pore water (6). In contrast to those in pore water, the PFCs in the sediment solid phase showed a different vertical profile. No TH-PFOS or short-chain PFCAs (C e 7) were detected in the sediment, although PFDS, PFOSA, and long-chain PFCAs (C g 11) were. These results correspond with the findings that with each CF2 moiety the distribution coefficient increase by 0.5-0.6 log units (11) and 0.87 log units (19), respectively. Ten PFCs were quantified in the sediment solid phase with a concentration range of <0.37-26.3 pg cm-3 for PFHxS, 2.63-28.7 pg cm-3 for PFOS, <1.86-6.07 pg cm-3 for PFDS, <0.37-16.0 pg cm-3 for PFOSA, <0.37-60.5 pg cm-3 for N-EtFOSAA, <0.37-1.07 pg cm-3 for PFOA, <0.37-2.51 pg cm-3 for PFNA, <0.37-1.61 pg cm-3 for PFDA, <0.37-12.6 pg cm-3 for PFUnDA, <0.37-3.34 pg cm-3 for PFDoDA, and <0.37-1.27 pg cm-3 for PFTeDA. These concentrations were 1-2 orders of magnitude lower than those measured in airdried sediment samples from the San Francisco Bay area by Higgins et al. (7). The concentrations of PFOA, PFDoDA, PFHxS, PFOS, and PFOSA found in air-dried sediment samples in the Ariake Sea and Tenjin and Katsura River in Japan were also 1-3 orders of magnitude higher (20, 21) than those found in the sediment solid phase in this study. In contrast to these studies, in only one sample PFOA was detected in the sediment solid phase from Tokyo Bay. The higher level of PFCs previously reported in sediments could be explained considering the effect of pore water. Both cores had a similar vertical profile, with a maximum PFC concentration at depths of 9-11 and 6-8 cm of 182 (sediment core A) and 132 pg cm-3 (sediment core B), respectively. Close to the surface, these PFC concentrations were slightly lower, and after reaching their maximum, with increasing sampling depth, they rapidly decreased by a factor of 6-9. In general, PFHxS, PFOS, PFOSA, N-EtFOSAA, and, in a lower proportion, PFUnDA accounted for the highest proportion in the sediment, while in the deeper layers only PFOS was found. PFOSA and N-EtFOSAA were found only in the upper layers (up to 35 cm in depth in core A to 46 cm in depth in core B), whereas the composition of PFHxS and PFOS increased with increasing depth. Interestingly, the proportion of PFCs increased in pore water with increasing sampling depth. In pore water, PFBA, PFOA, and PFNA dominated in the upper layers (<50 cm), but the composition changed in the deeper layers (g50 cm), in which in addition to PFOA, PFHxA became the predominant compound, whereas PFBA was only predominant in core B (Figure 3). Fluxes and Temporal Trends. Sediment concentration and flux, based on the dry sediment concentration, are shown for individual PFCs in sediment core A in Table 1. The fluxes were calculated using the dry sediment concentration multiplied by the density-corrected yearly sedimentation rate. The highest flux was observed for the PFCs, with 197 pg cm-2 yr-1 in 2001-2002. Before and after these time periods the PFCs decreased to 7.1 (1956-1958) and 87.7 pg cm-2 yr-1 (2006-2008). The greatest proportion of the flux was attributable to PFOS and EtFOSAA, whereas PFOS was largely responsible for the increasing flux and EtFOSAA for the reduction of the flux before and after 2001-2002. VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6971 Downloaded by GKSS FZ GEESTHACHT GMBH on October 8, 2009 | http://pubs.acs.org Publication Date (Web): August 19, 2009 | doi: 10.1021/es901213s FIGURE 3. Relative composition of PFOSA and N-EtFOSAA and individual PFSAs and PFCAs in pore water and sediment from two sediment cores (A and B) collected from Tokyo Bay, Japan. Note: The PFC concentrations, which were not detected, are shown as the sum of the half values of the method detection limit (MDL). The calculated flux in sediment core A was used for a rough estimation of the total flux of PFCs in Tokyo Bay sediments from 1956 to 2008, assuming that the sediment surface amount was 1000 km2. On the basis of this assumption, the total flux for the 52 year period was estimated to be 28.5 kg total for PFCs in Tokyo Bay sediments: 13.6 kg total for PFSAs, 11.6 kg total for PFOSA/N-EtFOSAA, and 3.2 kg total for PFCAs. Direct sources of exposure from manufacture and use were estimated as the major source for POSF (22) and PFCAs (5) in the environment. Public wastewater from around 35 million people and industrial wastewater (e.g., from textiles or paper fabrics, etc.) are be considered for the contamination of the Tokyo Bay with PFCs (4). It is also possible that aqueous film-forming foams (AFFFs) reach Tokyo Bay after their use at, for example, airports (23). But a large uncertainty exists about the contribution of the indirect sources (PFC precursors and/or impurities) on the total PFC contamination in the Tokyo Bay (22). Temporal trends were examined using ANOVA tests (SPSS 16.0 for Windows, 2007) to determine statistically significant differences (significance level R ) 0.05) of the concentration level and composition pattern between 1956 and 2008 for each analyte. Data were natural-logarithm 6972 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009 transformed prior to statistical analysis to meet assumptions of normality and homogeneity of variances. No significant trend was observed for PFHxS, PFDA, and PFDoDA (p > 0.073) (Figure S2 of the Supporting Information), whereas concentrations of PFOS (p < 0.0001), PFNA (p < 0.0001), and PFUnDA (p < 0.001) increased from 1956 to 2008, 1990 to 2008, and 1990 to 2008, respectively (Figure 4). Doubling times were calculated with t1/2 ) ln(2)/m, where m represents the slope of the natural logarithm transformed sediment concentration versus time. The doubling times for PFOS, PFNA, and PFUnDA were 16.1, 4.0, and 5.1 years, respectively. Similar doubling times have been found for PFNA and PFUnDA in polar bears (3.6-5.6 and 4.1-6.1 years, respectively) (24). Conversely, the PFOS doubling times reported here were longer than those reported for guillemot eggs from the Baltic Sea (7-10 years) (25), arctic ringed seals (3.0-7.1 years) (26), and polar bears (9.8-13.1 years) (24). In this study, however, the increasing trend of PFOS slowed between 2001 and 2008 and possibly reached a steady state level currently. The concentrations of N-EtFOSAA and PFOSA increased from 1985 to 2001 with doubling times of 4.5 and 6.3 years (p < 0.05), respectively, but after 2001, the concentration decreased TABLE 1. Concentrations (pg g-1 dry weight) and Estimated Fluxes (pg cm-2 yr-1) of Individual PFCs in Sediment Core A from Tokyo Bay, Japana depth time interval PFHxS PFOS PFDS PFOSA N-EtFOSAA PFOA PFNA PFDA PFUnDA PFDoDA PFTeDA (cm) represented conc. flux conc. flux conc. flux conc. flux conc. flux conc. flux conc. flux conc. flux conc. flux conc. flux conc. flux 0-3 2006-2008 56 13 96 23 <5 <0.2 3-6 2004-2006 46 13 128 37 <5 <0.3 6-9 2002-2004 18 6.4 97 34 23 8.2 9-11 2001-2002 20 7.5 117 43 25 9.2 13-15 1998-1999 43 17 95 37 <5 <0.4 17-19 1995-1997 18 8.9 39 19 <5 <0.5 21-23 1993-1994 29 16 24 13 <5 <0.5 25-27 1990-1991 42 19 25 11 <5 <0.5 29-31 1987-1989 20 9.6 18 9.0 <5 <0.5 33-35 1985-1986 18 9.7 9.9 5.3 <5 <0.5 37-39 1982-1983 <1 <0.6 60 34 <5 <0.6 41-43 1979-1981 <1 <0.6 15 9.2 <5 <0.6 45-47 1978-1979 <1 <0.6 70 42 <5 <0.6 49-51 1974-1975 <1 <0.7 18 12 <5 <0.7 53-55 1971-1973 <1 <0.7 13 8.2 <5 <0.7 57-59 1969-1970 <1 <0.7 12 7.8 <5 <0.7 61-63 1966-1967 <1 <0.8 43 33 <5 <0.8 65-67 1964-1965 <1 <0.8 18 13 <5 <0.8 69-71 1962-1963 <1 <0.8 11 8.2 <5 <0.8 73-75 1959-1960 <1 <0.8 13 10 <5 <0.8 77-79 1956-1958 <1 <0.7 10 7.1 <5 <0.7 a Flux estimates are based on the dry sediment method quantification limit (MQL). 38 9.0 63 15 6.8 1.6 14 3.4 6.4 1.6 64 15 48 14 101 29 <1 48 17 156 55 <1 51 19 247 92 <1 62 24 185 73 <1 17 8.4 73 35 <1 14 7.4 66 36 <1 16 7.3 84 38 <1 8.6 4.2 37 18 <1 5.0 2.6 14 7.6 <1 <1 <0.6 <1 <0.6 <1 <1 <0.6 <1 <0.6 <1 <1 <0.6 <1 <0.6 <1 <1 <0.7 <1 <0.7 <1 <1 <0.7 <1 <0.7 <1 <1 <0.7 <1 <0.7 <1 <1 <0.8 <1 <0.8 <1 <1 <0.8 <1 <0.8 <1 <1 <0.8 <1 <0.8 <1 <1 <0.8 <1 <0.8 <1 <1 <0.7 <1 <0.7 <1 <0.3 9.3 2.7 7.5 2.2 66 19 <0.4 4.6 1.6 4.9 1.7 40 14 <0.4 3.4 1.3 6.6 2.4 51 19 <0.4 3.7 1.4 5.4 2.1 42 17 <0.5 <1 <0.5 <1 <0.5 9.6 4.6 <0.5 1.1 0.6 <1 <0.5 13 7.2 <0.5 0.7 0.3 <1 <0.5 8.9 4.1 <0.5 <1 <0.5 <1 <0.5 <1 <0.5 <0.5 <1 <0.5 <1 <0.5 <1 <0.5 <0.6 <1 <0.6 <1 <0.6 <1 <0.6 <0.6 <1 <0.6 <1 <0.6 <1 <0.6 <0.6 <1 <0.6 <1 <0.6 <1 <0.6 <0.7 <1 <0.7 <1 <0.7 <1 <0.7 <0.7 <1 <0.7 <1 <0.7 <1 <0.7 <0.7 <1 <0.7 <1 <0.7 <1 <0.7 <0.8 <1 <0.8 <1 <0.8 <1 <0.8 <0.8 <1 <0.8 <1 <0.8 <1 <0.8 <0.8 <1 <0.8 <1 <0.8 <1 <0.8 <0.8 <1 <0.8 <1 <0.8 <1 <0.8 <0.7 <1 <0.7 <1 <0.7 <1 <0.7 concentration and sedimentation rate at each core section; <x 18 4.2 18 5.1 6.7 2.4 11 4.0 5.9 2.3 <1 <0.5 <1 <0.5 <1 <0.5 <1 <0.5 <1 <0.5 <1 <0.6 <1 <0.6 <1 <0.6 <1 <0.7 <1 <0.7 <1 <0.7 <1 <0.8 <1 <0.8 <1 <0.8 <1 <0.8 <1 <0.7 below the 3.7 0.9 6.6 1.9 <1 <0.4 <1 <0.4 <1 <0.4 <1 <0.5 <1 <0.5 <1 <0.5 <1 <0.5 <1 <0.5 <1 <0.6 <1 <0.6 <1 <0.6 <1 <0.7 <1 <0.7 <1 <0.7 <1 <0.8 <1 <0.8 <1 <0.8 <1 <0.8 <1 <0.7 respective Downloaded by GKSS FZ GEESTHACHT GMBH on October 8, 2009 | http://pubs.acs.org Publication Date (Web): August 19, 2009 | doi: 10.1021/es901213s significantly with half-lives of 2.8 and 13.5 years (p < 0.05), respectively (Figure 4 and Figure S2 of the Supporting Information, respectively). This may reflect increased production and emissions since the 1950s in the Tokyo Bay area and the phase out of POSF-based compounds by 3M in 2000, of which at least four companies in Japan were affected (22). Nearly the same temporal trend was observed for PFOSA in melon-headed whales (Peponocephala electra) stranded along the Japanese coast, with increasing concentration between 1982 and 2001/2002, followed by a decreasing concentration until 2006 (27). It should be noted that benthic organisms can change the contamination profiles of sediment cores because of bioturbation or bioaccumulation (28). It is also possible that N-EtFOSAA and PFOSA can further biodegrade to PFOS, which would increase the flux of this compound (29). No groundwater flow existed at both sampling stations to a 80 cm depth; however, horizontal pore water migration for shortchain PFCAs (C e 8) is possible. In addition, pore water may have caused the migration of PFCs that had collected in deeper layers. Thus, the sediment concentrations of PFOS and N-EtFOSAA in the deeper layers (>30 cm) may be artificially inflated by their subsequent deposition. This could explain why a flux of PFOS is seen until 1956. However, no evidence of downward migration in pore water was observed for PFHxS, PFDS, PFOSA, or long-chain PFCAs (C g 8). This suggests that any migration depends on the partitioning behavior of these compounds between pore water and sediment. Influence of Organic Carbon and pH on Sediment Concentrations. Physical and geochemical characteristics of the vertical sediment profile were variable. While the TOC decreased from 0.7% to 1.7%, the pH increased from 7.3 to 7.7 with the depth. The influence of organic carbon and pH on the sorption of PFCs has been investigated (11, 13). In this study, a positive correlation was found between organic matter and concentrations of PFOS, PFOSA, and PFUnDA (p < 0.0001, Figure 5) and, with lower significance, N-EtFOSAA, PFNA, and PFDoDA (p < 0.01; Figure S4 of the Supporting Information). This shows that organic matter may have an influence on the vertical distribution of PFCs in the studied sediment cores. Previous studies observed a significant correlation between the partition coefficients and organic content for PFOS, N-EtFOSAA, PFDA (11), and also for 8:2 fluorotelomer alcohol (13). The concentrations of PFOS, PFNA, and PFDoDA increased with decreasing pH (p < 0.05; Figure S5 of the Supporting Information). This corresponds with the results of Higgins et al., showing increasing sorption of PFCs with decreasing pH of approximately 0.37 log units per unit pH (11). Other factors, which were not investigated in this study such as the concentration of calcium cations, could have also an influence on the sorption capacity of sediment for PFCs (11). Partition Coefficients. The partitioning behavior of PFCs depends on their physicochemical characteristics as well as on sediment-specific parameters (11). N-EtFOSAA showed a good correlation between their concentrations in pore water and sediment (r2 ) 0.58, p < 0.0001; Figure S3 of the Supporting Information). No relationship was observed for the other PFCs or they were exclusively detected in pore water or sediment. Therefore, the partition coefficient was only calculated for N-EtFOSAA. The organic carbon normalized adsorption coefficient (KOC) was calculated with the organic carbon fraction (fOC) as described in the following. KOC)(cs/caq) 100/fOC (1) where cs is the adsorbed PFC on sediment in pg g-1, caq is the mass concentration of PFC in the aqueous phase pg cm-3, and fOC is the organic carbon fraction in percentage (30). The log KOC of N-EtFOSAA is with 2.99 ( 0.21 cm3 g-1 in the same range as determined experimental by Higgins et al. (log KOC 3.23 ( 0.18 L/kgOC, n ) 5 (11)). Further investigations on the adsorption and solution processes of PFCs are necessary. Environmental Significance. Our study showed the partition behavior of PFCs between pore water and sediment in two sediment cores. Pore water concentrations of PFCs were determined for the first time. PFSAs, N-EtFOSAA, and PFOSA seemed to bind more strongly to sediment than PFCAs, whereas only short-chain PFCAs (C e 7) could be found exclusively in pore water. These results corroborate the laboratory findings of Higgins et al. (11), and thus show that the perfluorocarbon chain length and functional group VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6973 Downloaded by GKSS FZ GEESTHACHT GMBH on October 8, 2009 | http://pubs.acs.org Publication Date (Web): August 19, 2009 | doi: 10.1021/es901213s FIGURE 4. Temporal trends of PFOS, N-EtFOSAA, PFNA, and PFUnDA in sediment core A from Tokyo Bay, Japan. FIGURE 5. Dependence of PFOS, PFOSA, and PFUnDA concentrations in sediment on sediment fraction organic carbon (fOC). influence partitioning behavior of PFCs in the real environment. In addition, an increasing sorption was found with increasing organic matter and decreasing pH, which correspond with experimental data (11, 19). However, other factors like geochemical parameters (e.g., metal cations, etc.) (11) or benthic organisms (e.g., degradation, bioturbated mixing, etc.) could have an influence on the partitioning behavior of PFCs in sediment. The presence of long-chain PFCAs (C g 8), PFSAs, PFOSA, and N-EtFOSAA in sediment suggests that they are bioavailabe to benthic organisms (28) and that aquatic sediment possibly acts as a sink for these 6974 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009 compounds. Future work could link the pore water and sediment concentrations of PFCs to bioconcentration factors, uptake routes, and possible adverse effects. The temporal trend data presented here provide a basis for future trend studies of PFCs in sediment cores. Acknowledgments The authors acknowledged the German Federal Environmental Foundation and International Bureau from the Federal Ministry of Education and Research (project JPN 08/A02) for supporting the project. L.W.Y.Y. and P.K.S.L. acknowledge funding from the Hong Kong Research Grants Council (CityU 160408). We thank Dr. Guruge K.S., National Institute of Animal Health, Japan. for the help of sample preparation, and Dr. Wang Yuan, Ms. Mak Y.L., Ms. Lai F.Y, and Mr. Jin L. from the City University of Hong Kong, Dr. Li P. from AIST, and Tokyo Kyuei Co. for the help of sampling in Tokyo Bay. Supporting Information Available Additional information about the sampling campaign, chemicals, analysis, method recovery rates, overview of PFC concentrations, temporal trends, and dependence of PFC concentration in sediment on fraction organic carbon and pH value. This material is available free of charge via the Internet at http://pubs.acs.org. Literature Cited (1) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204. (2) Austin, M. E.; Kasturi, B. S.; Barber, M.; Kannan, K.; Mohan Kumar, P. S.; Mohan Kumar, S. M. J. Neuroendocrine effects of perfluorooctane sulfonate in rats. Environ. Health Perspect. 2003, 111, 1485-1489. (3) Oakes, K. D.; Sibley, P. K.; Solomon, K. R.; Mabury, S. A.; Van der Kraak, G. J. Impact of perfluorooctanoic acid on fathead minnow (Pimephales promelas) fatty acyl-CoA oxidase activity, circulating steroids, and reproduction in outdoor microcosms. Environ. Toxicol. Chem. 2004, 23, 1912-1919. (4) Kissa, E. Fluorinated surfactants and repellents. Marcel Dekker 2001, 97. (5) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44. (6) Yamashita, N.; Kannan, K.; Taniyasu, S.; Horii, Y.; Petrick, G.; Gamo, T. A global survey of perfluorinated acids in oceans. Mar. Pollut. Bull. 2005, 51, 658-668. (7) Higgins, C. P.; Field, J. A.; Criddle, C. S.; Luthy, R. G. Quantitative determination of perfluorochemicals in sediments and domestic sludge. Environ. Sci. Technol. 2005, 39, 3946-3956. (8) Giesy, J. P.; Kannan, K. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342. (9) Yeung, L. W. Y.; So, M. K.; Jiang, G.; Taniyasu, S.; Yamashita, N.; Song, M.; Wu, Y.; Li, J.; Giesy, J. P.; Guruge, K. S.; Lam, P. K. S. Perfluorooctanesulfonate and related fluorochemicals in human blood samples from China. Environ. Sci. Technol. 2006, 40, 715- 720. (10) Kannan, K.; Koistinen, J.; Beckmen, K.; Evans, T.; Gorzelany, J. F.; Hansen, K. J.; Jones, P. D.; Helle, E.; Nyman, M.; Giesy, J. P. Accumulation of perfluorooctane sulfonate in marine mammals. Environ. Sci. Technol. 2001, 35, 1593-1598. (11) Higgins, C. P.; Luthy, R. G. Sorption of perfluorinated surfactants on sediment. Environ. Sci. Technol. 2006, 40, 7251-7256. (12) Westall, J. C.; Chen, H.; Zhang, W.; Brownawell, B. J. Sorption of linear alkylbenzenesulfonates on sediment materials. Environ. Sci. Technol. 1999, 33, 3110-3118. (13) Liu, J.; Lee, L. S. Solubility and sorption by soils of 8:2 fluorotelomer alcohol in water and cosolvent systems. Environ. Sci. Technol. 2005, 39, 7535-7540. (14) Jahnke, A.; Berger, U.; Ebinghaus, R.; Temme, C. Latitudinal gradient of airborne polyfluorinated alkyl substances in the marine atmosphere between Germany and South Africa (53 N-33 S. Environ. Sci. Technol. 2007, 41, 3055-3061. (15) Taniyasu, S.; Kannan, K.; Yeung, L. W. Y.; Kwok, K. Y.; Lam, P. K. S.; Yamashita, N. Analysis of trifluoroacetic acid and other short-chain perfluorinated acids (C2-C4) in precipitation by liquid chromatography-tandem mass spectrometry: Comparison to patterns of long-chain perfluorinated acids (C5-C18). Anal. Chim. Acta 2008, 619, 221-230. (16) Taniyasu, S.; Kannan, K.; So, M. K.; Gulkowska, A.; Sinclair, E.; Okazawa, T.; Yamashita, N. Analysis of fluorotelomer alchols, fluorotelomer acids, and short- and long-chain perfluorinated acids in water and biota. J. Chromatogr. A 2005, 1093, 89-97. (17) Powley, C. R.; George, S. W.; Ryan, T. W.; Buck, R. C. Matrix effect-free analytical methods for determination of perfluorinated carboxylic acids in environmental matrixes. Anal. Chem. 2005, 77, 6353-6358. (18) Yamashita, N.; Kannan, K.; Taniyasu, S.; Horii, Y.; Okazawa, T.; Petrick, G.; Gamo, T. Analysis of perfluorinated acids at partsper-quadrillion levels in seawater using liquid chromatographytandem mass spectrometry. Environ. Sci. Technol. 2004, 38, 5522-5528. (19) Liu, J.; Lee, L. S. Effect of fluorotelomer alcohol chain length on aqueous solubility and sorption by soils. Environ. Sci. Technol. 2007, 41, 5357-5362. (20) Nakata, H.; Kannan, K.; Nasu, T.; Cho, H.-S.; Sinclair, E.; Takemura, A. Perfluorinated contaminants in sediments and aquatic organisms collected from shallow water and tidal flat areas of thr Ariake Sea, Japan: Environmental fate of perfluorooctane sulfonate in aquatic ecosystems. Environ. Sci. Technol. 2006, 40, 4916-4921. (21) Senthilkumar, K.; Ohi, E.; Sajwan, K.; Takasuga, T.; Kannan, K. Perfluorinated compounds in river water, river sediment, market fish, an wildlife samples from Japan. Bull. Environ. Contam. Toxicol. 2007, 79, 427-431. (22) Paul, A. G.; Jones, K. C.; Sweetman, A. J. A first global production, emission, and environmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol. 2009, 43, 386-392. (23) Moody, C. A.; Field, J. A. Perfluorinated surfactants and the environmental implications of their use in fire-fighting foams. Environ. Sci. Technol. 2000, 34, 3864-3870. (24) Smithwick, M.; Norstrom, R. J.; Mabury, S.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G. Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American Arctic, 1972-2002. Environ. Sci. Technol. 2006, 40, 1139-1143. (25) Holmstrom, K.; Jarnberg, U.; Bignert, A. Temporal trends of PFOS and PFOA in guillemot eggs from the Baltic Sea, 19682003. Environ. Sci. Technol. 2005, 39, 80-84. (26) Butt, C.; Muir, D. C.; Stirling, I.; Kwan, M.; Mabury, S. A. Rapid response of Arctic ringed seals to changes in perfluoroalkyl production. Environ. Sci. Technol. 2007, 41, 42-49. (27) Hart, K.; Kannan, K.; Isobe, T.; Takashani, S.; Yamada, T.; Miyazaki, A.; Tanabe, S. Time trends and transplacental transfer of perfluorinated compounds in melon-headed whales strandes along the Japanese coast in 1982, 2001/2002, and 2006. Environ. Sci. Technol. 2008, 403, 215-221. (28) Higgins, C. P.; McLeod, P. B.; MacManus-Spencer, L. A.; Luthy, R. G. Bioaccumulation of perfluorochemicals in sediments by the aquatic oligochaete Lumbriculus variegatus. Environ. Sci. Technol. 2007, 41, 4600-4606. (29) Rhoads, K. R.; Janssen, E. M.-L.; Luthy, R. G.; Criddle, C. S. Aerobic biotransformation and fate of n-ethyl perfluorooctane sulfonamidoethanol (n-EtFOSE) in activated sludge. Environ. Sci. Technol. 2008, 42, 2873-2878. (30) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry. Wiley-Interscience, Hoboken, NJ, 2003. ES901213S Downloaded by GKSS FZ GEESTHACHT GMBH on October 8, 2009 | http://pubs.acs.org Publication Date (Web): August 19, 2009 | doi: 10.1021/es901213s VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6975 F' SEVTFR Environmental Pollution 158 (2010) 1467-1471 Contents lists available at ScienceDirect Environmental Pollution journal homepage: www.elsevier.com/locate/envpol ENVIRONMENTAL POLLUTION Polyfluoroalkyl compounds in landfill leachates Jan Busch a' b, Lutz Ahrens a'*, Renate Sturm', Ralf Ebinghaus a Department for Environmental Chemistry, GKSS Research Centre Geesthacht, Geesthacht, Germany b Leuphana University of Luneburg, Luneburg, Germany The first comprehensive survey of polyfluoroalkyl compounds (PFCs) in landfill leachates. ARTICLE INFO Article history: Received 14 August 2009 Received in revised form 13 December 2009 Accepted 16 December 2009 Keywords: Landfill Leachate PFCs PFOS PFOA ABSTRACT Polyfluoroalkyl compounds (PFCs) are widely used in industry and consumer products. These products could end up finally in landfills where their leachates are a potential source for PFCs into the aqueous environment. In this study, samples of untreated and treated leachate from 22 landfill sites in Germany were analysed for 43 PFCs. EPFC concentrations ranged from 31 to 12,819 ng/L in untreated leachate and 4-8060 ng/L in treated leachate. The dominating compounds in untreated leachate were perfluorobutanoic acid (PFBA) (mean contribution 27%) and perfiuorobutane sulfonate (PFBS) (24%). The discharge of PFCs into the aqueous environment depended on the cleaning treatment systems. Membrane treatments (reverse osmosis and nanofiltrations) and activated carbon released lower concentrations of PFCs into the environment than cleaning systems using wet air oxidation or only biological treatment. The mass flows of EPFCs into the aqueous environment ranged between 0.08 and 956 mg/day. 2009 Elsevier Ltd. All rights reserved. 1. Introduction Polyfluoroalkyl compounds (PFCs) are persistent, bioaccumulative and potential toxic pollutants (Brooke et al., 2004). Due to their unique chemical properties they are used for instance as surfactants and coatings. The allocation of European production of PFCs is estimated to be 49% for textile leather and carpet goods, 33% for paper and packaging, 15% for industrial surfactants additives and coatings, and 3% for aqueous fire fighting foams (AFFF) (Schultz et al., 2003). Perfluorinated acids, such as perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA), are soluble and can be transported in water, while neutral precursor compounds can transported via the atmospheric to remote regions (Ellis et al., 2004). Finally neutral precursor compounds can be degraded to perfluoroalkyl carboxylic acids (PFCAs) and perfluoroalkyl sulfonates (PFSA5) in biota (Tomy et al., 2004), sludge (Boulanger et al., 2005b; Rhoads et al., 2008), wastewater treatment plants (WWTPs) (Becker et al., 2008) and in the atmosphere (D'Eon et al., 2006; Martin et al., 2006). PFCs are widespread distributed and can be found in industrial as well as in remote regions (Giesy and Kannan, 2001). Emissions to the environment can occur during production processes and application of perfluorooctylsulfonyl fluoride (POSF) * Corresponding author. Tel.: +49 4152872353; fax: +49 4152872332. E-mail address: ec.gc.ca (L. Ahrens). 0269-7491/$ -- see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2009.12.031 based products and from consumer products, which contain POSFbased products (Prevedouros et al., 2006). Total POSF production between 1972 and 2002 is estimated to be 122,500 t, with a maximum of 4650 tin 2000 (Paul et al., 2009). Total production of PFCAs is estimated to be produced 4400-8000 t in total, from which 3200-7300 t is estimated to be released into the environment by direct emissions (i.e., fluoropolymer manufacturing and processing, use of AFFF and from consumer and industrial products) and indirect emissions (i.e., residual impurities or fluorotelomerbased AFFF) (Prevedouros et al., 2006). Direct sources for PFCs into the aqueous environment are street runoff with EPFC concentrations up to 380 ng/L (Murakami et al., 2009), surface runoff with EPFC concentrations of 1.1-81.8 ng/L (Kim and Kannan, 2007) and municipal and industrial WWTPs with EPFC concentrations of 30.5-266.3 ng/L (Ahrens et al., 2009a). The discharge of EPFCs from WWTPs into the aqueous environment was estimated to be in g per day range (Bossi et al., 2008; Clara et al., 2008), which results in .a 80 g/day/person (Huset et al., 2008). Indirect emissions are estimated to result for 85% from losses of consumer products like carpets, clothing, paper and packaging during use and disposal (Paul et al., 2009). The usage of PFOS and derivates is banned by the European Union for various applications (ERC, 2006) and voted to be regulated under Annex B of the Stockholm Convention (Stockholm Commission, 2009). A voluntary phase out of industrial production of ammonium perfluorooctanoate (APFO) in USA started in 2002. 1468 J. Busch et al. / Environmental Pollution 158 (2010) 1467-1471 Table 1 Overview on tested landfill sites with estimated annual amount of leachate in m3/year, treatment process, status, landfill class, and type of sample. Landfill site 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Amount of leachate (m3/year) unknown 46,000 26,280 7,000c 8000 175,200 28,700 9000 8000 200 20,000 55,000 10,000 20,000 8000 16,500 19,500 131,400 27,500 26,500 29,000 41,100 Treatment processa external BIO BIO, RO BIO, RO, WAO BIO (nitrification) WAO AC AC AC NF external NF RO flotation, BIO, AC RO BIO, AC flotation, BIO, AC AC BIO, RO flotation, BIO, AC external RO Status inactive, closed 2005 active active inactive, closed 1999 active active inactive, inactive, inactive, inactive, inactive, active active active active active active active inactive, active active closed closed closed closed closed closed 1979 1979 2004 1999 1986 1999 active Landfill classb class I class II class II Mixture of effluents from 3 sites (class II-III) class II class I, partial class II old landfill, partial unknown old landfill, partial unknown class II class II class II class II class II class II class II class II class II class II class II class II class II, only filter cake from a WWTP class II waste waste Sample of leachate (sample number) untreated (1) treated (1) treated (1) treated (1) treated (1) treated (1) and untreated (1) treated (3) treated (3) and untreated (1) treated (4) treated (1) untreated (1) treated (3) treated (1) treated (4) and untreated (1) treated (1) treated (4) treated (1) and untreated (1) treated (1) treated (1) treated (1) untreatedd (1) treated (1) a AC activated carbon, BIO biological treatment, NF nanofiltration, RO reverse osmosis, WAO wet air oxidation. b Classification according to German law (AbfAblV, 2001). Organic contamination of waste in `class II' is higher than in `class I' waste. `Class III' landfills are for high contaminated waste. `Old landfills' are not classified because they were installed before 1993 and are not under the regulation of the AbfAblV. c Sum of 3 landfill sites. d The leachate is untreated but fulfils the criteria for discharging into municipal WWTP. Nevertheless, used consumer products are in landfills subject to chemical reactions and degradation processes (Lema et al., 1988), which might lead to attaining of PFCs in leachates, which is suggested by different authors (Hekster et al., 2002; Schultz et al., 2003; Boulanger et al., 2005a; Prevedouros et al., 2006; Paul et al., 2009). Little information exists about PFCs in landfill leachates, whereas, to the knowledge of the authors, only four studies about PFCs in landfill leachates are available, which cover in total 15 landfill sites (Bossi et al., 2008; Woldegiorgis et al., 2008; Kallenborn et al., 2004; 3M, 2001). These studies reported data of only a few PFCs and give no detailed information on tested landfill sites, hence conclusions cannot be drawn on whether landfills are a source of PFCs into the aqueous environment. In addition, no detailed methods (e.g., using of mass-labelled internal standards (IS), quality control) was described which might be critical because of the complex matrix of landfill leachates. The aim of this study was investigate the occurrence of PFCs in untreated and treated leachate from 22 landfill sites in Germany. In Fig. 1. PFC concentrations in treated leachate of 19 landfill sites. Shown are the six most abundant compounds in ng/L. Note: Other PFCs are aggregated as `other'. The abbreviations correspond to the treatment system of the landfill site: AC activated carbon, BIO biological treatment, FLO flotation, NF nanofiltration, RO reverse osmosis, WAO wet air oxidation. J. Busch et at / Environmental Pollution 158 (2010) 1467-1471 1469 order to avoid matrix effects a sample volume of only 5 mL and 50 mL was used for untreated and treated leachate, respectively. The application of the final method to analyse PFCs in landfill leachate is described. Special attention is focused on the influence of different treatment systems (i.e., activated carbon, biological treatment, nanofiltration, reverse osmosis and wet air oxidation) on the concentration level and contribution of individual PFCs. Furthermore, a comparison of PFC concentrations in untreated and treated leachate was accomplished. Finally, short time variations of PFC concentrations in treated leachate and their emission into the aqueous environment was investigated. Table 2 Mean and median concentrations and range of EPFC in leachate after different cleaning treatment systems, in ng/L. Treatmenta Mean Median Concentration range R0 AC NF WAO BIO 5 42 8 723 2 940 2 3302 2 6041 24 22 3302 6041 15.2-129 9.26-4079 1992-4610 4023-8059 a AC = activated carbon, BIO = biological treatment, NF = nanofiltration, RO = reverse osmosis, WAO = wet air oxidation. 2. Materials and methods 2.1. Standards and reagents In this study, 43 PFCs, including 16 PFCAs, 7 PFSA5, 3 perfluoroalkyl sulfinates (PFSiAs), 3 perfluoroalkyl phosphoric acids (PFPA5), 4 perfluoroalkyl sulfonamides (FASAs), 3 perfluoroalkyl sulfonamidoethanols (FASEs), 3 fluorotelomer carboxylic acids (FTCAs), 3 fluorotelomer unsaturated carboxylic acids (FTUCAs) and 6:2 fluorotelomer sulfonate (6:2 FTS) were included in the method plus 20 mass-labelled IS. A detailed list of the analytes, acronyms, formula, purity, supplier and precursor and product ions for the mass spectrometer is given in Table S1 in the Supplementary material or can be found elsewhere (Ahrens et al., 2009c). Methanol (Suprasolv), acetonitrile (LiChrosolv), ammonium hydroxide (25%, Suprapur) and ammonium acetate were purchased from Merck (Darmstadt, Germany). Pure water was produced by a MilliQ Plus 185 system (Millipore, Germany). 2.2. Sampling sites Untreated and treated landfill samples were collected in 250 mL polypropylene (PP) bottles at 22 landfills sites in Germany. Untreated leachate samples were taken from 6 landfill sites before any treatment and one untreated leachate sample was taken from a landfill site which deposits only filter cake from WWTPs. In addition,19 samples were taken from treated landfill leachate after the treatment process but before the discharge into downstream waters or municipal wastewater. From four landfill sites (i.e., landfill sites 6, 8, 14, 17) samples before and after the treatment process were taken and at six landfill sites three samples (i.e., landfill sites 7, 8, 12,) or four samples (i.e., landfill sites 9, 14,16) were taken with a sampling rate of 2-4 weeks. All samples were stored at 4 C and were extracted within four weeks after sampling. A description of the landfill sites including estimated amount of leachate, cleaning treatment process of leachate, classification according to waste content and sample type is presented in Table 1. 2.3. Sample extraction and analysis The extraction of treated and untreated leachate samples was performed using solid-phase extraction (SPE) as described elsewhere (Taniyasu et al., 2005; Ahrens et al., 2009b) with a few modifications. Briefly, an aliquot of 5 mL of untreated leachate and 50 mL of treated leachate, respectively, was used for analysis. The samples were filtered using glass fibre filters (GFF, GC/C, Whatman, 0 47 mm, 1.2 m), and adjusted to a pH-value between 7 and 8 using ammonium hydroxide, since the mean recovery rates were best at this pH-value (results not shown). The aliquots were spiked with 100 L of 20 mass-labelled IS (see Table S2 in the Supplementary material, c = 100 ng/mL) in order to correct losses and matrix effects. Oasis WAX cartridges (Waters, 150 mg, 6 cc, 30 m) were used for SPE, which were preconditioned using 5 mL 0.1% ammonium hydroxide in methanol, 5 mL methanol and 5 mL Millipore water. After loading (flow = 1 drop/second) the cartridges were washed with 5 mL 0.1% ammonium hydroxide in Millipore water and dried by centrifugation (2 min at 3000 rpm). Elution was done by 14 mL of acetonitrile and 5 mL 0.1% ammonium hydroxide in methanol. Both eluates were combined and reduced under a gentle nitrogen stream to 150 L and spiked with 50 L of an injection standard (lnjS) (i.e., d5-EtFOSAA, c = 0.4 g/mL). For instrumental analysis, high performance liquid chromatography and tandem mass spectroscopy (HPLC-MS/MS) was used (for details see Ahrens et al. (2009c)). For some compounds (e.g., PFOS) more than one peak could be identified. These peaks occur due to the detection of isomers (Loganathan et al., 2007) and are not quantified due to the lack of standards. For four compounds, PFPS, PFNS, PFPeDA and PFHpA no standards were available. These compounds were quantified by parameters of the corresponding shorter and longer-chain PFCs and are therefore only estimations. 2.4. Quality control Quality control and assurance included the using of 20 internal standards, recovery rates, method blanks, mass detection limits (MDL), mass quantification limits (MQL), control standards, reproducibility and the calculation of the matrix effect. Blanks were prepared using Millipore water at least every eight samples together with the samples. The concentrations of the samples were corrected by the blank concentration. Blanks were found in the samples processed using 50 mL Millipore water for nine substances (i.e., PFHxS, PFOS, PFOSi, PFBA, PFPA, PFOA, PFDA, PFUnDA and PFOSA) in pg per liter range while in the 5 mL Millipore blank samples higher blank values of PFOA and PFBS were observed in low ng per liter range, because of the low sample volume. However, due to the high PFC concentration level in the leachate samples (mean EPFC concentration = 6086 4638 ng/L) the concentration in the blank samples were negligible. Recovery rates of IS in untreated leachate samples ranged between 44% for 13C4-PFBA and 70% for 13C4PFOSi (mean 60 19%, n = 57). Recovery rates in treated leachate ranged between 36% for 13C4-PFBA and 65% for 13C2-10:2-FTUCA (mean 52 17%, n = 6) (for details see Table S3 in the Supplementary material). The MDL and MQL were calculated in the samples with a signal to noise ratio of three and ten, respectively. The MQL ranged between 0.05 ng/L for PFHxSi and 22.8 ng/L for PFDPA (for details see Table S4 in the Supplementary material). 3. Results and discussion In 6 untreated leachates 25 of 43 PFCs were detected. EPFC concentrations ranged from 30.5 ng/L (landfill site 9) to 12,922 ng/L (landfill site 14) (mean concentration = 6086 4638 ng/L). The dominating compounds were PFBA (mean contribution 27%), PFBS (24%), PFHxA (15%), PFOA (12%), PFPA (6.0%), PFHpA (4.0%), 6:2 FTS (3.7%), PFOS (2.7%), and PFHxS (2.3%). In treated leachate almost all target compounds were detected (39 of 43 PFCs) with a similar mean contribution as for the untreated leachates. EPFC concentrations ranged from 3.97 ng/L (landfill 9) to 8060 ng/L (landfill 5) (mean concentration = 1336 2139 ng/L) (for details see Table S5 in the Supplementary material). A comparison of individual PFC concentration in treated leachates related to the treatment system is shown in Fig. 1. Effluent samples from reverse osmosis showed the lowest EPFC concentration (15.2-129 ng/L), followed by activated carbon (9.26-4079 ng/L) and nanofiltration (621-1257 ng/L). Wet air oxidation (19924610 ng/L) and biological treatment (4023-8059 ng/L) showed the highest concentrations (see Table 2). An explanation for this observation is possibly the different efficiency of the treatment systems to remove PFCs from the leachate. Reverse osmosis and nanofiltration are membrane-based treatments working with semipermeable membranes, which separate the leachate into a clean permeate and a contaminated Table 3 Comparison of EPFC concentrations in treated and untreated leachate from four landfill sites in ng/L, including the calculation of concentrations change as remaining percentage after treatment to the corresponding untreated leachate. Landfill site 6 8 14 17 Untreated leachate 1889 31 12,819 8370 Treated leachate 1993 9 4079 20 Concentration change +5.5% -70.5% -68.2% -99.8% Treatmenta WA0 AC AC AC Status active inactive inactive inactive a WAO = wet air oxidation. AC = activated carbon. 1470 J. Busch et al. / Environmental Pollution 158 (2010) 1467-1471 Table 4 Ranges and mean concentration of individual PFCs in landfill leachate in this study compared with literature data in ng/L.a (3M, 2001) (n 3) (Woldegiorgis et al., 2008) (n 4) (Kallenborn et al., 2004) (n 6) PFBS PFHxS PFOS PFDS PFBA PFHxA PFHpA PFOA PFNA PFDA PFUnA PFOSA NA NA <25-52.7 (17.7) NA NA NA NA ND-48.1 (16.9) NA NA NA NA <0.5-110 12-1800 32-1500 <1-0.28 <12-30 <7-310 <20-260 38-1000 <18-100 <20-220 <59 <2-7 (37.3) (518) (555) (0.07) (7.5) (77.5) (197.5) (537) (43.5) (82.5) (2.75) 5.64-112 (51.5) 12.4-143 (77.0) 32.8-187 (82.5) NA NA 26.4-697 (228) NA 92.4-516 (293) 4.7-61.5 (34.8) NA NA NQ-3.28 (1.17) a NA not analysed. ND not detected. <x below the respective method quantification limit (MQL). (Bossi et al., 2008) (n 2) NA <0.2-3.1 (0.8) <1.5-3.8 (1.1) NA NA NA NA <2-5.8 (2.9) <0.8 <1.6 <2.2 <0.3 This study (n 20) <0.39-1356 (220) <0.24-178 (22.2) 0.01-235 (30.9) ND <3.36-2968 (458) <0.37-2509 (234) <0.12-280 (48.1) <0.40-926 (145) <3.63-80.1 (7.29) <0.21-55.1 (5.98) <0.11-2.98 (0.36) <0.15-14.0 (2.77) residue. The membrane pores of the reverse osmosis are smaller than in nanofiltration and might be therefore more effective for PFC removal than nanofiltration. Activated carbon treatment is based on adsorption of contaminants to carbon and described as a possible cleaning technique for reduction of PFC concentrations in aqueous media, due to the high adsorption capacity (OchoaHerrera and Sierra-Alvarez, 2008). However, the adsorption rates of hacigtihvaPtePdFcCacrobnocnernatnragteidonbsetwweereeno6b8se.2r%veadnadt9la9n.8d%fi(lsl eseiteTasb1l4ea3n)dan1d8 after the treatment with activated carbon (i.e., 1452 ng/L and 4079 ng/L, respectively). This could be explained by the fact that the sorption capacity is limited depending on the mass flow and concentrations of the contaminants (Yu et al., 2009). Wet air oxidation uses Ozone to create OH-radicals in order to break-down contaminants by oxidation. A comparison of untreated and wet air oxidation treated leachate from landfill site 6 showed a higher concentration of PFCs in treated leachate than in untreated leachate (see Table 3). This could be a result of insufficient removal of PFCs by oxidation (Lema et al., 1988; Silva et al., 2004) or the degradation of precursor compounds to perfluoroalkyl acids b(Riohlooagdicsaelt atlr.,e2a0te0d8). lTehaechhaigtehes(tmPaFxCimcounmcentPraPtiFoCns cwoenrceefnoturantdioinn level 8059 ng/L). This could be explained by the persistence of perfluoroalkyl acids against biodegration and the possible biodegradation of precursor compounds to perfluoroalkyl acids, which was reported for biological treatment in WWTPs (Schultz et al., 2006; Sinclair and Kannan, 2006). Overall, lacking reduction of PFC concentrations and a possible degradation of precursor compounds suggests wet air oxidation and biological treatments are not proper suitable treatments for reducing PFC concentrations in landfill leachates. No significant changes in the relative contribution of PFCs were observed in the treated leachate in comparison to the untreated leachate. Discharge of PFCs into the aqueous environment was calculated by the mean SPFC concentration multiply by the amount of leachate. This is only a rough calculation assuming constant discharge of PFCs during the time. Short time variations in the concentration level were investigated at the landfills sites 7, 8, 9, 12, 14, 16 over a period of several months (see Table S5 in the Supplementary material). The results indicP ate only low variations in low contaminated leachate (e.g., mean PFC concentration of 4.0 2.4 ng/L (n 4P, landfill site 9)) and highly contaminated l(enach4a)t)e. T(hee.ge.,stmimeaatned dPisFcChacrognesceonftPratPiFoCnoof fall4e0x7a9mine6d9l0anndgfi/lLl sides ranged between 0.08 and 956 mg/day with a mean discharge of 135 mg/day. An interpolation of the mass flow of the existing landfills in Germany (w1700 landfill sites in 2009) using the mean PFC discharge of this study results in 240 g/day SPFCs of all landfill sites in Germany. Much higher mass flows were estimated for German WWTPs with 30-2200 mg/day for PFOA and 20-2800 mg/ day for PFOS (Becker et al., 2008). Furthermore, Schultz et al. (2006) reported mass flows form WWTPs in the United States of 2000 mg/ day for 8 PFCs. The low discharge of PFCs from landfills is a result of the low amount of leachate, however, the emissions of some landfill sites could be problematic because of the high concentration level. Interestingly, a correlation was found between the PFC concentration level in untreated leachate and the year of closing of the landfill sites (R 0.67). The SPFC concentration in active landfills ranged between 1889 ng/L and 12,819 ng/L, whereas the landfill site 8 (closed 1979) had only a SPFC concentration of 31 ng/L. This could be a result of the increasing deposition of waste over the time which contains PFCs. The reported leachate concentrations of other studies ranged between a few and hundreds of ng per liter (3M, 2001; Woldegiorgis et al., 2008; Kallenborn et al., 2004; Bossi et al., 2008) (Table 4). A comparison of the concentration level with this study should be considered carefully because of the small sample size and missing detailed methods description of the other studies. However, the concentration levels in this study are in the same range as of the other studies except for PFBS and PFBA, which were w10-100 times higher in this study (<0.39-1356 ng/L for PFBS and <3.36-2968 ng/L for PFBA) in comparison to the reported concentrations from Sweden (<0.5-112 ng/L for PFBS (Woldegiorgis et al., 2008; Kallenborn et al., 2004) and <12-30 ng/L for PFBA (Woldegiorgis et al., 2008)). Since data on landfills tested in other studies (i.e., deposited waste or cleaning treatment) are not available, presumably differences can be explained by different usage of PFCs and different regulation and treatment processes on landfills in different countries. 4. Conclusions Leachate of landfill sites were highly contaminated with PFCs with maximum SPFC concentrations of 12,819 ng/L in untreated leachate and 8060 ng/L in treated leachate. The shorter-chain PFCs (C 6) were more abundant than longer-chain compounds, whereas the mean contribution of PFBS and PFBA represent approximately the half of the SPFC proportion. However, the greatest influence on the concentration level is due to the treatment system. Lowest PFC concentration were observed after membrane treatments (reverse osmosis and nanofiltration) and activated carbon, while the highest PFC concentrations were found after biological treatment and wet air oxidation. Overall, the discharge of the leachate from landfills of this study represents only J. Busch et al. / Environmental Pollution 158 (2010) 1467-1471 1471 w1% of mass flows from WWTPs. This result indicates the leachate as a minor source for PFCs into the aqueous environment, however, extremely high concentration levels (thousand of ng per liter SPFCs) at some landfill sites could be problematic for the local environment. In addition, leachate from landfill sites, where PFCs manufactures deposite their waste, are probably high contaminated with PFCs and could be a risk for the environment. The data presented here highlighted the occurrence of PFCs in untreated and treated leachate from landfills. Future work should focus on the correlation of the individual PFC concentration depending on the deposited waste type and long-term monitoring programs. Acknowledgements We kindly acknowledge the landfill operators for cooperation and the provision of samples. Appendix. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.envpol.2009.12.031. References 3M, 2001. Environmental Monitoring - Multi-City-Study Water, Sluge, Sediment, POTW Effluent and Landfill Leachate Samples. 3M Environmental Laboratory. AbfAblV, 2001. German law to deposite munucipial waste. Abfallablagerungsverordnung - AbfAblV. Ahrens, L., Felizeter, S., Sturm, R., Xie, Z., Ebinghaus, R., 2009a. Polyfluorinated compounds in wastewater treatment plant effluents and surface waters along the River Elbe, Germany. Mar. Pollut. Bull. 58, 1326-1333. Ahrens, L., Plassmann, M., Xie, Z., Ebinghaus, R., 2009b. Determination of polyfluoroalkyl compounds in water and suspended matter in the River Elbe and North Sea, Germany. Front. Environ. Sci. Eng. China 3, 152-170. Ahrens, L., Siebert, U., Ebinghaus, R., 2009c. Total body burden and tissue distribution of polyfluorinated compounds in harbor seals (Phoca vitulina) from the German Bight. Mar. Pollut. Bull. 58, 520-525. Becker, A.M., Gerstmann, S., Frank, H., 2008. Perfluorooctane surfactants in waste waters, the major source of river pollution. Chemosphere 72, 115-121. Bossi, R., Strand, J., Sortkjaer, O., Larson, M.M., 2008. Perfluorinated compounds in Danish wastewater treatment plants and aquatic environments. Environ. Int. 34, 443-450. Boulanger, B., Peck, A.M., Schnoor, J.L., Hornbuckle, K.C., 2005a. Mass budget of perfluorooctane surfactants in Lake Ontario. Environ. Sci. Technol. 39, 74-79. Boulanger, B., Vargo, J., Schnoor, J.L., Hornbuckle, K., 2005b. Evaluation of perfluorooctane surfactants in a wastewater treatment system and in a commercial surface protection product. Environ. Sci. Technol. 39, 5524-5530. Brooke, D., Footitt, A., Nwaogu, T.A., 2004. Environmental risk evaluation report: perfluorooctanesulphonate (PFOS). Environ. Agency, 1-46. Clara, M., Scheffknecht, C., Scharf, S., Weiss, S., Gans, O., 2008. Emissions of perfluorinated substances (PFAS) from point sources - identification of relevant branches. Water Sci. Technol. 58, 59-66. D'Eon, J.C., Hurley, M.D., Wallington, T.J., Mabury, S.A., 2006. Atmospheric chemistry of n-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 40, 1862-1868. Ellis, D.A., Martin, J.W., Mabury, S.A., De Silva, A.O., Hurley, M.D., Sulbaek Anderson, M.D., Wallington, T.J., 2004. Degradation of fluorotelomer alcohols: a likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 38, 3316-3321. ERC, 2006. Directive 2006/122/EC of the European Parliament and of the Council on December 12th, 2006. Giesy, J.P., Kannan, K., 2001. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 35, 1339-1342. Hekster, F.M., De Voogt, P., Pijnenburg, A., Laane, R.W.P.M., 2002. Perfluoroalkylated substances - aquatic environment assessment. In: RIKZ (Ed.). RIKZ. Huset, C.A., Chiaia, A.C., Barfosky, D.F., Jonkers, N., Kohler, H.-P.E., Ort, C., Giger, W., Field, J.A., 2008. Occurrence and mass flows of fluorochemicals in the Glatt Valley Watershed, Switzerland. Environ. Sci. Technol. 42, 6369-6377. Kallenborn, R., Berger, U., Ja rnberg, U., 2004. Perfluorinated alkylated substances (PFAS) in the Nordic environment. Tema Nord, 1-107. Kim, S.-K., Kannan, K., 2007. Perfluorinated acids in air, rain, snow, surface runoff; and lakes: relative importance of pathways to contamination of urban lakes. Environ. Sci. Technol. 41, 8328-8334. Lema, J.M., Mendez, R., Blazquez, R., 1988. Characteristics of landfill leachates and alternatives for their treatment: a review leachate. Water Air Soil Pollut. 40, 223-250. Loganathan, B.G., Sajwan, K.S., Sinclair, E., Kumar, K.S., Kannan, K., 2007. Perfluoroalkyl sulfonates and perfluorocarboxylates in two wastewater treatment facilities in Kenntucky and Georgia. Water Res. 41, 4611-4620. Martin, J.W., Ellis, D.A., Mabury, S.A., Hurley, M.D., Wallington, T.J., 2006. Atmospheric chemistry of perfluoroalkanesulfonamides: kinetic and product studies of the OH radical and Cl atom initiated oxidation of n-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 40, 864-872. Murakami, M., Shinohara, H., Takada, H., 2009. Evaluation of wastewater and street runoff as sources of perfluorinated surfactants (PFCs). Chemosphere 74, 487-493. Ochoa-Herrera, V., Sierra-Alvarez, R., 2008. Removal of perfluorinated surfactants by sorption onto granular activated carbon, zeolite and sludge. Chemosphere 72, 1588-1593. Paul, A.G., Jones, K.C., Sweetman, A.J., 2009. A first global production, emission, and environmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol. 43, 386-392. Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 40, 32-44. Rhoads, K.R., Janssen, E.M.-L., Luthy, R.G., Criddle, C.S., 2008. Aerobic biotransformation and fate of n-ethyl perfluorooctane sulfonamidoethanol (n-EtFOSE) in activated sludge. Environ. Sci. Technol. 42, 2673-2878. Schultz, M.M., Barofsky, D.F., Field, J.A., 2003. Fluorinated alkyl surfactants. Environ. Eng. Sci. 20, 487-501. Schultz, M.M., Higgins, C.P., Huset, C.A., Luthy, R.G., Barofsky, D.F., Field, J.A., 2006. Fluorochemical mass flows in a municipal wastewater treatment facility. Environ. Sci. Technol. 40, 7350-7357. Silva, A.C., Dezotti, M., Sant'Anna Jr., G.L., 2004. Treatment and detoxification of a sanitary landfill leachate. Chemosphere 22, 207-214. Sinclair, E., Kannan, K., 2006. Mass loading and fate of perfluoroalkyl surfactants in wastewater treatment plants. Environ. Sci. Technol. 40, 1408-1414. Stockholm Commission, 2009. Press Release (May, 9th 2009): Governments Unite to Step-up Reduction on Global DDT Reliance and Add Nine New Chemicals under International Treaty. Stockholm Comission, Geneva. Taniyasu, S., Kannan, K., So, M.K., Gulkowska, A., Sinclair, E., Okazawa, T., Yamashita, N., 2005. Analysis of fluorotelomer alchols, fluorotelomer acids, and short- and long-chain perfluorinated acids in water and biota. J. Chromatogr. A 1093, 89-97. Tomy, G.T., Tittlemier, S.A., Palace, V.P., Budakowski, W.R., Braekevelt, E., Brinkworth, L., Friesen, K., 2004. Biotransformation of n-ethyl perfluorooctanesulfonamide by rainbow trout (Oncorhynchus mykiss) liver microsomes. Environ. Sci. Technol. 38, 758-762. Woldegiorgis, A., Anderson, J., Remberger, M., 2008. Results from the National Swedish Screening Programme. IVL Swedish Environmental Research Institute, Stockholm. Yu, Q., Zhang, R., Deng, S., Huang, J., Yu, G., 2009. Sorption of perfluorooctane sulfonate and perfluorooctanoate on activated carbon and resin: kinetic and isoterm study. Water Res. 43, 1150-1158. Environ. Sci. Technol. 2010, 44, 5534-5539 Global Pilot Study of Legacy and Emerging Persistent Organic Pollutants using Sorbent-Impregnated Polyurethane Foam Disk Passive Air Samplers SUSIE GENUALDI, SUM CHI LEE, MAHIBA SHOEIB, ANYA GAWOR, LUTZ AHRENS, AND TOM HARNER* Environment Canada, Science and Technology Branch, 4905 Dufferin Street, Toronto, Ontario M3H 5T4 Received March 26, 2010. Revised manuscript received June 9, 2010. Accepted June 18, 2010. Sorbent-impregnated polyurethane foam (SIP) disk passive air samplers were deployed alongside polyurethane foam (PUF) disk samplers at 20 sites during the 2009 spring sampling period of the Global Atmospheric Passive Sampling (GAPS) Network. The SIP disk samplers consisted of PUF disks impregnated with finely ground XAD-4 resin. The addition of XAD-4 greatly improves the sorptive capacity of the PUF disk samplers for more volatile and polar chemicals, and allows for linear-phase sampling over several weeks for these compounds. The SIP and PUF disks were analyzed for polychlorinated biphenyls (PCBs), neutral polyfluoroalkyl compounds (PFCs), and ionic PFCs. Correlations between sampler-derived air concentrations for PCBs in the PUF and SIP disks samplers were significant (p < 0.05). The SIP disks effectively captured 4--50% more of the low molecular weight PCBs than the PUF disks samplers, and the PUF disks also had limitations for timeweighted passive sampling of neutral PFCs in air. Theoretical uptake curves for PUF disks showed rapid equilibration occurring in just hours for 8:2 FTOH and in a few days for MeFOSE, while theoretical curves for SIP disks showed superior sampling profiles for the neutral PFCs. PFCs were measured on SIP disks at all sites with 8:2 FTOH being the dominant compound detected and urban centers (n = 3) having the highest total neutral PFC concentrations ranging from 51.7 to 248 pg/m3. A positive correlation was found between the FTOHs and FOSAs/ FOSEs (p < 0.001, Pearson correlation) indicating similar contamination sources. The SIP disk appears to be a promising passive air sampler for measuring both emerging and legacy POPs on a global scale. They can also be used as a complement to the PUF disk sampler for capturing broader classes of compounds, or as a replacementfor PUF disks entirely, especially when longer than quarterly deployment periods are desired. Introduction The Global Atmospheric Passive Sampling (GAPS) study has been measuring persistent organic pollutants (POPs) worldwide on both spatial and temporal scales since 2005 (1-3). The GAPS Network currently has 55 sites, mainly remote * Correspondin author e-mail: fax: ec.gc.ca; phone: 5534 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 14, 2010 background locations with a few urban and agricultural sites. POPs are of global concern because of their environmental persistence, toxicity, and potential to bioaccumulate (4). They also have the ability to undergo long-range atmospheric transport and accumulate in remote locations such as the Arctic (4). The Stockholm Convention on POPs under the United Nations Environment Programme (UNEP) identified a group of POPs termed the "dirty dozen" to be removed from the global environment (5). These POPs include organochlorine pesticides (OCPs), industrial chemicals such as polychlorinated biphenyls (PCBs), and their byproducts, mainly dioxins and furans (5). Recently in 2009, nine additional POPs were added to the Stockholm convention, including OCPs, polychlorinated diphenyl ethers (PBDE5), perfluorooctane sulfonate (PFOS), and perfluorooctane sulfonyl fluoride (PFOS-F) (6). Polyfluoroallcyl compounds (PFCs) including PFOS and perfluorooctanoate (PFOA) have been measured in biota in remote regions ( 7-10). It is thought these ionic compounds are a result of more volatile precursor compounds (neutral PFCs) that are reaching these regions by either long-range atmospheric transport followed by degradation to ionic PFCs such as PFOS and PFOA or by direct transport through ocean currents (11-19). To monitor the new compounds in air, which is one of the two core media for monitoring POPs under the Global Monitoring Plan of the Stockholm Convention (20), a new type of passive air sampler that can measure PFOS and its volatile precursors (e.g., fluorotelomer alcohols (FTOHs), perfluorooctane sulfonamides (FOSAs), and sulfonamidoethanols (FOSEs)) is necessary. The sorbent-impregnated polyurethane foam disk (SIP) sampler has been shown to have a much higher sorptive capacity (-2 orders of magnitude for the FTOHs) than the polyurethane foam (PUF) disk sampler for the measurement of FTOHs and FOSAs and FOSEs (21). This sampler was effective at measuring concentrations of FTOHs, FOSAs, and FOSEs in indoor air in Canada, which typically have much higher concentrations than those observed in outdoor air (21). SIP disk samplers appear to be a promising complement or alternative to the PUF disk sampler currently being used in the GAPS Network. The objectives of this research were to compare the performance of the SIP disk to the PUF disk sampler under field conditions to effectively sample POPs, and to assess the ability of the SIP disk sampler to capture a wider range of target compounds including FTOHs, FOSAs, and FOSEs. Experimental Section Sample Collection. SIP and PUF disk samplers were concurrently deployed at 20 sites during the 2009 spring sampling period (April to June) of the Global Atmospheric Passive Sampling (GAPS) Network (Figure SI.1 and Table SI.1 in the Supporting Information). The sampling locations consisted of 12 background, 4 polar, 3 urban, and 1 agricultural site. Individual sites were chosen based on the interests of Canada's Chemical Management Plan (North America) and also on potential source regions. The climate classification of the sites chosen for the pilot study were temperate, cold, polar, and dry regions, but there were no sites representative of the tropical region. Previous studies of PUF disk samplers deployed in the tropical regions (1, 3, 22) have shown they are effective at sampling POPs under these conditions. However, further studies are necessary to verify the performance of the SIP disk sampler under tropical conditions. The PUF and SIP disks were individually housed inside 10.1021/es1009696 2010 American Chemical Society Published on Web 06/25/2010 prewashed, presolvent rinsed stainless steel chambers. Details on the sampling apparatus have been previously provided (22). Sample Preparation. PUF disks (14 cm diameter 1.35 cm thick; surface area 365 cm2, mass 4.40 g, volume 207 cm3, Tisch Environmental, Cleves, OH) were precleaned by first washing in water followed by Soxhlet extraction for 24 h with acetone followed by another 24 h with petroleum ether. The PUF disks were then dried in a desiccator for 24 h and placed in clean, solvent-rinsed glass jars. Details on the preparation of SIP disk samplers have been previously reported (21). Briefly, XAD-4 (Supelco, Bellefonte, PA) was cleaned using successive sonications with methanol, dichloromethane, and hexane and then finely ground using a ball mill to a particle size of 0.75 m. The ground XAD was further Soxhlet extracted for 30 h using methanol, dichloromethane, and hexane. SIP disks were prepared by consecutively dipping precleaned PUF disks in an XAD-4/hexane (6.4 g/L) slurry and were uniformly coated with an average XAD-4 mass of 435 ( 68 mg per disk (n ) 75) (21). In comparison to the XAD-2 resin based passive air sampler, which uses 20 g of XAD-2 resin (spheres) (23), only 0.5 g of finely ground XAD-4 is impregnated onto a PUF disk (in making a SIP disk) to increase its capacity to measure volatile compounds. Finely ground XAD has a higher sorptive capacity compared to the XAD spheres (of the same mass) and is also easier to clean/extract which helps to maintain lower blanks levels (21). Prior to deployment, PUF disks were spiked with 250-500 ng per disk of each of the following depuration compounds (DCs): 13C12 PCB 3, 13C12 PCB 9, 13C12 PCB 15, 13C12 PCB 32, PCB 30, PCB 107, PCB 198, and d6 -HCH. These compounds consist of labeled and unlabeled PCBs not typically found in air, and range in log KOA values from 6.57 to 10.86 (24). SIPs were not spiked with depuration compounds in this study and were assumed to have the same sampling rates as the PUF disk codeployed at each location, given their identical geometry and that uptake is air-side controlled (25). Details on the sample extraction and quantification can be found in the Supporting Information. QA/QC. Field blanks were collected at all 20 sites for both PUF and SIP disk samples. The blank values for PCBs were below the limit of quantification (LOQ) for all congeners except PCB 8. In all cases where blanks had values of PCB 8 above the LOQ, measurements of PCB 8 in the corresponding samples were all below the LOQ. Therefore no blank correction was necessary. The SIP disk blanks were below the LOQ for all compounds except 8:2 FTOH and 10:2 FTOH and the concentration of the blanks ranged from 2 to 30% of the concentrations measured in the samples. For each site, the blank value was used as the new LOQ for that site and only samples with values greater than the LOQ were reported. Estimated method detection limits (EDLs) were calculated using EPA method 8280A and ranged from 0.41 to 1.2 pg/m3 for PCBs, 0.43 to 1.1 pg/m3 for the neutral PFCs, and 0.0087 to 0.18 pg/m3 for the ionic PFCs. All SIP and PUF disk samples were recovery corrected for neutral and ionic PFCs using the surrogate standards; further details on the surrogates chosen for each native compound have been previously reported (21). Method recovery values for the neutral PFCs, ionic PFCs, and PCBs ranged 29-134%, 14-89%, and 73-123%, respectively. Due to the high percent recovery for PCBs, it was not necessary to recovery correct the PCB concentrations. Sample Volumes derived from Depuration Compounds. Site-specific sampling rates were calculated for each site based on the loss of depuration compounds (DCs) added to the PUF disks prior to deployment. Only DCs having retention percentages less than 60% were used in this calculation, which ensures the loss of the DCs was due to depletion and not analytical variability (3). The percentage of depuration compounds present on the PUF disk at the end of the deployment period was greater than 7% at every site. If a depuration compound is completely lost at the end of the deployment period, that compound is not included in the calculation. DCs were not added to SIP disks since most would not meet this retention percentage due to the increased sorptive capacity of the SIP disk sampler. The high molecular weight depuration compounds (PCB-107 and PCB-198) are expected to have 100% recovery due their high KOA values (10.73 and 11.87, respectively), as demonstrated in previous studies (1, 2, 22). To account for loss during the analytical method, depuration compounds are normalized to PCB-198. Previous studies indicate that the recovery of PCB-198 from the PUF disk samples prior to correction agrees well with the method recovery values (22). In this study, the method recovery of PCBs was 103 ( 19%, which is close to the average recovery of 83 ( 15% for PCB-198 before correction. For the PUF disks, each PCB congener was adjusted for the volume of air (Vair) sampled during the deployment time based on the KPSMAIR partition coefficient between the passive sampling medium (PSM) and air of each congener using eq 1 (22). { [ ] } Vair ) (KPSM-A) (VPSM) 1 - exp - kA 1 t KPSM-A Dfilm (1) Where KPSM-A is equal to KPSM-AIR multiplied by the density of the passive sampling medium (DPSM in g/m3), VPSM is the volume of the passive sampling medium (m3), kA is the airside mass transfer coefficient (m/day), Dfilm is the effective film thickness (m), and t is time (days). Site-specific kA values were derived from the loss of DCs at each sampling location. Sampling rates R (m3/day) were calculated by multiplying the kA values by the surface area of the PUF disk sampler and can be found in Table SI.1 in the Supporting Information. The same R values are applied to the SIP disks since they have the same geometry as the PUF disk sampler, which implies the uptake rate (kinetics) of contaminants into the sampler is the same. This surrogate approach is required because SIP disks have a much greater sorptive capacity (equilibrium phase) and retain DCs more strongly. Losses of DCs from the SIP disks over the 3-month deployment period are not sufficient for deriving R. Further details on these calculations and previous uptake study results involving both the SIP and PUF disk samplers for PCBs and neutral PFCs have been described elsewhere (1, 3, 21, 22, 26). For the SIP disk samplers, for all PCB congeners, site-specific sampling rates determined using the PUF disk samplers were applied, assuming mass transfer is air-side controlled. SIP disks were also not expected to reach equilibrium for PCBs based on their greater sorptive capacity (21). The average site-specific sampling rates during this study ranged from 6.34 ( 4.03 m3/day. For the ionic PFCs, an average sampling rate of 4 m3/day was used. This is the average gas-phase sampling rate observed under the GAPS Network for the same sampler geometry (3) and is consistent with results from an unpublished uptake study for PFOS and PFOA, conducted in our laboratory. Further calibration studies are necessary to accurately calculate the sampling rate for these compounds. The air concentrations given for ionic PFCs represent results from gas-phase sampling only. Because many of these compounds are found in the particle phase (27),which has a lower sampling rate compared to the gas phase (28), the derived air concentrations are likely to be underestimating the true air concentrations. Results and Discussion PCB Concentrations. To compare the performance of the SIP disk sampler to that of the PUF disk sampler to capture VOL. 44, NO. 14, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5535 FIGURE 1. Comparison between SIP and PUF disk samplers using linear regressions of PCBs measured as (A) pg/sampler and (B) air concentrations in pg/m3. The dashed lines represent a 1:1 relationship between PUF and SIP disk samplers. FIGURE 2. Polychlorinated biphenyl (PCB) concentrations (pg/m3) measured in air using (A) PUF disk and (B) SIP disk air samplers. Starred sampling locations (*) indicate urban sites. legacy POPs, the PCBs accumulated in the PUF and SIP disks were compared using linear regressions. Three separate regressions were performed between the two samplers for low (di-, tri-), medium (tetra-, penta-), and high (hepta-, octa-, nona-, deca-) molecular weight PCBs. In Figure 1A, PCBs are compared between the two samplers in units of pg/sampler, while in Figure 1B the PCBs have been converted to air concentrations (pg/m3) using site-specific sampling rates and an additional correction was made for the PUF disks for the volume of air sampled based on the KPSM-AIR partition coefficient for each congener. In Figure 1A and B, r2 is close to unity (0.94, 0.99, respectively) in the regressions for the high molecular weight PCB congeners compared to the r2 of the regressions for the low molecular weight congeners (0.84, 0.91, respectively). From previous calibration studies, it is expected that the high molecular weight PCBs will remain in the linear sampling phase for the entire deployment period in both the PUF and SIP disk samplers, while some of the low molecular weight PCBs have the potential to reach equilibrium in the PUF disks before the end of the deployment (1, 25). When the amount of PCBs accumulated in the PUF and SIP disks was converted from pg/sampler (Figure 1A) to pg/m3 (Figure 1B), the slope of each of the regressions became close to 1: di- and tri- (0.50 to 0.53), tetra- and penta- (0.78 to 0.85) and hepta-, octa-, nona-, deca- (0.91 to 1). This correction more accurately calculates the sampler-derived air concentrations for PCBs in the PUF disk samplers, and also confirms the ability of the SIP disk to effectively capture PCBs as well as the PUF disk air sampler. The concentrations of PCBs measured in air using PUF and SIP disk samplers on a spatial scale can be seen in Figure 2A and B and Tables SI.3 and SI.4 in the Supporting Information. The air concentrations (pg/m3) compare quite well between the two samplers for the higher molecular weight PCBs with a percent difference of 5.5-11% for the penta- through hepta- congers and 28-70% for the di- to tetra- congeners. The SIP disk samplers show concentrations of the di-, tri-, and tetra- PCBs measured on the west coast of North America (sampling sites 4, 13, 17) that are 4-50% 5536 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 14, 2010 FIGURE 3. (A) Neutral polyfluoroalkyl compounds (PFCs) and (B) ionic PFCs measured in air (pg/m3) using SIP disk air samplers. Starred sampling locations (*) indicate urban sites. FIGURE 4. Theoretical uptake profiles for MeFOSE and 8:2 FTOH in (A) PUF disk and (B) SIP disk samplers at the Toronto, Ontario sampling location. Stars indicate the volume of air sampled at the end of the deployment period (90 days). higher than those measured by the PUF disk samplers, due to the higher sorptive capacity of the SIP disks to capture chemicals with lower log KOA values. Using a two sample t-test, the mean concentrations of PCBs measured at all sites were compared between the SIP and PUF samplers for the following congener groups: di- and tri- chlorobiphenyls (p ) 0.18), tetra- and penta- chlorobiphenyls (p ) 0.56), and hexa- to nona- chlorobiphenyls (p ) 0.95). The two-sided p-values were all much greater than 0.05, indicating that there is no statistically significant difference between the means of each group. The SIP disks compare well to the PUF disks under field conditions for the measurement of legacy POPs, and appear to be a promising sampler for the measurement of the more volatile PCB congeners. PFC Concentrations. Concentrations of PFCs in air were measured globally using SIP disk air samplers (Figure 3, Table SI.5). The dominant neutral PFC measured at all sites was 8:2 FTOH. The three urban sites in Toronto, ON Canada, Paris, France, and Sydney, FL (Figure 3, starred 6, 11, 18) had some of the highest concentrations with the sums of all neutral PFCs ranging from 51.7 to 248 pg/m3. A positive correlation was found between the neutral PFC classes FTOHs and FOSAs/FOSEs (p < 0.001, Pearson Correlation, SPSS version 16) indicating common contamination sources. Seven ionic PFCs (C4, C6, C8 PFSAs, C8-C11 PFCAs) were detected, with PFOS as the dominant compound. The highest sum of ionic PFC concentrations was also observed at the urban site Paris, France with a total concentration of 245 pg/m3. The PFCAs (C8-C11) were detected only at the Paris site, with concentrations ranging from 0.296 to 4.49 pg/m3 (Table SI.5). Previous studies have shown that air masses associated with urban centers are sources of PFCs (29-31). A fourth background site in Groton, CT (Figure 3, 20), located in the Northeastern U.S., also had high concentrations of neutral PFCs with a sum of 107 pg/m3. Models created from emissions associated with the manufacture, use, and disposal of DuPont fluorotelomer-based products have also shown that the Northeastern United States has the highest ground level air concentrations of 8:2 FTOH (32). PFOS was found frequently in the SIP disk samplers, whereas C8-C11 PFCAs were only detected at the Paris site. The ionic PFCs could be originating from either fine particles collected with the SIP disk samplers (28) and/or degradation of neutral PFCs (FTOHs, FOSAs, FOSEs) to the ionic PFCs during the sampling period of 90 days (33). Ionic PFCs are typically found in the particulate phase of air samples (34). Levels of PFSAs in this study are similar to those reported in the particulate phase from Northern Europe, and levels of PFCAs at the Paris site are similar to those found in the particulate phase in Kjeller (Norway), Manchester (UK), and Mace Head (Ireland) (30). Lower PFOS levels of 0.1-2.5 pg/ m3 were found in the particulate phase from a latitudinal gradient in the marine atmosphere between Germany and South Africa in the Atlantic Ocean (34). More work is required to investigate the partitioning of neutral and ionic PFCs in the gaseous and particulate phases, and also their potential to degrade on the SIP disk sampler. Uptake of PFCs in PUF and SIP Disk Samplers. PFCs accumulated in PUF disk samplers were quite low and only detected at eight sampling locations: Groton, CT; Paris, VOL. 44, NO. 14, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5537 France; Whistler, BC, Canada; Barrow, AK; Toronto, ON, Canada; Storhofdi, Iceland; Malin Head, Ireland; and Sydney, FL (Table SI.6). PFOS was the only ionic PFC detected in the PUF disk samplers, while two neutral PFCs (MeFOSE and 8:2 FTOH) dominated and both were measured at the Toronto, ON site. To compare the effectiveness of the SIP disk sampler over the PUF disk sampler to collect PFCs, theoretical uptake profiles were created for MeFOSE and 8:2 FTOH for the Toronto, ON sampling location (Figure 4). These curves were generated using the site specific kA value and KPSM-AIR partition coefficients previously determined from an uptake study (21). During the deployment period, the effective volume of air sampled in the PUF disk sampler was 0.940 m3 for 8:2 FTOH and 43.8 m3 for MeFOSE, while the SIP disk sampler was able to sample the equivalent of about 107 m3 for 8:2 FTOH and 282 m3 for MeFOSE. The SIP disk sampler sampled in the linear range for the entire sampling period for MeFOSE and 60 days for 8:2 FTOH, which is much longer than the PUF disk sampler which reached equilibrium within hours for 8:2 FTOH and in 50 days for MeFOSE. Longer linear-phase sampling periods are desirable for passive samples so that a more accurate, time-weighted air concentration is derived. Implications. The SIP disk passive air sampler is effective at collecting the more volatile chemicals such as FTOHs, FOSAs, and FOSEs, and has derived air concentrations comparable to that of the PUF disk sampler for legacy POPs. This sampler appears to be a promising alternative to the traditional PUF disk samplers deployed worldwide in the GAPS Network. The use of SIP disk samplers would allow for a wider range of chemicals to be measured and monitored over time, which would support the Global Monitoring Plan of the Stockholm Convention on POPs. Acknowledgments We thank all of the partners under the GAPS network who participated in the pilot study, and also Christine Spencer for help with the analysis of the ionic PFCs. Funding was provided by the Chemicals Management Plan, Pesticide Science Fund, and the Northern Contaminants Program. Supporting Information Available Additional details on sampling sites and PCB and PFC concentrations. This information is available free of charge via the Internet at http://pubs.acs.org. Literature Cited (1) Pozo, K.; Harner, T.; Wania, F.; Muir, D. C. G.; Jones, K. C.; Barrie, L. A. Toward a global network for persistent organic pollutants in air: Results from the GAPS study. Environ. Sci. Technol. 2006, 40 (16), 4867-4873. (2) Lee, S. C.; Harner, T.; Pozo, K.; Shoeib, M.; Wania, F.; Muir, D. C. G.; Barrie, L. A.; Jones, K. C. Polychlorinated naphthalenes in the Global Atmospheric Passive Sampling (GAPS) study. Environ. Sci. Technol. 2007, 41 (8), 2680-2687. (3) Pozo, K.; Harner, T.; Lee, S. C.; Wania, F.; Muir, D. C. G.; Jones, K. C. Seasonally Resolved Concentrations of Persistent Organic Pollutants in the Global Atmosphere from the First Year of the GAPS Study. Environ. Sci. Technol. 2009, 43 (3), 796-803. (4) UNECE Protocol on Persistent Organic Pollutants under the 1979 Convention on Long-Range Transboundary Air Pollution; United Nations Economic Commission for Europe. http:// www.unece.org/env/lrtap/full%20text/1998.POPs.e.pdf. (5) UNEP preparation of an International Legally Binding Instrument Implementing International Action on Certain Persistent Organic Pollutants; United Nations Environment Programme. (6) The 9 new POPs, COP 4 - Geneva. http://www.pops.int. (7) Giesy, J. P.; Kannan, K. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35 (7), 1339- 1342. (8) Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A. Identification of long-chain perfuorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38 (2), 373-380. 5538 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 14, 2010 (9) Smithwick, M.; Muir, D. C. G.; Mabury, S. A.; Solomon, K. R.; Martin, J. W.; Sonne, C.; Born, E. W.; Letcher, R. J.; Dietz, R. Perflouroalkyl contaminants in liver tissue from East Greenland polar bears (Ursus maritimus). Environ. Toxicol. Chem. 2005, 24 (4), 981-986. (10) Smithwick, M.; Norstrom, R. J.; Mabury, S. A.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G. Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American Arctic, 1972-2002. Environ. Sci. Technol. 2006, 40 (4), 1139-1143. (11) Wania, F. A global mass balance analysis of the source of perfluorocarboxylic acids in the Arctic ocean. Environ. Sci. Technol. 2007, 41 (13), 4529-4535. (12) Dreyer, A.; Weinberg, I.; Temme, C.; Ebinghaus, R. Polyfluorinated Compounds in the Atmosphere of the Atlantic and Southern Oceans: Evidence for a Global Distribution. Environ. Sci. Technol. 2009, 43 (17), 6507-6514. (13) Shoeib, M.; Harner, T.; Vlahos, P. Perfluorinated chemicals in the Arctic atmosphere. Environ. Sci. Technol. 2006, 40 (24), 7577- 7583. (14) Stock, N. L.; Furdui, V. I.; Muir, D. C. G.; Mabury, S. A. Perfluoroalkyl contaminants in the Canadian arctic: Evidence of atmospheric transport and local contamination. Environ. Sci. Technol. 2007, 41 (10), 3529-3536. (15) Jahnke, A.; Ahrens, L.; Ebinghaus, R.; Temme, C. Urban versus remote air concentrations of fluorotelomer alcohols and other polyfluorinated alkyl substances in Germany. Environ. Sci. Technol. 2007, 41 (3), 745-752. (16) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40 (1), 32-44. (17) McMurdo, C. J.; Ellis, D. A.; Webster, E.; Butler, J.; Christensen, R. D.; Reid, L. K. Aerosol enrichment of the surfactant PFO and mediation of the water - Air transport of gaseous PFOA. Environ. Sci. Technol. 2008, 42 (11), 3969-3974. (18) Yamashita, N.; Taniyasu, S.; Petrick, G.; Wei, S.; Gamo, T.; Lam, P. K. S.; Kannan, K. Perfluorinated acids as novel chemical tracers of global circulation of ocean waters. Chemosphere 2008, 70 (7), 1247-1255. (19) Armitage, J. M.; MacLeod, M.; Cousins, I. T. Comparative Assessment of the Global Fate and Transport Pathways of LongChain Perfluorocarboxylic Acids (PFCAs) and Perfluorocarboxylates (PFCs) Emitted from Direct Sources. Environ. Sci. Technol. 2009, 43 (15), 5830-5836. (20) Guidance for a Global Monitoring Programme for Persistent Organic Pollutants; UNEP Chemicals: Geneva, Switzerland, June 2004. (21) Shoeib, M.; Harner, T.; Lee, S. C.; Lane, D.; Zhu, J. SorbentImpregnated Polyurethane Foam Disk for Passive Air Sampling of Volatile Fluorinated Chemicals. Anal. Chem. 2008, 80 (3), 675-682. (22) Pozo, K.; Harner, T.; Shoeib, M.; Urrutia, R.; Barra, R.; Parra, O.; Focardi, S. Passive-sampler derived air concentrations of persistent organic pollutants on a north-south transect in Chile. Environ. Sci. Technol. 2004, 38 (24), 6529-6537. (23) Wania, F.; Shen, L.; Lei, Y.; Teixeira, C.; Muir, D. C. G. Development and calibration of a resin-based passive sampling system for monitoring persistent organic pollutants in the atmosphere. Environ. Sci. Technol. 2003, 37, 1352-1359. (24) Harner, T.; Bidleman, T. Measurements of octanol-air partition coefficients for polychlorinated biphenyls. J. Chem. Eng. Data 1996, 41, 895-899. (25) Shoeib, M.; Harner, T. Characterization and comparison of three passive air samplers for persistent organic pollutants. Environ. Sci. Technol. 2002, 36 (19), 4142-4151. (26) Chaemfa, C.; Barber, J. L.; Gocht, T.; Harner, T.; Holoubek, I.; Klanova, J.; Jones, K. C. Field calibration of polyurethane foam (PUF) disk passive air samplers for PCBs and OC pesticides. Environ. Pollut. 2008, 156 (3), 1290-1297. (27) Harada, K.; Nakanishi, S.; Sasaki, K.; Furuyama, S.; Saito, N.; Yamakawa, K.; Koizumi, A. Particle Size Distribution and Respiratory Deposition Estimates of Airborne Perfluorooctanoate and Perfluorooctanesulfonate in Kyoto Area, Japan. Bull. Environ. Contam. Toxicol. 2006, 76 (2), 306-310. (28) Klanova, J.; Eupr, P.; Kohoutek, J.; Harner, T. Assessing the influence of meteorological parameters on the performance of polyurethane foam-based passive air samplers. Environ. Sci. Technol. 2008, 42 (2), 550-555. (29) Piekarz, A. M.; Primbs, T.; Field, J. A.; Barofsky, D. F.; Simonich, S. Semivolatile Fluorinated Organic Compounds in Asian and Western U.S. Air Masses. Environ. Sci. Technol. 2007, 41 (24), 8248-8255. (30) Barber, J. L.; Berger, U.; Chaemfa, C.; Huber, S.; Jahnke, A.; Temme, C.; Jones, K. C. Analysis of per- and polyfluorinated alkyl substances in air samples from Northwest Europe. J. Environ. Monit. 2007, 9 (6), 530-541. (31) Kim, S.-K.; Kannan, K. Perfluorinated Acids in Air, Rain, Snow, Surface Runoff, and Lakes: Relative Importance of Pathways to Contamination of Urban Lakes. Environ. Sci. Technol. 2008, 41 (24), 8328-8334. (32) Yarwood, G.; Kemball-Cook, S.; Keinath, M.; Waterland, R. L.; Korzeniowski, S. H.; Buck, R. C.; Russell, M. H.; Washburn, S. T. High-resolution atmospheric modeling of fluorotelomer alcohols and perfluorocarboxylic acids in the north American troposphere. Environ. Sci. Technol. 2007, 41 (16), 5756-5762. (33) Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J. Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40 (3), 864-872. (34) Jahnke, A.; Berger, U.; Ebinghaus, R.; Temme, C. Latitudinal gradient of airborne polyfluorinated alkyl substances in the marine atmosphere between Germany and South Africa (53 degrees N-33 degrees S). Environ. Sci. Technol. 2007, 41 (9), 3055-3061. ES1009696 VOL. 44, NO. 14, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5539 Cite This: Environ. Sci. Technol. 2018, 52, 4340-4349 Article pubs.acs.org/est Per- and Polyfluoroalkyl Substances in Swedish Groundwater and Surface Water: Implications for Environmental Quality Standards and Drinking Water Guidelines Laura Gobelius, Johanna Hedlund, Wiebke Durig, Rikard Troger, Karl Lilja, Karin Wiberg, and Lutz Ahrens*, Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences (SLU), Box 7050, SE-750 07 Uppsala, Sweden Swedish Environmental Protection Agency (Naturvar dsverket), Valhallavag en 195, 115 53 Stockholm, Sweden *S Supporting Information ABSTRACT: The aim of this study was to assess per- and polyfluoroalkyl substances (PFASs) in the Swedish aquatic environment, identify emission sources, and compare measured concentrations with environmental quality standards (EQS) and (drinking) water guideline values. In total, 493 samples were analyzed in 2015 for 26 PFASs (26PFASs) in surface water, groundwater, landfill leachate, sewage treatment plant effluents and reference lakes, focusing on hot spots and drinking water sources. Highest 26PFAS concentrations were detected in surface water (13 000 ng L-1) and groundwater (6400 ng L-1). The dominating fraction of PFASs in surface water were perfluoroalkyl carboxylates (PFCAs; 64% of 26PFASs), with high contributions from C4-C8 PFCAs (94% of PFCAs), indicating high mobility of shorter chain PFCAs. In inland surface water, the annual average (AA)-EQS of the EU Water Framework Directive of 0.65 ng L-1 for PFOS (linear and branched isomers) was exceeded in 46% of the samples. The drinking water guideline value of 90 ng L-1 for 11PFASs recommended by the Swedish EPA was exceeded in 3% of the water samples from drinking water sources (n = 169). The branched isomers had a noticeable fraction in surface- and groundwater for perfluorooctanesulfonamide, perfluorohexanesulfonate, and perfluorooctanesulfonate, highlighting the need to include branched isomers in future guidelines. INTRODUCTION Per- and polyfluoroalkyl substances (PFASs) are an emerging class of chemicals with unique physicochemical properties, such as extremely high environmental persistence and surfactant characteristics due to their stable carbon-fluorine bonds of the hydrophobic part of the molecule and a hydrophilic head.1 Furthermore, they are appreciated for their stain- and waterrepellent properties and have widely been used as surfactants, e.g., in textiles, packaging material, and aqueous film forming foams (AFFFs).2 The broad range of applications as well as their highly persistent and bioaccumulative characteristics1 have led to a ubiquitous distribution in the environment.3 PFASs have been detected in waters,4 soils,5,6 plants,7 wildlife,3 and humans8 in populated as well as in remote environments.9 Main point sources for PFASs are fire training sites where PFAS-containing AFFFs have been used (often located at airports and military training grounds),10,11 discharges from manufacturing industries and sewage treatment plants (STPs),12,13 and landfill leachate.14,15 Additionally, various diffuse sources, mainly related to the urban environment, e.g., surface runoff, and atmospheric deposition contribute to the environmental burden.16,17 Short-chained PFASs with a perfluorocarbon chain length of <C7 for perfluoroalkyl carboxylates (PFCAs) and <C6 for perfluoroalkanesulfonates (PFSAs) (including their precursors) show a relatively high water solubility and increased (environmental) mobility.2 In contrast, long-chain PFASs are more likely to accumulate in biota or sorb onto surfaces, i.e., partition to soils and sediments.18 The PFASs precursors are generally volatile and can be transported via long-range transport19 or transformed to more stable PFASs such as PFCAs and PFSAs.20 The aqueous environment is heavily affected by PFASs, and concentrations in surface water and groundwater that serve as drinking water source(s) pose a risk for human health.21-23 Human exposure to PFASs can also occur via (indoor) inhalation of fine dust particles and intake of contaminated food (particularly seafood and fish).24,25 Adverse health effects Received: Revised: Accepted: Published: November 8, 2017 March 9, 2018 March 12, 2018 March 12, 2018 2018 American Chemical Society 4340 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article include increased likelihood of cancer and disorders related to hepato- and immunotoxicity.26 However, the majority of restrictions and regulations is referring to the two best studied substances, perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA). PFOS is listed in the Stockholm Convention on Persistent Organic Pollutants (POPs) from 2009,27 while PFOA is restricted by the EU as a substance of very high concern (SVHC) in 201328 on the European level. Similarly, the main US manufacturer (3M) of PFOS declared in 2000 to voluntarily phase out the production of PFOS-related chemicals and the US EPA PFOA Stewardship Program encouraged the main perfluorochemical-producing companies to voluntarily phase out PFOA until 2015.29,30 As a consequence, environmental concentrations and human exposure patterns are shifting to shorter chained surrogates for PFOS and PFOA.31,32 Various guideline values have been established or implemented on international and national level for PFASs in the aquatic environment.33-41 Accordingly, the EU introduced an annual average environmental quality standard (AA-EQS) of 0.65 ng L-1 for PFOS (sum of linear and branched isomers) in inland surface water and a maximum acceptable concentration (MAC-EQS) of 36 000 ng L-1 for PFOS.42 Furthermore, the Swedish Geotechnical Institute (SGI) proposed a groundwater threshold value of 45 ng L-1 for PFOS.34 The United Kingdom was among the first nations to introduce drinking water guidelines for PFOS (300 ng L-1) and PFOA (10 000 ng L-1) in 2008,35 followed by the United States Environmental Protection Agency (U.S. EPA) with provisional health advisory values of 200 and 400 ng L-1 for PFOS and PFOA, respectively.38 The Netherlands set a limit of 530 ng L-1 for PFOS in drinking water.37 The U.S. EPA revised their health advisory value in 2016; a combined concentration of 70 ng L-1 for PFOA and PFOS should not be exceeded.39 However, most guidelines focus only on PFOS and PFOA with a few exceptions such as the Swedish drinking water guideline with a threshold value of 90 ng L-1 for 11PFASs (i.e., perfluoroalkyl chains C3-C9 PFCAs, i.e., PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, and PFDA, perfluoroalkyl chains C4, C6, and C8 PFSAs, i.e., PFBS, PFHxS, and PFOS, and 6:2 fluorotelomer sulfonate (6:2 FTSA)) from 2016,43 the Danish drinking water guideline with a threshold of 100 ng L-1 for 12PFASs, i.e., 11PFASs plus perfluorooctanesulfonamides (FOSA), from 201541 and the German drinking water regulation for 7 PFASs, i.e., PFOS: 100 ng L-1; PFOA: 100 ng L-1; PFBA: 10 000 ng L-1; PFHxA: 6000 ng L-1; PFNA: 60 ng L-1; PFBS: 6000 ng L-1; and PFHxS: 100 ng L-1, amounting to 22 360 ng L-1 for 7PFASs, from 2017.36 The high degree of variation in the guidelines (threshold values, included substances, and specifications about isomers and transformation products) indicates a large uncertainty about potential adverse health effects and a need for further research. A better understanding of potential PFASs emission sources and levels of PFASs in the aquatic environment, particularly in drinking water source areas, would help the regulative process of reaching water quality status in compliance with EQSs and other guideline values and thereby establish a better protection of the ecosystem and human health. The aim of this study was to assess 26 PFASs in the Swedish aquatic environment including surface water (n = 279), groundwater (n = 161), landfill leachate (n = 10), STP effluents (n = 13), and reference lakes (n = 10) for implications for EQS and (drinking) water guidelines with focus on hot spots and drinking water source areas. Specific objectives include (i) establishment of baseline concentrations of PFASs in the aquatic environment including drinking water source areas, (ii) identification of PFAS sources, (iii) spatial distribution of PFASs, and (iv) comparison of measured PFAS concentrations with guideline values. EXPERIMENTAL SECTION Chemicals. The analyzed PFASs comprise PFCAs with perfluorocabon chain lengths C3-C13, C15, and C17 (PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTriDA, PFTeDA, PFHxDA, and PFOcDA), C4, C6, C8, and C10 PFSAs (PFBS, PFHxS, PFOS, PFDS), and the PFAS precursors methyl- and ethylperfluorooctanesulfonamides (FOSAs: FOSA, MeFOSA, EtFOSA), methyl- and ethylperfluorooctanesulfonamidoethanols (FOSEs: MeFOSE, EtFOSE), methyl- and ethylperfluorooctanesulfonamidoacetic acid (FOSAAs: FOSAA, MeFOSAA, EtFOSAA) and 6:2 FTSA (purchased from Wellington Laboratories). Additionally, 13C4 PFBA, 13C2 PFHxA, 13C4 PFOA, 13C5 PFNA, 13C2 PFDA, 13C2 PFUnDA, 13C2 PFDoDA, 18O2 PFHxS, 13C4 PFOS, 13C8 FOSA, d3-N-MeFOSA, d5-N-EtFOSA, d7-N-MeFOSE, d9-N-EtFOSE, d3-N-MeFOSAA, and d5-N-Et-FOSAA were used as mass labeled internal standards (for detailed information on the chemical standards and characteristics of included PFASs, see Tables S1 and S2). Study Design and Sampling. Sampling was conducted mainly between May and August 2015 by 21 regional county administration boards (CABs) following a standardized sampling protocol (for details see Figure S1, Tables S3-S6, and text in the Supporting Information). The samples were categorized into surface water (n = 289), groundwater (n = 161), reference lakes (n = 10), landfill leachate (n = 10), and STP effluents (n = 13). The sampling locations were selected based on (i) potential vicinity of PFASs hot spots and (ii) the importance as a drinking water source area. Additionally, the samples can be classified into drinking water (n = 169) and nondrinking water samples (n = 304). The sampling was conducted by grab sampling using 1 L polypropylene (PP) bottles with subsequent storage at 4 C until analysis. For quality control, 10 triplicate samples and 20 field blanks were collected. The field blanks were obtained by opening the PP bottles at the sampling site for 1 min, rinsing the empty bottles with methanol, transferring the methanol to a vial and concentrating it under a nitrogen stream for PFAS analysis (more details to be found in the text in the Supporting Information). Sample Preparation and Analysis. All samples were analyzed according to Ahrens et al. (2009)4 with some modifications. Briefly, 500 mL of filtered sample (1.2 m glass fiber filter (GFF), GE Healthcare Life Sciences, Whatman, UK) were spiked with 100 L of internal standard (IS) mix (4 ng mL-1 for each of the 16 IS) and extracted with solid-phase extraction (SPE) using Oasis WAX cartridges (6 cc, 500 mg, 60 m, Waters Corporation, USA). Prior to extraction, the cartridges were preconditioned with 4 mL of a 0.1% ammonium hydroxide in methanol solution, 4 mL of methanol and 4 mL Millipore water. The flow rate of the water sample through the cartridge was adjusted to 1 drop s-1, thereafter the cartridge was washed with 4 mL 25 mM ammonium acetate buffer in Millipore water. Subsequently, cartridges were dried in a centrifuge for 2 min at 3000 rpm and the extracts were eluted with 4 mL of methanol and 8 mL of a 0.1% ammonium hydroxide solution in methanol. Finally, the extracts were 4341 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article Figure 1. Boxplot diagram with whiskers showing the 26PFAS concentrations [ng L-1] in groundwater (n = 161), surface water (n = 289), reference lakes (n = 10), STP effluent (n = 13) and landfill leachate (n = 10). The maximum concentrations are given at the top, and individual values can be obtained from Table S13. concentrated under a nitrogen stream to 1 mL for analysis using high-performance liquid chromatography coupled to a tandem mass spectrometer (HPLC-MS/MS; Agilent Technologies 1200 Series, Palo Alto, CA, USA, coupled to an Agilent 6460 Triple Quadrupole System, Palo Alto, CA, USA) (for details see Ahrens et al., 20094). PFASs were quantified using the isotope dilution method. Branched isomers of PFHxS, PFOS and FOSA (expressed as branched) were semiquantified with the corresponding linear (L) standard (for detailed information on the sample preparation and instrumental analysis see text in the Supporting Information). Quality Assurance and Quality Control. As part of the quality assurance and quality control, field blanks (n = 20), lab blanks (n = 26), PP-bottle blanks (n = 10), and PP-tube blanks (n = 10) (in total n = 66 blanks) (for preparation of the blanks, see Supporting Information text), sample triplicates (n = 10), method detection limits (MDLs), and recoveries were evaluated (for details see Tables S7-S9). The calibration curves showed a good linearity with a squared correlation coefficient (R2) of generally >0.9. Field blank concentrations were generally <2% of the concentrations measured in the samples. MDLs were calculated based on the mean lab blank concentration plus 3 times the standard deviation of the lab blank. If the compound was not detected in any of the blanks, then the lowest calibration point was used as MDL. The MDLs ranged from 0.03 ng L-1 (PFPeA) to 1.8 ng L-1 (6:2 FTSA). The relative standard deviation for the triplicate samples (three different 1-L samples from the same site) was on average 39%. The mean recoveries of the 16 mass labeled IS ranged from 25% (d5-N-EtFOSA) to 130% (d5-N-Et-FOSAA) (n = 473). RESULTS AND DISCUSSION Distribution of PFASs in Surface Water, Groundwater, Landfill Leachate and STP Effluent. At least one PFAS was detected above MDL in 90% of the samples, demonstrating the ubiquitous occurrence of PFASs across Sweden in all types of waters (Tables S11 and S12). The mean 26PFASs concentration (for all samples; n = 493) was 94 ng L-1, whereas the median concentration was considerably lower with 2.4 ng L-1 due to a few samples with exceptionally high levels (up to 13 000 ng L-1 in surface water). This large variation was expected, since the sampling was focused on PFAS-impacted sites as well as drinking water source areas and background levels in remote lakes. Among the five categories of water investigated (Figure 1), landfill leachate was the most polluted with a mean 26PFASs concentration of 490 ng L-1 (median 440 ng L-1, n = 10) and a maximum of 1300 ng L-1 (Figure S2). The leachate was highly polluted at almost all sampled sites, which is in accordance with Busch et al.14 and Benskin et al.;15 however, the mean is not necessarily representative for Sweden as a whole due to the low number of sampling sites (n = 10). The STP effluents showed a relatively low 26PFASs mean of 35 ng L-1, but the second highest 26PFASs median concentration (26 ng L-1; n = 13), with values up to 110 ng L-1 (Figure S2). STP effluents in other countries have shown higher values, e.g., STP effluent discharges to the river Elbe, Germany, with 30-270 ng L-1 for 39PFASs,44 in the San Francisco, CA, area, with 80-160 ng L-1 for 20PFASs,45 and in China with 1000-1600 ng L-1 for 14PFASs.46 STP effluents containing industrial PFASs discharges have been shown to exhibit high PFAS concentrations44,47 compared to STPs that only processed domestic wastewater. In surface water (lakes and rivers), the mean 26PFAS concentration was considerably high with 110 ng L-1 (n = 289) and a median 26PFAS concentration of 3.9 ng L-1, demonstrating high variability and skewed data. The high variation can be explained by the sampling strategy (preferential sampling near potential PFAS sources), but other factors, such as hydrological conditions and precipitation, are also known to significantly impact PFASs levels in streams, rivers and lakes.48,49 Loos et al. investigated persistent organic pollutants in over 100 European rivers and found slightly higher concentrations of PFOS (on average 39 ng L-1, median 6.0 ng L-1) and PFOA (on average 12 ng L-1, median 3.0 ng L-1)50 than those found in this study with a mean L-PFOS (L referring to the linear isomers) concentration of 24 ng L-1 (median 0.40 ng L-1) and 6.2 ng L-1 for PFOA (median 0.80 ng L-1). McLachlan et al. investigated 7 different PFASs in 14 European river mouths and detected the highest levels for PFOA (on average 20 ng L-1), followed by PFHpA (on average 2.9 ng L-1), PFHxA (on average 1.9 ng L-1), and PFNA (on average 1.7 ng L-1),51 while we found mean concentrations of 6.2, 2.9, 13, and 1.7 ng L-1 for PFOA, PFHpA, PFHxA, and PFNA, respectively, in Swedish rivers (n = 279). The surface water concentrations from the remotely located reference lakes (n = 10) had the lowest mean 26PFAS concentration (3.4 ng L-1; median 1.4 ng L-1) of all five categories (Figure 2). The occurrence of 4342 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article Figure 2. Surface water sampling locations in Sweden and median PFAS composition profiles (n = 263) are shown if available. Sampling locations from drinking water source areas are confidential and thus not marked on the map (n = 16). The reference lakes (n = 10) are marked with blue diamonds and their composition profile is shown in the bottom right corner. Figure 3. Groundwater sampling locations in Sweden and median PFAS composition profiles (n = 8) are shown if available. Sampling locations from drinking water source areas are confidential and thus not marked on the map (n = 149). ND = not detected; NA = not available. PFASs in these lakes can be explained by atmospheric deposition rather than point sources and can be considered as background levels. The low concentrations confirm earlier hypotheses and findings of spatial proximity to point sources correlating with elevated PFAS concentrations.3,52 This is consistent with our surface water data showing slightly higher 26PFAS concentrations in southern Sweden, where the population density is highest, in comparison to northern Sweden (Figure 2). The mean groundwater (n = 161) concentration of 26PFASs was 49 ng L-1 (median 0.04 ng L-1) with values up to 6400 ng L-1, which is the second highest concentration detected in this study, and most probably due to 4343 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article impact from an adjacent fire training site. Maximum PFAS concentrations in this study were in the same range as previously observed at PFAS-contaminated areas impacted by fire training sites (3000-110 000 ng L-1 for PFAS53 and 17 ng L-1 for 8PFASs)22 or landfills (4400 ng L-1 for 8PFASs).22,53 However, the low median for the 26PFASs shows that the majority of groundwater samples were unaffected by PFAS sources (Figure 3). However, a few elevated concentrations can be found in the southern part of Sweden, most probably correlating with increased anthropogenic activity. Groundwater studies focusing on unaffected sites are scarce.54,55 Loos et al. reported a median PFOA concentration of 1 ng L-1 and a mean PFOS concentration of 4 ng L-1, whereas PFHxS and PFHpA were below the MDL (<0.4 ng L-1 for both PFOA and PFOS) in a screening study of groundwater in 23 European countries (n = 164).54 Munoz et al. detected a median 22PFASs concentration of 0.56 ng L-1 in groundwater (n = 80) in the tropical French Overseas Territories of French Guyana, Guadeloupe, Martinique, Mayotte, and Reunion with a maximum concentration of 638 ng L-1, which could be related to fire-fighting operations using AFFF or industrial activities.55 The predominant substances were the short-chained PFCAs and 6:2 FTSA, i.e., similar to the PFASs composition of groundwater found in this study. The composition profiles of PFASs varied with the different water types (Figure S1), while spatial trends were not clearly visible in the surface water (Figure 2) or groundwater composition profiles (Figure 3). The groundwater profile differed the most from other profiles, with an even distribution of PFCAs (40%), PFSAs (30%), and PFAS precursors (29%). L-FOSA and L-PFHxS had the highest fractions of individual PFASs in groundwater with 22 and 19% of the mean composition, respectively. PFHxS was also frequently found in groundwater in France;56 however, there is no data available on FOSA. Therefore, further research is suggested.54,56 In the other water types, namely, surface water (lakes and rivers), landfill leachate, STP effluent, and the reference lakes, the PFCAs were the dominating group with 63, 65, 53, and 77%, respectively (Figure S3). Similar observations of PFCAs as the dominant PFAS class were made for surface water (70% for PFCAs of the 39PFASs44 in Germany), landfill leachate (15- 56% (flow-through) and 86% (recirculated)15 in Canada), and STP effluents (70%44 in Germany; 30-60% for PFCAs of 20PFASs45 in the USA; 66%47 in Denmark). Data on background concentrations in aquatic environments is rather limited, with one study from Japan reporting PFOA values around 0.1 ng L-1 in remote lakes and rivers57 and a Scandinavian screening study reporting values between nondetected and 21 ng L-1 for 7PFASs in remote areas;58 however, none of these studies provided composition profiles. Out of the PFCAs, PFOA (15 and 26%) and PFHxA (17 and 25%) were the most dominant in surface water and landfill leachate, respectively, while being the second and third biggest fractions in STP effluent, only outnumbered by 6:2 FTSA (27%). Houtz et al. also found PFHxA, PFOA, and 6:2 FTSA to be dominant in STP effluent from the San Francisco, CA, area and concluded that the 6:2 FTSA derived from PFAScontaining AFFF.45 Interestingly, 6:2 FTSA is one of the dominant PFAS in the reference lakes (17%), although 6:2 FTSA cannot be considered as volatile (log vapor pressure = -0.96 Pa).1 Nevertheless, long-range atmospheric transport (including particle bound transport) of 6:2 FTSA cannot be ruled out and needs to be further investigated. Other than 6:2 FTSA, the background lakes contained high proportions of PFHpA (20%) and PFOA (17%). There was also a noteworthy difference of L-PFOS and PFOA in surface water with average proportions of 14 and 15%, respectively, in comparison to groundwater, where the proportions were three times lower (5.3 and 5.6%, respectively). This phenomenon has been described earlier by Higgins and Luthy, who showed that PFOS and PFOA have a higher potential for partitioning to particles and for bioaccumulation; thus the low fractions in groundwater imply that L-PFOS and PFOA adsorbed to soil on their way to the aquifer or were taken up by biota.59 Moreover, surface water is directly exposed to various direct and diffuse sources, while groundwater is protected by the soil layer that influences the composition of PFASs in the aquifer.22 Source Related Occurrence of PFASs. Six different source categories were established, for which source data were available (n = 360) from the CABs or Swedish EPA, whereas sites with unknown or several potential sources were excluded (n = 113) to avoid misinterpretation (Table S13). The categories fire training sites, unspecific industry, STP effluent, landfill/waste disposal, skiing, and urban area were based on point sources identified in literature11,13-15,17,22,31,45,47,60-62 and were evaluated separately for surface water and groundwater. Independent of the emission source characterized in this study, the surface water samples (n = 235) were dominated by PFCAs (on average, 74%), while PFCAs (on average, 35%), PFSAs (on average, 32%), and PFAS precursors (on average, 31%) were evenly distributed in the groundwater samples (Figure 4). The groundwater data (n = 30) were analyzed in a principal component analysis (PCA; Simca v.14, Umetrics; Figure S4). The PCA loading plot showed that L-FOSA was closely linked to landfills/waste disposal sites, whereas fire training sites were related to short-chained PFASs, i.e., PFBS, PFHxS, and PFBA, as well as L-PFOS. The source categories landfill, urban area, and unspecific industry are closely associated with each other and share similar contamination profiles but were not related to any specific PFASs due to the fact that most PFASs were strongly associated with each other. The maximum surface (n = 142) and groundwater (n = 41) PFAS concentrations (13 000 and 6400 ng L-1, respectively) were associated with firefighting training sites including air fields and military areas. Consequently, this source category showed the highest contamination with on average 200 ng L-1 26PFASs (median 7.9 ng L-1) and 184 ng L-1 (median 2.7 ng L-1), respectively. PFSAs (42%) constituted the main fraction in the AFFF-impacted groundwater, in accordance with a previous study,60 followed by PFCAs (34%) and PFAS precursors (24%). In the surface water samples, PFCAs (54%) were dominating, followed by PFSAs (34%) and PFAS precursors (11%). For individual PFASs, L-PFHxS (29% of the 26PFASs) was the dominant PFAS in groundwater, followed by L-FOSA (13% of the 26PFASs), whereas in surface water L-PFOS (19% of the 26PFASs) was the dominant PFAS, followed by PFHxA (15% of the 26PFASs). Similarly, Anderson et al. found PFOS and PFHxS to be the most common PFASs at AFFF-impacted sites in various media,60 and Awad et al. observed a high correlation between these two substances in ground water and surface water at an AFFF-impacted site, suggesting a common origin.11 Sites potentially impacted by landfills and waste disposal caused the second highest pollution with mean (median) 26PFAS concentrations of 64 ng L-1 (3.6 ng L-1) and 13 ng L-1 (2.0 ng L-1) in surface water (n = 20) and 4344 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article Figure 4. Composition profiles of PFASs for different source categories for surface water and groundwater. The surface water categories are fire training sites (n = 142), unspecific industry (n = 45), STP effluent (n = 14), landfill/waste disposal sites (n = 20), skiing areas (n = 6), and urban areas (n = 8). The groundwater categories are fire training sites (n = 41), unspecific industry (n = 61), landfill/waste disposal sites (n = 4), skiing areas (n = 3), and urban areas (n = 16). PFCAs are colored in blue and gray, PFSAs in green and PFAS precursors in orange/red. groundwater (n = 4), respectively. Despite not being in the proximity of drinking water source areas, discharges from such places still pose a risk for the aquatic environment.14,22,47 In the few landfill impacted groundwater samples represented in this study (n = 4), the composition profile was dominated by PFSAs (49%), followed by PFCAs (41%) and PFAS precursors (9.3%), with L-PFHxS (37%) and PFOA (18%) being the predominant compounds. In contrast, the mean composition of the surface water was strongly dominated by PFCAs (72%), mainly PFOA (24%) and PFHxA (22%), followed by PFSAs (19%) and PFAS precursors (9.5%). This is in accordance with previous studies that identified the shorter chain (C3-C8) PFCAs (58% of the 43PFASs)14 and C4-C8 PFCAs (83% for recirculated and 15-56% for flow through of the 24PFASs)15 as the dominant PFASs in leachate. This indicates that the PFAS composition profile can change during the leaching through soil, due to the higher mobility of shorter than longer chain PFASs and higher mobility of PFCAs than of the PFSAs,18,22 whereas the leachate probably enters the surface water directly and therefore the composition remains rather unchanged. However, the PFAS concentration and composition in leachate water strongly depends on the landfill content and the amounts disposed.14,15 26PFAS concentrations at sites impacted by STP effluent (only for surface water; n = 14) were relatively low (mean 9.4 ng L-1; median 2.7 ng L-1; maximum 36 ng L-1) in comparison to landfill impacted sites (mean 64 ng L-1; median 3.6 ng L-1; maximum 760 ng L-1 for surface water and 13 ng L-1, 2 ng L-1, and 49 ng L-1 for groundwater, respectively), which is lower than the individual PFAS concentrations of up to 24 ng L-1 (PFOA) detected in STP effluents in Denmark.47 Furthermore, in the same study, PFOS and PFOA were found to be the main components of the effluents (28 and 56% of the 6PFASs), and PFHxA or PFHpA (that constituted the main fraction in our effluentimpacted samples) were not detected.47 PFCAs (70%) were the dominant PFAS class in STP effluent-impacted sites, followed by PFSAs (20%) and PFAS precursors (9.6%) with PFHxA (25%) and PFHpA (15%) as the dominant PFASs. This is in accordance with Houtz et al., who found 15 out of 20 PFASs in wastewater effluent, with PFHxA as the predominant compound (median 24 ng L-1). However, PFHpA was detected at very low levels (5 ng L-1).45 Generally, Houtz et al. and Bossi et al. detected considerably higher concentrations of PFASs in the effluents than we did in the receiving waters.45,47 This can potentially be attributed to dilution effects, the distance to the source, different wastewater treatment techniques or lower inflow concentrations. Nevertheless, STPs have previously proven as important sources of PFASs, as conventional wastewater treatment techniques are ineffective in removing PFASs.13 PFASs are known to be released from STPs via several pathways including the effluent water itself, volatilization during the treatment process and subsequent sludge application as agricultural fertilizer.13 Sites impacted by unspecific industry (n = 104) comprise the manufacture of PFASs-containing products (e.g., textile industry) and processes like metal plating. In our study, this rather unspecific category still seems to be a reasonable indicator for PFASs in the surrounding environment,31,61 leading to elevated PFAS levels, as observed in our samples with maximum 26PFAS concentrations of 210 ng L-1 in groundwater and 173 ng L-1 in surface water. Here, the groundwater samples (n = 61) showed a lower mean 26PFASs concentration (5.7 ng L-1; median 0.30 ng L-1) than the mean in the surface water (13 ng L-1; median 3.5 ng L-1; n = 45). Furthermore, the groundwater composition was characterized by PFCAs (43%), followed by PFSAs (29%) and PFAS precursors (26%) with L-FOSA (23%) representing the main PFAS. In the surface water, PFCAs clearly dominated (73%), followed by PFSAs (19%) and PFAS precursors (6.9%). As mentioned above, this indicates that the PFAS composition 4345 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article Figure 5. Concentrations of PFOS and 26PFASs in drinking water (n = 169), surface water (n = 289) and groundwater (n = 161) and comparison with applicable guideline values. Note the logarithmic scale. The drinking water thresholds refer to the Swedish drinking water guideline, the surface water thresholds refer to the EU WFD, while the groundwater threshold was recommended by the Swedish SGI. profile may change during the leaching through soil depending on the physicochemical properties of individual PFASs.18,59 Consequently, the three most prevalent PFASs in surface water were PFOA (22%), PFHxA (17%) and PFHpA (14%). These compounds were also found at notable levels in the river Rhine watershed influenced by unspecific industries with PFOA as the predominant one.31 The sites impacted by cities (urban areas; n = 24) were more likely impacted by diffuse sources, e.g., street runoff, waxes, paints, and clothes, depending on the population density and industrial activities.17 The samples in this category showed only slightly elevated PFAS concentrations (mean 26PFASssurface water 1.0 ng L-1; mean 26PFASsgroundwater 3.1 ng L-1) in comparison to previously mentioned categories. While the median 26PFAS concentrations in the urbanimpacted water were rather similar, i.e., 0.90 ng L-1 for surface water and 0.80 ng L-1 for groundwater, the maximum groundwater concentration (22 ng L-1) was notably higher than the maximum surface water concentration (1.9 ng L-1). Additionally, the groundwater (n = 16) composition was similar for the PFSAs (41%) and PFCAs (39%) but lower for the PFAS precursors (19%), with L-PFHxS (20%) and L-FOSA (19%) as the main PFASs. In the surface water (n = 8), PFCAs (89%) clearly dominated over PFSAs (10%) with PFNA (26%), PFHxA (19%), and PFOA (19%) as the dominant PFASs and PFAS precursors not being present. The lowest PFASs burden occurred at sites impacted by skiing (n = 9), although skiing waxes are reported as a source for PFASs.62,63 The low values can likely be explained by the fact that the skiing areas are quite remotely located with a rather low impact of human activity. In addition, skiing waxes contain mainly long-chained PFASs63 with a higher potential to partition to particles rather than water,59 leading to low groundwater concentrations for 26PFASs (on average and median 0.20 ng L-1, maximum 0.40 ng L-1, n = 3). 26PFAS concentrations in surface water (n = 6) were only slightly higher (mean 13 ng L-1; median 2.2 ng L-1, maximum 57 ng L-1) in comparison to groundwater levels. The dominant PFASs in groundwater were L-FOSA (75%), PFNA (13%), and PFHpA (8%) (no PFSAs were detected), while PFOA (25%), PFHxA (22%), and PFNA (19%), were dominant in surface water with PFSAs constituting 14% (no PFAS precursors were detected). All 13 PFASs detected in >10% of the groundwater (n = 161) samples, i.e., PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFBS, PFHxS, PFOS, and FOSA, showed a significant correlation (p < 0.05) with each other, except FOSA, indicating a common source (Table S14). Additionally, all 11 PFASs detected in >10% of the surface water (n = 289) samples, i.e., PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFBS, PFHxS, PFOS, 6:2 FTSA, and FOSA, showed significant correlation with each other (p < 0.01), indicating a common or similar source (Table S15). Linear and Branched Isomers. The compounds FOSA, PFHxS, and PFOS were studied separately as branched or linear isomers (L-) for surface water and groundwater (Figure S5). Isomeric composition has previously been used to trace emission sources.15,64 The surface water samples generally showed a higher proportion of branched isomers than the groundwater samples, which is in contradiction to results from Kar rman et al.,64 who concluded the opposite and suggested higher water solubility of branched isomers. However, the higher proportion of branched isomers in surface water was only statistically significant (p < 0.05; two-sided Student's t 4346 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article test) for FOSA and PFOS. While branched-FOSA was solely occurring in the surface water samples (2% of the FOSA), branched-PFOS had the highest fraction of branched isomers in both, surface water (20% of the PFOS) and groundwater (8% of the PFOS). Kar rman et al. and Houde et al. found higher proportions of branched-PFOS in water samples, ranging between 39 and 42% and 44-57% of the PFOS concentration, respectively.64,65 The branched-PFHxS occurred with 5 and 7% of the PFHxS in groundwater and surface water, respectively. As the branched isomers are not explicitly mentioned in the drinking water guidelines referred to earlier,35,37-39,41,43 they were not considered in our evaluation of the drinking water safety. While the low fractions of branched isomers in our samples suggest a relatively low risk for drinking water, the results from Kar rman et al. and Houde et al. showed that branched isomers should be included in the drinking water guidelines.64,65 For the Water Framework Directive (WFD), another 4% (n = 12) of the surface water samples exceeded the AA-EQS threshold value (0.65 ng L-1) compared if only L-PFOS was considered. Ultimately, branched isomers should be considered in legislation and guidelines since they showed partly a high fraction in the aquatic environment, i.e., up to 20% for branched-PFOS in surface water, and have a similar toxicity potential compared to the linear isomer.66 Comparison of Measured Concentrations with Guideline Values. The sum of branched and linear isomers of PFOS was found at levels above the MDL (0.21 ng L-1) in 63% of the surface water samples (n = 176 out of 289), of which 74% (n = 123) exceeded the AA-EQS of 0.65 ng L-1, while all measured concentrations stayed well below the MAC-EQS value of 36 000 ng L-1 of the EU WFD (Figure 5). However, the results should be treated with caution since the value of 0.65 ng L-1 is close to the MDLPFOS in this study (0.21 ng L-1). There is thus an uncertainty in whether this threshold was exceeded only at that particular time of sampling or over long-term. Irrespectively, it might be questionable if PFOS concentrations of >0.65 ng L-1 are threatening the ecosystems.67-69 According to Loos et al., the median PFOS concentration in European rivers is 6.0 ng L-1. Hence, many European inland surface waters were exceeding the AA-EQS,50 while the median L-PFOS surface water concentration in this study was 0.40 ng L-1. To the best of the author's knowledge, there is no equally comprehensive study on PFASs in European inland surface waters available for comparison, however, it is not known how the concentrations have developed and if the data from Loos et al. still reflects current concentrations. Besides, the EU WFD only focuses on PFOS, while our results show that L-PFOS constituted only a small fraction (14%) of the 26PFASs in surface water, with a similar fraction of other PFASs, namely, PFHxA (17%), PFOA (15%), and PFHpA (14%). Thus, providing that sufficient toxicity data is available, it could be argued that other PFASs including branched and linear isomers as well as PFAS precursors need to be included in EU WFD and other regulations. None of the groundwater samples (n = 164) exceeded the guideline value for PFOS (45 ng L-1) recommended by SGI.34 If assuming the same toxicity as PFOS for all 26 PFASs, then 10 samples would exceed 45 ng L-1 for 26PFASs. This indicates that groundwater guideline values should include more PFASs, as discussed for the surface water, to protect ecosystem and human health. For the drinking water source areas, 5 out of 169 samples (3.0%) exceeded the threshold value of 90 ng L-1 for 11PFASs (PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFBS, PFHxS, PFOS, and 6:2 FTSA) of the Swedish National Food Agency43 (Figure S6). The number of samples would not increase if 26PFASs were considered in the drinking water threshold value of 90 ng L-1. However, L-FOSA, the substance with the highest PFAS fraction (20%) and a high detection frequency (31%) is not included in the Swedish drinking water guideline43 and should be considered in future assessments. The few (n = 5) samples that exceeded the Swedish drinking water guideline value originated from groundwater, with four of them most probably being impacted by fire training sites, while the fifth sample was located close to an agricultural area where PFAS-contaminated sludge could have been applied as fertilizer, a PFAS impact that has been observed previously.49 A spatial trend for PFASs could not be identified in the drinking water samples, but it is interesting to note that four counties with a low population density (<20 people per km2) had composition profiles dominated by FOSA, while all other composition profiles consisted of mainly PFCAs and PFSAs (Figure S6). The Swedish limit value for sensitive citizens of 900 ng L-1 was not exceeded in any of the samples from drinking water source areas. While the Swedish and Danish drinking water guideline values, i.e., 90 and 100 ng L-1 for 11PFASs and 12PFASs, respectively, were exceeded in a few cases, no drinking water sample from our study exceeds the drinking water recommendations from the U.S. EPA, United Kingdom, The Netherlands, and Germany. However, there are to the best of the author's knowledge no comprehensive screening studies available for other countries, but the current study emphasizes the need for increased knowledge about health risks associated with intake of PFASs and achieving a more coherent view in protecting aquatic ecosystems and human health. Furthermore, as monitoring data are scarce, there are large uncertainties about spatial and temporal variation of PFASs in the aquatic environment. For source tracing and better understanding of limit exceedance, regular monitoring of PFASs in the ecosystem and in drinking water source areas is urgently needed. Sites potentially impacted by PFASs from firefighting training as well as landfill leachate, STP effluents and biosolid impacted agricultural areas are of high concern and extra efforts should be made to identify more sources that pose risks to the environment and human health. ASSOCIATED CONTENT *S Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.7b05718. Explanatory text about sampling and blank preparation, map of counties, hot spot sampling locations and composition profiles, composition profiles of PFASs in various types of water, PCA for groundwater samples, fraction of linear and branched isomers, drinking water composition profiles for respective counties, table with chemicals, purity, manufacturer, table with most important PFASs characteristics, sampling protocols in Swedish and English, blank concentrations and MDLs, triplicates and recoveries, sample IDs and locations, individual PFAS concentration, list of samples included in source classification, Pearson coefficients and significance for PFASs in groundwater and surface water (PDF) 4347 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article AUTHOR INFORMATION Corresponding Author *E-mail: @slu.se. Phone: ORCID Laura Gobelius: 0000-0002-0407-2621 Lutz Ahrens: 0000-0002-5430-6764 Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS We kindly acknowledge Jelena Rakovic and Elm Andersson for assistance with the lab analysis, Emil Back for the compilation of the ArcGIS maps, the Swedish EPA for funding of the project (NV-02893-1S), and all County Administration Boards for the selection of sampling sites and the conduction of the sampling. REFERENCES (1) Wang, Z.; MacLeod, M.; Cousins, I. T.; Scheringer, M.; Hungerbuehler, K. Using COSMOtherm to Predict Physicochemical Properties of Poly- and Perfluorinated Alkyl Substances (PFASs). Environ. Chem. 2013, 8 (4), 389-398. (2) Buck, R. C.; Franklin, J.; Berger, U.; Conder, J. M.; Cousins, I. T.; de Voogt, P.; Jensen, A. A.; Kannan, K.; Mabury, S. A.; van Leeuwen, S. P. J. Perfluoroalkyl and Polyfluoroalkyl Substances in the Environment: Terminology, Classification, and Origins. Integr. Environ. Assess. Manage. 2011, 7 (4), 513-541. (3) Giesy, J. P.; Kannan, K. Global Distribution of Perfluorooctane Sulfonate in Wildlife. Environ. Sci. Technol. 2001, 35 (7), 1339-1342. (4) Ahrens, L.; Barber, J. L.; Xie, Z.; Ebinghaus, R. Longitudinal and Latitudinal Distribution of Perfluoroalkyl Compounds in the Surface Water of the Atlantic Ocean. Environ. Sci. Technol. 2009, 43 (9), 3122-3127. (5) Filipovic, M.; Laudon, H.; McLachlan, M. S.; Berger, U. Mass Balance of Perfluorinated Alkyl Acids in a Pristine Boreal Catchment. Environ. Sci. Technol. 2015, 49 (20), 12127-12135. (6) Braunig, J.; Baduel, C.; Heffernan, A.; Rotander, A.; Donaldson, E.; Mueller, J. F. Fate and Redistribution of Perfluoroalkyl Acids through AFFF-Impacted Groundwater. Sci. Total Environ. 2017, 596597, 360-368. (7) Yoo, H.; Washington, J. W.; Jenkins, T. M.; Ellington, J. J. Quantitative Determination of Perfluorochemicals and Fluorotelomer Alcohols in Plants from Biosolid-Amended Fields Using LC/MS/MS and GC/MS. Environ. Sci. Technol. 2011, 45 (19), 7985-7990. (8) Glynn, A.; Berger, U.; Bignert, A.; Ullah, S.; Aune, M.; Lignell, S.; Darnerud, P. O. Perfluorinated Alkyl Acids in Blood Serum from Primiparous Women in Sweden: Serial Sampling during Pregnancy and Nursing, And Temporal Trends 1996-2010. Environ. Sci. Technol. 2012, 46 (16), 9071-9079. (9) Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A. Identification of Long-Chain Perfluorinated Acids in Biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38 (2), 373-380. (10) Baduel, C.; Paxman, C. J.; Mueller, J. F. Perfluoroalkyl Substances in a Firefighting Training Ground (FTG), Distribution and Potential Future Release. J. Hazard. Mater. 2015, 296, 46-53. (11) Awad, E.; Zhang, X.; Bhaysar, S. P.; Petro, S.; Crozier, P. W.; Reiner, E. J.; Fletcher, R.; Tittlemier, S. A.; Braekevelt, E. Long-Term Environmental Fate of Perfluorinated Compounds after Accidental Release at Toronto Airport. Environ. Sci. Technol. 2011, 45 (19), 8081-8089. (12) Davis, K. L.; Aucoin, M. D.; Larsen, B. S.; Kaiser, M. A.; Hartten, A. S. Transport of Ammonium Perfluorooctanoate in Environmental Media near a Fluoropolymer Manufacturing Facility. Chemosphere 2007, 67 (10), 2011-2019. (13) Alder, A. C.; van der Voet, J. Occurrence and Point Source Characterization of Perfluoroalkyl Acids in Sewage Sludge. Chemosphere 2015, 129, 62-73. (14) Busch, J.; Ahrens, L.; Sturm, R.; Ebinghaus, R. Polyfluoroalkyl Compounds in Landfill Leachates. Environ. Pollut. 2010, 158 (5), 1467-1471. (15) Benskin, J. P.; Li, B.; Ikonomou, M. G.; Grace, J. R.; Li, L. Y. Per- and Polyfluoroalkyl Substances in Landfill Leachate: Patterns, Time Trends, and Sources. Environ. Sci. Technol. 2012, 46 (21), 11532-11540. (16) Ahrens, L.; Bundschuh, M. Fate and Effects of Poly- and Perfluoroalkyl Substances in the Aquatic Environment: A Review. Environ. Toxicol. Chem. 2014, 33 (9), 1921-1929. (17) Zushi, Y.; Masunaga, S. Identifying the Nonpoint Source of Perfluorinated Compounds Using a Geographic Information System Based Approach. Environ. Toxicol. Chem. 2009, 28 (4), 691-700. (18) Gellrich, V.; Stahl, T.; Knepper, T. P. Behavior of Perfluorinated Compounds in Soils during Leaching Experiments. Chemosphere 2012, 87 (9), 1052-1056. (19) Jahnke, A. Polyfluorinated Alkyl Substances (PFAS) in the Marine Atmosphere - Investigations on Their Occurrence and Distribution in Coastal Regions; GKSS-Forschungszentrum Geesthacht GmbH: Geesthacht, Germany, 2007. (20) Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Hurley, M. D.; Andersen, M. P. S.; Wallington, T. J. Atmospheric Lifetime of Fluorotelomer Alcohols. Environ. Sci. Technol. 2003, 37 (17), 38163820. (21) Jakobsson, K.; Kronholm Diab, K.; Lindh, C.; Persson, B.; Jonsson, B. Exponering for pedluorerade amnen (PFAS) i dricksvatten i Ronneby kommun. Region Skane: Malmo, Sweden, 2014. (22) Eschauzier, C.; Raat, K. J.; Stuyfzand, P. J.; De Voogt, P. Perfluorinated Alkylated Acids in Groundwater and Drinking Water: Identification, Origin and Mobility. Sci. Total Environ. 2013, 458-460, 477-485. (23) Hu, X. C.; Andrews, D. Q; Lindstrom, A. B.; Bruton, T. A.; Schaider, L. A.; Grandjean, P.; Lohmann, R.; Carignan, C. C.; Blum, A.; Balan, S. A.; Higgins, C. P.; Sunderland, E. M. Detection of Polyand Perfluoroalkyl Substances (PFASs) in US Drinking Water Linked to Industrial Sites, Military Fire Training Areas, and Wastewater Treatment Plants. Environ. Sci. Technol. Lett. 2016, 3 (10), 344-350. (24) Moriwaki, H.; Takata, Y.; Arakawa, R. Concentrations of Perfluorooctane Sulfonate (PFOS) and Perfluorooctanoic Acid (PFOA) in Vacuum Cleaner Dust Collected in Japanese Homes. J. Environ. Monit. 2003, 5 (5), 753-757. (25) Domingo, J. L. Health Risks of Dietary Exposure to Perfluorinated Compounds. Environ. Int. 2012, 40, 187-195. (26) Lau, C.; Anitole, K.; Hodes, C.; Lai, D.; Pfahles-Hutchens, A.; Seed, J. Perfluoroalkyl Acids: A Review of Monitoring and Toxicological Findings. Toxicol. Sci. 2007, 99 (2), 366-394. (27) Stockholm Convention on Persistent Organic Pollutants (POPs) Annex B - Restrictions; Perfluorooctane Sulfonic Acid, Its Salts and Perfluorooctane Sulfonyl Fluoride; Decision SC-4/17; Stockholm Convention: Chatelaine, Switzerland, 2009. (28) Inclusion of Substances of Very High Concern in the Candidate List (Decision of the European Chemicals Agency); Document ED/69/2013 SVHC Article 57d; ECHA - European Chemicals Agency: Helsinki, 2013. (29) EPA and 3M Announce phase out of PFOS. Published 05/16/ 2000. https://archive.epa.gov/epapages/newsroom_archive/ newsreleases/33aa946e6cb11f35852568e1005246b4.html (accessed September 29, 2017). (30) PFOA Stewardship Program; U.S. EPA: Washington, DC, 2010. (31) Moller, A.; Ahrens, L.; Surm, R.; Westerveld, J.; van der Wielen, F.; Ebinghaus, R.; de Voogt, P. Distribution and Sources of Polyfluoroalkyl Substances (PFAS) in the River Rhine Watershed. Environ. Pollut. 2010, 158 (10), 3243-3250. (32) D'Agostino, L. A.; Mabury, S. A. Identification of Novel Fluorinated Surfactants in Aqueous Film Forming Foams and 4348 DOI: 10.1021/acs.est.7b05718 Environ. SO. Technol. 2018, 52, 4340-4349 Environmental Science & Technology Article Commercial Surfactant Concentrates. Environ. Sci. Technol. 2014, 48 (1), 121-129. (33) European Commission. Directive 2013/39/EU of the European Parliament and of the Council - Amendments to the WFD, List of Priority Substances (PFOS). Off. J. Eur. Communities: Legis. 2013, 226, 1-17. (34) Pettersson, M.; Lan dell, M.; Ohlsson, Y.; Berggren Kleja, D.; Tiberg, C. Preliminara riktvarden for hog fluorerade amnen (PFAS) i mark och grundvatten; Statens geotekniska institut: Linkoping, Sweden, 2015. (35) Atkinson, C.; Blake, S.; Hall, T.; Kanda, R.; Rumsby, P. Survey of the Prevalence of Perfluorooctane Sulfonate (PFOS), Perfluorooctanoic Acid (PFOA) and Related Compounds in Drinking Water and Their Sources; WRc: Swindon, U.K., 2008. (36) Umweltbundesamt. Fortschreibung der vorlaufigen Bewertung von per- und polyfluorierten Chemikalien (PFC) im Trinkwasser - Empfehlung des Umweltbundesamtes nach Anhor ung der Trinkwasserkommission; Springer-Verlag: Berlin Heidelberg, 2017. (37) Moermond, C. T. A.; Verbruggen, E. M. J.; Smit, C. E. Environmental Risk Limits for PFOS - A Proposal for Water Quality Standards in Accordance with the Water Framework Directive; rivm - National Institute for Public Health and the Environment: Bilthoven, the Netherlands, 2010. (38) Provisional Health Advisories for Perfluorooctanoic Acid (PFOA) and Perfluorooctane Sulfonate (PFOS); U.S. EPA: Washington, DC, 2009. (39) Fact Sheet: PFOA & PFOS Drinking Water Health Advisories; U.S. EPA: Washington, DC, 2016. (40) RISKHANTERINGSRAPPORT - Risker Vid For orening Av Dricksvatten Med PFAS; Livsmedelsverket: Uppsala, 2016. (41) Administrative overvejelser og fastlggelse af grnsevrdier for perfluorerede alkylsyreforbindelser (PFAS-forbindelser), inkl. PFOA, PFOS og PFOSA i drikkevand, samt jord og grundvand til vurdering af forurenede grunde; Miljministeriet: Copenhagen, 2015. (42) European Commission. Directive 2000/60/EC: EU Water Framework Directive - Integrated River Basin Management in Europe. Off. J. Eur. Communities: Legis. 2000, 327, 1-73. (43) Riskhantering - PFAS i dricksvatten och fisk https://www. livsmedelsverket.se/livsmedel-och-innehall/oonskade-amnen/ miljogifter/pfas-poly-och-perfluorerade-alkylsubstanser/riskhanteringpfaa-i-dricksvatten (accessed October 18, 2017). (44) Ahrens, L.; Plassmann, M.; Xie, Z.; Ebinghaus, R. Determination of Polyfluoroalkyl Compounds in Water and Suspended Particulate Matter in the River Elbe and North Sea, Germany. Front. Environ. Sci. Eng. China 2009, 3 (2), 152-170. (45) Houtz, E. F.; Sutton, R.; Park, J.-S.; Sedlak, M. Poly- and Perfluoroalkyl Substances in Wastewater: Significance of Unknown Precursors, Manufacturing Shifts, and Likely AFFF Impacts. Water Res. 2016, 95, 142-149. (46) Zhang, C.; Yan, H.; Li, F.; Zhou, Q. Occurrence and Fate of Perfluorinated Acids in Two Wastewater Treatment Plants in Shanghai, China. Environ. Sci. Pollut. Res. 2015, 22 (3), 1804-1811. (47) Bossi, R.; Strand, J.; Sortkjr, O.; Larsen, M. M. Perfluoroalkyl Compounds in Danish Wastewater Treatment Plants and Aquatic Environments. Environ. Int. 2008, 34 (4), 443-450. (48) Ahrens, L. Polyfluoroalkyl Compounds in the Aquatic Environment: A Review of Their Occurrence and Fate. J. Environ. Monit. 2011, 13 (1), 20-31. (49) Exner, M.; Far ber, H. Perfluorinated Surfactants in Surface and Drinking Waters. Environ. Sci. Pollut. Res. 2006, 13 (5), 299-307. (50) Loos, R.; Gawlik, B. M.; Locoro, G.; Rimaviciute, E.; Contini, S.; Bidoglio, G. EU-Wide Survey of Polar Organic Persistent Pollutants in European River Waters. Environ. Pollut. 2009, 157 (2), 561-568. (51) Mclachlan, M. S.; Holmstrom, K. E.; Reth, M.; Berger, U. Riverine Discharge of Perfluorinated Carboxylates from the European Continent. Environ. Sci. Technol. 2007, 41 (21), 7260-7265. (52) Nguyen, M. A.; Wiberg, K.; Ribeli, E.; Josefsson, S.; Futter, M.; Gustavsson, J.; Ahrens, L. Spatial Distribution and Source Tracing of per- and Polyfluoroalkyl Substances (PFASs) in Surface Water in Northern Europe. Environ. Pollut. 2017, 220, 1438-1446. (53) Moody, C. A.; Field, J. A. Perfluorinated Surfactants and the Environmental Implications of Their Use in Fire-Fighting Foams. Environ. Sci. Technol. 2000, 34 (18), 3864-3870. (54) Loos, R.; Locoro, G.; Comero, S.; Contini, S.; Schwesig, D.; Werres, F.; Balsaa, P.; Gans, O.; Weiss, S.; Blaha, L.; Bolchi, M.; Gawlik, B. M. Pan-European Survey on the Occurrence of Selected Polar Organic Persistent Pollutants in Ground Water. Water Res. 2010, 44 (14), 4115-4126. (55) Munoz, G.; Labadie, P.; Botta, F.; Lestremau, F.; Lopez, B.; Geneste, E.; Pardon, P.; Devier, M.-H.; Budzinski, H. Occurrence Survey and Spatial Distribution of Perfluoroalkyl and Polyfluoroalkyl Surfactants in Groundwater, Surface Water, and Sediments from Tropical Environments. Sci. Total Environ. 2017, 607-608, 243-252. (56) Lopez, B.; Ollivier, P.; Togola, A.; Baran, N.; Ghestem, J.-P. Screening of French Groundwater for Regulated and Emerging Contaminants. Sci. Total Environ. 2015, 518-519, 562-573. (57) Saito, N.; Harada, K.; Inoue, K.; Sasaki, K.; Yoshinaga, T.; Koizumi, A. Perfluorooctanoate and Perfluorooctane Sulfonate Concentrations in Surface Water in Japan. J. Occup. Health 2004, 46 (1), 49-59. (58) Kallenborn, R.; Berger, U.; Jar nberg, U. Perfluorinated Alkylated Substances (PFAS) in the Nordic Environment; Nordic Council of Ministers: Copenhagen, 2004. (59) Higgins, C. P.; Luthy, R. G. Sorption of Perfluorinated Surfactants on Sediments. Environ. Sci. Technol. 2006, 40 (23), 7251- 7256. (60) Anderson, R. H.; Long, G. C.; Porter, R. C.; Anderson, J. K. Occurrence of Select Perfluoroalkyl Substances at U.S. Air Force Aqueous Film-Forming Foam Release Sites Other than Fire-Training Areas: Field-Validation of Critical Fate and Transport Properties. Chemosphere 2016, 150, 678-685. (61) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, Fate and Transport of Perfluorocarboxylates. Environ. Sci. Technol. 2006, 40 (1), 32-44. (62) Plassmann, M. M.; Berger, U. Trace Analytical Methods for Semifluorinated N-Alkanes in Snow, Soil, and Air. Anal. Chem. 2010, 82 (11), 4551-4557. (63) Plassmann, M. M.; Berger, U. Perfluoroalkyl Carboxylic Acids with up to 22 Carbon Atoms in Snow and Soil Samples from a Ski Area. Chemosphere 2013, 91 (6), 832-837. (64) Karrman, A.; Elgh-Dalgren, K.; Lafossas, C.; Moskeland, T. Environmental Levels and Distribution of Structural Isomers of Perfluoroalkyl Acids after Aqueous Fire-Fighting Foam (AFFF) Contamination. Environ. Chem. 2013, 8 (4), 372-380. (65) Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K. R.; Muir, D. C. Biological Monitoring of Polyfluoroalkyl Substances: A Review. Environ. Sci. Technol. 2006, 40, 3463. (66) O'Brien, J. M.; Austin, A. J.; Williams, A.; Yauk, C. L.; Crump, D.; Kennedy, S. W. Technical-Grade Perfluorooctane Sulfonate alters the expression of more transcripts in cultured chicken embryonic hepatocytes than linear Perfluorooctane Sulfonate. Environ. Toxicol. Chem. 2011, 30 (12), 2846-2859. (67) Ding, G.; Peijnenburg, W. J. G. M. Physicochemical Properties and Aquatic Toxicity of Poly- and Perfluorinated Compounds. Crit. Rev. Environ. Sci. Technol. 2013, 43 (6), 598-678. (68) Xia, J.; Niu, C. Acute Toxicity Effects of Perfluorooctane Sulfonate on Sperm Vitality, Kinematics and Fertilization Success in Zebrafish. Chin. J. Oceanol. Limnol. 2017, 35 (4), 723-728. (69) Ji, K.; Kim, Y.; Oh, S.; Ahn, B.; Jo, H.; Choi, K. Toxicity of Perfluorooctane Sulfonic Acid and Perfluorooctanoic Acid on Freshwater Macroinvertebrates (Daphnia Magna and Moina Macrocopa) and Fish (Oryzias Latipes). Environ. Toxicol. Chem. 2008, 27 (10), 2159-2168. 4349 DOI: 10.1021/acs.est.7b05718 Environ. Sci. Technol. 2018, 52, 4340-4349 FLSEVIFP Environmental Pollution 220 (2017) 1438-1446 Contents lists available at ScienceDirect Environmental Pollution journal homepage: www.elsevier.com/locate/envpol ENVIRONMENTAL POLLUTION Spatial distribution and source tracing of per- and polyfluoroalkyl substances (PFASs) in surface water in Northern Europe* Minh A. Nguyen , Karin Wiberg, Erik Ribeli, Sarah Josefsson, Martyn Futter, Jakob Gustaysson, Lutz Ahrens Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences (SLU), Box 7050, SE-75007 Uppsala, Sweden CrossMark ARTICLE INFO Article history: Received 10 June 2016 Received in revised form 21 September 2016 Accepted 29 October 2016 Available online 11 November 2016 Keywords: Per- and polyfluoroalkyl substances (PFASs) Source tracing Long-range atmospheric transport Surface water ABSTRACT The impact of point and diffuse sources for 26 per- and polyfluoroalkyl substances (PFASs) in northern Europe were investigated by studying Swedish rivers (n = 40) and recipient seawater (Baltic Sea and Kattegat; n = 18). Different composition profiles were observed in the rivers, with ten rivers having a remarkably high fraction of perfluoroalkane sulfonic acids (PFSAs; 65% of the EPFASs) as compared to other rivers (19%) suggesting major impact of one or several source types dominated by PFSAs. Population density and low latitude (south) were strongly correlated to the widely used perfluorooctanesulfonic acid (PFOS), perfluorooctanoic acid (PFOA) as well as to perfluorohexanesulfonic acid (PFHxS). Significant relationships between several PFCAs and PFSAs (i.e. perfluorobutanoic acid (PFBA), perfluoroheptanoic acid (PFHpA), PFOA, perfluorobutanesulfonic acid (PFBS), and PFHxS) and dissolved organic carbon (DOC) were detected (p < 0.05), indicating chemical binding and co-transport with DOC in fresh water and seawater. Partial least squares regression analysis showed that perfluoroalkyl carboxylic acids (PFCAs) were related to latitude according to their perfluorocarbon chain length (C3, C7, C8, Cg, C10 and C11), with longer chains associated with higher latitudes. This suggests the presence of mechanisms promoting higher prevalence of longer chained PFCAs in the north, e.g. precursor degradation, and/or aerosol associated stabilization of PFCAs and their precursors. 2016 Elsevier Ltd. All rights reserved. 1. Introduction Per- and polyfluoroalkyl substances (PFASs) are synthetic chemicals of concern due to their extreme persistence and high bioaccumulative potential (Giesy et al., 2010; Martin et al., 2003). Some PFAS5 are potentially carcinogenic and endocrine disrupting chemicals (EDCs), and their toxicity depends on functional group and perfluorocarbon chain length (Ulhaq et al., 2013a; Ulhaq et al., 2013b). PFAS5 have both hydrophobic and hydrophilic properties and are used as surfactants in e.g. textiles, carpets, leather, paper products, and fire-fighting foams (Ahrens et al., 2009a; Swedish EPA, 2006). They can be emitted from point sources (e.g. fire- fighting training sites, sewage treatment plants (STPs)) (Ahrens et al., 2015; Becker et al., 2008) or derive from diffuse sources (e.g. atmospheric deposition, run-off) (Kim and Kannan, 2007; * This paper has been recommended for acceptance by von Hippel Frank A. * Corresponding author. E-mail address: Bslu.se (M.A. Nguyen). http://dx.doi.org/10.1016/j.envpol.2016.10.089 0269-7491/O 2016 Elsevier Ltd. All rights reserved. Taniyasu et al., 2013), and can undergo long-range transport to remote areas through atmosphere and water currents (Ahrens et al., 2009a; Ahrens et al., 2015; Becker et al., 2008; Davis et al., 2007; Paul et al., 2009; Prevedouros et al., 2006; Shoeib et al., 2006). As a consequence, PFASs are ubiquitously distributed in the abiotic environment, wildlife and humans (Giesy and Kannan, 2001; Yamashita et al., 2008; Yeung et al., 2008). Concerns for human health and environmental effects have led to bans or restrictions of the use of some PFASs. For example, PFOS has been listed as a persistent organic pollutant (POP) under the Stockholm Convention since 2009 (UNEP, 2006), and PFOA and PFNA and their salts have been added to the candidate list of substances of very high concern for authorisation by the European chemicals agency (ECHA) since 2013 and 2015 (ECHA, 2013, 2015), respectively. The occurrence of PFAS5 has been studied in rivers and lakes in Europe (Loos et al., 2010; Munoz et al., 2015; Moller et al., 2010) with detected concentrations up to 1400 ng L-1 (Krka River in Slovenia) (Loos et al., 2009). PFASs in surface waters are eventually transported from land to the sea (McLachlan et al., 2007). One of the most contaminated seas globally is the semi-enclosed Baltic Sea M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 1439 (Ahrens et al., 2010; Helsinki Commission, 2010; Filipovic et al., 2013). Source apportionment studies are needed in this area due to the large influence from rivers, a population of 85 million in the catchment area, and conflicting evidence as to the importance of point and diffuse source PFASs inputs (UNEP, 2005). A mass balance study indicates PFAS pollution in the Baltic Sea is mainly the result of diffuse sources (e.g. atmospheric deposition) (Filipovic et al., 2013); however, other studies indicate that point sources (e.g. STP effluents) dominate fluxes to the sea (Bossi et al., 2008; Moller et al., 2010). The aim of this study was to evaluate the impact of point and diffuse sources on PFAS pollution in northern Europe. We studied river and seawater samples collected from 44 Swedish river sites and 18 sites in the receiving Baltic Sea and Kattegat (the latter located at the Swedish west coast) (Fig. 1). The levels and composition profiles of 26 PFASs were used to assess their spatial trends and fluxes, and their correlations with water chemistry and demographic, geographic and hydrological factors including catchment characteristics (size, population, population density, and run off) and sampling site characteristics (latitude, water flow, levels of dissolved organic carbon (DOC), salinity, and water temperature). 2. Materials and methods 2.1. Sampling The riverine sampling sites were located along the Swedish coast, close to the mouths of rivers that contribute most of the water discharge to the Baltic Sea and Kattegat (Brandt et al., 2008). The sites correspond to sites in the National Environmental Monitoring and Assessment Program for freshwater in Sweden (Folster et al., 2014). The samples were collected at a depth of approximately 0.4 m in October 2013 (Fig. 1; R1 to R40). In addition to the 40 river mouth samples, 4 upstream river sites were also included (Vindeln River e a tributary to Ume River (R9B and R9C), Ume River at Gubbole (R9A) and Gota River near Trollhattan (R37A)). Duplicate samples were taken at sites R4, R7, R15, R17, and R23. Seawater samples were collected in the Baltic Sea during a cargo ship tour (TransPaper Cargo Ship) in AugusteSeptember 2013 at 18 off-shore sites (Fig. 1, S1 to S18) in the Bothnian Bay (n 5, north), the Bothnian Sea (n 3, central), the Baltic Proper (n 6, south), and the west coast of Sweden (the Kattegat, n 4). Duplicate samples were taken at sites S6 and S13. The seawater samples were collected at approximately 2 m depth. Samples were collected in 1 L polypropylene (PP) bottles prerinsed with methanol. In the field, the sampling bottles were rinsed three times with river/seawater prior to sampling. Directly after sampling, the bottles were stored dark and cool until analysis. All samples were analyzed within 30 days of the sampling. A separate set of field samples were collected for suspended particulate matter (SPM) determination. Water temperature for riverine sites was measured or collected from the SLU database (Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, 2016) (Table S1, Supporting Information). Seawater temperature, pH, and salinity were measured at all marine sampling sites (Tables S2, Supporting Information). DOC and other data (water discharge, catchment size, and total population in catchment) for rivers water and catchments were collected from the SLU database (Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, 2016) and from the Swedish Meteorological and Hydrological Institute (2016) (Table S1, Supporting Information), while DOC in sea samples was measured in the laboratory. Fig. 1. The river (R1-R40), Baltic Sea (S1-S14), and Kattegat (S15-S18) sampling sites. Black dots (C) represent river sampling sites and blue triangles ( ) sea sampling sites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 2.2. Chemicals Target analytes included perfluoroalkyl carboxylic acids (PFCAs) with perfluoroalkyl carbon chain length of C3-C13, C15, and C17 (i.e. PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTriDA, PFTeDA, PFHxDA, and PFOcDA), C4, C6, C8, and C10 perfluoroalkane sulfonic acids (PFSAs) (i.e. PFBS, PFHxS, PFOS, and PFDS), 6:2 fluorotelomer sulfonate (6:2 FTSA), perfluorooctanesulfonamide (FOSA), N-methyl and N-ethyl FOSAs (MeFOSA and EtFOSA), 2-(perfluorooctanesulfonamido) acetic acid (FOSAA), methyl and ethyl FOSAAs (MeFOSAA and EtFOSAA), and N-methyl and N-ethyl perfluorooctane sulfonamido ethanol (MeFOSE and EtFOSE) (compilation of all target analytes and their log KOW values can be found in Table S3, Supporting Information). The mass-labelled internal standard (IS) mix included [13C4]-PFBA, [13C2]-PFHxA, [13C4]-PFOA, [13C5]-PFNA, [13C2]-PFDA, [13C2]PFUnDA, [13C2]-PFDoDA, [18O2]-PFHxS, [13C4]-PFOS, [13C8]-FOSA, [d3]-N-MeFOSA, [d5]-N-EtFOSA, [d3]-N-MeFOSAA, [d5]-N-EtFOSAA, [d7]-N-MeFOSE, and [d9]-N-EtFOSE, and the injection standard (InjS) was [13C8]-PFOA. All reference and mass-labelled compounds were purchased from Wellington Laboratories, Ontario, Canada. 1440 M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 2.3. Sample preparation and instrumental analysis Water samples were filtered using pre-weighed glass microfiber filters (Whatman GF/F; diameter 47 mm, 0.7 mm pore size) to differentiate between the aqueous and the particulate phase of PFASs. Prior to filtration, the glass equipment and the GF/Fs were heated at 400 C overnight. The GF/Fs were stored in the freezer (20 C) until extraction. Separate samples for SPM determination were filtered using the same type of filters, which were then equilibrated in a desiccator and measured gravimetrically. Prior to extraction, each GF/F filter and aqueous phase sample was spiked with 100 mL of the IS mix (20 pg mL1). For the aqueous phase, solid phase extraction (SPE) was performed as described by Ahrens et al. (2009b) with some modification. Briefly, 1 L of water was extracted using pre-conditioned Oasis weak anion exchange (WAX) cartridges (6 mL, 500 mg, 60 mm). The sample was washed with 4 mL of 25 mM ammonium acetate buffer (pH 4) in Millipore water and then eluted with 4 mL of methanol, followed by 4 mL of methanol with 0.1% ammonium hydroxide. The extracts were combined in pre-rinsed (methanol) 15 mL PP tubes, concentrated to 1 mL and transferred to 1.5 mL glass vials. For the particulate phase, GF/F filters were extracted in pre-rinsed (methanol) 15 mL PP tubes. The first extraction was performed by shaking at 200 rpm with 6 mL of methanol for 30 min. The second and third extractions were carried out by shaking at 200 rpm with 4 mL of methanol for 15 min each time. The combined extracts were concentrated to 1 mL and transferred to 1.5 mL glass vials. Pre-conditioned Oasis weak anion exchange cartridges (n 5) and pre-heated GF/F filters (n 6) were used as blanks. Millipore water could not be used for blanks due to elevated PFAS levels in the municipal drinking water. Before instrumental analysis, each of the extracts was spiked with 10 mL of InjS (200 pg mL1). In total, 26 PFASs were analyzed using high- performance liquid chromatography coupled to tandem mass spectrometry (HPLC-MS/MS; Agilent Technologies LC 1200 series coupled to 6460 Triple Quad system) in accordance with the method described by Ahrens et al. (2009a). The isotope dilution method was applied for the quantification of PFASs using a fivepoint calibration curve ranging from 0.05 to 40 ng mL1 using Agilent QQQ Mass Hunter software. 2.4. Statistical analysis Prior to statistical analysis, PFAS concentrations below the method detection limit (MDL) were replaced by MDL/2. Since some PFASs were not detected in all samples, individual PFASs with a detection frequency <50% were excluded from the calculations. The significance level was set to a 0.05. Similarity between rivers was investigated by cluster analysis based on PFAS compositional profiles (Past 3.10, (Hammer et al., 2001)). The paired group (UPGMA) clustering algorithm and Chord similarity index was chosen for evaluating the similarity of PFAS composition profiles in all rivers. Pearson correlation was used for assessing relationships between the different PFASs and for correlation of individual PFASs with DOC. In these analyses, the PFAS concentrations were log 10 transformed to reduce the skewness of data (Wiberg et al., 2002). Relations between PFAS levels and water chemistry, demography, geography, and hydrology were investigated by principal component analysis (PCA) and partial least squares regression (PLS) (SIMCA; MKS Umetrics AB). PLS was used to investigate the relations between predictive variables (X e PFASs concentrations) and objects (sampling sites), and to study the relation between predictive (X) and response variables (Y e water chemistry, demography, geography and hydrology). PCA was applied to get an overview of overall trends and patterns among all objects (sampling sites) and variables that were defined in both rivers and sea, i.e. PFASs levels, water chemistry and geographical characteristics. In the PCA and PLS analyses, PFAS concentrations as well as total population, runoff, population density, and salinity were log 10 transformed. Data were scaled to unit variance (UV), and following Wiberg et al. (2002), relationships were considered only if the correlation factor R2Y was >0.67 (PLS) and R2X > 0.9 (PCA). 2.5. Quality assurance and quality control The MDLs were calculated from the blanks (average of blanks 3 * standard deviation (s)). The samples were not blank corrected. If the PFASs were not detected in the solvent blank, the lowest calibration curve point was used as MDL. The blank concentrations ranged between non-detected and 0.12 ng L1 for the aqueous phase samples and were always non-detected for the particulate phase (GF/F) samples (Tables S4 and S5, Supporting Information). The MDLs ranged from 0.03 to 0.76 ng L1 for the aqueous phase and was 0.03 ng L1 for the particulate phase (GF/F) (Table S4 and S5, Supporting Information). Average recoveries of the ISs for the different PFASs ranged from 56 to 111% with standard deviations of 20e31% (Table S6, Supporting Information). Duplicate samples showed average standard deviations of 36% for the river samples and 44% for the marine samples (Table S7 and S8, Supporting Information). However, separate samples were taken one after each other, and a possible source of variation for the duplicate samples could arise from the difference in sampling time and depth between the collected samples. Information on blank concentrations in aqueous phase samples and particulate phase samples, recovery of ISs, and standard deviation of duplicate samples in river and sea samples can be found in Tables S4S8, Supporting Information. 3. Results and discussion 3.1. Spatial trends and composition profiles of PFASs in Swedish rivers and recipient seas Out of the 26 analyzed PFASs, 13 and 12 PFASs were detected in the river and seawater samples, respectively (Fig. 2, Table 1, Tables S9 and S10, Supporting Information). The contribution of PFASs in the particulate phase in comparison to the aqueous phase was insignificant (<1%) both in river and sea samples (Tables S11 and S12, Supporting Information). Therefore, data evaluation was sPolPeFlAy SbacsoendcoennttrhaetionPFsAinS concentrations in the aqueous the rivers ranged from 1.0 to phase. 60 ng The L1 (average 9.9 15 ng L1, median 4.0 ng L1; Fig. 2 and Table S9, Supporting Information). PPFAS concentrations in the recipient seas were approximately a factor of two lower than in the river samples, ranging from 1.2 to 14 ng L1 (average 4.5 3.3 ng L1, median 3.5 ng L1; Fig. 2 and Table S10, Supporting Information), athnedrciovnercsenwteraretioPnFBs Sva(2ri1e%d olefstshtehP anPiFnArSivs)erfos.llPorwedeodmbiynPaFnHt PxFSA(S1s8%in) and PFBA (14%). In contrast, the long chained PFASs (C7 for PFCAs and C6 for PFSAs) were predominant in the Baltic Sea and Kattegat with PFNA contributing 34%, followed by PFOS (19%) and PFOA (13%). One possible explanation for the higher fraction of PFNA in the sea might be that the contribution of PFNA (and other longer chain PFCAs) is high in atmospheric deposition (Filipovic et al., 2013), while in the terrestrial environment, PFNA and other longer chain substances are retained and do not readily reach the rivers. Long chained PFASs (e.g. PFOS and PFOA) have largely been replaced by short chained PFASs and polyfluoroalkyl products of equivalent length in industrial and consumer products since 2002 M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 1441 Fig. 2. Levels (ng L1) and composition profiles of individual PFASs in (A) 44 Swedish river samples (representing 40 rivers) from north to south along the Swedish east coast (R1 R30), and then from south to north along the Swedish west coast (R31R40) and (B) the Bothnian Bay (S1 (north) to S5 (south)), the Bothnian Sea (S6 (north) to S8 (south)), the Baltic Proper (S9 (north) to S14 (south)) and the Kattegat (S15 (south) to S18 (north)). (Moller et al., 2010; Wang et al., 2013b). In addition, shorter chained PFASs have higher water solubility and reach water bodies more readily than longer chained PFASs (Ahrens et al., 2015). This may explain the contrasting dominance of the short chained PFASs (i.e. PFBS, PFHxS and PFBA) in the river samples as compared to the sea samples observed in our study. The residence time of the seawater in Baltic Sea is around 30 years (Helsinki Commission, 2010), and consequently historical sources will be reflected there for decades, in contrast to river water that will respond to changes in source patterns more quickly. 1442 M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 Table 1 Detection frequency (%), concentration range (ng L1) and average concentrations (in brackets) of detected PFASs in river (n 40) and seawater samples (n 18). PFAS Rivers Detection frequency (%) Concentration range (ng L1) (Average) Recipient sea Detection frequency (%) Concentration range (ng L1) (Average) PFBA 100 PFHxA 27 PFHpA 48 PFOA 82 PFNA 100 PFDA 80 PFUnDA 86 PFDoDA 50 PFTeDA 23 PFBS 59 PFHxS 77 PFOS 98 FOSAA 0 FOSA 59 a NA not available 0.47e3.7 (2.1) 0.51e4.2 (2.4) 0.36e1.7 (1.0) 0.21e4.2 (2.2) 0.090e5.8 (2.9) 0.024e4.4 (2.2) 0.018e1.8 (0.91) 0.016e0.82 (0.42) 0.093e1.5 (0.80) 0.030e19 (9.5) 0.051e18 (9.0) 0.040e6.9 (3.5) NAa 0.032e0.46 (0.25) 100 0.34e0.67 (0.51) 0 NAa 30 0.61e1.0 (0.81) 100 0.21e1.3 (0.76) 100 0.14e5.7 (2.9) 40 0.045e0.83 (0.44) 0 NAa 5 0.045 40 0.016e0.072 (0.044) 65 0.062e0.57 (0.32) 100 0.11e1.7 (0.91) 100 0.11e2.5 (1.3) 5 0.061 75 0.019e0.051 (0.035) PFOS is listed as a priority substance in the European Union Water Framework Directive (WFD) with an annual average environmental quality standard (AA-EQS) of 0.65 ng L1 for inland surface waters and 0.13 ng L1 for seawater (EU, 2013; Weiss et al., 2015). The results from the current study showed that the AA-EQS value was exceeded at 12 of 44 river sites and 17 of 18 Baltic Sea and Kattegat sites (Figs. S2 and S3, Supporting Information). The spatial distribution of SPFASs showed that northern rivers (located on the east coast of Sweden) had generally lower levels of PFASs (R1R12, average 6.5 ng L1, median 1.8 ng L1) than rivers on the south-east and south-west coast (East: R22R30, average 13 ng L1, median 5.7 ng L1 and West: R31R40, average 5.9 ng L1, median 4.2 ng L1, respectively) (Fig. 2). A similar spatial trend was observed in the recipient sea, where levels of PFASs were generally lower in the northern sub-basins of the Baltic Sea than in Baltic Proper and the Kattegat (Fig. 2). This may be explained by higher riverine input of PFASs in the south with its higher population density compared to the north. The concentrations of PFASs in Swedish rivers (average SPFASs 9.9 ng L1, median 4.0 ng L1; n 40) were comparable with those in the Ganges in India (n 14; average SPFASs 8.9 ng L1, median 7.7 ng L1) (Sharma et al., 2016), but were generally lower than levels in other studied rivers in the world, such as SPFASs concentrations in rivers of Hyogo prefecture in Japan (n 41; average SPFASs 431 ng L1, median 12 ng L1) (Takemine et al., 2014), the Rhine in France, Germany and Switzerland (n 38; average SPFASs 36 ng L1, median 26 ng L1) (Moller et al., 2010), and the Han River in China (n 23) (average SPFASs 204 ng L1, median 180 ng L1) (Wang et al., 2013a). Composition profiles showed that the northern rivers (R1-R12) generally had higher fractions of long perfluorocarbon chained PFCAs (i.e. C8-C11) than rivers in the south (R22eR30) and the west (R31-R40), where higher fractions of C5-C7 PFCAs were observed (Fig. 2). This might indicate a stronger relative influence of atmospheric deposition in the north due to long-range transport of long chained PFCAs and their precursors (Shoeib et al., 2006; Young et al., 2007), in contrast to a stronger relative influence of contemporary point sources in the south due to the replacement of C8-based PFASs by shorter chained PFASs (Moller et al., 2010). A distinct composition profile was observed in ten of the rivers (R6, R7, R9A, R13, R16, R18, R20, R25, R30, and R32, see Table S1, Supporting Information), with PFBS (29% of the PFASs) and PFHxS (25%) as the dominant compounds, which can be compared to the other sampling sites showing dominance of PFBA (36%), and PFOA (14%). Besides a different composition profile, the PFASs concentrations were on average 7.5 times higher (30 ng L1) in these ten river sites as compared to the other rivers (average of 3.9 ng L1). The PFSA and PFCA concentrations showed a significant linear relationship for the ten river sites (p < 0.0005) with a slope of 1.8 (i.e. PFSAs dominant), whereas the slope for the other river sites was only 0.34 (i.e. PFCAs dominant; p < 0.00001) (Fig. 3A). The ten rivers with high fraction of PFSA grouped together in the cluster analysis (Fig. 3B), indicating impact of similar kind of point source(s) such as STP effluents, landfills or firefighting training areas (Ahrens et al., 2009b; Ahrens et al., 2015; Becker et al., 2008; Busch et al., 2010; Moller et al., 2010). In addition, among these ten rivers, four out of the five sites with the highest SPFAS levels (from 27 to 60 ng L1; R9A, R13, R16, and R25) showed very close resemblance in the PFAS composition (Fig. 2) and made up a subgroup in the cluster analysis (Fig. 3B), possibly impacted by a common source type. Based on information from a recent inventory of PFAS sources a nationwide compilation comprising >2000 potential PFAS sources and their locations (Swedish EPA, 2016), a map including potentially significant PFAS sources for the ten rivers with elevated concentrations was compiled (Fig. S1, Supporting Information). However, due to the large number of potential sources near the studied rivers, it was not possible to identify the main contamination source(s). 3.2. PFAS fluxes from Swedish rivers into the Baltic Sea and the Kattegat Estimated riverine fluxes of PFASs, calculated by the PFAS concentrations and river water discharges (data from Swedish Meteorological and Hydrological Institute (2016)), are shown in Table S13 (Supporting Information). In these calculations, upstream sampling sites (i.e. R9A, R9B, R9C, and R37A) were not included. Considering that concentration values are based on one grab sampling, the calculation of the fluxes is a cautious estimate assuming similar PFAS concentrations over the year, and not varying with river discharge. The total riverine input of SPFASs from the investigated Swedish rivers into the Kattegat (n 10) was 0.50 kg day1 (170 kg yr1), and the corresponding value for the Baltic Sea (n 30) was 2.8 kg day1 (1000 kg yr1), with R13 (420 kg yr1), R37 (83 kg yr1), and R22 (78 kg yr1) as major choignhtrP ibuPtFiAnSg rivers (Table S11, Supporting Information). Rivers concentrations (e.g. R16 with 60 ng L1 and R25 with with 52 ng L1), did not always show high fluxes (30 and 55 kg yr1, respectively) due to their low water discharge (16 and 34 m3 s1, respectively) (Table S11, Supporting Information). M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 1443 Fig. 3. (A) Rivers with predominantly PFSA (above 4.5 ng L1 in red color) and less-predominantly PFSA concentrations (below 4.5 ng L1 in blue color) plotted against PFCA concentrations (dotted line represents 1:1 ratio). (B) Cluster analysis based on PFAS concentrations of 10 river sites with predominantly PFSA concentrations (red) and other river sites with less-predominantly PFSA concentrations (blue) based on their concentration levels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Previous estimates of PFAS fluxes into the Baltic Sea were reported to be in the same magnitude as in this study, although different sampling sites and calculation approaches were used. The total flux of four PFASs (PFHxA, PFHpA, PFOA and PFNA) into the Baltic Sea was found to be 336 kg yr1 based on measurements and including discharges from six rivers (Latvia (n 1 river; River Daugava), Poland (n 2; the Oder and Vistula rivers), and Sweden (n 3; Kalix, Vindeln and Dal rivers)) (McLachlan et al., 2007). Our corresponding flux for these PFASs, including only Swedish rivers into the Baltic Sea and Kattegat (n 40), was somewhat lower (229 kg yr1) and thus in good agreement. Furthermore, a mass balance modeling study estimated the flux of PFHxA to 50e260 kg yr1, PFOA 396e473 kg yr1 and PFOS 759e793 kg yr1 into the Baltic Sea based on discharges from all surrounding countries (Sweden, Germany, Denmark, Finland, Russia, Poland, Latvia, Estonia, and Lithuania) (Filipovic et al., 2013). The modelled fluxes were higher than our measured/calculated fluxes for PFHxA (15 kg yr1), PFOA (78 kg yr1) and PFOS (105 kg yr1), while the flux of PFDA for Swedish rivers in our study (65 kg yr1) was in the lower end of the modelled fluxes for all Baltic Sea countries (53e207 kg yr1). Considering that our study only included Swedish rivers, the two studies are in good agreement. 3.3. Source tracing PLS models were calculated separately for river and sea samples. The modeling was done in a stepwise manner to determine which environmental variables (Y-variables: latitude, catchment size, catchment population, run off, water flow, DOC, salinity, and water temperature) could be connected to PFAS concentrations (X-variables). For the river model, most environmental factors did not show cumulative correlation factors R2Y (cumulative) > 0.67 (Table S14, Supporting Information), which was the pre-set cutoff for correlation. In the final model, only latitude and population density were included with R2Y (cumulative) of 0.72 and 0.67, respectively (score plots of the PLS for the rivers and recipient sea are shown in Fig. 4A and B). The latitude ranged from 55.780 to 65.853 north, and the population density from 0.24 to 364 individuals km2. The PLS analysis visualizes how the variables relate to each other, and confirm previous conclusions from the composition profile plots (Fig. 2A and B). In the rivers, shorter chained PFCAs (i.e. PFBA and PFOA) were shown to correlate with low latitude (south), which also has a high population density with on average 41 individuals km2 (R22-R40), while long chained PFCAs (i.e. PFNA, PFDA, PFUnDA, PFDoDA) showed the opposite tendency and correlated with high latitude (north), which is a less densely populated area of Sweden with on average 1.3 individuals km2 (R1-R12) (Fig. 4A). The PFCAs were even orderly associated to latitude according to their perfluorocarbon chain length (C3, C7, C8, C9, C10 and C11), with longer chains connected to higher latitude. These results support earlier observations on that PFCAs and their precursors are transported by long range air transport to remote Fig. 4. Partial least squares regressions (PLS) plots of PFAS concentrations (log10-transformed; orange circles) and the correlating environmental factors (log10-transformed population density and salinity; green squares) in (A) river (n 44) and (B) seawater samples (n 18). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 1444 M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 regions as shown from analysis of ice and snow samples from the Arctic (Cai et al., 2011). The reason for the observed spatial distribution pattern, i.e. higher relative abundance of longer chain PFCAs in the north, suggests a number of mechanisms promoting higher prevalence of longer chained PFCAs in the north, e.g. precursor degradation, and aerosol associated stabilization of PFCAs and their precursors in the atmosphere (Ahrens et al., 2012; Arp and Goss, 2009). The gradual sequence should not be over interpreted and directly compared to global distillation in a similar manner as the classical persistent organic pollutants (POPs), as both C3 and C7 (PFBA and PFOA) are associated to high population density. The PFSAs (C4, C6 and C8) did not display a corresponding spatial distribution pattern, but rather the opposite, with longer chains more strongly correlated to southern latitudes. This pattern is more challenging to interpret as few PFSA representatives were targeted (C4, C6, C8, C10) and detected (C4, C6, C8), and as the source strength from contemporary sources for two of them, C6 and C8 (PFHxS and PFOS), likely dominate largely over secondary sources (such as long-range air transport) in most areas. In agreement with Murakami et al. (2008) and Pistocchi and Loos (2009), we observed strong positive correlations between PFOS, PFOA and population density (south; Fig. 4A). In addition, we found a similar correlation for PFHxS and PFBA. For the recipient seas, all factors including DOC, latitude, salinity and water temperature showed a high cumulative correlation factor (R2Y (cumulative) > 0.67; Fig. 4B; Table S15, Supporting Information). Among these, latitude, salinity, and water temperature were closely related to each other, while inversely associated with DOC. PFOA, PFNA, PFBS, PFHxS, and PFOS clustered and were associated with low latitude. The higher concentrations in southern basins likely reflect a higher discharge of historically commonly used PFASs into the sea in these areas due to higher emissions in densely populated areas (Table S13, Supporting Information). There was low variation in the concentrations of the different sea sites (Fig. 2B), with the exception of FOSA. This might be related to the fact that FOSA originates from the degradation of volatile precursors (Martin et al., 2006; Young and Mabury, 2010) or is degraded to, for example PFOS (Rhoads et al., 2008). On the other hand, FOSA was correlated with the PFOA, PFHxS, and PFOS in the rivers (Fig. 4A), which may indicate common sources. A PCA model was calculated to be able to investigate overall trends and patterns among all objects (sampling sites) and all variables that were defined in both river and sea samples (variables: PFBA, PFOA, PFNA, PFBS, PFHxS, PFOS, FOSA, DOC and latitude) (Fig. S4, Supporting Information). PC1 explained 47% and PC2 22% of the variance (biplot in Fig. S4, Supporting Information). Again, it was shown that PFOS, PFOA and PFHxS were correlated with low latitude (south) (Fig. S4, Supporting Information) as shown in PLS models (Fig. 4A and B), and that FOSA had a deviating pattern. The river and sea sites were separated from each other, and this was driven by high DOC concentrations in the rivers. Additionally, now with all objects included, the PCA showed that PFBA was separated from all other PFASs (Fig. S4, Supporting Information) and was significantly associated with DOC (R2 0.49, p < 0.00001, log10-transformed data; Fig. 5). Using linear regression, several other shorter chain PFCAs and PFSAs, including PFHpA, PFOA, PFBS, and PFHxS, showed significant positive correlations to DOC (p < 0.05, log10 transformed data, Fig. 5 and Table S16, Supporting Information). This indicates that PFASs may bind to and be co-transported by DOC in water. PFCAs and PFSAs are negatively charged in natural waters by proton dissociation of the acid (Vierke et al., 2013). As the correlation was observed only for the shorter (i.e., more hydrophilic substances), one possible explanation for the observed phenomenon could be that these PFASs readily bind to positively charged ions that are complex bound to DOC (e.g. Ca2) (Du et al., 2014; Higgins and Luthy, 2006), while the longer chained PFASs rather partition to even more hydrophobic phases in the water, such as the organic carbon fraction of suspended particulate matter. However, further investigations on the sorption mechanisms and the possible influence of the character of DOC on the binding strength are needed to understand the observed relationship between PFBA, PFHpA, PFOA, PFBS, Fig. 5. Log DOC plotted against A) log PFBA, B) PFOA, C) PFBS, and D) PFHxS in rivers (n 40 for PFBA, n 33 for PFOA, n 24 for PFBS, and n 30 for PFHxS) and Baltic Sea and the Kattegat samples (n 18 for PFBA, PFOA, PFOS and n 12 for PFBS). M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 1445 PFHxS, and DOC. 4. Conclusions obsDerisvteindcitncroivmerpsowsititiohnhipgrhofiPlePsFAwSitchonhcigehntfrraatciotinosnaosfcPoFmSpAasrwedetroe other rivers suggesting source specific patterns of PFASs. This finding may be used for source tracing (e.g. fire-fighting training sites, effluents from sewage treatment plants and atmospheric deposition) of PFASs in the rivers; however, sampling along the course of the rivers is needed due to a multitude of potential source sites. Significant relationship between PFBA, PFHpA, PFOA, PFBS, PFHxS and DOC was detected (p < 0.05) for rivers and recipient seas. This may indicate chemical binding and co-transport of those PFASs with DOC in fresh water and sea water. Levels of PFCAs in rivers were related to the latitude of the sampling sites according to their perfluorocarbon chain length, with longer chains associated with higher latitudes. A number of possible mechanisms promoting a higher prevalence of longer chained PFCAs in the north were suggested. Further investigations on mechanisms driving the longrange atmospheric transport of long chain PFCAs and the spatial distribution of PFCAs observed are needed. Acknowledgement This work was funded by the Swedish Environmental Protection Agency (contract nr. 2213-13-026), the Oscar and Lili Lamm's Foundation (contract nr. DO2011-0033), and the Swedish Research Council Formas (contract nr. 216-2011-427). We thank Elin Lavonen (SLU), Matyas Ripszam (Ume University), the Captain and crew of the TransPaper Cargo ship, Christian Demandt (SLU), and all members working in the Swedish National Environmental Monitoring and Assessment Program for assistance in the sample collection. We are grateful to Professor Paul Geladi (SLU) for statistical evaluation support. We also would like to thank Salar Valinia (SLU) for helping with GIS software and Wiebke Drig (SLU), Rikard Troger (SLU), and Joachim Audet (SLU) for valuable discussions. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2016.10.089. References Ahrens, L., Barber, J.L., Xie, Z., Ebinghaus, R., 2009a. Longitudinal and latitudinal distribution of perfluoroalkyl compounds in the surface water of the Atlantic ocean. Environ. Sci. Technol. 43, 3122e3127. Ahrens, L., Felizeter, S., Sturm, R., Xie, Z., Ebinghaus, R., 2009b. Polyfluorinated compounds in waste water treatment plant effluents and surface waters along the River Elbe, Germany. Mar. Pollut. Bull. 58, 1326e1333. Ahrens, L., Gerwinski, W., Theobald, N., Ebinghaus, R., 2010. Sources of polyfluoroalkyl compounds in the North sSea Baltic sea and Norwegian sea: evidence from their spatial distribution in surface water. Mar. Pollut. Bull. 60, 255e260. Ahrens, L., Harner, T., Shoib, M., Lane, D.A., Murphy, J.G., 2012. Improved characterization of GasParticle partitioning for per- and polyfluoroalkyl substances in the atmosphere using annular diffusion denuder samplers. Environ. Sci. Technol. 46, 7199e7206. Ahrens, L., Norstrom, K., Viktor, T., Cousins, A.P., Josefsson, S., 2015. Stockholm Arlanda Airport as a source of per- and polyfluoroalkyl substances to water, sediment and fish. Chemosphere 129, 33e38. Arp, H.P.H., Goss, K.-U., 2009. Gas/Particle partitioning behavior of perfluorocarboxylic acids with terrestrial aerosols. Environ. Sci. Technol. 43, 8542e8547. Becker, A.M., Gerstmann, S., Frank, H., 2008. Perfluorooctane surfactants in waste waters, the major source of river pollution. Chemosphere 72, 115e121. Bossi, R., Strand, J., Sortkjr, O., Larsen, M.M., 2008. Perfluoroalkyl compounds in Danish wastewater treatment plants and aquatic environments. Environ. Int. 34, 443e450. Brandt, M., Ejhed, H., Rapp, L., May 2008. Naringsbelastningen p O stersjon och Vasterhavet 2006. Swedish Environment Protection Agency Report, pp. 1e43, 5815. Busch, J., Ahrens, L., Sturm, R., Ebinghaus, R., 2010. Polyfluoroalkyl compounds in landfill leachates. Environ. Pollut. 158, 1467e1471. Cai, M., Zhao, Z., Yin, Z., Ahrens, L., Huang, P., Cai, M., Yang, H., He, J., Sturm, R., Ebinghaus, R., Xie, Z., 2011. Occurrence of perfluoroalkyl compounds in surface waters from the North pacific to the Arctic ocean. Environ. Sci. Technol. 46, 661e668. Davis, K.L., Aucoin, M.D., Larsen, B.S., Kaiser, M.A., Hartten, A.S., 2007. Transport of ammonium perfluorooctanoate in environmental media near a fluoropolymer manufacturing facility. Chemosphere 67, 2011e2019. Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, 2016. http://www.slu.se/vatten-miljo/datahantering (accessed 20.09.16.). Du, Z., Deng, S., Bei, Y., Huang, Q., Wang, B., Huang, J., Yu, G., 2014. Adsorption behavior and mechanism of perfluorinated compounds on various adsorbentsda review. J. Hazard. Mater. 274, 443e454. ECHA, 2013. In: Inclusion of Substances of Very High Concern in the Candidate List for Eventual Inclusion in Annex XIV. ED/69/2013, Helsinki. ECHA, 2015. In: Inclusion of Substances of Very High Concern in the Candidate List for Eventual Inclusion in Annex XIV. ED/79/2015, Helsinki. EU, 2013. Directive 2013/39/EU: Priority Substances in the Filed of Water Policy. Filipovic, M., Berger, U., McLachlan, M.S., 2013. Mass balance of perfluoroalkyl acids in the Baltic sea. Environ. Sci. Technol. 47, 4088e4095. Folster, J., Johnson, R.K., Futter, M.N., Wilander, A., 2014. The Swedish monitoring of surface waters: 50 years of adaptive monitoring. Ambio 43 (1), 3e18. Giesy, J.P., Kannan, K., 2001. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 35, 1339e1342. Giesy, J.P., Naile, J.E., Khim, J.S., Jones, P.D., Newsted, J.L., 2010. Aquatic toxicology of perfluorinated chemicals. Rev. Environ. Contam. Toxicol. 1e52. Hammer, ., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4, 9. Helsinki Commission, 2010. Hazardous substances in the Baltic Sea: An Integrated Thematic Assessment of Hazardous substances in the Baltic Sea. Higgins, C.P., Luthy, R.G., 2006. Sorption of perfluorinated surfactants on sediments. Environ. Sci. Technol. 40, 7251e7256. Kim, S.-K., Kannan, K., 2007. Perfluorinated acids in air, rain, snow, surface runoff, and lakes: relative importance of pathways to contamination of urban lakes. Environ. Sci. Technol. 41, 8328e8334. Loos, R., Gawlik, B.M., Locoro, G., Rimaviciute, E., Contini, S., Bidoglio, G., 2009. EUwide survey of polar organic persistent pollutants in European river waters. Environ. Pollut. 157, 561e568. Loos, R., Locoro, G., Contini, S., 2010. Occurrence of polar organic contaminants in the dissolved water phase of the Danube River and its major tributaries using SPE-LC-MS2 analysis. Water Res. 44, 2325e2335. Martin, J.W., Mabury, S.A., Solomon, K.R., Muir, D.C.G., 2003. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 22, 196e204. Martin, J.W., Ellis, D.A., Mabury, S.A., 2006. Atmospheric chemistry of perfluoroalkanesulfonamides: kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-Ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 40, 864e872. McLachlan, M.S., Holmstrom, K.E., Reth, M., Berger, U., 2007. Riverine discharge of perfluorinated carboxylates from the european continent. Environ. Sci. Technol. 41, 7260e7265. Moller, A., Ahrens, L., Surma, R., Westerveld, J., Wielen, F.v.d., Ebinghaus, R., Voogt, P.d, 2010. Distribution and sources of polyfluoroalkyl substances (PFAS) in the River Rhine watershed. Environ. Pollut. 158, 3243e3250. Munoz, G., Giraudel, J.-L., Botta, F., Lestremau, F., Devier, M.-H., Budzinski, H., Labadie, P., 2015. Spatial distribution and partitioning behavior of selected polyand perfluoroalkyl substances in freshwater ecosystems: a French nationwide survey. Sci. Total Environ. 517, 48e56. Murakami, M., Imamura, E., Shinohara, H., Kiri, K., Muramatsu, Y., Harada, A., Takada, H., 2008. Occurrence and sources of perfluorinated surfactants in rivers in Japan. Environ. Sci. Technol. 42, 6566e6572. Paul, A.G., Jones, K.C., Sweetman, A.J., 2009. A first global production, emission, and environmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol. 43, 386e392. Pistocchi, A., Loos, R., 2009. A map of european emissions and concentrations of PFOS and PFOA. Environ. Sci. Technol. 43, 9237e9244. Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 40, 32e44. Rhoads, K.R., Janssen, E.M.-L., Luthy, R.G., Criddle, C.S., 2008. Aerobic biotransformation and fate of N-Ethyl perfluorooctane sulfonamidoethanol (N-EtFOSE) in activated sludge. Environ. Sci. Technol. 42, 2873e2878. Sharma, B.M., Bharat, G.K., Tayal, S., Larssen, T., Becanova, J., Karaskova, P., Whitehead, P.G., Futter, M.N., Butterfield, D., Nizzetto, L., 2016. Perfluoroalkyl substances (PFAS) in river and ground/drinking water of the Ganges River basin: emissions and implications for human exposure. Environ. Pollut. 208, 704e713. Shoeib, M., Harner, T., Vlahos, P., 2006. Perfluorinated chemicals in the arctic atmosphere. Environ. Sci. Technol. 40. Swedish EPA, 2006. Perfluorinated Substances and Their Uses in Sweden. 1446 M.A. Nguyen et al. / Environmental Pollution 220 (2017) 1438e1446 Swedish EPA, 2016. Hogfluorerade Amnen (PFAS) Och Bekampningsmedel - En Sammantagen Bild Av Forekomsten I Miljon Redovisning Av Ett Regeringsuppdrag. Swedish Meteorological and Hydrological Institute, 2016. http://vattenwebb.smhi. se/ (accessed 20.09.16.). Takemine, S., Matsumura, C., Yamamoto, K., Suzuki, M., Tsurukawa, M., Imaishi, H., Nakano, T., Kondo, A., 2014. Discharge of perfluorinated compounds from rivers and their influence on the coastal seas of Hyogo prefecture. Jpn. Environ. Pollut. 184, 397e404. Taniyasu, S., Yamashita, N., Moon, H.-B., Kwok, K.Y., Lam, P.K.S., Horii, Y., Petrick, G., Kannan, K., 2013. Does wet precipitation represent local and regional atmospheric transportation by perfluorinated alkyl substances? Environ. Int. 55, 25e32. Ulhaq, M., Carlsson, G., O rn, S., Norrgren, L., 2013a. Comparison of developmental toxicity of seven perfluoroalkyl acids to zebrafish embryos. Environ. Toxicol. Pharmacol. 36, 423e426. Ulhaq, M., O rn, S., Carlsson, G., Morrison, D.A., Norrgren, L., 2013b. Locomotor behavior in zebrafish (Danio rerio) larvae exposed to perfluoroalkyl acids. Aquat. Toxicol. 144e145, 332e340. UNEP, 2005. Global International Waters Assessment Baltic Sea, GIWA Regional Assessment 17. UNEP, 2006. Perfluorooctane Sulfonate e Risk Profile. Vierke, L., Berger, U., Cousins, I.T., 2013. Estimation of the acid dissociation constant of perfluoroalkyl carboxylic acids through an experimental investigation of their water-to-air transport. Environ. Sci. Technol. 47, 11032e11039. Wang, B., Cao, M., Zhu, H., Chen, J., Wang, L., Liu, G., Gu, X., Lu, X., 2013a. Distribution of perfluorinated compounds in surface water from Hanjiang River in Wuhan, China. Chemosphere 93, 468e473. Wang, Z., Cousins, I.T., Scheringer, M., Hungerbhler, K., 2013b. Fluorinated alternatives to long-chain perfluoroalkyl carboxylic acids (PFCAs), perfluoroalkane sulfonic acids (PFSAs) and their potential precursors. Environ. Int. 60, 242e248. Weiss, J., Boer, J.d., Berger, U., Muir, D., Ruan, T., Torre, A., Smedes, F., Vrana, B., Clavien, F., Fiedler, H., 2015. PFAS Analysis in Water for the Global Monitoring Plan of the Stockholm Convention-set-up and Guidelines for Monitoring. Wiberg, K., Bergman, A., Olsson, M., Roos, A., Blomkvist, G., Haglund, P., 2002. Concentrations and enantiomer fractions of organochlorine compounds in Baltic Species hit by reproductive impairment. Environ. Toxicol. Chem. 21, 2542e2551. Yamashita, N., Taniyasu, S., Petrick, G., Wei, S., Gamo, T., Lam, P.K.S., Kannan, K., 2008. Perfluorinated acids as novel chemical tracers of global circulation of ocean waters. Chemosphere 70, 1247e1255. Yeung, L.W.Y., Miyake, Y., Taniyasu, S., Wang, Y., Yu, H., So, M.K., Jiang, G., Wu, Y., Li, J., Giesy, J.P., Yamashita, N., Lam, P.K.S., 2008. Perfluorinated compounds and total and extractable organic fluorine in human blood samples from China. Environ. Sci. Technol. 42, 8140e8145. Young, C.J., Furdui, V.i., Franklin, J., Koerner, R.M., Muir, D.C.G., Mabury, S.A., 2007. Perfluorinated acids in arctic snow: new evidence for atmospheric formation. Environ. Sci. Technol. 41, 3455e3461. Young, C.J., Mabury, S.A., 2010. Atmospheric perfluorinated acid precursors: chemistry, occurrence, and impacts. Rev. Environ. Contam. Toxicol. 208, 1e110.