Article pubs.acs.org/est
Role of Snow Deposition of Perfluoroalkylated Substances at Coastal Livingston Island (Maritime Antarctica) Paulo Casal,† Yifeng Zhang,‡ Jonathan W. Martin,‡ Mariana Pizarro,† Begoña Jiménez,§ and Jordi Dachs*,† †
Institute of Environmental Assessment and Water Research, Spanish National Research Council (IDAEA-CSIC), Barcelona, Catalonia 08034, Spain ‡ Division of Analytical and Environmental Toxicology, Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, Alberta T6G 2G3, Canada § Department of Instrumental Analysis and Environmental Chemistry, Institute of Organic Chemistry, Spanish National Research Council (IQOG-CSIC), Madrid 28006, Spain S Supporting Information *
ABSTRACT: Perfluoroalkyl substances (PFAS) are ubiquitous in the environment, including remote polar regions. To evaluate the role of snow deposition as an input of PFAS to Maritime Antarctica, fresh snow deposition, surface snow, streams from melted snow, coastal seawater, and plankton samples were collected over a three-month period (December 2014−February 2015) at Livingston Island. Local sources of PFASs were significant for perfluoroalkyl sulfonates (PFSAs) and C7−14 perfluoroalkyl carboxylates (PFCAs) in snow but limited to the transited areas of the research station. The concentrations of 14 ionizable PFAS (∑PFAS) in freshly deposited snow (760−3600 pg L−1) were 1 order of magnitude higher than those in background surface snow (82−430 pg L−1). ∑PFAS ranged from 94 to 420 pg L−1 in seawater and from 3.1 to 16 ng gdw−1 in plankton. Ratios of individual PFAS concentrations in freshly deposited snow relative to surface snow (CSD/CSnow), snowmelt (CSD/CSM), and seawater (CSD/CSW) were close to 1 (from 0.44 to 1.4) for all perfluorooctanesulfonate (PFOS) isomers, suggesting that snowfall does not contribute significantly to PFOS in seawater. Conversely, these ratios for PFCAs ranged from 1 to 33 and were positively correlated with the number of carbons in the PFCA alkylated chain. These trends suggest that snow deposition, scavenging sea-salt aerosol bound PFAS, plays a role as a significant input of PFCAs to the Maritime Antarctica.
■
pollutants (POPs).19,20 Research stations and tourism around coastal regions are known localized sources of PFASs in Antarctica, as PFCAs and PFSAs have been reported indoors and nearby Antarctic research stations in a number of matrixes such as dust,20 snow, lakes, and seawater.19 However, the wider relevance of these local sources to the regional environment for the various PFASs, relative to long-range transport and deposition, remains unclear. PFASs have been reported in seawater from oceanic2,13,14 and coastal regions19,27,28 and are noticeably influenced by marine currents.25,26 However, the Antarctic Circumpolar Current acts as a natural barrier encircling Antarctica due to the weak north−south exchange of seawater that is thought to restrict the marine transport of PFASs.29 PFASs can also undergo atmospheric transport and deposition, and for inland areas as well as for neutral PFASs, such as perfluoroalkyl sulfonamidoalcohols (FOSEs), this is arguably the main long-
INTRODUCTION The occurrence of perfluoroalkyl substances (PFASs), such as perfluoroalkyl carboxylates (PFCAs) and perfluoroalkanesulfonates (PFSAs), has raised concern as they are persistent in the environment and bioaccumulate and biomagnify in terrestrial1 and aquatic food webs.2−6 Furthermore, a wide range of toxic effects has been reported for some of the most frequently studied PFASs.7,8 Despite these concerns, PFASs are high production volume chemicals and have many industrial and commercial applications due to their unique oleophobic and hydrophobic properties and high chemical and thermal stabilities.9 Ionizable PFASs, including PFCAs and PFSAs, are ubiquitous in the environment. Their sources are strictly anthropogenic, and environmental concentrations are highest around populated areas,10 such as in rivers.11,12 Nevertheless, PFAS can also be detected in remote regions including the open2,13,14 and deep oceans,15 in high altitude snow,16,17 in pristine terrestrial environments,18 and in polar regions.14,19−24 Despite its geographical isolation, the Antarctic maritime region is not exempt from the influence of anthropogenic chemical hazards, including PFASs and other persistent organic © 2017 American Chemical Society
Received: Revised: Accepted: Published: 8460
May 16, 2017 June 22, 2017 June 30, 2017 June 30, 2017 DOI: 10.1021/acs.est.7b02521 Environ. Sci. Technol. 2017, 51, 8460−8470
Article
Environmental Science & Technology
Figure 1. Sampling sites at Livingston Island. Three near-field and nine far-field surface snow samples were collected in NS1−NS3 and S1−S9, respectively. Four fresh snow deposition samples were collected in SD1−SD4. Three snowmelt samples were collected in SM1−SM3. Seawater and plankton samples were taken at the Raquelias site (62° 39,438′ S, 60° 23,306′ W) and Johnsons site (62° 39,556′ S, 60° 22,132′ W). The Raquelias site is located in the South Bay at 1 km from the coast and the JC1 research station, and Johnsons is located in a small bay only 200 m from the Johnsons glacier. The water column depth is 30 and 15 m at Raquelias and Johnston sampling sites, respectively. Both sampling sites presented a small halocline in the first 3−6 m.
range transport mechanism.21,30 These neutral PFASs may act as precursors of the ionizable PFASs through atmospheric oxidation31 or may be metabolized to ionizable PFASs following deposition to aquatic and terrestrial environments.21,32 Although PFCA and PFSA concentrations in remote snow and soils are clearly dependent on the atmospheric transport and deposition of PFASs, the extent to which atmospheric deposition and oceanic currents influence remote coastal areas remains unclear. Remarkably high concentrations of neutral PFAS have been reported for the Southern Ocean around the Antarctic peninsula,30,33 and these could lead to an introduction of neutral and ionizable PFAS to Antarctic waters through atmospheric deposition. Once PFASs reach coastal seawaters, they are bioavailable to aquatic food webs. Phytoplankton and zooplankton not only are at the base of the marine food web but also influence the marine fate of PFASs through the biological pump.2 Plankton concentrations and bioaccumulation factors (BAFs) have not previously been reported for PFASs in the Southern Ocean or in Maritime Antarctica. The isomeric composition of certain ionizable PFASs, specifically of perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS), may help to identify environmental sources or fate processes.2,22,34 Due to differences in the historical (mixture of linear and branched isomers) and contemporary (exclusively linear) methods of PFOA production, PFOA isomer profiles provide information on the manufacturing origins.2,22,34 Conversely, PFOS has always been produced as a mixture of linear and branched isomers. To the best of our knowledge, PFOS and PFOA isomeric compositions have not been reported in environmental samples from Maritime Antarctica. The aim of the study was to examine the relevance of atmospheric snow deposition as a source of PFASs to Maritime Antarctica. Specific objectives were as follows: (i) to investigate the occurrence of PFCAs and PFSAs in surface snow, snow deposition, snowmelt, seawater, and plankton in a coastal
Antarctic area at Livingston Island (South Shetlands), including plankton bioaccumulation, (ii) to evaluate sources using the PFOS and PFOA isomeric composition in these matrices, and (iii) to assess the transfer of PFCAs and PFSAs from the atmosphere to coastal waters through snow deposition and snowmelt.
■
MATERIALS AND METHODS Site Description and Sampling. Livingston Island is located in the Southern Ocean at 62° 39′ S, 60° 23′ W (Figure 1), as part of the South Shetland Archipelago, Antarctica. The island has an area of 798 km2 and is mostly covered by glaciers, with the highest elevation of 1700 m. The Spanish Antarctic research station Juan Carlos I (JC1) is situated in the North Bay side of the Hurd Peninsula. The only other research station at Livingston Island belongs to Bulgaria (Saint Kliment Ohridski) and is located in the same North Bay. With only two research stations, this area is less populated and less visited by cruise tourism than other islands in the South Shetlands archipelago. Plankton and seawater sampling was conducted from a rigid inflatable boat at two sampling sites: Johnsons and Raquelias (Figure 1). Before sample collection, CTD (Conductivity, Temperature, Depth) profiles were taken to evaluate water temperature, salinity, turbidity, fluorescence, and photosynthetic active radiation (Table S1, SI). Plankton sampling was performed by 3 vertical hauls from the bottom of the sampling site to the surface using a conical plankton nylon net with a 50 μm mesh size. Sampling depth was 14 m at Johnsons and 30 m at Raquelias. The samples were filtered through precombusted and preweighted glass fiber filters (47 mm, GF/ D Whatman). Visible zooplankton (i.e., krill was observed occasionally) was manually removed before filtration. The filters containing plankton samples were wrapped in precombusted aluminum foil and stored at −20 °C in airtight plastic bags until further analysis. Surface seawater samples (0.5 m depth) were collected in 2 L polypropylene bottles and 8461
DOI: 10.1021/acs.est.7b02521 Environ. Sci. Technol. 2017, 51, 8460−8470
Article
Environmental Science & Technology
Separation, identification, and quantification of PFASs was performed by ultrafast liquid chromatography tandem mass spectrometry (UFLC-MS/MS) on a Shimadzu LC-20AD UFLC coupled to a AB Sciex API 5000 triple quadrupole mass spectrometer (Applied Biosystems-MDS Sciex, Concord, ON, Canada) operating in negative ion mode with multiple reaction monitoring. A 10 μL aliquot of each sample was injected onto an Ascentis Express F5 column (2.7 μm, 90 Å, 10 cm × 2.1 mm, Sigma-Aldrich, Canada) maintained at 50 °C. Starting HPLC conditions were 90% water adjusted to pH 4 with acetic acid (A) and 10% pure methanol (B). Initial conditions were held for 1 min, then ramped to 60% B by 2 min, increased to 100% B by 23 min, 100% B by 5 min, and then 10% B and equilibrated for 10 min. Two sets of calibration standards were injected for quantification: one for quantitative analysis of total PFASs (i.e., sum of all isomers for each PFAS) and a second for PFOA and PFOS isomer-specific analysis. The PFAS calibration curve was derived from 9 concentrations ranging from 0.05 to 20 ppb. The calibration curve for the isomeric composition of PFOA and PFOS had 10 points from 0.05 to 40 ppb. Both calibration curves were highly linear for all compounds and isomers. Quality Assurance and Quality Control (QA/QC). Sampling materials (i.e., polypropylene bottles, a stainless steel shovel, and steel trays) were rinsed with methanol previous to sample collection. Procedural and field blanks were analyzed to monitor potential contamination during sampling and sample treatment. In total, 5 field and 4 procedural blanks were analyzed for plankton samples, and 4 field and 4 procedural blanks were analyzed for seawater and snow samples. The limits of quantification (LODs) were defined as the mean concentration of procedural and field blanks plus three times the standard deviation of the blank response (Table S3, SI). In plankton samples, the LOD of perfluorononanoate (PFNA) was in the same range as concentrations in field samples, thus PFNA was not reported for plankton. For analytes not detected in blanks, LOD was determined based on the signal-to-noise ratio of 3.0 (Table S3, SI). Tables S4−S6 (SI) show the frequency detection in the snow, seawater, and plankton samples. Linear and branched PFOS were found in the majority of samples. Only 1m-PFOS was not detected in any sample. Conversely, while the branched PFOA isomers were detected in plankton, these isomers were below LOD in the snow and surface seawater samples. Tables S4 and S5 (SI) show the maximum percentage of branched PFOA in each sample derived from the LOD and the measured linear PFOA concentrations. Triplicate recovery experiments were performed by spiking water and plankton samples with all target compounds and isomers, in addition to the internal standards. Average recoveries ranged from 73 ± 13% to 113 ± 31% for water samples and from 83 ± 34% to 133 ± 17% for plankton samples (Table S3, SI). Statistical Analysis. All statistical analyses were performed with SPSS Statistics version 22.0 (IBM Corp.), and significance was set to p < 0.05. Correlations among concentrations and environmental variables (biomass, water temperature, salinity, turbidity, fluorescence, photosynthetically active radiation (PAR), snow density, and snow age) were performed by Spearman rank-order analysis.
transported to the JC1 station to be processed by solid phase extraction (SPE). Seawater samples were then spiked with 10 pg mass labeled internal standards (Table S2, SI) and extracted using an established SPE method.13,35 Briefly, OASIS WAX cartridges were conditioned with 4 mL of methanol, 4 mL of methanol containing 0.1% ammonia, and 4 mL of chromatographic-grade water. Samples were then loaded, and SPE cartridges were washed with 4 mL of chromatographic-grade water, dried under vacuum, and stored at −20 °C in sealed bags until their elution and further treatment in the laboratory. Surface snow samples were collected with a stainless steel shovel into polypropylene bottles from 3 sampling sites in transited areas within the perimeter of the JC1 research station (near-field snow, NS1−NS3) and 9 far-field sampling sites (0.1−11 km) outside the research station perimeter (snow, S1−S9) (Figure 1). Only the top 2−3 cm of snow was collected. The sampling of surface snow took place between 1 and 6 days after a precipitation event (snow age was calculated as the time elapsed from the last precipitation event and when the snow was sampled). Additionally, four freshly deposited snow samples (SD1−SD4) were collected immediately after four different snowfalls in stainless steel trays previously placed outside the perimeter of the JC1 research station (Figure 1). These four samples integrated the entire snow deposition event. Three natural snowmelt samples (SM1−SM3) were collected in polypropylene bottles from surface streams (Figure 1), unequivocally originating from melting snow on days when ambient temperatures exceeded 0 °C. After collection, snow and snow deposition samples were left to melt at 4−6 °C for 24−48 h at the JC1 station in sealed polypropylene bags and bottles. Snow water samples followed the same SPE extraction method as surface seawater samples. The sample volume was 2 L (water equivalents) for all samples. Chemical Analysis. Plankton and water field blanks were collected during the sampling campaign at the research station and consisted of GF/D filters and OASIS WAX cartridges that followed the same filtration/elution procedures and transport than field samples. Trace analysis of plankton, seawater, snow, snowmelt, snow deposition, and field blank samples were conducted in an ultraclean laboratory at the University of Alberta, Canada. OASIS WAX cartridges were thawed, pH conditioned with 4 mL of ammonium acetate buffer (25 mM, pH 4), and vacuum-dried to remove the remaining water. The target compounds were eluted with 4 mL of methanol, followed by 4 mL of methanol containing 0.1% ammonia, reduced under nitrogen to near dryness, and reconstituted with 0.2 mL of 50:50 methanol/water. All plankton and field blank samples were extracted following an established procedure.2 Briefly, filters were lyophilized, weighed, and transferred to 15 mL polypropylene vials. Each sample was spiked with mass labeled internal standards (Table S2, SI). After addition of 5 mL of acetonitrile, the vials were vortexed, sonicated for 20 min, and centrifuged for 10 min at 3000 rpm. The supernatant was collected, and the extraction was repeated with another 5 mL of acetonitrile. The two extracts were combined and reduced under a gentle stream of nitrogen and transferred to a 1.5 mL microcentrifuge tube containing 25 mg of ENVI-Carb and 50 μL of glacial acetic acid. The solution was mixed thoroughly and then centrifuged for 10 min at 12000 rpm. The supernatant was collected again, diluted in 50 mL of HPLC-grade water and extracted using SPE-WAX as described for water samples. 8462
DOI: 10.1021/acs.est.7b02521 Environ. Sci. Technol. 2017, 51, 8460−8470
Article
Environmental Science & Technology
Figure 2. PFAS concentrations in snow samples (pg L−1) (A) and relative composition (B). Near-field (NS1−NS3) and far-field (S1−S9) surface snow samples are located inside and outside the JC1 research station area, respectively. SD1−SD4 are fresh snow deposition samples, and SM1− SM3 are snowmelt samples.
■
concentrations in NS1−NS3 (300 ± 420 pg L−1) were also significantly higher (Kruskal−Wallis, p < 0.013) than in S1−S9 (12 ± 10 pg L−1) and were dominated by a high contribution of PFOS (81−100% contribution to ΣPFSAs), which was not observed in samples S1−S9. Even though ΣPFCAs did not present significant differences between NS1−NS3 (1000 ± 550 pg L−1) and S1−S9 (250 ± 120 pg L−1), individual PFCA homologues, C7−C13, were significantly higher (Kruskal− Wallis, p < 0.013) in NS1−NS3 compared to S1−S9. Wild et al.20 also reported high ΣPFASs concentrations, especially PFOS, in indoor dust from the Casey research station in coastal East Antarctica.
RESULTS AND DISCUSSION
PFASs Occurrence in Snow. Near-field surface snow collected from transited areas within the perimeter of the JC1 research station (NS1−NS3), other far-field surface snow (S1− S9), snow deposition (SD1−SD4), and snowmelt (SM1−SM3) had notable differences in concentration (Figure 2A, Table S6, SI) and PFAS profiles (Figure 2B). ΣPFAS concentrations in surface snow from the JC1 transited areas (NS1−NS3, 1300 ± 960 pg L−1) were significantly higher (Kruskal−Wallis, p < 0.013) than in other surface snow samples collected further afield (S1−S9, 260 ± 120 pg L−1), suggesting a localized effect of the research station as a PFAS source. ΣPFSAs 8463
DOI: 10.1021/acs.est.7b02521 Environ. Sci. Technol. 2017, 51, 8460−8470
Article
Environmental Science & Technology
Figure 3. PFAS concentrations in plankton (ng gdw−1) and surface seawater (pg L−1) at Raquelias (R) and Johnsons sites (J) from December 2014 to February 2015.
Surface snow samples S1−S9 had ΣPFASs concentrations ranging from 80 to 430 pg L−1, and these concentrations were not correlated with the distance to the JC1 station, which ranged from 100 m (S2 and S3) to 11 km (S6 and S7) (Figure 1, Figures S1−S4, SI). This is evidence of a very limited and local influence of the research station as a source of PFASs, thus not influencing the levels in samples S1−S9 located outside the perimeter of the research station. It is noteworthy that not even the PFOS concentrations in snow (samples S1−S9) showed a significant correlation with distance to the JC1 station. PFAS concentrations in surface snow (S1−S9) at Livingston Island were consistently lower than those reported for nearby King George Island (1100 to 2500 pg L−1),19 which likely has a larger local influence from several research stations, greater cruise tourism, and even a small airport. Individual PFASs in surface snow outside the JC1 perimeter (S1−S9), namely PFOS, PFOA, PFNA and PFDA, had snow phase concentrations in the same range as for snow in the high Arctic (i.e., 14−54 pg L−1 for PFOA).23 The highest PFAS concentrations in all snow or snowmelt samples (760−3600 pg L−1) were detected in fresh snow deposition samples (SD1−SD4), suggesting an important role of long-range atmospheric transport as a deposition source of PFASs to Livingston Island. More specifically, individual PFCA homologues were present at the highest average concentrations in SD1−SD4 among all the snow samples, higher than in surface snow collected within the perimeter of the JC1 research station. Conversely, PFSAs contributed
