Deposition History of Polychlorinated Biphenyls to the

Article pubs.acs.org/est

Deposition History of Polychlorinated Biphenyls to the Lomonosovfonna Glacier, Svalbard: A 209 Congener Analysis Olga Garmash,†,‡ Mark H. Hermanson,*,‡ Elisabeth Isaksson,§ Margit Schwikowski,∥ Dmitry Divine,§ Camilla Teixeira,⊥ and Derek C. G. Muir⊥ †

School of Industrial Engineering, Tampere University of Applied Sciences, FI-33520 Tampere, Finland Department of Arctic Technology, University Center on Svalbard, NO-9171 Longyearbyen, Svalbard § Norwegian Polar Institute, NO-9296 Tromsø, Norway ∥ Paul Scherrer Institute, 5232 Villigen PSI, Switzerland ⊥ Aquatic Contaminants Research Division, Environment Canada, Burlington, Ontario L7R 4A6, Canada ‡

S Supporting Information *

ABSTRACT: A 37 m deep ice core representing 1957−2009 and snow from 2009 to 2010 were collected on the Lomonosovfonna glacier, Svalbard (78.82° N; 17.43° E) and analyzed for 209 polychlorinated biphenyl (PCB) congeners using high-resolution mass spectrometry. Congener profiles in all samples showed the prevalence of tetra- and pentachlorobiphenyls, dominated in all samples by PCB-44, PCB-52, PCB-70 + PCB-74, PCB-87 + PCB-97, PCB-95, PCB-99, PCB-101, and PCB-110. The ∑PCB flux varied over time, but the peak flux, ∼19 pg cm−2 year−1 from 1957 to 1966, recurred in 1974−1983, 1998−2009, and 2009−2010. The minimum was 5.75 pg cm−2 year−1 in 1989−1998, following a 15 year decline. Peak ∑PCB fluxes here are lower than measured in the Canadian Arctic. The analysis of all 209 congeners revealed that PCB-11 (3,3′-dichlorobiphenyl) was present in all samples, representing 0.9−4.5% of ∑PCB. PCB-11 was not produced in a commercial PCB product, and its source to the Arctic has not been well-characterized; however, our results confirm that the sources to Lomonosovfonna have been active since 1957. The higher fluxes of ∑PCB correspond to periods when average 5 day air mass back trajectories have a frequency of 8−10% and reach 60° N or beyond over northern Europe and western Russia or the North Sea into the U.K.

■

INTRODUCTION

compounds and during thermal processes, such as waste incineration.5,6 PCBs are subject to long-range atmospheric transport (LRAT) because of high vapor pressures and environmental persistence of many congeners and are found in air throughout the world, including the polar regions.7,8 The atmospheric lifetime of PCBs in the gas phase is generally defined by the reaction rate with OH radicals9: The photolysis of PCBs is inefficient because of the low absorption of energy in ultraviolet B (UV-B) wavelengths.10 Once PCBs (and other gas-phase organic contaminants) are transported to the Arctic, their atmospheric oxidation is significantly decreased because of the low UV-B inputs over several months, low humidity, low tropospheric ozone concentrations, and low temperatures (T), resulting in low production of OH•.11 Once in the high-latitude atmosphere, PCBs can be deposited onto glacier surfaces by condensation or scavenging

Polychlorinated biphenyls (PCBs) were produced and sold as various commercial mixtures and used predominantly as dielectric fluids in electrical equipment from 1929 until the 1970s. PCB usage decreased rapidly after it was found to be persistent and bioaccumulative following apparent environmental release. The U.S.A. ceased PCB production in 1977; European countries continued until 1984, while Russian producers did not stop before 1993.1 About 50% of PCBs were made in the U.S.A. Major European producers were located in Germany, France, and U.K. (U.S. manufacturer), contributing 27.2% of the global manufacturing,1 while minor production was in Italy, Spain, Czechoslovakia, and Poland. Russian output contributed 13.2%; however, it is possible that PCB production data for U.S.S.R. (Russia) and other eastern European countries are not reliable, even though these regions could be important PCB sources to various regions of the world, including the Arctic.1 Commercial mixtures can be recognized by their specific congener profiles, which are documented in various studies.2−4 PCBs can also be produced as byproducts during manufacturing of other chlorinated © XXXX American Chemical Society

Received: May 31, 2013 Revised: September 22, 2013 Accepted: September 27, 2013

A

dx.doi.org/10.1021/es402430t | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

of particles with sorbed PCBs by snowfall.12 Once frozen in ice, PCB decomposition processes are minimized. PCBs are thus preserved when the contaminated snow or firn becomes ice. Analyzing contaminants in ice cores from high-elevation glaciers can reveal the history of atmospheric transport and net deposition of contaminants from industrial and agricultural areas to the Arctic. Our objective is to measure the net atmospheric fluxes of all 209 PCB congeners to the Lomonosovfonna glacier, the highest elevation ice on Svalbard, by sampling an ice core with a historic record and a seasonal snow pit showing the most recent annual pre-summer net inputs. At high elevation, above the atmospheric boundary layer, all inputs are from LRAT. The history covered in the Lomonosovfonna core, from 1957 to 2010, covers periods of PCB production (up to ∼1977), end of production (1977− 1993),1 end of “primary” emissions (2001),8 and presumed dominance of land-based emissions, for example, shown to be active near the original PCB production facility.13 Previous ice core studies on Svalbard have shown high inputs of various brominated flame retardants and pesticides that likely have little or limited use on Svalbard.11,14,15 This is the first record of atmospheric PCB input history on Svalbard. Atmospheric PCBs on and near Svalbard were first measured from 1980 to 1983 [pentachlorobiphenyl (PeCB) only]16 and have been measured regularly with active sampling on Svalbard since 1993.17,18 Passive sampling, representing 1 year of accumulation, has also occurred on Svalbard.19 In all of the investigations, the number of PCB congeners reported has been limited to 31 or fewer.19 Analyzing all 209 PCB congeners in glacier ice makes it possible to identify the most persistent and abundant atmospheric congeners accumulating in the Arctic and to compare distributions and inputs over time, giving qualitative and quantitative results. A comparison of historic PCB inputs with atmospheric transport geography to the Arctic helps identify regions that are likely sources.

Figure 1. Map of Svalbard (with inset of the North Atlantic region showing the location of Svalbard).

the site in 2000 to identify background contamination from shipping, handling, and the laboratory. The XAD-2 sample and blank columns were shipped to EC, where the XAD-2 was removed, spiked with a known amount of 1,3,5-tribromobenzene (1,3,5-TBB)21 and with 29 13Clabeled recovery and 3 cleanup standards, and then sequentially extracted in CH3OH and CH2Cl2. These extracts were combined and washed with a 3% NaCl solution and dried on anhydrous Na2SO4 (pre-cleaned by extraction in CH2Cl2). The extract was volume-reduced and fractionated on a column of 10% H2O-deactivated silica gel (pre-cleaned by extraction in CH2Cl2). The first elution from the column, 20 mL of nhexane, contained the nonpolar compounds, including the PCBs, which were analyzed by high-resolution gas chromatography−mass spectrometry (GC−MS) with a SPB Octyl GC column by AXYS Analytical (Sidney, British Columbia, Canada) using United States Environmental Protection Agency (U.S. EPA) Method 1668A. The resulting PCB congener mass data (pg sample−1) were corrected for the recovery of 1,3,5-TBB (see Table S1 of the Supporting Information), blank-subtracted, and normalized to the original sample volume. The 13C PCB recoveries appear in Table S2 of the Supporting Information. The average blank concentrations (pg sample−1) and average sample signal/blank ratios are in Table S3 of the Supporting Information. Because the data were blank-subtracted, the congener values in the average blank (where available) become the detection limits; otherwise, the instrument detection limits are used. The recovery and signal/blank results met our quality control (QC) requirements. The analysis of PCBs using this method, with a SPB-Octyl column, results in coelution of 50 congeners. In addition, in this analysis, there were 50 other congeners that appeared not >2 times in seven samples at concentrations ≪1% of ∑PCB that were removed from the results (see Table S4 of the Supporting Information). Our 209 congener analysis yielded 109 peaks. Among coeluting congeners (except 70 + 74, 87 + 97, and 135 + 151), the most stable or abundant congener in commercial mixtures2 was chosen to represent the peak, as shown in Table S3 of the Supporting Information. The International Union of

■

EXPERIMENTAL SECTION In March 2009, we drilled a 37 m long ice core representing ∼1957−2009 from Lomonosovfonna (78.82° N; 17.43° E), the highest elevation glacier on Svalbard [1202 meters above sea level (m.a.s.l.)], above the atmospheric boundary layer (500 m.a.s.l. in the summer and 1000 m.a.s.l. in the winter) (Figure 1).20 The core diameter was 8 cm. A snow pit sample covering the period from 2009 to 2010 was collected 5 °C and then were pumped through stainless-steel columns containing Amberlite XAD-2 resin, which absorbed contaminants of interest from the melt. The meltwater volume was measured after pumping. The XAD-2 had been cleaned using sequential solvent extraction and packed into the columns in a high-efficiency particulate air (HEPA)- and carbon-filtered air clean room by Environment Canada (EC), Burlington, Ontario, Canada. Along with the seven samples, there were four XAD-2 columns used for pumping deep core samples (dated from before PCB production) from an earlier, deeper core drilled at B



Article

Figure 2. (A) PCB congener profile representing a bulk snow sample accumulated during 2009−2010. Note that no congeners from the nonachloro homologue were observed in any sample. (B) PCB congener profile from the ice core representing 1998−2009.

33, all congeners have a higher percentage of ∑PCB than in the 1998−2009 ice sample (Figure 2B). PCB-8 and PCB-11 in the snow sample are among the dominant congeners. This pattern of more low-molecular-mass congeners in the snow sample is the result of deposition during colder winter months, an effect that could be temporary as volatile PCB congeners move back into the atmosphere during warmer summer months.23 Even these deposited amounts of MoCB, DiCB, and TriCB are likely to be a small portion of what is in the atmosphere during the winter. Cousins et al.24 noted that, at −20 °C, only 10% of DiCBs sorb onto particulates compared to almost 100% of HxCBs, which predicts that lower chlorinated PCB congeners are likely to remain in the gas phase even at low T. This effect can be different among other congeners, however, because the vapor pressures of DiCB vary depending upon the structure (discussed below).25 Heavier PCB molecules deposit more easily by dry deposition and snow scavenging because of lower vapor pressures and are therefore more likely to be retained within the snowpack. Our expectation is that some of the lower molecular mass PCB congeners that make up a large fraction of ∑PCB in surface snow at Lomonosovfonna will move back to the atmosphere during summer when air T is higher and the snowpack transforms to firn.26 Domaine et al.27 observed that, when the snow area index (SAI, a function of the snow surface area and density, which changes during transformation to firn) was high and T was low, PCB-28 and

Pure and Applied Chemistry (IUPAC) numbering system is used. Ice core samples were dated using the seasonal cycles of the oxygen isotope ratio (δ18O), which is known to peak during the summer.22 The accumulation rate for the ice core was calculated using meltwater pumped (liters) and ice core dating (years) for every sample. The average accumulation rate was found to be ∼0.6 m weq, with a decreasing trend toward the top of the core. The reported PCB flux values, in pg cm−2 year−1, were calculated by dividing the mass of congener by the surface area of the core (50.27 cm2) and years represented by the volume of the sample. Reporting flux eliminates effects of varying snow accumulation rates.

■

RESULTS AND DISCUSSION

PCB Congener Profiles. To estimate possible sources and temporal trends of PCB congener inputs at Lomonosovfonna, it is important to look at the results in a qualitative way, in this case the percentage of ∑PCB congener profiles. Panels A and B of Figure 2 show congener profiles in the two upper samples in this study, snow sample 2009−2010 and ice core sample 1998−2009, respectively. When comparing these profiles, it is evident that the snow sample (Figure 2A) has more of the lower chlorinated PCBs, including monochloro, dichloro, and trichloro homologues (MoCB, DiCB, and TriCB); up to PCBC



Article

sovfonna results in a PCB profile that is distinctively dominated by TeCB and PeCB homologues. Because of the higher volatility of some congeners with low molecular mass, their deposition to the snow is expected to be limited, which we observed in the percentage of ∑PCB congener profiles in air/snow/ice when our sample results were compared to the air percentage of ∑PCB measured at the Zeppelin station in Ny-Ålesund,31 located at 474 m.a.s.l. about 120 km west from Lomonosovfonna (see locations in Figure 1); air samples had much higher percentages of TriCBs, whereas snow and ice samples had higher percentages of all other congeners from TetraCB and higher (see Figures S8 and S9 of the Supporting Information). It is also clear that the difference in congener profile between air/snow was less than between air/ice samples, which again tells us that the PCB congener profile in the snow changes as snow converts to firn and ice. In addition to the congeners shown in Figure 2A, three other MoCB, DiCB, and TriCB congeners, PCB-5, PCB-10, and PCB-34, were detected only in this sample, where they represented 0.14% of ∑PCB. After these congeners are deposited to snow, they apparently are re-emitted to the atmosphere as firn or ice is formed. Our qualitative analysis also includes comparing homologue distributions among the samples (see Figure S10 of the Supporting Information). TeCBs and PeCBs have the largest fractions of ∑PCB in our samples: 24−31 and 30−46%, respectively. The highest amount of PeCBs was observed in the ice core samples 1998−2009 and 1974−1983, while the TeCB fraction was similar among all of the samples. Villa et al.32 found the same range of TeCB of ∑PCB in an alpine glacier but observed a lower contribution of PeCB (26−38%). Some congeners within a homologue at Lomonosovfonna varied consistently together in all samples: When the fluxes of PCB-99 and PCB-101 are compared to each other, r = 0.995 and p < 0.05 (see Figure S11 of the Supporting Information). Congeners 28 and 31 varied together with r = 0.995 and p < 0.05. Among common HxCB congeners, PCB-138 closely follows the contribution of PCB-153 with r = 0.933 and p < 0.05. These results indicate that there is a high probability that the PCBs found in Lomonosovfonna originate from similar types of sources over the period of 1957−2010, with some notable exceptions. A total of eight commercial PCB mixtures were chosen for comparison to our samples to identify if there is a relationship between our observed congener profile and particular products that were likely used in the regions contributing air masses to Svalbard, including Sovol (on the basis of average data from refs 3, 4, and 33), Aroclor 1242 lot A3, Aroclor 1254 lots A4 (late production), and Aroclor 1254 G4 (early production),2 and Clophen 40, Clophen 50, and Clophen 60,34 Results show that our congener profiles are more consistent with Aroclor 1254 lot G4 (early production), followed by Clophen 40 and Clophen 50. None of these three products had large fractions of global production; the European sources contributed not >17% of the global total.1 The five major PCB congeners in our samples with global production estimates by Breivik et al.1 (PCB-52, PCB-70, PCB-101, PCB-110, and PCB-118) make up 30−35% of ∑PCB in our samples but 14% of global production of all PCBs. Considering the low production levels of these congeners, our PCB congener profiles from Lomonosovfonna are not related so much to levels of production but apparently more to the persistence of various

PCB-180 retention were both >90% and, when SAI was low and T was high (condition of a warmer snowpack), PCB-28 retention dropped to 28.7%, while PCB-180 retention remained >90%. Similarities in panels A and B of Figure 2 congener profiles are also apparent. Between PCB congeners 35 and 101 (nearly all of them TeCB and PeCB congeners), the profiles vary similarly (r = 0.970; p < 0.05), although the snow sample has a lower percentage of ∑PCB of these congeners than the ice core sample from 1998 to 2009 because of the higher DiCB and TriCB. The pattern from hexachloro to decachloro homologues (from HxCB to DeCB) above PCB-103 is very similar between the samples (r = 0.975; p < 0.05). Congeners 44, 52, 70 + 74, 87 + 97, 95, 99, 101, and 110 are dominant in all samples, representing 39−52% of ∑PCB; all but PCB-99 are among the 10 most abundant peaks in all samples (with the only departure of PCB-99 being the surface snow sample) (see Figure 2B and Figures S2−S7 of the Supporting Information). The most persistent congeners in the environment are often considered to be those with no vicinal H substituents on at least one phenyl ring,28 which, in this group, includes PCB-74, PCB-97, PCB-99, and PCB-101; the other dominant congeners are considered to be less stable but usually in biotic samples, which are dominated by congeners that are more resistant to metabolism, e.g., PCB-118, PCB-138, PCB153, and PCB-180.29 However, the most abundant congeners at Lomonosovfonna are dominant over 53 years, showing high abiotic environmental persistence even with some vicinal H substituents. The ice core sample representing accumulation from 1989 to 1998 has a high amount of MoCB and DiCB congeners mostly because of high contribution of congeners 3 and 11, respectively. PCB-3 represents 4.7% of the ∑PCB in this sample, in comparison to 0.08−0.65% in other ice core samples and 0.8% in the snow sample. It is likely that such a comparatively large amount of PCB-3 reflects a high concentration of this congener in the air at the time of deposition. PCB-11 in this sample has the greatest relative abundance among ice core samples, being 3.4% in comparison to 0.90−2.4% in other samples. While this sample has the lowest overall flux, the high fractions of PCB-3 and PCB-11 suggest the possibility of a source from a thermal process5,6,30 (PCB 11 is discussed below). The ice core sample from 1983 to 1989, with the third lowest ∑PCB flux, shows a greater contribution of hexa, hepta, and octa homologues (HxCBs, HpCBs, and OcCBs) in comparison to other samples, apparently because the TeCB and PeCB amounts are lower. Changes in the PCB profile over the ice core can be explained by the change in the emission intensity at the source or the change of the atmospheric transport pattern shown by 5 day air mass back trajectories (discussed below). PCB vapor pressures, which play a role in deposition and volatilization, are a function of the number of chlorine atoms in the PCB molecule,29 with certain exceptions among mono- and non-ortho-substituted congeners.25 Lower chlorinated congeners should be more easily mobilized and transported over longer distances than congeners with high molecular mass; they would mobilize slowly and deposit at lower latitudes, with only a small fraction reaching the Arctic. This could be the reason why a low amount of HxCB and higher chlorinated congeners is observed at Lomonosovfonna. Alternatively, the specific type of source that contributes to the contamination at LomonoD



Article

Sweden, and Estonia, in late 2012 by the European Environment Agency (http://www.eea.europa.eu/data-and-maps/ figures/change-in-pcb-emissions-1). One potential source of atmospheric PCBs on Svalbard is soils in the existing and former coal-mining communities. A survey of soils in these and other locations shows that two of the coal-mining sites have soil ∑PCB concentrations (seven congeners) up to 28.7 ppm, although the median values are not >0.29.35 The latter value is less than the U.S. EPA action level for residential soils (1.0 ppm), while the former is just over the maximum action level range for industrial soils (10−25 μg g−1)36 Because of low air temperatures (mean annual daily air temperature = −7.5 °C), it is expected that PCB volatilization from soils will be limited. In any case, the Lomonosovfonna sampling site would not be affected by these sources because it is above the atmospheric boundary layer.20 To see if there is any connection between air mass movement and PCB input at Lomonosovfonna, we calculated 5 day backward air mass trajectories that were averaged over the time period of our samples and divided into five clusters (see panels A and B of Figure 4 and Figures S12−S18 of the Supporting Information). The trajectories show some similar results throughout the years for major air mass sources to Lomonosovfonna from within the Arctic: Most of the air masses (∼40−57% frequency) during 5 days before reaching

congeners in source regions, emissions from those sources, likelihood of transport and deposition, and persistence in glacial conditions. Total PCB Flux. Quantitatively, the ∑PCB flux at Lomonosovfonna fluctuates by nearly a factor of 4 from 1957 to 2010, shown in Figure 3. The highest flux was observed

Figure 3. Flux of total PCBs to the Lomonosovfonna ice core and surface snow, 1957−2010.

during four historic periods, 1957−1966, 1974−1983, 1998− 2009, and 2009−2010, with approximately the same value of ∼19 pg cm−2 year−1. There was a drop from this level to the lowest ∑PCB flux of 5.8 pg cm−2 year−1 over 15 years after 1983, which is consistent with the decline in industrial production. Deposited PCBs show a significant increase after 1998, which is at least partially sustained in the surface snow sample, which, again, may show a decline in the PCB content as the snow is converted to firn during the summer. Total PCB flux to Lomonosovfonna is affected by two global processes: environmental release of persistent congeners and the movement of air masses from remaining sources of those PCBs to Lomonosovfonna. Because PCB production around the world peaked in about 1970 and then declined until total phase out in 1993,1 we expect that the PCB transported to the Arctic would also decrease, especially in recent years. This is not the case here: At the time of the peak production 1966− 1974, we observe a 2-fold drop in the PCB flux compared to 1957−1966 and then an increase in the period 1974−1983. After 1983, the PCB input decreased until 1998, when the greatest increase in this record occurred. The transport to and deposition of PCB at Lomonosovfonna is expected to be related to not only production and use but also distribution processes that could include “secondary” sources in more recent years, which include PCBs that may be newly released to the environment and available to the atmosphere in urban and industrial areas. The existence of these sources is sometimes not well-documented: occasionally, large new sources are discovered, as estimated during 1990−2010 for Portugal,

Figure 4. (A) 5 day average air mass back trajectories to Lomonosovfonna, 2009−2010. (B) 5 day average air mass back trajectories to Lomonosovfonna, 1989−1998. E



Article

sovfonna. Similar to our study, the results from the Agassiz Ice Cap showed neither a distinct trend of PCB deposition over a period of time nor a relationship to global PCB production. PCB-11. Analysis of 209 PCB congeners in our study revealed a high fraction of PCB-11 (3,3′-dichlorobiphenyl), particularly in three of seven samples, where it shows a peak of more than 2% of ∑PCB. This congener does not appear in most PCB mixtures of commercial origin2−4 and, therefore, is not commonly included in most PCB analyses. Very small concentrations of PCB-11 are reported by Anderson et al.38 in Aroclors 1016, 1221, 1232, and 1242 and by Kim et al.6 in Kaneclor (KC 300, KC 400, KC 500, and KC 600), where the highest mass percentage (0.278%) is reported in Aroclor 1221. In the Lomonosovfonna surface snow sample, PCB-11 represents 4.5% of ∑PCB (see Figure 2), while in ice core samples representing years 1989−1998 and 1957−1966, it contributes 3.4 and 2.4%, respectively. In the rest of the ice core, while contributing less to ∑PCB, PCB-11 is still the dominant congener among all MoCBs and DiCBs (0.9−1.9% of ∑PCB). Figure 5 shows the historic flux of PCB-11 in the Lomonosovfonna snow and ice profile. Deposition in the

Lomonosovfonna originated over the Arctic Ocean and Greenland, which we assume are not sources of atmospheric PCBs. The other air masses come from western and eastern Europe and western Russia and are likely to pass urban areas, where the production and use of PCB mixtures took place and where sources are most likely to remain.7 The four periods of maximum PCB flux, including that to the surface snow sample, occur when there are long southern trajectories reaching 60° N or beyond into the North Sea and near or into the U.K. (for example, 2009−2010, trajectory 2 in Figure 4A and Figure S18 of the Supporting Information), or through western Russia and Finland (see trajectory 3 in Figure S16 of the Supporting Information), which have frequencies ranging from 8 to 11%. Breivik et al.1 showed that atmospheric PCB emissions at latitudes from 30° to 60° N and in a longitudinal zone from 0° to 20° E (both in regions of these trajectories) are among the highest on Earth. On an annual basis, back trajectories from areas corresponding to these latitudes and longitudes reaching Lomonosovfonna have frequencies of 29−40 days per year. The exception to this trend is 1983−1989, which has a low relative ∑PCB flux but a long southern trajectory into western Norway but not into the North Sea (see trajectory 5 in Figure S15 of the Supporting Information). The trend of high contaminant loads to the Arctic coming from the North Sea/U.K. region was also observed by Harner et al.37 but only for coplanar PCB and polychlorinated naphthalenes. During 1974−1983, 11% of air mass trajectories extended over western Russia and Finland and reached 60° N (see trajectory 3 in Figure S16 of the Supporting Information). Considering that there was PCB production during that time, we observe one of the highest ∑PCB fluxes at Lomonosovfonna. The low ∑PCB flux during years 1966−1974 is related to no trajectory extending over Europe south of the Arctic Circle (see trajectory 2 in Figure S17 of the Supporting Information), effectively missing urban areas, even though the frequency was comparatively high (25%). This is similar to the trajectory for the lowest PCB flux period from 1989 to 1998, in which the European (southern) trajectory reached only to the Kola Peninsula (∼67° N), even though the trajectory frequency was 17% (see trajectory 2 in Figure 4B and Figure S14 of the Supporting Information). This sample also included the highest proportions of PCB-3 and PCB-11 (see above), which could be from thermal processes (smelting) found on the Kola Peninsula.30 Gregor et al.21 investigated PCB inputs to the Agassiz Ice Cap in the Canadian Arctic, (80.83° N; 72.94° W; 1600 m.a.s.l., also above the atmospheric boundary layer), the only other investigation known to the authors that presents the historic PCB flux from LRAT to snow in the Arctic. Snow pit and firn samples in this study represented years from 1963−1964 to 1992−1993, production and end of production periods, with total PCB flux ranging from 9.1 to 93 pg cm−2 year−1. Unlike Lomonosovfonna, PCB congeners at Agassiz Ice Cap were dominated by MoCBs and DiCBs. The analysis included 60 PCB congeners and included four of the top nine congeners in our samples (PCB-44, PCB-87, PCB-101, and PCB-110). ∑PCB flux to the Agassiz Ice Cap is significantly higher than that to Lomonosovfonna but still in the same order of magnitude. Gregor et al. did not identify potential sources of PCB but noted that the lack of a clear trend in the data resulted from variability in source and delivery functions, which also appear in PCB data and associated trajectories to Lomono-

Figure 5. Flux of PCB-11 congener in Lomonosovfonna ice core and surface snow, 2009−2010.

earliest sample (1957−1966) is the largest throughout the ice core (0.45 pg cm−2 year−1). Then, the flux of PCB-11 levels off during 1966−1998 at about 0.18 pg cm−2 year−1. An increase in PCB-11 flux to 0.31 pg cm−2 year−1 is observed in the upper ice core sample representing 1998−2009. The PCB-11 flux increases significantly in the surface snow sample to 0.83 pg cm−2 year−1. As noted before, variations in low chlorinated PCB concentrations between air/snow/ice suggest that the snowpack of 2009−2010 may lose some PCB-11 back to the atmosphere, even though its vapor pressure is lower than most DiCBs.25 F



Article

for PCB-13 (+coeluting PCB-12) and r = 0.986 and p < 0.01 for PCB-15). In the flue gas results, however, PCB-13 (+coeluting PCB-12) was more concentrated than either PCB-11 or PCB-15 in six of nine samples, unlike our ice core results, which are dominated by PCB-11. More studies on PCBs originating from thermal processes (including incinerators) would be useful to determine their extent as a source of PCB-11 to the eastern Arctic. In general, the sources of PCB-11 to Lomonosovfonna are not wellcharacterized.

The analysis of the historical deposition of PCB-11 leads to a conclusion that this congener has been emitted to the atmosphere somewhere at lower latitudes at least since 1957. From the 5 day air mass back trajectories (panels A and B of Figure 4 and Figures S12−S18 of the Supporting Information), there is a clear indication that the three higher flux periods were all coincident with trajectories extending far south into the North Sea. The two highest PCB-11 flux periods have this trajectory extending into the U.K. (Figure 4A and Figure S18 of the Supporting Information). The historical flux of PCB-11 closely varied with PCB-13 (+coeluting PCB-12) with r = 0.965 and p < 0.01 and PCB-15 with r = 0.919 and p < 0.01. These congeners are considered to be the most stable DiCBs based on chemical and physical properties39 and also have the lowest vapor pressures of all DiCBs because they have no ortho Cl substituents.25 In comparing PCB-11 variance with other non-ortho PCB in our data set (PCB-2, PCB-3, PCB-35, PCB-37, PCB-39, PCB-77, and PCB-79), all are either not correlated with PCB-11 or not significant. The first reference to PCB-11 in the environment appeared regarding seals collected in Nova Scotia in 1995,40 although it had earlier been identified as a residue of municipal waste incineration.30 Since then, it has been detected in air,41−43 with a strong association to human populations.44 It has been found in wildlife, water, and sediment,45−47 sometimes in high concentrations relative to other PCB congeners. In most of the studies of PCB-11 cited by Choi et al.,19 the production of diarylide yellow pigments or disposal of products with the pigments is considered to be the major source. This pigment is common in consumer paper and plastic,46 which could become airborne when the product decomposes or, perhaps, when incinerated. PCB-11 can also be formed from dechlorination of PCB-77 and PCB-126, congeners found in very small amounts in commercial mixtures, making that source unlikely.46 The high fraction of PCB-11 in the Lomonosovfonna samples is an indication that emissions at the source(s) must be very large. This has been suggested once before in an Arctic atmospheric investigation: 1 year of continuous passive air sampling (one sample) results at Ny-Ålesund from 2005 to 2006 show PCB-11 to be the most concentrated PCB congener observed at a site not affected by local contamination, making up 8.7% of ∑PCB of 206 congeners (of which 31 were reported).19 Considering possible sources to Lomonosovfonna, PCB-11 was found in five samples of flue gas and ash from a laboratoryscale fluidized-bed reactor from municipal solid waste (MSW) incineration facilities in Sweden.5 It was dominant among MoCB and DiCB in one of two analyzed ashes, about 7% of ∑PCB, and was detected in another ash sample as well as in flue gases but at lower concentrations. Two other dominant non-ortho Cl-substituted DiCB congeners found, PCB-13 (+coeluting PCB-12) and PCB-15, are present in Lomonosovfonna samples, vary consistently with PCB-11 (see above), and have similar environmental stability. Among these flue gas and ash results, PCB-11 was dominant among these DiCB in all but one sample. An investigation of waste incineration plants in Japan revealed amounts of PCB-11 ranging from 0.23 to 2.8% of ∑PCB in flue gases,6 which is below but consistent with the range that we observed at Lomonosovfonna (0.90−4.5%). The variance of PCB-11 with PCB-13 and PCB-15 in nine flue gas samples was also strong and significant (r = 0.994 and p < 0.01

■

ASSOCIATED CONTENT

S Supporting Information *

Five tables of QC and other data and 18 figures of PCB congener profiles, air/snow and air/ice PCB congener comparisons, homologue distributions, PCB-99 and PCB-101 flux comparison, and seven air mass back trajectories. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Telephone: +47-79023351. E-mail: markhermanson@me. com. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study was partly supported by the U.S.−Norway Fulbright Foundation (awarded to Mark H. Hermanson). Svetlana Divine assisted with air mass back trajectories. Carmen Vega provided ice core dating. Drilling was supported in part by the Norwegian Polar Institute.

■

REFERENCES

(1) Breivik, K.; Sweetman, A.; Pacyna, J. M.; Jones, K. C. Towards a global historical emission inventory for selected PCB congenersA mass balance approach. 3. An update. Sci. Total Environ. 2007, 377, 296−307. (2) Frame, G. M.; Cochran, J. W.; Bowadt, S. S. Complete PCB congener distributions for 17 Aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congenerspecific analysis. J. High Resolut. Chromatogr. 1996, 19, 657−668. (3) Wyrzykowska, B.; Bochentin, I.; Hanari, N.; Orlikowska, A.; Falandysz, J.; Yuichi, H.; Yamashita, N. Source determination of highly chlorinated biphenyl isomers in pine needlesComparison to several PCB preparations. Environ. Pollut. 2006, 143, 46−59, DOI: 10.1016/ j.envpol.2005.11.018. (4) Kannan, N.; Shultz-Bull, D. E.; Petrick, G.; Duinker, J. C. High resolution PCB analysis of Kanechlor, Phenoclor and Sovol mixtures using multidimensional gas chromatography. Int. J. Environ. Anal. Chem. 1992, 47, 201−215. (5) Jansson, S.; Lundin, L.; Grabic, R. Characterisation and fingerprinting of PCBs in flue gas and ash from waste incineration and in technical mixtures. Chemoshphere 2011, 85, 509−515, DOI: 10.1016/j.chemosphere.2011.08.012. (6) Kim, K. S.; Hirai, Y.; Kato, M.; Urano, K.; Masunaga, S. Detailed PCB congener patterns in incinerator flue gas and commercial PCB formulations (Kanechlor). Chemosphere 2004, 55, 539−553, DOI: 10.1016/j.chemosphere.2003.11.056. (7) Pozo, K.; Harner, T.; Wania, F.; Muir, D. C. G.; Jones, K. C.; Barrie, L. Toward a global network for persistent organic pollutants in air: Results from the GAPS study. Environ. Sci. Technol. 2006, 40, 4867−4873.

G



Article

(8) Hung, H.; Lee, S. C.; Wania, F.; Blanchard, P.; Brice, K. Measuring and simulating atmospheric concentration trends of polychlorinated biphenyls in the Northern Hemisphere. Atmos. Environ. 2005, 39, 6502−6512. (9) Anderson, P. N.; Hites, R. A. OH radical reactions: The major removal pathway for polychlorinated biphenyls from the atmosphere. Environ. Sci. Technol. 1996, 30, 1756−1763. (10) Matykiewiczová, N.; Klánová, J.; Klán, P. Photochemical degradation of PCBs in snow. Environ. Sci. Technol. 2007, 41, 8308− 8314. (11) Ruggirello, R. M.; Hermanson, M. H.; Isaksson, E.; Teixeira, C.; Forsström, S.; Muir, D. C. G.; Pohjola, V.; va de Wal, R.; Meijer, H. A. J. Current use and legacy pesticide deposition to ice caps on Svalbard, Norway. J. Geophys. Res.: Atmos. 2010, 115, No. D18308, DOI: 10.1029/2010JD014005. (12) Wania, F.; Hoff, J. T.; Jia, C. Q.; Mackay, D. The effects of snow and ice on the environmental behaviour of hydrophobic organic chemicals. Environ. Pollut. 1998, 102, 25−41. (13) Hermanson, M. H.; Scholten, C. A.; Compher, K. Variable air temperature response of gas-phase atmospheric polychlorinated biphenyls near a former manufacturing facility. Environ. Sci. Technol. 2003, 37, 4038−4042. (14) Hermanson, M. H.; Isaksson, E.; Forsström, S.; Teixeira, C.; Muir, D. C. G.; Pohjola, V. A.; van de Wal, R. Deposition history of brominated flame retardant compounds to an ice core from Holtedahlfonna, Svalbard, Norway. Environ. Sci. Technol. 2010, 44, 7405−7410. (15) Hermanson, M. H.; Isaksson, E.; Teixeira, C.; Muir, D. C. G.; Compher, K. M.; Li, Y.-F.; Igarashi, M.; Kamiyama, K. Current-use and legacy pesticide history in the Austfonna ice cap, Svalbard, Norway. Environ. Sci. Technol. 2005, 39, 8163−8169. (16) Oehme, M.; Ottar, B. The long range transport of polychlorinated hydrocarbons to the Arctic. Geophys. Res. Lett. 1984, 11, 1133−1136. (17) Oehme, M.; Haugen, J.-E.; Schlabach, M. Seasonal changes and relations between levels of organochlorines in Arctic ambient air: First results of an all-year-round monitoring program at Ny-Ålesund, Svalbard, Norway. Environ. Sci. Technol. 1996, 30, 2294−2304. (18) Hung, H.; Kallenborn, R.; Breivik, K.; Su, Y.; BrorströmLundén, E.; Olafsdottir, K.; Thorlacius, J. M.; Leppänen, S.; Bossi, R.; Skov, H.; Manø, S.; Patton, G. W.; Stern, G.; Sverko, E.; Fellin, P. Atmospheric monitoring of organic pollutants in the Arctic under the Arctic Monitoring and Assessment Program (AMAP): 1993−2006. Sci. Total Environ. 2010, 408, 2854−2873. (19) Choi, S.-D.; Baek, S.-Y.; Chang, Y.-S.; Wania, F.; Ikonomou, M. G.; Yoon, Y.-J.; Park, B.-K.; Hong, S. Passive air sampling of polychlorinated biphenyls and organochlorine pesticides at the Korean Arctic and Antarctic research stations: Implications for long-range transport and local pollution. Environ. Sci. Technol. 2008, 42, 7125− 7131. (20) Esau, I.; Repina, I. Wind climate in Kongsfjorden, Svalbard and attribution of leading wind driving mechanisms through turbulenceresolving simulations. Adv. Meterol. 2012, No. 568454, DOI: 10.1155/ 2012/568454. (21) Gregor, D. J.; Peters, A. J.; Teixeira, C.; Jones, N.; Spencer, C. The historical residue trend of PCBs in the Agassiz Ice Cap, Ellesmere Island, Canada. Sci. Total Environ. 1995, 160/161, 117−126. (22) McMorrow, A.; van Ommen, T. D.; Morgan, V.; Curran, M. A. J. Ultra-high-resolution seasonality of trace-ion species and oxygen isotope ratios in Antarctic firn over four annual cycles. Ann. Glaciol. 2004, 39, 34−40. (23) Hansen, K. M.; Halsall, C. J.; Christensen, J. H. A dynamic model to study the exchange of gas-phase persistent organic pollutants between air and a seasonal snowpack. Environ. Sci. Technol. 2006, 40, 2644−2652. (24) Cousins, I. T.; Beck, A. J.; Jones, K. C. A review of the processes involved in the exchange of semi-volatile organic compounds (SVOC) across the air-soil interface. Sci. Total Environ. 1999, 228, 5−24.

(25) Falconer, R.; Bidleman, T. F. Vapor pressures and predicted particle/gas distributions of polychlorinated biphenyl congeners as functions of temperature and ortho-chlorine substitution. Atmos. Environ. 1994, 28, 547−554. (26) Herbert, B. M. J.; Halsall, C. J.; Villa, S.; Jones, K. C.; Kallenborn, R. Rapid changes in PCB and OC pesticide concentrations in Arctic snow. Environ. Sci. Technol. 2005, 39, 2998−3005. (27) Domaine, F.; Gallet, J.-C.; Bock, J.; Morin, S. Structure, specific surface area and thermal conductivity of the snowpack around Barrow, Alaska. J. Geophys Res. 2012, 117, No. D00R14, DOI: 10.1029/ 2011JD016647. (28) Hermanson, M. H.; Johnson, G. W. Polychlorinated biphenyls in tree bark near a former manufacturing facility in Anniston, Alabama. Chemosphere 2007, 68, 191−198, DOI: 10.1016/j.chemosphere.2006.11.068. (29) Hansen, L. G. The ortho Side of PCBs: Occurrence & Disposition; Kluwer Academic Publishers: Dordrecht, Netherlands, 1999. (30) Takasuga, T.; Inoue, T.; Ohi, E.; Umetsu, N.; Ireland, P.; Takeda, N. Characterisation of PCBs formed during thermal processes. Organohalogen Compd. 1994, 19, 173−176. (31) Aas, W.; Breivik, K. Heavy Metals and POP Measurements, 2010; Norwegian Institute for Air Research: Kjeller, Norway, Aug 2012; EMEP/CCC-Report 3/2012. (32) Villa, S.; Negrelli, C.; Maggi, V.; Finizio, A.; Vighi, M. Analysis of a firn core for assessing POP seasonal accumulation on an Alpine glacier. Ecotoxicol. Environ. Saf. 2006, 63, 17−24. (33) Takasuga, T.; Senthilkumar, K.; Matsumura, T.; Shiozaki, K.; Sakai, S.-I. Isotope dilutoin analysis of polychlorinated biphenyls (PCBs) in transformer oil and global commercial PCB formulations by high resolution gas chromatography−high resolution mass spectrometry. Chemosphere 2006, 62, 469−484. (34) Schulz, D. E.; Petrick, G.; Duinker, J. C. Complete characterization of polychlorinated biphenyl congeners in commercial Aroclor and Clophen mixtures by multidimensional gas chromatography−electron capture detection. Environ. Sci. Technol. 1989, 23, 852−859. (35) Ottesen, R. T.; Bogen, J.; Finne, T. E.; Andersson, M.; Dallmann, W. K.; Eggen, O. A.; Jartun, M.; Lundkvist, Q.; Pedersen, H. R.; Volden, T. Geochemical Atlas of Norway. Part 2: Geochemical Atlas of Spitsbergen; Geological Survey of Norway and Norwegian Energy and Water Resources Dirctorate: Norway, 2010. (36) United States Environmental Protection Agency (U.S. EPA). Guidance on Remedial Actions for Superfund Sites with PCB Contamination; Office of Emergency and Remedial Response, U.S. EPA: Washington, D.C., Aug 1990; EPA/540/G-90/007. (37) Harner, T.; Kylin, H.; Bidleman, T. F.; Halsall, C.; Strachan, W. M. J.; Barrie, L. A.; Fellin, P. Polychlorinated naphthalenes and coplanar polychlorinated biphenyls in Arctic air. Environ. Sci. Technol. 1998, 32, 3257−3265. (38) Anderson, J. W. Determination of congeners of polychlorinated biphenyls in reference materials. J. High Resolut. Chromatogr. 1991, 14, 369−372. (39) Zhai, Z.-C.; Wang, Z.-Y.; Chen, X.-H.; Wang, L.-S. DFT calculation on 204 polychlorinated biphenyls: their thermodynamic function and implication of Cl substitute position. J. Mol. Struct.: THEOCHEM 2005, 714, 123−131, DOI: 10.1016/j.theochem.2004.08.060. (40) Addison, R. F.; Ikonomou, M. G.; Stobo, W. T. Polychlorinated dibeno-p-dioxins and furans and non-ortho polychlorinated biphenyls in grey seals (Halichoerus grypus) from Sable Island, Nova Scotia, in 1995. Mar. Environ. Res. 1999, 47, 225−240. (41) Hu, D.; Martinez, A.; Hornbuckle, K. C. Discovery of nonAroclor PCB (3,3′-dichlorobiphenyl) in Chicago air. Environ. Sci. Technol. 2008, 42, 7873−7877. (42) Hu, D.; Lehmler, H.-J.; Martinez, A.; Wang, K.; Hornbuckle, K. C. Atmospheric PCB congeners across Chicago. Atmos. Environ. 2010, 44, 1550−1557. H



Article

(43) Yongjian, Y.; Yingming, L.; Hongtao, S.; Pu, W.; Qinghua, Z.; Guibin, J. Contamination patterns of PCB-11 in the soil from several Chinese cities. Environ. Chem. 2011, 30, 1764−1768. (44) King, T. L.; Yeats, P.; Hellou, J.; Niven, S. Tracing the source of 3,3′-dichlorobiphenyl found in samples collected in and around Halifax harbour. Mar. Pollut. Bull. 2002, 44, 590−596. (45) Basu, I.; Arnold, K. A.; Venier, M.; Hites, R. A. Partial pressures of PCB-11 in air from several Great Lakes sites. Environ. Sci. Technol. 2009, 43, 6488−6492. (46) Rodenburg, L. A.; Guo, J.; Du, S.; Cavallo, G. J. Evidence for unique and ubiquitous environmental sources of 3,3′-dichlorobiphenyl (PCB 11). Environ. Sci. Technol. 2010, 44, 2816−2821, DOI: 10.1021/ es901155h. (47) Litten, S.; Fowler, B.; Luszniak, D. Identification of a novel PCB source through analysis of 209 congeners by US EPA modified method 1668. Chemosphere 2002, 46, 1457−1459.

I


Deposition History of Polychlorinated Biphenyls to the

Recommend Documents