Temperature Dependence of the Air Concentrations of Polychlorinated

Polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs) were quantified in 67 high volume air samples taken concurrently in a fore...
0 downloads 0 Views 300KB Size
Environ. Sci. Technol. 2007, 41, 4655-4661

Temperature Dependence of the Air Concentrations of Polychlorinated Biphenyls and Polybrominated Diphenyl Ethers in a Forest and a Clearing Y U S H A N S U , †,§ F R A N K W A N I A , * ,† YING DUAN LEI,† TOM HARNER,‡ AND MAHIBA SHOEIB‡ Department of Chemical Engineering & Applied Chemistry and Department of Physical & Environmental Sciences, University of Toronto Scarborough, 1265 Military Trail, Toronto, Ontario, Canada M1C 1A4, and Science and Technology Branch, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4

Polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs) were quantified in 67 high volume air samples taken concurrently in a forest and a clearing in southern Ontario, Canada from October 2001 to November 2002. Air concentrations were comparable between the two sites. Gaseous PCBs ranged from 6.4 to 150 pg‚m-3, and gaseous PBDEs ranged from below method detection limit (BDL) to 55 pg‚m-3 (with two extreme events up to 290 pg‚m-3). Particulate PBDEs ranged from BDL to 40 pg‚m-3. Gaseous concentrations of PCBs and PBDEs were highly temperature dependent, suggesting a relatively strong influence of re-evaporation. Air concentrations of highly chlorinated PCBs in the forest were more temperature dependent than those in the clearing, whereas no difference was observed for the lesschlorinated PCBs. Forest filtering may have enriched highly chlorinated PCBs in the forest soil relative to the soil in the clearing, resulting in a higher contribution of reevaporation for highly chlorinated PCBs at the forest. Compared to measurements conducted a decade earlier at a nearby site, PCB air concentrations were generally less temperature dependent, indicative of a reduction in the contribution of re-evaporation in the region. Furthermore, a significant correlation was found between temperature dependence and degree of chlorination, which had not been apparent in the previous study. This is presumably because depuration from soils occurred slower for highly chlorinated PCBs, resulting in their relatively higher abundance in terrestrial surfaces and, therefore, higher contribution from re-evaporation. Contrasting with the PCBs, the temperature dependence of PBDE air concentrations did not differ between congeners or between forest and clearing site. This could be a result of different usage and emission history: PCBs were banned approximately three decades ago, whereas PBDEs are currently still in use. * Corresponding author phone: (416) 287-7225; e-mail: [email protected]. † University of Toronto Scarborough. ‡ Environment Canada. § Current address: Science & Technology Branch, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4. 10.1021/es070334p CCC: $37.00 Published on Web 06/02/2007

 2007 American Chemical Society

Consequently, the influence of primary emissions on air concentrations is expected to be more important for PBDEs than for PCBs.

Introduction Previously observed seasonal variations of semivolatile organic compound (SOC) concentrations in air (1-6) have been attributed to changes in ambient temperature, application/emission patterns, OH radical concentrations, and other environmental factors, such as wind speed, wind direction, and snowmelt (2, 3, 5-7). In the environment, SOCs partition between air and the Earth’s surface, comprised of soils, vegetation, and water bodies, rendering air/surface exchanges an important environmental pathway. Some persistent SOCs can reach the remote Arctic regions by repeated cycling between air and surfaces (8), a process that has been termed “grass-hopping” (9). Previous studies have shown that forest canopies can effectively take up SOCs from air and falling leaves transfer them to forest soils (10). Modeling studies suggest that this forest filter effect reduces air concentrations at the expense of higher soil concentrations (11, 12). Ultimately, it increases the residence time of SOCs in the environment, because degradation of SOCs is generally slower in soils than in the atmosphere (12). Temperature is one of the major factors controlling SOC seasonality in air and the dependence of air concentrations on temperature can be quantitatively expressed as

ln (P/Pa) ) m/T + b

(1)

where P is the partial pressure of a SOC in air (Pa), T is ambient temperature, and m and b are fitting parameters (2, 3, 6, 13, 14). When the partial pressure in air is in equilibrium with the surface phases, this is a form of the Clausius-Clapeyron equation and the slope m can be interpreted as the enthalpy of the air-surface transition process. As equilibrium cannot always be assumed, the slope m, or an “apparent enthalpy of air-surface exchange” derived from that slope, is often used to distinguish the relative importance of primary advective inputs and local secondary emissions (13, 14). Different congeners of the polychlorinated biphenyls (PCBs), widely applied in electrical equipment prior to the 1970s (16), display a large range in physical-chemical properties (15). In response to regional and ultimately global bans on their usage, air concentrations of PCBs have declined, as documented by long-term monitoring programs in the Great Lakes (2) and the Arctic (17). Polybrominated diphenyl ethers (PBDEs), with a molecular structure similar to the PCBs, are mainly used as flame retardants. The primary technical mixtures include penta-, octa-, and deca-BDE (18). In contrast to PCBs, PBDEs are currently used and environmental levels are increasing globally (19). In this study, high-volume (HiVol) air samples were collected every 12 days over a 1-year period at a rural site in the Great Lakes region and analyzed for PCBs and PBDEs. The purpose of this study was to investigate temperature dependence and air/surface exchange processes of PCBs and PBDEs. Concurrent air samples were collected under a deciduous forest canopy and in a nearby clearing to study the effect that a forest has on SOC deposition and reevaporation. Furthermore, by comparing the behavior of two groups of SOCs, which were either banned decades ago or VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4655

are currently still in use, we hoped to gain insights into air/ surface exchange at different times of a SOC’s emission history.

Materials and Methods Sampling. The field sampling sites at Borden are about 75 km north of Toronto and 25 km south of Georgian Bay (44°19′ N/80°56′ W), in the Great Lakes region of southern Ontario, Canada. The forest site is a mixed deciduous forest of red maple and large-tooth aspen, whereas the meadow site is 3 kilometers away (20). Approximately 750-900 m3 of air were sampled over 24 h simultaneously at both sites by using a sampling train consisting of two glass fiber filters (GFFs) and two polyurethane foam (PUF) plugs. The sampling was done once every 12 days between October 2001 and November 2002 to coincide with the Integrated Atmospheric Deposition Network in the Great Lakes. In total, there were 35 sampling events (20). Temperatures were recorded continuously at a height of 1.7 m by a meteorological station located at the forest site. Half-hour air temperatures were averaged on a daily basis. The daily average temperature during the 1-year sampling varied over a 30 °C range (20). Sample Extraction, Cleanup, and Analysis. GFFs and PUFs were Soxhlet-extracted overnight with dichloromethane and petroleum ether, respectively. Extracts were cleaned and fractionated on silicic acid/alumina columns. Prior to instrumental analysis, 100 ng of mirex was added as an internal standard. All samples were quantified for 56 PCBs and 13 PBDEs by gas chromatography-mass spectrometry (GC-MS) in selected ion monitoring mode. Splitless injection onto a DB-5MS column was employed. Details on extraction, cleanup and instrumental analysis, including GC temperature program and target and qualifier ions, are given in the Supporting Information and in ref 20. Quality Assurance/Quality Control (QA/QC). Prior to Soxhlet extraction, recovery surrogates were added to the GFF and PUFs consisting of a mixture of 13C12-labeled PCBs (PCB-28, -52, -101, -138, -153, -180, and -209). PUF recoveries ranged from 72 ( 9% to 96 ( 11% (n ) 68), while those for the GFF were in the range of 62 ( 7% to 90 ( 8% (n ) 70) (Table S1). Since the analytical method cannot distinguish between labeled and native PBDEs, recovery experiments were done by spiking clean PUFs with 10 ng of mixed PBDEs, which were extracted and cleaned up as samples. Recoveries ranged from 78 ( 5% to 95 ( 1% for 13 PBDEs (n ) 5). Method detection limits (MDLs) of gaseous compounds were 0.063-0.58 pg‚m-3 for PCBs and 0.052-3.1 pg‚m-3 for PBDEs (n ) 35). MDLs of particle-bound compounds were 0.0630.58 pg‚m-3 for PCBs and 0.052-2.8 pg‚m-3 for PBDEs (n ) 35). Laboratory blanks yielded either nondetectable amounts or systematically lower values than the field blanks. No breakthrough to the back PUF or adsorption on the back GFF was found for PCBs or PBDEs. Data were corrected for blank levels, but not recovery. Detailed QA/QC information is given in the Supporting Information and in ref 20.

Results Air Concentrations of PCBs and PBDEs. Fifty-six PCB congeners were quantified in the PUFs, but not detected in the GFFs. Gaseous concentrations are presented for only 18 selected PCBs as other congeners were below MDLs in over 80% of the PUFs. These 18 PCBs have between two and seven chlorines per molecule. Arithmetic mean and median gaseous concentrations (CG) and their ranges are listed in Table S2 for individual congeners and total PCBs (Σ18PCB). During the 1-year sampling period, concentrations of Σ18PCB ranged from 6.4-110 pg‚m-3 (mean 42 pg‚m-3 and median 31 pg‚m-3) and from 7.9 to 150 pg‚m-3 (mean 44 pg‚m-3 and median 35 pg‚m-3) in clearing and forest, respectively. This 4656

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 13, 2007

range compares well with other HiVol measurements conducted in early 2002 in rural southern Ontario, i.e., BDL (i.e., below MDL) to 110 pg‚m-3 (5). Gaseous PCBs at the sampling site were dominated by the tri-CBs (PCB-17, -18, and -28) and penta-CBs (PCB-87, -95, -99, -101, -110, and -118), with moderate contributions from the di-CB (PCB-8), tetra-CBs (PCB-44 and -52), and hexa-CBs (PCB-138, -149, -151, and -153), and minor fractions of the hepta-CBs (PCB180 and -187). Among 13 quantified PBDEs, nine congeners were above the MDLs in more than 20% of the samples. These nine PBDEs range from tri- to hexa-BDE. The major components of technical “penta-BDE” (BDE-47, -99, and -100) were dominant in the gas phase. Mean and median concentrations of individual PBDEs and their ranges are listed in Table S2. Total gaseous concentrations (CG) of PBDEs (Σ9PBDE) in the clearing and forest ranged from 0.082 to 120 pg‚m-3 (mean 17 pg‚m-3 and median 9.2 pg‚m-3) and from BDL to 290 pg‚m-3 (mean 22 pg‚m-3 and median 7.9 pg‚m-3), respectively. When two sampling events with “extreme” peak concentrations (nos. 20 and 23, Figure 1B) are excluded, the maximum gaseous Σ9PBDE is 55 pg‚m-3 at the clearing and 36 pg‚m-3 in the forest. Total particle-bound concentrations (CP) of Σ9PBDE ranged from BDL to 40 pg‚m-3 (mean 5.2 pg‚m-3 and median 2.4 pg‚m-3) at the clearing and from BDL to 31 pg‚m-3 (mean 3.2 pg‚m-3 and median 0.36 pg‚m-3) in the forest. Concentrations previously measured with HiVol samplers in the Great Lakes region are comparable: 5-52 pg‚m-3 in 1997-1999 (21) and BDL to 34 pg‚m-3 in January-April 2002 (5). Also, air concentrations of PBDEs measured by passive air sampling were in the same range: 10-70 pg‚m-3 in 2002-2003 (22) and 10-30 pg‚m-3 in 20002001 (23). Two “extreme” peak PBDE concentrations were recorded during sampling event nos. 20 and 23 (Figure 1B). Event no. 23 was quite unusual in that it also displayed very high concentrations of polycyclic aromatic hydrocarbons (PAHs) (20). This particular sampling period started on 2 July 2002, the day after a national holiday celebrated with fireworks across Canada. It is possible that emissions of PBDEs and PAHs were much higher than usual on that day. Similar spikes in PBDE and PAH air concentrations were observed during bonfire night in the UK (24). During event no. 20, the air mass passed through the Greater Toronto area prior to arriving at Borden (Figure S1), suggesting that elevated PBDEs were likely a result of urban sources. Sampling of air masses that had recently passed over industrial/urban areas is likely also the cause of elevated PBDE concentrations during event nos. 28 and 34. Back-trajectories indicate that air had passed through southeastern Toronto a few hours before reaching Borden during event no. 28, and the air mass of event no. 34 largely originated in the eastern United States (Figure S1). A strong correlation between air concentrations of PCBs and PBDEs was reported previously for the Great Lakes region (21). Although the linear relationship between PCB and PBDE concentrations was weak in the current study, a significant correlation was found between the logarithm of the concentration of Σ18PCB and Σ9PBDE (Figure 2F). Concentrations of PCBs were significantly higher than those of the PBDEs (p < 0.001, paired t test), which is also consistent with what has been reported previously for the area (5, 21, 22). Concentrations of individual PCB and PBDE congeners are listed in Tables S3-S6 of the Supporting Information. Neither the concentrations of the PCBs, nor those of the PBDEs were significantly different between forest and clearing. This is not unreasonable considering the close proximity of the two sampling sites. Modeling studies that predict a notable decrease in SOC air concentrations as a result of the forest filter effect refer to the regional (11) or even global scale (12), and cannot be readily applied to the

FIGURE 1. Total concentrations of gaseous PCBs (Σ18PCB), gaseous and particle-bound PBDEs (Σ9PBDE) at the clearing and forest site, and ambient temperatures during 35 sampling events. local scale of the current field study. A recent study in the Italian Alps (25) claims to have observed quantifable differences in the air concentrations of PCBs in forests and clearings in close proximity. However this study relied on a passive air sampling methodology that is known to be influenced by variable wind speed conditions (22). The differences in the amounts of PCBs sequestered in the air samplers were therefore more likely the result of lower wind speeds and therefore slower uptake in the samplers deployed in the forest. Temperature Dependence of Air Concentrations. The time trend of the concentrations of gaseous Σ18PCB, and gaseous and particle-bound Σ9PBDE in clearing and forest are shown in Figure 1. Generally, elevated gaseous PCB concentrations were observed when temperatures were high, whereas low concentrations occurred during cold periods (Figure 1A). Concentrations of gaseous Σ9PBDE appeared to correlate with ambient temperature as well, generally showing high concentrations during summer and low concentrations during winter and spring (Figure 1B). Figure 1C indicates that no apparent relationship

was seen between particle-bound Σ9PBDE and ambient temperature. Table S7 in the Supporting Information lists regression results for eq 1 expressing the temperature dependence of the vapor concentrations of individual PCB and PBDE congeners and of different homologue groups at the two sites. The slope m was significant at the 99% confidence level for all PCB congeners except PCB-180 and -187, both of which had low air concentrations and fewer samples above the MDLs (Table S2). Similarly, the slope m of the temperature dependence relationship was significant for all PCB homologues except for hepta-CB, which is composed of PCB-180 and 187 (Table S7). The temperature dependence of gaseous PBDEs was not as strong as for the PCBs: the regressions were significant at the 95% confidence level for BDE-28, -47, -66, -99, and -100 at both sites, and for BDE-17 at the forest site. When data were pooled into homologue groups, significant correlations were only found for tri-BDE at both sites and penta-BDE at the forest site (Table S7). VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4657

FIGURE 2. Slope m for PCBs (A) and PBDEs (D); coefficient of determination R2 for PCBs (B) and PBDEs (E); plots of slope m against chlorine number of PCBs (C); and correlation between concentrations of Σ18PCB and Σ9PBDE on a logarithm scale (F). Comparing the regressions for different PCB congeners (Figure 2A), we find that slopes m become steeper with increasing degree of chlorination. The coefficient of determination R2 also increases with PCB chlorine number (Figure 2B), indicating that the amount of variance explained by temperature is larger for highly chlorinated PCBs than less chlorinated ones. When the slopes m were averaged within each homologue group and plotted against the number of chlorines (Figure 2C), a strong correlation was found, which is consistent with findings based on multi-year measurements in the Great Lakes region (6) and in other locations (13). Interestingly, such a correlation was not observed when PCBs were measured more than a decade earlier (1988/1989) (1, 13) at Egbert (44°14′N/79°47′W), which is located approximately 10 km southeast of Borden (Figure 2C). In contrast to the PCBs, the slopes m derived for PBDEs are not steeper with increasing halogenation (Figure 2D), and R2 decreases with bromine number (Figure 2E). When temperature dependence slopes m and R2 for 16 PCBs are compared between forest and clearing (Table S7, Figure 2A and B), no differences are found for di-, tri-, and -tetra-CB. However, slopes of highly chlorinated PCBs are significantly steeper in the forest than in the clearing (paired Student’s t test with two-pair distribution, p < 0.0001). R2 of the highly chlorinated PCBs are also significantly higher at the forest (p < 0.001) (Figure 2B). A stronger dependence of the slope m on the degree of chlorination was found at the forest than the clearing (Figure 2C). Different from the PCBs, the slopes m for the PBDEs did not show significant differences between the two sites (Figure 2D), and neither did R2 (Figure 2E). Influence of Temperature on PCB Composition. At 25 °C, the subcooled liquid vapor pressure (PL) of PCBs ranges over 3 orders of magnitude (1.48 × 10-1 Pa for PCB-8 and 1.08 × 10-4 Pa for PCB-180) (15). Moreover, the temperature dependence of PL is quite different: the heat of vaporization ranges from 70.6 kJ‚mol-1 for PCB-8 to 94.1 kJ‚mol-1 for PCB180 (15). This means that the gas/condensed phase exchange 4658

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 13, 2007

of highly chlorinated PCBs is more dependent on temperature than that of less chlorinated congeners. When air concentrations are driven more by re-evaporation than advective inflow, variations of temperature can lead not only to changes in absolute levels, but also to variation in the composition of a PCB mixture. Air concentrations of each homologue were divided by total PCB concentrations to derive the relative abundance (%) of a homologue group during each sampling event. The relative abundance and ambient temperature (t, °C) were linearly regressed, yielding significant correlations for all homologue groups except for tetra- and hepta-CB (Figure 3A and B, Table S8). The relative abundance of highly chlorinated hexa- and penta-CB increased with temperature at the expense of less-chlorinated di- and tri-CB. Lack of temperature dependence suggests no strong variation in the relative abundance of tetra-CB over the 1-year period, whereas a weak temperature dependence of the relative abundance of hepta-CB is presumably caused by low air concentrations at the sampling sites (Table S2). To further illustrate the temperature effect on PCB composition, samples were binned into five segments based on the temperature during sampling. Concentrations for events falling into the same temperature range were averaged and relative abundances were calculated using those averages. The relative abundance of di- and tri-CB decreased and that of pentaand hexa-CB increased at higher temperatures (Figure 3C and D). There is no apparent difference between the clearing and forest. These compositional changes are consistent with increasing heats of vaporization with increasing degree of chlorination. In contrast to the PCBs, no significant correlation between composition and ambient temperature was found for PBDEs.

Discussion Air concentrations of highly chlorinated PCBs showed a stronger dependence on ambient temperature than those of less chlorinated congeners (Figure 2A). However, no such

FIGURE 3. Relationship between relative abundance of PCB homologues in percent and temperatures at clearing (A and C) and forest (B and D) site. trend was observed at the nearby site of Egbert a decade earlier (Figure 2C) (7). Moreover, air concentrations of highly brominated PBDEs were no more temperature dependent than those of less brominated ones in this current study (Figure 2D). It appears that the temperature dependence of the PCBs varied temporally and is also different from that of the PBDEs. The data further indicate that air concentrations of highly chlorinated PCBs were more temperature dependent in the forest than in the clearing (Figure 2A), whereas no such difference is observed for either less chlorinated PCBs or PBDEs (Figure 2D). If the slope m is interpreted as a measure of the relative importance of re-evaporation and advective inputs for air concentrations (13, 14), a steeper slope implies a higher contribution of re-evaporation relative to advection at the sampling site. This allows us to draw several conclusions about the air-surface exchange of SOCs at Borden. Effect of Forest on Temperature Dependence of PCB Air Concentrations. Previous studies have shown that forests can effectively take up airborne SOCs of intermediate volatility (i.e., those with octanol-air partitioning coefficient log KOA ranging from 7 to 11) and further transport them to soils with falling leaves (10-12). Consequently, higher concentrations of these SOCs are usually found in forest soils than in nearby clearings (26). Considering their KOA values and the meteorological conditions at Borden, di- and tri-CBs (i.e., PCB-8, -15, -17, -18, and -28) are expected to reach equilibrium distribution between air and forest canopy during the lifetime of deciduous leaves, and tetra-CBs (i.e., PCB-44 and -52) will be close to reaching equilibrium (10, 27). The uptake capacity of the forest canopy is, however, much larger for highly chlorinated PCBs (i.e., penta-, hexa-, and heptaCB), leading to their continuous uptake in the canopy during the summer season. Therefore, enhancement of deposition from air to forest soils is larger for highly chlorinated PCBs than for less chlorinated ones.

Significantly steeper slopes and higher R2 (Figure 2A and B) suggest that for highly chlorinated PCBs the importance of secondary emissions was higher in the forest than at the clearing during the sampling period. The distance between the two sampling sites is approximately 3 km, and air concentrations measured at either site were not significantly different (Table S2, Figure 1A). PCBs had been banned for use about three decades ago (16) and current air concentrations are believed to be strongly influenced by secondary emissions globally (26). During periods of active PCB use, primary emission dominated; air concentrations were high; and exchange between air and soil was net depositional. Because of the forest filter effect, deposition of highly chlorinated PCBs to forest soils was enhanced relative to open areas during this period of usage and primary emissions. Accordingly, the inventory of highly chlorinated PCBs is expected to be higher in forest soils than in soils in open areas, as is indeed observed in European soils (26). After global bans on usage, primary emissions and air concentrations declined and air/soil exchange of PCBs reversed from net deposition to net evaporation, at least during summer time. The extent of re-evaporation depends on how much had been deposited in the past, being presumably higher from soils that had received higher deposition. Due to elevated deposition of highly chlorinated PCBs to forest soils, their fugacity gradient between soil and air is expected to be higher in the forest than in the clearing, which would explain why re-evaporation of highly chlorinated PCBs is more important in the forest than in the clearing (as is reflected in slopes m for highly chlorinated PCBs that are steeper in the forest than in the clearing, Figure 2A). It also explains why the importance of re-evaporation of less chlorinated PCBs is similar in forest and clearing (i.e., why the slopes m for less chlorinated PCBs are not significantly different, Figure 2A). These congeners are not subjected to forest filtering and the soil inventory and the soil/air fugacity gradient are likely VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4659

similar in the forest and the clearing. Fugacity in forest soils could be reduced because of higher fugacity capacity (due to higher organic matter content) compared to nonforest soils. However, field measurements indicate that PCB concentrations normalized by organic matter were higher in forest soils than nearby non-forest soils (25), suggesting larger fugacity gradients with the former. Change of PCB Temperature Dependence in Time. Air concentrations of PCBs were measured at Egbert in 1988/ 1989 (1, 7). Temperature dependence slopes m were generally steeper in the earlier study than the ones reported here, especially for less chlorinated PCBs. Moreover, slopes were similar for different PCB homologues in the Egbert study, whereas in Borden the slope was steeper with increasing degree of chlorination (Figure 2C). Comparison between the two studies suggests that the slope m varies with time, and so does the correlation between slope m and chlorine number. Strong temperature dependence (i.e., steep slopes and high R2) indicates that re-evaporative emissions were important for all PCBs in the earlier study (7), and similar slopes imply that the evaporation potential was comparable among different PCBs (13). In the time period of 1988/1989, the net direction of air/surface transfer for all PCBs was from soil to air. Since less chlorinated congeners are more degradable and more volatile than highly chlorinated ones (15), depuration rates from soils should be faster for less chlorinated PCBs. Accordingly, the reservoir in soils is expected to have declined faster for less chlorinated than for highly chlorinated PCBs, and eventually highly chlorinated PCBs may have become relatively enriched in soil. Therefore, re-evaporative emissions are expected to remain more important for highly chlorinated PCBs than less chlorinated PCBs. The above hypothesis is consistent with the observations in this study, that the decrease in slope was larger for less chlorinated PCBs than for highly chlorinated ones (Figure 2C). Moreover, larger energies of gas/condensed phase transfer for highly chlorinated PCBs than less chlorinated ones can also lead to steeper slopes for highly chlorinated PCBs (15), which is supported by the strong correlation between relative abundance of PCBs and ambient temperature (Figure 3A and B). Regional air concentrations of PCBs decreased approximately by a factor of 3 during the past decade (1). However, reduction factors were not much different between PCB congeners (Table S9). If air concentrations of PCBs were exclusively driven by re-evaporation from local surfaces, reductions in air concentrations should have been larger for less chlorinated PCBs than highly chlorinated ones. Borden and Egbert are in a rural area in the vicinity of a large urban region and the advective input of PCBs is a distinct possibility. Advective inputs would need to be larger for less chlorinated PCBs than for highly chlorinated ones to result in a similar reduction in the air concentrations of different PCBs. This is again consistent with observations in this study, namely shallower slopes and lower R2 for less chlorinated than highly chlorinated PCBs (Figure 2A and B). In summary, reevaporative emissions of PCBs declined in the past decade in the sampling region and the decrease was larger for less chlorinated PCBs. Comparison of Temperature Dependence of PCBs and PBDEs. In contrast to the PCBs, the slopes m derived for PBDEs did not become steeper with increasing halogenation (Figure 2D), which is similar to what had been observed for PCBs in the earlier study (7). Furthermore, R2 was lower for highly brominated PBDEs than for less brominated ones, indicating that re-evaporative emissions were more important for the latter. Differences between PCBs and PBDEs may reflect the fact that PCBs were banned for use more than three decades ago (16), whereas PBDEs are currently still in use (18). Air concentrations of PCBs are declining continu4660

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 13, 2007

ously (2, 17), whereas PBDE concentrations are increasing in the environment (19). Although steeper slopes m were found for highly chlorinated PCBs at the forest than the clearing, the m for the PBDEs did not show a significant difference between the two sites (Figure 2D), and neither does R2 (Figure 2E). Based on their log KOA values (10-12), the forest filter effect should enhance the deposition of PBDEs to forests, which may suggest that re-evaporation potential from surfaces was stronger for PCBs than PBDEs at the sampling site. It is conceivable that differences in the temperature dependence between less and highly brominated PBDEs, and between forest and clearing will become apparent after primary emissions are no longer dominant.

Acknowledgments We are grateful to B. Gevao for helping with sampling and R.M. Staebler of Environment Canada for providing meteorological data. The study was supported by the Canadian Foundation for Climate and Atmospheric Sciences.

Supporting Information Available Detailed description of quality assurance/quality control and instrumentation; recoveries of PCBs and PBDEs; tables with the concentrations of PCBs and PBDEs; regression results between air concentrations and relative abundance and temperature; annual average concentrations of PCBs in this and previous study; and back-trajectories for sampling event nos. 20, 23, 28, and 34. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Hoff, R. M.; Muir, D. C. G.; Grift, N. P. Annual cycle of polychlorinated biphenyls and organohalogen pesticides in air in Southern Ontario. 1. air concentration data. Environ. Sci. Technol. 1992, 26, 266-275. (2) Hillery, B. R.; Basu, I.; Sweet, C. W.; Hites, R. A. Temporal and spatial trends in a long-term study of gas-phase PCB concentrations near the Great Lakes. Environ. Sci. Technol. 1997, 31, 1811-1816. (3) Simcik, M. F.; Basu, I.; Sweet, C. W.; Hites, R. A. Temperature dependence and temporal trends of polychlorinated biphenyl congeners in the Great Lakes atmosphere. Environ. Sci. Technol. 1999, 33, 1991-1995. (4) Gouin, T.; Thomas, G. O.; Cousins, I.; Barber, J.; Mackay, D.; Jones, K. C. Air-surface exchange of polybrominated diphenyl ethers and polychlorinated biphenyls. Environ. Sci. Technol. 2002, 36, 1426-1434. (5) Gouin, T.; Harner, T.; Daly, G. L.; Wania, F.; Mackay, D.; Jones, K. C. Variability of concentrations of polybrominated diphenyl ethers and polychlorinated biphenyls in air: implications for monitoring, modeling and control. Atmos. Environ. 2005, 39, 151-166. (6) Carlson, D. L.; Hites, R. A. Temperature dependence of atmospheric PCB concentrations. Environ. Sci. Technol. 2005, 39, 740-747. (7) Hoff, R. M.; Muir, D. C. G.; Grift, N. P. Annual cycle of polychlorinated biphenyls and organohalogen pesticides in air in Southern Ontario. 2. atmospheric transport and sources. Environ. Sci. Technol. 1992, 26, 276-283. (8) Fellin, P.; Barrie, L. A.; Dougherty, D.; Toom, D.; Muir, D.; Grift, N.; Lockhart, L.; Billeck, B. Air monitoring in the Arctic: results for selected persistent organic pollutants for 1992. Environ. Toxicol. Chem. 1996, 15, 253-261. (9) Wania, F.; Mackay, D. Tracking the distribution of persistent organic pollutants. Environ. Sci. Technol. 1996, 30, 390A-397A. (10) Horstmann, M.; McLachlan, M. S. Atmospheric deposition of semivolatile organic compounds to two forest canopies. Atmos. Environ. 1998, 32, 1799-1809. (11) Wania, F.; McLachlan, M. S. Estimating the influence of forests on the overall fate of semivolatile organic compounds using a multimedia fate model. Environ. Sci. Technol. 2001, 35, 582590. (12) Su, Y.; Wania, F. Does the forest filter effect prevent semivolatile organic compounds from reaching the arctic? Environ. Sci. Technol. 2005, 39, 7185-7193.

(13) Wania, F.; Haugen, J.-E.; Lei, Y. D.; Mackay, D. Temperature dependence of atmospheric concentrations of semivolatile organic compounds. Environ. Sci. Technol. 1998, 32, 10131021. (14) Hoff, R. M.; Brice, K. A.; Halsall, C. J. Nonlinearity in the slopes of Clausius-Clapeyron plots for SVOCs. Environ. Sci. Technol. 1998, 32, 1793-1798. (15) Li, N.; Wania, F.; Lei, Y. D.; Daly, G. L. A comprehensive and critical compilation, evaluation, and selection of physicalchemical property data for selected polychlorinated biphenyls. J. Phys. Chem. Ref. Data 2003, 32, 1545-1590. (16) Breivik, K.; Sweetman, A.; Pacyna, J. M.; Jones, K. C. Towards a global historical emission inventory for selected PCB congeners - a mass balance approach: 2. emissions. Sci. Total Environ. 2002, 290, 199-224. (17) Hung, H.; Halsall, C. J.; Blanchard, P.; Li, H. H.; Fellin, P.; Stern, G.; Rosenberg, B. Are PCBs in the Canadian Arctic atmosphere declining? Evidence from 5 years of monitoring. Environ. Sci. Technol. 2001, 35, 1303-1311. (18) Alaee, M.; Arias, P.; Sjo¨din, A.; Bergman, Å. An overview of commercially used brominated flame retardants, their applications, their use patterns in different countries/regions and possible modes of release. Environ. Int. 2003, 29, 683-689. (19) Hites, R. A. Polybrominated diphenyl ethers in the environment and in people: a meta-analysis of concentrations. Environ. Sci. Technol. 2004, 38, 945-956. (20) Su, Y.; Lei, Y. D.; Wania, F.; Shoeib, M.; Harner, T. Regressing gas/particle partitioning data for polycyclic aromatic hydrocarbons. Environ. Sci. Technol. 2006, 40, 3558-3564. (21) Strandberg, B.; Dodder, N. G.; Basu, I.; Hites, R. A. Concentrations and spatial variations of polybrominated diphenyl ethers and

(22)

(23)

(24)

(25)

(26)

(27)

other organohalogen compounds in Great Lakes air. Environ. Sci. Technol. 2001, 35, 1078-1083. Gouin, T.; Harner, T.; Blanchard, P.; Mackay, D. Passive and active air samplers as complementary methods for investigating persistent organic pollutants in the Great Lakes Basin. Environ. Sci. Technol. 2005, 39, 9115-9122. Harner, T.; Shoeib, M.; Diamond, M.; Ikonomou, M.; Stern, G. Passive sampler derived air concentrations of PBDEs along an urban-rural transect: spatial and temporal trends. Chemosphere 2006, 64, 262-267. Farrar, N. J.; Smith, K. E. C.; Lee, R. G. M.; Thomas, G. O.; Sweetman, A. J.; Jones, K. C. Atmospheric emissions of polybrominated diphenyl ethers and other persistent organic pollutants during a major anthropogenic combustion. Environ. Sci. Technol. 2004, 38, 1681-1685. Jaward, F. M.; Di Guardo, A.; Nizzetto, L.; Cassani, C.; Raffaele, F.; Ferretti, R.; Jones, K. C. PCBs and selected organochlorine compounds in Italian mountain air: the influence of altitude and forest ecosystem type. Environ. Sci. Technol. 2005, 39, 34553463. Meijer, S. N.; Steinnes, E.; Ockenden, W. A.; Jones, K. C. Influence of environmental variables on the spatial distribution of PCBs in Norwegian and U.K. soils: Implications for global cycling. Environ. Sci. Technol. 2002, 36, 2146-2153. Su, Y.; Wania, F.; Harner, T.; Lei, Y. D. Deposition of polybrominated diphenyl ethers, polychlorinated biphenyls, and polycyclic aromatic hydrocarbons to a boreal deciduous forest. Environ. Sci. Technol. 2007, 41, 534-540.

Received for review February 9, 2007. Revised manuscript received April 16, 2007. Accepted May 1, 2007. ES070334P

VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4661