Fractionation of Stable Isotope-Labeled Organic Pollutants as a

of perdeuterated/unlabeled compound mixtures. In contrast, isotope fractionation of 13C6-labeled SOCs was much lower. A field tracer-release study was...
0 downloads 0 Views 124KB Size
Environ. Sci. Technol. 2004, 38, 3871-3876

Fractionation of Stable Isotope-Labeled Organic Pollutants as a Potential Tracer of Atmospheric Transport Processes REBECCA M. DICKHUT,* TIRUPPONITHURA V. PADMA, AND ALESSANDRA CINCINELLI† School of Marine Science, College of William and Mary, Gloucester Point, Virginia 23062

To test the potential for using stable isotope fractionation to examine the atmospheric transport of semivolatile organic compounds (SOCs), we conducted simplified distillation experiments in the laboratory and a tracerrelease experiment using mixtures of stable isotope-labeled (D and 13C) and unlabeled SOCs. Perdeuterated phenanthrene and R-hexachlorocyclohexane were transported more slowly via air-water gas exchange in our laboratory experiments, resulting in significant isotope fractionation of perdeuterated/unlabeled compound mixtures. In contrast, isotope fractionation of 13C6-labeled SOCs was much lower. A field tracer-release study was then conducted by spiking a seawater retention pond with a mixture of D10-labeled, 13C2-labeled, and unlabeled phenanthrene and examining isotope fractionation of the mixture after airwater gas exchange. No preferential fractionation of D10vs 13C2-labeled phenanthrene was observed in the pond water; however, greater fractionation of D10- vs 13C2-labeled phenanthrene was observed in air samples collected within a 1-100 m radius of the pond. Thus, stable isotope tracers may provide a means of examining the atmospheric transport and air-earth exchange rates of POPs in an environmental realistic setting.

Introduction Many classes of semivolatile organic compounds (SOCs) including polychlorinated biphenyls (PCBs), dibenzo-pdioxins (PCDDs), dibenzofurans (PCDFs), and organochlorine pesticides are considered persistent organic pollutants (POPs) (1). POPs degrade slowly under typical environmental conditions and also undergo bioaccumulation and bioconcentration due to their lipophilic nature. These chemicals exert various sublethal effects such as endocrine disruption, immunological and neurological dysfunctions, reduced reproductive success, genotoxicity, and teratogenicity (2-6). Researchers have hypothesized that POPs volatilize at equatorial latitudes, undergo atmospheric transport, and condense in colder climates (1, 7-9). Evidence, such as inverted latitudinal profiles (i.e., low concentrations near source regions with increased concentrations in polar regions) for hexachlorocyclohexanes (HCHs) and hexachlorobenzene in seawater (10-11) and terrestrial vegetation * Corresponding author phone: (804)684-7247; fax: (804)684-7786; e-mail: [email protected]. † Present address: Department of Chemistry, University of Florence, Italy. 10.1021/es034282p CCC: $27.50 Published on Web 06/05/2004

 2004 American Chemical Society

(12-13) support this hypothesis. Likewise, the accumulation of pesticides in air, water, soil, and biota from Arctic, Antarctic, and Scandinavian regions remote from any primary source (14-16) supports the idea of global atmospheric transport of these xenobiotic compounds. The physicochemical characteristics of an SOC including subcooled liquid vapor pressure, octanol-air partition coefficient, and condensation temperature affect the rate and extent of its atmospheric transport (1, 9). Consequently, it is proposed that specific compounds from various classes of SOCs deposit in different geographical regions during global atmospheric transport poleward (1, 9). This geographic separation of moderately volatile (subcooled liquid vapor pressure PL ) 10-4 to 1 Pa) SOCs during global atmospheric transport is similar to fractional distillation, with highly volatile components succumbing readily to gas-phase transport, while less volatile components undergo atmospheric transport to a lesser extent. Higher latitudes will therefore preferentially accumulate the more volatile fraction of a contaminant mixture, whereas lower latitude source regions will be left with relatively greater concentrations of the nonvolatile components. Such fractionation has been observed for PCBs in the Baltic Sea region (17). Similarly, we hypothesize that SOCs containing heavy stable isotopes are atmospherically transported to a lesser extent than those containing light isotopes, since heavier isotopes render substances less volatile. This phenomenon has been established for meteoric water (18, 19). In this case, water vapor generated by evaporation of ocean water in low latitude equatorial regions becomes increasingly depleted in heavy isotopes (18O, D) during atmospheric transport poleward. This well-documented example of latitudinal isotopic fractionation demonstrates the potential for using compound-specific stable isotopes for tracking the atmospheric transport of a material. Here we report the results of laboratory experiments and a field tracer study conducted using isotopically labeled and unlabled SOCs to investigate the feasibility of using stable isotope fractionation to examine the atmospheric transport and air/earth exchange of SOCs. We selected phenanthrene and R-hexachlorocyclohexane (R-HCH) as model SOCs, since both are classified as possessing “relatively high mobility” according to Wania and Mackay (1) with the potential for long-range atmospheric transport.

Experimental Section Laboratory Experiments. Air-water exchange of SOCs was studied in the laboratory using a series of temperaturecontrolled (40, 20, 0 °C) gas washing bottles containing 5001000 mL of artificial seawater (salinity ) 35 ppt), which were connected to each other via an air stream. A humidifier was placed upstream of the first gas washing bottle to prevent volatilization of water, and a sorbent trap was attached at the end of the system in order to collect any gas-phase SOCs that were present in the outflow from the last (0 °C) bottle. At the start of each experiment a labeled and unlabeled POP pair (e.g., D10phenanthrene and phenanthrene), were spiked into the water in the first (40 °C) bottle at a concentration lower than one-half solubility. The bottles were then sealed except for the air inlet and exit paths, and the air stream was turned on. SOC migration through the system was monitored by subsampling the water in each of the gas washing bottles. Water samples (5-50 mL) were extracted three times with 2-5 mL of hexane after the addition of a surrogate standard (D10anthracene). Excess water was then removed from the extracts by passing the hexane over Na2SO4. After the addition VOL. 38, NO. 14, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3871

of an internal standard (D10acenaphthene) and solvent reduction under purified N2, both the isotopically light and heavy compounds were quantified by gas chromatography/ electron impact mass spectrometry (GC/MS) on a 30 m × 0.25 mm i.d. DB-XLB column (J&W Scientific) using selective ion monitoring. From these data the ratio of the isotopically heavy (labeled) and light (unlabeled) POPs was calculated for each reservoir in order to determine if there was fractionation during air/water exchange. Experiments (between 8 and 21 days) were conducted using D10phenanthrene, 13C6phenanthrene, D6R-HCH, and 13 C6R-HCH, (Cambridge Isotopes, MA) along with the corresponding unlabeled form of each SOC. In the case of 13 C6R-HCH alone, a replicate experiment was conducted using a system designed to increase the number of air/earth exchanges by including diatomaceous earth columns (50 cm × 2.5 mm i.d.) after each seawater bottle. Sampling was conducted by drawing a sample out of the air inlet path (once air flow was stopped) using a syringe. The recovery of spiked compounds calculated using a mass balance ranged from 66% to >90% for the experiments reported here. Tracer-Release Study. The tracer-release experiment was conducted on the campus of the Virginia Institute of Marine Science. A seawater retention pond (∼40 m3, 45 m2) was spiked with a mixture of equal amounts of perdeuterated (D10-labeled), 13C2-labeled, and unlabeled phenanthrene by first dissolving the chemicals in 100 mL of methanol and diluting the methanol mixture to 2 L with distilled water. The methanol/water mixture was then released into the pond by submersing and emptying the jars containing the spike mixture and then mixed into the pond using gentle agitation. The nominal concentration of each compound in the pond following spiking was 100 ng/L; the aqueous solubility of phenathrene is 1 mg/L (20). Following spiking the pond was aerated to facilitate air/ water exchange using an air compressor connected to two large aquarium stones with Teflon tubing. Water samples were collected daily from the pond by submersing 4 L glass bottles just under the surface. Integrated (24 h) air samples were also collected at distances between 1 and 100 m of the pond using semiquantitative air samplers consisting of blower motors connected to aluminum ducts that held precleaned (21) polyurethane foam (PUF) plugs. Air flow through these samplers was not calibrated but was sufficient to obtain detectable quantities of phenanthrene on the PUF plugs in 24 h. Since our objective was to measure isotope ratios in air rather than absolute concentrations, it was not necessary to sample air quantitatively. Pond water samples (3-15 L) were extracted with hexane in separatory funnels, and PUF plugs were Soxhlet extracted with acetone and petroleum ether (24 each) after the addition of D10anthracene as a surrogate standard. All sample extracts were then reduced in volume to 1 mL by rotary evaporation and cleaned using silica column chromatography (21). The cleaned extracts were subsequently reduced in volume using rotary evaporation, and D10acenaphthene was added as an internal standard. After further volume reduction with purified N2, the extracts were analyzed by GC/MS as described above. Measured amounts of 13C2phenanthrene were corrected for natural abundance levels using the following equation 13

C2phenanthrenetracer ) 13C2phenanthrenemeasured 0.012phenanthrenemeasured

Data Reporting. Isotope ratios for both the laboratory and field tracer-release experiments are reported in del (δ) notation using the following equation

δ ) [[Rsample - Rinitial]/Rinitial]1000 3872

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 14, 2004

FIGURE 1. Loss of SOCs from 40 °C seawater due to air-water exchange in the laboratory experiments. where Rsample is the heavy/light SOC ratio in the sample and Rinitial is the initial heavy/light SOC ratio in water at the start of the experiment. Thus, negative δ values indicate that Rsample < Rinitial and that the SOCs are fractionated in the sample being depleted in the isotopically heavy compound compared to the initial mixture. Likewise, positive δ values indicate that the SOCs are fractionated in the sample, but in this case the isotopically heavy compound is enriched compared to the initial mixture.

Results and Discussion In all of the laboratory experiments the change in concentration with time in each bottle showed the same pattern. An exponential decrease in concentration for both labeled and unlabeled SOCs was always observed in the first (40 °C) bottle (Figure 1). This decline was associated with a concurrent increase in concentrations of both the unlabeled and labeled compounds in the second (20 °C) bottle, followed by a slow release and steadily increasing concentrations of the SOCs in the third (0 °C) bottle. The concentration profiles of the SOCs in the laboratory experiments demonstrate that atmospheric transport of SOCs via the gas phase occurred in the system similar to that predicted by Wania and Mackay (1) for chromatographic-like transfer of “relatively high mobility” POPs from warm to cold regions. The laboratory experiments further demonstrate that equilibrium was likely reached for air/water partitioning of the SOCs in the first bottle. Mackay et al. (20) demonstrated that plots of natural log water concentration versus time are linear for bubble stripping experiments in which air/water equilibrium is reached. This was the case in all of our laboratory experiments (Figure 1). In addition, the slopes of these plots are directly related to the Henry’s Law constants

FIGURE 2. Rayleigh plots for fractionation of SOCs in seawater following air-water exchange in the laboratory experiments at 40 °C.

FIGURE 3. Changes in δD10 and δ13C6 versus time due to sequential air/water exchange of phenanthrene and isotopically labeled phenanthrenes in laboratory experiments.

of the compound (20). For our experiments, D10phenanthrene and D6R-HCH had significantly lower slopes (p < 0.025) for plots of natural log concentration versus time than phenanthrene and R-HCH, respectively, indicating significantly lower Henry’s Law constants for D-labeled versus unlabeled SOCs. In contrast, the slope of natural log concentration versus time for 13C6R-HCH was not significantly different (p > 0.10) from R-HCH, indicating no significant difference in the air/water partition coefficients for 13C6R-HCH and R-HCH. However, the p-value for 13C6phenanthrene versus phenanthrene was marginally significant (p ) 0.053), indicating perhaps a slightly lower Henry’s Law constant for 13C phenanthrene versus phenanthrene. 6 The laboratory experiments can also be used to determine if isotopically heavy SOCs (i.e., D- and 13C-labeled SOCs) volatilize more slowly and are enriched in water compared to the corresponding unlabeled SOCs. In a Rayleigh distillation, a process in which the isotopic composition of a material undergoing reaction varies as a function of the extent of the reaction (22), a plot of ln[(δ + 1000)/(δi + 1000)] versus ln F, where F is the fraction of compound remaining in solution and δi is the initial δ value of the SOC mixture in the water, should yield a linear plot with a slope of R - 1, where R is the fractionation factor the process (23). Rayleigh distillation was demonstrated in our laboratory experiments using D10-enriched phenanthrene (Figure 2). In this case, the air-water enrichment factor ( ) (R - 1) × 1000) (23) for D10phenanthrene is -185‰, corresponding to significant enrichment of D10phenanthrene in the water and depletion in the gas phase. Such isotopic fractionation is theoretically predicted for volatilization of isotopically heavy versus light compounds; however, “inverse” isotope fractionation (en-

richment of the isotopically heavy component in the vapor phase) is usually observed during vaporization of hydrocarbons from the pure liquid, which is attributed to lower intermolecular binding energies of isotopically heavy molecules in the liquid phase (23). Therefore, it is likely that the molecular mass effects dominate the air/water exchange of D10phenanthrene rather than intermolecular binding energies. Rayleigh plots for D6R-HCH, 13C6phenanthrene, and 13C R-HCH also appeared linear; however, linear regressions 6 of these plots were only significant at the 89-93% confidence levels (Figure 2). In general, the other laboratory experiments support the D10phenanthrene results with fractionation and enrichment factors indicative of preferential enrichment of the isotopically heavy compounds in the water. Enrichment factors for the 13C6-labeled compounds were also lower than for the D-labeled compounds as illustrated by the lower slopes for plots of ln[(δ + 1000)/(δi + 1000)] versus ln F (Figure 2). Enrichment of D10phenanthrene, and to a lesser extent 13C phenanthrene, in the source bottle in the laboratory 6 experiments is also illustrated in plots of δ versus time (Figure 3, top). Note that δD increases with time at 40 °C, which is a direct effect of the significantly lower Henry’s Law constant for D10phenanthrene compared to phenanthrene. 13C phenanthrene also becomes enriched in the source 6 water but only after 9 days, indicative of the smaller difference in the Henry’s Law constants of phenanthrene and 13C phenanthrene. Fractionation of D phenanthrene was 6 10 also evident with subsequent transfer through the experimental system (i.e., at 20 and 0 °C) with the isotopically heavy compound becoming increasingly depleted relative to the unlabeled phenanthrene (Figure 3). This implies that less of VOL. 38, NO. 14, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3873

FIGURE 4. Decrease in phenanthrene and isotopically labeled phenanthrene concentrations in pond water with time during the tracer-release experiment. the perdeuterated phenanthrene was transported through the system and retained in the water during sequential air/ water exchange and/or that a greater proportion of unlabeled phenanthrene was retained in the second and third lower temperature bottles. In contrast, 13C6phenanthrene was initially enriched relative to the unlabeled phenanthrene with subsequent air/water exchanges and decreasing temperature. The reason for this conflicting trend, also observed in a replicate experiment, is unclear but may indicate a greater decrease in the Henry’s Law constant of 13C6phenanthrene with temperature compared to that of phenanthrene, leading to larger initial accumulation in the low-temperature bottles. However, with time this trend was reversed. At the conclusion of this experiment in which both deuterium and 13C-labeled phenanthrene were included, D10- and 13C6phenanthrene were both enriched at the source (40 °C bottle) but the enrichment of D10phenanthrene (δ ) 1358‰) was about twice that of 13C6phenanthrene (δ ) 626‰). The greater fractionation of deuterium-labeled phenanthrene compared to 13C-labeled phenanthrene is similar to that observed for D compared to 18O in the global distillation of meteoric water (18, 19). Our laboratory experiments supported the hypothesis that isotope fractionation occurs during air/water gas exchange of SOCs. Moreover, based on the results of these experiments, we hypothesized that D-labeled SOCs would show greater fractionation than 13C-labeled SOCs during volatilization and atmospheric transport in the environment. Specifically, we hypothesized that D10phenanthrene would fractionate (become depleted) due to retarded air/earth exchange relative to phenanthrene as well as due to mixing with phenanthrene in background air. In contrast, 13C2phenanthrene would exhibit depletion largely due to mixing with phenanthrene in background air. Note that in this case 13C2phenanthrene was selected rather than 13C6phenanthrene in order to minimize isotopic fractionation of this tracer. To test our hypothesis, we conducted a tracer-release experiment in which phenanthene, D10phenanthrene, and 13C2phenanthrene were spiked into a seawater retention pond in approximately equal amounts, and fractionation of these compounds was monitored in both the pond water and air within 100 m following air-water exchange and atmospheric transport. Consistent with the results of the laboratory experiments, over the course of the tracer-release experiment, the concentration of phenanthrene and the isotopically labeled phenanthrenes decreased exponentially in the pond water (Figure 4). Variability in replicate measurements was initially large (33-36%), indicating that the tracers were not well mixed throughout the pond at the start of the experiment; however, variability in the pond water concentrations of the tracers decreased to 17-19% within 24 h. Unlike the laboratory experiments, the concentrations of the isotopically labeled phenanthrenes decreased more rapidly in the pond 3874

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 14, 2004

FIGURE 5. Water/air fugacity ratios for phenanthrene during the tracer-release experiment. Error bars depict the propagated error of (42% in the calculated values. water compared to phenanthrene. Rate constants for loss of D10phenanthrene and 13C2phenanthrene from the pond were 0.873 ( 0.09 d-1 (r2 ) 0.96, p ) 0.0006) and 0.889 ( 0.094 d-1 (r2 ) 0.96, p ) 0.0007), respectively, whereas the rate constant for loss of phenanthrene from the pond was 0.421 ( 0.116 d-1 (r2 ) 0.77, p ) 0.022). The slower loss of phenanthrene from the pond compared with the isotopically labeled compounds may be due in part to addition of phenanthrene to the pond in conjunction with a rain event that occurred between sampling on days 3 and 4. Nonetheless, loss of phenanthrene from the pond was significantly slower from day 4 through day 6, during which time no wet deposition of phenanthrene occurred. Since the seawater retention pond had a sandy bottom, the sediments were not disturbed, and the pond was aerated throughout the experiment, air/water rather than sediment/ water exchange of the tracers was promoted. Thus, sorptive losses of the phenanthrene(s) were expected to be minimal, and atmospheric exchange of the SOCs was expected to dominate; however, other processes such as microbial degradation may also play a role in removing the phenanthrene(s) from the pond. Using an average gas-phase concentration (24.4 ( 6.7 ng/m3) previously measured for phenanthrene in the area (24), the measured water concentrations of phenanthrene, and a Henry’s Law constant of 3.98 Pa m3mol-1 (20), the water/air fugacity ratio (fw/fa; 25) for phenanthrene was calculated (Figure 5). The fw/fa for phenanthrene during the first 2 days of the tracer-release experiment was >1, indicating the potential for net volatilization of phenanthrene from the pond into the air. However, fw/fa was e1 after the second day of the tracer-release experiment, indicating a low potential for volatilization or the potential for net absorption of phenanthrene from the air. If losses of phenanthrene from the pond were entirely due to air/water exchange, the system would proceed toward equilibrium (fw/fa ) 1). Thus, the tendency toward fw/fa e 1 may indicate that other loss processes such as microbial degradation may also play a role in removing the phenanthrene from the pond. In contrast to phenanthrene, water concentrations of the isotopically heavy compounds remained steady only between days 3 and 4 (Figure 4), during which time volatilization of these compounds may have been balanced by wet deposition input due to the rain event. The continued loss of D10- and 13C phenanthrene from the pond from days 4 to 6 indicates 2 that the isotopically heavy compounds were not at equilibrium and that volatilization of these compounds continued to occur. However, as with phenanthrene, other loss processes such as microbial degradation may also play a role in removing the isotopically heavy phenanthrene(s) from the pond. One result of the slower loss of unlabeled phenanthrene from the pond compared with D10- and 13C2phenanthrene is fractionation whereby the pond water becomes increasingly depleted in isotopically heavy phenanthrenes (δ < 0, Figure

FIGURE 7. Change in D10phenanthrene (δD10) and 13C2phenanthrene (δ13C2) relative to phenanthrene in air with distance from the pond source during the tracer-release experiment.

FIGURE 6. Change in D10phenanthrene (δD10) and 13C2phenanthrene (δ13C2) relative to phenanthrene in pond water and air during the tracer-release experiment. Pond water symbols: day 1 (W1, upper right) through day 6 (W6, lower left). Air symbols: 1-7 correspond to sampling day; (A) 1 m, (B) 10 m, (C) 100 m from the pond. 6, top). This is in contrast to the laboratory experiments in which the isotopically heavy compounds became enriched in the source water. Moreover, no preferential fractionation of D10- vs 13C2phenanthrene was observed in the pond water, as both of these compounds were lost from the pond at similar rates. As with the pond water in the tracer-release experiment, air samples collected within 1-100 m were depleted in the isotopically heavy phenanthrenes (Figure 6, bottom). However, in contrast to the pond water, only air samples collected within 1 m of the pond during the first few days of the tracerrelease experiment exhibited similar fractionation of both D10- and 13C2phenanthrene. Air samples collected from 10 to 100 m around the pond and later in the experiment tended to be substantially more depleted in D10- compared to 13C2phenanthrene. The overall much greater depletion of the isotopically heavy phenanthrenes in air compared to pond water (e.g., note the axes on the top compared to the bottom graph of Figure 6) is presumably due to an increase in unlabeled phenanthrene from outside sources in air. However, although additional sources of phenanthrene to the local atmosphere would account for the greater depletion of both heavy isotopes in the air compared to water samples, this would not create preferential fractionation of D10phenanthrene compared to 13C2phenanthrene as observed for many of the air samples. Thus, the greater loss of D10phenanthrene from the atmosphere compared with 13C phenanthrene must be due to differential rates of air2 earth exchange (e.g., partitioning to terrestrial surfaces surrounding the pond including vegetation and surface films) or differential rates of transformation processes. Distance from the pond source was the dominant factor controlling δD10 and δ13C2 of phenanthrene in air through the first 4 days of the tracer-release experiment (Figure 7). For example, δ13C2 was significantly (p < 0.05) correlated with distance from the pond on days 2 and 5, whereas δD10

FIGURE 8. Change in D10phenanthrene (δD10) and 13C2phenanthrene (δ13C2) relative to phenanthrene in air at a sampling site located 100 m downwind of the pond during days 5-7 of the tracer-release experiment. was significantly (p < 0.05) correlated with distance from the pond on days 2-4. In both cases, δ values decreased with distance, indicating greater depletion of isotopically heavy phenanthrene with increased distance from the pond. However, after day 5, no significant differences in δD10 and δ13C2 of phenanthrene in air were observed relative to distance from the pond source. Nonetheless, significant changes in δ values were observed at some sites after day 5 of the experiment. In particular, phenanthrene δD10 and δ13C2 values in air significantly (p < 0.05) increased with time at a site located 100 m away from the pond source after day 5 (Figure 8). This site was predominantly downwind of the experimental area during the latter half of the experiment, and air samples at this site exhibited consistent increases in isotopically heavy phenanthrene with time from days 5 to 7. Moreover, the rate of increase in δ13C2 at this downwind site during the latter half of the experiment was ∼8× higher than that for δD10, illustrating the slower transport of D10 -relative to 13C2-labeled phenanthrene during air/earth exchange. Our objective in conducting this research was to initiate development of an isotopic tracer technique that could be used to examine the atmospheric transport and fate of POPs. The differential fractionation of D10- vs 13C2-labeled phenanthrene during atmospheric transport demonstrated through this research provides a means of distinguishing atmospheric dispersion of an SOC from losses due to air/earth exchange or transformation in the environment. Thus, the experiments described above indicate that isotopic tracers may indeed be useful for evaluating the atmospheric transport and fate of POPs; however, further development of the method is required. For example, lower detection limits will be needed to detect the isotopic tracers at greater distances from the VOL. 38, NO. 14, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3875

source. This may be accomplished using stable isotope ratio mass spectrometry rather than electron impact mass spectrometry as was done here. Moreover, greater sampling resolution including samples of terrestrial surfaces would yield information on the influence of meteorological conditions on the atmospheric transport and exchange rates of SOCs with terrestrial surfaces. Finally, measurement and tracking of natural abundance levels of isotopes associated with POPs would be a more widely applicable technique for evaluating the environmental transport and fate of these compounds as opposed to following the distribution of isotopically enriched tracers as was done here. Measurement of natural abundance levels of isotopes in POPs could also potentially prove to be a useful technique for tracking the long-range atmospheric transport of POPs. However, it is important to note that the results of the current study may not be directly applicable to fractionation of natural abundance levels of POPs in the environment. Fractionation of the isotopically enriched compounds used in this study is a result of the large mass differences (2-10 amu) between the compounds, which is not the case for natural abundance level isotopes associated with POPs. Consequently, additional research should also focus on evaluating the fractionation of natural abundance levels of isotopes associated with POPs during atmospheric transport.

Acknowledgments Thanks to Robert McDaniel II, Michele Cochran, Daymond E. Levine, and Melissa Schinck who assisted with the experiments. This work was partially funded by the National Science Foundation Award No. 0074948. VIMS contribution number 2602.

Literature Cited (1) Wania, F. W.; Mackay, D. Environ. Sci. Technol. 1996, 30, 390396. (2) Gardner, G. R.; Pruell, R. J.; Malcom, A. R. Mar. Environ. Res. 1992, 34, 59-63. (3) Dodson, S. I.; Hanazato, T. Environ. Health Pespect. 1995, 103, 7-11.

3876

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 14, 2004

(4) Facemire C. F.; Gross, T. S.; Guillette, L. J., Jr. Environ. Health Pespect. 1995, 103, 79-86. (5) Lahvis, G. P.; Wells, R. S.; Kuehl, D. W.; Stewart, J. L.; Reinhart, H. L.; Via, C. S. Environ. Health Pespect. 1995, 103, 67-72. (6) Sleidernik, H. M.; Everaarts, J. M.; Goksoyr A.; Boon, J. P. Environ. Toxicol. Chem. 1995, 14, 679-687. (7) Rappe, C. In Ecological Problems of the Circumpolar Area; International Symposium, Lulea, Sweden, June 1971; Bylund, E.; Linderholm, H.; Rune, O., Eds.; Norbottens Museum: Lulea, Sweden, 1974; pp 29-32. (8) Goldberg, E. D. Proc. R. Soc. London, Ser. B. 1975, 198, 277289. (9) Wania, F. W.; Mackay, D. Ambio 1993, 22, 10-18. (10) Tanabe, S.; Tatsukawa, R.; Kawano, M.; Hidaka H. J. Oceanogr. Soc. Jpn. 1982, 38, 137-147. (11) Iwata, H.; Tanabe, S.; Sakal, N.; Tatsukawa, R. Environ. Sci. Technol. 1993, 27, 1080-1098. (12) Calamari, D.; Bacci, E.; Focardi, S.; Gaggi, C.; Morosini, M.; Vighi, M. Environ. Sci. Technol. 1991, 25, 1489-1495. (13) Simonich, S. L.; Hites, R. A. Science 1995, 269, 1851-1854. (14) Larsson, P.; Okla, L.; Woin, P. Environ. Sci. Technol. 1990, 24, 1599-1601. (15) Bidleman, T. F.; Falconer, R. L.; Walla, M. D. Sci. Total. Environ. 1995, 160/161, 55-63. (16) Weber, K.; Goerke, H. Chemosphere 1996, 33, 377-392. (17) Agrell, C.; Okla, L.; Larsson, P.; Backe, C.; Wania, F. Environ. Sci. Technol. 1999, 33, 149-1156. (18) Craig, H. Science 1961, 133, 1702-1703. (19) Ingraham, N. L.; Taylor, B. E. J. Hydrol. 1986, 85, 183-197. (20) Mackay, D.; Shiu, W. Y.; Sutherland, R. P. Environ. Sci. Technol. 1979, 13, 333-337. (21) Dickhut, R. M.; Gustafson, K. E. Mar. Pollut. Bull. 1995, 30, 385-396. (22) Libes, S. An Introduction to Marine Biogeochemistry; John Wiley and Sons: New York, 1992; p 73. (23) Wang, Y.; Huang, Y. Appl. Geochem. 2003, 18, 1641-1651. (24) Gustafson, K. E. Ph.D. Dissertation, The College of William and Mary, Virginia Institute of Marine Science, 1996. (25) Falconer, R. L.; Bidleman, T. F.; Gregor, D. J. Sci. Total Environ. 1995, 160/161, 65-74.

Received for review March 30, 2003. Revised manuscript received April 9, 2004. Accepted April 29, 2004. ES034282P