Environ. Sci. Technol. 2004, 38, 4474-4483
Using Passive Air Samplers To Assess Urban-Rural Trends for Persistent Organic Pollutants. 1. Polychlorinated Biphenyls and Organochlorine Pesticides T O M H A R N E R , * ,† M A H I B A S H O E I B , † MIRIAM DIAMOND,‡ GARY STERN,§ AND BRUNO ROSENBERG§ Meteorological Service of Canada, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4, Department of Geography, University of Toronto, Toronto, Ontario, Canada M5S 3G3, and Freshwater Institute, Fisheries and Oceans Canada, Winnipeg, Manitoba, Canada R3T 2N6
Passive air samplers were used to investigate urbanrural differences of polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs) over an integrated time period. Samplers consisting of polyurethane foam (PUF) disks and semi-permeable membrane devices (SPMDs) were housed in protective chambers and deployed at six sites for a 4 month duration in the summer of 2000. The sampling transect originated in downtown Toronto and extended ∼75 km northward into a rural region. Results for the two types of samplers agreed well with one another. Higher blank levels were encountered for the SPMDs, especially for the OCPs, whereas blanks were very low for the PUF disks. Passive sampler-derived air concentrations were consistent with previous measurements of PCBs and OCPs in the region. The largest urban-rural gradient was observed for PCBs (∼5-10 times). Chlordanes also showed an urban-rural gradient, possibly reflecting past usage of chlordane on residential lawns and emissions from treated house foundations. Other OCPs exhibited a ruralurban gradient (dieldrin, endosulfan 1, and DDT isomers), which was attributed either to off-gassing from previously treated agricultural soils (dieldrin and DDTs) or to continued usage in agriculture (endosulfan 1). The results of this study demonstrated the feasibility of using such devices to determine air concentrations of persistent organic pollutants (POPs) and to assess their spatial distribution for timeintegrated samples. Data such as this is essential for: model validation and for process research and addressing international monitoring strategies on POPs.
Introduction Many semi-volatile organic compounds (SOCs) continue to persist in the environment long after their use has been discontinued. Furthermore, many of these SOCs are also toxic, bioaccumulative, and susceptible to long-range atmospheric transport. International bodies such as the UN-ECE (United Nations Economic Commission for Europe) and UNEP * Corresponding author telephone: (416)739-4837; fax: (416)7395708; e-mail:
[email protected]. † Environment Canada. ‡ University of Toronto. § Fisheries and Oceans Canada. 4474
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004
(United Nations Environment Program) have designated compounds that meet all of these criteria to be persistent organic pollutants (POPs) and have established priority lists for the elimination of the “worst” POPs. Processes have also been implemented for prioritizing additional chemicals. Chemicals currently designated include industrial chemicals such as the polychlorinated biphenyls (PCBs) and several organochlorine pesticides (OCPs). For the most part, these chemicals are no longer used in industrialized regions; however, they continue to show high residues in environmental compartments, including the atmosphere. It is appropriate therefore to assess the spatial concentrations of POPs in the atmosphere. This will serve to improve our understanding of the role of atmosphere as a transport conduit and to assess risk associated with exposure and human health concerns. Such exposure-related health concerns are especially relevant in the urban environment that is a source region of some POPs (e.g., PCBs). In recent years, levels and trends in POP concentration have been investigated through various national or multinational air sampling networks; for example, the Canada-US, Integrated Atmospheric Network (IADN) (1, 2), which has operated in the Great Lakes basin since 1992. Data from this and similar networks has been invaluable for determining long-term trends, for identifying chemicals that behave as POPs, and for assessing predictive capability of chemical fate models (2, 3). The limitation of such networks is that they have long intervals between sample events (e.g., 12 d in the case of IADN). This raises the possibility that, due to fluctuating air concentrations, the two samples collected in one month may not accurately represent the month as a whole. Relatively, few investigations have been published for typical urban areas (e.g., residential), where the majority of the population resides. This has resulted in limited information required for conducting human risk assessments. For instance, there is no published information on air concentrations of PCBs or OCPs in the city of Toronto over the past decade. Insights to urban sources has come from the work of Gingrich et al. (4), who have sampled atmospherically derived organic film (5, 6) on outdoor windows and observed higher loadings and film thickness in urban samples for several classes of POPs including PCBs, OC pesticides, and most recently, polybrominated dipheny ethers (PBDEs) (7). Wong et al. (8) have used experimental and forest soils to show a similar trend for polycyclic aromatic hydrocarbons (PAHs). Cities are also important because they are a focus and hence emission source of many industrial chemicals which are then susceptible to transport to other regions. Cities such as Toronto that are situated on the shores of the Great Lakes may also contribute significantly to loadings to the lakes through atmospheric deposition. These influences have been well demonstrated by the AEOLOS (Atmospheric Exchange Over Lakes and Oceans) project and measurements conducted in Chicago on the shore of Lake Michigan (9). The general lack of air concentration measurements for POPs in urban areas is partly attributed to difficulties and high costs associated with conventional high-volume air sampling. There is an obvious incentive for developing simple, cost-effective integrative samplers capable of assessing air burdens of POPs in urban areas. It is envisioned that such an effort would target information gaps and complement the results obtained through existing, active air surveillance networks. In this study, two types of passive air samplerss semipermeable membrane devices (SPMDs) and polyurethane foam (PUF) disksswere tested. Samplers were de10.1021/es040302r CCC: $27.50
2004 American Chemical Society Published on Web 08/03/2004
FIGURE 1. Schematic and photo of the passive air sampling chamber. The map shows sampling locations on an urban-rural transect. ployed at eight sites along an urban-rural gradient originating in the city of Toronto and extending approximately 75 km northward. The operation of the samplers is described, and passive sampler-derived air concentrations are calculated, evaluated, and used to assess urban-rural differences for several categories of POPs.
Theory As the name suggests, passive air samplers operate without the aid of a pump. They consist of an accumulating medium that has a high retention capacity for the target analytes. The uptake of POPs by SPMDs and PUF disks has been described in an earlier study (10) and was shown to be air-side controlled and thus a function of the air-side mass transfer coefficient (MTC), kA. This parameter is a weak function of temperature but strongly related to wind speed, with higher values at higher flow rates. This calibration of the SPMDs and PUF was conducted in a still indoor environment, and the resulting values of kA were comparable to values for diffusivity in air. During the field deployment, this low-wind environment was preserved by housing the samplers in protective chambers (Figure 1). In addition to dampening the wind speed effect on uptake, the chambers also protect the sampling medium from precipitation, coarse aerosols, and ultraviolet radiation, which may degrade target compounds. For POPs that are partitioned between the gas phase and associated with atmospheric particulate matter, the samplers have been shown to sample mainly the gaseous component (11). The maximum uptake of the SPMD or PUF disks is defined by their retention capacity or passive sampler medium-air partition coefficient, KPSM-A. Previous work has shown that KPSM-A is well correlated to the octanol-air partition coefficient, KOA, which is known for many POPs as a function of temperature including PCBs (12-14) and OCPs (15). The complete uptake profile is described by
CPSM ) KPSM-ACAIR(1 - exp -[(APSM)/(VPSM) (kA/KPSM-A)]t) (1) where CPSM and CA are concentrations (mass cm-3) of analyte in the passive sampler medium and air, respectively; APSM and VPSM are the planar surface area (cm2) and volume (cm3) of the sampling medium; kA is in cm d-1; and t is the exposure time (in d) (10). Values for the various calibration parameters
TABLE 1. Calibration Parameters (10) for Calculating Equivalent Sample Volumes for PUF Disks and SPMDs Using Eq 2a d-1)
kA (cm [cm VPSM (cm3) APSM (cm2) KPSM
a
s-1]
PUF disk
SPMD
9500 [0.11] 210 370 log KPUF-A ) FPUF(0.6366 log KOA - 3.1774)
11 200 [0.13] 8.5 495 log KSPMD-A ) FSPMD(0.8113 log KOA - 4.8367)
See text for definition of eq 2 parameters.
including the relationship between KPSM-A and KOA are presented in Table 1 for PUF disks and SPMDs. It is also possible to interpret KPSM-A as the equivalent volume of air that contains the same mass of analyte as 1 unit volume of passive sampling medium under equilibrium conditions (i.e., KPSM ) VAIR/VPSM ) CPSM/CAIR). Thus by analogy and replacing terms in eq 1, the equivalent air sample volume is
VAIR ) KPSM-AVPSM(1 - exp -[(APSM)/(VPSM) (kA/KPSM-A)]t) (2) In Figure 2, the equivalent air sample volume is plotted as a function of time for a range of KOA. Table 1 lists the constants that were used in eq 2. For chemicals with KOA values larger than 108.5 to 109, the sampling rate remains linear over the first 100 d and is mainly defined by the airside MTC. For compounds of lower KOA, the PUF disk and SPMD become saturated in less than 100 d. The time required for a chemical to reach saturation (or to attain its maximum equivalent air volume) is inversely proportional to KPSM (or KOA). The estimation of equivalent air volume can potentially be improved with the addition of “depuration” compounds to the sampling medium prior to deployment. These are isotopically labeled chemicals or unlabeled chemicals that do not exist in the atmosphere. Ockenden et al. (16) describe the use of “permeation reference compounds” (PRCs) for SPMDs. Because uptake is air-side controlled, the depuration of these chemicals can be related to an effective sample volume. Although depuration compounds (or PRCs) were VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4475
FIGURE 2. Time and KOA dependence of equivalent sample air volume for PUF disk samplers; calculated using eq 2 and parameters listed in Table 1. not used in this study, that took place during the early development of the samplers, they were used in a subsequent study with PUF disks. The results from that study (17) confirm the accuracy of the PUF disks sampling rates used here (10).
Methodology Sample Collection. PUF disks (14 cm diameter; 1.35 cm thick; surface area, 365 cm2; mass, 4.40 g; volume, 207 cm3; density, 0.0213 g cm-3; PacWill Environmental, Stoney Creek, ON) and SPMDs (EST Labs, St. Joseph, MO) were housed in stainless steel domed chambers (Figure 1). Samplers were deployed over the period July-October 2000 for approximately 120 d at seven sites along an urban rural transect extending ∼75 km north of downtown Toronto. The SPMD and PUF disks were held in SPMD “spider” carriers (EST Labs, St. Joseph, MO). A field blank was collected by installing and removing a PUF disk and SPMD. PUF disks and SPMDs were handled using solvent-rinsed tongs. Sampling chambers were prewashed and solvent-rinsed with acetone prior to installation of the passive sampling media. Seven sites were selected along an urban-rural transect (Figure 1). Three sites were located in south Toronto (Junction Triangle, Gage Building, South Riverdalesurban 1-3, respectively) and are classified as urban, high-density residential/industrial. Another urban site in north Toronto (Downsview) is considered medium-density, residential/ industrial (urban 4). Further north, two suburban sites (Richmond Hill and Aurorassuburban 1 and 2, respectively) are characterized as low-density residential/industrial. Duplicate SPMDs were deployed at the Richmond Hill site. The last site on the transect (Egbert) was rural and situated in an agricultural/farming region. PUF disks were not deployed at the urban site in South Riverdale. Mean temperatures at each site were assessed from archived data available through the National Air and Water Monitoring Activities (NatMon) Archive (http://ib.tor.ec.gc.ca/natmon/DataPub/historical_data/). Meterological stations nearest the sampling site were selected for hourly temperature data. Mean temperatures for the sample period ranged from 17.8 °C for Toronto to 15.5 °C at Egbert. Because this temperature range was fairly small, a common temperature of 17 °C was used for all calculations. Previous analysis has shown that the PUF disk uptake rate has a weak dependence on temperature (10). PUF disks were precleaned with water and then extracted by Soxhlet for 24 hstwice with acetone and then once with petroleum ether (PE). Prior to and after sample collection, 4476
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004
PUF disks were stored in solvent-rinsed, 1000-mL glass jars having Teflon-lined lids. PUF samples were stored cool and in the dark prior to analysis. SPMDs were kept frozen prior to and after deployment and stored in aluminum cans. Analysis. PUF disk samples were extracted in a Soxhlet extraction apparatus for 24 h using petroleum ether. Prior to extraction, PUF disks were fortified with PCB-30, which served as a surrogate for assessing method recoveries for each sample. Extracts were volume reduced to 2 mL by rotary evaporation and nitrogen blow down. After harvesting (sample collection), SPMDs were stored in the freezer in sealed aluminum containers and shipped to EST laboratory under ice for dialysis in hexane and gel permeation chromatography (GPC). Prior to dialysis, SPMDs were thoroughly brushed to remove any particle or dust stuck to the outer membrane. SPMDs were also fortified with PCB30 prior to dialysis by injection into the interior of the SPMD tube. Extracts were concentrated to about 0.5 mL using UHP nitrogen gas, filtered through sodium sulfate, and transferred to vials for GPC cleanup using dichloromethane. Sample extracts were solvent exchanged and concentrated to 1 mL in hexane. Concentrated extracts from PUF disk and SPMD samples were cleaned up and separated into three fractions using florisil (1.2% deactivated) column chromatography. Analysis and quantification details are presented elsewhere (18). Briefly, each fraction was concentrated to 200 µL under nitrogen, after which aldrin was added to monitor analytical instrument variability. Analysis was performed on a Varian 3400 gas chromatograph equipped with an electron capture detector (GC-ECD) and a DB-5 capillary column (60 m long, 0.25 mm i.d., 0.25µm stationary phase). Hydrogen was used as the carrier gas (2 mL min-1). PCBs and OCPs were quantified using external standard solutions (Ultra Scientific and Cerilliant). Samples were analyzed for 90 PCB congeners and 43 OCPs (18).
Results and Discussion QA/QC. Recoveries for PCB-30 surrogate that was added to each sample were 63-99% (mean 83%) for SPMDs and 6595% (mean 87%) for PUF disks. These values were satisfactory, and no correction was applied to the samples. A previous passive sampling calibration study showed good method recoveries (62-110%) for isotopically labeled PCB surrogates (13C PCBs 77, 81, 105, and 126) that were added to PUF disks and SPMDs prior to extraction (10). PUF extraction efficiencies were assessed previously by spiking blank PUFs (in triplicate) with OCPs and PCBs at environmentally realistic levels. Recoveries averaged greater than 90% (18). Quality assurance measures included analyzing standard reference materials (NIST) and routine participation in inter-laboratory check studies. All data have been blank corrected. For PUF disks, levels of PCB and selected OCPs in the field blank were below 1% of sample amounts with the exception of R-HCH at 3%. Blanks levels for SPMDs were higher and ranged from 8% to 10% for trichloro to heptachloro PCBs to a high value of 39% for octachloro PCBs. SPMD blanks were very high for OCPss7% for R-HCH, 90% for TC, and 34% for p,p′-DDT. Consequently, results for OCPs in SPMD samples are not presented. PCBs. At all sampling locations, air concentrations were derived from the amount of chemical sequestered by the passive sampler and the equivalent air volumes calculated using eq 2 (and Table 1). Results for a condensed PCB list (n ) 13) is summarized in Table 2, including congener KOA values and equivalent sample volumes. The condensed list represents the more dominant congeners from different PCB homologue classes and accounts for approximately 37% of ∑PCB for 90 congeners. A condensed list was used to simplify the presentation and interpretation of results.
TABLE 2. Passive Sampler-Derived Air Concentrations (pg m-3) for PCBs along an Urban-Rural Transect for the Period July-October 2000a PCB congener 31
28
52
49
44
95
101
110
149
118
153
138
∑PCB
180
°C)b
log KOA (17 8.33 8.40 8.72 8.76 8.91 9.30 9.51 9.83 10.04 10.05 10.24 10.48 11.05 sample vol (m3) 256 271 335 341 362 401 414 426 431 432 435 438 441 Gage PUF 34.7 41.2 69.4 26.6 31.1 89.0 63.8 45.5 43.3 23.3 39.5 31.4 7.8 SPMD 24.1 32.0 43.4 18.2 19.4 63.7 49.9 37.1 34.4 18.9 34.8 28.5 9.4 Junction PUF 25.1 28.7 55.6 18.7 18.4 54.9 39.7 29.8 18.3 17.1 18.1 15.8 2.6 S. Riverdale PUFc SPMD 15.3 23.5 21.5 11.2 12.2 31.3 24.9 20.2 11.8 12.2 13.7 11.2 2.4 Downsview PUF 15.8 18.4 24.3 9.2 12.9 23.4 16.9 13.4 7.9 7.9 8.7 6.7 1.0 SPMD 8.8 13.8 12.8 6.2 6.2 16.3 14.5 8.9 5.3 5.5 3.9 4.2 0.7 R. Hill (A) R. Hill (B) Aurora Egbert
d
PUF SPMD PUF SPMD PUF SPMD PUF SPMD
547 414 343 211 167 107
9.6 5.8
11.5 10.1
13.6 3.4
4.9 0.9
6.9 3.7
15.4 8.8
9.3 5.8
9.2 4.5
5.7 3.4
4.3 2.8
7.9 2.7
5.0 2.3
1.1 0.5
104 54.6
6.1 8.3 2.7 11.5 3.8
11.0 11.1 8.1 13.1 8.3
5.2 7.9 0.0 13.4 nad
3.2 3.8 1.7 6.2 4.1
4.2 4.8 0.0 9.5 4.0
10.6 5.6 7.7 23.1 8.7
6.9 7.3 3.8 9.1 3.8
5.8 5.7 2.6 8.0 3.2
3.7 3.6 2.7 4.5 1.9
3.5 2.6 1.9 4.1 1.9
3.8 5.8 1.7 8.1 3.6
3.0 3.6 1.7 4.8 1.4
0.7 0.7 0.4 0.7 0.3
67.8 70.7 34.8 116 45.0
a Log K b Ref 13. c No PUF sample available for this period. OA values (17 °C) and equivalent air sample volumes are given for each congener. na, data not available.
FIGURE 3. Comparison of SPMD and PUF disk-derived air concentrations for individual PCB congeners (n ) 13 at each site) for the integration period July-October 2000. Figure 3 compares SPMD- and PUF-derived air concentrations for all sites using the condensed PCB list. In general, PUF-derived air concentrations exceeded SPMD values by approximately 25%. The offset between the two data sets (i.e., slope of 0.77 vs 1) is likely due to small deviations in the selected air-side mass transfer coefficients (Table 1) that are based on a previous calibration study (10). Figure 4 shows ∑PCB (n ) 13) values for PUF and SPMD passive samplers along the seven site transect. There is a large gradient (∼5-10 times) in concentration ranging from ∼500 pg m-3 at the downtown urban sites to ∼50-100 pg m-3 at the rural end of the transect. The corresponding values for ∑PCB for all congeners (based on the scale-up factor of 1/0.37 or 2.7) are 1350 and 135-270 pg m-3, respectively. These findings demonstrate the continuing role of urban/ industrial areas in Toronto as emission sources of PCBs. Previous studies have also investigated urban areas as sources of PCBs. Gingrich et al. (4) observed much higher PCB levels in urban window films in Toronto compared to films collected in suburban and rural areas. The PCB profile
at the urban location was also enriched in the higher molecular weight PCBs and similar to Aroclor mixtures. Results from the IADN in the Great Lakes Basin show that air concentrations of PCBs in Chicago are approximately an order of magnitude greater than at rural background sites over the period 1993-2000. In 2000, ∑PCB in Chicago was on the order of about 1000 pg m-3 while at background stations values were in the range 100-200 pg m-3 (2). This is consistent with the results of the passive sampling transect presented here. Hafner and Hites (19) recently used a probabilistic source contribution model to show that atmospheric PCBs in the Great Lakes Basin primarily originate from urban sources. Sources of PCBs in urban areas include offgassing from PCB-treated construction materials in older homes and leakage from closed systems such as older electrical equipment (e.g. transformers that contain large quantities of PCB fluids) (20). Release of contaminated indoor air to the outside environment is also a major contributor (4, 5, 21, 22). This transfer pathway is of special concern for new classes of “domestic” POPs such as the brominated flame retardants and fluorinated surfactants (7, 11, 23), which were not investigated here. If cities act as a source of PCB it is likely that not only the air concentration will change with distance away from the source (as demonstrated in Figure 4) but that the composition may also change. This can also be inferred from the results for the urban, suburban, and rural window films discussed above (4, 5). Because higher chlorinated PCBs are more associated with the particle phase, they will be preferentially removed from the atmosphere through particle deposition. This occurs to a greater extent in urban/industrial areas because of higher particle concentrations and deposition fluxes (24), resulting in an urban fractionationsan air profile that becomes enriched with lower chlorinated congeners (less volatile and less associated with the particle phase) with distance from the urban source region. This urban fractionation effect is demonstrated in Figure 5 over the urbanrural transect. Percent composition of the PCB profile (based on the condensed list) was determined at each site and normalized to the percent composition for each congener at urban site 1 (Gage building). Thus for the Gage site, all values are unity. At other sites, values less than unity indicate depletion relative to the Gage building and values greater VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4477
FIGURE 4. PUF disk- and SPMD-derived air concentrations across an urban-rural transect for ∑PCB (13 congeners) for the integration period July-October 2000. than unity show enrichment. PCB congeners are listed on the x-axis and arranged according to volatility, from the lower chlorinated, less particle associated on the left to the higher chlorinated, more particle associated on the right. The observed trend shows depletion of the higher chlorinated PCBs and enrichment of the lower chlorinated congeners, with distance from the urban source region. The idea of chemical fractionation in the environment is not new. Wania and Mackay (25) predicted that, on a global scale, there will be preferential transport (contribution) of lighter PCB congeners with increasing latitude (distance from southern sources). Evidence of this latitudinal fractionation comes from Meijer et al. (26), who used SPMD type passive samplers to investigate PCB air concentrations on a transect from the south of the UK to the tip of Norway. Global fractionation of chemicals is primarily driven by temperature, surface-air exchange, and the different chemical retention capacities of various climate zones. However, we believe that this is the first study to demonstrate an “urban fractionation effect” for PCBs by sampling air at several sites along a relatively short (∼75 km) urban-rural transect. In the case of urban fractionation, we believe that higher particle deposition in the source region is driving the fractionation. It is also possible that the changes in the PCB pattern are occurring over a much larger spatial scale and associated with different regional air mass influences for the sites along the transect. OCPs. Table 3 summarizes passive sampler derived air concentrations for OCPs at all sample sites based on the amount of chemical collected by the passive sampler and the equivalent air volumes calculated using eq 2 (and Table 1). Although the passive samplers were not previously calibrated for OCPs (10) we assume that similar uptake rates and KOA-based capacity values (Table 1) apply. As discussed in ref 10, uptake by the sampler is controlled by molecular diffusivity in the air boundary layer. Although there is a dependence on chemical molecular weight, it is weak and does not substantially alter sampling rates for the range of PCBs and OCPs investigated. Compound-specific KOA values at 17 °C and equivalent sample volumes are also given. As discussed previously, concentrations of OCPs in the SPMD blank were exceptionally high; therefore, only PUF disk results 4478
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004
are presented. Figure 6a-f summarizes concentration trends for several OCPs along the urban-rural transect. R- and γ-HCH. Technical HCH is a mixture of several isomers of which R-HCH comprises 60-70% and γ-HCH comprises 10-12% (27). HCH was the dominant OCP detected in arctic air and water (28). As a result of ecosystem health concerns, technical HCH was banned in many countries during the 1970s. In North America, technical HCH use in Mexico continued until 1993. China and the Soviet Union continued to be large consumers of technical HCH until bans on use in the mid-1980s and 1990, respectively. After 1990, usage of HCH continued in India, Vietnam, and other isolated parts of the world (27). This continued use coupled with cycling of previously used R-HCH is believed to be the main source of R-HCH to the atmosphere. Because it is fairly volatile (log KOA ) 7.93, Table 2) and degrades slowly in the atmosphere (29), R-HCH is easily transported in the atmosphere and attains fairly uniform air concentrations globally (25). Release of γ-HCH to the environment continues through the use of the pesticide lindane, a purified form of the active isomer. Usage of γ-HCH in many parts of the world has declined in recent years. In North America, greatest usage occurs in the prairie region of western Canada where γ-HCH is used as a seed treatment (30, 31). Longrange transport of γ-HCH from the prairies is the dominant source of atmospheric γ-HCH in the Great Lakes Basin (31). Figure 6a and Table 3 summarize passive sampler-derived air concentrations for R- and γ-HCH. Of the two isomers, concentrations (pg m-3) for R-HCH are slightly higher and range from 40 to 60 while γ-HCH is in the range of 20-50. Concentrations for both isomers are uniform across the transect, and results for R-HCH are consistent with background concentrations for the Great Lakes Basin (2). The observed R-HCH/γ-HCH ratio along the transect (1.1-2.3, Table 3) is also fairly uniform and typical of background air (32). Chlordanes. Technical chlordane consists mainly of transchlordane (TC), cis-chlordane (CC), and trans-nonachlor (TN) in the proportion 1.00:0.77:0.62, respectively (33), and was used in agriculture (mainly corn) and on lawns and gardens
FIGURE 5. Urban fractionation effect on PCBs showing enrichment of lower molecular weight PCBs with increasing distance from the urban source region.
TABLE 3. Passive Sampler-Derived Air Concentrations (pg m-3) for OCPs along an Urban-Rural Transect for the Period July-October 2000a r-HCH γ-HCH log KOA (17 °C) b sample vol (m3) Gage Junction Downsview R. Hill (A) Aurora Egbert a
7.95 164 55.2 40.8 41.4 51.7 39.8 61.5
8.21 227 52.2 29.1 25.6 22.3 23.9 40.2
TC
CC
TN
9.40 408 35.9 18.9 8.7 12.6 9.2 16.4
9.45 411 33.1 17.1 8.9 14.3 10.1 19.1
9.86 427 20.4 13.1 8.1 12.2 9.1 15.1
p,p′DDE
dieldrin 9.29 401 35.6 26.2 19.4 26.4 13.6 75.8
o,p′DDT
10.21 434 60 33 30 39 27 305
p,p′DDT
9.93 10.30 429 436 21.7 27.0 13.0 14.4 8.0 10.5 13.7 17.4 9.7 11.7 40.3 52.9
endo1
TOX
r-/γHCH
TC/ CC
9.09 384 370 408 405 465 254 817
9.45 411 23.2 15.8 11.5 11.2 11.0 21.2
1.1 1.4 1.6 2.3 1.7 1.5
1.08 1.10 0.98 0.89 0.91 0.85
p,p′-DDT/ p,p′-DDT/ p,p′-DDE o,p′-DDT
0.45 0.43 0.35 0.45 0.43 0.17
1.2 1.1 1.3 1.3 1.2 1.3
Log KOA values (17 °C) and equivalent air sample volumes are given for each pesticide. b Ref 15.
until its cancellation in 1973, when its use was limited to termite control (until 1988 when it was de-registered for use as a termiticide) (34).
Figure 6b shows air concentrations of TC and CC in the range of 10-40 pg m-3, respectively. Meijer et al. (35) reported TC and CC concentrations (pg m-3) of 8.2 and 11 in October VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4479
4480
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004
FIGURE 6. PUF disk-derived air concentrations across an urban-rural transect for individual organochlorine pesticides for the integration period July-October 2000: (a) r- and γ-HCH, (b) trans- and cis-chlrodane, (c) dieldrin, (d) toxaphene, (e) p,p′-DDE and p,p′-DDT, (f) endosulfan 1 (r isomer). VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4481
2000 in the ambient air of an agricultural region in southern Ontario. These values are consistent and remarkably similar to the passive sampler-derived concentrations at nonurban sites along the transect (Table 3). Higher passive sampler-derived concentrations at the urban sites suggest that Toronto is a source of chlordanes, possibly resulting from previous use of chlordanes on lawns/ gardens and house foundations. Home garden soils in the Corn Belt region were reported to have higher concentrations of chlordanes as compared to agricultural soils in the region (34). In the environment, TC degrades more quickly than CC resulting in a TC/CC ratio of
