Environ. Sci. Technol. 2005, 39, 5763-5773
Using Passive Air Samplers To Assess Urban-Rural Trends for Persistent Organic Pollutants and Polycyclic Aromatic Hydrocarbons. 2. Seasonal Trends for PAHs, PCBs, and Organochlorine Pesticides A N N E M O T E L A Y - M A S S E I , †,‡,| 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, De´partement de Ge´ologie, UMR CNRS 6143, Faculte´ des Sciences et Techniques, Universite´ de Rouen, France, and Freshwater Institute, Fisheries and Oceans Canada, Winnipeg, Manitoba, Canada R3T 2N6
This is the second of two papers demonstrating the feasibility of using passive air samplers to investigate persistent organic pollutants along an urban-rural transect in Toronto. The first paper investigated spatial trends for polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs). This second paper investigates the seasonality of air concentrations for polycyclic aromatic hydrocarbons (PAHs), PCBs, and OCPs along this transect. Air samplers, consisting of polyurethane foam (PUF) disks housed in stainless steel domed chambers, were deployed for three 4-month integration periods from June 2000 to July 2001. The seasonal variations of derived air concentrations for PAHs, PCBs, and OCPs reflected the different source characteristics for these compounds. PAHs showed a strong urban-rural gradient with maximum concentrations at urban sites during the summer period (July-October). These high summer values in Toronto were attributed to increases in evaporative emissions from petroleum products such as asphalt. PCBs also exhibited a strong urban-rural gradient with maximum air concentrations (∼2-3 times higher) during the spring period (April-June). This was attributed to increased surface-air exchange of PCBs that had accumulated in the surface layer over the winter. R-HCH was fairly uniformly distributed, spatially and temporally, as expected. This pattern and the derived air concentration of ∼35 to ∼100 pg m-3 agreed well with high volume air data from this region, adding confidence to the operation of the passive samplers and showing that site-to-site differences in sampling rates was not an issue. For other OCPs, highest concentrations were observed during the spring period. * Corresponding author phone: (416)739-4837; fax: (416)739-5708; e-mail:
[email protected]. † Environment Canada. ‡ University of Toronto. § Fisheries and Oceans Canada. | Universite ´ de Rouen. 10.1021/es0504183 CCC: $30.25 Published on Web 06/24/2005
2005 American Chemical Society
This was associated with either (i) their local and/or regional application (γ-HCH, endosulfan) and (ii) their revolatilization (chlordanes, DDT isomers, dieldrin, and toxaphene). Principal component analysis resulted in clusters for the different target chemicals according to their chemical class/source type. The results of this study demonstrate how such a simple sampling technique can provide both spatial and seasonal information. These data, integrated over seasons, can be used to evaluate contaminant trends and the potential role of large urban centers as sources of some semivolatile compounds to the regional environment, including the Great Lakes ecosystem.
Introduction Polycyclic aromatic hydrocarbons (PAHs), organochlorine pesticides (OCPs), and industrial organochlorines such as polychlorinated biphenyls (PCBs) are compounds which, because of their physical-chemical properties, are persistent and mobile in the environment (1-4). These chemicals also tend to bioaccumulate and be toxic, and therefore they pose risks to biota and humans (5). Because the atmosphere plays a major role in the cycling of these organic pollutants, international regulation has focused on reducing emissions to air. Thus, many programs were launched to measure levels in air and deposition of these pollutants: for instance, the Integrated Atmospheric Deposition Network (IADN) in North America (6) and the European Monitoring and Evaluation Program (EMEP) in Europe (7). Moreover, international bodies such as the United Nations Economic Commission for Europe (UNECE) and the United Nations Environment Program (UNEP) have established priority lists for the elimination of the worst persistent organic pollutants (POPs), including PCBs and some OCPs. Although most of these contaminants are no longer in use in industrialized regions, they continue to persist in the environment, including the atmosphere. To respond to this increasing need for cost-effective and simple tools for assessing concentrations in air simultaneously at multiple sites, the use of passive sampling methods to monitor airborne contaminants has greatly increased over the past few years. Information obtained from such campaigns (8-13), on various scales, helps to address questions regarding sources of PAHs and POPs, their seasonal pattern, and chemical signature and to improve our understanding of the role of the atmosphere in the transport of these contaminants. Only a few studies have employed passive samplers in cities (8-10), which are known to be important emission sources of many contaminant classes to their surrounding regions (11-13). This is especially relevant for cities such as Toronto that are situated on the shore and may act as a source to the Great Lakes system. Furthermore, the assessment of risk associated with exposure and human health concerns is especially relevant in these urban areas where the majority of population resides. In the city of Toronto, few data about atmospheric concentrations of POPs and PAHs have been published (16). However, measurements in various compartments such as atmospherically derived organic film on impervious surfaces (17) and soils (18) showed high concentrations for several classes of compounds such as OCPs, PCBs, PAHs, and PBDEs (polybrominated diphenyl etherssa class of flame retardants that is ubiquitous in the environment) (19). In this study, passive samplers comprised of polyurethane foam (PUF) disks were deployed at seven sites along an VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5763
FIGURE 1. Location of sampling sites across the Greater Toronto Area. urban-rural transect in Toronto. The operation of the samplers and the method used to evaluate passive samplerderived concentrations were described in detail in a previous companion paper (16). That paper reported on the spatial distribution of PCBs and OCPs from July to October 2000. The objective of this second paper is to investigate seasonality in ambient air concentration and spatial distribution of PCBs, OCPs, and PAHs during 1 whole year.
Methodology Sample Collection. Atmospheric samples were collected at seven sites along an urban-rural transect extending ∼75 km north of downtown Toronto (Figure 1). The three first sites (Junction Triangle, Gage Building, and South Riverdale ) urban 1, 2, and 3, respectively) are located in south Toronto and are described as urban, high-density residential/ industrial. Located 16 km north of downtown Toronto, the Downsview site (urban 4) is considered as urban, medium density, and residential/industrial. Further north, two suburban sites (Richmond Hill and Aurora ) suburban 1 and 2) are characterized as low-density residential/industrial. Last, the rural site of Egbert (44°13′57′′ N/79°46′53′′ W) is situated in an agricultural/farming region. Air samplers were deployed for three 4-month integration periods from June 2000 to July 2001. Sample collection dates and number of days for the sampling periods are outlined in Table 1. Mean temperatures at each site were assessed through the National Air and Water Monitoring Activities Archive. Meteorological stations nearest the sampling sites were selected for temperature data. However, because the temperature differences for the three sampling periods were fairly small between sites, a common temperature of 16 °C, -3 °C, and 15 °C was used for all calculations for the sampling periods 1, 2, and 3, respectively. The passive air samplers consisted of PUF disks housed in stainless steel, domed chambers in order to reduce the influence of wind speed on uptake rate and also to protect the PUF disks from precipitation, direct particle deposition, and UV sunlight. A detailed description of sampling, preparation, workup, and theory of the PUF disk samplers is given by Harner et al. (16). Briefly, the uptake of POPs by PUF disk samplers is air-side controlled with an outdoor sampling rate of approximately 3.5 m3 day-1. This sampling rate for PUF disks has been confirmed in more recent studies using depuration compounds added prior to deployment (9). The PUF disk housings have also been shown to be effective in dampening the wind-effect on sampling rate (20). Prior to deployment, PUF disks comprised of virgin, untreated foam were precleaned by Soxhlet extraction using 5764
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
TABLE 1. Passive Sample Collection Informationa start date
end date
number of sampling days
2000/06/27 2000/10/31 2001/03/20 2000/06/28 2000/11/01 2001/03/20
2000/10/31 2001/03/20 2001/07/23 2000/11/01 2001/03/20 2001/07/23
126 141 125 126 140 125
2000/10/31 2001/03/20 2000/06/22 2000/11/03 2001/04/03 2000/07/25 2000/10/31 2001/04/03 2000/06/27 2000/11/03 2001/04/02 2000/07/05 2000/10/31 2001/04/03
2001/03/20 2001/07/23 2000/11/03 2001/04/03 2001/07/23 2000/10/31 2001/04/03 2001/07/23 2000/11/02 2001/04/02 2001/07/24 2000/10/31 2001/04/03 2001/07/23
140 125 134 155 111 98 155 111 128 152 113 118 154 111
sampling period site Junction Triangle ) urban 1 1 2 3 Gage Building ) urban 2 1 2 3 South Riverdale ) urban 3 1 2 3 Downsview ) urban 4 1 2 3 Richmond Hill ) suburban 1 1 2 3 Aurora ) suburban 2 1 2 3 Egbert ) rural 1 1 2 3
a Period 1 ) July to Oct, 2000 (summer-fall). Period 2 ) Nov 2000 to March 2001 (fall-winter). Period 3 ) April to June, 2001 (springsummer).
acetone and then petroleum ether. After cleaning, the PUF disks were desiccated under vacuum to remove excess solvent and stored cool and in the dark in solvent-rinsed glass jars having Teflon-lined lids. A separate sampling train consisting of a single glass fiber filter and two PUF plugs was used at the Gage Building (urban 2) and Egbert (rural 1) for total suspended particle determinations (TSP, µg of particles m-3 air). The filters were preweighed after equilibration in a constant humidity chamber for 48 h at 20 °C over a saturated sodium chloride solution. The same procedure was used after sampling to ensure that any changes in filter mass were attributed only to particulate matter and not to differences in water content. The instrument provided measurements from a 24-h sampling period on 3- or 6-day sampling cycles. Analysis. Samples were analyzed for 90 PCB congeners, 43 OCPs, and 19 PAHs. Prior to extraction, PUF disks were fortified with PCB-30, which served as a surrogate for assessing method recoveries for each sample. PUF disk samples were Soxhlet extracted for 24 h with petroleum ether, concentrated to ∼1 mL under a gentle stream of clean, dry
N2, and then separated into two fractions. Half the extract underwent cleanup and analysis for PAHs, while the other half was for POPs. For POPs, analysis and quantification details are presented elsewhere (19, 21, 22). Briefly, the extract was cleaned up into three fractions using florisil (1.2% deactivated) column chromatography. These three fractions were analyzed for several POPs classes including PCBs, OCPs, and PBDEs for which results are presented elsewhere (19). Each fraction was concentrated to ∼200 µL under nitrogen, after which aldrin was added to monitor analytical instrument variability. Following cleanup on an alumina:silicic acid column (2: 3), PAHs were analyzed on a Hewlett-Packard 5890 gas chromatograph equipped with a mass selective detector (Model HP5970) using a DB5-MS (30 m × 0.25 mm i.d.; 0.25 µm film thickness) capillary column (J&W Scientific). Quantification was performed using the internal standard method, utilizing five deuterated PAHs (100 ng mL-1) spiked into the cleaned up extracts. PAH calibration standards of 25, 50, 100, and 250 ng mL-1 containing all the target PAHs were prepared from stock solution supplied by Supelco Chromatography Products (Oakville, Ontario). Quality Control/Quality Assurance. All analyses were monitored using strict quality assurance and control measures. For PCBs and OCPs, quality assurance was identical to those given in ref 16. For PAHs, laboratory and field blanks consisted of pre-extracted PUF disks extracted and analyzed as samples. Practical detection limits were 0.24-8.82 pg m-3, depending on the compound. PAH recoveries were 84-114%. Consequently, reported values were not recovery corrected but were blank corrected using the mean blank value. Blank values were less than 1.4% of sample amounts. Instrument efficiencies were estimated using quality control standards after every five samples run on the GC-MS. Peaks were only integrated when the signal-to-noise ratio was g 3.
Results and Discussion Air concentrations for the target chemicals were derived from the amount accumulated in the PUF disk and the effective air volume. The eq 2 in ref 16 was used to estimate the effective air volumes (Vair, m3). Vair and resulting air concentrations are given in Tables 2-4 for PAHs, PCBs, and OCPs, respectively. For most compounds, the effective air volumes (Vair, m3) were on the order of ∼380 m3 on average, based on a linear sampling rate 3.5 m3 d-1 and a deployment time of ∼ 120 days. Lower molecular weight chemicals had lower values for Vair as they approach saturation (equilibrium) in the PUF disk during the sampling period. This is due to their lower octanol-air partitioning coefficient (KOA) values and hence lower PUF-air partition coefficient (KPUF-A) (16). PAHs. PAHs are produced by incomplete combustion of fossil fuels or organic matter and are considered as ubiquitous contaminants in the environment. PAHs have a mainly anthropogenic origin (automobile traffic, domestic heating, thermal power stations, and industrial emissions). They are recognized as mutagenic compounds and known to be carcinogenic in animals and humans (23). Table 2 provides a summary of the PAH concentrations for each period and for each analyzed compound. Σ 17 PAH concentrations ranged from 11.5 to 61.4 ng m-3 for the period 1 (July-Oct 2000), from 8.34 to 18.5 ng m-3 for the period 2 (Nov 2000-March 2001), and from 3.53 to 18.8 ng m-3 for the period 3 (April-June 2001) (Figure 2). These values were similar to those measured for urban sites worldwide. In Paris, Ollivon et al. (24) showed concentrations that ranged between 3 and 15 ng m-3 (Σ 8 PAHs); in North America, Cotham and Bidleman (25) reported values (Σ 13 PAHs) of 43, 57, 93, and 195 ng m-3 in Columbia, Portland, Denver, and Chicago, respectively. In Toronto, polymer-coated glass samplers (POGs) deployed in the CN Tower (up to 360 m) showed a
strong vertical gradient of PAHs with higher concentrations near ground level (26). In all of these studies, phenanthrene was the most predominant PAH, followed by fluorene and acenaphthene. PAH concentrations showed a strong urban-rural gradient, with total concentrations (Σ 17 PAHs) up to ∼5 times higher in urban sites than in the rural one (Figure 2). This has been showed previously in air (25, 27, 28), sediments (29), organic films on impervious surfaces (17, 30), soils (18), and atmospheric deposition (31). The gradient reflects PAH emission sources which are well-known to be proportional to the population density (31). However, PAH profiles were different for urban and rural sites with proportions of the higher molecular weight PAHs (e.g. benzo[ghi]perylene) decreasing along the urban-rural transect (Figure 2). This urban-rural fractionation effect was reported in the previous study for PCBs and is caused by the greater volatility (transport potential) of the low molecular weight PAHs. High molecular weight PAHs that are associated with particles to a greater extent will be deposited closer to the source. The significance of this fractionation is that it occurs over a relatively short distance of ∼75 km. A seasonal fluctuation in concentrations was also evident (Figure 2). For all sites, the highest PAH concentrations (Σ 17 PAHs) were found during period 1 (July-Oct), while the two other periods (Nov-March and April-June) showed similar concentrations. Σ PAHs was dominated by phenanthrene which exists primarily in the gaseous phase in ambient air. Since PUF disks samplers have been shown to sample mainly the gas phase (32), the seasonal variation of Σ PAHs is thus linked to phenanthrene concentrations. Higher summertime atmospheric concentrations for phenanthrene were previously observed in Toronto from 1987 to 1997 (33). The enrichment in phenanthrene during the summer could be due to increases in evaporative emissions from petroleum products such as asphalt, coal tar sealant, and roofing tar (33, 34). This can also explain why this seasonal trend is more marked for the urban areas (Figure 2), where this kind of surface is more prevalent. Although combustion-derived PAH emissions may be elevated during the colder months, PAH gas-phase concentrations will be reduced by partitioning to particles and snow which is enhanced at cold temperature. This effect will be greatest in urban areas where particle concentrations are highest. For instance, total suspended particulate (TSP) in the ambient air along the urban-rural transect was 75.2 and 18.2 µg m-3 in March 2000 for urban 2 and rural 1, respectively. PCBs. The emission of PCBs in the atmosphere peaked in late 1960s (35). Nowadays, the use of these chemicals is prohibited or restricted in many countries. Emissions to ambient air are likely due to revolatilization of previously emitted compounds (36, 37) and to continued release from point sources such as old industrial/urban areas where they were previously heavily used and still exist (35). The concentrations for 13 of the more dominant congeners are listed in Table 3. Σ 13 PCB concentrations ranged from 72 to 550 pg m-3 for period 1 (July-Oct 2000), from 65.6 to 506 pg m-3 for period 2 (Nov 2000-March 2001), and from 129 to 1350 pg m-3 for period 3 (April-June 2001). For all seasons, PCB concentrations showed a strong decrease with distance from the urban area (Figure 3), confirming the continuing role of urban/industrial areas in Toronto as emission sources of PCBs (16), possibly due to the outgassing of PCBs from buildings (38). Seasonal variability is also apparent. PCB air concentrations were ∼2-3 times higher during the spring-summer (period 3; mean temperature ) 15 °C) than during the periods 1 and 2 (16 and -3 °C, respectively) which showed similar values. This is consistent with a spring pulse effect (39) VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5765
5766 9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
TABLE 2. Passive Sampler-Derived Air Concentrations (ng m-3) for PAHs along the Urban-Rural Transect in Toronto in 2000-2001a ACY urban 1 urban 2 urban 3 urban 4 suburban 1 suburban 2 rural 1 Vair, m3
ACE
FLU
PHE
7.09 5.11
11.8 8.51
13.8 12.1
24.3 23.6
4.30 1.80 1.86 0.91 25
5.05 3.07 2.60 1.37 29
7.84 6.01 5.00 3.89 78
ANT
FTH
PYR
BaA
0.63 0.77
1.97 2.28
1.28 1.70
0.08 0.21
11.1 8.30 6.28 4.63 183
0.30 0.15 0.10 0.09 183
0.72 0.56 0.38 0.35 380
0.50 0.34 0.20 0.19 380
0.02 0.01 0.01 0.01 431
urban 1 urban 2 urban 3 urban 4 suburban 1 suburban 2 rural 1 Vair, m3
0.63 0.52 0.64 0.51 0.20
2.78 1.80 2.31 1.68 0.95
4.44 3.67 4.33 4.11 2.72
6.33 5.63 5.96 5.61 4.11
0.12 0.08 0.15 0.08 0.02
1.95 1.69 1.85 1.35 1.11
1.62 1.40 1.58 1.10 0.75
0.05 0.06 0.05 0.03 0.01
0.22 135
0.94 155
2.46 294
3.37 400
0.05 400
0.71 477
0.45 477
0.01 490
urban 1 urban 2 urban 3 urban 4 suburban 1 suburban 2 rural 1 Vair, m3
0.49 0.30 0.57 0.21 0.02 0.09 0.02 64
2.68 2.18 3.58 0.96 0.05 0.39 0.05 75
2.56 1.63 2.88 1.21 0.31 0.71 0.31 180
8.28 7.24 6.53 3.68 1.66 2.32 1.66 307
0.20 0.21 0.13 nd nd nd nd 307
2.15 1.81 1.33 0.74 0.86 0.44 0.86 430
1.32 1.26 0.82 0.43 0.41 0.20 0.41 430
0.14 0.21 0.09 0.01 0.01
