Environ. Sci. Technol. 2001, 35, 2156-2165
Seasonal and Species Differences in the Air-Pasture Transfer of PAHs KILIAN E. C. SMITH,* GARETH O. THOMAS, AND KEVIN C. JONES Environmental Science Department, Institute of Environmental and Natural Sciences, Lancaster University, Lancaster, LA1 4YQ, U.K.
A field plot was established at a semirural site in the U.K. to investigate the atmospheric transfer of PAHs to different pasture species over the whole growing season. The PAHs displayed a range of partitioning behaviors in the atmosphere from exclusively gas phase to exclusively particle bound, resulting in different modes of deposition to the plant surface. The different pasture species had different plant and sward characteristics, e.g., leaf morphologies, yields, etc. For the majority of PAHs, the plant species displayed a seasonality in concentrations, with concentrations being higher in the winter than in the summer. For the lighter PAHs, this seasonality was absent with soil outgassing and/or summer sources of PAHs being implicated. Airplant transfer factors (scavenging coefficients, with units m3/g dw) typically ranged between 4 and 52 during the summer, increasing to 8-88 during winter. Despite different plant and sward characteristics, the mixtures and concentrations of PAHs were similar for all the plant species. This indicates that there was little difference in the interception and retention behavior of the gas- and particle-phase PAHs. The implications of this for food chain transfer and airvegetation modeling are discussed.
Introduction Vegetation, such as pasture grass, plays an important role in the movement of semivolatile organic chemicals (SOCs) through the environment. This results from its ability to modify air-soil exchange processes (1, 2), to scavenge and retain SOCs from the atmosphere (3-5), and subsequently to introduce them into natural and agricultural terrestrial food chains (6, 7). In addition, vegetation is a potential site where biotic and abiotic losses of SOCs can occur (e.g., metabolism of SOCs by the surface microflora; photodegradation) (8, 9). Atmospheric deposition and air-surface exchange processes impact on the concentrations of all vegetation types to varying extents. These processes are complex varying spatially, temporally, between compounds, and with plant species. There is a need to understand these processes to reliably model food chain transfer and the regional/global air-surface exchange of SOCs (3, 5-7). Polycyclic aromatic hydrocarbons (PAHs) are one class of SOCs. They are ubiquitous and persistent environmental contaminants (10), and some PAH compounds are mutagenic and carcinogenic (11, 12). Within the class of compounds, they display a wide range of gas-particle partitioning behavior in the atmosphere, ranging from totally gas phase * Corresponding author e-mail:
[email protected]; telephone: +44 1524 593597; fax: +44 1524 593985. 2156
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 11, 2001
FIGURE 1. Influence of plant and environmental variables on the uptake of SOCs from the atmosphere. to entirely particle bound (13, 14). This partitioning is controlled by the prevailing environmental conditions, such as temperature, relative humidity, and mass/nature of the aerosol, so that for a given compound it can vary throughout the season (15-18). PAH sources (i.e., incomplete combustion) are spatially and seasonally variable with higher ambient levels occurring during winter (19). Under certain conditions, some PAHs are susceptible to degradation processes such as photolysis that can also potentially influence vegetation concentrations (20). Vegetation interception and retention of atmospherically derived PAHs is therefore expected to be dependent on the gas-particle distribution of a compound in the air and the nature of the deposition process (i.e., wet, dry particulate, dry gaseous). The relative importance of different modes of deposition to the vegetation will be influenced by characteristics of the sward, architecture and morphology of the individual plant species, and environmental variables such as air concentration, temperature, rainfall, wind speed, etc. (21). These variables are not independent but interact with one another, with the result that the contamination of vegetation is a dynamic and active process with the input/loss mechanisms varying both spatially and temporally (Figure 1). There is still considerable uncertainty about the mechanisms of atmospheric contamination of different vegetation types by SOCs. Concentration differences between different plant species are reported as ranging from being minor to being several orders of magnitude (3, 5, 22-26). It can be difficult to compare different studies from the literature, as they employ different experimental procedures and report the plant concentrations using a variety of different units. Additionally, it is known that environmental variables, for example, air concentration (27) and temperature (27-30), have a significant influence on plant SOC concentrations; these parameters are not always reported. In summary, it was hypothesized that there would be differences in the concentrations and mixtures of PAHs in 10.1021/es000200a CCC: $20.00
2001 American Chemical Society Published on Web 04/28/2001
TABLE 1. Species, Description of Morphology, and Sowing Rate of Pasture Species Used in the Experiment species name
common name
morphology
sowing rate (g/m2)
Festuca ovina Holcus lanatus Koeleria macrantha Lolium perenne Mix, lawn seed Trifolium repens
sheep’s fescue Yorkshire fog crested hair grass perennial rye grass mixture of fescues white clover
tufted, smooth thread-like leaves tufted, leaves gray downy all over fine waxy leaves tufted, smooth flat leaves lawn mix of smooth leaved fescues broad-leafed herb
50 50 50 50 45 50
different plant species, which would have profound implications for food chain exposure to these chemicals and airvegetation exchange modeling. A detailed field-based study was therefore set up to quantify the atmospheric transfer of PAHs to different pasture species over the whole growing season at a semirural site located away from any specific point sources. The aims of the study were to (i) investigate differences in air-plant transfer for a number of different pasture species, addressing the influence of plant/sward characteristics (i.e., leaf morphologies, different growth rates, etc.); (ii) investigate these processes for a group of compounds displaying a wide range of gas-particle partitioning behavior in the atmosphere. The external environmental variables such as air concentrations, temperature, and rainfall were the same for all species in this study. The data from the study also provide a data set for the testing of models.
Experimental Section Field Site and Experiment. The field plot was set up at a field station situated in a semirural location in northwest England and owned by Lancaster University. A meteorological site and ambient SOC monitoring equipment have been based at the site for a number of years. This site has also been used extensively for previous investigations into the environmental behavior of SOCs. Several papers have been published reporting ambient air concentrations and gasparticle partitioning behavior of PAHs and other SOCs together with studies on the atmosphere-pasture transfer of PCBs (2, 31-33). For this study, the turf and top 3 cm of soil were removed from the plot, the soil was dug over, and NPK fertilizer was added. The experiment was set up as a randomized bock design, with 30 blocks each consisting of 6 subplots of 0.7 × 0.7 m dimension. Each subplot was planted with one of six pasture species in mid-September 1996. Table 1 lists the species, their sowing rates, and morphological characteristics. Koeleria macrantha was very slow growing and did not provide sufficient mass for PAH analysis. The seeds had germinated by mid-September 1996, but there was insufficient yield to allow a harvest until October 1996. Subsequent harvests were taken approximately once a month until July 1997. Harvesting of each species was by hand-held shears, and the grass was carefully cut ∼2 cm above ground level to avoid soil contamination during the sampling process. For each time point, three blocks selected at random were destructively harvested to give three replicates for each species. Immediately after harvest, the yield (g/m2) was recorded, a subsample was taken for dry weight determination (see below), and the grass was stored unwashed in polythene bags in a freezer at -25 °C until analysis as described below. Concurrently, PAH air samples were taken every 1-2 weeks with a conventional high-volume air sampler, equipped with a glass fiber filter and polyurethane foam (PUF) plug to separate the gas and particle phases (∼600 m3 air sampled). Detailed meteorological data were available throughout the whole experimental period from the weather station operating at the site.
Extraction and Cleanup. Grass. Approximately 25 g of grass was frozen using liquid nitrogen and ground up with ∼50 g of sodium sulfate. A total of 50 ng of each individual deuterated surrogate PAH was added, and the grass was Soxhlet extracted in DCM for 4 h. Cleanup was by a silicaalumina precolumn followed by size exclusion chromatography (SEC). After final concentration of the sample, deuterated PAH internal standards were added prior to analysis by GC/MS. This method has been described in detail elsewhere (34). For each replicate, a subsample was taken for dry weight (dw) analysis. The sample was dried in an oven at 120 °C until a constant weight was recorded. Air. The glass fiber filters and PUF plugs were Soxhlet extracted for 12 h in hexane. Fractionation and cleanup was by a 3-g silica column. After final concentration of the sample, deuterated PAH internal standards were added prior to analysis by GC/MS. Air concentrations are reported from fluorene onward only. Deuterated PAH surrogate recovery standards were not used in the air analysis. Recoveries were tested by spiking glass fiber filters and PUF plugs with a full range of PAHs. All the PAHs had recoveries of >97%, except anthracene (62%). Reproducibility of the method was measured with the use of an in-house homogenized air extract. All the PAHs quantified had a coefficient of variation of < 8%, except chrysene (9%). This has been described in detail elsewhere (31). Instrumental and Data Analysis. The GC/MS analyses were performed using a HP 5890 series II GC equipped with a 30-m HP5MS column (0.25 mm i.d. and 0.25 µm film thickness) and a 2-m deactivated HP retention gap (0.53 mm i.d.). This was connected to a HP 5972 MSD operating in selected ion-monitoring mode (SIM) with the electron energy at 70 eV, the EI source held at 176 °C, and the interface temperature at 300 °C. Cool on-column injection of 1 µL of sample was performed using an automatic HP 7673 injector. The injector program was 60 °C for 0.05 min, 25 °C min-1 to 300 °C, which was held for 7.5 min. The oven temperature program was 60 °C for 1 min, 20 °C min-1 to 130 °C, and 6 °C min-1 to 310 °C, which was held for 15 min. The injector pressure was 11 psi for 1 min, 0.5 psi min-1 to 27.5 psi, and 20 psi min-1 up to a final pressure of 35 psi held until the end of the run. Helium was used as the carrier gas. The compounds analyzed and the abbreviations used in the text are now listed. Those compounds in italics were used as the internal standards; the compounds in bold were used as the surrogate standards: naphthalene-d8, Nap-d8; naphthalene, Nap; 2-methylnaphthalene, 2mNap; 1-methylnaphthalene-d8, 1mNap-d8; 1-methylnaphthalene, 1mNap; biphenyl, Biph; 2,6-dimethylnaphthalene, 2,6dmNap; acenaphthylene, Acy; acenaphthene-d10, Ace-d10; acenaphthene, Ace; 2,3,6-trimethylnaphthalene, 2,3,6tmNap; fluorene-d10, Fl-d10; fluorene, Fl; phenanthrene-d10, Phen-d10; phenanthrene, Phen; anthracene-d10, Anth-d10; anthracene, Anth; 1-methylphenanthrene, 1mPhen; fluoranthene-d10, Fluo-d10; fluoranthene, Fluo; pyrene-d10, Py-d10; pyrene, Py; pterphenyl-d14, pTerph-d14; benz[a]anthracene-d10, BaA-d10; benz[a]anthracene, BaA; chrysene, Chry; benzo[b]fluoranthene, BbF; benzo[k]fluoranthene, BkF; benzo[e]pyrene, BeP; VOL. 35, NO. 11, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2157
FIGURE 2. Total PAH, phenanthrene, and benzo[ghi]perylene concentrations in the atmosphere over the experimental period. benzo[a]pyrene-d12, BaP-d12; benzo[a]pyrene, BaP; perylened12, Per-d12; perylene, Per; 1,3,5-triphenylbenzene, 135TPB; indeno[123-cd]pyrene, IP; dibenz[ah]anthracene, DahA; benzo[ghi]perylene-d12, BghiP-d12; and benzo[ghi]perylene: BghiP. A detailed list of the ions monitored for the PAH analysis is given ref 34. Quantification. The PAHs analyzed are listed above. Each ring class of PAHs had its own individual surrogate standard and internal standard. The surrogate standard solution was applied to the sample immediately prior to extraction to produce recovery data for QA purposes. The internal standard solution was added to the cleaned-up sample extract prior to transfer to the GC vial to correct for sample volume and instrument variation. The parent and surrogate analytes were quantified against the appropriate internal standard. An internal standard quantification procedure was used using HP Environmental Data Analysis software. Recoveries for the grass analysis (n ) 150) were as follows: 1mNap-d8, 92% (RSD 14%); Fl-d10, 97% (RSD 6%); Anth-d10, 99% (RSD 8%); Py-d10, 106% (RSD 7%); pTerphd14, 107% (RSD 7%); BaP-d12, 128% (RSD 11%); and BghiPd12, 110% (RSD 10%). The high recoveries yet good reproducibility seen for the heavier compounds, in particular BaPd12, are a matrix-specific phenomenon. The reasons for this have been discussed elsewhere (34). The data below are not recovery corrected. In each batch of six samples, one was a blank. The data are all blank-corrected using the appropriate batch blank value. Individual compounds in the 0.2-1 ng/g fw were readily quantifiable by this method. Method detection limits have been defined in detail in ref 34. All data are expressed on a dry weight basis only. Data in terms of bulk yield are more widely available in the literature to allow for comparison and are also more useful in terms of food chain modeling. Expression of the results on a lipid weight basis is itself fraught with problems and therefore not done here. There are difficulties in relating the amount of extractable material (usually operationally defined by extraction of the vegetation using organic solvents, e.g., refs 24 and 27) to a meaningful physical entity such as the plant SOC sorption compartment. In addition, the expression of the results on a lipid basis is only valid when the contaminant concentration varies in direct proportion to the lipid content, otherwise erroneous conclusions may be reached (35, 36). In particular, this is an issue for those particle-bound compounds for which plant uptake is not directly affected by the quantitative and qualitative nature of the cuticular lipids. 2158
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 11, 2001
TABLE 2. Concentration and Percentage Particle-Bound PAHs in the Atmosphere over the Experimental Perioda air concn (pg/m3)
fluorene phenanthrene anthracene 1-methylphenanthrene fluoranthene pyrene benz[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[e]pyrene benzo[a]pyrene perylene indeno[123-cd]pyrene dibenz[ah]anthracene benzo[ghi]perylene
particle bound (%)
min
max
min
max
370 2760 90 190 620 310 20 90 40 30 30 20
