Environ. Sci. Technol. 2005, 39, 4784-4792
PCDD/Fs in Norwegian and U.K. Soils: Implications for Sources and Environmental Cycling ASHRAF HASSANIN,† ROBERT G. M. LEE,† EILIV STEINNES,‡ AND K E V I N C . J O N E S * ,† Environmental Science Department, Institute of Environmental and Natural Sciences, Lancaster University, Lancaster, LA1 4YQ, United Kingdom, and Department of Chemistry, Norwegian University of Science and Technology, NO-7491, Norway
This paper presents data on polychlorinated dibenzo-pdioxins and dibenzofurans (PCDD/Fs) in a set of wellcharacterized, undisturbed surface (0-5 cm) and subsurface background soils from the U.K. and Norway. The soils have been used previously to investigate the latitudinal distribution, fractionation, cold condensation, and “hopping” of other classes of persistent organic pollutants (POPs). The mono- to octa-CDD/F homologues were quantified. Woodland soils contained higher concentrations (on a dry and soil organic matter (SOM)-basis) than grassland soils, consistent with previous studies. The absolute concentrations of all the PCDDs and most of the PCDFs significantly decreased with latitude, generally supporting the idea of a “southern source region” and a “remote/ receiving northern region”. There was little evidence of “fractionation” and minimal influence of PCDD/F “hopping” on PCDD/F distribution. The %SOM content had a rather minor influence on soil PCDD/F composition. These findings contrast with the trends seen in these soils for hexachlorobenzene (HCB), polychlorinated biphenyls (PCBs), and polybrominated diphenyl ethers (PBDEs). Possible reasons for these differences are discussed and may include influences of/proximity to diffusive combustion sources and/ or sources of variable homologue emissions, formation/ conversion processes for PCDD/Fs in soils, or strong soilPCDD/F partitioning. These soils, from regionally remote/ background locations in Europe contained between 0.2 and 78 pg ΣTEQ/g DW. Some therefore exceed recommended levels of contamination for certain land uses by some European countries. These recommendations seem unrealistic and prohibitively restrictive in light of the dataset presented here.
Introduction Much of the environmental research on persistent organic pollutants (POPs) over recent years has focused on efforts to understand the relative importance of different primary and secondary sources, and the fate and behavior of POPs in the global environment. This is often motivated by the desire to better control sources, and to identify global source/sink * Corresponding author phone: (44)-1524-593972; e-mail:
[email protected]. † Lancaster University. ‡ Norwegian University of Science and Technology. 4784
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 13, 2005
regions, because of the potential for POPs to undergo longrange atmospheric transport (LRAT) and regional/global scale redistribution after their release into the environment (1-5). Previous studies have established that background soils play an important role in the global fate and distribution of POPs; they are a major environmental reservoir for these chemicals due to their high storage capacity (3-5). Background soils receive/exchange POPs with the atmosphere (6); as POPs are released into the environment, the burden in soils therefore becomes a function of the balance between inputs and losses (7). Our previous studies on polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), and hexachlorobenzene (HCB) in global background soils have provided evidence for the global fractionation and redistribution of POPs (3-5) in a way that is broadly consistent with expectations based on compound physicochemical properties (1, 2). The datasets from these studies have been used to help test/validate environmental fate models (e.g., 8). They have also highlighted the important role of soil organic matter (SOM) as a sink for POPs which can undergo repeated airsurface exchange (9, 10) and led to a suggestion that, over time, “infinitely persistent” compounds which can undergo repeated air-surface exchange would move toward a “global equilibrium distribution” with SOM. However, our studies to date have focused on the “more environmentally mobile POPs”. HCB is the POP with the greatest LRAT potential we have studied in regional/global soils; it has a log octanol-air partition coefficient (Koa) of ∼7.4 at 25 °C (11) and global background soil HCB concentrations were strongly positively correlated to soil OM content. This relationship was generally less marked for heavier PCBs and PBDEs in European and global background soils; the PCBs and PBDEs determined in the soils have log Koa values in the range 5.8-10.3 and 9.312.0, respectively (12, 13). In this study, we wanted to extend our work to other POPs, namely the polychlorinated dibenzo-p-dioxins (PCDDs) and -furans (PCDFs). We hypothesized that the PCDD/Fs may have rather different patterns of distribution, fate, and behavior because (a) they have many different potential historical and ongoing sources to the environment, many of which are combustion-related (14); their source distribution may therefore be more varied; (b) the range of sources generates a wide range of PCDD/Fs homologues and isomers; and (c) their properties span a greater rangesfrom log Koa ranging from ∼4.2 for the mono-CDD to ∼12.8-13 for the octa-CDD/Fs. An environmental implication of this property difference is perhaps best illustrated by the typical gas/ particle distribution of this array of POPs under ambient conditions. HCB, the lighter PCBs, and the tri-CDD/Fs are present in the atmosphere almost exclusively as gases, for example, while octa-CDD (OCDD) and octa-CDF (OCDF) are almost exclusively particle-bound (15, 16). Such property differences may therefore be expected to influence the global scale transport and cycling of these compounds. This paper therefore presents data on the PCDD/Fs in background soils, utilizing a set of well-characterized, undisturbed surface (0-5 cm) and subsurface background soils from the U.K. and Norway (i.e., they have not received any pesticides, sewage sludge, amendments, etc., just atmospheric deposition) (3-5). This set of soils, collected in 1998, is interesting for a number of reasons: (a) Norway is a net recipient of airborne pollutants, receiving inputs predominantly from the west and south, because of the net air flows and its position relative to more populated/ 10.1021/es0505189 CCC: $30.25
2005 American Chemical Society Published on Web 06/02/2005
FIGURE 1. Location map of sample sites. industrialized countries; (b) the U.K. is a densely populated and industrialized country, which is a significant source of several classes of POPs in Europe; (c) the soils selected for study cover a wide range of latitudes (from 50°6′ to 70°47′N) and climatic conditions, different land use/vegetative (deciduous, coniferous, grassland) covers, and a wide range of SOM contents (8-97%); (d) the soils have been previously analyzed for the other POPs, as mentioned above (3-5), thereby providing a useful context/contrast for the PCDD/F data. We discuss the data with respect to potential sources, global recycling/fractionation, and environmental variables which may be affecting POP fate and behavior.
Materials and Methods Soil Sampling and Preparation. The soil samples were taken from rural/remote sites, i.e., away from cities, roads, or other human activity. They were collected using a hand-held corer that was cleaned before and after each sample using moss or other vegetation from the site. In addition, the first two cores were always discarded. The sampling depth was 0-5 cm after removal of the litter layer and in some cases an additional sample was collected from 5-10 cm depth. Three cores, taken over an area of several m2, were bulked together to form one sample. The samples were wrapped in two layers of aluminum foil, sealed in two plastic bags, and stored in a cool box. Upon arrival in Lancaster, the samples were immediately transferred to a freezer where they were stored until analysis. Figure 1 and the Supporting Information give information on the location of the sampling sites. In total, 55 surface soils were analyzed. Paired surface and subsurface samples were analyzed from 38 sites. The properties of the soils (SOM, bulk density, site temperature data, etc.) have been presented and discussed elsewhere (3-5). Sample Extraction, Cleanup, and Analysis. The 3 cores were mixed together and a sub-sample of approximately 15 g of soil (wet weight) was taken and mixed with Na2SO4 to remove water. The sample was then transferred to a preextracted glass thimble and spiked with 15 µL of an isotope dilution/recovery standard containing 13C12-labeled PCDD/ Fs. The samples were extracted for 20 h using toluene (99.8%
purity), then ∼1 mL of nonane was added to the extract and the toluene was evaporated using a rotary evaporator. The extracts were transferred with hexane to clean 20-mL vials containing 1 mL of sulfuric acid (H2SO4) and left overnight for lipid and wax digestion. The extracts were then further cleaned using multilayer silica gel columns (30-mm i.d.) [filled from the bottom with 1 cm of Na2SO4, 1.5 g of activated silica gel, 3 g of silica gel treated with NaOH (1 M), 1.5 g of activated silica gel, 6 g of silica gel treated with 44% concn. H2SO4, by weight, 1.5 g of activated silica gel, topped with 1 cm of Na2SO4], with 200 mL of hexane applied as the eluent. To remove the remaining lipids, the concentrated samples were then passed through gel permeation chromatography (GPC) using 6 g of Biobeads SX 3 with hexane/DCM 1:1 (v/v) as the eluting solvent. The concentrated samples were finally separated into three fractions using 4.5 g of basic alumina chromatography; 12 mL of 7% DCM in hexane was used to elute the PCBs, followed by 6 mL of toluene to elute the non-ortho PCBs (co-PCBs), and finally 30 mL of hexane/DCM 1:1 (v/v) was used to elute the PCDD/Fs fraction. After concentration under a stream of nitrogen, the samples were transferred to injection vials, and 15 µL of nonane with a 37Cl4-labeled PCDD was added as an injection standard. Finally, the extracts were reduced in volume to 15 µL under a stream of nitrogen. All samples were analyzed by GC-HRMS using HP6890 GC and a VG Autospec Ultima mass spectrometer, tuned to a minimum of 10 000 resolving power and using electron impact ionization (EI+) in selected ion monitoring mode. Samples were introduced using splitless injection; the PCDD/F homologues and individual congeners were separated using a 60m DB5MS capillary column, in two different runs. Quantification was achieved using the isotope dilution method. The %SOM was determined by loss in weight on ignition at 450 °C overnight. Quality Assurance/Quality Control (QA/QC). A number of steps were taken to obtain data that would allow an assessment of the accuracy and reliability of the data. Analytical blanks, consisting of anhydrous sodium sulfate kept in jars similar to those used for samples and stored with the samples until analysis, were included at a rate of one VOL. 39, NO. 13, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4785
blank for every 5 samples. Recoveries were monitored by quantifying 20 13C12-labeled PCDD/F isotope dilution standards, using the injection standard as an internal standard. The average recoveries ranged between 79 and 101%. The method detection limit was calculated as 3 times the standard deviation of the concentrations found in the analytical blanks. If the concentrations in the blanks were below the instrumental detection limit, then the method detection limit was defined as equal to the instrumental detection limit. The criteria for the quantification of analytes were a retention time found within 2 s of that of the standard, isotope ratio found within 20% of that of the standard, and a signal-to-noise ratio of at least 3. The reproducibility of the method was very good; 13 replicates from different soil types and locations were analyzed along with other samples at different times. The detected compounds gave an average relative standard deviation (RSD) of 10.5% (range 7-15%).
Results and Discussion General Comments on the PCDD/Fs Levels and Distribution. Our analytical method is unusual in that we can also quantify the mono-, di-, and tri-chlorinated homologues. They are included in the tabulated data and some of the Figures. However, they are excluded from some others, for example, where they would affect the scaling/axes and obscure the trends of other homologues. Table 1 presents a summary of the data on a dry weight (DW) and an organic matter (OM) basis. There is a division between the U.K. grassland soils (16 sites), the U.K. woodland soils (17 sites), and the Norwegian woodland soils (21 sites), which will be discussed in more detail below. However, the variation in the whole dataset is a factor of ∼60, ranging between 70 and 4440 pg ∑(1-8)CDD/Fs per g DW (see Table 1). Low concentrations (e.g., 4000 pg/g) was from the southern part of Norway, at site 18 (58°33′N, 8°25′E) (see Figure 1). This may have been affected by proximity to a potential point source (a nickel refinery); studies have reported its influence on the PCDD/F concentrations found in the local aquatic environment (17, 18). The soils studied here are from regionally remote/ background locations in Europe. The data therefore highlight the variability in European soil “background values”. Their total toxic equivalency (∑ TEQ) content ranged between 0.2 and 78 pg/g DW. Guideline values have been set for PCDD/F concentrations in order to categorize soils as “clean” or “contaminated” or to define acceptable levels of contamination for certain land uses. Germany, Sweden, and The Netherlands have suggested that soils “for the most sensitive uses” should not contain >1-100 pg TEQ/g (14). In Finland, soils that exceed the level of 500 pg TEQ/g DW are recommended to be remediated to achieve target concentrations of 20 pg TEQ/g DW (19). These recommendations seem unrealistic and prohibitively restrictive in light of the European background values presented here. General Homologue Pattern in the Soils. Figure 2a shows the ∑(4-8)PCDD/F homologue pattern averaged for all the soils. It is dominated by OCDD, with the amounts of other PCDD homologues decreasing with decreasing level of chlorination. The PCDFs increased in abundance with decreasing level of chlorination. This pattern has been reported previously as “typical” for European surface soils and air (e.g., 20, 21). It is also very similar to the homologue pattern observed in archived Rothamsted surface soil collected in the 1800s (22-24). The pattern will have been influenced by the array of atmospheric sources, which have controlled the cumulative atmospheric deposition signal over past decades/centuries (25), by the possible formation of 4786
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 13, 2005
FIGURE 2. Average PCDD/Fs patterns: (a) general pattern in U.K./ Norway soils, and (b) U.K./Norway with land uses (the error bars represent the standard deviation). PCDD/Fs in soils (21), and by the loss processes (i.e., volatilization, leaching, degradation) which affect the composition of POPs in surface soils (7). Land Cover. Table 1 and Figure 3 show separate data for the U.K. grassland, and U.K./Norwegian woodland samples. Coniferous and deciduous forests were both sampled. In general, the grassland soils had the lowest concentrations: 70-790 (mean ) 280) pg ∑(1-8)CDD/Fs/g (DW), compared to 130-3450 (mean ) 1000) pg ∑(1-8)CDD/Fs/g in U.K. woodland. The Norwegian woodland samples contained from 21 to 1650 (mean ) 480) pg ∑(1-8)CDD/Fs/g, if site 18 is excluded. Previous studies have shown that woodland surface soils generally contain higher concentrations of POPs (including PCDD/Fs) than grassland soils (3-5). This may be explained by differences in scavenging/deposition mechanisms between woodland and grassland soils (3-5, 26, 27), and/or the turnover/mixing processes operating in the different soil systems. Bioturbation is minimal in coniferous soils, for example, but occurs in grasslands. The SOM content and composition differ between these soil types, but even if the data are expressed per unit OM content, the woodland soils generally contain higher concentrations than the grasslands (see Figure 3b). Figure 4 highlights this point further, presenting the average ∑(4-8)PCDD/F homologue concentrations in the three land use types on a DW and OM basis. Concentrations per g SOM are affected by a complex integral of many factors, such as the relative rates of organic matter supply (e.g., litter fall), the depth distribution and mixing rates of SOM, and the relative rates of SOM and compound degradation. Long-Range Atmospheric Transport of PCDD/Fs. Compound physical-chemical properties play an important role in their long-range atmospheric transport (LRAT) (15, 16). The LRAT of PCDD/Fs will be influenced by their vapor/ particle partitioning, which will affect their susceptibility to (a) chemical transformation, including photolytic destruction and atmospheric reaction (e.g., with hydroxyl radicals); (b) wet deposition; and (c) dry particle and dry gaseous deposi-
TABLE 1. Summary Data on PCDD/F Congeners and Homologues in Surface Soils (0-5 cm Depth) Expressed on Both Dry Weight Basis and Organic Matter Basis GL UK (pg g-1 DW) (n)16) mean
min
GL UK (pg g-1 OM) (n)16) max
mean
min
WL UK (pg g-1 DW) (n)17) max
VOL. 39, NO. 13, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF
3 2 2 3 2 2 0.3 13 1 17
0.4 1 1 1 1 1 0.1 3 0.3 4
13 9 8 12 8 8 2 50 5 81
12 9 9 14 9 9 2 57 5 75
2 2 2 2 2 2 0.1 9 1 11
62 30 29 58 29 31 17 200 18 320
2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD
0.1 1 1 2 3 16 78
0.03 0.2 0.3 1 1 6 25
1 2 3 5 9 44 200
1 3 4 7 11 71 330
0.1 1 1 1 4 16 116
3 9 11 23 31 220 820
44
440
630
173
Σ Isomers total mono-furans total di-furans total tri-furans total tetra-furans total penta-furans total hexa-furans total hepta-furans OCDF total mono-dioxins total di-dioxins total tri-dioxins total tetra-dioxins total penta-dioxins total hexa-dioxins total hepta-dioxins OCDD Σ P(4-8)CDD/F Σ P(1-8)CDD/F I-TEQ
9
4787
a
150 19 41 5 13 17 19 19 17
230 280 4
DL: method limit of detection.
b
min
PCDD/Fs congeners 18 2 8 1 7 1 8 1 7 1 7 1 1 0.1 34 4 3 0.3 42 4
max
mean
73 28 27 37 27 29 3 170 18 300
42 18 16 19 15 15 1 81 7 100
2 7 7 14 20 120 380
1 5 5 10 20 90 410
min
WL Norway (pg g-1 DW) (n)21)
WL Norway (pg g-1 OM) (n)21)
max
mean
min
max
mean
min
max
6 2 2 2 2 2 0.1 8 1 7
190 74 71 98 70 76 7 460 48 800
9 4 4 6 5 4 0.4 23 3 30
0.2 0.1 0.2 0.2 0.1 0.1 0.03 1.1 0.9 0.9
30 20 20 27 20 20 1 97 10 130
12 7 6 9 7 7 1 37 5 49
1 0.4 0.5 1 1 0.4 0.03 3 1 4
42 28 27 37 27 28 2 130 13 180
0.1 1 1 2 3 15 64
4 19 19 36 50 310 1100
0.3 1 1 3 4 22 89
0.03 0.02 0.08 0.06 0.1 1 4
1 5 6 11 17 90 330
0.4 2 2 4 6 33 130
0.03 0.1 0.2 0.2 0.5 3 16
1 7 8 15 23 125 450
0.5 2 2 5 9 41 170
0.04 0.4 0.4 1 1 7 39
1750
360
66
1300
860
130
3300
291
8.5
2171
399
35
2275
91 130 280 210 180 160 120 100
2 20 13 35 23 20 13 7
340 1000 1800 910 790 780 730 800
20 19 24 54 43 43 34 30
0.4 2 1 3 2 1 1 1
58 61 90 150 190 190 140 130
29 25 32 73 62 64 53 49
2 6 3 3 4 4 4 4
70 64 94 190 260 270 200 180
ND 250 93 81 130 190 230 380
ND 120 37 54 99 170 200 410
ND 3 3 12 19 31 33 64
ND 690 160 210 350 490 620 1100
5 8 3 10 20 37 46 89
5 3 0.2 1 1 1 2 4
5 19 13 56 75 140 170 330
16 10 4 13 28 53 68 130
6 6 1 1 3 5 7 16
25 26 13 58 100 190 240 450
2500 3400 48
1700 2300 30
290 450 5
6700 8000 126
560 475 10
15 20 0.2
4100 4440 78
750 855 14
55 75 0.8
60 220 17 55 77 74 77 81
68 230 22 52 70 81 84 75
10 14 5 16 15 14 15 11
240 1600 110 240 220 300 280 320
PCDD/Fs homologues 37 2 120 59 6 480 110 3 540 93 11 350 75 11 300 68 9 300 52 7 280 42 4 300
