Environ. Sci. Technol. 2001, 35, 1103-1110
Partitioning, Extractability, and Formation of Nonextractable PAH Residues in Soil. 1. Compound Differences in Aging and Sequestration GRANT L. NORTHCOTT* AND KEVIN C. JONES Department of Environmental Science, Institute of Environmental and Natural Sciences, Lancaster University, Lancaster, LA1 4YQ United Kingdom
This study was carried out to assess the influence of physicochemical properties on PAH sequestration in sterile sewage sludge-amended arable soil. Radiolabeled phenanthrene (14C-9-Phe), pyrene (14C-4,5,9,10-Pyr), and benzo[a]pyrene (14C-7-B[a]P) were spiked and aged for up to 525 days in sterile soil microcosms. The degree of compound sequestration at various sampling times was determined by their extractability with organic solvents and release from soil residues by base saponification extraction. The amount of PAH extractable by butanol and dichloromethane decreased with compound aging in the soil. The decrease in PAH extractability with aging, and the formation of nonextractable bound residues, increased with compound molecular weight, Kow and Koc. The amount of total extractable PAH determined by sequential dichloromethane soxtec and methanolic saponification extraction decreased from 98%, 97%, and 94% at day 10 to 95%, 91%, and 77%, respectively for 14C-9-Phe, 14C-4,5,9,10-Pyr, and 14C-7B[a]P after 525 days aging. During the same aging period there was an increase in the amount of PAH released from the soil by base saponification extraction, suggesting a progressive diffusion of PAHs into hydrolyzable and recalcitrant organic matter and mineral phases of soil. Calculated half-lives for the apparent loss of PAHs by sequestration in this experiment were dependent on the method used to extract them from soil. These half-lives ranged from 96 to 1789 days depending on the compound, and are in agreement with values obtained from previous spiking experiments using nonsterile soils. These results suggest that a considerable fraction of PAHs assumed degraded in previous studies may have been sequestered within the organic carbon and, to a lesser extent, mineral phases of soil.
Introduction The sorption of hydrophobic organic compounds (HOCs) to soil and sediment is an important process controlling their environmental fate and effects. Sorbed compounds are less available for partitioning and leaching in groundwater and exhibit reduced bioavailability, toxicity, and genotoxicity * Corresponding author phone: (+44) 1524 593300; fax: (+44) 1524 593985; e-mail:
[email protected]. 10.1021/es000071y CCC: $20.00 Published on Web 02/10/2001
2001 American Chemical Society
compared to dissolved counterparts. Organic compounds that persist in soil exhibit declining extractability and bioavailability to microorganisms and other soil organisms (including earthworms and invertebrates), with increasing contact time or “aging” (1-3). In the past it was assumed that these observations were due to the degradation of contaminants by microbial processes in soil. However, studies utilizing isotopically labeled compounds have demonstrated that significant amounts of compound are retained in soil as nonavailable and nonextractable sequestered residues (4). Furthermore, the proportion of sequestered residues increases with increasing soil contact time or aging. Aging is associated with a continuous diffusion and retention of compound molecules into remote and inaccessible regions within the soil matrix, thereby occluding the compounds from abiotic and biotic loss processes. Various mechanisms have been proposed to describe these processes and are discussed extensively elsewhere (5-7). The sorption of low polarity nonionic organic contaminants in soil and sediment with organic carbon (OC) >0.1% is controlled by organic matter (OM) interactions (8, 9) as solute adsorption on soil minerals is strongly suppressed by the strong dipole interaction between minerals and water molecules (10). This dominant role has been demonstrated by studies that establish the importance of both the amount and physical and chemical properties of OM for organic compound sequestration (11). PAHs and other HOCs generated by human activities accumulate in large amounts in sewage sludge which may be applied in this matrix to agricultural soils (12, 13). The sorption behavior of HOCs in soil is an important consideration in determining the effects of sewage sludge application to agricultural land. Sewage sludge can be repeatedly applied to agricultural soils to improve their nutrient status, organic matter content, aggregate properties, and crop production. However, the advantages provided by sewage sludge application have to be considered against a corresponding increase in the heavy metal and organic contaminant burden of soils and the potential for contaminant transfer to field crops and food chains. Contaminant sequestration, either in the sludge itself or in the amended soil, may be an important process for mediating contaminant availability, therefore allowing for the maintenance, or increase, of current sewage sludge application rates. This paper is the first of two examining the sequestration of PAHs in sewage sludge-amended arable soil. The objectives of this study were to assess the influence of PAH physicochemical properties and the effect of time, on their sequestration in sterile arable soil. Soil microcosms individually spiked with radiolabeled phenanthrene (14C-9-Phe), pyrene (14C-4,5,9,10-Pyr), and benzo[a]pyrene (14C-7-B[a]P) were sterilized, stored, and sampled over a period of 525 days. At various times the distribution of aged PAHs was determined by solvent extraction using butanol (BuOH) and dichloromethane (DCM), methanolic saponification extraction (MSE), and total sample oxidation of remaining residues. A unique feature of this study is the simultaneous application of this range of techniques for determining changes in the distribution of PAHs aged in soil for a period of up to 525 days. We discuss the aging phenomena as a loss mechanism by comparing the half-lives we determined for apparent compound loss in our study, with those from comparable studies of PAH loss processes in soil. VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1103
Materials and Methods Chemicals. Nonradiolabeled phenanthrene (Phe), pyrene (Pyr), and benzo[a]pyrene (B[a]P) and the 14C-radiolabeled analogues, 14C-9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[a]P were obtained from Sigma Aldrich Co. Ltd., U.K. Nonradiolabeled Phe (96% purity), Pyr (99%), and B[a]P (>98% purity) were obtained as crystalline solids while 14C-9-Phe (g98% radiopurity), 14C-4,5,9,10-Pyr (95% radiopurity), and 14C-7-B[a]P (g98% radiopurity) were supplied in 1 mL of toluene with total activities of 3.7 × 106 Bq (100 uCi). Acetone, butanol (BuOH), dichloromethane (DCM), and methanol (MeOH) were obtained from MerckBDH U.K. and Rathburn Chemicals Ltd., U.K. Sample oxidizer reagents, Carbosorb E (trapping solution) and Permafluor-E (scintillation fluid), the organic based combustion aid solution Combustaid, sample combustion cones and pads, 14C-Spec-Chec solution, Ultima Gold XR (UGXR) scintillation cocktail, and 23 mL plastic scintillation vials were obtained from Canberra Packard, U.K. Anhydrous sodium sulfate and analytical grade potassium hydroxide were obtained from MerckBDH, U.K. Compound spike solutions were prepared by adding aliquots of each radiolabeled stock solution to 10 mL of the corresponding nonradiolabeled 10,000-ppm PAH standard in acetone. The activity of the PAH spike solutions was determined by oxidizing 5 µL aliquots of each solution (n ) 5) and liquid scintillation counting (LSC) of the trapped activity.
Experimental Section Soils. Field-wet subsurface (5-20 cm) soil (sandy loam) was collected in November 1997 from a rural arable field near Kirkham in the Wyre Valley, Lancashire (U.K.-O.S. sheet 102 [397338]). This field had been amended with dewatered sewage sludge cake 10 months previously. The sampling and pretreatment of the soil is fully described elsewhere (14). DW of the prepared soil was 82.0 ( 0.2% as determined by drying to a constant weight at 105 °C (n ) 20). Soil total carbon, inorganic carbon, and organic carbon was 2.25 ( 0.09%, 0.15 ( 0.02%, and 2.10 ( 0.09%, respectively (n ) 20) (Carlo Erba CHNSO Elemental Analyzer EA 1108-R). Sewage Sludge. Anaerobically digested sewage sludge from the NorthWest Water Davey Hulme Wastewater Treatment Plant (WWTP) in Manchester, U.K. was concentrated by centrifuging at 10,000 rpm (15,300g) for 25 min (Beckman J2-21 at 4 °C) in polyethylene centrifuge tubes. The sampled liquid anaerobic sewage sludge and centrifuge concentrated sewage sludge solids contained 2.3% and 14.9% dry weight of solids, respectively, as determined by drying to constant weight at 105 °C. Organic matter and carbon content of the centrifuge concentrated sludge was 8.8 ( 0.3% (loss-onignition at 400 °C) and 5.1 ( 0.2%, respectively (conversion factor of 1.724) (15). Soil Spiking Procedures. Soil was spiked using a fully validated and described procedure (14). Briefly, 250 g of fieldwet soil and 6.7 g of centrifuge concentrated sewage sludge was added to a blender containing 10 mL of acetone and 250 µL of the appropriate radiolabeled PAH spiking standard. The contents were blended to mix, transferred to glass microcosms, and sterilized by gamma-irradiation. Concentrated sewage sludge was added to the soil at an equivalent loading of 40 tonnes/hectare, representing the highest level of sludge application to soil in the U.K. Unspiked soil microcosm blanks were prepared using an equivalent volume of acetone. Sewage sludge-amended soil microcosm replicates were spiked to a concentration of 10 mg kg-1 for nonradiolabeled Phe, Pyr, and B[a]P and 79.6 (4776 DPM.g-1), 96.1 (5766 DPM.g-1), and 89.9 Bq.g-1 (5394 DPM.g-1) for 14C9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[a]P, respectively. The soil microcosms were stored under standard laboratory conditions with mean winter and summer temperatures in the laboratory of 18 ( 2 °C and 22 ( 2 °C. 1104
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
FIGURE 1. Combined sample extraction and analysis scheme. Soil Extraction Procedures. Replicate samples from the soil microcosms were subjected to BuOH shake extraction, DCM soxtec extraction, and methanolic saponification extraction (MSE) using the procedures summarized in Figure 1. Full details regarding the validation and optimization of the extraction methods are available elsewhere (16). Briefly, wet soil was extracted with BuOH in 50 mL Teflon Oak Ridge Centrifuge tubes on a flat bed shaker and an aliquot removed for LSC. For DCM extractions the soil was mixed and ground with anhydrous granular sodium, transferred into Whatman cellulose extraction thimbles, and extracted using a TecatorSoxtec System HT 1043 Extraction Unit. The thimbles were immersed in 40 mL of boiling DCM for 30 min then raised for five and a half hours extraction with freshly distilled DCM. After extraction an aliquot of DCM was removed for LSC. The DCM extracted soil residues were left in a fumehood overnight to evaporate residual solvent, weighed, and subsampled for MSE. MSE was carried out by mixing 2 g of soxtec extracted soil residue with 9 mL of methanolic-KOH lye (14:1, MeOH: 2 M KOH, pH 12) in 50 mL Teflon test tubes and heating in a water bath at 90-95 °C for 5 h (17). The tubes were cooled and centrifuged at 5000 rpm for 30 min, and 5 mL of supernatant was removed for LSC. The saponified solid residue was recovered by filtration through Whatman 1 filter paper, rinsed with MeOH, and dried overnight in a fumehood. Activity of Spiked Soil, Extracted Soil Residues, and Soil Solvent Extracts. Compound activity in soil and extracted soil residues was determined by total sample oxidation using a Packard 307 Sample Oxidizer. Trapping efficiency was established by combusting aliqouts of 14C-Spec-chec solution (>98%). 14CO2 was eluted from the sample oxidizer with 10 mL of Carbosorb E and combined with 10 mL of Permafluor scintillation cocktail in 23 mL plastic scintillation vials. The activity of radiolabeled compound in BuOH, DCM, and methanolic KOH solvent extracts was determined by hand pipetting 5-10 mL of solvent extract into 23 mL plastic scintillation vials containing 10-17 mL of UGXR. The vials were capped, shaken, and stored overnight in the dark to stabilize and counter the effects of induced chemilumines-
cence and photoluminescence. Sample vial activity was counted with a Canberra Packard Tri-Carb 250CA liquid scintillation analyzer for a period of 10 min or sufficient time to provide 50× background count values. The effect of variable chemical and color quenching was corrected by preparing individual quench correction curves for each solvent type using solvent or methanolic saponification extracts of unspiked blank soil. Blank samples for total sample oxidation and solvent extraction were prepared on each sampling occasion using samples from corresponding aged unspiked soil microcosm blanks (acetone only). Measured activity was corrected by subtraction of appropriate blanks and adjusted for sample oxidizer trapping efficiency.
Results and Discussion Triplicate microcosms spiked with 14C-9-Phe, 14C-4,5,9,10Pyr, and 14C-7-B[a]P and a single blank spiked control were sampled immediately after being returned to the laboratory 10 days after spiking and thereafter at 21, 53, 170, 259, and 525 days. The total spike activity within each sampled microcosm was determined by total oxidation of multiple subsamples (n ) 6 or 10). Total Activity and Spike Recovery. Mean spike recovery for multiple sampled individual soil microcosms was excellent with values of 101 ( 7%, 104 ( 7%, and 102 ( 5% for 14C9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[a]P, respectively (mean ( 95% confidence interval, n ) 17). The stability of soil moisture content and PAH spike activity during the 18 month aging period demonstrated the integrity of the microcosm sealing and storage methods and maintenance of sterile conditions. Full details of the mean activity of spiked compound, spike homogeneity (RSD), and stability in each soil microcosm are reported elsewhere (14). While the overall mean spike activity calculated for all microcosms was very good, some individual microcosms exhibited a wide range of activity and spike homogeneity. This is not unexpected, given the heterogeneous nature of soil matrices and the difficulty of introducing spiked compounds into soils in a reproducible manner. Due to variation in spike activity and heterogeneity of distribution between microcosms, the results for compound extractability and residue formation are expressed as a percentage of the total soil spike activity measured for each individual microcosm. Solvent Extractability of Soil Aged PAHs. The quantity of compound extracted from the soil with BuOH, DCM, and MSE, described as the percent extracted, decreased with increasing aging and compound molecular weight (MW) (see Figure 2A-C). Inspection of Figure 2A-C shows that the initial decrease in the amount of extracted compound also increased with compound MW and Koc. The amount of 14C-9-Phe and 4C-4,5,9,10-Pyr extracted by BuOH, DCM, and DCM/MSE appears to be approaching a constant amount after 525 days aging, while 14C-7-B[a]P appears to be decreasing further. The amount of 14C-9-Phe extracted by BuOH and DCM becomes significantly different (p ) 0.05) from that extracted at day 10, after 259 and 170 days, respectively (Table 1). Similar significant differences in BuOH and DCM extractability occur for 14C-4,5,9,10-Pyr after 259 and 259 days respectively, and for 14C-7-B[a]P after 21 and 53 days respectively. The DCM extraction results for 14C-9-Phe and 14C-4,5,9,10-Pyr at day 170 are inconsistent with the overall trend of reducing extractability, being higher that the amount extracted at day 53. This partly obscures a more statistically significant reduction in compound extractability after 53 days aging. MSE of Soil Aged PAHs. A significant amount of each PAH could only be recovered from the aged soil samples by MSE, and sequential MSE resulted in increased extractability for all PAHs at all sampling times compared to soxtec extraction by DCM alone. The absolute amount of PAH
FIGURE 2. Extraction of 14C-9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[a]P aged in sewage sludge-amended arable soil. The symbols represent the mean, and error bars the 95% confidence interval, of nine replicate measurements. If error bars are not distinguishable they are obscured by the symbol. A. BuOH; B. DCM; C. MSE. extracted by MSE generally increased with aging (Table 1), except for the samples on day 10. The amount of each PAH extracted by MSE on day 10 was greater than that at day 21. The samples at day 10 were taken immediately after the microcosms were returned to the laboratory following sterilisation. This result may be an artifact resulting from soil disturbance during spiking, sterilization, and transport. Variability of Extraction Data. The variability of the data obtained by the three extraction methods is excellent with 95% confidence intervals < 5% (Table 1). The BuOH extraction data exhibited higher variability (1.4-4.8%) than that obtained for DCM soxtec extraction (0.3-1.8%) and MSE (0.10.8%). This may be explained by BuOH extraction being a nonexhaustive equilibrium extraction. Nonexhaustive solvent extractions typically display higher variability than equivalent VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1105
TABLE 1. Extraction of 14C-9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[A]P from Aged Arable Crop Soilf % extracted aging (days)
BuOH (n ) 9)
10 21 53 170 259 525
99.3 ( 3.3aAc,d 95.0 ( 1.4aA 97.3 ( 1.5aA 95.8 ( 1.5aA 92.5 ( 2.0bA 91.3 ( 2.1bA
10 21 53 170 259 525 10 21 53 170 259 525
DCM (n ) 9)
MSE (n ) 6)
totala
% MSEb
14C-9-Phe 90.5 ( 1.2aB 94.4 ( 0.5bA 91.6 ( 1.8cB 92.5 ( 0.3cB 89.8 (0.3dB 88.9 ( 0.5eA
7.7 ( 0.1a 3.4 ( 0.1b 5.5 ( 0.1b n/ae 5.9 ( 0.1c 5.8 ( 0.2d
98.2 ( 0.1a 97.8 ( 0.1b 97.1 ( 0.1c n/ae 95.7 ( 0.1d 94.7 ( 0.3e
7.8 3.5 5.7 n/ae 6.2 6.1
98.1 ( 4.6aA 93.8 ( 1.7aA 93.1 ( 4.8aA 90.3 ( 3.2aA 87.9 ( 2.7aA 87.0 ( 2.7aA
14C-4,5,9,10-Pyr 90.4 ( 0.8aB 91.7 ( 0.9aA 87.4 ( 0.8bB 88.4 ( 1.2bA 84.2 ( 0.7cB 82.4 (1.0dB
6.7 ( 0.1a 5.0 ( 0.1b 7.6 ( 0.2c n/ae 7.7 ( 0.2d 8.3 ( 0.3e
97.0 ( 0.1a 96.6 ( 0.1b 95.1 ( 0.2c n/ae 91.9 ( 0.2d 90.7 ( 0.4e
6.9 5.1 8.0 n/ae 8.3 9.1
89.6 ( 1.5aA 82.4 ( 2.5bA 79.9 ( 4.0bA 76.0 ( 2.0bA 73.0 ( 2.1bA 68.5 ( 2.6bA
14C-7-B[a]P 83.1 ( 0.9aB 83.0 ( 1.1aA 77.1 ( 0.6bA 73.2 ( 0.7cB 68.1 ( 1.0dB 60.3 ( 1.4eB
10.6 ( 0.2a 9.7 ( 0.5b 12.0 ( 0.3c n/ae 15.0 ( 0.8d 17.0 ( 0.3e
93.7 ( 0.2a 92.8 ( 0.5b 89.1 ( 0.3c n/ae 83.0 ( 0.8d 77.3 ( 0.4e
11.3 10.5 13.5 n/ae 18.0 22.0
a Total ) sum of DCM and MSE. b Percentage of MSE relative to total. c Values in a column followed by the same undercase letter are not significantly different from previous value (p ) 0.05). d Values in a row followed by the same capital letter are not significantly different from previous value (p ) 0.05). e n/a, not determined. f Values are the mean ( 95% confidence interval.
exhaustive extractions (16). The BuOH extracts produced a higher level and wider range of radioactivity quenching than the other solvent extracts, resulting in a higher level and variability of quench correction during LSC. Regardless, the variability of the BuOH data, along with that for DCM soxtec and MSE, is less than that observed for the homogeneity of spiked compound in the soil microcosms (14). PAH Stability in Soil Microcosms. PAHs in solution are sorbed rapidly to the external surfaces of soil, primarily to organic materials. This rapidly sorbed fraction is readily available for rapid desorption, volatilization and photochemical reactions, biodegradation by microorganisms, toxicity to various organisms, and extraction by organic solvents. This rapid fraction decreases in size due to the above-mentioned loss mechanisms and by diffusion into remote sites within the soil matrix, thereby undergoing a slow conversion process to an increasingly sequestered and unavailable fraction. This fraction requires extraction by astringent and vigorous means and is greatly reduced in bioavailability to micro- and higher organisms. The amount of extractable contaminant in soil is reduced by the abovementioned processes as well as chemical oxidation and reduction, covalent binding to components of the soil matrix, and polymerization reactions. Since these chemical reactions transform the parent compound to new structures, they do not represent sequestration of the parent molecule itself. The degradation of PAHs with three or more aromatic rings in soil by abiotic reactions is not an important loss mechanism, and photolysis at the soil surface is the dominant abiotic degradation reaction for PAHs (18, 19). It is unlikely that biodegradation and photodegradation of spiked PAHs occurred in the sealed glass Kilner jar soil microcosms which were sterilized by gamma-irradiation and stored in the dark within tape sealed cardboard boxes. Maintenance of soil moisture content and radiolabeled spike activity for the duration of the experiment provides further evidence that compound loss by volatilization and biotransformation did not occur (14). Therefore, the formation of nonextractable PAH residues during the aging experiment is attributed to abiotic sorption processes. 1106
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
PAH Extractability by BuOHsthe “Mild” Extraction Solvent. A number of studies investigating the extractability of Phe (and other compounds) from soil and soil fractions has observed a decline in extractability with increased compound aging (1, 3, 20-22). Phe has been a popular model compound for many of these studies as it is polyaromatic and relatively biodegradable by soil microorganisms. These studies have observed a significant reduction in BuOH extractability of Phe over aging periods from 13 to 230 days (89% after 27 days (1), 70% after 230 days (20)) and quantitative, or near quantitative, recovery of Phe following sequential DCM extraction. In contrast, we observed a much smaller decrease in the extractability of 14C-9-Phe with BuOH and DCM over 525 days aging and did not achieve quantitative recovery of 14C9-Phe using these solvents in our experiment (Table 1). We did not carry out sequential BuOH and DCM soxtec extractions of the soil samples but instead carried out BuOH shake and DCM soxtec extractions on separate soil subsamples. A comparison of these results (Table 1) shows that BuOH generally extracted a higher amount of each PAH than DCM Soxhlet extraction (p ) 0.05). However, the variability of BuOH extraction data is greater than the DCM Soxhlet extraction results. If the DCM extraction data exhibited a similar level of variability to the BuOH data, the amount of compound extracted by both solvents would in many instances, become statistically indiscernible. It is also possible that drying the soil by grinding with anhydrous sodium sulfate prior to DCM extraction reduced compound extractability by the collapse of soil micropores and dehydration and compression of SOM. Another possible explanation is that BuOH is able to extract a fraction of SOM and associated PAHs, that is not DCM extractable. This may include water or alcohol soluble components of the soil humic substances. The key result from the comparison of compound solvent extractability is that unlike many previous published studies, we have not observed large differences in the amount of Phe extracted by BuOH and DCM. The differences in Phe extractability between previous studies and our own may arise from differences in soil OC.
The amount of organic material in the soil (Lima loam) used in the referenced experiments (1-3, 20, 22, 23) is expressed as either organic matter (4 to 11%) or OC (8.7%). Using a conversion factor of 1.724 (15) this approximates a range of 2.32 to 8.7% OC. In comparison, the arable crop soil used in our experiment had an OC level of 2.10% which increased to 2.17% OC upon the addition of centrifuged sewage sludge. The extractability of Phe aged in soil with high OC is substantially reduced compared to soils with lower levels of OC (1). The OC levels reported for the Lima loam cover a reasonably wide range of values. We expect the OC level of the Lima loam would have to be approaching the high end of this range (8.7% OC) to explain the large difference in Phe extractablity. Unfortunately the cited studies do not report details of the methods used to measure OC in the Lima loam soils, making it difficult to construct meaningful comparisons between these and our own results based on the amount of soil OC. Differences in methodological procedures between the referenced Lima loam studies and our own may explain the dissimilarity in observed compound extractability. These include the following. (1) Inexhaustive BuOH extraction: Short extraction times including brief manual shaking (1-3) and vortex mixing for 30 s have been employed (24). We have demonstrated that a minimum 6 h shake extraction is necessary to achieve a constant amount of extracted compound (16). If shorter extraction times are used sequential DCM extraction will recover the fraction of compound from soil that remains extractable by BuOH. (2) Different soil: BuOH ratios: The mass of soil and solvent volume (16) and the concentration of spiked compound (25) affect the amount of compound extracted by BuOH. We used a soil:BuOH ratio of 1:10 (partitioning experiment standard), but other ratios including 1:1.875 (2), 1:1.25 (1, 3, 20, 21, 24), and 1:5 (24) have been used. (3) Spiking Phe into air-dried soil, which is subsequently rehydrated: This practice should be avoided as it can introduce artifacts into experiments (26). (4) Calculating total mass balances: The reported studies did not employ total sample oxidation of spiked soils or solvent extracted soil residues to provide conclusive compound mass balances or to verify the existence of nonextracted compound residues (1-3, 20-24). The fraction of PAH extracted from soil by BuOH has been likened to the labile compound fraction, that is, compound in aqueous solution and that sorbed to surfaces but accessible to extractant solutions. The BuOH extractable fraction of Phe has been correlated with the microbially bioavailable compound fraction (1, 3, 21). Therefore, it is plausible to suggest that the reduction in BuOH extractability of aged PAHs that we observed indicates a reduction in compound bioavailability with aging. PAH Extractability by “Vigorous” DCM Soxtec Extraction and MSE. Our observation of decreasing DCM soxtec extractability with increasing aging of PAHs in soil is smaller in magnitude than, but in agreement with, the findings of previous studies. For example, PAH extractability by acetone sonication extraction decreased to 5% for naphthalene and approximately 40% for anthracene, Phe, fluoranthene, and Pyr after 25 days incubation in a nonsterile loam soil (1.1%) (17). MSE of the solvent extracted soil residues released significant quantities of the nonextracted PAHs, up to 8090% of the original added amount. Naphthalene and B[a]P binding to two nonsterile soils (2.2 and 2.9% OC) reached a maximum after 30 days aging, when 97% of Naph and 5060% of B[a]P were nonextractable by sequential ethanol shake and DCM Soxhlet extraction (27). In another study the extractability (DCM Soxhlet) of Pyr in forest soil (3.5% OC) decreased from 57% to 28% over 270 days incubation in nonsterile soils, while showing a slight increase from 72 to 78% in sterile soil (28).
FIGURE 3. 14C-9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[a]P residues remaining in aged soil after combined soxtec extraction and MSE. The columns represent the mean, and error bars the 95% confidence interval, of six replicate measurements. In comparison, the extractability of PAHs by DCM soxtec in our sterile soil decreased to 88.9, 82.4, and 60.3% for 14C9-Phe, 14C-4,5,9,10-Pyr, and 14C-7-B[a]P, respectively, after a total of 525 days aging (Table 1). The extractability of 14C4,5,9,10-Pyr after 525 days (82.4%) and 14C-7-B[a]P after 170 days aging (73.2%) is similar to published recoveries of Phe aged in sterile soils after 3.7 months incubation (1) and Pyr aged for 270 days (28). A number of explanations may exist for the large difference in compound extractability between these studies and our own. First, many of these other experiments have been conducted on nonsterile soils where microbial processes may enhance compound sequestration and/or partially degrade parent compounds to polar derivatives that react and chemically bind to SOM (29, 30). Polar metabolites of 14C4,5,9,10-Pyr were identified in aged nonsterile forest soil exhibiting significantly lower compound extractability than a sterile control (28). Second, two of the referenced studies spiked compounds into air-dried soil. Spiking organic contaminants into air-dried soil enhances compound sequestration by providing active sites and regions that are normally occupied by water molecules (26). Finally, the solvent extraction methods employed included acetone sonication and DCM Soxhlet extraction which may have reduced extraction efficiencies compared to the optimized DCM soxtec extraction procedure we employed (16). As noted here and in previous studies, MSE released significant quantities of PAH that remained nonextractable after DCM soxtec extraction (17, 29). The absolute amount of PAH extracted by MSE increased with aging and compound MW (Table 1). However, significant quantities of PAH were retained within the soil after this astringent extraction and increased with compound MW and aging (Figure 3). The total recovery of PAH from the soil with combined DCM soxtec and MSE was almost quantitative for 14C-9-Phe aged for 10, 21, and 53 days and 14C-4,5,9,10-Pyr aged for 10 and 21 days (Table 1). This is comparable to previously published results where quantitative recovery of Phe, Pyr, and low MW PAHs aged for 21 days in soil was only achieved after MSE of solvent extracted soil residues (17). At longer aging times, quantitative recovery of 14C-9-Phe and 14C-4,5,9,10-Pyr was not achieved, while 14C-7-B[a]P was not quantitatively extracted at any of the sampled aging times. This result demonstrates the enhanced formation of nonextractable residues of higher molecular weight and polycondensed PAHs in soil. Sequestration of PAHs in Soil and SOM. The dissolution of SOM by methanolic hydrolysis releases quantities of nonsolvent extractable organic contaminants that are either bound by chemical bonds to or strongly associated and/or VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1107
occluded within SOM (4, 17, 29). Hydrolyzing SOM splits labile ester bonds, resulting in a limited breakdown of the macromolecular humic network, improving solvent accessibility and extractability. Organic contaminants that are bound to the SOM by hydrolyzable bonds will similarly be released under these conditions. The introduction of alkaline conditions also causes an extension of the humic macromolecular polymeric structure as a result of mutual repulsion between negatively charged carboxyl, phenolic, and hydroxyl functional groups. This provides improved solvent accessibility within the humic macromolecular structure, promotes the release of diffusion retarded or sequestered contaminant fractions, and reduces hydrogen-bonding ability to promote contaminant release. The MSE lye has a pH of 12 and under these heated saponification reaction conditions humic substances will also be extracted from soil. These alkaline conditions may also dissagregate and solubilize some mineral phases, notably silicates and ferric oxides, thereby releasing any associated sequestered residues. PAHs will only undergo oxidative coupling to SOM if reactive metabolites are produced by the microbial degradation of parent compound. These partially oxidized products can then become incorporated into SOM as bound residues by the formation of enzymatically catalyzed covalent ether and/or ester bonds. There was no evidence of microbial activity in the spiked and gamma-irradiated soil microcosms as determined by the preservation of spiked activity, absence of extractable compound transformation products, and heterotrophic plate counts (14). It seems reasonable, therefore, to assume that the radioactive residues remaining in soil after sequential DCM soxtec and MSE are parent PAHs sequestered within recalcitrant SOM associated with soil humin, and not covalently bound metabolites. Soil humin contains a significant quantity of soil OC. Reported values range from 10-20% (31) and up to as high as 45% total soil OC (32). This organic material is responsible for the retention of significant quantities of organic contaminants by humin, particularly nonpolar semivolatile PAHs, PCBs, and chlorinated pesticides (27, 33-35). Recent studies, using the methyl isobutyl ketone (MIBK) humic fractionation procedure, have revealed that a significant proportion of PAH residues (and other nonionic semivolatile organic compounds) sequestered within humin are associated with bound humic acid (BHA) and bound lipid fractions, with minor quantities retained by the remaining mineral residue (27, 35). These humic fractions have been overlooked by traditional alkaline solution methods for humic substance extraction. Carbon-13 cross polarization magic angle spinning (13C CPMAS) NMR studies of these humin fractions have shown that aliphatic carbon content increases in the order HA < mineral-OM < BHA, and aromatic carbon content increases in the reverse order (33). The accumulation of PAHs in the BHA and bound lipid fractions of humin is consistent with observations that PAHs preferentially adsorb to environmental organic materials with relatively higher aromaticity and lower polarity (11, 36). Surface area and pore size measurements of humin have shown that pore sizes range from ∼2 to ∼360 nm and are dominated by micropores (2-50 nm pore diameter), with organic matter blocking most of the small pores (31). Simple organic molecules (such as PAHs) are nanometers in size themselves and can diffuse into these micropores. In contrast, fungal hyphae and bacteria are orders of magnitude larger (µm) (37) and are unable to penetrate into these microporous structures and degrade sequestered compound (38). Nonextractable PAH residues in soil are extremely stable with degradation rates comparable to the turnover rate of natural humic substances. Recent studies have demonstrated the stability of PAH bound residues in soil. The addition of humus degrading microorganisms (white rot fungi and bacteria), 1108
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
TABLE 2. Estimated Abiotic Pseudo-First-Order Rate Constants and Parameters for the Apparent Loss of PAHs from Sterilized Soil over the 525 Day Aging Period extraction
% extracted
1/2 life (days)
[PAH]IFb
R2
DCM MSE
89.9 94.7
14C-9-Phe 3.87E-04 4.96E-03
1789 140
72.85 94.56
0.47 0.99
DCM MSE
82.4 90.7
14C-4,5,9,10-Pyr 2.73E-03 254 7.23E-03 96
79.6 90.62
0.85 0.997
DCM MSE
60.3 77.3
14C-7-B[a]P 2.73E-03 254 3.33E-03 208
53.15 74.11
0.98 0.98
ka (day-1)
a k (day-1) ) rate of apparent compound loss in days. amount of PAH extractable from soil at infinity.
b
[PAH]IF )
organic material amendments, and physical treatment of soil did not lead to enhanced release and/or biodegradation of nonextractable PAH residues (39, 40). Further evidence for the sequestration of PAHs in humin has been obtained by 13C CPMAS NMR and flash pyrolysis-GCMS experiments, which demonstrated that 13C-Pyr residues in sediment humin were parent compounds and not covalently bound reaction products (34). We believe that the PAH compound fraction remaining in the soil residues after MSE in our experiment is the parent compound sequestered within the organic matter phase of humin, primarily that described as BHA and bound lipid fractions. MSE removes reactive hydrolyzable OM and associated PAHs from soil, while an increasing amount of compound remains associated with recalcitrant OM that is primarily associated with soil humin. It is also possible that other phases in the soil, including some minerals and sootlike combustion derived carbon, are also contributing PAH sequestration in the aged soil. More significantly, the fraction of compound released by MSE is overlooked by traditional methods of solvent extraction and is therefore excluded from source and environmental inventories and contaminant risk assessments. Half-Lives for the Apparent Loss of PAHs in Sterile Aged Soil by Sequestration. The plots of compound extractability versus aging exhibit an exponential decrease that can be modeled as a pseudo-first-order rate process. Standard firstorder decay expressions assume that the process being modeled continues until it approaches zero. However, the plots for DCM soxtec and MSE of PAHs from soil (Figure 2B,C) suggest that the amount of extracted compound approaches an amount significantly greater than zero. Therefore, we applied a modified exponential expression to our data incorporating an additional constant representing the amount of PAH that remains extractable from soil at infinity (eq 1)
[PAH]t ) [PAH]IF + [PAH]0 × e-kt
(1)
where [PAH]t is the amount of PAH at time t or the extractable amount of PAH at time t, [PAH]IF is the amount of PAH extractable from the soil at t ) infinity (the corresponding nonextractable fraction is 1 - [PAH]IF), [PAH]0 is the amount of PAH at t ) 0 or the amount of PAH originally spiked into the soil, k is the pseudo-first-order rate constant, and t is the time in days. The data for PAH extractability using DCM soxtec and MSE was modeled against this expression using the Windows based software package SigmaPlot 5.0 (SPSS Inc.) by setting a maximum of 100 iterations. The results of this modeling analysis appear in Table 2.
TABLE 3. Comparison of Half-Lives for the Apparent Loss of PAHs Aged in Sterilized Soil (Our Study) with Others Determined for PAHs Spiked in Soil or Analyzed in Sewage Sludge-Amended Soil spiked soil
sludged soil
compd
our study
(41)
(19)
(44)
(18)
(42)
(43)
(41)
Phe Pyr B[a]P
1789/140 254/96 254/208
14 51 112
16-35 199-260 229-309
