Environ. Sci. Technol. 2003, 37, 5778-5782
Pyrene Degradation in the Rhizosphere of Tall Fescue (Festuca arundinacea) and Switchgrass (Panicum virgatum L.) YEN-CHIH CHEN,† M. KATHERINE BANKS,* AND A. PAUL SCHWAB‡ School of Civil Engineering, Purdue University, West Lafayette, Indiana 47907
A growth chamber study was conducted to investigate the fate of pyrene in the rhizosphere of tall fescue (Festuca arundinacea) and switchgrass (Panicum virgatum L.). For this study, 14C-labeled pyrene was used, and distribution of 14C activity was assessed after plant establishment. After 190 days of incubation, 37.7 and 30.4% of 14C-pyrene was mineralized in the soil planted with tall fescue and switchgrass, respectively, while 4.3% mineralization was observed for the unplanted control. Only 7.6 and 8.7% of pyrene was recovered from the soil in the two planted treatments, while 31.5% of pyrene remained in the unplanted control. Significant amounts of 14C were observed for all treatments and controls in the humic/fulvic fraction of soil at the end of the experiment. This research indicates the potential for pyrene mineralization in planted systems, although the ultimate fate of degradation byproducts is uncertain.
Introduction Phytoremediation is defined as “the use of plants and associated microorganisms to degrade or immobilize contaminants in soil and groundwater” (1). This remediation approach is an in-situ, nondestructive, and relatively costeffective technique. Mechanisms of phytoremediation for hazardous organic contamination in soil include (1) direct uptake by the plant, followed by metabolism, volatilization, or accumulation of contaminants; (2) microbial degradation stimulated by root exudates; and (3) co-metabolism of contaminants in the rhizosphere (1, 2). Laboratory and greenhouse experiments have demonstrated enhanced removal of PAHs (polycyclic aromatic hydrocarbons) when compared to nonplanted controls (36), and phytoremediation field trials have resulted in accelerated reduction of PAHs and other petroleum hydrocarbons (7, 8) in the rhizosphere. Recently, Liste and Alexander (9) reported enhanced degradation of pyrene by nine different plant species, including three field crops, three horticultural plants, and three pine seedlings. Pyrene was * Corresponding Author: (765) 496-3424; fax: (765) 496-3449; e-mail:
[email protected]. † Current address: Department of Civil and Environmental Engineering, Bucknell University, Lewisburg, PA 17837; e-mail:
[email protected]. ‡ Department of Agronomy, Purdue University, West Lafayette, IN 47907. 5778
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
reduced by 74% in vegetated soil compared to less than 40% in unplanted soil. Quantification of the parent compound only provides information about the overall disappearance of the target contaminant. The ultimate fate of the contaminant and the bioavailability of resulting byproducts are unclear if only total concentrations are assessed. Previous studies have shown bound residues to be a significant end point for PAHs in soil, especially in the humic/fulvic acid and humin phases (10, 11). The formation of these bound residues is significant during remediation in the presence of soil organic carbon. Limited information is available about the fate of PAHs during phytoremediation. However, the formation of bound residues is expected to be high due to the substantial input of organic material during plant growth. The ultimate fate of PAHs and byproducts in soils undergoing phytoremediation is an important component of end-point assessment. For this research, radiolabeled pyrene was used to evaluate the fate and bioavailability of PAHs in the rhizosphere of tall fescue (Festuca arundinacea) and switchgrass (Panicum virgatum L.) in a controlled growth chamber. We were able to quantify not only the amount of parent compound transformed into CO2 (mineralization) but also the portion incorporated into bound soil residues. The results from this work will advance our understanding of phytodegradation end points and potential human health risk for sites undergoing phytoremediation.
Experimental Approach Soil Preparation. A silty loam soil with no detectable PAHs was used in this study. The soil was collected from prairie grassland in Shades State Park, Indiana. The top 5-20 cm of soil was collected by soil cores. The soil was first sieved through an 8 mm sieve before exposure to pyrene. The soil was contaminated with 14C-pyrene (Sigma Scientific) and unlabeled pyrene resulting in a total concentration of 50 mg of pyrene/(kg of dry soil). The labeled contaminant was pyrene-4,5,9,10-14C with 58.7 mCi/mmole (98% purity, SigmaAldrich, St. Louis, MO). For the contamination process, pyrene was dissolved in acetone and sprayed onto the soil. The soil was mixed and air-dried for 3 days to allow for evaporation of the acetone. The soil properties are shown in Table 1. Exactly 1.2 kg of soil was placed in each pot. The initial amount of radioactive pyrene in each pot was determined by soil combustion with a Packard Sample Oxidizer (Model 307, Packard Bioscience Co., Meriden, CT). The combustion efficiency was tested and found to be 99.9 ( 1.8%. Plant Establishment. Two prairie grasses, tall fescue (Festuca arundinacea (F. arundinacea)) and switchgrass (Panicum virgatum (P. virgatum) L.), were selected for this research on the basis of prior performance in field studies (12). Plant seeds were germinated and grown for 1 additional week in perlite before transferring to the growth chambers. The seedlings were placed in an opening between the two chambers and sealed with a 1:4 wax and petroleum jelly mix at the top root section. Growth Chamber Design. Nine chambers (three fescue, three switchgrass, and three controls) were used in this study. Figure 1 illustrates the plant chamber. The foliar chamber had a height of 60 cm and a diameter of 30 cm, which resulted in a total volume of 44.5 L. Air flow was maintained by vacuum through the chamber to reduce 14C loss. A flow rate of 100 mL/min was determined to be sufficient for CO2 input and moisture removal. The root chamber had a similar design with a total volume of 1.5 L and 10 mL/min for flow rate. A 10.1021/es030400x CCC: $25.00
2003 American Chemical Society Published on Web 11/12/2003
TABLE 1. Soil Chemical and Physical Properties organic matter nitrate-Na phosphorusa potassiuma magnesium calcium sulfur zinc a
4.1% 143 mg/kg 28 mg/kg 132 mg/kg 464 mg/kg 2183 mg/kg 19 mg/kg 2.6 mg/kg
manganese copper iron boron soil pH sodium total CEC soluble salts
12 mg/kg 1.6 mg/kg 125 mg/kg 0.8 mg/kg 5.4 18 mg/kg 19 cmolc/kg 1 dS/m
Concentrations represent available fractions.
5 cm plug of wax and petroleum jelly (1:4) was used to seal the opening between the foliar chamber and the root chamber. Tensiometers were placed through the wall of the root chamber and used for soil moisture detection. During the experimental period, the soil moisture potentials ranged between 0 and -5 kPa. Water and fertilizer were applied by injection through a stopper to prevent leakage. Miracle-Gro was applied once every other week at the manufacturer’s recommended level for healthy plant growth (approximate concentrations introduced: N, 130 mg; P, 22 mg; K, 22 mg). A 16 h light period per day was maintained with 12 fluorescent lamps. After 3 weeks of plant growth, one tall fescue plant and one switchgrass plant did not survive due to poor root development. Therefore, final replicates of the planted treatments were 2, while there were 3 unplanted control replicates. Gas-Phase Sampling. Two sorbents were used to trap volatile organic compounds (VOCs) in the system. Glass tubes of 8 mm o.d. and 10 cm length were used for sorbent packing. Two sections containing 0.5 g of charcoal were used in one tube with glass wool separating the sections. Two sections of 0.25 g of XAD-2 (Supelco, Bellefonte, PA) were packed using the same method. The XAD-2 was prewashed with methanol and deionized water. Charcoal was used to trap polar VOCs, and XAD-2 was used to trap nonpolar VOCs. For the root chamber, one tube of charcoal and one tube of XAD-2 were used. Two additional charcoal tubes were used for the foliar chamber. If the last section of the tube showed a radioactive response, an extra tube was added to prevent overflow. Tubes were capped and kept at 4 °C before analysis. All sections were combusted using an oxidizer and counted with liquid scintillation for 14C (Tri-Carb 2900TR, Packard Bioscience). A 2.5 min combustion time was chosen for sample analysis. A recovery test was conducted along with each analysis event with a minimum recovery of 98%. A 50 mL aliquot of 2 N NaOH was used for CO2 trapping (10 mL for the root chamber: the removal efficiency was tested with an IR CO2 detector). A 1 mL aliquot of the solution was mixed with Hionic-Fluor (Packard Bioscience) and assessed to determine the total 14CO2 evolved. Trapping sorbents and solutions were renewed every 5 days. Soil Fractionation. To evaluate the presence of pyrene and byproducts in the soil, a modified serial extraction method was used (10). Soil was extracted into seven different fractions, including aqueous phase, organic phase, humic/ fulvic acid, post-base organic phase, strong acid phase, poststrong acid organic phase, and residuals. Soil extractions were performed by shaking soil with solvent for 24 h, settling for 16 h, and centrifuging at 10000g for 20 min. All aqueous samples were liquid-liquid extracted by 15 mL of dichloromethane. Residuals constitute the remaining matrix after all extractions, and radioactivity was determined by combustion through an oxidizer. The extraction procedure is as follows. A 10 g amount of soil was extracted in the dark with 25 mL of 20% H3PO4/H2O solution for 24 h in a 40 mL Nalgene Oak Ridge Teflon FEP centrifuge tube (Fisher Scientific, Hanover Park, IL). The soil was then extracted 3 times with 15 mL of dichloromethane/methanol (3:1) for 24 h each.
Humic/fulvic acid was removed by two extractions with 15 mL of 0.1 M NaOH and purged with nitrogen. A 20 mL aliquot of strong acid (10% HF/HCl) was added to liberate the strongly bound pyrene. The solution was mixed and then allowed to sit in the dark for 7 days before centrifugation. Post-base and post-acid organic extractions were also performed after base and acid extractions. The remaining soils were freeze-dried and combusted for the residual phase. All liquid phases were analyzed for radioactivity by placing 1 mL of solution into 10 mL of Ultima Gold or Hionic-Fluor (Packard Bioscience) for organic and base solutions, respectively, and counted using a liquid scintillation counter. For pyrene analysis, all liquid fractions were either liquidliquid extracted or replaced with suitable amounts of acetonitrile before HPLC analysis. Extracts were analyzed by HPLC chromatography on a Varian 9012 solvent delivery system and a Supelco LC-PAH column (250 mm × 4.6 mm, i.d.) using an elution gradient from 60/40 acetonitrile/water to 100% acetonitrile at 18 min with a flow of 1 mL/min. Pyrene was detected with Varian 9050 UV-vis detector at 245 nm. The HPLC data were analyzed with Varian Star System Control software. Peaks of interest were manually collected and counted by liquid scintillation to determine pyrene byproducts. Plant Analyses. To evaluate the amount of pyrene and byproducts absorbed into plant tissue, all plant parts were harvested and divided into leaves and roots sections. All roots were washed before analyses. Both shoots and roots were freeze-dried and weighed for determination of dry biomass. Triplicate plant subsamples (0.5 g) were combusted using an oxidizer and 14C quantified. Statistical significance was evaluated using SAS version 8.12 (SAS Institute, Cary, NC) with one-way ANOVA and least significant difference (LSD) for comparison of treatment means with p e 0.05.
Results Mineralization of Pyrene. Over the 190 day experimental period, approximately 37% of the added 14C was evolved as CO2 from the root chamber planted with tall fescue, 30% from the soil planted with switchgrass, and 4% from the unplanted control. The differences are statistically significant between planted treatments and the unplanted control (p e 0.05). Figure 2 illustrates the cumulative evolution of 14CO2 in the treatments. To statistically analyze the data set, transformation of the data was necessary. An Arsin transformation (sin-1 xvalue) was performed for the cumulative 14CO analysis. Regression of the data and least significant 2 index (LSI) were then performed, and there was no significant difference between tall fescue and switchgrass treatments at p e 0.05. However, both planted treatments were significantly different from the unplanted control. The two planted treatments had a distinctive lag period for the first 20 days of the study. The 14CO2 rapidly evolved after the lag period and continued for approximately 120 days before a plateau developed. The lag period was most likely the result of limited root influence in the early stages of the experiment. The 14CO2 in the leaf chambers was 5 and 2% of the initial 14C for tall fescue and switchgrass, respectively. The amount recovered from the control chamber was less than 0.5%. Similar statistical transformations were performed, and the two treatments showed significant differences throughout the experimental period. The radioactivity recovered from plant biomass was 8% of the initial 14C for tall fescue and 5% for switchgrass. By the end of the experiment, the total dry weight of tall fescue was 8.4 ( 0.7 and 3.2 ( 0.3 g for leaves and roots, respectively, and 7.3 ( 0.6 and 3.0 ( 0.5 g for switchgrass leaves and roots. The overall mass balance based on 14C activity is shown in Table 2. The total 14C recovery ranges from 92 to 107%. The unplanted control showed minor reduction of 14C from VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5779
FIGURE 1. Diagram of the growth chamber. Letters denote the following: charcoal (A, F, J, K, S, W); flow meter (B, C, L, N); water condenser (D, Q); XAD-2 (E, R); NaOH (G, H, T, U); empty vial (I, V); irrigation port (M); leachate collection (N); tensiometer (O); growth light (X); and vacuum pump (Y).
FIGURE 2. Cumulative 14CO2 from the root chambers. Percent CO2 of the total initial pyrene is shown. Error bars represent 1 standard deviation (9, tall fescue; 0, switchgrass; 2, unplanted control; n ) 3 for control; n ) 2 for tall fescue and switchgrass).
TABLE 2. Overall Mass Balance (%) of 14C in Planted and Unplanted Growth Chambers after 190 daysa,b control CO2 (root chamber) CO2 (leaf chamber) VOCs plant biomass soil total recovery
4.3 ( 1.4
