Remediation of Polycyclic Aromatic Hydrocarbon Compounds in

Feb 11, 2005 - which coincided with the propagation of the tree roots to the water table region. Remediation was limited to naphthalene and several th...
0 downloads 0 Views 643KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 1598-1605

Remediation of Polycyclic Aromatic Hydrocarbon Compounds in Groundwater Using Poplar Trees MARK A. WIDDOWSON,* SANDRA SHEARER,† RIKKE G. ANDERSEN, AND JOHN T. NOVAK The Charles E. Via Jr. Department of Civil and Environmental Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0105

A seven-year study was conducted to assess the effectiveness of hybrid poplar trees to remediate polycyclic aromatic hydrocarbon (PAH) compounds in soil and groundwater at a creosote-contaminated site. A reduction in the areal extent of the PAH plume was observed in the upper half of the 2-m-thick saturated zone, and PAH concentration levels in the groundwater declined throughout the plume. PAH concentrations began to decline during the period between the third and fourth growing seasons, which coincided with the propagation of the tree roots to the water table region. Remediation was limited to naphthalene and several three-ring PAHs (acenaphthylene and acenaphthene). PAH concentrations in soil and aquifer sediment samples also declined over time; however, levels of four-ring PAHs persisted at the lower depths during the study period. The naphthalene to total PAH concentration ratio in the most contaminated groundwater decreased from >0.90 at the beginning of the second growing season to approximately 0.70 at the end the study. Remediation in the lower region of the saturated zone was limited by the presence of a 0.3-m-thick layer of creosote present as a dense nonaqueous phase liquid (DNAPL). The nearly steady-state condition of the PAH concentrations observed during the last three years of the study suggests that the effectiveness of the phytoremediation system is limited by the rate of PAH dissolution from the DNAPL source.

Introduction Phreatophytic plants, particularly hybrid poplar trees, have been used for the in situ treatment of soil and groundwater contaminants and to control groundwater flow (1). Hybrid poplar trees are desirable for phytoremediation applications due to their high water uptake rates, easy propagation, deep root systems, and tolerance to high concentrations of contaminants (2). Phreatophytic phytoremediation systems can be effective in the capture of mobile aqueous-phase contaminants (3) and in the reduction of contaminant mass flux from source zones (4). Mechanisms for remediation of organic compounds by phreatophytes include (a) direct uptake and translocation and (b) rhizospheric degradation * Corresponding author phone: (540) 231-7153; fax: (540) 2317532; e-mail: [email protected]. † Present address: CH2M Hill, 2525 Airpark Drive, Redding, CA 96001. 1598

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

and filtration (1). Compounds that are removed from soil and groundwater through direct uptake (a) may undergo metabolism or transformation within the plant or (b) are passively volatilized (or some combination thereof). Rhizosphere bioremediation is attributed to (a) stimulation of microbial activity and contaminant transformation from the release of exudates and enzymes into the rhizosphere and (b) enhanced populations of bacteria and mycorrhizal fungi (5). Organic contaminants may sorb to roots, with the extent of sorption depending on the hydrophobicity of the compound (6,7). Researchers have investigated the efficacy of hybrid poplar trees and associated remediation mechanisms for a number of organic compounds. Varying degrees of uptake and metabolism have been demonstrated for atrazine (8), monoaromatic hydrocarbons (BTEX) compounds (7), trichloroethene (7, 9), MTBE (10), perchlorate (11), and explosives (12). These compounds exhibit a wide range of transpiration stream concentration factors (TSCF) and root concentration factors (RCF), which are indicators of susceptibility to direct uptake and sorption to roots, respectively (7). Although grasses have been shown to be effective in the remediation of surface soils contaminated with polycyclic aromatic hydrocarbon (PAH) compounds (13-16), the applicability of hybrid poplar trees to PAH plumes is a novel approach (17). Because aerobic and anaerobic biodegradation are known mechanisms for the remediation of PAHcontaminated aquifer sediment (18), rhizospheric bioremediation of dissolved-phase PAH compounds may be enhanced at sites where poplar roots reach the water table. PAH compounds exhibit a range of physical and chemical properties but may be characterized as moderately to strongly hydrophobic, as indicated by medium to high octanol-water partition coefficients (Kow) and low to medum solubility values in water. Table 1 lists the water solubility and Kow values for ten PAHs, listed in order by molecular weights, along with the predicted TSCF and RCF values based on the relationships developed by Burken and Schnoor (7) for hybrid poplar trees. Only the two-ring compound naphthalene is in the range characteristic of moderately hydrophobic chemicals (log Kow between 1.0 and 3.5) that are likely to be bioavailable to rooted, vascular plants (19). This suggests that plant uptake translocation is a possible mechanism for the removal of naphthalene from groundwater by poplar-based phytoremediation systems but is not likely for higher-ringed PAHs. This paper describes the results of long-term monitoring of soil and groundwater remediation at a PAH-contaminated site where hybrid poplar trees were planted. Creosote, derived from the coating of rail ties, was the source of PAH compounds present in soil and groundwater beneath the site. The objective of the study was to assess the effectiveness of the phytoremediation system to remediate PAH compounds in the groundwater and sorbed to the soil and aquifer sediment. To determine the effectiveness of this approach, a seven-year program to monitor PAH concentrations in groundwater and soil was implemented.

Materials and Methods Study Site. The study site was located in north-central Tennessee within the Appalachian Plateaus Physiographic Province (Figure 1). Coal tar creosote was used during railroad tie treatment operations from the early 1950s until 1973, which was stored in an above ground storage tank and an adjacent holding pond. In the early 1990s, creosote contamination was discovered along the banks of an adjacent creek, and a groundwater interceptor trench was constructed 10.1021/es0491681 CCC: $30.25

 2005 American Chemical Society Published on Web 02/11/2005

TABLE 1. Physical and Chemical Properties for Ten PAH Compounds PAH compound naphthalene acenaphthylene acenaphthene fluorene phenanthrene anthracene fluoranthene pyrene chrysene benzo[b]fluoranthene

water molecular solubility log weight (mg/L)a Kowa TSCFb 128 152 154 166 178 178 202 202 228 252

30.0 3.47 3.93 1.98 1.29 0.07 0.26 0.14 0.002 0.0012

3.37 4.33 4.07 4.18 4.46 4.45 5.33 5.32 5.61 6.57

0.56 0.21 0.29 0.25 0.17 0.17 0.03 0.03 0.02 0.00

RCFb 7.17 20.56 14.90 17.03 24.33 24.01 81.43 80.27 122.26 504.76

a Tiehm et al. (1997). b Calculated using relationships developed by Burken and Schnoor (1998).

to prevent further contamination from reaching the creek (Figure 1). In 1997, the responsible party received permission from state regulators to upgrade the interceptor trench and implement a system of hybrid poplar tree that was designed by an engineering firm. Immediately prior to the planting of trees, contaminated soil from the trench excavations was spread on the upper surface of the site, which included discarded coal in the northwest portion of the site. This layer of discarded coal up to 0.6 m thick covered approximately 35% of the phytoremediation system upgradient of the PAH contaminant plume. Below the excavated soil and coal layer, the site was underlain by a shallow surficial aquifer consisting of variably thick layers of fine to medium sand, sandy clay, and clay. The surficial aquifer was bounded below by dense shale present at a depth varying from 3.0 to 3.5 m below land surface. The saturated thickness of the surficial aquifer varied from 1.0 to 2.0 m. Hydraulic conductivity varied from 0.1 to 5.0 m/d, and the fraction of organic carbon in aquifer sediment varied from 0.3 to 1.0%. Groundwater flowed from west to east at an estimated velocity equal to 0.02 m/day. The aquifer was supplied by recharge originating in the area of an industrial facility to the west of the site. Annual precipitation at the site (on-site rain gauge) varied from 104 to 161 cm with only one year (2003) above the average annual precipitation of 152 cm. As shown in Figure 1, a phytoremediation system was planted in 13 rows, consisting of 1026 two- and three-yearold hybrid poplar trees. Eleven months later (April 1998), 120 additional trees were installed between the initial tree system and the creek. During this period, a groundwater monitoring network was installed. Immediately following the first planting of poplar trees, site characterization led to the discovery of creosote-based dense nonaqueous phase liquid (DNAPL) present at the base of the aquifer, varying from 0 to 30 cm in vertical thickness. PAH Sampling and Analysis. Groundwater samples were collected from 26 multilevel groundwater samplers (MLS); 17 of which were installed in 1997, with the remaining 9 installed over the next two years. The MLS design was modified from multilevel groundwater samplers described in LeBlanc et al. (20). Each MLS was constructed with 5-cmdiameter schedule 40 polyvinyl chloride pipe, which serves as the structural center and placed in a 10-cm borehole. The length of each MLS varied, depending on the depth to bedrock. Groundwater samples were obtained from eight sampling ports spaced at 0.305-m vertical intervals starting at 0.082 m above bedrock. Peristaltic pumps operating at land surface were used for the collection of groundwater samples through 0.64-cm color-coded plastic tubing. Sampling ports were constructed by cutting the end of each tube at an angle, covering the ports with a nylon material that

served as a filter, and mounting the port on the outside of the pipe using a cable tie. The annulus around each MLS was filled with filter sand from bedrock to 1 m below land surface followed by a bentonite-concrete seal. Groundwater samples were collected in 40-mL volatile organic analysis (VOA) amber vials and stored at 4 °C until extraction. Fisher optima grade methylene chloride was added to the groundwater using a 1-mL gastight syringe in a 30:1 sample-to-solvent ratio in 40-mL amber VOA vials. Vials were shaken for 90 s to allow contact time between the solvent and sample, after which the methylene chloride was transferred to gas chromatograph (GC) vials for PAH analysis using a 1.0-mL pipet. Groundwater samples were analyzed using a Shimadzu GC 14A GC using a J&W Scientific DB5-MS fused silica capillary column and a flame ionization detector (FID). Helium was used as the carrier gas and the auxiliary gas, and hydrogen was used as the fuel source for the FID. Sample injection volume was 2 µL. External standards were used for quantification of PAHs with a detection limit of 9.5 µg/L. Soil and sediment samples were collected annually along two transects that ran parallel with tree rows 1 and 6. Each transect began outside of the plume and ended in highly contaminated areas. Samples were collected from depths of 1-3 m below land surface using a hand auger. Five grams of air-dried soil were combined with 15 mL of methylene chloride in 40-mL VOA vials. Samples were agitated for 36 h and then transferred to GC vials for PAH analysis. The PAH compounds chosen for quantification were based on the analysis of the 16 priority pollutants in soil and groundwater samples collected in 1997 prior to the planting of the phytoremediation system. Ten PAHs (Table 1) represented the majority of PAH contamination at the site: one two-ring PAH (naphthalene), five three-ring PAHs (acenaphthene, acenaphthylene, anthracene, fluorene, and phenanthrene), three four-ring PAHs (fluoranthene, pyrene, and chrysene), and one five-ring PAH (benzo[b]fluoranthene).

Results and Discussion Phytoremediation System Monitoring. Groundwater was sampled prior to the start of each transpiration period (e.g., early spring) and during the middle of this period (midsummer) subject to the condition of the water table and availability of groundwater. Periods of drought during the study period, combined with the effects of the phreatophytes, typically reduced the saturated thickness to less than 1 m from mid-summer to late autumn, which often limited the number of samples collected and prevented more frequent monitoring of groundwater quality. The height of each poplar tree was monitored five times between system installation (April 1997) and November 2000 using surveying rods and clinometers. The system of hybrid poplar trees grew rapidly in the area of the site coinciding with the PAH contamination but grew poorly in areas where a layer of crushed coal was present below the surface soil. Between the start of the third growing season (March 1999) and later that year (November 1999), the average tree height throughout the system increased from 2.84 to 4.89 m. Root depth was investigated in September 2000 when a temporary trench was constructed parallel to the row of poplar trees near MW6. The root system for the two trees examined extended down to 2.0 m below land surface. Water level data from piezometers instrumented with recording pressure transducers during this same period showed fluctuations induced by phreatophytic comsumpitve use. By use of White’s equation, transpiration rates were calculated as 8 L/d per tree in the summer months (21). PAH Concentrations. Table 2 provides a summary of the 18 groundwater sampling events beginning in November 1997 and ending in July 2004. In general, the number of samples VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1599

FIGURE 1. Map of study site including the phytoremediation system of hybrid poplar trees, groundwater monitoring network, and maximum areal extent of PAH contamination of groundwater and creosote DNAPL.

TABLE 2. Groundwater Sampling Events, Maximum Total PAH and Naphthalene Concentrations, and Occurrence of PAH Compounds in Excess of Solubility max concn (µg/L) exceeding solubility no. MLS no. samples sampled analyzed

date Nov. 1997 Mar. 1998 Jun. 1998 Jan. 1999 Jun. 1999 Jul. 1999 Dec. 1999 Apr. 2000 Jul. 2000 Mar. 2001 Jul. 2001 Mar. 2002 Jul. 2002 Dec. 2002 Apr. 2003 Aug. 2003 Mar. 2004 Jul. 2004

9 11 6 20 18 17 18 19 22 22 9 22 14 19 23 18 24 23

20 38 28 74 49 59 70 78 67 96 17 78 29 108 80 40 119 77

total PAH 14 470 13 470 14 890 18 530 16 200 47 440 11 050 17 690 16 970 57 170 10 200 13 850 16 220 19 410 54 710 39 940 12 730 13 330

naphthano. no. lene compounds samples 13 000 12 470 13 890 17 530 14 200 38 440 10 040 14 870 15 430 54 660 9 580 12 460 14 250 16 060 27 590 12 060 8 924 12 960

0 2i,j 1d 1f 2f,i 7a,e,f,g,h,i,j 1f 4f,h,i,j 4f,h,i,j 5a,f,h,i,j 3f,i,j 6e,f,g,h,i,j 5f,g,h,i,j 6e,f,g,h,i,j 8c,d,e,f,g,h,i ,j 5d,e,f,g,h 5f,g,h,i,j 3e,g,h

0 1 1 4 5 10 1 7 8 7 3 6 9 6 10 8 9 5

a Naphthalene. b Acenaphthylene. c Acenaphthene. d Fluorene. e Phenanthrene. f Anthracene. g Fluoranthene. h Pyrene. i Chrysene. j Benzo[b]fluoranthene.

collected per event increased as the monitoring network developed but also reflect transient hydrologic conditions and groundwater availability. The maximum concentration of total PAH in groundwater samples varied from 10 200 to 57 170 µg/L. The maximum concentrations were always 1600

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

present in the center of the DNAPL zone at ML11 and ML7 at the lower ports. Concentrations greater than the average maximum PAH concentration (>20 000 µg/L) were associated with groundwater samples collected in the immediate proximity to free-phase creosote. The number of samples that exceeded aqueous solubility for at least one PAH compound was also variable (0-10 samples per event). The concentration of nine of the ten PAH compounds in groundwater exceeded aqueous solubility over the study period, with acenaphthylene as the sole exception. This is the result of cosolvency created by the relatively high concentration of naphthalene that persisted throughout the site. In addition, total monoaromatic hydrocarbon compounds were detected at ML7 (at levels as high as 755 µg/L total BTEX using EPA Method 602) that may have contributed to concentration in excess of solubility of the nine PAH compounds, including naphthalene. In an effort to determine the effect of the phytoremediation system on the PAH groundwater concentrations, the total PAH concentrations for each sampling event were segregated between shallow and deep and averaged using an arithmetic mean. This approach provided a means to compare time trends in PAH concentrations in the upper 1 m of the saturated zone, and the potential impact of the tree roots, to time trends in PAH concentrations in the lower 1 m where the DNAPL was present. The deep sample set was derived from three ports located between 0.076- and 0.68-m above bedrock. The shallow sample set consisted of up to five samples derived from ports located between 0.99- and 1.9-m above bedrock. The distribution of depth-averaged total PAH concentrations in groundwater are presented for four of the sampling

FIGURE 2. Depth-averaged total PAH concentrations (µg/L) in the upper region of the saturated zone located at heights between 0.99 and 1.9 m above bedrock collected in January 1999, December 1999, March 2001, and April 2003. events: January 1999, December 1999, March 2001, and April 2003. The January 1999 sampling event occurred after the second growing season and provided the first complete representation of groundwater contamination. The first three sampling events did not provide the same spatial representation. Comparison of total PAH concentrations over the first four sampling events at the same locations showed no significant changes over time, suggesting that after two growing seasons, the poplar trees had no direct impact on the groundwater remediation. Therefore, the January 1999 sampling event is representative of the baseline groundwater contaminant distribution, before which the impact of the phytoremediation system was negligible. The December 1999 sampling event occurred between growing seasons 3 and 4, and the March 2001 and April 2003 sampling events occurred at the beginning of growing seasons 5 and 7, respectively.

Figure 2 depicts the distribution of the total PAH concentrations (depth-average) in the shallow groundwater for the four sampling events. The upper plots (January and December 1999 sampling events) reflect a more widespread plume with higher concentrations when compared to the plumes depicted for the March 2001 and April 2003 sampling events (lower plots). By March 2001, the periphery of the plume in the shallow zone had grown smaller and was accompanied by reduction in the total PAH concentrations throughout of the plume. For example, at ML-12 located on the edge of the plume, total PAH concentrations in the shallow ports decreased from 570 µg/L in March 1998 to 103 µg/L in January 1999, 67 µg/L in December 1999, 4 µg/L in March 2001, and below detection in April 2003. At the plume center (ML-7), total PAH concentrations showed some increase in the shallow ports from March 1998 to January 1999 (12 700 VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1601

FIGURE 3. Depth-averaged total PAH concentrations (µg/L) in the lower region of the saturated zone located at heights between 0.076 and 0.68 m above bedrock collected in January 1999, December 1999, March 2001, and April 2003. and 13 100 µg/L, respectively). Total PAH concentrations then decreased to 5 980 µg/L in December 1999 and to 1 900 µg/L in March 2001 but then rebounded to 4 110 µg/L in April 2003. Figure 3 depicts the PAH plume in the deeper zone over the same period. In comparison to Figure 2, the size of the PAH plume decreased to a lesser extent, and the total PAH concentrations did not decrease to the same levels. Remediation is most evident in the eastern (far left) region of the PAH plume, centered around ML-2, and along the entire southern edge, delineated using the transect of samplers from ML-2 to ML-19. By examination of the time trend at ML-12 again, the average total PAH concentration in the deep ports increased from 1 430 µg/L in March 1998 to 2 200 µg/L in 1602

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

January 1999 but then decreased to 1 140 µg/L in December 1999. PAHs were not detected in the deep ports of ML-12 in March 2001, but a rebound to 450 µg/L was noted in April 2003. No reduction in contaminant levels was evident in the deepest ports of ML-7 (plume center), with total PAH concentrations fluctuating over time from 9 320 to 31 200 µg/L. The fingerlike elongation of the plume to the north shown in the April 2003 plot is a consequence of the DNAPL migrating along the bedrock during the monitoring period. Parts a and b of Figure 4 provide the time series concentration data in groundwater vs depth at the plume edge (ML-12) and plume center (ML-7), respectively, for the period between January 1999 and April 2003. At ML-12, the concentrations of total PAH progressively decreased, reaching

FIGURE 4. Total PAH concentrations in groundwater (a and b) and sediment and soil (c and d) vs height above bedrock at the plume edge (left column) and plume center (right column). nondetectable to 0.90) over time until December 2002. During the last three sampling events, the naphthalene to total PAH ratio decreased to 0.83, 0.81, and 0.70, respectively. Unlike the concentration trend in the upper port, the naphthalene and total PAH concentrations in the lower port remained greater than 10 000 µg/L (with the exception of April 2003). Although the naphthalene-total PAH ratios began to decline at different times (Figure 5), the rates of change over time were similar at each sampling port. For the upper port, the rate of change in the naphthalene-total PAH ratio starting at December 1999 was 20 times greater than the rate of decline during the initial period of decline (2.3 × 10-5 d-1 compared to 4.6 × 10-4 d-1, respectively), and the two rates were statistically different. Similarly, in the lower port, the rate of decrease in the naphthalene-total PAH ratio from March VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1603

FIGURE 5. Total PAH and naphthalene concentrations in groundwater over time in the plume center (ML-7) at sampling ports located 0.991 m above bedrock (upper plot) and 0.686 m above bedrock (lower plot).

FIGURE 6. Naphthalene, acenaphthene, and pyrene concentrations vs time and regression trend lines in groundwater at ML-7 from the sampling port located 0.991 m above bedrock (a) and at ML-12 from the bottom-most sampling port (b). 1998 to July 2002 was 9.7 × 10-6 d-1 but later increased to 3.4 × 10-4 d-1, resulting in statistically different rates. Parts a and b of Figure 6 show the time trend in concentrations at ML-7 (upper port) and ML-12 (deepest port) for naphthalene, acenaphthene, and pyrene (two-, three- and four-ring PAHs, respectively). The plot further demonstrates the selective removal of PAH compounds 1604

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

(naphthalene > acenaphthene > pyrene), showing decreasing rates of removal and concentration changes as the molecular weight of the compounds increases. At ML-7, the first-order rates of mass loss for naphthalene and acenaphthene are 8.1 × 10-4 d-1 and 2.9 × 10-4 d-1, respectively, and were statistically different. Similar and statistically different rates (1.3 × 10-3 d-1 and 2.5 × 10-4 d-1 for naphthalene and

acenaphthene, respectively) were observed at ML-12. In the case of pyrene, no discernible trend was apparent over the six-year sampling period at both ML-7 and ML-12. Implications for Application of Poplar Trees at Other PAH Sites. The results of this study lead to several questions regarding the applicability of phreatophytes to other sites, including: (a) What plant-based remediation mechanisms are responsible for the selective removal of PAH compounds at this site? (b) Is phytoremediation an applicable technology for PAH contamination, particularly with a DNAPL source? With regard to the first question, it appears that PAH reduction, especially naphthalene but also other PAHs, were reduced beginning at the time the roots entered the saturated zone. The marked decrease in the naphthalene concentrations exhibited at ML-7 and ML-12 (and elsewhere) in the upper sampling ports began to occur sometime between the July 1999 and December 1999 sampling events and continued thereafter. Once the concentration of PAHs dropped, they remained at lower levels, with variability being found in the measurements, reflecting the impact of hydrogeologic conditions (e.g., NAPL dissolution, recharge, etc.). The positive effect of the rhizosphere on aerobic and anaerobic biodegradation of PAHs at the site has been demonstrated (17). Microbial enumerations, coupled with microcosm data, from Robinson et al. (17) showed 1-2 orders of magnitude higher bacterial numbers in soil samples collected from the phytoremediation system relative to control locations with much higher levels of actinomycetes (47-78% compared to 0-19%) and increased PAH degradation rates. This indicates that some of the loss of contamination is due to microbial growth in the rhizosphere. Preliminary data suggests that other remediation mechanisms (i.e., plant uptake and phytovolatilization and direct volatilization coupled with vadose zone biodegradation) attributable to the poplar trees at this site are contributing to the mass loss of PAHs. The nearly steady-state conditions in the PAH concentrations during the last 3 years of the study period were thought to be the effect of the creosote DNAPL. Because the DNAPL continues to supply the more soluble PAHs back to the groundwater through dissolution, the dissolution rate appeared to be generally the same as the degradation rate so the concentration in the groundwater remained static, with the upper region being quite low while the lower region remains high. Although the PAH concentrations were not changing greatly after this initial decline, the PAHs in the upper and lower regions of the saturated zone showed “enrichment” in the higher ring PAHs. This was indicative of the depletion of the more soluble PAHs, especially naphthalene but also some of the three-ring PAHs. Although the phytoremediation system was effective in reducing the concentrations of naphthalene and three-ring PAHs, it was doing so at a rate that suggests that the time of remediation cannot be estimated by extrapolation of the groundwater monitoring data. At this site, the phytoremediation system appeared to be minimizing the loss of PAH compounds to the adjacent creek by a combination of enhanced removal and hydraulic control. However, the primary reason phytoremediation can be used at this site is that an interceptor trench was used to collect water before it could move offsite during the months when the trees were

not transpiring water. Without the trench, PAHs would be moving offsite, especially during the winter months. Thus, the use of phytoremediation as the sole remediation measure for sites does not appear to be appropriate where significant DNAPL sources are present and where site remediation objectives cannot be met without effective hydraulic control or other containment measure to ensure that PAHs do not pose a significant risk.

Acknowledgments We thank ARCADIS, Oak Ridge Tennessee office, for technical support on this project. This work was supported in part through the EPA Midwest Hazardous Substance Research Center, Purdue University.

Literature Cited (1) Schnoor, J. L. Phytoremediation of Soil and Groundwater; GWRTAC Report TE-02-01, 2002; p 37. (2) Aitchison, E. W.; Kelley, S. L.; Alvarez, P. J. J.; Schnoor, J. L. Water Environ. Res. 2000, 72 (3), 313-321. (3) Ferro, A.; Chard, J.; Kjelgren, R.; Chard, B.; Turner, D.; Montague, T. Int. J. Phytorem. 2001, 3 (1), 87-104. (4) Landmeyer, J. E. Int. J. Phytorem. 2001, 3 (1), 61-85. (5) Schnoor, J. L.; Licht, L. A.; McCutcheon, S. C.; Wolfe, N. L.; Carreira, L. H. Environ. Sci. Technol. 1995, 29 (7), 318A. (6) Briggs, G. G.; Bromilow, R. H.; Evans, A. A. Pestic. Sci. 1982, 13, 495-504. (7) Burken, J. G.; Schnoor, J. L. Environ. Sci. Technol. 1998, 32, 3379-3385. (8) Burken, J. G.; Schnoor, J. L. Uptake and Metabolism of Atrazine by Poplar Trees. Environ. Sci. Technol. 1997, 31 (5), 1399-1406. (9) Newman, L. A.; Wang, X.; Muiznieks, I. A.; Ekuan, G.; Ruszaj, M.; Cortellucci, R.; Domroes, D.; Karscig, G.; Newman, T.; Crampton, R. S.; Hashmonay, R. A.; Yost, M. G.; Heilman, P. E.; Duffy, J.; Gordon, M. P.; Strand, S. E. Environ. Sci. Technol. 1999, 33, 2257-2265. (10) Hong, M. S.; Farmayan, W. F.; Dortch, I. J.; Chiang, C. Y.; McMillan, S. K.; Schnoor, J. L. Environ. Sci. Technol. 2001, 35, 1231-1239. (11) van Aken, B.; Schnoor, J. L. Environ. Sci. Technol. 2002, 36, 2783-2788. (12) Yoon, J. M.; Oh, B.; Just, C. L.; Schnoor, J. L. Environ. Sci. Technol. 2002, 36, 4649-4655. (13) Aprill, W.; Sims, R. C. Chemosphere 1990, 20, 253-265. (14) Reilley, K. A.; Banks, M. K.; Schwab, A. P. J. Environ. Qual. 1996, 25, 212-219. (15) Banks, M. K.; Lee, E.; Schwab, A. P. J. Environ. Qual. 1999, 28, 294-298. (16) Robinson, S. L.; Novak, J. T.; Widdowson, M. A.; Crosswell, S. B.; Fetterolf, G. J. ASCE J. Environ. Eng. 2003, 129, 232-240. (17) Robinson, S. L.; Novak, J. T.; Widdowson, M. A.; Elliot, M. In Natural Attenuation of Environmental Contaminants; Leeson, A., Foote, E. A., Banks, M. K., Magar, V. S., Eds.; Battelle Press: Columbus, OH, 2001; pp 1-8. (18) Brauner, J. S.; Widdowson, M. A.; Novak, J. T.; Love, N. G. Bioremed. J. 2002, 6, 9-24. (19) Dietz, A. C.; Schnoor, J. L. Environ. Health Perspect. 2001, 109, 163-168. (20) LeBlanc, D. R.; Stephen, S. P.; Garabedian, P.; Hess, K. M.; Gelhar, L. W.; Quadri, R. D.; Stollenwerk, K. G.; Wood, W. W. Water Resour. Res. 1991, 27 (5), 895-910. (21) Panhorst, E. MS Thesis, Virginia Tech, 2000; p 69.

Received for review June 3, 2004. Revised manuscript received November 16, 2004. Accepted December 6, 2004. ES0491681

VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1605