Direct Volatilization of Naphthalene to the Atmosphere at a

Jul 26, 2006 - A vertical flux chamber was used to measure direct surface volatilization of naphthalene over nine months at a creosote-contaminated si...
8 downloads 3 Views 287KB Size
Environ. Sci. Technol. 2006, 40, 5560-5566

Direct Volatilization of Naphthalene to the Atmosphere at a Phytoremediation Site LINSEY C. MARR,* ELIZABETH C. BOOTH, RIKKE G. ANDERSEN, MARK A. WIDDOWSON, AND JOHN T. NOVAK The Charles E. Via, Jr. Department of Civil and Environmental Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061

Phytoremediation systems are known to reduce groundwater contamination by at least three major mechanisms: plant uptake, phytovolatilization, and enhanced rhizosphere bioremediation. The potential for such systems to enhance a fourth remediation pathwaysdirect surface volatilization of contaminants through the subsurface and into the atmosphereshas not yet been investigated in the field. A vertical flux chamber was used to measure direct surface volatilization of naphthalene over nine months at a creosote-contaminated site in Oneida, Tennessee, where a phytoremediation system of poplar trees was installed in 1997. A maximum flux of 23 µg m-2 h-1 was measured in August 2004, and naphthalene removal by the direct volatilization pathway is estimated to be 50 g yr-1 at this site. Results suggest that direct volatilization fluxes are most strongly affected by the groundwater level (thickness of the saturated zone), soil moisture, and changes in atmospheric pressure. At this site, transpiration and canopy interception resulting from the phytoremediation system significantly reduce the saturated thickness, increasing the vertical concentration gradient of naphthalene in the groundwater and thus increasing the upward diffusive flux of naphthalene through the subsurface. The presence of the trees, therefore, promotes direct volatilization into the atmosphere. This research represents the first known measurement of naphthalene attenuation by the direct volatilization pathway.

Introduction Direct volatilization of contaminants from groundwater into the gaseous phase in the vadose zone is a potentially significant remediation pathway, especially for species with high vapor pressures. After volatilizing, pollutants may be biodegraded rapidly under the relatively more aerobic conditions of the vadose zone and/or may be transported upward and released into the atmosphere. In addition to its role as a remediation pathway, direct volatilization may also serve as an important route of human exposure to hazardous chemicals. Known as vapor intrusion, the transport of gases from soil into a building is driven by advection and/or diffusion through cracks in the building, through the foundation itself, or through intentional openings such as drains (1). Direct volatilization and vapor intrusion have the potential to be a source of contaminant exposure for people in buildings situated near hazardous waste sites. * Corresponding author phone: 540-231-6071; e-mail: lmarr@ vt.edu. 5560

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

Previous measurement and modeling studies have shown direct volatilization to be an important mechanism of mass transfer for some groundwater contaminants (2-9). Field, laboratory, and modeling studies have demonstrated that they can migrate through the vadose zone and into the atmosphere, mainly by diffusive mechanisms (10-12). The diffusion of volatile compounds through the vadose zone is determined by physical and chemical properties of the contaminant; soil characteristics such as porosity, moisture content, and organic matter content; biological and chemical degradation; and water flux through the subsurface (8, 12, 13). In some cases, advection driven by changes in atmospheric pressure can also contribute to direct volatilization (2, 6, 10). Previous work has shown that the upward flux of trichloroethylene (TCE) vapors through the subsurface is influenced by atmospheric pressure, temperature, and soil moisture (6, 7, 10). Phytoremediation is a cost-effective and environmentally friendly technology for the treatment of subsurface contamination using vegetation and has been successfully applied at sites containing a variety of compounds (14, 15). Organic chemicals in groundwater systems are subject to two major phytoremediation mechanisms: plant uptake and enhanced rhizosphere bioremediation (16). Recent research has found that phytovolatilization, or the uptake of contaminants followed by their release into the atmosphere, is also a significant pathway for contaminant removal and remediation (15, 17-20). The potential for phytoremediation systems to enhance an additional remediation pathway, that of direct volatilization of contaminants through the vadose zone and into the atmosphere, has not yet been investigated in the field. In laboratory experiments, the presence of alfalfa plants accelerated the flux of methyl tert-butyl ether (MTBE) upward through soil gas and into the atmosphere by a factor of 4 relative to unplanted controls, and Zhang et al. attributed the observation to the plant-induced upward movement of water (21, 22). At phytoremediation sites, the relative importance of direct volatilization to the atmosphere compared to plant uptake, phytovolatilization, and other remediation mechanisms is unknown; and the presence of phreatophytes, plants whose roots extend down to the water table, may enhance direct volatilization. To address this issue, field experiments were conducted at a creosote-contaminated site where hybrid poplar trees were installed to remediate polycyclic aromatic hydrocarbons (PAHs) (23). The study focuses on naphthalene because it comprises the majority of the creosote and, as one of the lighter components, has the greatest potential for volatilization. The specific objectives of this research are to measure the rate of direct volatilization of naphthalene to the atmosphere at the site, to identify the most important meteorological and subsurface parameters that affect the magnitude of direct volatilization, and to assess whether the presence of the phytoremediation system enhances direct volatilization of naphthalene through the subsurface.

Materials and Methods Site Description. The study area is located in Oneida, Tennessee, approximately 60 miles northwest of Knoxville. Beginning in the early 1950s, railroad ties were treated with creosote at the site, and treatment continued intermittently until 1973. Sources of groundwater and soil contamination included an above-ground storage tank and an unlined holding pond that contained the discharged creosote. To control the movement of contaminated groundwater and to aid in subsurface remediation, a phytoremediation system 10.1021/es060087+ CCC: $33.50

 2006 American Chemical Society Published on Web 07/26/2006

FIGURE 1. Map of study site showing sampling locations and groundwater concentrations of naphthalene (µg L-1) at 0-0.7 m above bedrock in March 2004. consisting of 1120 poplar trees was installed in 1997 and 1998. The system, shown in Figure 1, also includes an interceptor trench and oil-water separator to prevent further contamination of the nearby creek. Groundwater samples show that most of the contamination is present as a dense nonaqueous phase liquid (DNAPL) located 2.4-3.7 m below ground surface. The soil consists of three layers: a mixture of soil, gravel, and coal comprises the top layer (∼1 m thick); silty sandy clay comprises the middle layer (1.5-2 m thick); and silty sand comprises the bottom layer (1.2-1.5 m thick) that lies on top of shale bedrock. A seven-year monitoring program at the site has shown that PAH concentrations in the groundwater began to decline after the poplar trees’ third growing season and that remediation was limited to the two- and three-ring PAHs, such as naphthalene and acenaphthene (23). Groundwater naphthalene concentrations at 1 m above bedrock have remained relatively constant over the last three years (20012004) because the dissolution rate from the DNAPL is at equilibrium with the loss rate. Flux Chamber. A dynamic chamber is used to measure the flux of naphthalene from the unsaturated zone into the atmosphere (2, 3, 6, 24-26). The body of the flux chamber consists of a stainless steel tray with an open bottom and dimensions of 24 × 29 × 4.4 cm. When installed 1.9 cm into the ground, the chamber has an effective interior height of 2.5 cm. The dimensions are similar to those of a chamber that has been validated extensively in laboratory experiments and found to recover an average of 95% of advective and diffusive fluxes (24). Air is circulated through the system at a flow rate of 2 L min-1 by a personal sampling pump (SKC, Inc. PCXR8). From the pump, air passes through a drying column containing 570 g of anhydrous calcium sulfate (Drierite), enters the chamber through a Teflon manifold containing 10 equally spaced 0.3-cm holes, passes through the chamber headspace and over the soil, exits via two 0.6cm diameter outlets, flows through two XAD-2 sorbent tubes (SKC, Inc.) in series that trap the naphthalene, and finally circulates back to the pump.

The flow rate of 2 L min-1 is chosen to maintain low concentrations inside the chamber while obtaining high sorbent tube efficiency. In laboratory tests, Tillman and Smith (24) found that their flux chamber with a flow rate of 0.125 L min-1 impeded diffusive flux by only 1.4% during step changes in loading and affected the subsurface gas-phase concentration gradient, and thus the flux, by 4%. The use of a higher flow rate should further minimize the impact of the flux chamber on diffusive flux due to nonzero headspace concentrations. To assess uncertainty in the method, we collected replicate samples with four chambers side-by-side in both laboratory and field settings and found a standard deviation of 7% in the measured flux. Flux Chamber Sampling. As shown in Figure 1, flux chambers were placed in areas of high (ML7 and ML11), medium (ML22), and low (ML16) groundwater naphthalene concentrations, and at a fifth location with concentrations below the detection limit (BDL) of 9.5 µg L-1. Before connecting the sorbent tubes, we ran the pump for ∼5 min to allow the system to reach steady-state. To minimize the effects of radiant heating and rain while maintaining ambient conditions around the chamber, we propped a plastic tub over the entire system. We calibrated flow rates in the field before and after each experiment (Bios Dry Cal DC Lite). Sampling durations ranged from hours to several days, with the majority of samples collected over a period of 3 days in order to collect sufficient mass of naphthalene for analysis. After sampling, the sorbent tubes were removed, capped, wrapped with aluminum foil, and stored at 4 °C for no longer than 24 h before analysis. Following the National Institute of Occupational Safety and Health’s (NIOSH) analytical method 5515 for PAHs, we extracted samples in toluene and analyzed them using gas chromatography with flame ionization detection (Hewlett-Packard 5890) with a DB-5 capillary column (Agilent J&W). The injector, detector, and oven temperatures were set to 305 °C, 310 °C, and 80 °C, respectively, with a temperature ramp of 10 °C min-1. Twentyfive percent of the samples consisted of field and laboratory blanks, including background ambient air at the site. The VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5561

FIGURE 2. Naphthalene surface fluxes measured in areas of high, medium, low, and below detectable (BDL) groundwater contamination. Samples were collected over a period of 1-4 days each month. Uncertainty in the measurement is estimated to be 7% based on replicate samples. flux F was calculated using eq 1:

F)

CeVe At

(1)

where Ce is the concentration of naphthalene in the extract (µg L-1), Ve is the total volume of the extract (L), A ) 0.070 m2 is the surface area covered by the chamber, and t is the sampling duration. Groundwater, soil, and meteorological variables were recorded during each sampling trip. Groundwater samples were collected from multi-level sampling wells (MLSs) at the locations shown in Figure 1. Soil moisture was measured using probes (Watermark Soil Moisture Sensor) that were installed in the locations indicated in Figure 1 at various depths below ground surface. The probes report soil water potential in units of pressure, and these results were converted to volumetric soil moisture percentages using the Van Genuchten relationship (27):

θw - θwr ) (1 + (Rψ)n)-m η - θwr

(2)

where θw is the volumetric water content (volume of water per total volume of solids and pores); θwr ) 0.100 is the residual water content for sandy clay; η is the porosity, measured at 0.37; R ) 0.027 cm-1 and n ) 1.23 are parameters for sandy clay; m ) 1-1/n; and ψ is the capillary pressure head, or the measured soil water potential in units of pressure divided by the density of water and gravitational acceleration. A weather station at the site recorded temperature, humidity, atmospheric pressure (to the nearest 0.1 in. of mercury, equivalent to 340 Pa), wind, and precipitation at 30-min intervals.

Results Figure 2 shows the results of 35 surface flux measurements collected between March and October 2004, once per month. Samples were collected at ML16 only during the last five months and at the BDL site on only three occasions: July, August, and October. The highest measured flux was 23 µg m-2 h-1 at ML7 in August. While flux values were roughly the same at all locations during the spring (March-June), relatively larger fluxes occurred in areas of high groundwater contamination (ML7 and ML11) during the summer and fall (August-November). ML22 was located in an area considered to have a medium level of contamination and showed intermediate fluxes of naphthalene to the atmosphere. Fluxes at the BDL site did not exceed 2 µg m-2 h-1. Although naphthalene was not present in the groundwater at this location, it may have been present in the contaminated soil that was excavated during the initial remediation and then 5562

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

spread over the top ∼0.3 m of the entire site. Fluxes at all locations were lowest during the July sampling event, which coincided with considerable rainfall. Over time, the plume shifted in space (23), so although ML11 corresponded to the area of highest groundwater concentration in Figure 1 (March 2004), this relationship may not always hold. A table containing detailed information about sampling dates and conditions, exact flux values, and corresponding meteorological and hydrogeological parameters is available in the Supporting Information. For the sampling dates when meteorological data are available (May-September), we compare temperature, relative humidity, and atmospheric pressure to fluxes at the three most contaminated locations: ML7, ML11, and ML22. Infiltrating rainfall may have affected fluxes, with 2.4 cm of rain falling in the 24 h before sampling on 6 July and an additional 0.5 cm falling during the sampling period. In August, 1.3 cm of rain fell during the last 5 h of the three-day sampling period for ML16 and ML22. Flux measurements at the other sites in August ended a day earlier and were not affected by rain. Because rain may have suppressed fluxes in July (fluxes shown in Figure 2 are much lower in July), these data points are excluded from the meteorological analysis. Figure 3 shows that fluxes did not appear to be related to temperature or humidity, except perhaps that the highest fluxes, which occurred in August, corresponded with the coolest temperatures. The figure does not include data from March or April, when temperatures and humidities were likely to have been lower than the range shown. The largest fluxes all occurred when the change in atmospheric pressure was negative, i.e., it decreased, during the sampling period. When the change in pressure was positive, fluxes were always below 5 µg m-2 h-1. The linear correlation coefficients r of flux with temperature, humidity, and change in pressure were -0.67, 0.18, and -0.49, respectively. Figure 4 shows the relationship between fluxes and level of the water table between March and October. Bedrock is at 432.8 m MSL (meters above mean sea level) at ML7, 432.3 m at ML11, and 433.7 m at ML22. As groundwater levels vary seasonally, data labels indicating the month of the measurement are also shown. The water table was at its highest in winter, i.e., March; decreased by more than 1 m throughout the spring and summer; and began to rise again in October. Coinciding with some of the lowest water table elevations, the highest fluxes occurred in August and September. When the water table was higher than 434.6 m, fluxes were always less than 4 µg m-2 h-1. Figure 5 shows the relationship between fluxes and soil moisture between April and October in areas of high and medium groundwater contamination. Soil moisture was measured at three depths; results from the medium depth (1.2-1.5 m below ground surface) are presented. Volumetric soil moisture at the medium depth varied between 24 and 37% and increased with depth. Soil moisture measurements were not made in March or July and were not available at ML11. Because soil moisture showed no seasonal dependence, the data points are not labeled by month. The linear correlation coefficient between flux and soil moisture for the data points shown in Figure 5 is -0.54. Higher fluxes occurred when soil moisture was less than 28%.

Discussion At this site, fluxes of naphthalene from the groundwater to the atmosphere exhibit a strong seasonal pattern, with higher magnitudes during the summer and early fall. This seasonality is similar to that found for TCE volatilization to the atmosphere at Picatinny Arsenal in New Jersey (6), although there, the fluxes are much higher, up to 1200 µg m-2 h-1. Larger gas-phase fluxes of TCE than of naphthalene would be expected because of TCE’s higher vapor pressure and

FIGURE 3. Naphthalene fluxes at ML7, ML11, and ML22 versus average ambient temperature, relative humidity, and change in atmospheric pressure during the sampling periods in May-September.

FIGURE 4. Naphthalene surface fluxes versus water table elevation (m above mean sea level) in areas of high and medium groundwater contamination. Data labels indicate the month of the measurement.

FIGURE 5. Naphthalene surface fluxes versus volumetric soil moisture measured 1.2-1.5 m below ground surface in areas of high and medium groundwater contamination. greater recalcitrance, all other parameters being equal. Like the measurements shown here, the Picatinny Arsenal’s highest fluxes of those not affected by rainfall occur in August. Meteorological Factors. Precipitation seems to have a significant short-term influence on the flux of contaminants through the subsurface. The lowest fluxes at the sites with high and medium contamination (ML7, ML11, and ML22) occurred in July, when at least 1 cm of rain fell in the hours immediately preceding the sampling period. Poulsen et al. (28) and Valsaraj et al. (29) suggest that as precipitation, soil moisture, and/or relative humidity increase, water molecules can displace hydrophobic organic compounds that are sorbed to soil or sediment, causing them to enter the void spaces. As a result, fluxes of contaminants will temporarily increase. Infiltrating rainwater can also displace gases in the subsurface voids; and precipitation is often accompanied by a drop in barometric pressure, which will result in large advective flows from the subsurface into the atmosphere. Furthermore, the

infiltration of clean rainwater results in a net transfer of naphthalene from the vapor phase to the aqueous phase (6). The combination of these mechanisms may flush gas-phase naphthalene from the unsaturated zone during rain events and thereby suppress fluxes after the rainfall. For some of the same reasons, rainfall during the last few hours of sampling may cause an increase in fluxes. In August, 1.3 cm of rain fell during the last 5 h of three-day samples at ML22 and ML16, and may have contributed to the large fluxes at these sites in areas of medium and low groundwater contamination. Other meteorological factors may also influence direct volatilization. Due to the temperature dependence of gasliquid partitioning, i.e., Henry’s constant, warmer temperatures should be associated with higher gas-phase concentrations and thus higher fluxes. However, Figure 3 shows that the highest fluxes are observed together with the lowest average temperatures, suggesting that temperature is not a controlling factor at this site. Although not examined in this work, subsurface temperatures may have a greater influence on volatilization rates and may be enhanced by heating from biological activity. Figure 3 also shows that fluxes do not appear to be related to ambient humidity. Variations in atmospheric pressure have been found to enhance the transport of vapors in the subsurface (30-33), a mechanism commonly referred to as barometric pumping. Choi et al. (10) report that advective fluxes of TCE from the subsurface, induced by changes in atmospheric pressure, occasionally exceed diffusive fluxes. While measured naphthalene fluxes do not show an obvious relationship to changes in pressure (Figure 3), all of the largest fluxes correspond to periods when atmospheric pressure drops. Therefore, advective transport from the subsurface to the atmosphere may be occurring at this site. The absence of pressure measurements and gas-phase naphthalene concentrations preclude a modeling study of naphthalene transport through the unsaturated zone. Hydrogeological Parameters. Figure 5 shows that naphthalene fluxes to the atmosphere are inversely related to soil moisture. This relationship can be explained by Fick’s first law

Jdiff ) -Dea

∂Ca ∂z

(3)

where Jdiff is the diffusive flux [M L-2 T-1], Dea is the effective diffusion coefficient of naphthalene in air, and Ca is the concentration of naphthalene in air. The effective diffusion coefficient can be estimated using Millington and Quick’s empirical model (34, 35),

Dea ) Da

θa2/3

(4)

η2

VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5563

where Da is the gas-phase diffusion coefficient and θa is the volumetric air content of soil. As soil moisture increases and air content decreases, the effective diffusion coefficient will decrease and flux will, therefore, decrease. The relatively small changes in soil moisture, however, cannot alone account for the 2-5-fold increase in fluxes in August and September. Diffusive processes have been found to dominate vadosezone transport of volatile compounds in field, laboratory, and modeling studies (10-12). The diffusion of volatile compounds through the unsaturated zone is determined by physical and chemical properties of the contaminant; soil characteristics such as porosity, moisture content, and organic matter content; biological and chemical degradation; adsorption to soil; and water flux through the subsurface (8, 12, 13). In the groundwater, because the hydrodynamic dispersion coefficient is 2 orders of magnitude smaller than the diffusion coefficient in the situation examined here (36), diffusion will be the dominant vertical transport process in the saturated zone, too. While meteorological factors can influence direct volatilization to the atmosphere, we hypothesize that at this site, the groundwater level (i.e., the saturated thickness) is the most important variable affecting seasonal variations in vertical naphthalene fluxes. Figure 4 clearly shows that the highest fluxes are observed only when the water table is below 434.6 m. Because diffusion coefficients in the aqueous phase are so much smaller than in the gaseous phase, transport of naphthalene upward from the DNAPL is controlled by diffusion in the groundwater. According to Fick’s law (eq 3 applied to water rather than air), the diffusive flux is proportional to the vertical concentration gradient. At the bottom of the saturated zone, directly above the DNAPL source, the concentration of naphthalene will be constant at its solubility under equilibrium conditions. At the top of the saturated zone, the concentration of naphthalene will be near zero, as it diffuses much more rapidly up through the unsaturated zone and into the atmosphere. Because the boundary conditions at the top and bottom of the saturated zone remain relatively unchanged while the saturated thickness at this site can vary by a factor of 2, the saturated thickness or groundwater level will determine the vertical flux of naphthalene through the subsurface. During the sampling campaign, the shallowest water table, 90 cm above the top of the DNAPL, was observed in September, and the deepest water table, 190 cm above the DNAPL, was found in March. This difference would cause an approximate doubling of the vertical concentration gradient in the groundwater between March and September, which would correspond to a doubling in flux. For comparison, the measured fluxes at the land surface at ML7 are 3.0 times higher in September than in March. As described below, the presence of the phytoremediation system is largely responsible for the seasonal fluctuations in the groundwater level. Incorporating biodegradation into the transport process is likely only to amplify the differences between fluxes under conditions of a low versus high water table. Due to the greater availability of oxygen in the unsaturated zone, the biodegradation rate is more rapid here than in the saturated zone. The linear shape of vertical concentration profiles of naphthalene in the saturated zone (23) supports the hypothesis that biodegradation is far less significant than diffusive transport in the saturated zone. On a mass basis though, biodegradation is likely to be greater in the saturated zone. Additional factors may also contribute to the influence of the saturated thickness on volatilization of contaminants to the atmosphere. Natural fluctuations in the water table elevation cause contaminants to redistribute vertically over an area termed the “smear zone,” where high transient concentrations may enhance diffusion upward through the unsaturated zone (4, 37). On a seasonal basis, during the late 5564

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

summer and early fall, residual contaminants may remain within the unsaturated zone as the water table retreats. Exposure of residual contamination to air-filled void spaces under dry weather conditions creates ideal conditions for large contaminant fluxes. Phytoremediation Enhances Direct Volatilization. Phytoremediation systems are known to reduce groundwater contamination by plant uptake, phytovolatilization, and rhizosphere bioremediation. They also act as a solar driven pump-and-treat system, increasing the flux of water through the vadose zone during high transpiration periods and lowering the groundwater table (19, 38, 39). The results presented here indicate that phreatophyte-based phytoremediation systems may also enhance direct volatilization of contaminants to the atmosphere. Poplar trees at this site draw down the water table during periods of high transpiration and control the recharge rate through interception of rainfall by the leaf canopy. These actions create larger vertical concentration gradients in the groundwater, and therefore, larger naphthalene fluxes, due to a decrease in the saturated thickness and the presence of a creosote DNAPL at a relatively shallow depth at the site. Additionally, the up-and-down movement of the groundwater table generated by seasonal and diurnal cycles in the trees’ transpiration rate exposes residual contaminants in the capillary fringe and may also enhance fluxes. Historical measurements indicate that the phytoremediation system directly affects the water table elevation at this site. We have been monitoring water table levels at the site since 1998 through 2005 and have compared levels inside and outside the phytoremediation system. We have observed clear differences in groundwater levels in response to rainfall/ recharge events; the differences are attributable to canopy interception and transpiration during the summer by the poplar trees. A detailed hydrological analysis of the site will be presented in a future publication. Phyto-enhanced direct volatilization will be most evident at sites where the depth to the DNAPL or other source of contamination is relatively shallow. At sites where a phreatophyte-based phytoremediation system is horizontally or vertically distant from the highest level of contamination, the decrease in the saturated thickness resulting from transpiration and/or canopy interception may not increase volatilization of VOCs. Such would be the case at sites where (a) the vertical thickness of the vadose zone is much greater than the saturated thickness, (b) the saturated thickness is relatively large and a DNAPL source is not located near the land surface, or (c) the phytoremediation system is positioned downgradient of the source and aqueous concentrations are well below solubility. Plants can also affect transport of volatile compounds through the subsurface in other direct and indirect ways (19). For example, plants can take up contaminants in their roots, translocate them, and serve as a source from which contaminants can diffuse out of the roots and into the soil gas. By displacing soil, root growth can create “preferential pathways” through which diffusive transport will be more rapid. Through evapotranspiration, plants can dry out the soil and cause an increase in the effective diffusion coefficient. While the present study investigates the bulk effect of plants on direct volatilization, a more controlled laboratory experiment and/or extensive modeling study is needed in order to elucidate the individual mechanisms by which plants enhance soil gas fluxes. Naphthalene Remediation. The area associated with groundwater naphthalene concentrations greater than 10 µg L-1 (Figure 1) is approximately 1000 m2. Within this area, the average naphthalene flux to the atmosphere is 15 µg m-2 h-1 during the summer and 3 µg m-2 h-1 during the rest of the year. Therefore, the direct volatilization pathway to the

atmosphere accounts for the removal of 50 g yr-1 of naphthalene from the groundwater. Andersen (40) has estimated removal of naphthalene at the site by various remediation mechanisms, including direct volatilization from the soil to the atmosphere, phytovolatilization, degradation in the saturated zone and degradation in the vadose zone (41). According to these estimates, direct volatilization and phytovolatilization combined account for 0.003-0.02% of the losses of naphthalene from the site. Estimates with and without trees show that phytoremediation increases the amount of volatilization by a factor of ∼4. The majority of naphthalene loss at this site is due to saturated zone biodegradation. Forthcoming work will compare in greater detail the relative contribution of direct volatilization versus biodegradation and phytovolatilization for remediation of naphthalene at this site. Measurement of fluxes of volatile groundwater contaminants to the atmosphere by the method presented here is a valuable tool for assessing remediation progress and the importance of different removal mechanisms at contaminated sites. Volatilization to the atmosphere may be especially important at sites where the saturated thickness is small, and vertical diffusive flux through the saturated zone will, therefore, be higher.

Acknowledgments This research was supported by the Via Endowment at the Virginia Tech Department of Civil and Environmental Engineering and by the Midwest Hazardous Substance Research Center at Purdue University. We thank Julie Petruska and Jody Smiley for their technical expertise and the landowner for allowing us to conduct research at the field site.

Supporting Information Available A table listing all measured naphthalene fluxes to the atmosphere and corresponding meteorological (temperature, humidity, pressure, change in pressure) and subsurface (water table elevation, soil moisture) conditions during each sampling period in 2004. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Olson, D. A.; Corsi, R. L. Fate and transport of contaminants in indoor air. Soil Sediment Contam. 2002, 11, 583-601. (2) Tillman, F. D.; Choi, J.-W.; Smith, J. A. A comparison of direct measurement and model simulation of total flux of volatile organic compounds from the subsurface to the atmosphere under natural field conditions. Water Resour. Res. 2003, 39, SBH 5.1-5.11 (doi: 10.1029/2003WR002098). (3) Jellali, S.; Benremita, H.; Muntzer, P.; Razakarisoa, O.; Schafer, G. A large-scale experiment on mass transfer of trichloroethylene from the unsaturated zone of a sandy aquifer to its interfaces. J. Contam. Hydrol. 2003, 60, 31-53. (4) Lahvis, M. A.; Baehr, A. L.; Baker, R. J. Quantification of aerobicbiodegradation and volatilization rates of gasoline hydrocarbons near the water table during natural attenuation conditions. Water Resour. Res. 1999, 35, 753-765. (5) Lahvis, M. A.; Baehr, A. L.; Baker, R. J. Evaluation of volatilization as a natural attenuation pathway for MTBE. Ground Water 2004, 42, 258-267. (6) Smith, J. A.; Tisdale, A. K.; Cho, H. J. Quantification of natural vapor fluxes of trichloroethene in the unsaturated zone at Picatinny Arsenal, New Jersey. Environ. Sci. Technol. 1996, 30, 2243-2250. (7) Choi, J.-W.; Smith, J. A. Geoenvironmental factors affecting organic vapor advection and diffusion fluxes from the unsaturated zone to the atmosphere under natural conditions. Environ. Eng. Sci. 2005, 22, 95-108. (8) Stauffer, P. H.; Birdsell, K. H.; Witkowski, M. S.; Hopkins, J. K. Vadose zone transport of 1,1,1-trichloroethane: Conceptual model validation through numerical simulation. Vadose Zone J. 2005, 4, 760-773.

(9) Lahvis, M. A.; Baehr, A. L. Estimation of rates of aerobic hydrocarbon degradation by simulation of gas transport in the unsaturated zone. Water Resour. Res. 1997, 32, 2231-2249. (10) Choi, J.-W.; Tillman, F. D.; Smith, J. A. Relative importance of gas-phase diffusive and advective trichloroethene (TCE) fluxes in the unsaturated zone under natural conditions. Environ. Sci. Technol. 2002, 36, 3157-3164. (11) Jellali, S.; Muntzer, P.; Razakarisoa, O.; Schafer, G. Large scale experiment on transport of trichloroethylene in a controlled aquifer. Transp. Porous Media 2001, 44, 145-163. (12) Jury, W. A.; Russo, D.; Streile, G.; Abd, H. Evaluation of volatilization by organic chemicals residing below the soil surface. Water Resour. Res. 1990, 26, 13-20. (13) Batterman, S.; Kulshrestha, A.; Cheng, H. Y. Hydrocarbon Vapor Transport in Low Moisture Soils. Environ. Sci. Technol. 1995, 29, 171-180. (14) Davis, L. C.; Banks, M. K.; Schwab, A. P.; Narayanan, M.; Erickson, L. E.; Tracy, J. C. In Biodegradation Technology Developments: Principles and Practice; Sikdar, S. K., Irvine, R. L., Eds.; Technomic: Lancaster, PA, 1998; Vol. 2, pp 182-220. (15) Schnoor, J. L. Phytoremediation of Soil and Groundwater; Ground-Water Remediation Technologies Assessment Center: Pittsburgh, PA, 2002. (16) Anderson, T. A.; Guthrie, E. A.; Walton, B. T. Bioremediation in the rhizosphere. Environ. Sci. Technol. 1993, 27, 2630-2636. (17) Newman, L. A.; Strand, S. E.; Choe, N.; Duffy, J.; Ekuan, G.; Ruszaj, M.; Shurtleff, B. B.; Wilmoth, J.; Heilman, P.; Gordon, M. P. Uptake and biotransformation of trichloroethylene in hybrid poplars. Environ. Sci. Technol. 1997, 31, 1062-1067. (18) Burken, J. G.; Schnoor, J. L. Predictive relationships for uptake of organic contaminants by hybrid poplar trees. Environ. Sci. Technol. 1998, 32, 3379-3385. (19) Narayanan, M.; Erickson, L. E.; Davis, L. C. Simple plant-based design strategies for volatile organic pollutants, Environ. Prog. 1999, 18, 231-242. (20) Ma, X. M.; Burken, J. G. TCE diffusion to the atmosphere in phytoremediation applications. Environ. Sci. Technol. 2003, 37, 2534-2539. (21) Zhang, Q.; Davis, L. C.; Erickson, L. E. An experimental study of phytoremediation of methyl-tert-butyl ether (MTBE) in groundwater. J. Hazard. Subst. Res. 2000, 2, 4.2.1-4.2.19. (22) Zhang, Q.; Davis, L. C.; Erickson, L. E. Plant uptake of methyl tert-butyl ether (MTBE) from ground water. Pract. Periodical Hazard., Toxic, Radioact. Waste Manage. 2001, 5, 136-140. (23) Widdowson, M. A.; Shearer, S.; Andersen, R. G.; Novak, J. N. Remediation of polycyclic aromatic hydrocarbon compounds in groundwater using poplar trees. Environ. Sci. Technol. 2005, 39, 1598-1605. (24) Tillman, F. D.; Smith, J. A. Design and laboratory testing of a chamber device to measure total flux of volatile organic compounds from the unsaturated zone under natural conditions. J. Contam. Hydrol. 2004, 75, 71-90. (25) Reichman, R.; Rolston, D. E. Design and performance of a gas flux chamber. J. Environ. Qual. 2002, 31, 1774-1781. (26) Gao, F.; Yates, S. R.; Yates, M. V.; Gan, J. Y.; Ernst, F. F. Design, fabrication, and application of a dynamic chamber for measuring gas emissions from soil. Environ. Sci. Technol. 1997, 31, 148-153. (27) Carsel, F. F.; Parrish, R. S. Developing joint probability distributions of soil water retention characteristics. Water Resour. Res. 1988, 24, 755-769. (28) Poulsen, T. G.; Yamaguchi, T.; Moldrup, P.; de Jonge, L. W.; Rolston, D. E. Predicting volatile organic vapor sorption from soil specific surface area and texture. J. Environ. Qual. 2000, 29, 1642-1649. (29) Valsaraj, K. T.; Ravikrishna, R.; Choy, B.; Reible, D. D.; Thibodeaux, L. J.; Price, C. B.; Yost, S.; Brannon, J. M.; Myers, T. E. Air emissions from exposed contaminated sediments and dredged material. Environ. Sci. Technol. 1999, 33, 142-149. (30) Massmann, J.; Farrier, D. F. Effects of atmospheric pressures on gas transport in the vadose zone. Water Resour. Res. 1992, 28, 777-791. (31) Elberling, B.; Larsen, F.; Christensen, S.; Postma, D. Gas transport in a confined unsaturated zone during atmospheric pressure cycles. Water Resour. Res. 1998, 34, 2855-2862. (32) Auer, L. H.; Rosenberg, N. D.; Birdsell, K. H.; Whitney, E. M. The effects of barometric pumping on contaminant transport, J. Contam. Hydrol. 1996, 24, 145-166. VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5565

(33) Takle, E. S.; Massmann, W. J.; Brandle, J. R.; Schmidt, R. A.; Zhou, X. H.; Litvina, I. V.; Garcia, R.; Doyle, R.; Rice, C. W. Influence of high-frequency ambient pressure pumping on carbon dioxide efflux from soil. Agric. Forest Meteorol. 2004, 124, 193-206. (34) Millington, R. J. Gas diffusion in porous media. Science 1959, 130, 100-102. (35) Millington, R. J.; Quirk, J. P. Permeability of porous solids. Trans. Faraday Soc. 1961, 57, 1200-1207. (36) Gelhar, L. W.; Welty, C.; Rehfeldt, K. R. A critical review of data on field-scale dispersion in aquifers. Water Resour. Res. 1992, 28, 1955-1974. (37) Lee, J. Y.; Cheon, J. Y.; Lee, K. K.; Lee, S. Y.; Lee, M. H. Factors affecting the distribution of hydrocarbon contaminants and hydrogeochemical parameters in a shallow sand aquifer. J. Contam. Hydrol. 2001, 50, 139-158. (38) Al-Yousfi, A. B.; Chapin, R. J.; King, T. A.; Shah, S. I. Phytoremediationsthe natural pump-and-treat and hydraulic barrier system. Pract. Periodical Hazard., Toxic, Radioact. Waste Manage. 2000, 4, 73-77.

5566

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

(39) Robinson, B.; Green, S.; Mills, T.; Clothier, B.; van der Velde, M.; Laplane, R.; Fung, L.; Deurer, M.; Hurst, S.; Thayalakumaran, T.; van den Dijssel, C. Phytoremediation: using plants as biopumps to improve degraded environments. Aust. J. Soil Res. 2003, 41, 599-611. (40) Andersen, R. G. In situ characterization and quantification of phytoremediation removal mechanisms for naphthalene at a creosote-contaminated site; Department of Civil and Environmental Engineering, Virginia Tech: Blacksburg, VA, 2006. (41) Pitterle, M. T.; Andersen, R. G.; Novak, J. N.; Widdowson, M. A. Push-pull tests to quantify in situ degradation rates at a phytoremediation site. Environ. Sci. Technol. 2005, 39, 93179323.

Received for review January 16, 2006. Revised manuscript received May 23, 2006. Accepted June 23, 2006. ES060087+