Characterization of NAPL Source Zone Architecture and Dissolution

Characterization of the impacts of porous media heterogeneity on the NAPL source zone architecture and dissolution kinetics in 3-D at the continuum-sc...
0 downloads 0 Views 504KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 3672-3678

Characterization of NAPL Source Zone Architecture and Dissolution Kinetics in Heterogeneous Porous Media Using Magnetic Resonance Imaging C H A N G Y O N G Z H A N G , * ,† CHARLES J. WERTH,† AND ANDREW G. WEBB‡ Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, 205 North Mathew Avenue, MC-250, Urbana, Illinois 61801, and Department of Bioengineering, Pennsylvania State University, 313 Hallowell Building, University Park, Pennsylvania 16802

A direct visualization method using magnetic resonance imaging (MRI) was developed to characterize sand grain size distribution, nonaqueous phase liquid (NAPL) source zone architecture, and aqueous flowpaths in a threedimensional (3-D) flowcell (26.5 cm × 10.5 cm × 10.5 cm) packed with a heterogeneous distribution of five different sand fractions. All images were acquired at a resolution of 0.1875 cm × 0.1875 cm × 0.225 cm. A 1H image of pore water resolved the heterogeneous permeability field; grain size differences as small as 0.1 mm could be distinguished. A time series of 1H images of water doped with the paramagnetic tracer MnCl2 were acquired and used to obtain voxel-scale breakthrough curves. Water preferentially flowed through coarse sands before NAPL release. After NAPL release, the flow bypassed NAPL zones, and bypassing was more evident for high NAPL saturation zones. A time series of 19F images of NAPL were acquired and used to determine voxel-scale NAPL saturation (Sn) during dissolution. Results show that 93% of NAPL mass was in the coarsest sand, most NAPL was trapped as pools and not as residual ganglia, NAPL saturation increased with depth, and the NAPL dissolution front moved vertically from the top to the bottom of the flowcell during the first 170 pore volumes of water flushed. NAPL component effluent concentrations initially increased due to the development of flow in zones with decreasing NAPL saturation. Flowpath images suggest that this occurs as NAPL transitions from pools (Sn > 0.15) to residual ganglia. The results highlight the importance of flow bypassing and provide the opportunity to develop more accurate NAPL dissolution models.

Introduction Nonaqueous phase liquids (NAPLs), such as gasoline and chlorinated solvents, have frequently been released into the subsurface through spills, improper disposal, and leaking * Corresponding author phone: (217)333-6967; fax: (217)333-6968; e-mail: [email protected]. † University of Illinois at Urbana-Champaign. ‡ Pennsylvania State University. 3672

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007

storage tanks. Once in the subsurface, NAPLs migrate downward due to gravity and become trapped in the pores of soils and groundwater sediments by capillary forces. Because of their low solubility (Cs), NAPLs can dissolve slowly into groundwater and cause contamination for decades. NAPL dissolution kinetics in groundwater is significantly affected by NAPL source zone architecture, which depends to a large extent on porous media heterogeneity. Previous researchers (1-6) have found that coarse sand lenses form preferential channels for NAPL flow, whereas fine lenses act as capillary barriers to NAPL migration, causing NAPLs to pool above them. Even very subtle variations in texture (i.e., grain size) can give rise to lateral flow. Hence, porous media heterogeneity can lead to a complex NAPL source zone architecture. Over the past decade and a half, extensive research has focused on characterization of NAPL source zone architecture and dissolution kinetics in one-dimensional (1-D) columns and two-dimensional (2-D) flow systems. In early work, in situ polymerization of styrene in flowthrough systems was used to investigate the influence of grain size and uniformity on NAPL morphology at residual (defined as stable, discontinuous distribution) saturation (Sn) (7, 8) and the influence of NAPL morphology on dissolution kinetics (9). NAPL ganglia size and uniformity were related to grain size and uniformity, and the importance of these parameters on NAPL dissolution was demonstrated. However, this method is very timeconsuming and limited to the characterization of NAPL morphology at a single point in time during dissolution. In more recent work, a variety of advanced in situ imaging techniques was used to characterize NAPL distribution and morphology during dissolution. At the pore-scale, highresolution synchrotron X-ray microtomography (10) and magnetic resonance imaging (11, 12) were used to measure the location and surface area of NAPL ganglia trapped in 1-D columns. Light-transmission (13-15) and fluorescence microscopy (16) were also used to measure key parameters related to NAPL morphology and distribution in micromodel representations of pore structures. Results from these studies revealed that, among other things, the NAPL dissolution rate is directly related to the NAPL surface area except when flow bypassing occurs. At the continuum-scale, both light-transmission imaging and dual-energy γ-attenuation were used to visualize NAPL source zone architecture (5) and in some cases during dissolution (17, 18), in 2-D heterogeneous sandbox experiments. Several important findings are that the majority of the NAPL mass was in the form of pools (defined as continuous distribution), that residual ganglia accounted for only a small faction of the total mass in heterogeneous media (5, 17), and that NAPL concentrations measured at local sampling ports directly downgradient from source zones containing residual NAPL were much higher than average concentrations exiting the system due to dilution effects (18). Neither light-transmission nor γ-attenuation has been extended to three dimensions due to the limited penetration depth of the energy sources. A limited number of studies has quantified the effects of NAPL source zone architecture on water flowpaths and the subsequent effects on dissolution kinetics. In 2-D systems, this was addressed at the continuum-scale indirectly from model fits of NAPL dissolution rates (18-22) and directly using tracer visualization (18, 21, 23). Results indicate that high NAPL saturations cause flow bypassing and reduce NAPL dissolution rates. Water flowpaths have been measured in 1-D columns (24) and in several cases in the presence of 10.1021/es061675q CCC: $37.00

 2007 American Chemical Society Published on Web 04/11/2007

NAPL (11, 25) using a phase encoding MRI technique. Pure phase encoding techniques, however, are extremely slow and so not practical for many imaging applications. Characterization of the impacts of porous media heterogeneity on the NAPL source zone architecture and dissolution kinetics in 3-D at the continuum-scale are critical to improve the remediation efficiency and prediction of NAPL impacts on groundwater quality. The objectives of this work are to develop a 3-D visualization method using MRI to characterize the influence of porous media heterogeneity on NAPL saturation distribution in the source zone at the continuum-scale and to elucidate the impacts of source zone architecture on aqueous phase flowpaths and long-term dissolution kinetics. MRI was previously used to identify the location of NAPL in three dimensions; the sum of MR signal intensity was found to be proportional to the total NAPL mass remaining in the system (26). However, it has not been used to quantify the local (i.e., voxel-scale) saturation distribution.

Materials and Methods The porous media, NAPL, and paramagnetic tracer for MRI experiments are first discussed in this section; further rationale for their selection is provided in the Supporting Information. Porous Media. Five translucent, hydrophilic silica sand fractions (Accusand, Unimin Corp., Le Seur, MN) with different particle size distributions and hydraulic conductivities (see Table 1 in Supporting Information) were used. These sands have been used in previous multiphase flow and NAPL entrapment studies (10, 27-29). NAPL. A denser-than-water NAPL (i.e., DNAPL), 1,3,5trifluorobenzene (TFB) (97%, Sigma-Aldrich, Milwaukee, WI), was used. Relevant chemical and physical properties are similar to the common groundwater contaminant, trichloroethylene (see Table 2 in Supporting Information). The 19F nuclei in TFB can be imaged separately from the 1H in water using MRI (12). Paramagnetic Tracer. MnCl2 (99.999%, Sigma-Aldrich, Milwaukee, WI) was added to water as a paramagnetic tracer to visualize aqueous phase flowpaths. The pH of water with MnCl2 was adjusted to ∼2.5 to prevent Mn2+ sorption. Column breakthrough experiments showed that at this pH, sorption of Mn2+ by sand is negligible (30). Experimental Aquifer. A 3-D flowcell (see Supporting Information) was built for MR imaging. The inner dimensions are 25.5 cm × 9 cm × 8.5 cm; the central portion of the flowcell contained a spatially correlated random permeability field 14 cm × 8 cm × 8 cm in size, with uniform sand block sizes of 1 cm3. A sequential indicator simulator (SIS) (31) was used to generate the permeability field using an equal number of five different 1 cm3 blocks of uniform hydraulic conductivity, with correlation lengths of ∼2.1 cm (x), 1.1 cm (y), and 1.2 cm (z) (see Supporting Information). Each of the five different block types are represented by the five Accusand fractions. All flowcell components were constructed from nonferromagnetic materials to avoid distortions in the local magnetic field. There is one inlet, and the effluent is directed into nine separate outlets that draw from equal crosssectional areas. The area surrounding the heterogeneous zone was uniformly filled with 50:70 Accusand; this consisted of a 0.5 cm thick layer on the top, bottom, and walls of the flowcell, a 4.5 cm long zone adjacent to the flowcell inlet, and a 3.5 cm long zone adjacent to the flowcell outlet. The 50:70 Accusand formed a capillary barrier to NAPL migration. After packing, the flowcell was sealed at the top. Multiple pore volumes (PV) of CO2 gas were flushed through the flowcell to facilitate gas displacement with approximately 10 PV of

water. Further details of the flowcell are provided in the Supporting Information. NAPL Entrapment and Dissolution Experiments. Approximately 22.5 mL of NAPL was released into the experimental aquifer under a constant head (∼20 cm water) and allowed to equilibrate for ∼10 h. NAPL dissolution began by continuously flushing water via a constant head (∼10 cm) at a Darcy velocity of 1 m/day (see Figure 2 in Supporting Information). As the NAPL dissolved, effluent samples from the nine sampling ports (combined or individually) were collected in hexane and analyzed with a Hewlet Packard (HP) 6890 gas chromatograph (GC) equipped with a flame ionization detector (FID). 3-D images of NAPL were taken at approximately 24 h intervals (i.e., every 10 PV) during dissolution. Approximately 5-10% of NAPL dissolved during this interval. No NAPL mobilization was observed from images taken during dissolution. The entrapment and dissolution experiments were repeated in a duplicate flowcell; differences in measured effluent concentrations, total number of voxels containing NAPL, and average voxel-scale NAPL saturation were all within 5%. Aqueous Phase Flowpaths. Flowpaths were imaged before and after NAPL was released into the flowcell, and after partial NAPL removal during dissolution, by introducing a step input of a MnCl2 solution. The inlet chamber of the flowcell was first flushed with ∼10 PV of tracer solution by opening the valve on top of the inlet chamber and closing all nine downgradient ports. Next, the tracer solution was continuously flushed through the flowcell by opening all nine downgradient ports and closing the valve on top of the inlet chamber, under the same constant head used during NAPL dissolution. Successive 1H images of aqueous phase flowpaths were acquired while the tracer solution was flushed through the aquifer. Imaging Methods. To determine the range of MnCl2 concentrations that yields a linear dependence on MR signal intensity, calibration experiments were performed by imaging vials filled with sand and different concentrations of MnCl2 in water. To determine the imaging parameters that yield a linear dependence of MR signal intensity on NAPL mass, calibration experiments were performed by imaging four different diameter capillary tubes filled with different volumes of NAPL. Details of imaging equipment, imaging methods, and image processing methods are described in the Supporting Information.

Results Heterogeneous Permeability Field. 1H imaging results and a 2-D slice of the permeability field generated using SIS are shown in Figure 1. There is a linear dependence (R2 ) 0.97) of transverse relaxation time (T2) on the particle size (Figure 1a). Differences in T2 between different sand fractions are at least 70 ms. The results indicate that differences in the mean grain diameter as small as 0.1 mm can be distinguished using this method. A T2-weighting spin-echo multi-slice (SEMS) imaging sequence (see Supporting Information) was used to capture the 2-D 1H image of water in the pore spaces of sand packed in the experimental aquifer (Figure 1b). In general, the image adequately captures the sand distribution packed according to the permeability distribution generated using SIS (Figure 1c), but some noise is apparent. NAPL Source Zone Architecture. 19F imaging results are shown in Figure 2. There is a linear correlation between NAPL mass and signal intensity obtained for different diameter capillary tubes filled with NAPL (Figure 2a, R2 ) 0.98); the same imaging parameters (TR ) 3000 ms and TE ) 25 ms) also resulted in a linear correlation for the entire flowcell experiment (Figure 2b, R2 ) 0.97). These results indicate that imaging parameters were appropriately chosen. The linear relationship for the entire flowcell experiment also indicates VOL. 41, NO. 10, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3673

FIGURE 1. (a) 1H transverse relaxation time vs sand particle size; (b) T2-weighted 1H image of the heterogeneous permeability field in layer 4 (TR ) 2100 ms and TE ) 25 ms); and (c) heterogeneous permeability field in layer 4 generated using SIS. that sorption to flowcell components did not measurably affect NAPL mass. The linear relationship was used to calculate voxel-scale NAPL saturation in the experimental aquifer (Figure 2c,d) through uniform calibration of the maximum voxel-scale intensity, which varied from image to image due to changes in the magnetic susceptibility. This approach yielded a NAPL mass for any image that was within 7% of the NAPL mass determined from effluent measurements. Detailed discussion of this calculation is in the Supporting Information. Several quantitative measures were determined from the 3-D images as dissolution occurred. These are (i) the mass distribution of NAPL among the different grain sizes, (ii) the maximum NAPL source zone dimensions, (iii) the voxelscale frequency distribution of the NAPL saturation, and (iv) the NAPL mass ratio of residual ganglia to pools. Dobson et al. (32) showed that residual saturations of a NAPL (nhexadecane) in Accusand 12:20, 20:30, 30:40, and 40:50 are 0.15, 0.17, 0.18, and 0.20, respectively. We assumed that these residual saturation values apply to TFB as well and that NAPL in a voxel is uniformly distributed. As a result, these values were used as threshold saturations for the appropriate sand fractions (i.e., voxels with Sn at or below these values contained residual NAPL, and voxels with Sn above these values contained NAPL pools). Although different threshold Sn values between 0.15 and 0.20 for Accusand 12:20 only 3674

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007

marginally changed the NAPL ganglia-to-pool ratios (the difference is 0.15 are connected) are length (x) ∼12 cm, width (y) ∼5.6 cm, and height (z) ∼8 cm. In the x and y directions, NAPL is mainly located in the center of the aquifer (Figure 3b). Vertically (z), the NAPL saturation increases with depth. During the first 170 PV of water flushed (corresponding to 75% of NAPL mass removal), the source zone height decreased faster than the length or width (Figure 3c). After 170 PV, greater than 99% of the remaining NAPL was located in the bottom third of the flowcell, and the source zone length and width decreased at a greater rate than height. A histogram of the NAPL saturation categories and the NAPL ganglia-to-pool mass ratios during dissolution are shown in Figure 4. The largest number of voxels is generally in the lowest (e0.15) and highest (g0.75) NAPL saturation categories. Hence, a large number of voxels contains either residual ganglia or high saturation NAPL pools. The number of voxels decreases for each of the five NAPL saturation groups during dissolution. The relative frequency of NAPL saturation at the voxel-scale is generally constant. However, the ratio of the number of voxels with Sn e 0.15 to voxels with Sn g 0.75 increased from 99% of the NAPL was in the bottom third of the flowcell), preferential dissolution occurred in the longitudinal/transverse direction. Previous authors (35, 36) identified preferential dissolution along higher permeability flow paths; nonuniform dissolution fronts or dissolution fingering developed in a 2-D system that initially contained uniformly distributed residual NAPL. In this study, preferential dissolution was evident in a 3-D heterogeneous flow field that contained a heterogeneous distribution of both

FIGURE 6. (a) Concentration breakthrough curves of Mn2+ in two voxels located at y ) 4.3 cm, z ) 1.8 cm, x1 ) 1.88 cm (voxel 1), and x2 ) 5.1 cm (voxel 2); (b) distribution of normalized mean arrival time; and (c) normalized mean arrival time vs NAPL saturation.

FIGURE 5. (a) 1H signal intensity vs Mn2+ concentration and (b) 2-D images of tracer flowpaths before and (c) after NAPL release, both at the same point in time. NAPL residual ganglia and pools. This implies that under field conditions, preferential NAPL dissolution is likely to occur along preferential flow paths. Third, the ratio of NAPL mass as residual ganglia to NAPL mass in pools increases during dissolution. This suggests that the relative flow through NAPL contaminated zones increases over time, resulting in higher NAPL-water interfacial areas and causing local effluent concentrations measured in the middle and bottom center ports to initially increase and then decrease during dissolution. This agrees with the results of previous authors (19, 20, 37) who measured an initial increase in the effluent concentration during dissolution, either in a 1-D system with NAPL distributed in a core and surrounded by annular clean sand or in 2-D systems that contained one zone with NAPL surrounded by clean sand. Further studies using pore-scale imaging techniques such as X-ray synchrotron microtomography (10) are needed to elucidate the mechanism that control the change in NAPL configuration during dissolution. The high spatial and temporal resolution of water flow paths and NAPL saturation determined with MRI present an

FIGURE 7. Average and local (T2, M2, and B2) effluent concentrations. T2, M2, and B2 are the top center, middle center, and bottom center sampling ports. excellent opportunity to develop and test NAPL entrapment and dissolution models, including those that incorporate relatively permeability and/or capillary pressure-NAPL saturation functions. Existing models have either not been tested against experimental data or they have been tested against only simple experiments (e.g., 1-D column experiments and 2-D experiments with one or a few emplaced NAPL source zones) that do not represent the range of heterogeneity encountered in the field. VOL. 41, NO. 10, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3677

Acknowledgments Support for this research was provided by the National Science Foundation CAREER Award for C.J.W. (BES-9733428) and by the University of Illinois at Urbana-Champaign Research Board. We thank Boris Odintsov at the Biomedical Imaging Center, University of Illinois for imaging assistance and Hongkyu Yoon at the Department of Civil and Environmental Engineering, University of Illinois for helping with the SIS algorithm. We also thank four anonymous reviewers for their helpful comments.

Supporting Information Available Tables of Accusand properties and physical properties of TFB in comparison to TCE, schematic illustration of experimental setup, and details of MRI methods. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Schwille, F. Dense chlorinated solvents in porous and fractured media: model experiments; Translated from the German by Pankow, J. F., Lewis Publishers: Boca Raton, FL, 1988. (2) Kueper, B. H.; Abbot, W.; Farquhar, G. Experimental observations of multiphase flow in heterogeneous porous media. J. Contam. Hydrol. 1989, 5 (1), 83-95. (3) Illangasekare, T. H.; Ramsey, J. L.; Jensen, K. H.; Butts, M. B. Experimental study of movement and distribution of dense organic contaminants in heterogeneous aquifer. J. Contam. Hydrol. 1995, 20 (1), 1-25. (4) Illangasekare, T. H.; Armbruster, E. J.; Yates, D. N. Nonaqueous phase fluids in heterogeneous aquiferssExperimental study. J. Environ. Eng. 1995, 121 (8), 571-579. (5) Glass, R. J.; Conrad, S. H.; Peplinski, W. Gravity-destabilized nonwetting phase invasion in macroheterogeneous porous media: Experimental observations of invasion dynamics and scale analysis. Water Resour. Res. 2000, 36 (11), 3121-3137. (6) Hofstee, C.; Walker, R. C.; Dane, J. H. Infiltration and redistribution of perchloroethylene in stratified water-saturated porous media. Soil Sci. Soc. Am. J. 1998, 62 (1), 13-22. (7) Conrad, S. H.; Wilson, J. L.; Mason, W. R.; Peplinski, W. J. Visualization of residual organic liquid trapped in aquifers. Water Resour. Res. 1992, 28 (2), 467-478. (8) Mayer, A. S.; Miller, C. T. The influence of porous medium characteristics and measurement scale on pore-scale distributions of residual nonaqueous phase liquids. J. Contam. Hydrol. 1992, 11 (3-4), 189-213. (9) Powers, S. E.; Abriola, L. M.; Webber, W. J. An experimental investigation of nonaqueous phase liquid dissolution in saturated subsurface systems: Steady state mass transfer rates. Water Resour. Res. 1992, 28 (10), 2691-2705. (10) Schnaar, G.; Brusseau, M. L. Pore-scale characterization of organic immiscible-liquid morphology in natural porous media using synchrotron X-ray microtomography. Environ. Sci. Technol. 2005, 39 (21), 8403-8410. (11) Johns, M. L.; Gladden, L. F. Magnetic resonance imaging study of the dissolution kinetics of octanol in porous media. J. Colloid Interface Sci. 1999, 210 (2), 261-270. (12) Zhang, C.; Werth, C. J.; Webb, A. G. A magnetic resonance imaging study of dense nonaqueous phase liquid dissolution in angular porous media. Environ. Sci. Technol. 2002, 36 (15), 3310-3317. (13) Kennedy, C. A.; Lennox, W. C. A pore-scale investigation of mass transport from dissolving DNAPL droplets. J. Contam. Hydrol. 1997, 24, 221-246. (14) Jia, C.; Shing, K.; Yortsos, Y. C. Visualization and simulation of nonaqueous phase liquid solubilization in pore networks. J. Contam. Hydrol. 1999, 35 (4), 363-387. (15) Sahloul, N. A.; Ioannidis, M. A.; Chatzis, I. Dissolution of residual nonaqueous phase liquids in porous media: Pore-scale mechanism and mass transfer rates. Adv. Water Resour. 2002, 25, 33-49. (16) Chomsurin, C.; Werth, C. J. Analysis of pore-scale nonaqueous phase liquid dissolution in etched silicon pore networks. Water Resour. Res. 2003, 39 (9), 1265; doi:10.129/2002WR001543. (17) Oostrom, M.; Hofstee, C.; Walker, R. C.; Dane, J. H. Movement and remediation of trichloroethylene in a saturated heterogeneous porous medium 1. Spill behavior and initial dissolution. J. Contam. Hydrol. 1999, 37 (1-2), 159-178. 3678

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007

(18) Brusseau, M. L.; Nelson, N. T.; Oostrom, M.; Zhang, Z. H.; Johnson, G. R.; Wietsma, T. W. Influence of heterogeneity and sampling method on aqueous concentrations associated with NAPL dissolution. Environ. Sci. Technol. 2000, 34 (17), 36573644. (19) Powers, S. E.; Nambi, I. M.; Curry, G. W. Nonaqueous phase liquid dissolution in heterogeneous systems: Mechanisms and a local equilibrium modeling approach. Water Resour. Res. 1998, 34 (12), 3293-3302. (20) Nambi, I. M.; Powers, S. E. NAPL dissolution in heterogeneous systems: An experimental investigation in a simple heterogeneous system. J. Contam. Hydrol. 2000, 44 (2), 161-184. (21) Brusseau, M. L.; Zhang, Z. H.; Nelson, N. T.; Cain, R. B.; Tick, G. R.; Oostrom, M. Dissolution of nonuniformly distributed immiscible liquid: Intermediate-scale experiments and mathematical modeling. Environ. Sci. Technol. 2002, 36 (5), 10331041. (22) Saba, T.; Illangasekare, T. H. Effect of groundwater flow dimensionality on mass transfer from entrapped nonaqueous phase liquid contaminants. Water Resour. Res. 2000, 36 (4), 971979. (23) Barth, G. R.; Illangasekare, T. H.; Rajaram, H. The effect of entrapped nonaqueous phase liquids on tracer transport in heterogeneous porous media: Laboratory experiments at the intermediate scale. J. Contam. Hydrol. 2003, 67 (1-4), 247268. (24) Baumann, T.; Petsch, R.; Niessner, R. Direct 3-D measurement of the flow velocity in porous media using magnetic resonance tomography. Environ. Sci. Technol. 2000, 34 (19), 4242-4248. (25) Okamoto, I.; Hirai, S.; Ogawa, K. MRI velocity measurements of water flow in porous media containing a stagnant immiscible liquid. Meas. Sci. Technol. 2001, 12, 1465-1472. (26) Chu, Y. J.; Werth, C. J.; Valocchi, A. J.; Yoon, H. K.; Webb, A. G. Magnetic resonance imaging of nonaqueous phase liquid during soil vapor extraction in heterogeneous porous media. J. Contam. Hydrol. 2004, 73 (1-4), 15-37. (27) Schroth, M. H.; Ahearn, S. J.; Selker, J. S.; Istok, J. D. Characterization of Miller-similar sands for laboratory hydrologic studies. Soil Sci. Soc. Am. J. 1996, 60, 1331-1339. (28) Conrad, S. H.; Glass, R. J.; Peplinski, W. J. Bench-scale visualization of DNAPL remediation processes in analogue heterogeneous aquifers: Surfactant floods and in situ oxidation using permanganate. J. Contam. Hydrol. 2002, 58 (1-2), 13-49. (29) Imhoff, P. T.; Mann, A. S.; Mercer, M.; Fitzpatrick, M. Scaling DNAPL migration from laboratory to the field. J. Contam. Hydrol. 2003, 64 (1-2), 73-92. (30) Zhang, C. Y. Impacts of source zone architecture on nonaqueous phase liquid dissolution and cleanup: A magnetic resonance imaging study. Ph.D. Dissertation, University of Illinois at Urbana-Champaign, Urbana, IL, 2006. (31) Deutsch, C. V.; Journel, A. G. GSLIB, Geostatistical Software Library and User’s Guide; Oxford University Press: New York, 1992. (32) Dobson, R.; Schroth, M. H.; Oostrom, M.; Zeyer, J. Determination of NAPL-water interfacial areas in well-characterized porous media. Environ. Sci. Technol. 2006, 40 (3), 815-822. (33) Zenge, M. O.; Ladd, M. E.; Vogt, F. M.; Brauck, K.; Barkhausen, J.; Quick, H. H. Whole-body magnetic resonance imaging featuring moving table continuous data acquisition with highprecision position feedback. Magn. Reson. Med. 2005, 54 (3), 707-711. (34) Kleinberg, R. L.; Sezginer, A.; Griffin, D. D.; Fukuhara, M. Novel NMR apparatus for investigating an external sample. J. Magn. Reson. 1992, 97, 466-485. (35) Imhoff, P. T.; Thyrum, G. P.; Miller, C. T. Dissolution fingering during the solubilization of nonaqueous phase liquids in saturated porous media 2. Experimental observations. Water Resour. Res. 1996, 32 (7), 1929-1942. (36) Imhoff, P. T.; Farthing, M. W.; Gleyzer, S. N.; Miller, C. T. Evolving interface between clean and nonaqueous phase liquid (NAPL)contaminated regions in two-dimensional porous media. Water Resour. Res. 2002, 38 (6), 1093; 10.1029/2001WR000290. (37) Geller, J. T.; Hunt, J. R. Mass-transfer from nonaqueous phase organic liquids in water-saturated porous media. Water Resour. Res. 1993, 29 (4), 833-845.

Received for review July 13, 2006. Revised manuscript received February 19, 2007. Accepted February 23, 2007. ES061675Q