Remediation of DNAPL-Contaminated Subsurface Systems Using

Jan 4, 2000 - They will remain stationary as long as the capillary forces are ... mobilization occurs, thereby not relying upon mass-transfer limited ...
1 downloads 0 Views 2MB Size
Research Communications Remediation of DNAPL-Contaminated Subsurface Systems Using Density-Motivated Mobilization C. T. MILLER,* E. H. HILL, III, AND M. MOUTIER Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, North Carolina 27599-7400

Dense nonaqueous phase liquids (DNAPLs), especially when present as trapped free-phase pools, are long-lived in the subsurface environment and extremely difficult to remove. Despite vigorous research efforts over the last two decades, all current DNAPL pool remediation strategies suffer from a combination of inefficiency, increased risk of contaminant spreading due to uncontrolled mobilization, and or high treatment costs. This work reports results from two novel strategies that use density-motivated approaches to mobilize DNAPLs. Experiments were conducted in onedimensional bench-scale columns and two-dimensional flow cells containing heterogeneous media and DNAPL pools to quantify removal efficiency. Results show that greater than 90% removal of the total DNAPL mass (including both pools and associated residuals) is possible. Further, the removal process is potentially safe and rapid: since it is a stable displacement strategy, it does not rely upon any mass transfer limited steps, and it can be applied using a total of less than one pore volume of flushing solution.

Introduction Restoration of environmental systems contaminated with hazardous wastes is a costly problem, averaging about $9 million per site in 1996 with an expected $750 billion in restoration remaining to be accomplished over the next 30 years in the United States alone (1). Contamination of subsurface environments by nonaqueous phase liquids that are denser than water (DNAPLs), such as chlorinated solvents, is an important subset of this overall restoration problem because such contaminants are believed to be a significant risk to human health when present at relatively low concentrations in the aqueous phase. Remediation of DNAPL-contaminated systems is one of the most difficult problems in the broad field of the environmental sciences, because DNAPLs are long-lived in the subsurface environment and extremely difficult to removes so difficult that some researchers have described the situation as “remediation in perpetuity” (2). These opinions aside, vigorous research over the last two decades has continued to pursue an effective strategy. These efforts have included direct pumping of the NAPL phase (3) and flushing of contaminated systems with water (2, 4), wateralcohol mixtures (5-9), water-surfactant mixtures (10-13), steam (14, 15), and a gas phase (16-18). Significant efforts * Corresponding author phone: (919)966-2643; fax: (919)966-7911; e-mail: casey•[email protected]. 10.1021/es990808n CCC: $19.00 Published on Web 01/04/2000

 2000 American Chemical Society

have also been expended to understand and promote biological transformations of DNAPLs, such as the chlorinated solvents tetrachloroethylene and trichloroethylene (19-23). Despite all of these investigations, an efficient and safe approach for removing DNAPLss that is, one requiring relatively few pore volumes (PVs) of an effective flushing solution and presenting limited risk for spreading the contamination to previously uncontaminated portions of the systems has yet to be advanced. The objectives of this work are (1) to describe two novel strategies for the remediation of DNAPL-contaminated porous medium systems, especially those containing DNAPL pools; (2) to investigate the effectiveness of these two strategies in heterogeneous porous medium systems; and (3) to suggest modifications of these basic approaches that warrant further consideration.

Background A significant reason for the ineffectiveness of DNAPL-removal strategies is that subsurface systems are heterogeneous in nature with respect to both the physical and chemical characteristics of the porous medium (24). As a result, DNAPLs released into the subsurface follow complex patterns of flow. These patterns create regions where the DNAPL is continuous across a large number of pores and result in local DNAPL saturations above residual. We define these continuous DNAPL regions as pools. They will remain stationary as long as the capillary forces are greater than the sum of gravitational and viscous forces (11). DNAPL pools are difficult to remove using conventional approaches, because these approaches rely upon mass transfer from the stationary DNAPL to a mobile flushing phase; rates of mass transfer from DNAPL pools to the mobile phases can be relatively slow, leading to inefficient removal (25-29). When enhanced flushing solutions such as cosolvents or surfactants are used, large reductions in interfacial tensions can occur, sometimes leading to mobilization in the vertical direction, which can spread the contaminant to previously uncontaminated portions of the system (5, 10, 11, 30-32).

Approach Our approaches involve modifying the balance of forces such that mobilization occurs, thereby not relying upon masstransfer limited processes. By considering a static balance between the buoyancy and capillary forces acting on a continuous DNAPL “blob” located within a porous medium and by assuming a zero contact angle between the NAPL and solid grains, one can derive (11) an equation of the form

σn-a g (Fn - Fa)gl r

(1)

where σn-a is the NAPL-aqueous phase interfacial tension, r is an effective pore size arising from local variations in entry pressures, Fn is the NAPL density, Fa is the aqueous phase density, g is gravitational acceleration, and l is a characteristic blob or pool length in the vertical direction. If buoyancy driving forces are able to overcome the capillary resisting forces either through density modification or through reductions in interfacial tensions, then the static balance will not hold, and the blob will be mobilized. Applying this idea to ensure safe mobilization, we demonstrate the use of dense brine solutions to control DNAPL migration. The two approaches investigated include VOL. 34, NO. 4, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

719

FIGURE 1. One-dimensional experimental apparatus and TCE volumetric saturations as determined by X-ray analysis before and after the NaI flush.

FIGURE 2. Demonstration of DNAPL pool removal using a dense brine flushing solution within a column.

FIGURE 3. Two-dimensional experimental apparatus. 720

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 4, 2000

FIGURE 4. Demonstration of DNAPL pool removal using a dense brine flushing solution within a two-dimensional, heterogeneous domain.

TABLE 1. Porous Media Properties property

Accusand 20/30 (33)

ASTM C778

U.S. Silica F-52

d50 (mm) uniformity coeff (d60/d10) hydraulic conductivity (cm/s) air entry pressure (cm H2O)

0.713 ( 0.023 1.19 ( 0.028 0.250 ( 0.005 8.66

0.347 ( 0.03 1.64 ( 0.04 0.021 ( 0.002 25.9 ( 0.5

0.263 ( 0.03 1.73 ( 0.04 0.0081 ( 0.0004 38.5 ( 0.5

(1) a two-fluid-phase approach in which the DNAPL is displaced upward due to buoyancy forces, and (2) a threefluid-phase approach in which the DNAPL is displaced downward under gravitational forces and collected on top of a dense brine layer that limits vertical transport. Three

different kinds of experiments were performed to investigate these two approaches. Table 1 lists the porous media used in these experiments, along with their physical properties. The most important properties for the purposes of these studies are grain size VOL. 34, NO. 4, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

721

TABLE 2. Physical Properties of the Fluidsa property

NaI solution

TCE w/ORO

1% SDSS solution

composition (by wt) density @20 °C (g/cm3) viscosity @20 °C (mPa‚s)

60% NaI, 40% H2O 1.7957 ( 0.0002 2.838 (35)

99.99% TCE, 0.01% ORO 1.4639 ( 0.0002 0.577 ( 0.004 (36)

1% SDSS, 99% H2O 0.9987 ( 0.0002 0.989 ( 0.007 @26 °C (34)

a

Where SDSS is a 1:1 by weight mixture of sodium diamyl- and sodium dioctyl-sulfosuccinates and ORO is the biological stain Oil Red O.

FIGURE 5. Demonstration of DNAPL removal from a two-dimensional, heterogeneous domain using desaturation in conjunction with a dense brine barrier. and air entry pressure, which determine the equilibrium distribution of DNAPL. Fluid properties used in these studies are listed in Table 2. Densities were measured using an Anton Paar DMA-48 density meter (Graz, Austria), and viscosities 722

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 4, 2000

were interpolated from published values (35, 36) as a function of salt concentration. Trichloroethylene (TCE), a typical chlorinated solvent often found at Superfund sites, was the DNAPL used for all studies. Sodium iodide (NaI) solutions

were used as the dense brine solutions for all experiments, although many alternatives exist and could be used in practice, with economic and local geochemical conditions determining the optimal choice for any given application.

Results and Discussion In the first type of experiment, a series of three investigations were performed to quantify the extent of TCE removal from pooled porous medium systems. As shown in Figure 1, these experiments were performed in a 25-cm long, 2.54-cm inside diameter glass column that was packed with 4.8-cm of ASTM C778 sand underlying 19.6-cm of Accusand 20/30. Figure 2 and the Supporting Information depict a sequence of events including (a) a distilled, deionized (DDI) water-saturated initial condition; (b) creation of a TCE pool by the C778A20/30 capillary barrier; (c,d) an upward flush with NaI solution at a rate of 0.09 mL/min that mobilized the TCE; (e) a downward flush with DDI water that removed the brine solution; and (f) the remaining TCE present only as a residual. Due to resets during the experiment and to the roll-over of the digital clock, the times shown in Figure 2 and the other images should only be used to correlate with the movies contained in the Supporting Information. Fluid saturations were monitored using X-ray attenuation methods (10, 31), which employed a 160 keV Pantak X-ray generation unit (Tronix, Branford, CT) and a high-performance germanium detector and multichannel analyzer (EG&G Ortec, Oak Ridge, TN). The X-ray unit was run at 90 keV and 10 mA with a beam collimated to a 5 × 5-mm projected area and a 5 s live counting time. The removal of NAPL from the pooled regions was fast and dramatic, extracting between 65.3 and 74.0% of the NAPL mass initially contained within each pool. The TCE remaining after brine displacement was distributed throughout the vertical column, and both visual inspection and X-ray analysis revealed no evidence of DNAPL pools, as shown in Figure 1. A second two-fluid-phase experiment was performed using a two-dimensional flow cell, as shown in Figure 3a, with an internal thickness of 2 ( 0.2 mm. Figure 4 and the Supporting Information show a sequence of images describing (a) the cell originally saturated with DDI water; (b) TCE added from a single syringe point near the top of the cell over a period of about 25 min and allowed to redistribute over a period of 5 h; (c) conditions after flushing with approximately 0.5 PV of NaI solution injected at a rate of 1 mL/min through the stainless steel screen along the bottom of the cell; (d) the state after approximately 1 PV of a NaI solution was added; (e) a downward flush using DDI water to remove the NaI solution; and (f) TCE remaining after the brine has been removed. Using a needle (located in the cavity above the sand) to capture the mobilized TCE, 54.2% of the total 6.510 ( 0.01 g of the initially injected TCE was recovered as free-phase. It is important to note that, as with the column experiments, the DNAPL pools were completely mobilized during the brine injection procedure as a direct result of buoyancy forces. In addition, the remaining TCE was distributed throughout the medium with a morphology that had a relatively large ratio of surface area to volume, which is an ideal arrangement for efficient removal using conventional flushing approaches, such as cosolvents or surfactants. A third experiment was performed in the two-dimensional flow cell using a three-fluid-phase approach. The cell was packed with the same media as shown in Figure 3b and with the same internal dimensions. The panes in Figure 5 and the Supporting Information show (a) an initial water-saturated condition; (b) stable TCE pools resulting after injection from the top of the cell using a needle and an equilibration time of 5 h; (c) a thin brine layer established at the bottom of the domain; (d) the upper portion of the cell dewatered; (e) a 0.3 PV pulse of surfactant added at the top of the cell; and (f)

surfactant injection, which mobilized both DNAPL and water toward the extraction needle. From image analysis (that is, by relating DNAPL mass to areal extent within the digitized images) we estimate that greater than 90% of the initial TCE injected into this heterogeneous system was removed. Further, the addition of the surfactant resulted in water removal as well, leaving the small remaining fraction of TCE in a state of relatively large surface area per volume and low water saturation, which is ideal for removing the remaining fraction of this volatile contaminant via gas-phase extraction. In each of the three types of experiments performed, efficient removal of TCE was observed from systems consisting of heterogeneous porous media, considering no more than one PV of flushing solution was used in any case. In all cases, both the observed mobilization of pools and the trapping of residual ganglia can be predicted from eq 1 by applying appropriate lengths for the characteristic length l. For the three-fluid-phase experiment, less than 0.5 PV of total flushing solution was used, including both brine and the surfactant. Further, these results are only first attempts at density-enhanced DNAPL remediation methods. Additional work to fine-tune the sequence of flushing solutions and stages of the processes are expected to yield improved results. Following the three-fluid-phase process with a stage of vapor extraction, for example, appears to be an especially promising approach. Areas for future investigation include developing refined flushing strategies and solutions, extending the processes to larger scales and three-dimensional systems, developing a mathematical model to describe the process, and conducting field-scale trials of the processes. If these strategies are as successful at the field scale, as the experiments performed to date would suggest, densityenhanced strategies could result in a relatively economical and safe method for remediating DNAPL source zones.

Acknowledgments The work was supported by Grant 5 P42 ES05948-02 from The National Institute of Environmental Health Sciences.

Supporting Information Available Additional images and movies depicting all three of the experiments. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Oppelt, E. T. Technical Report EPA/600/R-98/154; United States Environmental Protection Agency, Government Printing Office: Washington, DC, 1999. (2) Mackay, D. M.; Cherry, J. A. Environ. Sci. Technol. 1989, 23(6), 630-636. (3) Pankow, J. F.; Cherry, J. A. Dense Chlorinated Solvents and Other DNAPLs in Groundwater: History, Behavior, and Remediation; Waterloo Press: Portland, OR, 1996. (4) Bartow, G.; Davenport, C. Ground Water Monitoring Remediation 1995, 15(2), 140-146. (5) Brandes, D.; Farley, K. J. J. Water Environ. Res. 1993, 65(7), 869878. (6) Roeder, E.; Brame, S. E.; Falta, R. W. In Non-Aqueous Phase Liquids (NAPLs) in Subsurface Environment: Assessment and Remediation; Reddi, L. N., Ed.; New York, 1996; pp 333-344. (7) Rao, P. S. C.; Annable, M. D.; Sillan, R. K.; Dai, D.; Hatfield, K.; Graham, W. D.; Wood, A. L.; Enfield, C. G. Water Resources Res. 1997, 33(12), 2673-2686. (8) Lunn, S. R. D.; Kueper, B. H. Water Resources Res. 1997, 33(10), 2207-2219. (9) Lunn, S. R. D.; Kueper, B. H. Environ. Sci. Technol. 1999, 33(10), 1703-1708. (10) Okuda, I.; McBride, J. F.; Gleyzer, S. N.; Miller, C. T. Environ. Sci. Technol. 1996, 30(6), 1852-1860. (11) Pennell, K. D.; Pope, G. A.; Abriola, L. M. Environ. Sci. Technol. 1996, 30, 1328-1335. (12) Mason, A. R.; Kueper, B. H. Environ. Sci. Technol. 1996, 30(11), 3205-3215. VOL. 34, NO. 4, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

723

(13) Willson, C. S.; Hall, J. L.; Imhoff, P. T.; Miller, C. T. Environ. Sci. Technol. 1999, 33(14), 2440-2446. (14) Forsyth, P. A. Intl. J. Numerical Methods Fluids 1994, 19(12), 1055-1081. (15) Sittler, S. P.; Swinford, G. L.; Gardner, D. G. In Proceedings of the 1992 Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Restoration; The American Petroleum Institute and The Association of Ground Water Scientists and Engineers: 1992; pp 413-426. (16) Larkin, R. G.; Hemingway, M. P. In Proceedings of the 1991 Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Restoration; 1991; pp 191204. (17) Ahlfeld, D.; Dahmani, A.; Hoag, G.; Farrel, M.; Ji, W. In Proceedings of the Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation Conference; Ground Water Publishing Company; Dublin, OH, 1994; pp 175190. (18) Bass, D. H.; Brown, R. A. In Proceedings of the 1995 Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation; Ground Water Publishing Company; Dublin, OH, 1995; pp 621-636. (19) Lollar, B. S.; Slater, G. F.; Ahad, J.; Sleep, B.; Spivack, J.; Brennan, M.; MacKenzie, P. Org. Geochem. 1999, 30(8A), 813-820. (20) Ndon, U. J.; Randall, A. A. Water Res. 1999, 33(11), 2715-2720. (21) Middeldorp, P. J. M.; van Aalst, M. A.; Rijnaarts, H. H. M.; Stam, F. J. M.; de Kreuk, H. F.; Schraa, G.; Bosma, T. N. P. Water Sci. Technol. 1998, 37(8), 105-110. (22) Rothmel, R. K.; Peters, R. W.; St Martin, E.; Deflaun, M. F. Environ. Sci. Technol. 1998, 32(11), 1667-1675. (23) Chu, K. H.; Alvarez-Cohen, L. Appl. Environ. Microbiol. 1998, 64(9), 3451-3457.

724

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 4, 2000

(24) Thompson, A. F. B. Water Resources Res. 1993, 29(11), 37093726. (25) Whelan, M. P.; Voudrias, E. A.; Pearce, A. J. Contaminant Hydrol. 1994, 15(3), 223-237. (26) Pearce, A. E.; Voudrias, E. A.; Whelan, M. P. J. Environ. Eng.ASCE 1994, 120(5), 1191-1206. (27) Hunt, J. R.; Sitar, N.; Udell, K. S. Water Resources Res. 1988, 24(8), 1247-1258. (28) Hunt, J. R.; Sitar, N.; Udell, K. S. Water Resources Res. 1988, 24(8), 1247-1258. (29) Johnson, R. L.; Pankow, J. F. Environ. Sci. Technol. 1992, 26(5), 896-901. (30) Pennell, K. D.; Abriola, L. M.; Weber, J., Jr. Environ. Sci. Technol. 1993, 27(12), 2332-2340. (31) Imhoff, P. T.; Gleyzer, S. N.; McBride, J. F.; Vancho, L. A.; Okuda, I.; Miller, C. T. Environ. Sci. Technol. 1995, 29(8), 1966-1976. (32) Grubb, D. G. Ph.D. Dissertation, University of California at Berkeley, 1995. (33) Schroth, M. H.; Ahearn, S. J.; Selker, J. S.; Istok, J. D. J. Soil Sci. Soc. Am. 1996, 60(5), 1331-1339. (34) Hall, J. L. Technical Report 1449; Department of Environmental Science and Engineering, University of North Carolina at Chapel Hill: 1997. (35) Lengyel, S. J.; Tams, J.; Giber, J.; Holderyth, J. Acta Chim. Hung. 1964, 40, 125. (36) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics; CRC Press: Boston, 1997.

Received for review July 16, 1999. Accepted November 24, 1999. ES990808N