Hydrology [H]

 CC:711  Wednesday  1400h

Recent Advances in Remediation of NAPL-Contaminated Sites II

Presiding:  S Ghoshal, McGill University; D O'Carroll, University of Western Ontario


Understanding Processes Contribution to Plume Loading at DNAPL Sites Through Intermediate Scale Testing: Next Step in Translating to Field

* Illangasekare, T H (tissa@mines.edu), Center for Experimental Study of Subsurface Environmental Processes, Colorado School of Mine, 1600 Illinois St, Golden, CO 80401, United States
Soga, K (ks207@cam.ac.uk), Department of Engineering Cambrideg University, Trumpington St, Cambrideg, United Kingdom
Nikolopoulos, P (pn228@ca..ac.uk), Department of Engineering Cambrideg University, Trumpington St, Cambrideg, United Kingdom

In the last several decades, a considerable effort has been expended to improve the understanding of the fundamental processes associated with the behavior of chlorinated solvents in the subsurface with the goal of predicting concentrations in dissolved plumes in the field. Both under natural and conditions that are imposed during remediation, source zones containing entrapped solvents in the form of dense non-aqueous phase liquids (DNAPLs) are affected by the physical, geochemical and biological conditions and transformations that occur in complex subsurface soil-water environments. Research has been conducted in laboratory test systems ranging in scales from small columns to intermediate scale tanks to develop and validate up-scaling methods to transfer this knowledge to field systems. These investigations involved natural dissolution of entrapped DNAPLs, surfactant enhance dissolution, mass transfer under chemical oxidation, enhanced bio- activated dissolution and source and plume treatment using nano iron particles. In addition, plume loading as a result of rebound of diffused mass from low permeability zones was investigated. The tanks were packed to represent different configurations of soil heterogeneity and texture architecture, and different DNAPL entrapment architecture was created. Comprehensive data sets that relate different source zone conditions to plume generation were developed to validate some of the up-scaling methods. Even though these up-scaling methods as developed provide the correct theoretical foundation, practical application at field sites are constrained by the fact that the necessary field-scale parameters are not easy to obtain. This paper explores practical methods to transfer the knowledge obtained from the laboratory to field systems to make decisions on source zone and plume management. Issues such as how to incorporate effects of source zone architecture, parameters of the heterogeneity and source zone hydraulics into up-scaling in practical ways are explored. New strategies for site characterization to obtain information to predict mass flux generation from source zones are discussed.


Rebound of a Coal Tar Creosote Plume Following Partial Source Zone Treatment With Permanganate

Thomson, N (nthomson@uwaterloo.ca), University of Waterloo, Department of Civil and Environmental Engineering 200 University Avenue West, Waterloo, Ont N2L 3G1, Canada
* Fraser, M (michelle.fraser@stantec.com), Stantec Consulting Ltd., 49 Frederick Street, Kitchener, Ont N2H 6M7, Canada
Lamarche, C
EM: , Conestoga College, 299 Doon Valley Drive, Kitchener, ONT N2G 4M4, Canada
Barker, J (jfbarker@uwaterloo.ca), University of Waterloo, Deparment of Earth & Environmental Sciences 200 University Avenue West, Waterloo, Ont N2L 3G1, Canada
Forsey, S
EM: , University of Waterloo, Deparment of Earth & Environmental Sciences 200 University Avenue West, Waterloo, Ont N2L 3G1, Canada

The objective of this study was to investigate the potential of partial permanganate treatment to reduce the ability of a coal tar creosote source zone to generate a multi-component plume at the pilot-scale over both the short-term (weeks to months) and the long-term (years). A network of ~280 14-point multilevel samplers was used to monitor the dissolved plumes and mass discharge of ten compounds emanating from an emplaced coal tar creosote source for greater than 10 years. Bench scale experiments demonstrated that eight of the ten study compounds were readily oxidized by permanganate. Based on mass balance estimates, the 125 Kg of permanganate delivered to the source zone over 35 days would be capable of transforming at most 10% of the residual NAPL. This was sufficient to produce a short-term (after 150 days) decrease in mass discharge of greater than 35% for all monitored compounds except biphenyl, dibenzofuran, and fluoranthene. Pre- and post-treatment soil core data indicated a highly variable spatial distribution of mass within the source zone and provided no insight into the mass removed. The down-gradient plume was monitored approximately 1, 2 and 4 years following treatment. Once treated, oxidized compounds displayed a reduced plume mass and mass discharge while they migrated through the monitoring network. The data collected at 1 and 2 years post- treatment showed a decrease in mass discharge of 10 to 60% and/or total plume mass of 0 to 55%. Once the treated compounds migrated beyond the monitoring network (4-years post treatment) the mass discharge and plume mass of these compounds returned to pre-treatment values or higher. Non-reactive compounds displayed no significant change in mass discharge or plume mass. In the long term, reduction of mass discharge and total plume mass was within the error associated with mass discharge and total plume mass estimates, even at this highly monitored site.


Modelling of Electrical Resistance Heating in Porous Media

* Krol, M M (krol@ecf.utoronto.ca), University of Toronto, 35 St. George St., Toronto, ON M5S 1A4, Canada
Sleep, B E (sleep@ecf.utoronto.ca), University of Toronto, 35 St. George St., Toronto, ON M5S 1A4, Canada
Johnson, R L (rjohnson@ebs.ogi.edu), Oregon Health & Science University, 20000 NW Walker Road, Beaverton, OR 97006, United States

Electrical Resistance Heating (ERH) is an innovative remediation technology for sites contaminated with chlorinated solvents. The technique works by heating the subsurface, changing the vapour pressures, viscosities, and densities of contaminants and water, enhancing contaminant removal. Although significant energy savings can result from limiting temperature to sub-boiling levels during ERH, it is not clear how this impacts remediation. A two-dimensional (2D) finite difference model was developed to study this scenario, using a 2D tank experiment to validate results. To assess the usefulness of this technology, the effect of heating on water flow due to viscosity and buoyancy effect was examined and compared to experimental results. The effect of electrode placement and presence of different soil layers on heating was also examined. Due to the dependence of water viscosity and density on temperature, buoyancy induced upward flow was observed in the areas of localized heating. In addition, the numerical model indicates that electrode placement within a low permeability layer increased the diffusion of contaminants out of the layer and focused flow through the heated zone above the low permeability layer.


High Resolution Numerical Simulation of DNAPL Source Zone Remediation in Heterogeneous Porous Media

* Kueper, B (kueper@civil.queensu.ca), Queen's University, Department of Civil Engineering, Kingston, ON K7L 3N6, Canada
West, M (m.r.west@cogeco.ca), Queen's University, Department of Civil Engineering, Kingston, ON K7L 3N6, Canada

High resolution numerical simulation was used to evaluate the performance of DNAPL mass removal technologies in spatially correlated random permeablity fields. A number of template sites were created varying according to mean permeability, permeability variance, DNAPL type, and DNAPL release volume. A reactive transport model (RT3D) was coupled with a multiphase flow model (DNAPL3DRX) to simulate DNAPL migration and redistribution followed by treatment using either hydraulic displacment, chemical oxidation with potassium permanganate, surfactant flushing, or enhanced bioremediation. The various technologies were compared to each other on the basis of DNAPL mass removal, mass flux reduction, and concentration reduction. All technologies were also compared to a no treatment scenario involving ten years of dissolution. It is concluded that the performance of any particular technology is dependent on delivery, geology and DNAPL distribution.


Application of Steam Flushing to Removal of a DNAPL Mixture of Volatile and Very Low Volatility Organic Compounds

* Sleep, B E (sleep@ecf.utoronto.ca), Civil Engineering, University of Toronto, 35 St. George St., Toronto, ON M5S1A4, Canada
Zhang, Z (zzhang998@yahoo.ca) AB: A laboratory study was performed to assess the effectiveness of steam flushing for removal of a dense nonaqueous phase liquid (DNAPL) mixture of monochlorobenzene (MCB) and DDT. MCB has a boiling point of 131 C, while DDT has a melting point of approximately 110 C, but is soluble in MCB. In the study 600 mL of a MCB-DDT DNAPL mixture was injected into a medium sand layer above a capillary barrier (silt) in a cell of dimension 110 cm long by 60 cm high by 10 cm thick. After the DNAPL became immobilized as a pool on the capillary barrier steam was flushed through the system at an inlet temperature of 135 C and a pressure of 100 kPa, to represent removal of DNAPL from approximately 10 m below the water table. Approximately 3.5 pore volumes of steam (as condensate) were flushed through the DNAPL zone above the capillary barrier over an 11 hour period. Temperatures in the DNAPL zone exceeded the melting point of pure DDT and reached the boiling point of MCB. Initial DNAPL removal was due to hydraulic displacement (as indicated by DNAPL composition close to the source DNAPL composition), followed by primarily MCB removal by steam distillation. A total of 255 mL of DNAPL was recovered. Soil sampling following steam flushing showed levels of both MCB and DDT remaining in the original DNAPL zone that were consistent with the presence of DNAPL. Significant levels of MCB and DDT were also found below the capillary barrier, indicating that downward movement of MCB and DDT occurred, likely due to desaturation of the capillary barrier during steam flushing. In addition, heating to temperatures associated with significant depths below the water table enhances the mobility of the MCB-DDT DNAPL by reducing the DNAPL viscosity and preventing the solidification of DDT that might otherwise occur with MCB removal.


Air Sparging Versus Gas Saturated Water Injection for Remediation of Volatile LNAPL in the Borden Aquifer

* Barker, J (jfbarker@sciborg.uwaterloo.ca), University of Waterloo, Dept. of Earth & Environmental Sciences, Waterloo, ON N2L3G1, Canada
Nelson, L (leif_nelson@hotmail.com), WorleyParsons, 100 - 4500 16 Ave. NW, Calgary, AB T3B0M6, Canada
Doughty, C (Cynthia.Doughty@ontario.ca), Ontario Ministry of Environment, SW Region 733 Exeter Road, London, ON N6E1L3, Canada
Thomson, N (nthomson@uwaterloo.ca), University of Waterloo, Dept. of Civil & Environmental Engineering, Waterloo, ON N2L3G1, Canada
Lambert, J (lambert_j_m@yahoo.com), University of Waterloo, Dept. of Earth & Environmental Sciences, Waterloo, ON N2L3G1, Canada

In the shallow, rather homogeneous, unconfined Borden sand aquifer, field trials of air sparging (Tomlinson et al., 2003) and pulsed air sparging (Lambert et al., 2009) have been conducted, the latter to remediate a residual gasoline source emplaced below the water table. As well, a supersaturated (with CO2) water injection (SWI) technology, using the inVentures inFusion system, has been trialed in two phases: 1. in the uncontaminated sand aquifer to evaluate the radius of influence, extent of lateral gas movement and gas saturation below the water table, and 2. in a sheet pile cell in the Borden aquifer to evaluate the recovery of volatile hydrocarbon components (pentane and hexane) of an LNAPL emplaced below the water table (Nelson et al., 2008). The SWI injects water supersaturated with CO2. The supersaturated injected water moves laterally away from the sparge point, releasing CO2 over a wider area than does gas sparging from a single well screen. This presentation compares these two techniques in terms of their potential for remediating volatile NAPL components occurring below the water table in a rather homogeneous sand aquifer. Air sparging created a significantly greater air saturation in the vicinity of the sparge well than did the CO2 system (60 percent versus 16 percent) in the uncontaminated Borden aquifer. However, SWI pushed water, still supersaturated with CO2, up to about 2.5 m from the injection well. This would seem to provide a considerable advantage over air sparging from a point, in that gas bubbles are generated at a much larger radius from the point of injection with SWI and so should involve additional gas pathways through a residual NAPL. Overall, air sparging created a greater area of influence, defined by measurable air saturation in the aquifer, but air sparging also injected about 12 times more gas than was injected in the SWI trials. The pulsed air sparging at Borden (Lambert et al.) removed about 20 percent (4.6 kg) of gasoline hydrocarbons, mainly pentane and hexane, from the residual gasoline via sparging. A similar mass was estimated to have been removed by aerobic biodegradation. The extent of volatile recovery needs to be better defined and so post-sparging coring and analysis of residual LNAPL is underway. Impressively, the second SWI trial recovered more than 60 percent of the pentane-hexane from the NAPL. In both field experiments there was potential for minor additional recovery if the system had been operated longer. Comparison of efficiency of the pulsed air sparging and SWI systems is difficult in that the initial LNAPL residuals have different chemistry, but similar distribution, different volumes of gas were used, and biodegradation accounted for a significant removal of hydrocarbons only in the air sparging system. The SWI trial recovered an impressive portion of the volatile LNAPL, while using considerably less gas than the air sparging system, but the SWI delivery system was both more complex and more expensive than the air sparging system. Additional trials are underway in more complex aquifers to further assess the performance of the SWI technology, including costs and practical limitations.


In Situ Development of Gas Saturation by Supersaturated Water Injection in Porous Media

* Ioannidis, M (mioannid@cape.uwaterloo.ca), University of Waterloo, 200 University Ave. W., Waterloo, ON N2L3G1, Canada
Zhao, W (wzhao@uwaterloo.ca), University of Waterloo, 200 University Ave. W., Waterloo, ON N2L3G1, Canada
Li, m (m46li@uwaterloo.ca), University of Waterloo, 200 University Ave. W., Waterloo, ON N2L3G1, Canada
Enouy, R (renouy@uwaterloo.ca), University of Waterloo, 200 University Ave. W., Waterloo, ON N2L3G1, Canada
Unger, A (aunger@uwaterloo.ca), University of Waterloo, 200 University Ave. W., Waterloo, ON N2L3G1, Canada

Gas saturation can develop in situ within initially water-saturated porous media by injecting a gas- supersaturated aqueous phase. Supersaturated water injection (SWI) is a novel technology with several potential applications to groundwater and soil remediation. These include supply of reactive gases for in situ bioremediation and recovery of immobile non-aqueous phase liquids trapped in the form of ganglia within the pore space. In this presentation we provide insight into the fundamental physics of gas cluster growth and mobilization using pore-scale simulation. We employ a 2D pore network model which idealizes the pore space as a lattice of cubic chambers connected by tubes of square cross-section. Following heterogeneous nucleation, growth of gas phase clusters is driven by convective diffusion of solute from the bulk aqueous phase and is characterized by a ramified pattern controlled by capillary and buoyancy forces. Gas cluster coalescence, mobilization under the action of buoyancy forces and fragmentation resulting from capillary instabilities are accounted for in the model. Novel results on gas-phase cluster growth pattern and mobilization dynamics, relative permeability and macroscopic mass transfer rate coefficient are obtained. Modeling of the SWI process at the macroscopic (Darcy) scale is subsequently addressed using a multiphase, multi- component continuum model. This model is calibrated against experimental data obtained from columns packed with sand to obtain the dependence of gas relative permeability on gas saturation.