A theoretical model of air and steam co-injection to prevent the downward migration of DNAPLs during steam-enhanced extraction

When steam is injected into soil containing a dense volatile non-aqueous phase liquid contaminant the DNAPL vaporized within the heated soil region condenses and accumulates ahead of the steam condensation front. If enough DNAPL accumulates, gravitational forces can overcome trapping forces allowing the liquid contaminant to flow downward. By injecting air with steam, a portion of the DNAPL vapor remains suspended in equilibrium with the air, decreasing liquid contaminant accumulation ahead of the steam condensation front, and thus reducing the possibility of downward migration. In this work, a one-dimensional theoretical model is developed to predict the injection ratio of air to steam that will prevent the accumulation of volatile DNAPLs. The contaminated region is modeled as a one-dimensional homogeneous porous medium with an initially uniform distribution of a single component contaminant. Mass and energy balances are combined to determine the injection ratio of air to steam that eliminates accumulation of the contaminant ahead of the steam condensation front, and hence reduces the possibility of downward migration. The minimum injection ratio that eliminates accumulation is defined as the optimum injection ratio. Example calculations are presented for three DNAPLs, carbon tetrachloride (CCl4), trichloroethylene (TCE), and perchloroethylene (PCE). The optimum injection ratio of air to steam is shown to depend on the initial saturation and the volatility of the liquid contaminant. Numerical simulation results are presented to validate the model, and to illustrate downward migration for ratios less than optimum. Optimum injection ratios determined from numerical simulations are shown to be in good agreement with the theoretical model.     

Removal of NAPLs from the unsaturated zone using steam: prevention of downward migration by injecting mixtures of steam and air

Steam injection for remediation of porous media contaminated by nonaqueous phase liquids has been shown to be a potentially efficient technology. There is, however, concern that the technique may lead to downward migration of separate phase contaminant. In this work, a modification of the steam injection technology is presented, where a mixture of steam and air was injected. In two-dimensional experiments with unsaturated porous medium contaminated with nonaqueous phase liquids, it was demonstrated how injection of pure steam lead to severe downward migration. Similar experiments, where steam and air were injected simultaneously, resulted in practically no downward migration and still rapid cleanup was achieved. The processes responsible for the prevention of downward migration when injecting steam-air mixtures were analyzed using a nonisothermal multiphase flow and transport model. Hereby, three mechanisms were identified and it was demonstrated how the effectiveness of these mechanisms depended on the air-to-steam mixing ratio.     

Dip-angle influence on areal DNAPL recovery by co-solvent flooding with and without pre-flooding

A two-dimensional (2D) laboratory model was used to study effects of gravity on areal recovery of a representative dense non-aqueous phase liquid (DNAPL) contaminant by an alcohol pre-flood and co-solvent flood in dipping aquifers. Recent studies have demonstrated that injection of alcohol and co-solvent solutions can be used to reduce in-situ the density of DNAPL globules and displace the contaminant from the source zone. However, contact with aqueous alcohol reduces interfacial tension and causes DNAPL swelling, thus facilitating risk of uncontrolled downward DNAPL migration. The 2D laboratory model was operated with constant background gradient flow and a DNAPL spill was simulated using tetrachloroethene (PCE). The spill was dispersed to a trapped, immobile PCE saturation by a water flood. Areal PCE recovery was studied using a double-triangle well pattern to simulate a remediation scheme consisting of an alcohol pre-flood using aqueous isobutanol ( approximately 10% vol.) followed by a co-solvent flood using a solution of ethylene glycol (65%) and 1-propanol (35%). Experiments were conducted with the 2D model oriented in the horizontal plane and compared to experiments at the 15 degrees and 30 degrees dip-angle orientations. Injection was applied either in the downward or upward direction of flow. Experimental results were compared to theoretical predictions for flood front stability and used to evaluate effects of gravity on areal PCE recovery. Sensitivity experiments were performed to evaluate effects of the alcohol pre-flood on PCE areal recovery. For experiments conducted with the alcohol pre-flood and the 2D model oriented in the horizontal plane, results indicate that 89-93% of source zone PCE was recovered. With injection oriented downward, results indicate that areal PCE recovery was 70-77% for a 15 degrees dip angle and 57-59% for a 30 degrees dip angle. With injection oriented upward, results indicate that areal PCE recovery was 57-60% at the 30 degrees dip angle, which was similar to PCE recovery for injection in the downward flow direction. Lower areal PCE recovery at greater dip angles in either direction of flow was attributed to DNAPL swelling and migration, flood front instabilities and bypassing of the displaced fluid past the extraction wells during the alcohol pre-flood. Additional results demonstrate that the use of an alcohol pre-flood can be beneficial in improving DNAPL recovery in the horizontal orientation, but pre-flooding may reduce areal recovery efficiency in dip-angle orientations. This study also demonstrates the use of theoretical perturbation (fingering) analysis in predicting NAPL recovery efficiency for flooding processes in remediating aquifers with dip angles.     

Removal of volatile and semivolatile organic contamination from soil by air and steam flushing

A soil core, obtained from a contaminated field site, contaminated with a mixture of volatile and semivolatile organic compounds (VOC and SVOC) was subjected to air and steam flushing. Removal rates of volatile and semivolatile organic compounds were monitored during flushing. Air flushing removed a significant portion of the VOC present in the soil, but a significant decline in removal rate occurred due to decreasing VOC concentrations in the soil gas phase. Application of steam flushing after air flushing produced a significant increase in contaminant removal rate for the first 4 to 5 pore volumes of steam condensate. Subsequently, contaminant concentrations decreased slowly with additional pore volumes of steam flushing. The passage of a steam volume corresponding to 11 pore volumes of steam condensate reduced the total VOC concentration in the soil gas (at 20 degrees C) by a factor of 20 to 0.07 mg/l. The corresponding total SVOC concentration in the condensate declined from 11 to 3 mg/l. Declines in contaminant removal rates during both air and steam flushing indicated rate-limited removal consistent with the persistence of a residual organic phase, rate-limited desorption, or channeling. Pressure gradients were much higher for steam flushing than for air flushing. The magnitude of the pressure gradients encountered during steam flushing for this soil indicates that, in addition to rate-limited contaminant removal, the soil permeability (2.1 x 10(-9) cm2) would be a limiting factor in the effectiveness of steam flushing.     

Field study of TCE diffusion profiles below DNAPL to assess aquitard integrity

An area where a free-product accumulation of trichloroethylene (TCE) dense non-aqueous phase liquid (DNAPL) occurs at the bottom of a 10-m-thick surficial sand aquifer was studied to determine the integrity of the underlying, 20-m-thick, clayey silt aquitard formed of glaciolacustrine sediment. TCE concentration-versus-depth profiles determined from aquitard cores collected at five locations indicated penetration of detectable TCE 2.5 to 3.0 m into the aquitard. Two of the profiles show persistent DNAPL at the aquitard interface, while two others indicate that DNAPL, present initially, was completely dissolved away producing concentration declines at the aquitard interface. The fifth profile suggests shallow DNAPL penetration (<0.5 m) into the aquitard, however, this penetration, which was likely caused by cross-contamination during core collection or cone penetrometry (CPT) of the aquitard interface, did not increase the maximum depth of TCE penetration. Combining the field profiles with one-dimensional model simulations, downward migration of the aqueous TCE front, defined as the EPA MCL of 5 microg/l, which was below the analytical detection limit, was projected to a distance between 4 and 5 m below the top of the aquitard. Using a single set of estimated aquitard parameter values, simulations of aqueous TCE migration into the aquitard provided a good fit to four of the field profiles with a migration time of 35 to 45 years, consistent with the history of TCE use at the site. These simulations indicate aqueous TCE migration is diffusion-dominated with only small advective influence by the downward groundwater velocity of 2 to 3 cm/year or less in the aquitard due to pumping of the underlying aquifer to supply water to the facility in the past 50 years. The applicability of the parameter values was confirmed by in situ diffusion experiments of 1-year duration, in which stainless steel cylinders containing DNAPL were inserted into the aquitard. The diffusion-dominated nature of the profiles indicates that the aquitard provides long-term protection of the underlying aquifer from contamination from this DNAPL zone. Simulations of long-term migration of the TCE solute front indicate breakthrough to the lower aquifer at 1200 years for the no advection scenario and at 500 years if the strong downward hydraulic gradient persists. However, even after breakthrough, the mass flux through the aquitard to the underlying aquifer remains relatively low, and when considered in terms of potential impacts to pumping wells, concentrations are not expected to increase significantly above present-day MCLs. The use of contaminant profiles of different time and distance scales, in addition to hydraulic data, dramatically improves the ability to assess aquitard integrity, and provides improved transport parameter values for estimating contaminant arrival times and fluxes. The apparent lack of deep preferential pathways for TCE migration, such as open fractures, is probably due to the softness of the silty aquitard deposit and minimal physical or chemical weathering of the aquitard provides long-term protection of the underlying aquifer from contamination from this DNAPL zone. Simulations of long-term migration of the TCE solute front indicate breakthrough to the lower aquifer at 1200 years for the no advection scenario and at 500 years if the strong downward hydraulic gradient persists. However, even after the breakthrough, the mass flux through the aquitard to the underlying aquifer remains relatively low, and when considered in terms of potential impacts to pumping wells , concentrations are not expected to increase significantly above present-day MCLs. The use of contaminant profiles of different time and distance scales, in addition to hydraulic data, dramatically improves the ability to assess aquitard integrity, and provides improved transport parameter values for estimating contaminant arrival times and fluxes. The apparent lack of deep preferential pathways for TCE migration, such as open fractures, is probably due to the softness of the silty aquitard deposit and minimal physical or chemical weathering of the aquitard.     

Aquifer washing by micellar solutions: 1: Optimization of alcohol–surfactant–solvent solutions

Phase diagrams were used for the formulation of alcohol–surfactant–solvent and to identify the DNAPL (Dense Non Aqueous Phase Liquid) extraction zones. Four potential extraction zones of Mercier DNAPL, a mixture of heavy aliphatics, aromatics and chlorinated hydrocarbons, were identified but only one microemulsion zone showed satisfactory DNAPL recovery in sand columns. More than 90 sand column experiments were performed and demonstrate that: (1) neither surfactant in water, alcohol–surfactant solutions, nor pure solvent can effectively recover Mercier DNAPL and that only alcohol–surfactant–solvent solutions are efficient; (2) adding salts to alcohol–surfactant or to alcohol–surfactant–solvent solutions does not have a beneficial effect on DNAPL recovery; (3) washing solution formulations are site specific and must be modified if the surface properties of the solids (mineralogy) change locally, or if the interfacial behavior of liquids (type of oil) changes; (4) high solvent concentrations in washing solutions increase DNAPL extraction but also increase their cost and decrease their density dramatically; (5) maximum DNAPL recovery is observed with alcohol–surfactant–solvent formulations which correspond to the maximum solubilization in Zone C of the phase diagram; (6) replacing part of surfactant SAS by the alcohol n-butanol increases washing solution efficiency and decreases the density and the cost of solutions; (7) replacing part of n-butanol by the nonionic surfactant HOES decreases DNAPL recovery and increases the cost of solutions; (8) toluene is a better solvent than D-limonene because it increases DNAPL recovery and decreases the cost of solutions; (9) optimal alcohol–surfactant–solvent solutions contain a mixture of solvents in a mass ratio of toluene to D-limonene of one or two. Injection of 1.5 pore volumes of the optimal washing solution of n-butanol–SAS–toluene–D-limonene in water can recover up to 95% of Mercier DNAPL in sand columns. In the first pore volume of the washing solution recovered in the sand column effluent, the DNAPL is in a water-in-oil microemulsion lighter than the excess aqueous phase (Winsor Type II system), which indicates that part of the DNAPL was mobilized. In the next pore volumes, DNAPL is dissolved in a oil-in-water microemulsion phase and is mobilized in an excess oil phase lighter than the microemulsion (Winsor Type I system). The main drawback of this oil extraction process is the high concentration of ingredients necessary for DNAPL dissolution, which makes the process expensive. Because mobilization of oil seems to occur at the washing solution front, an injection strategy must be developed if there is no impermeable limit at the aquifer base. DNAPL recovery in the field could be less than observed in sand columns because of a smaller sweep efficiency related to field sand heterogeneities. The role of each component in the extraction processes in sand column as well as the Winsor system type have to be better defined for modeling purposes. Injection strategies must be developed to recover ingredients of the washing solution that can remain in the soil at the end of the washing process. ©1997 Elsevier Science B.V.     

Displacement and sweep efficiencies in a DNAPL recovery test using micellar and polymer solutions injected in a five-spot pattern

Soil washing with micellar solutions is a promising alternative for the remediation of DNAPL source zones. As with any flushing technology, the success of soil washing with micellar solutions depends in a very large part on the ability of the solution to contact the contaminant (sweep efficiency) and then on the efficiency of contaminant removal once this contact is made (displacement efficiency). We report here on a field test where a micellar solution was used to recover a DNAPL in an open five-spot pattern in which polymer solutions were also injected before and after the washing solution to improve sweep efficiency. The washing solution formulation was optimised in the laboratory prior to the test to obtain good dissolution capacity. For a high-concentration and low-volume soil flushing remediation test such as the one performed (0.8 pore volumes of actual washing solution injected), slug sizing of the washing solution is critical. It was evaluated by an analytical solution. In a five-spot pattern, the displacement efficiency of the washing solution was observed to vary in the porous medium as a function of the radial distance from the injection well because: (1) the volume of the washing solution flowing through a section of the test cell changes (maximum close to the injection well and minimal at the pumping wells); (2) the in situ velocity changes (maximum at the wells and minimum between the wells) and; (3) the contact time of the washing solution with the NAPL changes as a function of the distance from the injection well. The relative importance of the recovery mechanisms, mobilisation and dissolution, was also observed to vary in the test cell. The reduced velocity increased the contact time of the washing solution with the DNAPL enhancing its dissolution, but the decrease of the capillary number caused less mobilisation. The washing process is much more extensive around the injection well. The use of an injection-pumping pattern allowing a complete sweep of the remediated area is essential. Following a comprehensive characterisation, modeling is an efficient tool to design the injection-pumping scheme and to optimise injection and pumping rates providing the best areal sweep. The vertical sweep can be controlled by using a polymer solution (Xanthan gum). The polymer solution also has a positive effect on front stability between the solutions injected. The injection rate of the polymer solution that follows the washing solution must be kept minimal initially to prevent dilution of the washing solution by fingering.     

A new concept: the use of neutrally-buoyant microemulsions for DNAPL remediation

Even in the absence of mobilization of dense nonaqueous phase liquid (DNAPL), the microemulsion that forms when the surfactant solubilizes a dense contaminant such as trichloroethylene will be more dense than water and tends to migrate downward. This paper addresses the issue of migration with a new concept: surfactant enhanced aquifer remediation at neutral buoyancy. Laboratory results of surfactant remediation in two-dimensional model aquifers show that downward migration of microemulsion containing solubilized dense contaminants can be reduced to an acceptable level, even in the absence of capillary barriers in the aquifer. One model experiment was designed to exhibit a small degree of vertical migration and full capture of the microemulsion at the extraction well. The second experiment was designed to demonstrate the effect of large buoyancy forces that lead to excessive downward migration of the microemulsion. Density measurements of aqueous solutions containing sodium dihexyl sulfosuccinate surfactant, isopropanol, trichloroethylene, and sodium chloride are presented. A companion paper presents the results of the flow and transport calculations needed for this approach to surfactant flooding.     

PCE dissolution and simultaneous dechlorination by nanoscale zero-valent iron particles in a DNAPL source zone

While the capability of nanoscale zero-valent iron (NZVI) to dechlorinate organic compounds in aqueous solutions has been demonstrated, the ability of NZVI to remove dense non-aqueous phase liquid (DNAPL) from source zones under flow-through conditions similar to a field scale application has not yet been thoroughly investigated. To gain insight on simultaneous DNAPL dissolution and NZVI-mediated dechlorination reactions after direct placement of NZVI into a DNAPL source zone, a combined experimental and modeling study was performed. First, a DNAPL tetrachloroethene (PCE) source zone with emplaced NZVI was built inside a small custom-made flow cell and the effluent PCE and dechlorination byproducts were monitored over time. Second, a model for rate-limited DNAPL dissolution and NZVI-mediated dechlorination of PCE to its three main reaction byproducts with a possibility for partitioning of these byproducts back into the DNAPL was formulated. The coupled processes occurring in the flow cell were simulated and analyzed using a detailed three-dimensional numerical model. It was found that subsurface emplacement of NZVI did not markedly accelerate DNAPL dissolution or the DNAPL mass-depletion rate, when NZVI at a particle concentration of 10g/L was directly emplaced in the DNAPL source zone. To react with NZVI the DNAPL PCE must first dissolve into the groundwater and the rate of dissolution controls the longevity of the DNAPL source. The modeling study further indicated that faster reacting particles would decrease aqueous contaminant concentrations but there is a limit to how much the mass removal rate can be increased by increasing the dechlorination reaction rate. To ensure reduction of aqueous contaminant concentrations, remediation of DNAPL contaminants with NZVI should include emplacement in a capture zone down-gradient of the DNAPL source.     

Removal of DNAPL contamination from the saturated zone by the combined effect of vertical upward flushing and density reduction

A common aspect of innovative remediation techniques is that they tend to reduce the interfacial tension between the aqueous and non-aqueous phase liquids, resulting in mobilization of the organic contaminant. This complicates the remediation of aquifers, contaminated with Dense Non-Aqueous Phase Liquids (DNAPLs), as they are likely to migrate downwards, deeper into the aquifer and into finer layers. A possible solution is the use of swelling alcohols, which tend to reduce the density difference between the aqueous phase and the DNAPL. To avoid premature mobilization upon the initial contact between the DNAPL and the alcohol, several researchers have proposed the use of vertical upward flow of the alcohol. In this paper, we present an equation, which describes the upward mobilization of both continuous and discontinuous DNAPLs and so the important parameters governing the upward controlled mobilization of the DNAPL. The need and required magnitude of this specific discharge was investigated by conducting four column experiments in which the initial density of the DNAPL and the permeability was varied. It was shown that the required flow velocities increase with the permeability of the porous medium and the initial density difference between the aqueous phase and the DNAPL. Whenever the specific discharge falls below the critical value, the DNAPL moves downward. A second set of column experiments looked at the impact of permeability of porous medium on the solubilization and mobilization of DNAPL during alcohol flooding. Columns, packed with coarse or fine sand, containing a residual trichloroethylene (TCE) or perchloroethylene (PCE) saturation were flushed with the alcohol mixture at a fixed specific discharge rate. The induced pressure gradients in the aqueous phase, which were higher in the fine sand, resulted for this porous medium in extensive mobilization of the DNAPL against the direction of the buoyancy force. The density of the first NAPL coming out of the top of the fine sand was close to that of the pure DNAPL. In the coarser sand, the pressure gradients were sufficient to prevent downward migration of the DNAPL, but upward mobilization was minimal. The predominant removal mechanism in this case was the much slower solubilization.     

Mobilization and entry of DNAPL pools into finer sand media by cosolvents: two-dimensional chamber studies

Two-dimensional chamber studies were conducted to determine qualitative and quantitative performance of cosolvents targeted at pooled dense non-aqueous phase liquid (DNAPL) (perchlorethylene, PCE) residing above a fine-grain capillary barrier. Downward mobilization of DNAPL, up gradient along an overriding cosolvent front, was observed. This produced significant pooling above a fine-grain layer that in some cases lead to entry into the capillary barrier beneath. Entry pressure calculations using physical and hydrogeologic parameters provided an excellent prediction of breakthrough of DNAPL into the capillary barrier. Calculations predict approximately 0.5 m of DNAPL would be necessary to enter a Beit Netofa clay, under extreme cosolvent flooding conditions (100% ethanol). Gradient injection of cosolvent did not appear to provide any benefit suggesting a rapid decrease in interfacial tension (IFT) compared to the rate of DNAPL solubilization. Use of a partitioning alcohol (tertiary butyl alcohol, TBA) resulted in DNAPL swelling and reduced entry into the capillary barrier. However, the trapping of flushing solution, containing PCE, could potentially lead to longer remediation times.     

Natural remobilization of multicomponent DNAPL pools due to dissolution

Mixtures of dense nonaqueous phase liquids (DNAPLs) trapped in the subsurface can act as long-term sources of contamination by dissolving into flowing groundwater. If the components have different solubilities then dissolution will alter the composition of the remaining DNAPL. We theorized that a multicomponent DNAPL pool may become mobile due to the natural dissolution process. In this study, we focused on two scenarios: (1) a DNAPL losing light component(s), with the potential for downward migration; and (2) a DNAPL losing dense component(s), with the potential for upward migration following transformation into a less dense than water nonaqueous phase liquid (LNAPL). We considered three binary mixtures of common groundwater contaminants: benzene and tetrachloroethylene (PCE), PCE and dichloromethane (DCM), and DCM and toluene. A number of physical properties that control the retention and transport of DNAPL in porous media were measured for the mixtures, namely: density, interfacial tension, effective solubility, and viscosity. All properties except density exhibited nonlinear relationships with changing molar ratio of the DNAPL. To illustrate the potential for natural remobilization, we modelled the following two primary mechanisms: the reduction in pool height as mass is lost by dissolution, and the changes in fluid properties with changing molar ratio of the DNAPL. The first mechanism always reduces the capillary pressure in the pool, while the second mechanism may increase the capillary pressure or alter the direction of the driving force. The difference between the rate of change of each determines whether the potential for remobilization increases or decreases. Static conditions and horizontal layering were assumed along with a one-dimensional, compositional modelling approach. Our results indicated that for initial benzene/PCE ratios greater than 25:75, the change in density was sufficiently faster than the decline in pool height to promote DNAPL breakthrough into the adjacent porous medium. In contrast, there was no potential for natural remobilization of a PCE-DCM mixture, primarily because the densities of the components are not sufficiently different. Dissolution of a DCM-toluene mixture decreased the density, reducing the tendency for downward displacement. However, the ultimate transformation from a DNAPL to an LNAPL may induce upward displacement. These results suggest that at sites with DNAPL pools containing a mix of components of sufficiently different densities and relative solubilities, natural remobilization may be an active mechanism, with implications for site evaluation and remediation.     

Bench-scale visualization of DNAPL remediation processes in analog heterogeneous aquifers: surfactant floods and in situ oxidation using permanganate

We have conducted well-controlled DNAPL remediation experiments within a 2-D, glass-walled, sand-filled chamber using surfactants (Aerosol MA and Tween 80) to increase solubility and an oxidant (permanganate) to chemically degrade the DNAPL. Initial conditions for each remediation experiment were created by injecting DNAPL as a point source at the top of the chamber and allowing the DNAPL to migrate downward through a water-filled, heterogeneous, sand-pack designed to be evocative of a fluvial depositional environment. This migration process resulted in the DNAPL residing as a series of descending pools. Lateral advection across the chamber was used to introduce the remedial fluids. Photographs and digital image analysis illustrate interactions between the introduced fluids and the DNAPL. In the surfactant experiments, we found that DNAPL configured in a series of pools was easily mobilized. Extreme reductions in DNAPL/water interfacial tension occurred when using the Aerosol MA surfactant, resulting in mobilization into low permeability regions and thus confounding the remediation process. More modest reductions in interfacial tension occurred when using the Tween 80 surfactant resulting in modest mobilization. In this experiment, capillary forces remained sufficient to exclude DNAPL migration into low permeability regions allowing the excellent solubilizing properties of the surfactant to recover almost 90% of the DNAPL within 8.6 pore volumes. Injection of a potassium permanganate solution resulted in precipitation of MnO2, a reaction product, creating a low-permeability rind surrounding the DNAPL pools. Formation of this rind hindered contact between the permanganate and the DNAPL, limiting the effectiveness of the remediation. From these experiments, we see the value of performing visualization experiments to evaluate the performance of proposed techniques for DNAPL remediation.     

Hydraulic characterization for steam enhanced remediation conducted in fractured rock

To explore the viability of Steam Enhanced Remediation (SER) in fractured rock a small-scale steam injection and water/vapour extraction pilot study was conducted at the former Loring Air Force Base in northern Maine, USA. A detailed well testing program was undertaken to assist in the design of the injection and extraction well array, and to assess the possibility of off-site heat and contaminant migration. A structurally complex limestone having low matrix porosity and a sparse distribution of fractures underlies the study site. To characterize the groundwater and steam flow pathways, single-well slug tests and more than 100 pulse interference tests were conducted. The results of the well testing indicate that the study site is dominated by steeply dipping bedding plane fractures that are interconnected only between some wells in the injection/extraction array. The SER system was designed to take advantage of interconnected fractures located at depth in the eastern end of the site. An array of 29 wells located in an area of 60 by 40 m was used for steam injection and water/vapour extraction. The migration of heat was monitored in several wells using thermistor arrays having a 1.5 m vertical spacing. Temperature measurements obtained during and after the 3 month steam injection period showed that heat migration generally occurred along those fracture features identified by the pulse interference testing. Based on these results, it is concluded that the pulse interference tests were valuable in assisting with the design of the injection/extraction well geometry and in predicting the migration pathways of the hot water associated with the steam injection. The pulse interference test method should also prove useful in support of any other remedial method dependant on the fracture network for delivery of remedial fluid or extraction of contaminants.     

A multi-objective optimization framework for surfactant-enhanced remediation of DNAPL contaminations

The occurrence of Dense Non-Aqueous Phase Liquid (DNAPL) contaminations in the subsurface is a threat for drinkwater resources in the western world. Surfactant-Enhanced Aquifer Remediation (SEAR) is widely considered as one of the most promising techniques to remediate DNAPL contaminations in-situ, be it with considerable additional costs compared to classical pump-and-treat remediations. A cost-effective design of the remediation set-up is therefore essential. In this work, a pilot SEAR test is executed at a DNAPL contaminated site in Belgium in order to collect data for the calibration of a multi-phase multi-component model. The calibrated model is used to assess a series of scenario-analyses for the full-scale remediation of the site. The remediation variables that were varied were the injection and extraction rate, the injection and extraction duration, and the surfactant injection concentrations. A constrained multi-objective optimization of the model was applied to obtain a Pareto set of optimal remediation strategies with different weights for the two objectives of the remediation: (i) the maximal removal of DNAPL and (ii) a total minimal cost. These Pareto curves can help decision makers to select an optimal remediation strategy in terms of cost and remediation efficiency. The Pareto front shows a considerable trade-off between the total remediation cost and the removed DNAPL mass.     

Laboratory investigation of flux reduction from dense non-aqueous phase liquid (DNAPL) partial source zone remediation by enhanced dissolution

This study investigated the benefits of partial removal of dense nonaqueous phase liquid (DNAPL) source zones using enhanced dissolution in eight laboratory scale experiments. The benefits were assessed by characterizing the relationship between reductions in DNAPL mass and the corresponding reduction in contaminant mass flux. Four flushing agents were evaluated in eight controlled laboratory experiments to examine the effects of displacement fluid property contrasts and associated override and underride on contaminant flux reduction (R(j)) vs. mass reduction (R(m)) relationships (R(j)(R(m))): 1) 50% ethanol/50% water (less dense than water), 2) 40% ethyl-lactate/60% water (more dense than water), 3) 18% ethanol/26% ethyl-lactate/56% water (neutrally buoyant), and 4) 2% Tween-80 surfactant (also neutrally buoyant). For each DNAPL architecture evaluated, replicate experiments were conducted where source zone dissolution was conducted with a single flushing event to remove most of the DNAPL from the system, and with multiple shorter-duration floods to determine the path of the R(j)(R(m)) relationship. All of the single-flushing experiments exhibited similar R(j)(R(m)) relationships indicating that override and underride effects associated with cosolvents did not significantly affect the remediation performance of the agents. The R(j)(R(m)) relationship of the multiple injection experiments for the cosolvents with a density contrast with water tended to be less desirable in the sense that there was less R(j) for a given R(m). UTCHEM simulations supported the observations from the laboratory experiments and demonstrated the capability of this model to predict R(j)(R(m)) relationships for non-uniformly distributed NAPL sources.     

Investigation of polychlorinated biphenyl removal from contaminated soil using microwave-generated steam

A feasibility study of polychlorinated biphenyl (PCB) removal from contaminated soils using microwave-generated steam (MGS) was performed. Initial experimental results show that MGS effectively removed PCBs from contaminated soil with an overall removal efficiency of greater than 98% at a steam-to-soil mass ratio of 3:1. Removal efficiency was found to be dependent upon the amount of steam employed, expressed as a mass ratio of steam applied to soil mass. Evaporation was identified as a major mechanism in removing PCBs from the soil. Rapid expansion and evaporation of pore water by microwave dielectric heating accelerated evaporation rates of PCB molecules. Increased solubility of PCBs into the heated aqueous phase is also hypothesized. Together these effects increase mass-transfer rates, thus enhancing removal of PCBs from the soil.     

Assessing impacts of partial mass depletion in DNAPL source zones: II. Coupling source strength functions to plume evolution

Analytical solutions, describing the time-dependent DNAPL source-zone mass and contaminant discharge rate, derived previously in Part I [Falta, R.W., Rao, P.S., Basu, N., this issue. Assessing the impacts of partial mass depletion in DNAPL source zones: I. Analytical modeling of source strength functions and plume response. J. Contam. Hydrol.] are used as a flux-boundary condition in a semi-analytical contaminant transport model. These analytical solutions assume a power relationship between the flow-averaged source concentration, and the source DNAPL mass; the empirical exponent (gamma) is a function of the flow field heterogeneity, DNAPL architecture, and the correlation between them. The DNAPL source strength terms can account for partial source remediation, either at time zero, or at some later time after the DNAPL release. The transport model considers advection, retardation, three-dimensional dispersion, and sequential first-order decay/production of several species. A separate solution is used to compute the time-dependent mass of each contaminant in the plume. A series of examples using different values of gamma shows how the benefits of partial DNAPL source remediation can vary with site conditions. In general, when gamma>1, relatively large short-term reductions in the plume concentrations and mass occur, but the source longevity is not strongly affected. Conversely, when gamma<1, the short-term reductions in the plume concentrations and mass are smaller, but the source longevity can be greatly reduced. In either case, the source remediation effort is much more effective if it is undertaken at an early time, before much contaminant mass has entered the plume. If the remediation effort is significantly delayed, the leading parts of the plume are not affected by the source remediation, and additional control or remediation of the plume itself is required.     

DNAPL distribution in the source zone: effect of soil structure and uncertainty reduction with increased sampling density

This paper focuses on parameters describing the distribution of dense nonaqueous phase liquid (DNAPL) contaminants and investigates the variability of these parameters that results from soil heterogeneity. In addition, it quantifies the uncertainty reduction that can be achieved with increased density of soil sampling. Numerical simulations of DNAPL releases were performed using stochastic realizations of hydraulic conductivity fields generated with the same geostatistical parameters and conditioning data at two sampling densities, thus generating two simulation ensembles of low and high density (three-fold increase) of soil sampling. The results showed that DNAPL plumes in aquifers identical in a statistical sense exhibit qualitatively different patterns, ranging from compact to finger-like. The corresponding quantitative differences were expressed by defining several alternative measures that describe the DNAPL plume and computing these measures for each simulation of the two ensembles. The uncertainty in the plume features under study was affected to different degrees by the variability of the soil, with coefficients of variation ranging from about 20% to 90%, for the low-density sampling. Meanwhile, the increased soil sampling frequency resulted in reductions of uncertainty varying from 7% to 69%, for low- and high-uncertainty variables, respectively. In view of the varying uncertainty in the characteristics of a DNAPL plume, remedial designs that require estimates of the less uncertain features of the plume may be preferred over others that need a more detailed characterization of the source zone architecture.     

Characterization of DNAPL from the U.S. DOE Savannah River Site

The composition of chlorinated hydrocarbon DNAPLs (dense non-aqueous phase liquids) from field sites can be substantially different than the material originally purchased for use as a solvent. Waste management practices at the U.S. Department of Energy's (DOE) Savannah River Site (SRS) included co-disposal of a wide range of organic and inorganic wastes. In 1991, a clear, orange-colored DNAPL was found in two wells near the SRS M-area settling basin. Waste effluent from the fuel and target fabrication facilities that were discharged to this settling basin included acids, caustics, metals and chlorinated solvents. The characterization of the SRS DNAPL suggests that numerous constituents partitioned into the DNAPL during its use as a solvent, co-disposal and ultimate migration through the subsurface. Trace constituents in the DNAPL include metals, from processing operations or co-disposal practices and subsurface minerals, high molecular weight hydrocarbons and alkyl esters, and acids. This complex mixture results in DNAPL-water interfacial properties that are substantially different than would be expected from a simple mixture of PCE and TCE. Under conditions when there is a high DNAPL to water volume ratio, a semi-rigid film accumulates on water droplets suspended in the DNAPL. It is concluded that the array of precipitated metal species comprising this film contributes to the interfacial tension that is over an order of magnitude lower than expected for a "clean" PCE/TCE mixture.