PROPERTIES AND PROCESSES
The physical and chemical properties of chlorinated solvents govern their behavior in the environment. Important fluid properties include density, viscosity, solubility, and volatility. Important properties governing fluid-fluid and fluid-solid interactions include interfacial tension, wettability, capillary pressure, residual saturation, and relative permeability. Comprehensive descriptions of these properties and the physics of DNAPL movement are in Mercer and Cohen (1990), Cohen and Mercer (1993), and Pankow and Cherry (1996). The effects of these properties on the occurrence, fate, and transport of chlorinated DNAPL are summarized in this section. As the scope of this report is limited to chlorinated solvents, within the context of this report, the term DNAPL is used to describe the immiscible or nonaqueous phase of chlorinated solvents. As such, discussions of DNAPL within this report only apply to liquids with high density and low viscosity. Liquids with high density and high viscosity such as creosote and PCB oils are also DNAPL's, but because they are much less mobile in the subsurface, they behave differently and are not addressed in this report.
Pankow and Cherry (1988, 1996) recognized that chlorinated solvents have physical, chemical, and biological properties that make this class of compounds particularly likely to cause ground-water contamination (table 1). Properties for selected chlorinated solvents are listed in table 2.
The high densities and low viscosities of chlorinated solvents relative to water allow them to move readily downward as DNAPL through the subsurface under the influence of gravity. The exact pathway of downward migration is influenced by such factors as interfacial tension, capillary pressure, hydraulic gradients, structural controls, and the type and nature of openings in an aquifer.
Liquid interfacial tension between DNAPL and water develops because of the difference between the greater mutual attraction or cohesive forces of like molecules within each fluid and the lesser attraction of dissimilar molecules across the immiscible fluid interface (Cohen and Mercer, 1993). Completely miscible liquids have an interfacial tension of zero.
Wettability refers to the preferential spreading of one fluid over solid surfaces in a two-fluid system and depends on interfacial tension. The interaction of adhesive forces between the two fluids and the solid surface and cohesive forces within the fluids will usually result in one fluid having a greater affinity for the solid. The fluid with the greater affinity for the solid surface is the wetting fluid. In a DNAPL/water system, water is usually the wetting fluid and will preferentially coat the aquifer solids (fig. 2). Wettability is controlled by the properties of the two fluids and the composition of the solid surface.
The difference between the wetting fluid pressure and nonwetting fluid pressure is the capillary pressure. Capillary pressure depends on interfacial tension, wettability, and pore size or fracture aperture. Effects of capillary pressure explain much of the distribution and movement of subsurface DNAPL (Cohen and Mercer, 1993). Because the adhesive forces between the wetting fluid and the solid surface are greater than the adhesive forces between the wetting and nonwetting fluid, the resulting capillary pressure draws the wetting fluid into smaller openings in porous media and smaller aperture spaces in fractures. The nonwetting fluid is repelled from the smaller spaces. The pressure head in the nonwetting fluid must exceed the capillary force to displace the wetting fluid and enter an opening.
Capillary pressure increases as the size of the opening decreases. Fine-grained layers with small pore openings and fractures with small apertures have higher capillary pressures and can restrict DNAPL movement. The entry pressure for a nonwetting DNAPL to penetrate a layer in porous media is equal to the capillary pressure that must be overcome to enter the pores of the layer. Similarly, the entry pressure for a DNAPL to penetrate a fracture is equal to the capillary pressure that must be overcome to enter the largest aperture of the fracture. This entry pressure can be directly related to the height or thickness of DNAPL which must accumulate above a layer or fracture to penetrate the capillary barrier. Therefore, macropores, large-aperture fractures, and coarse-grained layers with relatively large openings are preferential pathways for DNAPL movement.
In the vadose zone, capillary phenomena are the dominant mechanisms controlling the movement and distribution of DNAPL. In this situation, DNAPL can be either the wetting or nonwetting fluid, depending on the moisture content of the media. In dry media where DNAPL and air occupy the pore spaces, the DNAPL is usually the wetting fluid and will coat the geologic media and preferentially occupy the smaller pore spaces. Where DNAPL is the wetting fluid, capillary forces enhance DNAPL entry into fine-grained media.
More typically, the vadose zone is partly saturated or water-wet, resulting in DNAPL, water, and air occupying the pore spaces. Water will usually be the wetting fluid with respect to DNAPL and will coat the geologic media and preferentially occupy the smaller pore spaces (fig. 2). DNAPL will be wetting with respect to air and will coat the water, situating itself between the water and air phases. In this situation, capillary forces will act as a barrier to DNAPL entry into fine-grained media. When a dry medium with DNAPL as the wetting fluid is invaded by water, DNAPL is displaced from the solid surfaces by the water.
After a DNAPL mass has moved through the vadose zone, a part of the DNAPL will be retained by capillary forces as residual DNAPL. Values of residual saturation of chlorinated solvents in the vadose zone typically range from 0.01 to 0.10 for dry sands and 0.02 to 0.20 for moist sands (Schwille, 1988; Poulsen and Kueper, 1992; and Cohen and Mercer, 1993). DNAPL may be retained as films, wetting pendular rings, wedges surrounding aqueous pendular rings, and as nonwetting blobs in pore throats and bodies (Cohen and Mercer, 1993).
Two processes in the vadose zone work to deplete a DNAPL mass: volatilization into the air phase and dissolution into water. Direct volatilization of DNAPL into the soil gas is generally the most significant mechanism for depletion of chlorinated DNAPL's from the vadose zone (Pankow and Cherry, 1996). The higher the vapor pressure of a compound, the more readily it will volatilize (table 2). Additionally, DNAPL that dissolves into water in the vadose zone would also be available to volatilize into the soil gas or sorb to solid surfaces. Subsequently, sorbed contaminant may be remobilized through volatilization or dissolution.
Vapor-phase contamination, whether from direct volatilization or dissolution and then volatilization, is a source for a dissolved-phase plume in the ground water, either from dissolution into infiltrating recharge water or diffusion at the water-table surface. The vapor-phase contamination will move by diffusion and sink by density-driven advection. The higher the relative vapor density, the greater the tendency for the vapor-phase contamination to sink (table 2) . These processes will spread the source for the dissolved ground-water plume over a larger area. Diffusive loss of vapors to the atmosphere can occur, but will be limited if the ground surface is covered with vegetation or finer-grained layers which will restrict vapor movement (Pankow and Cherry, 1996).
Below the water table, where DNAPL and water occupy pore space, DNAPL is usually the nonwetting fluid and must overcome capillary forces to enter the smaller pore spaces occupied by the water. DNAPL will continue to move downward under the force of gravity until a finer-grained layer presents a capillary barrier. The DNAPL will then be diverted laterally, seeking a path downward, or will pool at the barrier until significant pressure builds to penetrate the capillary barrier.
DNAPL pools in porous media typically are wide and shallow. Compared with residual DNAPL, pools of DNAPL have less surface area per volume in contact with ground water. DNAPL pools can persist for long periods of time. The rate of dissolution from a DNAPL pool is controlled by the vertical dispersion and subsequent removal of the dissolved phase by the moving ground water. Johnson and Pankow (1992) concluded that because the vertical mixing process is quite weak, the lifetime of chlorinated DNAPL pools will typically be on the order of decades to centuries.
As in the vadose zone, the trailing edge of the DNAPL mass will leave residual DNAPL trapped by capillary forces as isolated blobs and ganglia. Values of residual saturation of chlorinated solvents in porous media in the ground-water zone have been measured in the range from 0.15 to 0.40 (Anderson, 1988). Residual saturation values in the ground-water zone are normally greater than the values in the vadose zone because the fluid density ratio (DNAPL to air as compared to DNAPL to water) favors greater drainage in the vadose zone and, as the nonwetting fluid in the ground-water zone, DNAPL is held in the larger pore spaces (Cohen and Mercer, 1993). The dominant natural process to remove residual DNAPL below the water table is dissolution into ground water. The residual DNAPL provides a source for dissolved-phase ground-water plumes.
DNAPL movement in fractures is controlled by the same properties and processes as in granular material. The concepts of wettability and capillary pressure apply to DNAPL movement in a fracture as they do in a granular material. If the capillary pressure at the leading edge of the DNAPL exceeds the entry pressure of a water-saturated fracture, the DNAPL will displace the water and enter the fracture. In most cases, water will be the wetting fluid and will coat the fracture walls, and DNAPL will be the nonwetting fluid and will fill the larger aperture spaces. The minimum pool height required to overcome the entry pressure is proportional to the DNAPL-to-water interfacial tension and inversely proportional to the difference in fluid densities and fracture aperture (Kueper and McWhorter, 1991). Using values of density and interfacial tension typical of chlorinated solvents, fracture apertures on the order of 2 to 100 microns (μ) can be invaded with DNAPL pool heights in the range of 0.15 to 1.0 m (Kueper and McWhorter, 1991; Kueper and others, 1992; Pankow and Cherry, 1996). Fractures in this aperture range have been measured in fractured, unlithified clay deposits (McKay and others, 1993). Dissolution-enlarged fractures in carbonate aquifers have apertures of a scale significantly larger than this, commonly on the order of a millimeter to tens of centimeters. Therefore, the entry pressure for dissolution-enlarged fractures in carbonate rocks will be easily overcome by even a thin pool of DNAPL. In large open fractures [greater than 1 centimeter (cm)], capillary forces will be insignificant and DNAPL will drain freely under the influence of gravity. In fractures filled with residuum, DNAPL entry and movement is controlled by the pore size of the material filling the fracture.
Migrating DNAPL will not uniformly fill a fracture but will preferentially migrate along the larger aperture pathways that present the least capillary resistance. Progressively smaller aperture fractures will be invaded due to the increased fluid pressure at the base of the DNAPL accumulation if the DNAPL extends vertically as a continuous phase (Pankow and Cherry, 1996).
Large vertical accumulation of DNAPL will most likely occur in settings with small fracture apertures and may be less common in settings with dissolution-enlarged fractures. Once a DNAPL has entered a fracture network, it will most likely continue to drain into the network until the DNAPL source is depleted. DNAPL will move into an intersecting fracture if the local capillary pressure at the advancing front exceeds the entry pressure of the intersecting fracture.
DNAPL migration will be predominantly downward due to gravity, but significant lateral flow can occur along horizontal bedding planes or fractures in response to the closing of fractures with depth. Local structure, including degree of fracture interconnection and distribution of fracture apertures within individual fracture planes, will control DNAPL migration (Kueper and others, 1992). The preferential pathways for DNAPL migration will not necessarily be the same as for ground-water flow. Numerical models and laboratory studies (Schwille, 1988; Pruess and Tsang, 1990; Kueper and McWhorter, 1991; Murphy and Thomson, 1993) have indicated that rates of DNAPL movement in single, small-aperture [less than 1 millimeter (mm)], rough-walled fractures range from minutes to hours per 1 m length of fracture (Pankow and Cherry, 1996). If these rates are applied to field conditions, DNAPL could sink through hundreds of meters of fractured rock in a matter of days to weeks, depending on fracture openings and interconnection. Rates in dissolution-enlarged fractures would be even faster.
Once the supply of the DNAPL to a fracture has been depleted, the DNAPL will redistribute itself as residual DNAPL and pools. In laboratory experiments, Schwille (1988) showed increased solvent retention in fractures when the aperture was reduced. Pools are distinct from residual DNAPL in that they are formed when the leading edge of the migrating DNAPL can no longer overcome capillary resistance (Pankow and Cherry, 1996). This situation can occur where a fracture pinches down to a smaller aperture or where the fracture is filled with fine-grained material . Substantial amounts of DNAPL can be retained as pools in fractures. These pools could be remobilized if the balance of forces holding them static changes. In large conduits and dissolution-enlarged fractures where DNAPL would be expected to drain freely, large amounts of DNAPL can pool in depressions, particularly in horizontal to gently dipping fractures or bedding-plane openings. DNAPL pools in fractured media can be a network of small interconnected fractures filled with solvent or DNAPL accumulations in depressions of dissolution-enlarged fractures.
DNAPL pools can persist for long periods of time. Pools that fill interconnected fractures have limited surface area in contact with moving ground water. The rate of dissolution from a DNAPL pool is controlled by vertical dispersion or diffusion (depending on the location of the pool), the solubility of the DNAPL, and subsequent removal of the dissolved phase by the moving ground water (Hunt and others, 1988). Because both dispersion and diffusion are quite slow, the lifetime of chlorinated DNAPL pools will typically be on the order of decades to centuries. DNAPL pools in depressions of dissolution-enlarged fractures present more surface area to moving ground water than do pools filling interconnected fractures, but DNAPL removal is still limited by vertical mixing. Ground-water flow in dissolution-enlarged fractures may be turbulent, enhancing mixing with DNAPL pools and decreasing the DNAPL pool lifetime.
Matrix diffusion is an important mechanism in DNAPL depletion where DNAPL is pooled in interconnected fractures and the matrix porosity is high (as is typical in fractured clays, shales, and some sedimentary rocks) (VanderKwaak and Sudicky, 1996). In matrix diffusion, DNAPL slowly dissolves into the adjacent water wetting the fracture and then diffuses into the porous matrix (fig. 3). Diffusion of DNAPL into water held in matrix porosity decreases the DNAPL mass held in fractures (Parker and others, 1994) and slows the movement of the concentration front in fractured aquifers (Vogel and Giesel, 1989). Pankow and Cherry (1996) show that in media with small fracture apertures (less than 1 mm) and significant matrix porosity (greater than 5 percent) the total void space in the matrix of fractured porous media is commonly orders of magnitude larger than the void space provided by the fracture network, and matrix diffusion can account for the complete disappearance of DNAPL from fractures. They further show that for a clay with matrix porosity of 37 percent and typical fracture apertures of 1 to 100 μ, TCE would disappear into the clay matrix on the order of days to a few years; for sedimentary rocks with matrix porosity of 10 percent and fracture apertures of 10 to 100 μ, years or decades would be required. Given the same matrix porosity, time for complete removal increases as fracture aperture increases due to higher DNAPL volume in relation to surface area. In karst formations with larger dissolution-enlarged fractures (greater than 1 mm), removal of DNAPL by matrix diffusion will be even slower. The dense Paleozoic carbonate rocks that occur in Tennessee have limited matrix porosity, typically less than 3 percent (Brahana and others, 1988). Higher matrix porosity occurs locally in weathered halos surrounding dissolution-enlarged fractures. In Tennessee karst, the regolith, fractured shaley limestones, and residual fill of conduits may contain very small fractures (1 to 100 μ) and have enough matrix porosity so that matrix diffusion can be important.
In formations with large fracture porosities where the matrix/fracture mass-storage capacity ratio is less than one, complete removal of DNAPL cannot occur by diffusion into the matrix alone because the pore-water volume is insufficient (Pankow and Cherry, 1996). Although matrix diffusion may be more effective in removing DNAPL mass in fractures than dissolution into ground water, once the DNAPL mass is depleted, the mass diffused into the matrix will provide a continuing source for a dissolved ground-water plume as it diffuses back out of the matrix.
The properties and processes discussed here control how chlorinated solvents behave in the subsurface. Various processes, such as volatilization, dissolution, and matrix diffusion, serve to deplete the DNAPL through mass transfer to other phases. In most cases, these processes do not remove any of the contaminant mass from the subsurface environment; the ultimate fate of the DNAPL is to be a persistent source of dissolved-phase contamination (fig. 4). As the dissolved-phase contamination migrates, dispersion and sorbtion reduce contaminant concentrations. Under the right conditions, chemical and microbial transformations of chlorinated solvents can result in natural or enhanced attenuation of dissolved-phase plumes. Various reactions including hydrolysis, reductive dechlorination, co-metabolism, and nucleophilic substitution are currently the subject of field research (Vogel and McCarty, 1985; Henry and Grbic-Galic, 1994; Campbell and others, 1997).
| Next | Title Page | Table of Contents | List of Illustrations | List of Tables | Conversion Factors and Vertical Datum | Glossary |