Groundwater contamination by nonaqueous liquids such as organic solvents and petroleum hydrocarbons frequently occurs as a result of surface spills, tank leaks, and improper disposal practices. This first of two papers examines the physics governing the emplacement and movement of a separate phase in porous media, the role of sorption, and the conditions necessary to mobilize a separate phase. The movement of the separate phase is controlled by capillary forces, and ganglia displacement by groundwater is not possible under reasonable hydraulic gradients. In addition, because of mass transfer limitations in liquid phase dissolution, groundwater extraction at contaminated sites is shown to be ineffective in removing the nonaqueous contaminant within a reasonable time frame. Therefore other means of mobilizing the trapped second phase are needed, steam displacement is proposed and steam front propagation through contaminated porous media is evaluated. The results of laboratory experiments supporting some of these analytical results are presented in the second paper (Hunt et al., this issue).

No abstract

The axisymmetric creeping motion of a neutrally buoyant deformable drop flowing through a circular tube is analysed with a boundary integral equation method. The fluids are immiscible, incompressible, and the bulk flow rate is constant. The drop to suspending fluid viscosity ratio is arbitrary and the drop radius varies from 0.5 to 1.15 tube radii. The effects of the capillary number, viscosity ratio, and drop size on the deformation, the drop speed, and the additional pressure loss are examined.Drops with radius ratios less than 0.7 are insensitive to substantial variation in capillary number and viscosity ratio, and computed values of drop speed and extra pressure loss are in excellent agreement with small deformation theories (Hestroni et al. 1970; Hyman & Skalak 1972a). For this drop size range, significant deformation will result only for Ca > 0.25. The onset of a re-entrant cavity is predicted at the trailing end of the drop for Ca ≈ 0.75. Drop speed and meniscus shape become independent of drop size for radius ratios as small as 1.10. The extra pressure loss can be positive or negative depending mainly on the viscosity ratio, however a relatively inviscid drop can cause a positive extra pressure loss when capillary forces are significant. Computed values for extra pressure loss and drop speed are in good agreement with the experimental data of Ho & Leal (1975) for drops of sizes comparable with the tube radius.

An analysis of the movement of nonaqueous liquids such as organic solvents and gasoline in part 1 (Hunt et al., this issue) showed that separate phase liquids are not completely displaced by groundwater flow under typical pumping conditions and that removal of the compounds by dissolution is mass transfer limited. Steam displacement was proposed as a more efficient cleanup strategy for separate phase organic liquids present in porous media. This paper presents the results of a series of laboratory sand column experiments designed to confirm the theoretical analyses presented in part 1 (Hunt et al., this issue). Experiments with trichloroethylene, a benzene‐toluene mixture, and a commercial gasoline, showed that water flow at rates as high as 15 m d−1 could not displace the separate phase liquids when present in a sand matrix in quantities corresponding to a column average saturation of 2.5%. Steam injection, on the other hand, displaced the contaminants as a separate phase just ahead of the steam front, producing a concentrated, small‐volume waste stream. Analysis of the laboratory data consisting of pressure gradients, temperature profiles, and water flow velocities shows that the laboratory results are consistent with the theoretical predictions. Moreover, computations of energy requirements show that steam displacement of separate phase contaminants is economically attractive.

Heat and mass transfer characteristics of a sand-water-steam system heated at the top and cooled at the bottom were studied. It was found that at steady-state conditions the system segregated into three regions. The top region was conduction-dominated with the voids containing a stationary superheated steam. The middle region was convection-dominated, nearly isothermal, and exhibited an upward flow of the liquid by capillary forces and a downward flow of steam due to a slight pressure gradient. The bottom portion contained a stationary compressed liquid and was also conduction dominated. The length of the two-phase convection zone was evaluated through the application of Darcy’s equations for two-phase flow and correlations of relative permeabilities and capillary pressure data. The model was in excellent agreement with the observed results, predicting a decreasing two-phase zone length with increasing heat flux. The thermodynamics of the two-phase zone were also analyzed. It was found that the vapor phase was in a superheated state as described by the Kelvin equation for vapor pressure lowering. Also, it was evident that the liquid must also be superheated for thermodynamic equilibrium to result. A stability analysis demonstrated that the superheated liquid can exist in an unconditionally stable state under conditions typical of porous systems. The degree of liquid superheat within the two-phase zone of these experiments was obtained.