Experimental assessment of multicomponent gas transport flux mechanisms in subsurface systems.

Abstract

Prediction of vapor component migration in subsurface systems requires an evaluation of the relative importance of different modes of gaseous transport. A comprehensive experimental program was undertaken to explore gas transport mechanisms in a variety of soil materials representative of natural systems. These soils included: three dry uniform materials, a sea sand, an Ottawa sand, and Kaolinite clay; five graded mixtures of these uniform soils; and wet sea sand samples. Some of these soils were tested at different bulk densities. The experimental program consisted of two sets of experiments: those conducted to characterize porous media transport parameters and those conducted to examine multicomponent diffusional transport. The first set of experiments encompassed single gas (permeability) and binary gas measurements. Results of these experiments were used to develop correlations for the effective Knudsen radii and the diffusibility factors as functions of medium properties. Experimental measurements demonstrated the significance of non-equimolar fluxes in binary diffusional transport and the potential importance of Knudsen diffusion in the presence of an aqueous phase, which reduced the effective area of gas transport. In the second set of experiments, multicomponent transport of methane and trichloroethylene through air was examined in a specially designed test cell apparatus under isobaric and non isobaric conditions. This diffusion cell represents an open system, in which component gases pass by both edges of the soil sample. Experimental measurements are used to evaluate the applicability of three diffusional transport models: Fick's law, the Stefan-Maxwell equations, and the Dusty Gas Model (DGM). Measured fluxes from the multicomponent experiments were compared to fluxes computed using each model in conjunction with the measured diffusion parameters. Comparisons indicated that for isobaric conditions all models can accurately predict the measured fluxes provided the measured effective diffusion coefficients are used. For the non-isobaric conditions examined, significant slip flux transport was observed and all models underestimated the measured fluxes.