Mobilization of residual tetrachloroethylene during alcohol flushing of clay-containing porous media

摘要

Mobilization of residual DNAPL in a contaminated aquifer may be problematic fobr the application of enhanced dissolution remediation techniques such as alcohol or surfactant flushing. The objective of this study was to investigate the effects of clay fines on residual saturation, permeability and the onset of mobilization of PCE during alcohol flushing. Column experiments were conducted with soils containing 0/100, 2/100, 5/100, 10/100, and 20/100 clay that was completely mixed within a well--sorted Ottawa sand. Column characterization included pressure--saturation relation- ships, pore--size distribution, permeability, and residual saturation. The columns were flushed with isopropyl alcohol (IPA) in concentration increments to determine when mobilization of PCE occurred. Trapping numbers were calculated. Results showed that increasing amounts of clay resulted in decreased hydraulic conductivity, increased residual PCE saturation and a larger pore size distribution. Residual PCE was found to begin to mobilize in the clay--mixed media at a range of trapping numbers below those reported in the literature tbr homogeneous well-sorted sand. This suggests that the potential to mobilize the residual PCE is greater in the mixed media. This may be explained by the formation of multi-pore branched ganglia in the clay--containing media as opposed to singlet formation in a well--sorted media. The multi--pore ganglia may mobilize more readily than singlets due to the added force of head acting on the PCE interface where the trapping is occurring. Using a narrow range of trapping numbers determined in homogeneous media to predict mobilization may not be appropriate in mixed media because mobilization was noted at a variety of trapping numbers. However, relationships between the trapping number at mobilization and various soil parameters such as the amount of clay, residual DNAPL saturation, and pore size distribution, indicate that these, coupled with the trapping number, may aid in our ability to predict mobilization.