Contaminant Mass–Removal Behavior during Chemical Oxidation in a Three-Dimensional, Fractured Sandstone Network Experiment

Abstract

A three-dimensional (3D) bench-scale, low-porosity fractured-sandstone network experiment was used to evaluate the behavior and effectiveness of in situ chemical oxidation (ISCO) using potassium permanganate (KMnO4) to deplete tetrachloroethylene (PCE) dense nonaqueous phase liquid (DNAPL). The purpose of the 3D fracture experiment was to compare mass removal during steady state DNAPL dissolution conditions to ISCO, and to determine how the presence of fracture intersections affected the formation of reaction byproducts in the experimental 3D fracture network. DNAPL dissolution in the fracture network was evaluated and described using an effective parameter: the bulk mass transfer coefficient (KL). Residual DNAPL saturation and subsequent steady state DNAPL dissolution conditions were established during ambient groundwater flow through the experimental network, and then ISCO was applied. PCE mass removal rates during steady state DNAPL dissolution conditions were compared to mass removal rates achieved during ISCO experiments. The effective DNAPL-water interfacial area (ai) was characterized and evaluated using conservative and interfacial area tracer tests. The ai was used to compare availability of DNAPL-water contact during dissolution studies as well as pre- and post-ISCO applications. The results of this research indicate that, in the experimental 3D fracture network, mass removal rates during steady state DNAPL dissolution conditions were generally greater than mass removal rates during ISCO experiments over the same timeframe. Research data indicate that, similar to what has been observed in fracture plane studies, the formation of reaction products [manganese dioxide (MnO2) and carbon dioxide (CO2)] limited contact of the oxidant with the network DNAPL, and thus limited mass removal during ISCO. The reduction in DNAPL-water contact due to reaction byproduct formation was supported by the lack of quantifiable ai following ISCO. The reaction byproducts are believed to have altered the flow paths during ISCO, which limited the DNAPL-oxidant contact and reduced the PCE mass removal rates. Experimental results suggest that stopping the application of ISCO when PCE mass removal is no longer effective (i.e., when the generated chloride concentration is within 10% of the background chloride concentration), the ISCO will successfully remove more PCE than dissolution alone. Subsequent experiments that evaluate DNAPL entrapment due to reaction byproduct formation, not simply flow path alteration, are needed to better understand if this as a viable option to address DNAPL in fracture network settings.