Plume Development and Mass Flux Following Surfactant-Based Treatment of Heterogeneous PCE-DNAPL Source Zones

The zones of contamination at typical chlorinated solvent-contaminated sites can be divided into two regions: a source zone in which free-phase contaminants (dense nonaqueous phase liquids, or DNAPLs) are present, and a solute plume containing only dissolved-phase contaminants. Currently, pump-and-treat is the most common method for achieving dissolved-phase plume containment; however, it is widely recognized that this approach is generally ineffective for source zone mass removal. As a result, a number of innovative technologies for in situ DNAPL source zone treatment have been developed, but mass removal using these methods is often incomplete. In addition, the effects of partial source zone mass removal on subsequent dissolved-phase plume development and contaminant flux remain poorly understood. To address these issues laboratory-scale experiments were conducted in a two dimensional (2-D) aquifer cell having overall dimensions of 150 cm (length) by 48 cm (height) by 1.4 cm (internal thickness) and containing both source zone and down-gradient "plume" regions. The aquifer cells were packed under water-saturated conditions with Accusand (either 20/30 sieve size or a mixture of 50% 20/30 and 50% 40/50 sieve sizes). Within the source zone, three layers of F-70 Ottawa sand lenses were emplaced to mimic heterogeneous regions of lower permeability media. Following tetrachloroethene (PCE) release and redistribution in the source zone, a solubilizing surfactant solution containing 4% Tween 80 was used to achieve sequential PCE mass removals ranging from 30% to 80%. At the conclusion of each surfactant flood, down-gradient contaminant concentrations and mass fluxes were monitored at a hydraulic gradient of 1x10^-3. The PCE-DNAPL distributions in the source zone were quantified using light transmission prior to and following each surfactant flood. PCE-DNAPL distribution was expressed in terms of a ganglia to pool ratio (G:P), for which the volume of PCE above residual saturation (S_r = 11%) was considered to be "pooled". Results from three aquifer cell experiments are reported here; the first two cells contained highly-pooled source zones having an initial G:P values of 0.26:1 (80% pooled) and 0.50:1 (70% pooled), while the third contained a moderately-pooled source zone with a G:P value of 1.6:1 (40% pooled). For the first highly-pooled cell, flux-averaged effluent PCE concentrations cell decreased from 150 mg/L to 70 mg/L after 45% PCE mass removal, with a subsequent reduction to 5 mg/L following 80% PCE mass removal. Similar behavior was observed for the second highly pooled cell, with effluent PCE concentrations decreasing from 100 mg/L to 50 mg/L following 60% PCE mass removal and subsequently to 20 mg/L following 80% PCE mass removal. Effluent PCE concentrations in the moderately pooled aquifer cell were not statistically different from the initial value of 100 mg/L after 40% PCE mass removal, but subsequently decreased to approximately 30 mg/L following 75% PCE mass removal. Differences in post-treatment plume development and mass flux between the experiments were attributed to (a) preferential removal of PCE mass that existed as entrapped ganglia and were readily solubilized during the surfactant flood, and (b) the persistence of DNAPL pools which accounted for 40 to 80% of the initial PCE mass. Results obtained from these studies provide direct experimental evidence of the potential impacts of DNAPL source zone architecture and partial mass removal on plume development and reductions in contaminant mass flux.