Experimental upscaling analyses for a surfactant-enhanced in-situ chemical oxidation (S-ISCO) remediation design

Surfactant-enhanced in-situ chemical oxidation (S-ISCO) is an emerging innovative remediation technology for the treatment of dense non-aqueous phase liquids (DNAPLs). S-ISCO combines the solubilization of contaminants by means of surfactants with the chemical oxidation by an oxidizing agent, thus, potentially increasing the efficiency of the state-of-the-art ISCO technique. Scientific investigations are needed to enable the technology transfer for potential field applications based on the development of a remediation design under well-defined boundary conditions. For this purpose, experimental upscaling analyses were performed using the special infrastructure of the research facility for subsurface remediation (VEGAS). Batch tests showed that oxidation of the selected surfactant E-Mulse 3® (EM3) by activated persulfate (Na-PS) reduced the solubilization of the model contaminants 1,4-DCB, naphthalene, and PCE. As a consequence, the processes of contaminant solubilization and degradation were temporally and spatially separated in the developed remediation design. A proof of concept was provided by performing an S-ISCO medium-scale experiment (100 cm length, 70 cm height, 12.5 cm width), with 1,2-DCB as model DNAPL contaminant to be treated. A groundwater circulation well (GCW) was used to inject a 60 g/L Na-PS solution and to effectively mix the reagents. Sampling of the experiment's outflow and the soil material after treatment showed that neither rebound effects nor residual mass loadings on the soil material could be detected after termination of the S-ISCO treatment. To further evaluate the S-ISCO remediation design under field-like conditions, a large-scale S-ISCO experiment was conducted (6 m length, 3 m height, 1 m width), allowing for an extensive sampling campaign to monitor relevant processes. An efficient contaminant removal from the former source zone could be reached by surfactant solubilization, decreasing contaminant levels from initially over 2000 mg/L 1,2-DCB to final concentrations below 5 mg/L 1,2-DCB. The heterogeneously distributed contaminant degradation, implemented by a three-filter GCW, was attributed to density-induced migration processes that impeded an optimal reaction zone. A density-dependent numerical transport could qualitatively match the observations. By comparing different simulation scenarios, an adapted operation of the GCW was established that provides for a more efficient distribution of the density-influenced oxidant injection.