Conducted in conjunction with an Environmental Protection Agency (EPA)-funded project, the overall objective of this ESTCP project was to demonstrate that electrical resistivity imaging (ERI) could be used to monitor and facilitate an in situ chemical oxidation (ISCO) demonstration designed to treat RDX with permanganate at the Nebraska Ordnance Plant.

ERI is a geophysical technique that rapidly and economically collects data using an array of electrodes on the surface and then creates a model and an image of subsurface electrical resistivity. In appropriate environments and with appropriate materials, ERI shows the distribution of materials of interest, either contaminants, amendments, or both. With a good ERI image, investigators can drill for specific, observed, subsurface targets.

The pilot-scale demonstration consisted of a grid of 12 wells, specifically an extraction well (EW-1), two injection wells (IW-1 and IW-2), and nine monitoring wells. To create a permanganate curtain across the injection wells, sodium permanganate (NaMnO₄) was injected into the field via a proportional mixing-injection trailer system. Groundwater was extracted from a center extraction well via a submersible pump at a rate of 151.6 L/min (40 gpm) and delivered to an intake manifold located onboard the trailer system. Approximately 1707.2 L (451 gal) of 40% (w/w) sodium permanganate, spiked with potassium bromide, was pumped at 3.79 L/min (1 gpm) to an intake manifold where extracted groundwater and sodium permanganate were mixed at a ratio of 40:1. The mixed eluent was then gravity fed into each of two neighboring injection wells, IW-1 and IW-2, at approximately 77.7 L/min (20.5 gpm).

The site was monitored with ERI before, during, and after the injection. The ERI applied at the site uses arrays of easily placed, noninvasive surface electrodes to measure apparent subsurface electrical resistivity. This data was then processed and correlated to subsurface properties that caused changes in the electrical properties of the subsurface such as grain size or formation fluids.

Results showed that permanganate was effective in reducing groundwater RDX concentrations under pilot-scale conditions and that ERI was successful in imaging the initial size and distribution of the injected permanganate plume. ERI also quantitatively mapped the hydraulic conductivity distribution. RDX concentrations temporally decreased in wells closest to the injection wells as the permanganate migrated down gradient. RDX degradation rates of 0.12 1/d and 0.087 1/d were observed in wells immediately downgradient of the injection wells. These rates were lower than what was observed under laboratory batch conditions at 11.5ºC (0.20 1/d) and likely a result of a lower initial permanganate concentration (6,000 versus 15,000 mg/L). RDX concentrations decreased between 70 and 80% in monitoring wells immediately downgradient from the injection wells. Monitoring wells farther down gradient did not show a true breakthrough or significant changes in RDX concentrations.

Results from spatial monitoring of permanganate verified that monitoring wells only captured fingers of permanganate and that the groundwater sampling procedures likely mixed treated with non-treated groundwater during pumping. This mixing of permanganate-treated and untreated groundwater would explain why initial permanganate concentrations were less than target values. Moreover, the observed decreases in RDX concentrations during permanganate breakthrough (73-80%) may have been higher if researchers could have selectively sampled only the permanganate-treated groundwater. Despite problems encountered in getting the permanganate uniformly distributed across the injection well and throughout the well screen interval, pilot-scale results provide proof-of-concept that permanganate can degrade RDX in situ and support permanganate as a possible remedial treatment for RDX-contaminated groundwater.

RDX concentrations did not decline below the EPA health advisory for drinking water nor did the permanganate injection influence all the anticipated wells downgradient from the injection zone. The ERI data sets provided an explanation for this failure. When the permanganate was injected, small differences in the ability of the material to conduct water forced the injectate to flow upgradient away from the monitoring wells. Furthermore, the ERI data showed that the permanganate solution descended below the monitoring wells.

The ERI demonstration successfully observed the initial permanganate flow, although with unexpected time and monetary expense. The demonstration did not evolve as planned and anticipated: locating and tracking the permanganate required an additional data collection event and additional processing. ERI protocols can be altered to more economically and efficiently respond to unanticipated results.

ERI geophysical techniques provide the advantages of economically acquiring large, spatially extensive data sets to track the distribution and flow of injectate. Site assessment following groundwater remediation efforts typically involves discrete point sampling using wells or multilevel piezometers along anticipated flowpaths. Small variations in hydraulic conductivity can divert groundwater flow away from anticipated flowpaths, frustrating efforts to monitor remediation efforts with pre-placed wells. Without a dense network of multilevel piezometers throughout and surrounding the area of interest, point sampling cannot reliably determine the spatial distribution of contaminant nor the flow of injectate.