Permittivity and electrical resistivity measurements and estimations during the recovery of DNAPL in saturated porous media: 2D tank experiments

International audiencePumping experiments were performed in a 2D tank in order to estimate the recovery yield and rate of pure heavy chlorinated organic compounds (DNAPL; dense non-aqueous phase liquids). Several operating configurations were considered: permeability of the saturated zone, pumping flow rates, addition of surfactant (to reduce capillary effects), and thermal treatment (to reduce DNAPL viscosity). The experiments were monitored with permittivity (ε) and electrical resistivity (ρ) measurements. The experiments were also monitored with photography allowing, based on image interpretation, to accurately convert optical densities into water saturations (Sw) and to validate the geophysical data collected. Average Sw were determined in all detection areas from permittivities and electrical resistivities. These average Sw were used to calculate theoretical permittivities using the complex refractive index model (CRIM) and theoretical electrical resistivities using Archie's law. We found a good correlation between measured permittivities and the estimated permittivities with image. Conversely, the correlation was less accurate between measured and estimated electrical resistivities except for low electrical resistivities (i.e. high Sw). It is therefore not possible to accurately quantify water saturations using electrical resistivity monitoring alone. However, the accuracy is sufficient (especially for high values of Sw) to highlight differences between the three considered treatment technologies (i.e. without enhancement, with chemical and/or thermal enhancements). A combination of monitored electrical resistivity (which provides an integrated vision of the cone of depression) and monitored permittivity (which provides accurate but spatially limited information), and imaging allows a more accurate calibration of the relationship between resistivities and Sw and quantification of remediation rates and yields