Comparison of thermal and chemical enhanced recovery of DNAPL in saturated porous media: 2D tank pumping experiments and two-phase flow modelling

Abstract

Pumping experiments were performed in a 2D tank in order to estimate the recovery yield of pure heavy chlorinated organic compounds (DNAPL; dense non-aqueous phase liquids) by varying different parameters: permeability of the saturated zone, pumping flow rates, addition of surfactant and heating. Surfactant was added to decrease capillary forces involved in the entrapment of DNAPL in porous media while temperature was increased to reduce DNAPL viscosity (and hence increase its mobility). Chemical enhancement was performed with the addition of Sodium Dodecyl Benzene Sulfonate (SDBS) (at its Critical Micelle Concentration, to avoid DNAPL dissolution) and thermal enhancement was performed at 50 °C (to avoid DNAPL volatilization). The experiments were monitored with photography allowing, on the basis of image interpretation, to convert optical densities (OD) into water saturations (Sw). Image interpretations were compared with modelling results. The two-phase flow modelling was performed with the pressure-pressure formulation using capillary pressure and relative permeability functions based on the van Genuchten–Mualem equations. Measured volumes of DNAPL recovered as well as the displacement of the DNAPL-water interface (radius and height of the cone of depression) are consistent with the modelling results. Furthermore, chemical enhancement results in a significant increase in the recovery rates of DNAPL. The observed improvement in the recovery of DNAPL with chemical enhancement is due to the fact that: (i) the residual saturation inside the cone of depression is lower and (ii) the cone of depression radius and height increase. Thermal enhancement had no beneficial effect on DNAPL recovery rate or yield. This study shows that it is possible to accurately determine water and DNAPL saturations by image interpretation during pumping tests in a 2D tank in the laboratory. For field-scale applications, the two-phase flow model allows to determine remediation yields as well as the volumes of the cone of depression according to the different operating conditions.