Pore-Scale Analysis of DNAPL Dissolution and Biomass Distribution

Abstract

A comparison of equilibrium and non-equilibrium dissolution of tetrachloroethylene (PCE) was conducted to ascertain how PCE saturation, individual blob properties (volume, surface area, sphericity), and PCE occupied pores are affected by two distinct dissolution regimes. One-dimensional columns were imaged at various dissolution stages using high resolution (~10 um) synchrotron x-ray tomography (XT) and image subvolumes were analyzed using a series of grain, pore network structure, and blob analysis algorithms. An analysis of algorithm-generated data was conducted to determine grain and pore statistics, PCE saturation and individual blob properties, and correlations between PCE blobs and pore network structure. Grain and pore data demonstrated an accurate and consistent segmentation of grains and pores across experiments and a consistent packing between columns. PCE removal rates with pore volumes flushed in both equilibrium and nonequilibrium experiments were consistent until arrival of the primary dissolution front in equilibrium columns. Arrival of the primary dissolution front and number of pore volumes required to completely remove PCE from equilibrium experiments matched well with theoretical predictions. Nonequilibrium dissolution rates varied during the course of the experiment, with increased dissolution observed near the conclusion of the experiments. Blob properties within the equilibrium columns remained relatively constant for all dissolution steps prior to the arrival of the primary dissolution front. Changes in pore-level blob properties in the nonequilibrium experiments were correlated to small perturbations in the mass transfer rates. Deviations in mass transfer rates within the subvolumes occurred over relatively short timescales and are most likely due to the inability to image and analyze a representative elementary volume (REV). XT was also used to investigate the feasibility of imaging biomass within a porous media system. Lugol’s iodine was used to dope the biomass and mass attenuation histograms were compared to those of an undoped biomass system and an abiotic system imaged with and without Lugol’s iodine filling the pore space. After pre-processing with an anisotropic diffusion program, the biomass could be identified within void space of the biotic columns. This insight will aid in the development of XT in exploring the effects of biomass on pore-scale aqueous flow paths.