Dissolution of nonaqueous phase liquids in saturated subsurface systems.

Abstract

Nonaqueous phase liquids (NAPLs) entrapped in the subsurface comprise long-term sources of groundwater contamination as they slowly dissolve. The objective of this research work is to gain insight into phenomena controlling NAPL-water interphase mass transfer processes. Experimental efforts focused on one-dimensional column experiments to measure both steady-state (initial) and time-dependent NAPL dissolution rates. The use of styrene as a primary NAPL enabled the in-situ polymerization of NAPL blobs and subsequent characterization of the entrapped NAPL. Experimental measurements indicate that effluent concentrations are less than values predicted by models based on the local equilibrium assumption for high aqueous phase velocities, characteristic of pump-and-treat remediation, and at low NAPL saturations. Dissolution rates are shown to depend on the pore-structure characteristics of a porous medium. Differences in entrapped NAPL blob size and shape distributions explain trends in effluent concentrations. Larger blobs trapped in coarse or well-graded media have less specific surface area, and correspondingly slower mass transfer rates. Phenomenological models are developed from steady-state and transient dissolution data. Correlations are proposed to quantify a modified Sherwood number in terms of several variables to describe the dependence of mass transfer rates on aqueous phase velocity and NAPL-water specific surface area. Verification with independent data sets indicates that the model based on Reynolds number, normalized median grain size and uniformity index provides the best predictive capability for steady-state dissolution rates. Two conceptual models describing the reduction in mass transfer rates as surface area is reduced are considered. The theta model assumes that the NAPL exists in a lumped domain; the sphere model describes blob distributions as a set of spherical globules with a range of diameters. Both conceptual models are better able to simulate styrene dissolution data and predict TCE dissolution than others published in the literature. Numerical simulations explore the potential sensitivity of NAPL dissolution processes to blob size distributions at the field-scale. Field-scale data are inadequate, however, to permit assessment of the appropriateness of the local equilibrium assumption at this scale.