Sorbed-Phase Remediation under Diffusion-Limited Conditions: The Role of Equilibrium Driving Forces and Initial Conditions

For sorbed organic contaminants, slow diffusion from regions that are impermeable to fluid convection can be a principal limitation to contaminant "availability" during remediation, whether by pump-and-treat or transformation mechanisms, and whether under natural or "enhanced" conditions. Moreover, and as highlighted in the summary description of this AGU session, "efforts to account for diffusion limitations in the field are fraught with error because of physical and chemical heterogeneities, and nonlinear responses of diffusion to system perturbations." In particular, the definitions of rate parameters (coefficients) will depend strongly on the rate mechanisms that are assumed, and numerical estimates of their values will be dependent not only upon the heterogeneous chemistry and physics of the systems studied, but also upon assumptions made in regard to (1) the equilibrium thermodynamics (driving force for mass transfer), (2) the initial conditions of the system (e.g., at the on-set of a desorption study), and (3) the details of the experimental boundary conditions over the course of the observations. In this presentation, we will review some of our own experimental and modeling efforts that have aimed to quantify and "account" for diffusion effects in comparably well-characterized systems for which we have attempted to independently understand the above factors. Examples include organic chemical sorption and desorption in batch systems (finite bath uptake and infinite bath release from aquifer sands and surface soils), column systems (column studies involving macropore transport through sorbing solids of low-permeability), and a field scenario (diffusive contaminant release from a fine-grained aquitard into an overlying aquifer). Our emphasis will be on: (1) illustrating how issues of sorption nonlinearity, spatial heterogeneity, and sorption non-equilibrium (at the on-set of desorption) can interplay to cause very complex desorption/remediation behavior; and (2) demonstrating that such behavior is very difficult to interpret (much less predict) if the complicating factors are not properly accounted through independent experimental investigation. In light of our results, we pose questions regarding the extent to which "interpretive" modeling of phenomenological results is useful and to which prediction is possible. Although we doubt that geologic heterogeneities can ever be characterized to a level that allows mechanistic (or predictive) modeling of field-scale behavior, at least a qualitative understanding of the conceptual issues is nonetheless important in remediation design and risk assessment. For any "new" site of soil or sediment contamination, understanding of organic contaminant fate will always be incomplete without careful evaluation of the heterogeneity of the solid matrix and (we suggest) specific study of sorption on selected materials through the use of "probe" chemicals. Although the selection of the most important solids for study will always be a major question, continued basic research with selected materials can help guide the selection.