MODELING NONLINEAR SORPTION OF TRICHLOROETHENE IN NATURAL SORBENTS WITH KEROGEN

Sorption of chlorinated solvents affects the distribution of these compounds in the subsurface environment, hindering remediation efforts. The nonlinear behavior observed in sedimentary aquifer materials that contain carbonaceous matter, such as kerogen, adds complexity to sorption modeling because the sorption distribution coefficient (Kd) varies as a function of sorbate concentration. This behavior is difficult to predict and is inadequately simulated with commonly used linear and two-parameter isotherm models. This study addresses these challenges by comparing the performance of a suite of partitioning, adsorption, and dual-mode isotherm models.

Trichloroethene sorption isotherms were measured for four sedimentary sorbents collected near Borden, Ontario and Vestal, New York. These materials were selected as model sorbents because they contain kerogen. Kerogen is an important natural sorbent because it is the most common sedimentary carbonaceous matter on earth and it’s resistance to degradation allows it to be present in autochthonous and allochthonous sedimentary deposits. The measured sorption data were nonlinear across aqueous sorbate concentrations spanning nearly five orders of magnitude, a range typical of chlorinated solvent contamination sites.

The isotherms were fit with 11 sorption models including linear, Freundlich, Langmuir, and Polanyi-type models. The Polanyi-type models were ranked as the most plausible for each isotherm based on the corrected Akaike Information Criterion. The estimated linear partitioning coefficients for the Polanyi-partition model were within a factor of four of that predicted on the basis of the fraction of organic carbon and an octanol-water partitioning linear free energy relationship. The estimated sorption capacity terms were constant within a factor of three when normalized by the fraction of organic carbon. The contribution of the A and B fitting parameters showed similarity among the studied sorbents. These results suggest that the Polanyi-type models are the most accurate for simulating nonlinear sorption in kerogen across broad sorbate concentrations. Furthermore, it may be feasible to predict Polanyi-type model parameters with reasonable accuracy using measurable sorbent properties, reducing the need for extensive sorption data.