Membrane behavior and coupled solute transport through a geosynthetic clay liner

The ability of a geosynthetic clay liner (GCL) to act as a semi-permeable membrane that restricts migration of solutes, and the influence of this membrane behavior on solute transport through a GCL, are evaluated in this research. A coupled solute transport model that accounts for membrane behavior (i.e., salt-sieving, chemico-osmotic counter flow) in clay soils is developed. In addition, laboratory chemico-osmotic/diffusion and column tests are conducted on specimens of a GCL subjected to potassium chloride (KCl) solutions. Measured chemico-osmotic efficiency coefficients (ω) range from 0.14 to 0.63, and effective salt-diffusion coefficients $D*s$ range from 2.38 × 10−10 m2/s to 7.83 × 10−11 m2/s for 10-mm-thick GCL specimens subjected to KCl source concentrations ranging from 0.0039 M to 0.047 M. The results indicate that two coupling effects limit KCl diffusion through the GCL: (1) an explicit coupling effect due to salt-sieving, as predicted by the coupled flux transport theory; and (2) implicit coupling characterized by a concentration-dependent decrease in $D*s$ that ts related to the apparent tortuosity factor (τ_a). These effects are correlated with measured ω values for the GCL. Comparison of experimental results from column tests with model simulations based on independently measured values of ω and $D*s$ (or τ_a) indicate that implicit coupling is more significant than explicit coupling for reducing solute flux through the GCL under diffusion-dominated conditions.; The results of theoretical simulations indicate that lower values of solute flux and longer solute transit times through a soil membrane are predicted by coupled solute transport theory relative to advective-dispersive theory that neglects chemico-osmotic efficiency. The coupled transport model also provides better agreement with measured data from the column tests than advective-dispersive theory. Use of advective-dispersive theory to simulate solute transport through GCLs appears reasonable under diffusion-dominated conditions, provided that implicit coupling is included by using an appropriate value of $D*s$ (or τ_a) that corresponds to the solute concentration.