Change in pore size distribution of compacted soil layers and its effect on solute breakthrough curves

It is often useful to predict contaminant migration from waste containment systems, such as landfills, as part of the assessment of the overall impact of such systems on the receptor environment. In many instances, material properties, for example, those of the liner, are assumed to be constant. This study was conducted to evaluate the accuracy of considering constant material and transport parameters in the modelling of sodium and chloride breakthrough curves through a compacted soil layer using the commercial software, Pollute v.7. Experiments were conducted with three different mixtures of glass beads and varying amounts of kaolinite (30, 40 and 50% by weight). The base line hydraulic conductivity K of the samples was established using distilled water as permeant. The observed values of K were 8.2X10-11 m/s, 1.28X10-10 m/s and 1.48X10-10 m/s for the 30, 40 and 50% kaolinite, respectively. These values did not change when the permeant was changed from distilled water to 0.04 M NaCl Effective diffusion coefficient of 3.5-8.5 x 10-10 m2/s was obtained for sodium and 1.9-4 x 10-10 m2/s for chloride. These results also showed that diffusion of both ions in the soils was affected by the percentage of clay fraction. The greater the amount of clay, the lower the diffusion coefficient obtained. Moreover, the diffusion coefficient of sodium was approximately two times that of chloride and this trend was visually apparent from the shape of the breakthrough curves for Na+ and Cl-. Modelling with constant porosity overestimated the concentration of both ions. The pore size distribution of each mixture was determined from mercury intrusion porosimetry testing before and after hydraulic conductivity test. The results showed a decrease of 24%, 13% and 12% in the porosity of the 30, 40 and 50% kaolinite mixture. Sensitivity analysis carried out by decreasing the porosity of the mixture by these percentages did not alter breakthrough curves noticeably. On the other hand, sensitivity analysis based on changes in the distribution coefficient and diffusion coefficient showed a considerable change in model outputs. It was concluded that although the porosity changed during hydraulic conductivity test, it did not eliminate the discrepancy between experimental results and modelling results, In fact, the model was found to be more sensitive to change in diffusion coefficient and distribution coefficient. Therefore, more studies are required to monitor these parameters during hydraulic conductivity testin