Cs sorption to potential host rock of low-level radioactive waste repository in Taiwan: Experiments and numerical fitting study

Highlights: → We demonstrated experiments and numerical fittings of Cs sorption to argillite step-by-step in this study. → Both clay model and iron oxide model were used and their influences on the goodness of fitting results were reported. → The number of sorption sites dominated the goodness of fitting results. → Tracer experiments could be replaced in light of reducing the amount of radioactive waste produced. - Abstract: A reliable performance assessment of radioactive waste repository depends on better knowledge of interactions between nuclides and geological substances. Numerical fitting of acquired experimental results by the surface complexation model enables us to interpret sorption behavior at molecular scale and thus to build a solid basis for simulation study. A lack of consensus on a standard set of assessment criteria (such as determination of sorption site concentration, reaction formula) during numerical fitting, on the other hand, makes lower case comparison between various studies difficult. In this study we explored the sorption of cesium to argillite by conducting experiments under different pH and solid/liquid ratio (s/l) with two specific initial Cs concentrations (100 mg/L, 7.5 x 10-4 mol/L and 0.01 mg/L, 7.5 x 10-8 mol/L). After this, numerical fitting was performed, focusing on assessment criteria and their consequences. It was found that both ion exchange and electrostatic interactions governed Cs sorption on argillite. At higher initial Cs concentration the Cs sorption showed an increasing dependence on pH as the solid/liquid ratio was lowered. In contrast at trace Cs levels, the Cs sorption was neither s/l dependent nor pH sensitive. It is therefore proposed that ion exchange mechanism dominates Cs sorption when the concentration of surface sorption site exceeds that of Cs, whereas surface complexation is attributed to Cs uptake under alkaline environments. Numerical fitting was conducted using two different strategies to determine concentration of surface sorption sites: the clay model (based on the cation exchange capacity plus surface titration results) and the iron oxide model (where the concentration of sorption sites is proportional to the surface area of argillite). It was found that the clay model led to better fitting than the iron oxide model, which is attributed to more amenable sorption sites (two specific sorption sites along with larger site density) when using clay model. Moreover, increasing s/l ratio would produce more sorption sites, which helps to suppress the impact of heterogeneous surface on Cs sorption behavior under high pH environments.