Sorption processes have been largely studied and modelled by considering single mineral surfaces (iron oxides, clay etc). In the case of a complex surface such as claystone, it seems important to be able to predict the behaviour of the radioactive elements in contact with the natural site. The modelling part becomes very complex if the whole minerals composing the geological media have to be taken into account. In that way, this study aims at testing various mineral mixtures, to identify the most reactive phase controlling the adsorption processes, in order to get a “simplified” system for a model calculation. Various mineral assemblages (goethite, pyrite, calcite and montmorillonite) were studied here, in contact with an anionic specie (selenium). Experiments were carried out using single, binary, ternary or quaternary solid mixtures, in order to evaluate the major reactive mineral phase. A surface complexation model was then considered to fit the experimental data. Stoichiometries and constants were extracted using FITEQL by fitting experimental results obtained on single solid in NaNO3, and used in a predictive calculation with PHREEQC for mixtures in synthetic grounwater. By doing this, we wanted to verify if the assumption that surface reactivity of single minerals keeps the same in a different environment can be used for predictive calculations.

Results obtained have shown that the presence of calcite and pyrite does not influence the total amount of selenium adsorbed. Then a simplified model was proposed, taking into account only the major phase responsible for the sorption of Se.