Modeling of Cs+ diffusion and retention in the DI-A2 experiment (Mont Terri). Uncertainties in sorption and diffusion parameters

In the DI-A2 experiment several non-reactive and reactive tracers were injected as a pulse in a packed-off borehole in the Opalinus Clay. Unlike the previous DI-A1 test, the design of the Teflon filter in the injection borehole forced the water to flow through the filter and the open space between the filter and the borehole wall (the filter itself did not act as a diffusion barrier between the circulating solution and the rock). The decrease in tracer concentration in the liquid phase was monitored during a period of a year. Afterwards, the borehole section was overcored and the tracer profiles in the rock were analyzed. A main interest of this experiment was to understand the chemical behavior of sorbing tracers: Cs⁺ (stable), ⁸⁵Sr²⁺, ⁶⁰Co²⁺ and Eu³⁺ (stable). The complete dataset (except for Eu³⁺ because of strong sorption to experimental equipment) was analyzed in a previous study with a 2D diffusion–reaction model and the derived diffusion and sorption parameters were compared with laboratory data. As in DI-A1, a difference by a factor of about 2 for sorption (magnitude of the Freundlich isotherm) was obtained between in situ and laboratory batch sorption experiments.Recent experimental and modeling studies have shown equivalent Cs⁺ sorption on intact and disaggregated Opalinus Clay samples. In view of these developments, new modeling of Cs⁺ diffusion and retention in the DI-A2 experiment has been performed using CrunchFlow. The calculations include transport by diffusion and a multisite cation exchange model to account for the retention of Cs⁺. The new results show that upscaling of Cs⁺ sorption from laboratory to field is no longer required. However, a difference in sorption by a factor of about 2 is still explained by the use of different versions of the same cation exchange model (a small difference in the selectivity coefficient for one type of site). This uncertainty in sorption leads to an uncertainty in the effective diffusion coefficient (D_e) for Cs⁺, also by a factor of 2 (2–4 × 10^?10 m²/s). Clearly, the values of D_e obtained are correlated with the strength of sorption in the model, with stronger sorption leading to larger D_e values. Discrimination between the two versions of the exchange model is not possible when using only the results of the in situ test. Additionally, during early times (t < 10 days) the drop in Cs⁺ concentration in the circulation system is slower than expected. Due to the experimental setup, this slow decrease in concentration cannot be caused by the filter in the contact between borehole and rock. Poor mixing in the circulation system could explain this effect.