Coupled THCM model of a heating and hydration concrete-bentonite column test

Abstract Radioactive waste disposal in deep geological repositories in clay formations envisage a compacted bentonite engineered barrier and a concrete liner. The alkaline conditions caused by the degradation of concrete could affect the performance of the engineered barrier. The geochemical interactions occurring at the concrete-bentonite interface (B-CI) for the non-isothermal unsaturated conditions prevailing at repository post-closure have been studied by CIEMAT with a heating and hydration concrete-bentonite column test. The column consists of a 3 cm thick concrete sample emplaced on top of a 7.15 cm block of compacted bentonite. The column was hydrated through the concrete at a constant pressure with a synthetic clay porewater while the bottom of the column was heated at 100 °C. Here we report a coupled thermo-hydro-chemical-mechanical (THCM) model of the column test, which lasted 1610 days. The model was solved with a THCM code, INVERSE-FADES-CORE. Experimental observations show calcite and brucite precipitation in the concrete near the hydration boundary, portlandite dissolution and calcite and ettringite precipitation in the concrete, calcite and sepiolite precipitation in the bentonite near the B-CI, calcite dissolution in the bentonite far from the B-CI and gypsum precipitation in the bentonite near the heater. Model results attest that advection is relevant during the first months of the test. Later, solute diffusion becomes the dominant transport mechanism. Calcite and brucite precipitate in the concrete near the hydration boundary because the concentrations of dissolved bicarbonate and magnesium in the hydration water are larger than the initial concentrations in the concrete porewater. Calcite and brucite precipitate in both sides of the B-CI. Sepiolite precipitates in the bentonite near the B-CI. The model predicts portlandite and C1.8SH dissolution in the concrete. Ettringite and C0.8SH precipitate near the hydration boundary while ettringite dissolves in the rest of the concrete at very small rates. The porosity changes occur at the hydration boundary and at both sides of the B-CI due to mineral dissolution/precipitation. The porosity reduces to zero in a 0.03 cm thick zone in the concrete near the B-CI due to brucite and calcite precipitation. The high pH front (pH > 8.5) diffuses from the concrete into the bentonite and penetrates 1 cm at the end of the test after 1610 days. Model results are sensitive to grid size. Mineral precipitation and the thickness of the zone affected by mineral precipitation in the bentonite near the B-CI increase when the grid size increases while pore clogging in the concrete near the B-CI is computed only for grid sizes smaller than 0.018 cm. The non-isothermal conditions play an important role in mineral precipitation. The reduction in porosity in the B-CI for constant temperature is smaller than that of the non-isothermal run. The model reproduces the on-line measured temperature and relative humidity data as well as the water content and porosity data collected at the end of the test. Model results capture the main trends of the mineralogical observations, except for ettringite and CSH phases for which the predicted precipitation is smaller than the observed values. Model results improve when the specific surface of ettringite is increased by a factor of 10.