Reactive transport modelling of shale-bentonite interactions in a hypersaline environment

Fully-coupled reactive transport models were developed to simulate the evolution of the interface between the bentonite/sand mixture and the Georgian Bay Shale, which forms the wall rocks of the shaft overlying the Cobourg Formation (limestone), within which it is proposed to locate the repository. The highly saline nature of the shale porewater led to a Pitzer approach being used to model mineral solubilities. The 'base case' simulation suggests that there will be rapid partial replacement of Namontmorillonite with Ca-montmorillonite in the bentonite/sand, with K-montmorillonite then replacing Ca-montmorillonite over longer timescales (tens of thousands of years). Over 100,000 years, minor alteration of the primary minerals at the shaft seal-rock interface occurred in this model, resulting in a reduced porosity alteration zone with a thickness of a few centimetres. The main alteration product in the bentonite/sand was saponite, whereas in the shale, it was analcime (which was included as a representative sodium-rich framework silicate). A number of variant cases were also developed that explored the effects of model discretisation and geochemical assumptions (controls on dissolved silica concentrations) on calculated system evolution. The variation in discretisation led to minor differences in the evolution of the system, especially with regard to porosity. The assumption that quartz buffers the initial dissolved silica concentrations instead of amorphous silica (the 'base case' assumption) resulted in kaolinite forming along with saponite in the bentonite/sand, and kaolinite forming in preference to analcime in the host rock. However, none of the simulations resulted in a significant increase in bentonite-sand porosity over 100,000 years and most of the thickness of the bentonite/sand shaft seal remained unaltered. Hence the models suggest that the performance of the DGR bentonite/sand shaft seals in contact with hypersaline porewater should not be detrimentally affected.