Scientists at the Mont Terri research site, which was established in 1996, have been studying the behavior of nuclear waste in contact with engineered and natural geological barriers. Much focus has been on Opalinus clay, a dense, watertight material considered to be a promising barrier for containing radioactive waste. The site's long-term experiments provide datasets that help validate new scientific models.
The MIT study: Researchers tested a new software model developed by Tournassat and Steefel called CrunchODiTi, an advanced version of CrunchFlow, an earlier program. According to the researchers, the software can account for electrostatic effects, an essential factor in simulating the interactions between negatively charged clay minerals and other components like cement used in engineered barriers.
Unlike earlier models, which struggled to match experimental data, CrunchODiTi simulates these interactions in three dimensions with high resolution, thanks to its parallel processing capability on high-performance computers.
Spotlight: The team focused on a specific 13-year-old experiment at Mont Terri involving interactions between cement and the surrounding claystone. Over time, a mix of positively and negatively charged ions was introduced to the cement borehole, and researchers observed changes in the thin, 1-centimeter “skin” at the interface between cement and clay.
By comparing simulation outcomes to the experimental data from this area, the researchers found strong alignment between the two, confirming the model’s accuracy. The results suggest significant physical and chemical changes—such as mineral precipitation and porosity clogging—occur at the interface, impacting the long-term movement of radionuclides.
Results: The findings have important implications for nuclear waste disposal planning. The improved model allows for more-accurate long-term predictions of radionuclide behavior, informing material choices and safety designs for potential geological repositories. While the current focus is on clay formations, the model could also be adapted to evaluate other materials like salt.
The researchers plan to expand their work by incorporating new data and potentially developing machine learning-based models to reduce computational costs. Their goal is to provide robust, scientifically validated tools that can guide public policy and earn public trust in nuclear waste storage solutions.
