Full pore size distribution of shale by combination of innovative gas adsorption isotherms, mercury intrusion porosimetry and laboratory nano-X ray tomography
Résumé
The pore size distribution of clay rocks and soils is crucial for understanding and modelling cation exchange, gas migration, solute diffusion, and hydro-mechanical behavior. This is resulting of the proportion of bulk water and the water within the double layer at the surface of minerals, as well as capillary and osmotic pressures, which are controlled by the pore sizes. Conventional methods, such as mercury intrusion porosimetry and gas adsorption isotherms, are commonly used to characterize the pore network across a wide range of scales. However, isotherms are applied to crushed powder samples which is inappropriate for assessing the preserved microstructure. Then, the two types of pore size distribution provided never successfully match to our knowledge. For this reason, nitrogen adsorption experiments were conducted on preserved rock blocks using an innovative approach with a kinetic control of adsorption equilibrium [1]. Pore size analysis was performed by simulating isotherms with Grand Canonical Monte Carlo kernels. This approach was applied to the Callovo-Oxfordian shale, which is planned to host the French nuclear waste repository. For the first time, this study provides a complete set of comparable data between mercury intrusion porosimetry and gas adsorption isotherms, accurately measuring pore throat and pore body size distributions. However, it is important to note that this coupling only probes pores ranging from the smallest micropores to macropores up to a size of 600nm. Therefore, a full description of the pore network is not yet achieved. The largest pores missing were imaged using laboratory nano X-ray tomography on a 1.5 mm wide core with a voxel size of 295 nm. Advanced in-house algorithms were applied to denoise and restore the numerical volume, and to accurately detect the pores. The segmentation of the pores allowed for the estimation of their size distribution, which was used to build a complete balance of the pore network. To illustrate the impact of the enhanced pore size distributions achieved, a multiscale model implemented in the Code_Aster Finite Element software [2] was used to simulate an experimental swelling pressure curve that was recorded on the Callovo-Oxfordian shale. The results showed a significant improvement in consistency with the experimental data. [1] Matskova N., Prêt D., Gaboreau S., Cosenza P., Brechon R., Gener I., Fialips C.I., Dubes G., Gelin F. (2017) Proceeding of the Unconventional Resources Technology Conference held in Austin, Texas, USA, 24-26 July 2017. DOI 10.15530-urtec-2017-2689299. [2] Mhamdi Alaoui H., Giot R., Prêt D., Cosenza P., Hedan S. (2023) Computers and Geotechnics, 61, 105612.