Abstract The interplay between the strength of crystalline rocks and fluid diffusion is fundamental to the dynamics of the crust and that of fault systems. We measured the strength of thermally cracked Westerly granite as a function of strain‐rate, confining pressure and fluid pressure. Experiments were conducted at the same initial pore fluid pressure with either nominally drained or undrained boundary conditions. The samples internal pore pressure was measured with pressure transducers attached to the surface of the samples. A systematic decrease in the internal pore pressure was observed when approaching peak stress, leading to an increase in the effective stress and strength of the samples. Although peak brittle strength increased with strain‐rate, it remained consistent with a simple Mohr‐Coulomb failure criterion when calculating the effective stress from the internal pore pressure. Performing additional experiments with undrained boundary conditions revealed that the Mohr‐Coulomb envelope could be determined from a single experiment, demonstrated by the stress trajectory being almost tangential to the failure criterion during the final stage of the experiment. Finally, we demonstrate that the onset of dilatant strengthening (and pore pressure drop) was reached for a critical volumetric (dilatant) strain rate. Because of the competition between fluid diffusion and dilatancy, this critical volumetric strain‐rate, derived through dimensionless analysis, is proportional to permeability, and consequently, exhibits pressure dependence. Overall, our results are summarized with a simple conceptual model, where the brittle strength is bounded by the drained (lower) and undrained (upper) limits, the transition between both being permeability and strain‐rate dependent.
Lin et al. (Wed,) studied this question.