The competition between fluid diffusion and volume dilatancy during the failure process of thermally cracked Westerly granite

Pore fluid pressure is known to significantly influence the mechanical strength of rocks. On one hand, an increase in pore fluid pressure may favor crack growth and thus trigger failure, while on the contrary, a decrease in pore fluid pressure will inhibit crack propagation and stabilize failure. On the other hand, whether pore fluid decrease or increases depends on the combination of: 1) pore-fluid pressure boundary conditions; 2) pore fluid pressure diffusion timescale relative to deformation timescales; 3) the latter governing the evolution of permeability and storage capacity within the system. Hence, whether pore fluid pressure will stabilize failure or not, via dilatant strengthening, will depend on a number of parameters, amongst which boundary conditions (drained or undrained), initial permeability, and strain rate must be included.

Here, we perform a comprehensive investigation into the mechanical strength of thermally cracked Westerly granite, in order to provide insights on pore fluid pressure evolution during rock failure. We conducted triaxial loading experiments on heat-treated Westerly granite samples (heated to 700 °C). In order to investigate the mechanical and hydraulic responses throughout the entire failure process, experiments were performed under both drained and undrained boundary conditions, at different strain rates (10^-4, 10^-5, 10^-6, and 10^-7 s^-1) and initial effective confining pressures (5, 20, and 40 MPa). All experiments were conducted at an initial pore pressure of 50 MPa. During each experiment, stress, strain, as well as pore pressure (using 8 in-situ pore pressure transducers) were monitored. Acoustic emission and the evolution of elastic P-wave velocities were also recorded. So far, our experimental results demonstrate that: 

• Water-saturated granites under undrained conditions show higher strength than drained ones, due to dilatant strengthening from reduced pore pressure at failure.
• Under drained conditions, the onset of pore pressure drop is governed by the competition between fluid diffusivity and volumetric strain rate.
• Dilatant strengthening under drained conditions is strain-dependent, with larger pore pressure drops at high (10^-4 /s) vs. low (10^-7 /s) loading rates.  

Our results highlight the importance of dilatant strengthening during the failure of crystalline rock. For instance, it is possible that a number of former studies realized under nominally drained water saturated conditions may have underestimated the effect of water weakening, due to important – yet unobserved at the time- dilatant strengthening happening, resulting in strength of dry and saturated specimen being almost equivalent.

Dilatant strengthening being most efficient at high strain rate, we can safely extrapolate that it is most efficient just before or just after failure. In particular, by stabilizing failure, it may explain long foreshock and aftershock sequences, as seem to be the case in at least some of the recorded AE sequences during our experiments. Finally, whether thermal pressurization is or not able to balance out (because of high velocity frictional heating) dilatant strengthening during dynamic rupture remains to be investigated.