The Opalinus Clay (OPA) formation is considered a suitable host rock candidate for nuclear waste storage. However, the sealing integrity and long-term safety of OPA are potentially compromised by pre-existing natural or artificially induced faults. Therefore, characterizing the mechanical behavior and microscale deformation mechanisms of faults and the surrounding rock is relevant for predicting repository damage evolution. In this study, we performed triaxial tests using saw-cut samples of the shaly and sandy facies of OPA to investigate the influence of pressure and mineral composition on the deformation behavior during fault reactivation. Dried samples were hydrostatically pre-compacted at 50 MPa and then deformed at constant strain rate, drained conditions and confining pressures (pc) of 5–35 MPa. Mechanical data from triaxial tests was complemented by local strain measurements to determine the relative contribution of bulk deformation and fault slip, as well as by acoustic emission (AE) monitoring, and elastic P-wave velocity measurements using ultrasonic transmissions. With increasing pc, we observe a transition from brittle deformation behavior with highly localized fault slip to semi-brittle behavior characterized by non-linear strain hardening with increasing delocalization of deformation. We find that brittle localization behavior is limited by pc at which fault strength exceeds matrix yield strength. AEs were only detected in tests performed on sandy facies samples, and activity decreased with increasing pc. Microstructural analysis of deformed samples revealed a positive correlation between increasing pc and gouge layer thickness. This goes along with a change from brittle fragmentation and frictional sliding to the development of shear zones with a higher contribution of cataclastic and granular flow. Friction coefficient at fault reactivation is only slightly higher for the sandy (µ ~ 0.48) compared to the shaly facies (µ ~ 0.4). Slide-hold-slide tests performed after ~ 6 mm axial shortening suggest stable creeping and long-term weakness of faults at the applied conditions. Our results demonstrate that the mode of fault reactivation highly depends on the present stress field and burial history.
