Loading path controls the seismic potential of strongly cemented healing faults

Abstract The seismic potential of faults has traditionally been assessed on the basis of frictional properties and geometric configurations, a framework derived largely from studies of cohesionless faults. However, this framework may not apply to strongly cemented and healed faults, such as those filled with quartz or calcite veins, whose reactivation involves cohesive failure and off-fault damage rather than purely frictional sliding. Through triaxial experiments on precut sandstone and shale samples with varying fault dip angles, we show that stress path, rather than fault dip angle, dominates the seismic potential of fault reactivation. Under the load-weakening path that simulates normal faulting, the seismic moment released by strongly cemented faults is two orders of magnitude higher than that under the load-strengthening path that simulates reverse faulting, whereas variations in dip angle yield only a onefold to twofold difference. Remarkably, under the load-weakening path, the mechanically unfavorable 60° dip angle triggers the most violent instability. Rupture not only fractures the cement but also propagates into the surrounding rock, generating a supersized seismic moment and indicating significant energy overspill. We propose the rate of cement cohesion loss as the controlling mechanism. Rapid cohesion loss under the load-weakening path forces energy to overspill into the host rock, whereas slow loss under the load-strengthening path confines dissipation within the fault zone. Finally, the observation of nearly constant seismic moment across experiments with different unloading rates further supports the dominant role of stress path in controlling the seismic potential of fault reactivation.