The strength and mechanical behaviour of quartz slip interfaces: an experimental investigation

Abstract

An experimental study has been undertaken to explore the strength, mechanical behaviour and microstructural evolution of bare interfaces in quartz sandstone during slip. These experiments were designed to simulate fault processes with increasing depth in the continental crust. Two main aspects have been explored: (1) the effect of temperature and confining pressure on the behaviour and stability of favourably-oriented faults, and (2) the influence of reactivation angle on the mechanical behaviour and associated microstructural evolution of a fault zone. Experiments were conducted on Fontainebleau sandstone using a triaxial deformation apparatus, at normal stresses comparable to that in the continental seismogenic regime and over small slip displacements. The first suite of experiments was conducted at temperatures of 400-927°C and confining pressures of 50-200MPa. Experiments reveal complex transitions in fault behaviour between stick-slip and stable sliding regimes. Mechanical results are coupled with microstructural analysis using multiple techniques (including high resolution FE-SEM, and FIB-TEM) that provide insights into fault surface processes down to the nano-scale. Significant findings include the identification of a partially amorphous layer formed during aseismic creep and the generation of pure-silica frictional melt (pseudotachylyte) during high temperature seismic slip events. The pseudotachylyte is recognisable by the formation drawn-out glass filaments and fractured glass patches on the fault surfaces, forming a discontinuous layer up to 2µm thick and covering 10-60% of the fault surface. At normal stresses > 200MPa, frictional melt develops within the first 50µm of rapid slip, correlating with changes in slip acceleration and velocity. High temperature hydrothermal treatment of melt-covered fault surfaces indicates that the pseudotachylyte has a short lifespan (<1 hour) in the presence of high temperature, reactive fluids. The second suite of experiments explores reactivation of fault surfaces inclined between 25º and 70º to the maximum shortening direction, representing faults that vary from optimally-oriented to severely-misoriented for failure. These faults have been reactivated in both dry and fluid-saturated conditions, using two different loading mechanisms. ‘Stress-driven failure’ involves increasing the axial load at constant rate until failure, whereas ‘fluid-driven failure’ is achieved by maintaining a constant axial load and increasing pore fluid pressure until slip occurs. While the initial reactivation of faults obeys frictional theory, continued reactivation is strongly influenced by the microstructural evolution of the fault surface, most notably through the development of frictional melt. Rapid-slip events form a locally-continuous layer of frictional melt in both the dry and water-saturated samples. The presence of pseudotachylyte increases fault cohesive strength through a process termed ‘melt-welding’. Melt-welded regions serve as a nucleation point for the development of off-fault damage and on the most unfavourably-oriented faults, cause lock-up and the failure of a new, more favourably-oriented fault. This work provides new insights into the behaviour and microstructural development of fault surfaces during the early stages of seismic instability. These results have implications for the interpretation of slip processes in natural fault zones, and also more generally for understanding slip mechanics, weakening distances and coseismic fault strength within the continental seismogenic regime.