Recent experimental investigations of friction in rock demonstrate pronounced reduction in the apparent coefficient of friction at seismic slip rates (>0.1 m/s). Here we present the capabilities of a novel apparatus under development that can accommodate both double-direct and triaxial shear configurations to investigate friction at low to high-speed (up to 2 m/s) sliding conditions. The high acceleration capability of the loading-system is complementary to other high-speed testing machines, and is particularly well suited for study of rate-dependence of friction at high normal stress and of transient behavior associated with dynamic weakening. Preliminary friction experiments at normal stress of approximately 5 MPa have been conducted to investigate transient friction associated with slip velocity steps from low to high speed sliding (>0.5 m/s). Results show that, independent of surface roughness, friction increases concurrently with a step increase to high velocity and then decreases with slip consistent with the direct and evolution effects documented for rate-state friction behavior at quasi-static slip rates. Friction during sustained highspeed sliding is characterized by a steady and substantial weakening that likely reflects the accumulation of fine wear product along the surfaces.
Experiments designed to investigate the frictional behavior of rock generally involve sliding the surfaces of two blocks of rock past each other. In these tests the rock blocks do not fail or break but slide on either fractured or pre-cut, prepared surfaces with or without an intervening layer of granular material called gouge by analogy with natural faults (e.g., ). The coefficient of sliding friction is represented by the stress across the sliding surface, specifically the shear stress parallel to the sliding direction divided by the normal stress on the surface. Friction tests can be conducted on unconfined or confined samples. The normal load in unconfined tests is limited by the fracture strength of the rock blocks, typically on the order of 50 MPa. In confined tests, samples are sealed in an impermeable membrane and surrounded by a pressurized gas or a fluid at a confining pressure, Pc. Confined tests simulate the high pressure condition at depth in the earth where an earthquake occurs. Samples also may be permeated by a fluid, normally water, at a pore fluid pressure, Pp.
Recent experimental investigations of friction in rock and gouge appropriate to earthquake slip using high-speed apparatus demonstrate pronounced reduction in the apparent coefficient of friction at seismic slip rates greater than 0.1 m/s, which likely reflects lubrication by generation of melt or ultrafine wear particles, thermalchemical processes such as flash heating and thermal decomposition, and dynamic thermal fluid pressurization within thin slip zones. Recent past and current development of testing devices is aimed at improving our ability to investigate coseismic friction weakening at all conditions of earthquake rupture (specifically slip velocity, normal stress, fluid states, ambient temperature) and develop relations for scaling (spatial and temporal) laboratory results to natural earthquake faulting.