Experimental challenges to reproduce seismic fault motion

This presentation briefly reviews scientific and technical development in the studies of intermediate to high-velocity frictional properties of faults and summarizes remaining technical challenges to reproduce nucleation to growth processes of large earthquakes in laboratory. Nearly 10 high-velocity or low to high-velocity friction apparatuses have been built in the last several years in the world and it has become possible now to produce sub-plate velocity to seismic slip rate in a single machine. Despite spreading of high-velocity friction studies, reproducing seismic fault motion at high P and T conditions to cover the entire seismogenic zone is still a big challenge. Previous studies focused on (1) frictional melting, (2) thermal pressurization, and (3) high-velocity gouge behavior without frictional melting. Frictional melting process was solved as a Stefan problem with very good agreement with experimental results. Thermal pressurization has been solved theoretically based on measured transport properties and has been included successfully in the modeling of earthquake generation. High-velocity gouge experiments in the last several years have revealed that a wide variety of gouges exhibit dramatic weakening at high velocities (e.g., Di Toro et al., 2011, Nature). Most gouge experiments were done under dry conditions partly to separate gouge friction from the involvement of thermal pressurization. However, recent studies demonstrated that dehydration or degassing due to mineral decomposition can occur during seismic fault motion. Those results not only provided a new view of looking at natural fault zones in search of geological evidence of seismic fault motion, but also indicated that thermal pressurization and gouge weakening can occur simultaneously even in initially dry gouge. Thus experiments with controlled pore pressure are needed. I have struggled to make a pressure vessel for wet high-velocity experiments in the last several years. A technical difficulty was how to absorb hydrodynamic shock due to abrupt fault motion in the vessel, and this was overcome by pressurizing water in the vessel, acting as pore fluid, using pressurized gas (in other words using gas as a cushion). I will report preliminary experimental results on high-velocity rock-on-rock friction under pore-water pressure. Other technical challenges are (i) how to produce step-change in velocity to see if the framework of rate-and-state friction law holds in high-velocity regime, (ii) how to conduct high-velocity friction experiments in hydrothermal conditions to study frictional properties relevant to slow slip and low-frequency tremors, and (iii) how to conduct high-velocity friction experiments at high normal stresses. The first task became possible with a low to high-velocity apparatus in Beijing and a few other machines, and I will show some preliminary results. There is no fundamental difficulty in (ii) since O-ring is enough to seal piston rotating at a high speed. However, (iii) will be the hardest because of severe thermal fracturing of host rocks that limits the axial stress. Use of aluminum sleeve made it possible to apply the normal stress to about 30 MPa, but new device and a high motor power is needed to go higher normal stress.