The effect of shear load on frictional healing in simulated fault gouge

We report on frictional strengthening (healing) in granular quartz gouge as a function of time of true stationary contact. To distinguish between the slip-dependent [Ruina, 1983] and time-dependent [Dieterich, 1979] friction constitutive laws, we designed tests similar to conventional slide-hold-slide (SHS) tests except that shear load was completely removed prior to holds. We find large healing values (0.033-0.054 for holds of 101-102 s) compared with quasi-static SHS experiments, and our data indicate time-dependent weakening in contrast with strengthening observed from SHS tests. Gouge layer compaction increases with increasing hold time, comparable to observations from conventional SHS tests. Our data indicate that purely time dependent processes have a minor influence on healing under the conditions tudied and/or that such effects are efficiently erased by particle rearrangement during removal/reapplication f shear load. The data are not adequately described by either the time (Dieterich) or slip (Ruina) dependent state evolution laws. the existing studies are limited to rock surfaces (initially without gouge) held in quasi-stationary contact. Under these conditions, shear load is non-zero during intervals of frictional restrengthening and both slipand time-dependent mechanisms may operate. To further distinguish between the evolution laws, and to extend detailed studies of friction evolution to granular fault gouge, we have performed experiments involving true stationary contact. In these tests shear load is totally removed prior to hold periods, and we directly compare the results with data from conventional slide-hold-slide (SHS) experiments. Our friction data for short times of true stationary contact (101-102 s) show significantly higher static friction levels than conventional SHS tests. In addition, we find negative healing rates, and thus decreasing static friction as a function of hold time in contrast to conventional SHS experiments. Comparison of our data with similar data for sliding between rock surfaces indicates ignificant differences, which may relate to processes inherent o shear within granular material.