Mineralogical Controls of Fault Healing in Natural and Simulated Gouges with Implications for Fault Zone Processes and the Seismic Cycle

The frictional strength and stability of tectonic faults is determined by asperity contact processes, granular deformation, and fault zone fabric development. The evolution of grain-scale contact area during the seismic cycle likely exhibits significant control on overall fault stability by influencing frictional restrengthening, or healing, during the interseismic period, and the rate-dependence of sliding friction, which controls earthquake nucleation and the mode of fault slip. We report on laboratory experiments designed to explore the affect of mineralogy on fault healing. We conducted frictional shear experiments in a double-direct shear configuration at room temperature, 100% relative humidity, and a normal stress of 20 MPa. We used samples from a wide range of natural faults, including outcrop samples and core recovered during scientific drilling. Faults include: Alpine (New Zealand), Zuccale (Italy), Rocchetta (Italy), San Gregorio (California), Calaveras (California), Kodiak (Alaska), Nankai (Japan), Middle America Trench (Costa Rica), and San Andreas (California). To isolate the role of mineralogy, we also tested simulated fault gouges composed of talc, montmorillonite, biotite, illite, kaolinite, quartz, andesine, and granite. Frictional healing was measured at an accumulated shear strain of ~15 within the gouge layers. We conducted slide-hold-slide tests ranging from 3 to 3000 seconds. The main suite of experiments used a background shearing rate of 10 μm/s; these were augmented with sub-suites at 1 and 100 μm/s. We find that phyllosilicate-rich gouges (e.g. talc, montmorillonite, San Andreas Fault) show little to no healing over all hold times. We find the highest healing rates (β ≈ 0.01, Δμ per decade in time, s) in gouges from the Alpine and Rocchetta faults, with the rest of our samples falling into an intermediate range of healing rates. Nearly all gouges exhibit log-linear healing rates with the exceptions of San Andreas Fault gouge and montmorillonite. These samples show no healing at hold times < 100s and then increasing healing at longer hold times. Our results show that healing rate correlates with compaction during holds and subsequent dilation upon reshear. An exception is the Rocchetta fault sample, which exhibits a high rate of healing but low amounts of compaction and dilation. We interpret this to be due to the calcite-rich nature of the sample, which likely results in chemically assisted healing. We find that samples predominately composed of strong, framework silicate minerals compact and dilate more than samples primarily composed of weak, platy minerals. Because velocity-weakening frictional behavior and time-dependent strength recovery are necessary conditions for unstable slip, our data indicates that the presence of phyllosilicate-rich fault gouge would promote stable sliding within the fault. XRD analysis is underway and we will evaluate the relationships between mineralogy and sliding friction, slip velocity, and frictional healing.