The strength of a natural serpentinite gouge was measured at 400°C and an effective pressure of 100 MPa. The strength of the gouge decreased with increasing pore pressure to 50 MPa and then was nearly constant between 50 and 125 MPa. The samples at 50-125 MPa pore pressure slid stably, whereas those at 3-25 MPa pore pressure showed stick-slip behavior. Lengthening the pre-heating time to 72,000 s of samples at 100 MPa pore pressure led to increases in both the strength and the likelihood of sliding unstably. With still longer pre-heating times, the strengths decreased again and sliding once more became stable. The strength of the gouge at 3 MPa pore pressure was unaffected by changes in pre-heating time. A possible explanation for the results is that excess pore pressures are being generated in the gouge, causing the true effective pres- sure to be lower than the apparent value.
An understanding of the behavior of gouge materials at depth in fault zones may be critical to earthquake prediction. Moore et al. (1983) described the effects of temperature and confining pressure at 3 MPa pore pressure on the strength and sliding stability of three sheet-silicate-rich gouges and a gouge composed of crushed Westerly granite. Similar studies at 100 MPa pore pressure have recently been completed (Moore et al., in preparation). Comparison of these results showed that at 400°C and 600°C the sheet-silicate- rich gouges were considerably stronger at 3 MPa pore pressure and 100 MPa confining pressure than at 100 MPa pore pressure and 200 MPa confining pressure. Because the imposed effective pressures in these experiments were nearly the same, the marked differences in strength suggested that the effective stress law might not hold at high temperatures. The experiments described in this paper test the effects on gouge strength of changing pore pressure at a constant effective pressure and temperature, using the serpentinite gouge of previous studies. Additional experiments test the importance of heating time on gouge strength. These results also apparently contradict the effective stress law; however, we propose that this law does hold and present a theory in which excess pore pressures generated at different times in the gouge layer reduce the true effective pressures. This theory will be tested in future experiments.
The serpentinite gouge material used is a natural gouge collected from the San Andreas fault zone near San Carlos, California. The gouge consists essentially of chrysotile, with trace amounts of calcite, chlorite, and an opaque mineral. The experimental assembly is described in detail in Moore et al. (1983). Each sample consisted of a 0.65-am-thick layer of gouge sandwiched between 30 ° polished sawcut surfaces in a 19-mm-diemeter granite cylinder. Heat was provided to the sample by a surrounding resistance heater, the deionized water that served as pore fluid was introduced to the sample through a hole drilled into the upper granite piece. Pressures and strains were computer-controlled and -recorded; force and displacement measurements were made outside the pressure vessel using a load cell and displacement transducer. After pore and confining pressures were applied, the samples were heated and then held at temperature and pressure for a specified period of time before loading.