Dilatancy of Quartz Gouge in Pure Shear

Dilatancy in artificial fault gouge consisting of quartz sand is produced under pure shear loading conditions at shear stresses of about 1/2 the frictional shear failure strength. The gouge, in a layer at 45 ø to the axis of a confined cylindrical sample, is maintained at constant normal stress and internal pore pressure while increasing the shear stress. Repeated cycles to failure result in compaction of the gouge up to shear strains of about 0.3 and porosity of about 15%, after which the gouge shows porosity parted from linear, increasingly negative dilation to positive dilation with increases of axial stress and constant confining pressure. They attributed the volume increase to the opening of microcracks within the granite cylinder. In loosely compacted soils or sands, dilatancy has long been known to occur (Youd, 1972) although only after considerable compaction of the original pore space takes place during the initial increments of inelastic strain. Intact specimens of marble, limestone, sandstone and talc subjected to increases upon shear and decreases upon unloading. 20% axial strain, even though macroscopically The rate of change of porosity with shear stress, d•/d•, is about 2 x 10-4MPa -1, at about 45 to 100 MPa effective normal stress, increasing at lower normal stress. The shear stress at the onset of dilatancy, is nearly independent of normal stress and confining pressure and is considerably lower than has been previously observed. Pure shear loading may be nearer to the conditions for natural loading of strike slip or normal faults than triaxial loading, and we conclude that dilatancy should ductile, show net dilatant volumetric strains following initial compaction when the confining pressures are less than 200 to 400 MPa (Edmonds and Paterson, 1971). At higher confining pressures (400 to 800 MPa) the plasticity of calcite and talc suppresses dilatant cracking but the sandstone becomes dilatant after about 10% longitudinal strain. Thus, even in the ductile regime, microcracking may occur in intact rock at sufficiently high shear stresses or shear strains, although the phenomenon is more pronounced at occur prior to earthquakes in natural fault zones. lower confining pressures. Dilatancy was seized upon by Nur (1972) and Introduction Scholz, et al. (1973) as a means for explaining apparently anomalous negative excursions in the Understanding the processes responsible for the ratio of shear to compressional wave travel times destabilization of faults and the ensuing earthin the region surrounding impending earthquakes in quakes probably contains the key to predicting Central Asia (Nersesov, et al., 1971), and the earthquakes. Some of the early optimism that timing of precursory phenomena relative to the earthquakes should be predictable arose from labomagnitude of the earthquakes. Pointing out that ratory observations of markedly non-linear stressundersaturated rocks have a Vp/Vs ratio about 5% strain behavior in rocks just prior to shear failless than saturated ones, Nur proposed that dilaure. F. C. Frank (1965) proposed that one of the tancy, causing undersaturation followed by diffumechanical processes known to precede shear failsion of pore water into the pore space to reure, namely dilatancy, should lead to instability. saturate the rocks, would lead to the observed Brace and his co-workers (Brace, et al., 1966; Schock, et al., 1973; Hadley, 1973; Jones, 1980; Scholz and Kranz, 1974) and others (Edmonds and Paterson, 1972; Zoback and Byerlee, 1975; Weeks, 1980; Teufel, 1981) examined dilatant volume changes in rock and gouge under various confining pressures and differential stresses. Brace, et al. (1966) found that at differential stresses about half the fracture strength of confined, intact Westerly granite, volumetric strains develocity anomalies. However, with few exceptions (Aggarwal, et al., 1973) subsequent observations have failed to establish the Vp/Vs anomaly as a phenomenon precursory to earthquakes elsewhere. Hadley (1973) noted that, experimentally, dilatancy was not observed in intact rocks until the stresses on the potential failure surface exceeded the frictional strength for sliding on preexisting cracks. Therefore, Hadley reasoned that frictional failure on faults would occur before Geophysical Monograph Series Mineral and Rock Deformation: Laboratory Studies Vol. 36 Copyright American Geophysical Union 2 DILATANCY OF QUARTZ GOUGE