Stress-stimulated current of dry rocks with constant clamping stress
Wednesday, 17 December 2014
A set of nominally dry rocks (gabbro, granite, limestone, marble, and sandstone) were subjected to asymmetric loading with a large hydraulic press. A pair of precision platens made from 1018 low carbon steel were used to apply uniaxial compressive stress (σ) to the sample, via a thin electrical insulator made from ultra-high molecular weight (UHMW) polyethylene. Self-adhesive copper electrodes were applied and burnished on the end faces and the stress-stimulated current (SSC) was monitored using a Keithley 617 instrument. A preload stress level of 5.5 MPa was applied to firmly clamp the assembly throughout the experiment. From this baseline, σ was increased to 22.25 MPa and held for 100 seconds before returning to the clamping stress level. This loading profile was repeated for four or more cycles, with a stress rate on the order of 5MPa/sec. After the first load cycle, the SSC transients (and SSV offsets) are reversible when σ returned to its baseline level. All samples showed alternating unipolar SSC transients at the beginning and end of each load cycle. SSC from limestone, Westerly granite and marble were at, or below, the measurement limit (±1 pA). All other samples except sandstone showed a negative SSC with increasing stress. For stress-stimulated voltage (SSV) there was a richer variety of transients observed such as unipolar, bipolar and more complex transient dynamics. Limestone was the only sample tested with no SSV transients although this particular rock had a major calcite inclusion in the sample. White granite tended to have the least stable SSC and SSV values. Of the six different rock samples tested under identical conditions, the SSC and SSV observed were not greater than –15 pA, presumably due to improved experimental procedures. The response for rocks with semiconductor properties (gabbro, granite) is the same as those without semiconductor properties (limestone, marble), although the values for marble were below the noise. For repetitive Δσ of 16.7 MPa gabbro, limestone, and three types of granite showed negative or zero SSC, and sandstone exhibited roughly +1 pA SSC. In conclusion, use of a constant clamping force was found to eliminate a large source of instability in the experiment. Stability improved after the first full cycle, and if subsequent loads were kept below the maximum load of the first cycle.