Electrical potential changes and acoustic emissions generated by fracture and fluid flow during experimental triaxial rock deformation

Natural electrical potential signals have been recorded from numerous seismically active areas around the world and therefore have been proposed as a potential earthquake prediction tool. The streaming potential is being used to locate sub-surface water reservoirs, to monitor steam fronts during enhanced oil recovery techniques, and to delineate the anisotropy of fractures in geothermal and oil reservoirs. The generating mechanism for these signals is still unclear although plausible theories include: - Piezoelectric fields produced through stress changes on piezoelectric materials, such as quartz, found in many rocks. - Electrokinetic currents induced through a pressure gradient and caused by electrical charge transport within a moving fluid. - Less well-established theories for instance involving current carrying mobile O' charges. To investigate the relative significance of these mechanisms, I have measured the direct current electrical potential and acoustic emissions during constant strain rate rock deformation under simulated crustal conditions of pressure and pore fluid pressure. Some sixty-one experiments were done on rock samples of quartz rich Darley Dale and Bentheim sandstone and quartz free basalt from Iceland. A computer and servo-controlled conventional triaxial cell was used to deform dry, water-saturated and brine-saturated rock samples at confining pressures between 20 and 200MPa, pore fluid pressures between 10 and 50MPa and strain rates from 10-4 s-1 to 10-6 s-1 I identify the primary sources of the electrical potential signals as being generated by (i) piezoelectricity in dry sandstone experiments and (ii) electrokinetic effect in saturated basalt experiments. I show that electrical potential signals from the other proposed methods are not detectable above the background noise level. It can therefore be argued that the electrokinetic effect is the main electrical potential generating mechanism within the upper crust.Both precursory and coseismic signals to dynamic rupture exist for pure water-saturated samples. However for brine saturated rock samples, the increased ionic mobility of the charges is suspected as the cause of absence of precursory signals, hicreasing effective pressure decreases the pre-seismic electrical potential anomaly and is found to increase the co-seismic electrical signal. These results are explained in terms of the number of electrical and hydraulic pathways available for electrical current flow. Variations of electrical potential difference and acoustic emission change markedly with strain between the compaction and dilatancy phases of rock deformation, and show dependence on both effective pressure and strain rate. Differences between streaming potential and hydraulic permeability during deformation are explained using a model of varying electrical and hydraulic tortuosity. Intergranular and intragranular cracking can be distinguished using the electrical potential generated from brine saturated rocks. The electrical and streaming potential signals occurring during deformation are found to reflect the accumulating and accelerating damage prior to fracture and the localisation of damage at dynamic fracture