An exploratory study of the seismic properties of thermally cracked, fluid-saturated aggregates of sintered glass beads

Abstract

Synthetic analogues for cracked crustal rocks have been prepared by sintering soda-lime-silica glass beads of ~ 300 micron diameter at temperatures near the glass transition, and subsequent thermal cracking induced by quenching from high temperature into water. The resulting microstructure involves a controllable concentration (0.03-0.15) of equant pores connected by a network of cracks of uniformly low aspect ratio a ~ 0.0007. Systematic studies of seismic wave speeds and attenuation in such media saturated with fluids ranging widely in viscosity, with both high-frequency ultrasonic and low-frequency forced-oscillation techniques, and related measurements of permeability, are expected to provide a more robust laboratory-based understanding of poroelastic relaxation. In the first stage of this project, cylindrical specimens of 15 mm diameter and 50-150 mm length were tested in both torsional and flexural oscillation. Such measurements were performed under conditions of independently controlled confining (Pc) and pore-fluid (Pf) pressures. Permeability was measured in situ by isolating the pore fluid reservoirs at either end of the specimen and observing the return to pore-pressure equilibrium following a small pore-pressure perturbation in one of the reservoirs. The same specimens were tested dry and argon-saturated, before and after thermal cracking, in order to isolate the effect of the crack network. The responses to torsional and flexural oscillation are essentially elastic with both shear modulus and Young's modulus E independent of oscillation period (1-100 s) and minimal strain-energy dissipation 1/Q < 0.002. The permeability and elastic moduli of the cracked material are each markedly pressure dependent for effective pressure Peff = Pc - Pf < 50 MPa â?? consistent with crack closure at pressures ~ Ea. Work in progress with fluid saturants of higher viscosity (water and glycerol) is targeting predicted porelastic transitions associated with local (squirt) and global (specimen-wide) fluid flow.