Flow strengths of quartz aggregates: Grain size and pressure effects due to hydrolytic weakening

Abstract

Novaculites and quartzite ranging in grain size from 1.2–211 μm have been experimentally deformed at confining pressures of 350–1620 MPa under conditions of constant displacement rate and constant deviatoric stress in order to determine the effects of both grain size and pressure on the rheology of quartzite. The amount of water available to the samples was varied so that flow strengths for the entire suite of samples could be compared at several, nominally equal water concentrations; samples were vacuum dried at 800°C for 12 hours, left as is, or sealed in Pt jackets with 0.03–0.4 wt % water added. Novaculites deformed at 800°C and 10−6 /s−1 in the presence of 0.4 wt % water show a continuous decrease in flow strength with increasing confining pressure over the range 350–1590 MPa. At high confining pressures 950–1600 MPa, constant displacement rate experiments show three distinct grain size effects, corresponding to the three levels of water concentration: (1) “grain boundary hardening” for vacuum‐dried samples, (2) grain size independent strength for as is samples, and (3) “grain boundary weakening” for samples deformed in the presence of water. Although grain boundary‐dislocation interactions may lead to grain boundary hardening and grain boundary sliding to weakening, the details of the mechanical data, in combination with microstructural observations, are inconsistent with existing models of intrinsic grain boundary effects. Instead, the strength‐grain size relations are interpreted as resulting from variations in structurally incorporated water, which, in turn, are controlled by diffusion of water to and from the grain boundaries. Finally, constant stress experiments show that the activation energy Q and stress exponent n for creep do not depend on grain size, so that the different strengths observed must be incorporated in the preexponential terms. However, both Q and n show a continuous decrease with increasing amounts of available water, from Q = 300 kJ/mol and n ≃ 4 for vacuum‐heated samples, to Q = 130 kJ/mol and n = 2.6 for water‐added samples.