Faults, fluids and friction : effect of pressure solution and phyllosilicates on fault slip behaviour, with implications for crustal rheology

Abstract

In order to model the mechanics of motion and earthquake generation on large crustal fault zones, a quantitative description of the rheology of fault zones is prerequisite. In the past decades, crustal strength has been modeled using a brittle or frictional failure law to represent fault slip at upper crustal levels, and power law creep equations to describe dislocation creep at lower crustal levels. It has long been argued, however, that such two-mechanism strength profiles may significantly overestimate crustal strength, in particular in zones of high strain. Indeed, geophysical observations (notably heat flow and stress orientation data) indicate that major fault zones are significantly weaker than predicted by classical two-mechanism strength profiles. Various explanations for the weakness of major faults have been proposed. These include mechanical processes such as the presence of elevated (superhydrostatic) fluid pressures, the transient reduction of friction by a dynamic reduction of normal stress during earthquakes. Apart from this, the inferred weakness of faults may be due to chemically related effects, i.e., due to the operation of fluid-assisted deformation mechanism such as pressure solution. This may act in concert with the growth and alignment of weak phyllosilicate minerals. The explanation for fault weakness remains a source of controversy, since little experimental work has been done which allows rigorous testing of the various hypotheses. The mechanical behaviour of major fault zones accordingly remains poorly understood. The aim of this thesis is to investigate the effect of solution transfer processes on the mechanical behaviour of fault zones, in order to arrive at a quantitative, mechanism-based understanding of fault slip behaviour under hydrothermal conditions. To this end, sliding experiments were carried out on simulated, gouge-bearing faults. The experiments were performed in a room-temperature ring-shear apparatus. This apparatus allowed control of normal stress and sliding velocity, and in addition allowed high shear strains necessary to achieve significant microstructural modifications. Normal stresses used range from 0.5 to 9 MPa, and sliding velocities of 0.005-10 Ilm/s were used. The experiments were performed using halite and halitelkaolinite gouge as rock analogue materials. Halite was used because the rates of solution transfer processes in halite are well-constrained and extremely rapid at low-stress conditions. This allowed us to do experiments under conditions where solution transfer processes and cataclasis dominated over dislocation creep processes in halite. Kaolinite was used a phyllosilicate constituent because of its simple chemistry, low ion exchange capacity and absence of interlayer water.