The influence of hydrogen, deformation geometry, and grain size on the rheological properties of olivine at upper mantle conditions

Abstract

Many important geophysical processes, including mantle convection and the associated movement of Earth's tectonic plates, are strongly dependent upon the rheological properties of Earth's upper mantle. Olivine is the most abundant mineral in the upper mantle and therefore largely controls the mechanical behavior of this region of Earth's interior. Many experimental investigations have been carried out to study the rheological properties of olivine single crystals, synthetically produced aggregates, and naturally occurring mantle rocks at asthenospheric temperatures. In contrast, relatively few studies have focused on measuring the rheological properties of olivine deforming at lithospheric temperatures. Furthermore, there are several unanswered questions about the microphysical processes that control deformation of olivine at upper mantle conditions. One outstanding question in the field of rock and mineral physics is "Do different microphysical processes control the rate of deformation of olivine at asthenospheric compared to lithospheric mantle conditions?" To address this question we carried out direct shear experiments on olivine single crystals at temperatures that span the transition from lithospheric to asthenospheric mantle conditions. The results of these experiments, which are presented in Chapter 2, demonstrate that the dependence of strain rate upon stress transitions from a power-law relationship at high temperatures to an exponential dependence at lower temperatures. This transition in rheological behavior is consistent with deformation that is controlled by the climb of dislocations at high-temperature conditions and deformation that is controlled by the glide of dislocations at low-temperature conditions. Furthermore, the direct shear geometry allows for isolation of the (001)[100] and (100)[001] dislocation slip systems, which cannot be individually activated in triaxial compression. At high-temperature conditions, crystals oriented for shear on the (001)[100] slip system are observed to be weaker than crystals oriented for shear on the (100)[001] slip system. At low-temperature conditions the opposite relationship is observed: crystals oriented for shear on the (100)[001] slip system are weakest. Another important outstanding question is "Do the mechanisms of hydrolytic weakening in olivine differ at asthenospheric compared to lithospheric mantle conditions?" In Chapter 3 we report the results of experiments carried out on olivine single crystals under hydrous conditions at both asthenospheric and lithospheric temperatures. For crystals deformed at high-temperatures and under hydrous conditions, the dependence of strain rate on stress follows a power-law relationship with a stress exponent (n) of ~2.5, consistent with deformation that is rate limited by diffusion of silicon through the olivine lattice. In contrast, crystals deformed at high-temperatures and under anhydrous conditions yield n values of ~3.5, consistent with deformation that is rate limited by diffusion of silicon through the cores of dislocations. At low temperature conditions, the strain rate of both hydrous and anhydrous crystals are equally well described by the same exponential dependence of stress. These observations demonstrate significant hydrolytic weakening occurs at asthenospheric temperatures, but hydrolytic weakening cannot be resolved at lithospheric temperatures for our experimental conditions. Lastly, we address a question about polycrystalline deformation: "What deformation mechanism is responsible for grain-size sensitive (GSS) power-law creep of olivine aggregates?" In Chapter 4 we compare strain rates measured during deformation experiments on olivine aggregates to strain rates calculated from a micromechanical model of intragranular slip. The micromechanical model uses the measured stress from deformation experiments and grain orientations determined from post-deformation electron backscatter diffraction measurements to approximate the contribution of dislocation creep to the strain rate. Olivine aggregates deform up to a factor of 4.6 times faster than the maximum possible rates determined from the micromechanical model of intragranular slip. The ratio of experimentally determined strain rates to those from the micromechanical model is strongly dependent upon grain size, but is independent of stress and strength of lattice-preferred orientation. These observations indicate that GSS power-law creep occurs in both weakly and strongly textured olivine aggregates at the studied conditions. We consider three explanations for the observed rheological behavior, (1) a combination of diffusion and dislocation creep, (2) the operation of dynamic recrystallization creep, and (3) the operation of dislocation-accommodated grain-boundary sliding. Our analyses indicate that the microstructural and mechanical behavior of olivine aggregates deforming in the grain-size sensitive power-law regime are most consistent with the operation of dislocation-accommodated grain-boundary sliding at the studied experimental conditions.