The Origins of Olivine Fabric Transitions and Their Effects on Seismic Anisotropy in the Upper Mantle

Convecting mantle plays a central role in the thermal and geochemical evolution of the Earth. It provides the principal force responsible for major geological features such as mountains and ocean basins. Plate tectonics and its violent consequences such as earthquakes and volcanoes are all manifestations of the dynamics of the convective mantle. Shearing forces generated by mantle convection leads to lattice preferred orientation (LPO) of the major upper mantle mineral phases. LPO that develops in this way is thought to be the principal cause behind seismic anisotropy in the upper mantle, which can consequently be used to chart convective flow of the mantle. Strong changes in seismic anisotropy occur in the top 300 km of the upper mantle where olivine is the principal mineral. In this study a solid media high pressure deformation apparatus, called the deformation-DIA or D-DIA, has been used to deform aggregates of San Carlos olivine in simple shear geometry at pressures between 3 and 8.5 GPa and temperatures from 1300-1500°C. As part of this project a high pressure and temperature solid-media cubic assembly was developed to facilitate these experiment that employed alumina pistons cut at 45° to shear the sample but minimized cold deformation of the sample by employing initially porous alumina in the sample column. Once stable high pressures and temperature were reached the cubic assembly was deformed by compressing two vertically oriented anvils of the D-DIA, while the four horizontally oriented anvils were maintained at a constant loading force. This assembly shortening led to shearing of the olivine sample. Recovered samples were analyzed for fabric development employing electron backscattered diffraction (EBSD) and microstructure was observed using transmission electron microscopy (TEM). Experiments were performed at each pressure and temperature as a function of strain rate and H2O content. In dry olivine deformation experiments performed at slower strain rates an A-type fabric dominated at all pressures and temperatures, implying deformation by dislocation glide through the (010)[100] slip system. At higher strain rates evidence for the B-type fabric was observed, suggesting increased activity of the (010)[001] slip system at higher stresses. Recrystallization grains size and dislocation densities were used to estimate stresses in the samples and a good correlation was observed between strain rate and estimated flow stresses. Dry experiments from 8.5 GPa and 1500°C exhibited no LPO, which may be an indication for deformation through diffusion accommodated grain boundary sliding at these conditions. No indication was found that pressure influences the dominant slip system in olivine, in contrast to previous studies. It is considered that previously reported incidences of pressure effects can in fact be attributed to the development of higher stresses in experiments performed at higher pressures. Fabrics in H2O bearing olivine deformed at similar conditions revealed the overriding dominance of the C-type fabric, developed through action of the (100)[001] slip system. Variations in pressure, temperature and strain rate had little influence on this fabric development. TEM observations confirmed the presence of dislocations with slip systems consistent with the development of the macroscopic fabrics. Viscoplastic self consistent modeling was employed to understand the development of fabric in the samples and to estimate the relative contributions of variations slip systems to the developed fabrics. These results are used to construct an olivine fabric map which is found to be consistent with some previous studies at lower pressures. It is argued that the decrease in seismic anisotropy observed in the top 300 km of the upper mantle cannot originate from a pressure induced change in the dominant olivine deformation fabric. Instead it is argued that changes in the H2O content of olivine with depth cause a shift in the dominant fabric from A-type to C-type, with a possible excursion through the E-type fabric, dominant slip system (001)[100], which was, however, not observed in this study. Modeling is used to show that this variation in fabric with depth can cause the observed weakening the seismic anisotropy in the upper mantle if the olivine H2O content increases from below 100 ppm at 50 km to 250 ppm at 300 km. Rather than implying an increased in the H2O content of the mantle with depth, however, it is argued that this change in olivine H2O content can be caused by changes in the H2O olivine-pyroxene partition coefficients with depth, for a fixed bulk mantle H2O content of 200 ppm. Similar deformation experiments performed on a peridotite assemblage at 8.5 GPa and 1300°C indicate identical olivine fabrics to those observed in monomineralic experiments at the same conditions. Fabrics for diopside and enstatite were found to be similar to those found in previously performed lower pressure experiments. Experiments on a piezoelectric single crystal of GaPO4 were performed in the D-DIA and 6-ram MAVO press at high pressures in order to measure charge on the crystal developed through the application of deviatoric stresses. Electrical charges were measured through the use of an operational amplifier. Experiments performed at room temperature using a developed cubic assembly were successful in measuring quantifiable electrical charges resulting from the advancement of the deformation anvils by as little as 0.5 µm. Although the piezoelectric constant for this material is not yet calibrated at high pressures, stresses were estimated from the measured charges and measureable values were in the range 4-350 MPa