Evidence from field, laboratory, and numerical studies indicates that rock strength varies with the direction of the applied force. This anisotropy in viscosity (the ability of a rock to flow) influences large-scale geodynamic process and is integral to the operation of plate tectonics. Anisotropic viscosity arising from pre-existing heterogeneities in mantle structure strongly correlates with both ancient and present-day plate boundary formation. Anisotropic viscosity also markedly affects the location and dynamics of formation of shear zones such as the San Andreas fault. Despite the major influence of anisotropic viscosity on large-scale geodynamic processes, there is a paucity of data quantifying its magnitude in deformed rocks. With our experiments, therefore, we will explore the extent to which anisotropic viscosity influences the weakening of such plate boundaries, enabling them to be localized and narrow, the essential feature of plate tectonics. Consequently, the research outlined in this proposal directly addresses the question "How anisotropic is the viscosity of mantle rocks, and how does this anisotropy evolve during deformation?" We emphasize that our experiments break totally new ground and will provide the first experimental results on viscous anisotropy. This research will impact not only our understanding of the dynamic evolution of Earth's interior, it will also provide important constraints on the mechanical properties of structural ceramic materials used in high-temperature applications.
The most important sources of anisotropy in viscosity are strong alignment of the grains that make up a rock and stress-driven segregation of melt into melt-rich layers. Our research utilizes an innovative experimental approach to determine the anisotropy in viscosity resulting from either a strong crystallographic preferred orientation or a pronounced layering of melt-rich bands. Experiments will be carried out on solid aggregates of olivine and partially molten aggregates of olivine plus basalt. A thin-walled cylindrical sample is first deformed to high strain in torsion to produce significant anisotropy in the microstructure and determine the shear viscosity parallel to either the dominant orientation of the easiest crystallographic glide plane or the layering defined by the melt-rich bands. This sample is then deformed to a small additional strain in tension to determine the viscosity normal to the easy glide plane or melt-rich bands, respectively. The degree of anisotropy in viscosity is expressed as the ratio of these two viscosities. Our initial results indicate that viscous anisotropy is both stronger and evolves on a different timescale than predicted by models based on results from traditional experimental approaches. One long-term goal of this research is to provide insight into the anisotropy in rheological, transport, and seismic properties of deformed partially molten regions of the lower crust and upper mantle. A second goal is to furnish constraints for theoretical analyses that are critical for understanding the underlying physical mechanisms involved in mantle dynamics. A third goal is to develop scaling laws for extrapolation of experimental results from laboratory conditions to mantle temporal and spatial scales.
