B. Sample assembly design

In addition to the need for improved anvil design, we will explore various modifications of sample assembly designs. Design of a sample assembly including the gasket is critical to achieve (homogenous) high pressure and temperature. For the RDA, it is particularly important because both P-T, and stress-strain distribution are heterogeneous, but developing a sample assembly is also critical to all areas of deformation experiments. All PIs will contribute to sample assembly designs and will communicate to maximize the efficiency of technique development. The D-DIA sample assembly is a cubic volume of solid pressure medium in which the test column of sample, thermocouple, furnace, pistons, insulators, etc. are located. As high-pressure rock deformation has developed from its infancy, different research groups have developed their own flavor of cell assemblies and have a considerable combined knowledge of the properties of cell materials. As the science advances, however, there is a need for better measurement precision, higher pressures and temperatures, better control of chemical environments (especially water). We will therefore develop and calibrate classes of cell assemblies for general use. The plan here is not to bring the art to mass production as has now been achieved for 6/8 multianvil assemblies but to bring about significant improvement in the quality of deformation experiments in the D-DIA. The principal shortcomings of cell assemblies currently in use are mechanical stability of the deformation column (the presence of thermocouple wires in the piston are a particular problem), temperature accuracy, the ability to “self gasket,” and control of water fugacity. The pressure medium in most D-DIA experiments to date has been either boron epoxy (BE), fine powdered boron in an epoxy binder, or mullite (3Al2O3•2SiO2). For high pressure studies of mantle rheology, neither material has the ideal properties of controllable water content and pressure efficiency. BE imparts demonstrably significant water content to olivine samples, such that it seems impossible to conduct a “dry” olivine experiment with BE. Mullite leaves samples bone dry, but is too friable to self-gasket” in a satisfactory manner, and causes premature anvil failure, thus limiting operating pressures to 7-8 GPa. MgO, used on occasion may have favorable gasketing and water properties, but has a high thermal conductivity that severely limits cell temperature. Coordinating with efforts of the other investigators, we plan to test the characteristics of a number of configurations, such as composite media (mullite interior, BE exterior) and independent gasketing as discussed above.