MR24A-01 INVITED 16:00h
Development of a rotational Drickmaer apparatus for large-strain deformation experiments under deep Earth conditions
Developing an apparatus for quantitative deformation experiments beyond the transition zone pressures ($>$15 GPa) is critical for understanding whole mantle geodynamics. A new type of torsion apparatus "Rotational Drickamer Apparatus (RDA)" has been developed in which large-strain deformation experiments have been performed at pressures and temperatures up to $\sim$15 GPa and $\sim$1700 K respectively. The apparatus consists of opposing anvils made of tungsten carbide that are supported by a pyrophyllite gasket. A sample is located between two anvils sandwiched between alumina or YAG disks. The sample space is heated by two disk heaters made of a mixture of TiC and diamond. Deformation is achieved by rotating one anvil relative to another through a Harmonic Drive $^{TM}$ gear box. Deformation experiments up to $\sim$15 GPa and $\sim$1700 K to shear strains of up to $\sim$2 have been performed for samples of (Mg,Fe)O and (Mg,Fe) $_{2}$SiO $_{4}$ olivine and wadsleyite at the rotation rates of 3-7 x 10$^{-4}$ rpm corresponding to the shear strain rates of 0-5 x 10$^{-4}$ s$^{-1}$. A conical window is made in a cylinder containing the anvils for the X-ray {\it in-situ} stress and strain measurements. In this apparatus, both uniaxial compression and shear deformation occur. To determine uniaxial stress and shear stress separately, X-ray diffraction measurements were performed at five different angles with respect to the rotation (compression) axis (i.e., $0\deg$, $\pm$$45\deg$, $\pm$$90\deg$). The sample thickness change and shear deformation were monitored by an imaging system in the synchrotron experiments. This apparatus allows quantitative studies of plastic deformation and deformation microstructural development at a fixed strain-rate under the pressure and temperature conditions equivalent to deep mantle exceeding $\sim$500 km depth.
MR24A-02 16:15h
The Rheology of Dry, Melt-Free Polycrystalline Fo90 Olivine
We have recently measured the shear modulus and attenuation of pure, fine-grained solution-gelation-derived olivine at seismic frequencies (Jackson et al., JGR, 2002). These data have been fitted to a Burgers type model that includes a term for Newtonian viscous behavior. However, at the relatively short time-scales and microstrains of the torsional forced oscillation experiments the viscosity is not very well constrained. Motivated by the need to better constrain the long-term viscous behavior we have begun a program to determine the finite strain rheology of these materials. A particular aim is to determine the activation energy for diffusion creep over a similar temperature range as for the torsional forced oscillation data. After reaction of the gel to olivine in a controlled atmosphere furnace, cold-pressed pellets are hot-pressed at a range of temperatures to produce a range of mean grain sizes. The subsequent triaxial compressive creep experiments are conducted in a Paterson-type gas medium apparatus at a confining pressure of 300 MPa and temperatures in the range from 1150 to $1350\deg$C. Each sample is deformed at a range of applied loads at a constant temperature. During each segment of a test the load is held constant to determine the strain rate for a given stress before stepping up in load for the next constant load segment. Preliminary observations indicate that grain growth is minor during the deformation tests; mean grain sizes vary with hot-pressing temperature from 3 to 6 micron. The deformation tests clearly show the transition from diffusion to dislocation creep at stresses of 100 to 150 MPa. When the diffusion creep data is normalized to a common grain size with a grain size exponent of 3, strain rates at a given stress for these genuinely melt-free samples are up to 2 orders of magnitude lower than those observed by Hirth and Kohlstedt (JGR, 1995) for nominally melt-free samples with grain sizes $>$ 10 micron prepared from natural olivine. However, our strain rates are similar to those reported by Beeman and Kohlstedt (JGR, 1993) for melt-free samples with grain sizes similar to ours. These initial results suggest that either the presence of melt has a larger effect on the strength of upper mantle rocks than has previously been reported or the grain size exponent has been overestimated.
MR24A-03 16:30h
Deformation Experiments at High Pressures in Diamond Anvil Cells: Texture Development in MgSiO3 Perovskite
With increasing interest in the geodynamics of the lower mantle and seismic anisotropy in the deep earth, a better understanding of the deformation behavior of lower mantle minerals is essential. Pressure conditions in the lower mantle can at present only be achieved with diamond anvil cells (DAC) and thus we have conducted DAC deformation experiments up to 50 GPa on (Mg,Fe)SiO$_{3}$ perovskite, the major lower mantle phase. In order to quantify the reproducibility and accuracy of the technique, we investigated copper up to 25 GPa as a reference material to verify texture type and homogeneity. Our previous DAC experiments documented texture development during phase transformations from olivine to ringwoodite, and to perovskite + magnesiowuestite. The new experiments explored enstatite transforming to perovskite and confirmed strong texture development with (001) lattice planes at high angles to the compression direction. These experiments furthermore document strengthening of textures with time and significant changes during in situ heating, with grain growth, corresponding to recrystallization. While conditions are still far from those in the lower mantle (stress, strain rate and temperature particularly) DAC experiments shed light on the deformation mechanisms in high pressure phases that can be used for modeling geodynamic processes in the deep earth.
MR24A-04 16:45h
Plastic Deformation of MgSiO$_{3}$ Perovskite: Cubic or Orthorhombic ?
The Earth's lower mantle consists mainly of (Mg, Fe)SiO$_{3}$ perovskite and (Mg, Fe)O. Perovskite which takes up at least 70 per cent of the total volume is likely to control the rheology of the lower mantle. Understanding the elementary deformation mechanisms (dislocations, slip systems) of perovskite is thus one of the central issues of deep mantle studies. This knowledge is also needed to interpret seismic anisotropy in the lower mantle. (Mg, Fe)SiO$_{3}$ perovskite is not stable at high temperature under ambient pressure. Deformation experiments on perovskite must be carried out under extreme pressure-temperature conditions, inducing severe limits on our progress in that field. Contrary to the ideal, high-symmetry, structure of CaTiO$_{3}$ which is cubic (space group Pm3m), (Mg, Fe)SiO$_{3}$ perovskite is distorted, resulting in an orthorhombic symmetry (space group Pbnm). Distortions consist in tilted octahedra and Mg atoms being displaced from the centres of their sites. Assessing the influence of these distortions on plastic shear anisotropy in perovskite is thus a major issue as cubic or orthorhombic symmetries can lead to very different crystal preferred orientations and seismic anisotropies. Considerable insight into the physical properties of minerals at high pressure has been obtained from calculations using computational methods. Mechanical properties represent however a complicated case as several length-scales from atomic to macroscopic must be considered. In minerals (with usually complex crystal chemistry), the choice of slip systems is largely controlled at the atomic scale within the dislocation cores. Modelling dislocation cores of minerals at the atomic scale requires many atoms to be simulated and is still very challenging. In this study, we propose an alternative approach based on the calculation of generalized stacking fault (GSF). A GSF is produced when a crystal is sheared in a given plane by a given displacement u. It is not stable in general and must be balanced by a restoring force. It is thus possible to calculate the intrinsic resistance opposed to plastic shear associated to a given slip system. This approach is applied here for the first time to (Mg, Fe)SiO$_{3}$ perovskite to provide a first estimate of plastic anisotropy in this important mineral.
MR24A-05 17:00h
A New Technique for In Situ X-ray Microtomography Under High Pressure
We have developed a new technique for in situ synchrotron microtomography to study texture evolution in multi-phase specimens under high pressure and temperature. Two critical issues in performing tomography experiments under pressure are (1) the limited X-ray access to the sample because of the highly absorbing materials, such as tungsten carbide and tool steel, typically used in the pressure vessel and (2) a high pressure compatible rotation mechanism to collect projections of the sample continuously from 0 to 180$\deg$. We addressed these issues by (1) employing an opposed-anvil high pressure cell, known as the Drickamer cell, with an X-ray transparent containment ring, to allow panoramic X-ray access, and (2) rotating the Dricakmer cell by Harmonic Drive$^{TM}$ gear reducers, with thrust bearings supporting the hydraulic load. The design of the rotation mechanism benefited from the rotational deformation apparatus developed by Yamazaki and Karato (Rev. Sci. Instrum., 72, 4207, 2001). We report results obtained from a test run performed under pressure with monochromatic synchrotron radiation. A sapphire sphere (1.0 mm dia.) was embedded in a powdered mixture of Fe and 9 wt.$%$ S alloy. The diameter of the sample chamber was 2 mm. Under pressure, the entire Drickamer cell was rotated to collect radiographs of the sample at various angles from 0 to 179.5$\deg$ in 0.5$\deg$ step size. Computational reconstruction of these projections provided three dimensional (3D) distribution of linear attenuation coefficient of the sample with a spatial resolution of 6 microns. The shape change in the sapphire sphere during compression was clearly observed. Using the program Blob3d, reconstructed 3D images of the sphere were separated from the surrounding Fe-S alloy. Volumes of the sphere were then accurately determined from the extracted images, by carefully defining the image intensity threshold. The errors in the volume measurement are about 0.3 to 0.7$%$, mostly due to shadowing by anvil deformation. The results, although performed using a solid sample, demonstrate the potential of measuring melt volume. Previous density measurements using X-ray radiography with only one dimensional data assumed that the shape of the sample remained unchanged throughout the experiment. In our new technique, this assumption is no longer required and density of melts can be inferred directly from the sample volume even when the molten sample is distorted. Other applications of this apparatus will be also discussed.
MR24A-06 17:15h
Material Composite Behavior Under High-Pressure
In situ x-ray diffraction techniques under relevant pressure and temperature conditions provide unique information about phase stability, elasticity and deformation behavior of Earth materials. Often samples consist of a calibrated standard intermixed with the material of interest. Accurate measurements of equation of state are based on two assumptions: that the equation of state of the calibrant is known precisely, and that the pressures of these two materials are the same. However, except under strict conditions of hydrostaticity, pressures are not necessary equal. To provide a detailed examination of the pressure relationship in the diamond anvil cell sample chamber, we analyzed two standard materials mixed together in a controlled geometry. Our samples consisted of unidirectional Al$_{2}$O$_{3}$ ceramic fibers ($\sim$ 1$\mu$m diameter) distributed in an Al metal matrix. This was ideal because both materials are existing high-pressure standards and the oxide/metal mixture is similar to many experiments. We conducted room temperature radial x-ray diffraction experiments using a diamond anvil cell at the X17C beamline at National Synchrotron Light Source. We studied two different fiber orientations with respect to the diamond anvil cell compression axis: one with fibers oriented vertically and the second, horizontally. In each case we measured the d-spacing of lattice planes as a function of rotation angle between principle stress axes and diffraction geometry. From these data, we calculated pressure and supported differential stress of both Al and Al2O3. We found that geometry plays an important role in determining the relative pressure and strength behavior of the two materials. At comparable pressures, in the vertical fibers case, P$_{Al}$ $\sim$ 8.7 GPa, P$_{Al2O3}$ $\sim$ 10.2 GPa and in the horizontal fibers case, P$_{Al}$ $\sim$ 9.6 GPa when P$_{Al2O3}$ $\sim$ 10.2 GPa. Thus, when the fibers are oriented vertically, aluminum pressure is always smaller than alumina pressure; whereas in the horizontal case; fibers and matrix pressures are almost the same (not more than 0.6 GPa in difference in our experiments). In addition, we present finite element modeling of behavior of composite materials in the diamond cell sample chamber that are in excellent agreement with experiments results. With this study we show that the geometry of samples in the diamond cell must be understood in order to properly interpret measurements. Our ultimate goal is to use this information to design samples that are optimized for better measurements of rheological behavior of Earth interior materials.
MR24A-07 17:30h
Pore Space Geometry and Seismic Anisotropy of Rocks: 3-D Experimental Investigation
Pressure-driven closing of the pores in the rock sample results in changes of its effective physical properties. We use 3-D ultrasonic pulse-transmission method to characterize the relationships between the spatial distribution of microcracks and elastic anisotropy of the rock. With the use of apparatus developed in Geophysical Institute, Prague, the P-wave velocities and amplitudes ({\IT V$_{P}$} and {\IT A$_{P}$}) are measured in 132 directions on spherical rock samples. The measurements are carried out at several steps of confining pressure within pressure-increasing and pressure-decreasing paths (0.1-400, 400-0.1 MPa). Then the directions of maximum and minimum velocities and amplitudes are found for which the measurements is repeated under continually changing pressure. As the measurements are repeated at the same position under various pressures, the data can be processed so that the change of the {\IT V$_{P}$} and {\IT A$_{P}$} between the individual pressure-steps and the hysteresis at particular pressure can be seen directly. The resulting differential diagrams show the magnitude of {\IT V$_{P}$} and {\IT A$_{P}$} changes in 3-D which are mainly due to the pressure-induced closing of microcracks, and respectively, the flexibility of the microcracks. Using the data measured at high confining pressure or those computed with averaging method we are able to distinguish the influence the deformation-induced lattice re-orientations from the pore-related properties. Numerous measurements carried out on various rock samples show that the anisotropic patterns of {\IT V$_{P}$} and {\IT A$_{P}$} changes due to the closing of oriented microcracks and other pores highly correlate with the macroscopic structural features of the rock (preferred grain-shape orientation, fracture cleavage, stretching lineation) and are sensitive to them. It is believed that in such cases where the structural features associated with porosity can not be observed directly, the above outlined method will be applicable as a tool for the examination of pore space geometry.
MR24A-08 17:45h
Effect of Partial Melt on P and S Wave Velocities in Olivine-Basalt Aggragates
The presence of relatively small amounts of melt may have great effects on the elastic properties of mantle. Most of these works were achieved either by experiments at low pressure (commonly less than 1.0GPa) and high temperature or by numerical simulation, so it is essential to study the effect of partial melt on the elastic property in mantle rocks at simultaneous high pressures and temperatures. The compressional and shear wave velocities (V$_{P}$ and V$_{S}$) for two synthetic aggregate samples of olivine plus 4% and 13% basalt (S1 and S2) have been measured up to 5.0GPa and 1530K(100-150km) using simultaneous ultrasonic interferometry in conjunction with in situ X-ray diffraction and X-radiographic imaging techniques. At 3.5GPa, V$_{P}$ and V$_{S}$ in the two samples showed an almost linear decrease as the temperature increasing, and above 1350K began to drop remarkably. At 1473K, V$_{P}$ and V$_{S}$ reduced about 2.1% in the sample with 13% basalt, three times more than those in the sample with 4% basalt (about 0.7%). The large increase of V$_{P}$/V$_{S}$ ratios in the two samples at the temperatures above 1350K suggests that the partial melting is involved. Variations in P and S wave velocities and V$_{P}$/V$_{S}$ ratios at above 1300K might be related to partial melting of the basalt in olivine samples. These results allow us to investigate the anomalous elastic behavior in the upper mantle, and provide important clues to reveal the relationship between mantle physical state and seismic observations.