MR21A-01 08:00h
Discovery of MgSiO$_{3}$ Post-Perovskite Phase Transition
Recent progress in X-ray diffraction (XRD) measurement in situ at high-pressure and -temperature in a laser-heated diamond-anvil cell (LHDAC) enables us to do a search for new high-pressure phases to the core-mantle boundary condition. MgSiO$_{3}$ perovskite is a principal mineral at least in the upper part of the lower mantle, but its stability and possible phase transition at greater depths have long been controversial. We first found the phase transition of MgSiO$_{3}$ perovskite in an experiment with multi-component natural mantle composition, in which XRD pattern significantly changed at 115 GPa and 2550 K. Later, we observed similar change in the XRD pattern of pure MgSiO$_{3}$ composition above 127 GPa and 2500 K. Based on the fifteen new diffraction peaks from pure MgSiO$_{3}$, we determined the crystal structure of post-perovskite phase by the molecular dynamics (MD) calculations (Murakami et al., Science, 2004). The new phase has SiO$_{6}$ octahedral sheet stacking-structure with orthorhombic symmetry (space group: Cmcm) and is isostructural with UFeS$_{3}$ and CaIrO$_{3}$. Stability of MgSiO$_{3}$ post-perovskite phase was confirmed later by theory. First-principles calculations further demonstrate that it can cause the DO_L discontinuity, S-wave anisotropy, and anti-correlation between S-wave and bulk-sound velocities in the lowermost mantle. Our group has been working on the post-perovskite phase transition in a variety of bulk chemical compositions. It occurs in both MgGeO$_{3}$ (Hirose et al., Am. Mine., 2004) and MnGeO$_{3}$ around 60 GPa at 1600 K, indicating that they are low-pressure analogues to MgSiO$_{3}$. In contrast, CdGeO$_{3}$ perovskite becomes less distorted with increasing pressure, and the perovskite structure is stable at least to 110 GPa and 2000 K. The Al-bearing MgSiO$_{3}$-rich post-perovskite phase also appears in a MORB composition above 108 GPa and 2200 K. The buoyancy relationships between the mantle and former basaltic crust can be complex at the post-perovskite phase transition in both the compositions. It has important implications for the fate of basaltic crust component in the deep interior.
MR21A-02 08:15h
In Situ X-ray Diffraction Study of MgSiO$_3$ Perovskite to the Core-Mantle Boundary Conditions
In order to understand the stability and crystal structure of MgSiO$_3$ perovskite in the deep mantle, in situ angle-dispersive X-ray diffraction measurements have been performed at 90-144 GPa and 300-2900 K in double-sided laser-heated diamond-anvil cells at the GSECARS sector of the Advanced Photon Source. MgSiO$_3$ glass was used as a starting material. Mixtures of MgSiO$_3$ and platinum (pressure standard) were loaded in 50 micron holes in rhenium gaskets and compressed with beveled diamond anvils (100 micron size cullet). In our study, the samples were surrounded by argon which serves as both pressure medium and thermal insulation. Together with double-sided laser heating, this helps to reduce temperature and stress gradients in the sample during laser heating. The occurrence of a new diffraction peak at 2.57-2.62 \AA, which we previously reported based on energy-dispersive X-ray diffraction measurements, was confirmed at 90-144 GPa. The pressure-induced d-spacing shift of this new peak is consistent with that expected for MgSiO$_3$ perovskite lines, indicating that the new diffraction line is associated with either a modification of the crystal structure of MgSiO$_3$ perovskite at 88 GPa or another structure with very similar compressibility. During 25 minutes of heating at 2000 K and 144$\pm$10 GPa, only the diffraction peaks of MgSiO$_3$ perovskite were observed together with the new line at 2.57-2.62 \AA\ and the lines from the pressure scale and the pressure medium. When we increase temperature to 2500 K, however, significant changes were observed: several new peaks and peak splittings, as well as intensity changes of some peaks. The major lines of MgSiO$_3$ perovskite were still observed in the diffraction patterns. Dominant peaks among the new diffraction features can be explained by those of the proposed post-perovskite phase (CaIrO$_3$-type: Murakami et al., 2004). Our results therefore support the possibility of a crystal structure modification in MgSiO$_3$ perovskite at 1800-2000-km depth conditions and a major phase transition in MgSiO$_3$ near the base of the mantle.
MR21A-03 INVITED 08:30h
MgSiO$_3$ post-perovskite at D" conditions
The thermoelastic properties of the new polymorph of MgSiO$_3$ with the CaIrO$_3$ structure and more stable than the Pbnm-perovskite phase at conditions close to those expected in the D" region has been investigated by first-principles computations and contrasted with those of the perovskite phase. Although we investigate only single and pure phases, the elasticity of aggregates containing predominantly these phases is expected to differ similarly, although in smaller magnitude. We therefore predict the major trends in seismic properties expected to occur in the presence of such phase change, such as velocity discontinuities, ratios of velocities and density anomalies, and anisotropy in aggregates with preferred orientation. Consequences of this model mineralogy for the D" region will be discussed. J. Tsuchiya and T. Tsuchiya thank JSPS for research fellowships. Research supported by NSF/EAR 0135533 (COMPRES), 0230319.
MR21A-04 08:45h
Theoretical and experimental discovery of a CaIrO$_{3}$-type MgSiO$_{3}$ phase in Earth's D" layer
Recent seismic studies have shown that the region above the core-mantle boundary (CMB), called the D" layer, contains strong seismic anomalies, such as seismic discontinuity, anisotropy, and anticorrelation between the shear and bulk sound velocities. For a long time, mineralogists and petrologists believed that the Earth's lower mantle is composed mainly of Fe- and Al-bearing MgSiO$_{3}$ perovskite. However, a high-pressure experimental study predicted a phase transition of MgSiO$_{3}$ perovskite using an analogy of the phase transition sequence of iron oxide, Fe$_{2}$O$_{3}$. A new high-pressure phase of Fe$_{2}$O$_{3}$ with an orthorhombic (Cmcm) CaIrO$_{3}$-type structure was stable above 60 GPa [1]. Recently, this idea was confirmed simultaneously by recent theoretical and experimental studies [e.g., 2]. We explored this idea using {\it ab initio} simulations and high-pressure X-ray diffraction experiments using a laser-heated diamond anvil cell. Theoretical calculations were based on density functional theory within the local density approximation (LDA) and the generalised gradient approximation (GGA). High-pressure experiments were performed using an angle-dispersive X-ray diffraction technique at the synchrotron beam line BL10XU, SPring-8 in Japan. The transition depth from the orthorhombic MgSiO$_{3}$ perovskite to CaIrO$_{3}$-type phase matched that of the observed seismic discontinuity at D" layer. A Clapeyron slope of this transition, calculated to be 8-10 MPa/K, agrees with both seismic observations and experimental results. The elastic properties of the CaIrO$_{3}$-type phase explain most of the previously puzzling properties of the D" layer: its seismic anisotropy, strongly undulating shear-wave discontinuity at its top, and possibly the anticorrelation between shear and bulk sound velocities. This new phase is therefore likely to be a major Earth-forming mineral, and its discovery will change our understanding of the deep Earth's interior. We will also discuss our latest, quite unexpected, results on this phase. [1] Ono et al, J. Phys. Chem. Solid 65 (2004)1527-1530. [2] Oganov and Ono, Nature 430 (2004) 445-448.
MR21A-05 INVITED 09:00h
Dynamical Consequences of the Post-Perovksite Phase Transition and Implications for Radiative Heat Transfer in Deep Mantle.
The recent discovery of a phase transition from the perovskite ( PV) to post-perovskite ( PPV ) in the deep lower mantle has important implications for plume dynamics and physical properties in the lower mantle. The proximity of this exothermic phase transition to the core-mantle boundary would exert a profound influence on the generation of lower mantle plumes. We have employed a two-dimensional Cartesian model within the framework of the extended Boussinesq approximation for handling phase transitions for both the spinel to perovskite and the PV to PPV transition. An aspect-ratio 10 box is employed for variable viscosity. For the backgorund state we have used a depth-dependent thermal expansivity and a depth-dependent viscosity with a high viscosity peak in the lower mantle . Both constant and temperature-dependent viscosity models have been considered. The important distinction we found here is the thermal conductivity model. For a constant thermal conductivity model we obtained many small-scale instabilities and the lack of mega plume-like structures. This result does not depend on the type of viscosity employed. However, coherent large upwellings are obtained with the introduction of a temperature-dependent power-law radiative thermal conductivity ( Matyska et al., 1994). Cold downwellings have great difficulties reaching the base of the mantle because of efficient thermal assimilation by radiative transfer. Our findings point to the potential importance of radiative transfer in the deep mantle suggested recently ( Badro et al., 2004 ) , if one were to reconcile with both the existence of the PPV transition and the presence of the two mega slow velocity structures in the lower mantle, inferred from seismic tomography. Matyska, C., Moser, J. and D.A. Yuen, The potential influence of radiative heat transfer on the formation of megaplumes in the lower mantle, Earth Planetary Sci Lett., 125, 255-266, 1994. Badro, J., Rueff, J-P., Vanko, G., Monaco, G., Fiquet, G., and F. Guyot, Electronic transitions in perovskite: possible nonconvecting layers in the lower mantle, Science, 305, 383-386, 2004.
MR21A-06 09:15h
The interaction between a thermo-chemical boundary layer and the post perovskite phase change near the core-mantle boundary
Two- and three-dimensional numerical simulations of compressible, thermo-chemical mantle convection including plate-like behavior are used to investigate the effects of the perovskite to post-perovskite phase transition at around 2700 km depth on the dynamics and observational (seismological) signatures of thermochemical mantle convection. Both cases with, and without a compositionally-dense layer of subducted oceanic crust, are considered. For purely thermal convection the exothermic post-perovskite phase change destabilizes the lower thermal boundary layer, increasing the heat flow, increasing interior mantle temperature, and increasing the number and time-dependence of upwelling plumes. The resulting weak, highly time-dependent upwellings also have a smaller horizontal spacing than the plumes that occur in the absence of the phase transition [Nakagawa and Tackley, 2004 published in GRL]. In the present study, we progress to a more realistic model that includes plate-like behavior of the upper boundary layer, which may alter the character of the downwellings reaching the D'' region, and the possibility of a layer of segregated subducted crust above the CMB, which, due to its different perovskite fraction, may result in interesting interactions as may occur in the 660-720 km depth region. Our analysis focuses on seismological observables, including topography of the discontinuity, the possibility of a second crossing of the discontinuity at greater depth [Hernlund et al., 2004], the temperature gradient in the lower 40 km which might be linked to ULVZ presence of absence, and horizontal wavelengths and planform. Also studied is the effect on heat transfer across the CMB, which is important for understanding core evolution.
MR21A-07 09:30h
Is there a Post-Perovskite layer beneath the African Superplume?
Most tomographic models display relatively fast velocity patches beneath regions involving past subduction. Synthetic seismograms generated for such models will produce a small triplication matching S, Scd and ScS data sets, if we allow a small jump in velocity of 1% or more, depending on the global model (Sidorin et al, 1999; Ni et al., 1999). The variation in observations from patch to patch suggests a positive Clapeyron Slope ($\gamma$) of about 6 megapasscal per Kelvin which produces the layer thickness varying from 300km (fast regions) to 60 km (slow regions). Assuming a larger $\gamma$ can eliminate the layer in the slower regions. \par Here we attempt to detect such a boundary beneath the African Superplume by examining data along a 2D section from South Sandwich Island to Arabia, including several arrays. Close examination of paths where ScS is nearly in the middle of the structure reveal simple looking waveforms where (ScS-S) waveform observations are similar to PREM synthetics. Perhaps, the jump associated with a possible phase change in velocity occurs near the onset of the thin CMB thermal-boundary-layer with a negative change producing near cancellation, or it is not present. Since this large structure seems unique in terms of relative sharp walls (lateral boundaries), it may have unique chemistry as well.
MR21A-08 09:45h
The post perovskite phase transition: implications for shear-induced poro-viscoelastic interactions at D''
The extent and nature of material transfer from the fluid outer core across the CMB into D'' remains enigmatic. Seismological and other geophysical evidence suggest that small scale heterogeneity and anisotropy on a scale of c. 10 km at the base of D'' may be due to the presence of a melt phase of undetermined composition, while geochemical evidence suggests some plume-related magmas may comprise 1-5% core derived material. Motivated by these observations, we present a mechanical model for the extraction of core melt upwards across the CMB into the mantle side region of D'' and subsequent interaction with the post-perovskite (PPV) phase transition. A strong requirement of the model is that the D'' region behaves as a poro-viscoelastic material on timescales comparable with the characteristic Maxwell relaxation time. Upwelling of outer core fluid can in principle be driven by a number of external deformation mechanisms including stresses associated with the new phase transition, loading by cold downwellings, or instabilities in the rotating outer core due to a `bumpy' CMB. Using new {\it ab-initio} estimates of the PPV elastic constants, we show that shear-enhanced dilation of a poro-viscoelastic D'' matrix has the potential to drive local fluid flow in the elastic limit on a timescale of 1-100 years. If loading rates locally exceed c. 10$^{-10}$ s$^{-1}$, calculated core metal flow rates are of the order 10$^{-4}$ ms$^{-1}$, far in excess of previous estimates based on static percolation or capillary flow. Provided this minimum required loading rate is maintained, core liquid metal could in principle be transported several 10's of km upwards into D'' on geologically short timescales, resulting in local rapid changes in electrical conductivity. Given the strong assumed dependence of the PPV phase transition on composition, periodic localised excursions of infiltrating Fe rich liquid metal from the outer core into the lowermost mantle may have profoundly affected the positioning of this transition over time.