A New Phase Change in Earth's Deep Mantle: What Does This Mean for the Rest of Us? I
Presiding: D Yuen, University of Minnesota; T Lay, University of California, Santa Cruz
U31A-01 08:30h
Ab Initio Molecular Dynamics Calculations on the High Temperature and Pressure Properties of the Postperovskite Phase and the Effect of Al3+, Fe2+ and Fe3+ on its Elasticity and Stability
We have used ab initio molecular dynamics to calculate the elastic properties of the MgSiO3 postperovskite phase under the appropriate conditions of the lowermost mantle. In addition, we have calculated the properties of the FeSiO3 and Al2O3 postperovskite end-members in order to evaluate the effect of Fe and Al on the stability and elastic properties of the postperovskite phase. We find the following: a) In contrast to perovskite itself, shear-wave anisotropy does not change appreciably with temperature. b) A transversely isotropic aggregate of postperovskite aligned with a vertical [010] can produce VSH > VSV of about 1 to 2%, however, this would require about 50% alignment of the grains to account for the seismic observations. A far greater degree of anisotopy (>15%) is found when the symmetry axis is [001]. This would require a much smaller amount of crystal alignment, but also a slip system that is not the intuitively obvious. c) The postperovskite phase has an anomalously low value for dlogVS/dlogVP of about 1. This is in contrast to perovskite (about 2.0) and values from seismological observations (2 to 3). d) We find that Al2O3 has little effect on the relative stabilities of perovskite and postperovskite. However, the Al2O3 postperovskite phase shows a remarkable amount of shear anisotropy. e) Calculations on Fe bearing perovskites and postperovskites suggest that Fe strongly stabilises postperovskite, in agreement with experiments. f) We also show how the postperovskite phase can account for the apparent strength and depth of the observed P- and S-wave D" discontinuity.
U31A-02 08:45h
Effects of chemistry on the properties and transformation of perovskite to the post-perovskite phases with implications for the lowermost mantle
We perform 0K static first-principles calculations on the perovskite (pv) and post-perovskite (ppv) phases in the MgSiO3 - FeSiO3 - Al2O3 system up to lower mantle pressures. We determine the relative stability and the elastic properties of the pure and some intermediate terms and discuss some implications for the Earth's lower mantle. We show that the addition of Fe2+ in MgSiO3 considerably decreases the pv-ppv transition pressure. The ppv phase of FeSiO3 is stable at all pressures with respect to the pv phase. FeSiO3 ppv is metallic, in antiferromagnetic configuration, and stable with respect to oxides (FeO and SiO2) above 115-120 GPa. The addition of Fe also increases the bulk modulus and decreases the shear modulus of MgSiO3. The addition of Al2O3 in MgSiO3 slightly increases the pv-ppv transition. The pv phase of Al2O3 is unstable at all pressures with respect to corundum, Rh2O3(II) and ppv structures. Both the bulk and the shear modulus of MgSiO3 pv and ppv decrease with the increase of Al content. Both Fe and Al decrease the seismic wave velocities of MgSiO3 pv and ppv. Our calculations suggest that the lower mantle might be locally enriched in Fe, forming low and ultra-low velocity zones that ensure the electromagnetic coupling between the mantle and the outer core.
U31A-03 09:00h
Impact of the Post-Perovskite Transition on the Structure of D
The recently discovered phase transition of magnesium perovskite to a high pressure post-perovskite (PPV) structure has a direct impact on the complex constitution of the D" layer directly above the core mantle boundary. Strong lateral variations in seismic wave velocities in D" resulting in seismic wave diffraction, observed as PKIKP precursors (Cleary and Haddon,1972, van den Berg et al, 1978), could be connected with the PPV transition. Besides seismic wave velocity, thermal conductivity is also material dependent as shown in (Giesting et al., 2004) and (Badro et al., 2004). Phase dependent thermal conductivity therefore results in complex structures of thermo-physical parameters of the D" layer and a change in the style of mantle convection expressed in the propensity of plume generation from the CMB (Matyska and Yuen, 2004). We have investigated the impact of the new phase transition in mantle convection models including thermal coupling of the core. We apply an extended Boussinesq finite element model including latent heat production of the exothermic PPV phase transition. We apply a continuous representation of the dominant mineral phase in the description of the phase dependent thermo-physical parameters. The conductivity is adapted from the composite model of Hofmeister (1999), where we apply a phase dependent control parameter scaling the radiative conductivity component. Radiative thermal conductivity depends strongly on the temperature and in the post-perovskite phase even more so, because of the high-spin to low-spin transition as hypothesized by Badro et al. (2004). This way our model accomodates strong lateral variations of the phase boundary such that isolated patches of the high pressure phase show up above the CMB for a sufficiently high initial core temperature and high values of the Clapeyron slope, like the reported 7.5 MPa/K for the PPV transition (Tsuchiya, 2004). These patches are reflected in strong lateral variations of the phase dependent parameters which may explain the highly irregular character of the D" layer. The results also show strong space-time variations of the core/mantle heat flux of up to 100% of the background value. Such fluctuations would be expected to have a significant impact on geodynamo processes.
U31A-04 INVITED 09:15h
Thermal Conductivity in the Deepest Mantle
The assumed dependence of thermal conductivity (k) on temperature strongly influences results from mantle convection models due to feedback in the temperature equation. To estimate k of post-perovskite, analogues are used. Experimental measurements of thermal diffusivity (D)using a laser-flash apparatus show that the lattice component(k = D x density x heat capacity) for minerals becomes independent of temperature above ca 1200-1900 K. This behavior is seen in 10 crystal structures, glasses, and lavas. Independent of phase, D asymptotes to ca 0.7 mm2/s. Heat capacity of 3R and density from PREM give klat near 1 W/m-K at deep mantle temperatures. Theoretically, klat(P,T)= [1+K'P/K]klat(T) where K is bulk modulus, providing ca 5 W/m-K near the core mantle boundary. Radiative transfer is important at high T. To first order, this process is pressure independent because absorption and emission characteristics roughly cancel. Important parameters are temperature, grain-size (d) due to scattering, and the absorption coefficient (A). Spectra of real materials are frequency dependent with intensities proportional to Fe content. As post-perovskite spectra are unknown, asymptotic limits are considered for expected d = 1 mm. For small, weakly absorbing grains with A roughly constant, kgrey = (16/3)d2 An2ST3 where S is the Stefan-Boltzmann constant and n is index of refraction. For mantle garnets, A nears 0.2/mm in the visible, which should proxy for the similar site in post-perovskite, providing krad about 2 W/m-K at 2500 K for a transparent mantle (low Fe content, any spin state). For dark grains (large dA, e.g., high Fe content), kgrey =(16/3)n2ST3/A. Estimating A as 10/mm, gives 1 W/m-K. Calculations based on spectral data have a weaker T dependence as neither the UV nor IR spectral regions participate, but other factors increase k, and thus our asymptotic limits constrain k for the deepest mantle with small grain size. Although the values of klat and krad are similar, the strong temperature dependence of radiative transfer means that this process controls mantle convection, and formation of large plumes. In particular, radiative transfer is relatively impeded in Fe-rich regions, leading to locally warmer temperatures and possibly upwellings.
U31A-05 09:30h
Effects of the post-perovskite phase change on the thermal evolution of the Earth's core
The heat flow through the core-mantle boundary is a key quantity for understanding the thermal evolution of the Earth??s core, as the geodynamo is presumably strongly affected by the temporal variation of CMB heat flow. A major challenge is to understand how this heat flux can have remained high enough to power the geodynamo over geological history without resulting in larger-than-observed cooling of the core and growth of the inner core. This problem has been approached using various coupled models of mantle convection and core heat balance: (1) Simple isochemical models using parameterized mantle convection and core heat balance [Buffett, 2002; Labrosse, 2003; Nimmo et al., 2004] have too rapid core cooling hence a too large inner core, (2) models with a global layer of dense material above the CMB [McNamara and van Keken, 2000] have a CMB heat flow that is too low for the geodynamo to occur, but (3) with discontinuous chemical layering [Nakagawa and Tackley, 2004], viable evolution solutions are obtained, with the best scenarios requiring 100-200 ppm radioactive potassium in the core [Nakagawa et al., 2004]. Recently, using high pressure experiments and ab initio calculations, a new perovskite to post-perovskite phase change was discovered near the CMB [Murakami et al., 2004; Oganov and Ono, 2004]. Dynamically, such a phase change results in small-scale instabilities in lower thermal boundary layer and higher CMB heat flow [Nakagawa and Tackley, 2004]. Furthermore, if the CMB is in the perovskite stability field then a double-crossing of the phase boundary may occur [Hernlund et al., 2005]. As the core and mantle cool with time, the location and thickness of the layer of post-perovskite phase will change [Nakagawa and Tackley, 2005]. In this study, a coupled model of thermo-chemical mantle convection including the post-perovskite phase change and a global heat balance in the core based on the entropy variation is used to assess the CMB heat flow, thermo-chemical structures in the D?? region (dense piles due to basaltic component and double-crossing due to the post-perovskite phase change) and their influence on the thermal history of the Earth??s core. The required amount of potassium in the core will be also discussed in this presentation.
U31A-06 09:45h
Initiation of Thermo-Chemical Plumes from the Post-Perovskite Layer
The recent discovery of a stability field for the post-perovskite phase in the deep mantle has strong consequences for the origin and petrology of lower-mantle plumes that develop in proximity of the core-mantle boundary. In order to study dynamics, geometry and mineralogical composition of such plumes we have implemented a self-consistent petrological-thermomechanical model of the lower mantle, wherein we have included density changes due to phase transitions involving post-perovskite in both the continuity and the momentum equations, and latent heat of mineral reactions and adiabatic and shear heating in the energy conservation equation. The model accounts for dependence of thermal conductivity, which has a strong radiative component, on temperature, pressure and composition (decreasing with increasing Fe-content). The petrological model covers the entire mantle and is derived by free energy minimization together with estimates for the thermodynamic properties of post-perovskite, perovskite, wuestite and various mid-mantle silicates. This approach precisely quantifies the thermodynamic and mechanical influence of the post-perovskite and perovskite forming phase transformations, which depend on temperature, pressure and composition .We have applied a well-tested marker-in-cell method and conservative finite-differences to solve governing equations in 2-D. During generation of thermal-chemical plumes temperature and mantle composition counteract introducing complexity in plume dynamics. We will test the various possibilities and conditions for mega- plume structures and investigate the stable and collapsing (oscillating) end-members that are expected to limit the range of behavior that may arise as result of lower mantle complexity. Also of interest is the fate of the descending down-wellings, as they will be heated up and softened by the post-perovskite transition.