Why post-perovskite should have a low viscosity and its dynamical consequences .

In the past 4 years the discovery of the post-perovskite (PPV) phase of ( Mg, Fe)SiO3 [1,2] has revolutionized the geosciences. Previous work on PPV has focussed mainly on aspects of equation of state, such as density, elastic constants, Clapeyron slopes and temperature intercepts. These studies have brought a consistent picture of the thermal and seismic structure of the lower mantle. However, the viscosity of PPV holds the key to any sort of predictability concerning the instabilities generated in the D" layer. However, viscosities of solids are extremely difficult to measure experimentally or to predict theoretically, due to a disparity of time and length scales and extremely small deformation rates. We propose an alternative approach, based on a combination of evidence from mineral physics, geomagnetism and geodynamics, to conclude that PPV viscosity is about 2-3 orders of magnitude less than that for perovskite and this has important dynamical consequences. One can show that the viscosity based on diffusion is proportional to the inverse of the electrical conductivity based on ionic processes of oxygen anions. Based on analogy with Al2O3, for which direct measurements of the conductivity under pressure were available since 1996 [3], Oganov and Ono proposed in 2005 that PPV should have a higher (by ~2 orders of magnitude) ionic electrical conductivity than perovskite [4,5]. This was confirmed by recent experiments of Ohta on (Mg,Fe)SiO3 [6] and suggests that there should be a concomitant decrease of the PPV viscosity. Geoid modelling work with laterally varying viscosity also reveals that the the long wavelength regions of the D" associated with descending slabs with PPV to have a viscosity around 100 to 1000 times smaller than the viscosity associated with regions having a lower seismic velocity anomalies. By means of a two-dimensional finite-element model, we show that the presence of a very low viscosity PPV lens above the CMB exerts a strong influence on the heat flux from the core, inducing sharper heat flow peaks than models without the PPV transition. The time dependence of the heat flow is also more chaotic with PPV transition. This finding may have important implications for core-mantle coupling and the character of the Earth's magnetic field. [1] Murakami M., Hirose K., Kawamura K., Sata N., Ohishi Y. (2004). Science 304, 855-858. [2] Oganov A.R. & Ono S. (2004). Nature 430, 445-448. [3] Weir S.T., Mitchell A.C., Nellis W.J. (1996). J. Appl. Phys. 80, 1522-1525. [4] Oganov A.R., Ono S. (2005). Proc. Natl. Acad. Sci. 102, 10828-10831. [5] Ono S., Oganov A.R., Koyama T., Shimizu H. (2006). Earth Planet. Sci. Lett. 246, 326-335. [6] Ohta K, Onoda S, Hirose K, Sinmyo R., Shimizu K., Sata N., Ohishi Y., Yasuhara A. (2008). Science 320, 89-91.