第七届青年地学论坛

State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang, China
*Corresponding author: F. Zhang (zhangfeiwu@vip.gyig.ac.cn)

The lower mantle is generally thought to consist of primarily three phases- magnesium silicate perovskite (MgSiO₃) in the form of bridgmanite, calcium silicate perovskite (CaSiO₃) and ferropericlase.  Of these the two perovskite structures have near identical chemical and crystal structures and thus at some point must form a single phase (CaMgSi₂O₆).  The formation of such a phase could have profound effects on the seismic, rheological and conductive properties of the lower mantle.

By using ab-initio calculations we determine that this is unlikely to happen for pure CaSiO₃ and MgSiO₃ except in hot regions of the deep mantle.  The introduction of other elements can change this story considerably however.  Ferrous iron and Ti both can induce the creation of phase mixtures while Al and ferric iron suppress it.  Mixing of these phases appears to be very fast compared to the timescales of diffusion meaning that the source, history and partitioning of defect elements introduced to the lower mantle are all important in determining its phase structure.  Thus complex histories need to be considered to determine the phase structure of the lower mantle.  Ti has a particularly strong effect and phase mixing provides a mechanism to segregate Ti into Ti-rich regions that will then phase mix.  Such an effect is not seen for ferrous iron as very large quantities of iron (>~20%) start to decrease the favourability of phase mixing.

We also predict that perovskite phase mixing should be seismically visible with similar signals as have been observed in Large Low Velocity Shear Provinces (LLSVPs). Thus we predict the possibility of strong seismic hetereogeneity in hot or Ti-rich areas of the mantle or areas with unusual histories.
 

bridgmanite,ab-initio,solid solution,LLSVPs