Fe-periclase reactivity at Earth's lower mantle conditions: Ab-initio geochemical modelling
- Authors: Merli, M., Bonadiman, C., Diella, V., Sciascia, L., Pavese, A.
- Publication year: 2017
- Type: Articolo in rivista (Articolo in rivista)
- OA Link: http://hdl.handle.net/10447/243157
Abstract
Intrinsic and extrinsic stability of the (Mg,Fe)O solid mixture in the Fe-Mg-Si-O system at high P, T conditions relevant to the Earth’s mantle is investigated by the combination of quantum mechanical calculations (Hartree- 26 Fock/DFT hybrid scheme), cluster expansion techniques and statistical thermodynamics. Iron in the (Mg,Fe)O binary mixture is assumed to be either in the low spin (LS) or in the high spin (HS) state. Un-mixing at solid state is observed only for the LS condition in the 23–42 GPa pressure range, whereas HS does not give rise to un-mixing. LS (Mg,Fe)O un-mixings are shown to be able to incorporate iron by subsolidus reactions with a reservoir of a virtual bridgmanite composition, for a maximum total enrichment of 0.22 FeO. At very high P (up to 130/3150 GPa/K), a predominant (0.7 phase proportion), iron-rich Fe-periclase mixture (Mg0.50Fe0.50)O is formed, and it coexists, at constrained phase composition conditions, with two iron-poor assemblages [(Mg0.90Fe0.10)O and (Mg0.825Fe0.175)O]. These theoretical results agree with the compositional variability and frequency of occurrence observed in lower mantle Fe-periclase from diamond inclusions and from HP-HT synthesis products. The density difference among the Fe-periclase phases increases up to 10%, between 24 and 130 GPa. The calculated bulk Fe/Mg partitioning coefficient between the bridgmanite reservoir and Fe-periclase, Kd, is 0.64 at 24 GPa; it then drops to 0.19 at 80 GPa, and becomes quasi-invariant (0.18–0.16) in the lowermost portion of the Earth’s mantle (80– 130 GPa). These Kd-values represent an approximate estimate for the Fe/Mg-partitioning between actual bridgmanite and Fe-periclase. Consequently, our Kd-values agree with experimental measurements and theoretical determinations, hinting that iron preferentially dissolves in periclase with respect to all the other iron-bearing phases of the lower mantle. The continuous change up to 80 GPa (2000 km depth) of the products (compositions and phase proportions) over the MgO-FeO binary causes geochemical heterogeneities throughout the lower mantle, but it does not give rise to any sharp discontinuity. In this view, anomalies like the ULVZs, explained with a local and abrupt change of density, do not seem primarily ascribable to the mixing behavior and reactivity of (Mg,Fe)O at subsolidus.