Spin crossover in Fe₂SiO₄ liquid at high pressure

Abstract

We combine spin-polarized density functional theory with first principle molecular dynamics (FPMD) to study the spin crossover in liquid Fe 2 SiO 4 , up to 300 GPa and 6000 K. In contrast to the much sharper transition seen in crystals, we find that the high-to low-spin transition occurs over a very broad pressure interval (>200 GPa) due to structural disorder in the liquid. We find excellent agreement with the experimental Hugoniot. We combine our results with previous FPMD calculations to derive the partial molar volumes of the oxide components MgO, FeO, and SiO 2 . We find that eutectic melts in the MgO-FeO-SiO 2 system are denser than coexisting solids in the bottom 600 km of Earth's mantle.

Spin crossover in liquid (Mg,Fe)O at extreme conditions

“…With increasing pressure, the liquid is driven into a mixed-spin phase of coexisting high-spin and low-spin ions, with the mean moment approaching approximately one half that of the highspin value towards P = 200 GPa. The crossover thus appears extremely broad, as was found recently for another Fe-bearing liquid planetary material Fe 2 SiO 4 [26] from direct DFT MD without considering the full Gibbs free energy. The breadth of the spin crossover region of at least hundreds of GPa in the liquid contrasts with that in the crystalline phase where the crossover is much sharper: High-spin and low-spin Fe coexist over a range of pressure of ∼50 to 150 GPa.…”

“…Spin crossover has been shown to accompany a solid-to-liquid phase transition in some * eero.holmstrom@aalto.fi systems [19], and a different threshold temperature for spin crossover in the solid phase compared to a liquid crystal phase of the same complex has been reported [20]. While the crossover in crystalline planetary materials has received much attention in recent years [21][22][23][24][25], the question of a spin crossover in the liquid phase of these materials is little explored to date [7,26].…”

Thermodynamics of the MgO‐FeO‐SiO2system up to 140 GPa: Application to the crystallization of Earth's magma ocean

“…Our prediction for the density of (FeO) l (blue dashed line) is in good agreement with recent calculations [ Ramo and Stixrude , ]. The blue shading depicts the uncertainties that our inversion predicts.…”

“…However, our liquid Fe 2 SiO 4 (purple dashed line of Figure ) remains significantly lighter that of Thomas et al [] (purple dotted line) along a 10 K GPa −1 geotherm. This seems contradictory with the fact that along Hugoniot paths, the molecular dynamics simulations of Ramo and Stixrude [] are said to be in good agreement with the shock experiments of Thomas et al []. Even using large Margules, our model has difficulties in predicting a (Fe 2 SiO 4 ) l with a density close to that of (FeO) l .…”

“… The melting curve of wüstite has been measured by Fischer and Campbell []. This observation constrains the reference potential of (FeO) l , along the P ( T ) melting curve and its relative value with respect to in the P ‐ T space. The density of liquid Fe 2 SiO 4 has been measured by Thomas et al [] using shock wave experiments and confirmed by molecular dynamics simulations of Ramo and Stixrude []. This observation constrains the EoS of (FeO) l and the Margules pressure derivative related to the mixing of (FeO) l with (SiO 2 ) l , i.e., the excess mixing volume. The Mg/Fe partitioning between melt and bridgmanite …”