Reactions between molten iron and silicate melt were investigated with mixtures of pure iron and silicates (1Fe + 3MgSiO3 enstatite and 1Fe + 3(Mg0.9Fe0.1)2SiO 4 olivine in volumetric ratio) as starting materials at pressures of 10-26 GPa and temperatures of about 2500øC. The results show that a certain amount of Si (up to about 2%) dissolves in molten iron from silicate melt and that the dissolution is enhanced with increasing pressure. Many small spherical blobs composed of SiO2 and FeO present in coalesced iron grains were interpreted as quenched immiscible liquid formed during cooling. Therefore O also dissolves in molten iron under the experimental conditions. No evidence for dissolution of Mg was obtained. The present study also indicates that Si and O are important light elements of Earth's core if core segregation occurred in the deep magma ocean. The chemical evolution of Earth's core is discussed on the bases of the current core formation model and the present experimental results. that an extensive amount of O can dissolve in molten iron at temperatures higher than 1700øC. Ringwood and Hibberson [1991] examined the relative solubilities of several oxides in molten iron at 16 GPa and concluded that under conditions higher than 16 GPa and 2000øC, oxide species could be extracted from the silicate in the sequence FeO, SiO2, and MgO 1Now at Mitsubishi Material Company, Ohmiya, Japan. Paper number 94JB02645. 0148-0227/95/94JB-02645505.00 with increasing pressure and temperature. Knittle and Jeanloz [1989, 1991] showed that molten iron reacts with (Mg, Fe)SiO3 perovskite to form stishovite and an iron alloy containing Si and O at temperatures of 2700ø-3000øC and pressures of 25-70 GPa. Goarant et al. [1992] observed that molten iron and iron-sulfide extract FeO from the iron containing silicate and oxide at about 70 and 130 GPa, resulting in the dissolution of O in the molten metallic phases.According to recent planet accretion models [e.g., Sasaki and Nakazawa, 1986] the terrestrial core was formed by the sinking of molten iron through a completely or partially molten silicate layer which extended to a depth of more than 1000 km (the magma ocean). In this mechanism of the core formation the reactions between molten iron and silicate melts should have played a decisive role in determining the chemistries of both the core and the mantle. However, detailed knowledge of such reactions has been sparse so far. Pioneering works using diamond cells [Goarant et al., 1992; Jeanloz, 1989, 1991] are of a rather qualitative nature because the experiments were carried out on very small-scale samples with correspondingly large pressure and temperature gradients.In this context, Ito and Katsura [1991] showed the dissolution of Si up to about 2% in molten iron from silicate melt at 24 GPa and 2550øC using a large-volume multianvil apparatus. In this article we report the partitioning of Si between molten iron and silicate melts more systematically and suggest that Si and O would be potential light elements in the outer cor...

Partitioning of K between molten iron and silicate melt was examined at about 26 GPa and 2600°C using a mixture of pure iron and natural microcline as the starting material. The dissolution of K in molten iron up to about 0.2 wt % was confirmed. The distribution coefficient of K between molten iron and silicate melt was about 3.8×10−2 (atomic ratio). This value suggests the presence of 5 ppm K in the earth's core, yielding 3×1010 W for the present day heat generation by K.

No abstract