Abstract Here we determined the P wave velocity of liquid Fe 84 Si 16 up to 56 GPa based on inelastic X‐ray scattering measurements in a laser‐heated diamond‐anvil cell. We found that silicon significantly increases the P wave velocity of liquid iron under high pressure. The equation of state (EoS) of liquid Fe‐Si is obtained from the present sound velocity measurements. When extrapolating the present data to core pressures with the EoS, the silicon concentration must be limited to less than wt.% in the outer core to explain its P wave velocity. In contrast, if silicon is the sole light element, to wt.% is necessary to account for the outer core density deficit. Thus, a liquid Fe‐Si cannot explain both the density and sound velocity simultaneously, suggesting that silicon is not the predominant light element in the core. Recent core formation models have predicted that the initial core contained 2 to 9 wt.% Si, which makes the P wave velocity much faster than that observed in the present‐day outer core. The core may have changed its composition by crystallizing SiO 2 or (Mg,Fe)SiO 3 . It is possible that the initial core with <4 wt.% Si and ~5 wt.% O segregated from silicate under moderately oxidizing conditions and its composition evolved into a silicon‐poor one (<1.9 wt.% Si), which is compatible with the seismological observations. Such moderate Si content in the original core could also account for the high MgO/SiO 2 and 30 Si/ 28 Si ratios in Earth's mantle relative to chondritic values.
Abstract Melting phase relations and crystal‐melt element partitioning in a mid‐oceanic ridge basalt bulk composition were studied to 135 GPa using laser‐heated diamond‐anvil cell techniques. Using field‐emission‐type electron microprobe (FE‐EPMA), transmission electron microscope (TEM), and laser ablation‐inductively‐coupled plasma mass spectrometer (LA‐ICP‐MS), we obtained comprehensive analyses of major and trace elements in coexisting melt and solid phases. CaSiO 3 ‐perovskite (Ca‐pv) was found to be the liquidus phase throughout the lower mantle pressure range. Whereas silica, followed by Mg‐perovskite, are the second and third crystallizing phases to pressures exceeding 100 GPa, postperovskite, closely followed by seifertite, succeed Ca‐pv at 135 GPa. The partitioning of trace elements between Ca‐pv and melts exhibited a strong pressure effect, possibly due to a combination of high compressibility of cations compared to the lattice site in Ca‐pv and melt compressional effects. The Ca‐pv/melt partition coefficients for Na and K ( D Na and D K ) increase with increasing pressure, with D Na close to unity and D K greater than unity at lowermost mantle pressures. Also, D Nd becomes larger (or identical within uncertainty) than D Sm in the deep lower mantle. Partial melt formed by 51% partial melting of mid‐oceanic ridge basalt at 135 GPa showed marked iron‐enrichment and should thus have negative buoyancy at the base of the mantle. The density of residual solid is almost identical to the PREM density, and therefore, it is likely to be involved in mantle convection and recycled to the surface.
Understanding effects of non-hydrostatic pressure on phase transitions in minerals relevant to the Earth's mantle is important to translate the observable seismic signals to not directly observable mineralogical models for the deep Earth. SiO2 can occur as a free phase in subducting slabs, which contain sedimentary layers and/or mid-ocean-ridge basalts. In this study, we report on the effect of deviatoric strain on the pressure-induced phase transition in SiO2 and its consequences on the seismic signal.
We have determined subsolidus phase relations in the Fe–FeS system up to 271 GPa using laser‐heated diamond‐anvil cell techniques. In situ synchrotron X‐ray diffraction (XRD) measurements performed at high pressure and high temperature demonstrate the coexistence of hexagonal close‐packed (hcp) Fe and tetragonal Fe 3 S up to 241 GPa and 2510 K. In contrast, the XRD data obtained above 250 GPa show that the hcp phase coexists with the CsCl (B2)‐type phase for three different Fe–S bulk compositions (10, 16, and 20 atm% S). Furthermore, chemical analyses using a scanning transmission electron microscope on a retrieved sample indicate that Fe 3 S sample decomposes into two phases at 271 GPa and 2530 K, consistent with the XRD data. Theory predicts the presence of extensive solid solution between Fe and FeS at inner core conditions, whereas our results suggest that the Fe–FeS system remains eutectic at least to 271 GPa.
