HR: 08:15h
AN: MR11A-02    [Abstracts]
TI: Density of Hydrous Ultramafic Silicate Melt under the Earth's Deep Upper Mantle Conditions
AU: * Jing, Z
EM: zhicheng.jing@yale.edu
AF: Yale University, Department of Geology & Geophysics, New Haven, CT 06511 United States
AU: Matsukage, K N
EM: kmatsu@mx.ibaraki.ac.jp
AF: Yale University, Department of Geology & Geophysics, New Haven, CT 06511 United States
AU: Matsukage, K N
EM: kmatsu@mx.ibaraki.ac.jp
AF: Ibaraki University, Department of Environmental Sciences, Mito, Ibaraki, 310-0056 Japan
AU: Karato, S
EM: shun-ichiro.karato@yale.edu
AF: Yale University, Department of Geology & Geophysics, New Haven, CT 06511 United States
AB: Density of silicate melts is a critical material property in our understanding of geochemical evolution of the Earth. Previous studies (e.g., Agee & Walker 1993; Suzuki et al., 1995) showed that the density of dry silicate melts can be higher than that of surrounding solids under deep upper mantle conditions. However, melts formed under such conditions likely contain some water (Bercovici & Karato, 2003), which will reduce the melt density. In this study, we performed sink/float experiments between 10 and 14GPa and at 2173K to determine the density of hydrous ultramafic silicate melts, using a Kawai-type multianvil apparatus. We choose a target melt composition based on the experimental study by Litasov & Ohtani (2002). With this chemical composition, olivine reacts with the melt above the liquidus, so we used diamond as the density marker. However, diamond is much denser than a melt with a typical mantle like Fe/Mg ratio. Therefore in this study we determined the density of melts with high Fe contents, and from the relation between Fe content and melt density, we inferred the melt density with Earth-like Fe/Mg. Four Fe-rich compositions with 5wt% water and different iron content were chosen as starting materials. Density crossovers between melts and diamond were observed for all compositions. The densities of four melts at 14GPa and 2173K were calculated using the Birch-Murnaghan equation of state. The pressure derivative of isothermal bulk modulus (K_t^') of the melts was estimated to be around 4. The density of mantle melt with mantle value of content and 5wt% water at 14GPa, 2173K was extrapolated to be ~3.42±0.4g/cm³. We compared our density results for hydrous melts with previous results on dry melts and found that water is more compressible than other components in melt. The estimated partial molar volume of water at 14GPa and 2173K is ~8±2cm³/mol, which is significantly lower than the value at low pressures. The conditions under which the density crossover between hydrous melts and the surrounding solids at 410 km were determined for different water content, temperature, and partial molar volume of water. Results showed that the critical water content is ~6wt% for T=1800K and V̅_H2O=8cm³/mol, although the uncertainties are still large. We recognize, however, that these initial results likely underestimated the melt density in Earth's mantle. The low-pressure data indicate that oxygen fugacity has important effects on melt density (e.g., Kress & Carmichael, 1991). In our experiments, we used a rhenium capsule to minimize the chemical reaction, which defines a relatively high oxygen fugacity. At lower oxygen fugacities, the oxidation state of iron is likely different, which will increase the density of melts. Calculations showed that at room pressure the density difference between sample buffered by Re-ReO₂ and by Mo-MoO₂ is ~0.5-2%, which is important to determine the conditions for density crossover. We will report on new results on the influence of oxygen fugacity on melt density which will provide a better estimate for the conditions for density cross-over.
DE: 8145 Physics of magma and magma bodies
DE: 8439 Physics and chemistry of magma bodies
SC: Mineral and Rock Physics [MR]
MN: Fall Meeting 2005