*Kazutaka Yamaguchi1, Takaaki Kawazoe1, Toru Inoue1, Takeshi Sakai2
(1.Hiroshima Univ., 2.Ehime Univ. )
Keywords:Wadsleyite, Ferric iron, Oxygen fugacity, Melting temperature, Mantle transition zone
It is known that water is supplied into the Earth's interior by subduction of oceanic plates. Water is not the only component brought to the Earth's interior by subduction of the oceanic plates. Oxygen is also brought into the Earth's interior. For example, ~200 kg/yr of oxygen is estimated to be brought to the Earth's interior at the Mariana subduction zone (Brounce et al., 2019). Some of the subducted oxygen are recycled to the Earth’s surface by the island-arc and back–arc volcanisms, but most of the subducted oxygen are supplied to the deep mantle. The amount is estimated to be ~180 kg/yr. In addition, Fe3+/ΣFe in melt inclusions ejected from hot-spot volcanoes has been measured. From the measurements, the maximum Fe3+/ΣFe of the melt inclusions from the hot-spot volcanoes was found to be 0.4 (Moussallam et al., 2014). This value is higher than the global MORB average, indicating that the oxygen fugacity in the Earth's interior is heterogeneous. Moreover, seismological observations indicate the existence of a low-velocity layer of seismic waves around 410 km depth (Tauzin et al., 2010). This low-velocity layer is not localized but widely distributed. This low-velocity layer is likely to be caused by partial melting of the mantle. In this study, we investigated the effect of high oxygen fugacity on the melting temperature of wadsleyite. A Kawai-type multi-anvil apparatus was used for the high-temperature and high-pressure experiments at Hiroshima University. The starting materials were a powder and single crystals of olivine from San Carlos, Arizona. Use of the single crystals can decrease water content in recovered samples by minimizing surface adsorbed water. Temperature and pressure conditions were 1300°C to 1600°C and 14 GPa to 17 GPa, respectively. Re-ReO2 and Mo-MoO2 oxygen fugacity buffers were used to control oxygen fugacity. Au was used as a capsule material to enclose the olivine starting materials and the buffer materials. After keeping the target temperature for desired duration, the samples were quenched by turning off the heating power supply. Microstructural observations were made with a scanning electron microscope. Phase identification was performed by micro-Raman spectroscopy and micro-focus X-ray diffraction. Chemical composition was measured by EPMA. The water content of the samples was measured by FTIR. At 15 GPa and 1400°C using a powder as the starting material, no quenched melt was observed in both samples buffered with the Re-ReO2 and Mo-MoO2 buffers. At 15 GPa and 1500°C using a powder as the starting material, quenched melt was observed in the sample with oxygen fugacity controlled by Re-ReO2 buffer. In contrast, no quenched melt was observed in the sample with oxygen fugacity controlled by Mo-MoO2 buffer. The water content of the samples was 0.4 wt%. The solidus temperature of wadsleyite under low oxygen fugacity conditions is estimated to be about 2300°C (Ohtani et al., 1998). In the present experiment, wadsleyite melted under high oxygen fugacity condition at 1500°C. Therefore, the results of this study indicate that the melting temperature of wadsleyite was decreased by ~800°C. It is estimated that this 0.4 wt% water lowers the melting temperature of wadsleyite by about 200°C (Litasov and Ohtani, 2003). The additional experiment was performed at 15 GPa and 1500°C using single crystals of olivine as the starting material to minimize water in the sample. As a result, no quenched melt was observed in both samples buffered by the Re-ReO2 and Mo-MoO2 buffers. We will perform more runs with the single crystals to constrain the effect of oxygen fugacity on the melting temperature of wadsleyite. The results of this study indicate that the melting temperature of wadsleyite was decreased by ~800°C with high oxygen fugacity and 0.4 wt% water.