Isoji Miyagi

A new hydrous silicate, a water reservoir, in the upper part of the lower mantle

We synthesized a new hydrous silicate in the pressure range from 20 GPa to 24 GPa at 800–1300°C. This phase, named tentatively as phase G, has a hexagonal unit cell, a=4.790 (3) A and c=4.344 (3) A, and V=86.3 (2) A³ and the atomic ratio Mg/Si=0.66±0.03. SIMS analysis revealed that it contains 14.5±2.0wt% water. Phase G has a chemical formula of Mg1.14Si1.73H2.81O6 and a density of 3.37g/cm³. Phase G coexists with periclase and Mg-perovskite under the lower mantle conditions, and thus it can be a reservoir of water in cold slabs penetrating into the lower mantle.

Decomposition of K-amphibole at high pressures and implications for subduction zone volcanism

Abstract The stability of K-amphibole has been studied at pressures 12.6–16.5 GPa, and at temperatures 940–1450°C. K-amphibole decomposes into an assemblage of clinoenstatite+diopside+an unknown phase (X)+stishovite+fluid at pressures above 16 GPa and temperatures below 1200°C. The phase boundary has a negative Clapeyron slope, and the high temperature assemblage is clinoenstatite+diopside+X+wadeite-type K 2 Si 4 O 9 +fluid at pressures 14–16 GPa. The X phase has a cation ratio of approximately K:Mg:Si=1:2:2, and contains 1.7±0.1 wt.% H 2 O as determined by SIMS measurements, leading to a formula of K 4 Mg 8 Si 8 O 25 (OH) 2 . The present results suggest that the decomposition of K-amphibole in the dragged hydrous peridotite layer at the base of the mantle wedge may produce certain amounts of H 2 O-rich fluid at 14–16 GPa (∼450 km depth), while some of H 2 O is trapped in the new hydrous phase X and is further carried into deeper regions of the mantle. The aqueous fluid released by the decomposition of K-amphibole should react with β -phase to form hydrous β -phase in the mantle transition region. Thus, the dehydration of K-amphibole would not cause any volcanic activities in the back arc regions, in contrast to the dehydrations of amphibole, chlorite and phlogopite, which are presumably responsible for the first and the second volcanic chains. However, some of the volcanic activities such as in Muriah, Indonesia, may be related to the dehydration of K-amphibole in unusually hot regions above the subducting slab.

Reply [to “Comment on ‘A new hydrous silicate, a water reservoir, in the upper part of the lower mantle’ by Eiji Ohtani, Hiroki Mizobata, Yasuhiro Kudoh, Toshiro Nagase, Haruo Arashi, Hisayoshi Yurimoto, and Isoji Miyagi”]

Phase G reported by Ohtani et al. [1997] and Kudoh et al. [1997a] is identical with the new phase D reported by Kuroda and Irifune [1998] and Yang et al. [1997]. However, there is ambiguity on existence of phase D defined by Liu [1987]. There are following possibilities on existence of Lius phase D; (1) the product was a mixture of phase G and the other contaminant phases, (2) it realy exists stably in different pressure and temperature conditions from those studied by our works, or (3) the phase is a metastable phase formed due to a short run duration and/or no sealing of water in the diamond anvil cell. Therefore, it is not appropriate to use the name phase D for the dense hydrous magnesium silicate phase G reported by us [Ohtani et al., 1997].