Thermal equation of state of superhydrous phase B to 27 GPa and 1373 K

Pressure–volume–temperature relations have been measured to 27 GPa and 1373 K for superhydrous phase B (Mg10Si3O14(OH)4) using synchrotron X-ray diffraction with a multi-anvil apparatus at the SPring-8 facility. The analysis of room-temperature data fitted to a third-order Birch–Murnaghan equation of state (EOS) yields V0 = 623.38 ± 0.39 Å3; K0 = 138.7 ± 3.0 GPa and K′ = 4.9 ± 0.3, when pressure was calibrated using the Au EOS of Anderson et al. [Anderson, O.L., Issak, D.G., Yamamoto, S., 1989. Anharmonicity and the equation of state for gold. J. Appl. Phys. 65, 1534–1543]. These values are consistent with subsequent thermal EOS analysis and previous estimations for superhydrous phase B. A fit of P–V–T data to high-temperature Birch–Murnaghan EOS yields V0 = 623.47 ± 0.37 Å3; K0 = 135.8 ± 2.6 GPa; K′ = 5.3 ± 0.2; (∂KT/∂T)P = −0.026 ± 0.003 GPa/K and zero-pressure thermal expansion α = a0 + a1T with a0 = 3.2 (1) × 10−5 K−1 and a1 = 1.2 (4) × 10−8 K−1. The Anderson–Grüneisen parameter is estimated to be δT = 5.4. A fit to the thermal pressure EOS gives V0 = 623.50 ± 0.36 Å3; K0 = 135.3 ± 2.3 GPa; K′ = 5.3 ± 0.2; (∂KT/∂T)V = −0.002 (2) GPa/K and α0 = 3.8 (2) × 10−5 K−1. The lattice dynamical approach using a Mie–Grüneisen–Debye EOS yields Grüneisen parameter γ0 = 1.33 ± 0.05 and q = 2.03 ± 0.35, if the Debye temperature θ0 is fixed at 860 K, as calculated from sound velocities. The analysis of axial compressibility and thermal expansivity indicates that the a-axis is more compressible (KTa = 126 ± 3 GPa) than the b-axis (KTb = 137 ± 1 GPa) and c-axis (KTc = 143 ± 1 GPa). The temperature dependence of KT is stronger for the b-axis, (∂KT/∂T)Pb = −0.033 (3) GPa/K, than for the a-axis (∂KT/∂T)Pa = −0.028 (7) GPa/K and c-axis (∂KT/∂T)Pc = −0.017 (3) GPa/K. The present EOS enables us to accurately estimate the density of superhydrous phase B in a pyrolitic composition under deep mantle conditions. The density reduction of hydrated subducting slab (∼1 wt.% H2O) at the top of the lower mantle due to the presence of ∼18% of superhydrous phase B would be 1.9–2.1%. Accordingly, a model slab that is homogeneously hydrated may be buoyant relative to the surrounding mantle rocks and would not penetrate to the deep lower mantle.