In situ reinvestigation of reaction phase A plus high-pressure clinoenstatite to forsterite plus water in the system MgO-SiO₂-H₂O (MSH)

Abstract

The dehydration reaction of phase A + high P clinoenstatite to forsterite + water was experimentally investigated at water-saturated conditions in the pressure range between 7.0 and 10.0 GPa by in situ reversal runs in a multi-anvil press at the synchrotron source of PETRA III in Hamburg. By using closed watertight X-ray transparent Ti capsules, its position is determined by reversal brackets at 8.3 GPa (700–760 ∘C), 8.6 GPa (700–740 ∘C), and 9.8 GPa (750–800 ∘C); thus, the equilibrium of the reaction corresponds ideally to the data reported by Wunder (1998). Optical investigations of the quenched product phases show strong grain coarsening of phase A and clinoenstatite, whereas nucleated forsterite from the breakdown of the aforementioned phases is very fine grained. This corresponds to recent experimental observations that the grain size of phases formed in hydration reactions are significantly larger than those from dehydration reactions. In addition, we performed three time-dependent in situ experiments at 9–10 GPa and 800–870 ∘C and monitored the reaction progress every 10 min to determine the kinetics of the forsterite formation from phase A + high P clinoenstatite. The growth of forsterite at these P–T conditions, already visible after 10 min, confirms the results of the bracketing experiments. However, the reaction is extremely slow, and even after more than 3 h, significant amounts of phase A and high P clinoenstatite are still present. This is in contradiction to other dehydration reactions of former experimental studies, e.g. the fast dehydration of serpentine, which completely dehydrates within 3 h, even at much lower temperatures, closely overstepping serpentine stability. Despite its reaction sluggishness, which would contradict the concept of earthquake initiation, the observed formation of nano-sized forsterite as a dehydration product may still indicate the potential of this reaction to cause mechanical instabilities and, thus, seismicity within cold subduction zones at depths of the Earth's mantle. Additionally, at depths exceeding serpentine dehydration, the phase A + high P/low P clinoenstatite breakdown to forsterite + water might induce geochemical and geophysical processes, including the formation of low-velocity zones within the overlying mantle wedge from the large amounts of fluid liberated by this water line reaction. After the breakdown of antigorite, the assemblage phase A + clinoenstatite might act as a bridge to transport water to larger depths during cold subduction, followed by the formation of other hydrous high P phases.