Effects of H 2O on the phase behaviour of the forsterite-enstatite system at high pressures and temperatures and implications for the Earth

Abstract

In addition to many of the naturally occurring hydrous minerals known in the system MgOSiO 2H 2O, various dense hydrous magnesium silicates (10 Å-phase, phases A, B, C and D) have been synthesized by high-pressure and high-temperature experiments in the past two decades. In conjunction with these high-pressure hydrous phases, the effects of H 2O on the observed phase transformations in the forsterite (Mg 2SiO 4)-enstatite (MgSiO 3) system at high pressures and high temperatures are investigated on the basis of both experimental data and theory. For forsterite, the modified spinel phase may exist up to 7.1 wt.% H 2O along with enstatite + phase A, up to 9.9 wt.% H 2O along with stishovite + phase A, and up to only 3.2 wt.% H 2O along with stishovite + phase B. The spinel phase of Mg 2SiO 4 is no longer stable when more than 3.2 wt.% H 2O is present. Although both stishovite and the ilmenite-type MgSiO 3 do not occur in dry Mg 2SiO 4 at high pressures, a small amount of H 2O would yield assemblages containing stishovite and the ilmenite phase in Mg 2SiO 4 at high pressures. The post-spinel phase assemblage, perovskite + periclase, may be stable, along with phase D, up to ∼ 11 wt.% H 2O. In the case of enstatite, except for the assemblage modified spinel + stishovite, which exists (along with phase A) up to 7.1 wt.% H 2O and for the perovskite-type MgSiO 3 (along with periclase and phase D) which exists up to ⋍ 15 wt.% H 2 O , all the high-pressure phase transformations observed in dry MgSiO 3 may not occur at high pressures if > 2.3 wt.% H 2O is present. Based on experimental observations, the various high-pressure hydrous phases are stable from as low as 40 kbar to more than 280 kbar at various temperatures. Thus, there is no lack of ‘storage phases’ for H 2O in the deeper interior of the Earth's mantle. According to the impact-induced dehydration model advanced recently by shock-wave experiments, it is estimated that the Earth's interior contains probably at least more than five times its present water contents in the near-surface geochemical reservoirs (hydrosphere + crust). So long as both forsterite (olivine) and H 2O coexist in the upper mantle, even in trace amounts, there must exist a ‘water line’, which is defined by the reaction 5Mg 2SiO 4 + 3H 2O ⇄ 3MgSiO 3 + Mg 7Si 2O 14H 6 forsterite water enstatite phase A in the upper mantle. The equilibrium boundary of the ‘water line’ has been, in fact, determined in two different phase assemblages by experiments, but was not noticed. If the Earth's surface was covered by large-scale melts during and shortly after its accretion, H 2O would be stored in phases A, B,… below the ‘water line’ and would liberate as water above the ‘water line’ during freezing. Today's oceans were most likely derived from water released in the mantle when the palaeogeotherm dropped below the solidus of mantle materials, stripping the free water above the ‘water line’ from depths near 350 km. Thus, the oceans were formed rapidly and early in the Earth's history, and slow degassing is still going on today. The low-velocity zone observed in today's mantle is interpreted as a region, containing free water, between the ‘water line’ and the stability field of amphiboles in the suboceanic mantle, and between the ‘water line’ and the stability field of phlogopite in the subcontinental mantle. In other words, the low-velocity zone is the free water zone and the ‘water line’ is the base of the low-velocity zone in the Earth's mantle.