Phase equilibria in subducting basaltic crust: implications for H sub 2O release from the slab

Fluids released from subducting slabs induce partial melting of the mantle wedge above the slab, which in turn is responsible for arc volcanism at the Earth's surface. The partially hydrated basaltic layer of the slab is a potential source of these fluids and a major reservoir for H sub 2O at depth. Constraining the stability domains of hydrous phases and the position of the dehydration reactions in this system in pressure-temperature (P-T) space is essential to describe and quantify the fluid release from subducting oceanic crust into the overlying mantle wedge. Experiments were conducted in the ranges of 2.2-3.4 GPa and 625-750 C to determine phase equilibria in an H sub 2O- saturated natural basalt at conditions relevant to subduction zones. The experimental duration was typically 1 month, although some experiments were replicated with a shorter run duration (1-2 weeks) in order to identify potentially metastable phases. A mixture of a natural mid-ocean ridge basalt (MORB) glass and mineral seeds was used as the starting material. Oxygen fugacity was buffered within +/-1.3 log units of nickel-bunsenite (NiNiO). The results obtained show that a calcic amphibole (barroisite) is stable from 2.2 to about 2.4 GPa. At 2.6 GPa, it is replaced by a sodic amphibole (near end-member glaucophane), which is stable up to 3 GPa at 625 C. This high-pressure assemblage constitutes a true analog of a natural amphibole-bearing eclogite and the first synthesis of glaucophane from a rock of basaltic composition. As opposed to the results of previous studies on basaltic compositions [A.R. Pawley, J.R. Holloway, Science 260 (1993) 664-667; S. Poli, Am. J. Sci. 293 (1993) 1061-1107; S. Poli, M.W. Schmidt, J. Geophys. Res. 100 (1995) 22299- 22314; M.W. Schmidt, S. Poli, Earth Planet. Sci. Lett. 163 (1998) 361-379], chloritoid is present only as a metastable phase in the pressure-temperature range investigated here. Metastability of chloritoid in earlier experiments, due to short run duration, is the most likely explanation for this difference, and suggests that chloritoid does not play an important role in the overall dehydration process of the basaltic layer in subduction zones. At pressures above the stability field of amphibole, zoisite/clinozoisite becomes the stable hydrous phase at temperatures above 645 C, whereas lawsonite is stable at lower temperatures. The positions of the zoisite-out and lawsonite-out reactions determined in this study indicate that, for an interme