Tracing fluid transfers in subduction zones: an integrated thermodynamic and <i>δ</i>¹⁸O fractionation modelling approach

Abstract

Oxygen isotope geochemistry is a powerful tool for investigating rocks that interacted with fluids, to assess fluid sources and quantify the conditions of fluid–rock interaction. We present an integrated modelling approach and the computer program PTLoop that combine thermodynamic and oxygen isotope fractionation modelling for multi-rock open systems. The strategy involves a robust petrological model performing on-the-fly Gibbs energy minimizations coupled to an oxygen fractionation model for a given chemical and isotopic bulk rock composition; both models are based on internally consistent databases. This approach is applied to subduction zone metamorphism to predict the possible range of δ18O values for stable phases and aqueous fluids at various pressure (P) and temperature (T) conditions in the subducting slab. The modelled system is composed of a mafic oceanic crust with a sedimentary cover of known initial chemical composition and bulk δ18O. The evolution of mineral assemblages and δ18O values of each phase is calculated along a defined P–T path for two typical compositions of basalts and sediments. In a closed system, the dehydration reactions, fluid loss and mineral fractionation produce minor to negligible variations (i.e. within 1 ‰) in the bulk δ18O values of the rocks, which are likely to remain representative of the protolith composition. In an open system, fluid–rock interaction may occur (1) in the metasediment, as a consequence of infiltration of the fluid liberated by dehydration reactions occurring in the metamorphosed mafic oceanic crust, and (2) in the metabasalt, as a consequence of infiltration of an external fluid originated by dehydration of underlying serpentinites. In each rock type, the interaction with external fluids may lead to shifts in δ18O up to 1 order of magnitude larger than those calculated for closed systems. Such variations can be detected by analysing in situ oxygen isotopes in key metamorphic minerals such as garnet, white mica and quartz. The simulations show that when the water released by the slab infiltrates the forearc mantle wedge, it can cause extensive serpentinization within fractions of 1 Myr and significant oxygen isotope variation at the interface. The approach presented here opens new perspectives for tracking fluid pathways in subduction zones, to distinguish porous from channelled fluid flows, and to determine the P–T conditions and the extent of fluid–rock interaction.