Experimental constraints on the ferric Fe content and oxygen fugacity in subducted serpentinites

ABSTRACT Serpentinites play an important role in the delivery of water into subduction zones. In addition, serpentinites also contain ferric Fe and can transport significant redox potential. We present high-pressure and high-temperature experiments and Mössbauer spectroscopy measurements on natural lizardite and antigorite samples equilibrated at various oxygen fugacities in order to quantify the relationship between the oxygen fugacity f(O2) and the Fe3+/Fetot ratio in these two phases. In antigorite, Fe3+ partitions into the octahedral site and is charge balanced by tetrahedral Al. In lizardite, tetrahedral Fe3+ is observed only at low temperature as well as under high f(O2), whereas Fe3+ prefers the octahedral site at temperatures exceeding 500 °C and at 3 to 5 GPa. Although metastable, lizardite remains in redox equilibrium in our experiments at conditions above the lizardite to antigorite phase transformation at 300 °C and demonstrates a similar stability to antigorite. The Al concentration of lizardite is found to be temperature dependent, and it was possible to reequilibrate the Fe3+/Fetot ratio of lizardite from 0.1 to 0.9 by using redox buffers such as Fe metal, graphite, graphite–calcite, Re–ReO2 and Ru–RuO2. Our experiments on antigorite demonstrate that antigorite does not adjust its Al concentration on experimental time scales. Since Fe3+ is charge balanced by Al, it was also not possible to manipulate the Fe3+/Fetot ratio of antigorite. The coexisting phases, however, show chemical equilibration with this antigorite composition. We have retrieved the standard Gibbs energy for Fe3+- and Al-endmembers of antigorite and lizardite and calculated the metamorphic evolution of subducting serpentinites. The lizardite to antigorite transformation does not cause a decrease in the bulk Fe3+/Fetot ratio under f(O2) buffered conditions, in contrast to observations from some natural settings, but does result in the formation of additional magnetite due to antigorite having a lower Fe3+/Fetot ratio than lizardite at equilibrium. If the f(O2) of antigorite serpentinite is buffered during subduction, such as due to the presence of graphite and carbonate, the bulk Fe3+/Fetot ratio decreases progressively. On the other hand, in a closed system where the bulk serpentinite Fe3+/Fetot ratio remains constant, the f(O2) increases during subduction. In this scenario, the f(O2) of an antigorite serpentinite with a typical Fe3+/Fetot ratio of 0.4 increases from the fayalite–magnetite–quartz to the hematite–magnetite f(O2) buffer during dehydration. These f(O2) results confirm earlier inferences that fluids produced by antigorite dehydration may not contain sufficient oxidised sulphur species to oxidise the mantle wedge. Sufficiently high levels of f(O2) to mobilise oxidised sulphur species may be reached upon antigorite dehydration, however, if closed system behaviour maintains a high bulk redox potential across the lizardite to antigorite phase transformation. Alternatively, oxidation of the mantle wedge might be achieved by oxidising agents from sources in subducted oceanic crust and sediments.