The influence of oceanic oxidation on serpentinite dehydration during subduction

Serpentinites are central to carry water and fluid-mobile elements down subduction zones. The breakdown of antigorite represents the most prominent aqueous fluid release, boosting fluid-mediated element cycling from the slab to the mantle wedge. At Cerro del Almirez, Spain, an antigorite dehydration reaction front is preserved in subducted serpentinites. Bulk rock and mineral major element chemistry linked to detailed petrography reveals that silicate mineral Mg# (100⁎[Mg/(Mg+Fe)]molar) are higher than bulk rocks due to the presence of magnetite, and olivine Mg# are lower in Chl-harzburgite than in Atg-serpentinite. The amount of magnetite is lower in Chl-harzburgite (1.4 vol%) than in Atg-serpentinite (2.8 vol%), resulting in reactive bulk rock compositions with Mg# of 92.7 and 96.0, respectively. Pseudosection modelling employing these reactive bulk compositions yields a small temperature field at 670 °C, 1.6 GPa where Atg-serpentinite and Chl-harzburgite coexist at the same metamorphic conditions. Thus, the antigorite dehydration front represents a compositional boundary rather than a thermal front (isograd). We suggest that this compositional boundary represents an oxidation front established upon serpentinisation at the ocean floor. Previous studies have shown that with increasing extent of serpentinisation increasing amounts of magnetite are formed concomitant with an increase in the Mg# of coexisting silicates. During subduction of such heterogeneous ultramafic rocks antigorite dehydration will be a continuous reaction that can occur over up to ∼40 °C, controlled by variations in Mg# of the reactant silicates. The difference in the amount of magnetite between Atg-serpentinite and Chl-harzburgite is thus not related to a change in redox budget of subducting serpentinites at Almirez imposed by the antigorite dehydration reaction. Rather, mass balance considerations suggest that ferric iron from antigorite may actually be the most prominent contribution to the redox budget of the antigorite dehydration reaction. Our findings also imply that direct comparison between Atg-serpentinites and Chl-harzburgites to infer geochemical changes associated with prograde dehydration reactions may lead to erroneous conclusions, including estimates on element loss mediated by aqueous fluid escape and associated changes in redox budget based on Fe3+/Fetot.