Metal saturation in the upper mantle

Abstract

The oxygen fugacity $f_{{\text{O}}_{\text{2}}}$of the Earth’s mantle is one of the fundamental variables in mantle petrology. Through ferric–ferrous iron and carbon–hydrogen–oxygen equilibria, $f_{{\text{O}}_{\text{2}}}$ influences the pressure–temperature positions of mantle solidi and compositions of small-degree mantle melts^1,2,3. Among other parameters, $f_{{\text{O}}_{\text{2}}}$ affects the water storage capacity and rheology of the mantle^4,5. The uppermost mantle, as represented by samples and partial melts, is sufficiently oxidized to sustain volatiles, such as H₂O and CO₂, as well as carbonatitic melts^6,7, but it is not known whether the shallow mantle is representative of the entire upper mantle. Using high-pressure experiments, we show here that large parts of the asthenosphere are likely to be metal-saturated. We found that pyroxene and garnet synthesized at >7 GPa in equilibrium with metallic Fe can incorporate sufficient ferric iron that the mantle at >250 km depth is so reduced that an (Fe,Ni)-metal phase may be stable. Our results indicate that the oxidized nature of the upper mantle can no longer be regarded as being representative for the Earth’s upper mantle as a whole and instead that oxidation is a shallow phenomenon restricted to an upper veneer only about 250 km in thickness.