Fe3+ Distribution and Fe3+/ΣFe-Oxygen Fugacity Variations in Kimberlite-Borne Eclogite Xenoliths, with Comments on Clinopyroxene-Garnet Oxy-Thermobarometry

Abstract

Listen

Translate article

Abstract The valence state of iron in mantle-derived melts, such as mid-ocean ridge basalt, is a useful proxy for oxygen fugacity (ƒO2). On subduction, oceanic crust undergoes metamorphic reactions that alter its initial ƒO2, generating compositional and redox heterogeneity in the supra-subduction zone and in the convecting mantle source of oceanic basalts. Kimberlite-borne eclogite xenoliths with ancient oceanic crustal protoliths represent an important archive to trace these processes in deep time. We determined, by Mössbauer spectroscopy, Fe3+/ΣFe for garnet (0.03–0.15, average 0.08; n = 13) and clinopyroxene (0.05–0.37, average 0.23; n = 11) in bimineralic eclogite xenoliths from the Udachnaya kimberlite pipe (Siberian craton), combined the results with high-quality literature data to estimate bulk Fe3+/ΣFe and ƒO2, assessed associated uncertainties and discussed petrological implications. The incorporation of Fe3+ in, and distribution between, eclogite minerals is controlled by chemical composition (X), temperature (T), pressure (P) and ƒO2. Therefore, Fe3+/ΣFe in garnet alone is an imperfect proxy for bulk Fe3+/ΣFe. Values for the distribution of Fe3+/ΣFe between clinopyroxene and garnet (D(Fe3+/ΣFe)cpx-grt) range from 1.4 ± 0.1 to 20 ± 13 and correlate strongly with the abundance distribution of another polyvalent element, vanadium (r2 = 0.80, n = 28), probably because both vary as a function of X-T–P-ƒO2. This allows to estimate Fe3+/ΣFe in clinopyroxene and bulk rocks (0.01–0.32, average 0.06, n = 159) for the majority of literature data where only garnet Fe3+/ΣFe is known. Low Fe3+/ΣFe is retained in many high-Al2O3 eclogites with plagioclase-rich cumulate protoliths (Eu/Eu*> > 1), despite extended residence in the cratonic lithosphere. Bulk Fe3+/ΣFe increases during mantle metasomatism but is not particularly linked to enrichment in highly incompatible elements. Low grossular content in garnet ensuing from mantle metasomatism, especially at low temperature, limits uptake of Fe3+, which is accommodated in clinopyroxene instead, leading to very high D(Fe3+/ΣFe)cpx-grt. Eclogite oxybarometry requires that Fe3+/ΣFe of only garnet be known, whereby the input temperatures should be calculated with all Fe as Fe2+. For temperatures projected onto regional steady-state geothermal gradients, ƒO2 values relative to the fayalite-magnetite-quartz buffer (∆logƒO2(FMQ)) range from −5.9 to −0.2 (average − 3.1, n = 174), and are too low to stabilise oxidised sulphur in all and oxidised carbon in the vast majority of samples. Thus, ancient oceanic crust was a sink rather than a source of oxygen. In particular CaO-rich eclogite xenoliths occasionally are corundum-bearing and SiO2-undersaturated, typically with silica-deficient clinopyroxene with cations per formula unit <1.97 for 6 O anions, resulting in overestimated ƒO2. For the remaining samples, cryptic metasomatism, with enrichment in highly incompatible elements, is clearly focused in the ƒO2 interval of FMQ−4 to FMQ−3, indicating that mantle metasomatism can have both an oxidising and a reducing effect on eclogite. Unmetasomatised bimineralic eclogites have lower ƒO2 than coesite-bearing ones, reflecting reduction during metamorphism and silica-consuming partial melting of ancient subducted oceanic crust.