FERRIC IRON IN OROGENIC LHERZOLITE MASSIFS AND CONTROLS OF OXYGEN FUGACITY IN THE UPPER MANTLE

Abstract

The bulk Fe2O3 contents and inter-mineral distributions of Fe3+ were investigated in a suite of samples from several orogenic lherzolite massifs, including Beni Bousera (Morocco), Ronda (Spain) and Lherz (France). Ferric iron contents were determined for each phase by Mössbauer spectroscopy and these results were combined with microprobe data and modal abundances to determine the bulk Fe2O3 contents. The notion that Fe3+ is moderately incompatible during partial melting in the mantle is supported by the observed decrease in bulk Fe2O3 content with increasing MgO, as well as by the generally lower Fe3+ content of clinopyroxene in samples with low modal abundances of this phase. The partitioning of Fe3+ between orthopyroxene and clinopyroxene is consistent with literature data for spinel peridotite xenoliths. The partitioning of Fe3+ between clinopyroxene and spinel is composition dependent, changing with the Cr/Al of spinel. Thus partitioning is expected to be different in the spinel peridotite and plagioclase peridotite facies of the upper mantle. The f{hook}O2 of orogenic massifs varies over several log units relative to the FMQ buffer. The values recorded by spinel-based and clinopyroxene-based oxygen barometry are generally comparable, indicating redox equilibrium between these two phases even at the relatively low temperature conditions existing in parts of the lithospheric mantle. Calculated bulk Fe2O3 contents range from 0.03 to 0.27 wt.%. The combination of modal abundance and major element composition means that orthopyroxene is a major contributor to the bulk Fe2O3 budget of peridotites, although clinopyroxene and spinel are much richer in Fe3+ on a per formula unit basis. Residual Cr-rich spinel is the dominant source of Fe3+ in plagioclase peridotites. In terms of the geochemical behaviour of Fe3+, it can be concluded that orogenic peridotites exhibit essentially the same behaviour as spinel peridotite xenoliths. In terms of the controlling factors of f{hook}O2 in the upper mantle, our data set records a certain degree of decoupling of f{hook}O2 from whole rock Fe2O3 content, even if a correlation between these two parameters is generally apparent. This decoupling is because, whereas both whole rock Fe2O3 content and f{hook}O2 are influenced by partial melting and melt extraction, additional processes such as metasomatism and phase changes can effectively reset f{hook}O2 without always causing a concomitant change in bulk Fe2O3 content. Modelling the oxidation of spinel reveals that the f{hook}O2 can be readily reset under initially reducing conditions, but the incorporation of progressively more and more Fe3+ in spinel is required to further raise f{hook}O2. A quasi-limit of Δlogf{hook}O2 = FMQ + 1 is expected. These results are consistent with the general redox behaviour observed for spinel peridotites. Our data imply that Fe3+-Fe2+ equilibria have an important, if not dominant influence on f{hook}O2 in the spinel peridotite facies of the upper mantle. © 2006 Elsevier B.V. All rights reserved.