The oxygen fugacity at which graphite or diamond forms from carbonate-bearing melts in eclogitic rocks

Abstract

The oxygen fugacity (fO2) at which carbonate-bearing melts are reduced to either graphite or diamond in synthetic eclogite compositions has been measured in multi-anvil experiments performed at pressures between 3 and 7 GPa and temperatures between 800 and 1,300 °C using iron–iridium and iron–platinum alloys as sliding redox sensors. The determined oxygen fugacities buffered by the coexistence of elemental carbon and carbonate-bearing melt are approximately 1 log unit below thermodynamic calculations for a similar redox buffering equilibrium involving only solid phases. The measured oxygen fugacities normalized to the fayalite–magnetite–quartz oxygen buffer decrease with temperature from ~−0.8 to ~−1.7 log units at 3 GPa, most likely as a result of increasing dilution of the carbonate liquid with silicate. The normalized fO2 values also decrease with pressure and show a similar decrease with temperature at 6 GPa from ~−1.5 log units at 1,100 °C to ~−2.4 log units at 1,300 °C. In contrast to previous arguments, the stability field of the carbonate-bearing melt extends to lower oxygen fugacity in eclogite rocks than in peridotite rocks, which implies a wider range of conditions over which carbon remains mobile in natural eclogites. The raised prevalence of diamonds in eclogites compared to peridotites may, therefore, reflect more effective scavenging of carbon by melts in these rocks. The ferric iron contents of monomineralic layers of clinopyroxene and garnet contained in the same experiments were also measured using Mossbauer spectroscopy. A preliminary model was derived for determining the fO2 of eclogitic rocks from the compositions of garnet and clinopyroxene, including the Fe3+/ΣFe ratio of garnet, using the equilibrium, $$\mathop { 5 {\text{CaFeSi}}_{2} {\text{O}}_{6} }\limits_{\text{cpx}} + \mathop {1/3{\text{Ca}}_{3} {\text{Al}}_{2} {\text{Si}}_{3} {\text{O}}_{12} }\limits_{\text{garnet}} + {\text{O}}_{2} = \mathop { 2 {\text{Ca}}_{3} {\text{Fe}}_{2} {\text{Si}}_{3} {\text{O}}_{12} }\limits_{\text{garnet}} + \mathop {1/3{\text{Fe}}_{3} {\text{Al}}_{2} {\text{Si}}_{3} {\text{O}}_{12} }\limits_{\text{garnet}} + \mathop {4{\text{SiO}}_{2} }\limits_{\text{coesite}}$$ The model, which reproduces the independently determined fO2 of the experimental data to within 0.5 log units, can be used to estimate the fO2 of ultrahigh-pressure metamorphic eclogites and cratonic eclogitic xenoliths. Although there are very few analyses of garnet Fe3+/ΣFe ratios from eclogite samples, the range in fO2 recorded by available eclogitic xenoliths is similar to that reported for peridotitic xenoliths and generally within the graphite/diamond stability field. Estimates for the average bulk Fe3+/ΣFe ratio of modern basaltic oceanic crust, however, are higher than the values for most of these xenoliths, and upon subduction, crustal carbon is likely to remain in the carbonate stability field to depths of at least 250 km. If eclogite xenoliths originated from subducted oceanic crust, then their generally lower fO2 most likely reflects either lower initial basaltic Fe3+/ΣFe ratios, loss of Fe2O3 through partial melting or the initial presence of organic carbon.