Back-scattered electron image of a sample produced in the multi-anvil press at RSES at 13 GPa and 1250°C. The lower part is quenched alkali-rich carbonatite melt (the bright phase is a mixture of ReO2 and Re metal, which maintains the oxygen state at high levels to prevent reduction of the carbonate to diamond). The upper 2/3 of the sample is majoritic garnet (bright phase), omphacitic clinopyroxene (mid-grey) and stishovite (dark grey). The black patches are holes in the polished sample surface. The image was obtained using a Hitachi FE-SEM in the Centre for Advanced Microscopy at ANU.
The Earth's Deep Carbon Cycle
The carbon cycle between various reservoirs in the Earths’ exosphere (atmosphere-hydrosphere-biosphere) exerts a critical control on climate on a range of time scales which are short relative to the age of the earth. However, the earth has a much deep carbon cycle, whereby carbon is recycled from exosphere into the deep mantle via subduction and returned to the exosphere during volcanism, on much longer time scales of millions or billions of years.
My colleagues, students and I are investigating this deep carbon cycle using a range of techniques, including high pressure experimental petrology using multi-anvil and piston-cylinder apparatuses, with sophisticated microbeam imaging and analysis of experimental run products. We are interested in the stability and melting temperatures of carbonate phases during deep subduction at pressures relevant to the sub-arc environment (≤ 6GPa), down to the uppermost part of the lower mantle (≤23 GPa).
As an example, the back-scattered electron image shows a multi-anvil experiment conducted at 13 GPa and 1250°C on an average mid-ocean ridge basaltic composition with a few wt% CaCO3 added. This is simulating deep subduction of altered, mafic oceanic crust, which commonly contains a small amount of calcite, added during hydrothermal alteration of the crust. Under these conditions, the sample has crystallized an assemblage of majoritic garnet + omphacitic clinopyroxene + stishovite, with a carbonatitic partial melt. This indicates that along relatively warm subduction geotherms, deeply subducted carbonate-bearing oceanic crust could produce carbonatitic melts in the deep upper mantle. The effects of these when they segregate from the subducting crust and move into overlying peridotite are currently being experimentally investigated.
Of great importance is the influence of oxygen fugacity on carbon’s behaviour under these P-T conditions. We have developed and applied a synchrotron-based method (Fe K-edge XANES) for determination of the Fe3+ content of mantle garnets in kimberlite-bourne peridotites (Berry et al. 2010; Yaxley et al. 2012; Hanger et al. 2014) and eclogites. This enables determination of the mantle oxygen fugacity in the cratonic lithosphere and deeply subducting oceanic crust, and investigation of its effects on diamond stability, partial melting and metasomatism.