Carbon-rich melts in the deep mantle

Adrian JONES

University College London

Carbon-rich melts in the deep mantle ADRIAN P JONES123 1Dept Earth Sciences, University College London, Gower Street, London WC1E 6BT UK The noble gas and radiogenic isotope geochemistry of global carbonatites (n>527 [1]) likely represent concentrations of asthenospheric mantle carbon and provide an indirect measure of deep carbon flux through time. If as geochemists propose [2], their sources are deeper than sub-lithospheric mantle, then carbonatites cannot be derived directly by oxidation of typical diamond reservoirs sampled by kimberlites. Recent surveys of nominally carbon-free mantle silicate minerals point to the increased significance of discrete carbon phases as deep mantle carbon reservoirs. In other words, deep carbon is unlikely to be stored in the same way as water; there is no deep carbon equivalent to oceans of nominally anhydrous silicates (NAH’s). Therefore, new data for the stabilities of deep mantle carbonate minerals is particularly important, since different structured Mg- and Ca-carbonate minerals are predicted to be stable through the deep mantle and in a Ca-carbonate phase may perhaps be stable at pressures of ~136 GPa at the core mantle boundary (CMB [3]). The physical properties of carbonate melts have only been measured at modest pressures (< 5.5 GPa), and are characterized by extraordinarily low viscosities compared to silicate melts [4], but there is, so far, insufficient data to derive a full equation of state for higher-P carbonate melts > 6 GPa in the deeper mantle. However, first order estimates can be derived from calculated carbonate mineral data in the lower mantle, and compared with recent physical data for silicate melts under different geoetherms applicable to Earth’s history, including magma oceans (Stixrude pers comm). Deep carbonate melts may be extremely important agents for selective transfer of components, and they may superced the role of metasomatic fluids widely recognised in relatively shallow mantle diamond. Simple dynamic constraints will be used to illustrate where deep carbonate melts are expected, and how these may be expected to compare or contrast with their shallow pressure counterparts (carbonatites).