An almost universal CO2 - CO32− carbon isotope fractionation function for high temperatures

•Carbon isotope fractionation factor CO2-carbonate.•C-isotope fractionation for all species containing the carbonate-group almost identical.•Consistent set of 13C/12C fractionation functions of geologically important species.•Degassing of MORB dominated by H2CO3, hence same δ13C value as mantle.•High pressure fluids dominated by H2CO3, no 13C/12C fractionation during subduction. High-temperature experiments on the CO2 - carbonate melt and carbonatite - nephelinite melt pairs are used to quantify equilibrium carbon isotope fractionation at conditions of 1 atm, 750–950 °C and 0.3–0.8 GPa, 1170 °C, respectively. Together with available experimental data, all results are within error consistent with a universal CO2 – CO32− fractionation function103lnαCO2/carbonate=6.41(11)·106T2−2.54(8)·1012T4(T in K, valid for >650 °C) that encompasses solid and molten carbonates and also carbonate components in silicate melts. The experiments hint towards a small fractionation of carbonatite vs. silicate melt of +0.39 ± 0.35 ‰ (1 σ), which however remains within error indistinguishable from zero. A generalized single fractionation function suggests that the nearest neighbours of carbon (i.e., O) dominate isotope fractionation while the next nearest neighbours only have a minor role, which greatly facilitates the understanding of carbon transfer in the deep Earth. If there is small to negligible 13C/12C-fractionation between solid carbonate, molten (ionic) carbonate liquid and covalent carbonated silicate melt, then H2CO3- and HCO3−-species in fluids or gases should also have low 13C/12C fractionation factors relative to the carbonate-group in depolymerized silicate melts or carbonates. Speciation models predict that the H2CO3- and HCO3−-species dominate over the CO2 species in COH-fluids at high pressures or high temperatures, i.e. during MOR basalt degassing or subduction zone devolatilization. A consequent small 13C/12C fractionation (of 0.1–0.3 ‰) applies to (i) continuous degassing of COH-fluids from mid-ocean ridge basalts, consistent with an observed difference of δ13CMORB and δ13Cmantle ≤1 ‰ and (ii) C-transfer of a carbonatite or COH-fluid from the subducted crust to the mantle, which then has a near-identical δ13C as the source. Finally, a combination of our results with available experimental values allows calculation of a consistent set of high-temperature (>650 °C) carbon fractionation factors for CO2, CO32− (carbonate, carbonatite or carbonated silicate melt), C