CO2 content of andesitic melts at graphite-saturated upper mantle conditions with implications for redox state of oceanic basalt source regions and remobilization of reduced carbon from subducted eclogite
We have performed experiments to determine the effects of pressure, temperature and oxygen fugacity on the CO2 contents in nominally anhydrous andesitic melts at graphite saturation. The andesite composition was specifically chosen to match a low-degree partial melt composition that is generated from MORB-like eclogite in the convective, oceanic upper mantle. Experiments were performed at 1–3 GPa, 1375–1550 °C, and fO2 of FMQ −3.2 to FMQ −2.3 and the resulting experimental glasses were analyzed for CO2 and H2O contents using FTIR and SIMS. Experimental results were used to develop a thermodynamic model to predict CO2 content of nominally anhydrous andesitic melts at graphite saturation. Fitting of experimental data returned thermodynamic parameters for dissolution of CO2 as molecular CO2: ln(K0) = −21.79 ± 0.04, ΔV0 = 32.91 ± 0.65 cm3mol−1, ΔH0 = 107 ± 21 kJ mol−1, and dissolution of CO2 as CO32−: ln(K0) = −21.38 ± 0.08, ΔV0 = 30.66 ± 1.33 cm3 mol−1, ΔH0 = 42 ± 37 kJ mol−1, where K0 is the equilibrium constant at some reference pressure and temperature, ΔV0 is the volume change of reaction, and ΔH0 is the enthalpy change of reaction. The thermodynamic model was used along with trace element partition coefficients to calculate the CO2 contents and CO2/Nb ratios resulting from the mixing of a depleted MORB and the partial melt of a graphite-saturated eclogite. Comparison with natural MORB and OIB data suggests that the CO2 contents and CO2/Nb ratios of CO2-enriched oceanic basalts cannot be produced by mixing with partial melts of graphite-saturated eclogite. Instead, they must be produced by melting of a source containing carbonate. This result places a lower bound on the oxygen fugacity for the source region of these CO2-enriched basalts, and suggests that fO2 measurements made on cratonic xenoliths may not be applicable to the convecting upper mantle. CO2-depleted basalts, on the other hand, are consistent with mixing between depleted MORB and partial melts of a graphite-saturated eclogite. Furthermore, calculations suggest that eclogite can remain saturated in graphite in the convecting upper mantle, acting as a reservoir for C.