The effect of oxygen fugacity on the solubility of carbon-oxygen fluids in basaltic melt

Abstract

The solubility of CO 2 CO fluids in a mid-ocean ridge basalt ( morb) has been measured at 1200°C, 500–1500 bar, and oxygen fugacities between NNO and NNO-4. High oxygen fugacities, and thus CO 2-rich fluids, were produced by using a starting material equilibrated at NNO, and Ag 2C 2O 4 as the fluid source. Low oxygen fugacities were achieved by using graphite capsules, and MgCO 3 as the fluid source. These graphite-saturated fluids have the lowest possible C/O 2CO ratio for a given pressure and temperature. Experiments were run in a rapid-quench internally heated pressure vessel. Fluid compositions were measured using a simple vacuum technique and by Raman spectroscopy of fluid inclusions. The two techniques yielded comparable results. Fourier transform micro-infrared spectroscopy was used to identify and measure concentrations of dissolved volatiles in double-polished wafers of the quenched glasses. Carbonate was the only carbon-bearing species identified. Raman spectroscopic analysis of inclusion-free areas of glass confirmed the absence of dissolved molecular CO 2, CO and carbon. The measured concentrations of dissolved CO 2 in the glasses were proportional to the fugacity of CO 2 during the experiments, calculated from the measured fluid compositions. The data were fit to the equation X CO 2 melt(ppm)= 0.492 fCO 2 (bar). The insolubility of CO, compared to CO 2, may be related to the fact that dissolution of CO requires reduction of another species in the melt, whereas dissolution of CO 2 does not. Due to the fact that CO will be an important component of natural C O fluids at low pressures and low oxygen fugacities, equilibrium dissolved CO 2 contents will be less than calculated assuming pure CO 2 fluids, but as the C/O 2CO ratio in a pure C O fluid at fixed pressure and temperature is a direct function of oxygen fugacity, measurement of the oxygen fugacity of quenched glasses or trapped fluids in natural samples should allow saturation concentrations to be calculated. Dissolved CO 2 contents of some morb are less than expected if they were in equilibrium with pure CO 2. These samples must, therefore, have been more reduced than average if they were fluid-saturated. Together with results from other studies of CO 2 and H 2O solubilities in basalt, the results of this study provide a comprehensive framework for modelling CO 2 solution in morb.