A CO₂ solubility model for silicate melts from fluid saturation to graphite or diamond saturation

By: James Eguchi, Rajdeep Dasgupta

Abstract: A model based on a thermodynamic framework for CO₂ concentrations and speciation in natural silicate melts at graphite/diamond-saturated to fluid-saturated conditions is presented. The model is simultaneously calibrated with graphite-saturated and fluid-saturated conditions allowing for consistent model predictions across the CCO buffer. The model was calibrated using water-poor (≤1 wt% H₂O) silicate melts from graphite- to CO₂-fluid-saturation over a range of pressure (P = 0.05–3 GPa), temperature (T = 950–1600 °C), composition (foidite-rhyolite; NBO = 0.02–0.92; wt% SiO₂ ~ 39–77, TiO₂ ~ 0.1–5.8, Al₂O₃ ~ 7.5–18, FeO ~ 0.2–24 MgO ~ 0.1–24, CaO ~ 0.3–14, Na₂O~1-5, K₂O ~ 0–6), and f_O2 (~QFM +1.5 to ~QFM −6). The model can predict CO₂ concentrations for a wide range of silicate melt compositions from ultramafic to rhyolitic compositions, i.e., melts that dissolve carbon only as carbonate anions CO₃^2– and those that dissolve carbon both as CO₃^2– and as molecular CO₂^mol as a function of pressure, temperature, and oxygen fugacity. The model also does a reasonable job in capturing CO₂ solubility in hydrous silicate melts with ≤2–3 wt% H₂O. New CO₂solubility experiments at pressures >3 GPa suggest that the newly developed CO₂ solubility model can be satisfactorily extrapolated to ~4–5 GPa. Above 5 GPa the model poorly reproduces experimental data, likely owing to structural change in silicate melt at pressures above 5 GPa. An Excel spreadsheet and a Matlab function are provided as online supplementary materials for implementing the new CO₂ solubility model presented here.