A Thermodynamic Approach for Modeling H2O-CO2 Solubility in Alkali-rich Mafic Magmas at Mid-crustal Pressures
Volatile solubility in magmas is dependent on several factors, including composition and pressure. Mafic (basaltic) magmas with high concentrations of alkali elements (Na and K) are capable of dissolving larger quantities of H2O and CO2 than low-alkali basalt. The exsolution of abundant gases dissolved in alkali-rich mafic magmas can contribute to large explosive eruptions. Existing volatile solubility models for alkali-rich mafic magmas are well calibrated below 200 MPa, but at greater pressures the experimental data is sparse. To allow for accurate interpretation of mafic magmatic systems at higher pressures, we conducted a set of mixed H2O-CO2 volatile solubility experiments between 400 and 600 MPa at 1200 °C in six mafic compositions with variable alkali contents. Compositions include magmas from volcanoes in Italy, Antarctica, and Arizona. Results from our experiments indicate that existing volatile solubility models for alkali-rich mafic magmas, if extrapolated beyond their calibrated range, over-predict CO2 solubility at mid-crustal pressures. Physically, these results suggest that volatile exsolution can occur at deeper levels than what can be resolved from the lower-pressure experimental data. Existing thermodynamic models used to calculate volatile solubility at different pressures require two experimentally derived parameters. These parameters represent the partial molar volume of the condensed volatile species in the melt and its equilibrium constant, both calculated at a standard temperature and pressure. We derived these parameters for each studied composition and the corresponding thermodynamic model shows good agreement with the CO2 solubility data of the experiments. A general alkali basalt solubility model was also constructed by establishing a relationship between magma composition and the thermodynamic parameters. We utilize cation fractions from our six compositions along with four compositions from the experimental literature in a linear regression to generate this compositional relationship. Our revised general model provides a new framework to interpret volcanic data, yielding greater depths for melt inclusion entrapment than previously calculated using other models, and it can be applied to mafic magma compositions for which no experimental data is available.
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- 0810 Post-secondary education;