Ti-in-quartz thermobarometry and TiO2 solubility in rhyolitic melts: New experiments and parametrization

Abstract The Ti-in-quartz thermobarometer has a wide potential for constraining crystallization pressure and temperature of quartz in natural geological systems. However, there is a long-lasting debate on the applicability of two models that were proposed previously, based on the equilibration of quartz with Ti-bearing aqueous fluids. In this study, the Ti-in-quartz thermobarometer was calibrated based on partitioning data of Ti between quartz and aluminosilicate melt in the pressure and temperature range of 0.5−4 kbar and 700−900 °C, which are conditions relevant for high-silica magmas stored at crustal depths. For seventeen experiments, in which both quartz, rutile and high-silica glass are present as experimental products (i.e., activity of TiO2 in silicate melt equals to unity), the Ti concentrations in quartz can be modeled with the following equation: log ⁡ C Ti Qtz = 5.3226 − 1948.4 / T − 981.4 ⁎ P 0.2 / T , in which C Ti Qtz is the Ti concentration (ppm) in quartz, T is temperature in kelvin and P is pressure in kbar. Based on the data from this study and a previous work of Hayden and Watson (2007) , we modeled the dependence of rutile (TiO2) solubility in silicic melt on temperature, pressure and melt composition, which can be expressed as log ⁡ ( S Ti liq ) = 6.5189 − 3006.5 / T − 461.0 ⁎ P 0.2 / T + 0.1155 ⁎ FM , in which S Ti liq is Ti solubility (ppm) at rutile saturation and FM is a parameter accounting for melt compositional effect, computed as FM = ( Na + K + 2 Ca + 2 Mg + 2 Fe ) / ( Si ⁎ Al ) , in which the chemical symbols denote molar fractions of each cation. Combining the two models presented above as well as some additional experimental data at activity of TiO2 log ⁡ ( C Ti Qtz / C Ti liq ) = − 1.1963 + ( 1058.1 − 520.4 ⁎ P 0.2 ) / T − 0.1155 ⁎ FM , in which C Ti liq is Ti concentration (ppm) in melt. Assuming an uncertainty of input temperature of ±25 °C, the corresponding pressure can be determined within ±0.2 kbar. However, the Ti concentrations in quartz and glass need to be determined with a high precision. Typical values of the ratio C Ti Qtz / C Ti liq in natural systems vary in the range from ∼0.09 to ∼0.13, corresponding to a change of pressure from ∼5 to ∼1 kbar assuming a temperature of ∼800 °C. The model above was applied to natural datasets obtained for several silicic eruptions (i.e. Oruanui Rhyolite, Early Bishop Tuff, Toba Tuff, Upper Bandelier Tuff). The analyses of quartz and glass inclusions in quartz indicate that the pre-eruptive magma storage pressures are mainly in the range 2–4 kbar. These pressures are consistent to or slightly higher than the maximum value estimated previously from the analysis of H2O-CO2 in glass inclusions, indicating a possible post-entrapment loss of hydrogen from melt inclusions and that gas saturation provides a minimum estimation of pressure.