Rutile solubility in H2O, H2O–SiO2, and H2O–NaAlSi3O8 fluids at 0.7–2.0 GPa and 700–1000 °C: Implications for mobility of nominally insoluble elements

The solubility of rutile was measured in H2O, H2O–SiO2 and H2O–NaAlSi3O8 fluids at 700–1000 °C, 0.7–2.0 GPa, in a piston-cylinder apparatus. Solubility was determined by weight loss using a double-capsule method. Rutile solubility in pure H2O shows isothermal increase with pressure (P), isobaric increase with temperature (T), and is low at all conditions investigated (6–118 ppm Ti). Rutile solubility in H2O is given by logcTi° = 6.173 − 5425/T + 178.4P/T, where cTi° is Ti concentration in ppm, T is in K, and P in GPa. This leads to thermodynamic properties of the reaction rutile = TiO2,aq of ΔSr° = 28.6 J/mol K, ΔHr° = 104 kJ/mol, and ΔVr° = − 3.4 cm3/mol. At 800 °C and 1 GPa, addition of SiO2 (up to quartz saturation) did not change rutile solubility relative to that in pure H2O. Determination of rutile solubility in H2O–NaAlSi3O8 fluids was complicated by incongruent dissolution of albite to paragonite or corundum + fluid; however, fluid compositions could be estimated within narrow limits using a mass-balance scheme. The solubility of rutile increases linearly with dissolved Na–Al silicate at fixed P and T, as described by cTi = cTi° + Bws where cTi is ppm Ti, ws is wt.% dissolved silicate and B is given by log B 6.512 − 1.665P − 6224/T + 2215P/T, with T and P again in K and GPa. sults help explain discrepancies among previous studies of rutile solubility in H2O at similar P and T. The new data agree within error with those of Tropper and Manning [Tropper, P., Manning, C.E., 2005. Very low solubility of rutile in H2O at high pressure and temperature, and its implications for Ti mobility in subduction zones. American Mineralogist 90, 502–505.], but give lower solubility than earlier piston-cylinder-based determinations due to suppression of new crystal growth in the present experiments. However, the new data yield higher solubilities than are predicted from a hydrothermal diamond-anvil study, probably because of our longer run times and more complete equilibration. Combination of predicted Ti concentrations in melt-saturated H2O with H2O-saturated albite melts suggests that the melt–vapor partition coefficient for Ti is constant at 9.5 ± 1.5 from 700 to 900 °C at 1 GPa and rutile saturation, implying that an H2O-rich magmatic vapor phase can transport significant Ti in mid- to deep-crustal settings. Because crustal and mantle fluids will contain alkalis, Al and Si, the results in H2O–NaAlSi3O8 fluids provide a better foundation for modeling high-P metasomatic processes than pure H2O values. The strong increase in rutile solubility with dissolved Na–Al silicate suggests that complexing with these constituents promotes Ti mobility and transport during fluid–rock interaction in the lower crust and upper mantle.