Ti-in-quartz: Evaluating the role of kinetics in high temperature crystal growth experiments

We present results from 25 hydrothermal quartz growth experiments, all conducted at 800 °C and 1 kbar but with varying starting materials and run times, to address discrepancies between calibrations of the titanium-in-quartz (TitaniQ) thermobarometer. In our experiments, a gold capsule is loaded with silica glass, water, and either rutile or anatase as the TiO2 source. In most experiments, there is also a large quartz seed crystal contained in an open inner capsule. The use of rutile versus anatase has a significant influence on the (re)crystallization pathways of the SiO2 and TiO2 components. When rutile is used, quartz overgrowths have abundant open cavities and complex zonations. The rutile does not completely dissolve because rutile is the stable TiO2 polymorph, and yet, new rutile forms at the quartz seed-overgrowth interface and on the outer surface of quartz crystals. This suggests crystallization of quartz near Ωrut ∼ 1, but wide-ranging Ti concentrations and zonations in quartz are indicative of kinetic effects. When powdered anatase is used, the quartz overgrowths look markedly different, lacking the open cavities and instead exhibiting step edges and terraces. The Ti concentrations in quartz from these experiments are also wide-ranging but reach larger values. Our results span the range of previous calibrations and indicate that Ti concentrations in quartz are sensitive to the TiO2/SiO2 ratio of the fluid as opposed to the absolute concentration (or activity) of dissolved TiO2. We present a kinetic model for quartz and rutile growth from a fluid where the input parameters are the initial degrees of supersaturation with respect to quartz and rutile, the total reactive surface area, and rate constants that link the degree of supersaturation to net precipitation rates. The model can explain many of the salient features of our experimental results, as well as those from previous studies, but requires that the rate constant multiplied by the reactive surface area for rutile is less than that of quartz, and that rutile solubility depends on the SiO2 concentration of the fluid, as documented in the recent literature. Complete quartz-rutile equilibrium may not have been established in any of the experimental studies, but low-pressure experiments with slowly grown quartz seem to be more reliable than extrapolations from high-pressure experiments for thermobarometry of shallow natural systems.