Reevaluating Reaction Rates from Diffusion Profiles in Minerals and Effects on Trace Element Thermometers.

The overgrowth of one mineral upon another (corona) can be associated with a diffusion profile in the reactant (core) crystal. In principle, these profiles can be used to infer reaction rates using a steady state reaction model (constant velocity movement of the reactant/product grain boundary). In practice, however, natural data do not conform to these models. Assuming mass conservation and equilibrium partitioning between the two reactant and product minerals, steady state reaction predicts an exponential profile. Rutile that was overgrown by titanite sometimes show quasi-exponentially decreasing profiles in trace elements, suggesting reaction rates might be inferred. However, the steady-state reaction model also predicts that: (1) The mass excess in the diffusion profile in rutile should be balanced by a mass deficit in titanite, (2) Titanite should be zoned and exhibit a compositional trend complementary to the profile in rutile, and (3) Trace element partitioning between titanite and rutile should be maintained as the corona develops and should be preserved at the rutile-titanite interface. None of these three predictions are observed in natural titanite overgrowths on rutile, rather titanite is generally unzoned and significant trace element mass deficits or excesses indicate partial to complete interaction with other matrix minerals. Reaction rates might still be obtainable if titanite and rutile rim compositions continuously equilibrate with the matrix during reaction. If so, however, temperatures from Zr-in-titanite (ZiT) and Zr-in-rutile (ZiR) should be identical. Yet commonly they are not, even accounting for potentially low activities of zircon and quartz at the reaction interface, rather T(ZiT) > T(ZiR). Explanations for the diffusion profiles in rutile range from steadily increasing or decreasing reaction rates, or even post-reaction development of a diffusion profile with a fixed boundary condition. While reaction rates cannot be quantified, the models help explain higher T(ZiT) in some rocks than expected from either ZiR or cation exchange thermometers: incomplete loss of Zr from the rutile-titanite reaction zone locally increased Zr content of titanite and apparent T(ZiT). These results suggest caution in interpreting T(ZiT) when titanite overgrows other Ti-bearing minerals.