Experimental calibration of clumped isotope reordering in dolomite

Dolomite clumped isotope compositions are indispensable for determining the temperatures and fluid sources of dolomitizing environments, but can be misleading if they have modified since formation. Carbonate Δ47 values are susceptible to resetting by recrystallization during diagenesis, and, even in the absence of dissolution and reprecipitation reactions, alteration by solid-state reordering during prolonged residences at elevated temperatures. In order to understand the potential of dolomite Δ47 values to preserve the conditions of dolomitization in ancient sections, the kinetic parameters of solid-state reordering in this phase must be determined. We heated mm-sized crystals of near-stoichiometric dolomite in a René-type cold seal apparatus at temperatures between 409 and 717 °C for 0.1-450 h. In order to prevent the decarbonation of dolomite to calcite, periclase, and CO2 at these conditions, the system was pressurized with CO2 to 0.45-0.8 kbar. Over the course of 31 temperature-time points and 128 individual Δ47 measurements of powdered dolomite crystals from these points, we observed the evolution of dolomite Δ47 values from the initial (unheated) composition of the crystals (0.452 ± 0.004‰, corresponding to a formation temperature of ∼145 °C) towards high-temperature equilibrium distributions. Complete re-equilibration occurred in the 563-717 °C experiments. As with previous heating experiments using calcite and apatite, dolomite Δ47 exhibited complex reordering behavior inadequately described by first-order Arrhenian-style models. Instead, we fit the data using two published models for clumped isotope reordering: the transient defect/equilibrium defect model of Henkes et al. (2014), and the exchange-diffusion model of Stolper and Eiler (2015). For both models, we found optimal reordering parameters by using global least-squares minimization algorithms and estimated uncertainties on these fits with a Monte Carlo scheme that resampled individual Δ47 measurements and re-fit the dataset of these new mean values. Because the exact Δ47-T relationship between 250 and 800 °C is uncertain, we repeated these fitting exercises using three published high-temperature Δ47-T calibrations. Regardless of calibration choice, dolomite Δ47 rate constants determined using both models are resolvably slower than those of calcite and apatite, and predict that high-grade dolomite crystals should preserve apparent equilibrium blocking temperatures of between ∼210 and 300 °C during cooling on geologic timescales. Best agreement between model predictions and natural dolomite marbles was found when using the exchange-diffusion model and the ab initio Δ63-T calibration of Schauble et al. (2006), projected into the Δ47 reference frame by Bonifacie et al. (2017). Therefore, we recommend modeling dolomite Δ47 reordering using the exchange-diffusion model and this parameter set. In simple heating scenarios, the two models disagree. The transient defect/equilibrium defect model suggests that dolomite fabrics resist detectable reordering at ambient temperatures as high as 180 °C for tens of millions of years, while the exchange-diffusion model predicts incipient partial reordering perhaps as low as 150 °C. In either case, barring later recrystallization, dolomite Δ47 values should be faithful recorders of the conditions of dolomitization in sedimentary sections buried no hotter than ∼150 °C for tens of millions of years.