Search for: All records

Oxygen isotopes are a well-known geochemical tool with applications to equilibrium thermometry, fluid tracing, and magma and ore petrogenesis. High-precision, high-spatial resolution oxygen isotope analysis by SIMS has also enabled the development of oxygen isotopes as a tool for geospeedometry. Here, we detail Fast Grain Boundary (FGB), an updated computational approach and software tool for determining time-temperature (T-t) histories through modeling of oxygen isotope diffusion. FBG models a rock system, rather than a single phase (cf. thermochronometry based on He, Ar, Pb), and has the potential to constrain continuous thermal histories over a wide range of temperatures, including at high temperature (500-800°C). The new FGB also allows for inversion of the FGB model to extract thermal histories from intragrain oxygen isotope zoning data using the Levenberg-Marquardt (LM) algorithm. Tests with synthetic datasets show that the LM algorithm is able to distinguish between simple linear cooling and more complex thermal histories containing, for instance, reheating events. Inversion of an actual oxygen isotope data set from titanite are consistent with the previously determined T-t path for the sample region, showing a brief period of >700 °C conditions, followed by cooling below 500 °C in <5 m.y.. However, the inversion suffers from a flat-bottomed minimum and does not produce a well-converged T-t path. These results point to analytical precision as a continuing challenge in recovering tightly constrained thermal histories for the real data set and emphasize the need for further development of high-precision microanalytical oxygen isotope standards. In the meantime, we use FGB modeling to explore sampling and analytical approaches that improve the resolution of inversion solutions for current analytical capabilities. For instance, inversion most successfully recovers a well constrained T-t path solution when SIMS analysis targets oxygen isotope gradients developed near grain rims, as opposed to oxygen isotope values in grain centers. Additional tests that probe the sensitivity of the inversion results to modal mineralogy and relative grain sizes suggest that careful targeting of samples in the field can enhance the recovery of unique T-t paths. 

We present an updated and expanded version of the Fast Grain Boundary (FGB) program, originally developed by [1]. In its current form, FGB forward models the oxygen-isotope compositional evolution of a rock resulting from diffusive oxygen isotope exchange. FGB is a tool for 1) constraining thermal histories (cooling rates and durations) from measured intragrain oxygen isotope zoning profiles, and 2) predicting oxygen isotope zoning that results from coupled volume and grain boundary diffusion. The FGB model is mass-balance constrained through exchange with a finite, grain-boundary reservoir and does not require a Dodson-like infinite reservoir assumption. The new FGB program code is written in Python and includes a graphical user interface. The inverse modeling capabilities for FGB are currently under development. We present preliminary results for thermal history inversion from a test case using oxygenisotope diffusion zoning data from titanite. Both the gradient descent and the Levenberg-Marquardt (LM) algorithms are applied to the FGB model in search of cooling histories that maximize agreement between the model output and recorded data. Various schemes of regularization are employed to ensure meaningful realizations of cooling histories. Additionally, to prevent local extrema entrapment, the results of these algorithms are compared to long-run brute force methods that have been implemented through Amazon’s cloud computing services. The observed zoning profiles in the example titanite dataset can be forward modeled with several arbitrary (and potentially biased) cooling histories. The preliminary results of inverse modeling reduce initial bias and suggest an episodic cooling history that may include an isothermal period or even a reheating event. These results demonstrate both the potential for oxygen-isotope zoning to preserve prevailing thermal events and the potential of the FGB model for recovering these events. [1] Eiler, Baumgartner & Valley (1994), Computers & Geoscience 20, 1415-1434.