EXPLORING MDD BEHAVIOR IN HEMATITE: WHAT CAUSES IT AND WHAT CAN WE DO WITH IT

Hematite shows potential to record thermal histories over a very wide range of temperatures through the combined (U-Th)/Ne, (U-Th)/He, and ⁴He/³He systems. A key feature that differentiates hematite from other phases used for thermochronology is that it is commonly found as a dense aggregate of crystallites that range in size from nm to um scales. This polycrystalline structure has been interpreted as the source for Multiple Diffusion Domain (MDD) behavior, which we investigate in two ways. First, physically crushing samples with previously observed highly reproducible MDD behavior alters the diffusive results by removing the largest diffusion domain. Second, a museum quality, cm-scale specular hematite sample from Minas Gerais, Brazil was confirmed to be a large single crystal through SEM and EBSD analysis. Material from this crystal was proton irradiated to produce ³He and diffusion experiments were conducted on crushed aliquots sieved to 294 and 680 µm. The single crystal aliquots showed none of the MDD behavior observed in other samples and when corrected for size, the diffusive results between the two aliquots are indistinguishable. Taken together these observations suggest that (1) MDD behavior does not occur in single crystals of hematite, and (2) the physical crystal size is the diffusion domain. Using the kinetics observed from the single crystal aliquots, we confirm that He is strongly retained in hematite with closure temperature ranging from ~60 to ~200˚C for 10 nm to 100 µm crystals. Given this wide thermal range, we explore the sensitivity of inverting the MDD data to infer thermal histories through forward models. We find that models perform very well for simple linear or isothermal cooling histories. For complex histories such as reheating, models are sensitive to the timing and temperature of maximum reheating and subsequent cooling of the most recent event, however the onset of heating or previous events are not resolved.