E. Bruce Watson, Daniele J Cherniak, Claire I O Nichols , Ben P Weiss

The ferrimagnetic mineral magnetite (Fe3O4) is abundant in banded iron formation (BIFs), and has the potential to provide U-Pb or Pb-Pb age information on these rocks because it incorporates small amounts of U during growth. Combined with age measurements, paleomagnetic studies of BIF magnetites may also yield insight into the history of Earth’s magnetic field and its relationship to early evolution of Earth’s interior and atmosphere. Reliable magnetite ages utilizing Pb isotopes require knowledge of Pb diffusion in the magnetite structure. For this reason, we undertook an experimental investigation of Pb diffusion in magnetite by diffusing Pb2+ ions into pre-polished slabs of natural magnetite oriented parallel to {001} or {111}. A mixture of PbSO4 and Fe2O3 was used as a surface powder source to supply Pb2+ diffusant at the sample surface and at the same time buffer the oxygen fugacity of the system at magnetite-hematite (MH) - a typical fO2 for banded iron formations (BIFs) due to the common presence of both iron oxides (and where Pb2+ is stable relative to other Pb valence states). Diffusion experiments spanned temperatures of 500−675°C and durations of 75 to 2035 h. Following each experiment, in-diffused Pb was depth-profiled using Rutherford backscattering spectroscopy (RBS) and Pb diffusivities were calculated from the profiles using an infinite half-space diffusion model. The following diffusion law for Pb2+ in magnetite is based upon 12 independent diffusivity measurements: DPb (m2·s-1) = (9×10-17 m2·s-1) exp(-98,000 J·mol-1)/RT) where the uncertainties in the pre-exponential constant and activation energy are ±6% and ±15%, respectively. Pb diffusion in magnetite over the temperature range of our study is orders of magnitude slower than projected for other divalent cations based on down-temperature extrapolation of previously measured diffusion laws (e.g., for Mn2+, Fe2+, Co2+, and Ni2+). This finding is encouraging in terms of the potential suitability of magnetite for U-Pb age determinations of BIFs and other magnetite-bearing rocks. Indeed, classical Dodson closure temperatures well above 500°C are not unrealistic in cases where magnetite crystals having large diffusion domains (e.g., >100µm in radius) are cooled relatively rapidly (e.g., at 100°C/MYr). This is of particular significance for paleomagnetic studies, since the Curie temperature of magnetite is 580°C and therefore the age of magnetization in magnetite-bearing rocks may be directly dated. However, slow cooling of magnetites having small diffusion domains can lead to Pb loss at temperatures of 200°C or lower. Pb mobilization is evaluated for various time-temperature scenarios that involve both heating and cooling as well as “closed-loop” time-temperature paths. We conclude that U-Pb or Pb-Pb age determinations of BIF magnetites are potentially reliable, but isotopic results should be assessed in concert with knowledge of the thermal history of the host rock and the effective grain size ofthe magnetites.