On the growth of Al2O3 scales

Understanding the growth of Al2O3 scales requires knowledge of the details of the chemical reactions at the scale–gas and scale–metal interfaces, which in turn requires specifying how the creation/annihilation of O and Al vacancies occurs at these interfaces. The availability of the necessary electrons and holes to allow for such creation/annihilation is a crucial aspect of the scaling reaction. The electronic band structure of polycrystalline Al2O3 thus plays a decisive role in scale formation and is considered in detail, including the implications of a density functional theory (DFT) calculation of the band structure of a Σ7 image bicrystal boundary, for which the atomic structure of the boundary was known from an independent DFT energy-minimization calculation and comparisons with an atomic-resolution transmission electron micrograph of the same boundary. DFT calculations of the formation energy of O and Al vacancies in bulk Al2O3 in various charge states as a function of the Fermi energy suggested that electronic conduction in Al2O3 scales most likely involves excitation of both electrons and holes, which are localized on singly charged O vacancies, image and doubly charged Al vacancies, image, respectively. We also consider the variation of the Fermi level across the scale and bending (“tilting”) of the conduction band minimum and valence band maximum due to the electric field developed during the scaling reaction. The band structure calculations suggest a new mechanism for the “reactive element” effect—a consequence of segregation of Y, Hf, etc., to grain boundaries in Al2O3 scales, which results in improved oxidation resistance—namely, that the effect is due to the modification of the near-band edge grain-boundary defect states rather than any blocking of diffusion pathways, as previously postulated. Secondly, Al2O3 scale formation is dominated by grain boundary as opposed to lattice diffusion, and there is unambiguous evidence for both O and Al countercurrent transport in Al2O3 scale-forming alloys. We postulate that such transport is mediated by migration of grain boundary disconnections containing charged jogs, rather than by jumping of isolated point defects in random high-angle grain boundaries.