The formation of an ultrathin aluminum oxide film at the ${\mathrm{Fe}}_{0.85}{\mathrm{Al}}_{0.15}$(110) surface (${\mathrm{A}}_2$ random alloy) has been studied by a variety of surface-sensitive techniques (x-ray photoemission, low-energy electron diffraction, surface x-ray diffraction, and scanning tunneling microscopy) supplemented by ab initio atomistic simulations. Since iron is not oxidized in the conditions used, the study focused on the coupling between aluminum oxidation and segregation processes. Compared to the bare surface, whose average composition (${\mathrm{Fe}}_{0.6}{\mathrm{Al}}_{0.4}$) is closer to the ${\mathrm{B}}_2$-CsCl structure over a $\sim 3$ nm depth, the oxidation hardly affects the subsurface segregation of aluminum. All the structural and chemical fingerprints point to an oxide film similar to that found on NiAl(110). It is a bilayer ($\sim 7.5\AA$ thick) with a composition close to ${\mathrm{Al}}_{10}{\mathrm{O}}_{13}$ and a large $(18.8\times 10.7){\AA }^2$ nearly rectangular unit cell; an almost perfect match between substrate periodicity and the $(1\times 2)$ oxide supercell is found. Nevertheless, microscopy reveals the presence of antiphase domain boundaries. Measured Al $2p$ and O $1s$ core-level shifts match calculated ones; their origin and the relative contributions of initial/final state effects are discussed. The ubiquity of the present oxide on different supports asks for the origin of its stability.

• Received 3 January 2020
• Revised 17 April 2020
• Accepted 30 June 2020

DOI:https://doi.org/10.1103/PhysRevMaterials.4.074409