We report an experimental and theoretical analysis of the root 3 x root 3-R30 degrees and 2 x 2 reconstructions on the MgO (111) surface combining transmission electron microscopy, x-ray photoelectron spectroscopy, and reasonably accurate density functional calculations using the meta-generalized gradient approximation functional TPSS [Tao, Perdew, Staroverov, and Scuseria, Phys. Rev. Lett. 91, 146401 (2003)]. The experimental data clearly show that the surfaces contain significant coverages of hydroxyl terminations, even after UHV annealing, and as such cannot be the structures which have been previously reported. For the 2 x 2 surfaces a relatively simple structural framework is detailed which fits all the experimental and theoretical data. For the root 3 x root 3 there turn out to be two plausible structures and neither the experimental nor theoretical results can differentiate between the two within error. However, by examining the conditions under which the surface is formed we describe a kinetic route for the transformation between the different reconstructions that involves mobile hydroxyl groups and protons, and relatively immobile cations, which strongly suggests only one of the two root 3 x root 3 structures.
