Electronic and atomic structure computation of disordered low index surfaces of [formula omitted]-alumina

Abstract

Electronic and atomic structure of low index polar (0 0 1, 0 1 1, 1 1 1) spinel γ -alumina surfaces have been investigated using a modified large-scale quantum semi-empirical simulations. Atomic structure optimization and electronic structure calculation were done in the direct space on periodic large unit cells in which random distribution of cationic vacancies close to the surface were included. Predicted electronic structures match more accurate ab initio Density Functional Theory (DFT) results on aluminum oxides. We found that the restructured surfaces behave as amorphous-like shapeless planes provided unit cell dimensions are much greater than primitive translations as periodicity constraints are weakened. Aluminum ions vacancies in the surface region are shown to induce a surface atomic disordering even at 0 K, correlated with the appearance of low coordinated ions, as proven experimentally and theoretically with classical Molecular Dynamics (MD) calculation on some aluminum oxide systems. The lowering of the surface charge density is obtained from two distinct mechanisms involving large ion movements. On the oxygen rich plane termination, this results from the formation of O 2 δ− entities. On the aluminum rich plane termination, the same effect is related to a loss of co-ordination number together with a reduction of cationic charge. Furthermore, large-scale simulations allow to statistically quantify specific relaxed surface ion densities for which the Local Density of States (LDOS) is evaluated. Their inferred electronic properties are then compared to available probe molecule adsorption experiments investigated by infrared spectroscopy. We show that surface electronic states are not a simple function of ion coordination. Our large-scale quantum semi-empirical calculation fills the gap between atomic structures predicted by classical MD approach with electronic structure results obtained by DFT on small systems.