The atomic and electronic structure, formation energy, and the energy barriers for migration have been calculated for the neutral O vacancy point defect ($F$ center) in cubic $\mathrm{Sr}\mathrm{Ti}{\mathrm{O}}_3$ employing various implementations of density functional theory (DFT). Both bulk and $\mathrm{Ti}{\mathrm{O}}_2$-terminated (001) surface $F$ centers have been considered. Supercells of different shapes containing up to 320 atoms have been employed. The limit of an isolated single oxygen vacancy in the bulk corresponds to a 270-atom supercell, in contrast to commonly used supercells containing $\sim 40–80$ atoms. Calculations carried out with the hybrid B3PW functional show that the $F$ center level approaches the conduction band bottom to within $\sim 0.5\mathrm{eV}$, as the supercell size increases up to 320 atoms. The analysis of the electronic density maps indicates, however, that this remains a small-radius center with the two electrons left by the missing O ion being redistributed mainly between the vacancy and the $3d\left(z^2\right)$ atomic orbitals of the two nearest Ti ions. As for the dynamical properties, the calculated migration energy barrier in the low oxygen depletion regime is predicted to be $0.4\mathrm{eV}$. In contrast, the surface $F$ center exhibits a more delocalized character, which leads to significantly reduced ionization and migration energies. Results obtained are compared with available experimental data.

Using the B3PW hybrid exchange-correlation functional within density-functional theory and employing Gaussian-type basis sets, we calculated the atomic and electronic structures and thermodynamic stability of three double-layered (DL) $\mathrm{Sr}\mathrm{Ti}{\mathrm{O}}_3\left(001\right)$ surfaces: (i) SrO-terminated, (ii) $\mathrm{Ti}{\mathrm{O}}_2$-terminated, and (iii) $(2\times 1)$ reconstruction of $\mathrm{Ti}{\mathrm{O}}_2$-terminated $\mathrm{Sr}\mathrm{Ti}{\mathrm{O}}_3\left(001\right)$ recently suggested by Erdman et al. [Nature (London) 419, 55 (2002)]. A thermodynamic stability diagram obtained from first-principles calculations shows that regular $\mathrm{Ti}{\mathrm{O}}_2$- and SrO-terminated surfaces are the most stable. The stability regions of $(2\times 1)$ DL $\mathrm{Ti}{\mathrm{O}}_2$- and DL SrO-terminated surfaces lie beyond the precipitation lines of SrO and $\mathrm{Ti}{\mathrm{O}}_2$ compounds and thus are less stable with respect to regular $\mathrm{Sr}\mathrm{Ti}{\mathrm{O}}_3\left(001\right)$ surfaces. Analysis of the stability diagram suggests that Sr precipitation on $\mathrm{Sr}\mathrm{Ti}{\mathrm{O}}_3$ surface never occurs. Our simulations show a substantial increase of Ti-O covalency on the DL surfaces as compared to the regular surfaces, which are themselves more covalent than the crystalline bulk. The implications of our calculated results for recent experimental observations are discussed.

We present the results of ab initio hybrid density-functional calculations of the atomic and the electronic structures of ${\text{LaMnO}}_3$ (LMO) and ${\text{La}}_{1-x_b}{\text{Sr}}_{x_b}{\text{MnO}}_3$ (001) surfaces. The total energies obtained from these calculations were used to analyze thermodynamic stability of the surfaces. We predict Sr and O vacancy segregation to the surface to occur with similar energies ($\sim 0.5\text{eV}$ per defect). In pure LMO only ${\text{MnO}}_2$ termination is thermodynamically favorable under typical operational conditions of a cathode in solid oxide fuel cells, whereas Sr doping makes La(Sr)O termination favorable. Finally, the role of Sr doping in cathode degradation is discussed.