It was noted that the migration of O in this material corresponded to an apparently simple jump between neighboring bridge sites. However, theoretical calculations had produced conflicting results and had failed to provide a satisfactory explanation for the observed activation energy of 2.5eV. Comprehensive first-principles calculations were presented here which demonstrated that the apparently simple O jump was actually a complex process which involved coupled barriers and could be properly described quantitatively in terms of an energy hyper-surface with a saddle ridge and an activation energy of about 2.5eV. Earlier calculations had corresponded to various points or lines on this hyper-surface.

M.Ramamoorthy, S.T.Pantelides: Physical Review Letters, 1996, 76[2], 267-70