Density functional calculations were used to study the ionic conductivity in γ-Li3PO4 and γ-Li2.88PO3.73N0.14. Starting from the crystal structure of γ-Li2.88PO3.73N0.14, a model cluster without defects (Li15PO10) was constructed; as well as another new oxynitride (Li14PO8N) in which Li and O defects were introduced as one O atom was replaced by N. In order to model the ionic conductivity, various pathways for Li motion were considered. The first of them involved Li+ motion between 2 crystallographic sites through the faces of adjacent LiO4 tetrahedron via an unoccupied octahedral site. The second one involved direct Li+ motion through the faces of adjacent LiO4 tetrahedra. Both mechanisms were unlikely, for the parent model cluster, because of the high computed energy barrier which was associated with Li+ mobility in the cluster. A reasonable energy barrier in the nitride cluster was obtained which involved Li+ and O2- defect creation and incorporated N. The barrier was calculated to be about 1.26eV for Li+ mobility through tetrahedral faces for the nitride structure; compared with 4.8eV in the parent cluster. Upon considering parameters such as Li–N covalency, ionic radius and tetrahedral distortion, nitridation was expected to enhance ionic conductivity. The magnitude of the ionic conductivity was related to the height of the energy barrier to Li+ jumps between different crystallographic sites.

Theoretical Approach to Ionic Conductivity in Phosphorus Oxynitride Compounds. H.Rabaâ, R.Hoffmann, N.C.Hernández, J.F.Sanz: Journal of Solid State Chemistry, 2001, 161[1], 73-9