Structure, optical absorption, and luminescence energy calculations of Ce3+ defects in LiBaF3

We address two remarkable features in the optical behavior of Ce3+ defects in LiBaF3: the fourfold splitting of the Ce3+ 5d manifold in a cubic system, and the unusually large Stokes shift, of around 1 eV (approximate to 9000 cm(-1)), between the energy of the lowest Ce3+ 4f --> 5d absorption line and its 5d --> 4f luminescence energy. To this end we investigated the electronic properties and the structure of several possible luminescence center configurations in LiBaF3:Ce3+, each consisting of a Ce3+ substitution at a Ba or Li site, plus an appropriate charge-compensating defect. Using a plane-wave pseudopotential density-functional-based method to optimize the geometry of a supercell consisting of 3 x 3 x 3 LiBaF3 unit cells, containing a single luminescence center, the equilibrium structures of these defect complexes were determined. We performed nb initio cluster calculations at the Hartree-Fock level to determine the optical-absorption energies of the Ce3+ 4f --> 5d transitions in these different geometries. Comparison of these energies with the results of optical-absorption measurements on LiBaF3:Ce3+ shows that the most likely luminescence center configuration consists of Ce3+ at a Ba site, charge compensated by the substitution of one of its nearest-neighboring Ba ions by a Li+ ion. For this configuration we have repeated the cluster and supercell calculations with Ce3+ in the [Xe]5d(1) excited-state electronic configuration to determine the Ce3+ 5d-->4f luminescence energy and to study effects that can explain the large Stokes shift in this material. These calculations predict an extensive lattice relaxation, induced by the excitation of the Ce3+ ion, and yield a Stokes shift of 0.61 eV (compared to 1 eV found from experiment). The origin of this large Stokes shift is identified as a strong coupling of the crystal-field splitting of the Ce3+ 5d manifold to the displacement of four of its F nearest neighbors.