Exciton Models in the Alkali Halides

Abstract

The common models for excitons in the alkali halides are criticized as to their goodness as zero-order models. Some new calculations on energies, oscillator strengths, and charge distributions are made with the models as they are commonly used. It is shown that the customary use of an LCAO (linear combination of atomic orbitals) method with the electron-transfer model gives qualitatively incorrect results, particularly for the wave functions and charge distributions, results which are not even internally consistent with the model. It is argued that the other model, an excitation model, gives at least semiquantitatively correct results, and is a satisfactory zero-order approximation. Finally, it is shown that in the alkali halides the LCAO method must inevitably be an inconvenient one for finding accurate charge distributions and wave functions for localized centers.