X-Ray Photoelectron Spectroscopy of Alkali Halides

We have measured ionization potentials for the removal of various electrons from the alkali chlorides and the sodium and cesium halides. These binding energies are compared with the predictions of a simple point charge model with and without corrections for polarization and repulsion. The simple model predicts and the data show that the spacings of the energy levels for a given ion are independent of what crystal it is in and are the same as for the free ion. The point charge model also allows us to calculate the difference between cation and anion energy levels in the same crystal. There is, however, a disagreement between the predicted and experimental values of this difference that ranges from about 1.8 eV for LiCl to −0.2 eV for RbCl. This discrepancy is markedly reduced by inclusion of polarization effects. The point charge model with polarization and repulsion corrections predicts absolute ionization potentials for the alkali and halide ions that differ in a systematic way from those observed. For the sodium halides, the difference between calculated and experimental energies decreases monotonically from about −4 eV for NaF to about 0.9 eV for NaI. The origin of these discrepancies is apparently due to charging of the samples and their trend is directly attributable to the size of the respective bandgaps. We show that the electrostatic model can be used to provide a comparison between experimental binding energies for inner electrons in a crystal and Hartree-Fock calculations for binding energies of inner electrons in free atoms. Finally, using the point charge model we show that the binding energy of an electron on a highly charged ion in an ionic crystal is only slightly different from the binding energy of the same electron in the neutral atom.