Atomic exchange potentials and angle resolved photoemission

Abstract

The usefulness of angle-resolved photoemission from core states of adatoms for surface structure studies has been questioned because of the total mismatch between theory and experiment for normal emission from the 4d levels of Te on Ni(100) as a function of photon energy. We point out that the theory was based on the use of an Xα potential for Te with a α value appropriate for ground state properties. We show that such a potential is unsuitable for scattering problems and that much lower a values are required to give a reasonable description of sharp peaks in the atomic photoionization cross section as a function of photon energy. We show that Xα calculations of the phase shifts give much improved correspondence with full Hartree-Fock calculations when α is reduced below its usual range of 1 to 23, and that a corresponding improvement in the matching of experimental and calculated photoelectron azimuthal distributions is found. A realistic exchange potential must be energy dependent, decreasing to zero at high enough energies. We show that the simplest reasonable approximation to the local density exchange potential, i.e. the Hartree-Fock approximation for the uniform electron gas, resolves the problem of the normal emission data from Te 4d on Ni(100).