Esca. results versus other physical and chemical data

Abstract

X-ray photoemission spectra yield quantities of very direct interest in physics and chemistry. In this paper the relations of these spectra to other data and concepts are discussed. Both initial-state and final-state properties may be studied: the former are treated first. Charge distributions in molecules alter the effective (Coulomb plus exchange) potential experienced by core electrons in molecular ground states, there by shifting their binding energies. The shifts can be calculated by abinitio methods or more directly by using potential models based on intermediate-level molecular-orbital theories such as INDO. One version, the ground-state potential model (GPM) yields good predictions of core-level shifts among atoms in similar environments. Alternatively, the measured shifts may be used to derive charges on individual atoms in molecules. It is more difficult to derive charges in solids in this way, but a characteristic splitting in the more tightly-bound valence bands yields a direct measure of ionicity in simple binary compounds of the zinc-blende and rocksalt structures. Atomic orbital composition of molecular orbitals can be deduced from photoemission spectra. In solids such as diamond and graphite comparison of photoemission spectra with x-ray emission spectra yields the atomic-orbital composition of the valence bands. Turning to final-state properties, the spectra are dominated by relaxation effects. Again a simple approach—the relaxation potential model (RPM)—predicts core-level shifts well for cases in which the atomic environments are varied substantially. Among ammonia and the methylamines, for example, the N(ls) shifts are predicted correctly by RPM, while GPM reverses the order. For paramagnetic molecules RPM predicts electron charge transfer toward the positive hole but usually spin transfer away, in agreement with experiment. Extra-atomic relaxation in metals, a many-body effect, is manifest both as a contribution to the binding energy and as line-shape asymmetry. Delocalized valence electrons also show relaxation shifts that can be understood as polari zation of the electron gas toward the “Coulomb hole”. Auger lines show larger relaxation shifts. Comparison of core-level or Auger shifts in nonmetallic solids separately is questionable because there is no reference level, but intercomparison of the two is meaningful. Finally, core-level binding-energy trends in series of simple alcohols, etc., agree quantitatively with proton affinities and core-level shifts in other functional groups. This suggests extending the concept of Lewis basicity to include lone pairs of core electrons. Thus, core-level shifts measure the chemical reactivity—a quantity of great chemical importance that depends on both initial- and final-state properties—rather directly. Rela xation energies are shown to be the dominant cause of trends in the lowest ionization potentials of simple alcohols and amines.