L2,3VVandMVVAuger spectra of copper

Abstract

Electron-excited integral $L_{2,3}\mathrm{VV}$ and $\mathrm{MVV}$ Auger spectra from clean copper surfaces are presented and compared with x-ray excited $L_{2,3}\mathrm{VV}$ spectra, with other electron-excited derivative $\mathrm{MVV}$ spectra, and with the results of atomic-model and band-theory calculations. Our $L_{2,3}\mathrm{VV}$ spectra agree well with the x-ray-excited results in both line shape and $L_3\mathrm{VV}-\mathrm{to}-L_2\mathrm{VV}$ integrated-intensity ratio. Recent atomic-model interpretations of these spectra by McGuire are briefly reviewed and found, especially for the $L_3\mathrm{VV}$ spectrum, to provide a very good explanation of the spectra: the multiplet splitting mechanism is responsible for the major peaks, and satellite-intensity contributions resulting from Coster-Kronig $L_1{L}_{2}V$, $L_1{L}_{3}V$, and $L_2{L}_{3}V$ transitions give rise to the low-energy portions of the spectra. Our $\mathrm{MVV}$ spectra also have sharp features with shapes that are in agreement with atomic-model calculations. The experimental $M_{2,3}\mathrm{VV}-\mathrm{to}-M_1\mathrm{VV}$ integrated-intensity ratio is much larger than 3 to 1, and varies with primary-electron-beam energy, indicating that satellite intensity contributions make up part of the $\mathrm{MVV}$ signals. The sharp features in these spectra, however, do not change with variations in primary-beam energy from 119 to 3000 eV indicating that they are not dependent upon satellite intensity. The experimental $M_3\mathrm{VV}-\mathrm{to}-M_2\mathrm{VV}$ integrated-intensity ratio also remains unchanged at 1.3 to 1 for the same variation in primary-beam energy. A broad high-energy shoulder forms part of the $\mathrm{MVV}$ spectra, especially for the $M_1\mathrm{VV}$ signal. This feature of the $\mathrm{MVV}$ line shapes is not in agreement with the atomic-model calculations. It is also not simply related to the undistorted valence-band density of states for copper. Comparison of our $M_1\mathrm{VV}$ signal with recent band-theory arguments of Cini and of Sawatzky indicates that the $\mathrm{MVV}$ line shapes are explainable in terms of distortions in the Auger core-valence-valence signals from narrow-valence-band metals that are expected to occur as a result of hole-hole interactions in the two-hole final-state configuration. Estimates of the hole-hole repulsion energy based on these comparisons are discussed in light of an earlier estimate of this interaction energy based on $L_3\mathrm{VV}$ data.