Use of “Shake-up” Satellites in Photoelectron Spectra for Analysis of Oxide Layers on Metals

Abstract

THE oxide films which protect most commercial, oxidation-resistant alloys from further oxidation are so thin (<0.5 µm) that they cannot be examined in section by conventional metallographic techniques. Similarly the use of the electron-probe microanalyser is limited to films thicker than 1 µm both because of its precision and because of the rather deep penetration of the high energy electron beam used to stimulate X-ray emission. Several diffraction techniques can be used to elicit the crystalline structure of the surface atomic layers, but this does not obviate the need for compositional data for determining which elements in complex alloys are responsible for their protective oxidation behaviour. Analysis by radio-activation, is useful for certain elements but is not yet widely adopted. Many workers are now attempting to obtain similar information by use of Auger and photoelectron (ESCA) spectroscopy1. The photoelectrons ejected by almost monochromatic X-rays are both characteristic of their parent element and show small shifts in their characteristic energy according to its chemical state. Thus, in principle, the emission spectral lines of the oxide layer can be distinguished from those of the metal. With layers of the order of 0.005 µm, however, it could be necessary to deconvolute them from the overlapping spectral lines of the underlying metal-with all the associated difficulties. It is for this reason that the announcement by Novakov2 that certain satellites of principal spectral lines are characteristic of individual compounds, particularly copper and nickel oxides, is of importance to the study of thin oxide layers. The satellite peaks are associated with discrete energy losses from certain photoelectrons by the so called “shake-up” processes3,4, that is the excitation of electrons in the valence band to higher energy states. The purpose of this communication is to show that the satellite peaks of copper and nickel oxides can be seen in the spectra of thin (0.1 µm) oxide layers on alloys of these metals as well as in those obtained from the pure oxides by Novakov2,5 and to give an example of their use in oxidation research.