The characterisation of non-evaporable getters by Auger electron spectroscopy: Analytical potential and artefacts

The surfaces of getter materials are particularly difficult to analyse because of their high chemical reactivity. The results obtained can be strongly influenced by the experimental set-up and procedures. In this paper the experimental influence on the Auger electron spectroscopy results is discussed, based on the measurements of more than 100 different non-evaporable getter (NEG) materials. There are four typical changes in the Auger electron spectra when a NEG becomes activated. The oxygen peak intensity decreases, the shape of the metal peaks changes, the carbon peak shape changes shape and intensity and a chlorine peak occurs. All these changes are affected by instrumental artefacts. The Zr-MNV peak shape changes occurring during the reduction of ZrO2 are well suited to determine the onset of NEG activation, while the slope with which the O-KLL peak intensity decreases in a certain temperature range is a better criterion for the determination of the temperature at which activation is complete. The O-KLL intensity and the Zr-MNV peak shape are influenced by the adsorption of residual gas, in particular of H2O, and by electron stimulated desorption of oxygen. The C-KLL peak shape changes from "graphitic" to "carbidic" when the NEG becomes activated. The changes of the C-KLL intensity observed in the present study are mainly caused by the transfer of carbon contamination from the sample holder onto the sample during NEG heating. The presence of chlorine at submonolayer coverage is easily detected by AES but not by XPS. The temperature at which a Cl-LMM peak appears in the NEG Auger electron spectra is correlated with the onset of NEG activation. The Cl-LMM intensity is strongly reduced by electron irradiation, in particular when AES measurements are carried out on hot samples.