Quantitative surface chemical mapping with Auger and backscattered electron signals

Abstract

The objective of surface analytical imaging methods is to map the quantitative distribution of the elements in the surface of a solid. Auger spectroscopy and imaging are often used for this purpose but the contrast in scanning Auger microscope (SAM) images of inhomogeneous, rough, surfaces show some features which are not due to the spatial variations of the surface chemical composition. Various schemes devised to reduce the effects on the image contrast of roughness and sub-surface inhomogeneity are reviewed in this paper. A model sample with a range of known surface inclinations and with known surface and sub-surface composition variations is used to demonstrate the effects of two general classes of correction schemes intended to reduce the contrast of such artefacts. The first class is based upon various arithmetic combinations of the number of electrons detected by a spectrometer set at the energy of an Auger peak and then at an energy (or a number of energies) on the spectral background above the peak. The extent to which these different combinations are successful in reducing topographical contrast and revealing otherwise hidden chemical contrast is summarised. It will be shown how a simple estimate of the Auger peak height leaves large topographical effects in the image contrast and that ratios of measurements at two energies provide a good first approximation for the removal of topographical artefacts. However, even those ratios which provide the best topographic compensation can overcorrect for variations in the sub-surface composition and its effects upon the Auger backscattering factor. In addition, they can introduce new problems for quantitative surface microanalysis. A second class of corrections involves the use of the signals from a set of four backscattered electron (BSE) detectors which collect BSE imaes at the same time as the energy-analysed images. The sum and difference signals from these detectors can be used, together with appropriate calibration, to compute three images in which the pixel intensities give the local surface inclinations and effective atomic numbers of the sub-surface material. These images can be used, in turn, to derive correction images for the Auger maps so as to yield a more accurate map of the surface chemical composition. The combination of the correlated Auger and BSE signals for each place on the surface requires the use of an instrument which collects these signals simultaneously from several detectors. Various methods of using such correlations are discussed. Surfaces with large variations in the contrast in the “raw” Auger images can be analysed with a precision of approximately ±5 at% after exploitation of such correlations.