Optimum experimental conditions for quantitative surface microanalysis by reflection electron energy-loss spectroscopy

Experimental conditions for obtaining high quality core-shell ionization edges in reflec- tion electron energy-loss spectroscopy (REELS) are investigated.Under the (600) specular-"mirror" reflection conditions and using the relative ionization cross-section measured from a MgO thin foil in the transmission geometry for collection semi-angle 03B2 = 1.2mrad, the chemical composition of MgO (100) surfaces is determined to be NO/NMg = 1.5 ± 0.15.This value is not significantly affected by varying the resonance diffraction conditions near the [001] zone axis, under which the spectra were acquired.An incorrect apparent composition will result if channeling effects along the [011] zone axis are not considered properly.Surface microanalysis is limited by the accuracy of the core-shell effective ionization cross-section (EICS), which depends not only on the property of a single atom but also on the dynamical elastic and inelastic scattering and channeling processes of electrons.An experimental method is outlined by which to measure the relative EICS from a thin foil specimen in the transmission case under the equivalent resonance conditions as in reflection geometry.1. Introduction.Reflection electron energy-loss spectroscopy (REELS) combines the techniques of reflection elec- tron microscopy (REM) and electron energy-loss spectroscopy (EELS) in a transmission electron microscope (TEM) [1].The REELS spectra are acquired from reflected electrons under sur- face resonance conditions (see [2] for a review), the electrons having travelled a certain distance along the surface before being reflected [3].Surface compositional microanalysis, an important application of REELS, usually requires the simultaneous detection of two or more atomic inner- shell ionization edges.As a result of strong dynamical scattering effects, however, the signal-tobackground (S/B) ratios of the K ionization edges located above 1 keV are limited in the REM geometry, which may compromise the accuracy of surface microanalysis.In REM, the electrons reflected from the surface can be classified as Bragg-but not resonance, resonance-but not Bragg, and Bragg-resonance.The optimum inelastic signal usually obtained in the last case.In practice, there are many ways to achieve surface resonance, such as axial and