ABSTRACT

We develop the methodology to predict spectral signal-to-noise ratios (SNRs) of electron energy-loss spectroscopy (EELS) by computing spectra taking account of valence electron excitations, inner-shell excitations and elastic scattering together with the noise characteristics of the detection system. The consideration of multiple scattering in our model allows us to apply it to a wide range of specimen thicknesses and collection angles. The detectability of an element is judged by the SNR. The calculated SNRs at the oxygen and magnesium K-edges in MgO agree well with experiment.