Quantifying the low-energy limit and spectral resolution in valence electron energy loss spectroscopy

While the development of monochromators for scanning transmission electron microscopes (STEM) has improved our ability to resolve spectral features in the 0–5 eV energy range of the electron energy loss spectrum, the overall benefits relative to unfiltered microscopes have been difficult to quantify. Simple curve fitting and reciprocal space models that extrapolate the expected behavior of the zero-loss peak are not enough to fully exploit the optimal spectral limit and can hinder the ease of interpreting the resulting spectra due to processing-induced artifacts. To address this issue, here we present a quantitative comparison of two processing methods for performing ZLP removal and for defining the low-energy spectral limit applied to three microscopes with different intrinsic emission and energy resolutions. Applying the processing techniques to spectroscopic data obtained from each instrument leads in each case to a marked improvement in the spectroscopic limit, regardless of the technique implemented or the microscope setup. The example application chosen to benchmark these processing techniques is the energy limit obtained from a silicon wedge sample as a function of thickness. Based on these results, we conclude on the possibility to resolve statistically significant spectral features to within a hundred meV of the native instrumental energy spread, opening up the future prospect of tracking phonon peaks as new and improved hardware becomes available.

► We examine and extrapolate spectral limits in EELS.
► We compare deconvolution and subtraction techniques.
► We probe the effect of sample thickness on the spectral limit and distributions.