On the quantitativeness of EDS STEM

Chemical mapping using energy dispersive X-ray spectroscopy (EDS) in scanning transmission electron microscopy (STEM) has recently shown to be a powerful technique in analyzing the elemental identity and location of atomic columns in materials at atomic resolution. However, most applications of EDS STEM have been used only to qualitatively map whether elements are present at specific sites. Obtaining calibrated EDS STEM maps so that they are on an absolute scale is a difficult task and even if one achieves this, extracting quantitative information about the specimen - such as the number or density of atoms under the probe - adds yet another layer of complexity to the analysis due to the multiple elastic and inelastic scattering of the electron probe. Quantitative information may be obtained by comparing calibrated EDS STEM with theoretical simulations, but in this case a model of the structure must be assumed a priori. Here we first theoretically explore how exactly elastic and thermal scattering of the probe confounds the quantitative information one is able to extract about the specimen from an EDS STEM map. We then show using simulation how tilting the specimen (or incident probe) can reduce the effects of scattering and how it can provide quantitative information about the specimen. We then discuss drawbacks of this method - such as the loss of atomic resolution along the tilt direction - but follow this with a possible remedy: precession averaged EDS STEM mapping.