摘要： Excitons and plasmonic interactions, which are effectively responsible for the transfer of energy within devices such as solar cells, LEDs and semiconductor circuits, have been understood in theory for decades. However, the photophysical behaviour within materials has always been rather difficult to understand and be directly observed. Surface structure, localised thickness variations and presence of edges are bound to influence the macroscopic properties of the materials. Understanding the local surface structure and chemistry of these materials at the nanoscale is crucial in order to reach the full potential of the materials for real-life applications. Hence, there is a need to fully characterise the physical and chemical properties from the bottom up i.e. at the level of individual atoms and to map the optoelectronic properties where they happen. Due to recent technical improvements we can now access parts of the low loss electron energy-loss (LL EEL) spectra which had previously been inaccessible. This opens up new possibilities to study nanomaterials not only at unprecedented energy but also – contrary to bulk optical techniques – with a spatial resolution at the nanoscale, as described by Zhou, Dellby [1]. Although some significant progress has been made recently in unravelling the physical origins of the LL EEL features as shown by Tizei, Lin [2], significant gaps in our understanding of the signals and their origins remain. In the study presented here, we used for the first time a combination of experimental monochromated LL STEM EEL spectroscopy and theoretical calculations using time-dependent density functional theory (TDDFT) as well as the Bethe-Salpeter equation (BSE) to study the optical properties of MoS2 at the nanoscale with the aim to understand the origins of the peaks and regional variations of the complete LL EEL spectrum. We report that we identified and resolved as well as mapped mid-bandgap excitonic signals at ~1.88eV and at ~2.08eV on MoS2 flakes using monochromated LL EELS (figure 1) and confirmed their origin by BSE calculations; we also identified and mapped several plasmonic peaks (figure 1) using LL EELS combined with TD DFT. Furthermore, we observed great spatial variations in the LL EELS signal when comparing the edge to inner regions of a flake, i.e. with increasing number of layers, and we show how these can be largely attributed to beam geometry effects. The effects of the experimental set-up on the low loss EELS signal will be discussed.