Monitoring strong coupling in nonlocal plasmonics with electron spectroscopies

Plasmon-exciton polaritons provide exciting possibilities to control light-matter interactions at the nanoscale by enabling closer investigation of quantum optical effects and facilitating novel technologies based, for instance, on Bose-Einstein condensation and polaritonic lasing. Nevertheless, observing and visualizing polaritons is challenging, and traditional optical microscopy techniques often lead to ambiguities regarding the emergence and strength of the plasmon-exciton coupling. Electron microscopy offers a more robust means to study and verify the nature of plexcitons, but it is still hindered by instrument limitations and resolution. A simple theoretical description of electron beam-excited plexcitons is therefore vital to complement ongoing experimental efforts. Here we apply analytic solutions for the electron-loss and photon-emission probabilities to evaluate plasmonexciton coupling studied either with the recently adopted technique of electron energy-loss spectroscopy, or with the so-far unexplored in this context cathodoluminescence spectroscopy. Foreseeing the necessity to account for quantum corrections in the plasmonic response, we extend these solutions within the framework of general nonlocal hydrodynamic descriptions. As a specific example, we study core-shell spherical plasmon-molecule hybrids, going beyond the standard local-response approximation through the hydrodynamic Drude model for screening and the generalized nonlocal optical response theory for nonlocal damping. We show that electron microscopies are extremely powerful in describing the interaction of emitters with the otherwise weakly excited by optical means higher-order plasmonic multipoles, a response that survives when quantum-informed models are considered. Our work provides, therefore, both a robust theoretical background and supporting argumentation to the open quest for improving and further utilizing electron microscopies in strong-coupling nanophotonics.