Theory of Pines' demon in multiband metals

In 1956, David Pines predicted the existence of an acoustic plasmon, dubbed a "demon", in metals with multiple and sufficiently distinct charge carrier species. Despite extensive searches, demons have not been observed directly until recent momentum-resolved electron energy-loss spectroscopy (M-EELS) experiments on Sr₂RuO₄. Here, we discuss the necessary and sufficient conditions for the existence of demons as a stable collective excitation in multiband metals. We derive expressions for the dispersion, intensity, and damping rate of the demon in terms of band structure parameters, thereby going beyond the heuristic assumptions in Pines' original arguments. Furthermore, we comment on the suitability of various experimental probes, including M-EELS, optical spectroscopy, and X-ray spectroscopy, for detecting demons, and we make predictions for other materials where demons may be present and observable in experiment.