Impossible Spectroscopy: Stellar populations in the Poisson regime with MUSE NFM (and beyond)

In an IFU observation of an unresolved stellar population, some pixels by chance sample more giant stars than average, while other pixels by chance contain fewer. This trivial statement about Poisson fluctuations can be harnessed to derive otherwise unobtainable information from the data. Applied to elliptical galaxies, the method isolates the spectra of cool RGB/AGB stars and the spectral variation along the giant sequence. Thus we can probe stars with super-solar metallicity and alpha-enhanced abundances, that are among the most uncertain ingredients in stellar population models. For younger populations, O- and C-rich giant spectra can in principle be extracted independently. The narrow-field mode of MUSE is uniquely suited to this technique, simultaneously providing many thousands of near-independent samples from the same underlying population. I will show a practical demonstration using MUSE-NFM observations of a bulge field in NGC 5128 (Cen A), with the recovered variation spectrum being a good but not perfect match to our predictions from stochastic stellar population models. I will describe ongoing efforts to extend model calculations over a wider range of age and metallicity, and into the near-IR for use with new ERIS data. Ultimately we aim to apply the method to the centres of true giant ellipticals, which will never be resolvable by more traditional techniques. MUSE observations point the way towards this goal, which should be achievable with HARMONI on the E-ELT.