As expected we saw that the Ca II triplet lines got broader for stars with higher luminosity classes (aka lower surface gravities); however, based on the excellent suggestion of George Wallerstein and his memory of a 1939 paper, we also checked out the H-alpha lines in the spectra and saw a strong correlation with surface gravity! The correlation comes from overpopulation of the metastable 2S level in non-LTE conditions; this overpopulation is stronger at low surface gravities and leads to an evolution of the H-alpha line that is similar to Ca II but actually tracks surface gravity effects more closely. In particular it’s actually better than Ca II at distinguishing between class III red giants and class I red supergiants. Right now properly identifying “foreground AGB stars” vs. “background red supergiants” is still quite difficult (and can lead to confusion in things like age, metallicity and membership debates); the use of H-alpha going forward should be a huge help in resolving these ambiguities!
The paper will be published in ApJ and can now be downloaded from astro-ph.