A new method for isolating M31 red giant stars: The discovery of stars out to a radial distance of 165 kpc

We present a method for isolating a clean sample of red giant branch stars in the outer regions of M31. Our study is based on an ongoing spectroscopic survey using the DEIMOS instrument on the Keck II 10 m telescope. The survey aims to study the kinematics, (sub) structure, and metallicity of M31's halo. Although most of our spectroscopic targets were photometrically screened to reject foreground Milky Way dwarf star contaminants, dwarf stars still constitute a substantial fraction of the observed spectra in the sparse outer halo. Our likelihood-based method for isolating M31 red giants uses five criteria: (1) radial velocity, (2) photometry in the intermediate-width DDO51 band to measure the strength of the MgH/Mg b absorption features, (3) strength of the Na I lambda 8190 absorption line doublet, (4) location within an (I, V-I) color-magnitude diagram, and (5) comparison of photometric (color-magnitude diagram based) versus spectroscopic (Ca II lambda 8500 triplet based) metallicity estimates. We also discuss other potential giant/ dwarf separation criteria: the strength of the K I absorption lines at 7665 and 7699 angstrom and the TiO bands at 7100, 7600, and 8500 angstrom. Training sets consisting of definite M31 red giants and Galactic dwarf stars are used to derive empirical probability distribution functions for each diagnostic. These functions are used to calculate the likelihood that a given star is a red giant in M31 versus a Milky Way dwarf star. Using our diagnostic method, we isolate 40 M31 red giants beyond a projected distance of R = 60 kpc from the galaxy's center, including three red giants at R similar to 165 kpc. The ability to identify individual M31 red giant stars gives us an unprecedented level of sensitivity in studying the properties of the galaxy's outer halo.