Metallicity and Kinematics of M31's Outer Stellar Halo from a Keck Spectrocscopic Survey

We present multislit CaII near-infrared triplet spectroscopy with the Keck 10-m telescope/LRIS of 99 M31 halo red giant candidates in the apparent mag range 20<I<22 in a SE minor axis field at R=19kpc. The spectra are used to isolate M31 halo red giants from foreground Milky Way dwarf stars, background compact field galaxies and M31 disk giants. The observed radial velocity distribution shows an equal mix of M31 halo giants, fit by a Gaussian of width 150km/s centered on -300km/s (M31 systemic velocity), and foreground Galactic dwarfs with v<~0km/s. A secure sample of 29 M31 red giant stars is identified on the basis of radial velocity (and, in 4 cases, B-I color). For this sample, there is rough agreement between photometric and spectroscopic [Fe/H] estimates. The [Fe/H] distribution of M31 halo giants has an rms >0.6 dex, spans the full 2dex range over which [Fe/H] measurement methods are calibrated, and has a mean/median value around <[Fe/H]>=-1.9 to -1.1dex and possibly higher since the high-[Fe/H] end of the distribution is not well constrained by our sample. Our data on M31 field halo giants suggest that they are more metal-rich than their Milky Way counterparts, similar to M31 and Milky Way globular clusters, consistent with a scenario in which the halo is built from accretion of small stellar subsystems. There are 4 stars in the secure M31 sample with unusually strong Ca lines (>~solar [Fe/H]) at a common velocity of -340km/s close to M31's systemic velocity, similar to what might be expected for M31 disk giants on the minor axis. The M31 disk would have to be large and/or warped to account for these stars. More likely, these stars are metal-rich debris from a past accretion event in M31's halo.