We present Keck/DEIMOS spectroscopy of stars in the recently discovered Milky Way satellites Hydra II, Pisces II, and Laevens 1. We measured a velocity dispersion of 5.4 (+3.6 -2.4) km/s for Pisces II, but we did not resolve the velocity dispersions
of Hydra II or Laevens 1. We marginally resolved the metallicity dispersions of Hydra II and Pisces II but not Laevens 1. Furthermore, Hydra II and Pisces II obey the luminosity-metallicity relation for Milky Way dwarf galaxies (<[Fe/H]> = -2.02 +/- 0.08 and -2.45 +/- 0.07, respectively), whereas Laevens 1 does not (<[Fe/H]> = -1.68 +/- 0.05). The kinematic and chemical properties suggest that Hydra II and Pisces II are dwarf galaxies, and Laevens 1 is a globular cluster. We determined that two of the previously observed blue stars near the center of Laevens 1 are not members of the cluster. A third blue star has ambiguous membership. Hydra II has a radial velocity <v_helio> = 303.1 +/- 1.4 km/s, similar to the leading arm of the Magellanic stream. The mass-to-light ratio for Pisces II is 370 (+310 -240) M_sun/L_sun. It is not among the most dark matter-dominated dwarf galaxies, but it is still worthy of inclusion in the search for gamma rays from dark matter self-annihilation.
We analyze Keck ESI spectroscopy of HVS17, a B-type star traveling with a Galactic rest frame radial velocity of +445 km/s in the outer halo of the Milky Way. HVS17 has the projected rotation of a main sequence B star and is chemically peculiar, with
solar iron abundance and sub-solar alpha abundance. Comparing measured T_eff and logg with stellar evolution tracks implies that HVS17 is a 3.91 +-0.09 Msun, 153 +-9 Myr old star at a Galactocentric distance of r=48.5 +-4.6 kpc. The time between its formation and ejection significantly exceeds 10 Myr and thus is difficult to reconcile with any Galactic disk runaway scenario involving massive stars. The observations are consistent, on the other hand, with a hypervelocity star ejection from the Galactic center. We show that Gaia proper motion measurements will easily discriminate between a disk and Galactic center origin, thus allowing us to use HVS17 as a test particle to probe the shape of the Milky Ways dark matter halo.
We report positions, velocities and metallicities of 50 ab-type RR Lyrae (RRab) stars observed in the vicinity of the Orphan stellar stream. Using about 30 RRab stars classified as being likely members of the Orphan stream, we study the metallicity a
nd the spatial extent of the stream. We find that RRab stars in the Orphan stream have a wide range of metallicities, from -1.5 dex to -2.7 dex. The average metallicity of the stream is -2.1 dex, identical to the value obtained by Newberg et al. (2010) using blue horizontal branch stars. We find that the most distant parts of the stream (40-50 kpc from the Sun) are about 0.3 dex more metal-poor than the closer parts (within ~30 kpc), suggesting a possible metallicity gradient along the streams length. We have extended the previous studies and have mapped the stream up to 55 kpc from the Sun. Even after a careful search, we did not identify any more distant RRab stars that could plausibly be members of the Orphan stream. If confirmed with other tracers, this result would indicate a detection of the end of the leading arm of the stream. We have compared the distances of Orphan stream RRab stars with the best-fit orbits obtained by Newberg et al. (2010). We find that model 6 of Newberg et al. (2010) cannot explain the distances of the most remote Orphan stream RRab stars, and conclude that the best fit to distances of Orphan stream RRab stars and to the local circular velocity is provided by potentials where the total mass of the Galaxy within 60 kpc is M_{60}~2.7x10^{11} Msun, or about 60% of the mass found by previous studies. More extensive modelling that would consider non-spherical potentials and the possibility of misalignment between the stream and the orbit, is highly encouraged.
Segue 2, discovered by Belokurov et al. (2009), is a galaxy with a luminosity of only 900 L_sun. We present Keck/DEIMOS spectroscopy of 25 members of Segue 2--a threefold increase in spectroscopic sample size. The velocity dispersion is too small to
be measured with our data. The upper limit with 90% (95%) confidence is sigma_v < 2.2 (2.6) km/s, the most stringent limit for any galaxy. The corresponding limit on the mass within the 3-D half-light radius (46 pc) is M_1/2 < 1.5 (2.1) x 10^5 M_sun. Segue 2 is the least massive galaxy known. We identify Segue 2 as a galaxy rather than a star cluster based the wide dispersion in [Fe/H] (from -2.85 to -1.33) among the member stars. The stars [alpha/Fe] ratios decline with increasing [Fe/H], indicating that Segue 2 retained Type Ia supernova ejecta despite its presently small mass and that star formation lasted for at least 100 Myr. The mean metallicity, <[Fe/H]> = -2.22 +/- 0.13 (about the same as the Ursa Minor galaxy, 330 times more luminous than Segue 2), is higher than expected from the luminosity-metallicity relation defined by more luminous dwarf galaxy satellites of the Milky Way. Segue 2 may be the barest remnant of a tidally stripped, Ursa Minor-sized galaxy. If so, it is the best example of an ultra-faint dwarf galaxy that came to be ultra-faint through tidal stripping. Alternatively, Segue 2 could have been born in a very low-mass dark matter subhalo (v_max < 10 km/s), below the atomic hydrogen cooling limit.
We obtain Keck HIRES spectroscopy of HVS5, one of the fastest unbound stars in the Milky Way halo. We show that HVS5 is a 3.62 +- 0.11 Msun main sequence B star at a distance of 50 +- 5 kpc. The difference between its age and its flight time from the
Galactic center is 105 +-18(stat)+-30(sys) Myr; flight times from locations elsewhere in the Galactic disk are similar. This 10^8 yr `arrival time between formation and ejection is difficult to reconcile with any ejection scenario involving massive stars that live for only 10^7 yr. For comparison, we derive arrival times of 10^7 yr for two unbound runaway B stars, consistent with their disk origin where ejection results from a supernova in a binary system or dynamical interactions between massive stars in a dense star cluster. For HVS5, ejection during the first 10^7 yr of its lifetime is ruled out at the 3-sigma level. Together with the 10^8 yr arrival times inferred for three other well-studied hypervelocity stars (HVSs), these results are consistent with a Galactic center origin for the HVSs. If the HVSs were indeed ejected by the central black hole, then the Galactic center was forming stars ~200 Myr ago, and the progenitors of the HVSs took ~100 Myr to enter the black holes loss cone.
Although there have been recent claims that there is a large dispersion in the abundances of the heavy neutron capture elements in the old Galactic globular cluster M92, we show that the measured dispersion for the absolute abundances of four of the
rare earth elements within a sample of 12 luminous red giants in M92 (less than or equal to 0.07 dex) does not exceed the relevant sources of uncertainty. As expected from previous studies, the heavy elements show the signature of the r-process. Their abundance ratios are essentially identical to those of M30, another nearby globular cluster of similar metallicity.
We derive the star formation histories of eight dwarf spheroidal (dSph) Milky Way satellite galaxies from their alpha element abundance patterns. Nearly 3000 stars from our previously published catalog (Paper II) comprise our data set. The average [a
lpha/Fe] ratios for all dSphs follow roughly the same path with increasing [Fe/H]. We do not observe the predicted knees in the [alpha/Fe] vs. [Fe/H] diagram, corresponding to the metallicity at which Type Ia supernovae begin to explode. Instead, we find that Type Ia supernova ejecta contribute to the abundances of all but the most metal-poor ([Fe/H] < -2.5) stars. We have also developed a chemical evolution model that tracks the star formation rate, Types II and Ia supernova explosions, and supernova feedback. Without metal enhancement in the supernova blowout, massive amounts of gas loss define the history of all dSphs except Fornax, the most luminous in our sample. All six of the best-fit model parameters correlate with dSph luminosity but not with velocity dispersion, half-light radius, or Galactocentric distance.
We present a new solution for the dispersive element in astronomical spectrographs, which in many cases can provide an upgrade path to enhance the spectral resolution of existing moderate-resolution reflection-grating spectrographs. We demonstrate th
at in the case of LRIS-R at the Keck 1 Telescope a spectral resolution of 18,000 can be achieved with reasonable throughput under good seeing conditions.
We present an abundance analysis based on high dispersion and high signal-to-noise ratio Keck spectra of a very highly microlensed Galactic bulge dwarf, OGLE-2007-BLG-349S, with Teff ~ 5400 K. The amplification at the time the spectra were taken rang
ed from 350 to 450. This bulge star is highly enhanced in metallicity with [Fe/H] = +0.51 pm 0.09 dex. The abundance ratios for the 28 species of 26 elements for which features could be detected in the spectra are solar. In particular, there is no evidence for enhancement of any of the alpha-elements including O and Mg. We conclude that the high [Fe/H] seen in this star, when combined with the equally high [Fe/H] derived in previous detailed abundance analysis of two other Galactic bulge dwarfs, both also microlensed, implies that the median metallicity in the Galactic bulge is very high. We thus infer that many previous estimates of the metallicity distribution in the Galactic bulge have substantially underestimated the mean Fe-metallicity there due to sample bias, and suggest a candidate mechanism for such. If our conjecture proves valid, it may be necessary to update the calibrations for the algorithms used by many groups to interpret spectra and broad band photometry of the integrated light of very metal-rich old stellar populations, including luminous elliptical galaxies.
