James Craig Ostheimer

The Metal-poor Halo of the Andromeda Spiral Galaxy (M31)

We present spectroscopic observations of red giant branch (RGB) stars over a large expanse in the halo of the Andromeda spiral galaxy (M31), acquired with the DEIMOS instrument on the Keck II 10 m telescope. Using a combination of five photometric/spectroscopic diagnostics?(1) radial velocity, (2) intermediate-width DDO51 photometry, (3) Na I equivalent width (surface gravity-sensitive), (4) position in the color-magnitude diagram, and (5) comparison between photometric and spectroscopic [Fe/H] estimates?we isolate over 250 bona fide M31 bulge and halo RGB stars located in 12 fields ranging from R = 12 to 165 kpc from the center of M31 (47 of these stars are halo members with R > 60 kpc). We derive the M31 spheroid (bulge and halo) metallicity distribution function and find it to be systematically more metal-poor with increasing radius, shifting from [Fe/H] = -0.47 ? 0.03 (? = 0.39) at R 60 kpc, assuming [?/Fe] = 0.0. These results indicate the presence of a metal-poor RGB population at large radial distances out to at least R = 160 kpc, thereby supporting our recent discovery of a stellar halo in M31 (structural component with an R-2 power-law surface brightness profile). This component has a distinct metallicity distribution from M31s bulge. If we assume an ?-enhancement of [?/Fe] = +0.3 for M31s halo, we derive [Fe/H] = -1.5 ? 0.1 (? = 0.7). Therefore, the mean metallicity and metallicity spread of this newly found remote M31 RGB population are similar to those of the Milky Way halo.

THE SPLASH SURVEY: SPECTROSCOPY OF 15 M31 DWARF SPHEROIDAL SATELLITE GALAXIES*

We present a resolved star spectroscopic survey of 15 dwarf spheroidal (dSph) satellites of the Andromeda galaxy (M31). We filter foreground contamination from Milky Way (MW) stars, noting that MW substructure is evident in this contaminant sample. We also filter M31 halo field giant stars and identify the remainder as probable dSph members. We then use these members to determine the kinematical properties of the dSphs. For the first time, we confirm that And XVIII, XXI, and XXII show kinematics consistent with bound, dark-matter-dominated galaxies. From the velocity dispersions for the full sample of dSphs we determine masses, which we combine with the size and luminosity of the galaxies to produce mass-size-luminosity scaling relations. With these scalings we determine that the M31 dSphs are fully consistent with the MW dSphs, suggesting that the well-studied MW satellite population provides a fair sample for broader conclusions. We also estimate dark matter halo masses of the satellites and find that there is no sign that the luminosity of these galaxies depends on their dark halo mass, a result consistent with what is seen for MW dwarfs. Two of the M31 dSphs (And XV, XVI) have estimated maximum circular velocities smaller than 12 km s^(–1) (to 1σ), which likely places them within the lowest-mass dark matter halos known to host stars (along with Bootes I of the MW). Finally, we use the systemic velocities of the M31 satellites to estimate the mass of the M31 halo, obtaining a virial mass consistent with previous results.

THE SPLASH SURVEY: INTERNAL KINEMATICS, CHEMICAL ABUNDANCES, AND MASSES OF THE ANDROMEDA I, II, III, VII, X, AND XIV DWARF SPHEROIDAL GALAXIES

We present new Keck/DEIMOS spectroscopic observations of hundreds of individual stars along the sightline to the first three of the Andromeda (M31) dwarf spheroidal (dSph) galaxies to be discovered, And I, II, and III, and combine them with recent spectroscopic studies by our team of three additional M31 dSphs, And VII, X, and XIV, as a part of the SPLASH Survey (Spectroscopic and Photometric Landscape of Andromedas Stellar Halo). Member stars of each dSph are isolated from foreground Milky Way dwarf stars and M31 field contamination using a variety of photometric and spectroscopic diagnostics. Our final spectroscopic sample of member stars in each dSph, for which we measure accurate radial velocities with a median uncertainty (random plus systematic errors) of 4-5 km s^(–1), includes 80 red giants in And I, 95 in And II, 43 in And III, 18 in And VII, 22 in And X, and 38 in And XIV. The sample of confirmed members in the six dSphs is used to derive each systems mean radial velocity, intrinsic central velocity dispersion, mean abundance, abundance spread, and dynamical mass. This combined data set presents us with a unique opportunity to perform the first systematic comparison of the global properties (e.g., metallicities, sizes, and dark matter masses) of one-third of Andromedas total known dSph population with Milky Way counterparts of the same luminosity. Our overall comparisons indicate that the family of dSphs in these two hosts have both similarities and differences. For example, we find that the luminosity-metallicity relation is very similar between L ~ 10^5 and 10^7 L_☉, suggesting that the chemical evolution histories of each group of dSphs are similar. The lowest luminosity M31 dSphs appear to deviate from the relation, possibly suggesting tidal stripping. Previous observations have noted that the sizes of M31s brightest dSphs are systematically larger than Milky Way satellites of similar luminosity. At lower luminosities between L = 10^4 and 10^6 L_☉, we find that the sizes of dSphs in the two hosts significantly overlap and that four of the faintest M31 dSphs are smaller than Milky Way counterparts. The first dynamical mass measurements of six M31 dSphs over a large range in luminosity indicate similar mass-to-light ratios compared to Milky Way dSphs among the brighter satellites, and smaller mass-to-light ratios among the fainter satellites. Combined with their similar or larger sizes at these luminosities, these results hint that the M31 dSphs are systematically less dense than Milky Way dSphs. The implications of these similarities and differences for general understanding of galaxy formation and evolution are summarized.

Exploring Halo Substructure with Giant Stars: I. Survey Description and Calibration of the Photometric Search Technique

We have begun a survey of the structure of the Milky Way halo, as well as the halos of other Local Group galaxies, as traced by their constituent giant stars. These giant stars are identified via large-area, CCD photometric campaigns. Here we present the basis for our photometric search method, which relies on the gravity sensitivity of the Mg I triplet+MgH features near 5150 ? in F?K stars, and which is sensed by the flux in the intermediate-band DDO51 filter. Our technique is a simplified variant of the combined Washington/DDO51 four-filter technique described by Geisler, which we modify for the specific purpose of efficiently identifying distant giant stars for follow-up spectroscopic study: We show here that for most stars the Washington T1-T2 color is correlated monotonically with the Washington M-T2 color with relatively low scatter; for the purposes of our survey, this correlation obviates the need to image in the T1 filter, as originally proposed by Geisler. To calibrate our (M-T2, M-DDO51) diagram as a means to discriminate field giant stars from nearby dwarfs, we utilize new photometry of the main sequences of the open clusters NGC 3680 and NGC 2477 and the red giant branches of the clusters NGC 3680, Melotte 66, and ? Centauri, supplemented with data on field stars, globular clusters and open clusters by Doug Geisler and collaborators. By combining the data on stars from different clusters, and by taking advantage of the wide abundance spread within ? Centauri, we verify the primary dependence of the M-DDO51 color on luminosity and demonstrate the secondary sensitivity to metallicity among giant stars. Our empirical results are found to be generally consistent with those from analysis of synthetic spectra by Paltoglou & Bell. Finally, we provide conversion formulae from the (M, M-T2) system to the (V, V-I) system, corresponding reddening laws, as well as empirical red giant branch curves from ? Centauri stars for use in deriving photometric parallaxes for giant stars of various metallicities (but equivalent ages) to those of ? Centauri giants.

THE SPLASH SURVEY: INTERNAL KINEMATICS, CHEMICAL ABUNDANCES, AND MASSES OF THE ANDROMEDA I, II, III, VII, X, AND XIV DWARF SPHEROIDAL GALAXIES {sup ,}

We present new Keck/DEIMOS spectroscopic observations of hundreds of individual stars along the sightline to the first three of the Andromeda (M31) dwarf spheroidal (dSph) galaxies to be discovered, And I, II, and III, and combine them with recent spectroscopic studies by our team of three additional M31 dSphs, And VII, X, and XIV, as a part of the SPLASH Survey (Spectroscopic and Photometric Landscape of Andromedas Stellar Halo). Member stars of each dSph are isolated from foreground Milky Way dwarf stars and M31 field contamination using a variety of photometric and spectroscopic diagnostics. Our final spectroscopic sample of member stars in each dSph, for which we measure accurate radial velocities with a median uncertainty (random plus systematic errors) of 4-5 km s^(–1), includes 80 red giants in And I, 95 in And II, 43 in And III, 18 in And VII, 22 in And X, and 38 in And XIV. The sample of confirmed members in the six dSphs is used to derive each systems mean radial velocity, intrinsic central velocity dispersion, mean abundance, abundance spread, and dynamical mass. This combined data set presents us with a unique opportunity to perform the first systematic comparison of the global properties (e.g., metallicities, sizes, and dark matter masses) of one-third of Andromedas total known dSph population with Milky Way counterparts of the same luminosity. Our overall comparisons indicate that the family of dSphs in these two hosts have both similarities and differences. For example, we find that the luminosity-metallicity relation is very similar between L ~ 10^5 and 10^7 L_☉, suggesting that the chemical evolution histories of each group of dSphs are similar. The lowest luminosity M31 dSphs appear to deviate from the relation, possibly suggesting tidal stripping. Previous observations have noted that the sizes of M31s brightest dSphs are systematically larger than Milky Way satellites of similar luminosity. At lower luminosities between L = 10^4 and 10^6 L_☉, we find that the sizes of dSphs in the two hosts significantly overlap and that four of the faintest M31 dSphs are smaller than Milky Way counterparts. The first dynamical mass measurements of six M31 dSphs over a large range in luminosity indicate similar mass-to-light ratios compared to Milky Way dSphs among the brighter satellites, and smaller mass-to-light ratios among the fainter satellites. Combined with their similar or larger sizes at these luminosities, these results hint that the M31 dSphs are systematically less dense than Milky Way dSphs. The implications of these similarities and differences for general understanding of galaxy formation and evolution are summarized.

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 M31s 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 λ8190 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 λ8500 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 A and the TiO bands at 7100, 7600, and 8500 A. 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 galaxys center, including three red giants at R ~ 165 kpc. The ability to identify individual M31 red giant stars gives us an unprecedented level of sensitivity in studying the properties of the galaxys outer halo.

Detection of the main-sequence turnoff of a newly discovered milky way halo structure in the triangulum-andromeda region

An upper main sequence (MS) and main-sequence turnoff (MSTO) feature appears in the color-magnitude diagram (CMD) of a large-area photometric survey of the southern half of M31 stretching to M33. Imaging in the Washington M, T2, DDO51 system allows us to remove the background M31/M33 giants from our CMD and more clearly see the dwarf star feature, which has an MSTO near M ~ 20.5. The corresponding stellar population shows little density variation over the 12° × 6° area of the sky sampled and is of very low surface brightness, Σ > 32 mag arcsec-2. We show that this feature is not the same as a previously identified MS+MSTO in the foreground of the Andromeda galaxy that has been associated with the tidal stream ringing the Milky Way disk at less than half the distance. Thus, the new stellar system is a separate, more distant entity, perhaps a segment of tidal debris from a disrupted satellite galaxy. It is most likely related to the structure with similar distance, location, and density uniformity seen as an excess of K and M giants in the Two Micron All Sky Survey reported in the companion paper by Rocha-Pinto and coworkers.

Global Properties of M31's Stellar Halo from the SPLASH Survey. I. Surface Brightness Profile

We present the surface brightness profile of M31s stellar halo out to a projected radius of 175 kpc. The surface brightness estimates are based on confirmed samples of M31 red giant branch stars derived from Keck/DEIMOS spectroscopic observations. A set of empirical spectroscopic and photometric M31 membership diagnostics is used to identify and reject foreground and background contaminants. This enables us to trace the stellar halo of M31 to larger projected distances and fainter surface brightnesses than previous photometric studies. The surface brightness profile of M31s halo follows a power law with index –2.2 ± 0.2 and extends to a projected distance of at least ~175 kpc (~2/3 of M31s virial radius), with no evidence of a downward break at large radii. The best-fit elliptical isophotes have b/a = 0.94 with the major axis of the halo aligned along the minor axis of M31s disk, consistent with a prolate halo, although the data are also consistent with M31s halo having spherical symmetry. The fact that tidal debris features are kinematically cold is used to identify substructure in the spectroscopic fields out to projected radii of 90 kpc and investigate the effect of this substructure on the surface brightness profile. The scatter in the surface brightness profile is reduced when kinematically identified tidal debris features in M31 are statistically subtracted; the remaining profile indicates that a comparatively diffuse stellar component to M31s stellar halo exists to large distances. Beyond 90 kpc, kinematically cold tidal debris features cannot be identified due to small number statistics; nevertheless, the significant field-to-field variation in surface brightness beyond 90 kpc suggests that the outermost region of M31s halo is also comprised to a significant degree of stars stripped from accreted objects.

DYNAMICS AND STELLAR CONTENT OF THE GIANT SOUTHERN STREAM IN M31. I. KECK SPECTROSCOPY OF RED GIANT STARS

This paper presents the first results from a large spectroscopic survey of red giant branch (RGB) stars in M31 using DEIMOS on the Keck 10 m telescope. A photometric prescreening method, based on the intermediate-width DDO51 band centered on the Mg b/MgH absorption feature, was used to select spectroscopic targets. RGB candidates were targeted in a small section of M31s giant southern tidal stream at a projected distance of 31 kpc from the galaxys center. We isolate a clean sample of 68 RGB stars by removing contaminants (foreground Milky Way dwarf stars and background galaxies) using a combination of spectroscopic, imaging, and photometric methods: radial velocity and the surface gravity-sensitive Na I doublet are particularly useful in this regard. About 65% of the M31 stars are found to be members of the giant southern stream, while the rest appear to be members of the general spheroid population. The mean (heliocentric) radial velocity of the stream in our field is -458 km s-1, blueshifted by -158 km s-1 relative to M31s systemic velocity, in good agreement with recent velocity measurements at other points along the stream. The intrinsic velocity dispersion of the stream is found to be 15 km s-1 (90% confidence limit). A companion paper by Font and coworkers discusses possible orbits, implications of the coldness of the stream, and properties of the progenitor satellite galaxy. The kinematics, and possibly the metallicity distribution, of the general spheroid (i.e., nonstream) population in this region of M31 indicate that it is significantly different from samples drawn from other parts of the M31 spheroid; this is probably an indication of substructure in the bulge and halo. The stream appears to have a higher mean metallicity than the general spheroid, [Fe/H] ~ -0.54 versus -0.74, and a smaller metallicity spread. The relatively high metallicity of the stream implies that its progenitor must have been a luminous dwarf galaxy. The Ca II triplet line strengths of the M31 RGB stars are generally consistent with photometric estimates of their metallicity (derived by fitting RGB fiducials in the color-magnitude diagram). There is indirect evidence of a population of intermediate-age stars in the stream.

Kinematic and chemical constraints on the formation of M31's inner and outer halo

The halo of M31 shows a wealth of substructures, some of which are consistent with assembly from satellite accretion. Here we report on kinematic and abundance results from Keck DEIMOS spectroscopy in the near-infrared calcium triplet region of over 3500 red giant star candidates along the minor axis and in off-axis spheroid fields of M31. These data reach out to large radial distances of about 160 kpc. The derived radial velocity distributions show an indication of a kinematically cold substructure around ~17 kpc, which has been reported before. We devise a new and improved method to measure spectroscopic metallicities from the calcium triplet in low signal-to-noise ratio spectra using a weighted co-addition of the individual lines. The resulting distribution (accurate to ~0.3 dex down to signal-to-noise ratios of 5) leads us to note an even stronger gradient in the abundance distribution along M31s minor axis and in particular toward the outer halo fields than previously detected. The mean metallicity in the outer fields reaches below –2 dex, with individual values as low as –2.6 dex. This is the first time such a metal-poor halo has been detected in M31. In the fields toward the inner spheroid, we find a sharp decline of ~0.5 dex in metallicity in a region at ~20 kpc, which roughly coincides with the edge of an extended disk, previously detected from star count maps. A large fraction of red giants that we detect in the most distant fields are likely members of M33s overlapping halo. A comparison of our velocities with those predicted by new N-body simulations argues that the event responsible for the Giant Stream is most likely not responsible for the full population of the inner halo. We show further that the abundance distribution of the Stream is different from that of the inner halo, from which it becomes evident, in turn, that the merger event that formed the Stream and the outer halo cannot have contributed any significant material to the inner spheroid. All these severe structure changes in the halo suggest a high degree of infall and stochastic abundance accretion governing the buildup of M31s inner and outer halo.