Searches for the Most MetalPoor Candidates from SDSS and SEGUE

1
Searches for the Most Metal-Poor Candidates from
SDSS and SEGUE

• Timothy C. Beers
• Department of Physics Astronomy
• Michigan State University
• JINA Joint Institute for Nuclear Astrophysics

SDSS
2
Direct Collaborators

• Daniela Carollo (INAF, Italy /JINA)
• Young Sun Lee (MSU / JINA)
• Sivarani Thirupathi (MSU / JINA)
• Brian Marsteller (MSU / JINA)
• John Norris (ANU / SEGUE)
• Masashi Chiba (Tohuku University, Japan)
• Carlos Allende Prieto (Univ. of Texas)
• Constance Rockosi (UCSC)
• Brian Yanny (FNAL)
• Heidi Newberg (RPI)
• Jeffrey Munn (USNO)

3
Why the Fascination with Large Numbers of MP
Stars ?

• Extremely MP stars have recorded the heavy
  element abundances produced in the first
  generations of stars
• The shape of the low-metallicity tail of the
  Metallicity Distribution Function (MDF) will
  (eventually) show structure that reveals the
  characteristic abundances of major epochs of star
  formation in early Galaxy
• Change in the nature of the MDF as a function of
  distance may reveal the assembly history of the
  MW
• Determination of the frequency of various
  elemental abundance signatures, e.g., enhancement
  of C/Fe, alpha/Fe, etc.
• Identification of relatively rare objects amongst
  MP stars, e.g., r-process / s-process enhanced
  stars

4
New Efforts for Finding Very Metal-Poor Stars

• Stellar observations at medium-resolution have
  been obtained during the course of the Sloan
  Digital Sky Survey (SDSS)
• Calibration of spectrophotometry / telluric bands
• Directed studies (e.g., BHB stars, C-rich stars)
• Failed QSO targets
• New stellar observations being obtained during
  the course of SDSS extension program SEGUE
• See http//www.sdss.org/dr6/start/aboutsegue.html

5
SEGUE The Sloan Extension for Galactic
Understanding and Exploration

• Use existing SDSS hardware and software to
  obtain
• 3500 square degrees of additional ugriz imaging
  at lower Galactic latitudes
• Stripes chosen to complement existing areal
  coverage includes several vertical stripes
  through Galactic plane
• Medium-resolution spectroscopy of 250,000
  optimally selected stars in the thick disk and
  halo of the Galaxy
• 200 spectroscopic plate pairs of 45 / 135 min
  exposures
• Objects selected to populate distances from 1 to
  100 kpc along each line of site
• Proper motions available (from SDSS) for stars
  within 5 kpc

6
SEGUE uses stellar probes of increasing absolute
brightness to probe increasing distances in the
disk, thick disk and Milky Way halo.
K III
d lt 100 kpc
BHB/BS
d lt 50 kpc
Streams and outer halo stars
MSTO/F
d lt 15 kpc
G
thin, thick disk stars
d lt 6 kpc
Inner and outer halo stars
KV
d lt 1 kpc
r 1.5kpc
Other spectroscopic surveys will not probe as
deep, for instance, Blue Horizontal Branch Stars
(BHBs) from a survey with Vlt 12 are from a
volume within 1.5 kpc of the sun.
8 kpc
7
Likely Numbers of Detected MP Stars from SEGUE

• Actual numbers will depend on the shape of the
  halo Metallicity Distribution Function
• Fe/H lt -2.0 20,000 (VMP)
• Fe/H lt -3.0 2,000 (EMP)
• Fe/H lt -4.0 200 ? (UMP)
• Fe/H lt -5.0 20 ? (HMP)
• Fe/H lt -6.0 2 ? (MMP)

8
SEGUE observing plan and status as of July 2007
SDSS Imaging scan
Declination -20 degrees
Planned SEGUE grid pointings (200)
Planned SEGUE scan (3500 sq deg)
Planned targeted SEGUE pointings(60)
Sgr stream planned scan
Completed SEGUE plate pointing
Completed SEGUE imaging
9
SSPP - Methodology I
10
SSPP - Methodology II
http//segue.uchicago.edu/documents/sspp0627.pdf
11
Validation with Globular Clusters
12
High-Res Observations To Date
Allende Prieto et al. (in preparation)
13
(No Transcript)
14
Sample SDSS-I Spectra with Fe/H -2.0
15
Sample SDSS-I Spectra with Fe/H -2.5
16
Sample SDSS-I Spectra with Fe/H lt -3.0
17
The Low-Metallicity Tail of the Metallicity
Distribution Function of SDSS-I Stars
N 4225 S/N gt 10/1
18
The Low-Metallicity Tail of the Metallicity
Distribution Function of SEGUE Stars
N 2414 S/N gt 10/1
19
Nature of the Galactic Halo(s) Conclusions First

• The structural components of the stellar
  populations in the Galaxy have been known for (at
  least) several decades
• Bulge / Thin Disk / Thick Disk (MWTD) / Halo
• New results from SDSS have now revised this list
• Halo ? Halos
• Inner Halo Dominant at R lt 10-15 kpc
• Highly eccentric
  (slightly prograde) orbits
• Metallicity peak at
  Fe/H -1.6
• Likely associated with
  major/major collision
  of massive components early in galactic
  history
• Outer Halo Dominant at R gt 15-20 kpc
• Uniform distribution of
  eccentricity
  (including highly retrograde) orbits
• Metallicity peak around
  Fe/H -2.2
• Likely associated with accretion from
  dwarf-like galaxies over an extended
  period, up to present

20
See Carollo et al. 2007 (astro-ph/0706.3005)

• The Dichotomy of the Galactic Halo of the Milky
  Way
• Daniela Carollo, Timothy C. Beers, Young Sun Lee,
  Masashi Chiba, John E. Norris , Ronald Wilhelm,
  Thirupathi Sivarani, Brian Marsteller, Jeffrey A.
  Munn, Coryn A. L. Bailer-Jones, Paola Re
  Fiorentin, Donald G. York

21
Source of Prima Facie Evidence for the Nature of
the Halo

• Based on analysis of spectroscopy, photometry,
  and astrometry (proper motions) for SDSS-I DR-5
  spectrophotometric and telluric calibration stars
  Initial sample 25000 stars
• Primarily F and early type G-type turn off stars
  in the thick-disk and halo(s) population of the
  Galaxy, especially useful for spectrophotometric
  calibration effort
• Color-based selection, with emphasis on bluer
  stars, ensures an adequate number of very
  low-metallicity (Fe/H lt -2.0) stars
• Apparent magnitude selection, with emphasis on
  brighter stars, ensures many are in the local
  volume (d lt 4 kpc 7 lt R lt 10 kpc), and hence
  have meaningful proper motions Reduced sample
  11000 stars
• Accurate estimates of radial velocity,
  metallicity, temperature, surface gravity, and
  distance are obtained for all these stars by the
  SDSS/SEGUE Stellar Parameter Pipeline (SSPP)

22
Spatial Distribution of Stars
23
Galactic Velocity Components

• Proper motions obtained from the re-calibrated
  USNO-B Catalog, typical accuracy 3-4 mas/yr (Munn
  et al. 2004)
• Used in combination with the measured radial
  velocities and estimated distances from the SSPP
  to derive the full space motion components (U, V,
  W) relative to the local standard of rest

24
UVW vs. Fe/H

• Note the clear presence of low velocity
  dispersion components in all three plots
• This is the thick-disk and metal-weak thick-disk
  population
• The high velocity dispersion component is the
  highly spatially overlapped inner-and outer-halo
  component

25
Fe/H vs. V Component
26
MDF for Retrograde Stars
27
Fe/H vs. Eccentricity / The History
ELS 1962
Fe/H -1.5
Fe/H 0
28
The orbital parameters were evaluated adopting a
Galactic potential of the Stackel form (Chiba
Beers 2000 for details). For the first time we
can clearly distinguish the presence of the thick
disk (and MWTD) population as a separate entity
from the local halo(s) stars.
29
Fe/H vs. Orbital Eccentricity
30
Nature of the Orbits
31
Summary of Kinematics

• . stars at z_max gt 5 kpc
• FW analysis for the same data as above
• Fe/Hmax min N V_rot error
• -1.00 -2.00 2041 -43. 7.
• -2.00 -5.00 549 -46. 14.
• . stars at z_max gt 10 kpc
• FW analysis for the same data as above
• Fe/Hmax min N V_rot error
• -1.00 -2.00 554 -75. 16.
• -2.00 -5.00 231 -61. 27.

32
Flattened Inside / Spherical Outside Inversion
from Kinematics to Density Prediction

• By making simplifying assumptions about nature of
  galactic potential, e.g., that the Jeans theorem
  applies
• One can invert motions to recover the underlying
  density field armchair cartography
• May Binney (1986)
• Sommer-Larsen Zhen (1990)
• Chiba Beers (2000)
• Note progression from flattened to spherical with
  decreasing metallicity

33
Origin of the Outer Halo ? (Belokurov et al.
2006)
34
Summary / Whats Next ?

• Clear demonstration that the halo is a misnomer
  there are two halos, an inner and an outer,
  with
• Different orbital distributions
• Different spatial distributions
• Different metallicity distributions
• One can now target outer halo stars in order to
  elucidate their chemical histories (a/Fe,
  C/Fe), and possibly their accretion histories
• One can now preferentially SELECT outer halo
  stars based on proper motion cuts in the local
  volume (SEGUE-II)
• One can now take advantage of the lower Fe/H,
  in general, of outer halo stars to find the most
  metal-poor stars (all three stars with Fe/H lt
  -4.5 have properties consistent with outer halo
  membership)
• One can soon constrain models for formation /
  evolution of the Galaxy that take all of the
  chemical and kinematic information into account
  (e.g., Tumlinson 2006)

35
Towards a Virtual Galaxy Courtesy J. Tumlinson
Work is beginning now to couple stochastic
chemical evolution to dark matter dynamics within
N-body simulations (Gadget2), to calculate
realistic observables in the full 6D
position/velocity and 20D chemical spaces of
modern surveys then full hydro.