HiRes Followup of Rave Stars: Stellar Parameters, Metallicities, etc'

1
Hi-Res Follow-up of Rave StarsStellar
Parameters, Metallicities, etc.
RAVE Project Meeting April 8, 2006

• Jon Fulbright
• Keck Fellow
• Johns Hopkins Univ.
• Greg Ruchti, Rosie Wyse
• Johns Hopkins Univ.
• Based on RAVE data and observations at the Apache
  Point 3.5-m telescope
• Partially funded by the Keck Foundation.

2MASS
2
Outline

• High-resolution spectra of RAVE stars with APO.
• Comparison of stellar parameters from RAVE
  second-year data (v04) vs. those derived from the
  high-resolution spectra.
• Quick-and-dirty correction for Teff m/H in
  v04.
• Three examples of chemical abundance science that
  could/can be done using RAVE spectra/database
• Al/Fe and Chemical Fingerprinting of stellar
  pops.
• Abundances of high-velocity/outer halo stars.
• A C-rich, very metal-poor RAVE star (found by
  accident)!

3
Observational Data

• Obtained 30 stars with the echelle spectograph
  on the Apache Point Obs. (APO) 3.5-m telescope in
  Nov. 2005.
• The echelle data have R 35,000 and cover nearly
  the entire visible wavelength region (380010000
  Å).
• For most stars, S/N 60120/pixel.
• Some of the targets were Greg Ruchtis
  high-velocity stars and the observations were
  part of a longer-term project to check for RV
  variabilityshorter exposures, but the S/N will
  be built up.

4
Analysis of Hi-Res Spectra

• Use gt100 Fe I / Fe II lines to determine
  parameters.
• Set Teff from excitation plot.
• Set log(g) from ionization equillibrum n(FeI)
  n(FeII)
• Set m/H to match Fe/H.
• Set vt from log(RW) plot.
• Also compare to photometric Teff values
• Use Alonso et al. (1996 1999) Teff vs. J-K
  relations.
• (J-K)TCS 0.949(J-K)2MASS 0.004
• Teff 0.5816 0.9134(J-K) 0.1443(J-K)2 (for
  giants)
• Teff (5040 K) / Teff sT(A99 calibration)
  125 K
• d(T)/d(J-K) (50 K)/0.01 magnot a great
  indicator.

¼
5
Analysis
Line lists from Fulbright (2000) for metal-poor
stars, Fulbright et al. (2006) for metal-rich
starsboth roughly consistent. Common to have gt
100 Fe I lines for each star, 5 to 10 Fe II
lines. Example at left T6478-00245-1 RAVE Teff
5750 K, logg 4.0, m/H 0.0 APO Teff
5573 K, logg 3.7, m/H 0.07 0.09 sT
50-100 K, slogg 0.2 dex sm/H 0.10.2, svt
0.1 km/s
6
Overview of Parameter Comparison

• High resolution (R gt 30,000) spectra with broad
  wavelength coverage give all the same information
  as the RAVE spectra RVs, m/H, parameters,
  etc.
• Stellar parameters Teff, log(g), m/H, vt.
• The methods used in the analysis are independent
  of what RAVE is doing, so it is the ultimate
  check of RAVE data products.
• Spectra contain abundance data on over 20
  elements.
• Downsides are 1) The spectra are time-consuming
  to obtain (10 min to hours per star) and 2) The
  analysis is still time-consuming (1 hour per
  star just to get the stellar parameters when
  things work well).

7
RAVE Parameters
The selected stars cover a range of stellar
parameters. All have Corr. gt 15, 2nd year
data. Sample isochrone shown on figure (Padova
group, 11 Gyr, Z 0.001) only to guide eye. RAVE
m/H cover 2.0 to 0.5, most either 0.5 or
0.0. One star has high rotational velocity.
Another star is a BHB or sdB star (Teff 8750
K). Both excluded from further analysis.
8
RAVE Parameters
Lines connect RAVE and APO points for same
star. Note many stars move along the giant branch
from TO to around the RGB clump. Metal-rich TO
stars can have line strengths line metal-poor
bright RGB starsdegeneracy in that part of
parameter space.
9
T(RAVE) vs. T(J-K) vs. T(APO)
For the 29 echelle stars ltTRAVE TEXgt 342
739 K ltTRAVE TJ-Kgt 424 872 K ltTJ-K
TRAVEgt 81 356 K For the 33941 good RAVE
v04 data points ltTRAVE TJ-Kgt 171 742 K
10
T(RAVE) vs. T(J-K)
RPM(K) Diagram of 2nd Year data (Corr. gt 15,
s(HK) lt 0.5) shows that these stars are mainly
TO stars and RGB clump stars.
Thin Disk Thick Disk Halo
½
11
Changes in Other Parameters
The correlated nature of these offsets is likely
a sign the degeneracy in stellar parameter space
over line strength.
12
Recalibrated RAVE MDF
13
Summary RAVE vs. APO

• Stellar parameters from the v04 catalog cannot be
  trusted.
• Using T(J-K) or other Teff vs. color relation may
  greatly improve the situation, but it assumes
  that most stars are either TO or RGB stars. Cool
  dwarfs, BHB stars, etc. can still be problematic.

• Want to take some more parameter calibrators,
  but need to move to science-driven observations
  to justify time to TAC.

14
Abundance Science I Al/Fe

• Study of MW bulge (Fulbright, Rich McWilliam et
  al. 2006, about to be submitted), plus the work
  of others for different stellar pops, shows that
  the Al/Fe vs. Fe/H plot discriminates between
  stellar populations. Mn and Cu also act the same
  way.
• The thick disk stars lie between the bulge and
  thin disk. Halo stars cover a wide range of
  Al/Fe over the metal-poor end.

Fulbright et al. (2006)
¾
15
Abundance Science I Al/Fe

• The 8772 and 8773 Å Al I lines are relatively
  strong (50-100 mÅ in the Fe/H gt -0.5 stars
  studied with APO data), fairly clear of blends,
  and behave similarly to the better-known red Al
  lines (e.g., 6696/8 Å, 5577 Å, etc.).
• Can this be done with RAVE data?
• Will require good vt values due to strength, but
  the 1 dex range of Al/Fe at a given Fe/H
  means even bad abundances are OK.

16
Abundance Science II Hi-vel stars
Fulbright (2000, 2002)
Local stars with high rest-frame velocities in
the Galaxy and have orbits w/Rapo gt 20 kpc have
lower a/Fe, etc. than other metal-poor halo
stars.
17
Abundance Science II Hi-vel stars
Fulbright (2000, 2002)
? APO data
?
?
?
?
?
?
?
Analyzed the spectra of 2 of the high-velocity
stars identified by G. Ruchti. The X/Fe values
shown above agree in form with Fulbright (2000)
results.
18
Abundance Science III C-stars
Noticed a strong C2 bandhead at 5165 Å for one of
the parameter calibrator stars. Taken because
it was suspected to be metal-pooronly one so
far! RAVE param. 4500 K/0.5/-2.0 APO param.
4660 K/0.6/-2.3 X/Fe ratios are normal for a
metal-poor star of its Fe/H. Exceptions
C/Fe 1.8 Ba/Fe 0.9
C2
19
Abundance Science III C-stars
High Ba/Fe suggests this is an
s-process-enhanced C-rich star. These stars are
probably binaries where a companion evolved
through the AGB phase and polluted the stars seen
nowkey to understanding the s-process. Cohen et
al. (2006) looked at the C-rich stars in the HES
surveystars with V mag of 14 to 16. This star
has V 11 mag.
Can we find C-rich star using features (CN?) in
the RAVE region? We will take a higher S/N
spectra of this star to find out.
20
(No Transcript)
21
Is it the Data Quality?
22
Observational Data II

• All data from two half-nights in Nov. 2005.
• Two half-nights in
• Feb. 2006clouded out.
• Three half nights next week 1116 Apr.
• Expect 2 half-nights
• per quarter.

• Want to take some more parameter calibrators,
  but need to move to science-driven observations
  to justify time to TAC.

23
Observation Plan

• For the first phase, obtain spectra for a wide
  range of stellar parameters, analyze, and compare
  to what RAVE has obtained.
• How can the RAVE results be improved?
• The data include the RAVE wavelength region How
  do the spectra compare? Can we obtain other
  elemental abundances from the RAVE region?

24
RAVE Parameters
Open Dots are the final APO high-res.
parameters. Note clustering around TO and red
giant-branch clump. One weakness of
ionization-based log(g)s is that main sequence
stars often end up with values closer to 4.0 than
the canonical value of 4.5.
25
(No Transcript)