Elemental Abundance Trends of the Elemental Abundance Trends of the - PowerPoint PPT Presentation

Elemental Abundance Trends of the Elemental Abundance Trends of the Elemental Abundance Trends of the Metal- -Poor Galactic Disks: Clues to Poor Galactic Disks: Clues to Metal Metal-Poor Galactic Disks: Clues to Formation of the Galaxy Formation of the Galaxy Formation of the Galaxy Gregory Ruchti (PhD thesis) Department of Physics and Astronomy Johns Hopkins University advisors: Rosemary Wyse Jon Fulbright QuickTime™ and a decompressor are needed to see this picture.

Motivation Motivation • Probe the evolution of the Milky Way Galaxy through • Probe the evolution of the Milky Way Galaxy through metal-poor (early?) disk stars metal-poor (early?) disk stars • Address questions such as: • Address questions such as: • How and when did the Galactic disk start to form? • How and when did the Galactic disk start to form? • How was it that a thin and a thick disk formed? • How was it that a thin and a thick disk formed? • What is the importance of mergers in disk formation and Galactic • What is the importance of mergers in disk formation and Galactic evolution? evolution? • How did the Galaxy evolve chemically? • How did the Galaxy evolve chemically? • Combine power of large RAVE sample to provide • Combine power of large RAVE sample to provide candidate metal-poor disk stars, with elemental candidate metal-poor disk stars, with elemental abundances from follow-up high-resolution spectroscopy abundances from follow-up high-resolution spectroscopy • Information on recent and past SFH, IMF, more than overall [M/H] • Information on recent and past SFH, IMF, more than overall [M/H] • Different theories of disk(s) formation and evolution make different • Different theories of disk(s) formation and evolution make different predictions for elemental abundance patterns and low-metallicity predictions for elemental abundance patterns and low-metallicity tail (starburst vs satellite accretion etc) tail (starburst vs satellite accretion etc)

dSph vs. field-star abundances complied by Koch Halo kinematics Different patterns reflect different SFH and chemical enrichment

What about the metal-poor disk(s)? What about the metal-poor disk(s)? • To date, published elemental • To date, published elemental abundances for only about 25 thick abundances for only about 25 thick disk stars with [Fe/H] < -1, and only disk stars with [Fe/H] < -1, and only ~10 with [Fe/H] < -1.5. ~10 with [Fe/H] < -1.5. • Estimated old ages and high • Estimated old ages and high [alpha/Fe] ratios consistent with [alpha/Fe] ratios consistent with rapid star formation and fixed IMF rapid star formation and fixed IMF • However, these stars are within a • However, these stars are within a few hundred parsecs of the Sun: a few hundred parsecs of the Sun: a small volume where thick disk rare. small volume where thick disk rare. • Additionally, cannot investigate • Additionally, cannot investigate possible radial and vertical possible radial and vertical metallicity or elemental abundance metallicity or elemental abundance gradients in this small volume. gradients in this small volume. • A larger sample, probing a larger • A larger sample, probing a larger volume, would greatly increase our volume, would greatly increase our e.g. Reddy & Lambert 2008 understanding of the metal-poor understanding of the metal-poor Thick disk: green, red and open thick disk . thick disk . ‘Hybrid’: black Halo: blue

Candidate Selection Candidate Selection • From RAVE database identify metal- poor stars with thin or thick disk kinematics as candidate sample • Mostly giants. • Use RAVE pipeline stellar parameters to derive preliminary distances (fits to isochrones) and hence space motions • Derive likely population assignment based on space motions, radial velocities and RPMD. • Candidate stars assigned to Galactic disk population if thin- or thick-disk population probabilities are 10 times greater than that of being a halo star. Same method applied to distinguish thin and thick disks.

Echelle Observations Echelle Observations • Obtained high resolution (R > 35,000) echelle spectra (S/N > 100) for ~500 (! – good weather!) candidate metal- poor thick disk stars. • Gives Fe, many other elements, and allows for improved stellar parameters and population assignment. • Echelle at Apache Point Observatory 3.5-m telescope. • Resolving Power ~ 37,500; λ = [3500, 9800] Å • 2-3 half-nights every quarter year. • MIKE on Magellan Clay Telescope in Chile. • Blue: Resolving Power ~ 42,000; λ = [3560, 5050] Å • Red: Resolving Power ~ 33,000; λ = [4850, 9400] Å • 2 nights in May 2007 and 2 nights in October 2008 • UCLES on Anglo-Australian Telescope in Australia. • Resolving Power ~ 40,000; λ = [4460, 7270] Å • 7 nights in June 2007, 5 nights in September 2008 • FEROS on 2.3-m Telescope at La Silla Observatory in Chile. • Resolving Power ~ 47,000; λ = [3900, 10,000] Å • 7 nights in February 2009

Elemental Abundances Elemental Abundances • Measure equivalent widths using the Automatic Routine for line Equivalent widths in stellar Spectra (ARES; Sousa et al. 2007). • Compared to hand-measurements, ~3 mÅ differences • Abundance analysis follows same methodology as Fulbright 2000: • One of the largest uniform samples of halo star abundances in the literature. • Differential halo-disk and disk-disk comparisons do not suffer any significant systematic effects. • Line list includes many interesting species: • Fe I and Fe II for stellar parameter estimation • Major alpha elements (Mg, Si, Ca, Ti, O) • Major neutron capture elements (eg. Ba, Eu, Y, Zr) • Odd-Z light elements (eg. Na, Al, K) • Fe-group elements (eg. V, Cr, Ni, Zn)

Elemental Abundances Elemental Abundances • Stellar parameters also derived during analysis: • Effective temperature set using excitation temperature method based on Fe I lines. • Surface gravity, log(g), set by minimizing difference between calculated abundance of iron from Fe I lines and Fe II lines. • Microturbulent velocity selected to minimize slope of Fe I iron abundance versus reduced width of each Fe I line. • A short-fall of this method is for log(g) < 1.0, ionization equilibrium fails. • For these log(g) values are typically lower than expected. • Typical errors: σ Teff ~200K, σ log(g) ~0.3, σ [Fe/H] ~0.1, σ [alpha/Fe] ~0.1 Check with Elizabeth re status of pipeline!!

0 Echelle [Fe/H] Histogram -1 Echelle [Fe/H] -2 -3

Distances Distances Under development • Distances are estimated using echelle stellar parameters: • Fit to grid of Padova isochrones defined in metallicity with two selected ages: 12 Gyr and 4 Gyr. • Used T eff , log(g), [Fe/H] (all echelle-based), and 2MASS J-K s as fit parameters. • To account for lifetimes in different evolutionary phases we applied the theoretical Luminosity Functions as a priori probabilities to our fits: most important for the low log g RGB. • How reliable are the isochrones? • Final distances estimated from the mean of the distances derived from the two ages. • Typical errors are ~ 20-30%

106 stars analysed so far to obtain elemental abundances Log g (echelle) ~2dex spread in [Fe/H] Log Teff (echelle) • Padova isochrones with Z = 0.001012 ([Fe/H] = -1.5, [alpha/Fe] = 0.3) -- Black = 12 Gyr; Red Dashed = 4 Gyr • Errors in Log(g) ~ 0.3 dex, Errors in Teff ~ 200 K

Population Assignments Population Assignments Under development Applied 5 ( 3 + 2) Membership Criteria with different distance dependencies: 1. Full space motion criterion • Similar to initial selection of sample to observe • Compute 3-D space motions and compare to standard velocity distributions: Thin disk: 〈 V Θ 〉 = -15 km/s, ( σ Π , σ Θ , σ Z ) = (39,20,20) • Thick disk: 〈 V Θ 〉 = -51 km/s, ( σ Π , σ Θ , σ Z ) = (63,39,39) • • Also possibility of higher lag for metal-poor thick disk Halo: 〈 V Θ 〉 = -220 km/s, ( σ Π , σ Θ , σ Z ) = (141,106,94) • • Candidate stars assigned to Galactic disk population if thin- or thick-disk population probabilities are 10 times greater than that of being a halo star. Same method applied between thin and thick disks.

2. Component dispersion limit criterion • Another assignment probability given by how much each velocity component differs from its standard mean velocity. If |V i - 〈 V i 〉 | ≥ 2.5 x σ Vi , then the star is assigned to the next • population. 3. RPMD criterion • Find which population in reduced proper motion space is closest to each star • A weighted average of these three criteria is computed • Set up monte carlo of 1000 points using normal distribution for each input parameter • Calculate mean and spread for each assignment using the monte carlo points -- spread used as the weight • Gives an initial estimate of the final population assignment.