Substructure in the Galaxy

Transcription

1 Substructure in the Galaxy Amina Helmi Kapteyn Astronomical Institute Groningen, NL

2 Is this how our Galaxy formed? Jeffrey Gardner

3 Hierarchical paradigm Main characteristic of model: mergers Can we find the signatures? Substructure and the search for tidal streams How did the building blocks look like? what is the link to today's nearby small galaxies (dsph)?

4 Mergers and substructure Can we find the signatures? Substructure and the search for tidal streams How did the building blocks look like? what is the link to today's nearby small galaxies (dsph)?

5 Substructure in the halo In outer halo tidal features spatially coherent over many Gyr -> wide-field surveys SDSS, 2MASS, yielding spectacular results: - Substructure appears to be common - Yet to quantify fraction of halo has been accreted Belokurov et al 2006 Juric et al 2006 Inner halo: predicted to be in place long ago Mixing is large, signatures of accretion -> moving groups (predicted ~ 500) Samples are still small, not many new recent discoveries

6 Substructure in the thick disk The thick disk could have resulted from a minor merger Hints of substructure in the thick disk Gilmore, Wyse & Norris find extra kinematic component with v r ~ 100 km/s Navarro, Helmi & Freeman revisited Arcturus group (Eggen 1971); present in any compilation of radial velocities of metal-weak stars; also in RAVE Gilmore et al 2002 Navarro et al 2004

7 Substructure in the thick disk The Nordström catalog shows excess of stars with high-eccentricities compared to smooth Galaxy model CMD is very distinct from disk Sharp transition in the properties of stars at ecc ~ 0.3 Clear presence red giant branch subgiant branch turn-off Helmi et al 2006

8 Thick disk groups Three groups can be identified on the basis of their slightly different chemistry (also slightly different kinematics) Isochrone fitting shows they are indeed different (location of the T.O.) Detailed chemical abundances also peculiar Thick disk: a pre-existing disk + debris from one disrupted satellite? Or, the superposition of debris from several satellites?

9 Mergers and the building blocks Can we find the signatures? Substructure and the search for tidal streams How did the building blocks look like? what is the link to today's nearby small galaxies (dsph)?

10 The dwarf spheroidals Dwarf spheroidals are the smallest nearby galaxies All contain ancient (> 10 Gyr) and metal-poor stars (< -1.6 dex) metallicity is built through time in the interiors of stars stellar atmospheres have the chemical abundance of ISM at formation time fossil record of the conditions of the very early Universe constrain IMF at high-z, could be linked to reionization age = 15 Gyr Sculptor Fornax Scl DART Sextans Carina

11 The dsph and the Galactic building blocks The oldest stars in the dsph (hence the most metal-poor) presumably formed about the same time as the first stars in the building blocks Therefore, one may compare the metallicity distribution of the stars at the metal-poor end both in the dsph and in the Galactic halo Clearly, one cannot/should not compare present-day dsph to the Galactic halo since the dsph have had a Hubble time to evolve on their own.

12 The dsph and the Galactic building blocks Compare BB to BB, BB to the most ancient components of presentday systems, but not to presentday systems that have evolved independently

13 DART and the dwarf spheroidals Dwarf galaxies Abundances and Radial velocities Team: Eline Tolstoy, Mike Irwin, Vanessa Hill, Kim Venn, Matthew Shetrone, Giuseppina Battaglia, Bruno Letarte, Pascale Jablonka, Nobuo Arimoto, Francesca Primas, Patrick François, Andreas Kaufer, Thomas Szeifert, Tom Abel, Kozo Sadakane ESO Large Program on the VLT/FLAMES Spectra for several 100s of stars in dwarf spheroidals HR detailed chemical abundance studies in the centre (Fnx, Scl, Sext, Car) LR metallicity and radial velocity from CaT across the system (Fnx, Scl, Sext) Fornax Sculptor Sextans Carina

14 LR data Select targets on the RGB from imaging survey WFI/ESO Oldest population also represented amongst RGB stars Across the whole galaxy (to avoid biases induced by the spatial distribution of the various stellar pops) In the case of Sextans, basically followed up every red giant branch star candidate down to V=20 In the case of Sculptor followed ~50% of the candidates Tolstoy et al 2004

15 LR data Members within ± 3 σ; clear separation from foreground S/N > 10 allows δ[fe/h] 0.1 dex; δv r ~ 2 km/s [Fe/H] derived from CaT EW calibration 513 * 933 * 202 * 364 *

16 Metallicity distribution Large variety in metallicity distribution (reflects widely varying star formation and chemical enrichment) Common denominator: no stars with [Fe/H] < -3 dex!

17 Metal-poor end of the metallicity df Simple model of chemical evolution N(<Z) = A(1 -exp{-(z-z 0 )/p}) A depends on initial gas available Z 0 initial abundance, p=yield For small Z, and since Z = Z 10 [Fe/H] N(<[Fe/H]) ~ a 10 [Fe/H] + b where a = A Z /p, b = -a 10 [Fe/H] 0 Exponential decline at low [Fe/H] understood The initial metallicity of the gas [Fe/H] 0 is very similar in all galaxies Sculptor -2.9 ± 0.2 Sextans -2.7 ± 0.1 Fornax -2.7 ± 0.3 Carina -2.7 ± 0.2

18 Implications and the Galactic halo Suggests uniform pre-enrichment 1 Mpc 3 volume Lowest metallicity coincides with mean of the IGM towards QSO at z ~ 3-5 (corresponding to T ~ Gyr ago) The Galactic halo does show very metal-poor (VMP) stars Ryan & Norris 1991 Christlieb 2004

19 Implications and the Galactic halo 130 HES giants with [Fe/H] < -2.5 dex, ~ 35 with [Fe/H] < -3 dex Bootstrap HES distribution to account for the different sample sizes Significantly different KS test probabilities < 10-3 If the same parent distribution, would have expected ~1/4 (i.e. 35/130) stars in the dsph < -3 dex. Although samples are small, we should have seen 10 stars with [Fe/H] < -3 dex. Helmi & DART 2006

20 The puzzle Dwarf spheroidals lack very metal-poor stars, [Fe/H] < -3 dex, despite their low mean metal abundance. Comparison to Galactic halo, leads to the conclusion that the dwarfs progenitors not the building blocks of galaxies like the Milky Way Possible scenarios: building blocks are high-σ peaks, dsph 1σ peaks collapse earlier and enrich the IGM, from which the dwarfs (at later z) form. dsph not related to the reionization of the Universe. IMF in building blocks different from dwarf spheroidals First stars top heavy in dsph, but low mass stars present in building blocks (high σ peaks, e.g. Nakamura & Umemura 1999).