Diapositiva 1

1
Cosmic stellar relics in the Galactic halo
Stefania Salvadori1, Raffaella Schneider2
Andrea Ferrara1 1SISSA/International School for
Advanced Studies, via Beirut 4, 34100 Trieste,
Italy 2INAF-Osservatorio Astrofisico di
Arcetri, Largo E. Fermi 5, 50125 Firenze, Italy
astro-ph/0611130
Metal poor stars represent the living fossils of
the first stellar generations. The Metallicity
Distribution Function (MDF) of long living metal
poor stars is observed both in the Galactic halo
and in nearby dwarf galaxies, satellites of the
Milky Way (MW). Using a Monte Carlo method based
on the semi-analytical Press Schechter
formalism, we develop a new code GAMETE (GAlaxy
MErger Tree Evolution) to reconstruct the
merger tree of the Galaxy and follow the
evolution of gas and stars along the hierarchical
tree. Our approach allows to compare the
observational properties of the MW with model
results exploring the imprint of undetectable
high redshift processes. By matching model
predictions to the observed MDF of metal poor
stars in the Galactic halo we (i) study the
metal enrichment history of the MW environment
(ii) set limits on the primordial Initial Mass
Function (IMF) and the critical metallicity (Zcr)
governing the transition from PopIII to PopII
star formation mode (iii) investigate the
statistical impact of second generation of stars
in up-to-date observed samples.
The model
Comparison with data
Hierarchical merger tree Along the lines of Cole
(1) and Volonteri, Haardt Madau (2) we
have developed a binary Monte Carlo algorithm
with accretion mass in order to reconstruct the
hierarchical merger tree of the Galaxy up to
z20. At each time-step halos can lose part of
their mass (corresponding to a cumulative
fragmentation into halos with Mmass and fragment into two progenitor halos. We
have assumed a mass resolution Mres M4(z)/10,
where M4(z) represents the halo mass which
corresponds to a virial equilibrium temperature
Tvir104K. At each redshift the mass below the
resolution limit accounts for the Galactic
Medium (GM) which represents the mass reservoir
into which halos are embedded.
The fiducial Model We compare the new MDF
determination for Galactic halo stars by Beers
Christlieb (12) with the simulated MDF obtained
using the fiducial model (Fig.3). For
completeness we have added to the sample the two
known hyper-metal poor stars (HMPS) (13,14,15).
The fiducial model provide a good fit to the
MDF but cannot account for the two HMPS.
Zcr 104 Z?
Zcr 104 Z?

• Star formation history
• Several simple but physically motivated
  prescriptions have been adopted in order to study
  the star formation history the chemical
  evolution along the hierarchical tree
• (i) Stars form in Lya cooling halos (Tvir 104K
  )
• (ii) M f(z) Mgas e ?t/tff(z) Mgas where tff
  is the free-fall time and e a free parameter
• (iii) According to the critical metallicity
  scenario (3,4,5,6,7,8,9) low-mass star
  formation is triggered by the presence of metals
  in the gas exceeding Zcr10-51Z?
• PopIII stars form if Z Zcr and with a
  reference mass mPopIII 200 M?
• PopII/I stars form if Z Zcr and according
  to a Salpeter IMF (0.1 M?
• (iv) Stars, once formed, evolve instantaneously
  (IRA approximation)
• (v) Mechanical feedback is active if SN explosion
  energy ESN ewNSN overcomes the binding
  energy of the halo. ew is the second free
  parameter of the model
• (vi) Gas metals ejected into the Interstellar
  Medium (ISM), an eventually into the GM through
  mechanical feedback, are instantaneously
  homogeneously mixed in it.

Fig.3 Left panel MDFs observed (points,
3) and simulated (histogram) using the
fiducial model and Zcr104 Z?, mPopIII 200M?.
The histogram represents the average value of the
MDF over 200 realizations of the merger tree
the shaded area the s Poissonian errors. Right
panel The same results plotted in terms of
cumulative number of stars below a given Fe/H
Model Calibration We have used the observed
properties of the MW to fix the best values of
the two model free parameters, e and ew which
respectively regulate the star formation and the
wind efficiency. We found that the calibration
procedure is completely independent of Zcr and
PopIII IMF. The global properties of the MW are
well reproduced both using a close-box model
(e0.5, ew0 .i.e. no wind) and a model
including mechanical feedback (e0.7,ew 0.2).
Constraints on Zcr value We explore the
sensitivity of our results to variation of Zcr
(Fig.4). The choice Zcr0 is equivalent to state
that stars, including metal-free ones, form at
all times according to a Salpeter IMF. Models
with Zcr 10-6 Z? can account for the two HMP
stars at the price of overpopulating the
metallicity desert 5.3
Model results
The Metallicity Distribution Function We use the
properties of the MDF in order to study the role
of feedback and discriminate between the two best
fit models (Fig.1). Close box models do not
reproduce the observed Fe/H range since no
stars with Fe/H -1.6 are formed. Feedback
models instead allow an efficient dilution of
metals. Low-Fe/H PopII stars form in halos
accreting GM gas enriched by earlier SN explosion
(Fig.2a). Mechanical feedback is required in
order to reproduce both the global properties of
the MW and the observed Fe/H range. We refer to
the case e0.7, ew0.2 as our fiducial model.
Zcr104 Z? mPopIII 200 M?
Fig.5 The same as Fig.3 but for values of
Zcr105 , 106 , 0.
Observed Fe/H range
Statistics of second generation stars We study,
for different Zcr models, the statistical impact
of second generation (2G) stars i.e. stars
that form out of gas enriched only though the
nucleosynthetic products of PopIII stars (Fig.5).
The expected number of 2G stars is negligible in
the observed sample, independent on the Zcr
value. In Beers Christlieb 12 sample (2876
stars) only 1-2 stars could retain the metal-free
nucleosynthetic imprint.
Fig.1 MDFs for models e 0.5, ew 0 (shaded
histogram) and e 0.7,, ew 0.2 (no shaded
histogram) Histograms are obtained as average
over 200 realizations.of the merger tree
The Galactic Medium We study the evolution of
GM iron and oxygen abundance together with the
specific contribution to O/H by PopIII and
PopII stars (Fig.2a). We find that PopIII stars
dominate the GM enrichment for 11 PopII stars are the dominant enrichment channel
at lower redshifts z amount of metals predicted in the GM at z0 have
been ejected by halos with M 6109M? (Fig.2b)
Fig.6 Impact of 2G stars for the fiducial model
and Zcr 104 ,10 6, 0. The highest histograms
show the total MDF while the smaller represent
the MDFs for 2G stars. The percentage show the
2G fraction with respect to the total in the
range in which they appear.
References
1Cole S. et al., 2000 MNRAS, 319, 168.
2Volonteri M., Haardt F. Madau P., 2003, ApJ,
582, 559. 3Bromm et al., 2001, MNRAS, 328, 969.
4Omukai K., 2001, ApJ, 534, 809. 5Omukai K.
et al., 2005, ApJ, 626, 627. 6Schneider R. et
al., 2002, ApJ, 571, 30. 7Schneider et al.,
2003, Nat, 422, 869. 8Schneider et al., 2006,
MNRAS, 369, 1437. 9Bromm V. Loeb A., 2004,
New Astron., 9, 353. 10Ganguly R. et al., 2005,
ApJ, 157, 251. 11Mori M. Umemura M., 2006,
Nat, 440, 644. 12Beers Christlieb 2006,
private communication. 13Christlieb N. et al.,
2002, Nat, 419, 904. 14Frebel A.. et al., 2005,
Nat. 434, 871. 15Christlieb N., Bessel M.
Eriksson K., 2006, in preparation
Fig.2a Evolution of the GM elemental abundances.
Solid lines show the Fe/H and O/H evolution
averaged over 200 realization. The shaded area
delimits the s dispersion region for Fe/H. The
point is the measured O/H in High velocity
clouds (10). Fig.2b Lower panel ratio of
metals ejected by halos (Mej ) as a function of
their mass and redshift with respect to the total
amount of metals predicted in the GM at z0
(MejTOT ). Curves represent Mej/MejTOT510(4,3,2
) isocontour. The minimum and maximum (Mmin
,Mmax) halo mass in which star formation can
develop are also shown. The two rectangles
identify the position of the maxima the
horizontal line show the PopIII stars termination
epoch.Top panel the shaded area shows the
cumulative contribution to MejTOT , integrated
over redshift, by halos with different mass
points with associated 1s error bars represent
stellar-to-total mass ratios in the corresponding
mass bin.