The First Billion Years of Galaxy Evolution Ranga-Ram Chary Spitzer Science Center/Planck Data Center California Institute of Technology, Pasadena, USA

1
The First Billion Years of Galaxy
EvolutionRanga-Ram CharySpitzer Science
Center/Planck Data CenterCalifornia Institute of
Technology, Pasadena, USA
2
z10
Colleagues and Collaborators Asantha Cooray
(UCI) Ian Sullivan (Caltech) GOODS team
(Earth) Edo Berger (Princeton) Len Cowie
(Hawaii) Avi Loeb (Harvard)
3
Outstanding Questions.

• What were the first sources that formed ?
• Which sources were responsible for reionizing the
  Universe?
• What was the reionization history of the Universe
  ?
• Did supermassive black holes form in the nucleus
  of dark matter halos before stars ?
• What was the luminosity function of the first
  galaxies and how does that relate to the dark
  matter halo mass function?
• How soon was dust produced in these galaxies ?
• What was the initial mass function of stars in
  the first galaxies ?
• What was the early metal enrichment history ?

4
Need 1 photon to start reionization

• Sensitive to
• Clumpiness of the gas
• Temperature of the gas
• Co-moving electron density

Need 10 photons to maintain ionized hydrogen
due to recombinations
5
Constraining Small HI Fractions at z6

• Fan et al., Becker et al., Djorgovski et al.
  spectra of Sloan quasars showing the
  Gunn-Peterson trough
• Evidence for N_Hlt10-3.5

6
WMAP constrains large HI fractions
WMAP 1s range
7
We can consider two reasonable scenarios which
fit all existing data
WMAP Tau 0.084/-0.016 zlt6 contributes
0.04 So 0.044/-0.016 measures reionization
history.
?z9 700 Myr
?z3 ?400 Myr
Now lets balance the ionizing photon budget and
stay out of debt.
8
Detecting zgt5 Galaxies/QSOs Lyman Break Technique
9
What were the sources of reionization ?
Yan Windhorst 04
But even galaxies are too few
Definitely galaxies QSOs are too few
10
Cosmic Variance might be a Problem
1 deg
A GOODS Field
11
Minimum Required Evolution of UV Luminosity
Density
Fast
Slow
12
Data/fits from Bouwens et al. 08
Solid lines Required for Reionization Dashed
lines Measured
Some uncertainty in the faint end slope of the
UVLF But clear evolution in L between 6 and 7
13
Predicted Evolution of the UVLF
Fast reionization Evolving L
Dark Matter Halo Mass Function
Slow reionization Evolving L
Fast reionization Constant L at zgt6
Slow reionization Constant L at zgt6
Chary Cooray 08
14
6 7 10 13 14
z Age(Gyr) 6 0.95 7 0.78 10 0.48 13 0.34
14 0.30 22 0.16
Fast Reionization
6 7 10 13 14 22 23 24
But L has to be fixed at zgt7! Otherwise faint
end slope is too steep
Slow Reionization
15
In principle, CIB measurements can reveal the
difference
16
Unfortunately Sky Background at Infrared
Wavelengths is Dominated By Zodiacal Light
CIRB (nW m-2 sr-1) ? ?I? 2.2
22.4?6.0 3.5 11.0?3.3 140 25?7 240 14?3
17
But to see the action, need the faint
galaxies.And to see faint galaxies, need cosmic
explosions which are x1000 brighter than the
galaxy.
J-band
R-band
GRB 050904 z6.295 Kawai et al., Haislip et al.
Can measure gas and metals in the galaxies.
18
Hint of a higher SFR at z6 from GRB Hosts
Chary, Berger, Cowie, ApJ, 671
19
Spitzer/GOODS data on 60 zgt5 Galaxies
ACS z-band
IRAC ch1
IRAC ch2
20
Spitzer detects about 50 of zgt5 objects, mostly
the brighter ones
All
Spitzer
HAEs
21
Massive Galaxies Have Larger Balmer Breaks
Older Stellar Population
See also Yan et al, Bunker et al., Eyles et al.,
Stark et al.
22
Detecting Ha from IRAC 4.5/3.6 micron ratios
technique works for QSOs
Jiang et al. Sloan QSOs
Flux Density
23
Old galaxies can reproduce the red color, but not
the UV flux
More Ha
Young galaxies
Old galaxies
24
Can fit SEDs and calculate SF History
Difference is from Ha in emission ? UV SFRlt10
M?/yr Ha SFR100 M?/yr Ha EW0.2µm
Indirect evidence for AV1 mag of dust within 1
Gyr of the Big Bang. But jury is out since no mm
detection (Combes group).
Chary et al. 05 Hu et al. 02 Schaerer et al.
25
Constraints on Reionization if Dust were important
?107 M?/Mpc3
?108 M?/Mpc3

• For the measured star-formation rate and stellar
  mass density, reionization must be a brief
  process, lasting 20-50 Myr which happened between
  z6-7 INCONSISTENT WITH WMAP !
• If reionization was a longer process starting at
  z9, stellar mass density must be higher than
  measured OR
• more ionizing photons per baryon i.e. a top-heavy
  stellar mass function

26
Maybe SFR was higher in the past due to
e-folding time of burst or due to dust
Key parameters Stellar Initial Mass
Function Age of starburst Mass of stars Metals
in stars
Nuisance Parameters Clumping Factor Escape
Fraction
Chary, 2008, ApJ
27
Balancing the Ionizing Photon Budget A Simple
Low Mass Cut in the IMF is Insufficient
So, a top-heavy IMF at zgt6 seems
necessary. Furthermore, dust at high-z has been
directly observed in QSOs.
28
Conclusions

• What were the first sources that formed ?
• Star-forming galaxies with a LF which is similar
  to the DM halo mass function
• Which sources were responsible for reionizing the
  Universe?
• Sub-L star-forming galaxies
• What was the reionization history of the Universe
  ?
• Slow reionization models are preferred where
  ?z(0.1,0.9)9
• How soon was dust produced in these galaxies ?
• Seems to be within 800 Myr from QSO spectroscopy
  and candidate HAEs. Evidence for SN dust ?
• What was the initial mass function of stars in
  the first galaxies ?
• Top-heavy, dN/dMM-1.6