High Redshift Starbursts

1
High Redshift Starbursts

• Mauro Giavalisco
• Space Telescope Science Institute
• and the GOODS team
• STScI/ESO/ST-ECF/JPL/SSC/Gemini/Boston U./U.
  Ariz./U. Fla./U. Hawai/UCLA/UCSC/IAP/Saclay/Yale/A
  UI

2
The Quest for the Early Galaxies
Giavalisco 2002 ARAA Ellis 1997 ARAA

• During the mid-90s, with improved
  instrumentation, the commissioning of the 8-m
  class telescopes, and the repair of HST, a number
  of influential deep galaxy surveys (CFRS, LBGS,
  HDF) uncovered two important pieces of evidence
• Normal, luminous galaxies (the bright end of the
  Hubble sequence) were essentially in place by z1
• Massive (M) galaxies formed prior to z1
• The universe was well populated with star-forming
  galaxies at z3
• At z1 these must be old and/or massive or both.
  Are these the progenitors of the bright galaxies?
• Earlier suggestions that the bulk of galaxies
  formation occurred at zlt1 and that essentially
  no galaxies are to be expected at redshifts zgt1
  (1993, actual quote) were dismissed.

3

Lilly et al. 1995
Abraham et al. 1996
4
Star-forming galaxies at z3 (Lyman Break
Galaxiess)
Steidel, Giavalisco, Dickinson, Pettini
Adelberger 1996
5
Efficient star formation at zgt2.5
Steidel, Adelberger, Giavalisco, Dickinson
Pettini 1999
6
Galaxy morphology at z3

• Smaller
• Regulars,
• Irregulars,
• Merging,
• Spheroids?
• Disks?
• No Hubble Seq.
• No l-dependence

Giavalisco et al. 1994 Giavalisco et al.
1996 Steidel, Giavalisco, Dickinson Adelberger
1996 Lowenthal et al. 1997 Dickinson 1998
Giavalisco 1998 Papovich, Giavalisco, Dickinson,
Conselice Ferguson 2003 Papovich, Dickinson,
Giavalisco, Conselice Ferguson 2004
7
UV-star formation rates
Some rates are relatively low, todays
spirals others are prodigiously
high Metallicity 1/10 to solar Still an
open issue
8
The birth of the GOODS

• No Hubble Sequence apparently observed at zgt2.
  When and how did it form?
• What kind of galaxies are LBGs
• Bursting dwarfs? Massive?
• What did they evolve into? How much stellar mass
  did they contribute?
• Up to which redshift are there LBGs? When did SF
  on galactic scale start?
• Are there other (non LBG selectable, I.e. non
  star-forming or very obscured) galaxies at zgt2?
• How does star formation occur and evolve?

9
The GOODS Treasury/Legacy Mission

• Aim to establish deep reference fields with
  public data sets from X-ray through radio
  wavelengths for the study of galaxy and AGN
  evolution of the broadest accessible range of
  redshift and cosmic time.
• GOODS unites the deepest survey data from NASAs
  Great Observatories (HST, Chandra, SIRTF), ESAs
  XMM-Newton, and the great ground-based
  observatories.
• Primary science goals
• The star formation and mass assembly history of
  galaxies
• The growth distribution of dark matter structures
  
• Supernovae at high redshifts and the cosmic
  expansion
• Census of energetic output from star formation
  and supermassive black holes
• Measurements or limits on the discrete source
  component of the EBL
• Raw data public upon acquisition reduced data
  released as soon as possible

10
A Synopsis of GOODS

• GOODS Ground
• ESO, institutional partner (PI C. Cesarsky),
  CDF-S
• Full spectroscopic coverage in CDF-S
• Ancillary optical and near-IR imaging
• Keck, access through GOODS CoIs
• Deep spectroscopic coverage
• Subaru, access through GOODS CoI
• Large-area BVRI imaging
• NOAO support to Legacy Treasury
• Very deep U-band imaging
• Gemini
• Optical spectroscopy, HDF-N
• Near-IR spectroscopy, HDF-S
• ATCA, ultra deep (5-10 mJy) 3-20 cm imaging, of
  CDF-S
• VLA, ultra deep HDF-N (Merlin, WSRT)
• JCMT SCUBA sub-mm maps of HDF-N

• GOODS Space
• HST Treasury (PI M. Giavalisco)
• B, V, i, z (3, 2.5, 2.5, 5 orbits)
• 400 orbits
• ?? 0.05 arcsec, or 0.3 kpc at 0.5ltzlt5
• 0.1 sq.degree
• 45 days cadence for Type Ie Sne at z1
• SIRTF Legacy (PI M. Dickinson)
• 3.6, 4.5, 5.8, 8, 24 µm
• 576 hr
• 0.1 sq.degree
• Chandra (archival)
• 0.5 to 8 KeV
• ?? lt 1 arcsec on axis
• XMM-Newton (archival)

11
(No Transcript)
12
GOODS/ACS B 27.5 V 27.9 i 27.0 z 26.7
HDF/WFPC2 B 27.9 V 28.2 I 27.6
?m 0.3-0.6 AB mag S/N10 Diffuse source,
0.5 diameter Add 0.9 mag for stellar sources
In 2-3 months we will release a new stack of
15 orbits in the z band, as well as 50 and
30 more exp. time in the i and V bands, in
both fields, plus source catalogs (GOODS)
13
GOODS galaxies at High Redshift
B435 V606 i775
z850

• Theory predicts that dark matter structures form
  at z20-30
• It does not clearly predict galaxies, because we
  do not fully understand star formation
• Empirical information on
• galaxy evolution needed to
• the highest redshifts
• GOODS yielded the deepest and largest quality
  samples
• of LBGs at z4 to 6

z4
14
LBG color selection
B-dropouts, z4
V-dropouts, z5
15
Galaxies at z6 (6.8 of the cosmic age)
Dickinson et al. 2003
16
Observed redshift distribution
24
Z5.78
Z6.24?
Curves from full numerical simulations Giavalisco
et al. 2004, 2005
V
Z5.83
Spectra from Bunker et al. 2003 Stanway et al.
2003 Vanzella et al. 2004 and the GOODS Team
17
LBG luminosity function
Apparently, very little evolution in the UV
luminosity function
18
The history of the cosmic star formation activity
We find that at z6 the cosmic star formation
activity was nearly as vigorous as it was at its
peak, between z2 and z3. NOTE soon, nearly
all GOODS will have three times the original
exposure time in z band, and 50 more in i band
(thanks to the Sne program). Measure at z6 will
significantly improve.
a-1.6 assumed
Giavalisco et al. 2004 Giavalisco et al. 2005, in
prep.
19
Still uncertainty on measures

• LF still not well constrained
• Clean z6 color selection still missing
• Cosmic variance still not understood
• Will use SST data to refine z6 sample
• Will triple exp time in GOODS

See also Bunker et al. 2004
Bouwens et al. 2004
20
SFR from X-ray emission
Lehmert et al 2005 See also Giavalisco 2002, ARAA
21
Star formation rates
z4 B-band dropouts
Dust obscuration correction Calzetti starburst
obscuration law BC synthetic SED Similar to
what observed at z3
22
SIRTF Imaging
20.0 20.7
0.11 26.3
0.21 25.6
1.66 23.4
1.35 23.6
GOODS sensitivity
5-s limiting flux µJy 5-s limiting AB mag
23
Stellar mass star formation

• Mass Rest-frame near-IR (e.g., rest-frame
  K-band at z3), provides best photometric measure
  of total stellar content
• Reduces range of M/L(l) for different stellar
  populations
• Minimizes effects of dust obscuration
• Star formation Use many independent indicators
  for to calibrate star formation (obscured open)
  in ordinary starbursts (e.g. LBGs) at z gt 2.
• mid- to far-IR (SIRTF/MIPS) rest-frame UV
  (e.g, U-band) radio (VLA, ATCA) sub-mm
  (SCUBA, SEST) nebular lines (spectroscopy)

Stellar mass fitting
Measuring star formation
24
Rest-optical -IR at z6

• SST IRAC detections of z6 galaxies
• gt stellar population dust fitting possible

ch1, 3.6mm lrest5300A
ch2, 4.5mm lrest6600A
Dickinson et al in prep
25
Luminosity Density versus Color and Redshift
Papovich et al. 2003
U- and B- dropouts have similar UV-Optical
color-magnitude "trends. Rest-frame UV
luminosity density roughly comparable at z 3
and 4. Increase of 33 in the rest-frame B-band
luminosity density from z 4 to 3. UV-Optical
color reddens from z 4 to 3, which implies an
increase in the stellar-mass/light
ratio. Suggests that the stellar mass is
increasing by gt 33 growth in B-band luminosity
density.
increase of 33
26
Implications for Galaxy Evolution
Dickinson, Papovich, Ferguson, Budavari 2003
27
Implications for Galaxy Evolution
Dickinson, Papovich, Ferguson, Budavari 2003
Stellar mass is building up We still need to
know how this growth depends on the total
mass Total mass of individual galaxies seems to
evolve less rapidly bottles form first, wine is
added later
GOODS Papovich et al. 2004
28
Morphology of Lyman Break Galaxies at z4
Sersic profile fits and Sersic indices Ravindran
ath et al. 2005
Irregulars (n lt 0.5)

Disks (0.5 gt n gt 1.0)
29
Morphology of Lyman Break Galaxies at z4
Bulges (n gt 3.0)
Central compact component / point sources? (n
5.0)
30
LBG morphology light profiles
We measured the light profiles and parametrized
them with the Sersic index
Ravindranath et al. 2005
31
Morphology of LBG
Theory predicts that when they form undisturbed,
galaxies are disks. Images show a distribution of
morphology. Both spheroid-like and disk-like
morphology are observed.
Ravindranath et al. 2005
z0 disks z0 spheroids
32
Morphology of LBG the GINI and M20 coefficients
mergers
spheroids
Both spheroids and disk, as well as transitional
morphologies, observed. Major mergers estimated
at 15-25, both at z4 and z1.4 (in agreement
with kinematics of close pairs with DEIMOS-DEEP
Lin et al. 2005)
Lotz, Madau, Giavalisco, Conselice Ferguson 2005
33
Local galaxies at high redshift
Statistics calibrated using local galaxies
Lotz et al. 2005
34
LBG morphology
Lotz et al. 2005
35
LBG morphology
Lotz et al. 2005
36
LBG morphology
Lotz et al. 2005
37
Infrequent morphological k-correction
WFPC2 (HDF) and NIC3 J and H images
Internal color dispersion consistent with
relatively young and homogeneous stellar
population
Dickinson 1998 Papovich, Giavalisco, Dickinson,
Conselice Ferguson 2004 Papovich, Dickinson,
Giavalisco Conselice Ferguson 2004
38
The Evolution of galaxy size

• First measures at these redshifts
• Testing key tenets of the theory
• Galaxies appear to grow hierarchically

RH(z)-2/3
Standard ruler
RH(z)-1
Ferguson et al. 2003
39
Galaxy Clustering at High Redshift

• Galaxies at high redshifts have strong spatial
  clustering, I.e. they are more clustered than the
  z0 halos de-evolved back at their redshift.
• High-redshift galaxies are biased, I.e. they
  occupy only the most massive portion of the mass
  spectrum (today, the bias of the mix is b1).
• Important
• evolution of clustering with redshift contains
  information on how the mass spectrum gets
  populated with galaxies as the cosmic time goes
  on.
• Clustering of star-forming galaxies contains
  information on relationship between mass and star
  formation activity

40
Clustering of star-forming galaxies at z3
r03.3/- 0.3 Mpc h-1 g -1.8 /- 0.15
Steidel et al. 2003 Adelberger et al. 1998
Giavalisco et al. 1998
41
Strong clustering, massive halos
g1.55 r0 3.6 Mpc h-1
Porciani Giavalisco 2002
Adelberger et al. 2004
42
local galaxies mgt2.5E10 MO mgt1.0E11 MO
LBGs
K20
EROs
sub-mm
SDSS QSOs
Somerville 2004
43
Clustering segregationmass drives LUV (SFR)
GOODS Ground
Lee et al. 2005
Adelberger et al. (1998, 2004) Giavalisco et al.
(1998) Giavalisco Dickinson (2001)
44
Clustering segregation at z4 and 5
Clustering segregation is detected In the GOODS
ACS sample at z4 Consistent with other
measures, e.g. Ouchi et al. 2004
Lee et al. 2005
45
Halo sub-structure at z4
We are observing the structure within the
halo. Break observed at 10 arcsec Note 10
arcsec at z4 is about 350 kpc. See also
Hamana et al. 2004
Lee et al. 2005
46
The Halo Occupation Distribution at z4
Consistent with Hamana et al. 2004 and Bullock et
al. 2001
ltNggt(M/M1)a MgtMmin
1-s
2-s
Lee et al. 2005
47
The Halo Occupation Distribution at z0
a 0.89 /- 0.05 M1 (4.74 /- 0.50) x 1013
MO Mmin 6.10 x 1012 MO
From SDSS data Zehavi et al. 2004
48
Halos and Galaxies at z3-5
Halo substructure we observe an excess of
faint galaxies around bright ones. massive halos
contain more than one LBG Bright Centers
z_850lt24.0 Faint centers 24.0lt z_850
lt24.7 Satellites z_850 gt25.0
Lee et al. 2005
49
Halos and Galaxies at z3-5
Clustering scaling in good agreement with
hierarchical theory Implied halo mass in the
range 5x1010 1012 MO 1-s scatter between mass
and SFR smaller that 100
Giavalisco Dickinson 2001 Porciani Giavalisco
2002 Lee et al. 2004, in prep.
50
EROs, orUV-faint galaxies at z2-3
Galaxies selected from near-IR photometry
(J-K)gt2.3 A fraction would NOT be selected by
LBG criteria (UV selection) However, overlap
with LBG not quantified and likely significant
(see Adelberger et al. 2004). They appear in
general more evolved, I.e. more massive (larger
clustering), with larger stellar mass, more
metal rich, and more dust obscured) than LBGs.
Occurrence of AGN also seems higher. At z3
these galaxies have about 50 of the volume
density of LBGs (highly uncertaint). However
they possibly contribute about up to 100 of
the LBG stellar mass density, because they have
higher M/L ratios
Van Dokkum et al. 2004
51
EROs
Kslt 22, R-Ksgt3.35
Moustakas et al. 2004
52
EROs

• ACS resolution is crucial to
• understand the nature of EROs
• Broad-band SED or statistical
• morphology cannot discriminate
• Evidence of massive galaxies at z1.2-1.5

Moustakas et al. 2004
53
HUDF/GOODS EROs
Yan et al. 2004
54
HUDF/GOODS EROs
Uses HUDF plus GOODS-SST data SED fitting
disfavour very dust obscured, star-forming
galaxies SED better reproduced by a
two-component composite populations an old,
evolved one, plus a low-intensity star-forming
one. Stellar mass relatively large 1010 1011
MO Evidence that similar objects exist at z7
(Mobasher et al. 2005)
Yan et al. 2004
55
LBGs at z5 and 6
Evidence of large stellar mass at z5, 6
Yan et al. 2005
56
LBGs at z5 and 6
Evidence of large stellar mass at z5, 6
Yan et al. 2005
57
An evolved, massive galaxy at z7?
HUDF GOODS-SST
Mobasher et al. 2005, submitted to Nature
58
NIR-selected galaxies
NIR selected galaxies with Klt20 VLT FORS
spectra SED fits show Mstar gt1011 MO Claims
that NIR selection yields more massive galaxies
than UV selection
Daddi et al. 2004
59
Different populations?
Near-IR selection picks up the high-end of the
distribution of masses (total and stellar)
Adelberger et al. 2004
60
Galaxies at z1-0
Cosmic variance
Todays stellar mass density
Evolution of the integrated mass density, Mgt1011
MO GOODS data Little evolution in the stellar
mass density from z1 to today Note that at z1
spirals dominated stellar mass density the
opposite at z0 morphology transformation
Bundy, Ellis Conselice 2005
61
Ravindranath et al. 2003

• Sersic indices nlt2
• Rest-frame MB lt-19.5
• Photometric redshifts

62
Disk galaxy evolution from GOODSRavindranath et
al. 2003
Number-densities are relatively constant to z1

• Tendency for smaller sizes at z1 (30 smaller)

63
The evolutionary link?
The expected evolution of clustering
(correlation length) suggests what the high
redshift galaxies might evolve into at later
epochs.
Giavalisco, 2002 ARAA
Adelberger et al. 2004
64
Summary

• GOODS exploring fundamental issues of cosmic
  origins
• Large-scale star formation in place at less than
  7 of the cosmic time
• SF galaxies observed to at least up z7
• Massive galaxy started very early in the cosmic
  evolution
• Cosmic star formation (as traced by UV light)
  varies mildly at 3ltzlt6
• Universe is as prolific a star former at z6 as
  it is at z3, after triplicating age
• Unclear proportion of obscured and evolved
  galaxies
• Obscured SF might contribute up to 100 of
  stellar mass density and star formation (2x)
• SF galaxies seem already diversified at z4.
  Evolved galaxies up to z7?
• Morphology mix includes spheroids, disks 14-25
  mergers at z1.4-5
• Direct evidence of growth of stellar mass from
  z4 to z1.
• Galaxies get smaller at zgt1 size evolution
  consistent with hierarchical growth
• Massive galaxies in place at z1 some galaxies
  are massive at z2-3
• Spatial clustering key to study relationship of
  star formation and dark matter
• Evidence of halo sub-structure at z4. Transition
  at r1 Mpc Mmin109 MO
• Spatial clustering depends on UV luminosity,
  decreases for fainter galaxies
• More massive halos host more star formation
  scaling consistent with CDM spectrum
• Implies relatively large total masses 5x1010
  1012 MO