Download Chapman

High Redshift Galaxies
Scott Chapman (Cambridge, IoA)
Aug22, 2007
Overview and Motivation
0.
Introductory remarks about galaxy formation
& evolution
I.
Direct observations of galaxy progenitors
forming at high redshift (z ≥ 1.5)
II. Protoclusters at high redshift (z ≥1.5)
Context: Hierarchical Galaxy Formation
(How/when are the galaxy components assembled?)
Big Bang … Cosmic Microwave Background …
… Galaxy Formation and Evolution … Fossil Records today!
Submm/Radio is a superb probe of obscured forming galaxies; well defined SEDs;
unimpeded by dust obscuration; “bolometric selection”
Andromeda (M31) is ideal laboratory to study L* galaxy components.
Galaxy Evolution - Gravity
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
GALAXY FORMATION and EVOLUTION: with baryons …
Dark matter
Bright galaxies
Low SFR
High SFR
N-body + semi-analytic
-gas cooling
-star formation
-SNe feedback
-galaxy mergers
within halos
•Success reproducing
observed parameters
at z=0.
•High-z objects of various
classes continue to be
problematic -- SMGs, DRGs,
(e.g., Sommerville et al. 2000;
Baugh et al. 2005)
(Kauffmann et al. 1998)
Progenitors of Massive Galaxies
•Until the mid-1990’s the only z>2 objects known were QSOs, radio galaxies,
and QS0 absorbers (DLA/LLS)
• How can we go about isolating more normal galaxies during the epoch of
star/galaxy formation?
• The study of high-redshift (let’s say z>1.5) galaxies has exploded in the last
~10 years, with multiple techniques for isolating high redshift galaxies,
making use of multi-wavelength data spanning from the radio to the X-ray
• As opposed to traditional magnitude-limited surveys, down to specific flux
limit, new results utilize several complementary selection techniques for
finding high-z galaxies, selecting overlapping yet complementary populations
--> must determine how they overlap/complement each other to describe
entire galaxy population at a given epoch.
“Foreground (z<2) Galaxies”
NB: IAB=25 
1 photon/year/cm2/A
From: LeFevre,
Vettolani et al 2003
 “Brute force” spectroscopy is inefficient for z>2!
The Laboratory for studying high-z galaxies
UKIRT
OVRO
CSO JCMT Subaru Keck-I/II IRTF
VLA
IRAM
CFHT
MERLIN
HST
Palomar
5-m
Chandra
Spitzer
Gemini
Ground-based vs. space-based?
.........................opt..NIR.................radio window.......
0.1nm 1
10 100 1mm 10 100
1cm 10
wavelength
1m
10 100m
Some sub-mm windows
from good sites
NB: in addition to transmissivity, the emissivity of the Earth’s
atmosphere is a big problem: sky at 2mm is 10,000 brighter
than at 0.5mm
(Spitzer Space
Telescope)
10-m mirror
0.85-m mirror
2-m astronomer
2.4-m mirror
(Hubble Space
Telescope)
(Keck Telescope)
Local Galaxies - e.g. Luminous
Elliptical Galaxies
… the most massive galaxies in the local Universe
•Pressure-supported, metal-rich stellar systems.
“Simple” galactic-scale stellar systems
•Extremely homogeneous and old (implies z>2 formation)
•Live in the highest density regions (clusters)
•Largest stellar mass of any galaxies
( up to 10 M*)
• Host the most massive Black Holes in
local Universe, >109Mo
HOW ARE THEY MADE ???
-observations of the high-z Universe
Formation Mechanisms
•
•
Wide range of proposed mechanisms for forming big galaxies:
– (pseudo-)monolithic collapse (Tmerge<TSF)
– major merger of two existing galaxies (Tmerge~TSF)
– an extended series of minor mergers (Tmerge>TSF)
Can we observationally distinguish between these scenarios?
Dust obscured, merging galaxies:
e.g.,The Antennae
• Distinct opt/UV and far-IR
luminosity
• 90% emitted at far-IR
ISO 15um
• Dust obscures UV;
absorbs and re-radiates
at longer wavelengths
(~100-200 microns)
ISOCAM
HST WFPC2
CSO/SHARC-2: 350um
The Extragalactic Background
?
FIRB
UV/Opt
• FIRB = opt/UV EBL -> half of the energy production (from SF or AGN)
over history of the Universe arises in highly obscured regions
• Much of the star formation in the Universe might be obscured:
resolve FIRB into constituent components
Components of FIRB
(to be studied directly and indirectly)
350
850
Dole et al. 2005
Submm is almost all high-z galaxies (80% z<4)
MidIR and FarIR is 80% z<1.7
Radio is extremely useful, BUT …
submm best (only?) way to probe the high-z components of FIRB
Studying the obscured activity
ground
Herschel
(2007)
Spitzer
SED of an ageing stellar
population of solar
metalicity with dust
f(24mm) vs Lbol
Papovich and Bell 2003
At “far-IR”
wavelengths, Spitzer
can only observe well at
24 mm …
what are the bolometric
corrections?
… they’re large!
And especially bad for
the most luminous,
dust-obscured
populations.
(leave aside for now)
Progenitors of Massive Galaxies
• Review of techniques
•(focus on submm/radio selection, UV, & nearIR)
• Some key questions and results
z>1.5 Submm Selection (~100 z’s)
(courtesy I. Smail)
(Smail 2005)
• Submm galaxies: dusty, most of luminosity comes out in submm waveband
• First detected by SCUBA/JCMT in 1997 at 850 mm (Smail+1997)
• Counts: ~1000/sq. degree at 5 mJy (limit of bright SCUBA survey);
a few 100 submm galaxies ID’d
• In principle, SCUBA sensitive to dusty galaxies to z>6 (negative K-correction)
… but we now know that almost no SCUBA galaxies are that distant!
Submm/Radio: bolometric selection of galaxies?
Identify SMGs:
(although Spitzer
proving very useful -- Pope+06)
-Known tight correlation with FIR
-Good spatial resolution
(MERLIN 0.3”, VLA 1.4”)
Quick Time™ and a
TIFF (LZW) decompressor
are needed to see this picture.
FarIR/radio correlation roughly
holds at z=2.5
(Kovacs et al. 2006)
Far-IR correlates with
synchrotron radio
(SNe: shock accelerates electrons)
1.4GHz
(Ivison et al. 2002)
SCUBA galaxies (SMGs)
High resolution RADIO and HST reveal large, merging galaxies.
Smail+1999,2004; Chapman+2003,2004)
greyscale
contours
JCMT/SCUBA
VLA
MERLIN
HST
HST
SCUBA galaxies (SMGs)
Redshift surveys have given us tremendous insight into a
hyper-luminous population of cool dust SB-dominated galaxies.
<z>=2.3
Chapman+03,05,
Pope+05
Keck 10m
Swinbank+04
Massive/merging
Neri+03, Greve+05
Genzel+02, Tacconi+05
IRAM
z>1.5 Submm Selection
(courtesy I. Smail)
(Chapman et al. 2005)
• Breakthrough: using radio (1.4 GHz) positions for optical spectroscopy
(ameliorates ambiguity from 15” SCUBA beam), 50% of sources have z’s
• Redshifts of ~100 SMGs enable study of physical properties
• Typical LIR=8x1012L and dust temperature Td~38K (confirmed by 350 mm
observations)
• Less than ~10% of submm galaxies are at z>>3
z>1.5 Submm Selection
(courtesy I. Smail)
(Chapman et al. 2005)
• IR luminosities correspond to sfr >1000 M/yr (Salpeter) if bulk of FIR is
powered by star formation. If SF lasts for ~108 yr, significant stellar mass
formed
• Much rarer than other samples, but higher inferred SFR, likely contribute
significantly to SFR density at high z
z>1.5 Submm Selection
(Chapman et al. 2003)
• Rest-frame UV spectra obtained with Keck/LRIS-B (like UV-selected samples)
• Spectra show features of star-formation and AGN (Ly, NV, CIV, SIV), restframe optical spectra sometimes [NII]/Hor broad H
• Raises question of AGN contribution to bolometric luminosity, Deep (Chandra
2 Msec image) X-ray emission indicates presence of Fe Kline, and absorbed
non-thermal continuum slope
• BUT: AGN appears not to be energetically important -> submm emission
dominated by star formation
z>1.5 Submm Selection
Note: galaxies w/
smaller star
formation rates
not detected yet
in CO
(Tacconi et al. 2006)
• Masses: stellar masses estimated from SED-fitting (how does AGN
affect stellar mass estimates?); dynamical masses estimated from H, COlinewidths; cold molecular gas masses estimated from CO line luminosities
• Dynamical masses w/in 10kpc, 2x1011 M(Swinbank et al. 2004, 2006),
molecular gas masses 5x1010 M (Tacconi et al. 2006)
z>1.5 Submm Selection
• Submm galaxies appear to be massive systems with
prodigious star-formation rates, may also be strongly clustered
(but uncertain because of small number of redshifts)
• Could be progenitors of QSOs, and massive galaxies at lower
redshift
•Interesting note: several years ago, a lot was made of the fact
that there was so little overlap between submm galaxies and
z~3 UV-selected galaxies. But now we know that most submm
galaxies are at z<3, and have similar redshift distribution to z~2
UV-selected galaxies. >~50% of submm galaxies have colours of
UV-selected galaxies (but bigger SFRs)
Photometric Pre-selection: UV
• ~50 objects/square arcmin
down to R=25. How do you pick
out the high-redshift galaxies?
• Lyman discontinuity at restframe 912 A gives z~3 galaxies
very distinctive observed UGR
colours
(Steidel et al. 1992, 1993, 1995,
1996, 2003)
z>1.5 Rest-UV Color Selection
• z~3 UGR Lyman Break criteria, adjusted
for z~2 (Adelberger et al. 2004)
• Spectroscopic follow-up with optimized UV-sensitive setup (Keck I/LRIS-B)
• ~1000 galaxies at z~3, >750 galaxies with spectroscopic redshifts at z=1.4-2.5,
in what was previously called the Redshift Desert
Measuring Redshifts: z~3
Meaure: Ly em/abs,
IS abs at z>2.5
• At z =1.4-2.5, these
features are in the near
UV … bonus: strong
rest-frame optical
emission lines have
shifted into the nearIR:
• formerly called THE
REDSHIFT DESERT
Redshift Desert
Low redshift
• Emission-line z
• [OII], [OIII],
Hb, H
• At z > 1.4,
[OII] moves past
9000 AA, while
Ly below 4000
AA at z<2.3: no
strong features in
the optical
SDSS galaxy at z=0.09
Redshift Desert
Low redshift
• Abs-line z
• 4000 AA break,
Ca H&K, Mg
• At z > 1.4,
4000AA break
moves past
9000 AA
SDSS galaxy at z=0.38
Keck/LRIS-B Efficiency
• LRIS-B +
400/3400 grism -->
40% efficiency from
3800-5000 AA
Unsmoothed
• Low night-sky
background in
near-UV (~3 mag
fainter than at 9000
AA)
(Steidel et al. 2004)
Measuring Redshifts: z~2
• Low- and highionization outflow
lines, Ly
• He II emission, CIII]
emission
• Fewer galaxies have
Ly emission (57%
have no Ly) than in
z~3 sample (cf. SMG!)
(Steidel et al. 2004)
Outflow Kinematics: z~3 vs. z~2
z~3
• Kinematic evidence for
large-scale outflows is a
generic feature of UV and
submm-selected galaxies at
z>2
z~2
•Signature of “feedback”,
perhaps fundamental to
understanding galaxy
formation
Unsmoothed
(Steidel et al. 2004)
z>1.5 Rest-UV Color Selection
• z~3 UGR criteria (Lyman Break), adjusted
for z~2 (Adelberger et al. 2004)
• Spectroscopic follow-up with optimized UV-sensitive setup (Keck I/LRIS-B)
• ~1000 galaxies at z~3, >750 galaxies with spectroscopic redshifts at z=1.4-2.5,
in what was previously called the Redshift Desert
Evolution of Clustering to z~1, 0
(Adelberger et al. 2004)
(Blain et al. 2004)
Follow evolution of DM halo clustering in simulation
• Matches early-type absorption line DEEP2 galaxies at z~1 (Coil et al. 2003)
& SDSS E’s at z=0.2
•Typical UV-selected galaxy at z=2-3 will evolve into an elliptical by z=0
Summary:
z>1.5 Rest-UV Colour Selection
(typical sfr~50-100M/yr)
(typical stellar mass ~few x 1010M, large range)
What does UV-selection mean in terms of physical properties?
Star-formation that’s only moderately (factor of few-100)
extinguished by dust, but a large range of stellar masses
(smaller range of baryonic masses), clustering implies that
these will correspond to early-type galaxies at z~0.
… z>1.5 Near-IR selection
• Extension of K20 survey
group (i.e., get z’s for
everything with K<20), use Bz, z-K color criteria to select
both star-forming galaxies and
passive galaxies at z>1.4
• Incomplete for fainter
objects with small Balmer
Breaks, weighted more
towards fairly massive objects
(Daddi et al. 2004)
• Significant overlap of
BzK/SF with UV-selected
samples
z>1.5 FIRES/J-K selection (~20 zs)
(Reddy et al. 2005)
(Franx et al. 2003)
• J-K>2.3 criteria meant to select massive evolved galaxies with significant
Balmer/4000 Å breaks at z>2; turns out selection also yields massive dusty
starbursts
•~25% appear to contain AGN (much higher than fraction of UV-selected
population)
• Only limited number of spectroscopic redshifts
z>1.5 Summary
• In addition to UV-selected, BzK, J-K, submm, there are
other techniques, such as the K-band/photo-z technique
of the Gemini Deep Deep Survey (GDDS), and new
Spitzer capabilities: IRAC (mass-selected), MIPS/24
micron (sfr-selected, analogous to SCUBA)
• Now that there are several groups using different
selection techniques to find galaxies at z~2, we need to
understand how the samples relate to each other (each
sample has certain benefits but is incomplete; e.g., UVselected sample has largest set of redshifts and spectra)
• Reddy et al. (2005) considered the overlap among
different samples, and contribution of each to the sfr
density at z~2-2.5
At the highest redshifts?
Hard stuff to study! (e.g., contamination)
Controversy at z~ 6?
Did the star formation rate change from z~ 6 to z~ 3? If so, did it
increase or decrease?
Each possibility (increase, decrease, stay the same) had
been claimed by various groups studying z~ 6 galaxies.
For z~ 6, Lyαis between I & z bands, so we are looking for I-drops
R. J. Bouwens, G.D. Illingworth, J. P. Blakeslee, M. Franx set
out to do the most careful analysis possible and thereby come
to some firm conclusions about such galaxies.
The Star Formation History of the Universe
• Millimeter/Radio galaxies form
most of the stars we see today
in big bursts
•2-5 billion years after the big
bang.
UV-selected
Radio/submm
(Chapman et al. 2005)
LESSONS:
•we still don’t even know
about all the galaxies in the
Universe!
•Local galaxies (M31) show
unexpected components!
•Simulations of how galaxies
form are not always very
predictive!
Key Questions
• What is the evolution in global sfr and stellar mass
density vs. z?
• What is the evolution in number density of galaxies as a
function of (stellar) mass and star-formation rate?
• What are the star-formation histories of galaxies
(burst/episodic, continuous), and how do they accumulate
their stellar mass?
• What are the origins of different morphological types?
• What is the chemical enrichment in galaxies vs. z, and
by how much do they enrich their surroundings (vs.
mass)?
• What is the effect of supernovae/AGN feedback on gas
in galaxies and the surrounding IGM?
• How do we make a continuous timeline of galaxies from
high redshift to z~0 (map one sample to another)?
Key Techniques/Goals
• New multi-wavelength technologies are helping us
address these questions, beyond ground-based optical
imaging and spectroscopy
• Wide-field near-IR imaging (stellar masses) and near-IR
spectroscopy (dynamical masses, sfr, chemical
abundances)
• Chandra X-ray observations (sfr and AGN)
• Spitzer/IRAC (stellar masses) and MIPS (dust
luminosity, sfr)
• HST ACS/NICMOS (morphologies)
• Full understanding of energetics and stellar and metal
content is a multi-wavelength endeavor
• Detailed comparison with numerical simulations and
semi-analytic models
Concluding Philosophical Comment
• Rich set of observations of galaxies in the early universe,
for statistical samples of galaxies selected with different
techniques (though it’s still a challenge to get spectra)
• There’s much more that I could have presented
• Field of z>1.5 galaxy evolution is completely different
since just 10 years ago, when UV-selection technique was
first implemented
• What’s happening in the next 10 years? (feedback,
connecting to samples at other redshift, building
mass/sfr-limited samples, disk galaxies)