HWR

1
II. The Formation and Evolution of Massive
GalaxiesorShattering the Monolith
Hans-Walter RixMax-Planck-Institute for
AstronomyHeidelberg
2
Overview

• Massive Galaxies today
• Massive galaxies since z1 not much happenin?
• Massive galaxies at high redshifts when did
  they form?
• Galaxy formation models vs. Observations

3
In the present-day universeStrateva..Baldry..Bant
on..Hogg..Kauffmann..SDSS 2001-2005

• Color-distribution of galaxies in bi-modal.
• Almost all (r-band) luminous galaxies are red.
• Luminous/massive galaxies are red because they
  dont have young stars.
• Locus of color-luminosity distribution of
  galaxies depends little on environment.
• Population density of galaxies within this locus
  depends strongly on environment density.

4
red
color
blue
bright
faint
luminosity
SDSS
Strateva et al 2001 Baldry et al. 2003
BlantonHogg
5
massive galaxies had most of their activity in
the past low mass galaxies had more extended SF
histories
present/past averaged star formation rate
Brinchmann et al. 2004
6
Color-Luminosity Distribution of Galaxies as a
Function of Environment
luminous
blue
7
Clues from Abundance Ratios a/FeThomas et al
2002,2005

• The most massive galaxies seem to have the most
  rapid enrichment!

8
In the present-day universeStrateva..Baldry..Bant
on..Hogg..Kauffmann..SDSS 2001-2005

• Color-distribution of galaxies in bi-modal.
• Almost all (r-band) luminous galaxies are red.
• Luminous/massive galaxies are red because they
  dont have young stars.
• Locus of color-luminosity distribution of
  galaxies depends little on environment.
• Population density of galaxies within this locus
  depends strongly on environment density.

9
Massive Galaxies since z1

• By now massive galaxies are red and dead. When
  did they stop forming stars? Why did they
  stop forming stars?
• Even after star-formation has ceased, the
  assembly can continue through merging. How much
  of this is actually happening?

10
Evolution of the Galaxy PopulationWolf et al
2003 Bell et al 2004

• Building on Autofib and CFRS Surveys (c.a. 1996)
• LV and U-V (rest-frame) from 3 disjoint COMBO-17
  fields
• 32.000 galaxies to z1.2
• Galaxy color distribution has been bi-modal
• (as in the local universe e.g. Strateva et al.
  2001)

11
The (Red) Color-Magnitude RelationBell,Rix, et
al 2004

• Color ridge of red sequence reddens towards the
  present epoch, consistent with passive fading of
  old stellar pops.
• As in clusters (van Dokkum, Franx), but here
  volume averaged!

12
COMBO-17 Early-type (red sequence)
galaxiesBell,Wolf,Rix, et al 2004

• But..B-band luminosity density of red-sequence
  galaxies stays constant with z
• stellar mass in red-sequence must increase (by a
  factor of 2)

13
COMBO-17 Early-type Galaxies

• How does the observed evolution agree with
  hierarchical models of galaxy formation?
• Cole et al. 2000 fiducial model using same color
  criteria as we use (0.25 mag bluewards of the CMR
  in rest U-V)

14
What galaxies are available to turn-off
star-formation and fade into the red
sequence?Borch,Meisenheimer,Rix et al 2005

• Use SED to estimate the M/L
• Red galaxies dominated the massive end to at
  least z1

Z0.7
15
Evolution of the Galaxy (Stellar) Mass
FunctionBorch, Meisenheimer, Rix et al 2005
16
What are these red galaxies at z1

• Red and dead? Or Red and Dusty?
• Decide by
• Morphologies
• Thermal-IR flux

GEMS (Galaxy Evolution from Morphologies and
SED) Rix et al 2004 Map 30x30 COMBO-17 field
(E-CDFS) with HST 170 HDFs in area Encompassing
UDF,GOODS HSTs largest color image
17
GEMS Rix et al 2004 around CDFS
18
(No Transcript)
19

• Boriss 3-fold image

20
Red Dusty or Dead?
21
Red/Dusty or Red/Dead ? (Bell et al 2003)

• Pick thin redshift slice (0.7) to eliminate
• differential band-shifting
• differential (1z)4 dimming
• Visual classification vs Sersic index

85 of the red sequence at z0.7 are early
types, i.e. red and dead

22
Thermal-IR (24mm) estimated star-formation in
massive (gt5x1010Mo) galaxies at z0.7 and
z0Bell, Rix, Papovich, et al. 2005
Many non-detections here!
UVIR based star-formation rate in massive
galaxies builds up the population only by a few
percent since z0.7!
23
Is/was massiveno new stars expected? No!
Semi-analytic models (Somerville et al 2005)
Observations
Red ?
Luminosity ?
Luminosity ?
24
BH growth and self-regulation

• Massive galaxies have massive central black holes
• thermal coupling of AGN energy with ISM
• BH growth self-regulated (produces MBH-M
  relation)
• AGN drives a wind that expels the gas from the
  galaxy

Di Matteo, Springel Hernquist 2005
25
Semi-analytic model with AGN feed-backSomerville
et al 2005

• cold gas ejected (and never re-accreted) if
  MBHgtMcrit
• standard merging prescription

still have a cooling flow problem!
Observations
Models
26
Did nothing happen to massive galaxies
(Mgt5x1010M0) since z1?

• Galaxies can grow in mass by
• forming new stars
• coalescence (merging) of pre-existing bits
• Can we estimate a merger rate for massive
  galaxies?
• Steps
• Devise quantifiable definition of ongoing
  merger
• Estimate a timescale for this phase incidence ?
  rate
• Quantify how the ongoing merger changes the
  probability of entering sampe e.g. merger ?
  star-formation? lumosity boost

27
Dry Mergers at zlt0.7Bell, Naab, Somerville,
Rix.. 2005

• Simulations
• Naab Burkert (2005)
• Each progenitor is a merger remnant
• Mass ratios 11 14
• Recognizable
• lt5 kpc separation
• Asymmetries / broad tails
• For 250 Myr independent of
• merger details (mass ratio, pericenter, etc)

28
Actually observed dry mergers at zlt0.7 in GEMS
Using 250 Myr timescale 0.9/-0.4 mergers /
massive red galaxy since z0.7 with mass ratios
41 --- 11
29
Comparison with semi-analytic models
early type red sequence
major mergers
GEMS dry mergers
dry major mergers (red sequence mergers)
(all MVlt-20.5)
Mbulge quenched model
Bell et al 2005
30
So, when did the stars in massive galaxies form?

• Lets back up first
• red sequence accounts for most stars in
  Mgt5x1010 at z0
• 30 of all stars (now) are in galaxies with
  Mgt5x1010
• Not many star formed since z1 in galaxies with
  Mgt5x1010

Present-day stellar mass function Bell et al. 2003
Early type Late type Total

• Only 10 of all current stellar mass existed at
  z3
• ? characteristic formation epoch z2-3 ?

31
Is there ared sequence at z2 ?(from FIRES)
4
2
0
-2
4
2
(M2200 V)rest
0
-2
4
2
0
-2
32
Distant Red Galaxies (from FIRES) (Franx et al
2003, van Dokkum et al 2004, Foerster-Schreiber
et al 2005, Labbe et al 2005)

• Typical values
• Selected in rest-frame optical
• M5x1010 3x1011 Mo
• SFR 50-150 Mo/yr
• For SFR e-t/t ? tfit500Myr
• ? SFRxt1011Mo
• Av2 mag
• Density 1/3 of Ly break

33
Census of massive galaxies at z2-3

• At this epoch, the majority of massive galaxies
  are seen to form stars vigorously M/ltSFRgt lt
  tHubble
• Extreme SFR phases or masses sub-mm galaxies
• Lower-mass/SFR end Ly-break galaxies
• Bulk(?) distant red galaxies
• SED-modelling indicates that building 1010.5Mo
  within few 100 Myrs may be a common event

34
Can models explain the abundance patterns?
Thomas and Kauffmann 1999
35
Semi-analytical, hierarchical model with SF
truncation in massive galaxies (to get the red
sequence right)
redshift
The more massive elliptical galaxies have a
shorter formation timescale!
De Lucia et al., in preparation
Lookback time (Gyr)
36
Massive galaxies (Mgt5x1010M0)

• Most massive galaxies are done forming stars now.
• This has largely been the case since z1-1.5
• This is a volume averaged statement!
• Stopping late star-formation, requires new
  ingredient.
• Red sequence being fed at lower masses
• Major merger rate for massive galaxy 1 since z1
• Massive and dead becomes rare at z2
• Build-up of massive galaxies seen directly
• Both Mgt5x1010M0 and M/ltSFRgt lt tHubble commonly
  seen.
• Whether build-up at the most massive end is rapid
  enough to explain a/Fe patterns, remains to be
  seen.

37
z1.1
38
z0.95
39
z0.8
40
z0.65
41
z0.55
42
z0.4
43
z0.3
44
z0.15
45
(No Transcript)
46
Evolution of the Galaxy Population
47
(No Transcript)