star formation downsizing: testing the role of mergers and agn

Star Formation Downsizing:Testing the Role of Mergers and AGN

Kevin Bundy(University of Toronto)

Richard Ellis (Caltech), Tommaso Treu (UCSB),

Antonis Georgakakis, Paul Nandra, Elise Laird (IC)

DEEP2 Team at UC Berkeley & Santa Cruz

UC BerkeleyJuly, 2007

Outline

• Introduction and MotivationA Ride on the Downsizing Bandwagon.

• Observations: Characterizing DownsizingThe quenching of star formation, the rise of early-

types.

• Are Major Mergers Enough?

• The Role of AGN Activity

Introduction and Motivation: A Ride on the Downsizing Bandwagon

Bimodal Galaxy Distribution

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Bell et al. 2003

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.QuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.Star

formingBlue

Late typeYoung

PassiveRed

Early typeOld

• Hubble Sequence - morphology shows dynamically distinct populations

• Gas content/integrated colors - different ages and star formation histories

Kauffmann et al. 2003

Old

Young

Early-type

Late-type

z = 0

Origin?

Evolution?

Bimodality & MassQuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

… Dark Matter …

Hierarchical CDM Assembly

z=18

z=6

z=1.4

z=0

Downsizing: How to Build a Bandwagon

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, see Heavens et al. 2004)

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, see Heavens et al. 2004)

• Evolution in M/L from the Fundamental Plane

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, but see Heavens et al. 2004)

• Evolution in M/L from the Fundamental Plane

Downsizing: How to Build a Bandwagon

Treu et al. 2005

HigherSFR

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, see Heavens et al. 2004)

• Evolution in M/L from the Fundamental Plane

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, see Heavens et al. 2004)

• Evolution in M/L from the Fundamental Plane

• Surveys: Cowie et al. 1996, Brinchmann & Ellis 2000, Bell et al. 2005 COMBO17, Bauer et al. 2005, Juneau et al. 2005, Borsch et al. 2006, Brown et al. 2006, …

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, see Heavens et al. 2004)

• Evolution in M/L from the Fundamental Plane

• Surveys: Cowie et al. 1996, Brinchmann & Ellis 2000, Bell et al. 2005 COMBO17, Bauer et al. 2005, Juneau et al. 2005, Borsch et al. 2006, Brown et al. 2006, …

Downsizing: How to Build a Bandwagon

Juneau et al. 2005

Downsizing: How to Build a Bandwagon

1. Start with a broad prediction from confident theorists.

2. Find observations that (you think) prove them wrong.

• Existence of massive, evolved galaxies at z~2 (e.g. FIRES)

• The most massive galaxies at z=0 have the oldest stellar pops (many examples, see Heavens et al. 2004)

• Evolution in M/L from the Fundamental Plane

• Surveys: Cowie et al. 1996, Brinchmann & Ellis 2000, Bell et al. 2005 COMBO17, Bauer et al. 2005, Juneau et al. 2005, Borsch et al. 2006, Brown et al. 2006, …

3. Give it a catchy name.

Downsizing: Should We Be Worried?

Defining Downsizing

1. Archeological Downsizing

• Age vs. mass at z=0

2. Assembly Downsizing

• Assembly rate vs. mass

3. Downsizing of Star Formation

• SF/type vs. mass and redshift

3. Downsizing of Star FormationSF/type vs. mass and redshift

The sites of star formation appear to shift from including high-mass galaxies at early epochs (z~1-2) to only lower-mass galaxies at later epochs.

3. Downsizing of Star FormationSF/type vs. mass and redshift

The sites of star formation appear to shift from including high-mass galaxies at early epochs (z~1-2) to only lower-mass galaxies at later epochs.

How do we reconcile downsizing in the context of the hierarchical CDM paradigm?

Downsizing through Gastrophysics

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Mergers

Cluster physics

AGN Feedback

Starbursts/SN

Downsizing through Gastrophysics

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Mergers

Cluster physics

AGN Feedback

Starbursts/SN

How do we understand mass and redshift dependence?

Observations: Characterizing Downsizing

The Palomar K-band + DEEP2 Redshift Survey

• DEEP2: 40,000 spec-z’s from DEIMOS on Keck II

80 Keck nights, z<1.5 over 3 deg2, R < 24.1

Spread over 4 fields, including the EGS

• Palomar K-band: 65 nights with WIRC on 200 inch

1.5 deg2 to K=20, 0.2 deg2 to K=21

• Combined: 12,000 redshifts with K-band detections

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.

Field 22 16:52 +34:00 Field 32 23:00 +00:00 Field 42 2:30 +00:00

EGS 14:16 +52:00

Key Physical Properties

1. Stellar Mass

• Palomar K-band, multi-band SED fitting

2. SFR Indicator (bimodality)

• (U-B) Restframe Color, C. Willmer

• Morphology (from GOODS, Bundy et al. 2005)

3. Environmental Density

• 3rd nearest neighbor, M. Cooper

Results:Galaxy

Stellar Mass Function

Mass

Nu

mb

er

Den

sity

• Little total evolution

Results:Galaxy

Stellar Mass Function

Partitioned by restframe (U-B) color into blue

(active) and red (quiescent) populations.

Mass

• Little total evolution

• Transformation to early-types

Nu

mb

er

Den

sity

Results:Galaxy

Stellar Mass Function

Partitioned by restframe (U-B) color into blue

(active) and red (quiescent) populations.

Mass

• Little total evolution

• Transformation to early-types

• Evolving transition mass, Mtr

Nu

mb

er

Den

sity

Red Fraction Growth Function

RedFraction

Highest M*

Lowest M*

Cosmic Age (Gyr)

Red Fraction Growth Function

RedFraction

Highest M*

Lowest M*

Cosmic Age (Gyr)

8% Gyr -1

9% Gyr -1

11% Gyr -1

16% Gyr -1

25% Gyr -1

Is quenching and downsizing a result of environment?

Extreme Environments

Mass

Low Density

Extreme Environments

Mass

Low Density

Extreme Environments

Mass

Low/High Density

Extreme Environments

Mass

• Moderate dependence on density

• Downsizing accelerated in dense regions

Low/High Density

What Have We Learned?

• Downsizing results from the quenching of star formation.

• Quenching is accelerated in dense environments but is apparent in all environments.

• We are therefore looking for internal (non-environmental) processes…

A Popular Picture

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Mergers

Cluster physics

AGN Feedback

Starbursts/SN

A Popular Picture

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Mergers

Cluster physics

AGN Feedback

Starbursts/SN

A Popular Picture

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Mergers

Cluster physics

AGN Feedback

Starbursts/SN

• Initial quenching of star formation (SF downsizing) and morphological transformation triggered by mergers.

• Mergers also fuel black holes… may initiate radio mode AGN feedback.

QuickTime™ and a decompressor

are needed to see this picture.

Mergers & Feedback

Springel, Hernquist, Hopkins

• We need to test the merger hypothesis.

• We need to test the AGN hypothesis.

• Connection to CDM halo assembly?

Is the picture correct?

Testing the Current Picture: Are Major Mergers Enough?

Merge!

One Approach: Dynamical Mass

125 GOODS-N Spheroidals, 8 hr Keck spectra, IR Masses(Treu et al. 2005, Bundy et al. 2005)

Do New Spheroidals Form via Major Merging?

(Astro-ph arXiv:0705:1007)

What’s the strategy?

Use dynamics to estimate Mvirial of halos hosting spheroidals.

Compare to expected assembly history of dark matter halos.

Estimating Spheroidal Halo Mass

• Assume simple isothermal+NFW profile motivated by lensing results.

• Normalization set by 2

Calibrate to M* in two z-bins and apply to the full GOODS spheroidal sample.

Gavazzi et al. 2007

Vir

ial

Mass

Stellar Mass

Spheroidal Halo Mass Function

Spheroidal Halo Mass Function

Spheroidal Halo Mass Function

Spheroidal Halo Mass Function

SDSS

Spheroidal Halo Mass Function

SDSS

New Spheroidals

Spheroidal Halo Mass Function

New Spheroidals

SDSS

Spheroidal Halo Mass Function

RecentHalo Mergers

Millennium Simulation

New Spheroidals

SDSS

What this tells us

• Apparently not enough major mergers to support rising abundance of spheroidals… !

• Other mechanisms involved: secular bulge growth, disk fading, role of S0 galaxies. (see Bower; DeLucia; Lotz)

• What about AGN/starburst feedback and M- relation?

Testing the Current Picture: The Role of AGN Activity

The Appeal of AGN• Widely recognized presence of SM black holes and

the M•- relation.

• Large available energy without need for SF.

• Cluster cooling flows.

• AGN “Downsizing” in Luminosity Function (e.g., Barger et al. 2005)

• Observations beginning to link AGN hosts with red early-types and post-starbursts. (Kauffmann et al. 2004, Grogin et al. 2005, Nandra et al. 2007, Pierce et al. 2007, Yan et al. 2006, Goto et al. 2006)

There are (at least) 2 ideas of how AGN feedback works.

Merger-Driven, Explosive Feedback

Springel, Hernquist, Hopkins, Robertson, Di Matteo

• Importance of merging... morphological transformation.

• What sets the mass dependence?

• What prevents gas from cooling and forming stars later?

• Can starbursts do the same thing? How would you tell?

Radio Mode AGN Feedback

• Halo gas pre-heated… how?

• Low AGN luminosity, but efficient coupling to hot gas.

• Now implemented in many semi-analytic models. (Granato et al. 2004, Croton et al. 2006, Bower et al. 2006, Scannapieco et al. 2005)

Key Questions

• Is there an observational link between evolution in AGN activity and star formation downsizing? Need M*

• Do AGNs cause quenching?

Chandra X-ray Observations from AEGIS

• 200 ks, covering the EGS, 0.5-10 keV, 1300 sources

• 170 X-ray sources with redshifts and K-band masses

• Primarily selects obscured AGN hosts, some QSOs

• ~50% more could be X-ray absorbed.

AGN Host Mass Functions

AGNHosts

AGN Host Mass Functions

Linking Quenching and AGN

QuenchingRate

Linking Quenching and AGN

Linking Quenching and AGN

AGN TriggerRate

AssumingtAGN = 1 Gyr = AGN /tAGN

Linking Quenching and AGN

Set QuenchingRate equal to Trigger

Rate

Linking Quenching and AGN

Set QuenchingRate equal to Trigger

Rate

HopkinsPrediction

(2005)

Evidence for a Link

Nandra et al. 2006

• If tAGN ~ Gyr, X-ray luminous AGN are likely to be associated with quenching.

• AGN hosts are mostly red, early-type, possibly post-starburst. (e.g., Yan et al. 2006, Nandra et al. 2007, Pierce et al. 2007, Grogin et al. 2005, Kauffmann et al. 2004)

Evidence for a Link

• If tAGN ~ Gyr, X-ray luminous AGN are likely to be associated with quenching.

• AGN hosts are mostly red, early-type, possibly post-starburst. (e.g., Yan et al. 2006, Nandra et al. 2007, Pierce et al. 2007, Grogin et al. 2005, Kauffmann et al. 2004)

• But estimated accretion rates show a large dispersion in both host mass and color, suggesting AGNs do not cause quenching. Refueling?

Summary and Conclusions

• Quenching of star formation leads to downsizing which is apparent in all environments, suggesting non-environmental mechanisms are important.

• Major mergers, however, may not be enough to explain the rising abundance of spheroidals.

• New evidence links mass dependent AGN activity with quenching, but argues against the notion that explosive AGN feedback causes quenching to occur.