Coevolution of black holes and galaxies at high redshift

David M Alexander (Durham)

The growth of black holes

Black Hole

Accretion Disk

Action: Star formation

Acti

on

: A

GN

acti

vit

y

Mbh/Msph~10-3

Some key issues in observational cosmology

The black-hole-host connection

Th

e ro

le o

f en

vir

on

me

nt?

Difficulty in Constructing a

Complete AGN Census

• Most AGNs are hidden at optical wavelengths by dust and gas

• Host-galaxy dilution for lower-lum AGNs

SDSS optical quasar selection identifies a small

fraction (~1-10%) of all AGNs

X-ray surveys provide a more complete census of AGN

activityC

entral en

gin

e

Obscuring “torus”

X-r

ays:

pen

etr

ate

h

igh

colu

mn

d

en

sit

ies

5 ks Chandra BootesMurray et al. (2005)

• Detection of even low-luminosity AGN out to high redshift

• Need deeper spectroscopy (~70% complete) although photozs now getting to good quality (e.g., Luo et al. 2010)

X-ray Surveys: great steps towards a complete census of AGN activity

2 Ms Chandra deep fieldsBrandt et al. (2001), Alexander et al. (2003);Giacconi et al. (2002); Luo et al. (2008)

Just CDF-N sources

High (quasar) luminosities

Moderate (Seyfert) luminosities

Low luminosities

(1) Below detection limit: stacked X-ray data of IR-bright z~2 galaxies - hidden AGNs

Particularly when allied with infrared identification of the most heavily obscured

AGNs

Daddi et al. (2007); see also Fiore et al. (2008, 2009), Donley et al. (2008), and others

Alexander et al. (2008)

(2) Spectroscopic identification of individual X-ray undetected luminous

AGNs

Steidel et al. (2002); Alexander et al. (2008)

IR spectra and SEDs

Optical spectroscopy

Growth of black holes

Note: the definition of high redshift here is “only” z~1-4 - the identification of large numbers of typical z>6 AGNs is challenging: only ~1 of z>6 AGN probably per deep X-ray field and none reliably identified to date (see Gilli and Brandt talks)

Key result from X-ray surveys: AGN “cosmic downsizing”

Ueda et al. (2003)

Fiore et al. (2003)

Hasinger et al. (2005)

Luminosity-dependent density evolution: high-luminosity AGNs (i.e., quasars) peaked at higher redshifts than typical AGNs

Also Cowie et al. (2003); Barger et al. (2005); La Franca et al. (2005); Aird et al. (2010) amongst others

Grew more rapidly in past

Growing more rapidly now

z~0: Heckman et al. (2004)

Downsizing in “active” black-hole masses?

In general this appears to be true - massive black holes (~108 solar masses) were growing more rapidly at z~1 than in the present day, where smaller

black holes (<107 solar masses) were most active (Heckman et al. 2004; Goulding et al. 2010)

However, current observational constraints do not yet allow robust conclusions

z~1: Babic et al. (2007)

Gro

wing

mor

e ra

pidl

y no

w

Gre

w mor

e ra

pidl

y in

pas

t

See also Ballo et al. (2007); Alonso-Herrero et al. (2008)

Constraints at higher redshifts?

Marconi et al. (2004)

Model result (Soltan type approach)

Higher-redshift constraints are key since many models predict rapid black-hole growth at high redshift

But these are challenging due to lack of spec-zs of a complete sample and black-hole-host galaxy mass relationship uncertainties: Brusa et al. (2009)

similar z~2 constraints to Alexander et al. (2008) above

Alexander et al. (2008)

Limited constraints for highly selected sample

z~2 ULIRGs vs SDSS

Tracing the black-hole-host mass relationship

Black-hole masses estimated with virial technique (see Vestergaard

talk)

Host-galaxy masses estimated from a variety of techniques:

Host vel disp (em and abs lines), CO line widths, absolute

magnitudes, stellar masses, and SED fitting

But general concensus is for modest evolution in MBH-MGAL ratio with

redshift - with relationship ~2-4x higher at z~2 and perhaps higher at z~2-6: factor ~4 based on Merloni et

al. (2010)

Evolution perhaps only significant at highest masses (>3x108 solar

masses): see Di Matteo talk (also Merloni et al. 2010)

Virial black-hole mass estimator: MBH=G-1 RBLR V2

BLR

Challenging to measure *both* black hole and host mass without

significant uncertainties (see Wang talk)

See also McLure et al. (2006); Peng et al. (2006a,b); Shields et al. (2006); Woo et al. (2006, 2008); Salviandar et al. (2007); Treu et al. (2007); Jahke et al. (2009); De Carli et al. (2010)

Merloni et al. (2010) result in COSMOS

Similar result for rapidly evolving z~2 SMGs

Estimated MBH using virial mass estimator for the few z~2 SMGs with

broad lines

Estimated MGAL for the majority which are host-galaxy dominated

Not significant difference in average MBH-MGALratio for various z~2

populations

But require constraints of more typical z~2-6 AGNs - want to test if there is a

black-hole mass dependence: need better spectroscopy, imaging, and

larger-area deep X-ray surveys

Alexander et al. (2008); Hainline et al. (2010)

Lamastra et al. (2010)

SAM results, including feedback

Tentative evidence for feedback inducing blow out?

Broad (800 km/s) high-velocity (200-500 km/s) [OIII] gas

Alexander et al. (2010); see Nesvadba et al. (2006,2007,2008) for results on rarer z~2 radio-loud AGNs

Di Matteo et al. (2005) simulation

~4-8 kpc extent of broad [OIII gas]Collapsed IFU spectrum of z~2.07 SMG

What role does environment play?

A key laboratory of black-hole growth mechanisms:

a distant protocluster?

400ks Chandra exposure of SSA22

z=3.09 SSA22 protocluster

Predicted to become a massive Coma-like

cluster by the present day

The galaxy density is ~6x higher than the field already at z~3.09

• The AGN activity per galaxy is larger in the protocluster compared to the field by a factor of 6.1+10.3 (enhanced at the 95% confidence level) for AGNs with LX > 3 × 1043 ergs s−1.

• Fraction of LAEs hosting AGNs appears to be positively correlated with the local LAE density (96% confidence level).

Enhanced black-hole growth compared to the field

LBGs

LAEs

Lehmer et al. (2009a)

LX > 3 × 1043 ergs s−1

Lehmer et al. (2009b)

More massive black holes active at earlier times?

Implication: the characteristic X-ray luminosity and “active” black-hole mass appears to be a function of environment as well as redshift

Lehmer et al. (2009a)

• If the AGN fraction is larger in the protocluster simply due to the presence of more massive SMBHs, then an average protocluster AGN would be more luminous than an average field AGN by the same factor

• For this to be the case, the SMBHs would have to be ≈3−10 times more massive in the protocluster than the field: likely 108-109 solar masses rather than 107-108 solar masses

Good agreement with that found for a z=2.3 protocluster (Digby-North et al. 2010)

Host galaxies appear more massive in protocluster

And AGN fraction declines to lower redshifts

Require more constraints for more protoclusters, particularly at high redshift

Need deep X-ray observations of other overdense regionsNeed wider area X-ray surveys to cover full range of environments

Martini et al. (2009)

AGN fraction in massive galaxy clusters

z~3.09 protocluster AGN fraction• Significant drop (1-2 orders of magnitude) in AGN fraction for similarly overdense regions at z<1

• AGN activity has been “switched off” in galaxy clusters/protoclusters since z~3 to z<1

• Need to trace this out from z~2-8 with X-ray observations of more protoclusters/overdense regions

Has heating of the ICM tentatively started?

~17% of extended Ly emitters host luminous AGN in the z~3.09 protocluster; AGNs are luminous enough to power the 10-100 kpc

extended Ly emission

QuickTime™ and aH.264 decompressor

are needed to see this picture.

Geach et al. (2009) Ly images of some protocluster AGNs

Potential heating mechanisms

AG

N p

ow

er

Star-formation power NASA press conference movie

• AGN cosmic downsizing possibly due to increase in active black hole with redshift

• typical active black holes at z~1 are ~108 solar masses; more work required to obtain reliable results at higher redshifts

• Evolution in black-hole-host mass relationship quite modest out to z~2 - may be stronger out to higher redshift but current samples are limited

• possible black-hole mass dependence - need constraints for typical black holes• possible evidence for z~2 feedback inducing blow out

• Environmental dependencies on black-hole growth: AGN fraction increases in z~2-3 protoclusters than compared to field - probably due to more massive “active” black holes

• AGN fraction significantly decreases compared to field at z<1 - possibly due to gas depletion or heating

• AGN activity in z~3 protocluster luminous enough to power 10-100 kpc Ly halos - early evidence for ICM heating?

SummarySummary