evolution of supermassive black holes from deep & large area x-ray surveys marcella brusa (max...

Evolution of Supermassive Black Holes from deep &large area X-ray surveys

Evolution of Supermassive Black Holes from deep &large area X-ray surveys

Marcella Brusa (Max Planck Institut fuer

Extraterrestrische Physik)

AGN 8 - Torino, May 2008

Outline Introduction: AGN evolution

What we know: observations & models

Still missing (among others..): high-redshift population Compton Thick AGN census

Summary & future


Main actors in this game XMM/Chandra COSMOS teams CDFS/GOODS/MUSIC team ELAIS-S1 team HELLAS2XMM team

.. and in particular thanks to: A.Comastri, F.Fiore, R.Gilli, A.Bongiorno, N.Cappelluti, F.Civano, G.Hasinger, K. Iwasawa, V. Mainieri, A. Merloni, M. Mignoli, S.Puccetti, M.Salvato, C. Vignali, G.Zamorani, C.Feruglio, A.Grazian, F.LaFranca, F.Pozzi, P.Santini…


introduction AGN evolution

Once upon a time… (1960s-70s): QSO/AGN are triggered by accretion on SMBH QSO/AGN are 1% of galaxy population: transient phase or rare objects?

[45 years of observations]

Nowadays: SMBH can be revealed in almost 100% of local

galaxies AGN transient phase!


MAIN OBSERVATIONAL RESULT: The discovery of dead SMBH in

most local bulges & tight correlation between MBH and bulge properties

Co-evolution of galaxies and SMBH

Large scale Galaxy properties strongly depend on BH mass

Marconi & Hunt 2003Marconi & Hunt 2003

Ferrarese & Merritt 2000 Gebhardt et al. 2000

Marconi & Hunt 2003


AGN and galaxy co-evolutionAGN and galaxy co-evolution Early on Strong galaxy interactions; Mergers between gas rich

galaxies drive gas which fuel both SF and BH activity;

Violent starbursts episodes;

Heavily obscured BH growth

When galaxies coalesce

Accretion peaks; SMBH becomes optically

“visible” (QSO phase) as AGN winds blow out gas

Later times SF & BH accretion

quenched; Dead quasars (or slowly accreting BH) in red

galaxies (passive evolution)[see also Granato+2004, Di

Matteo, Springel & Hernquist 05, Hopkins+07]

Li et al. 2007

To prove this scenario we need:

1) Complete SMBH census, 2) Physical models for AGN feedbacks 3) Observational constraints to these models

Li et al. 2007


How to study AGN evolution? Count and detect AGN at different phase/stages of

their evolution Surveys Accretion Luminosity is emitted over a broad range

of wavelengths, BUT the X-ray emission is the AGN footprint

X-ray surveys AGN come in 2 flavours: unobscured and

obscured,obscuration affects mostly the soft X-ray and optical wavelengths

Hard (>2 keV) X-ray surveys (unbiased) Multiwavelength follow-up (redshifts!)


Evolution of Obscured AGN – why bother? (1) XRB fit

Gilli, Comastri, Hasinger 07Red -> unobscured AGNBlue + Black -> obscured AGN

Obscured AGN are needed to reproduce the XRB peak

[e.g. Setti & Woltjer 1989…….Gilli, Comastri & Hasinger 2007]


IF all galaxies undergo an AGN phase and IF dead SMBH observed today are the remnants/witnesses of this

phase

The BH mass density obtained integrating the luminosity emitted by AGN over the cosmic time is expected to be similar to that measured in local bulges

Soltan’s argument: mass accreted is equal to the energy released (E = mc2…)

(Soltan 1982)

ρ(direct) ~ ρ ● ~ 3-5.5 x 105 M⊙ Mpc-3

[see e.g. Shankar et al. 2008Marconi et al. 2004, Fabian 2003, Fabian & Iwasawa 1999 etc]

Evolution of Obscured AGN – why bother? (2) BH growth

INCLUDING ALSO OBSCURED / CT AGN


XMM-COSMOS 2 deg2

-16

-15

-14

-13

Flu

x 2

-10

keV

(cg

s)

CDFN-CDFS 0.1deg2 Barger et al. 2003; Szokoly et al. 2004

EGS/AEGIS 0.5deg2Nandra et al. 2006

HELLAS2XMM 1.4 deg2Fiore et al. 20003Cocchia et al. 2006Champ 1.5deg2Silverman et al. 2005

XMM HBS ~25 deg2Della Ceca+04,08XBOOTES 9 deg2Murray et al. 2005, Brand et al. 2005

Survey of surveys - X-rays + multiwave

SEXSI 2 deg2 Eckart et al. 2006

Area(see Brandt & Hasinger 2005 ARA&A 43, 827)

E-CDFS 0.3deg2 Lehmer et al. 2005

ELAIS-S1 0.5deg2Puccetti et al. 06, Feruglio et al. 08

Lockman Hole 0.2 deg2 Brunner et al. 2008

ContiguousC-COSMOS 0.9 deg2


picture quite clear from optical/soft/hard X-ray surveys:

Luminosity-Dependent Density Evolution (LDDE)

[see also Ueda+03, Cirasuolo+05, Della Ceca+08]

Anti-hierarchical growth/Downsizing [Cowie et al. 1996, Merloni 2004, De Lucia et al. 2006]

Evolution

Fiore +03La Franca +05

Hasinger+05

Bongiorno+07


Fraction of absorbed sources: Luminosity dependence

fraction of obscured AGN is a strong function of L: most luminous, less obscured

Same result in DIFFERENT bands despite the very different selections!!

Hasinger et al. 2005

(see also La Franca +05, Treister+05, Della Ceca+08)

Black: X-ray Hasinger 2008

Green: IRMaiolino et al. 2007 Red: Optical Simpson et al. 2005


Fraction of absorbed sources: Redshift dependence

more debated! Seen in (some) data [e.g. La Franca+05, Treister+06, Hasinger08], not

seen in others (Ueda+03, Dwelly&Page 2006), not needed in XRB models (Gilli+07) but expected/predicted in feedback models (Menci+08)

La Franca +05 Hasinger 2008


Still missing (open problems) Compton Thick objects / census

number density and evolution of high-z population (z>3)

Interplay between AGN and SF / feedback models [see Marulli/Fontanot] Bolometric output/reprocessing [see Severgnini/Bellocchi/Lusso] BH mass / accretion rate evolution [see Labita] Elusive AGN (e.g. XBONGs/EXOs) [see Civano/Del

Moro/Lanzuisi] Role of the envinronment in triggering nuclear activity [e.g. results from AEGIS/Groth strip – Georgakakis+07,08]


The population of z>3 QSOs(XMM-COSMOS)

Brusa, Comastri, Gilli et al. 2008, ApJ, submitted


?

The population of z>3 QSORadio QSO (Wall et al., 2005) Optical QSOs (Schmidt et al., 1995, Fan et al. 2001,2004)

X-ray: What happens at z>3? decline or ~costant ?Soft X-ray ROSAT/Chandra/ XMM (Hasinger, Miyaji & Schmidt 2005) Chandra/ROSAT(Silverman et al. 2005/2007)

statistics still low at z~3-5 (NO statistics at z>6)

COSMOS


+ MAMBO, CFHT, Bolocam and (future) others

COSMOS major components (in order of appearance) :

HST/ACS (I-band – 590 orbits – I(AB)~27)

2002-2003: Subaru imaging (~25 nights - b,v,r,i,z=26/27) VLA (265 hours – 24 μJy) GALEX deep (200 ks, AB~25) XMM-Newton (800 ks – 10-15 cgs)

2004-2005: XMM-Newton (600 ks) VLT (540 hours) & Magellan (12+ nights) SPITZER-IRAC (200 hours - ~1 μJy)

2006: SPITZER-MIPS (200 hours - ~70 μJy) Chandra (1.8 Ms)

Cosmos Survey2 deg2 (PI: N. Scoville)>150 team members worldwide!

http://www.astro.caltech.edu/cosmos/

XMM-NewtonPI: G. Hasinger

Hasinger et al. 2007

soft 0.5-2.0 keVmedium 2.0-4.5 keV

hard 4.5-10.0 keV

http://cosmos.astro.caltech.eduApJS special issue vol. 172 (>550 pages!)


AGN redshift distribution in XMM-COSMOS

X-ray sample (AGN) Empty: specz+photzFilled: specz

Sources extracted from a flux limited sample in XMM-COSMOS (50% of the area coverage in at least one band)

1651 XMM sources - 10% problematic ID using IR+Chandra info ~670 “secure” spectroscopic

redshifts (40%), incompleteness especially for high-z and Type 2 AGN

Accurate photometric redshifts available (Salvato et al. 2008)


LESS than 10% catastrophic outliers

(to be compared with COMBO-17, Wolf et al. 2004)

434 objects with “super-secure”spectro-z (further refinement/analysis)

Photo-z computed using >30 bands: SDSS, Subaru including IB, CFHT, J, K, IRAC..

AGN photometric redshifts

Salvato et al. (in prep)


AGN redshift distribution in XMM-COSMOS

X-ray sample (AGN) Empty: specz+photzFilled: specz

z>3 sample: 40 objects (22 specz + 18 photoz)Additional 14 objects, no photoz available. Candidates very high-z AGN (EXOs, Koekemoer et al. 2004)

Sources extracted from a flux limited sample in XMM-COSMOS (50% of the area coverage in at least one band)

1651 XMM sources - 10% problematic ID using IR+Chandra info ~670 “secure” spectroscopic

redshifts (40%), incompleteness especially for high-z and Type 2 AGN

Accurate photometric redshifts available (Salvato et al. 2008)


Examples of images, spectra & photoz

B g v r I z IRAC 3.6

XID 54439z=4.241


X-ray properties

HR vs. redshift

Ratio of obscured/ unobscured objects in agreement with XRB models


Contribution to source countsLower bound: 22 spectro-z

Upper-bound: adding 10 EXOs Dashed line: Expectations from XRB models using Hasinger et al. 2005 LF

Solid line:Exponential decay introduced at z=2.7 (Schmidt+95)

Flat evolution completely ruled outTightest constraints to date (largest and cleanest sample)

z>4 point:Rhook & Haenelt 08 predict a factor of ~3 higher wrt dataModels: Gilli, Comastri & Hasinger 2007,

A&A


Space densitiesRed curve: predictions from XRB models logLx>44.2 AGN (unobs+obs)[Gilli+07 using Hasinger+05 and La Franca+05]

Dashed area: (rescaled) space density for optically selected bright QSO [Richards+2006, Fan+2001]

Blue curve: Silverman+08 LF, I<24 sample


The census of z~1-3 CT AGN/QSOs(CDFS+COSMOS)Fiore et al. 2008a, ApJ, 672, 94

Fiore, Puccetti, Brusa et al. 2008b, ApJ, submitted


Unveiling obscured accretion

X-ray surveys: miss most highly obscured (Compton Thick) AGN [e.g. Comastri 2004, Worsley et al. 2005] IR surveys: highly obscured AGN shine in the MIR – dust reprocessing [e.g. Martinez-Sansigre et al. 2005, Polletta et al. 2006] Goal: combining X-ray and IR surveys to get the SMBH census and compile

AGN bolometric LF X-ray-MIR surveys: CDFS/GOODS/MUSIC [Grazian et al. 2006, Brusa et al. in prep, CDFS team papers]

ELAIS-S1/SWIRE [Feruglio et al. 2008, La Franca et al. 2008]

COSMOS [XMM-COSMOS: Brusa et al. 2008; + C-COSMOS/S-COSMOS papers]

X-ray surveys: miss most highly obscured (Compton Thick) AGN [e.g. Comastri 2004, Worsley et al. 2005] IR surveys: highly obscured AGN shine in the MIR – dust reprocessing [e.g. Martinez-Sansigre et al. 2005, Polletta et al. 2006] Goal: combining X-ray and IR surveys to get the SMBH census and compile

AGN bolometric LF X-ray-MIR surveys: CDFS/GOODS/MUSIC [Grazian et al. 2006, Brusa et al. in prep, CDFS team papers]

ELAIS-S1/SWIRE [Feruglio et al. 2008, La Franca et al. 2008]

COSMOS [XMM-COSMOS: Brusa et al. 2008; + C-COSMOS/S-COSMOS papers]


All sources

BL AGN

NOT BL AGN

High MIR/O

MIR selection of CT AGN

XMM-COSMOS sample HR distribution [Brusa et al. 2008b]See also results from ELAIS-S1, Feruglio et al. 2008, Lanzuisi’s talk

Select candidate, luminous obscured AGN by imposing:

24 micron bright fluxes (luminous) +

optically faint red sources

(optically obscured) high

MIR/O ratio

Select candidate, luminous obscured AGN by imposing:

24 micron bright fluxes (luminous) +

optically faint red sources

(optically obscured) high

MIR/O ratio


Combining MIR/O and R-K criteria:selection of CT AGN at z~2

Fiore et al. 2008a

See also Daddi et al. 2007

GOODS CDFS field(1 Ms Chandra data)+MUSIC MW catalog(Spitzer+HST+VLT)

~110 obscured AGN candidates

Stack of Chandra images excluding X-ray detections in two different MIR/O and R-K bins

Tracks of obscured AGN

[Pozzi et al 2007]


A new population of CT AGN

Curves: model predictions from Gilli, Comastri & Hasinger (2007) for L> 42, 43, 44, 45

Daddi et al. 2007

Fiore et al. 2008

Martinez-Sansigre et al. 2005Polletta et al. 2006

The observed MIR luminosity and the observed HR imply (unobs) Lx>43 and NH>24 for ~80% of the sources

Low R-K or


COSMOS MIR AGN Fiore et al. 2008b

High AGN fraction (~65%) in MIPS selected samples (deeper X-ray data + more comprehensive analysis)


Summary

The fraction of obscured sources depend strongly on luminosity

The fraction of obscured sources evolves (increases) with redshift

Flat evolution at z>3 definitively ruled out (space densities & number counts)

Significant population of CT QSO at z~2 detected


What’s next? Compton Thick AGN/quasars: XMM ultra deep 1.3 Ms (this fall) exploit stacking… (Chandra 2 Ms) Simbol-X (high-energy) Herschel (reprocessing) High-z quasars: eROSITA (high-L) XEUS (low-L and very high-redshift) joint multiwavelength campaigns

evolution of supermassive black holes from deep & large area x-ray surveys marcella brusa (max...

Documents

agn phase

agn winds

agn evolutionwhat

etcevolution of obscured

introduction agn evolutiononce

agn footprint xray surveysagn

coevolution of galaxies

accreting bh