Cosmological evolution of supermassive black holes and AGN: a synthesis model for accretion and feedback Andrea Merloni Excellence Cluster Universe & Max-Planck Institut für Extraterrestrische Physik Garching, Germany Sebastian Heinz Univ. Wisconsin, Madison, USA “The inner Kiloparsec”, Ierapetra, Crete, 6/6/2008

Black Holes in the local Universe Ω*BH≈710-5 [Fukugita & Peebles (2007)] Density of stellar mass BHs increases w/galaxy mass How does the high- mass (Supermassive) peak grow? Stellar physics, SN explosions, GRB ΩSMBH≈2.710-6 Accretion over cosmological times, Active Galactic Nuclei, galaxy evolution Sgr A* M87 Ωbaryon≈4.510-2 ; Ωstars≈2.510-3

Outline: What do we know about evolution of SMBH population? Evolution of mass and accretion rate density Constraints on radiative efficiency and avg. BH spin An “AGN synthesis model” Eddington ratio distributions, AGN fractions, etc. Constraints on AGN downsizing Mapping the different accretion modes Radiative vs. kinetic energy output and feedback

Synthesized backgrounds and AGN census Marconi et al. 2004 Hard X-rays (2-10 keV) Soft X-rays (0.5-2 keV) Optical/UV Gilli, Comastri, Hasinger 2007

Bolometric corrections robustness Hopkins et al 2006 Crucial: varying X-ray bolometric correction with luminosity (Marconi et al 2004) Accretion rate BH growth rate Radiative efficiency

Note: radiative efficiency vs. accretion efficiency accretion efficiency (BH spin) radiative efficiency Non Spinning BH Maximally Spinning BH Determined by the complex physics of gas accretion

(Normalized SMBH mass density) Integral constraints Soltan (1982) first proposed that the mass in black holes today is simply related to the AGN population integrated over luminosity and redshift BHAR(z) = (Normalized SMBH mass density)

Accretion power and star formation Hopkins and Beacom (2006) Brusa et al. 2008 Best constraints on high-z (z>3) X-ray selected AGN evolution (XMM-COSMOS) Absorption-corrected X-ray LF, a, + contribution of Compton thick AGN from XRB synthesis model (~20-30%) of total grown mass Merloni and Heinz 2008

Convergence in local mass density estimates             Graham and Driver (2007)

Constraints on avg. radiative efficiency rad 0=BH,0/(4.2x105) i=BH(z=zi)/BH,0 (Input from seed BH formation models needed!) lost=BH,lost/BH,0 (Mass density of SMBH ejected from galactic nuclei due to GW recoil after mergers) 0.065/[0 (1+lost)] rad 0.069/[0 (1- i +lost)] Merloni and Heinz 2008

Perez-Gonzalez et al.(2007) SMBH vs TOTAL stellar mass densities Perez-Gonzalez et al.(2007) rad=0.07 1 3 z

Constraints on avg. SMBH spin a* 0.065/[0 (1+lost)] rad 0.069/[0 (1- i +lost)] (Not symmetrized!, see King’s talk)

AGN/SMBH downsizing: clues from X-rays AGN downsizing Ueda et al. 2003; Fiore et al. 2003; Barger et al. 2005; Hasinger et al. 2005

Continuity equation for SMBH growth Cavaliere et al. (1973); Small & Blandford (1992); Marconi et al. (2004) 1- Need to know simultaneously mass function M(M,t0) and accretion rate distribution 2- At any z (or t), it is possible to combine mass function and bolometric luminosity function to calculate accretion rate distribution Picture from Di Matteo et al. (2007)

Mass function must get narrower with z Solved continuity equation backwards in time L/Ledd function gets wider Merloni and Heinz 2008.

“Active” BH Fractions/Mass Sensitivity curves Consequences of broad accretion rate distributions: Number density and mass functions of AGN are very sensitive to survey selection function The concept of “active” BH must be defined observationally Merloni and Heinz (2008)

SMBH downsizing The anti-hierarchical evolution at low z seems reversed at high z Heckman et al. 2004 ~ 3 Log MBH 23,000 type 2 AGN at z<0.1 Growth Time

Radiated energy density by BH mass We can follow the history of progenitors of local black holes Merloni and Heinz (2008)

Mass weighted avg. Eddington ratio evolution We can follow the history of progenitors of local black holes

AGN downsizing: changing accretion modes SMBH accrete at lower (average) rates at later times Accretion theory (and observations of X-ray Binaries) indicate that The energy output of an accreting BH depends crucially on its accretion rate Low-accretion rate systems tend to be “jet dominated” ( In the recent (misleading!) Cosmology jargon: Quasar mode vs. Radio mode (explosive vs. gentle)

Radio cores scaling with M and mdot A “fundamental plane” of active BHs [Merloni et al. 2003; Falcke et al. 2004] LR/1.3 1038 M1.38 Open triangles: XRB Filled squares: AGN Very little scatter if only flat-spectrum low-hard state sources are considered (Körding et al. 2006) Maccarone et al. 2003

AGN feedback: evidence on cluster scale 1 Msec observation of the core of the Perseus Cluster with the Chandra X-ray Observatory True color image made from 0.3-1.2 (red), 1.2-2 (green), 2-7 (blue) keV photons First direct evidence of ripples, sound waves and shocks in the hot, X-ray emitting intracluster gas Radio maps reveal close spatial coincidence between X-ray morphology and AGN- driven radio jets Fabian et al. 2006

Low Power AGN are jet dominated By studying the nuclear properties of the AGN we can establish a link between jet power and accretion power The observed slope (0.50±0.045) is perfectly consistent with radiatively inefficient “jet dominated” models (see E. Churazov’s talk) Cyg X-1 Log Lkin/LEdd=0.49 Log Lbol/Ledd - 0.78 Merloni and Heinz (2007)

Accretion diagram for LMXB & AGN Model parameter HK (high-kinetic; RLQ) LK (low-kinetic; LLAGN, FRI) HR (high-radiative; RQQ) (Blandford & Begelman 1999, Merloni 2004, Körding et al. 2007)

SMBH growth Most of SMBH growth in radiatively efficient mode 20-26%   0.065-0.07 Marconi et al. (2004) Merloni 2004; Merloni et al. 2008

Core Radio/LKin relation: effects of beaming Log Lkin=0.81 Log L5GHz +11.9 Slope=0.81 Slope=0.54 Observed LR (beaming) Derived from FP relation Monte Carlo simulation: Statistical estimates of mean Lorentz Factor ~8 Not a distance effect: partial correlation analysis Pnul=2 10-4 Merloni and Heinz (2007)

AGN Kinetic Luminosity function Körding, Jester and Fender (2007); Merloni and Heinz (2008)

Kinetic Energy output and SMBH growth Körding, Jester and Fender (2007); Merloni and Heinz. (2008)

Kinetic efficiency of growing black holes

Conclusions SMBH grow with a broad accretion rate distribution (be very careful when discussing AGN fractions, AGN lifetimes, etc.) Most of SMBH growth occurred in radiatively efficient episodes of accretion. Very strong constraint on rad. efficiency (spin) The anti-hierarchical trend is clearly seen in the low-z evolution of SMBH mass function. Reversal at higher z? Feedback from “Low-luminosity AGN” are most likely dominated by kinetic energy The efficiency with which growing black holes convert mass into mechanical energy is 0.3-0.5% (but strongly dependent on BH mass and redshift). Do we need Quasar winds/outflow for feedback at all?

Hubble Heritage Project The M87 jet Hubble Heritage Project http://heritage.stsci.edu/2000/20/index.html