1. The relation among black holes, their host galaxies and AGN activity INAF ISTITUTO NAZIONALE DI ASTROFISICA NATIONAL INSTITUTE FOR ASTROPHYSICS Galaxies and Structures through Cosmic Times Venice, March 26-31, 2006 Alessandro Marconi INAF-Osservatorio Astrofisico di Arcetri

2. In collaboration with …  Andrea Comastri (INAF – Bologna, I)  Roberto Gilli (INAF – Bologna, I)  Günther Hasinger (MPE, Garching, D)  Leslie Hunt (INAF – IRA, Firenze, I)  Roberto Maiolino (INAF – Arcetri, Firenze, I)  Guido Risaliti (INAF – Arcetri, Firenze, I)  Marco Salvati (INAF – Arcetri, Firenze, I)

3. Supermassive Black Holes  Supermassive BHs (10 6 -10 10 M  ) are detected in 30-40 NEARBY (D<100 Mpc) galaxies ( e.g. Ferrarese & Ford 2005 ).  M BH correlates with L sph /M sph ( Kormendy & Richstone 1995, Magorrian et al. 1998, McLure & Dunlop 2002, Marconi & Hunt 2003) and σ e (Ferrarese & Merritt 2000, Gebhardt et al. 2000 ).  Supermassive BHs likely present in all galaxies.  Supermassive BHs are also expected as “Relics” of AGN activity.  Are the local supermassive BHs consistent with being AGN relics?

4. The Relation between Local Black Holes and AGN relics  Compare the mass density of local BHs with that of AGN relics ( e.g. Soltan 1982, Fabian & Iwasawa 1999, Elvis, Risaliti & Zamorani 2002 )  Compare the local BH Mass Function with the mass function of relic BHs ( e.g. Yu & Tremaine 2002, Ferrarese 2002, Marconi et al. 2004, Merloni 2004, Shankar et al. 2004 )

5. + M BH - σ e relation + Faber-Jackson relation relationGalaxy Velocity Function The Local BHMF from M BH – L bul / σ e e.g. Salucci et al. 1998, Marconi & Salvati 2001 e.g. Yu & Tremaine 2002, Black Hole Mass Function + M BH -L bul relation + Bulge/Total correction Bulge (Spheroid) Luminosity Function Galaxy per Morphological Type e.g. Ferrarese 2002, Aller & Richstone 2003 (SDSS, Sheth et al. 2003)

6. The local Black Hole Mass Function  Using M BH -L bul and M BH - σ e provide consistent BH mass functions (differences included in shaded area which indicates uncertainties)  ρ BH ≃ 4.1 +1.9 -1.4 ×10 5 M  Mpc -3 ( cf. Merritt & Ferrarese 2001, Ferrarese 2002, Shankar et al. 2004 )  In summary: 3-5 ×10 5 M  Mpc -3 ( see Ferrarese & Ford 2005 for a review)

7. The AGN BH Mass Function  Assume accretion onto BH as powering mechanism of AGN to link L AGN with M BH [L= λ M BH c 2 /t E = ε (dM/dt)c 2 ]  Use the continuity equation (Cavaliere et al. 1971) to relate the BH Mass function N(M BH ) to the AGN Luminosity function Φ (L)  Critical issues:  L is the TOTAL accretion luminosity  Φ (L) is the luminosity function of ALL AGNs (observations provide Φ only for a subset of the AGN population)

8. Local BHMF vs Relics BHMF  The relic BHMF is a function of the band in which AGN are selected.  Even the hard (2-10 keV) XLF does not sample the whole AGN pop  Heavily obscured Compton-thick AGN are missing  X-ray background spectrum Qso LF Hard-X LF Soft-X LF

9. X-ray Background constraints  XRB models provide the total numbers of Compton-thin + Compton- thick AGN  Two options explored:  M1: R = obscured/unobscured AGN ratio = constant  M2: R decreasing with luminosity Gilli, Comastri, Hasinger 2006 in prep.

10. Local BHMF vs Relic BHMF  Correction for Compton-Thick sources from XRB models  whole AGN pop considered  The only free parameters are the accretion efficiency and Eddington ratio  Assume:  ε =0.1 (L= ε d M /dt c 2 )  λ =1 (L= λ L Edd )

11. Radiative Efficiency and Fraction of Eddington luminosity  Efficiency and fraction of Eddington luminosity are the only free parameters!  Determine locus in ε - λ plane where there is the best match between local and relic BHMF!  ε =0.04-0.10 λ =0.08-0.5 which are consistent with common ‘beliefs’ on AGNs

12. Local BHMF vs Relic BHMF  Local and Relic BHMFs are in agreement without considering merging.  Either merging of BHs is negligible for z<3 or it does not modify significantly the BHMF (e.g. Granato et al. 2004, Menci et al. 2004, Haiman, Ciotti & Ostriker 2004). with best ε and λ values …

13. Anti-Hierarchical BH growth  This is qualitatively consistent with models of galaxy formation (e.g. Menci et al. 2003, Granato et al. 2003)  Big BHs form in deeper potential wells  they form first.  Smaller BHs form in shallower potential wells and are more subjected to feedback effects (star form., AGN),  they form later and take more time to grow.  See also Merloni 2004. 50% of final mass

14. Conclusions  The local BH mass density is ρ BH = 4.1±1.5 ×10 5 M  Mpc -3.  The local BH mass function and the BH mass function of AGN relics are in good agreement with standard ε and λ values ( ε ~ 0.1, λ ~ 1.0).  Merging of BH’s either is not important or it does not significantly alter the relic BHMF, at least at z<3.  The BH growth is anti-hierarchical: smaller BH’s, M BH < 10 7 M , grow at lower redshifts, z<1, with respect to more massive ones, z=1-3.  Local BH's grew during AGN phases in which accreting matter was converted into radiation with ε = 0.04-0.1 and emitted at a fraction λ = 0.08-0.5 of the Eddington luminosity. Marconi et al. 2004, 2006 in preparation

15. L bol /L Edd from the sample of SDSS quasars (Mc Lure & Dunlop 2004)

16. Bolometric Corrections Zheng et al., Telfer et al. QSO templates α ox from Elvis, Risaliti & Zamorani 2002 and α ox (L) from Vignali et al. 2004

17. M BH – host galaxy correlations  M BH - L Sph (Kormendy & Richstone 1995, McLure & Dunlop 2002, Marconi & Hunt 2003)  M BH - M Sph (Magorrian et al. 1998, Marconi & Hunt 2003, Häring & Rix 2004)  M BH - σ e (Ferrarese & Merritt 2000, Gebhardt et al. 2000)  M BH -n (Sersic index, Graham et al. 2003)  They are still debated (few direct BH mass determinations, ~30-40, even less ~20 fully (?) reliable) but they are commonly taken as paradigms!  The M BH - σ e / L Sph /M Sph relations:  Indicate close link between BH growth and galaxy evolution  Allow a demography of BHs

18. Take into account dispersion in AGN X-ray spectral slope

19. Duty cicle of active BHs  δ is the fraction of Active BHs  δ = Φ ( L, t) / N( M BH, t)  δ = 1 @ z=3 is the initial condition (negligible effect on BH Mass Function at z=0) Gilli et al. M2 new

20. Total Total Lifetime of active BHs  M BH e-fold time (Salpeter’s):  To grow a BH SEVERAL t Salp needed: 7 × t Salp 10 3  10 6 M  14 × t Salp 10 3  10 9 M   t Salp independent of M BH, longer τ BH at lower M BH indicates a more difficult growth of smaller BHs (feedback?).  Estimated AGN lifetimes range from 10 6 to 10 8 yr (AGNs from SDSS imply lifetimes > 10 8 yr; Miller et al. 2003). τ BH ~2 ×10 8 yr (>10 9 M  ) τ BH ~7 ×10 8 yr (<10 8 M  )

21. Evolution of BH and stellar mass density SFR ≈ 4000×BHAR See also Merloni, Rudnick & Di Matteo 2004

22. Consistency of M BH -L bul and M BH - σ e  M BH -L bul and M BH - σ e provide the same BHMF?  Check BHMF obtained with Φ ( σ ) and Φ (L) from the same sample [9000 E/S0 from SDSS; Bernardi et al, Sheth et al 2003]  Take into account intrinsic dispersion of correlations:  M BH - σ has rms ≤ 0.3 (Tremaine et al. 2002).  M BH - σ and M BH -L bul have similar dispersion, rms ~0.3 (Marconi & Hunt 2003).