The Origins & Evolution of the Quasar Luminosity Function

Transcription

1 The Origins & Evolution of the Quasar Luminosity Function Philip Hopkins 07/14/06 Lars Hernquist, Volker Springel, Gordon Richards, T. J. Cox, Brant Robertson, Tiziana Di Matteo, Yuexing Li, Sukanya Chakrabarti

2 A subset of recent quasar samples...

3 And yet... Theory: KH 00 Malbon+ 06 K Vittorini+ 05 Siana+ 06 ( ; 1450) ( ; i-band) ( ; 1450) ( ; 1450) ( ; 1216) Observations:

4 Richards+ 06

5 How to Address This? METHODOLOGY Given bolometric Phi(L), convolve over SEDs: FIRST INTEGRAL SPITZER 2MASS SDSS GALEX XMM/CHANDRA Elvis+ 94; Vanden Berk+ 01; Telfer+ 01; Ueda+ 03; Hatziminaoglou+ 05; Richards+ 06;

6 Methodology GIVEN A BOLOMETRIC PHI(L), CONVOLVE OVER KNOWN SEDs Vanden Berk+ 01 Telfer+ 02 Hatziminaoglou+ 05

7 Methodology GIVEN A BOLOMETRIC PHI(L), CONVOLVE OVER KNOWN SEDs Want the intrinsic SED (e.g. Marconi+ 04) Type 1, un-obscured/un-reddened, subtract host light BIG BLUE BUMP HOT DUST Hopkins, Richards, & Hernquist 06 POWER LAW + COMPTON REFLECTION

8 Methodology MORE THAN ONE SED Dependence on L Distribution of SED shapes Remove obscuration component Steffen+ 06 Steffen+ 06 Elvis+ 94 Richards+ 06

9 Methodology OBSCURATION Ueda+ 03 Then convolve over NH distributions Include Compton-thick Dependence on L: Ueda+ 03; Hasinger+ 04; Grimes+ 04; Simpson+ 05 Tozzi+ 06

10 Obtain Phi(L) specific to observed band, L_obs, and redshift: HRH 06

11 Luminosity- Dependence TEST DIFFERENT BOLOMETRIC CORRECTIONS, ETC.

12 LF vs. Redshift UV THROUGH IR

13 LF vs. Redshift UV THROUGH IR

14 What Do We Learn? ZERO-TH ORDER z_peak = /- 0.05

15 What Do We Learn? ZERO-TH ORDER Barcons+ 99; Gruber+ 00 Marconi+ 04; Shankar+ 04; Tremaine+ 02 See also Lauer+ 06 Marconi+ 04; Shankar+ 04

16 What Do We Learn? ZERO-TH ORDER From Phi(M_B<-26) Actual

17 What Do We Learn? FIRST ORDER

18 What Do We Learn? FIRST ORDER Little ambiguity in L-M mapping Model-independent Simulations Observed mdot Hopkins+ 06 Kollmeier+ 05 Vestergaard+ 04

19 What Do We Learn? FIRST ORDER Little ambiguity in interpretation at z < 2 High-z can t get bigger Observed mdot Observed clustering Local BHMF Porciani+ 04; Croom+ 05; Myers+ 06; Adelberger+ 06 Marconi+ 04; Shankar+ 04

20 What Do We Learn? FIRST ORDER Suggestive: sites of BH/spheroid co-formation? M_sph(50%) * Cimatti+ 05: Bundy+ 06: Borch+ 06: M_BH Hopkins, Bundy+ 06

21 What Do We Learn? FIRST ORDER High-z :: low-m_bh, or still building up? M_BH vs. M_BH_final

22 What Do We Learn? FIRST ORDER High-z :: low-m_bh, or still building up? Host masses? Clustering (in prep)

23 Recall...

24 Recall...

25 Luminosity- Dependent Density Evolution SECOND ORDER

26 What Do We Learn? SECOND ORDER Equivalently, slopes flatten with z

27 What Do We Learn? SECOND ORDER Smooth; Low z Faint End (X-ray LDDE ) Incompleteness? All bands/samples? Want faint-end slope over large-z from single surveys

28 What Do We Learn? SECOND ORDER Heckman+ 04 Faint End (X-ray LDDE ) Incompleteness? Lots of (very) low-m active BHs? No faint-end Lqso-Lhost correlation (Bahcall+; Hao+; Vanden Berk+) Local (Hao+05) faint-end slope Light-weighted BHMF Marconi+ 04; Shankar+ 04 radio counts (Haiman+ 04) X-ray counts (Volonteri+ 06)

29 What Do We Learn? SECOND ORDER Faint End (X-ray LDDE ) Incompleteness? Lots of (very) low-m active BHs? Change in effective duty cycle/lifetime for more massive BHs at low mdot Luminosity-Dependent Quasar Lifetimes Low-M_BH (slow decay) High-M_BH (fast decay) +

30 What Do We Learn? SECOND ORDER Mbh=10^8 3x10^8 3x10^9 Constrain Lifetimes + Feedback Physics

31 What Do We Learn? SECOND ORDER Bright End Binning? Probably important at z<1 z < < z < 0.3 Sazonov & Revnitsev 04 Very unlikely with SDSS DR3 (bin-bin dz<0.03)

32 What Do We Learn? SECOND ORDER Bright End Binning? Lensing? Richards+ 04, 06 see also Wyithe 04

33 What Do We Learn? SECOND ORDER Bright End Binning? Lensing? Effective bias? (distribution of bolometric corrections) Unlikely in optical Favored in *all bands* Large, optical surveys still the best bet: hope for IR?

34 What Do We Learn? SECOND ORDER Bright End (Systematics) Reflects shape of halo MF/buildup? z=1.5, 2, 3 z=8 Shape ~ Sheth-Torman time derivative? Lapi+ 06 z=6 z=5

35 What Do We Learn? SECOND ORDER Bright End (Systematics) Reflects shape of halo MF/buildup? Feedback again? Croton+ 06 Scannapieco & Oh 04

36 What Do We Not Know How to Interpret? Phi_star: what does it mean? Number of active systems? Duty cycles / lifetimes?

37 Summary The combined set of quasar observations has enormous constraining power that should be exploited -11 < phi < -2; 8 < L < 16; z = Need to be careful about combining observations Systematics now the dominant uncertainty at z<4 Constrains AGN physics: SEDs & NH depend on L, not z Break, luminosity density, shape change well-measured Encodes information about population buildup & feedback Cosmic Downsizing as manifest in QSOs Complex shape evolution Quasar lifetime not one number: Luminosity-dependent lifetimes

38 Host Light

39 Given the Conditional Quasar Lifetime, De-Convolve the QLF QUANTIFIED IN THIS MANNER, UNIQUELY DETERMINES THE RATE OF TRIGGERING Log(Time at L) Simple quasar lifetimes Log(L/L sun ) Log(L/L sun ) If every quasar is at the same fraction of Eddington, the active BHMF (and host MF) is a trivial rescaling of the observed QLF

40 Given the Conditional Quasar Lifetime, De-Convolve the QLF QUANTIFIED IN THIS MANNER, UNIQUELY DETERMINES THE RATE OF TRIGGERING Log(Time at L) Simple quasar lifetimes Observed luminosity function Log(L/L sun ) Log(L/L sun ) If every quasar is at the same fraction of Eddington, the active BHMF (and host MF) is a trivial rescaling of the observed QLF

41 Given the Conditional Quasar Lifetime, De-Convolve the QLF QUANTIFIED IN THIS MANNER, UNIQUELY DETERMINES THE RATE OF TRIGGERING Log(Time at L) Simple quasar lifetimes Observed luminosity function Log(L/L sun ) Log(L/L sun ) If every quasar is at the same fraction of Eddington, the active BHMF (and host MF) is a trivial rescaling of the observed QLF

42 Given the Conditional Quasar Lifetime, De-Convolve the QLF QUANTIFIED IN THIS MANNER, UNIQUELY DETERMINES THE RATE OF TRIGGERING Log(Time at L) Simple quasar lifetimes Log(L/L sun ) Formation rate/ triggering rate/ active BHMF Log(L/L sun ) Observed luminosity function If every quasar is at the same fraction of Eddington, the active BHMF (and host MF) is a trivial rescaling of the observed QLF

43 Log(Time at L) Simulated quasar lifetimes Formation rate vs. MBH Observed QLF Log(L/L sun ) Log(L/L sun ) Feedback-regulated lifetime drives a given QSO to lower L after blowout, and spends more time at low-l Much stronger turnover in formation/merger rate Faint-end QLF dominated by decaying sources with much larger peak luminosity/hosts

44 Feedback-driven Blowout Gives M-sigma Relation PREVENTS RUNAWAY BLACK HOLE GROWTH Di Matteo et al Black hole growth without feedback with feedback (Gebhardt et al. 2000; Ferrarese & Merritt 2000; Tremaine et al. 2002) Springel et al. 2004

45 Quasar Clustering is a Strong Test of this Model MOST FAINT QSOS ARE DECAYING BRIGHT QSOS - SHOULD BE IN SIMILAR HOSTS Weak dependence of clustering on observed luminosity (Croom et al. 2005, Adelberger & Steidel 2005, Myers et al. 2005) Observed trends with redshift (M_halo ~ 10^13 M_sun) Adelberger & Steidel 05 Myers et al. 05 Lidz et al Croom+ 05

46 Generalizing the Model NOT ALL AGN ARE MERGER-DRIVEN Almost any (ex. radio) AGN feedback will share key properties: Point-like Short input (~ t_salpeter) E~E_binding (defines when the feedback is important) Suggests analytical solutions for decay of accretion rates in feedback-driven winds or blastwaves Agrees well with simulations! Generalize to Seyferts Disk-dominated galaxy, central molecular clouds Calculate accretion rate(time) when a cloud collides with the BH

47 The Seyfert Luminosity Function A STOCHASTIC BUT FEEDBACK-REGULATED MODEL Hopkins & Hernquist 2006 Seyferts (disk-dominated; stochastic cloud fueling) Dead Hot gas/stellar wind fueled systems Post-Starburst Spheroids (post-merger lightcurve decay) Hao+ 05; Ueda+ 03;

48 The Seyfert Luminosity Function PREDICT THE EDDINGTON RATIO DISTRIBUTIONS FROM THIS FUELING MODE, AS BEFORE Yu+ 05

49 The Seyfert Luminosity Function CORRECTIONS TO THE M_BH-SIGMA RELATION Tremaine+ 02; Onken+ 04; Nelson+ 04; Peterson+ 04, 05; Barth+ 04, 05; Greene & Ho 05

50 The Seyfert Luminosity Function CONTRIBUTION AS A FUNCTION OF REDSHIFT Ueda+ 03

51