Dissecting the Dissecting the co- co- evolution evolution of Black Holes of Black Holes and galaxies and galaxies via basic accretion and via basic accretion and clustering models and the clustering models and the size distribution of size distribution of early-type galaxies early-type galaxies Francesco Shankar Francesco Shankar FERRARA AGN9 26/05/10 with: D. Weinberg, M. Bernardi, F. Marulli, with: D. Weinberg, M. Bernardi, F. Marulli, J. Moreno, Y. Shen, R. Sheth, S. Bonoli, J. J. Moreno, Y. Shen, R. Sheth, S. Bonoli, J. Miralda-Escude’, L. Ferrarese, M. Crocce, Z. Miralda-Escude’, L. Ferrarese, M. Crocce, Z. Haiman, C. Li, G. Kauffmann, S. White Haiman, C. Li, G. Kauffmann, S. White
Transcript
Dissecting the Dissecting the co-evolutionco-evolution of of Black Holes and galaxies Black Holes and galaxies via basic accretion and via basic accretion and
clustering models and the size clustering models and the size distribution of early-type distribution of early-type
galaxies galaxies
Francesco ShankarFrancesco Shankar
FERRARA AGN9 26/05/10
with: D. Weinberg, M. Bernardi, F. Marulli, with: D. Weinberg, M. Bernardi, F. Marulli, J. Moreno, Y. Shen, R. Sheth, S. Bonoli, J. Miralda-Escude’, J. Moreno, Y. Shen, R. Sheth, S. Bonoli, J. Miralda-Escude’,
L. Ferrarese, M. Crocce, Z. Haiman, C. Li, G. Kauffmann, L. Ferrarese, M. Crocce, Z. Haiman, C. Li, G. Kauffmann, S. WhiteS. White
SAMs are working hard tounderstand what is going on…
``our knowledge on the physics of accretion ontoBHs and their interaction with galaxies is still poor to drawfirm conclusions’’
Fontanot et al.
Malbon et al. Lapi, FS, et al.
A model-independent approach: The Continuity Equation
),(),(),(),(),(
tMStMStMntMMMt
tMnBHoutBHinBHBH
BH
BH
Important historical references : Cavaliere et al. (1971); Soltan (1982); Small & Blandford (1992); Salucci et al. (1999)
Ess
BHBHj BHjBH tt
t
MzMUzMptMM ;),(),(),( ,
BHM BHjBHBHj BHj MLzMzMUzMpzL ),(),(),,(),(
Redshift-dependent P(L/LEdd,z) distributions?
Mass-dependent P(L/LEdd,MBH,z) and radiative efficiency?
The median size decreases and σ increases at high-z, at FIXED Stellar Mass
)()Re()( 2 zzzM DYN
FS et al. in preparation
Gaskell 2009
FS, Bernardi, Haiman 2009
A “Cumulative” Test
SO…WHAT DID WE LEARN ABOUTSO…WHAT DID WE LEARN ABOUTHOW BHs EVOLVE? HOW BHs EVOLVE?
ACCRETION:ACCRETION: possible redshift and mass possible redshift and mass dependence in P(dependence in P(λλ), and mass ), and mass dependence in dependence in εε..
QUASAR CLUSTERINGQUASAR CLUSTERING: scatter L-dep. : scatter L-dep. (?), small-scales-> significant fraction of (?), small-scales-> significant fraction of satellites increasing with time (?); satellites increasing with time (?);
Negative Evolution in Mbh-Negative Evolution in Mbh-σσ!!!!!!
The sizes decrease and σ increase at high-z at FIXED Stellar Mass
High Clustering: Less Massive BHs in the Local Universe!
FS, D. Weinberg, J. Miralda-Escude’ 2009 FS 2009
dzdz
dt
L
zL
ttM
efBH log
),(
)10ln(
1),(
Continuity Equation : to reduce number density of low-mass BHscontinuosly decrease the Eddington Ratio in time!
A non-evolving Mbh-A non-evolving Mbh-σσ relation! relation!
35.1
54545.13
5
3
4
)1(~
~~)1(~
VIRH
VIRHBH
VzM
VzMM
CONCLUSIONSStarting From a Basic Model for the Triggering and Shining of QSOs we find:
1-Mass-dependent Light Curves favored
2-High clustering of z>3 QSOs : implies rapid growth/massive BH seeds; NO excess ‘’merger bias’’
3-High Clustering of AGNs in SDSS: implies flat BH mass function at the low-mass end
4-Scaling Relations : at high-z same Mbh-σ, but smaller sizes and higher σ for given stellar mass favors higher Mbh-Mstar!
The Clustering of “MERGING” Halos
S. Bonoli, FS, S. White, et al. 2009
We select the halos from the MSwhich have recently merged
Merger Rate enough to host all QSOs
First Ingredient: Merger/Virialization Rates of Halos
F&M rates consistentwith accurate theoretical estimates!
Fakhouri&Ma rates fits to the MS
J. Moreno
Another Application: The SDSS z>3 Quasar Clustering
FS, M. Crocce, J. Miralda-Escude’, P. Fosalba, D. Weinberg 2008
Duty cycle~1
n(Mbh=L/)~Ф(L)/P0
First Ingredient: Merger/Virialization Rates of Halos
F&M rates consistentwith accurate theoretical estimates!
Fakhouri&Ma rates fits to the MS
J. Moreno
The ratio <MBH/MSTAR> probably was nearly constantat all times at least up to z~1.5
Do BHs grow faster than Galaxies?
How do we cope with the How do we cope with the evolving evolving σσ in Ellipticals? in Ellipticals?
SO…WHAT DID WE LEARN ABOUTSO…WHAT DID WE LEARN ABOUTHOW BHs EVOLVE? HOW BHs EVOLVE?
ACCRETION:ACCRETION: can reproduce the local BH can reproduce the local BH mass function; preferred parameters are mass function; preferred parameters are 0.5<0.5<<1 and 0.07<<1 and 0.07<<0.1. Multi Edd. Ratios <0.1. Multi Edd. Ratios do not change Accreted BHMF.do not change Accreted BHMF.
The The Quasar clusteringQuasar clustering, independent , independent constraints on duty cycle, mean L-Mhalo constraints on duty cycle, mean L-Mhalo relation, and scatter, small-scales constraints relation, and scatter, small-scales constraints on the BH triggering mechanismson the BH triggering mechanisms
Constraints on co-evolution Constraints on co-evolution from evolution from evolution of QSO LF matchingof QSO LF matching
The ratio <MBH/MSTAR> probably was nearly constantat all times at least up to z~1.5
Do BHs grow faster than Galaxies?
ISSUES ABOUT WHICH IS THE MOST FUNDAMENTALRELATION, IF IT IS A 3-VARIABLE ONE, ETC…
WE FIND THE MBH-σ RELATION TO BE THE TIGHEST~0.18 dex
FS, L. Ferrarese 2009
An Application: The SDSS z~1.5 quasar clustering
FS, Shen, Weinberg, et al. 2009
Coupling with duty cyclefrom Continuity Equationbreaks some degeneracies!
First Results: large scatter!
Broad Eddington ratio Distributions III
Very Broad p() Very Broad p()+f(z)
SPECIAL MODELS: -dependent Bolometric Correction
Vasudevan & Fabian 2007
Low Radiative Efficiency+Low Eddington ratios
SimilarDownsizing
Harder to matchthe local BHMF:<0.1; <0.06
Broad Distributions IV: The Obscured Fraction
MBH-α
Babic et al.; Tozzi et al.;Alexander et al.
Hasinger; Akylas et al.
FS, D. Weinberg, J. Miralda-Escude’ 2009a
Broad Distributions III: The Obscured Fraction
MBH-αMBH
-α
SO FAR WE HAVE CONSIDERED MODELSWITH CONSTANT
EDDINGTON RATIO DISTRIBUTIONS
WE NOW ALLOW FOR THE ACCRETION RATES TO DEPEND ON REDSHIFT
(INCREASE/DECREASE WITH z)AND, AT FIXED TIME,
TO DEPEND ON BLACK HOLE MASS(INCREASE/DECREASE WITH MBH)
INPUT/CONSTRAINTS FOR MODELS
AGN LUMINOSITYFUNCTION
LOCAL BH MASS FUNCTION
PASSIVE BIAS : A SIGNATURE OF RAPID BH GROWTHAND MASSIVE SEEDs
a long delay ``lowers’’the bias at the shining
R. Angulo, M. Crocce
fs= fraction of AGNs which are satellites
Eastman et al.; Martini et al.; Sivakoff et al.
Low-z Clustering: RESULTSCoil et al.; Croom et al.; da Angela et al.; Francke et al.; Hennawi et al; Mountrichas et al.; Myers et al.; Padmanabhan et al.; Plionis et al; Porciani et al.; Shen et al.
Varying the Reference Model…L~MBH
The MThe Mbhbh--σσ Relation Relation
At peak
At shutdown
Observed locally
35.1
545423
5
3
4
)1(~
~~)1(~
VIRH
VIRHBH
VzM
VzMM
dzdz
dt
L
zL
ttM
efBH log
),(
)10ln(
1),(
L=LEdd (MBH )LEdd=1.3e38x(MBH/1e8) erg/s tef /
FS, D. Weinberg, J. Miralda-Escude’ 2008
P0=Ф(L)/n(Mbh=L/)
An Example of applying inputs fromAccretion+Clustering
into SAMs for Co-evolution
•Integrating the duty cycles in time: estimate of the mean lifetime of quasarsof a given final mass
Clumpy GAS at Tvir COLD GAS
RESERVOIR(low J)
STARS
IGMSMBH-QSO
SNae & QSO feedback
Radiative cooling
Radiation drag(SFR)
Viscous accretion
Collapse
Stellarevolution
Granato et al. 2004, 2006, FS et al. 2006
QSO ou
tflows
Within each virialized DM halo
SCUBA QSO phase
Passive Evolution
tdelaytvis
0
0
5.0
8
080
0 1010),(
t
BHBH M
MyrdttMP
FOR THE FUTURE:
Study of the Black Holes in MPA SAM:
-link light-curve to energy self-regulation
-compute the number of BH pairs and make predictions for LISA
-compute the statistics of radio sources
-probe more triggers for BH accretion
FS, F. Marulli, M. Bernardi, et al. 2009
But if Galaxies merge…what happens to SMBHs?
SO…WHAT DID WE LEARN ON SO…WHAT DID WE LEARN ON HOW BHs EVOLVE? HOW BHs EVOLVE?
Single-Single- models models can reproduce the local BH mass can reproduce the local BH mass function; preferred parameters are 0.5<function; preferred parameters are 0.5<<1 and <1 and 0.07<0.07<<0.1. <0.1.
Broad p(Broad p() distributions) distributions yield similar BH growth yield similar BH growth histories if histories if is independent of BH mass. is independent of BH mass.
The The high-z clusteringhigh-z clustering, especially at z=4 measured in , especially at z=4 measured in SDSS, requires very high host halo masses and SDSS, requires very high host halo masses and matching the AGN luminosity function requiresmatching the AGN luminosity function requires
0.4<0.4<//<1, i.e., <1, i.e., >0.2 if >0.2 if >0.5!>0.5!
If the If the Eddington ratio decreases with BH mass and zEddington ratio decreases with BH mass and z: : -match to the AGN fraction in the field and clusters -match to the AGN fraction in the field and clusters -match to small-scale clustering-match to small-scale clustering -match to the obscured fraction-match to the obscured fraction
MODELING THE low-z QUASAR CLUSTERING
1-The previous method breaks down at low-z: multiple quasars in halos!
2-We return to our previous accretion modeling tied to the observed AGN Luminosity Function: BH mass function predicted from continuity equation
3-We add the assumption that BH mass is a monotonic function (with scatter) of halo mass or the maximum mass of the subhalo
4-Scatter between L and MHALO can come either from scatter in MBH-MHALO or from a broad p(,z) distribution
5-Add new physical parameter Q=Ps/Pc! Small difference at large scales, significant at small ones
FS, M. Crocce, J. Miralda-Escude’, P. Fosalba, D. Weinberg 2008
Another Application: The SDSS z>3 Quasar Clustering
Soltan argument
maxmax
min0
2
2
),()1(
/)1(/
/
zL
LBoltotal
ACCBH
ACCBolBol
dLzLc
L
dz
dtdzK
dtdMdtdM
cdtdMKLL
maxmax
min0
2
232
2
),(4
)1()1(
)1()1()(
4),(),(
zS
S
Boltotal
A
LA
L
dSSzSnc
zdzc
K
zD
DzdtcDdV
D
LSdz
dz
dVdLzLdzdSzSn
independent of cosm. par. but very dependent on the Kbol/ε ratio!
SMBH from Merging/Dark Accretion or through Visible Accretion detected in the
AGN luminosity Functions?
Massive Dark Objects observed in all bulged-galaxies strong link with the host spheroid M/n/
What are MDOs? How and why are they connected with spheroids and DM? What is their role in shaping galaxies?
Dunlop & McLure; Ferrarese et al.; Gebhardt et al.; Graham et al.; Haring & Rix; Lauer et al.; Magorrian et al.; Marconi & Hunt; FS & L. Ferrarese 2009a,b
VC+DM profile VVIR(zvir=0)(Mvir )1/3
=kVc
+
SAMs are working hard tounderstand what is going on…
``our knowledge on the physics of accretion ontoBHs and their interaction with galaxies is still poor to drawfirm conclusions’’
Fontanot et al.
Malbon et al. Lapi et al.
OUTLINE OF THE TALKGOAL = EMPIRICALLY CONSTRAIN BLACK HOLE EVOLUTION IN A STATISTICAL SENSE
Local BH mass function and AGN lum. Functions:
Constraints on radiative efficiency,duty cycles, Eddington ratios
AGN Clustering:
Independent constraints on radiative efficiency, duty cycles, and dependencies on mass, redshift
Basic Modeling:
Evolution of sizes and velocities in Elliptical galaxies and the MBH-σ
For all relations used I convolve For all relations used I convolve with a Gaussian weightwith a Gaussian weight to account to account for intrinsic for intrinsic scatterscatter ! !
First Step: How many SMBH?How Massive?How many SMBH?How Massive?
Φ(L)→Φ(Lbulge)
MBH - Lbulge
Ф(MBH)
MBH - ()
Results: systematic uncertainties!
THE ACTIVE EVOLUTION OF BLACK HOLES: THE AGN LUMINOSITY FUNCTION
-radiative efficiency when coupled to AGN accretion history
Let’s have a look at what we think co-evolution is….
Evolution of Ellipticals in models:Wet phase+Dry phase
FS, Marulli, Bernardi, et al. 2009a
The ratio <MBH/MSTAR> probably was nearly constantat all times at least up to z~1.5-2
Do BHs grow faster than Galaxies?
FS, M. Bernardi, Z. Haiman 2008
Low-z Clustering: RESULTSCoil et al.; Croom et al.; da Angela et al.; Francke et al.; Hennawi et al; Mountrichas et al.; Myers et al.; Padmanabhan et al.; Plionis et al; Porciani et al.; Shen et al.
FS, M. Crocce, J. Miralda-Escude’, P. Fosalba, D. Weinberg 2008
Another Application: The SDSS z>3 Quasar Clustering
Good constraints from the small scales….
What do we learn fromhigh-z clustering?
1-DIFFICULT: high bias rare BHs, high duty cycles!
2-Reproducing the LF requires high and/or low : 0.4</<1; Shen et al. --> >0.5 --> >0.2
3-High halo mass and the limit duty cycle ≤ 1 leads to very rapid drop of quasar number counts at z>6
FS, M. Crocce, J. Miralda-Escude’, P. Fosalba, D. Weinberg 2008
Towards a successful Model:
-We already saw in PART I: mass dependence can match the obscured fraction
-A reasonable match to duty cycles and low-z clustering can be found if:
1-MBH-1/3
2- f(z)=[1-exp(z/2)]
3-Q=0.5-1
4-Scatter in MBH-Mhalo ~0.3-0.5 dex
fs= fraction of AGNs which are satellites
Eastman et al.; Martini et al.; Sivakoff et al.
Towards a successful Model II:
TO GET FINAL ANSWERS:TO GET FINAL ANSWERS:
FROM OBSERVATIONS:FROM OBSERVATIONS:Resolve systematics in the AGN LF knee and bright endResolve systematics in the AGN LF knee and bright endUnderstand biases in the Understand biases in the -distributions-distributionsSystematics in the mean bias valuesSystematics in the mean bias values
FROM THEORY:FROM THEORY:Convolve continuity Equation with BH merger rates Convolve continuity Equation with BH merger rates Get final BH mass function at all z consistentGet final BH mass function at all z consistent
with ALL observableswith ALL observablesPredict the BH mass function from SAM Predict the BH mass function from SAM
WILL WE BE ABLE TO OBSERVE z>6 QSOs?
1 QSO overthe whole sky
Same RedShift Distributions…but…more Accretion for the more Massive BHs
Very Broad p()Very Narrow p()
Broad Eddington ratio Distributions
Very Broad p() Very Broad p()+f(z)
We have checked we are using the right bias….
SPECIAL MODELS II: Low Accretion in ADAF modes
Broad Distributions III: The Obscured Fraction
Broad Distributions III: The Obscured Fraction
MBH-α
Babic et al.; Tozzi et al.;Alexander et al.
Hasinger; Akylas et al.
ANOTHER APPLICATION: THE z>3 QUASAR CLUSTERING
MBH ~ α (MHALO)β ---> Ф(MHALO,z) ---> Ф(MBH,z)
BHBH
BHef Md
Md
Ld
dt
dz
z
zMtzL 'log
'log
log),'()10ln(),(
tef~/
2-Assuming , we know how much energy is radiated and at which L
1-Assuming monotonic relation between BH and Halos
3-The parameters we use are: ,,,,
1.4
1
10),(
28.1
12
z
M
MzfM mBH
Scatter weakens the bias
The growth of the halo MF determines the BH growth
Low-z Clustering: Small and Large Scales
cP
sPQ
dmmLMNmsPmncPMn
dmmLMNmsPmnmbcPMnMb
Lb
M
M
,
,
)|][()(,)(,)(
)|][()(,)()(,)()(
)(
NO match at the small scales
Broad Eddington ratio Distributions
FS, D. Weinberg, J. Miralda-Escude’ 2009
…Same DOWNSIZING….
Duty cycle of AGNs: fraction of “Active’’ Galaxies
Broad Eddington ratio Distributions II
Very Narrow p() Very Broad p()Hopkins et al. LF+Broad p() peaked
at higher
SPECIAL MODELS: -dependent Bolometric Correction
Vasudevan & Fabian 2007
Low Radiative Efficiency+Low Eddington ratios
SimilarDownsizing
Harder to matchthe local BHMF:<0.1; <0.06
THE SUPERMASSIVE-BLACK HOLE “BUSINESS” IN 4 BULLETS:
1-The tightest local relations and their evolution with redshift/mass
2-The Demography of SMBHs, AGN statistics and Accretion
