Collaborators: Alex Beelen, Misty Bentz, Frank Bertoldi, Chris Carilli, Pierre Cox, Xiaohui Fan, Shai Kaspi, Dan Maoz, Hagai Netzer, Chris Onken, Pat Osmer,

Collaborators: Alex Beelen, Misty Bentz, Frank Bertoldi, Chris Carilli, Pierre Cox, Xiaohui Fan, Shai Kaspi, Dan Maoz, Hagai Netzer, Chris Onken, Pat Osmer, Chien Peng, Brad Peterson, Rick Pogge, Gordon Richards, Francesco Shankar, Adam Steed, Fabian Walter, David

Weinberg

Marianne Vestergaard

University of Arizona

First Steps Toward Constraining Supermassive Black-Hole Growth:Mass Estimates of Black Holes in

Distant Quasars

Drexel University, February 10, 2006Drexel University, February 10, 2006

Active Galactic Nuclei• Bright galaxies with a point-

source of non-stellar activity in nuclei

• They are rare – comprise only a few percent of bright galaxies

• The most powerful are called quasars.

• Quasar nuclei outshine their host galaxy light

~10 17 cm -- scale of oursolar system

(Francis et al. 1991)

(Elvis et al. 1994)

Supermassive Black Holes

• How are their mass measured?

• How do they grow?

• How are black holes and galaxies connected?

Black Holes and Galaxy Formation

• Black holes are likely ubiquitous in galaxy centers• MBH – σ* relationship

The M – σ Relationship

(Tremaine et al. 2002; See alsoFerrarese & Merritt 2000; Gebhardt et al. 2000)

Black HoleMass

Bulge Velocity Dispersion

M σ4


– Formation and evolution of bulges and black holes must be intimately connected

– When was it established? And how? – What came first, black hole or bulge (galaxy)?

• Black hole/star-formation feedback (theory)– Negative feedback kills star formation and black hole

growth by expelling gas (e.g., Springel, Di Matteo, & Hernquist 2005)


(Springel et al. 2005)

Black holeactivity

Star formationactivity

Time (Gyr)


– Formation and evolution of bulges and black holes must be intimately connected

– When was it established? And how? – What came first, black hole or bulge (galaxy)?

• Black hole/star-formation feedback (theory)– Negative feedback kills star formation and black hole

growth by expelling gas (e.g., Springel, Di Matteo, & Hernquist 2005)

– Positive feedback stimulate star formation (Silk 2005)

• Consequence: Galaxy bulges form later than supermassive black holes


I. Black Hole Mass a. Determinationsb. Distributions

II. Black Hole – Galaxy Connection

III. Black Hole Evolution

Talk Outline

Talk OutlineI. Black Hole Mass

a. Determinationsb. Distributions



mv2 – GmMBH /R = 0

Black Hole Mass

Mm

MBH = v2 R /G

Black Hole Mass

M

MBH = v2 R /G

Black Hole Mass

MBH = v2 R /GBlack Hole Mass

Insert figure from HST/ MW?

R

V

Why Study Quasar Black-Holes?

• Quiescent black holes (in normal galaxies) can only be studied in the nearby Universe

• Quasars are luminous and therefore ideal tracers of black holes to the highest observable redshifts

• Their host galaxies are prime targets for studying galaxy evolution in the early Universe

HST/STIS

8m telescope

30m telesc.

10010

Distance (Mpc)

(Ferrarese 2003)

109

108

Black Hole Mass

How Can MBH be Determined for Active Black Holes?

• Stellar kinematics

• Gas kinematics

• Reverberation mapping

(√)

(√)

√ √

Local Universe Higher-z

Possible Virial Estimators

Source Distance from central source

X-Ray Fe K 3-10 RS

Broad-Line Region 600 RS Megamasers 4 104 RS Gas Dynamics 8 105 RS Stellar Dynamics 106 RS

In units of the Schwarzschild radius RS = 2GM/c2 = 3 × 1013 M8 cm .

Mass estimates from thevirial theorem:

M = f (r V 2 /G)

wherer = scale length of regionV = velocity dispersionf = a factor of order unity, depends on details of geometry and kinematics

Note: the reverberation technique is independent of angular resolution

MBH = f v2 RBLR/G

Reverberation Mapping: RBLR= c τ

Virial Mass Estimates

t1 – t2 =

t = t1

t = t2

t = t3

t = t3 +

Reverberation Mapping Results

NGC 5548, the most closely monitored active galaxy

Continuum

Emission line

Light Curves

(Peterson et al. 2002)

MBH = f v2 RBLR/G

Reverberation Mapping:

–RBLR= c τ

• vBLR

Line width in variable (rms) spectrum

Virial Mass Estimates

t1– t2=

t = t1

t = t2

t = t3

t = t3 +

Reverberation Mapping

NGC 5548, the most closely monitored active galaxy

(Peterson et al. 1999)

• Velocity dispersion is measured from the line in the rms spectrum.– The rms spectrum

isolates the variable part of the lines.

– Constant components (like narrow lines) vanish in rms spectrum

Velocity Dispersion of the Broad Line Region and the Virial Mass

MBH = f v2 RBLR/G

f depends on structure and geometry of broad line region

(based on Korista et al. 1995)

MBH-: Comparison of Active and Quiescent Galaxies

• Reverberation masses appear to fall along the MBH - relation for quiescent galaxies

• The scatter is also similar: ≲ a factor of 3

Bulge velocity dispersion(Courtesy C. Onken)

Mass

AGNs

Gals

How Can Quasar MBH be Determined?

• Stellar kinematics

• Gas kinematics

• Reverberation mapping

(√)

(√)

√ √

Local Universe Higher-z

• Scaling relations √ √

Virial Mass Estimates MBH = f v2 RBLR/G

• Reverberation Mapping: RBLR=cτ, vBLR

Radius – Luminosity Relation: (Kaspi et al. (incl MV) 2005; Bentz,Peterson,

Pogge,MV,Onken 2006, ApJ, submitted)

• Scaling Relationships:

MBH FWHM2 L β

RBLR Lλ(5100Å)0.50

RBLR Lλ(1350Å)0.53

(see e.g. Vestergaard 2002)

Single-Epoch Mass Estimates - CIV

• 1 scatter = factor 2.3

(Vestergaard & Peterson 2006)

Mergs/s10

)1350(λL

km/s10

FWHM(CIV)105.4M

53.0

44λ

2

36

BH

Log [VP(CIV, single-epoch)/M]

Log

[ M

BH

(R

e ve r

b er a

tio n

)/ M

]

Scaling Relationships: (calibrated to 2004 Reverberation MBH)

• CIV:

1σ uncertainty: factor ~3.5

• Hβ:

Virial Mass Estimates: MBH=f v2 RBLR/G

Mergs/s10

)1350(λL

km/s10

FWHM(CIV)104.5M

0.53

44λ

2

36

BH

Mergs/s10

)5100(λL

km/s10

β) FWHM(H108.3M

0.50

44λ

2

36

BH

(see also Vestergaard 2002, and McLure & Jarvis 2002 for MgII)

( Vestergaard & Peterson 2006)

NGC 5548

Filled circles: 1989 data from IUE and ground-based telescopes.

Open circles: 1993 data from HST and IUE.

… Dotted line corresponds to virial relationship with M = 6 × 107 M.

Highest ionizationlines have smallestlags and largest Dopplerwidths.

Peterson and Wandel 1999

R (M/V) -1/2

Virial Relationships

(Peterson & Wandel 1999, 2000; Onken & Peterson 2002)

Emission lines:SiIV1400, CIV1549, HeII1640, CIII]1909, H4861, HeII4686

• All 4 testable AGNs comply:– NGC 7469: 1.2 107 M

– NGC 3783: 3.0 107 M

– NGC 5548: 6.7 107 M

– 3C 390.3: 2.9 108 M

• Scalings between lines: vFWHM

2(H) lag (H)

vFWHM2(CIV) lag (CIV)

• R-L relation extends to high-z and high luminosity quasars:– spectra similar (e.g., Dietrich et al 2002)

– luminosities are not extreme• R-L defined for 1042 – 1046 erg/s

(Vestergaard 2004)

(Dietrich et al 2002)

Radius – Luminosity Relation (Data from Kaspi et al. 2005)

Main goal: improve scalinglaws by reducing scatter

Improving the Scaling Relationships

Issues:• Host galaxy contamination

– HST imaging

• Accuracy of Single-epoch MBH estimates

– HST & ground-based study (HST archive project, PI: MV)

• Improved Masses and RBLR

– Improved monitoring of nearby sources

R-L relation scatter dominates scatter in mass scaling law

(Bentz, Peterson, Pogge, MV, Onken 2006)

Talk OutlineI. Black Hole Mass

a. Determinationsb. Distributions



Masses of Distant Quasars

• Ceilings at MBH ≈ 1010 M LBOL < 1048

ergs/s

• MBH ≈ 109 M beyond space density drop at z ≈ 3

(H0=70 km/s/Mpc; ΩΛ = 0.7)(Vestergaard 2004)

Quasars

• Dramatic space density drop at z ≳3

• Very luminous AGNs were much more common in the past.

• The “quasar era” occurred when the Universe was 10-20% its current age.

(Peterson 1997)



ergs/s


(H0=70 km/s/Mpc; ΩΛ = 0.7)(Vestergaard et al. in prep)



ergs/s


(DR3 Qcat: Schneider et al. 2005)(DR3 Qcat: Schneider et al. 2005) (Vestergaard et al. in prep)

Using MgII line to Estimate Black Hole Mass

• Bridge 0.8 ≲ z ≲ 1.3 gap• Will use SDSS to

calibrate MgII scaling law• Complications:

– FeII contamination of line and continuum

(Vestergaard & Wilkes 2001)

Requires template fitting

Talk Outline

I. Black Hole Masses



High Redshift Quasars and their Galaxies • UV, radio, X-ray properties similar at z > 3 (e.g., Constantin et al.

2002; Dietrich et al. 2002; Stern et al. 2000; Mathur et al. 2002)

• Black holes of distant quasars are very massive ~ (1-5)x 109 M

– Are their host galaxies also massive and old?

• Circumstantial evidence for intense star formation on galaxy scales associated with quasars at z 4:≳

– strong sub-mm/far-IR emission: ~108 M warm dust

– strong CO emission: ~1011 M of cold molecular gas (Ohta et al. 1996; Walter et al. 2003)

Dust and CO emission: large scale star formation rates 500 – 2000 M/yr (e.g., Omont et al. 2001, Carilli et al. 2001)

High Redshift Quasars and their Galaxies

• Some evidence for massive, old galaxies:

– z~2 quasar hosts have bulge luminosity consistent with old passively evolving stellar populations (Kukula et al. 2001)

– Low-z host galaxies are dominated by old (8-14Gyr) stellar populations (Nolan et al. 2001)

Quasar Host Galaxies at High Redshift• Conclusive test: mean age

and mass of stellar bulge

• Study of the most massive black holes at z ≳ 4– HST UV imaging: young stars

L(1500Å) → star formation rate

– HST Cy15 IR imaging: older stars

– Spitzer mid-IR: warm dust

– Sub-mm data: cooler dust

– CO imaging: cold molecular gas

• Goals: – Characterize stellar bulge:

mean age, mean mass, and star formation rate

– Determine MBH /MBulge

(Vestergaard 2004)

Redshift →

(Data from Bruzual & Charlot 2003)

Black Hole to Bulge Mass Ratio at High Redshift

(Peng et al. 2006, in prep)

Lensed Quasar Host Galaxy at Redshift 4.7

Original data PSF+Galaxy Model Galaxy residual

VLA CO (2-1) emission imagewith Einstein Ring (Carilli et al. 2003)

HST ACS UV image

Strong sub-mm source

Talk Outline

I. Black Hole Masses



Black Hole Growth in the Early Universe

Theoretical model predictions:

• Accretion only– Radiatively efficient– Radiatively inefficient

• Merger activity

• Obscured growth

• A combination of the above?

(Steed & Weinberg 2003)

Predicted evolution of black hole mass functions for different growth scenarios

Preliminary Mass Functions of Active Supermassive Black Holes

• Different samples show relatively consistent mass functions (shape, slope, normalization)

(Vestergaard & Osmer, in prep.; Vestergaard, Fan, Osmer et al., in prep.)

• Goal: constrain BH growth (with Fan, Osmer,

Steeds, Shankar, Weinberg)

(H0=70 km/s/Mpc; ΩΛ = 0.7)

• BQS: 10 700 sq. deg; B16.16mag

• LBQS: 454 sq. deg; 16.0BJ18.85mag

• SDSS: 182 sq. deg; i* 20mag

• DR3: 5000 sq. deg.; i* >15, 19.1, 20.2


• Different samples show relatively consistent mass functions (shape, slope)




(H0=70 km/s/Mpc; ΩΛ = 0.7)

• BQS: 10 700 sq. deg; B16.16mag



• DR3: 5000 sq. deg.; i* >15, 19.1, 20.2


• Different samples show relatively consistent mass functions (shape, slope)




(H0=70 km/s/Mpc; ΩΛ = 0.7)

• BQS: 10 700 sq. deg; B16.16mag



• DR3: 5000 sq. deg.; i* >15, 19.1, 20.2


• Locally mapped volume (R ≤ 100 Mpc):

MBH ≤ 3x109 M

• SDSS color-selected sample and DR3: (Fan et al. 2001, Schneider et al. 2005)

~9.5 quasars per Gpc3 with MBH ≥ 5x109 M

→ need ~25 times larger volume locally (R ≤ 290 Mpc)

(H0=70 km/s/Mpc; ΩΛ = 0.7)

Summary • >>> We can do physics with active galaxies and quasars <<<• MBH in Active Nuclei can be determined to within an accuracy:

– Low-z: ~factor of 3 (measured)– Higher z: ~factor of 4 (estimated!!)

• Black hole mass distributions:– <MBH> ≈ 109 M, even at 4 z 6≲ ≲– Maximum black hole mass at ~1010 M

• Black Hole Evolution and Galaxy Formation in Early Universe:– Ongoing study of galaxies at high redshift with the most

massive black holes (~1010 M)– MBH /MBulge ratio

– Mass functions of active black holes– Constrain growth of black holes and their galaxy bulges by

comparing these data with theoretical evolutionary models

The M – σ Relationship• Vittorini, Shankar, & Cavalier 2005, astro-ph/0508640 (BH

growth history from merger/feedback events; simulation)

• Robertson et al. 2005, astro-ph/0506038 (mergers simulation)

• Di Matteo, Springel, & Hernquist 2005, Nature, 433, 604 (merger induced BH growth and starformation; simulation)

• Springel, Di Matteo, & Hernquist 2005, MNRAS, 361, 776 (BH/star formation feedback; simulations)

• Miralda-Escude & Kollmeier 2005, ApJ 619, 30 (stellar capture)

• Sazonov et al. 2005, MNRAS 358, 168 (radiative BH feedback)

• King 2003, ApJ 596, L27 (supercritical accretion, outflows)

• Adams et al. 2003, ApJ 591, 125 (rotating BH collapse model)

• ….and many more…..

Secondary Mass Estimation MethodsVia MBH - *

bulge RelationMeasured *

bulge :

CaII 8498, 8542, 8662Å; z < 0.06

(Tremaine et al. 2002)(Ferrarese et al. 2001)

M 4.0

1 scatter ≈ 0.3 dex

M 4.72; MF00

AGNs


bulge Relation

[OIII]5007 FWHM *bulge

(Nelson & Whittle 1996; Nelson 2000)

(Boroson 2003)

Tremaine slope

Radio-louds

1 scatter ≈ 0.7 dex

• Line asymmetries

• Outflows• Radio sources

• (Interacting systems)


bulge Relation

Fundamental Plane: e, re *

bulge MBH

• Possibly significantly uncertain

- nuclear glare- bulge/disk decomposition (e.g., McLeod & Rieke 1995; Barth et al 2003)

-FP scatter (~0.6dex for RGs; e.g. Woo & Urry 2002)

- MBH - *bulge scatter

(Barth et al. 2003)

1 scatter = ? ( 0.7dex)

FundamentalPlane

log re

FP(,<e>)

Secondary Mass Estimation MethodsVia MBH - Lbulge Relation

(McLure & Dunlop 2001, 2002)

MR Lbulge

1 scatter ≈ 0.45 - 0.6 dex

• Nuclear glare

• Bulge/disk decomposition (e.g., McLeod & Rieke 1995; Barth et al 2003)

• Scaling relation scatter ?

MBH(dynamical)

MBH(scaling)

To first order quasar spectra look

similar at all redshifts

(Dietrich et al 2002)

Radius – Luminosity Relations

2HH

24

)H(

rn

L

cnr

QU

r L1/2

To first order, AGN spectra look the same

Same ionization

parameter Same density

[Kaspi et al (2000) data]

Radius-UV Luminosity Relationship for High-z Quasars

(Korista et al. 1997)

M = VFWHM2 RBLR/G

↑ ↑ ↓0.1109 M 4500km/s 33 lt-days

Ф RBLR-2 L

Log Ф Log n(H)

<L> ≈ 1047 ergs/s

Radius-UV Luminosity Relationship for High-z Quasars

(Dietrich et al. 2002)

M = VFWHM2 RBLR/G

Ф RBLR-2

Reverberation Mapping

• Kinematics and geometry of the broad-line region (BLR) can be tightly constrained by measuring the emission-line response to continuum variations.

• Can be done with

UV/optical lines.NGC 5548, the most closely monitored Seyfert 1 galaxy

Reverberation Mapping Results

• BLR sizes are measured from the cross-correlation time lags between continuum and emission-line variations.

• This gives the first moment of the transfer function.

NGC 5548, the most closely monitored Seyfert 1 galaxy

Continuum

Emission line

Reverberation Mapping Assumptions1 Continuum originates in a single central source.

– Continuum source (1013–14 cm) is much smaller than BLR (~1016 cm)

– Continuum source not necessarily isotropic

2 Light-travel time is most important time scale.• Cloud response instantaneous

rec = ( ne B)1 0.1 n101 hr

• BLR structure stable dyn = (R/VFWHM) 3 – 5 yrs

3 There is a simple, though not necessarily linear, relationship between the observed continuum and the ionizing continuum.

• In practice, programs have concentrated on solving the

velocity-independent (or 1-d) transfer equation:

The Transfer Equation

dtCtL )()()(

– Transfer function is line response to a -function outburst.

• It is most common to determine the cross-correlation

function and obtain the “lag”

dtt )(ACF)()(CCF

• Under these assumptions, the relationship between the

continuum and emission lines is:

Emission-linelight curve

“TransferFunction”

ContinuumLight Curve

dtCVtVL )(),(),(

The Transfer Equation• The aim of reverberation mapping is to solve for the

transfer function from the observables, the continuum

light curve C(t) and the emission-line light curve L(V,t).

• As noted earlier, currently we have been able to get

only the cross-correlation lag with any certainty.

Emission-linelight curve

“TransferFunction”

ContinuumLight Curve

dtCVtVL )(),(),(

dtt )(ACF)()(CCF

Reverberation Mapping Results• AGNs with lags for

multiple lines show that highest ionization emission lines respond most rapidly ionization stratification

• Combine lag with line width to get a “virial mass”.

Documents

Collaborators: Alex Beelen, Misty Bentz, Frank Bertoldi, Chris Carilli, Pierre Cox, Xiaohui Fan, Shai Kaspi, Dan Maoz, Hagai Netzer, Chris Onken, Pat Osmer,