Mitch Begelman JILA, University of Colorado GROWING BLACK HOLES.

Mitch Begelman

JILA, University of Colorado

GROWING BLACK HOLES

COLLABORATORS

• Marta Volonteri (Michigan)• Martin Rees (Cambridge)• Elena Rossi (JILA/Leiden)• Phil Armitage (JILA)• Isaac Shlosman (JILA/Kentucky)• Kris Beckwith (JILA)• Jake Simon (JILA)

EARLYQSOs with M>109M at z>6

OFTENOne per present-day galaxy

BLACK HOLES FORMED…

HOW DID THESE BLACK HOLES GET

THEIR START?

2 SCHOOLS OF THOUGHT:

• Pop III remnants – Stars form, evolve and collapse

– M*~103 M

– MBH~102 M

• Direct collapse– Massive gas cloud accumulates in nucleus– Supermassive star forms but never fully relaxes;

keeps growing until collapse

– M*>106 M

– MBH >104 M

Rees, Physica Scripta, 1978

Rees’s flow chart

32 years later …

Begelman & Rees, “Gravity’s Fatal Attraction” 2nd Edition, 2010

Begelman & Rees, “Gravity’s Fatal Attraction” 3nd Edition E-book?

Keeping up with the times…

• Pop III remnants – ~100 (?) M BHs form at z > 20– 105-6 M halos, Tvir ~ 102-3 K– Grow by mergers & accretion– Problems:

• Slingshot ejection from merged minihalos? • Feedback/environment inhibits accretion?

• Direct collapse– Initial BH mass = ? at z < 12 – 108-9 M halos, Tvir >104 K– Grow mainly by accretion – Problem:

• Fragmentation of infalling gas?

~ Smaller seeds, more growth time

Larger seeds, less growth time

TRADEOFFS:

STAGE I:

COLLECTING THE GAS

The problem: angular momentum

The solution: self-gravitating collapse

SELF-GRAVITATING COLLAPSE: A GENERIC MECHANISM:

• “Normal” star formation

• Pop III remnants

• Direct collapse-1

Sun4 yr M2.0 K,10 MT

G

T

G

vM

2/33

~

-1Sun

24 yr M1010~ K,1000100~ MT

-1Sun

45 yr M1010~ K,10010~ MT

DM

gas

DM

gas

(approx.) 25.0en. pot.

en. rot.

Halo with slight rotation Gas collapses if virialgas TT

“BARS

WITHIN

BARS”

Shlosman, Frank & Begelman 1989

Dynamical loss of angular momentum

through nested global gravitational instabilities

Wise, Turk, & Abel 2008

Collapsing gas in a pre-galactic halo:

R-2 density profile

Wise, Turk, & Abel 2008

Global instability, “Bars within Bars”:

Instability at distinct scales → nested bars

WHY DOESN’T THE COLLAPSING GAS FRAGMENT

INTO STARS?

IT’S COLD ENOUGH …

… BUT IT’S ALSO HIGHLY TURBULENT

Wise, Turk, & Abel 2008

Collapse generates supersonic turbulence, which inhibits fragmentation:

HOW TURBULENCE COULD SUPPRESS FRAGMENTATION

Begelman & Shlosman 2009

Razor-thin disk (Toomre approximation):

FRAGMENTATION SETS IN BEFORE BAR INSTABILITY

ROTATIONAL SUPPORT ⇨

⇦ F

RA

GM

EN

T S

IZE

THE KEY IS DISK THICKENING

BAR

FR

AG

ME

NT

S

kGkvt 22222

HOW TURBULENCE COULD SUPPRESS FRAGMENTATION

Begelman & Shlosman 2009

Disk thickened by turbulent pressure:

BAR INSTABILITY SETS IN BEFORE FRAGMENTATION

ROTATIONAL SUPPORT ⇨

⇦ F

RA

GM

EN

T S

IZE

THE KEY IS DISK THICKENING

BAR

FR

AG

ME

NT

S

WHY?

THICKER DISK HAS “SOFTER” SELF-GRAVITY

⇨ LESS TENDENCY TO FRAGMENT

(DOESN’T AFFECT BAR FORMATION)hk

kGkvt

1

22222

HOW TURBULENCE COULD SUPPRESS FRAGMENTATION

Begelman & Shlosman 2009

5% of turbulent pressure used for thickening :

ENOUGH TO KILL OFF FRAGMENTATION

ROTATIONAL SUPPORT ⇨

⇦ F

RA

GM

EN

T S

IZE

THE EFFECT IS DRAMATIC

BAR

FR

AG

ME

NT

S

MORE SIMULATIONS (WITH HIGHER RESOLUTION) NEEDED!

At

radiation trapped in infalling gas halts the collapse

Rapid infall can’t create a black hole directly…

AUyr 1

4~1-

SolM

MR

STAGE II:

SUPERMASSIVE STAR

SUPERMASSIVE

STARS

• Proposed as energy source for RGs, QSOs • Burn H for ~106 yr• Supported by radiation pressure fragile

• Small Pg stabilizes against GR to 106 M

• Small rotation stabilizes to 108-109 M

Hoyle & Fowler 1963

THINGS HOYLE & FOWLER DIDN’T KNOW

ABOUT SUPERMASSIVE STARS

• They are not thermally relaxed

… because they didn’t worry about how they formed

AU 4~ mR

INCOMPLETE THERMAL RELAXATION SWELLS THE STAR:

MR

THINGS HOYLE & FOWLER DIDN’T KNOW

ABOUT SUPERMASSIVE STARS

• They are not thermally relaxed • They are not fully convective

… because they didn’t worry about how they formed

STRUCTURE OF A SUPERMASSIVE STAR

CONVECTIVE CORE

matched to

RADIATIVE ENVELOPE

0 1 2 3 4 5 6 70 .0

0 .2

0 .4

0 .6

0 .8

1 .0

cT

T

Scaled radius

coreM

M*

const.3/4 P

POLYTROPE

)(3/23/4 rMP

“HYLOTROPE”

Thanks, G. Lodato & A. Accardi!

(hyle, “matter” + tropos, “turn”)

HYLOTROPE,

NOT

HELIOTROPE!!

FULLY CONVECTIVE

PARTLY CONVECTIVE

MAX. M

ASS

INCOMPLETE CONVECTION DECREASES ITS LIFE & MAX. MASS

THINGS HOYLE & FOWLER DIDN’T KNOW

ABOUT SUPERMASSIVE STARS

• They are not thermally relaxed • They are not fully convective • If made out of pure Pop III material they

quickly create enough C to trigger CNO

… because they didn’t worry about how they formed

METAL-POOR STARS BURN HOTTER

A BLACK HOLE FORMS

SMALL (< 103 M) AT FIRST …

… BUT SOON TO GROW RAPIDLY

STAGE III:

QUASISTAR

“QUASISTAR”

• Black hole accretes from envelope, releasing energy

• Envelope absorbs energy and expands • Accretion rate decreases until energy output =

Eddington limit – supports the “star”

Begelman, Rossi & Armitage 2008

SO THE BLACK HOLE GROWS AT THE EDDINGTON LIMIT, RIGHT?

BUT WHOSE LIMIT?

EDDINGTON

GROWTH AT EDDINGTON LIMIT FOR ENVELOPE MASS > 103-4 X BH MASS

EXTREMELY RAPID GROWTH

“QUASISTAR”

• Resembles a red giant • Radiation-supported convective envelope • Photospheric temperature drops as black hole

grows

Central temp. ~106 K

Radius ~ 100 AU

Tphot drops as BH grows

DEMISE OF A QUASISTAR

• Critical ratio: RM=(Envelope mass)/(BH mass) • RM < 10: “opacity crisis” (Hayashi track)• RM < 100: powerful winds, difficulty matching accretion

to envelope (details very uncertain)

Final black hole mass:

Sol

SolSol

MBH

M

MM

MRM

64

1-27

1010~

yr 110~

STAGE IV:

“BARE” BLACK HOLE

“Normal” growth via accretion & mergers

THE COSMIC CONTEXT

• Collapse occurs only in gas-rich & low ang. mom. halos• Need ang. mom. parameterλ~0.01-0.02 vs. meanλ~0.03-0.04

• Competition with Pop III seeds• Pre-existing Pop III remnants may inhibit quasistar formation • ... but pre-existing quasistars can swallow Pop III remnants

• Merger-tree models vs. observational constraints:• Number density of BHs vs. z (active vs. inactive)• Mass density of BHs vs. z (active vs. inactive)• BH mass function vs. z• Total AGN light (Soltan constraint) • Reionization

Volonteri & Begelman 2010

BLACK HOLE mass density

All BHs: (thin lines) Active BHs: (thick lines)

TOTAL AGN LIGHT POP III

ONLY

Volonteri & Begelman 2010

CAN SUPERMASSIVE STARS OR QUASISTARS BE DETECTED?

Quasistars peak in optical/IR: some hope?

Supermassive stars:

AGNmodest a to...similar K102~

erg/s104~4/15

eff

45

mT

L

…strong UV source (hard to distinguish from clusters of hot stars)

JWST quasistar counts

Tphot=4000 K Band: 2-10 mSens. 10 nJy

Lifetime ~106 yr

λspin<0.02

λspin<0.01

1/JWST field

1/JWST field

WHAT ABOUT M-σ?

• Do AGN outflows really clear out entire galaxies? – or is global feedback a “red herring”?

• Do BH grow mainly as Eddington-limited AGN or in smothered, “force-fed” states (e.g., following mergers)• if the latter, then BH growth could be coupled

to σthrough infall rate σ3/G• ... but what is the regulation mechanism?

To conclude …

BOTH ROUTES TO SUPERMASSIVE BLACK HOLE FORMATION ARE STILL IN PLAY

MASSIVE BLACK HOLE FORMATION BY DIRECT COLLAPSE LOOKS PROMISING

THE PROCESS INVOLVES 2 NEW CLASSES OF OBJECTS

QUASISTARS AT Z~6-10 MIGHT BE DETECTABLE WITH JWST

Requires self-gravitating infall without excessive fragmentation

Supermassive stars initial ⇨seedsQuasistars ⇨ rapid growth in massive cocoon

Many unsolved problems: Effects of mass loss? Late formation after mergers? Formation around existing black holes? ....

DIRECT COLLAPSE LOOKS PROMISING

CORE COLLAPSE OF SUPERMASSIVE STARS

QUASISTARS DETECTABLE?

RAPID GROWTH INSIDE MASSIVE COCOONS