THE FIRST SUPERMASSIVE BLACK HOLES Mitch Begelman JILA

THE FIRST SUPERMASSIVE BLACK HOLES? Mitch Begelman JILA, University of Colorado

COLLABORATORS • • Marta Volonteri (Cambridge/Michigan) Martin Rees (Cambridge) Elena Rossi (JILA) Phil Armitage (JILA)

NEED TO EXPLAIN: • Ubiquity of BHs in present-day galaxies • QSOs with M>109 M at z>6 – Age of Universe < 20 t. Salpeter Eddington-limited accretion would have to: – Start early – Be nearly continuous – Start with MBH >10 – 100 M

The Rees Flow Chart Begelman & Rees, MNRAS 1978

One more time, with calligraphy… Rees, Physica Scripta, 1978

18 years later… with 4 -color printing! Begelman & Rees, “Gravity’s Fatal Attraction” 1996

ORIGIN OF SEEDS: 2 APPROACHES • Pop III remnants – ~100 (? ) M BHs form at z~20 – Sink to center of merged haloes – Accrete gas/merge with other seeds • Direct collapse – Initial BH mass = ? – How is fragmentation avoided? – Growth mainly by accretion of gas • Problems faced by both: – angular momentum transport – Can be exceeded?

• WHAT ANGULAR MOMENTUM PROBLEM? • WHOSE LIMIT? EDDINGTON • SEEDS FROM DIRECT COLLAPSE? • OBSERVING SEED BH FORMATION

Halo with slight rotation DM Gas collapses if DM gas Dynamical loss of angular momentum through nested gas global gravitational instabilities “BARS WITHIN BARS” (Shlosman, Frank & Begelman 89)

WHEN IS GAS UNSTABLE? • Init. collapse with conserved ang. mom. – Mestel, rigid, exponential disk – Include NFW halo potential – Vary fraction of gas collapsing, fd • Choose halo ang. mom. parameter λspin from lognormal distribution • Stability threshold: f T/W > 0. 25 – Christodoulou et al. 1995, f = form factor

Probability of instability, lognormal distibution Max. halo spin parameter for instability, as function of gas infall fraction STABLE rigid Mestel UNSTABLE exponential

Global ang. mom. transport (“Bars within Bars”) : M yr-1 • Mass distribution isothermal, • Accumulated mass dominates for • BWB fails at ?

Local ang. mom. transport ( ): M yr-1 viscosity parameter (~1: Gammie 2001) • If T const. or incr. inwards, likely to reach center FRAGMENTATION UNLIKELY TO STOP INFALL

TEMP. /METALLICITY DEPENDENCE • – Metal-free, H 2 -dissociated gas falls in • (Metal rich and/or H 2 fragments? ) – Inflow rate, mass accumulation HIGH FAVORS DIRECT BH FORMATION • – Infall requires H 2 or pollution by Pop III – Inflow rate, mass accumulation LOW FAVORS POP III STAR FORMATION

WHOSE LIMIT? SUPPOSE A SEED BH SETTLES IN THE MIDDLE OF THE ACCUMULATED GAS Max. BH accretion rate is for the mass of the ENVELOPE BH

RAPID GROWTH OF A PREGALACTIC BH + FORMATION OF THE SEED? ACCUMULATED GAS “QUASISTAR” Pregalactic halo Could seed BH grow from ~10 to >105 MSol at ? (Begelman, Volonteri & Rees 06)

WHAT IS A “QUASISTAR”? • Self-gravitating structure laid down by infall • Radiation-dominated, rotation crucial • Pre-BH: – Entropy small near center, increases with r – Very different from the supermassive stars postulated by Hoyle and Fowler YOU’VE GOT TO BE KIDDING

WHAT IS A “QUASISTAR”? • Self-gravitating structure laid down by infall THAT’S • Radiation-dominated, rotation crucial BETTER • Pre-BH: – Entropy small near center, increases with r – Very different from the supermassive stars postulated by Hoyle and Fowler • Post-BH: – “Nuclear” energy source is BH accretion – Expands and becomes fully convective – Like radiation-dominated, metal-free red giant

QUASISTAR STRUCTURE : PRE-BH • Mass m* (M ) increases with time • Core with • Envelope M yr-1 – Entropy increases outward – convectively stable – Rotation increases binding energy • Outer radius • Core radius constant shrinks – Nuclear burning inadequate to unbind star • Core mass ~ 10 M constant • When core temp. rapid cooling by thermal neutrinos

CORE COLLAPSE AND FORMATION OF ~10 -20 M SEED BH + SUBSEQUENT ACCRETION AT EDDINGTON LIMIT FOR QUASISTAR MASS

QUASISTAR STRUCTURE : POST-BH • BH accretes adiabatically from quasistar interior • Adjusts so energy liberated • Radiation-supported convective envelope (w/rotation) – Central temp drops to ~ 106 K – Radius expands to ~ 100 AU – Convection becomes inefficient well inside photosphere, where , but photosphere temp. drops as BH grows • Teff < 5000 K opacity crisis

OPACITY CRISIS • Pop III opacity plummets below T~5000 K – Higher temp than for solar abundances

Mayer & Duschl 2005 Metal-free opacities

Mayer & Duschl 2005 Metal-free opacities Analogous to Hayashi track, but match to radiationdominated convective envelope Plausible range of photosphere densities

Mayer & Duschl 2005 Est. min. temp. If Tphot drops below minimum (~5000 K), flux inside quasistar exceeds Eddington limit, dispersing it.

OPACITY CRISIS • Pop III opacity plummets below T~5000 K – Higher temp than for solar abundances • Convection is relatively inefficient – Transition to radiative envelope at T~55, 000 K ~ 10 Tphot – κ ~ κes at transition, but much lower at photosphere • If Tphot < Tmin , flux exceeds Eddington limit at transition quasistar disperses • Connect BH accretion physics to quasistar structure, predict Tphot(m. BH, m*)

CONNECTION TO BH ACCRETION Once limiting temperature is reached, … dispersal is inevitable (and accelerates)

QUASISTAR MODELS WITH “TOY” OPACITY Begelman, Rossi & Armitage 2007

GROWING THE BLACK HOLE • Eddington-limited growth: • Ignoring opacity crisis: M yr-1 – Super-Eddington growth to ~107 M in 108 yr • Onset of opacity crisis: Teff ~ 5000 K when – Estimates sensitive to: • Value of Tmin • Details of BH accretion • Energy loss details: rotational funnel, photon bubbles….

CAN QUASISTARS BE DETECTED?

DETECTING A QUASISTAR • Most time spent as ~5000 K blackbody • Radiates at Eddington limit for 105 m 5 M • Max flux ~

JWST z=6 z=10 z=20 PEAK OF BLACKBODY

DETECTING A QUASISTAR • Most time spent as ~5000 K blackbody • Radiates at Eddington limit for 105 m 5 M • Max flux ~ • Better to observe @ 3. 5µm, on Wien tail • Corona/mass loss hard tail, easier detection

JWST z=6 z=10 z=20 WIEN TAIL – MASS LOSS MAY IMPROVE FURTHER

Nseeds (Mpc-3) HOW COMMON ARE QUASISTARS? 100 per L* galaxy 1 Cumulative comoving no. density of seeds …but their lifetimes are short

COMOVING DENSITY OF QUASISTARS Lifetime = 106 yr Metal enrichment model from Scannapieco et al. 2003 No enrichment L* galaxies Some enrichment Heavy enrichment

COMOVING DENSITY OF QUASISTARS Lifetime = 106 yr All 104 K haloes 100 per L* galaxy 104 K haloes with λ<0. 02 1 per L* galaxy Metal enrichment model from Scannapieco et al. 2003

DENSITY ON SKY • • 10 -3 seeds Mpc-3 (1/L* gal. ) • JWST FOV ~ 4. 8 arcmin 2 ~ 1 per random field 103. 5 QS deg. -2 Could be ~10 -100 x more common • How to distinguish from other objects? – Colors: ~pure blackbody (not dust reddened) • Observe on Wien tail • No lines (distinguish from T dwarfs) – Unresolved (distinguish from nearby starbursts) – Clustering (like 104 K haloes)

QUASISTAR DENSITY ON SKY Lifetime = 106 yr 10/JWST field 1/JWST field

WHAT HAPPENS NEXT? • If super-Edd. phase extends beyond opacity crisis, BH seeds could be as massive as 106 M • Worst case: super-Edd. phase ends at ~103 M • 10 t. Salpeter between z=10 and z=6 growth by (only) 20, 000 BUT – Exceeding LEdd by factor 2 squares growth factor! – Mergers can account for factor 10 -100 of growth

CONCLUSIONS I I. Angular momentum transport is fast if global gravitational instabilities set in (“Bars within Bars”) II. BH can grow at Eddington limit for the surrounding envelope, which can be cvcvcvcfor the BH III. BH seed can form in situ from the infalling envelope itself (aided by ν cooling) or can be captured Pop III remnant

CONCLUSIONS II IV. BH seeds grow inside a “quasistar” powered by BH accretion, with a radiation pressure-supported convective envelope V. Min. Teff of quasistar is ~5000 K, lifetime is > 106 yr VI. Quasistars could be common and may be detectable by JWST