Download Black Hole Accretion

Black Hole Event
Horizon
Ramesh Narayan
Can We Be Sure That We
Have Discovered Black Holes?

Not really…

Astronomers have shown that black hole candidates (BHCs) have
M  3M, so they are not NSs

The mass gives us a good reason to suspect that the objects may
be BHs

But we need to find some independent evidence that our BHCs
actually possess Event Horizons, before we can be sure that
they are really BHs

How to do this? Compare BH and NS systems and look for some
dramatic difference that implies lack of surface
Signatures of the Event
Horizon

Differences in quiescent luminosity (Narayan, Garcia &
McClintock 1997, Garcia et al. 2001; Narayan et al. 2002;…)

Differences in variability power spectra (Sunyaev & Revnivtsev
2000)

Differences in Type I X-ray bursts (Narayan & Heyl 2002)

Differences in X-ray colors (Done & Gierlinsky 2003)

Differences in thermal surface emission (McClintock, Narayan &
Rybicki 2003)
Quiescent
Transients
BHCs are at least 100 times
fainter than NSs (Narayan,
Garcia & McClintock 1997; Garcia et
al. 2001;…)
Such a large difference can be
explained if BHs have event
horizons and NSs do not.
Hard to explain otherwise
Spectra of Quiescent NS SXTs

Quiescent NS SXT spectra have two components:
soft thermal + power-law

Soft component typically has L ~ 1033 erg/s

Interpreted as heat from deep crustal nuclear
reactions (Brown, Bildsten & Rutledge 1998), which
is generated during accretion outbursts, and escapes
during quiescence

Predicted flux agrees with observations

If BHCs have surfaces, then they should show a
similar thermal component --- unavoidable
Spectra of Quiescent BH SXTs


Spectra are all consistent with pure power-law
Severe limit L < 1031 erg/s on thermal component in XTE J1118+480
(McClintock, Narayan & Rybicki 2003)

Predicted luminosity: 3x1032 to 1034 erg/s. Huge discrepancy

Obvious explanation: no surface in BHCs, i.e., event horizon
Type I X-ray Bursts




Discovered by Grindlay
et al. (1976)
Sudden brightening,
once every several hrs;
lasts about 10-100 s
Physics understood:
unstable nuclear burning
of accreted gas
Very common in NS XRBs
No Type I Bursts in BHCs!!

No BHC has ever shown a Type I burst

Obvious explanation: They have event horizons, so
material cannot pile up, and there can be no bursts

But, is the lack of a surface
the only reason why the
sources do not burst?

Not obvious…

Some NS XRBs don’t have
bursts
Modeling Type I Bursts
(Narayan & Heyl 2002, 2003)

We follow the accretion layer as it builds up. For each choice of
the accretion column:

We solve for the equilibrium profiles of density , temperature T,
radiative flux F, and H/He fractions X, Y, as a function of depth
(column density  g/cm2) in the accretion layer

We then carry out a formal linear stability analysis of the layer
(this is new):-
if stable, the object will have no bursts

if unstable, the object will have Type I bursts
The Governing Equations
There are 5 coupled partial
differential equations:
hydrostatic balance, radiative
transfer, energy conservation,
hydrogen burning, helium
burning (just like stellar
astrophysics!)
Plus equation of state, expressions
for radiative and conductive opacity,
and hydrogen and helium burning
rates
Parameters: gravity g, mass
accretion rate d/dt, initial H and
He fractions X0 and Y0, and core
temperature Tcore (inner bc)
P
 g , P  P(  , T ),

T
3
1
1
1

F,


,
3
 4acT
  rad  cond
ds
F
T
   H   He  
,
dt

dX
H
 * ,
dt
EH
dY
dX  He

 * ,
dt
dt EHe
d
  
  .
dt t

What is New in this Work?

We carry out a full stability analysis, using the
(complex) growth rate as an eigenvalue

We are more quantitative than previous work (e.g.,
Hansen & van Horn 1975; Fujimoto et al. 1981, 1987;
Paczynski 1983; Fushiki & Lamb 1988; Cumming &
Bildsten 2000;…)

But, we cannot study the burst itself --- that requires
full time-dependent simulations
Results for a 1.4M Neutron
Star

We have tried a variety of
accretion rates (0.001 -- 1 LEdd)
and NS radii (1.6 -- 4 Schw. radii,
i.e., 6.5 -- 16 km), for three
choices of the core temperature

We find that accreting NSs will
produce Type I bursts over a
range of accretion rates, but no
bursts when close to Eddington

Pretty good agreement with data
M=1.4M, Tcore=108K
Narayan & Heyl (2003)
Structure of Black
Hole Candidates

If BHCs are not black holes, then
what kind of objects are they?

BHCs may be just like NSs (though
more massive), with a hard surface
on which gas accumulates

BHCs may be made of porous (dark)
matter through which gas falls to
collect at the center
10M Object With a Hard
Surface

We tried a variety of accretion
rates (0.001 -- 1 Eddington)
and BH radii (9/8 -- 3 Schw.
radii, i.e., 33 -- 85 km), with
different choices of Tcore

We find that BH candidates
should exhibit Type I bursts
over a wide range of
conditions (wider than NSs!)
M=10M, Tcore=107.5K
Narayan & Heyl (2002)
So Why Don’t BH XRBs Burst?

Wrong fuel (no H, He) X

Wrong core temperature X

Wrong burst recurrence time X

Wrong accretion rate X

Rotation X

Magnetic fields X

Burning front propagation ?

Exotic matter ?
Exotic Matter?

What if our BH candidate has some exotic
interior? --- pion or kaon condensate or
hyperons or quark star matter (or something
even more exotic)

The burst phenomenon is limited to quite a low
density (typically ~106 g/cm3) on the surface.
Weird interior at 1014 -- 1015 g/cm3 has no
effect on surface bursts
Non-Interacting Exotic
Matter?




What if BH candidates are
made of non-interacting
particles?
Accreting gas may simply sink
to the center and form a
separate compact sphere
Would these objects burst?
Our calculations indicate Yes
(Yuan, Narayan & Rees 2003,
in preparation)
Fermion-Fermion and
Boson-Fermion Stars

Assume that BHC consists of two independent fluids
interacting via gravity




Fermionic dark matter (mf ~ 200 MeV) plus gas (this is a
fermion-fermion star)
Bosonic dark matter (mb ~ 10-18 MeV) plus gas (this is a
boson-fermion star)
Calculate surface gravity, redshift, etc. at the surface
of the gas sphere
Evaluate burst properties of the models (Yuan,
Narayan & Rees 2003)
Results
Bursts are unavoidable on fermion-fermion and boson-fermion
stars (Yuan et al. 2003)
So Why Don’t
BH XRBs Burst?

We have eliminated
virtually every other
explanation

The only remaining explanation
is that BHCs have event
horizons: no surface, no burst
Summary





Are BHCs really black holes with event horizons?
Several independent arguments suggest that the
answer is Yes. Absence of Type I X-ray bursts is an
especially strong argument.
Each argument has loopholes (hardly any for bursts)
If BHCs are not black holes, they must negotiate so
many different constraints, it becomes highly
implausible
We can be pretty confident that they are black holes!