Black Holes: A General Introduction
Black Holes: A General Introduction
Black Holes: A General Introduction
Jean-Pierre Luminet
Abstract. Our understanding of space and time is probed to its depths by black holes.
These objects, which appear as a natural consequence of general relativity, provide a
powerful analytical tool able to examine macroscopic and microscopic properties of the
universe. This introductory article presents in a pictorial way the basic concepts of
black holes theory, as well as a description of the astronomical sites where black holes
are suspected to lie, namely binary Xray sources and galactic nuclei.
Let me begin with an old Persian story. Once upon a time, the butterflies organized a summer school devoted to the great mystery of the flame. Many discussed
about models but nobody could convincingly explain the puzzle. Then a bold
butterfly enlisted as a volunteer to get a real experience with the flame. He flew
off to the closest castle, passed in front of a window and saw the light of a candle.
He went back, very excited, and told what he had seen. But the wise butterfly
who was the chair of the conference said that they had no more information
than before. Next, a second butterfly flew off to the castle, crossed the window
and touched the flame with his wings. He hardly came back and told his story;
the wise chairbutterfly said your explanation is no more satisfactory. Then
a third butterfly went to the castle, hit the candle and burned himself into the
flame. The wise butterfly, who had observed the action, said to the others: Well,
our friend has learned everything about the flame. But only him can know, and
thats all.
As you can guess, this story can easily be transposed from butterflies to scientists confronted with the mystery of black holes. Some astronomers, equipped
with powerful instruments such as orbiting telescopes, make very distant and
indirect observations on black holes; like the first butterfly, they acknowledge
the real existence of black holes but they gain very little information on their
real nature. Next, theoretical physicists try to penetrate more deeply into the
black hole mystery by using tools such as general relativity, quantum mechanics and higher mathematics; like the second butterfly, they get a little bit more
information, but not so much. The equivalent of the third butterfly would be
a spationaut plunging directly into a black hole, but eventually he will not be
able to go back and tell his story. Nevertheless, by using numerical calculations
such as those performed at the Observatoire de Meudon I will show you later,
outsiders can get some idea of what happens inside a black hole.
2
2.1
Let us begin to play like the second butterfly, and explore the black hole from
the point of view of theoretical physics. An elementary definition of a black hole
is a region of space-time in which the gravitational potential, GM/R, exceeds
the square of the speed of light, c2. Such a statement has the merit to be independent of the details of gravitational theories. It can be used in the framework
of Newtonian theory. It also provides a more popular definition of a black hole,
according to which any astronomical body whose escape velocity exceeds the
speed of light must be a black hole. Indeed, such a reasoning was done two centuries ago by John Michell and Pierre-Simon de Laplace. In the Philosophical
Transactions of the Royal Society (1784), John Michell pointed out that if the
semi diameter of a sphere of the same density with the sun were to exceed that
of the sun in the proportion of 500 to 1, (...) all light emitted from such a body
would be made to return towards it, and independently, in 1796, Laplace wrote
in his Exposition du Syst`eme du Monde: Un astre lumineux de meme densite
que la terre et dont le diam`etre serait deux cents cinquante fois plus grand que
celui du soleil, ne laisserait, en vertu de son attraction, parvenir aucun de ses
rayons jusqu`
a nous ; il est donc possible que les plus grands corps lumineux
de lunivers soient, par cela meme, invisibles. Since the density imagined at
this time was that of ordinary matter, the size and the mass of the associated
invisible body were huge - around 107 solar masses, corresponding to what
is called today a supermassive black hole. Nevertheless, from the numerical
figures first proposed by Michell and Laplace, one can recognize the well-known
basic formula giving the critical radius of a body of mass M :
RS =
M
2GM
km,
3
c2
M
(1)
where M is the solar mass. Any spherical body of mass M confined within the
critical radius RS must be a black hole.
These original speculations were quickly forgotten, mainly due to the development of the wave theory of light, within the framework of which no calculation
of the action of gravitation on light propagation was performed. The advent of
general relativity, a fully relativistic theory of gravity in which light is submitted to gravity, gave rise to new speculations and much deeper insight into black
holes.
To pictorially describe black holes in space-time, I shall use light cones. Let
me recall what a light cone is. In figure 1, a luminous flash is emitted at a
given point of space. The wavefront is a sphere expanding at a velocity of
c = 300 000 km/s, shown in a) at three successive instants. The light cone
representation in b) tells the complete story of the wavefront in a single spacetime diagram. As one space dimension is removed, the spheres become circles.
The expanding circles of light generate a cone originating at the emission point.
If, in this diagram, we choose the unit of length as 300 000 km and the unit of
time as 1 second, all the light rays travel at 45 .
t3
Time
t2
t3
lig matte
r
ht
t1
t2
1 sec
forbidden
Space
t1
Sp
ac
300000 km
Figure 1
Fig. 1 The light cone.
The light cone allows us to depict the causal structure of any space-time. Take
for instance the Minkowski flat space-time used in Special Relativity (figure 2).
At any event E of space-time, light rays generate two cones (shaded zone). The
rays emitted from E span the future light cone, those received in E span the past
light cone. Physical particles cannot travel faster than light: their trajectories
remain confined within the light cones. No light ray or particle which passes
Particle
Future
Time
Particle
of E
Particle
Light ray
Space
Time
Light
rays
Past
Space
of E
Figure 2
Figure 3
Fig. 2&3 The space-time continuum of Special Relativity and the soft space-time of
General Relativity.
through E is able to penetrate the clear zone. The invariance of the speed of light
in vacuum is reflected by the fact that all the cones have the same slope. This
is because the space-time continuum of Special Relativity, free from gravitating
matter, is flat and rigid. As soon as gravity is present, space-time is curved
and Special Relativity leaves room to General Relativity. Since the Equivalence
Principle states the influence of gravity on all types of energy, the light cones
follow the curvature of space-time (figure 3). They bend and deform themselves
according to the curvature. Special Relativity remains locally valid however: the
worldlines of material particles remain confined within the light cones, even when
the latter are strongly tilted and distorted by gravity.
2.2
Spherical collapse
distant
astronomer
singularity
r=0
R3
event
horizon
r=2
light
cones
outgoing
light
rays
R2
R1
E4
E3
Time
E2
E1
Space
ingoing
light
rays
surfac
e of the star
Figure 4
Fig. 4 A space-time diagram showing the formation of a black hole by gravitational
collapse.
big one, so that it is not destroyed too quickly by the tidal forces. On board the
ship, the commander sends a solemn salute to mankind, just at the moment when
the ship crosses the horizon. His gesture is transmitted to distant spectators via
television. The film on the left shows the scene on board the spaceship in proper
time, that is, as measured by the ships clock as the ship falls into the black hole.
The astronauts salute is decomposed into instants at proper time intervals of
0.2 second. Crossing of the event horizon (black holes have not a hard surface) is
not accompanied by any particular event. The film on the right shows the scene
received by distant spectators via television. It is also decomposed into intervals
of apparent time of 0.2 second. At the beginning of his gesture, the salute is
slightly slower than the real salute, but initially the delay is too small to be
noticed, so the films are practically identical. It is only very close to the horizon
that apparent time starts suddenly to freeze; the film on the right then shows
the astronaut eternally frozen in the middle of his salute, imperceptibly reaching
the limiting position where he crossed the horizon. Besides this effect, the shift
in the frequencies in the gravitational field (the so-called Einsteins effect) causes
the images to weaken, and they soon become invisible.
All these effects follow rather straightforwardly from equations. In General
Relativity, the vacuum space-time around a spherically symmetric body is described by the Schwarzschild metric:
1
2M
2M
dt2 + 1
dr2 + r2d 2 ,
ds2 = 1
r
r
(2)
4M r 3/2
3
2M
(3)
and is well-behaved at the event horizon. The apparent time (as measured by a
distant observer) is given by
p
r 1/2
r/2M + 1
+ 2M ln p
,
(4)
t = 4M
2M
r/2M 1
Proper time
in seconds
Film A
Film B
0.00
Apparent time
in seconds
0.00
0.20
0.20
0.40
0.40
0.60
0.60
0.80
Crossing
the horizon
0.80
1.00
1000.00
1.20
Figure 5
Fig. 5 The astronauts salute.
r
1)
2M
(5)
Distance
Horizon
Apparent time
Proper time
Singularity
Time
Figure 6
Fig. 6 The two times of a black hole.
2M
)dv2 + 2dvdr + r 2 d 2.
r
(6)
Now the light cones are perfectly well behaved. The ingoing light rays are given
by
dv = 0,
(7)
the outgoing light rays by
dv =
2dr
.
1 2M
r
(8)
The metric can be analytically continued to all r > 0 and is no more singular at
r = 2M . Indeed, in figure 4 such a coordinate system was already used.
2.3
A black hole may well form from an asymmetric gravitational collapse. However
the deformations of the event horizon are quickly dissipated as gravitational
radiation; the event horizon vibrates according to the socalled quasi-normal
modes and the black hole settles down into a final axisymmetric equilibrium
configuration.
The deepest physical property of black holes is that asymptotic equilibrium
solutions depend only on three parameters: the mass, the electric charge and
the angular momentum. All the details of the infalling matter other than mass,
electric charge and angular momentum are washed out. The proof followed from
efforts over 15 years by half a dozen of theoreticians, but it was originally suggested as a conjecture by John Wheeler, who used the picturesque formulation
a black hole has no hair. Markus Heuslers lectures in this volume will develop
this socalled uniqueness theorem.
horizon
time
Gravitational
waves
Space
Collapsing star
Figure 7
Fig. 7 Gravitational collapse of a star.
2
2
2
2
+(r2 + a2 + 2M rasin )sin2 d2 +
dr + d
ds2 = (1
(9)
10
From this formula we can see, however, that the black hole parameters cannot
be arbitrary. Electric charge and angular momentum cannot exceed values corresponding to the disappearance of the event horizon. The following constraint
must be satisfied: a2 + Q2 M 2.
When the condition is violated, the event horizon disappears and the solution
describes a naked singularity instead of a black hole. Such odd things should not
exist in the real universe (this is the statement of the socalled Cosmic Censorship Conjecture, not yet rigorously proved). For instance, for uncharged rotating
configuration, the condition Jmax = M 2 corresponds to the vanishing of surface
gravity on the event horizon, due to centrifugal forces; the corresponding solution is called extremal Kerr Solution. Also, the maximal allowable electric
charge is Qmax = M 1040e M/M , where e is the electron charge, but it is
to be noticed that in realistic situations, black holes should not be significantly
charged. This is due to the extreme weakness of gravitational interaction compared to electromagnetic interaction. Suppose a black hole forms with initial
positive charge Qi of order M . In realistic conditions, the black hole is not isolated in empty space but is surrounded by charged particles of the interstellar
medium, e.g. protons and electrons. The black hole will predominantly attract
electrons and repel protons with charge e by its electromagnetic field, and predominantly attract protons of mass mp by its gravitational field. The repulsive
electrostatic force on protons is larger than the gravitational pull by the factor
eQ/mp M e/mp 1018. Therefore, the black hole will neutralize itself almost
instantaneously. As a consequence, the Kerr solution, obtained in equation (9)
by putting Q = 0, can be used for any astrophysical purpose involving black
holes. It is also a good approximation to the metric of a (not collapsed) rotating
star at large distance, but it has not been matched to any known solution that
could represent the interior of a star.
The Kerr metric in Boyer-Lindquist coordinates has singularities on the axis
of symmetry = 0 obviously a coordinate singularity
and for = 0. One
There is a deep analogy between a rotating black hole and the familiar phenomenon of a vortex - such as a giant maelstrom produced by sea currents. If
we cut a light cone at fixed time (a horizontal plane in figure 8), the resulting
spatial section is a navigation ellipse which determines the limits of the permitted trajectories. If the cone tips over sufficiently in the gravitational field,
11
static
limit
navigation
circle
ship
SPACE-TIME
direction
of rotation
SPACE
Figure 8
Vortex
Figure 9
the navigation ellipse detaches itself from the point of emission. The permitted
trajectories are confined within the angle formed by the tangents of the circle,
and it is impossible to go backwards.
This projection technique is useful to depict the causal structure of spacetime around a rotating black hole (figure 9). The gravitational well caused by
a rotating black hole resembles a cosmic maelstrom. A spaceship travelling in
the vicinity is sucked towards the centre of the vortex like a boat. In the region
outside the so-called static limit (clear), it can navigate to whereever it wants. In
the zone (in grey) comprised between the static limit and the event horizon, it
is forced to rotate in the same direction as the black hole; its ability to navigate
freely is decreased as it is sucked inwards, but it can still escape by travelling in
an outwards spiral. The dark zone represents the region inside the event horizon:
any ship which ventured there would be unable to escape even if it was travelling
at the speed of light. A fair illustration is the Edgar Poes short story: A descent
into the maelstrom (1840).
The
static limit is a hypersurface of revolution, given by the equation r =
M + M 2 a2cos2 . As we can see from figure 10, it intersects the event horizon
at its poles = 0, but it lies outside the horizon for other values of . The
region between the static limit and the event horizon is called ergoregion. There,
all stationary observers must orbit the black hole with positive angular velocity.
The ergoregion contains orbits with negative energy. Such a property has lead to
the idea of energy extraction from a rotating black hole. Roger Penrose (1969)
suggested the following mechanism. A distant experimentalist fires a projectile
in the direction of the ergosphere along a suitable trajectory (figure 10). When
it arrives the projectile splits into two pieces: one of them is captured by the
black hole along a retrograde orbit, while the other flies out of the ergosphere
12
Recovered
fragment
Rotation
axis
Static
limit
Ergosphere
Event
Horizon
Captured
fragment
Disintegration
Initial
projectile
Figure 10
Fig. 10 Cross-section of a rotating black hole.
13
cannot be lowered by classical (e.g. non quantum) processes. The maximum extractable energy is as high a 29 per cent for rotational energy and 50 per cent
for Coulomb energy. It is much more efficient that, for instance, nuclear energy
release (0.7 per cent for hydrogen burning).
2.5
In black hole dynamics: the infinitesimal variations of the mass M , the charge
Q and the angular momentum J of a perturbed stationary black hole are
g
dA + H dJ + H dQ.
related by dM = 8
Second Law.
In thermodynamics, entropy can never decrease: dS 0.
In black hole dynamics, the area of event horizon can never decrease: dA 0.
14
Growth of a
black hole
Collision of two
black holes
A 1 > A1+ A2
t2
Formation of a
black hole
A2
t1
time
A1
space
Formation
Figure 11
Fig. 11 The irreversible growth of black holes.
This law implies for instance that the area of a black hole resulting from
the coalescence of two parent black holes is greater than the sum of areas of
the two parent black holes (see Figure 11). It also implies that black holes
cannot bifurcate, namely a single black hole can never split in two parts.
Third law.
In thermodynamics, it reflects the inaccessibility of the absolute zero of temperature, namely it is impossible to reduce the temperature of a system to
zero by a finite number of processes.
In black hole mechanics, it is impossible to reduce the surface gravity to zero
by a finite number of operations. For Kerr black holes, we have seen that
zero surface gravity corresponds to the extremal solution J = M 2 .
It is clear that the area of the event horizon plays formally the role of an
entropy, while the surface gravity plays the role of a temperature. However,
as first pointed out by Bekenstein, if black holes had a real temperature like
thermodynamical systems, they would radiate energy, contrarily to their basic
definition. The puzzle was solved by Hawking when he discovered the evaporation
of mini-black holes by quantum processes.
2.6
The details of Hawking radiation and the - not yet solved - theoretical difficulties linked to its interpretation are discussed by other lecturers (Gerard t Hooft,
15
Evaporation
+E
-E
+E
-E
III
Quantum
vacuum
II
-E
-E
Time
IV
horizon
+E
I
+E
Space
Figure 12
Fig. 12 The quantum evaporation of a mini black hole by polarisation of the vacuum.
Andreas Wipf and Claus Kiefer) in this volume. Therefore I shall only present
the basic idea in a naive pictorial way (figure 12). The black holes gravitational
field is described by (classical) general relativity, while the surrounding vacuum
spacetime is described by quantum field theory. The quantum evaporation process is analogous to pair production in a strong magnetic field due to vacuum
polarization. In the Fermi sea populated by virtual pairs of particles-antiparticles
which create and annihilate themselves, the four various possible processes are
depicted schematically in figure 12.
Some virtual pairs emerging from the quantum vacuum just annihilate outside the horizon (process I). Some pairs produced in the vicinity of the black
hole disappear completely in the event horizon (process IV). Some pairs are
splitted, one particle (or antiparticle) escaping the black hole, the other one being captured (processes II and III). The calculations show that the process II
is dominant, due to the (classical) gravitational potential which polarizes the
quantum vacuum. As a consequence, a black hole radiates particles with a thermal spectrum characterized by a blackbody temperature precisely given by the
formula suggested by the thermodynamical analogy:
T =h
M
g
= 107
K,
2
M
(11)
where h
is Plancks constant. We immediately see that the temperature is completely negligible for any astrophysical black hole with mass comparable or
greater to the solar mass. But for mini-black holes with mass 1015g (the typ-
16
ical mass of an asteroid), the Hawking temperature is 1012K. Since the black
hole radiates away, it looses energy and evaporates on a timescale approximately
given by
3
M
10
(12)
tE 10 years
1015grams
Thus, mini-black holes whose mass is smaller than that of an asteroid (and size
less than 1013cm) evaporate on a timescale shorter than the age of the universe.
Some of them should evaporate now and give rise to a huge burst of high energy
radiation. Nothing similar has ever been observed (-ray bursts are explained
quite differently). Such an observational constraint thus limits the density of
mini-black holes to be less than about 100/(lightyear)3 .
The black hole entropy is given by
S=
kB A
h
4
(13)
Space-time mappings
17
curved 2geometry with metric (1 2Mr(r) )1dr2 + r2 d2. Such a surface can be
2
2
visualized by embedding it in Euclidean 3space ds2 = dz 2 +
pdr + d . For a
noncollapsed star with radius R, the outer solution z(r) = 8M (r 2M ) for
r R 2M is p
asymptotically flat and matches exactly the nonsingular inner
solution z(r) = 8M (r)(r 2M (r)) for 0 r R (figure 13). For a black hole,
the embedding is defined only
p for r 2M . The corresponding surface is the
Flamm paraboloid z(r) = 8M (r 2M ). Such an asymptotically flat surface
exhibits two sheets separated by the Schwarzschild throat of radius 2M . The
two sheets can be either considered as two different asymptotically flat parallel
universes (whatever the physical meaning of such a statement may be) in which a
black hole in the upper sheet is connected to a timereversed white hole in the
lower sheet (figure 14), or as a single asymptotically flat spacetime containing a
pair of black/white holes connected by a socalled wormhole (figure 15). The
freedom comes from the topological indeterminacy of general relativity, which
allows us to identify some asymptotically distant points of spacetime without
changing the metric.
BLACK HOLE
"OUR"
UNIVERSE
star
Schwarschild
throat
"ANOTHER"
UNIVERSE
horizon
WHITE HOLE
Figure 13
Figure 14
However the embedding technique does not give access to the regions of
space-time inside the event horizon.
Kruskal diagram To explore inner space-time we use the maximal analytical
extension of the Schwarzschild metric. This is achieved by means of a coordinate
transformation discovered by Kruskal:
u2 v 2 = (
r
1)er/2M
2M
18
WHITE HOLE
BLACK HOLE
Horizon
Horizon
WORMHOLE
Figure 15
Fig. 15 A wormhole in space-time.
v
u
t
coth 4M
=
1
t
tanh 4M
< 2M
for r = 2M
> 2M
(15)
32M 3 r/2M
e
(dv2 + du2) + r2 d 2
(16)
r
In the (v, u) plane, the Kruskal space-time divides into two outer asymptotically flat regions and two regions inside the event horizon bounded by the future
and the past singularities. Only the unshaded region is covered by Schwarzschild
coordinates. The black region does not belong to space-time. In the Kruskal diagram (figure 16), light rays always travel at 45, lines of constant distance r
are hyperbolas, lines of constant time t pass through the origin. The interior of
the future event horizon is the black hole, the interior of the past event horizon
is the white hole. However it is clear that the wormhole cannot be crossed by
timelike trajectories: no trajectory can pass from one exterior universe to the
other one without encountering the r = 0 singularity.
Moreover, the Kruskal extension is a mathematical idealization of a spherical
black hole since it implicitly assumes that the black hole exists forever. However
in the physical universe, a black hole is not inscribed in the initial conditions of
the universe, it may form only from gravitational collapse. In such a case, one
gets a truncated Kruskal diagram (figure 17), which indicates that only the
future event horizon and the future singularity occur in a single asymptotically
flat spacetime. Such a situation offers no perspective to spacetime travelers!
ds2 =
2M
,t
=
19
riz
on
D
OUR
UNIVERSE
distant
observer
(r = const)
st
ho
F
on
riz
white
hole
onst
t=c
ho
black
hole
pa
onst
t=c
fu
tu
re
ANOTHER
UNIVERSE
r=
singularity
r=0
,t
2M
r=
r=0
singularity
Figure 16
Fig. 16 Exploring a spherical black hole using Kruskals map.
not bring much more information than the Kruskal one, but it turns out to be
the best available tool to reveal the complex structure of a rotating black hole.
Figure 18 shows the manyfingered universe of a Kerr black hole; it suggests
that some timelike trajectories (B, C) may well cross the outer EH and inner
IH event horizons and pass from an asymptotically flat external universe to
another one without encountering a singularity. This is due to the fact that the
singularity S is timelike rather than spacelike. Also, the shape of the singularity
is a ring within the equatorial plane, so that some trajectories (A) can pass
through the ring and reach an asymptotically flat spacetime inside the black
hole where gravity is repulsive. However, the analysis of perturbations of such
idealized Kerr spacetimes suggests that they are unstable and therefore not
physically plausible. Nevertheless the study of the internal structure of black
holes is a fascinating subject which is more deeply investigated in this volume
by Werner Israels lectures.
The fact that General Relativity does predict the existence of black holes and
that General Relativity is a reliable theory of gravitation does not necessarily
prove the existence of black holes, because General Relativity does not describe
the astrophysical processes by which a black hole may form.
20
BLACK HOLE
on
iz
Future
singularity
r=0
nt
re
or
e
Ev
EXTERNAL
UNIVERSE
tu
Fu
Formation
of black hole
STELLAR
SURFACE
A Beginning of
collapse
Equilibrium
STELLAR INTERIOR
Figure 17
Fig. 17 Truncated Kruskal map representing the collapse of a star into a black hole.
The basic process of stellar evolution is gravitational contraction at a rate controlled by luminosity. The key parameter is the initial mass. According to its
value, stars evolve through various stages of nuclear burning and finish their
lives as white dwarfs, neutron stars or black holes. Any stellar remnant (cold
equilibrium configuration) more massive than about 3M cannot be supported
by degeneracy pressure and is doomed to collapse to a black hole.
The figure 19 shows the stellar paths in a density-mass diagram according to
the most recent observational and theoretical data. Below 8M stars produce
white dwarfs, between 8 and 45M they produce neutron stars; black holes are
formed only when the initial mass exceeds 45M (we note on the diagram that
stars with initial mass between 20 and 40M suffer important mass losses at
21
antigravity
Universe
Universe
C
IH
IH
EH
antigravity
antigravity
Universe
IH
Universe
EH
B
Universe
another
EH
IH
IH
EH
another
antigravity
Universe
IH
another
EH
OUR
EH
Universe
EH
IH
time
antigravity
IH
EH
UNIVERSE
Universe
space
Figure 18
Fig. 18 Penrose map of a rotating black hole.
the stage of helium burning). Taking account of the stellar initial mass function,
one concludes that approximately 1 supernova over 100 generates a black hole
rather than a neutron star. Another possibility to form a stellar mass black hole
is accretion of gas onto a neutron star in a binary system until when the mass
of the neutron star reaches the maximum allowable value; then, gravitational
collapse occurs and a low mass black hole forms. Taking into account these
various processes, a typical galaxy like the Milky Way should harbour 10 7 108
stellar black holes.
22
10
7 9
60
20 40
120
H-ignition
He-ignition
C-ignition
10
O-ignition
w
dw hite
arf
s
Si-ignition
10
10
10
40
ne
u
sta tron 20
rs
14
les
ho
k
c
a
l
Figure 19
Fig. 19 The density-mass diagram of astronomical objects.
3.2
A massive black hole can form by continuous growth of a seed stellar mass
black hole, by gravitational collapse of a large star cluster or by collapse of a large
density fluctuation in the early universe (see next subsection). A well nourished
stellar mass black hole can grow to a supermassive black hole in less than a
Hubble time. Such a process requests large amounts of matter (gas and stars)
in the neighborhood, a situation than can be expected in some galactic nuclei.
A dense cluster of ordinary stars, such that the velocity dispersion vc v ,
where v 600 km/s is the typical escape velocity for main sequence stars,
first evolves through individual stellar burning; supernovae explosions lead to
the formation of compact remnants, e.g. neutron stars and stellar mass black
holes. A cluster of compact stars becomes relativistically unstable at sufficiently
high central gravitational redshift 1 + zc = (1 2M )1/2 1.5 (Zeldovich
23
and Podurets, 1965). Numerical simulations (e.g. Shapiro and Teukolsky, 1987,
Bisnovatyi-Kogan, 1988) confirm this scenario. Starting with 107 2.108 compact stars of 1 10M within a cluster radius r 0.01 0.1 pc and velocity
dispersion 800 2000 km/s, the evolution proceeds through three stages:
secular core collapse via the gravothermal catastrophe (long timescale)
short epoch dominated by compact star collisions and coalescences, leading
to the formation of black holes with mass M 90M
relativistic instability leading to a massive black hole surrounded by a halo
of stars.
3.3
Zeldovich in 1967 and Hawking in 1971 pointed out that it was in principle
possible to create a black hole with small mass (e.g. below the Chandrasekhar
limit) by applying a sufficiently strong external pressure. Such conditions could
have been achieved only in the very early universe. Gravitational forces may
locally halt the cosmic expansion of a clump of matter and reverse it into collapse
if the self-gravitational potential energy of the clump exceeds the internal energy:
GM 2
G2R5 pR3
R
(17)
Light cannot escape (classical) black holes but one can hope to detect them indirectly by observing the electromagnetic energy released during accretion processes.
24
Accretion of gas onto a compact star (neutron star or black hole in a binary system) releases energy in the X-ray domain, see S. Chakrabartis and J.
Tr
umpers lectures in this volume for the details. Search for stellar mass black
holes thus consists in locating rapidly variable binary X-ray sources which are
neither periodic (the corresponding X-ray pulsars are interpreted as rotating
neutron stars) nor recurrent (the corresponding X-ray bursters are interpreted
as thermonuclear explosions on a neutron stars hard surface). In spectroscopic
binaries, the Doppler curve of the spectrum of the primary (visible) star provides
the orbital period P of the binary and the maximum velocity v of the primary
projected along the line-of-sight. Keplers laws gives the following mass function
which relates observed quantities to unknown masses:
(Mc sin i)3
P v3
=
2G
(M + Mc )2
(18)
where Mc and M are the masses of the compact star and of the optical primary,
i the orbital inclination angle. A crucial fact is that Mc cannot be less than the
value of the mass function (the limit would correspond to a zero-mass companion
viewed at maximum inclination angle). Therefore the best black hole candidates
are obtained when the observed mass function exceeds 3M since, according
to the theory, a neutron star more massive than this limit is unstable and will
collapse to form a black hole. Otherwise, additional information is necessary
to deduce Mc : the spectral type of the primary gives approximately M , the
presence or absence of X-ray eclipses gives bounds to sin i. Hence Mc is obtained
within some error bar. Black hole candidates are retained only when the lower
limit exceeds 3M . At present day, about ten binary Xray sources provide
good black hole candidates. They can be divided into two families: the high
mass Xray binaries (HMXB), where the companion star is of high mass, and
the lowmass Xray binaries (LMXB) where the companion is typically below a
solar mass. The latter are also called Xray transients because they flare up
to high luminosities. Their mass properties are summarized in the table 1 below.
Other galactic Xray sources are suspected to be black holes on spectroscopic
or other grounds, see Chakrabartis lectures in this volume for developments. For
instance, some people argue that gamma-ray emission (above 100 keV ) emitted
from the inner edge of the accretion disc would attest the presence of a black hole
rather than a neutron star, because the high-energy radiation is scattered back by
the neutron stars hard surface and cools down the inner disc. If this is true, then
many gammaray novae in which no measurement of mass can be done (due to
the absence of optical counterpart or other limitations) are also good black hole
candidates. This is specially the case for Nova Aquila 1992 and 1 E 17407-2942,
two galactic sources which also exhibit radio jets. Such microquasars involving
both accretion and ejection of matter provide an interesting link between high
energy phenomena at the stellar and galactic scales, see J. Tr
umpers lectures.
25
Table 1 Stellar mass black hole candidates
mass function Mc /M
HMXB
3.5
M /M
Cygnus X-1
LMC X-3
LMC X-1
LMXB (Xray transients)
0.25
2.3
0.14
1121
5.6 7.8
4
2442
20
48
V 404 Cyg
A 0620-00
GS 1124-68 (Nova Musc)
GS 2000+25 (Nova Vul 88)
GRO J 1655-40
H 1705-25 (Nova Oph 77)
J 04224+32
6.07
2.91
3.01
5.01
3.24
4.65
1.21
1015
0.6
517
0.20.7
4.26.5
0.50.8
6-14
0.7
4.5 6.5
1.2
59
0.4
614 0.3 0.6
After the original speculations of Michell and Laplace, the idea of giant black
holes was reintroduced in the 1960s to explain the large amounts of energy released by active galactic nuclei (AGNs). This generic term covers a large family
of galaxies including quasars, radiogalaxies, Seyfert galaxies, blazars and so on,
for the classification see J. Tr
umpers, W. Collmars and V. Sch
onfelders lectures in this volume. The basic process is accretion of gas onto a massive black
hole. The maximum luminosity for a source of mass M , called the Eddington luminosity, is obtained by balance between gravitational attraction and radiation
pressure repulsion acting on a given element of gas. It is given by
L 1039
M
W
108M
(19)
The observed luminosities of AGNs range from 1037 1041 W, where the
higher values apply to the most powerful quasars. Then the corresponding masses
range from 106 1010M .
Due to constant improvements of observational techniques, it turned out in
the 1990s that most of the galactic nuclei (active or not) harbour large mass
concentrations. Today the detection of such masses is one of the major goals of
extragalactic astronomy. The most convincing method of detection consists in the
dynamical analysis of surrounding matter: gas or stars near the invisible central
mass have large dispersion velocities, which can be measured by spectroscopy.
It is now likely that giant black holes lurk in almost all galactic nuclei, the
energy output being governed by the available amounts of gaseous fuel. The
best candidates are summarized in Table 2.
26
For instance, our Galactic Centre is observed in radio, infrared, Xray and
gammaray wavelengths (other wavelengths are absorbed by dust clouds of the
galactic disc). A unusual radiosource has long been observed at the dynamical
centre, which can be interpreted as lowlevel accretion onto a moderately massive
black hole. However, a definite proof is not yet reached because gas motions are
hard to interpret. Recently Eckart and Genzel (1996) obtained a full three
dimensional map of the stellar velocities within the central 0.1 pc of our Galaxy.
The values and distribution of stellar velocities are convincingly consistent with
the hypothesis of a 2.5 106M black hole.
The nucleus of the giant elliptical M87 in the nearby Virgo cluster has also a
long story as a supermassive black hole candidate. Several independent observations are consistent with a 1 3 109M black hole accreting in a slow, inefficient
mode. A disc of gas is orbiting in a plane perpendicular to a spectacular jet; recent spectroscopic observations of the Hubble Space Telescope show redshifted
and blueshifted components of the disc, which can be interpreted by Doppler
effect as parts of the disc on each side of the hole are receding and approaching
from us.
The spiral galaxy NGC 4258 (M 106) is by far the best massive black hole
candidate. Gas motions near the centre has been precisely mapped with the 1.3
cm maser emission line of H2O. The velocities are measured with accuracy of 1
km/s. Their spatial distribution reveals a disc with rotational velocities following
an exact Keplers law around a massive compact object. Also the inner edge of
the disc, orbiting at 1080 km/s, cannot comprise a stable stellar cluster with the
inferred mass of 3.6 107M .
dynamics
maser
gas
gas
gas
stars
stars
stars
stars
stars
stars
stars
stars
27
The black hole in our galaxy and the massive black holes suspected in nearby
ordinary galaxies would be small scale versions of the cataclysmic phenomena
occurring in AGNs. But AGNs are too far away to offer a spectral resolution good
enough for dynamical measurements. Indeed, estimates of luminosities of AGNs
and theoretical arguments involving the efficiency of energy release in strong
gravitational fields invariably suggest that central dark masses are comprised
between 107 109 M . Variability of the flux on short timescales also indicates
that the emitting region has a small size; many AGNs exhibit large luminosity
fluctuations over timescales as short as one hour, which indicate that the emitting
region is smaller than one light-hour. Such large masses in such small volumes
cannot be explained by star clusters, so that accreting massive black holes remain
the only plausible explanation.
3.6
Stellar disruption
The accretion of gas at rate dM/dt and typical efficiency 0.1 produces a
luminosity
dM/dt
)
W
(20)
L 1039(
0.1 1M /year
By comparing the luminosity of the accretion model with the observed luminosities in AGNs, we conclude that the gas accretion rate must lie in the range
102 102 M /year. One is thus led to the question about the various gas production mechanisms able to fuel a giant black hole. An efficient process is mass
loss from stars passing near the hole. Current models of galactic nuclei involve a
massive black hole surrounded by a dense large cloud of stars. Diffusion of orbits
makes some stars to penetrate deeply within the gravitational potential of the
black hole along eccentric orbits. Disruption of stars can occur either by tidal
forces or by high-velocity interstellar collisions (figure 20). The collision radius
Rcoll 7 1018 108MM cm for a solartype star is the distance within which the
freefall velocity of stars becomes greater than the escape velocity at the stars
surface v (typically 500 km/s for ordinary stars); if two stars collide inside Rcoll
they will be partially or totally disrupted.
Also stars penetrating the critical tidal radius RT 6 1013 ( 108MM )1/3cm
for solartype stars will be ultimately disrupted, about 50 per cent of the released
gas will remain bound to the black hole. In some sense, the tidal encounter of a
star with a black hole can be considered as a collision of the star with itself. . .
In the collision process the factor = vrel /v plays a role analogous to the
penetration factor = RT /Rp in the tidal case (where Rp is the periastron
distance). As soon as 1 the stars are disrupted, and when 5 the stars
are strongly compressed during the encounter. Thus, in both cases, appears as
a crushing factor, whose magnitude dictates the degree of maximal compression
and heating of the star.
The first modelisation of the tidal disruption of stars by a big black hole was
done in the 1980s by myself and collaborators (see Luminet and Carter, 1986
28
collision
radius
tidal
radius
star
star 1
star 2
BH
vrel
Rp
RT
BH
Figure 20
Fig. 20 Tidal and collision radius.
and references therein). We have discovered that a star deeply plunging in the
tidal radius without crossing the event horizon is squeezed by huge tidal forces
and compressed into a short-lived, ultra-hot pancake configuration. Figure 21
shows schematically the progressive deformation of the star (the size of the star
has been considerably over-emphasized for clarity). The left diagram represents
the deformation of the star in its orbital plane (seen from above), the right one
shows the deformation in the perpendicular direction. From a to d the tidal forces
are weak and the star remains practically spherical. At e the star penetrates the
tidal limit. It becomes cigar-shaped. From e to g a mangle effect due to tidal
forces becomes increasingly important and the star is flattened in its orbital
plane to the shape of a curved pancake. The star rebounds, and as it leaves
the tidal radius, it starts to expand, becoming more cigar-shaped again. A little
further along its orbit the star eventually breaks up into fragments.
If the star chances to penetrate deeply (says with 10), its central temperature increases to a billion degrees in a tenth of a second. The thermonuclear
chain reactions are considerably enhanced. During this brief period of heating,
elements like helium, nitrogen and oxygen are instantaneously transformed into
heavier ones by rapid proton or alphacaptures. A thermonuclear explosion takes
place in the stellar pancake, resulting in a kind of accidental supernova. The
consequences of such an explosion are far reaching. About 50 per cent of the
stellar debris is blown away from the black hole at high velocity (propelled by
thermonuclear energy release), as a hot cloud able of carrying away any other
clouds it might collide with. The rest of the debris falls rapidly towards the hole,
producing a burst of radiation. Like supernovae, the stellar pancakes are also
crucibles in which heavy elements are produced and then scattered throughout
the galaxy. Thus, observation of highvelocity clouds and enrichment of the interstellar medium by specific isotopes in the vicinity of galactic nuclei would
29
break-up
cigar
pancake
i
Exit
tidal radius
g
Black
hole
f
e
Entry
a
f
e
cigar
tidal radius
a
sphere
Figure 21
Fig. 21 The disruption of a star by tidal forces of a black hole.
Imagine a black hole surrounded by a bright disc (Figure 22). The system is
observed from a great distance at an angle of 10 above the plane of the disc.
The light rays are received on a photographic plate (rather a bolometer in order to capture all wavelengths). Because of the curvature of spacetime in the
30
age
primary im
disc
upper side
of disc
black
hole
imag
secundary
photographic plate
System
lower side
of disc
neighborhood of the black hole, the image of the system is very different from
the ellipses which would be observed if an ordinary celestial body (like the planet
Saturn) replaced the black hole. The light emitted from the upper side of the
disc forms a direct image and is considerably distorted, so that it is completely
visible. There is no hidden part. The lower side of the disc is also visible as an
indirect image, caused by highly curved light rays.
The first computer images of the appearance of a black hole surrounded by
an accretion disc were obtained by myself (Luminet, 1978). More sophisticated
calculations were performed by Marck (1993) in Schwarzschild and Kerr space
times. A realistic image, e.g. taking account of the spacetime curvature, of the
blue- and redshift effects, of the physical properties of the disc and so on, can
be precisely calculated at any point of spacetime including inside the event
horizon. A movie showing the distortions observed along any timelike trajectory
around a black hole was produced (Delesalle, Lachi`eze-Rey and Luminet, 1993).
The figure 23 is a snapshot taken along a parabolic plunging trajectory. During
such a thought journey the vision of the third butterfly becomes accessible, all
external spectators can admire the fantastic landscape generated by the black
hole.
For a long time considered by astronomers as a mere theoretical speculation, black holes are now widely accepted as the basic explanation for Xray
massive binaries and galactic nuclei. Allowing for the elaboration of the most
likely models, black holes also respond to the principle of simplicity, according
to which among equally plausible models, the model involving the least number of hypotheses must be preferred. However, for such a wide acceptance to be
settled down, the basic picture of a black hole had to be drastically changed.
The conjunction of theoretical and observational investigations allowed for such
a metamorphosis of the black hole image, passing from the primeval image of a
naked black hole perfectly passive and invisible, to the more sophisticated image
of a thermodynamical engine well-feeded in gas and stars, which turns out to
31
be the key of the most luminous phenomena in the universe. Then the modern
astronomer won over to such a duality between light and darkness may adopt
the verse of the french poet Leon Dierx: Il est des gouffres noirs dont les bords
sont charmants.
References
[1] Begelman, M., Rees, M. (1996): Gravitys Fatal Attraction: Black Holes
in the Universe (New York: Scientific American Library)
[2] Chandrasekhar, S. (1992): The Mathematical Theory of Black Holes, (Oxford: Oxford University Press)
[3] Delesalle, L., Lachi`eze-Rey , M., Luminet J.-P. (1994): Infinitely Curved,
video VHS 52 mn, Arte/CNRS Audiovisuel
[4] DeWitt, C., DeWitt, B.S. (Eds.) (1973): Black Holes (Les Houches School,
Gordon and Breach: New York)
32
[5] Hawking, S.W., Israel, W. (Eds.) (1989): 300 Years of Gravitation (New
York, Cambridge: Cambridge University Press), pp. 199446
[6] Luminet, J.-P. (1979): Astron. Astrophys. 75, 228
[7] Luminet, J.-P., Carter, B. (1986): Astrophys. J. Suppl. 61, 219
[8] Luminet, J.-P. (1992): Black Holes (New York, Cambridge: Cambridge
University Press). German translation: Schwarze L
ocher (Vieweg, 1997)
[9] Marck, J.-A. (1993): Class. Quantum Grav. 13, 393
[10] Misner, C.W., Thorne,
Francisco: Freeman)
K.S., Wheeler,
[11] ONeill, B. (1995): The Geometry of Kerr Black Holes (Wellesley: Peters)
[12] Wheeler, J.A. (1990): A Journey into Gravity and Spacetime. (New York,
Cambridge: Cambridge University Press)
TROU NOIR
"NOTRE"
UNIVERS
Horizon
Gorge de
Schwarschild
"AUTRE"
UNIVERS
TROU BLANC