Blackhole thermodynamics

Jacob D. Bekenstein

Citation: Phys. Today 33(1), 24 (1980); doi: 10.1063/1.2913906

View online: http://dx.doi.org/10.1063/1.2913906
View Table of Contents: http://www.physicstoday.org/resource/1/PHTOAD/v33/i1
Published by the American Institute of Physics.

Additional resources for Physics Today

Homepage: http://www.physicstoday.org/
Information: http://www.physicstoday.org/about_us
Daily Edition: http://www.physicstoday.org/daily_edition

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
 Black-hole thermodynamics
Including black holes in the scheme of thermodynamics has disclosed
a deep-seated connection between gravitation, heat and the quantum that
may lead us to a synthesis of the corresponding branches of physics.

Jacob D. Bekenstein

To the physicist casually interested in spherically collapsing massive star (see field charge, quark color, and other den-
gravitation, a black hole is a passive object figure 2). This paper was the first to re- izens of the theorist's mind as parameters
that swallows anything near it and cannot gard the black hole as a central phenom- have been exhibited. But no extension of
be made to disgorge it; it absorbs but enon in astrophysics. But this is not the the charged Kerr solution has ever
cannot emit. At the close of the last aspect of interest here. Rather, we are emerged as a description of equilibrium
decade the experts shared this view. interested in the black hole as a state of black holes, and no "quantum number"
Recently, however, this simple picture has the gravitational field that resembles an other than M, Q, and L, has been found to
changed entirely. Perhaps no single de- ordinary object in many respects, espe- characterize the state of black holes. (For
velopment highlighted more the new cially in its interactions with the rest of example, its baryon and lepton numbers
views about black holes than the quantum the universe. would be unobservable outside the black
argument presented by Stephen Hawking By the 1960's the most general black hole, the magnetic pole strength is, as far
of Cambridge University in 1974 that a hole solution known was one describing a as we know, not a freely occurring quan-
black hole must radiate spontaneously rotating electrically charged hole in sta- tity and present theory regards all ob-
with a thermal spectrum. The impor- tionary state. It is parameterized by the servable hadrons—from which the black
tance of this phenomenon is not so much object's mass M, charge Q and angular hole would form—as "colorless.")
in possible practical applications, not even momentum L (see table 1), and is gener- This conclusion appears to be theory
in its astrophysical implications, but ally known as the charged Kerr black independent. Black-hole solutions of
rather in that it has confirmed earlier hole1 after Roy Kerr who discovered the several of general relativity's competitor
suspicions that gravitation, thermody- special case with Q = 0 in 1962. The gravitational theories (scalar-tensor
namics and the quantum world are deeply Schwarzschild hole has L = Q = 0. theories and supergravity) have been
interconnected. This connection, which found. They all belong to the charged
might be symbolized by the thermody- Black holes have no hair Kerr family. The principle "black holes
namic engine shown in figure 1, engenders In principle one would expect equilib- have no hair" evidently transcends the
hope that we may achieve a synthesis of rium black holes with more parameters bounds of general relativity, and may be
these three branches of physics in our describing shape and various other regarded as a general law of black-hole
time and bears witness to the profound properties. Yet in the late 1960's John A. physics.
unity of physics, a unity too often veiled Wheeler, then at Princeton University,
in an age of increasing specialization. suggested1 that in fact the charged Kerr Irreversibility of black holes
Black holes emerged as solutions of the black holes are the most general equilib- As a graduate student of Wheeler's at
gravitational field equations of Albert rium black holes states as far as exterior Princeton I found "black holes have no
Einstein's general relativity which de- properties are concerned. (See also the hair" distressing for a reason he brought
scribe regions of space-time invisible from article by Remo Ruffini and Wheeler, home to me in a 1971 conversation. The
their exterior. The first such known so- "Introducing the black hole," PHYSIC'S principle, he argued, allows a wicked
lution, found in 1916 by the noted physi- TODAY, January 1971, page 30.) He was creature—call it Wheeler's demon—to
cist and astronomer Karl Schwarzschild, led to this conjecture, which he whim- commit the perfect crime against the
represents the space-time geometry sically paraphrased as "black holes have second law of thermodynamics. It only
(gravitational field) of a spherical static- no hair," by uniqueness theorems of has to drop a package containing some
black hole; the only adjustable parameter Werner Israel and Brandon Carter re- entropy into a stationary black hole, thus
of the solution is the object's mass M. In garding the Schwarzschild and Kerr holes. decreasing the entropy in the part of the
a famous paper, published on the day Since then an impressive amount of evi- universe visible from the exterior. The
marking the outbreak of World War II, J. dence has piled up in favor of the conjec-
associated changes in M, Q, and L do not
Robert Oppenheimer and Harlan Snyder ture; in particular, it has become clear
uniquely reveal how much entropy is then
demonstrated that this Schwarzschild that there is no way to introduce quan-
tities like baryon and lepton numbers, inside the hole, so an exterior observer
solution describes the final state of a
strangeness, etc. as black-hole parame- with no inside information about the
Jacob D. Bekenstein is Professor of Physics at ters. True, over the years black-hole so- package can never be sure that the total
the Ben-Gurion University of the Negev, Beer- lutions having magnetic monopole, scalar entropy in the universe has not decreased.
Sheva, Israel. For him the second law is transcended—

24 PHYSICS TODAY / JANUARY 1980 0031-9228/80/010024-07/$00.50 © 1980 American Institute of Physics

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
made irrelevant. It loses its predictive
power, so that black holes seem to be
outside the province of thermodynamics.
This circumstance seemed disastrous, not
only because it would deprive us of the use
of model-free thermodynamic reasoning
in investigating the bizarre black holes,
but also because it could be seen as
throwing doubt on their very existence,
even in principle.
While mulling over this dilemma I was
struck by the possible relevance of a dis-
covery that Demetrios Christodoulou,
another of Wheeler's students, had made
not long before. While investigating the
efficiency of certain processes proposed
by Roger Penrose for extracting rotational
energy from a Kerr black hole and con-
verting it to mechanical energy of parti-
cles, Christodoulou noticed that the most
efficient processes are those associated
with reversible changes of the black hole.1
Less efficient processes are all connected
with the irreversible increase of a certain
"irreducible mass" M\r—the inextractable
part of the mass of the hole. Since in
thermodynamics reversible processes are
the most efficient ones for converting
energy from one form into another, there
was a clear thermodynamic ring to all this.
So black holes might be in harmony with
thermodynamics after all.
If there were any doubts as to the gen-
erality of the irreversibility, they were
dispelled by a theorem proved by Hawk-
ing: The surface area A of the boundary
or horizon of any black hole cannot de-
crease and will increase in a dynamical
process. For a Kerr hole A is propor-
tional to MiT2, so the area theorem implies
Christodoulou's result that all but very
idealized processes increase MiT. But the
theorem shows irreversibility to be a
property of all black hole processes, not
just those of near-equilibrium holes. And
it points to a formal analogy between
black hole area and entropy of a closed
system—both like to increase.

Black-hole entropy
It occurred to me that some monotonic
increasing function of A might play the
role of entropy of black holes—entropy on
the same footing as ordinary thermal en-
tropy. In my 1972 dissertation I dis-
cussed various questions connected with
such an identification, and showed how to
use it to defeat the schemes of Wheeler's
demon and make black hole physics con-
sistent with thermodynamics.
The first question was, what function
of A is to be identified with the entropy
S b h of the black hole. There are several
other conditions an entropy function
must meet besides the inability to de-
crease spontaneously: It must also be A gravitational-thermodynamic engine for converting heat to work, conceived by Robert Geroch,
additive for independent systems, for uses a black hole as heat sink. The box is filled with thermal radiation and lowered in the hole's
example. From the outset the most likely field to its horizon. The gravitational field does work on the box in this process. The radiation
prospects were proportionalities to A or is then allowed to escape into the hole, and the box is hauled back up at expense of less work than
to \/~A . The second appeared especially was obtained in the first stage. It can then be refilled with thermal radiation from the reservoir.
attractive, because it makes the entropy Because the box cannot be allowed to touch the horizon, the efficiency of conversion is less than
proportional to the mass for black holes 1 and not unity as originally thought. (Drawing by Louis Fulgoni) Figure 1

PHYSICS TODAY / JANUARY 1980 25

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
 connection becomes meaningless if one
goes to the classical limit h-*0. Thus it
"Time" Singularity's history reflects the deep connection of gravita-

I "Distance"
tion, thermodynamics and the quantum
world I alluded to.
Information and entropy
By now the reader of these lines must
be wondering about the physical meaning
of Sbh- It cannot be the thermal entropy
of the matter inside the hole. For con-
sider a sun-like star with mass 2 X 10:wg.
Its thermal entropy is 105H£ in order of
Horizon's history - magnitude. Were the star to collapse
completely, it would make a stationary
black hole with surface area of order 100
km'2 (see table 1). Our formula associates
with it an entropy of order \0'8k. Yet no
known dissipative process can generate
enough entropy during the collapse to
multiply the matter's entropy by a factor
102". Thus Sbh is something else than
thermal entropy. In statistical physics,
as well as other fields, entropy has come
to stand for missing information. With
that interpretation, the thermal entropy
of a cube of sugar simply measures our
ignorance as to the precise microscopic
state of the molecules in the cube, while
its macroscopic state is fully described by
chemical composition, temperature, vol-
ume and perhaps a few other variables.
Spherical collapse to a black hole. We show the spacetime histories of the collapsing body's The ignorance of the exterior observer
surface (illustrated as one-dimensional), of the resulting singularity and of the horizon expanding about the matter inside a black hole is
to engulf the body. The tilting of the light cone by gravitation explains why the region interior to deeper. "Black holes have no hair" allows
the horizon is invisible from outside. The spacelike hypersurfaces labeled 2 supplant, in relativity,
us only knowledge of M, Q and L for a
the concept of space at a given time. Figure 2
stationary hole; neither the microstate,
nor composition (baryons, photons. . .),
nor temperature, nor structure (shape,
with the same ratios of L/M2 and Q/M; quantum gravitation. He has argued that size. . .), nor anything else is—even in
for ordinary matter, also, the entropy is Lp\v must represent the smallest scale at principle—measurable by a distant ob-
proportional to mass, other things being which spacetime can be regarded as a server. The function S^h is the obvious
equal. But this choice leads to trouble. smooth manifold; at smaller scales it must candidate for quantifying this deeper ig-
When two black holes coalesce, Hawking's have a foam-like consistency; L\>\\ is often norance, so it is not surprising that it can
area theorem says the surface area must called the Planck-Wheeler length. be vastly larger than any reasonable es-
be at least as large as the sum of the areas Wheeler's suggestion led- to a formula timate of the thermal entropy.
of the two holes. However, the sum of the for black hole entropy independent of the
square roots of the original areas can ex- properties of matter The following reasoning shows this in-
ceed the square root of the final_area, so 2 terpretation of Sbh is not far fetched.
making Sbh proportional to V A would Sbh = ykA/LpW' When an object disappears down a black
endow it with unacceptable behavior. No Here ij is a pure number, presumably of hole, we lose information about its mi-
problem arises if Sbivis proportional to A. order unity, to be determined separately. crostate, composition, motion, and so on.
Thus I chose- Sbh to be proportional to This formula links a purely thermody- (The information about the motion is
the area of the horizon. namic quantity—entropy—to a purely clearly tied with properties of the hole in
Determination of the proportionality gravitational one—horizon area, and the whose field the object moves. But clearly
constant was more problematic. There
is an obvious constant with dimensions of
entropy—Boltzmann's constant k. What Table 1 The charged Kerr black hole
was missing was the scale of length whose
square should be used to reduce A to a Parameter mass M charge Q angular momentum 1.
dimensionless quantity. Wheeler sug-
gested the fundamental length Characteristic length m= GM/c2 2
q= (GQ /c )'' A 2
a = UMc
roughly size of hole
;l:l
Llnv = = 1.61 X 10- cm Quantity equivalent to 1 cm 1.35 X 102ag 3.49 X 1024esu 3 X 10 10 cm 2 /s
specific angular
which was introduced into physics by momentum
Max Planck after his discovery of the law Constraint m2 ^ q2 + a2
for the blackbody spectrum. Planck Formula for horizon area A = 47r|[(m+ ( m 2 - q2- a 2 ) 1 ' 2 ]'! + a2|
•
noted that this scale involves fundamen-
Potential surface tension 0 electric potential 4>
tal constants independent of particle angular velocity Q.
properties, in contrast to other lengths Definition (dMc2/<)A)LO (dMc2/c)Q)LA {8Mc2/dL)aA
such as the classical radius of the electron. Formula (m2 - q2 - a2)''Jc4/2<3A Q{4w/A - (i,walA)2y2 (UM)4ir/A
Later, Wheeler stressed the role Lpw C 4 /2G= 6.07 X 104e1 erg/cm
must play in the ultimate theory of

26 PHYSICS TODAY / JANUARY 1980

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
 the minimum of the total lost information that there were situations it could not
taken over all possible motions depends handle.
only on properties other than the motion
and should be independent of the hole's Black hole temperature
properties.) If the identification of Sbh Consider the expression for Sbh of a
as information hidden inside the black charged Kerr hole (see table 1); its total
hole is correct, the minimum possible differential with respect to M, Q and L
growth in S h h from injection of a given can be written as 2
• body into a hole in any orbit should be
independent of the hole's properties, the = d(Mc-) - - fidL
same for a big hole or a small one, a static where Tbh is defined in terms of the non-
or a rotating one. A priori, no mechanical negative quantity 8 of table 1 by Tbh =
reason is evident for why this should be so. L}>\\'26/r]k, and <J> and fi turn out to be the
' Yet detailed calculations'2^ showed that conventionally defined electric potential
' for injection into a stationary hole the and rotational angular frequency of the
minimum growth in S b h is 8-KTjkfj.cro/h hole's surface (see table 1), so that 3>dQ is
where n is the object's mass and r(> its ef- the work done on the hole by adding to it
fective radius. This result is independent charge dQ, and fidL is the work done by
of the hole's parameters, as required. addition of angular momentum dL. Of
Thus the interpretation is consistent. course, d(Mc2) is the corresponding
In statistical physics the entropy di- change in the hole's energy. Thus the
vided by k gives the logarithm of the expression above has exactly the form of
number of microstates compatible with Electrons and positrons
the usual expression for the first law of
Gravitational
the given macrostate. It is thus reason- and photons waves
thermodynamics (combined with the
t able to expect that exp(S|,h//j) gives the definition of entropy) for an ordinary
> number of possible interior configurations equilibrium system at temperature 7\,h.
of the black hole compatible with the The expression invites us to regard Sbh as
Interior configurations of a given black hole are
given exterior state whose black hole en- microstates of all possible structures and
a genuine entropy if we are willing to re-
tropy is Shh- Here "interior configura- compositions, a very few of which are shown. gard Tbh as a quantum temperature for an
tion" refers to each possible microstate of These "configurations" include states of dif- equilibrium black hole—quantum be-
composition and structure counted sep- ferent baryon and lepton numbers, so long as cause h appears in it. For the numerical
arately (see figure 3). This seemingly total mass, charge and angular momentum are value of 7\,h as function of black hole
unverifiable relation has, in fact, very preserved. They need not be just arrangements mass, see figure 5. Note the curious fact
palpable consequences to which I will of matter; they can also include gravitational that the smaller Me2, the higher Thh-
waves, for example, which are regarded as vac- Ordinarily the lower the energy content
come later. Just now I want to show how
uum systems in general relativity. Figure 3
it leads to an estimate of the constant rj. of a body, the colder it is, other things
Consider a hole with entropy Shh and, being equal.
consequently, exp(Sbh/^) possible interior infall of the package will be compensated, The physical meaning of T|,h proved
configurations. Inject into it the simplest or even overcompensated, by an increase elusive. One clearly cannot measure a
of systems, a generic elementary particle. in Shh- temperature by sticking a thermometer
This has several possible microstates In 1972 when I conjectured this gener- into a black hole; already when one brings
differing in spin, charge, and other alized second law, it was not clear that it it into the neighborhood of the hole, the
quantum numbers. So the number of would always work. In fact, nearly ev- instrument would be torn to pieces by
interior configurations after the injection erybody I discussed it with objected that tidal forces. My favorite argument for
will exceed 2 exp(SbhA) since each old it must often fail because the growth in bringing out the physics in 7\h revolved
configuration together with one micro- black-hole area is caused by the mechan- about a heat engine conceived by Robert
state makes a new configuration. Hence ical properties of the infalling body, which Geroch, illustrated in figure 1, that em-
the growth in black hole entropy exceeds are unrelated to its content. In my dis- ploys a black hole as a heat sink. A tem-
k In 2. But our previously mentioned sertation, and in later work when at the perature of order Tbh enters into the ex-
result assures us that the minimum University of Texas at Austin, I countered pression for the efficiency of this engine
growth in Sbh is Sin]k, because rQ is h/fic this objection by using the result about in the same way the heat sink's tempera-
for an elementary particle. Comparing the minimum increase in Sbh and by cal- ture enters the efficiency of a Carnot en-
the two values, we infer a lower bound (In culating the entropy content for various gine in a down-to-earth situation.3 But
2)/87r, or 0.028, for r/. This was the best simple systems. 2;l The calculations the argument could not nail down the
known estimate for r\ until Hawking de- suggested that there is a limit on the en- precise value of the hole's effective tem-
termined it to be V4, an order of magni- tropy that can be packed in a body of perature. Dennis Sciama of Oxford
tude larger. given mass and size. In every case University has pointed out that this de-
checked in which a small package of en- ficiency and some subtle conceptual
The generalized second law tropy falls into a stationary hole, the problems in the discussion were a pre-
Black-hole entropy is just what is generalized second law was found to work monition of the existence of Hawking's
needed to solve the paradox posed by (see figure 4). It could thus be taken as radiation. At that time, though, the
Wheeler's demon, for it makes possible a the second law of a black-hole thermo- argument was important in invalidating
generalization of the second law: "the dynamics. Geroch's claim that the process could
sum of the black-hole entropy and the Of course, a law is not proved by spe- violate the Kelvin statement of the second
ordinary thermal entropy outside black cific instances; it cannot be proved. law. It does not, and it also respects our
holes cannot decrease." By speaking only Rather it gains credibility with each ex- statement of the generalized second
about quantities determinable from out- ample in which it works. If, in addition, law.
side the holes, this law avoids the problem the proposed law passes the test posed by Regarding Thh as a black-hole tem-
the ordinary second law ran into. It is not a novel or unfamiliar situation, one as- perature created a severe problem for
transcended. It makes a statement sumes that it describes nature correctly. thermodynamics of black holes. Suppose
subject to verification: that the decrease We shall see that the generalized second a black hole is immersed in thermal ra-
in the outer world's entropy following the law passed just such a test after it seemed diation of temperature T, with T less than

PHYSICS TODAY / JANUARY 1980 27

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
 speculative. What induced me to stick
with it was the prospect of employing the
rich arsenal of thermodynamics and sta-
tistical physics to attack the issue of the
quantum aspects of black holes, a pros-
pect nonexistent in black-hole mechanics.
In those days in 1973 when I was often
told that I was headed the wrong way, I
drew some comfort from Wheeler's opin-
ion that "black-hole thermodynamics is
crazy, perhaps crazy enough to work."
Black-hole radiance
Early in 1974 Hawking announced his
discovery of the thermal radiance of black
holes.5 He had been studying the theory
of quantum phenomena in the neighbor-
hood of a microscopic black hole; years
earlier he had proposed that such objects
may have been formed in profusion in the
early cosmos. By applying the technique
of second quantization to a boson field
(such as the electromagnetic field)
evolving in the vicinity of a collapsing
spherical object, he found that the col-
lapse creates quanta of the field that es-
cape to infinity. As the collapse winds up
with the formation of a Schwarzschild
black hole, the emission, instead of dying
out as might be naively expected, attains
a form and rate independent of details of
the collapse. Its spectrum is thermal:
The generalized second law supplants the ordinary second law, which it transcends. The entropy the mean number of quanta emitted in
on the visible parts of the hypersurfaces decreases from 2 2 to 2 3 as the horizon engulfs the body one mode of frequency u> is given by
and from 2 4 to 2 5 as the entropy package thrown by Wheeler's demon falls in. However, the
black-hole entropy S bh increases during these intervals and the sum of the two entropies grows n = V[exp(h<jo/kT) - I]" 1
monotonically, as shown in the accompanying graph. Figure 4
where T is the hole's absorptivity (fraction
of incident waves absorbed) for classical
Tbh- A simple estimate shows the hole to nearly universal skepticism. Most people waves in the given mode, and T is a cer-
absorb more thermal entropy by simply inclined towards the viewpoint formu- tain characteristic temperature. (A
sucking in radiation than the growth in lated by James Bardeen, Carter and similar result holds for fermions.) The
black hole entropy can compensate for: Hawking at the Les Houches 1972 sum- emissivity of a black hole has the same
there is an apparent flagrant violation of mer school on black holes, and later Planck distribution required by Kirch-
the generalized second law. My attempts summarized in a lucid paper, "The Four hoff's laws for the emissivity of any hot
to clear up this paradox were forced and Laws of Black Hole Mechanics."4 They body of absorptivity T and temperature
inelegant.3 The same problem does not regarded the analogy between black holes T. Hawking thus found that the hole
arise for an ordinary hot body in a colder and thermodynamics as suggestive but radiates like a hot, non-black ("gray")
radiation bath because the body also ra- purely formal, reflecting no profound body. The luminosity is proportional to
diates and generates thermal entropy. kinship of the two subjects, and being M~2 and exceeds 1 watt for M < 6 X 1019g
Classically a black hole cannot radiate, so unconnected with the quantum. They (corresponding to a size less than 10"8
this observation would seem irrelevant. recognized 6 as the analog of temperature, centimeter).
However, the ubiquitous appearance of h but regarded the thermodynamic tem- Hawking also gave an argument show-
in 7\,h, Sbh and the generalized second law perature of a black hole as zero, because ing that charge and rotation do not alter
shows that one is here dealing with a apparently a black hole cannot be in the thermal nature of the emission."1
quantum issue. The possibility that a equilibrium with thermal radiation at fi- They only affect the form of the charac-
black hole could radiate by virtue of its nite temperature. They regarded the teristic temperature T. In every case it
quantum temperature was thus apparent horizon area A as an analog of entropy, is given by the same formula as T|,h with
very early. Yet the rudimentary under- but thought the thermodynamic entropy the choice TJ = V.».
standing of quantuhi processes in a black of a black hole to be infinite. The role of Hawking's discovery surprised every-
hole's vicinity available in 1973 gave no second law they gave to the area theorem, body; of all I was probably the most
evidence for or against such radiation. while rejecting the generalized second law pleased for it provided the missing pieces
Thus I chose to try to resolve the paradox as invalid, believing (erroneously, as is of black-hole thermodynamics. The re-
by more conventional means' and in so clear today) that it is possible to add en- sult verified the contention that a nonzero
doing lost the chance to make a striking tropy to a black hole without incre- quantum temperature is associated with
prediction with black-hole thermody- menting its area. Table 2 compares the a stationary black hole; it left no doubt
namics. Be that as it may, the paradox tenets of this "black-hole mechanics" with that the Hawking temperat ure is identical
was soon to find its resolution. those of black-hole thermodynamics. to Tbhi it fixed the value of the elusive r\,
There can be no question that black- and it revealed the physical meaning of
Black hole mechanics hole mechanics was, at the time, the Tbi,: it is the temperature of the quan-
This was the state of the art in mid- "common-sense" approach, and seemed tum radiation from the black hole.
1973. Black-hole thermodynamics, to stand on solid ground. By contrast, Hawking realized that his discovery
summarized in table 1, was received with black-hole thermodynamics was frankly solved the paradox posed by the black

28 PHYSICS TODAY / JANUARY 1980

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
hole in a colder radiation bath. He gave - 1060 physical field is characterized by a mul-
a simple proof6 that after the radiation 10'° T titude of virtual particle-antiparticle
emitted by the hole is assimilated into the
ambient bath, the sum of S h h and exterior v
*\ /
pairs continually appearing and annihi-
lating each other. The pairs cannot en-
radiation entropy is larger than before.
Thus the generalized second law does in
fact hold because the hole generates suf-
10s
\ /
A
dure; the creation of one out of nothing
entails a violation of the conservation of
energy, so it must disappear before the
ficient thermal entropy. (The moment- - 10" time allotted to it by the uncertainty
by-moment applicability of the law has 10°
s. principle runs out. But if a pair is created
been established by an explicit statistical near a black-hole horizon, the story can be
calculation of the radiation entropy.7) \ . different. Hawking points out that one
Having started as a vocal critic of the 10-1 of the members of the pair may have time
generalized second law, Hawking became
/ , 1
10!0 10" 10'° to tunnel through the horizon into one of
the person who made it fully consistent MASS (gram) the negative-energy particle orbits or
with the gedanken experiments. states that exist inside the hole due to the
Actually, an isolated radiating hole il- Black-hole temperature (color) and black-hole colossal binding effects of gravitation. Its
entropy (black) as functions of black hole mass companion is left with positive energy and
lustrates the power of the generalized (i) = V4). The solid lines are appropriate to a
second law even more strikingly. The may be able to escape from the black
Schwarzschild hole, that is, for q2 + a2 = 0;
radiation causes the hole's surface area to broken lines apply to a rotating or charged Kerr
hole's pull as a free, real particle. In the
decrease steadily—a flagrant (quantum) hole with q2 + a2 = 0.99 m 2 (this is nearly the final analysis, its energy comes from the
violation of the second law of black hole largest possible charge or spin a black hole can hole, so the process is effectively one of
mechanics (a classical theorem). How- have—see t a b l e ! ) . Figure 5 radiation by the hole. No overall viola-
ever, a detailed statistical calculation" tion of the conservation of energy occurs;
shows that the increase in exterior (ra- the ingoing particle carries negative en-
diation) entropy exceeds the decrease in ley, just before Hawking's discovery. ergy into the hole, whose mass is thereby
Sbh, so the generalized second law in its Unruh studied a second-quantized scalar decreased. This flux of negative energy
original formulation is obeyed (see figure field (a meson field) evolving in the through the horizon is what causes the
6). The Hawking radiation process thus background gravitational field of an breakdown of the area theorem.
provided a novel and unexpected test, eternal Kerr black hole. He found9 an The pair created near the horizon is in
which the (classical) area theorem failed, outflux of energy in the superradiant what one calls a pure quantum state (de-
but which the (quantum) second law modes for a quantum state containing no scribed by a wave function), the farthest
passed successfully. Ironically the area quanta incident on the hole. thing in the world from a thermal state,
theorem, one of the motives for the in- In the Zel'dovich-Unruh radiance the which is a mixed state (described by a
troduction of black-hole entropy, has energy comes from the hole's rotation; density matrix) of maximal entropy. But
fallen victim of the revolution in our un- there is no radiance for a Schwarzschild as Sciama has observed in another con-
derstanding of black holes. Its replace- hole. The area theorem is respected. By text, the disappearance of the particle (or
ment is a new law of apparently wide ap- contrast Hawking's radiance occurs for antiparticle) behind the horizon breaks
plicability, one joining together hitherto Schwarzschild as well as Kerr holes. The the correlations inherent in the pure state,
distinct aspects of nature—gravitation energy must evidently be drawn from the leaving a mixed state. The loss of infor-
and heat, gravitation and the quantum. "irreducible" mass (equal to the mass for mation into the hole is complete, so the
Schwarzschild holes) by virtue of the entropy of the new state is maximal.
Radiance and superradiance quantum violation of the area theorem Thus one understands the thermal nature
The theoretical necessity for Hawking's mentioned earlier. The Hawking radia- of the radiation. By the same token it is
emission has been verified by many tion emerges in every mode, and it is dis- clear why the Zel'dovich-Unruh radiation
workers and by varied approaches. All tinguished from the first type by its is not thermal. Its origin parallels that of
the explanations of the phenomenon are thermal nature. And it occurs only if the the Hawking radiance. But the relevant
technically complicated. In seeking a hole was formed by collapse while the negative energy states are now those in
physical intuition of the process it pays to Zel'dovich-Unruh emission comes only the ergosphere of a Kerr hole—the region
first inquire how it differs from the other from a hole that has existed since the in- girding the horizon. (These states are the
known type of black-hole radiance. In finite past. ones that make the Penrose energy-ex-
1971 the versatile physicist and astro- traction processes possible.) Because the
physicist Yakov B. Zel'dovich of the So- The role of the horizon trapped member of the pair does nothave
viet Academy of Sciences conjectured* Hawking's heuristic picture 1 " 1 of the to cross the horizon, no great loss of in-
that a Kerr hole should emit bosons process he deduced by a complicated formation ensues, and we don't get a
spontaneously in those modes whose an- mathematical argument is disarmingly thermal state. For this reason also there
gular frequency u> and "magnetic" quan- simple. We know that in the absence of is no violation of the area theorem.
tum number m satisfy the condition exterior fields the vacuum state of any The alert reader will have noticed the

(x> < mil

Zel'dovich, and independently Charles
Misner, had pointed out that classically Table 2 Black-hole mechanics versus black-hole thermodynamics
the hole will amplify radiation scattered
off it in these modes, which ever since Concept Black-hole mechanics Black-hole thermodynamics
then have been known as superrradiant entropy A is like entropy; the physical entropy is The black-hole entropy Sbh is ijc3kA/Gh
modes. It was natural to regard this infinite
classical "superradiance" as a manifes- temperature H is like temperature; the physical The black-hole temperature 7"bh is Ghh)c3k
tation of quantum stimulated emission, temperature is zero
and Zel'dovich inferred that it must be = dMc2- Tbh6Sbh = d/Mc2 - 4>dO - SidL
first law HdA 4>dO- ildL
accompanied by spontaneous emission as The sum of Sbt) and the entropy exterior to
second law For one black hole A cannot decrease;
in other contexts in physics. His conjec- when black holes coalesce the total black holes cannot decrease (generalized
ture was verified by William Unruh, then horizon area increases (area theorem) second law)
at the University of California at Berke-

PHYSICS TODAY / JANUARY 1980 29

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
 the generalized second law is respected in
the face of Hawking's radiance, something
I referred to earlier.
Statistical black hole physics has also
resolved an intriguing paradox connected
with black hole radiance. I mentioned
how Zel'dovich predicted the existence of
spontaneous emission by a Kerr hole in
superradiant modes by interpreting su-
perrradiance as stimulated emission, and
insisting that spontaneous and stimulated
emission go together. Now, according to
Hawking there is also spontaneous emis-
sion in non-superradiant modes, such as,
for example, all modes associated with the
Schwarzschild hole. If stimulated emis-
sion is associated with it, why is there no
superradiance in this case as in the pre-
vious one? Field theory did not clarify
the situation because it did not seem to
require the universal appearance of
stimulated emission in every non-super-
radiant mode.14
A palatable resolution of the paradox
emerged from a detailed statistical in-
vestigation of stimulated emission by
black holes carried out by Amnon Meisels
and myself at the Ben-Gurion Universi-
ty. 1S The central quantity in the analysis
is p{m |n), the probability that if n quanta
in a given mode are incident on a black
hole, m are returned outwards in that
The Hawking thermal radiation begins emerging when most of the collapsing body is inside the
horizon. The associated flow of negative energy through the horizon causes a steady decrease
mode by all possible processes. We cal-
of S bh from 2 3 on. However, the growing entropy of the radiation is sufficient to make the sum culated p(m\n) with the following as-
of the entropies inside and outside the hole increase, as shown in the graph. Figure 6 sumptions:
• The black hole is in a thermal radiation
bath
crucial role of the horizon. Whichever de Sitter universe, a cosmological model • The mean number of quanta the black
way we look at it, by black-hole thermo- describing a classically empty universe hole returns under the influence of the
dynamics or from the field-particle point with a high degree of symmetry, and thermal radiation incident on it is the sum
of view, it is the horizon that, by its in- which is separated into two regions by a of the spontaneous emission and a frac-
formation-hiding properties, places a horizon. Hawking and Gary Gibbons, tion 1 — F "reflected" from the incident
thermal stamp on the process. We see who discovered this effect,13 note that the mean number
this also in situations far from the realm area of that horizon has an entropy in- • The probability distribution has a
of black holes. For example, a detector terpretation identical to that for a black maximal entropy, for the same reasons I
tuned to the electromagnetic field and hole. These examples show that the mentioned in connection with the heu-
being accelerated uniformly in empty connection between gravitation and ristic explanation of Hawking's process.
spacetime (described by the Minkowski thermodynamics is likely to be a general Prakash Panangaden and Robert Wald
metric) will measure a quantum noise feature of nature, relevant beyond the have shown11' that this same probability
characteristic of thermal electromagnetic realm of black holes. can be derived from the field theory, al-
radiation at a temperature proportional beit more laboriously.
to its acceleration! The existence of this Stimulated emission Curiously, the probability p(m\n) is
intriguing phenomenon was foreshad- Hawking's analysis of black-hole radi- not just the composite of the probability
owed in work done1' in 1975 by Paul C. W. ance was an application of quantum-field distributions for spontaneous emission,
Davies of King's College, London, and theory in curved spacetime, a subject in p{m 10), and that for ordinary scattering
made theoretically compelling by Unruh1'- which research has boomed since his pi- of indistinguishable bosons. A third
a year later. As Sciama has pointed out, oneering work. These investigations have component distribution is needed to re-
one can again understand what is hap- led to increasing understanding of the produce it. This was only isolated re-
pening in terms of Hawking's pairs. influence of gravitation on quantum cently, and its form leaves no. doubt that
Every physics student knows that to an processes, and of the quantum nature of it describes stimulated emission in every
accelerated observer part of the spacetime gravitation. Yet the methods of this ap- mode: the mean number of quanta sent
is invisible because he outruns all signals proach are complex, and the results do not out by the corresponding process is pro-
coming from those parts. For such an lend themselves to description in simple portional to n, as one would expect for
observer there is a horizon. Thus the el- terms. However, if one is only interested stimulated emission. The proportional-
ements are present for producing thermal in understanding further the quantum- ity coefficient is none other than Ein-
radiation by a process analogous to that thermal properties of black holes, there is stein's coefficient of stimulated emission,
for a black hole. But here the energy is an alternative simpler approach, a direct fli.
ultimately drawn from the agent which outgrowth of black-hole thermodynamics, As in the case of atomic transitions, B \
accelerates the detector. which makes use of general statistical turns out to equal the mean number of
Another example of thermal radiation arguments, and which may thus be called spontaneously emitted bosons, which is
ultimately associated with a horizon is statistical black-hole physics. Its first just the Einstein coefficient of spontane-
that which all inertial observers see in the accomplishment was the verification that ous emission A. One can likewise calcu-

30 PHYSICS TODAY / JANUARY 1980

Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms
late the coefficient of absorption, Bj; it amplitude for evolution of a geometry 5. S. W. Hawking, Nature 248, 30 (1974);
obeys the relation (gravitational field) that contains a black Commun. Math. Phys. 43, 199 (1975).
hole region, for complex values of the 6. S. W. Hawking, Phys. Rev D 13, 191
coordinates. Curiously, this geometry is (1976).
We can now see that the classical ab- periodic in imaginary time, it, with a pe- 7. J. I). Bekenstein, Phys. Rev. D 12, 3077
sorption coefficient T is not just the riod T that is determined by M, Q, and L. (1975).
quantum absorption coefficient, but is Gibbons and Hawking were then able to H. Ya. B. Zel'dovich, Zh. Eksp. Teor. Fiz.
less than it by the stimulated emission exploit a trick well known to workers in Pis'ma 14, 270 (1971) (JETP Letters 14,
coefficient (after all, stimulated emission statistical physics: the formal amplitude 180(1971).
suppresses scattering). The explicit for a system to evolve back to its initial 9. W. Unruh, Phys. Rev. D 10, 3194 (1974).
forms of Bj and B- make it clear that for state in an imaginary time, it, equals its 10. S. W. Hawking, "The Quantum Mechanics
a mode satisfying the superradiance partition function (or statistical sum over of Black Holes," Scientific American 236,
condition, stimulated emission wins {Bt states) for the (real) temperature h/kt. January 1977, page 34.
> B\), so that the classical reflection Because of its periodicity the hole's ge- 11. P. C. W. Davies, J. Phys. A 8, 609(1975).
coefficient, 1 — T, is larger than unity. ometry evolves back to the initial state 12. W. Unruh, Phys. Rev. D 14, 870 (1976).
This is precisely what superradiance after "time" r, and thus a thermal parti- 13. ('.. Gibbons, S. W. Hawking, Phys. Rev. D
means! By contrast, if the mode does not tion function is defined for it. The tem- 15,2738(1977).
satisfy the superradiance condition, ab- perature turns out to be none other than 14. R. M. Wald, Phys. Rev. D 13, 3176
sorption dominates stimulated emission T|,h (with i) = '/)), and the entropy derived (1976).
(B| < Bf), 1 - 7 is less than unity, so that from the partition function is none other 15. J. D. Bekenstein, A. Meisels, Phys. Rev. D
classically the hole absorbs, and there is than S\,h. The agreement with the values 15, 2775 (1977); J. D. Bekenstein in The
implied by Hawking's radiance is proof Unification Problem {Proceedings of the
no superradiance. Thus a Schwarzschild
that the thermal properties of the gravi- Einstein Centennial Symposium, Jeru-
hole is capable of stimulated emission, but
tational field are very real and manifest salem), Y. Ne'eman, ed., Addison-Wesley,
not of superradiance. themselves in varied contexts. Cambridge (1980).
Counting interior configurations 16. R. M. Wald, P. Panangaden, Phys. Rev. D
Space limitations forbid a detailed Hi. 929(1977).
In quantum physics the ratio of B- survey of the applications and ramifica-
coefficients, B\IBt, equals the ratio of 17. U. H. Gerlach, Phys. Rev. D 14, 1479
tions of the thermodynamics of black
(1976).
degeneracy factors of the upper to the holes. Yet no description of the subject
lower energy level of the transition in could be representative without men- 18. G. Gibbons, S. W. Hawking, Phys. Rev. D
15,2752(1977).
question. For an equilibrium black hole, tioning some highlights. Hawking'5 has
the ratio can easily be shown"' to equal employed standard thermodynamic rea- 19. P. C. W. Davies, Proc. Roy. Soc. (London)
A 353,499 (1977).
the ratio of the quantity we have called soning to deduce the criterion for a black
the number of possible interior configu- hole to condense stably out of thermal 20. P. Candelas, D. Sciama, Phys. Rev. Lett.
rations, exp(Sbh/^), evaluated for the 38,1372(1977). D
radiation (when can droplets condense
hole's parameters before emission of a out of vapor?). Likewise, DaviesIU has
quantum in the mode in question, to the shown that the charged Kerr holes are
same number after emission. Thus if we
regard the hole as a conventional quan-
separated by the values of their parame-
ters into two distinct classes connected by
COLD
tum system, each of its states acts as if it
were degenerate, with a multiplicity
a second order phase transition like the A
point of He4. The role of irreversibility
REMEDIES r
exp(S'bh/fe). A more dramatic illustration in black-hole thermodynamics has been Cryogenic
of the interpretation of Si,h in terms of put in a new light by Philip Candelas and Temperature
interior configurations would be hard to Sciama-" who have demonstrated the Sensors
come by. The possible interior configu- close connection between quantum fluc- Cryo Resistor, Germanium
rations act as if they were all present si- tuations of the gravitational field at the Resistance Thermometer, an
multaneously in determining how the hole accepted standard for 14 years
horizon and the dissipation reflected in
- 1 . 5 - 100K calibration
radiates. Alternatively, if one has confi- the irreversibility, a connection analogous - 3 m K accuracy to NBS
dence in the interpretation of S(,h, the to that between noise and resistance in an —0.5mK reproducibility
result underlines the basic similarity of electrical resistor. — uniform sensitivity
black holes to conventional quantum ra- These and other examples underline Ruggedized Germanium
diating systems. the basic similarity of the bizarre black Resistance Thermometer
Convincing as it is, the interpretation holes to everyday objects in the quantum — same temperature characteristics
domain, a similarity which surfaces as the Cryo Resistor
of Shh has not yet been supplemented by — withstands shock and vibration
a general method for explicitly counting clearly in the thermodynamic viewpoint. — industrial and aerospace
interior configurations and demonstrating Because of this the dream of under- applications
the connection with Sbh. Ulrich H. Ger- standing the quantum aspects of the Carbon Composition
lach has analyzed17 a special model in gravitational field is closer to fulfillment Resistance Thermometer
which counting can be done schematically today than a decade ago. —available with matched
and has obtained rough agreement with characteristics
Sbh- A different approach, which, though References — inexpensive calibrations 1.5 - 300K
not formally a counting method, has great 1. C. W. Misner, K. S. Thome, J. A. Wheeler, —stable, accurate, dependable
potential, has been advocated by Gibbons Gravitation, Freeman, San Francisco We're cold but not impersonal.
and Hawking18 in the framework of a (1973). Contact us for a remedy to your
program for quantizing gravitation using 2. J. D. Bekenstein, Nuovo Cimento Lett.4, particular needs. Write or call collect.
Richard Feynman's "sum over histories" 737 (1972); Phys. Rev. D 7, 2333 (1973).
approach. Here the quantum amplitude :l. J. D. Bekenstein, Phys. Rev. I) <l, 3292
for a system to evolve from an initial to a (1974).
final state in time t is given by a certain 4. J. Bardeen, B. Carter. S. VV. Hawking,
phase factor summed over all conceivable Commun. Math. Phys. 31, llil (1973); see 2457 University Ave.
evolutions between the two states. also Black Holes, B. S. DeWitt, C. M. St. Paul, WIN 55114 USA
DeWitt eds. Gordon and Breach, New (612) 645-0072
Gibbons and Hawking computed the
York (1973). Circle No. 17 on Reader Service Card
PHYSICS TODAY / JANUARY 1980 31
Downloaded 16 Jul 2012 to 141.225.218.75. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms