Gravitation, Thermodynamics, and Quantum Theory 9 9 9 1 Robert M. Wald n Enrico Fermi Institute and Department of Physics a J University of Chicago 2 5640 S. Ellis Avenue 1 Chicago, Illinois 60637-1433 1 v December 10, 2001 3 3 0 1 0 Abstract 9 9 Duringthepast30years,researchingeneralrelativityhasbrought c/ to light strong hints of a very deep and fundamental relationship be- q tween gravitation, thermodynamics, and quantum theory. The most - r striking indication of such a relationship comes from black hole ther- g modynamics,where itappears thatcertainlawsofblackhole mechan- : v icsare,infact,simplytheordinarylawsofthermodynamicsappliedto i X a system containing a black hole. This article will review the present r status of black hole thermodynamics and will discuss some of the re- a lated unresolved issues concerning gravitation, thermodynamics, and quantum theory. 1 1 Introduction Atthe end of a century|particularlyone marking the end of a millennium| it is natural to attempt to take stock of the status of a (cid:12)eld of endeavor from as broad a perspective as possible. In the (cid:12)eld of physics, two theories| general relativity and quantum mechanics|were developed during the (cid:12)rst quarter of the present century. These theories revolutionized the way we think about the physical world. Despite enormous progress during the re- mainderofthiscenturyinexploringtheirconsequencesandintheapplication of these theories to construct successful \standard models" of cosmology and particle physics, at the end of this century we are still struggling to reconcile general relativity and quantum theory at a fundamental level. The revolutions in physics introduced by general relativity and quantum theory were accompanied by major changes in the language and concepts used to describethephysicalworld. In generalrelativity,itis recognizedthat space and time meld into a single entity|spacetime|andthat the structure of spacetime is described by a Lorentz metric, which has a dynamical char- acter. Consequently, simple Newtonian statements such as \two particles are a distance d apart at time t" become essentially meaningless in general relativity until one states exactly how t and d are de(cid:12)ned for the particular class of spacetime metrics under consideration. Furthermore, concepts such as the \local energy density of the gravitational (cid:12)eld" are absent in general relativity. The situation is considerably more radical in quantum theory, where the existence of \objective reality" itself is denied, i.e., observables cannot, in general, consistently be assigned de(cid:12)nite values. I believe that the proper description of quantum phenomena in strong gravitational (cid:12)elds will necessitate revolutionary conceptual changes in our view of the physical world at least comparable to those that occurred in the separate developments of general relativity and quantum theory. At present, the greatest insights into the physical nature of quantum phenom- ena in strong gravitational (cid:12)elds comes from the analysis of thermodynamic properties associated with black holes. This analysis also provides strong hints that statistical physics may be deeply involved in any fundamental conceptual changes that accompany a proper description of quantum gravi- tational phenomena. At the present time, string theory is the most promising candidate for a theory of quantum gravity. One of the greatest successes of string theory to 2 date has been the calculation of the entropy of certain classes of black holes. However, the formulation of string theory is geared much more towards the calculation of scattering matrices in asymptotically (cid:13)at spacetimes rather than towards providing a local description of physical phenomena in strong gravitational (cid:12)elds. Within the framework of string theory, it is di(cid:14)cult to imagine even how to pose (no less, how to calculate the answer to) lo- cal physical questions like, \What would an observer who falls into a black hole experience as he approaches what corresponds classically to the space- time singularity within the black hole?" Thus, string theory has not yet provided us with new conceptual insights into the physical nature of phe- nomena occurring in strong gravitational (cid:12)elds that are commensurate with some of its mathematical successes. It may well be that|even assuming it is a correct theory of nature|string theory will bear a relationship to the \ultimate theory of everything" that is similar to the relationship between \old quantum theory" and quantum theory. Therefore, I feel that it is very encouraging that, at the present time, intensive e(cid:11)orts are underway toward providing reformulations of string theory. However, to date, the these e(cid:11)orts have mainly been concerned with obtaining a formulation that would unify the (di(cid:11)erent looking) versions of string theory, rather than achieving new conceptual foundations for describing local quantum phenomena occurring in strong gravitational (cid:12)elds. Thus, at present, most of the physical insights into quantum phenomena occurring in strong gravitational (cid:12)elds arise from classical and semiclassical analyses of black holes in general relativity. In this article, I will review classical and quantum black hole thermodynamics and then discuss some unresolved issues and puzzles, which may provide some hints as to the new conceptualfeatures that maybepresentin thequantumdescriptionof strong gravitational(cid:12)elds. Inthediscussion,Iwillnotattempttoprovideabalanced account of researchinthis area, but rather willmerelypresentmyown views. 2 Classical black hole mechanics Undoubtedly, one of the most remarkable and unexpected developments in theoretical physics to have occurred during the latter portion of this century was the discovery of a close relationship between certain laws of black hole physics and the ordinary laws of thermodynamics. It was (cid:12)rst pointed out 3 by Bekenstein [1] that the area nondecrease theorem of classical general rel- ativity [2] is analogous to the ordinary second law of thermodynamics, and he proposed that the area, A, of a black hole (times a constant of order unity in Planck units) should be interpreted as its physical entropy. A short time later, Bardeen, Carter, and Hawking [3] extended the analogy between black holes and thermodynamics considerably further by proving black hole analogs of the zeroth and (cid:12)rst laws of thermodynamics. In this section, I will give a brief review of the laws of classical black hole mechanics. First, we review the de(cid:12)nition of a black hole and some properties of sta- tionary black holes. In physical terms, a black hole is a region where gravity is so strong that nothing can escape. In order to make this notion precise, one must have in mind a region of spacetime to which one can contemplate escaping. For an asymptotically (cid:13)at spacetime (M;g ) (representing an iso- ab lated system), the asymptotic portion of the spacetime \near in(cid:12)nity" is such a region. The black hole region, , of an asymptotically (cid:13)at spacetime, B (M;g ), is de(cid:12)ned as ab M I ( +); (1) (cid:0) B (cid:17) (cid:0) I where + denotes future null in(cid:12)nity and I denotes the chronological past. (cid:0) I The event horizon, , of a black hole is de(cid:12)ned to be the boundary of . In H B particular, is a null hypersurface. Note that the entire future history of H the spacetime must be known before the location of can be determined, H i.e., possesses no distinguished local signi(cid:12)cance. H If an asymptotically(cid:13)at spacetime(M;g ) contains a blackhole, , then ab B is said to be stationary if there exists a one-parameter group of isometries B on (M;g ) generated by a Killing (cid:12)eld ta which is unit timelike at in(cid:12)nity. ab The black hole is said to be static if it is stationary and if, in addition, ta is hypersurface orthogonal. The black hole is said to be axisymmetric if there exists a one parameter group of isometries which correspond to rotations at in(cid:12)nity. A stationary, axisymmetric black hole is said to possess the \t (cid:30) (cid:0) orthogonality property" if the 2-planes spanned by ta and the rotational Killing (cid:12)eld (cid:30)a are orthogonal to a family of 2-dimensional surfaces. A null surface, , whose null generators coincide with the orbits of a K one-parameter group of isometries (so that there is a Killing (cid:12)eld (cid:24)a normal to ) is called a Killing horizon. There are two independent results (usually K referred to as \rigidity theorems") that show that in wide variety of cases of interest, the event horizon, , of a stationary black hole must be a Killing H 4 horizon. The (cid:12)rst, due to Carter [4], states that for a static black hole, the staticKilling(cid:12)eldta mustbenormal to thehorizon, whereasfor a stationary- axisymmetric black hole with the t (cid:30) orthogonality property there exists a (cid:0) Killing (cid:12)eld (cid:24)a of the form (cid:24)a = ta+(cid:10)(cid:30)a (2) which is normal to the event horizon. The constant (cid:10) de(cid:12)ned by eq.(2) is called the angular velocity of the horizon. Carter’s result does not rely on any (cid:12)eld equations, but leaves open the possibility that there could exist stationary black holes without the above symmetries whose event horizons are not Killing horizons. The second result, due to Hawking [5] (see also [6]), directlyproves that in vacuum or electrovac general relativity,the event horizonof anystationary blackholemustbea Killinghorizon. Consequently, if ta fails to be normal to the horizon, then there must exist an additional Killing (cid:12)eld, (cid:24)a, which is normal to the horizon, i.e., a stationary black hole must be nonrotating (from which staticity follows [7], [8]) or axisymmetric (though not necessarily with the t (cid:30) orthogonality property). Note that (cid:0) Hawking’stheoremmakesnoassumptions ofsymmetriesbeyondstationarity, but it does rely on the properties of the (cid:12)eld equations of general relativity. Now,let beanyKillinghorizon(notnecessarilyrequiredtobetheevent K horizon, , of a black hole), with normal Killing (cid:12)eld (cid:24)a. Since a((cid:24)b(cid:24) ) also b H r is normal to , these vectors must be proportional at every point on . K K Hence, there exists a function, (cid:20), on , known as the surface gravity of , K K which is de(cid:12)ned by the equation a((cid:24)b(cid:24) ) = 2(cid:20)(cid:24)a (3) b r (cid:0) It follows immediately that (cid:20) must be constant along each null geodesic generator of , but, in general, (cid:20) can vary from generator to generator. It K is not di(cid:14)cult to show (see, e.g., [9]) that (cid:20) = lim(Va) (4) where a is the magnitude of the acceleration of the orbits of (cid:24)a in the region o(cid:11) of where they are timelike, V ( (cid:24)a(cid:24) )1=2 is the \redshift factor" of a K (cid:17) (cid:0) (cid:24)a, and the limit as one approaches is taken. Equation (4) motivates the K terminology \surface gravity". Note that the surface gravity of a black hole is de(cid:12)ned only when it is \in equilibrium", i.e., stationary, so that its event horizon is a Killing horizon. 5 Inparallelwiththetwoindependent\rigiditytheorems"mentionedabove, there are two independent versions of the zeroth law of blackhole mechanics. The (cid:12)rst, dueto Carter [4](seealso [10]), states that forany blackhole which is static or is stationary-axisymmetric with the t (cid:30) orthogonality property, (cid:0) the surface gravity (cid:20), must be constant over its event horizon . This re- H sult is purely geometrical, i.e., it involves no use of any (cid:12)eld equations. The second, due to Bardeen, Carter, and Hawking [3] states that if Einstein’s equation holds with the matter stress-energy tensor satisfying the dominant energy condition, then (cid:20) must be constant on any Killing horizon. Thus, in the second version of the zeroth law, the hypothesis that the t (cid:30) orthogo- (cid:0) nality property holds is eliminated, but use is made of the (cid:12)eld equations of general relativity. A bifurcate Killing horizon is a pair of null surfaces, and , which A B K K intersect on a spacelike 2-surface, (called the \bifurcation surface"), such C that and are each Killing horizons with respect to the same Killing A B K K (cid:12)eld (cid:24)a. It follows that (cid:24)a must vanish on ; conversely, if a Killing (cid:12)eld, C (cid:24)a, vanishes on a two-dimensional spacelike surface, , then will be the C C bifurcation surface of a bifurcate Killing horizon associated with (cid:24)a (see [11] for further discussion). An important consequence of the zeroth law is that if (cid:20) = 0, thenin the\maximallyextended"spacetimerepresentinga stationary 6 black hole, the event horizon, , comprises a branch of a bifurcate Killing H horizon [10]. This result is purely geometrical|involving no use of any (cid:12)eld equations. Asaconsequence,thestudyofstationaryblackholeswhichsatisfy the zeroth law divides into two cases: \degenerate" black holes (for which, by de(cid:12)nition, (cid:20) = 0), and black holes with bifurcate horizons. The (cid:12)rst law of black hole mechanics is simply an identity relating the changes in mass, M, angular momentum, J, and horizon area, A, of a sta- tionary blackhole when itis perturbed. To (cid:12)rst order, the variations of these quantities in the vacuum case always satisfy 1 (cid:14)M = (cid:20)(cid:14)A+(cid:10)(cid:14)J: (5) 8(cid:25) In the original derivation of this law [3], it was required that the pertur- bation be stationary. Furthermore, the original derivation made use of the detailed form of Einstein’s equation. Subsequently, the derivation has been generalized to hold for non-stationary perturbations [7], [12], provided that the change in area is evaluated at the bifurcation surface, , of the unper- C 6 turbed black hole. More signi(cid:12)cantly, ithas been shown [12] that thevalidity of this law depends only on very general properties of the (cid:12)eld equations. Speci(cid:12)cally, a version of this law holds for any (cid:12)eld equations derived from a di(cid:11)eomorphism covariant Lagrangian, L. Such a Lagrangian can always be written in the form L = L(g ;R ; R ;:::; ; ;:::) (6) ab abcd a bcde a r r where denotes the derivative operator associated with g , R denotes a ab abcd r the Riemann curvature tensor of g , and denotes the collection of all ab matter (cid:12)elds of the theory (with indices suppressed). An arbitrary (but (cid:12)nite) number of derivatives of R and are permitted to appear in L. In abcd this more general context, the (cid:12)rst law of black hole mechanics is seen to be a direct consequence of an identity holding for the variation of the Noether current. The general form of the (cid:12)rst law takes the form (cid:20) (cid:14)M = (cid:14)S +(cid:10)(cid:14)J +:::; (7) bh 2(cid:25) where the \..." denote possible additional contributions from long range matter (cid:12)elds, and where (cid:14)L S 2(cid:25) n n : (8) bh ab cd (cid:17) (cid:0) (cid:14)R ZC abcd Here n is the binormal to the bifurcation surface (normalized so that ab C n nab = 2), and the functional derivative is taken by formally viewing the ab (cid:0) Riemann tensor as a (cid:12)eld which is independent of the metric in eq.(6). For the case of vacuum general relativity,where L = Rp g, a simple calculation (cid:0) yields S = A=4 (9) bh and eq.(7) reduces to eq.(5). As already mentioned at the beginning of this section, the black hole analog of the second law of thermodynamics is the area theorem [2]. This theorem states that if Einstein’s equation holds with matter satisfying the null energy condition (i.e., T kakb 0 for all null ka), then the surface area, ab (cid:21) A, of the event horizon of a black hole can never decrease with time. In the context of more general theories of gravity, the nondecrease of S also has bh been shown to hold in a class of higher derivative gravity theories, where 7 the Lagrangian is a polynomial in the scalar curvature [13], but, unlike the zeroth and (cid:12)rst laws, no general argument for the validity of the second law of black hole mechanics is known. However, there are some hints that the nondecrease of S may hold in a very general class of theories of gravity bh with positive energy properties [14]. Takentogether,thezeroth,(cid:12)rst, andsecondlaws1 ofblackholemechanics in general relativity are remarkable mathematical analogs of the correspond- ing laws in ordinary thermodynamics. It is true that the nature of the proofs of the laws of black hole mechanics in classical general relativity is strikingly di(cid:11)erentfrom thenature of the arguments normallyadvanced for thevalidity of the ordinary laws of thermodynamics. Nevertheless, as discussed above, the validity of the laws of black hole mechanics appears to rest upon general features of the theory (such as general covariance) rather than the detailed form of Einstein’s equation, in a manner similar to the way the validity of the ordinary laws of thermodynamics depends only on very general features of classical and quantum dynamics. In comparing the laws of blackhole mechanicsin classical general relativ- ity with the laws of thermodynamics, the role of energy, E, is played by the mass, M,oftheblackhole;theroleoftemperature,T,isplayedbyaconstant times the surface gravity, (cid:20), of the black hole; and the role of entropy, S, is played by a constant times the area, A, of the black hole. The fact that E and M represent the same physical quantity provides a strong hint that the mathematicalanalogy betweenthe laws of blackhole mechanicsand the laws of thermodynamics might be of physical signi(cid:12)cance. However, in classical general relativity, the physical temperature of a black hole is absolute zero, so there can be no physical relationship between T and (cid:20). Consequently, it also would be inconsistent to assume a physical relationship between S and A. As we shall now see, this situation changes dramatically when quantum 1It should be noted that I have made no mention of the third lawof thermodynamics, i.e.,the \Planck-Nernst theorem", which states that S 0 (or a \universalconstant") as ! T 0. The analogof this law fails in black hole mechanics, since there exist \extremal" ! black holes of (cid:12)nite A which have (cid:20)=0. However, I believe that the the \Planck-Nernst theorem" should not be viewed as a fundamental law of thermodynamics but rather as a property of the density of states near the ground state in the thermodynamic limit, which isvalidfor commonlystudied materials. Indeed, examples can be givenofordinary quantum systems that violate the \Planck-Nernst theorem" in a manner very similar to the violations of the analog of this law that occur for black holes [15]. 8 e(cid:11)ects are taken into account. 3 Quantum black hole thermodynamics The physical temperature of a black hole is not absolute zero. As a result of quantum particle creation e(cid:11)ects [16], a black hole radiates to in(cid:12)nity all species of particles with a perfect black body spectrum, at temperature (in units with G = c = h(cid:22) = k = 1) (cid:20) T = : (10) 2(cid:25) Thus, (cid:20)=2(cid:25) truly is the physical temperature of a black hole, not merely a quantity playing a role mathematically analogous to temperature in the laws of black hole mechanics. In fact, there are two logically independent results which give rise to the formula (10). Although theseresults are mathematicallyverycloselyrelated, it is important to distinguish clearly between them. The (cid:12)rst result is the original thermal particle creation e(cid:11)ect discovered by Hawking [16]. In its most general form, this result may be stated as follows (see [11] for further discussion): Consider a a classical spacetime (M;g ) describing a black hole ab formed bygravitational collapse, suchthat the blackhole \settles down" to a stationary (cid:12)nal state. By the zeroth law of black hole mechanics, the surface gravity,(cid:20), of the black hole (cid:12)nal state willbe constant overits eventhorizon. Consider a quantum (cid:12)eld propagating in this background spacetime, which is initially in any (non-singular) state. Then, at asymptotically late times, particles of this (cid:12)eld will be radiated to in(cid:12)nity as though the black hole were a perfect black body2 at the Hawking temperature, eq. (10). It should be noted that this result relies only on the analysis of quantum (cid:12)elds in the regionexteriortotheblackhole,anditdoes notmakeuseofanygravitational (cid:12)eld equations. The second result is the Unruh e(cid:11)ect [17] and its generalization to curved spacetime. In its most general form, this result may be stated as follows (see [18], [11] for further discussion): Consider a a classical spacetime (M;g ) ab 2Iftheblackholeisrotating,thethespectrumseenbyanoberveratin(cid:12)nitycorresponds to what would emerge from a \rotating black body". 9 that contains a bifurcate Killing horizon, = , i.e., there is a one- A B K K [K parameter group of isometries whose associated Killing (cid:12)eld, (cid:24)a, is normal to . Consider a free quantum (cid:12)eld on this spacetime. Then there exists K at most one globally nonsingular state of the (cid:12)eld which is invariant under the isometries. Furthermore, in the \wedges" of the spacetime where the isometries have timelikeorbits, this state (if it exists) is a KMS (i.e., thermal equilibrium) state at temperature (10) with respect to the isometries. Note that in Minkowski spacetime, any one-parameter group of Lorentz boosts has an associated bifurcate Killing horizon, comprised by two inter- secting null planes. The unique, globally nonsingular state which is invariant under these isometries is simply the usual (\inertial") vacuum state, 0 >. j In the \right and left wedges" of Minkowskispacetime de(cid:12)ned by the Killing horizon, the orbits of the Lorentz boost isometries are timelike,and, indeed, these orbits correspond to worldlines of uniformly accelerating observers. If we normalize the boost Killing (cid:12)eld, ba, so that Killing time equals proper time on an orbit with acceleration a, then the surface gravity of the Killing horizon is (cid:20) = a. An observer following this orbit would naturally use ba to de(cid:12)ne a notion of \time translation symmetry". Consequently, when the (cid:12)eld is in the inertial vacuum state, a uniformly accelerating observer would describe the (cid:12)eld as being in a thermal equilibrium state at temperature a T = (11) 2(cid:25) as originally found by Unruh [17]. Although there is a close mathematical relationship between the two re- sults described above, it should be emphasized these results refer to di(cid:11)erent states of the quantum (cid:12)eld. In the Hawking e(cid:11)ect, the asymptotic (cid:12)nal state of the quantum (cid:12)eld is a state in which the modes of the quantum (cid:12)eld that appear to a distant observer to have propagated from the black hole re- gion of the spacetime are thermally populated at temperature (10), but the modes which appear to have propagated in from in(cid:12)nity are unpopulated. This state (usually referred to as the \Unruh vacuum") would be singular on the white hole horizon in the analytically continued spacetime containing a bifurcate Killing horizon. On the other hand, in the Unruh e(cid:11)ect and its generalization to curved spacetimes, the state in question (usually referred to as the \Hartle-Hawking vacuum") is globally nonsingular, and all modes 10