arXiv:2103.08340v3 [gr-qc] 30 Aug 2021 stecneto neethorizon. event an well as of effects concept strong-field relativit the some general as explore of to tests one allows classical and four straightforward the a of enables relativity simple, discussion is general metric in Their studied courses. spacetimes of examples ular i.I apn for happens It met Schwarzschild the in ric. arise inter- not the does particles that, infalling gravitational Beyond of phenomenon absent. esting are fields gravitational oe,dsrbdb h esnrNrsrmmetric, Reissner-Nordström the by described holes, erc unott xii ti oeacsil places. accessible more in it non-stationary exhibit Some to out horizon. turn event metrics the inside well only uso fisesnilfaue nSc I rcdn the results. preceding its dis- II, and a Sec. give model in our will features we of presentation essential Rather, its article. of in this cussion upon of touched part na- hardly main in be the electromagnetic will radiation mostly sta- Hawking is ture), a the (which of by itself models describable radiation are fully metrics Vaidya not radia- spacetime Since is Hawking by that metric. evaporating tionary object phenomenological hole an in black way tion, a natural a of in models arises It interest. gravita- either. of ones properties interesting the studying waves, in tional effects mass- discussed time-dependent i.e., be spectacular dust, may While null mass emits particles. symmetric or less spherically absorbs a that of distribution exterior the describing e nifligpril,bfr h atrcncosthe cross can latter the before the particle, un- in infalling evaporates hole an appearing black the der that term found With they mass metric, time. this global the the on made dependent metric Schwarzschild they which in h erci ttoay oeet of effects so stationary, is metric The cwrshl lc holes black Schwarzschild ipetm-eedn erci h adametric, Vaidya the is metric time-dependent simple A n20,At n rumn osdrdatymodel, toy a considered Trautmann and Aste 2005, In academic of purely not is metric Vaidya the that Note h adamti:epce n nxetdtat fevapor of traits unexpected and expected metric: Vaidya The eutn rmHwigrdain o fthe of not radiation, Hawking from resulting mtigHwigrdain hsmti sepce ocapt multiple to small expected a is to metric up coordinates This radial radiation. Hawking emitting iea xeln prxmto otesaeiegoer i geometry spacetime the to ( approximation excellent an give ocmee o hr ie narc ihaphoton. horizo a evolving with radiation, race Hawking a holes, Black in time, Keywords: appa short demonstr the a given, to for is enough interpretation compete, An close to infinity. rest to from escape released may e particle as a is horizon of Third, hole properties its black the the cross of on horizon takes hole radius Schwarzschild black Fir the the Second, developed. of expl are evaporation first predictions a complete as interesting serve Several may model the fields. classroom, the In horizon). 5 M h non adamti sitoue samdlfranon-r a for model a as introduced is metric Vaidya ingoing The h adamti sntdvi fsome of devoid not is metric Vaidya the .INTRODUCTION I. ≥ M charged ⊙ ) ntttfrPyi,Ot-o-urceUiestt Mag Otto-von-Guericke-Universität, Physik, für Institut xeta xrml ag itne rmtehrzn(exce horizon the from distances large extremely at except , 1,2 peial ymti black symmetric spherically r mn h otpop- most the among are repulsion inside uisPenc n lu Kassner Klaus and Piesnack Julius time-dependent fi,bcuelgtcnecp rmtesrnigaprn h apparent shrinking the from escape can light because it, of fradially of Dtd uut3,2021) 30, August (Dated: ( 3 > but 1) y, 4 6 - fteShazcidrdu.Frlre ai,i will it radii, larger For radius. Schwarzschild the of oio.I a one out pointed was It horizon. tog n ecno aclt h akrato be- reaction metric back evolving the become the and flux should calculate radiation of Hawking cannot effects the disappearance tween we quantum final and where – strong, presumed hole, at – black radiation the Hawking the to of do close We dependence times time way. the consistent dynam- know mechanically black a quantum not evaporating producing a an from in describe cry hole would far of that a construction spacetime still The ical is metric photons. a as such such massless, particle emitted be the compati- must and light, a achieved, of that is on speed causality means with the not this bility at Because constant moves energy surface. is radiated null met- the parameter Vaidya a a mass on but to the spacelike leads which time in EF ric, the on dependent mass coordinates across (GP) continuous Gullstrand-Painlevé not is there). it infinity to that (diverging fact horizon the coordinate the time to the due to unrelated outside is in- horizon event coordinate the time side Schwarzschild the meaningful; ically iklti E)coordinates (EF) Finkelstein shells. radial respective attraction their the inside time, masses same different different the of at with There- experience, observers should velocity. stationary radii finite How- coordinate a two at positions. away fore, radiated radial is different for energy at instantaneously ever, the sitting change on observers will features. depends it all unrealistic only, parameter of coordinate mass the number time the a hit is since with this then example, Clearly, model For and toy time. hole) a proper the still finite the of in to singularity dissolution close central final (except problem of any time Schwarz- an without the towards radius cross falling ordinarily schild observer will well-defined an hole a black model, evaporating has this mak- least In at so meaning. time-dependent the spacetime, mass Schwarzschild across the the continuous ing of are horizon pa- coordinates mass event These time-dependent a introduce rameter. to point starting a nipoeetwssgetdi e.7 where 7, Ref. in suggested was improvement An hspolmmyb ovdb sn Eddington- using by solved be may problem This ns t atce rpe al nuhbefore enough early dropped particles st, naprn oio,adtetu event true the and horizon, apparent an tn htsc atcewudb able be would particle a such that ating h aeo srpyia lc holes black astrophysical of case the n rto fnnsainr gravitational non-stationary of oration ethrzni togyrple and repelled strongly is horizon rent r h hsc ftesaeiefor spacetime the of physics the ure sl swt neenlbakhole. black eternal an with as asily eug Germany deburg, ttn nhre lc hole black uncharged otating dn h omcparticle cosmic the eding 7 tn lc holes black ating htti oe sntphys- not is model this that 10,11 nta.Mkn the Making instead. orizon. 8,9 eetknas taken were s 2 non-perturbatively. Even a semiclassical self-consistent uum state in Minkowski spacetime (at past null infinity solution does not seem to be available, in which the met- I −) will, after the spacetime has become curved due to ric would be a classical field arising from the quantum the formation of a black hole, be associated with a state of mechanical expectation value Tµν of the stress-energy non-zero particle content near the event horizon. Some of tensor.12 Nevertheless, the simplesth i guess at what the these particles, mostly photons, escape to future null in- semiclassical time-dependent metric of an evaporating finity I +, and the calculation proceeds by following the black hole should look like, subject to the condition of field at I + backwards in time through the interaction a minimum of realism regarding energy transport, would with the collapsing star to I −, using analytic contin- be a Vaidya metric, leaving one with the freedom to spec- uation techniques to evaluate the spectral properties of ify largely arbitrary time-dependent mass functions.12 the so-created radiation, which turns out to be thermal As will be discussed below, if the dynamical horizon of with temperature and entropy given by the Bekenstein the description is to correspond to a future horizon, the expressions. The field calculation may be visualized in metric must be an ingoing Vaidya metric. Such a metric terms of a wave picture, with the properties of the in- with a mass parameter that decreases as a function of going wave coming from I − modified by its interaction time corresponds to a black hole losing mass by infalling with the event horizon. Detailed considerations involve negative-energy null dust. This feature, while acceptable negative frequency components of the wave falling into near the horizon (as will be seen in the next section), the black hole and positive frequency components being is not very realistic far from it, where Hawking radia- amplified, allowing part of them to escape to I +. tion clearly is outgoing and consists of positive-energy It should be noted that Hawking himself provided an particles. Then, to add more realism to the model, the alternative picture of the process already in the extended ingoing Vaidya metric near the horizon should be comple- version of his first explanation of the effect.20 It focuses mented by an outgoing one at larger radial coordinates, on the particle content of the quantum fields involved with the surface joining the two metrics corresponding rather than on the decomposition of the field into waves to a pair creation location.13 A few years ago, a model with amplitudes given by annihilation and creation op- for an evaporating black hole without any spacetime sin- erators. In this picture, Hawking described a negative gularities was presented that employs these features.14 energy flux inward across the horizon, coming from the Since our purpose is to present a model that is useful surrounding region in which, by virtue of quantum fluc- in exploring physical effects in the classroom, we stick to tuations, pairs of virtual particles appear and disappear. the simplest case discussed by Hiscock12 containing some One of these typically has positive and the other nega- elements of realism, i.e., just an ingoing Vaidya metric. tive energy (so as to satisfy energy conservation in the Local results near the horizon then should have validity, long run). He then stated:20 The negative particle is in but the assessment of results obtained at large radii needs a region which is classically forbidden but it can tunnel additional scrutiny. through the event horizon to the region inside the black The remainder of the paper is organized as fol- hole where the Killing vector which represents time trans- lows. After the explanatory Sec. II on Hawking radia- lations is spacelike. In this region the particle can exist tion, we consider, in Sec. III, the maximally extended as a real particle with a timelike momentum vector even Schwarzschild metric in a diagram using Kruskal-Sze- though its energy relative to infinity as measured by the keres (KS) coordinates.15,16 Discussing which regions of time translation Killing vector is negative. The other par- the full spacetime are covered by ingoing and outgoing ticle of the pair, having a positive energy, can escape to EF coordinates, we decide which of the two sets is the infinity where it constitutes a part of the thermal emis- best starting point in constructing a Vaidya metric. Sec- sion described above. [...] Instead of thinking of neg- tion IV gives the equations of motion for a test particle ative energy particles tunnelling through the horizon in falling radially into the model black hole and simplifies the positive sense of time one could regard them as posi- them to a form that can be compared with the Newtonian tive energy particles crossing the horizon on past-directed limit. These equations are solved numerically in Sec. V. world-lines and then being scattered on to future-directed The existence of repulsive effects of the time-dependent world-lines by the gravitational field. Although Hawk- gravitational field is demonstrated analytically. Finally, ing emphasized that this is a heuristic picture, not to some conclusions are given in Sec. VI. be taken too literally, it may be more appealing than the picture of positive-frequency and negative-frequency wave components, where the energy balance is not read- II. HAWKING RADIATION ily visible with the overall wave consisting predominantly of positive-frequency components. Hawking provided a physical mechanism for Beken- Of course, when talking of particles here, we mean stein’s idea that the area of the event horizon of a black quantum particles without committing ourselves to their hole corresponds to its entropy and its surface gravity to manifestation in a localized or wavy manner. In fact, its temperature17,18 (each with the appropriate propor- the particles of Hawking radiation are not very point- tionality constant). In two papers exploring this quan- like. The wavelength of the maximum of the spectrum tum mechanical mechanism,19,20 he presented a calcu- of Hawking radiation for a solar-mass black hole is about lation based on quantum field theoretical considerations, 47 km at infinity, whereas the Schwarzschild radius is a according to which a field operator associated with a vac- mere 3 km. So these photons would be detectable with 3 antennas rather than with photomultipliers, i.e., they one should expect an infaller to take longer for the evap- would be considered waves rather than particles. More- orating black hole than for the static one. But then the over, they are so poorly localized that it is impossible black hole must disappear, before the observer can fall to indicate a single value for the redshift they have un- into it. This of course begs the question how an event dergone since their creation near the black hole horizon horizon could form in the first place, because the surface (because the redshift is significantly different for differ- of the star will, on collapse, behave similarly to the in- ent parts of the wave). A wave emerging at about two faller. On the other hand, observers can fall into eternal Schwarzschild radii from the center of the geometry will black holes within a finite interval of their proper time have only a roughly 40% higher frequency at that po- and Hawking radiation is a weak effect for stellar-mass sition than at infinity, meaning that its most probable black holes, which suggests that crossing of the event wavelength would still be more than 30 km. Photons horizon by an observer should happen essentially in the with such wavelengths might easily tunnel distances on same way as for a static black hole. But it is a contradic- the order of 10 km, i.e., three Schwarzschild radii. Hence, tion for an observer to both fall and not fall into a black in adopting the particle picture, we should not imagine hole. This paradox has been resolved in Ref. 7 modeling the radiation to come from the surface constituted by the the metric of an evaporating black hole appropriately, horizon. Its location of creation may be washed out with by use of a time coordinate that does not diverge at the an uncertainty on the order of the Schwarzschild radius. horizon and is resolved in this paper the same way (but In fact, the “quantum atmosphere”, in which Hawking using a different time coordinate). radiation is created, has recently been estimated, from a In such a model, there must be a region of negative en- 1+1 dimensional calculation, to extend in radial coordi- ergy density outside the horizon, as transpires from the nate from 1.5 to 2 Schwarzschild radii.21 quote reproduced above from Hawking’s 1975 paper.20 In Newer calculations of Hawking radiation have been classical general relativity, negative energy density is con- presented22 which are much closer in spirit to the picture sidered forbidden and solutions requiring it are viewed as of tunneling quantum particles than Hawking’s original unrealizable, but in the presence of quantum mechanics, one. One advantage of calculations that are manifestly they cannot be discarded automatically. Our considera- based on a tunneling process is that it is not necessary to tions in the following will be based on a classical model, consider the complete collapse geometry before formation but negative energy density is admitted in order to mimic of the black hole. While the details of the collapse do not the quantum effect of negative-energy virtual particles. influence the final result, the picture usually drawn, hav- ing null trajectories in a Penrose diagram that connect I − with I + (and pile up near the horizon) may favor III. THE SCHWARZSCHILD SPACETIME IN misconceptions such as the idea that all of the Hawking TERMS OF EDDINGTON-FINKELSTEIN radiation comes from the time before the surface of the COORDINATES collapsing star crosses the horizon,23 which would mean that the black hole remains incipient forever, i.e., that In this section, we illustrate the idea that different co- it does not actually form. A nice compact discussion of ordinate systems may cover different parts of the space- features of Hawking radiation may be found in Ref. 24. time manifold. The necessity to capture certain features Experimental verification of Hawking radiation from a (such as a future event horizon) in the description may typical black hole with a mass exceeding that of our sun then restrict the choice of possible coordinates. is virtually impossible, because the effect is so tiny. It The maximally extended Schwarzschild solution, corre- has been suggested25 that sonic black holes (in which a sponding to an eternal spherically symmetric black hole – fluid takes on supersonic speeds so that there should be and a bit more – is describable via KS coordinates. The Hawking-like sound emission from the “sonic horizon”) resulting diagram given in Fig. 1 will also be useful in might allow experimental access to the phenomenon. exhibiting the difference between the two EF coordinate Indeed, success in detecting Hawking radiation from a systems. Bose-Einstein condensate as a black-hole analog was re- The metric is27 ported recently.26 4r3 Returning to real black holes, an object falling towards ds2 = s e−r/rs dT 2 dX2 r2dΩ2 , (1) the event horizon seems to get slower and freeze there due r − − to the fact that photons take longer and longer to escape where r = 2GM/c2 defines the Schwarzschild radius from the vicinity of the horizon. What is more, it turns s (M is the black hole mass, G Newton’s gravitational out that the object will take infinite Schwarzschild time constant, c the vacuum speed of light) and dΩ2 = to actually reach the horizon, so the slowing-down would, dϑ2 + sin2 ϑdϕ2 abbreviates the line element on a unit it appears, not just be an optical illusion. That a black sphere. The radial coordinate r of the Schwarzschild hole dissolves via Hawking radiation in finite time then metric is expressible via X and T with the help of leads to an apparent paradox: If it takes infinite time X2 T 2 = r 1 er/rs . for an observer to fall into an eternal black hole with − rs − fixed event horizon, falling into an evaporating black hole The diagram of Fig. 1 may be read as giving a two- should also take infinite time. After all, the attracting dimensional section of the spacetime at fixed ϑ and ϕ. mass decreases and the horizon recedes, so if anything, One of its interesting features is that radial light rays 4
In practice, identification is always made with region II, T r = rs because the horizon between regions IV and I, sometimes t = ∞ called antihorizon,31 is permeable from IV to I only and thus corresponds to a white hole (into which nothing can r =0 A fall, but from which stuff may be ejected). The vacuum solution, described by the metric (1) is not applicable in- ∗ II t = t side a star, so observers in region I will see an illusory B horizon31 (the surface of the collapsing star) instead of I an antihorizon. Once a black hole has formed, the vac- uum part of the metric will resemble that of an eternal III X black hole with only regions I and II explorable. IV A pair of photons, one escaping one infalling, is in- dicated in the figure, the first drawn as a solid arrow; r =0 the second as a dashed arrow.32 This is to illustrate the point that for an observer A, whose approximate posi- r = r tion is marked in the figure, Hawking radiation seems to s 31 t = −∞ come from the illusory horizon, because he will not see the infalling photon and the trajectory of the escaping one is parallel to the horizon in the KS diagram. How- FIG. 1. Kruskal-Szekeres diagram of the maximally extended ever, an observer B who is closer to the illusory horizon Schwarzschild metric. Loci of constant Schwarzschild time co- will not see that particular photon, because it does not ordinate t correspond to straight lines through the origin (a ∗ line t = t is drawn explicitly). Loci of constant Schwarzschild really come from the surface of the collapsing star. For radial coordinate r are given by hyperbolas having the an- a shrinking black hole, observer A will see radiation that gle bisectors of the axes as asymptotes (the lines r = 0 are was emitted at a slightly higher temperature than that 33 drawn) or, in the limit r = rs, by these bisectors them- seen by B, and the difference in photon content was cre- selves. Two sets of Schwarzschild coordinates are necessary ated in the spacetime interval between the two observers, to cover all four regions. The solid hatch lines correspond to which typically will mean that it was created after the ob- (non-equidistant) constant null coordinates v in the ingoing servation made at B’s spacetime position but before the Eddington-Finkelstein metric, the dashed ones to constant observation by A. An observer falling freely through the coordinates u in the outgoing EF metric. Other features are horizon should not detect any Hawking radiation at all at explained in the text. the crossing point, if the equivalence principle continues to hold at the horizon, as suggested by Ref. 34.35 Let us now discuss EF coordinates. One way to obtain are parallel to either of the two bisectors of the pair of them from Schwarzschild coordinates is to first introduce coordinate axes, so it is easy to discuss the exchange of the tortoise coordinate light signals of observers in the geometry.28 Event hori- r r∗ = r + r ln 1 , (2) zons are given in Fig. 1 by the future and past light s r − s cones of the origin. The Schwarzschild time coordinate which approaches as r rs, and then to introduce, is timelike only in the regions I and III (X > T and −∞ → X < T , respectively) but spacelike in the| regions| as new time coordinate, either − | | II and IV (T > X and T < X , respectively, with r∗ 2 2 | | − | | v = t + , (3) T X < 1), because the absolute value of the slope c of any− straight line through the origin is bigger than 1 yielding the ingoing EF metric with coordinates v, r, ϑ, in these two regions.29 Correspondingly, the coordinate and ϕ or r behaves like a spatial coordinate in regions I and III, but like a time coordinate in regions II and IV. r∗ u = t , (4) Customarily, region I (r > r , < t < ) is − c s −∞ ∞ considered as describing the outside of a Schwarzschild which gives the outgoing EF metric with coordinates u, black hole. The complete set of Schwarzschild coordi- r, ϑ, and ϕ. It is then easy to determine the coordinate nates, including r < r , 1 1 r T = e−cu/2rs + 1 er/rs+cu/2rs , (6) r, as M(v) diminishes with increasing v. This must be so −2 2 r − for the ingoing Vaidya metric, because with increasing r, s more and more negative-energy null dust is between the respectively. We note that all values 2 that rs(0) = 2GM0/c ). In Ref. 12, it is argued that the for the free fall of a particle in the gravitational field of dependency (9), based on a fixed-background calculation, a point mass (or a spherically symmetric mass distribu- cannot hold down to mass zero, as this would lead to an tion with radius smaller than r). Since the time deriva- infinite flux of radiated particles. At least near the end of tives are with respect to proper time rather than absolute evaporation, the functional law must then be modified, Newtonian time, the dynamics will look Newtonian only an effect that we have studied but which has little impact at sufficiently small velocities, as long as the difference on the results given here, so we skip its discussion. between the rate of proper time of the particle and that The equations of motion for a test particle falling freely of an external observer does not become visible. in the Vaidya spacetime may be obtained from the La- If we now take a temporal variation of the mass term rs grangian into account, then we have both contributions, the first being a generalized Newtonian law with time-dependent r (v) ; 2 = 1 s c2v˙2 2 c v˙r˙ r2 ϑ˙2 + sin2 ϑϕ˙ 2 , mass term (rs(v) M(v)) and the standard 1/r depen- L − r − − dence, whereas the second term contains the time deriva- (14) tive of the mass and decays as 1/r. For diminishing mass, the second term is positive, so it is a repulsive contribu- in which a dot denotes a derivative with respect to the tion to the acceleration of the test particle. proper time τ of the particle. The Euler-Lagrange equa- In a Newtonian universe, we would expect only the tions read first term to be present (with the advanced time in the d ∂ ∂ argument of rs replaced by absolute time). This should L L =0 . (15) be true even when the mass varies with time, because dτ ∂q˙ − ∂q Newtonian gravity spreads instantaneously. Hence, what Writing them out for q = ϕ and q = ϑ we find that we see in general relativity instead is an effect of gravity they are solved by ϕ = const. and ϑ = const. Therefore, traveling at a finite velocity. purely radial geodesics exist and to find these, we may Let us estimate the relative sizes of the second and first consider an effective Lagrangian, obtained from Eq. (14) terms. For v < vl, the ratio between their magnitudes is by dropping the last two terms. The equations for the ′ ′ 2 2 2 two other coordinates then are (r dr /dv = r (v)): rs(v) rv˙ rv˙ rv˙ s,v s s Q = | | = . (20) ≡ r (v)c 3c(v v) ≈ 3cv r r r s l − l 1 s cv¨ + s cv˙ r˙ r¨ s,v cv˙ 2 =0 , (16) − r r2 − − 2r The last approximation is valid for black holes having rs v¨ + cv˙ 2 =0 , (17) a mass bigger than that of the sun and up to times v 2r2 well exceeding the current lifetime of the universe (13.8 9 67 74 × and the definition of the effective Lagrangian will be use- 10 y), because we have vl > 2.1 10 y =6.6 10 s. From Eq. (18), we obtain the estimates× v˙ < 1/× 1 rs ful for simplifications of the equations of motion: − r √ rs for negative r˙ and v˙ < (1 + 2)/(1 r ) for positive 2 − p ds 2 rs(v) 2 2 r Closer to the horizon, the static approximation may be insufficient, because the second term of Eq. (19) can be- 10 rs(v) come large. But there, the ingoing Vaidya metric should r =10.0 be good, because it faithfully reproduces the negative en- 0 8 r0= 8.0 ergy density near the horizon. According to a fairly rigor- r = 6.0 37 0 ous calculation in 1+1 dimensions, the sum of the out- r0= 4.0 6 r = 2.0 going and ingoing fluxes in the radial rest frame, which is 0 proportional to T +T , becomes zero at r 1.6r (and r/r^ uu vv ≈ s is negative for smaller r), so the Vaidya metric with its 4 negative energy density should be good up to about this radial coordinate, and this should be true independent of the mass of the black hole. This is also consistent with 2 Ref. 21. Hence, for black holes of realistic size, approxi- mating the metric via the ingoing Vaidya metric will be 0 viable for all interesting regions of spacetime. Moreover, 0 20 40 60 80 100 ^ it has the advantage of analytic accessibility that is lost, v/v if instead we patch two or more metrics together13 in order to gain a tiny bit of accuracy at larger r. FIG. 2. Trajectories of radially infalling particles in the Vai- In order to render the time-dependent Schwarzschild dya metric (8). rˆ = rs(0), vˆ = rs(0)/c. The unit rˆ has been left out in the legend for brevity, i.e., r0/rˆ has been renamed radius visible in our figures in the next section, we have to r0. to consider black holes with much smaller masses than will be found in astrophysics. For these, we expect the ingoing Vaidya metric to be a good approximation up fixed (at the value vˆ = rs(0)/c). By choosing a nondi- to r = 1.6rs and a decent approximation out to a few more Schwarzschild radii, so answers obtained from it mensional value for the time vl of evaporation, defined should be qualitatively correct. Results will become qual- by rs(vl)=0, we therefore fix a physical quantity, which itatively incorrect, once the second term of Eq. (19) be- is the initial mass of the black hole. To see this, consider comes bigger than the first one at large r, due to the v cv πG M 2 l = l = 2560M 2 = 2560π 0 , (21) fact that the second term falls off with a weaker power vˆ r (0) 0 ~c m2 of r than the first. The dominant second term will then s p have the wrong sign, because it would be attractive for −8 where mp = ~c/G =2.176 10 kg is the Planck mass. an outgoing Vaidya metric. For the values v /vˆ = 100, 10×, and 5 used in our figures, p l we have M0 = 0.112 mp, 0.0352 mp, and 0.0249 mp, re- spectively, extremely small masses indeed. Black holes V. NUMERICAL SOLUTION AND ANALYTIC with such a small mass cannot arise from direct gravita- DISCUSSION tional collapse. They might be the result of a density fluc- tuation in some violent subatomic process, of a kind that To gain understanding about the dynamics in this met- could possibly have arisen very shortly after the big bang. ric, the equations of motion should be integrated numer- These black holes would, however, have long since van- ically for exemplary situations. They are simple enough ished. Yet, it is conceivable that some primordial black to make this a nice practical exercise. holes starting out with masses around 1.8 1011 kg, ap- In a number of cases, we solved the geodesic equa- proaching the end of their life today, might× have current tions beyond the evaporation time, where the derivative masses close to a Planck mass. of r (v) has a singularity (v v)−2/3. This poses an The reason for considering such unrealistic cases here s ∝ l − obstacle to direct numerical integration. Solvers for or- is that for a black hole of about one solar mass, the 80 dinary differential equations with time step control will nondimensional vl would be on the order of 10 , and not normally cross that point, in attempting to resolve the Schwarzschild radius would look constant in our fig- the singularity. Therefore, a numerical approach avoid- ures. To see a particle actually survive the evaporation ing the appearance of diverging quantities was developed. of the black hole, we must tune its starting distance and Its details will not be presented here, in order to keep the time so that it closes in on the horizon position only near paper concise.38 the point of complete evaporation (or later). In Fig. 2, the trajectories of a particle released Before looking at this kind of behavior, let us briefly from rest at different heights and approaching the clarify whether the Schwarzschild radius rs(v) is still an Schwarzschild radius well before evaporation are pre- event horizon, when rs is decreasing as a function of time. sented. The parameters are the same as in Fig. 1 of The equation of motion for an outgoing radial light ray Ref. 7 and the outcome is similar. The particle falls into follows from Eq. (8) by setting dΩ = 0, ds = 0, and the black hole in all cases. dv =0: Our results are given in terms of nondimensional vari- 6 dr 1 rs(v) ables, with rˆ rs(0) chosen as the length unit and the = c 1 . (22) ≡ dv 2 − r speed of light set equal to 1. Then the time unit is also 8 This has been integrated in Fig. 3 for a few initial values r (v) r = r0 close to rs(0). 4 s r0=4.0 r0=3.5 3.0 r (v) s r0=3.0 r =1.2 0 3 r0=2.5 r =1.0 2.5 0 r0=2.0 r0=0.9 r0=1.5 r =0.88 ^ 2.0 0 r/r 2 r0=0.87 r =0.8 ^ 0 r/r 1.5 1 1.0 0 0.5 0 5 10 15 20 v/v^ 0.0 0 1 2 3 4 5 6 v/v^ FIG. 4. Trajectories missing and hitting the black hole. FIG. 3. Light sent radially outward from positions near the initial Schwarzschild radius. happen, the second term of Eq. (19) must become larger A light ray sent from the Schwarzschild radius rs(0) than the first one, due to r getting large, the ingoing will obviously escape, having zero coordinate velocity Vaidya metric will not be a good model for an evaporat- at v = 0 and positive velocity at any later time. In- ing black hole anymore, so we will not consider this case terestingly, some rays starting their journey with r < here. rs(0) manage to cross the shrinking Schwarzschild radius, Let us instead determine the range of r0 values, for which therefore is not a true event horizon anymore, but which the force on the particle is repulsive already in the rather an apparent horizon. An apparent horizon is a initial phase of its fall. Since we release it from rest, we limiting trapped null surface; i.e., outward-directed light may assume that r˙ 0. Using (18), we then express v˙ 2 ≈ rays emitted outside the apparent horizon will move out- by functions of r and v alone and find from Eq. (19) ward, while outward-directed light rays emitted inside it r (v)c2 r′ (v)c will move inward. r¨ = s + s . (23) − 2r2 2(r r) Escaping light rays from inside start falling inward but s − do so more slowly than the horizon shrinks, so they are Setting r¨ =0, we obtain a quadratic equation for r with passed by the latter and afterwards move outward. Light the roots40 sent from sufficiently far inside this apparent horizon will, 2 2 2 rsc rs c rs c rsc however, fall into the singularity, so there still exists an r01 = + + =3cvl , (24) 2 ≪ event horizon, separating events from which null infinity −2rs,v s4rs,v rs,v rs≈ r −rs,v can never be reached from those that may send signals to r c r2c2 r2c r null infinity. That horizon, however, does not exist indef- s s s s r02 = 2 + rs 1+ . −2r − 4r r rs,v≈≪c 3cv initely; it is gone after the time vl. Moreover, if quantum s,v s s,v s,v l mechanics leads to avoidance of the singularity as is gen- (25) erally believed, possibly replacing it with a fuzzy region of fluctuating spacetime that contains quantum fields at Repulsion dominates for r0 > r01 and r0 < r02. The high energy density, whatever was caught in that high- first result is irrelevant for stellar-mass evaporating black holes, because it is exorbitantly large, bigger than 6.3 energy density region would get out again (albeit trans- 67 50 × formed into photons or other elementary particles), once 10 ly, well beyond 10 times the particle horizon of the the region dissolved, and escape to future null or time- universe. like infinity. But then there would be no event horizon On the other hand, the second result tells us that there by definition. is repulsion in the ingoing Vaidya metric with decreasing mass slightly outside, but near, the Schwarzschild radius. Let us next consider a situation, where it is possible for And this close to the horizon, the ingoing Vaidya metric a particle to miss the singularity in the black hole. This can be safely assumed to yield a decent description! is depicted in Fig. 4. Particles starting from a radius We first verify the validity of the analytic prediction slightly exceeding 3.2 rs(0) will not cross the apparent by numerical simulation. Fig. 5 gives the trajectories, for horizon. The inward velocity of both the particle starting vl/vˆ = 10, of a particle starting at a radius that exceeds at 3.5 r (0) and that starting at 4 r (0) decreases a little s s rs(0) by just one percent and of one starting only one just before complete evaporation, which is obviously due thousandth of rs(0) above the apparent horizon. Eq. (25) to the repulsive term in (19). predicts initial repulsion for rs < r0 < 1.033 rs here. The model can even produce overall repulsion after an The first particle starts moving away from the appar- initial phase of inward falling.39 However, since for this to ent horizon, being repelled more strongly than attracted, 9 to infinity. Indeed, we find in simulations (not shown) 5 that a particle which is started slightly inside the initial Schwarzschild radius with a coordinate velocity not sig- 4 nificantly below u0 = c rs(0)/r0 1 manages to cross the apparent horizon− and to escape− to future timelike p 3 infinity. We expect this result to be robust with regard to a ^ r/r modification of the metric outside of a neighborhood of rs(v) 2 r0= 1.010, particle rs. Any outside metric patched to the ingoing Vaidya r = 1.001, particle 0 metric must not add an additional horizon. Light that r0= 1.001, light 1 reaches the outer metric will therefore escape to future infinity. A particle accelerated to relativistic velocities by the inner metric will then generally also escape to infinity 0 0 5 10 15 20 on entering the outer metric. v/v^ VI. CONCLUSIONS FIG. 5. Trajectories of particles released from rest at r0 = 1.01 rs(0) and r0 = 1.001 rs(0). The second particle escapes to infinity. The path of a light ray starting from the same r0 The Vaidya metric is an appealing tool to present as the second particle is given for comparison. certain effects of non-stationary gravitational fields in classroom, having applications to black-hole evaporation. Simple analytical results demonstrate interesting phys- ical effects, in particular gravitational repulsion in the but as its distance from rs(v) increases, it is pulled back, presence of a shrinking apparent horizon. The numerical turns around, and finally falls into the singularity. How- studies presented here are not too advanced, and can be ever, the second particle, starting significantly closer to assigned to students as exercises. the apparent horizon, experiences such a strong repulsion Use of the ingoing Vaidya metric with a decreasing that it actually escapes from the black hole. In order to mass term is motivated by its utility in modeling a space- gain some understanding, we have also plotted the tra- time geometry with a shrinking future (apparent) hori- jectory of an outgoing light ray starting together with the zon. We have argued that this kind of description is second particle. What becomes immediately clear is that reasonable near and outside the apparent horizon, be- the two trajectories are pretty close to each other, so the cause negative energy density is expected to arise there particle reaches a speed that is close to the speed of light. due to quantum effects,20,37 and the Vaidya metric con- It is then easy to form a meaningful mental picture by sidered describes an inflow of negative energy carried remembering the river model of a black hole, proposed by by null dust. If we take Hawking’s view that this is 41 Hamilton and Lisle. According to this idea, space about the same as an outflow of positive-energy particles along a black hole may be viewed as flowing inward across the past-directed geodesics, then we can say that what the event horizon, where it reaches the speed of light. This ingoing Vaidya metric fails to account for is the final explains nicely why nothing can cross the event horizon scattering of these quanta off the gravitational field that from the inside out – motion relative to space cannot ex- makes their path future-directed again. This would cer- 42 ceed the speed of light. With the Vaidya metric, inflow- tainly be bad, if the purpose was to describe these par- ing space reaches the speed of light at the time-dependent ticles, i.e., Hawking radiation itself. However, we were 43 apparent horizon rather than at the event horizon, but mainly interested in physical effects of the modifications otherwise the situation is similar. Positioning a parti- of the metric brought about by Hawking radiation. For cle at rest slightly above the apparent horizon actually most of the time a black hole exists, Hawking radiation means giving it a very high outward speed with respect is a small effect. The metric at some distance from the to the inflowing space. The closer to the apparent hori- horizon will essentially be the Schwarzschild metric; i.e., zon this is done, the closer the outward speed is to the rs changes so slowly that the second term on the right- speed of light. A particle extremely close to the apparent hand-side of Eq. (19) is negligible and it also does not horizon should therefore move like a photon. Since out- really matter whether the argument of the first term is going light will always escape when emitted outside the the advanced time v, the retarded time u or even the horizon, a massive particle close enough to it will do the Schwarzschild time t. same. This is no longer true near the end of the black hole’s Note that it is impossible to start a particle from rest life, because then Hawking radiation becomes a strong inside the apparent horizon, as its velocity with respect effect and even may dominate the metric. We must then to the inflowing space would have to exceed c. However, model it more accurately in the whole spacetime. It is it is possible to give a particle slightly inside, and close to, therefore concluded that results (not discussed here) such the apparent horizon an initial condition corresponding as net repulsion after full evaporation39 are an artifact of to a sufficiently small inward velocity so that the shrink- the ingoing Vaidya metric in most cases. This effect is ing horizon will catch up with it, allowing it to escape not likely to occur for a black hole at the end of its life. 10 A more sophisticated approach, properly reproducing because of the presence of negative energy density. Re- the (future-directed) outflow of positive energy far from pulsion in the Reissner-Nordström metric is explicable the horizon would be to keep the ingoing Vaidya metric because of negative pressure (which is stress),3 exerted for radii up to a small multiple of rs and use an outgoing by the electric field. In the equations of motion, an elec- Vaidya metric for larger r.13 There would be matching trostatic potential term appears, representing negative conditions for the places where the two metrics meet. energy. This then suggests that it is the negative energy Realizing this approach might be a bit too involved for density in both the Reissner-Nordström solution and the the classroom, and it would not provide any quantitative Vaidya metric that is responsible for the appearance of gains for the description of realistic astrophysical black repulsion. A difference, however, is that the electrostatic holes. energy term in the Reissner-Nordström geodesic equa- The properties of evaporating black holes obtained tions exceeds the mass energy of the black hole inside the within our model and discussed here should be robust inner horizon. By contrast, a volume integral of the neg- against model variations aiming at a generally more ative energy density of the Vaidya metric, constructed to quantitative description. Two of these are pretty intu- represent an evaporating black hole, will be much smaller itive. The first is that, for particles falling towards the than the remaining mass energy of the black hole for all black hole sufficiently long before evaporation, there is times except very briefly before complete evaporation. no problem crossing the horizon. Hawking radiation does Moreover, the repulsion will happen only for particles al- not prevent them from falling in. This would be expected most coordinate stationary near the horizon. An infalling from a comparison of the time scales on which Hawking particle having already acquired a substantial inward ve- radiation becomes relevant with those for an infaller to locity at the horizon will not experience repulsion (v˙ re- cross the horizon, if a time coordinate is used that does mains finite at r = rs in (18) for r<˙ 0). In the Reissner- not become singular at the horizon. Second, the nature of Nordström case, the repulsion is velocity independent. the horizon changes. The event horizon of the stationary These considerations suggest that the presence of nega- limit of the Vaidya metric (which is the corresponding tive energy density alone does not explain the repulsion EF metric) turns into an apparent horizon. That is, it effect in the case considered here, although it may be a remains a trapped null surface, but light may eventually necessary condition for radial repulsion. escape from it due to its shrinking in time. Contrary On the other hand, there is no need to invoke energy to apparent horizons of growing black holes (the normal considerations at all to understand repulsion the way we case), this apparent horizon is outside of the event hori- presented it in discussing Fig. 5. If the apparent horizon zon, obviously because the standard condition of positive is receding, no matter what the reason, then a coordinate energy density everywhere is violated. stationary particle close enough to it will have essentially Moreover, it is expected that a counterintuitive aspect the speed of outgoing light relative to the inflowing space found here will prevail in a generally more quantitative in the river model of a black hole.41 Such a relativistic description, viz. the strong repulsion experienced by a particle will be ejected and is likely to escape from the particle released from rest just above the horizon. Such black hole. In this view, the repulsion is not due to a a particle may escape to infinity and this effect, predicted repelling force but rather an inertial effect. By contrast, from the Vaidya metric (but also from a time-dependent a particle that has a sufficiently large inward velocity generalization of the GP metric7) is likely to be a quali- slightly outside the horizon is much slower than light with tatively correct result. Once the particle has been accel- respect to the inflowing space and thus will not be able erated to nearly the speed of light in a thin shell above to avoid being captured. the horizon, it will also escape in a model, where the Finally, the system considered here seems to be the outer part of the metric has been changed into some- simplest one exhibiting radial gravitational repulsion out- thing more realistic. Because this outer metric will not side a horizon. In principle, our prediction of particles contain a second horizon, any relativistic particle should being ejected violently from near the horizon could be not have problems escaping, just as a light ray would. tested in a sonic black-hole analog,25 set up with a re- Arguably, the effect is not so counterintuitive after all, ceding horizon. 1 K. Schwarzschild, “Über das Gravitationsfeld eines pulsion,” Gen. Rel. Grav. 20, 123–129 (1986) Massenpunktes nach der Einsteinschen Theorie,” in 4 P. C. Vaidya, “The Gravitational Field of a Radiating Sitzungsberichte der Königlich-Preußischen Akademie der Star,” Proc. Indian Acad. Sci. A 33, 264–276 (1951) Wissenschaften (Reimer, Berlin, 1916) pp. 189–196, En- 5 A. Einstein and N. Rosen, “On gravitational waves,” J. glish translation: On the Gravitational Field of a Mass Franklin Institute 223, 43–54 (1937) Point According to Einstein’s Theory, S. Antoci and A. 6 A. Aste and D. Trautmann, “Radial fall of a test particle Loinger, arXiv: physics/9905030v1 onto an evaporating black hole,” Canad. J. Phys. 83, 1001– 2 J. Droste, “The field of a single centre in Einstein’s theory 1006 (2005) of gravitation, and the motion of a particle in that field,” 7 K. Kassner, “Radially falling test particle approaching Ned. Acad. Wet., SA 19, 197–215 (1917) an evaporating black hole,” Canad. J. Phys. 97, 267–276 3 Ø. Grøn, “Poincaré Stress and the Reissner-Nordström Re- (2019) 11 8 P. Painlevé, “La mécanique classique et la théorie de la constant value of that coordinate is spacelike (timelike). relativité,” C. R. Acad. Sci. (Paris) 173, 677–680 (1921) 30 The single event (X,T ) = (0, 0) has an infinity of time 9 A. Gullstrand, “Allgemeine Lösung des statischen Einkör- coordinates t, just as the pole of a sphere has an infinity perproblems in der Einsteinschen Gravitationstheorie,” of ϕ values in spherical coordinates. Arkiv. Mat. Astron. Fys. 16, 1–15 (1922) 31 A. J. S. Hamilton, “Hawking radiation inside a Schwarz- 10 A. S. Eddington, “A comparison of Whitehead’s and Ein- schild black hole,” Gen. Rel. Grav. 50, 50 (48pp) (2018) stein’s Formulæ,” Nature 113, 192 (1924) 32 The representation is symbolical, because pair creation 11 D. Finkelstein, “Past-Future Asymmetry of the Gravita- should, as noted before, not be considered a pointlike tional Field of a Point Particle,” Phys. Rev. 110, 965–967 event. (1958) 33 However, since A is at a larger r coordinate than B, the 12 W. A. Hiscock, “Models of evaporating black holes. I,” radiation he sees will be more strongly redshifted than that Phys. Rev. D 23, 2813–2822 (1981) seen by B. 13 W. A. Hiscock, “Models of evaporating black holes. II. Ef- 34 D. Singleton and S. Wilburn, “Hawking Radiation, Un- fects of the outgoing created radiation,” Phys. Rev. D 23, ruh Radiation, and the Equivalence Principle,” Phys. Rev. 2823–2827 (1981) Lett. 107, 081102 (2011) 14 S. A. Hayward, “Formation and Evaporation of Nonsingu- 35 A freely falling observer is in a local inertial system, so local lar Black Holes,” Phys. Rev. Lett. 96, 031103 (2006) experiments do not allow her to detect the black hole hori- 15 M. D. Kruskal, “Maximal Extension of Schwarzschild Met- zon on crossing it. If the horizon is a condition for Hawk- ric,” Phys. Rev. 119, 1743–1745 (1960) ing radiation in an almost stationary spacetime, then a 16 G. Szekeres, “On the singularities of a Riemannian mani- freely falling observer cannot see radiation from that hori- fold,” Publ. Math. Debrecen 7, 285–301 (1960) zon, once she reaches it. Radiation from a distant horizon 17 J. D. Bekenstein, “Black Holes and Entropy,” Phys. Rev. is not forbidden by the equivalence principle, as demon- D 7, 2333–2346 (1973) strated by the simple fact that an observer at infinity does 18 J. D. Bekenstein, “Generalized second law of thermody- observe Hawking radiation. Ref. 34 gives support to the namics in black-hole physics,” Phys. Rev. D 9, 3292–3300 validity of the equivalence principle at the horizon only for (1974) certain accelerated observers, by showing that Hawking 19 S. W. Hawking, “Black Hole explosions?” Nature 248, 30– radiation and Unruh radiation have the same temperature 31 (1974) for them. Note that the calculation of a non-zero Hawking 20 S. W. Hawking, “Particle Creation by Black Holes,” Com- temperature31 for a freely falling observer passing the hori- mun. Math. Phys. 43, 199–220 (1975) zon is insufficient to prove observability of the radiation. 21 R. Dey, S. Liberati, Z. Mirzaiyan, and D. Pranzetti, “Black What has to be proven in addition is that its intensity (or hole quantum atmosphere for freely falling observers,” its gray-body factor) is non-zero. Phys. Lett. B 797, 134828 (1–7) (2019) 36 T. Opatrný and L Richterek, “Black hole heat engine,” Am. 22 M. K. Parikh and F. Wilczek, “Hawking radiation as tun- J. Phys. 80, 66–71 (2012) neling,” Phys. Rev. Lett. 85, 5042–5045 (2000) 37 P. C. W. Davies, S. A. Fulling, and W. G. Unruh, “Energy- 23 U. H. Gerlach, “The mechanism of blackbody radiation momentum tensor near an evaporating black hole,” Phys. from an incipient black hole,” Phys. Rev. D 14, 1479–1508 Rev. D 13, 2720–2723 (1976) (1976) 38 Interested readers can find them in an earlier version of 24 M. Visser, “Essential and inessential features of Hawking this article, deposited on arXiv: 2103.08340v2 [gr-qc]. radiation,” Int. J. Mod. Phys. D 12, 649–661 (2003) 39 J. Piesnack, “Freier Fall von Testteilchen im Gravitations- 25 W. G. Unruh, “Experimental Black-Hole Evaporation?” feld verdampfender Schwarzer Löcher,” Bachelor Thesis, Phys. Rev. Lett. 46, 1351–1353 (1981) Otto-von-Guericke Universität Magdeburg (2020) 26 40 J. Steinhauer, “Observation of quantum Hawking radiation The approximation rs ≪ r may be directly applied in (23) and its entanglement in an analogue black hole,” Nature to obtain the approximate result for r01. It also implies Physics 12, 959–965 (2016) rs,v ≪ c, which has to be used in the equation for r02, not 27 T. Mueller and F. Grave, “Catalogue of Spacetimes,” satisfying rs ≪ r02. arXiv: 0904.4184v3 [gr-qc] (2010) 41 A. J. S. Hamilton and J. P. Lisle, “The river model of black 28 K. Kassner, “Why ghosts don’t touch: a tale of two adven- holes,” Am. J. Phys. 76, 519–532 (2008) turers falling one after another into a black hole,” Eur. J. 42 This is true for classical objects. For quantum tunneling, Phys. 38, 015605 (2017) the speed of light is not an insurmountable barrier. 29 A coordinate is timelike (spacelike) if a hypersurface of 43 This holds for other metrics with a receding apparent hori- zon as well.7