Acceleration of particles by black holes as a result of deceleration: ultimate manifestation of kinematic nature of BSW eect

Oleg B. Zaslavskii
Kharkov V.N. Karazin National University, 4 Svoboda Square, Kharkov, 61077, Ukraine

arXiv:1202.0565v3 [gr-qc] 19 May 2012

The recently discovered so-called BSW eect consists in the unbound growth of the energy Ec.m. in the centre of mass frame of two colliding particles near the black hole horizon. We consider a new type of the corresponding scenario when one of two particles (critical) remains at rest near the horizon of the charged near-extremal black hole due to balance between the attractive and repulsion forces. The other one hits it with a speed close to that of light. This scenario shows in a most pronounced way the kinematic nature of the BSW eect. In the extremal limit, one would gain formally innite Ec.m. but this does not happen since it would have require the critical massive particle to remain at rest on the null horizon surface that is impossible. We also discuss the BSW eect in the metric of the extremal Reissner-Nordstr om black hole when the critical particle remains at rest near the horizon.
PACS numbers: 04.70.Bw, 97.60.Lf

I.

INTRODUCTION

The recent nding of the eect of the unbound growth of the energy Ec.m. in the centre of mass frame due to collisions of particles near the black hole horizon [1] (BSW eect) attracts now much attention. Both manifestations of this eect in dierent situation are being studied in detail and also the very nature of the eect itself is under investigation. It was observed in [2] that the underlying physical reason of the BSW eect can be explained in kinematic terms. Namely, it turns out that, roughly speaking, a rapid particle collides with a slow one near the horizon, this leads to the growth of the relative velocity and, as a

Electronic address: [email protected]

2 result, to the unbound growth of the corresponding Lorentz gamma - factor, so the energy Ec.m. becomes unbound near the horizon. This general circumstance was also conrmed in thorough analysis of the BSW eect in the Kerr metric [3]. Nonetheless, some doubts remain concerning the possibility to give an alternative explanation. If something is being accelerated to unbound energies, one is tempted to ask, what source does this, and what is the physical underlying reason of such an eect. The aim of the present work is to reveal the kinematic nature of the BSW eect in the most pronounced way. To this end, we consider the situation when one of two colliding particles is motionless while the other one moves (as usual) with a nite energy in the frame of a distant observer. In a sense, this is the ultimate and clear manifestation of the kinematic nature of the eect under discussion that does not require to search for further hidden dynamic factors. The model which we discuss shows the key issue as clear as possible: the role of gravitation in producing the BSW eect of the unbound growth of Ec.m. (acceleration of particles) consists not in acceleration but in deceleration of one of two particles (in the sense that its velocity is reduced to zero)! To achieve our goal, we consider the spherically symmetric metric of a charged black hole that admits the equilibrium of a particle that remains motionless. In other words, we want to balance the gravitation force by electrical repulsion. Apart from this, it is important that such a point be located in the vicinity of the horizon. For deniteness, we consider the innermost stable equilibrium point which is the counterpart of the innermost stable circular orbit for the Kerr metric [4]. Such orbits were discussed recently due to their potential astrophysical signicance [5], [6]. (See their generalization to dirty rotating black holes [7].) There exists also their analog in the magnetic eld where the BSW eect was studied recently in [8]. The simplest choice would seem to be the Reissner-Nordtr om (RN) black hole but for this metric the orbit with the required properties exists for indierent equilibrium only (see Sec. V below). Therefore, for the analog of inner stable orbits we take the charged black hole with nonzero cosmological constant . It turns out that it is required that < 0, so we mainly deal with the Reissner-Nordtr om - anti-de Sitter one (RN-AdS) which is sucient for our purposes. It is also worth noting that interest to black holes with the cosmological constant < 0 revived in recent years due to AdS/CFT correspondence [9]. In addition, we consider also another type of orbit a particle in the state of indierent equilibrium

3 in the metric of the extremal ReissnerNordstr om black hole.

II.

EQUATIONS OF MOTION

Consider the space-time describing a charged black hole with the cosmological constant. Its metric can be written as ds2 = f dt2 + dr 2 + r 2 d2 , f (1) (2)

2m Q2 r 2 + 2 . f =1 r r 3 horizon lies at r = r+ where f (r+ ) = 0. The electric potential = Q +C r

Throughout the Letter we assume that the fundamental constants G = c =h = 1. The

(3)

where the C is the constant of integration. It is assumed that we work in the gauge where the only nonvanishing component A0 = . For the asymptotically at case, say, for the Reissner-Nordstr om or Kerr-Newman metric, it is usually chosen C = 0 to have = 0 at innity (see, e.g., eq. 3.63 of [10]). In the absence of asymptotic atness, its choice becomes conditional. It is worth stressing that physically relevant quantities contain not the potential itself but the dierence with respect to some reference point (innity or horizon). For example, in black hole thermodynamics, the potential enters the action in the form
(r )(r+ ) f

that is nonsingular at the horizon (see eq. 4.15 of [11]). In equations of motion

(see below) only the combination E q appears where q is the particles charge. If we change the potential according to + C , the corresponding shift in the energy E E + qC . For convenience, we choose C = 0 in (3). We restrict ourselves by radial motion since this case is the most interesting in the context under discussion. As is known, under the presence of the electromagnetic eld, dynamics of the system is described by the generalized momentum P related to the kinematic one p = mu by the relation p = P qA where u =
dx d

is the four-velocity of a test

massive particle, is the proper time, A is the vector potential. Due to staticity, the energy E = P0 of a particle moving in this metric is conserved, P0 is the time component of the generalized momentum P . Then, using also the relation u0 = g 00 u0 , we obtain (dot

4 denotes the derivative with respect to the proper time ) = u0 = X , t mf X = E q. > 0, so that E q > 0. We assume that t m2 r 2 = Vef f = X 2 m2 f . Now, we are interested in equilibrium solutions r = r0 = const, Vef f (r0 ) = 0. (7) (6) (4) (5)

Additionally, we require that they possess the following properties: (i) r0 is a perpetual turning point, (ii) it lies near the horizon, r+ r0 . Condition (i) means that, in addition to (7), equation
Vef f ( r0 ) = 0

(8)

should hold. Eqs. (7), (8) ensure that not only r but also all higher derivatives vanish. It follows from (6), (7) that for a particle with r = 0, X ( r0 ) = m f (r0 ). (9)

It is instructive to elucidate for which types of black holes equations (7) and (8) are selfconsistent near the horizon, so that equilibrium points exist there in agreement with requirement (ii). Physical motivation for considering this requirement comes from our main goal investigation of the BSW eect since this eect occurs just in the vicinity of the horizon. If we take the derivative of the eective potential Vef f in eq. (6) and take into account also the relation (7), we obtain 1 ( r0 ) = m Vef 2 f f ( r0 ) qQ m2 f ( r0 ) . 2 r0 2 (10)

1 2 Vef f (r0 ) m where we used the fact for the metric (1) = 2 f (r+ ). Thus if = 0,

Let us consider the limit r0 r+ , so f (r0 ) 0. Then, it follows from (10) that

eq. (8) cannot be satised in the horizon limit. Therefore, for nonextremal black holes the equilibrium points cannot exist near the horizon (although they can exist elsewhere at a nite distance from the horizon). This generalizes previous observations [3], [7] made for rotating black holes. However, if 0, the equilibrium points close to the horizon do exist as will be shown below.

5
III. PROPERTIES OF EQUILIBRIUM POINT

For the Kerr metric [4] and, in general, for axially-symmetric rotating black holes [7], there are so-called innermost stable orbits (ISCO) which correspond to the threshold of stability. We consider now their analogs in our case, so we must add to (7) and (10), also equation Vef f (r0 ) = 0. For brevity, we will call this an innermost stable equilibrium point (ISEP). We are interested in the near-horizon region where we can expand f in the Taylor series with respect to x = r0 r+ : f = 2x + Dx2 + Cx3 .... (12)

(11)

From now on, we assume that is a small parameter, so a black hole is a near-extremal. Then, this leads to an interplay between two small quantities and x. We assume the condition Dx which one can check a posteriori that for the solutions obtained. Then, the procedure for the description of the equilibrium points is mathematically similar to that for the description of circular orbits in the background of rotating black holes [7]. In both cases, we are interested in solutions for which r = 0 and which are on the threshold of stability. Therefore, I omit technical details (which are connected with simple but rather cumbersome calculations) and give the main results of eqs. (7), (8), (11). It turns out that x3 H 3 2 , where H3 = and the constants in (12)
3 3 r+ 2 4()(1 2r+ )

(13)

(14)

(15)

D= C=

1 2, 2 r+ 8 2 . + 3 r+ 3 r+

(16) (17)

6 As in the extremal limit 0 we must have f > 0 in the vicninity of the horizon from the outside, the coecient D > 0. Then, in combination with H > 0, this entails that < 0. By substitution of (14) into (12) we obtain f 3 4
1/3 2 1/6 2/3 ()1/3 (1 2r+ ) . dr f

(18)

Although r0 r+ , the proper distance l =

between the particle and the horizon

does not vanish and, moreover, it grows unbound when 0: x0 1 1 1 ln . l ln 3 D This is in full analogy with the rotating case [4], [7]. (19)

IV.

COLLISIONS WITH UNBOUND ENERGY

Now, we consider the collision of two particles. To avoid unnecessary complications due to possible Coulomb repulsion between particles having the charge of the same sign, we can assume that the particle falling towards a black hole is neutral. Assuming, for simplicity, that both particles have the same mass m, the energy is given by the formula [12]
2 Ec.m. X1 X2 Z 1 Z 2 = 1 + 2m2 f m2

(20)

where Zi = Xi2 m2 f , i = 1, 2. (21)

Let, for deniteness, a motionless particle have i = 1. The unbound growth of the energy Ec.m. occurs if one particle has on the horizon (X1 )+ = 0 (we call it critical) whereas for the other one (X2 )+ = 0 (we call it usual) - see [12] for details. In our case, Z1 = 0 according to eq. (9), so the formula simplies:
2 Ec.m. X2 . = 1 + 2m2 m f

(22)

Particle 1 has X1 has the form

f , so it is near-critical near the horizon f 0 which is just the

case we are dealing with. Then, we obtain for the collision near the horizon that the energy 2mX2 A 3
1

Ec.m.

(23)

7 where it follows from (18) that A= 4 3

(24)

The dependence Ec.m. 1/3 is similar to that for rotating black holes (cf. eq. 5.1 of [5] and eq. 89 of [7]).

V.

DEGENERATE CASE: BSW EFFECT FOR A PARTICLE AT REST IN THE EXTREMAL REISSNER-NORDSTROM METRIC

In investigation of the BSW eect, one is led to deal with dierent limiting transitions: r0 r+ , (X1 )+ 0 that requires certain care. It was demonstrated earlier (see eqs. 11 and 15 of [13] and eqs. 8, 10 of [12]) that these limits do commute and give Ec.m. in both cases for collisions of particles moving towards the horizon of an extremal black hole. In our case there are two distinctions from the aforementioned situation: (i) a black hole is nonextremal with small but nonzero , (ii) the point of collision cannot be considered as an independent parameter since it coincides with the equilibrium point of particle 1 whose location r0 is controlled by according to (14), (15). Therefore, one cannot make permutation between r0 r+ and 0 that represents now a self-consistent indivisible procedure. Meanwhile, one may ask what happens to the points of equilibrium if the limit = 0 is taken from the very beginning that corresponds to an extremal black hole. Formally, r0 r+ in this limit. However, on the horizon which is light-like surface, the time-like trajectories cannot exist, so the solution r0 = r+ for it is fake (see the detailed analysis in Sec. III C of [5]). The real trajectory is not strictly static and asymptotically approaches the horizon [14], [15], [13], [5]. It would seem that in such circumstances the questions about ISEP in the near-horizon region do not make sense at all. Nonetheless, there is an exceptional case when ISEP degenerates into points of indierent equilibrium. This happens just for extremal Reinssner-Nordstrom black holes. Then, in eq. (2), Q = M = r+ , = 0. By direct check, it is easy to see that eqs. (7), (8) lead to the consequences that E = m = q . But for these values of particles parameters, the eective potential Vef f = 0 for any r . Actually, this means that a particle can be at rest at any position r0 due to balance between gravitational attraction and electric repulsion, so equilibrium is indierent. (More on the properties of equilibrium in the Reissner-Nordstr om space-time can be found in [16]).

8 Now, eq. (22) gives us an exact expression

(25)

r0 1/2 ) . In other words, for an arbitrary r0 . Thus we can see that for r0 r+ , Ec.m. (1 r +

we place particle 1 in any point at rest and inject another particle from the outside (say, from innity). When the location of particle 1 approaches the horizon, Ec.m. , so we again obtain the BSW eect.

VI.

KINEMATIC CENSORSHIP

There is one more question concerning the possibility to take the limit 0. It follows from our formulas for the BSW eect at ISEPs that for any small but nonzero the energy Ec.m. is large but nite. Can one simply take the value = 0 from the very beginning and gain an innite energy? In any real physical event the actual energy that can be released must be nite. With respect to collisions of particles, it can be named kinematic censorship. Therefore, the energy Ec.m. can be as large as one likes but it cannot be literally innite. To understand, how this kinematic censorship is realized in our case, one should take into account explanations from the previous Section. We would like to stress it once again that the orbit r0 = r+ to which formally tends the ISEP is not suitable since a trajectory of a massive particle cannot lie on the light-like horizon surface. In the example with indierent equilibrium in the extremal Reinssner-Nordstrom background, the situation is even more clear: we can place particle 1 at any position r0 > r+ which is as close to r0 as one likes but it cannot coincide with r0 nonetheless.

VII.

ROLE OF GRAVITY IN BSW EFFECT

The results obtained concern charged black holes and represent a counterpart of those for the circular orbits in the background of rotating black holes [5], [7]. Meanwhile, the basic point does not have an analogue in the case of the Kerr metric. Indeed, the circular orbits in the near-horizon region of the Kerr black hole lie in the ergosphere, so equilibrium is not possible there. The perpetual turning point in [4] is related to circular orbit, so a particle necessarily has the nonzero angular velocity with respect to a distant observer. Meanwhile,

9 in the present work a particle located near the horizon has both zero velocity and zero angular momentum. The BSW eect with participation of such a particle sharpens and reveals its kinematic nature. Naively, one could think that the particle is accelerated during the infall into the black hole and the problem seemed to be to explain why and how this happens. However, we see that the real picture is quite dierent. One of two colliding particles is kept xed to remain at rest near the horizon while the other one starts its motion from the outside. As the all process is essentially nonlocal, one cannot attribute the growth of relative energy to the action of force that exerts on some united object. Moreover, in the collision under discussion particle 2 that comes from the outside is typical (usual) in that it has an arbitrary nite individual energy and zero charge, so the relation X+ = 0 typical of the critical particle is not satised for it. As a result, its velocity in the static frame approaches that of light [2], and this becomes true for any such a particle. As a result, the relative velocity of two usual particles remains nite and the BSW eect is impossible (see [2] for details in general and analysis for the Kerr metric in [3]). To gain the BSW eect, one should select such a particle that approaches the velocity less than that of light near the horizon. In the present case, it is the particle with literally zero velocity that remains zero all the time before collision. In other words, the role of gravitation consists here not in acceleration of particle 1 but in keeping it in rest due to balancing electric repulsion, so the eventual outcome of the unbound energy Ec.m. is obtained in a sense as a consequence of deceleration or, more precisely, arrest of one of particles! Moreover, in the case of the Reissner-Nordstr om-anti-de Sitter black hole the combined action of gravitation, electricity and the cosmological term is to arrest particle 1 in such a way that it remains there on the threshold of stability. (One can take a particle
with Vef f > 0 instead of (11) to have it strictly stable with respect to radial displacement.)

Another example with the same qualitative features is indierent equilibrium of a charged particle in the metric of the extremal Reinssner-Nordstr om black hole. It is this particle that remains at rest is now a critical one. It is in accord with the general principle that for the BSW eect to occur, one of particles should be critical and the other one should be usual [2], [13]. Thus the case considered above shows in a most pronounced way that the kinematic nature of the BSW enables one to obtain unbound energies because of dierent action of gravity on essentially dierent (in a kinematic sense) kinds of particles.

10 It is worth noting that in [3] the collision between the infalling particle and the particle at rest was discussed (after eq. 34). Meanwhile, there is a qualitative dierence between both situations. In [3], the particle was kept at rest in the Schwazrschild background that was possible by hand only and required an almost innite force near the horizon. Meanwhile, in the example considered in our work, one can check that the value of acceleration a if nite. Indeed, one can calculate the scalar a2 = a a where a = u; u , semicolon denotes the covariant derivative. It is easy to nd that for the metric (1), a2 =
(f )2 4f
2 12r+ 2 r+

is

nite due to the near-extremal character of a black hole. Thus gravity in combination with electric repulsion and the cosmological constant, ensures the BSW eect on a motionless particle in a self-consistent way. In doing so, the particle waiting at rest plays the role of a target which is hit by the infalling particle.

VIII.

CONCLUSION AND PERSPECTIVES

The type of the BSW eect discussed in the present Letter is somewhat dierent from the original one considered in the pioneering paper [1]. It makes the kinematic nature of the BSW eect especially pronounced. Inclusion of backreaction and radiation into general scheme can change the details of the eect signicantly [17], [14] but one can expect that the main qualitative features of the BSW eect still persist just due to its kinematic nature. Moreover, the fact that in the scenario discussed in the present work, a near-critical or critical particle remains at rest, suggests that for it gravitational radiation (mentioned in aforementioned papers as a factor acting against the BSW eect) is absent now at all. The role of backreaction on the metric is more subtle but, anyway, the scenario of collisions considered in the present paper simplies the picture and can be useful for further analysis. More detailed study of the generic BSW eect with account for all these factors is a nontrivial problem that needs separate treatment.

[1] M. Banados, J. Silk, S.M. West, Phys. Rev. Lett. 103 (2009) 111102. [2] O. B. Zaslavskii, Phys. Rev. D 84 (2011) 024007. [3] A. A. Grib, Yu. V. Pavlov and O. F. Piattella, Int. J.Mod. Phys. A 26 (2011) 3856. [4] J. Bardeen, W. H. Press, and S. A. Teukolsky, Astrophys. J. 178 (1972) 347.

11
[5] T. Harada and M. Kimura, Phys. Rev. D 83 (2011) 024002. [6] T. Harada and M. Kimura, Phys.Rev.D 84 (2011) 124032. [7] O. B. Zaslavskii, arXiv:1201.5351. [8] V. P. Frolov, Phys. Rev. D 85, 024020 (2012). [9] J. M. Maldacena, Adv. Theor. Math. Phys. 2 (1998) 231. [10] V.P. Frolov, I.D. Novikov, Black Hole Physics: Basic Concepts and New Developments, Kluwer Academic, Amsterdam, 1998. [11] H.W. Braden, J.D. Brown, B.F. Whiting, J.W. York, Phys. Rev. D 42 (1990) 3376. [12] O. B. Zaslavskii, Pisma v Zh. Eksp. Teor. Fiz. 92, 635 (2010); JETP Lett. 92 (2010) 571. [13] O. B. Zaslavskii, Phys. Rev. D 82 (2010) 083004. [14] T. Jacobson and T. P. Sotiriou, Phys. Rev. Lett. 104 (2010) 021101. [15] A. A. Grib and Y.V. Pavlov, Pisma v Zh. Eksp. Teor. Fiz. 92 (2010) 147 [JETP Lett. 92 (2010) 147]. [16] W. Bonnor, Class. Quant. Grav. 10 (1993) 2077. [17] E. Berti, V. Cardoso, L. Gualtieri, F. Pretorius, and U.Sperhake, Phys. Rev. Lett. 103 (2009) 239001.