Arxiv:2108.12408V1 [Gr-Qc] 27 Aug 2021

Energy Extraction from Static Black Hole Binaries

Lucas Timotheo Sanches1, ∗ and Maur´ıcio Richartz2, † 1Centro de CiênciasNaturais e Humanas, Universidade Federal do ABC (UFABC), 09210-170 Santo André,SãoPaulo, Brazil 2Centro de Matemática, Computa¸cãoe Cogni¸cão, Universidade Federal do ABC (UFABC), 09210-170 Santo André,SãoPaulo, Brazil We define and sketch the generalized ergosphere of the Majumdar-Papapetrou spacetime. In particular, we demonstrate the existence of closed orbits of negative energy that live outside the event horizon of such a spacetime. Relying on the Penrose process mechanism, we use these orbits to illustrate the possibility of energy extraction from a binary black hole by particle scattering. We also analyze the efficiency of the process, and construct explicit examples that optimize the extraction of energy.

I. INTRODUCTION interaction allows for highly efficient energy extraction schemes, such as the one introduced in Ref. [8] to model The best mathematical description of rotating black the emission of ultra-high energy cosmic rays from su- holes is given by the theory of General Relativity and, in permassive black holes. Furthermore, recent numerical particular, by the Kerr metric [1–4]. Unlike static black simulations of plasma and jets around Kerr black holes holes, a Kerr black hole is characterized by the existence indicate the important role that negative energy parti- of a very peculiar region around its event horizon, known cles and the Penrose process play on the total energy as the ergosphere. Particles that reach the ergosphere can flux coming from the black hole’s jets [9]. still avoid the event horizon and, hence, are not doomed The aim of our work is to investigate energy extrac- to end at the spacetime singularity. Nevertheless, any ob- tion via the Penrose process in a black hole binary sys- server lying inside the ergosphere is unavoidably dragged tem. Any study involving astrophysical binary systems along by the rotational motion of the black hole. must acknowledge the fact that there is no known ex- Particles inside the ergosphere may have negative en- act analytical solution of Einstein’s equations describing ergies (according to a static observer at infinity). Relying such a system. Even if numerical or approximate analyt- on this property, Penrose and Floyd devised a mechanism ical solutions (obtained, for example, through Numerical (known as the Penrose process) to extract energy from Relativity or Post Newtonian methods) are employed, a rotating black hole [5]. The idea consists in sending a the fact that the associated metric is non-stationary lim- particle from infinity towards the black hole and assumes its the applicability of the standard concept of an ergo- that, once inside the ergosphere, it decays into two other sphere. Recognizing these difficulties, in this paper we particles. If one of the fragments is counter-rotating with investigate the possibility of energy extraction from the the black hole and has negative energy, it will be captured Majumdar-Papapetrou (MP) metric [10, 11], which is an by the black hole, meaning that the other fragment will exact solution of Einstein’s equations that describes a escape to infinity. Due to the conservation of the four- static binary of extremally charged black holes. momentum, the escaping fragment will have more energy and more angular momentum than the incident particle. Despite its simplicity, the MP solution has recently Rotational energy and angular momentum are, thus, ef- been used as a surrogate model for black hole binaries fectively extracted from the black hole. in order to understand how single black hole phenomena To illustrate the importance of the Penrose mechanism, transpose to a binary system. For instance, Ref. [12] has we highlight a few of the many recent research endeav- employed the MP metric to understand the connection ours that attempt to establish a relation with observ- between quasinormal modes and light rings in the context able astrophysical phenomena. The collisional version of black hole binaries. Refs. [13, 14], on the other hand, of the process, for instance, considers multiple particles have computed the shadows cast by a MP binary to bet- that collide and scatter in the ergoregion, allowing ar- ter understand chaotic scattering in a binary black hole bitrarily high center-of-mass energies. This process can system. The resulting shadow shares many similarities arXiv:2108.12408v1 [gr-qc] 27 Aug 2021 potentially act as a mechanism to eliminate dark matter and qualitative features with the shadows computed in particles near a supermassive black hole [6]. The mag- Ref. [15] using a numerically simulated binary black hole netic Penrose process [7], on the other hand, considers background. Ref. [16] has also applied the MP metric charged particles and black holes surrounded by mag- to analyze particle scattering around a black hole binary netic fields (originated, for instance, by plasma accre- and asserted its effectiveness in approximating a head on tion disks around the black hole). The electromagnetic collision in the limit of large separations and small approach speeds. We follow here an analog approach in order to gain physical intuition and qualitative insights about energy extraction from black hole binaries by the ∗ [email protected] Penrose mechanism. In particular, we extend the con- † [email protected] cept of a particle dependent generalized ergosphere [17], 2 which enables the extraction of electromagnetic energy forces and, hence, do not follow geodesics. If the electro- from Reissner-Nördstrom(RN) black holes, to the MP magnetic interaction is attractive, negative energy tra- solution and study how the energy extraction efficiency jectories and energy extraction are, in principle, possi- is affected by the presence of a companion black hole. ble. For a single charged black hole described by the Our paper is organized as follows. In Sec. II A we in- RN metric, the fact that the Penrose process is viable troduce the MP spacetime and derive the equations of was shown in Ref. [20]. Building on the notion of a gen- motion for a charged massive particle. In Sec. II B the eralized ergosphere [17, 20], we shall study the motion notion of a generalized ergosphere is laid down and ex- of charged particles around the MP spacetime to investi- plored, while the existence of negative energy orbits is gate the possibility of negative energy motion and energy studied in Sec. II C. In Sec. III we show how a Penrose extraction. process can be set up in the MP spacetime and we give explicit examples of such a process. In Sec. IV we analyze how much energy can be extracted and how it depends A. Motion of charged particles in MP on the several parameters describing the system. The last section is dedicated to our final remarks. The motion of a massive charged test particle (with charge-to-mass ratio µ) in a spacetime with metric gµν , subject to the electromagnetic potential Aµ, is deter- II. THE MAJUMDAR-PAPAPETROU mined by the Lagrangian SPACETIME 1 µ ν α L = gµν x˙ x˙ − µAαx˙ , (2.4) The MP spacetime is a static electrovacuum solu- 2 tion of Einstein’s equations that represents a set of ex- where xµ = xµ(λ) denotes the position of the particle at tremal black holes whose mutual gravitational attrac- proper time λ and the dots represent derivatives with re- tion is cancelled by their mutual electromagnetic repul- spect to λ. Adopting Weyl’s cylindrical coordinates, and sion [10, 11, 18]. For two black holes of masses M1 and taking into account the explicit form of the MP metric, M2 and electric charges Q1 = M1 and Q2 = M2, in equi- the Lagrangian above can be recast as [21] librium and separated by a distance 2a along the z-axis, the MP line element, in Weyl’s cylindrical coordinates 1 t˙2 1 (t, ρ, φ, z), is given by [19] L = − + U 2 ρ˙2 + ρ2φ˙2 +z ˙2 − µ 1 − t.˙ 2 U 2 U (2.5) dt2 2 2 2 2 2 2 Since the Lagrangian does not depend explicitly on t ds = − 2 + U(ρ, z) dρ + ρ dφ + dz , (2.1) U(ρ, z) and φ, two constants of motion (as measured by freely- falling observers at infinity) can be readily identified. The where constant associated with the time symmetry is the energy M M divided by the mass m of the particle: U(ρ, z) = 1 + 1 + 2 . (2.2) pρ2 + (z + a)2 pρ2 + (z − a)2 ∂L t˙ 1 E = − = + µ 1 − . (2.6) ˙ 2 The electromagnetic potential Aµ associated with the ∂t U U MP solution is The constant associated with the angular symmetry is 1 the angular momentum per unit mass (with respect to A = 1 − δ , (2.3) µ U µt the z axis): ∂L where δµν is the Kronecker delta. We note that Weyl’s L = = U 2ρ2φ.˙ (2.7) coordinates describe only the exterior of the black holes. ∂φ˙ The event horizons and the black holes themselves are collapsed into the points ρ = 0, z = ±a. We denote the Plugging the constants of motion into the 4-velocity normalization condition, i.e.x ˙ µx˙ = −1, and solving for total mass of the binary system by MT = M1+M2 and its µ E, yields mass ratio by MR = M2/M1. Without loss of generality, we assume that 0 ≤ MR ≤ 1. s Being described by a static metric, the MP spacetime 1 L2 1 E = µ 1 − + + +ρ ˙2 +z ˙2, (2.8) does not possess an ergosphere (in the usual sense of a U ρ2U 4 U 2 spacetime region outside the black hole where the Killing µ vector (∂t) becomes spacelike). Consequently, the en- where the negative root has been ignored due to the fact ergy associated with geodesic motion in the MP space- that E must be positive for particles at infinity when µ = time is always positive, meaning that energy extraction 0 (see Ref. [17] for a detailed discussion about positive by free particles is impossible. Charged particles, how- and negative root states for E). We note that Eq. (2.8) ever, interact with charged black holes through Lorentz reduces to the expression found in Ref. [20] for a RN 3 black hole if the mass of one of the black holes is taken to the minimum possible energy is associated with particles be zero and an appropriate coordinate system, centered at rest. Lettingρ ˙ =z ˙ = 0 and L = 0 (i.e. φ˙ = 0), around the other black hole, is adopted. For convenience we conclude that the minimum energy (per unit mass) and later use, we rewrite Eq. (2.8) as allowed for a particle sitting at (t, ρ, φ, z) is

ρ˙2 +z ˙2 = E2 (ρ, z) − V (ρ, z), (2.9) 1 1 eff eff E = µ 1 − + . (2.15) min U U where This minimum energy will be negative if the following 1 Eeff (ρ, z) = E − µ 1 − , (2.10) two conditions are satisfied. Since U ≥ 1, µ must be U negative, meaning that the charge of the particle must be opposite to the charge of the black holes. Addition- and ally, the resting particle must be inside the spatial region L2 1 determined by Veff (ρ, z) = 2 4 + 2 . (2.11) ρ U U 1 M 1 + M + R > − R , (2.16) Note that the quantities defined above are subject to the pρ2 + (z + 1)2 pρ2 + (z − 1)2 µ following constraints: where ρ = ρ/a, z = z/a and µ = µMT /a are dimension- 2 Eeff (ρ, z) ≥ 0,Eeff (ρ, z) ≥ Veff (ρ, z). (2.12) less quantities. The parameter µ can be understood as a measure of the potential energy (per unit particle mass) The equations of motion are obtained directly from associated with the electromagnetic interaction between the Euler-Lagrange equations. After using Eqs. (2.6) and the particle and the binary. The inequality above deter- (2.7) to eliminate t˙ and φ˙, one is left with mines the generalized ergosphere of the MP spacetime: particles with opposite charge in relation to the black hole and located inside the region determined by Eq. (2.16) L2(U + ρ∂ U) 2ρ ˙z∂˙ U − (E2 +z ˙2 − ρ˙2)∂ U ρ z ρ will have negative energies if their velocities are suffi- ρ¨ − 3 5 + ρ U U ciently small. Particles outside, on the other hand, will µ E − µ have positive energies regardless of their velocities. − µ − 2E + ∂ U = 0, (2.13) U U ρ Note that the notion of a generalized ergosphere depends not only on the geometry of the spacetime, but also and on properties of the particle (through the charge-to-mass ratio µ), as in Refs. [17, 20]. In particular, the shape of L2∂ U 2ρ ˙z∂˙ U − (E2 − z˙2 +ρ ˙2)∂ U z ρ z the ergosphere depends only on the parameters M and z¨ − 2 5 + R ρ U U µ. In order to understand this dependence, we plot the µ E − µ generalized ergosphere for nine pairs M -µ. Each pair − µ − 2E + ∂ U = 0, (2.14) R U U z corresponds to a point in the parameter space of Fig. 1 and is labeled by a letter (A-I). The φ = 0 section of which reduce to the equations of motion found in Ref. [12] the associated ergospheres are shown in Fig. 2. Due to for neutral particles. Together with Eqs. (2.6) and (2.7), the axisymmetry of the problem, the ergospheres are the the two equations above fully determine the trajectory of solids of revolution obtained by the rotation of the regions a massive charged particle in the MP spacetime once ap- shown in Fig. 2 with respect to the z-axis. Note that the propriate initial conditions have been specified. In fact, generalized ergosphere can be either a single connected we first specify the energy E, the angular momentum L, region (for parameters inside the red section of Fig. 1) and the initial values for ρ, z, andz ˙. The initial value or the disjoint union of two connected regions1 (for pa- forρ ˙ is determined from Eq. (2.8). Using this informa- rameters in the green section of Fig. 1). The boundary tion, we can solve the system of equations (2.13)-(2.14) separating connected and disconnected ergospheres, rep- to obtain ρ(λ) and z(λ). The final step is the integration resented by the black line in Fig. 1, corresponds to the of Eqs. (2.6) and (2.7), subject to the initial data for t saturation of Eq (2.16). and φ, to find t(λ) and φ(λ). What information can we extract from Eq. (2.16) and Figs. 1 and 2? Naturally, the ergosphere becomes more evenly distributed around the black holes when the mass B. Generalized Ergosphere ratio MR approaches one. This is seen in the figures by following the points A → B → C, or D → E → We want to know if charged particles can have negative energies in the MP spacetime. From Eq. (2.8) it is evident that, for fixed µ, the energy is completely determined by the angular momentum L, and the values of ρ, z,ρ ˙ 1 In such a case we shall refer to each connected region as a single andz ˙ at any given instant of time. At a fixed position, ergosphere. 4

F, or G → H → I. Furthermore, if −2 < µ < −1, we see that an increase in the mass ratio may produce a single ergosphere from two disjoint ones, as in D → E. Similarly, for any mass ratio, an increase in the absolute value of µ can also induce the merger of the ergospheres, as in D → G, B → E and C → F. In fact, no matter what the mass ratio is, if −µ is sufficiently small, each black hole will be surrounded by its own ergosphere. In contrast, if −µ is sufficiently large, there will be a single ergosphere surrounding the binary black hole. Finally, we remark that, according to Eq. (2.16), the effects of the charge of the particle, of the total mass of the binary and of the separation parameter are all combined in µ. Therefore, as far as the visualization of the ergosphere is concerned, the effect of increasing |µ| is exactly the same as increasing MT or decreasing a.

FIG. 1. The parameter space MR-µ. The generalized ergosphere exists if and only if µ < 0. It consists of a single connected region for any point inside the red section. For points in the green section, on the other hand, the ergosphere is the disjoint union of two connected regions. The black curve marks the boundary between connected and disconnected ergospheres. The ergospheres associated with the black labeled 0 0.2 0.4 0.6 0.8 1 dots are plotted in Fig. 2.

FIG. 3. Energy levels inside the generalized ergosphere of a MP black hole with MR = 1/2 and µ = −5/4, corresponding to the point E in Figs. 1 and 2. The color bar represents Emin/µ. From left to right, the panels correspond respectively to µ = −0.01, µ = −5/12, and µ = −1.5. The horizontal and vertical axes in each panel are ρ and z, respectively. The red dots indicate the location of the black holes.

The minimum energy (per unit mass) at a given point of the spacetime, in contrast, depends on the explicit values of µ, MR and MT /a. To illustrate this dependence, we plot in Fig. 3 the energy levels inside the ergosphere determined by MR = 1/2 and µ = −5/4 (which corre- ponds to the point E in Figs. 1 and 2) for diﬀerent com- binations of µ and MT /a. The color bar in Fig. 3 indi- cates the value of Emin/µ, which, according to Eq. (2.15), varies from zero (at the boundary of the ergosphere) to one (at the black holes). As |µ| decreases and MT /a increases (while µ and MR are kept ﬁxed), the shape of the ergosphere remains unchanged, but the energy levels become more evenly distributed around the black holes.

C. Negative energy trajectories

FIG. 2. The φ = 0 section of the ergosphere for different Trajectories associated with negative energies are con- points (labeled A-I) in the parameter space of Fig. 1. In each fined within the ergosphere defined by Eq. (2.16). For a plot the horizontal and vertical axes are ρ and z, respectively. generic set of initial conditions, such paths will eventually The red dots indicate the location of the black holes. end up at either one of the black holes. This is not much different than the case of a single RN or a Kerr black 5 hole, in which particles with negative energies that start outside the event horizon necessarily enter the black hole and reach the spacetime singularity. Instead of focusing on this type of trajectory, in this section we concentrate on trajectories which have no analogs in standard (Kerr and RN) black hole spacetimes [22, 23]. We want to investigate whether or not closed orbits of negative energy that live outside the black holes are allowed2. We follow the procedure described in the last paragraph of section II A to solve the equations of motion. To simplify our work, we restrict our attention to two classes of planar motion and show that, by fine tuning the initial data, one FIG. 4. Example of closed trajectories of negative energy is able to find closed orbits of negative energies in the MP (black curves) in the plane containing the black holes. The spacetime. These closed orbits of negative energies are blue region is the generalized ergosphere of the spacetime, unstable, in the sense that generic perturbations of the while the red dots indicate the location of the black holes. initial conditions result in trajectories that end at one of The horizontal and vertical axes in each panel are ρ and z, the black holes. respectively. Left planel: trajectory corresponding to L = 0, E = −2/10, a = M1 = M2 = 1, µ = −5, ρ(0) = 0, z(0) = The first class of orbits assumes zero angular momen- 5.314064237978... (fine tuned), andz ˙(0) = 0. The associated tum (L = 0) so that the trajectories are restricted to a period of revolution is λ ≈ 22.64. Right planel: trajectory plane that contains both black holes (here we choose the corresponding to L = 0, E = −2/100, a = 1, M1 = 2M2 = 2, φ = 0 plane without loss of generality). For a given µ = −5, ρ(0) = 0, z(0) = 12.832155724084... (fine tuned), and set of parameters a, M1, M2, E, and µ, after fixing z˙(0) = 0. The associated period of revolution is λ ≈ 43.28. ρ(0) = 0 andz ˙(0) = 0, we fine tune the initial position z(0). The initial valueρ ˙(0) is determined from Eq. (2.8). To illustrate, we show in Fig. 4 two exam- andρ ˙(0) produce infinitesimal variations on the original ples of eight-shaped orbits obtained with this scheme. trajectory (if z(0) andz ˙(0) are kept fixed). The left panel of Fig. 4 represents an orbit with energy To analyze these orbits we resort to Eqs. (2.9)-(2.11). 2 (per unit mass) E = −2/10 and charge-to-mass ratio Given a, E and L, the critical points ρ where Eeff (ρ, 0) = µ = −5 for an equal mass MP binary (M1 = M2 = 1) Veff (ρ, 0) represent circular closed orbits. Typically, how- with separation parameter a = 1. The trajectory starts ever, the particle will be confined inside a compact sec- on the z-axis, at z(0) = 5.314064237978... (fine tuned). tion of the z = 0 plane, in a characteristic “zoom-whirl” One complete revolution of the trajectory, with corre- orbit [12, 27], moving between a minimum radius ρmin sponding period of λ ≈ 22.64, is shown. The right and a maximum radius ρmax. In Fig. 5, we show the panel of Fig. 4, on the other hand, exhibits an orbit of effective energy and the effective potential for a MP E = −2/10 and µ = −5 for a MP binary with mass spacetime with a = 1 and M1 = M2 = 1, when the ratio MR = 1/2 (M1 = 2,M2 = 1) and separation pa- particle’s charge-to-mass ratio, angular momentum (per rameter a = 1. The trajectory starts on the z-axis, at unit mass), and energy (per unit mass) are, respectively, z(0) = 12.832155724084... (fine tuned). One complete µ = −5, L = 12.85869, and E = −5/100. The par- revolution of the trajectory, with correponding period of ticle is constrained to move between ρmin ≈ 0.543 and approximately λ ≈ 43.28, is shown. ρmax ≈ 3.857. The corresponding trajectory, assuming The second class of orbits is only possible for equal the starting point to be ρ(0) = 2 and φ(0) = 0, is also mass binaries and comprises trajectories that are con- shown in Fig. 5. fined to the z = 0 plane, which is the plane equidistant to the black holes. Particles with negative energy whose motion is constrained to this plane will evolve in a per- III. PENROSE PROCESS petual oscillatory motion, either in straight lines (when L = 0) or in more complicated precessing orbits (when Following Penrose’s original proposal [5] and its exten- L 6= 0). Even though these trajectories are unstable to sion to RN black holes [17, 20], we now investigate the generic perturbations, on the plane itself they are sta- possibility of energy extraction from a MP binary black ble. In other words, infinitesimal perturbations of ρ(0) hole. The mechanism we shall explore is analogous to the most standard one: a negatively charged particle is sent towards the binary black hole and, once inside the generalized ergosphere, breaks up into two fragments, one of 2 We note, however, that if the Kerr black hole is immersed in an which escapes to infinity with more energy than the origi- external magnetic field, closed orbits of negative energy are allowed outside the event horizon [24, 25]. We also note that stable nal particle. We assume that the incident particle follows circular orbits of negative energy exist around a Kerr naked sin- a trajectory T(0), which starts outside the ergosphere and gularity (and that the Penrose process can effectively take place ends inside it, at the break-up point (ρ∗, φ∗, z∗). From using such orbits) [26]. the break-up point, two other trajectories, labeled T(1) 6

12 We remark that all derivatives in (3.4) are evaluated at the break-up point. We also note that the break-up of 10 the incident particle naturally imposes restrictions on the

8 masses of its fragments. By squaring Eq. (3.2) and using the fact that the four-momentum is future-pointing and 6 timelike, we obtain the inequality [28]

4 2 2 2 m(1) + m(2) < m(0). (3.6) 2 0 1 2 3 4 5

FIG. 5. Left panel: eﬀective energy (black curve) and effective potential (purple dashed curve) for L = 12.85869, E = −5/100, a = 1, M1 = M2 = 1, µ = −5. Right panel: an example of a closed trajectory of negative energy in the z = 0 symmetry plane. The blue region is the generalized ergosphere of the spacetime. The trajectory (black curve) is gener- ated by setting the parameters to L = 12.85869, E = −5/100, a = 1, M1 = M2 = 1, µ = −5, ρ(0) = 2, and φ(0) = 0. The associated period of revolution is λ ≈ 35.

FIG. 6. Left panel: Penrose process in the ρ-z plane of a MP spacetime with M = 2, M = 1, and a = 1. Right and T(2), start. Each trajectory T(i) is a timelike path 1 2 xµ (λ) parametrized by its proper time λ. To ﬁx nota- panel: Penrose process in the z = 0 plane of a MP spacetime (i) with M = M = 1 and a = 1. In each plot, the incoming tion, let m , µ , E , L , and P µ denote, respec- 1 2 (i) (i) (i) (i) (i) trajectory T(0) (black curve) splits at the black point into the tively, the mass, the charge-to-mass ratio, the energy per negative energy orbit T(1) (red curve) and the trajectory of unit mass, the angular momentum per unit mass (with the escaping fragment T(2) (purple curve). The parameters respect to the z-axis), and the 4-momentum of the par- that generate these trajectories are shown in Tables I and II. ticle i on trajectory T(i). We assume that fragment 1 The blue region is the generalized ergosphere (for particle 1) and the red dots indicate the location of the black holes. remains inside the ergosphere (meaning that E(1) < 0), while fragment 2 escapes back to inﬁnity. The quantities that characterize each particle are related by conservation equations. Charge conservation, i m(i) µ(i) E(i) L(i) ρ˙(i) z˙(i) for instance, yields 0 1.00000 -0.08345 1.00000 0 0.60000 0.09699

µ(0)m(0) = µ(1)m(1) + µ(2)m(2). (3.1) 1 0.70000 -5.00000 -0.02000 0 0.41207 0.00000 2 0.23248 14.69629 4.36171 0 1.34012 0.41719 The conservation of the four-momentum applied at the break-up point, on the other hand, reads TABLE I. Parameters that generate the trajectories T(0), T(1), µ µ µ and T(2) of the Penrose process shown in the left panel of P(0) = P(1) + P(2). (3.2) Fig. 6. The derivativesρ ˙(i) andz ˙(i) are evaluated at the break- up point. Each component of the vector equation (3.2) above corresponds to a diﬀerent conservation equation with straight- forward physical interpretation. The zero-component is just the conservation of total energy, i.e. i m(i) µ(i) E(i) L(i) ρ˙(i) z˙(i) 0 1.00000 -0.27698 1.00000 0.00000 1.00000 0 E(0)m(0) = E(1)m(1) + E(2)m(2). (3.3) 1 0.10000 -5.00000 -0.05000 12.85870 1.36059 0 The spatial components are the conservations of linear 2 0.33342 0.66890 3.01423 -3.85662 2.59116 0 momenta, i.e.

( TABLE II. Parameters that generate the trajectories T(0), m ρ˙ = m ρ˙ + m ρ˙ (0) (0) (1) (1) (2) (2) , (3.4) T(1), and T(2) of the Penrose process shown in the right panel m(0)z˙(0) = m(1)z˙(1) + m(2)z˙(2) of Fig. 6. The derivativesρ ˙(i) andz ˙(i) are evaluated at the break-up point. and the conservation of angular momentum, i.e. From Eq. (3.3), the energy carried away by the escap- L(0)m(0) = L(1)m(1) + L(2)m(2). (3.5) ing fragment is E(2)m(2) = E(0)m(0) − E(1)m(1). Since 7 we have assumed that E(1) < 0, the escaping particle will 1. Maximum efficiency carry away more energy than the incident particle had. If the negative energy fragment collapses to one of the black First of all, the minimum energy for an incident par- holes, it will directly decrease the energy associated with ticle coming from infinity, according to Eq. (2.8), is the black hole, as in Penrose’s original proposal. If, on E(0) = 1 and corresponds to the particle having zero the other hand, the negative energy fragment remains in kinetic energy at infinity. For that reason, we shall as- a closed orbit (such as the ones described in Sec. II C), sume from now on that L(0) = 0 and E(0) = 1. Secondly, where does the extra energy for the escaping fragment according to Eq. (2.15) and the discussion in the first come from? Since the final state (binary black hole and paragraph of Sec. II B, the energy per unit mass of par- bound particle fragment) is less energetic than the binary ticle 1 is most negative if the particle is initially at rest. itself, we conclude that the extracted energy is associated Therefore, at the break-up point we set with the binding energy of the binary. In other words, the presence of the bound particle fragment reduces the ˙ ρ˙(1) =z ˙(1) = φ(1) = 0, (4.2) energy required to form a MP binary and the excess energy is transferred to the escaping fragment. meaning that the associated angular momentum per unit In Fig. 6 we exhibit examples of Penrose processes mass and energy per unit mass are, respectively, L(1) = 0 that include some of the negative energy orbits shown and in Sec. II C. The left panel exhibits a Penrose process 1 1 that takes place in the ρ-z plane of a MP spacetime with E = Emin = µ 1 − + , (4.3) (1) (1) (1) U U M1 = 2, M2 = 1, and a = 1. The incoming parti- ∗ ∗ cle breaks-up at ρ = 0 and z = 5.314064237978... (fine tuned). The right panel, on the other hand, exhibits a where we have defined U∗ = U(ρ∗, z∗) for simplicity. Penrose process that occurs in the z = 0 plane of a MP Let us now study the allowed values for m(1). The conservation of linear momenta, Eq. (3.4), yields the relation spacetime with M1 = M2 = 1 and a = 1. The incoming particle breaks-up at ρ = 2 and φ = 0. The parameters 2 2 ! 2 2 ! that generate these examples and satisfy Eqs. (3.1)-(3.6) ρ˙(0) +z ˙(0) ρ˙(1) +z ˙(1) m2 = m2 + m2 , (4.4) are given in Tables I and II. (2) (0) 2 2 (1) 2 2 ρ˙(2) +z ˙(2) ρ˙(2) +z ˙(2)

where all derivatives are evaluated at the break-up point. 2 2 After replacingρ ˙(i) +z ˙(i) using Eqs. (2.9)-(2.11), and em- IV. ENERGY EXTRACTION EFFICIENCY ploying the conservation equations (3.1), (3.3) and (3.5), the expression above reduces to The eﬃciency η of the Penrose process can be deﬁned q m = m2 − 2m m α + m2 , (4.5) as the ratio between the energy output and the energy (2) (0) (0) (1) (0) (1) input. Using Eq. (3.3), we have where E(2)m(2) − E(0)m(0) E(1)m(1) 1 η = = − . (4.1) α(0) = U∗ 1 − µ(0) 1 − . (4.6) E(0)m(0) E(0)m(0) U∗

Note that the expression between brackets in the defini- A natural question arises: what is the maximum effi- tion of α(0) is precisely the effective energy of the inci- ciency of the Penrose process in a MP spacetime? Since dent particle evaluated at the break-up point as given by η is directly proportional to E(1) and inversely propor- Eq. (2.10). The inequalities given in Eq. (2.12), together tional to E(0), in order to maximize the efficiency of the with the fact that U∗ ≥ 1, therefore imply that process we need to make the absolute value of E(1) as large as possible and E(0) as small as possible. We also α(0) ≥ 1, (4.7) want the negative energy fragment to be as massive as possible in comparison to the mass of the incident parti- and cle. In other words, we want to extract as much energy as possible starting with as little energy as possible. We µ(0) ≤ 1. (4.8) shall assume that the MP spacetime is fixed (meaning that M1, M2 and a are known), the break-up point has Eq. (4.5) and the fact that the masses are positive, been specified as (ρ∗, z∗, φ∗) and the charge-to-mass ratio together with the constraint imposed by Eq. (3.6), also µ(1) is known. Given these hypotheses, we will determine imply that how much energy can be extracted from a MP black hole and how the remaining parameters must be chosen in m(1) 0 < < α(0). (4.9) order to optimize the process. m(0) 8

Since the radicand in Eq. (4.5) must be positive, the 0.5 bound given by Eq. (4.9) can be further reﬁned to yield 0.4

0.3 q 2 0 < m(1) < m(0) α(0) − α(0) − 1 . (4.10) 0.2

0.1 Hence, the allowed range for m(1) is maximized when the inequalities (4.7) and (4.8) are saturated. In fact, when 0.0 µ(0) → 1 one can choose m(1) → m(0), thus maximizing -0.1 the ratio m(1)/m(0) that appears in Eq. (4.1). Conse- quently, the eﬃciency of the Penrose process in a MP -0.2 spacetime is bound from above according to -3 -2 -1 0 1 2 3 b min η < η = −E(1) . (4.11)

We remark that the upper bound above is a function b FIG. 7. Efficiency bound η as a function of z∗ for ρ∗ = 0 of µ(1), the break-up point coordinates ρ∗ and z∗, and (solid curve), ρ∗ = 1/2 (dashed curve), and ρ∗ = 1 (dotted the MP parameters (M1, M2, and a). With the help of curve) when MR = 1/2, MT = 3, and µ(1) = −5/4. The asso- Eq. (4.3), we now investigate in detail how these quanti- ciated ergosphere corresponds to the point E in Figs. 1 and 2, tities affect the efficiency bound. while the associated energy levels are shown are shown in the middle panel of Fig. 3. The red dots indicate the efficiency when the particle breaks-up at the black holes. The black dot A. Dependence on the parameters shows the maximum efficiency ηb = 0.03161 for the processes represented in Fig. 11 and Tables III and IV. b The dependence of η on the charge-to-mass ratio µ(1) b is simple: η increases linearly with µ(1). The dependence of ηb on the break-up point can be understood with the help of the energy levels shown in Fig. 3: the efficiency fixed), ηb will also increase if the break-up point is closer bound increases as the break-up point approaches either to the lighter black hole and will decrease if the break-up one of the black holes. A more detailed analysis is shown point is closer to the heavier black hole. This behaviour b in Fig. 7, where we plot η as a function of z∗ for selected is related to the fact that the growth of MR produces an values of ρ∗ when MR = 1/2, MT = 3, and µ(1) = −5/4 expansion of the ergosphere around the lighter compan- (corresponding to the energy levels shown in the mid- ion and a reduction of the ergosphere around the heavier dle panel of Fig. 3). Note that when the break-up point companion, as seen in the bottom panel of Fig. 8. Note is outside the ergosphere, the efficiency bound becomes that the efficiency bound is independent of the mass ra- negative, meaning that the escaping fragment will carry tio if the break-up point is equidistant to the black holes. away less energy than the incident particle had. The On the other hand, as shown in Fig. 9, when MT /a in- Penrose process is most efficient if the break-up occurs creases (while the other parameters are kept fixed), the exactly at either one of the event horizons, as indicated ergosphere expands and ηb increases. We note that in by the red dots in Fig. 7. When this happens, the up- the limit MT /a → ∞, no matter where the break-up oc- per bound is −µ(1), which is in agreement with results curs, the efficiency bound approaches its maximum, i.e. b obtained for the RN metric [28, 29]. η → −µ(1). This is explained by the fact that the dis- To study the dependence of the efficiency bound on tance between the break-up point and the black holes the masses of the black holes, we plot ηb as a function of becomes negligible in comparison to the size of the ergo- the mass ratio MR when MT /a = 3 (top panel of Fig. 8) sphere when MT /a → ∞. and as a function of MT /a when MR = 1/2 (top panel b of Fig. 9). In both cases we have set the charge-to-mass Finally, we investigate the behaviour of η when µ(1) ratio as µ(1) = −5/12. Each curve in these plots is as- and MT /a vary simultaneously, but their product, i.e. sociated with a different break-up point. The break-up µ(1) = µ(1)MT /a, is kept fixed, meaning that the shape points are shown in the bottom panels of Figs. 8 and 9. of the ergosphere is fixed (only the energy levels inside of b In the bottom panel of Fig. 8, we also exhibit the gen- it change). In Fig. 10 we plot η as a function of µ(1) and eralized ergosphere associated with a few selected values MT /a for three different break-up points when MR = of MR (which are chosen to reproduce the ergospheres 1/2 and µ(1) = −5/4, so that the resulting ergosphere labeled D, E and F in Figs. 1 and 2). Analogously, the corresponds to the one labeled E in Figs. 1 and 2. The ergospheres depicted in the bottom panel of Fig. 9 (and break-up points are chosen to be (ρ∗, z∗) = (d, 1 − d), the associated values of MT /a) correspond to the ergo- with d = 1/2, d = 1/3, and d = 1/4. We observe that, as spheres labeled B, E and H in Figs. 1 and 2. |µ| increases and MT /a decreases, the efficiency bound As shown in the top panel of Fig. 8, when the mass increases, approaching an asymptotic value in the limit ratio MR increases (while the other parameters are kept µ(1) → −∞ and MT /a → 0. 9

0.20 0.35

0.15 0.30

0.25 0.10 0.20

0.05 0.15

0.10 0.00 0.05

-0.05 0.00 0.0 0.2 0.4 0.6 0.8 0 5 10 15 20 25 30

2 2

1 1

0 0

-1 -1

-2 -2

-2 -1 0 1 2 -2 -1 0 1 2

b b FIG. 8. Top panel: eﬃciency bound η as a function of MR FIG. 9. Top panel: eﬃciency bound η as a function of MT /a when MT /a = 3 and µ(1) = −5/12. Bottom panel: gen- when MR = 1/2 and µ(1) = −5/12. Bottom panel: generalized ergosphere for selected values of MR (correspond- eralized ergosphere for selected values of MT /a (corresponding to the points D, E and F in Fig. 1) when MT /a = 3 ing to the points B, E and H in Fig. 1) when MR = 1/2 and µ(1) = −5/12. The red dots indicate the location of and µ(1) = −5/12. The red dots indicate the location of the black holes. The purple dots are the locations of the the black holes. The purple dots are the locations of the break-up points (ρ∗, z∗) = (1/2, 1/2), (ρ∗, z∗) = (1/2, 0), and break-up points (ρ∗, z∗) = (1/2, 1/2), (ρ∗, z∗) = (1/2, 0), and (ρ∗, z∗) = (1/2, −1/2) and are labeled X, Y and Z, respec- (ρ∗, z∗) = (1/2, −1/2) and are labeled X, Y and Z, respectively. Each curve in the top panel corresponds to one of the tively. Each curve in the top panel corresponds to one of the break-up points shown in the bottom panel. break-up points shown in the bottom panel.

B. Examples Following the discussion leading to the inequality We now give concrete examples of energy extraction in (4.11), we set L(0) = L(1) = L(2) = 0, meaning that a MP binary black hole spacetime whose eﬃciency ap- all trajectories are restricted to the plane φ = φ∗. The proaches the theoretical maximum given by Eq. (4.11). energy E(1) is determined by Eq. (4.3), while E(0) = 1. The general procedure that we follow is outlined below. According to Eqs. (4.8) and (4.10), we choose First, we choose the MP parameters M1, M2 and a. Sec- ond, we choose the charge-to-mass ratio µ(1) and a break- up point (ρ∗, z∗, φ∗) that is inside the generalized ergosphere of the spacetime. Without loss of generality, we µ(0) = 1 − ε, (4.12) set m(0) = 1. m(1) = 1 − ν, (4.13) 10

be remedied by assuming, from the start, that the es- 5/40 5/32 5/24 5/16 5/8 caping fragment is massless (however, this would modify Eq. (2.8) and the present analysis). Nevertheless, it is 0.6 possible, in principle, to have inﬁnitesimally small values for ε and ν. Finally, we point out that not every angle θ0 produces trajectories that are consistent with the 0.4 assumptions of a Penrose process. More precisely, only certain ranges of θ0 give rise to trajectories T(0) and T(2) that, respectively, start and end inﬁnitely far away from

0.2 the black holes.

0 -10 -8 -6 -4 -2 0

b FIG. 10. Efficiency bound η as a function of µ(1) (bottom axis) and MT /a (top axis) for MR = 1/2 and µ(1) = −5/4. The fact that µ(1) is fixed implies that µ(1) and MT /a are inversely proportional to each other. The curves, from bottom to top, correspond to the break-up points (ρ∗, z∗) = (d, 1 − d), with d = 1/2, d = 1/3, and d = 1/4 respectively. The black dot shows the maximum efficiency ηb = 0.03161 for FIG. 11. Examples of Penrose processes that approach the the processes represented in Fig. 11 and Tables III and IV, maximum theoretical efficiency in a MP spacetime. The ef- i.e. when d = 1/2, µ = −5/12 and M /a = 3. (1) T ficiencies of the processes in the left and right panel are, respectively, 90% and 99% of the theoretical maximum. In both cases the incoming trajectory T (black curve) splits at the where ε and ν are small positive parameters satisfying (0) black point into the negative energy orbit T(1) (red curve) and s ! the trajectory of the escaping fragment T(2) (purple curve). 2 The parameters that generate these trajectories are shown in ν > ε(U∗ − 1) 1 + − 1 . (4.14) ε(U∗ − 1) Tables III and IV. The blue region is the generalized ergosphere (for particle 1) and the red dots indicate the location of the black holes. Since m(1) has been fixed, the mass m(2) can be determined by Eq. (4.5). The quantities µ(2) and E(2) are determined from Eqs. (3.1) and (3.3), respectively. At the break-up point, according to Eq. (4.2), we have i m(i) µ(i) E(i) L(i) ρ˙(i) z˙(i) ρ˙(1) =z ˙(1) = 0 for the negative energy fragment. For the incident particle, on the other hand, the choices for µ(0) 0 1.00000 0.99990 1.00000 0 0.00609 0.00158 and E(0) imply that 1 0.90000 -0.41667 -0.03161 0 0.00000 0.00000 2 0.09756 14.09301 10.54183 0 0.06243 0.01621 2 2 U∗ − 1 ρ˙(0) +z ˙(0) = ε 2 [2 + ε(U∗ − 1)] . (4.15) U∗ TABLE III. Parameters that generate the trajectories T , (0) By choosing the angle θ(0) = Arg ρ˙(0) + i z˙(0) between T(1), and T(2) of the Penrose process shown in the left panel the velocitiesρ ˙(0) andz ˙(0), equation Eq. (4.15) can be of Fig. 11. The derivativesρ ˙(i) andz ˙(i) are evaluated at the break-up point. used to determineρ ˙(0) andz ˙(0) individually at the break- up point. The conservation of linear momentum then fixes the values ofρ ˙(2) andz ˙(2) through Eq. (3.4). At this point, the trajectories T(0), T(1), and T(2) have all been determined and the efficiency of the associated Penrose i m(i) µ(i) E(i) L(i) ρ˙(i) z˙(i) process is η = (1 − ν)ηb. In order to maximize the effi- 0 1 0.99999 1.00000 0 0.00193 0.00050 ciency of the Penrose process, one must, therefore, set η 1 0.99000 -0.41667 -0.03161 0 0.00000 0.00000 to be as small as possible. Note, however, that one cannot 2 0.00685 206.13521 150.50460 0 0.28104 0.07297 choose ε = 0, because this would implyρ ˙(0) =z ˙(0) = 0 andρ ¨(0) =z ¨(0) = 0 at the break-up point, contradicting the fact the trajectory T(0) starts infinitely far away. Sim- TABLE IV. Parameters that generate the trajectories T(0), ilarly, ν cannot be chosen to saturate inequality (4.14), T(1), and T(2) of the Penrose process shown in the right panel of Fig. 11. The derivativesρ ˙ andz ˙ are evaluated at the otherwise m(2) would be exactly zero, contradicting the (i) (i) fact that all trajectories are timelike. If desired, this can break-up point. 11

We conclude by showing in Fig. 11 two explicit exam- efficiency of the Penrose processes in MP black holes, ples of the procedure outlined above for the MP space- understanding how it is influenced by several parameters time with M1 = 2, M2 = 1, and a = 1. The charge-to- and providing a prescription for maximizing the extrac- mass ratio of the negative energy fragment is chosen as tion of energy. µ(1) = −5/12, so that the associated generalized ergo- Thanks to the success of the LIGO-Virgo collabora- sphere is the one identified by the letter E in Figs. 1 and tion, we live in an age where direct gravitational wave 2. The break-up point is chosen as (ρ∗, z∗) = (1/2, 1/2), detections provide evidence on the abundance of compact fixing the trajectory T(1) and its energy (per unit mass) binaries in our Universe. For example, in the first half E(1) = −0.03161 (remember that the initial conditions of LIGO’s latest observation run, more than 30 binary are chosen so that E(1) is the minimum possible). Ac- system candidates were identified [30]. Nevertheless, the cording to Eq. (4.11), the efficiency bound is ηb = 0.03161 possibility of energy extraction from such systems by the (which corresponds to the black dots in Figs. 7 and 10). Penrose process is yet to be explored. In view of that, The trajectories T(0) and T(2) are specified by the choices the MP metric can give a fair idea of what happens in of ε, ν, and θ(0). In the left panel of Fig. 11 we ex- a “snapshot” of the collision process between two black hibit the Penrose process for ε = 10−4, ν = 10−1, and holes and can be considered a first step towards a better θ(0) = 0.25403, whose efficiency is 90% of the theoret- understanding of effects such as the Penrose process ical maximum. In the right panel of Fig. 11 we ex- in the non-static case. And a better understanding hibit the Penrose process for ε = 10−5, ν = 10−2, and of the Penrose process in black hole binaries may also θ(0) = 0.25403, whose efficiency is 99% of the theoretical contribute to a better understanding of superradiant maximum. The parameters that generate these examples effects [31–35] in similar scenarios. In fact, even though and satisfy Eqs. (3.1)-(3.6) are given in Tables III and IV. some interesting ideas regarding superradiance by binary systems have been proposed [36–38], further research is needed. V. FINAL REMARKS ACKNOWLEDGMENTS We have demonstrated the possibility of energy extraction by the Penrose process in a static binary black hole This research was partially financed by the Coor- described by the MP spacetime. Relying on the concept dena¸cãode Aperfei¸coamento de Pessoal de N´ıvel Supe- of a particle dependent generalized ergosphere, we have rior (CAPES, Brazil) - Finance Code 001. M. R. ac- shown that closed orbits of negative energy exist outside knowledges support from the Conselho Nacional de De- of the event horizons of a MP binary and can be used senvolvimento Cient´ıficoe Tecnológico(CNPq, Brazil), in energy extraction processes. We have also studied the Grant No. FA 315664/2020-7.

[1] M. Visser, in Kerr Fest: Black Holes in As- [12] T. Assump¸cão,V. Cardoso, A. Ishibashi, M. Richartz, trophysics, General Relativity and Quantum Gravity and M. Zilhão,Phys. Rev. D 98, 064036 (2018). Christchurch, New Zealand, August 26-28, 2004 (2007) [13] J. Shipley and S. R. Dolan, Class. Quant. Grav. 33, arXiv:0706.0622 [gr-qc]. 175001 (2016), arXiv:1603.04469 [gr-qc]. [2] C. Bambi, Mod. Phys. Lett. A26, 2453 (2011), [14] J. O. Shipley, Strong-field gravitational lensing by black arXiv:1109.4256 [gr-qc]. holes, Ph.D. thesis, Sheffield U. (2019), arXiv:1909.04691 [3] S. A. Teukolsky, Class. Quant. Grav. 32, 124006 (2015), [gr-qc]. arXiv:1410.2130 [gr-qc]. [15] A. Bohn, W. Throwe, F. Hébert, K. Henriksson, [4] E. Berti, General Relativity and Gravitation 51, 140 D. Bunandar, M. A. Scheel, and N. W. Taylor, Class. (2019), arXiv:1911.00541 [gr-qc]. Quant. Grav. 32, 065002 (2015), arXiv:1410.7775 [gr-qc]. [5] R. Penrose and R. M. Floyd, Nature Physical Science [16] D. Bini, A. Geralico, G. Gionti, W. Plastino, and N. Ve- 229, 177 (1971). landia, (2019), arXiv:1906.01991 [gr-qc]. [6] J. D. Schnittman, Gen. Rel. Grav. 50, 77 (2018), [17] D. Christodoulou and R. Ruffini, Phys. Rev. D 4, 3552 arXiv:1910.02800 [astro-ph.HE]. (1971). [7] A. Tursunov and N. Dadhich, Universe 5, 125 (2019), [18] J. B. Hartle and S. W. Hawking, Communications in arXiv:1905.05321 [astro-ph.HE]. Mathematical Physics 26, 87 (1972). [8] A. Tursunov, Z. Stuchl´ık, M. Koloˇs, N. Dadhich, [19] O. Semerákand M. Basovn´ık,Phys. Rev. D 94, 044006 and B. Ahmedov, Astrophys. J. 895, 14 (2020), (2016). arXiv:2004.07907 [astro-ph.HE]. [20] G. Denardo and R. Ruffini, Physics Letters B 45, 259 [9] K. Parfrey, A. Philippov, and B. Cerutti, Phys. Rev. Lett. (1973). 122, 035101 (2019), arXiv:1810.03613 [astro-ph.HE]. [21] J. Ryzner and M. Zofka,ˇ Class. Quant. Grav. 32, 205010 [10] S. D. Majumdar, Phys. Rev. 72, 390 (1947). (2015), arXiv:1510.02314 [gr-qc]. [11] A. Papaetrou, Proc. Roy. Irish Acad.(Sect. A) A51, 191 [22] A. A. Grib, Y. V. Pavlov, and V. D. Vertogradov, Mod. (1947). Phys. Lett. A 29, 1450110 (2014), arXiv:1304.7360 [gr- 12

qc]. [30] R. Abbott et al. (LIGO Scientiﬁc, Virgo), (2020), [23] O. B. Zaslavskii, Mod. Phys. Lett. A 36, 2150120 (2021), arXiv:2010.14527 [gr-qc]. arXiv:2006.02189 [gr-qc]. [31] C. Misner, Bulletin of the American Physical Society 17, [24] A. Prasanna and N. Dadhich, Il Nuovo Cimento B (1971- 472 (1972). 1996) 72, 42 (1982). [32] Y. B. Zel’Dovich, JETP Lett. 14, 180 (1971). [25] F. de Felice, F. Sorge, and S. Zilio, Classical and Quan- [33] A. A. Starobinsky, Sov. Phys. JETP 37, 28 (1973). tum Gravity 21, 961 (2004). [34] J. D. Bekenstein, Phys. Rev. D7, 949 (1973). [26] Z. Stuchl´ık, Bulletin of the Astronomical Institutes of [35] R. Brito, V. Cardoso, and P. Pani, Lect. Notes Phys. Czechoslovakia 31, 129 (1980). 906, pp.1 (2015), arXiv:1501.06570 [gr-qc]. [27] J. Levin and G. Perez-Giz, Phys. Rev. D 77, 103005 [36] J. G. Rosa, Phys. Lett. B749, 226 (2015), (2008), arXiv:0802.0459 [gr-qc]. arXiv:1501.07605 [gr-qc]. [28] M. Bhat, S. Dhurandhar, and N. Dadhich, Journal of [37] J. G. Rosa, Phys. Rev. D95, 064017 (2017), Astrophysics and Astronomy 6, 85 (1985). arXiv:1612.01826 [gr-qc]. [29] S. Parthasarathy, S. Wagh, S. Dhurandhar, and N. Dad- [38] L. K. Wong, Phys. Rev. D 100, 044051 (2019), hich, The Astrophysical Journal 307, 38 (1986). arXiv:1905.08543 [hep-th].