Beruflich Dokumente
Kultur Dokumente
volume V ∼ m5 . For a stellar black hole, this is larger III. FORMULATION OF THE PROBLEM
than our universe.
There is a lot of available real estate inside a black Consider the geometry of a collapsed object. We work
hole, according to classical general relativity! in ingoing Eddington-Finkelstein coordinates (v, r, θ, φ).
The result can be extended to other spherically sym- For simplicity, we take a null spherical shell of energy
metric geometries, like the non-singular black hole metric m collapsing along the v = 0 surface. Before this sur-
considered in [9] and [10]; in Appendix A we treat the face, spacetime is flat. After this surface, the geome-
Reissner-Nordström case. It can also be extended to the try is a Schwarzschild black hole and the line element
case of a Kruskal black hole. We do so in Appendix B. is the standard Schwarzschild geometry in Eddington-
The volume turns out to be infinite (as expected) but the Finkelstein coordinates
difference V (∆v) = V (v2 ) − V (v1 ) can be appropriately
defined and is finite: it grows linearly in ∆v = v2 − v1 ds2 = −f (r)dv 2 + 2dvdr + r2 dΩ2 , (3)
when ∆v is large.
In the last section, we present some considerations on where f (r) = 1 − 2m/r and dΩ2 = sin2 θ dφ2 + dθ2 . The
the relevance of these results for the “information para- relation with theR Schwarzschild coordinates t, r, θ, φ, is
dox” discussion. given by v = t + fdr (r) = t + r + 2m ln |r − 2m|.
The horizon is at r = 2m. It is foliated by spheres
Sv defined by r = 2m and constant v. The sphere Sv
II. VOLUME INSIDE A SPHERE is defined physically as the one crossed by a light signal
sent by a stationary observer at large (with respect to m)
Consider a (metric) 2d sphere S immersed in flat distance r from the hole, at proper time t = v − r.
Minkoswki spacetime. Let R be its radius and A = πR2 For each Sv we are interested in the spherically sym-
its area. We say that it encloses the volume V = 34 πR3 . metric 3d surface Σv which is bounded by Sv and has
What does this mean? It means that there is a 3d space- maximal 3d volume. The volume of this surface is called
like surface Σ bounded by S which has volume V . But V (v). Our objective is now to compute V (v). This a well
there are a lot of spacelike surfaces bounded by S in defined problem.
spacetime: which one do we mean, to define the inte-
rior volume? The answer can be given in two equivalent
IV. THE MAXIMISATION PROBLEM
manners:
(i) Σ lies on the same simultaneity surface as S. A 3d spherically symmetric surface Σ, can be thought
of as the direct product of a 2-sphere and a curve γ in
(ii) Σ is the largest spherically symmetric surface the v-r plane.
bounded by S.
Σ ≡ γ × S2 (4)
These two characterizations of Σ are equivalent. To see γ 7→ (v(λ), r(λ)). (5)
this, we can chose (without loss of generality) coordinates
(x, y, z, t) where S is given by t = 0, r2 ≡ x2 + y 2 + z 2 = The curve γ is given here in parametric form, with an
R2 . A spherically symmetric surface Σ bounded by S is arbitrary parameter λ. We choose λ = 0 on the horizon
defined by the function t = t(r), r ∈ [0, R], with t(R) = 0. (r = 2m) and call λf (f for ‘final’) the value of λ at r = 0.
Its volume is Thus, the initial and final endpoints of γ are given by
s
Z R 2 r(0) = 2m, r(λf ) = 0, (6)
2 dt(r)
V = dr 4πr 1− (2) v(0) = v, v(λf ) = vf . (7)
0 dr
The surface Σ is coordinatized by λ, θ, φ. The line
which is maximised by t(r) = 0 (because any variation element of the induced metric on Σ is
adds a contribution in the timelike directions and reduces
the volume), namely by the Σ on the simultaneity sur- ds2Σ = −f (r)v̇ 2 + 2v̇ ṙ dλ2 + r2 dΩ2
(8)
face. The “space inside S” is therefore the largest spher-
ically symmetric space bounded by S. where the dot indicates differentiation by λ. We hereafter
Let us now move to a curved spacetime. Given a sphere do not consider the flat part inside the horizon (under the
S in a spherically symmetric geometry, what is the vol- null shell) which would have contributed to the volume by
ume inside it? Lacking flatness, simultaneity surfaces about 4π(2m)3 /3. We are interested in the asymptotic
have no special significance, in general. But the second behavior of the volume, when v >> m and it will be seen
definition of the space inside S extends immediately, and that this contibution becomes negligible. The condition
we adopt it here from now. This allows us to define the that Σ is spacelike reads
volume V enclosed in any 2d sphere S in a spherically
symmetric spacetime. − f (r)v̇ 2 + 2v̇ ṙ > 0 (9)
3
from the requirement that ds2 |Σ > 0 for all coordinate Equations (14) and (16) are all we need to analyse the
values. The proper volume of Σ is given by geodesics. They can be recast in the form1
p
ṙ = − r−4 A2 + r4 f (r)
Z λf
(17)
Z q
VΣ [γ] = dλ dΩ r4 (−f (r)v̇ 2 + 2v̇ ṙ) sin2 θ 1
0 S2 v̇ = (18)
Z λf p A + r4 ṙ
= 4π dλ r4 (−f (r)v̇ 2 + 2v̇ ṙ). (10) It can be easily seen that A has to be negative for the
0
geodesic to be spacelike. Then, v̇ and ṙ are both negative
The surface Σv that extremizes the volume is deter- and there are only positive terms in (11). Integrating (17)
mined by the curve γv that extremizes this integral and we get
by the specification of v and vf . Z 2M
VΣ r4
= λf = dr p . (19)
4π 0 A2 + r4 f (r)
V. GEODESICS IN THE AUXILIARY
MANIFOLD Equation (19) shows that there is a restriction imposed
on A
The last equation shows that finding Σv is the same as A2 + r4 f (r) > 0
solving for the equations of motion with langrangian 27 4
⇒ A2 > −rV4 f (rV ) = m ≡ A2c (20)
p 16
L(r, v, ṙ, v̇) = r4 (−f (r)v̇ 2 + 2v̇ ṙ). (11)
The last condition comes from inspecting the polynomial
It is useful to note that this can be rewritten as −r4 f (r). It has roots at r = 0 and r = 2M , is otherwise
q positive in that range and reaches a maximum at rV =
3
L(r, v, ṙ, v̇) = g̃αβ dxα dxβ (12) 2 m.
We integrate numerically equations (17) and (18) be-
low in Section VII; but we can already directly derive
and can be thought of as the line element of an (auxiliary)
the essential lesson by noticing the following: inserting a
2d curved spacetime. That is, the metric g̃αβ is given by
constant radial value for r, (17) and (18) become
ds2Maux = r4 (−f (r)dv 2 + 2dvdr) (13) A2 = −r4 f (r) (21)
1
where α, β, ... can take the two values v and r. Finding v̇ = (22)
A
Σv is equivalent to finding the geodesics of this auxiliary
metric. Since −r4 f (r) > 0 in the range 0 < r < 2M , every
Furthermore, equation (10) shows that the proper constant r provides a solution. In other words, the r =
length of the geodesic in the auxiliary metric (times 4π) const. surfaces are spacelike geodesics of the auxilliary
is precisely the volume of Σ. manifold or equivalently stationary (maximal) points of
The condition (9) that Σ be spacelike, suggests that the volume functional (10). Integrating the v̇ equation
L > 0, since r is positive. The Langrangian appears to we get
vanish at r = 0, which is the final point for the geodesic
λf = A(vf − v). (23)
but, as can be seen from (17), ṙ becomes infinite. Thus,
γ is a spacelike geodesic in Maux . We now recognize that Thus, the r = const. surface with the largest volume
a well suited parametrization is to take λ as the proper betwen two given v is when A is largest. That is, for r =
length in Maux . After the extremization, we set rV which gives A = Ac . These considerations provide the
basis for the derivation of the asymptotic volume. This
L(r, v, ṙ, v̇) = 1 is done in the next section.
V = 4πλf . (15) We are now interested in the volume for large v. The
proper length of the geodesic should increase monoton-
The metric g̃αβ has a Killing vector, ξ µ = (∂v )µ ∝ ically from S which is situated at the endpoint (v, 2m)
(1, 0). Since γ is an affinely parametrized geodesic in
Maux , the inner product of ξ with its tangent ẋα = (v̇, ṙ),
is conserved
1 The plus sign choice in (17) would correspond to geodesics out-
4 side the horizon
r (−f (r)v̇ + ṙ) = A (16)
4
1 p
rV = 3M + 9M 2 − 8Q2 (A3)
4
ACKNOWLEDGMENTS
Then,
CR thanks Don Marolf for useful exchanges. MC
thanks Alejandro Perez for extensive discussions dur- A2c > −rV4 f (rV ) (A4)
ing the past year and Goffredo Chirco, Aldo Riello and
Ernesto Frodden for early discussions on this subject. It is then evident that the analysis presented here is es-
MC also acknowledges support from the Educational sentially the same in this spacetime. The difference is
Grants Scheme of the A.G.Leventis Foundation for the that instead of looking at the region between the singu-
academic years 2013-2014 and 2014-2015. larity and the horizon, we look at the region between the
two horizons. If one wishes to include the volume in the
region r < r− then a different analysis must be carried
Appendix A: Volume in Reissner-Nordström out for that part since r re-acquires its interpretation as a
spacelike coordinate. The situation for that region would
In this appendix, we check the reasoning presented be similar to that of Minkowski space (see Appendix C),
in this paper in a less trivial example. This spacetime, i.e. it would yield a constant contribution that becomes
describing a non-rotating spherically symmetric charged negligible for large times.
black hole, can be described by (3) with
f (r) = 1 − 2M/r + Q2 /r2 . (A1)
Appendix B: Volume in Kruskal
There are two horizons on this spacetime located at the
zeros of f , given by
A straightforward application of the definition of vol-
ume that we have given in the case of a Kruskal space-
p
r± = M ± M 2 − Q2 . (A2)
time gives infinite volume for any sphere Sv on the hori-
The outer horizon, at r+ , is an event horizon and the zon. We may obtain a finite volume by requiring Σ to be
inner horizon, at r− , is an apparent horizon. It is easily bounded by two spheres, Sv and Su , one on each of the
two outgoing horizons. Fixing Su , the volume V (u, v) of
this surface clearly satisfies
4 After the appearance of this paper on the arXiv, the result we √
point out has been developed for the spinning case by Bengtsson V (u, v2 ) − V (u, v1 ) ∼ 3 3πm2 (v2 − v1 ) (B1)
and Jacobson in [18] and for other cases by Yen Chin Ong in
[19]. in the asymptotic region.
7
Appendix C: Volume in Minkowski space from which we deduce that dr = dv. Thus, the max-
imal volume hypersurfaces are given by t = v − r =
It is instructing to see how the definition we gave in constant as expected. Notice also that the Eddington-
the introduction and the results in the paper work in Finkelstein diagram of flat space is the standard depiction
the trivial case of flat space. We can again define a null of Minkowski space in polar coordinates with the angular
coordinate v = t + r. This is the same as setting f (r) = 1 directions suppressed. That is, the vertical axis is now
in the line element (3) or most of the formulas given in t̄ = v − r = t and the EF-diagram is simply a t vs r plot
the paper. It is direct from the integral (19) that the with the usual 45 degrees causal structure.
maximum volume is given by A = 0 and that the result The situation is of course very different. In this
is the usual expression for the volume of a sphere with case, there is a finite maximum volume to be achieved,
radius 2M . The equations (17) and (18) become when A = Ac = 0. By the characterization we gave in
Section II, given a sphere Σ, we look for the spacelike
1 surface Σ0 that spans the interior of the sphere and
ṙ = (C1) has the largest volume. These will always be the
r2
1 t = constant surfaces.
v̇ = 2 (C2)
r
[1] M. Parikh, “Volume of black holes,”Physical Review D [10] C. Rovelli and F. Vidotto, “Planck stars,”
73 (June, 2006) 124021, arXiv:0508108 [hep-th]. arXiv:1401.6562.
[2] D. Grumiller, “The Volume of 2D Black Holes,” [11] F. Estabrook, H. Wahlquist, S. Christensen, B. DeWitt,
arXiv:0509077 [gr-qc]. L. Smarr, and E. Tsiang, “Maximally Slicing a Black
[3] W. Ballik and K. Lake, “The volume of stationary Hole,”Physical Review D 7 (May, 1973) 2814
black holes and the meaning of the surface gravity,” [12] R. Bousso, “A covariant entropy conjecture,”Journal
arXiv:1005.1116. of High Energy Physics 1999 (July, 1999) 004–004,
[4] W. Ballik and K. Lake, “Vector volume and black arXiv:9905177 [hep-th].
holes,”Physical Review D 88 (Nov., 2013) 104038, [13] H. M. Haggard and C. Rovelli, “Black hole fireworks:
arXiv:1310.1935. quantum-gravity effects outside the horizon spark black
[5] B. S. DiNunno and R. A. Matzner, “The volume inside a to white hole tunneling,” arXiv:1407.0989.
black hole,”General Relativity and Gravitation 42 (May, [14] A. Barrau and C. Rovelli, “Planck star phenomenology,”
2009) 63–76, arXiv:0801.1734. arXiv:1404.5821.
[6] T.K. Finch, “The volume of the Schwarzschild black hole [15] A. Barrau, C. Rovelli and F. Vidotto, “Fast Radio Bursts
with respect to assorted time slicings,” (Nov., 2012), and White Hole Signals,” arXiv:1409.4031.
arXiv:1211.4337 [hep-th]. [16] A. Perez, “No firewalls in quantum gravity: the role of
[7] M. Cvetič, G. W. Gibbons, D. Kubizák, and C. N. Pope, discreteness of quantum geometry in resolving the infor-
“Black hole enthalpy and an entropy inequality for the mation loss paradox,” arXiv:1410.7062.
thermodynamic volume,”Physical Review D 84 (July, [17] E. Bianchi, “Black hole entropy from graviton entangle-
2011) 024037, arXiv:1012.2888. ment,” arXiv:1211.0522.
[8] G. W. Gibbons, “What is the Shape of a Black Hole?,” [18] I. Bengtsson and E. Jacobson “Black Holes: Their large
arXiv:1201.2340. interiors,” arXiv:1502.01907.
[9] S.A. Hayward“Formation and evaporation of non- [19] I.C. Ong “Never Judge a Black Hole by its Area,”
singular black holes,” arXiv:gr-qc/0506126. arXiv:1503.01092.