Extremal limit of the regular charged black holes in nonlinear electrodynamics

Jerzy Matyjasek∗
Institute of Physics, Maria Curie-Sklodowska University
pl. Marii Curie-Sklodowskiej 1, 20-031 Lublin, Poland
(Dated: November 12, 2018)
The near horizon limit of the extreme nonlinear black hole is investigated. It is shown that
resulting geometry belongs to the AdS2 × S2 class with different modules of curvatures of subspaces
and could be described in terms of the Lambert functions. It is demonstrated that the considered
class of Lagrangians does not admit solutions of the Bertotti-Robinson type.

PACS numbers: [Link], [Link], [Link]

Still growing interest in the extremal black holes that is Let us consider general relativistic formulation of non-
arXiv:gr-qc/0403109v1 28 Mar 2004

seen recently is motivated by their unusual and not fully linear electrodynamics described by a class of a gauge
understood nature. The problems of entropy, semiclasi- invariant Lagrangians L(F ), where F = Fab F ab and
cal configurations, interactions with matter, or informa- Fab = ∂a Ab − ∂b Aa . According to the well-known theo-
tion paradox – to name a few – have not been resolved rem [17, 18] if L(F ) has a Maxwell asymptotic for weak
yet. Apart from their global structure and behaviour, fields, i.e., L(F ) ∼ F and LF = dL/dF → const as
the near-horizon region is also of interest. Indeed, apply- F → 0, then any static and spherically symmetric so-
ing appropriate limiting procedure in the geometry of the lutions to the coupled equations of the nonlinear elec-
extreme and near extreme black holes one can generate trodynamics and general relativity with nonzero electric
new exact solutions [1, 2, 3, 4, 5, 6]. charge cannot have a regular center. It does not mean
On the other hand, of the equal importance are the that this no go theorem forbids existence of regular black
questions of the nature of singularities that reside, hidden holes in general. Simple example of a regular magnetic
to external observer, in the centers of most black holes. solution has been presented by Bronnikov in Ref. [15].
According to the widespread opinion, singularities that It bears a formal resemblance to the solution found ear-
plague such models are symptoms of illness of the theory lier by Ayón–Beato and Garcı́a [14] describing electri-
rather than its health and the role of the regular solutions cally charged configuration. It should be noted however
in our understanding of the black hole physics could not that the Ayón–Beato and Garcı́a solution has been con-
be overestimated. One of the methods which can be used structed with the use of different Lagrangians valid in
in constructions of non-singular models is replacing black different regions [15]. Attempts to circumvent the con-
hole interior by a regular core. This idea appeared almost ditions of the no go theorem have been undertaken by
forty years ago, in mid sixties [7, 8, 9] and is actively Burinskii and Hildebrandt [19]. They demonstrated that
investigated today (see Ref. [10] for a discussion and a modifications of the Ayón–Beato and Garcı́a model, in
comprehensive list of references). The exact analytical which electric field does not extend to the central region
solution for a specific sources has been constructed in are in concord with the no go theorem
Refs. [11, 12, 13]. The coupled system of equations obtained from the
Among various regular models known to date, espe- total action
cially intriguing are the solutions to the coupled equa- 1
Z
√
tions of nonlinear electrodynamics and general relativ- S = d4 x −g [R − L(F )] (1)
16π
ity found by Ayón–Beato and Garcı́a [14] and by Bron-
nikov [15]. The latter describes magnetically charged for the static and spherically symmetric configuration de-
black hole, and provides an interesting example of the scribed by the line element of the form
system that could be both regular and extremal. More-
over, its simplicity allows exact treatment as the location 1
ds2 = −f (r)dt2 + dr2 + r2 dΩ2 , (2)
of the horizons can be expressed in terms of the Lambert f (r)
functions [16].
In this note we shall investigate the extremal magnet-
ically charged black hole with the special emphasis put 2m(r)
f (r) = 1 − , (3)
on the near horizon geometry and its relation with the r
exact solutions of the Einstein field equations with topol-
with nonlinear magnetic field reduce to simple quadra-
ogy AdS2 × S2 . To begin with, however, we remind a few
ture
basic facts.
1
Z
m(r) = L(F )r2 dr + C, (4)
4
∗ Electronic address: matyjase@[Link] where C is an integration constant. Integrating Eq. (4)
 2

with function L(F ) given by merge at

"   #
1/4 4w0
−2 F xextr = (15)
L(F ) = F cosh a , (5) 1 + w0
2
into a degenerate event horizon. Numerically one has
where F = 2Q2 /r4 and a is a free parameter, one obtains
xextr = 0.871, and qextr = 1.056. (16)
3/2 1/2
 
|Q| a|Q|
m(r) = C − tanh . (6) The three types of solutions therefore are: the regular
2a r
Employing boundary condition m(∞) = M and setting Wi(ξ)
a = |Q|3/2 /2M yields

Q2 ξ
 
2M − 0.4 − 0.2 0.2 0.4 0.6 0.8 1
f (r) = 1 − 1 − tanh . (7)
r 2M r −1
x+
One of the most attractive features of this solution 2
−2
is possibility to express the location of the horizons in
terms of the Lambert functions, Wi (ξ). The Lambert −3 1.5
functions [20] defined by means of the general formula
1
−4

0.5
exp(W (ξ))W (ξ) = ξ, (8) −5
q
have two real branches, W0 (ξ) and W−1 (ξ) with a branch −6 0.2 0.4 0.6 0.8 1
point −1/e. The run of the real branches of the Lambert
functions is displayed in Fig. 1. Since the value of the
FIG. 1: Location of the horizons x+ (upper curve) and x−
principal branch of the Lambert function at 1/e plays an
(bottom curve) plotted against q. The run of the real branches
important role in our considerations we put of the Lambert function W0 (ξ) (upper curve) and W−1 (ξ)
(lower curve) is displayed for comparison. The degenerate
W0 (1/e) = w0 . (9) 1/2
event horizon at x+ = 2w0 corresponds to the branch point
To solve equation f (r) = 0 let us introduce a new of the Lambert function at ξ = −1/e.
‘radial’ coordinate x and a new parameter q by means of
r = M x and Q2 = q 2 M 2 , respectively. Subsequently black hole with the inner and event horizons for q < qextr ,
introducing a new unknown function W̃ the extremal black hole for q = qextr , and the regular
configuration for q > qextr . At large distances as well
4q 2 as for small charges the geometry (2) resembles that of
x=− , (10)
4W̃ − q 2 the Reissner-Nordström with one notable distinction: for
q > 1 the Reissner-Nordström solution describes naked
one arrives at singularity whereas the regular geometry (2) could be
q2 interpreted as a particle like solution. Qualitative be-
exp(W̃ )W̃ = − exp(q 2 /4). (11) haviour of f (r) is displayed in Fig. 2.
4
As the result of simple manipulations one can relate the f(r)
exact location of the event horizon r+ (= M x+ ) and the 1
inner horizon r− (= M x− ) with the the real branches of
0.8
the Lambert functions
0.6
4q 2
x+ = − , (12) 0.4
4W0 (− q4 exp(q 2 /4)) − q 2
2

0.2
and r
1 2 3 4 5
4q 2 − 0.2
x− = − 2 . (13)
4W−1 (− q4 exp(q 2 /4)) − q 2
FIG. 2: Qualitative behaviour of the function f (x) for vari-
The horizons r+ and r− for ous values of the parameter q. Top to bottom the curves are
for q > qextr , q = qextr , and q < qextr . The double root
qextr = 2w0 ,
1/2
(14) corresponds to the extremal event horizon.
 3

To investigate the near-horizon geometry of the regular the Einstein fiels equations reduce to two independent
magnetic black hole further let us postpone analyses of equations:
the solution (2) for a while and consider the line element
of the form 1 1
= L(F ) (27)
r02 2
1
ds2 = −dt2 + dy 2 + r02 dΩ2 .


(17) and
hy 2
 
Other useful representations obtained through a redefi- dL(F ) 1
h = 2 F23 F 23 − L(F ) . (28)
nition of coordinates according to formulas dF 4

h1/2 t = eτ coth χ, Now we shall select the particular form of the La-
grangian. Let us choose L(F ) as given by (5) with
h1/2 y = eτ sinh−1 χ (18)
|Q|3/2
and a = , (29)
2α
sinh2 χ = Rh − 1, where α is some positive parameter of dimension of
τh = T (19) length. Introducing a new coordinate x and a new param-
eter q by means of the formulas r0 = αx and Q = αq,
are Eqs. (27) and (28), after some manipulations, could be
rewritten as
1
ds2 = − sinh2 χdt2 + dχ2 + r02 dΩ2


(20)
q2
 2
h −2 q
cosh = 1 (30)
x2 2x
and
and
dR2
ds2 = − R2 h − 1 dT 2 + + r02 dΩ2 , (21)


 2
R2 h − 1 q
q2 q2 q 4 sinh 2x
 
h = 4 2 cosh−2 −  . (31)
respectively. Topologically it is AdS2 × S2 , i.e., a direct x α 2x 2x5 α2 cosh3 q2
2x
product of the two-dimensional anti-de Sitter geometry
and two-sphere of constant curvature. The curvature One can relegate hyperbolic functions combining
scalar for the line element (17) is simply a sum of the Eqs. (30) and (31)
curvatures of the subspaces AdS2 and S2 :
1/2
q2 x2

1
R = KAdS2 + KS2 , (22)
2
− 1− 2 = hα2 . (32)
x 2x q
where KAdS2 = −2h and KS2 = 2/r02 .
In general the problem should be treated numerically:
We intend to solve the coupled system of equations of
for a particular choice of q the equation (30) gives con-
the nonlinear electrodynamics and general relativity. For
crete values of x and hence h. First, observe that depend-
the line element (17) the Einstein tensor is given by
ing on q, Eq. (30) has two, one or has no solutions at all.

1 1
 Numerical calculations indicate that for qc = 1.325 there
b
Ga = diag − 2 , − 2 , h, h , (23) is only one solution at x = 0.735, whereas for q < qc
r0 r0
Eq. (30) has two solutions. Moreover, it should be noted
whereas the electromagnetic tensor compatible with as- that there is a particular combination of q and x express-
sumed symmetries is simply ible in terms of the Lambert functions satisfying Eq. (30).
Indeed, it could be easily checked that q = qextr and
F̃ = Q sin θdθ ∧ dφ, (24) x = xextr (given respectively by Eqs. (14) and (15))
comprises such a solution with h given by
where Q is, as before, the magnetic charge. It could be
easily verified that 1 (1 + w0 )3
h = . (33)
32α2 w02
2Q2
F = Fab F ab = > 0. (25) Results of numerical calculations are displayed in Fig. 3.
r04

As the stress-energy tensor of the nonlinear electrody- Although Bertotti-Robinson line element [21, 22] is a
2
namics is special case of (17) with h = 1/r+ it does not belong to a
  family of solutions of Eq. (30). It could be easily shown
b 1 dL(F ) bc 1 b that vanishing of the trace of the stress-energy tensor
Ta = Fac F − δa L(F ) , (26)
4π dF 4 (26) is, by (25), possible only for Lagrangians L(F ) ∝ F,
 4

x 2
Since A−1 > r+ , the line element does not belong to the
Bertotti-Robinson class, as expected. This result could
0.8
be easily anticipated as the stress-energy tensor of the
0.6 nonlinear electromagnetic field has nonvanishing trace at
the event horizon. It should be noted that putting α =
0.4 M in (17) with (33) one obtains (37).
0.2 It is evident that the procedure of constructing the
near horizon geometry is insensitive to the nature of the
q
0.2 0.4 0.6 0.8 1 1.2 central region of the black hole. One expects, therefore,
that similar consideration could be carried out for the
FIG. 3: Curvature radii of S 2 (solutions of Eq. (30)) plotted electrically charged Ayón–Beato and Garcı́a solution as
against parameter q. modified by Burinskii and Hildebrandt.
Finally let us return to the problem of boundary con-
and, consequently, L(F ) in the form given by (5) does ditions and restrict discussion to the solutions describ-
not admit solutions of the Bertotti-Robinson type. ing black holes. Assuming that the exact location of
Now, let us return to the black hole solution. The the event horizon is known, the boundary condition
extremal black hole is described by a line element (2) m(r+ ) = r+ /2 gives
with
  
2M 2M w0 r+ |Q|3/2

a|Q|1/2

f (r) = 1 − 1 − tanh . (34) m(r) = + tanh
r r 2 2a r+
3/2 1/2
 
It could be easily shown that x+ given by (15) is the |Q| a|Q|
− tanh . (38)
degenerate event horizon. In order to investigate the ge- 2a r
ometry of the vicinity of the event horizon and to obtain
uniform approximation we introduce new coordinates t̃
and y : Denoting first two (constant) terms in the right hand
side of the above equation by MH and demanding a =
t ε |Q|3/2 /2MH results in the equation which relates r+ , Q,
t̃ = , r = r+ + , (35)
ε Ay and MH
where
3 Q2
 
1 (1 + w0 ) 2MH
A = . (36) 1− 1 − tanh = 0, (39)
32M 2 w02 r+ 2MH r+

Expanding the function f (r) in terms of ε, retaining

quadratic terms and subsequently taking the limit ε = 0 and the line element (2) with (7). The location of the
we obtain inner horizon is given by Eq. (13) and the near horizon
1 geometry is, of course, identical to the one considered
ds2 = 2
−dt2 + dy 2 + r+ dΩ2 .


2
(37) previously.
Ay

[1] P. Ginsparg and M. J. Perry, Nucl. Phys. B222, 245 [7] A. D. Sakharov, Sov. Phys. JETP 22, 21 (1966).
(1983). [8] E. B. Gliner, Sov. Phys. JETP 22, 378 (1966).
[2] O. B. Zaslavskii, Phys. Rev. D56, 2188 (1997), gr- [9] J. M. Bardeen, in GR 5 Proceedings (Tbilisi, 1968).
qc/9707015. [10] I. Dymnikova, Gen. Rel. Grav. 24, 235 (1992).
[3] M. Caldarelli, L. Vanzo, and S. Zerbini, The ex- [11] I. Dymnikova, Int. J. Mod. Phys. D12, 1015 (2003), gr-
tremal limit of d-dimensional black holes (2000), hep- qc/0304110.
th/0008136. [12] A. Borde, Phys. Rev. D55, 7615 (1997), gr-qc/9612057.
[4] N. Dadhich, On product spacetime with 2-sphere of con- [13] M. Mars, M. M. Martı́n-Prats, and J. M. M. Senovilla,
stant curvature (2000), gr-qc/0003026. Class. Quantum Grav. 13, L51 (1996).
[5] O. J. C. Dias and J. P. S. Lemos, Phys. Rev. D68, 104010 [14] E. Ayon-Beato and A. Garcia, Phys. Lett. B464, 25
(2003), hep-th/0306194. (1999), hep-th/9911174.
[6] V. Cardoso, O. J. C. Dias, and J. P. S. Lemos, Nariai, [15] K. A. Bronnikov, Phys. Rev. D63, 044005 (2001), gr-
Bertotti-Robinson and anti-Nariai solutions in higher di- qc/0006014.
mensions (2004), hep-th/0401192. [16] J. Matyjasek, Phys. Rev. D63, 084004 (2001), gr-
 5

qc/0010097. [20] R. M. Corless, G. H. Gonnet, D. E. G. Hare, D. J. Jeffrey,

[17] K. A. Bronnikov, V. N. Melnikov, G. N. Shikin, and K. P. and D. E. Knuth, Adv. Comput. Math. 5, 329 (1996).
Staniukowicz, Ann. Phys. 118, 84 (1979). [21] B. Bertotti, Phys. Rev. 116, 1331 (1959).
[18] K. A. Bronnikov, Phys. Rev. Lett. 85, 4641 (2000). [22] I. Robinson, Bull. Acad. Polon. 7, 351 (1959).
[19] A. Burinskii and S. R. Hildebrandt, Phys. Rev. D65,
104017 (2002), hep-th/0202066.