Exact solutions in F (R) theory of gravity

Changjun Gao
Key Laboratory for Computational Astrophysics, National Astronomical Observatories,
Chinese Academy of Sciences, Beijing, 100012, China and
State Key Laboratory of Theoretical Physics, Institute of Theoretical Physics,
Chinese Academy of Sciences, Beijing 100190, China

You-Gen Shen
Shanghai Astronomical Observatory, Chinese Academy of Sciences, Shanghai 200030, China
(Dated: September 2, 2016)
We find a new method for looking for the static and spherically symmetric solutions in F (R)
theory of gravity. With this method, a number of new solutions in terms of the analytic functions
arXiv:1602.08164v6 [gr-qc] 1 Sep 2016

are obtained. We hope this investigation may be of some help in the searching for some other
solutions in F (R) theory of gravity.

PACS numbers: 98.80.Jk, 04.40.Nr, 04.50.+h, 11.25.Mj

I. INTRODUCTION terms of known analytic functions are obtained. We hope

the investigation may be of some help in the searching
The searching for exact solutions in the F (R) theory for other new solutions in F (R) theory of gravity with or
of gravity is important but challenging because the equa- without energy-momentum tensor. Throughout this pa-
tions of motion are of fourth order derivatives. Although per, we adopt the system of units in which G = c = ~ = 1
the searching is challenging, many remarkable achieve- and the metric signature (, +, +, +).
ments have been made. We here briefly review these
achievements.
II. EXACT SOLUTIONS
For the specific choice of R1+ , a class of exact static
spherically symmetric solution has been presented in [1].
A. Ansate for the metric
Using the method of Lagrange multiplier, reference [2]
presents a Lagrangian derivation of the equation of mo-
tion for the static spherically symmetric spacetimes in In general, the metric of a four dimensional, static and
F (R) theory of gravity. It is found the corresponding spherically symmetric spacetime takes the form of
equations of motion are simply of first order derivative 1
and thus some new solutions are obtained. Reference ds2 = U (x) dt2 + dx2 + f (x) d2 , (1)
A (x)
[3] constructs some new static spherically symmetric in-
terior solutions in R1+ theory. With the method of where x plays the role of radial variable and d2 is
Noether symmetries, reference [4] find some new static the line element for the two dimensional unit sphere.
spherically symmetric solutions in F (R) theory. Refer- U (x), A(x), f (x) are three functions to be determined.
ence [5] constructs the spherically symmetric solutions of One can always express A(x) as follows
F (R) gravity using their input function method. Refer-
ence [6] presents a static axially symmetric vacuum solu- A (x) = U (x) B (x) f (x)2 , (2)
tion for F (R) gravity in the Weyls canonical coordinates.
Finally, using the so-called generator method, reference with B (x) a new function to be determined. The metric
[7] presents some static spherically symmetric solutions becomes
in F (R) theory in n dimensional spacetimes. For more 1
works on static solutions in F (R) gravity, we refer to the ds2 = U (x) dt2 + 2 dx
2
+ f (x) d2 .
U (x) B (x) f (x)
references [7-29]. (3)
In this paper, we develop an alternative method for Let
looking for static and spherically symmetric solutions in
dx
Z
F (R) theory of gravity. Similar to reference [5], we appro- r= , (4)
priately fix one of the unknown functions initially. Then B
the equations of motion become remarkably simple and
then we find the metric can be written as
solvable. Using this method, a number of new solutions in
1
ds2 = U (r) dt2 + 2 dr
2
+ f (r) d2 . (5)
U (r) f (r)
Electronic address: gaocj@bao.ac.cn The advantage of this form is that the determinant of the
Electronic address: ygshen@center.shao.ac.cn metric is independent of the radial variable, r. We shall
 2

see later the equations of motion are greatly simplified Then we obtain a class of exact solutions.
with this form of metric. Given the metric, the Ricci 2(+1)
scalar R is calculated to be f (r) = b0 r 22 +2+3 ,
2(+1)

3 2 K (r) = b
0r
22 +2+3 ,

R = U f 2 + 3f U f + 2f U f + f 2 U , (6) 2(21)
224+1 2+2
2 f U (r) = d0 r 22 +2+3 + c1 r +2+3 + c2 r 22 +2+3 ,
2(21)
222+2 2
where the prime denotes the derivative with respect to r. R = d1 r +2+3 + d2 r 2 +2+3
,
2

2
2 + 2 + 3
d0 ,
(22 2 1) (42 + 2 + 1) b30
B. Exact solutions ( 1)
d1 6 ,
b0 (22 2 1)


The Lagrangian density of F (R) theory of gravity takes c2 b20 23 2 5 2
d2 6 2 ,
the form (22 + 2 + 3)
(13)
L = F (R) . (7)
where , b0 , c1 , c2 are constants. We see the expression
of K(r) and R(r) are obtained. Taking Eq. (11) into
Compared to struggling with the Einstein equations, consideration, we can derive the expression of F (R). In
it is much more convenient to directly vary the action the next, we shall investigate some specific solutions with
different exponent in order that the expression of F (R)
Z
are analytic functions.
S= d4 x gL , (8) To understand the solutions more explicitly, lets
rewrite the metric in the Schwarzschild coordinate sys-
tem. Rescale proper distance s and time variable t as
with respect to U and f , respectively. By so doing, we follows, s2 b0 s2 ,t2 b0 t2 and let
find the equations of motion are
22 +2+3
r = x +1 , (14)
 
2


 3 2
f K 3f Kf +K f + 2f f = 0 , (9) we obtain
2
i  
42 2 42 +2+1 42 42
h
(2f U K) K (3U f ) ds2 = x +1 d0 + c1 x +1 + c2 x +1 dt2
 
2  2 +2+1 2 42
1
+K 3U f + 2U f + 2f U + 2 = 0 , (10) +d3 d0 + c1 x 4
+ c2 x 4
dx2
f +1 +1

K = K (r) F,R . (11)

+x2 d2 , (15)

Here F,R dR dF
. We have only two equations of motion with
(9) and (10) but three functions to be determined: U , f
2
and K. Therefore the system of equations (9) and (10) 1 22 + 2 + 3
d3 3 , (16)
are not closed. It seems reasonably to initially specify b0 ( + 1)2
a certain form of F (R) by hand. However, by this way, K (x) = x2 , (17)
the equations of motion turn out to be fourth order dif- 2(12)
ferential equations and it is very difficult to find their R = b0 d1 x2 + b0 d2 x +1 . (18)
solution.
Rescale t, c1 , c2 as follows
Observing equations (9) and (10), we see if we fix not
the expression of F (R), but the expression of K(r) ini- t2 t2 /d3 , c1 c1 d3 , c2 c2 d3 , (19)
tially, then we can solve for f (r) from Eq. (9). Once
K(r) and f (r) are given, the function U (r) can be ob- the metric is simplified to be
tained from Eq. (10). Above is exactly the strategy what  
42 2 2 +2+1 2 42
we shall take in the following calculations. 2
ds = x +1 c0 + c1 x 4 +1
+ c2 x 4 +1
dt2
In order to obtain the analytic solutions, we specify K
1
as follows

42 +2+1 42 42
+ c0 + c1 x +1 + c2 x +1 dx2

K (r) = f (r) . (12) +x2 d2 , (20)
 3

where given by
h c1 i
d0 ( + 1)
2 s2 = x 1 + + c2 x dt2 + 2x2 d2
c0 = . (21) x
d3 (2 2 1) (42 + 2 + 1)
2 h c1 i1
+ 1+ + c2 x dx2 , (25)
x
We see there are three parameters in the solution, i.e.,
, c1 , c2 . We shall see later c1 and c2 play the role of  
2R
q
mass-like term and (cosmological constant)-like term, F (R) = 9c2 81c22 + 20R
25
respectively. Up to this point, we are able to understand  12
the solutions and construct the expression of F (R).
q  q
81c22 + 20R + 6c2 81c22 + 20R + 9c2 (26)
,

which is also endowed with a deficit angle. When c2 = 0,

C. some examples
we have
 c1  2  c1 1 2
1. F = R 2 ds2 = x 1 + dt + 1 + dx + 2x2 d2 (. 27)
x x
Let c1 = 2M with M a positive constant. We have
Consider the first term in the expression of g00 and let 1
it be a constant, we obtain = 0, 21 . In the case of
  
2 2M 2 2M
= 0, the solution is given by ds = x 1 dt + 1 dx2 + 2x2 d2 (28)
.
x x
ds2 = 1 + c1 x1 + c2 x2 dt2 + x2 d2
  This spacetime has an event horizon at x = 2M and a sin-
1 2 gularity at x = 0 which is the same as the Schwarzschild
+ 1 + c1 x1 + c2 x2

dx , spacetime. But different from the schwarzschild space-
K (x) = 1 , time, it has a deficit angle. Furthermore, the physical
meaning of M may be not the mass of some object and
F (R) = R 6c2 . (22)
it is still unclear for us.
The metric is exactly the well-known Schwarzschild-de
Sitter solution provided that 81c2 21
4. F = R
2
2

c2 = /3 , c1 = 2M . (23)
Let the first term in the expression of Ricci scalar R
Here , M are the cosmological constant and the black disappear, we obtain = 1. In this case, the solution is
hole mass respectively. given by
 
c1 4
ds = x 1 + 7 + c2 x dt2 + x2 d2
2
x2 7
2. F = 2 12c2 R  1
c1
+ 1 + 7 + c2 x dx2 , (29)
When = 12 , we find the solution is given by x2

81c22 21
x2 F (R) = . (30)
ds = 1 + c1 x2 + c2 x2 dt2 + d2
2
 
R 2
2
 2 2 1
 2 In this case, the spacetime has the same property as the
+ 1 + c1 x + c2 x dx , case of = 1/4.
K (x) = x ,
2
R = 12c2 2 , 2 3
x 5. F = 9
(R) 2
p
F (R) = 2 12c2 R . (24)
When = 12 , the solution is given by
This is the massless Reissner-Nordstrom-de Sitter black h c1 i 1
hole with a deficit angle. ds2 = x4 1 + 2 dt2 + x2 d2
x 2
h c1 i1 2
+ 1 + 2 dx , (31)

h
81c2
i
x
2 +20R+6c2
h p i
2R
3. F = 25 9c2 81c22 + 20R h i1
81c2
2 +20R+9c2
2 2 3
F (R) = (R) 2 . (32)
9

Let the second term in the expression of g00 be a con- This spacetime has an event horizon x = c1 , a curva-
stant, we obtain = 41 . In this case, the solution is ture singularity at x = 0 and a deficit angle.
 4

6. F = 16
R
30
c
49 2
In Fig. 1, we plot the evolution of exponents k, m with
respect to the dimensionless constants . When = 0,
When = 2, the solution is given by we have k = 1 and m = 2 which corresponds to the
Schwarzschild-de Sitter solution.
h c1 c2 i 1 In Fig. 2, we plot the position of event horizons with re-
ds2 = x4 1 + 7 + 2 dt2 + x2 d2
x x 7 spect to for different c1 and c2 . Curves A, B, C, D, E
h c1 c2 i1 2 correspond to c1 = 1.795, 1.995, 2.039, 2.075, 2.175 and
+ 1 + 7 + 2 dx , (33) c2 = 0.026, 0.031, 0.036, 0.039, 0.049, respectively. The
x x
figure shows that, in the case of A, there exists two hori-
zons for every . The smaller one represents the black
16 30 hole event horizon and the larger one represents the cos-
F (R) = c2 . (34)
R 49 mic event horizon. The case of B is exactly the same as
A. With the increasing of c1 and c2 , the black hole event
The structure of this spacetime is the same as the case
horizon expands and the cosmic event horizon shrinks.
of = 21 .
For some special value of , two horizons coincide which
is the case of C. Continue increasing c1 and c2 , we find
D. general properties: singularity, event horizons both the black hole event horizon and cosmic event hori-
and deficit angle zon vanishes and the black hole singularity is naked in
some intervals of (the case of D and E). In this case,
the conjecture of cosmic censorship is violated.
In this section, let us study the structure of spacetime
generally from Eq. (15). We find in general, there are
two event horizons and one curvature singularity in the
spacetimes. Furthermore, a deficit angle always exists 4
in the solutions except for the Schwarzschild-de Sitter
2
solution. Now let us show these points one by one.
In the first place, equation (18) shows x = 0 is always 0 m
the singularity of spacetime because of the divergence of
the first term at x = 0. Secondly, the spacetime has 2
a deficit angle of 1/d3 except for the schwarzschild-de
Sitter solution (d3 =1). In the following we turn to the 4 k
horizons. The equation of event horizons could be derived
from equation (15) as follows 6

42 +2+1 42 42 8
d0 + c1 x +1 + c2 x +1 =0. (35)
10
Let c1 , c2 absorb non-vanishing d0 and
0 1 2 3 4
2 2
4 + 2 + 1 4 4 2
k , m , (36)
+1 +1 FIG. 1: The evolution of exponents k, m with respect to the
dimensionless constants . When = 0, we have k = 1
we have and m = 2 which corresponds to the Schwarzschild-de Sitter
solution.
1 + c1 xk + c2 xm = 0 . (37) .
When

< 1 , (38)
III. CHECKING THE SOLUTIONS IN THE
we find JORDAN FRAME

k>0, m>0. (39)

On the other hand, when
> 1 ,
k, m would evolve with as shown in Fig. 1. The struc-
ture of event horizons for < 1 is similar to the case
of > 1. So we may pay our attention to > 1.
In section IIB, by varying the action with respect to
U and f , we derive two equations of motion, Eq. (9) and
Eq. (10). Then a class of solutions are obtained from
(40) the system of equations. Here we should emphasize that,
in the applying of this method, some unnecessary solu-
tions which actually do not satisfy the Einstein equations
might be present. The reason is simply. We obtain two
equations of motion, Eq. (9) and Eq.(10) with the varia-
 5

by
7
1 

R00 = AA B x + 2AA Bx A 2
Bx + 4AA B ,
4A2 x
6 1 

R11 = AA B x + 2AA Bx A 2 Bx + 4A2 B ,
4A2 x
5 A 1 

R22 = A Bx + AB x + 2AB 2A . (46)
2Ax2
x4 B
E The combinations of Eqs. (43-45) yield three equations
D C of motion as follows
3
C D E
BK BA K

R00 R11


B K + BK + = 0 ,(47)
2 2 2A
A

BK BK
K R11 R22 + BK = 0 , (48)
x 2
0.2 0 0.2 0.4 0.6

BA K BK
K R22 R00 + =0. (49)
FIG. 2: The position of event horizons with respect to for
2A x
different c1 and c2 . Substituting Eqs. (15-17) into Eqs. (47-49), we find the
. solutions satisfy all the Einstein equations.

IV. UNDERSTANDING THE SOLUTIONS IN

tion method. But Einstein equations show us with three THE EINSTEIN FRAME
equations of motion. Eq. (9) and Eq. (10) are exactly
some combinations of the three Einstein equations. So
It is well-known that F (R) gravity can be written in
solutions satisfied by the Einstein equations are certainly
the Einstein frame where it emerges as General Relativity
satisfied by Eq. (9) and Eq. (10). However, the converse
plus a scalar field. So in order to under the solutions
may be not true. Therefore, one should check wether the
better, lets analyze the solutions in the Einstein frame.
solutions meet all the Einstein equations. In this section,
The action for modified gravity in the absence of matter
we shall complete this examination.
is given by
In the Jordan frame, the generalized Einstein equations
corresponding to action (7) take the form
1
Z
S= d4 x gF (R) , (50)
1 22
g 2 F,R = 0 . (41)


F,R R F g
2
where 2 = 8. The action is equivalent to [30, 31]
In the Schwarzschild coordinate system, the metric for
1 h
Z i

the static and spherically symmetric spacetime is S = 2 d4 x g F () + F () (R ) , (51)
2
1
ds2 = A (x) dt2 + dx2 + x2 d2 . (42)
where F () = dF/d. On can verify that the equation
B (x)
of motion for the scalar field gives
Then Eqs. (41) turn out to be =R, (52)

F BK 2BK which reproduces the original action. Implemented with
KR00 + BK + + = 0 , (43) a conformal transformation,
2 2 x

F BA K 2BK
KR11 + + =0, (44) F () g = g , (53)
2 2A x

F BK BA K the action is reduced to the Einstein gravity plus the
KR22 + BK + +
2 2 2A scalar field which is minimally coupled to gravity [32-35]

BK 
1 4 3 g
Z
+ =0, (45) F
F (54)
x S = d x g R
22 2F 2
1  
where K F,R and prime denotes the derivative with
2 F F . (55)
respect to x. The components of Ricci tensor are given F
 6

Define a canonical scalar field by solutions are

6
r
3
= ln 2x2 ,
= 2
ln F , (56) 8
2
1 6
V = e ,
we find 22
ds2 = g dx dx
1 
2 4

 c1  2  c1 1 2
  = x 1 + dt + 1 + dx
1 1
Z
V () , x x x
S= d4 x
g R g

22 2 2 4
1

1   + 2x2 d2 . (60)
V ( ()) = 2 2 F F . (57) x
2 F
We have checked they satisfy all the Einstein equations,
The corresponding Einstein equations are given by Eq. (58).
  
2 1 2
G = + g V . (58)
2
81c2 21
3. F = R
2
2
In the next, we shall transform the solutions obtained
in section II-C, from the Jordan frame to the Einstein
frame. The first solution is for the Schwarzschild-de Sit- In this case, we have
ter spacetime and it is trivial. So we shall start from the
second one.
6
= ln x ,

3 6/2  
/ 6
V = e 12c 2 + 7e ,
42
1. F = 2 12c2 R
ds2 = g dx dx
7 9 11
x 2 + c1 + c2 x 2 2 x2
In this case, we derive the scalar field, the scalar field = dt + 7 dx2
potential and the metric as follows x x + c1 + c2 x
2 9/2

4 4 2
+ x d , (61)
7
6 x2
= ln ,
4 2
1 6  8  and they satisfy all the Einstein equations, Eq. (58).
V = e 1 + 12c e 3 ,
2
22
ds2 = g dx dx
! 3
2 2 2 2 2 2x 4. F = 2
(R) 2
3
= x 1 + c1 x + c2 x dt + d2 9
2 2 4

2
1 2 In this case, we have
+ x 1 + c1 x2 + c2 x2 dx . (59)
2

6 x2
We have checked they satisfy all the Einstein equations, = ln ,
4 2
Eq. (58). 3 6/3
V = e ,
22
ds2 = g dx dx
 
2.
5
F = 4 5 4 R 4
5 2 4
 c1  2  c1 1 2
= x 1 + dt + 1 + dx
x x x

For the third solution presented in subsection C, we 2
+ xd2 . (62)
find the expressions of both the scalar potential and the 2
Einstein tensor are rather lengthy. So for simplicity, we
shall take c2 = 0 in this subsection. The corresponding and find they satisfy all the Einstein equations, Eq. (58).
 7

5. F = 16
R
30
c
49 2
but without the black hole carrying an extra (indepen-
dent) scalar charge. This is sometime referred to as hair
In this case, we have of the second kind.

6 x2
= ln ,
7 V. CONCLUSION AND DISCUSSION

1 2 6  
/ 6
V = e 3 15c 2 196e ,
492
2 In conclusion, we have developed a new method for
ds = g dx dx looking for static and spherically symmetric solutions in
1  c1 c2  1 6 2 F (R) theory of gravity. The method consists of a right
= x8 1 + 7 + 2 dt2 + x d
49 x x 343 form of metric, Eq. (5) and the right form of input func-
1 x4 tions, f (r) or K(r). With this method, the equations of
+ dx2 . (63) motion become remarkably simple and solvable.
49 1 + xc17 + xc22
We have constructed a number of new solutions in
and find they satisfy all the Einstein equations, Eq. (58). terms of known analytic functions. It is found that there
Observing all the solutions presented above, we notice are always two event horizons and one curvature singular-
two remarkable properties in them. In the first place, the ity in the spacetimes. Furthermore, a deficit angle always
scalar field is not asymptotically vanishing. Secondly, the exists in the solutions except for the Schwarzschild-de
spacetime is not asymptotically flat. Hawking has argued Sitter solution. The other interesting points include the
that black holes cannot carry scalar charge, provided that black hole spacetime is always asymptotically non-flat
the scalar field minimally couples to the metric [36]. That and the scalar field always asymptotically infinite. The
is to say scalar field has to be constant in the black hole two points violate the assumptions in the no-hair theo-
spacetime. The conclusion has been generalized to the rem. So the theorem doe not hold for these solutions.
standard scalar-tensor theories [3739]. So the presence Finally, choosing different initial input functions, f (r) or
of the solutions obtained in this paper seems to violate K(r), one is expected to run into other new solutions.
the no-hair theorem. Then is there any inconsistency
with the theorem? The answer is no. We remember
that one of the key assumptions in no-hair theorem is
the spacetime of black holes should be asymptotically Acknowledgments
flat and the scalar field asymptotically vanishing. How-
ever, none of our solutions are asymptotically flat and the We thank the two referees for the valuable suggestions.
scalar field is asymptotically infinite. Actually, Sotiriou The work is supported by the Chinese MoST 863 pro-
and Zhou [40] have pointed out that, in principle, one gram under grant 2012AA121701, the NSFC under grant
scalar field could always have a nontrivial configuration 11373030, 10973014, 11373020 and 11465012.

[1] T. Clifton and J. D. Barrow, Phys. Rev. D 72,
103005 (2005).
[2] L. Sebastiani and S. Zerbini, Eur. Phys. J. C 71,
1591 (2011).
[3] A. Shojai and F. Shojai, Gen. Rel. Grav. 44, 211 (2012).
[4] S. Capozziello, N. Frusciante and D. Vernieri, Gen. Rel.
Grav. 44, 1881 (2012).
[5] T. Multamki and I. Vilja, Phys. Rev. D 74,
064022 (2006).
[6] A. C. Gutirrez-Pieres and C. S. Lopez-Monsalvo, Phys.
Lett. B 718, 1493 (2013).
[7] Z. Amirabi,. M. Halilsoy, and S. Habib Mazharimousavi,
[arXiv:gr-qc/1509.06967].
[8] R. Goswami, A. M. Nzioki, S. D. Maharaj and S. G.
Ghosh, Phys. Rev. D 90, 084011 (2014).
[9] S. H. Hendi, B. Eslam Panah, and S. M. Mousavi, Gen.
Rel. Grav. 44, 835 (2012).
[10] S. E. P. Bergliaffa and Y. E. C. de Oliveira Nunes, Phys.
Rev. D 84, 084006 (2011).
[11] A. de la Cruz-Dombriz, A. Dobado and A. L. Maroto,
Phys. Rev. D 80, 124011 (2009).
[12] J. A. R. Cembranos, A. de la Cruz-Dombriz and
P. J. Romero, Int. J. Geom. Meth. Mod. Phys.
11,1450001(2014).
[13] M. A. Resco, A. de la Cruz-Dombriz, F. J. Llanes
Estrada, V. Z. Castrillo, [arXiv:gr-qc/1602.03880]
[14] F. Lobo and M. Oliveira, Phys. Rev. D 80,
104012 (2009).
[15] T. R. P. Carames and E. R. Bezerra de Mello, Eur. Phys.
J. C 64, 113 (2009).
[16] A. Azadi, D. Momeni, and M. Nouri-Zonoz, Phys. Lett.
B 670, 210 (2008).
[17] S. Capozziello, A. Stabile, and A. Troisi, Class. Quant.
Grav. 25, 085004 (2008).
[18] K. Kainulainen, J. Piilonen, V. Reijonen and D. Sunhede,
Phys. Rev. D 76, 024020 (2007).
[19] G. Cognola, M. Rinaldi, L. Vanzo and S. Zerbini, Phys.
Rev. D 91, 104004 (2015).
[20] S. Nojiri and S. D. Odintsov, Class. Quant. Grav. 30,
125003 (2013).
[21] S. Nojiri and S. D. Odintsov, Phys. Lett. B 735,
376 (2014).
 8

[22] S. H. Hendi, B. Eslam Panah and C. Corda, Can. Journ. 2793 (1983).
Phys. 92(1), 76 (2014). [31] D. Wands, Class. Quant. Grav. 11, 269 (1994).
[23] S. H. Hendi, B. Eslam Panah and R. Saffari, Int. J. Mod. [32] B. Whitt, Phys. Lett. B 145, 176 (1984).
Phys. D 23, 1450088 (2014). [33] J. D. Barrow and S. Cotsakis, Phys. Lett. B 214,
[24] E. Babichev and D. Langlois, Phys. Rev. D 81, 515 (1988).
124051 (2010). [34] K. Maeda, Phys. Rev. D 39, 3159 (1989).
[25] E. Babichev and D. Langlois, Phys. Rev. D 80, [35] G. Magnano and L.M. Sokolowski, Phys. Rev. D 50,
121501 (2009). 5039 (1994).
[26] S. H. Hendi and D. Momeni, Eur. Phys. J. C 71, 1823 [36] S. W. Hawking, Comm. Math. Phys. 25, 167 (1972).
(2011). [37] T. P. Sotiriou and V. Faraoni, Phys. Rev. Lett. 108,
[27] R. Myrzakulov, L. Sebastiani,S. Vagnozzi and S. Zerbini, 081103 (2012).
[arXiv:gr-qc/1510.02284]. [38] J. D. Bekenstein, Phys. Rev. D 51, 6608 (1995).
[28] H. Lu, A. Perkins, C.N. Pope, K.S. Stelle, Phys. Rev. [39] J. D. Bekenstein, arXiv:gr-qc/9605059.
D 92, 124019 (2015). [40] T. P. Sotiriou, S. Y. Zhou, Phys. Rev. Lett. 112,
[29] H. Lu, A. Perkins, C.N. Pope, K.S. Stelle, Phys. Rev. 251102 (2014).
Lett. 114, 171601 (2015).
[30] P. Teyssandier and Ph. Tourrenc, J. Math. Phys. 24,