a

r
X
i
v
:
g
r
-
q
c
/
0
5
0
1
1
0
5
v
1


3
1

J
a
n

2
0
0
5
Self sustained traversable wormholes?
Remo Garattini

Universit`a degli Studi di Bergamo, Facolt`a di Ingegneria,

Viale Marconi 5, 24044 Dalmine (Bergamo) ITALY.
We compute the graviton one loop contribution to a classical energy in a traversable wormhole
background. Such a contribution is evaluated by means of a variational approach with Gaussian
trial wave functionals. A zeta function regularization is involved to handle with divergences. A
renormalization procedure is introduced and the nite one loop energy is considered as a self-
consistent source for the traversable wormhole.
I. INTRODUCTION
A wormhole can be represented by two asymptotically at regions joined by a bridge. To exist, they must satisfy
the Einstein eld equations, like a black hole. While black holes are generally well accepted astrophysical objects,
wormhole are yet to be discovered. One very simple and at the same time fundamental example of wormhole is
represented by the Schwarzschild solution of the Einsteins eld equations. One of the prerogatives of a wormhole
is its ability to connect two distant points in space-time. In this amazing perspective, it is immediate to recognize
the possibility of traveling crossing wormholes as a short-cut in space and time. Unfortunately, although there is no
direct experience, a Schwarzschild wormhole does not possess this property. It is for this reason that in a pioneering
work Morris and Thorne[1] and subsequently Morris, Thorne and Yurtsever[2] studied a class of wormholes termed
traversable. Unfortunately, the traversability is accompanied by unavoidable violations of null energy conditions,
namely, the matter threading the wormholes throat has to be exotic. Classical matter satises the usual energy
conditions. Therefore, it is likely that wormholes must belong to the realm of semiclassical or perhaps a possible
quantum theory of the gravitational eld. Since a complete theory of quantum gravity has yet to come, it is important
to approach this problem semiclassically. On this ground, Hochberg, Popov and Sushkov considered a self-consistent
solution of the semiclassical Einstein equations corresponding to a Lorentzian wormhole coupled to a quantum scalar
eld[3]. On the other hand, Khusnutdinov and Sushkov xed their attention to the computation of the ground state of
a massive scalar eld in a wormhole background. They tried to see if a self-consistent solution restricted to the energy
component appears in this conguration[4]. Motivated by these works, we are interested to repeat the approach of
Ref.[4] but using gravitons instead of scalars. In particular, we wish to study the one loop contribution of the gravitons
to the total energy. This is quite similar to compute the Casimir energy on a xed background. It is known that,
for dierent physical systems, Casimir energy is negative. This is exactly the feature that the exotic matter should
possess. In particular, we conjecture that quantum uctuations can support the traversability as eective source
of the semiclassical Einsteins equations. The rest of the paper is structured as follows, in section II we dene the
eective Einstein equations, in section III we introduce the traversable wormhole metric, in section IV we give some
of the basic rules to perform the functional integration and we dene the Hamiltonian approximated up to second
order, in section V we study the spectrum of the spin-two operator acting on transverse traceless tensors, in section
VI we regularize and renormalize the one loop energy contribution and we speculate about self-consistency of the
result. We summarize and conclude in section VII.
II. THE EFFECTIVE EINSTEIN EQUATIONS
We begin with a look at the classical Einstein equations
G

= T

, (1)
where T

is the stress-energy tensor, G

is the Einstein tensor and = 8G. Consider a separation of the metric

into a background part, g

, and a perturbation, h

,
g

= g

+ h

. (2)

Electronic address: Remo.Garattini@unibg.it

2
The Einstein tensor G

can also be divided into a part describing the curvature due to the background geometry
and that due to the perturbation,
G

(g

) = G

( g

) + G

( g

, h

) , (3)
where, in principle G

( g

, h

) is a perturbation series in terms of h

. In the context of semiclassical gravity,

Eq.(1) becomes
G

= T

ren
, (4)
where T

ren
is the renormalized expectation value of the stress-energy tensor operator of the quantized eld. If the
matter eld source is absent, nothing prevents us from dening an eective stress-energy tensor for the uctuations
as
1
T

ren
=
1

( g

, h

)
ren
. (5)
From this point of view, the equation governing quantum uctuations behaves as a backreaction equation. If we x
our attention to the energy component of the Einstein eld equations, we need to introduce a time-like unit vector u

such that u u = 1. Then the semi-classical Einsteins equations (4) projected on the constant time hypersurface
become
G

( g

) u

= T

ren
= G

( g

, h

) u

ren
. (6)
To further proceed, it is convenient to consider the associated tensor density and integrate over . This leads to
2
1
2
_

d
3
x
_
3
gG

( g

) u

=
_

d
3
xH
(0)
=
1
2
_

d
3
x
_
3
g G

( g

, h

) u

ren
, (7)
where
H
(0)
=
2
_
3
g
G
ijkl

ij

kl

1
2
_
3
gR
(3)
(8)
is the background eld super-hamiltonian and G
ijkl
is the DeWitt super metric. Thus the uctuations in the Einstein
tensor are, in this context, the uctuations of the hamiltonian. To compute the expectation value of the perturbed
Einstein tensor in the transverse-traceless sector, we use a variational procedure with gaussian wave functionals. In
practice, the right hand side of Eq.(7) will be obtained by expanding
E
wormhole
=
|H

|
|
=
_

H
(0)

+ H
(1)

+ H
(2)

+ . . .

_
|
(9)
and retaining only quantum uctuations contributing to the eective stress energy tensor. H
(i)

represents the hamil-

tonian approximated to the i
th
order in h
ij
and is a trial wave functional of the gaussian form. Then Eq.(7)
becomes
H
(0)

=
_

d
3
xH
(0)
=
_

H
(1)

+ H
(2)

+ . . .

_
|
. (10)
The chosen background to compute the quantity contained in Eq.(7) will be that of a traversable wormhole.
1
Note that our approach is very close to the gravitational geon considered by Anderson and Brill[6]. The relevant dierence is in the
averaging procedure.
2
Details on sign conventions and decomposition of the Einstein tensor can be found in Apeendix A
3
III. THE TRAVERSABLE WORMHOLE METRIC
In Schwarzschild coordinates, the traversable wormhole metric can be cast into the form
ds
2
= exp(2(r)) dt
2
+
dr
2
1
b(r)
r
+ r
2
_
d
2
+ sin
2
d
2

. (11)
where (r) is called the redshift function, while b (r) is called the shape function. Proper radial distance is related to
the shape function by
l (r) =
_
r
rt
dr

_
1
b(r

)
r

, (12)
where the plus (minus) sign is related to the upper (lower) part of the wormhole or universe. Two coordinate patches
are required, each one covering the range [r
t
, +). Each patch covers one universe, and the two patches join at r
t
,
the throat of the wormhole dened by
r
t
= min {r (l)} . (13)
Instead of Eq.(11), we consider the special case where 0 and b (r) = r
2
t
/r. Then, we get
ds
2
= dt
2
+
dr
2
1
r
2
t
r
2
+ r
2
_
d
2
+ sin
2
d
2

(14)
and with the help of Eq.(12), we obtain
l (r) =
_
r
rt
dr

_
1
r
2
t
r
2
=
_
r
2
r
2
t
= r
2
= l
2
+ r
2
t
(15)
and Eq.(14) becomes
ds
2
= dt
2
+ dl
2
+
_
r
2
t
+ l
2
_ _
d
2
+ sin
2
d
2

. (16)
The new coordinate l covers the range < l < +. The constant time hypersurface is an Einstein-Rosen bridge
with wormhole topology S
2
R
1
. The Einstein-Rosen bridge denes a bifurcation surface dividing in two parts
denoted by
+
and

. To concretely compute Eq.(10), we consider on the slice deviations from the wormhole
metric of the type
g
ij
= g
ij
+ h
ij
, (17)
where g
ij
is extracted from the line element (16) whose form becomes
ds
2
= dt
2
+ g
ij
dx
i
dx
j
. (18)
IV. ENERGY DENSITY CALCULATION IN SCHR

ODINGER REPRESENTATION
In order to compute the quantity

d
3
x
_
3
g G

( g

, h

) u

ren
, (19)
we consider the right hand side of Eq.(10). Since H
(1)

is linear in h
ij
and h, the corresponding gaussian integral
disappears and since
_
3
g

(g

) u

= H, (20)
4
it is clear that the hamiltonian expansion in Eq.(10) does not coincide with the averaged expanded Einstein tensor of
Eq.(19) because Eq.(20) involves a tensor density. Therefore, the correct setting is
_

d
3
x
_
3
g G

( g

, h

) u

ren
=
_

d
3
x
_
3
g
_

H
(2)

_
3
g
(2)
H
(0)

_
|
, (21)
where
_
3
g
(2)
is the second order expanded tensor density weight. Following the same procedure of Refs.[8, 9], the
potential part of the right hand side of Eq.(21) becomes
_

d
3
x

g
_

1
4
hh +
1
4
h
li
h
li

1
2
h
ij

i
h
l
j
+
1
2
h
l

i
h
li

1
2
h
ij
R
ia
h
a
j
+
1
2
hR
ij
h
ij
+
1
4
Rh
ij
h
ij

1
4
hRh
_
. (22)
The term
_

d
3
x

g
_
1
4
Rh
ij
h
ij

1
4
hRh
_
, (23)
makes the dierence between the hamiltonian expansion and the Einstein tensor expansion. To explicitly make
calculations, we need an orthogonal decomposition for both
ij
and h
ij
to disentangle gauge modes from physical
deformations. We dene the inner product
h, k :=
_

gG
ijkl
h
ij
(x) k
kl
(x) d
3
x, (24)
by means of the inverse WDW metric G
ijkl
, to have a metric on the space of deformations, i.e. a quadratic form on
the tangent space at h, with
G
ijkl
= (g
ik
g
jl
+ g
il
g
jk
2g
ij
g
kl
). (25)
The inverse metric is dened on co-tangent space and it assumes the form
p, q :=
_

gG
ijkl
p
ij
(x) q
kl
(x) d
3
x, (26)
so that
G
ijnm
G
nmkl
=
1
2
_

i
k

j
l
+
i
l

j
k
_
. (27)
Note that in this scheme the inverse metric is actually the WDW metric dened on phase space. The desired
decomposition on the tangent space of 3-metric deformations[10, 11] is:
h
ij
=
1
3
hg
ij
+ (L)
ij
+ h

ij
(28)
where the operator L maps
i
into symmetric tracefree tensors
(L)
ij
=
i

j
+
j

2
3
g
ij
( ) . (29)
Thus the inner product between three-geometries becomes
h, h :=
_

gG
ijkl
h
ij
(x) h
kl
(x) d
3
x =
_

g
_

2
3
h
2
+ (L)
ij
(L)
ij
+ h
ij
h

ij
_
. (30)
5
With the orthogonal decomposition in hand we can dene the trial wave functional
{h
ij
(

x )} = N exp
_

1
4l
2
p
_

hK
1
h
_

x,y
+

(L) K
1
(L)
_

x,y
+

hK
1
h
_
Trace
x,y
_
_
, (31)
where N is a normalization factor. We are interested in perturbations of the physical degrees of freedom. Thus we
x our attention only to the TT tensor sector reducing therefore the previous form into
{h
ij
(

x )} = N exp
_

1
4

hK
1
h
_

x,y
_
. (32)
Therefore to calculate the energy density, we need to know the action of some basic operators on [h
ij
]. The action
of the operator h
ij
on | = [h
ij
] is realized by
h
ij
(x) | = h
ij
(

x ) {h
ij
} . (33)
The action of the operator
ij
on |, in general, is

ij
(x) | = i

h
ij
(

x )
{h
ij
} . (34)
The inner product is dened by the functional integration:

1
|
2
=
_
[Dh
ij
]

1
{h
ij
}
2
{h
kl
} , (35)
and by applying previous functional integration rules, we obtain the expression of the one-loop-like Hamiltonian form
for TT (traceless and transverse) deformations
H

=
1
4
_

d
3
x

gG
ijkl
_
(16G) K
1
(x, x)
ijkl
+
1
(16G)
(
2
)
a
j
K

(x, x)
iakl
_
. (36)
The propagator K

(x, x)
iakl
comes from a functional integration and it can be represented as
K

x ,

y )
iakl
:=

h
()
ia
(

x ) h
()
kl
(

y )
2()
, (37)
where h
()
ia
(

x ) are the eigenfunctions of

2
, whose eigenvalues will be denoted with
2
(). denotes a complete
set of indices and () are a set of variational parameters to be determined by the minimization of Eq.(36). The
expectation value of H

is easily obtained by inserting the form of the propagator into Eq.(36)

E (
i
) =
1
4

i=1
_
(16G)
i
() +

2
i
()
(16G)
i
()
_
. (38)
By minimizing with respect to the variational function
i
(), we obtain the total one loop energy for TT tensors
E
TT
=
1
4

_
_

2
1
() +
_

2
2
()
_
. (39)
The above expression makes sense only for
2
i
() > 0, i = 1, 2. The meaning of
2
i
will be claried in the next section.
Coming back to Eq.(10), we observe that the value of the wormhole energy on the chosen background is
_

d
3
xH
(0)
=
1
16G
_

d
3
x

gR
(3)
=
r
t
2G
. (40)
and to one loop the self-consistent equation for TT tensors becomes
r
t
2G
= E
TT
. (41)
Note that the self-consistency on the hamiltonian as a reversed sign with respect to the energy component of the
Einstein eld equations. This means that an eventual stable point for the hamiltonian is an unstable point for the
eective energy momentum tensor and vice versa.
6
V. THE TRANSVERSE TRACELESS (TT) SPIN 2 OPERATOR FOR THE TRAVERSABLE
WORMHOLE AND THE W.K.B. APPROXIMATION
In this section, we evaluate the one loop energy expressed by Eq.(39). To this purpose, we begin with the operator
describing gravitons propagating on the background (16). The Lichnerowicz operator in this particular metric is
dened by
_

2
h
TT
_
j
i
:=
_

T
h
TT
_
j
i
+ 2
_
Rh
TT
_
j
i
R
_
h
TT
_
j
i
, (42)
where the transverse-traceless (TT) tensor for the quantum uctuation is obtained by the following decomposition
h
j
i
= h
j
i

1
3

j
i
h +
1
3

j
i
h =
_
h
T
_
j
i
+
1
3

j
i
h. (43)
This implies that
_
h
T
_
i
i
= 0. The transversality condition is applied on
_
h
T
_
j
i
and becomes
j
_
h
T
_
j
i
= 0. Thus

T
h
TT
_
j
i
=
S
_
h
TT
_
j
i
+
6
r
2
_
1
b (r)
r
_
, (44)
where
S
is the scalar curved Laplacian, whose form is

S
=
_
1
b (r)
r
_
d
2
dr
2
+
_
4r b

(r) r 3b (r)
2r
2
_
d
dr

L
2
r
2
(45)
and R
a
j
is the mixed Ricci tensor whose components are:
R
a
i
=
_
b

(r)
r
2

b (r)
r
3
,
b

(r)
2r
2
+
b (r)
2r
3
,
b

(r)
2r
2
+
b (r)
2r
3
_
. (46)
The scalar curvature is
R = R
j
i

i
j
= 2
b

(r)
r
2
(47)
We are therefore led to study the following eigenvalue equation
_

2
h
TT
_
j
i
=
2
h
j
i
(48)
where
2
is the eigenvalue of the corresponding equation. In doing so, we follow Regge and Wheeler in analyzing
the equation as modes of denite frequency, angular momentum and parity[12]. In particular, our choice for the
three-dimensional gravitational perturbation is represented by its even-parity form
h
even
ij
(r, , ) = diag
_
H (r)
_
1
b (r)
r
_
1
, r
2
K (r) , r
2
sin
2
L(r)
_
Y
lm
(, ) , (49)
with
_
_
_
H (r) = h
1
1
(r)
1
3
h(r)
K (r) = h
2
2
(r)
1
3
h(r)
L(r) = h
3
3
(r)
1
3
h(r)
. (50)
From the transversality condition we obtain h
2
2
(r) = h
3
3
(r). Then K (r) = L(r). For a generic value of the angular
momentum L, representation (49) joined to Eq.(44) lead to the following system of PDEs
_

_
_

l
+ 2
_
b

(r)
r
2

b(r)
r
3

b

(r)
r
2
__
H (r) =
2
1,l
H (r)
_

l
+ 2
_
b

(r)
2r
2
+
b(r)
2r
3

b

(r)
r
2
__
K (r) =
2
2,l
K (r)
, (51)
7
where
l
is

l
=
_
1
b (r)
r
_
d
2
dr
2
+
_
4r b

(r) r 3b (r)
2r
2
_
d
dr

l (l + 1)
r
2

6
r
2
_
1
b (r)
r
_
. (52)
Dening reduced elds
H (r) =
f
1
(r)
r
; K (r) =
f
2
(r)
r
, (53)
and passing to the proper geodesic distance from the throat of the bridge dened by
dx =
dr
_
1
b(r)
r
, (54)
the system (51) becomes
_

_
_

d
2
dx
2
+ V
1
(x)
_
f
1
(x) =
2
1,l
f
1
(x)
_

d
2
dx
2
+ V
2
(x)
_
f
2
(x) =
2
2,l
f
2
(x)
(55)
with
_

_
V
1
(x) =
l(l+1)
x
2
+r
2
t
+ U
1
(x)
V
2
(x) =
l(l+1)
x
2
+r
2
t
+ U
2
(x)
, (56)
and
_

_
U
1
(x) =
_
6
(x
2
+r
2
t
)
2
x
2

r
2
t
(x
2
+r
2
t
)
2
_
U
2
(x) =
_
6
(x
2
+r
2
t
)
2
x
2
+
3r
2
t
(x
2
+r
2
t
)
2
_
. (57)
In order to use the WKB approximation, we dene two r-dependent radial wave numbers k
1
(x, l,
1,nl
) and
k
2
(x, l,
2,nl
)
_

_
k
2
1
(x, l,
1,nl
) =
2
1,nl

l(l+1)
(x
2
+r
2
t
)
U
1
(x)
k
2
2
(x, l,
2,nl
) =
2
2,nl

l(l+1)
(x
2
+r
2
t
)
U
2
(x)
. (58)
The number of modes with frequency less than
i
, i = 1, 2, is given approximately by
g (
i
) =
_

i
(l,
i
) (2l + 1) dl, (59)
where
i
(l,
i
), i = 1, 2 is the number of nodes in the mode with (l,
i
), such that

i
(l,
i
) =
1

_
+

dx
_
k
2
i
(x, l,
i
). (60)
Here it is understood that the integration with respect to x and l is taken over those values which satisfy k
2
i
(x, l,
i
)
0, i = 1, 2. Thus the total one loop energy for TT tensors is given by
E
TT
=
1
4
2

i=1
_
+
0

i
d g (
i
)
d
i
d
i
=
2

i=1
_
+

dx
_
x
2
+ r
2
t
_
_
1
4
2
_
+

Ui(x)

2
i
_

2
i
U
i
(x)d
i
_
8
=
_
+

dx
_
x
2
+ r
2
t
_
[
1
+
2
] , (61)
where
_

1
=
1
4
2
_
+

U1(x)

2
1
_

2
1
U
1
(x)d
1

2
=
1
4
2
_
+

U2(x)

2
2
_

2
2
U
2
(x)d
2
. (62)
VI. ONE LOOP ENERGY REGULARIZATION AND RENORMALIZATION
In this section, we proceed to evaluate the one loop energy. The method is equivalent to the scattering phase shift
method and to the same method used to compute the entropy in the brick wall model. We use the zeta function
regularization method to compute the energy densities
1
and
2
. Note that this procedure is completely equivalent to
the subtraction procedure of the Casimir energy computation where zero point energy (ZPE) in dierent backgrounds
with the same asymptotic properties is involved. To this purpose, we introduce the additional mass parameter in
order to restore the correct dimension for the regularized quantities. Such an arbitrary mass scale emerges unavoidably
in any regularization schemes. Then we have

i
() =
1
4
2

2
_
+

Ui(x)
d
i

2
i
(
2
i
U
i
(x))

1
2
(63)
If one of the functions U
i
(x) is negative, then the integration has to be meant in the range where
2
i
+U
i
(x) 0. In
both cases the result of the integration is
3
=
U
2
i
(x)
64
2
_
1

+ ln
_

2
U
i
(x)
_
+ 2 ln 2
1
2
_
, (64)
where the absolute value has been inserted to take account of the possible change of sign. Then the total regularized
one loop energy is
E
TT
(r
t
, ; ) =
_
+

dx
_
x
2
+ r
2
t
_
[(
1
() +
2
())] . (65)
The result of the integration over the x coordinate leads to the following expression
E
TT
(r
t
, ; ) =
1
16
_

a
r
t

b
r
t

c
r
t
ln (r
t
)
_
, (66)
where a = 23.35740218, b = 276.6026775 and c = 212.0575042. Then the self consistent equation (41) can be written
in the form
r
t
2G
=
1
16
_
a
r
t
+
b
r
t
+
c
r
t
ln (r
t
)
_
. (67)
In order to deal with nite quantities, we renormalize the divergent energy by absorbing the singularity in the classical
quantity. In particular, we re-dene the bare classical constant G
1
G

1
G
0
+
a
8
2
r
2
t
. (68)
Therefore, the remaining nite value for the eective equation (67) reads
r
t
2G
0
=
1
16
_
b
r
t
+
c
r
t
ln (r
t
)
_
. (69)
3
Details of the calculation can be found in the Appendix B.
9
This quantity depends on the arbitrary mass scale . It is appropriate to use the renormalization group equation to
eliminate such a dependence. To this aim, we impose that[14]

d
d
_
r
t
2G
0
()
_
=
d
d
_
1
16
_
b
r
t
+
c
r
t
ln (r
t
)
__
, (70)
namely
r
t
2

G
1
0
()

c
16r
t
= 0. (71)
Solving it we nd that the renormalized constant G
0
should be treated as a running one in the sense that it varies
provided that the scale is changing
1
G
0
()
=
1
G
0
(
0
)
+ K ln
_

0
_
or G
0
() =
G
0
(
0
)
1 + G
0
(
0
) K ln
_

0
_, K =
c
8
2
r
2
t
; (72)
where
0
is the normalization point. We substitute Eq.(72) into Eq.(69) to nd

2G
0
(
0
)
=
1
16
_
b
r
2
t
+
c
r
2
t
ln (r
t

0
)
_
, (73)
where we have divided by r
t
. In order to have only one solution
4
, we nd the extremum of the r.h.s. of Eq.(73) and
we get
c 2b
2c
= ln ( r
t

0
) = r
t
=
1

0
exp
_
c 2b
2c
_
(74)
and
1
G
0
(
0
)
=
c
2
0
16
2
exp
_

c 2b
c
_
. (75)
With the help of Eqs.(74) and (75), Eq.(72) becomes
1
G
0
()
=
1
G
0
(
0
)
+
c
8
2
r
2
t
ln
_

0
_
=
c
2
0
16
2
exp
_

c 2b
c
__
1 + 2 ln
_

0
__
=
1
G
0
(
0
)
_
1 + 2 ln
_

0
__
. (76)
It is straightforward to see that we have a constraint on /
0
. Indeed we have to choose
>
0
exp
_

1
2
_
= .6065306597
0
, (77)
otherwise G
0
() becomes negative. We have now two possibilities:
1. we identify G
0
(
0
) with the squared Planck length. As a consequence we obtain
r
t
=
1
4
_
cG
0
(
0
) = 1.158822606l
p
=
0
= .3860531213m
p
. (78)
4
Note that in the paper of Khusnutdinov and Sushkov[4], to nd only one solution, the minimum of the ground state of the quantized
scalar eld has been set equal to the classical energy. In our case, we have no external elds on a given background. This means that it
is not possible to nd a minimum of the one loop gravitons, in analogy with Ref.[4]. Moreover the renormalization procedure in Ref.[4]
is completely independent by the classical term, while in our case it is not. Indeed, thanks to the self-consistent equation (41), we can
renormalize the divergent term.
10
2. We identify
0
with the Planck scale and we get
r
t
= .4473670842l
p
=
_
G
0
(
0
) = .3860531213l
p
. (79)
If we compare our result with that obtained by Khusnutdinov and Sushkov[4], we see that in their work they
nd that the wormhole radius is
r
w
.0141l
p
(80)
and its mass
m
w
11.35m
p
(81)
is trans-planckian. We recall that these values have been obtained by looking at a massive scalar eld quantized
in a wormhole background playing the role of the exotic matter. In our case, we suppose that the graviton
quantum uctuations play the role of the exotic matter and even if we x the renormalization point at the
Planck scale as in Ref.[4], we nd a radius r
t
> r
w
.
VII. SUMMARY AND CONCLUSIONS
Motivated by the works of Anderson and Brill, in searching for gravitational geons[6] and of Khusnutdinov and
Sushkov[4], we have considered the possible existence of a self-consistent solution of the semiclassical Einsteins
equations in a traversable wormhole background. In particular, we have xed our attention to the graviton quantum
uctuations around such a background. The uctuations, contained in the perturbed Einstein tensor, play the role
of the exotic matter considered in Ref.[4]. A variational approach with the help of gaussian trial wave functionals
has been used to compute the one loop term. The adopted procedure is quite close to that of Anderson and Brill
except for the averaging process which, in our case, involves a variational calculation. To handle the divergencies we
have used a zeta function regularization which, in this context, is formally equivalent to a Casimir energy subtraction
procedure. A renormalization of the Newtons constant G has been performed to absorb the ultraviolet divergencies.
To avoid dependences on the renormalization scale a renormalization group equation has been computed. While the
renormalization process is not new in the context of semiclassical Einstein eld equations, to our knowledge it is the
use of the renormalization group equation that seems to be unknown, especially concerning self-consistent solutions
and traversability. The interesting point is the obtained consistency with a wormhole radius r
t
greater than radius
r
w
of Eq.(80). It is likely that this result be correlated with the graviton themselves. Despite of this, the obtained
traversability has to be regarded as in principle rather than in practice because of the obtained wormhole
radius size. Nevertheless, the calculation is far to be complete because only gravitons have been taken into account
and the traversable wormhole metric considered is restricted to Eq.(14). A possible improvement of our evaluation
could come from including a charge[15] or by looking at a Schwarzschild-like metric, but even with a dierent choice
of the wave functionals entering in the variational approach.
APPENDIX A: EINSTEIN EQUATIONS AND THE HAMILTONIAN
Let us consider the Einstein equations
G

= R

1
2
g

R = 8GT

. (A1)
R

is the Ricci tensor and R is the scalar curvature. If u

is a time-like unit vector such that g

= 1,then
the Einstein tensor G

becomes
G

= R

1
2
g

R = R

+
1
2
R. (A2)
By means of the Gauss-Codazzi equations[16],
R = R
(3)
2R

K
2
K

, (A3)
where K

is the extrinsic curvature and R

(3)
is the three dimensional scalar curvature. For a time-like vector, we
take the lower sign and Eq.(A2) becomes
G

=
1
2
_
R
(3)
+ K
2
K

_
. (A4)
11
If the conjugate momentum is dened by

=
_
(3)
g
2
(Kg

) , (A5)
then
K
2
K

=
_
2
_
(3)
g
_
2 _

2
2

_
(A6)
and
_
(3)
g
2
G

=
_
(3)
g
2
R
(3)
+
2
_
(3)
g
_

2
2

_
= H
(0)
, (A7)
namely Eq.(8) with the reversed sign.
APPENDIX B: THE ZETA FUNCTION REGULARIZATION
In this appendix, we report details on computation leading to expression (63). We begin with the following integral
() =
_

_
I
+
=
2
_
+
0
d

2
(
2
+U(x))

1
2
I

=
2
_
+
0
d

2
(
2
U(x))

1
2
, (B1)
with U (x) > 0.
1. I+ computation
If we dene t = /
_
U (x), the integral I
+
in Eq.(B1) becomes
() =
2
U (x)
2
_
+
0
dt
t
2
(t
2
+ 1)

1
2
=
1
2

2
U (x)
2
B
_
3
2
, 2
_
1
2

2
U (x)
2

_
3
2
_
( 2)

_

1
2
_ =

4
U (x)
2
_

2
U (x)
_

( 2)

_

1
2
_, (B2)
where we have used the following identities involving the beta function
B (x, y) = 2
_
+
0
dt
t
2x1
(t
2
+ 1)
x+y
Re x > 0, Re y > 0 (B3)
related to the gamma function by means of
B (x, y) =
(x) (y)
(x + y)
. (B4)
Taking into account the following relations for the -function
( 2) =
(1 + )
( 1) ( 2)
,
_

1
2
_
=

_
+
1
2
_

1
2
, (B5)
and the expansion for small
(1 + ) = 1 + O
_

2
_
,
_
+
1
2
_
=
_
1
2
_

_
1
2
_
( + 2 ln2) + O
_

2
_
12
x

= 1 + lnx + O
_

2
_
, (B6)
where is the Eulers constant, we nd
() =
U
2
(x)
16
_
1

+ ln
_

2
U (x)
_
+ 2 ln2
1
2
_
. (B7)
2. I computation
If we dene t = /
_
U (x), the integral I

in Eq.(B1) becomes
() =
2
U (x)
2
_
+
0
dt
t
2
(t
2
1)

1
2
=
1
2

2
U (x)
2
B
_
2,
3
2

_
1
2

2
U (x)
2

_
3
2

_
( 2)

1
2
_ =
1
4

U (x)
2
_

2
U (x)
_

_
3
2

_
( 2) , (B8)
where we have used the following identity involving the beta function
1
p
B
_
1

p
,
_
=
_
+
1
dtt
1
(t
p
1)
1
p > 0, Re > 0, Re < p p Re (B9)
and the reection formula
(z) (1 z) = z(z) (z) (B10)
From the rst of Eqs.(B5) and from the expansion for small

_
3
2

_
=
_
3
2
_
(1 ( 2 ln 2 + 2)) + O
_

2
_
x

= 1 + lnx + O
_

2
_
, (B11)
we nd
() =
U
2
(x)
16
_
1

+ ln
_

2
U (x)
_
+ 2 ln2
1
2
_
. (B12)
[1] M. S. Morris and K. S. Thorne, Am. J. Phys. 56, 395 (1988).
[2] M. S. Morris, K. S. Thorne, and U. Yurtsever, Phys. Rev. Lett. 61, 1446 (1988).
[3] D. Hochberg, A. Popov and S. V. Sushkov, Phys. Rev. Lett. 78 (1997) 2050, gr-qc/9701064.
[4] N. R. Khusnutdinov and S. V. Sushkov, Phys. Rev. D 65 084028, (2002), hep-th/0202068.
[5] M. Visser, Lorentzian Wormholes (AIP Press, New York, 1995) 64.
[6] P. R. Anderson and D. R. Brill, Phys.Rev. D 56 (1997) 4824, gr-qc/9610074 .
[7] R. Arnowitt, S. Deser, and C. W. Misner, in Gravitation: An Introduction to Current Research edited by L. Witten (John
Wiley & Sons, Inc., New York, 1962).
[8] R. Garattini, Phys.Rev. D 59 104019, (1999), hep-th/9902006.
[9] R. Garattini, Int. J. Mod. Phys. A 18 (1999) 2905, gr-qc/9805096.
[10] M. Berger and D. Ebin, J. Di. Geom. 3, 379 (1969).
[11] J. W. York Jr., J. Math. Phys., 14, 4 (1973), Ann. Inst. Henri Poincare A 21 (1974) 319.
[12] T. Regge and J. A. Wheeler, Phys. Rev. 108, 1063 (1957).
[13] M. Bordag, U. Mohideen and V.M. Mostepanenko, Phys. Rep. 353, 1 (2001).
[14] I.O. Cherednikov, Acta Physica Slovaca, 52, (2002), 221.
[15] S. W. Kim and H. Lee, Phys. Rev. D 63 064014, (2001), gr-qc/0102077; S. W. Kim, Scattering Problem of Scalar Wave
in Wormhole Geometry, gr-qc/0102078.
[16] S.W. Hawking and G.F.R. Ellis, The Large Scale Structure of Spacetime (Cambridge Univ. Press, 1973).
[17] I.S. Gradshteyn and I.M. Ryzhik, Table of Integrals, Series, and Products (corrected and enlarged edition), edited by A.
Jerey (Academic Press, Inc.).