Gen Relativ Gravit (2011) 43:2631–2647

DOI 10.1007/s10714-011-1185-7

RESEARCH ARTICLE

Quasinormal frequencies of asymptotically

flat two-dimensional black holes

A. López-Ortega · I. Vega-Acevedo

Received: 15 November 2010 / Accepted: 14 April 2011 / Published online: 1 May 2011
© Springer Science+Business Media, LLC 2011

Abstract We discuss whether the minimally coupled massless Klein-Gordon and

Dirac fields have well defined quasinormal modes in single horizon, asymptotically
flat two-dimensional black holes. To get the result we solve the equations of motion
in the massless limit and we also calculate the effective potentials of Schrödinger type
equations. Furthermore we calculate exactly the quasinormal frequencies of the Dirac
field propagating in the two-dimensional uncharged Witten black hole. We compare
our results on its quasinormal frequencies with other already published.

Keywords Quasinormal modes · Two-dimensional black holes ·

Witten black hole

1 Introduction

Two-dimensional gravity theories are widely studied [1]. Although the simplifica-
tions in two-dimensional gravitational models remove some relevant features of
higher dimensional gravitational systems, these also allow more detailed analysis of
some physical phenomena. Furthermore these simplifications allow to address some

A. López-Ortega (B)
Centro de Investigación en Ciencia Aplicada y Tecnología Avanzada, Unidad Legaria,
Instituto Politécnico Nacional, Calzada Legaria # 694,
Colonia Irrigación, Delegación Miguel Hidalgo, 11500 Mexico, D.F., Mexico
e-mail: [email protected]

I. Vega-Acevedo
Departamento de Física, Escuela Superior de Física y Matemáticas, Instituto Politécnico Nacional,
Unidad Profesional Adolfo López Mateos, Edificio 9, 07738 Mexico, D.F., Mexico
e-mail: [email protected]

123
2632 A. López-Ortega, I. Vega-Acevedo

conceptual problems in a simple framework [1,2]. Thus among the reasons for ana-
lyzing two-dimensional gravity models we enumerate:
i) The models of two-dimensional gravity appear as effective gravity theories in
string theory, in spherically symmetric or axial reduction of higher dimensional
gravitational systems, and as a limit case in other physical systems [1].
ii) These two-dimensional models reproduce qualitatively many features of the phys-
ical phenomena of higher dimensional systems as Hawking evaporation, black
hole thermodynamics, and critical collapse [1,2].
iii) Two-dimensional gravity is an appropriate framework for testing some ideas about
quantum gravity [1].
Two-dimensional gravity theories have black hole solutions [1] and these two-
dimensional solutions are useful to test some ideas on how to solve several problems
that we find in the physics of black holes, as the origin of their entropies and the
information loss problem [1]. Taking into account the usefulness of these two-dimen-
sional black hole solutions, it is appropriate to study other aspects of their behavior,
for example, how these two-dimensional black holes react when we slightly perturb
them with a test field.
It is well known that when we perturb a black hole with a test field, it oscillates with
well defined modes that satisfy the appropriate radiation boundary conditions at the
horizon and at the asymptotic region. These decaying modes are called quasinormal
modes (QNM in what follows) and their complex frequencies are named quasinormal
frequencies (QNF in what follows) [3,4].
Near the horizon the QNM boundary condition usually imposed is that the field
must be purely ingoing. This condition arises from the fact that classical black holes
do not emit radiation. Far from the event horizon the QNM boundary condition that the
field must satisfy depends on the asymptotic structure of the black hole. For example,
for asymptotically flat black holes we impose that the field must be purely outgoing
far from the event horizon, whereas for asymptotically anti-de Sitter black holes we
usually impose that the field (or its flux) goes to zero at infinity [3,4].
The computations of the QNF are useful to know some properties of the black hole
as for example its classical stability [3,4] and the determination of its entropy quanta
[5–7]. Furthermore since the QNF depend on the physical parameters of the black hole
as its mass, electric charge, and angular momentum, they give us a tool for determin-
ing these physical parameters. Considering the relevance of the two-dimensional black
holes [1,2], we think that an useful addition to the known results is the calculation of
their QNF.
Taking into account the simplicity of lower dimensional systems we expect that for
some lower dimensional black holes their QNF can be calculated exactly. Recently
appeared several examples of lower dimensional black holes whose QNF are exactly
computed, for example, we are aware of the gravitational systems studied in [8–26].
In these references are calculated exactly the QNF for the three-dimensional BTZ
black hole [8–13], for the three-dimensional dilatonic black holes [14–17], for the
three-dimensional warped AdS(3) black holes [18–20], for the three-dimensional de
Sitter spacetime [21,22], for the two-dimensional de Sitter spacetime [23], and for the
two-dimensional Witten black hole [23–26].

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2633

Furthermore in two dimensions we expect to find relevant results for a family of

black hole solutions. For example, in [23] are calculated exactly the QNF for a family
of two-dimensional black holes. We also note that in [27] Kettner et al. compute
the asymptotic QNF of the Klein-Gordon field non-minimally coupled to the dilaton
propagating in a family of black holes that are solutions to the equations of motion
obtained from actions with power law potentials [28].
Here we investigate whether the massless Klein-Gordon and Dirac fields have well
defined QNF in single horizon, asymptotically flat two-dimensional black holes. Fur-
thermore we calculate exactly the QNF of a test Dirac field in Witten black hole. The
results about the QNF of the Dirac field in Witten black hole extend those obtained
for the Klein-Gordon field [24,25] and the massless Dirac field [26].
This paper is organized as follows. In Sect. 2 we analyze whether the massless
Klein-Gordon and Dirac equations have solutions that satisfy the boundary conditions
of the QNM for single horizon, asymptotically flat two dimensional black holes. For
both fields we also calculate the effective potentials of Schrödinger type equations
and use these potentials to discuss the existence of their QNF. In Sect. 3 we exactly
calculate the QNF of the Dirac field propagating in the two-dimensional Witten black
hole. We also compare our results with other already published on the QNF of this
two-dimensional black hole. Finally in Sect. 4 we summarize and discuss the obtained
results.

2 Quasinormal modes of massless fields in two-dimensional black holes

For a generic single horizon, asymptotically flat two-dimensional black hole in this
Section we study whether the massless Klein-Gordon and Dirac equations have solu-
tions that satisfy the boundary conditions of the QNM. In what follows we consider
two-dimensional black holes with line element

dr 2
ds 2 = f dt 2 − , (1)
g

with f and g being functions of the r coordinate. We assume that the functions f
and g allow the existence of an event horizon at r = r H and when r → +∞ the
spacetime is asymptotically flat.
Furthermore we define the appropriate tortoise coordinate x for spacetime (1) by

dr
x= √ , (2)
fg

and we assume that it satisfies

lim x → −∞, lim x → +∞, (3)

r →r H r →+∞

thus x ∈ (−∞, +∞) as r ∈ (r H , +∞).

An example of black hole that satisfies the conditions that we have imposed on the
background is the two-dimensional Schwarzschild black hole for which

123
2634 A. López-Ortega, I. Vega-Acevedo

2M
f (r ) = g(r ) = 1 − , (4)
r
where r is the radial coordinate, M is related to the mass of the black hole, and it has
an event horizon at r H = 2M. From formula (2) we find that for the two-dimensional
Schwarzschild black hole the tortoise coordinate is equal to

x = r + 2M ln(r − 2M), (5)

and we note that it satisfies conditions (3).

For these black holes we define the QNM of test fields as the oscillations that satisfy
the boundary conditions:
a) The oscillations are purely ingoing near the event horizon.
b) The oscillations are purely outgoing at infinity.
Let us begin with the minimally coupled Klein-Gordon field that for mass different
from zero satisfies the equation
 
 + m 2 Φ = 0, (6)

where  denotes the two-dimensional d’Alembertian and m is the mass of the Klein-
Gordon field. For solving the Klein-Gordon equation in the two-dimensional metric
(1) we propose a separable solution of the form

Φ(r, t) = R(r )e−iωt , (7)

to find that Eq. (6) reduces to the radial differential equation

  
ω2 g d  dR
R+ gf − m 2 R = 0. (8)
f f dr dr

Using the tortoise coordinate (2) for the line element (1) we get that in the massless
limit Eq. (8) may be written as

d2 R
+ ω2 R = 0, (9)
dx 2
whose solutions are

R = Aeiωx + Be−iωx , (10)

where A and B are constants.

If the tortoise coordinate satisfies the previously enumerated conditions in formu-
las (3), then the first term in (10) is an outgoing wave near the event horizon and at
infinity. The second term in formula (10) is an ingoing wave near the event horizon
and at infinity. Thus the solutions (10) can not satisfy the boundary conditions of

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2635

the QNMs for asymptotically flat black holes and we obtain that the massless Klein-
Gordon field does not have well defined QNM in a single horizon, asymptotically flat
two-dimensional black hole.
Now we study the massless Dirac field. We first note that in metric (1) the Dirac
equation reduces to a pair of coupled partial differential equations and that these
equations decouple in the massless limit. To see this fact we write the Dirac equation

i/∇ Ψ = mΨ, (11)

in the two-dimensional metric (1) and use the method of Ref. [29] to reduce the Dirac
equation to ordinary differential equations. As usual, Ψ is a two-spinor, ∇
/ denotes the
Dirac operator, and m is the mass of the fermion field. The Dirac operator is given
by ∇/ = γ μ ∇μ where ∇μ denotes the covariant derivative, the matrices γ μ satisfy
γ γ ν + γ ν γ μ = 2g μν , and in two dimensions they are square 2 × 2 matrices.
μ

See [26,30,31] for other studies on the Dirac field propagating in two-dimensional
spacetimes.
To simplify the Dirac equation first we factor out the function f in metric (1). Thus
 
dr 2
ds 2 = f dt 2 − = f ds̃ 2 , (12)
fg

where

dr 2
ds̃ 2 = dt 2 − . (13)
fg

Note that under a conformal transformation of metric (1), that is

gμν = f g̃μν , (14)

if we change the two spinor, the Dirac operator, and the mass in the form

1 1 1
Ψ = Ψ̃ , ∇
/Ψ = ∇
/̃ Ψ̃ , m = m̃, (15)
f 1/4 f 3/4 f 1/2

then the quantities Ψ̃ , ∇

/̃ Ψ̃ , and m̃ satisfy the Dirac equation in spacetime (13) [29].
Using the tortoise coordinate (2) we get that metric (13) takes the form

ds̃ 2 = dt 2 − dx 2 , (16)

and exploiting the chiral representation for the γ μ matrices

   
0 1 0 −1
γ0 = γt = , γ1 = γr = , (17)
1 0 1 0

123
2636 A. López-Ortega, I. Vega-Acevedo

we find that the Dirac equation reduces to the pair of coupled partial differential
equations

∂t Ψ̃2 − ∂x Ψ̃2 = −im f Ψ̃1 ,

∂t Ψ̃1 + ∂x Ψ̃1 = −im f Ψ̃2 , (18)

for the components Ψ̃1 and Ψ̃2 of the two-dimensional spinor

 
Ψ̃1
Ψ̃ = . (19)
Ψ̃2

In the massless limit the previous system of coupled equations simplifies to the
decoupled equations

∂t Ψ̃2 − ∂x Ψ̃2 = 0,
∂t Ψ̃1 + ∂x Ψ̃1 = 0, (20)

or written in an equivalent form


∂t Ψ̃2 − f g∂r Ψ̃2 = 0,

∂t Ψ̃1 + f g∂r Ψ̃1 = 0. (21)

Taking the components Ψ̃1 and Ψ̃2 of the two spinor as

Ψ̃s (x, t) = Rs (x)e−iωt , s = 1, 2, (22)

we obtain that Eqs. (20) transform into

dR2
+ iω R2 = 0,
dx
dR1
− iω R1 = 0, (23)
dx
whose solutions are

R1 = Aeiωx , R2 = Be−iωx , (24)

with A and B constants.

Thus in the massless limit the function R1 represents a purely outgoing wave near
the horizon and far from the horizon, in a similar way the function R2 represents a
purely ingoing wave near the horizon and far from the horizon. Therefore the mass-
less Dirac field can not satisfy the boundary conditions for the QNM of generic single
horizon, asymptotically flat two-dimensional black holes.
To get these results we only assume that the two-dimensional black holes (1) have
a single horizon, they are asymptotically flat, and their tortoise coordinates (2) satisfy
conditions (3).

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2637

The previous results show that in single horizon, asymptotically flat two dimensional
black holes, the massless Klein-Gordon and Dirac fields do not have modes that satisfy
the boundary conditions of the QNM. Notice that the previous result is valid only for
the QNM that satisfy the conditions a) and b) and that for the two dimensional black
holes whose QNM satisfy other boundary conditions, we may find that the massless
Klein-Gordon and Dirac fields have well defined QNF frequencies. Furthermore this
result points out that in two dimensional black holes it is necessary to analyze in detail
the process of calculation for the QNF of the massive and massless fields, because it
is possible that the QNF of the massless field can not be obtained from the QNF of
the massive field by a limiting procedure.
We notice that an equivalent result is obtained for the Dirac and Klein-Gordon
fields propagating in the three-dimensional extreme BTZ black hole [13]. Moreover a
similar result is true for the Klein-Gordon field that propagates in a five-dimensional
dilatonic black hole. For this spacetime in Ref. [25] it is shown that for the s wave of the
massless Klein-Gordon field we can not satisfy the boundary conditions of the QNM.
For single horizon, asymptotically flat two-dimensional black holes that are solu-
tions to the equations of motion derived from actions with power law potentials [28],
in Ref. [27] Kettner et al. calculate the asymptotic QNF of the massless Klein-Gordon
field. They find well defined asymptotic QNF and their conclusions do not contradict
the previous results because in [27] they calculate the asymptotic QNF of the massless
Klein-Gordon field non-minimally coupled to the dilaton field and here we analyze
the minimally coupled Klein-Gordon field.
To support our previous results, in what follows we present the effective potentials
[32] for the radial equations of the Klein-Gordon and Dirac fields propagating in two
dimensional spacetime with metric (1). For the massive Klein-Gordon field we note
that the radial differential equation (8) may be transformed into

d2 R
+ ω2 R = V R, (25)
dx 2

where the effective potential V is equal to

V = m 2 f. (26)

For the Dirac field, if we make the ansatz (22), then we find that the system of
coupled differential equations (18) reduces to

dR2 
+ iω R2 = i f m R1 ,
dx
dR1 
− iω R1 = −i f m R2 . (27)
dx

Changing the functions R1 and R2 by

R1 = eiπ/4 R̃1 , R2 = e−iπ/4 R̃2 , (28)

123
2638 A. López-Ortega, I. Vega-Acevedo

we obtain that the new functions R̃1 and R̃2 satisfy the coupled differential equations

d R̃2 
+ iω R̃2 = − f m R̃1 ,
dx
d R̃1 
− iω R̃1 = − f m R̃2 . (29)
dx

Following Chandrasekhar [32] we define


W = −m f, Z ± = R̃1 ± R̃2 , (30)

to find that the functions Z ± satisfy Schrödinger type equations

d2 Z ±
+ ω2 Z ± = V± Z ± , (31)
dx 2

where

dW m√ df m√ df
V± = W 2 ± = m2 f ∓ g =V∓ g . (32)
dx 2 dr 2 dr

Notice that for the massive Klein-Gordon field the effective potential (26) does not
depend on the function g which appears in the black hole metric (1), in contrast, for
the Dirac field the effective potentials (32) depend on the function g. We also note
that in the massless limit the effective potentials (26) and (32) go to zero. Thus in
metric (1) the radial functions of the massless Klein-Gordon and Dirac fields satisfy
free Schrödinger type equations. This fact is consistent with the absence of QNM for
the massless Klein-Gordon and Dirac field in single horizon, asymptotically flat two
dimensional black holes. Furthermore when in [27] the coupling function between the
Klein-Gordon field and the dilaton is a constant, we get that in radial Schrödinger type
equation the effective potential goes to zero.
Finally, from the coupled equations (27) for the Dirac field we can get decoupled
equations for the radial functions R1 and R2 . The decoupled equations can be written
in a similar form
 √   
d2 Rs √ d g dRs √ d 1 ω 2 Rs
g + g ± iω g √ Rs + − m 2 Rs = 0, (33)
dr 2 dr dr dr f f

with s = 1, 2 and take the upper sign for R2 and the lower sign for R1 . As far as we
know the results of this Section have not been published.

3 Quasinormal frequencies of the Dirac field in Witten black hole

A well studied two-dimensional spacetime is the uncharged Witten black hole

[1,33–35], whose metric and dilaton are

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2639

dr 2 r
ds 2 = 1 − e−r dt 2 − , φ = φ0 − , (34)
1 − e−r 2

where φ0 is a constant.1 This black hole is asymptotically flat and it has a horizon at
r H = 0 [33,34].
The propagation of test fields in the two-dimensional black hole (34) is analyzed in
several papers, for example [23–26,36–39], mainly to study its classical and quantum
stability. As the Witten black hole has a single horizon and is asymptotically flat, its
QNM are the oscillations that satisfy the boundary conditions a) and b) of the previous
section, (but see formula (59) below).
In Witten black hole for the coupled to scalar curvature massive Klein-Gordon field
its QNF are calculated by Becar et al. [24], but as noted in [25], there are additional
QNF to those found in [24] (also these can be written in a simpler form [25]). The
QNF of the non-minimally coupled to scalar curvature massive Klein-Gordon field
are equal to [24,25]

i  4m 2
ω=− 2n + 1 ± 1 − 4ζ − √ , (35)
4 2n + 1 ± 1 − 4ζ

where ζ is the coupling constant to scalar curvature. For the Witten black hole the
QNF of the massless Dirac field are calculated in [26], here we calculate the QNF
of the Dirac field and comment on the existence of the QNF for the massless Dirac
field at the end of this Section. Furthermore notice that in [40] the QNF of the charged
Witten black hole [35] are calculated using the WKB method.
The metric of the Witten black hole takes the form (1) with

f (r ) = g(r ) = 1 − e−r , (36)

therefore proposing separable solutions as those of formula (22), from (18) we find
that in Witten black hole the Dirac equation simplifies to

dR2 
(1 − e−r ) + iω R2 = im 1 − e−r R1 ,
dr
dR 
(1 − e−r )
1
− iω R1 = −im 1 − e−r R2 . (37)
dr

From this system of coupled equations for the radial functions R1 and R2 we can
get the decoupled equations (see also Eq. (33))

d2 R s e−r dRs iωνe−r ω2

(1 − e−r ) + − R s + Rs − m 2 Rs = 0, (38)
dr 2 2 dr 2(1 − e−r ) 1 − e−r

1 Sometimes this black hole is called CGHS dilaton black hole [23].

123
2640 A. López-Ortega, I. Vega-Acevedo

where

1 for s = 2,
ν= (39)
−1 for s = 1.

As for the Klein-Gordon field [24,38] in Eq. (38) we make the change of variable

z = 1 − e−r , (40)

to find that these equations transform into

 
d2 Rs 1 1 dRs iων Rs ω 2 Rs m 2 Rs
+ − − 2 + − = 0. (41)
dz 2 2z 1−z dz 2z (1 − z) z 2 (1 − z)2 z (1 − z)2

To solve these differential equations we propose that the functions Rs take the form

Rs (z) = z αs (1 − z)βs Fs (z). (42)

If the parameters αs and βs satisfy the equations

αs iων
αs2 − + ω2 − = 0, βs2 + ω2 − m 2 = 0, (43)
2 2

thus
⎧ ⎧√
⎨ iων + 21 , ⎨ m 2 − ω2 ,
αs = βs = (44)
⎩ ⎩ √ 2
−iων, − m − ω2 ,

then the functions Fs are solutions of the differential equations

   
d2 Fs 1 3 dFs
z(1 − z) 2 + 2αs + − 2αs + 2βs + z
dz 2 2 dz
 
iων βs
+ 2ω2 − αs − − m 2 − 2βs αs − Fs = 0. (45)
2 2

We can transform the previous equations to the hypergeometric form [41,42]

d2 F dF
z(1 − z) 2
+ (c − (a + b + 1)z) − abF = 0, (46)
dz dz

with parameters

as = αs + βs + 21 , bs = αs + βs , cs = 2αs + 21 . (47)

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2641

In what follows we study in detail the component Ψ̃2 (ν = 1) and choose the values
of the parameters α2 and β2 as

α2 = iω + 21 , β2 = m 2 − ω2 . (48)

If the parameter c2 is not integer, then the solutions of Eq. (45) with s = 2 are [41,42]

F2 (z) = C̃1 2 F1 (a2 , b2 ; c2 ; z) + C̃2 z 1−c2 2 F1 (a2 − c2 + 1, b2 − c2 + 1; 2 − c2 ; z),

(49)

with C̃1 , C̃2 constants, 2 F1 (a, b; c; z) is the hypergeometric function and hence the
radial function R2 becomes
√
m 2 −ω2
R2 = C̃1 z iω+1/2 (1 − z) 2 F1 (a2 , b2 ; c2 ; z)
√
+C̃2 z −iω (1 − z) m −ω 2 F1 (a2 − c2 + 1, b2
2 2
− c2 + 1; 2 − c2 ; z). (50)

Notice that the variable z satisfies

z = 0 at r = r H , z → 1 as r → ∞, (51)

therefore we find that near the horizon the radial function R2 behaves as

R2 ≈ C̃1 z iω+1/2 + C̃2 z −iω ≈ C̃1 eiωx+x/2 + C̃2 e−iωx , (52)

where x is the tortoise coordinate (2) for the Witten black hole and is given by

x = ln(er − 1). (53)

From (22) we point out that the time dependence of the Dirac field is exp(−iωt),
therefore we obtain that in the near horizon expression (52) for R2 , the first term is
an outgoing wave, whereas the second term is an ingoing wave. Near the horizon the
boundary condition of the QNM imposes that the field must be purely ingoing, thus
we take C̃1 = 0. Hence the radial function R2 takes the form
√
R2 = C̃2 z −iω (1 − z) m 2 −ω2
2 F1 (A2 , B2 ; C2 ; z), (54)

with

A2 = a2 − c2 + 1, B2 = b2 − c2 + 1, C2 = 2 − c2 . (55)

As is well known, when c is not a negative integer and c − a − b is not an integer

the hypergeometric function 2 F1 (a, b; c; z) satisfies Kummer’s property [41,42]

123
2642 A. López-Ortega, I. Vega-Acevedo

Γ (c)Γ (c − a − b)
2 F1 (a, b; c; z) = 2 F1 (a, b; a + b + 1 − c; 1 − z)
Γ (c − a)Γ (c − b)
Γ (c)Γ (a + b − c)
+ (1 − z)c−a−b
Γ (a)Γ (b)
×2 F1 (c − a, c − b; c + 1 − a − b; 1 − z), (56)

where Γ (a) denotes the gamma function

Taking into account Kummer’s property we write the radial function R2 as

√
m 2 −ω2 Γ (C 2 )Γ (C 2
− A2 − B2 )
R2 = C̃2 z −iω (1 − z)
Γ (C2 − A2 )Γ (C2 − B2 )
√
×2 F1 (A2 , B2 ; A2 + B2 − C2 + 1; 1 − z) + C̃2 z −iω (1 − z)− m −ω
2 2

Γ (C2 )Γ (A2 + B2 −C2 )

× 2 F1 (C 2 − A2 , C 2 − B2 ; C 2 − A2 − B2 + 1; 1 − z).
Γ (A2 )Γ (B2 )
(57)

As r → +∞ (z → 1) the previous expression simplifies to (note that as z → 1, we

can use the approximation 1 − z ≈ e−x )
Γ (C2 )Γ (C2 − A2 − B2 ) −√m 2 −ω2 x Γ (C2 )Γ (A2 + B2 − C2 ) √m 2 −ω2 x
R2 ≈ e + e .
Γ (C2 − A2 )Γ (C2 − B2 ) Γ (A2 )Γ (B2 )
(58)
To determine precisely what is the ingoing or outgoing wave, we assume that
√
ω2 − m 2
> 0, (59)
ω
to find that as r → +∞ the first term of (58) is an ingoing wave, whereas the sec-
ond is an outgoing wave [43]. The boundary condition of the QNM imposes that we
have only outgoing waves at infinity. Taking into account the properties of the gamma
function we must satisfy the condition

C2 − A2 = −n, or C2 − B2 = −n, n = 0, 1, 2, . . . (60)

From the previous equations we find that for the component Ψ˜2 of the Dirac field
its QNF are equal to

i  2  i 
1 2

ω2 = − n − m2 , ω2 = − n+ 2 − m2 . (61)
2n 2 n+ 1
2

Notice that for the first set of the previous QNF we must take n = 1, 2, 3, . . .. Using
a similar method we find that for the component Ψ˜1 of the Dirac field its QNF are

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2643

i   i  
1 2
ω1 = − (n + 1)2 − m 2 , ω1 = − n+ − m2 , (62)
2(n + 1) 2 n+ 1
2
2

with n = 0, 1, 2, . . ..
Thus for the components Ψ̃1 and Ψ̃2 we find that their QNF are purely imaginary
as those of Refs. [14]–[23]. We also note that Ψ̃1 and Ψ̃2 have one set of QNF that
coincides with one set of the QNF for the minimally coupled Klein-Gordon field of
the same mass (35), but both components Ψ̃1 and Ψ̃2 have an additional set of QNF,
and this additional set is different from the QNF of the minimally coupled massive
Klein-Gordon field. This result illustrates that even in two dimensions the Klein-Gor-
don and Dirac fields have different responses to the gravitational fields.
We say that the QNM are stable when their amplitudes decrease as time increases.
From time dependence taken in formulas (7) and (22) we obtain that the QNM are
stable for ω I < 0, where ω I denotes the imaginary part of the QNF. For the Witten
black hole we find that the QNM of the component Ψ̃1 are stable when

n > m − 21 , (63)

and the QNM of the component Ψ̃2 are stable when

n > m for n ≥ 1 and 1

2 > m for n = 0. (64)

For the Klein-Gordon field in Witten black hole the stability of its QNM is studied
in [24].
Thus for the Dirac field we may expect that at least the fundamental QNM is unsta-
ble, and depending on the value of the mass additional QNM may be unstable. Nev-
ertheless we recall that in the method used to calculate QNF (61) and (62) we impose
the condition (59), thus we must verify whether these QNF satisfy this condition. For
the component Ψ2 , from the first set of QNF (61) follows
√
ω2 − m 2 n2 + m 2
= 2 > 0, for n > m, (n ≥ 1), (65)
ω n − m2

and for the second set of QNF (61) we get

√
ω2 − m 2 (n + 1/2)2 + m 2
= > 0, for n + 1
> m, (n ≥ 0). (66)
ω (n + 1/2)2 − m 2 2

Hence only the stable QNM (those that fulfill the stability condition (64)) satisfy our
assumption (59).
So this fact points out that the modes that fail to keep the stability condition (64),
in addition they do not satisfy condition (59) that we use to calculate the QNM of the
Witten black hole. A similar result is valid for QNF (62) of the component Ψ̃1 . We think
that these modes must be studied in more detail, because in [24] for the Klein-Gordon
field are found some unstable QNM. Moreover the results by Azreg-Ainou [37] point

123
2644 A. López-Ortega, I. Vega-Acevedo

out that the Witten black hole has some classical instabilities. Nevertheless for the
Dirac field we obtain that the possible unstable modes do not fulfill our assumption (59).
Since the Witten black hole is a single horizon, asymptotically flat two-dimensional
black hole and its tortoise coordinate (53) satisfies conditions (3), from the results
explained in Sect. 2 we expect that the minimally coupled massless Klein-Gordon and
Dirac fields do not have well defined QNF. Nevertheless analytical expressions for the
QNF of the massless Dirac field are presented in Ref. [26] (see formula (37) in [26]).
Thus on the existence of the QNF for the massless Dirac field in Witten black hole
our results of Sect. 2 contradict those of [26].
To solve this issue we note that in Ref. [26] there is an inconsistency. To write
solutions (29) of [26] Becar et al. assume that their parameter c is not an integer [42].
Furthermore to write their equation (34) they assume that their parameter c is not a
negative integer and that their quantity c − a − b is not an integer because they use
Kummer’s property of the hypergeometric function [42]. From the expressions for
the QNF of the massless Dirac field given in [26] [see their formula (37)], and using
the expressions for the parameters a, b, and c that appear in [26] [see their formulas
(26)-(28)] we find that for frequencies (37) of [26], their quantities c and c − a − b are
integers. This fact shows that Becar et al. contradict their assumptions and therefore
the frequencies that they give in [26] are not QNF of the massless Dirac field.
The previous comments imply that the Witten black hole does not have well defined
QNM when the quantity c of Ref. [26] is not an integer. It is convenient to extend the
analysis of [26] to integral values of c. For c equal to an integer it is expected that
one solution of Eq. (25) in [26] involves logarithmic terms [42]. Nevertheless for this
problem the solutions can be found in a simple closed form.
First we note that to obtain decaying modes we must take c = N with n ∈ Z, N ≤ 0.
Furthermore when c = N we note that the quantities a, b, and c − a − b of Eq. (25)
in [26] are integers and we notice that for this value of c one of the quantities a, b
is always a negative integer. As a consequence we expect the simplification of the
solutions [42].
In fact, when c is a non-positive integer we find that Eq. (25) of [26] transforms
into2

z(1 − z)g (z) + (c − 2cz)g (z) − c(c − 1)g(z) = 0, (67)

whose solutions are

g(z) = C̃1 z 1−c + C̃2 (1 − z)1−c , (68)

where C̃1 and C̃2 are constants (as previously).

Therefore the component φ2 (z) of formula (24) in [26] is equal to

φ2 (z) = z 1/4 (C̃1 z iω (1 − z)−iω + C̃2 z −iω (1 − z)iω ). (69)

2 In formulas (67)–(69) we follow the conventions of Ref. [26].

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2645

From this expression for φ2 (z) we get that the first term is an outgoing wave near the
horizon and at infinity. The second term is an ingoing wave near the horizon and at
infinity. Hence these solutions do not satisfy the boundary conditions of the QNM,
and also when the quantity c is an integer we do not find QNF for the massless Dirac
field in Witten black hole.
Thus according to the results of Sect. 2 we believe that in Witten black hole the
minimally coupled massless Klein-Gordon and Dirac fields do not have well defined
QNF. Notice that for the Dirac field propagating in Witten black hole for QNF (61)
and (62) the limits m → 0 are well defined, nevertheless the frequencies we get in
these limits are not QNF.
For the coupled to scalar curvature massive Klein Gordon field propagating in
Witten black hole the effective potential is [38]

m 2 (er − 1) ζ (er − 1) m 2 ex ζ ex
V = + = + . (70)
er e2r e x + 1 (e x + 1)2

Following Chandrasekhar [32] we find that in Witten black hole for the Dirac field its
effective potentials take the form

m (1 − e−r )1/2 m 2 ex m e x/2

V± = m 2 (1 − e−r ) ∓ = ∓ . (71)
2 er ex + 1 2 (e x + 1)3/2

We point out that for the minimally coupled fields the effective potentials (70) and
(71) go to zero as the mass of the field goes to zero (it is expected from the results of
Sect. 2). Furthermore notice that the effective potentials V and V± satisfy

lim V = lim V± = 0, lim V = lim V± = m 2 . (72)

x→−∞ x→−∞ x→+∞ x→+∞

4 Summary

We find that the minimally coupled Klein-Gordon and Dirac fields do not have modes
that satisfy the standard boundary conditions for the QNM of single horizon, asymp-
totically flat two-dimensional black holes. Thus if we are interested in calculating
the QNF of the Klein-Gordon and Dirac fields in a black hole of this family, then
we must analyze the massive fields (as in the previous section and [23–25]), fields
non-minimally coupled to scalar curvature (as in [23,24]) or to dilaton (as in [27]).
This result shows that for two-dimensional black holes when we know the QNF of
a massive field and take the massless limit, we must verify whether the frequencies
obtained in this limit are QNF. We believe that the consequences of this result deserve
further research.
In [26] for the Witten black hole Becar et al. assert that the QNM of the massless
Dirac field are stable, but based on the results of Sect. 2 we believe that in the massless
case the QNF of this fermion field are not well defined, and we also show that there
are inconsistencies in the method used in Ref. [26]. Therefore the conclusion of Ref.
[26] on the stability of the QNM for the massless Dirac field in Witten black hole is

123
2646 A. López-Ortega, I. Vega-Acevedo

not valid, because the QNM are not defined in the massless limit. Furthermore for
both components of the Dirac field we find that the modes not satisfying the stability
conditions (63) and (64), at the same time they do not fulfill the assumption (59) and
it is probable that these modes are not QNM. Notice that for the Klein-Gordon field
propagating in Witten black hole some classical and quantum instabilities are known
(see Refs. [23–25,37–39]).

Acknowledgments We thank to Dr. R. Cordero Elizalde his interest and support of this project. This work
was supported by CONACYT México, SNI México, EDI-IPN, COFAA-IPN, and Research Projects SIP-
20110729 and SIP-20111070. I. Vega-Acevedo acknowledges financial support from CONACYT research
grant no. J1-60621-I.

References

1. Grumiller, D., Kummer, W., Vassilevich, D.V.: Phys. Rept. 369, 327 (2002). [arXiv:hep-th/0204253]
2. Grumiller, D., Meyer, R.: Turk. J. Phys. 30, 349 (2006). [arXiv:hep-th/0604049]
3. Kokkotas, K.D., Schmidt, B.G.: Living Rev. Rel. 2, 2 (1999). [arXiv:gr-qc/9909058]
4. Berti, E., Cardoso, V., Starinets, A.O.: Class. Quant. Grav. 26, 163001 (2009). [arXiv:0905.2975
[gr-qc]]
5. Hod, S.: Phys. Rev. Lett. 81, 4293 (1998). [arXiv:gr-qc/9812002]
6. Kunstatter, G.: Phys. Rev. Lett. 90, 161301 (2003). [arXiv:gr-qc/0212014]
7. Maggiore, M.: Phys. Rev. Lett. 100, 141301 (2008). [arXiv:0711.3145 [gr-qc]]
8. Birmingham, D., Sachs, I., Solodukhin, S.N.: Phys. Rev. Lett. 88, 151301 (2002). [arXiv:hep-th/
0112055]
9. Cardoso, V., Lemos, J.P.S.: Phys. Rev. D 63, 124015 (2001). [arXiv:gr-qc/0101052]
10. Birmingham, D.: Phys. Rev. D 64, 064024 (2001). [arXiv:hep-th/0101194]
11. Decanini, Y., Folacci, A.: Phys. Rev. D 79, 044021 (2009). [arXiv:0901.1642 [hep-th]]
12. Sachs, I., Solodukhin, S.N.: JHEP 0808, 003 (2008). [arXiv:0806.1788 [hep-th]]
13. Crisostomo, J., Lepe, S., Saavedra, J.: Class. Quant. Grav. 21, 2801 (2004). [arXiv:hep-th/0402048]
14. Fernando, S.: Gen. Relativ. Gravit. 36, 71 (2004). [arXiv:hep-th/0306214]
15. López-Ortega, A.: Gen. Relativ. Gravit. 37, 167 (2005).
16. Fernando, S.: Phys. Rev. D 77, 124005 (2008). [arXiv:0802.3321 [hep-th]]
17. Fernando, S.: Phys. Rev. D 79, 124026 (2009). [arXiv:0903.0088 [hep-th]]
18. Chen, B., Xu, Z.b.: JHEP 0911, 091 (2009). [arXiv:0908.0057 [hep-th]]
19. Chen, B., Xu, Z.b.: arXiv:0901.3588 [hep-th]
20. Li, R., Ren, J.R.: arXiv:1008.3239 [hep-th]
21. Du, D.P., Wang, B., Su, R.K.: Phys. Rev. D 70, 064024 (2004). [arXiv:hep-th/0404047]
22. López-Ortega, A.: Gen. Relativ. Gravit. 38, 743 (2006). [arXiv:gr-qc/0605022]
23. Zelnikov, A.: JHEP 0807, 010 (2008). [arXiv:0805.4031 [hep-th]]
24. Becar, R., Lepe, S., Saavedra, J.: Phys. Rev. D 75, 084021 (2007). [arXiv:gr-qc/0701099]
25. Lopez-Ortega, A.: Int. J. Mod. Phys. D 18, 1441 (2009). [arXiv:0905.0073 [gr-qc]]
26. Becar, R., Lepe, S., Saavedra, J.: Int. J. Mod. Phys. A 25, 1713 (2010)
27. Kettner, J., Kunstatter, G., Medved, A.J.M.: Class. Quant. Grav. 21, 5317 (2004). [arXiv:gr-qc/0408042]
28. Cadoni, M.: Phys. Rev. D 53, 4413 (1996). [arXiv:gr-qc/9510012]
29. Lopez-Ortega, A.: Lat. Am. J. Phys. Educ. 3, 578 (2009). [arXiv:0906.2754 [gr-qc]]
30. Mann, R.B., Morsink, S.M., Sikkema, A.E., Steele, T.G.: Phys. Rev. D 43, 3948 (1991)
31. Morsink, S.M., Mann, R.B.: Class. Quant. Grav. 8, 2257 (1991)
32. Chandrasekhar, S.: The Mathematical Theory of Black Holes. Oxford University Press, Oxford (1983)
33. Witten, E.: Phys. Rev. D 44, 314 (1991)
34. Mandal, G., Sengupta, A.M., Wadia, S.R.: Mod. Phys. Lett. A 6, 1685 (1991)
35. McGuigan, M.D., Nappi, C.R., Yost, S.A.: Nucl. Phys. B 375, 421 (1992). [arXiv:hep-th/9111038]
36. Hsu, R.R., Huang, G., Lin, W.F., Lee, C.R.: Class. Quant. Grav. 10, 505 (1993). [arXiv:hep-th/9205023]
37. Azreg-Ainou, M.: Class. Quant. Grav. 16, 245 (1999). [arXiv:gr-qc/9902005]
38. Frolov, V.P., Zelnikov, A.: Phys. Rev. D 63, 125026 (2001). [arXiv:hep-th/0012252]

123
Quasinormal frequencies of asymptotically flat two-dimensional black holes 2647

39. Frolov, V.P., Massacand, C.M., Schmid, C.: Phys. Lett. B 279, 29 (1992)
40. Li, X.Z., Hao, J.G., Liu, D.J.: Phys Lett. B 507, 312 (2001). [arXiv:gr-qc/0205007]
41. Abramowitz, M., Stegun, I.A.: Handbook of Mathematical Functions, Graphs, and Mathematical
Table. Dover Publications, New York (1965)
42. Wang, Z.X., Guo, D.R.: Special Functions. World Scientific Publishing, Singapore (1989)
43. Ohashi, A., Sakagami, M.a.: Class. Quant. Grav. 21, 3973 (2004). [arXiv:gr-qc/0407009]

123