0% found this document useful (0 votes)
12 views

Discussing the Spectrum of the Kalb-Ramond Field Coupled to 3D Gravity

The document discusses the dynamical mass generation for the Kalb-Ramond (K-R) field in three-dimensional gravity, highlighting that torsion prevents 1-loop mass corrections. It explores the coupling of the K-R field with gravity, showing that in the absence of torsion, a mass gap is produced for the 2-form field. The analysis includes the calculation of 1-loop corrections and the implications of torsion on the mass spectrum of the theory.

Uploaded by

gustavo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
12 views

Discussing the Spectrum of the Kalb-Ramond Field Coupled to 3D Gravity

The document discusses the dynamical mass generation for the Kalb-Ramond (K-R) field in three-dimensional gravity, highlighting that torsion prevents 1-loop mass corrections. It explores the coupling of the K-R field with gravity, showing that in the absence of torsion, a mass gap is produced for the 2-form field. The analysis includes the calculation of 1-loop corrections and the implications of torsion on the mass spectrum of the theory.

Uploaded by

gustavo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

Discussing the Spectrum of the

Kalb-Ramond Field Coupled to


arXiv:hep-th/0203158v1 18 Mar 2002

3D-Gravity
J.L. Boldoa∗, J.A. Helayël-Netob,c† and N. Panzac
a
UFES, Universidade Federal do Espı́rito Santo
CCE, Departamento de Fı́sica
Campus Universitário de Goiabeiras 29060-900
Vitória, ES, Brazil
b
CBPF, Centro Brasileiro de Pesquisas Fı́sicas
Rua Xavier Sigaud 150, 22290-180 Urca
Rio de Janeiro, Brazil
c
UCP, Universidade Católica de Petrópolis
Rua Barão do Amazonas 124, 25685-070
Petrópolis, Brazil

December 17, 2018

Abstract
The mechanism of dynamical mass generation for the gauge field
is studied through 1-loop. We find out that torsion is an obstruction
to the appearance of a 1-loop mass correction. Contrary, if torsion is
not present, a mass gap is generated for the 2-form field.


E-mail: [email protected]

E-mail: [email protected]
1 Introduction
The Abelian 2-form gauge potential, usually referred to as Kalb-Ramond
field (K-R), exhibits a number of interesting features in four dimensional
space-time (4D) [1]; special emphasis is given to the fact that it mixes up
with the electromagnetic gauge potential so as to yield a massive spin-1
excitation based on a [U (1)]2 -symmetry, without the need for Higgs scalars
[2]. More recently, the coupling between the K-R and Maxwell fields has
been reconsidered in connection with the issue of Dirac-like monopoles with
massive photons [3]. Also, interesting discussions on the possible non-Abelian
extension of the K-R field are set in the works of Ref. [4].
Reassessing the K-R field in three dimensional space-time (3D) brings
about some peculiarities that show up as a by-product of three dimensional
space-time. For example, the Abelian gauge field dual to the K-R potential
in 3D is subject to a particular gauge symmetry that selects its longitudinal
component as the physical propagating mode, whereas its transverse part
appears as a compensating mode, and so it can always be gauged away.
Once the mass-shell condition is imposed, no physical propagating degree
of freedom survives and we have the peculiarity that the K-R gauge field
carries no on-shell degrees of freedom in 3D [5]. We understand that this is a
consequence of setting its dynamics on a flat space-time. So, our propose is
to set out this paper in order to illustrate how the coupling of the K-R field
to the 3D counterpart of the Einstein-Hilbert gravity excites a massive scalar
mode in the spectrum of the theory. This is a result that already emerges at
classical level.
The main motivation of our work is to compute 1-loop corrections to
the K-R field self-energy, so as to understand how these effects change the
mass spectrum set up at tree-level. We pursue our investigation in the cases
of ordinary (torsionless) gravity (Section 2) and, later on, we go over to the
general case of gravity with torsion (Section 3) [6]. We find some peculiarities
at the end of our analysis and we shall comment on them in our Conclusive
Comments (Section 4).

1
2 The Riemannian Case
The model we contemplate accounts for the non-minimal coupling of 3D-
gravity to the K-R field according to the Lagrangian density given below:

L = LEH + Lgauge + Lint , (2.1)

where the first term is the usual Einstein-Hilbert Lagrangian:


1√
LEH = gR. (2.2)
κ2
Notice the absence of the overall minus sign that appears in 4D-gravity: in
3D, this must be the choice in such a way to avoid that the graviton becomes
a ghost.
The second term describes the Lagrangian for the antisymmetric gauge
field Bµν ,
1√
Lgauge = gGµνλ Gµνλ , (2.3)
6
where Gµνλ is a 3-form written in terms of the potential Bµν as follows:

Gµνλ = ∇µ Bνλ + ∇λ Bµν + ∇ν Bλµ . (2.4)

The last term accounts for the interaction between gravity and the gauge
field:

Lint = ξ gǫµνλ ∇µ Bνλ R, (2.5)
µνλ
with ǫµνλ = ε√g , εµνλ being the totally antisymmetric symbol, while ∇µ
denotes the covariant derivative. The coupling constant ξ carries dimension
1
of (mass)− 2 .
To analyse the espectrum of the model, we consider the linearised ap-
proximation of the full theory. This consists in the expansion of the metric
field around the flat background as below:

gµν = ηµν + κhµν , (2.6)

where hµν is the perturbation associated to the graviton field.


As for the 2-form gauge potential, it is more convenient to parametrise it
in terms of its dual, Bµ = 3!1 ǫµνλ B νλ . Doing so, the gauge transformation for
Bµν can be rephrased as

δBλ = ǫλ µν ∇µ χν , (2.7)

2
where χµ is the 3-vector gauge parameter.
In order to invert the wave operator that mixes hµν and Bµ in the bilinear
piece of the action, we fix the De Donder gauge for hµν ,
1
LDonder = Fµ F µ , (2.8)

where  
λ 1 λ
Fµ = ∂λ hµ − δµ h . (2.9)
2
On the other hand, the gauge transformations (2.7) for Bµ suggests the
following gauge-fixing term for the Abelian symmetry associated to Bµν :
2
Lgf = β ǫµνλ ∇ν Bλ . (2.10)

With these elements, the bilinear piece of the action can be cast under
the form:
1X α
L= φ Oαβ φβ , (2.11)
2 αβ

where φ = (hµν , B µ ) and O is the differential wave operator to be inverted


to give us the graviton and the gauge field propagators. This operator can
be decomposed in four sectors, namely,
 
Aµν,κλ Bµν,κ
O= , (2.12)
Cµ,κλ Dµ,ν

where
 (2)  (1) (α + 1)  (0−s)  (0−w)
Aµν,κλ = Pµν,κλ − Pµν,κλ − Pµν,κλ − P
2√ 2α 2α 4α µν,κλ
2  (0−sw) (0−ws)

+ Pµν,κλ + Pµν,κλ ,

Bµν,κ = −2κξθµν ∂κ ,
Cµ,κλ = 2κξ∂µ θκλ , (2.13)
Dµ,ν = −2 (βθµν + ωµν ) ,
where the P ′ s are the Barnes-Rivers projector operators for symmetric second-
rank tensors in D=3 (see Ref. [7] for details), while θµν and ωµν are respec-
tively the transverse and longitudinal operators for vectors.

3
The propagators are given by

h0 |T [Φα (x) Φβ (y)]| 0i = i O−1 αβ
δ 3 (x − y) , (2.14)

where wave operator is inverted by using of the multiplicative relations sat-


isfied by the projectors [7]. In so doing, the propagators in momentum space
read as follows:

2i 2 [p2 − m2 (α + 1)] (0−w)
hhhi = 2 −P (2) + αP (1) + P
p (p2 − m2 )
h 
m2 (0−s)
√ (0−sw) (0−ws)
i
− 2 P + 2 P +P ,
(p − m2 )

1
hhµν Bκ i = (θµν + 2ωµν ) pκ , (2.15)
4κξp2 (p2 − m2 )

i im2
hBµ Bν i = θµν − ωµν ,
2βp2 2p2 (p2 − m2 )

where m2 = 8κ12 ξ2 and we have suppressed the indices in hµν and in the Pµν,κλ
projectors.
The above propagators display poles at p2 = 0 and p2 = m2 > 0, so that
tachyons do not show up. The next step is a investigation of the possibility
that negative-norm state might be present.
A necessary criterium for unitarity at tree-level is analysed by saturat-
ing the propagator with external currents J µν and J µ (for the graviton and
the gauge field, respectively), compatible with the gauge symmetries of the
Lagrangian. Ghosts are absent in the model if the imaginary part of the
residue of the current-current transition amplitude taken at the propagator
poles is non-negative. For the graviton, this amplitude in momentum space
is written as

A ≡ J ∗µν (p) hT [hµν (−p) hκλ (p)]i J κλ (p) , (2.16)

where, by virtue of the transversality of J µν (p), only the spin projectors P (2)
and P (0−s) survive. This fact enable us to write the imaginary part of the
residue of the transition amplitude at the pole p2 = 0 as

I m(ResA) = 2 |J µν |2 − |Jνν |2 . (2.17)

4
Now, defining the following set of independent vectors in momentum
space, 
pµ ≡ p0 , p , pµ ≡ (p0 , −p) , ǫµ ≡ (0, ǫ) , (2.18)
that satisfy the conditions:
2
p̄.p = p0 + (p)2 > 0 , (2.19)
p.ǫ = p̄.ǫ = 0,
µ
ǫ ǫµ = −1,

the symmetric tensor current J µν (p) can be decomposed according to:

Jµν (p) = a (p) pµ pν + b (p) p(µ p̄ν) + c (p) p(µ ǫν) +


d (p) p̄µ p̄ν + e (p) p̄(µ ǫν) + f (p) ǫ(µ ǫν) . (2.20)

Substituting (2.20) into (2.17), and making use the relations (2.18) and
(2.19), we find that I m(ResA) = 0, showing that the massless pole does not
propagate. For the pole at p2 = m2 , one obtains
2
I m(ResA) = Jµµ , (2.21)

that is always positive-definite. From this result, we find only one on-shell
degree of freedom for the massive graviton.
For the vector field, the transition amplitude is given by

A = J ∗µ (p) hT [Bµ (−p) Bν (p)]i J ν (p) . (2.22)

Expanding the current J µ (p) with respect to the basis (2.18),

J µ (p) = a (p) pµ + b (p) pµ + c (p) εµ , (2.23)

and by making use of its conservation law,

ǫµνλ pν Jλ = 0 , (2.24)

we can show that the massless pole is again non-dynamical, while the residue
of the amplitude at the massive pole give us:
1 2 2
|a| m , (2.25)
2

5
which ensures one physical degree of freedom. Therefore, the non-minimal
coupling of the K-R field to gravity results in a appearance of a dynamical
massive pole in the longitudinal sector of Bµν .
We now turn into the calculation of the 1-loop mass correction for the
gauge field propagator. The relevant vertices to our discussion are shown
in figure 1, while all possible self-energy corrections are depicted in figure 2.
The vertices of the figures 1.a and 1.b come from the Lagrangians (2.3) and
(2.5) by expanding the metric up to order κ, while the one of figure 1.c is
obtained from (2.3) up to order κ2 .

(a) (b) (c)

Figure 1

(a) (b) (c)


Figure 2

From the Feynman rules displayed in figure 1, the contribution coming


from the loop integrals can be evaluated. Due to algebraic complexity of
the vertices and propagators, the explicit evaluation of these diagrams only
was possible by means of the software FORM. Since we are concerned with
the contribution to the vector field mass, and we checked that a Chern-
Simons-like term is not generated, we can calculate the diagrams by setting
all momenta appearing on the external legs to zero. One therefore, obtains
the following contribution to the 1-loop shift in the gauge field mass:
2 3π 2
∆m 1−loop
= √ . (2.26)
16 2κξ 3

6
3 The Non-Riemannian Case
Now, following the same procedure as the one in the previous section, we
propose to extend our analysis to the case where torsion is non-vanishing.
In three-dimensional space-time, the torsion field can be covariantly split
according to its SO(1,2) irreducible components:
1
Tµνλ = εµνλ ϕ + (ηνλ tµ − ηµλ tν ) + εµνα X α λ , (3.27)
2
namely, a trace part tµ = Tµν ν , a totally antisymmetric part ϕ = 3!1 εµνλ Tµνλ
and a traceless rank-2 symmetric tensor Xµν [7].
A possible gauge-invariant action may be so chosen that there is no cou-
pling at all between the K-R field and torsion, i.e., the former “neither yields
nor feels torsion”. Thus, let us consider the action
Z  
3 √ 1 1 µνλ e µB ,
µ
S= dx g R + Gµνλ G + 2ξR∇ (3.28)
κ2 6

where the tilde on ∇µ means that we are considering only the Riemaniann
part of the covariant derivative. Varying the total action (3.28) with respect
to the independent fields, we get the following result for the trace part of the
torsion:
e µ∇
4κ2 ξ ∇ e ν Bν
tµ = − , (3.29)
e ν Bν
1 + 2κ2 ξ ∇
while the variations with respect to the fields ϕ and Xµν give us

ϕ=0,
(3.30)
Xµν = 0 .

The torsion irreducible components act as mere auxiliary fields and one can
eliminate them by means of their respective algebraic field equations. There-
fore, substituing this results back into the action, one obtain a Lagrangian
expressed exclusively in terms of Bµ and hµν :

1 √ e 1√ µνλ √ ee µ
L = g R + gG µνλ G + 2ξ g R ∇µ B (3.31)
κ2 6
√ ∇e λ∇e µB µ∇e λ∇
e ν Bν
+8κ2 ξ 2 g .
e µB µ
1 + 2κ2 ξ ∇

7
From the bilinear sector of the Lagrangian above supplemented by the
gauge fixing terms (2.8) and (2.10), we obtain the following propagators:
  
2i p2
hhhi = 2 −P + αP + 2 (α + 1) + 2 P (0−w)
(2) (1)
p m
 h 
p 2
(0−s)
√ (0−sw) (0−ws)
i
+ 1+ 2 P + 2 P +P , (3.32)
m

 
i 1
hBµ Bν i = 2 θµν + ωµν .
2p β

It is remarkable to notice now that the poles are located at p2 = 0,


contrary to the previous case, where torsion was not present. It is just
the last term of the action (3.31) the responsible for the suppression of the
massive pole. The mixing between hµν and Bµ arising from the mentioned
term appears with the right coefficient that eliminates the massive pole.
Proceeding analogously to what we did in the previous Section, the mass-
less poles are shown to be non-dynamical for both fields. From this result,
one concludes that the coupling of the K-R field to gravity with non-vanishing
torsion may result in a theory without on-shell degrees of freedom at tree-
level as long as the action (3.31) is concerned. Moreover, by computing the
Feynman rules for this action and calculating the 1-loop self-energy graphs,
we checked that no mass is dynamically generated by radiative corrections.

4 Concluding Remarks
As a final result, we can state that the coupling of the 3D K-R field with
Einstein-Hilbert gravity excites a dynamical mode of the rank-2 gauge po-
tential, whenever torsion is not considered. The inclusion of torsion as a
non-dynamical field changes this result, as we checked in Section 3. How-
ever, had we considered non-minimal couplings with higher powers in the
curvature, the situation would be completely changed, as torsion would be-
come dynamical and its mixing with the graviton [7] would for sure trigger
a dynamical mass for the gauge field on the basis of the results worked out
in Section 2.

8
To conclude, we would like to mention that, if gravity were described by
a pure Chern-Simons term, even in the torsionless case the K-R field would
remain non-dynamical on-shell, since no mass would be generated neither at
the classical nor at the 1-loop level.

References
[1] M. Kalb and P. Ramond, Phys.Rev. D9 (1974) 2273;

[2] E. Cremmer and J. Scherk, Nucl.Phys. B72 (1974) 117; C.R. Hagen,
Phys.Rev. D19 (1979) 2367; T.J. Allen, M.J. Bowick and A. Lahiri,
Mod.Phys.Lett. A6 (1991) 559;

[3] W.A. Moura Melo, J.A. Helayël-Neto and N. Panza, Int.J.Mod.Phys.


A14 (1999) 3949; H.R. Christiansen, M.S. Cunha, J.A. Helayël-Neto,
L.R.U. Manssur and A.L.M.A. Nogueira, Int.J.Mod.Phys. A14 (1999)
147, Int.J.Mod.Phys. A14 (1999) 1721; C.N. Ferreira, J.A. Helayël-Neto
and M.B.D.S.M. Porto, Nucl.Phys. B620 (2002) 181;

[4] J. Barcelos-Neto, A Cabo, and M.B.D. Silva, Z. Phys. C72 (1996) 345;
A. Lahiri, Phys. Rev. D55 (1997) 5045; M. Henneaux, V.E.R. Lemes,
C.A.G. Sasaki, S.P. Sorella, O.S. Ventura and L.C.Q. Vilar, Phys.Lett.
B410 (1997) 195;

[5] J.L. Boldo, J.A. Helayël-Neto and N. Panza, accept for publication in
Class.Quantum Grav.;

[6] F.W. Hehl, P. von der Heyde, G.D. Kerlick, and J.M. Nester, Rev.
Mod.Phys. 48 (1976) 3641; F.W. Hehl, J.D. McCrea, E.W. Mielke, and
Y. Ne’eman, Phys.Rep. 258 (1995) 1; De Sabbata and M. Gasperini,
“Introduction to Gravitation”, World Scientific, Singapore (1985);

[7] J.L. Boldo, L.M. de Moraes and J.A. Helayël-Neto, Class. Quantum Grav.
17 (2000) 813.

You might also like