On spin-1 massive particles coupled to a Chern-Simons field

M. Gomes a , L. C. Malacarne b and A. J. da Silva a
a Instituto

arXiv:hep-th/9908181v1 27 Aug 1999

de F´ısica, USP
C.P. 66318 - 05315-970, S˜
ao Paulo - SP, Brazil.
b Departamento de F´
ısica, Universidade Estadual de Maring´
Av. Colombo, 5790 - 87020-900, Maring´
a - PR, Brazil.

We study spin one particles interacting through a Chern-Simons field. In the
Born approximation, we calculate the two body scattering amplitude considering three possible ways to introduce the interaction: (a) a Proca like model
minimally coupled to a Chern-Simons field, (b) the model obtained from (a)
by replacing the Proca’s mass by a Chern-Simons term and (c) a complex
Maxwell-Chern-Simons model minimally coupled to a Chern-Simons field. In
the low energy regime the results show similarities with the Aharonov-Bohm
scattering for spin 1/2 particles. We discuss the one loop renormalization
program for the Proca’s model. In spite of the bad ultraviolet behavior of
the matter field propagator, we show that, up to one loop the model is power
counting renormalizable thanks to the Ward identities satisfied by the interaction vertices.

In the recent years much work has been devoted to the study of the properties of the
Chern-Simons (CS) field [1]. This was motivated not only by its potential applications
to condensed matter physics but also because these studies have unveiled some new and
interesting aspects of the dynamics of relativistic quantum physics. In particular, it has
been noticed that in some circumstances the CS field plays a stabilizing role providing
theories with improved ultraviolet behavior [2]. However, most of these investigations have
been restricted to the cases of spinless and spin 1/2 particles. The reasons behind this
fact are the notorious difficulties found in the conventional treatment for higher spin fields
in four dimensions. The troublesome aspects include noncausal propagation and lack of
renormalizability. It is certainly worthwhile to study the interaction of a CS and spin one
matter fields so that the origin of the difficulties could be better understood and perhaps
new and safer routes could be found. With this in mind, we would like to present here the
results of some investigations concerning the dynamics of spin one fields interacting through
a CS term.
As a first observation, we note that a free, spin one particle of mass M can be described
alternatively by the Proca Lagrangian,


up to one loop the model is renormalizable. The model (2) is equivalent to the self-dual model. simulated in the field theory approach by a quartic (φ∗ φ)2 interaction. we also study the Born approximation for the two body scattering amplitudes. or by the Maxwell-Chern-Simons (MCS) Lagrangian.4]. We will pursue the investigation of the spin effects on the perturbative AB scattering by considering spin one particles. as vacuum polarization and anomalous magnetic momentum. The paper contains also an appendix with details of the calculations. [5]. One advantage of such procedure is that it incorporates some purely quantum field effects. We may recall that for spinless particles the Born approximation found in the perturbative method only agrees with the expansion of the exact result if a contact interaction. e. First of all. In both approaches the AB scattering was discussed from a first quantized viewpoint. In this investigation we will study the two body scattering amplitude for the cases of minimal coupling of (1) to a CS field and when in (1) the Proca mass M 2 φ†µ φµ is replaced by a complex CS term. Here we consider the problem from the perspective of the theory of quantum fields. Our work is organized as follows. The two formulations are not entirely inequivalent. A discussion of our results is presented in section IV. Nevertheless. 2 2 (2) Whereas the first formulation encompasses both modes of spin ±1. [3. We will also consider the case of minimal coupling of (2) to a CS field. LP = − Fµν 2 (1) where Fµν = ∂µ φν − ∂ν φµ . Previous work in this direction started either with a complex Proca field minimally coupled to the electromagnetic field [11] or with a linearized Yang Mills equation [12]. and this last model is like a square root of the Proca’s. on the other hand. It is also known that in the spin 1/2 case no new interaction is needed. thanks to the Ward identities satisfied by the basic interaction vertices. II. such equivalence does not in general persists whenever the models are coupled to other dynamical fields [6]. let us examine some of the kinematic aspects of the asymptotic theories. turns out to be power counting renormalizable to all orders of perturbation. is included from the beginning [8. Analogously to the scattering of lower spin particles we will expect to find similarities with the Aharonov-Bohm (AB) scattering [7]. the latter.. the role of the quartic interaction being played by the magnetic Pauli term [10]. POLARIZATION VECTORS AND FEYNMAN RULES As a preliminary step toward our study of the AB scattering of two spin one particles. In section III we discuss in detail the one loop renormalization parts for the Proca model and also determine the anomalous magnetic moment.1 † µν F − M 2 φ†µ φµ . The MCS model. 9]. represents only a single mode of spin M/|M|. There. In section II we present the polarization vectors and Feynman rules for the models mentioned above. in spite of the bad behavior of the propagator of the matter field. 1 † µν M µνρ † LM CS = − Fµν F + ǫ [φµ ∂ν φρ + φµ (∂ν φρ )† ]. as the low energy limit of a fully quantized relativistic theory of spin one particles interacting through a CS field. Eq. (2). however. i. As we will show. For the Proca model minimally coupled to a CS field we will discuss the one loop renormalization program and calculate the anomalous magnetic moment of the matter field. being a 2 .

Eq. As discussed in [13]. In the Landau gauge (λ → ∞). (2) one has just one polarization which can be taken as εα (k) = (ε0 (k). 1. ǫij pj ) |~p| ǫµ2 = ( |~p| wp pi . i |M where ε0 (k) = k. s = ±1 >= |0> 2 (7) are eigenstates of (6).~ ε(0) ε(0) M being the and εi (k) = εi (0) + |M |(w ki with εµ (0) = √12 0. (5) where T µν = Fρ†µ F ρν + Fρµ F †ρν − g µν LP . In this space. Aµ . M M |~p| (4) The creation operators a† and b† allow us to construct the Fock space of the asymptotic states. the analytic expressions accompanying these rules are: CS field propagator: kρ 1 . 1.transversal field.~ |M | | p +|M |) polarization vector in the particle’s rest frame. εi (k)) . is given by JS = −i Z 2 d2 p X ǫij ǫiλ′ (p)ǫjλ (p) [a†λ′ (p)aλ (p) + b†λ′ (p)bλ (p)]. 2wp λ. Minimally coupling the Proca field to a CS field. leads us to the Lagrangian 1 θ λ LP = − G†µν Gµν − M 2 φ†µ φµ + εµνρ Aµ ∂ ν Aρ + (∂µ Aµ )2 . 2wp λ=1 λ (3) √ where wp = p~2 + M 2 and the polarization vectors satisfy the transversality condition pµ ǫµλ = 0. the Proca field described by (1) can be expanded in plane waves as φµ = 1 2π Z 2 d2 p X ǫµ [aλ e−ipx + b†λ eipx ]. in the case of the MCS model. ∂µ φµ = 0.λ′ =1 (6) In the particle’s rest frame we can see that a†1 (0) ± ia†2 (0) √ | p = 0. ). ~ ~ (8)   k. A convenient choice is ǫµ1 = (0. 2 2 2 (9) where Gµν = D µ φν −D ν φµ and D µ = ∂ µ −ieAµ . The Feynman rules associated to the above Lagrangian are depicted in Fig. Dµν (k) = − ǫµνρ 2 θ k + iǫ Matter field propagator: 3 (10) . we found that the spin part of the angular momentum operator J= Z d2 x εij xi : T 0j : .

s3 )Γµαβ (p1 . p)∆σβ (p). To make contact with the Aharonov-Bohm scattering. The tree approximation for this process is h i h Mf i = ε∗β (p3 . From the above expressions we can verify that the scattering amplitude vanishes unless spin is conserved. These properties will be helpful to discuss the ultraviolet behavior of the Green functions. Mf i (s. θ 4 (19) . p′) = pα Γµαβ (p. s) = ǫα1 (p) + isǫα2 (p) √ . p3 )εα (p1 . ~p′ ) = 4ie2 |M| −isα e 2 [s + 2i cot(α)] . in leading order. 0) = 0. ~p. s3 ↔ s4 ) . 2 (18) with ǫα1 (p) and ǫα2 (p) as in (4). −~p ′ ) and s3 . 1 1 (16) e and The expressions (14) and (15) are typical of gauge theories being similar to the ones found in scalar QED. After expanding in powers of |~p|/M. ~p) and p2 = (wp . i (17) where q = p1 − p3 and εα (p.pα pβ −i αβ g − ∆ (p) = 2 p − M 2 + iǫ M2 " αβ # (11) and interaction vertices (p and p′ denote the matter field’s momenta) h Γµαβ (p. s1 ) Dµν (q) ε∗ρ (p4 . −~p) and spins s1 and s2 .. dpµ (14) d µαβ Γ1 (p. s4 . The energy of the incoming particle is wp = m2 + ~p2 . We will then denote the momenta and spins of the outgoing particles by √ p3 = (wp .e. s2 ) +(p3 ↔ p4 . p4 = (wp . i. we get. are circularly polarized vectors. p − q) = Γµναβ 2 dpν (15) p′β Γµαβ (0. p′ ) = −ie (p + p′ )µ g αβ − pβ g µα − p′α g µβ 1 h i Γµναβ = ie2 g µβ g να + g µα g νβ − 2g µν g αβ . s1 = s3 and s2 = s4 or s1 = s4 and s2 = s3 . p4 )εσ (p2 . We assume that in the center of mass frame the incoming particles have momenta p1 = (wp . the spins si can either take the values ±1 and |~p| = |~p ′ |. respectively. s4 )Γνσρ 1 1 (p2 . let us study the low energy approximation for the scattering of two vector particles. 2 The above propagators and vertices obey the identities e i (12) (13) d∆αβ (p) = ∆αρ (p)Γµρσ 1 (p. ~p ′ ).

The model is renormalizable only up to one loop. Using this propagator we found a scattering amplitude. (8). the corresponding theory will be in principle renormalizable. to be done in the next section. Had we employed a topological mass instead of the Proca’s. (19) is a Pauli interaction between each vector particle and the magnetic field produced by the other. In the anti-parallel case these effects cancel each other. Of course these comments do not apply if the mass has a topological origin. shows that the degree of superficial divergence is actually lowered. In spite of the improved behavior. Let us now consider a model in which the mass of the vector particles has a topological origin. i. (21) As this propagator has a better ultraviolet behavior than (11). One could use (9) but with the Proca mass replaced by a topological one. In the case of minimal coupling we get a result which contains an additional numerical factor. 4ie2 |M| [s + 2i cot(α)] . The two possible couplings give different cross sections and. 1/4 . radiative correction should produce diverse cross sections even in the cases associated to (19) and (22) where the corresponding Lagrangians differ only by the mass terms. Actually.e. Besides that. the model still suffers from renormalization problems due to the divergence of graphs with Nφ equal to four and six. One can then envisage two possibilities to introduce the coupling to the CS field. Taking into consideration this fact we conclude that the effective degree of divergence for a generic one loop graph γ is 1 d(γ) = 3 − NA − Nφ .. In the case of the Lagrangian with a Proca mass. as established by (23). This is apparent from an inspection of the asymptotic behavior of the matter field vector propagators. However. the Feynman rules would be the same unless for the propagator for the matter field which would become ∆αβ M CS pα pβ −i pρ αβ g − = 2 + iMεαβρ 2 2 2 p − M + iǫ p p " # . there is no way to select a preferred one. the origin of the constant term in Eq. If the corresponding counterterms are added to (9) then the relations (14-16) will not be able to guarantee renormalizability even at one loop. at the present. Similarly to the spin 1/2 case.where α is the scattering angle and s = s1 + s2 is the total spin of the incoming particles. 2 (23) where NA and Nφ are the number of external lines belonging to the CS and to the matter vector field. θ (22) which differs from (19) just by a phase factor. higher order loops will contain nonrenormalizable divergences. in the front of (22). the longitudinal term in the propagator spoils renormalizability. a more careful analysis. 5 . M 2 φ†µ φµ → M µνρ † M ε φµ ∂ν φρ + εµνρ φµ (∂ν φρ )† 2 2 (20) or just consider the MCS field described by (2) minimally coupled to a CS field. one should use the polarization vector given in Eq. In such situation.

p)Dµν (k). any one loop graph with nCS and nφ internal CS and matter field lines. We obtain Σαβ (p) = e2 F (p)εαβρ pρ . because of Eq. ONE LOOP RENORMALIZATION As we will show now. After establishing (23) we shall now examine each case of divergence. 2nCS + NCS = V1 + 2V2 . We shall examine separately each one of them. and containing V1 and V2 trilinear and quadrilinear vertices has the degree of superficial divergence given by d(γ) = 3 − nCS + V1 = 3 − Nφ + V1 . due to (16) the contraction of the longitudinal part of the matter field propagator with the vertex Γ1 produces a result whose degree in the loop momentum is reduced by V1 . p + k)∆ (p + k. α 1 1 (2π)3 (28) Although the degree of divergence indicates a quartic divergence. (14) and (15) imply that the sum of graphs with NCS external CS lines will contain th the NCS power of the external momenta and therefore in such sum the degree of divergence is effectively reduced by NCS . However the radiative corrections to the two and to the four point vertex function of the matter field. (M + p ) − F (p) = 2Mp2 2M p2 M − p2 6 (30) . By power counting. 8πθ (29) where (from now on we assume M to be positive) " √ !# M + p2 1 (3M 2 + p2 ) 2 2 2 2 2 √ (M − p ) Log √ . Notice that logarithmically divergent parts are odd in the integration variables and therefore vanish under symmetric integration. Symmetric integration show immediately that the contributions of the graphs 2(b) and 2(c) vanish. for the cases under consideration. (16). 2. This implies that. Besides that. with d(γ) ≥ 0. up to one loop. 2nφ + Nφ = 2V1 + 2V2 .III. already acts as a renormalization. i. the vacuum polarization and the trilinear vertex all have d(γ) > 0. (9) have an effective degree of divergence as given in (23). We will employ dimensional regularization which. e. (25) (26) (27) Formula (23) follows now from the following observations: 1. as specified by (23). the one loop contributions to the amplitudes for the theory defined by Eq. the six point vertex function of the matter field and the four point function with external lines associated to two CS field and two matter fields do not generate counterterms. 2. Let us begin by the selfenergy contributions which in one loop correspond to the graphs shown in Fig. 2 (24) where we used nCS + nφ = V1 + V2 . The amplitude associated to 2(a) is (details of this calculation are presented in the Appendix) αβ Σ (p) = Z d3 k ναα′ µβ ′ β ′ β ′ (p + k)Γ Γ (p. the above expression diverges only quadratically.

∆αβ R Σ′ (p2 ) −i αβ g − 1 − = 2 p − M 2 − Σ′ (p2 ) p2 " ! pα pβ − iΘ(p)εαβρ pρ . dx q M 2 − p2 x(1 − x) 7 (39) . 2 8πθ (p − M 2 ) (34) One sees that the needed counterterm corresponds to a CS term for the matter field. (2π)3 2 Z (36) After some calculations whose details are relegated to the appendix. k)∆ = αα 1 1 (2π)3 Z d3 k µναβ = Γ ∆αβ (k) . Π(p2 ) = i (p2 − 4M 2 ) h 2 2 4M − (p + 4M )I . Log 2Mp2 2M p2 M − p2 (32) In this situation we obtain the renormalized propagator. (35) with Πµν A Πµν B d3 k µβα να′ β ′ ′ (k)Γ (k. p + k)∆β ′ β (p + k) Γ (p + k. (37) which is transversal as required by current conservation. Thus we define a renormalized amplitude by αβ αβρ Σαβ pρ R (p) = Σ (p) − δMε 2 e FR (p)εαβρ pρ . In the last expression. Let us now focus our attention on the vacuum polarization contributions to the CS field two point function. To secure that the physical mass is M one still have to proceed a finite renormalization. 2 M # (33) where Σ′ (p2 ) = e2 p2 FR2 (p) 8πθ (p2 − M 2 ) and Θ(p) = e2 FR (p) . The associated graphs have been drawn in Fig. = 8πθ where δM = e2 M 2πθ (31) and " √ !# (M 2 − p2 ) (M 2 − p2 )(3M 2 + p2 ) M + p2 2 2 √ √ FR (p) = 3M − p − .Although finite the above expression does not vanish at p2 = M 2 . we get Πµν pµ pν ie2 Π(p2 ) g µν − 2 = 8π p ! . We have µν Πµν = Πµν A + ΠB . 3. 0 4M 2 (38) with I0 = Z 0 1 1 .

(44) + pµ 2 2 3M M M 3M 2  ′ Γµαβ a (p. performing the calculations on the matter field’s mass shell we get −e3 M 5 βµ ασρ 5 αµ βσρ 14 αβ µσρ g ε pσ p′ρ + g ε pσ p′ρ − g ε pσ p′ρ 2 2 8πθ 4M 4M 4M 2     17 αµρ 15 αµρ ′ 15 βµρ 17 βµρ ′ ′α ε p − ε p + p ε p + ε p + pβ − ρ ρ ρ ρ 24M 2 8M 2 8M 2 24M 2     1 1 αβρ 2 αβρ ′ 2 αβρ ε pρ + 2 εαβρ p′ρ + p′µ ε pρ + ε pρ . 4. p )D (k) ∆ Γ σρ αβ 1 (2π)3 2 (43) These expressions are in general very complicated but in the low momentum regime a great simplification occurs.For low momentum Πµν approaches the expression. Indeed. 2 8πθ 3M 3M 2   (45) (46) It can be easily verified that these results satisfy current conservation as expressed in the Ward identity e dΣαβ (p) . Γµαβ (p. p′ )Dσρ (k) . p ) = d3 k h ραα′ Γ (p. p′ ) = c   4 ′α αµρ ′ 2 ′µ αβρ ′ −e3 M −2εαβµ − p ε pρ − p ε pρ . analogously to what happens in the spin 1/2 case. p′ ) = b Γµαβ (p. Πµν =  ie2  µν 2 g p − pµ pν . The lowest order contributions to the trilinear vertex come from the first three graphs shown in Fig. p ) = Γµαβ (p. p′ ) b −e3 M 4 β αµρ 2 µ αβρ = −2εαβµ + p ε pρ − p ε pρ 2 8πθ 3M 3M 2 Γµαβ (p. p + k)∆ D (k) αβ σρ 2 (2π)3 1 (42) i d3 k h µραα′ σβ ′ β ′ ′ ′ ′ ′ (p + k)Γ (p + k. p + k)∆α′ ρ′ (p + k) (2π)3 1 Z ′ ′ i ′ Γ1µρ σ (p + k. p′ ) = c Z Z (41) i d3 k h ραα′ µσβ ′ β ′ ′ Γ (p + k)Γ (p. p′ + k)∆σ′ β ′ (p′ + k)Γ1σβ β (p′ + k. The corresponding analytic expressions are respectively ′ Γµαβ a (p. 6πM (40) implying that the effective low momentum Lagrangian contains a Maxwell term.


µαβ .

p) + Γµαβ (p. a (p. p) . = Γµαβ (p. to disentangle the various contributions. here it is also convenient to do the calculations 8 . Usually this done by coupling the matter field to an external electromagnetic field. p) + Γb c d(p)µ p2 =M 2 (47) We are now in a position to calculate the vector meson anomalous magnetic moment. Besides that.

−~p) = ε∗β (−~p. −e3 M 1 1 − 2 pβ εαiρ pρ + 2 g iβ εασρ pσ p′ρ 8πθ M M   3 1 iα βσρ ′ −e M 1 α βiρ iαβ p ε pρ + 2 g ε pσ pρ . −~p) = ε∗β (−~p. The gyromagnetic factor may be changed (sometimes g = 2 is desirable [15]) if the magnetic term [16] Lmag = −ieγ(∂µ Aν − ∂ν Aµ )φ∗µ φν (51) is added to (9). s)Γiαβ p. s)Γiαβ (~p. (16) the ultraviolet behavior of the Green functions is definitely wrecked. −~p) . s) . After these remarks let us proceed toward the computation of the anomalous magnetic moment. From the expressions for the three graphs given above we have Γia−c (~p. (49) and the limit prescription singles out the term proportional to ~q. as the new vertex (51) does not obey Eq. p′ )∆α′ β ′ (p′ )ΣβR β (p′ ) . Γµαβ (p. it is easy to check that the magnetic moment for a spin one particle becomes e(1 + γ)/(2M). the Breit frame. −~p)εα (~p. p′ ) = Γµαα 1 e However. if one insists in having g = 2 another approach becomes mandatory. The above result means that the gyromagnetic factor is g = 1. we get 9 (55) (56) . −~p) = 8πθ M 2 M Γiαβ p. In such case. s) = a+b+c εα (~ 3e3 ij ǫ qj . 2M 2M (50) where s = ±1 is the spin. saturating with external polarizations. p ) = ΣR (p)∆α′ β ′ (p)Γ1 and ′ h ′ ′ i (53) i (54) (p. p′ ) d (p. 2M q µ = lim q→0 (48) where Γi (~p. Nevertheless. For this reason we shall not anymore consider the possibility of adding (51). In the tree approximation we get µ= e e gs = ± . 8πθ (52) We still have to add the contributions of the graphs 4(d) and 4(e) which corresponds to the analytic expressions h ′ µβ β ′ αα Γµαβ (p. Γe (~p. Thus. It is then found that the magnetic moment is given by [14] 1 qj ǫij 2 Γi (~p. −~p) = d (~   so that.in a particular frame where ~p + p~ ′ = 0. for small momenta and in the Breit frame.

in the low energy regime we get Mas1=1.s2 =1. s) Γiαβ p. It is also worthwhile to remark that. (63) . DISCUSSION The calculation of the radiative corrections to the propagators and vertices done in the previous section allow us to incorporate vacuum polarization and anomalous magnetic moment effects in the two body scattering amplitudes computed earlier. (59) It should be noticed that the graphs with self energy corrections contribute significantly to the final result. the magnetic moment for vector particles of spin ±1 is e2 e 1± µ=± 2M 8πθ " # . In fact.s2 =−1. 8πθ (58) Thus.s2 =−1. 8πθ (57) The sum of Eqs (52) and (57) gives Γi (~p.s4 ie4 M = 8πθ2  1 + s4 [(−1 + s3 ) 1 − cos α −2s3 cos α + (1 + s3 ) cos2 α + 2i sin α − i(1 + s3 ) sin α cos α 10 i  .s3 .s4 ie4 M 1 + s4 = (cos α − i sin α) 2 8πθ 1 − cos α  h −2 + (1 − s3 ) cos α + (1 + s3 ) cos2 α − i(1 + s3 ) sin α cos α Mas1 =−1. −~p) εα (~p. In our case. as in the spin 1/2 case.s4 ie4 M = 8πθ2  −1 + s4 [(1 + s3 ) 1 − cos α 2 −2s3 cos α − (1 − s3 ) cos α + 2i sin α − i(1 − s3 ) sin α cos α Mas1 =−1. −~p) = e3 ij ε qj . as can be seen from (29) a CS term is produced.h i ε∗β (−~p.s3 .s3 . This is similar to what happens for the case of spin 1/2 particles in the Coulomb gauge and results from the fact that in both cases the interaction with the CS field modifies the free propagators in an essential way. IV.s4 ie4 M −1 + s4 = (cos α + i sin α) 2 8πθ 1 − cos α i  (60) i  (61) i  (62)  h 2 2 − (1 + s3 ) cos α − (1 − s3 ) cos α − i(1 − s3 ) sin α cos α Mas1=1. −~p) + Γiαβ p. the anomalous magnetic moment has the same sign for both spin up and spin down situations [10].s2 =1.s3 . up to one loop. s) = − d (~ e (~ 2e3 ij ǫ qj .

s3 . Let us begin considering Eq. As argued in the introduction. For the vacuum polarization diagram. 5a. 5b is obtained from the above expressions by exchanging s3 and s4 . We have indicated other possible scheme as the use of the complex MCS model minimally coupled to a CS term. 5e and 5f) and box (Fig. Fig. 5d. (64) The next diagram. k → k + px one finds −ie2 Z 1 Z d3 k Lαβ Σ (p) = dx . These graphs present a logarithmic ultraviolet divergence (by Lorentz covariance the would be linear divergence does not appear). in this work we have studied some properties of spin one particles interacting through a CS field. we found the result Mcs1 . Although this model is power counting renormalizable. ACKNOWLEDGMENTS This work was supported in part by Conselho Nacional de Desenvolvimento Cient´ıfico e Tecnol´ogico (CNPq). 5c. APPENDIX A: In this appendix we shall present some details of the calculations of the vector meson self-energy and of the CS polarization tensor. this is a complex system but we found some simplifications which allowed us to study its behavior up to the one loop level.s2. the algebraic structure is very much cumbersome that no practical results are possible even at the one loop level.for the contributions from the vertex corrections indicated in Fig. In such situation the scheme adopted by us seems to be the most useful although not being easily generalizable to higher orders. θ (2π)3 [k 2 − C 2 ]2 0 αβ (A2) where C 2 = M 2 (1 − x) − p2 x(1 − x) and n Lαβ = [3M 2 + xp2 ](εσβρ kσ k α + εασρ kσ k β )pρ + [(1 − x)M 2 p2 o − (1 − x)(pµ k µ )2 (x2 − x3 )p4 ]εαβσ pσ + 2xp2 εαβρ kρ kσ pσ /M 2 . 5g and 5h) diagrams. Besides that they are so much intricate that even using dimensional regularization and employing an algebraic computer program the integrals did not turned out to be feasible. There are two denominators so that employing Feynman’s trick 1 = AB Z 0 1 dx 1 [(A − B)x + B]2 (A1) and changing the integration variable. (28). does not contribute at leading order of |~p|/M. 11 (A3) .s4 = 2ie4 M [(1 + s1 s3 ) cos α − i(s1 + s3 ) sin α] 3πθ2 [(1 + s2 s4 ) cos α − i(s2 + s4 ) sin α]. To summarize. Coordena¸c˜ao de Aperfei¸coamento de Pessoal de N´ıvel Superior (CAPES) and Funda¸c˜ao de Amparo `a Pesquisa do Estado de S˜ao Paulo (FAPESP). The result for Fig. Fig. To complete the one loop calculation it is still necessary to compute the contributions of the triangle (Fig.

The parametric integration can be done exactly producing the result (37). results Πµν = e2 Z 1 Z d3 k N2µν dx . M2 0 (2π)3 [k 2 − a2 ]2 (A7) where. Let us now turn to the polarization tensor. dx 8πθM 2 0 M 2 (1 − x) − p2 x(1 − x) (A4) The computation of the remaining parametric integration is straightforward and produces the result (30). dx 2 [M − p2 x(1 − x)]1/2 (A9) with N3µν = M 2 (2 − 12x + 8x2 )[pµ pν − p2 g µν ] + p2 [pµ pν (−x + 7x2 − 6x3 ) + p2 g µν (−x − x2 + 2x3 )] .Using dimensional regularization one can perform the momentum integration to get Σαβ (p) = e2 Z 1 6M 4 + [p4 − M 2 p2 − 6M 4 ]x + [M 2 p2 − 7p4 ]x2 + 6p4 x3 αβσ q ε pσ . translating the integration variable. as before. M 2 (2π)3 [k 2 − M 2 ][(k + p)2 − M 2 ] (A5) where N1µν = [k 2 (pµ k ν + k µ pν ) − 2k µ k ν k α pα ] + [8M 2 k µ k ν − (4M 2 g µν + pµ pν )k 2 + k α pα (2pµ k ν + 2k µ pν − 2g µν k β pβ ) + p2 (−3k µ k ν + 2g µν k 2 )] + [3M 2 (pµ k ν + k µ pν − 6g µν k α pα )] + [M 2 (4M 2 g µν + pµ pν − 3p2 g µν )] . 12 (A10) . (A8) With the help of dimensional regularization this becomes µν Π ie2 = 8πM 2 Z 0 1 N3µν . From (35) and (36) and proceeding analogously to what we have done before we arrive at Πµν = e2 Z d3 k N1µν .k → k − px and deleting the terms odd in k. a2 = M 2 − p2 x(1 − x) and N2µν = M 2 [4M 2 g µν + pµ pν (1 − 6x + 8x2 ) + p2 g µν (−3 + 6x − 4x2 )] + k µ k ν [8M 2 − 3p2 + 2p2 x] + k 2 [−4M 2 g µν − pµ pν (1 + 2x) + 2p2 g µν ] + [2k µ pν k α pα + 2pµ k ν k α pα − 2g µν (k α pα )2 ] . (A6) Employing again Feynman’s formula (A1).

B185. [7] Y. Phys. Tsai. 253 (1994). Nucl. Gomes. B338. M. C. Townsend. Gomes. Phys. Schonfeld. 3750 (1995). D 42. Yang Phys. Phys. Paul and A. K. 884 (1993). Rev. Li. J. J. A11. Phys. Hagen. 115. [3] P. Rev. [6] M. C. Rev. Lett. (1997). Rev. Chen and M. 48. Lett. S. J. da Silva. 128. Phys. 58. Phys. Lozano. B279. D 7. R. Rev. O. L. Phys. Mod. R. Ann. M. Phys. B171 244 (1986). D 46. Goldhaber. Phys. V. Schwinger. ibid Phys Rev. Rev. Phys. A. Phys. 38 (1984). Schaposnik. [5] S. Glashow. 485 (1959). M. Nucl. B139 371. R. J. Lett 70. R. Rev. Gomes. Phys. 3529 (1992). Rev. Rothe and K. Lett. 514 (1987). Phys. Phys. L. Hagen. S. 2825 (1996). K. Bergman and G. R. Gomes and A. N. Rev. E. Phys. 10. D 56. C. Kim and Y. J. 107 (1959). [16] H. D 52. 045002 (1999). Rev. R. P. S. 953 (1940). 6339 (1997). Lett. D. H. Nascimento and A. (1984). Gomes and A. D 55. D 57. Malbouisson and A. Deser. Lett. 3524 (1990). J. Rev. 2250 (1997). 140. Phys. da Silva. Girotti. [4] R. 3579 (1998). Int. I. 3623. Porrati. [9] M. Lett. Pilch and P. L. Bohm. Phys. V. Phys. Batalin and E. D 59. Khare. [2] W. D 55. O. Horner and A. S. 157 (1981). [8] O. Fradkin. Malacarne and A. Rev. 439. 137 (1998). Templeton. 5951 (1997). H. J. da Silva. S. H. B136. Phys. 885 (1962) 13 . 229. Rev. [14] K. 275 (1995). F. Nucl. da Silva. T. [12] M. M. Phys. Lett. Phys. J. Pinheiro and A. Phys. Phys. Girotti. Fradkin and F. Ferrara and M. Rev. Ann. Van Nieuwenhuizen. Aharonov and D. da Silva. Hagen and S. 372 (1982). Mod Phys. D 56. Fleck. M. da Silva. Alves. S. D. A 6 3119 (1991). Ramaswamy. Rothe.S. L. S. J. Deser. Devecchi. [13] F. [10] C. Corben and J. 975 (1982) and Ann.REFERENCES [1] S. 416 (1994). K. [15] S. Gomes. Jackiw. Jackiw and S. [11] C. A. 242. M. J. Phys. 3710 (1973). J. J. S. Phys. Banerjee. Lee and C.

FIGURES FIG. 14 . Feynman rules for the Proca model minimally coupled to a CS field. 3. 1. 4. 2. Graphs contributing to the vector meson anomalous magnetic moment. FIG. 5. FIG. Graphs contributing to the one-loop correction to the vector meson propagator. One loop vacuum polarization graphs associated to the CS field. One loop contributions to the vector meson scattering. FIG. FIG.