π 0 → 2γ decay under strong magnetic fields in the NJL model

Máximo Coppola ,1 Daniel Gomez Dumm ,2 and Norberto N. Scoccola 1, 3

1
Departamento de Fı́sica Teórica, Comisión Nacional de Energı́a Atómica,
Av. Libertador 8250, (1429) Buenos Aires, Argentina
2
IFLP, CONICET—Departamento de Fı́sica, Facultad de Ciencias Exactas,
Universidad Nacional de La Plata, C.C. 67, 1900 La Plata, Argentina
3
CONICET, Rivadavia 1917, (1033) Buenos Aires, Argentina
arXiv:2507.13560v1 [hep-ph] 17 Jul 2025

Abstract
0
We study the anomalous π → γγ decay under an external uniform magnetic field in
the framework of a two-flavor Nambu-Jona-Lasinio model. It is seen that the full decay
width gets strongly reduced with the external field, and that the differential width is almost
independent of the direction of the outgoing photons. We also find that the result for the
total width can be very well approximated by a simple expression obtained at the lowest
order in the chiral limit, which is just a direct extension to finite magnetic fields of the well
known B = 0 result that follows from the anomalous Wess-Zumino-Witten action. As a
result, the magnetic suppression of the width can be understood in terms of the evolution
of both the neutral pion mass and decay constant with the magnetic field.
 2

I. INTRODUCTION

In the past few decades there has been an increasing interest in the effect of intense magnetic
fields on the physics of strong interactions [1–4]. To a great extent this is due to the impact that
such large magnetic fields might have in the study of the early Universe [5, 6], in the analysis of
high energy noncentral heavy ion collisions [7, 8] and in the description of compact stellar objects
like magnetars [9, 10]. In fact, magnetic fields have been shown to induce several interesting
phenomena in quantum chromodynamics (QCD), such as the chiral magnetic effect [11–13], the
enhancement of the QCD quark-antiquark condensate (magnetic catalysis) [14], the decrease of
critical temperatures for chiral restoration and deconfinement QCD transitions (inverse magnetic
catalysis) [15], etc.

Moreover, external magnetic fields can be used to probe QCD dynamics. In this context, the
influence of these fields on the properties of light hadrons has attracted a lot of attention. The
corresponding analyses are usually involved, since first-principle theoretical calculations require to
deal in general with QCD in a nonperturbative regime, and consequently a variety of theoretical
approaches have been considered. For example, the effect of intense external magnetic fields on
light pseudoscalar meson masses has been studied in the framework of Nambu-Jona-Lasinio (NJL)-
like models [16–34], quark-meson models [35–38], chiral perturbation theory (ChPT) [39–41], path
integral Hamiltonians [42, 43], effective chiral confinement Lagrangians [44, 45] and QCD sum
rules [46]. In addition, several results for the spectrum of these mesons in the presence of back-
ground magnetic fields have been obtained from lattice QCD (LQCD) calculations [15, 47–52].

The effect of a magnetic field on the decay widths of light pseudoscalar mesons is a related topic
that has also been considered in the literature. However, most of the existing works have been
devoted to the analysis of the leptonic decays of charged pions [53–58]. The main conclusion of
Refs. [55, 58], where actual predictions for these processes are given, is that the presence of the
external field leads to a great enhancement of the corresponding decay widths. On the other hand,
the influence of a magnetic field on the anomalous decay π 0 → γγ has received far less attention.
In the past few years this problem has been studied in Refs. [59, 60] in the framework of ChPT,
where the validity of the analysis is restricted to values of eB up to ∼ m2π . At this point we should
emphasize the importance of the π 0 → γγ decay in the context of chiral theories. In fact, it was
through a careful analysis of the corresponding decay amplitude (in vacuum) that anomalies were
first discovered [61, 62], and this led to the realization that the U(1)A symmetry is not conserved
in quantum field theories like QCD. Having this in mind, it is clear that the understanding of how
 3

the π 0 → γγ decay width gets modified by the presence of an external magnetic field appears to
be a quite relevant issue.

⃗ on the π 0
The aim of the present work is to study the influence of a strong magnetic field B
anomalous decay using a two-flavor NJL model. We remind that the NJL model is an effective
chiral quark model of QCD in which gluon degrees of freedom are integrated out in favor of
some local quark-antiquark interactions that lead to the dynamical breaking of chiral symmetry.
The model is non-renormalizable and, therefore, a regularization procedure has to be adopted
in order to deal with divergent quantities. The choice of this procedure may be considered as
part of the definition of the model itself. In this work we adopt a magnetic field independent
regularization (MFIR) method [63–65], which has been shown to be convenient when dealing with
magnetized systems [66]. While in the NJL model quark condensates emerge from the usual mean
field approximation, mesons are treated as excitations whose masses are obtained by considering
second order corrections in the corresponding bosonized action. The quark propagator in the
presence of a uniform magnetic field, considered in this work in the Schwinger form [2, 3], is used
to evaluate the π 0 polarization function and its decay amplitude to two photons. As an additional
ingredient, we consider the case of B-dependent effective quark-quark coupling constants; this
possibility has been previously explored in effective models [67–69] in order to reproduce the inverse
magnetic catalysis effect observed at finite temperature in LQCD calculations.

This paper is organized as follows. In Sec. II we present the theoretical formalism, introducing
the magnetized two-flavor NJL-like lagrangian to be used in our calculations. Expanding the
bosonized action in powers of fluctuations around mean field values up to quadratic order, we
derive expressions for certain relevant quantities associated to the π 0 → γγ decay width, such as
quark masses, quark propagators, the quark-pion coupling strength and the neutral pion mass. In
Sec. III we calculate the magnetized π 0 → γγ decay width from the third order expansion of the
bosonized action. We also show that in the chiral limit the resulting coupling constant gπ0 γγ has a
simple expression that turns out to be a direct extension to finite magnetic field of the well-known
B = 0 result that follows from the gauged Wess-Zumino-Witten action. In Sec. IV we present and
discuss our numerical results, while in Sec. V we provide a summary of our work, together with
our main conclusions. We also include appendixes A, B and C to provide some technical details of
our calculations.
 4

II. NJL MODEL UNDER STRONG MAGNETIC FIELDS: MEAN FIELD

PROPERTIES, π 0 MASS AND COUPLING TO QUARKS

The Lagrangian density of the NJL two-flavor model, including couplings with an electromag-
netic field, is given by
h
2
2 i 1
L = ψ̄ (iD
/ − m0 ) ψ + G ψ̄ψ + ψ̄ iγ5⃗τ ψ − Fµν F µν , (1)
4

where ψ = (u d)T , τi are Pauli matrices and m0 is the current quark mass, which is assumed to
be equal for u and d quarks. The interaction between fermions and the electromagnetic field Aµ is
driven by the covariant derivative

Dµ = ∂µ + i Q̂Aµ , (2)

where Q̂ = diag(qu , qd ), with qu = 2e/3 and qd = −e/3, e > 0 being the proton electric charge.
In addition, Eq. (1) involves the electromagnetic tensor F µν = ∂ µ Aν − ∂ ν Aµ . The gauge field
Aµ = Aµ + aµ contains both the external classical field Aµ and dynamical (i.e. propagating)
degrees of freedom aµ . We consider the particular case in which the external electromagnetic field
⃗ orientated along the positive 3-axis.
is given by an homogeneous stationary magnetic field B
Since we are interested in studying a meson decay, it is convenient to bosonize the fermionic
theory, introducing a scalar field σ(x) and pseudoscalar fields ⃗π (x) while integrating out fermion
fields. The bosonized action can be written as
Z
1
d4 x σ(x) σ(x) + ⃗π (x) · ⃗π (x) + SEM ,

 
Sbos = −i ln det iD − (3)
4G

with
h i
iDx,x′ = δ (4) (x − x′ ) iD
/ − m0 − σ(x) − i γ5 ⃗τ · ⃗π (x) , (4)

where a direct product to an identity matrix in color space is understood. The pure electromagnetic
field contribution to the effective action is given by

B 2 (4) 1
Z Z
1 4 µν
 
SEM = − d x Fµν F = − V + d4 x aµ (x) η µν ∂α ∂ α − ∂ µ ∂ ν aν (x) , (5)
4 2 2

where the first and second terms in the rhs correspond to the classical and propagating contribu-
tions, respectively. For the Minkowski metric, the convention η µν = diag(1, −1, −1, −1) is adopted.
Since we consider a static and uniform external magnetic field, in what follows the classical con-
tribution to SEM will be dropped.
 5

We proceed by expanding the bosonized action in powers of fluctuations of the bosonic fields
around their mean field (MF) values. We assume that the scalar field σ(x) has a nontrivial trans-
lational invariant MF value σ̄, while vacuum expectation values of pion fields πi (x) are π̄i = 0 so
as to respect the parity symmetry of vacuum. Thus, we write

MF, u MF, d


Dx,x′ = diag Dx,x ′ , Dx,x′ + δDx,x′ . (6)

The MF piece, which is diagonal in flavor space, includes the interaction of quark fields with the
classical electromagnetic field Aµ . One has

MF, f
= −i δ (4) (x − x′ ) i/∂ − Qf A(x)
 
Dx,x ′ / −M , (7)

where M = m0 + σ̄ is the quark effective mass and f = u, d. On the other hand, δDx,x′ reads
h i
δDx,x′ = iδ (4) (x − x′ ) Q̂ a(x)
/ + δσ(x) + i γ5 ⃗τ · δ⃗π (x) . (8)

Replacing in the bosonized effective action and expanding in powers of meson fluctuations
around the MF values, we get

MF quad cub
Sbos = Sbos + Sbos + Sbos + ... (9)

The MF action per unit volume is given by

MF
σ̄ 2 −1
Z
Sbos iNc X 4 4

MF, f
= − − d x d y TrD ln Sx,y , (10)
V (4) 4G V (4)
f =u,d

MF, f MF, f

−1
where Nc is the number of colors, TrD stands for trace in Dirac space, and Sx,y = iDx,y is
the MF quark propagator in the presence of the classical magnetic field. We choose to use for this
propagator the Schwinger form, given by
Z
MF, f
Sx,y = eiΦf (x,y) e−i p(x−y) S̄f (p) , (11)
p

where
Z ∞   
2 2 2 tan(σBf )
S̄f (p) = − i dσ exp −iσ M − p || + p⃗⊥ − iϵ
0 σBf
 
1 2 p⃗⊥ · ⃗γ⊥
× (p || · γ || + M ) (1 − sf γ γ tan(σBf )) − , (12)
cos2 (σBf )

with Bf = |qf B| and sf = sign(qf B). Here we have defined “parallel” and “perpendicular” four-
vectors

pµ|| ≡ (p0 , 0, 0, p3 ) , µ
p⊥ ≡ (0, p1 , p2 , 0) , (13)
 6

and we have introduced a shorthand notation for the momentum integral

d4 p
Z Z
≡ . (14)
p (2π)4

We have also used the notation p || · γ || = p0 γ 0 − p3 γ 3 and p⃗⊥ · ⃗γ⊥ = p1 γ 1 + p2 γ 2 . The gauge
dependence of the propagator is encoded in the so-called Schwinger phase Φf (x, y) associated to
the classical external field
Z y  
1
Φf (x, y) = qf dξµ Aµ (ξ) + FAµν (ξν − yν ) , (15)
x 2

where FAµν is the corresponding electromagnetic strength tensor.

MF
As usual, from the minimization condition ∂Sbos /∂ σ̄ = 0 one obtains a gap equation, which can
be written as

X Z
M = m0 + 2 i GNc TrD S̄f (p) . (16)
f =u,d p

This is a gauge-independent expression, as it should. It involves a momentum integral that is diver-

gent and has to be properly regularized. As stated in the Introduction, we use the magnetic field
independent regularization (MFIR) scheme: for a given unregularized quantity, the corresponding
(divergent) B → 0 limit is subtracted and then added in a regularized form. Thus, each quantity
can be separated into a “B = 0” term and a “magnetic” contribution. The divergent “B = 0”
piece is then regularized by introducing a 3D momentum cutoff Λ. Notice that, in general, the
“B = 0” term still depends implicitly on B (e.g. through the values of the dressed quark masses
M ); hence, it should not be confused with the value of the studied quantity at vanishing external
field. The resulting regularized form of the gap equation can be found e.g. in Ref. [29].
Next, we consider quadratic terms in the expansion of Sbos in powers of δDx,x′ . We omit the
description of charged pion contributions, since charged pions cannot mix neither with the photon
nor the π 0 . In contrast, neutral pions are in general expected to mix with photons [59]. Here we
consider only lowest-order dynamical electromagnetic contributions to the π 0 → γγ decay; thus, we
dismiss possible π 0 − γ mixing effects. As usual, the π 0 effective propagator can be obtained from
a resummation of quark loop chains, using the random phase approximation. Since we are dealing
with a neutral particle, the contributions of Schwinger phases associated with quark propagators
[see Eq. (11)] cancel out, leading to gauge invariant polarization functions that depend only on the
difference x − x′ (i.e., they are translationally invariant). If one performs a Fourier transformation,
the conservation of momentum implies that the polarization function is diagonal in the momentum
 7

basis. Thus, in this basis we have

Z  
1 0 1
Sπ 0 = − δπ (−q) − Ππ0 (q || , q⊥ ) δπ 0 (q) ,
2 2
(17)
2 p 2G
where
X
Ππ0 (q 2|| , q⊥
2
TrD iS̄f (p+ ) iγ5 iS̄f (p− ) iγ5 .
 
) = −iNc (18)
f

Here p± = p ± q/2, while S̄f (p) is given in Eq. (12). Once again, employing the MFIR scheme we
regularize the above polarization function by separating Πreg
π0
= Πmag
π0
+ Π0,reg
π0
. The corresponding
expressions are given in App. A. Choosing the frame in which the π 0 meson is at rest, its mass can
be obtained by solving the equation
1
− Πreg
π0
(m2π0 , 0) = 0 . (19)
2G
As final remark, we recall that the pion field wave function has to be renormalized. Notice that
the external magnetic field introduces a spatial anisotropy reflected in the momentum dependence
2 ), which distinguishes parallel from perpendicular components. We introduce a (parallel)
Ππ0 (q 2|| , q⊥
1/2
“wave function renormalization constant” Z || , defined by fixing the residue of the two-point
function at the pion pole. This constant, and the related quark-pion coupling constant gπ0 qq , are
given by [29]
∂Ππ0
Z ||−1 = gπ−2
0 qq = . (20)
∂q 2|| q2 2
|| = m 0π
2
q
⃗⊥ =0

III. π 0 → γγ DECAY WIDTH

A. Contributions to the decay amplitude

For a pion of momentum q decaying into two photons of momenta p and k, the decay width is
given by
d3 p d3 k |iT |2
Z
1 1 X
Γπ0 γγ (B) = , (21)
2Eπ 2 (2π)3 2p0 (2π)3 2k 0 VT
λp ,λk =±

where λp , λk are the helicities of the outgoing photons, while V and T are the volume and time
interval in which the interaction is active. Note that a factor 1/2 has been included to account
for the identical nature of the outgoing photons. The transition amplitude is given by the relevant
Ŝ-matrix element between the initial and final states

iT = ⟨γ(p, λp )γ(k, λk )|(Ŝ − 1)|π 0 (q)⟩ . (22)

 8

By properly expanding the fields in terms of creation and annihilation operators we get
Z
∗ ∗
iT = ϵµ (p, λp ) ϵν (k, λk ) d4 x d4 y d4 z eipz eiky e−iqx ⟨0| T[J µ (z)J ν (y)J5 (x)] |0⟩ , (23)

where ϵµ (p, λp ), ϵµ (k, λk ) are the photon polarization vectors, T is the time-ordering operator and
the quark currents are given by

J µ (x) = ψ̄(x)(−Q̂γ µ )ψ(x) , J5 (x) = ψ̄(x)(−gπ0 qq iγ5 τ3 )ψ(x) . (24)

To obtain the connected 3-point correlation function in Eq. (23), we first add to the Lagrangian in
Eq. (1) auxiliary terms that include source functions αµ (x) and α5 (x),

L̂ = L + αµ (x)J µ (x) + α5 (x)J5 (x) , (25)

and then we take functional derivatives of the corresponding extended bosonized action with respect
to the auxiliary sources. In this way we have

δ 3 Ŝbos
⟨0| T[J µ (z)J ν (y)J5 (x)] |0⟩ = (−i)2 . (26)
αµ (z) αν (y) α5 (x) αµ =α5 =0

After taking trace over color, flavor and coordinate spaces, we get
Z
iT (q, p, k) = ϵµ (p, λp )∗ ϵν (k, λk )∗ d4 x d4 y d4 z eipz eiky e−iqx
X h µν i
× iGf (x, y, z) + iGfνµ (x, z, y) , (27)
f =u,d

where
h i
iGfµν (x, y, z) = iNc
X
MF, f
TrD (−gπ0 qq κf iγ5 ) iSx,z (−qf γ µ ) iSz,y
MF, f
(−qf γ ν ) iSy,x
MF, f
. (28)
f =u,d

Here, the factors κu = −κd = 1 arise from the Pauli matrix τ3 in the quark axial current J5 (x).
The contributions Gfµν (x, y, z) and Gfνµ (x, z, y) to the sum in Eq. (27) correspond to the Feynman
diagrams depicted in Fig. 1.
Taking into account the form of the quark propagators given by Eq. (11), we get

Gfµν (x, y, z) = − gπ0 qq Nc κf qf2 exp i (Φf (x, z) + Φf (z, y) + Φf (y, x))
 

Z
e−is(x−z) e−ir(z−y) e−it(y−x) TrD iγ5 iS̄f (s) γ µ iS̄f (r) γ ν iS̄f (t) .
 
× (29)
rst

It is easy to see that the sums of Schwinger phases appearing in the amplitude are such that
⃗
gauge-dependent pieces cancel out. Recalling that we consider an homogeneous magnetic field B
 9

p p
z z
µ µ

q q
π0 x π0 x

ν ν
y y
k k

Diagram I Diagram II

FIG. 1. One-loop Feynman diagrams that contribute to the π 0 → γγ decay. Diagrams I and II correspond
to the amplitudes iGfµν (x, y, z) and iGfνµ (x, z, y), respectively. Internal (external) arrows indicate the charge
(momentum) flux. The interaction points in configuration space are indicated with labels x, y and z.

orientated along the 3-axis, we obtain the gauge-independent result

Φf (x, z) + Φf (z, y) + Φf (y, x) = qf (xµ − yµ )F µν (zν − xν )/2

= sf Bf ϵij3 (zi − xi )(xj − yj )/2 . (30)

At this point we note that the two amplitudes in Eq. (27) are related by a charge conjugation
ˆ namely
transformation C,

Cˆ Gfνµ (x, z, y) Cˆ−1 = Gfµν (x, y, z) . (31)

Therefore, we can write

iT (q, p, k) = − i gπ0 qq Nc ϵ∗µ (p, λp ) ϵ∗ν (k, λk )

X X Z
κf qf2 G(εsf ) (r, s, t) TrD iγ 5 iS̄εf (t) γ ν iS̄εf (r) γ µ iS̄εf (s) ,
 
× (32)
f =u,d ε=±1 rst

where the subscript εf indicates that sf has to replaced by εsf in the corresponding propagator,
as dictated by charge conjugation. The function G(εsf ) (s, r, t) is defined by
Z
G(εsf ) (r, s, t) = d4 x d4 y d4 z ei(pz+ky−qx) e−is(x−z)−ir(z−y)−it(y−x)
 
εsf Bf
× exp i ϵij3 (zi − xi )(xj − yj ) . (33)
2
After integration over space coordinates one gets
4
G(εsf ) (r, s, t) = (2π)10 δ (4) (q − p − k) δ (2) (p || + s || − r || ) δ (2) (k || + r || − t || )
Bf2
 
2εsf  1 1 1 2 2 2 1 1 1 2 2 2

× exp i (p + s − r )(k + r − t ) − (k + r − t )(p + s − r ) . (34)
Bf
 10

As can be seen, while parallel momentum is conserved at each vertex, the conservation of per-
pendicular momenta gets broken owing to Schwinger phases, which mix the interaction points.
Nevertheless, momentum is still conserved for the overall π 0 → γγ process (as indicated by the
first Dirac delta), allowing us to define the decay amplitude M(q, p, k) as

iT (q, p, k) = (2π)4 δ (4) (q − p − k) iM(q, p, k) . (35)

It is convenient to write M(q, p, k) by indicating explicitly the contraction with photon polarization
vectors. We obtain

iM(q, p, k) = ϵµ (p, λp )∗ ϵν (k, λk )∗ Ruµν − Rdµν ,

(36)

where Rfµν is given by

qf2
Z
Rfµν = − i 16 π 2 Nc gπ0 qq (2π)4 δ (2) (p || + s || − r || ) δ (2) (k || + r || − t || )
Bf2 rst
 i
X 2εsf h 1 1 1 2 2 2 1 1 1 2 2 2
× exp i (p + s − r )(k + r − t ) − (k + r − t )(p + s − r )
Bf
ε=±1

× TrD iγ 5 iS̄εf (s) γ µ iS̄εf (r) γ ν iS̄εf (t) .

 
(37)

B. Decay width calculation

Taking into account Eq. (21) and the definitions in Eqs. (35) and (36), the π 0 → γγ decay width
is given by
d3 p d3 k
Z
1 1
Γπ0 γγ (B) = (2π)4 δ (4) (q − p − k)
2Eπ 2 (2π)3 2p0 (2π)3 2k 0

2
ϵµ (p, λp )∗ ϵν (k, λk )∗ Ruµν − Rdµν
X
× . (38)
λp ,λk =±

This expression can be conveniently worked out in the pion rest frame, i.e. taking q µ = ( mπ0 , ⃗0 ).
Momentum conservation leads then to p⃗ = −⃗k and p0 + k 0 = mπ0 . We assume that outgoing
photons can be taken at tree level in QED, neglecting higher order corrections that could arise
from the presence of the external magnetic field. Hence, we consider the usual on-shell conditions
p2 = k 2 = 0.
The calculation of Rfµν in the pion rest frame is outlined in App. B. After some work we arrive
at the result
gπ0 qq e2 Nc M 0 0µνλ
Rfµν = i p ϵ pλ If , (39)
π2
 11

where we have used the convention ϵ0123 = +1. The proper-time integral If is given by

qf2 Bf ∞ ∞ ∞
1 + t1 t3 − (t1 + t3 )t2
Z Z Z
exp − (τ1 + τ2 + τ3 ) M 2
 
If = dτ1 dτ2 dτ3
e2 0 0 0 (τ1 + τ2 + τ3 ) Tf
" #
4τ1 τ3 (p0 )2 + (τ1 + τ3 )τ2 p2|| (t1 + t3 )t2 p⃗⊥2
× exp − , (40)
τ1 + τ2 + τ3 Tf Bf

where we have introduced the definitions ti = tanh(τi Bf ) and Tf = t1 + t2 + t3 + t1 t2 t3 .

The contraction with photon polarization vectors can be carried out taking into account the
expression for ϵµ (p, λp ) in App. C. One has

ϵµ (p, ±)∗ ϵν (k, ±)∗ ϵ0µνλ pλ = ∓ ie∓ 2iϕ p0 ,

ϵµ (p, +)∗ ϵν (k, −)∗ ϵ0µνλ pλ = ϵµ (p, −)∗ ϵν (k, +)∗ ϵ0µνλ pλ = 0 , (41)

where ϕ is an arbitrary phase. Therefore, from Eq. (39), we get

 2

2 gπ0 qq Nc M
Rdµν
X
∗ ∗
ϵµ (p, λp ) ϵν (k, λk ) Ruµν − = 2 m4π0 α 2
(Iu − Id )2 , (42)
π
λp ,λk =±

where α = e2 /(4π) is the fine structure constant.

As it is usually done, we can express the decay width in terms of an effective pion-to-two-photon
coupling gπ0 γγ , which will now depend on the external field. Writing

π 2 3 2
Γπ0 γγ (B) = α mπ0 gπ0 γγ (B) , (43)
4

and integrating over phase space in Eq. (38), we get

 2 Z 1
gπ0 qq Nc M
gπ2 0 γγ (B) = d cos θ (Iu − Id )2 , (44)
2π 2 0

⃗ and the three-momentum p⃗ of one of the outgoing

where θ is the angle between the magnetic field B
photons. One has then p2|| = p⃗⊥2 = m2π0 sin2 θ/4. Finally, to perform the integrals in Eq. (40) we
find it convenient to introduce new variables x, y and z, defined by

τ1 = xz , τ2 = yz , τ3 = (1 − x − y)z . (45)

In this way, the proper-time integrals can be rewritten as

qf2 Bf ∞ 1 1−x
Z Z Z
z  
If = dz 1 + t1 t3 − (t1 + t3 )t2
dx dy
e2 0 0 0 Tf

m2π0
    
2 2 2 (t1 + t3 )t2
× exp −z M − x(1 − x − y)mπ0 + sin θ zy(1 − y) − . (46)
4 Tf Bf
 12

As expected, the decay shows only an axially symmetric angular distribution, since full rota-
tional symmetry is broken by the external magnetic field. The corresponding differential decay
width reads
 2
dΓπ0 γγ (B) π 2 3 gπ0 qq Nc M
2
= α mπ0 Iu − Id , (47)
d cos θ 4 2π 2

where the dependence with θ is given by Eq. (46).

Our expression for the decay width can be evaluated in the limit of vanishing external field,
where ti → τi Bf and Tf → zBf . In this case the θ-dependence in Eq. (46) vanishes, recovering
spherical symmetry. We obtain
Z ∞ Z 1 Z 1−x
Nc M 2 −x(1−x−y)m2
gπ0 γγ (0) = gπ0 qq dz dx dy e−z [M π0
]
6π 2 0 0 0
Nc M

1  m 0  2
π
= gπ0 qq arcsin , (48)
3π 2 mπ 0 2M

in agreement with the result obtained in Ref. [70]. Notice that in that reference the usual Feynman
form of quark propagators was considered.

C. Chiral limit

As stated, it is also interesting to consider the anomalous π 0 decay in the chiral limit. Taking
the limit mπ0 → 0 in Eq. (46) we have

qf2 Bf ∞ 1 1−x
Z Z Z
z ch 2
Ifch = dz dx dy [1 + t1 t3 − (t1 + t3 )t2 ] e−z (M ) , (49)
e2 0 0 0 Tf

where M ch is the (B-dependent) effective quark mass obtained in the chiral limit m0 = 0. This
integral can be analytically performed, leading to

qf2 1
Ifch = . (50)
e2 2 (M ch )2

Replacing this expression into Eq. (44), and noting that in this case the integral over cos θ can be
trivially performed, we get

Nc
gπch0 γγ (B) = , (51)
12 π fπch
2
0 (B)

where we have made use of the generalized Goldberger-Treiman relation fπch ch

0 gπ 0 qq = M
ch , which

is shown to be valid at finite B (in the chiral limit) within the NJL model [29]. It is worth
mentioning that, in fact, three (vector and axial-vector) pion-to-vacuum form factors arise for the
 13

π 0 in the presence of the external magnetic field [56]. Here we denote by fπch
0 the one that reduces

to the usual decay constant fπ at B = 0. Moreover, Eq. (51) turns out to be a direct extension
to finite magnetic field of the well-known B = 0 result that follows from the anomalous gauged
Wess-Zumino-Witten action introduced in Refs. [71, 72].
In this way, the expression for the magnetized π 0 → γγ decay width is found to be given by

α2 m3π0 (B)
Γch
π 0 γγ (B) = , (52)
64π 3 [fπch
0 (B)]
2

where the kinematical factor m3π0 is taken to be nonzero, keeping the lowest order in the chiral
expansion. Clearly, for B = 0 this equation reduces to the well known chiral result [73].

IV. NUMERICAL RESULTS

In this section we provide numerical results arising from our calculation of the π 0 → γγ decay
width. The values of related quantities such as mπ0 and gπ0 qq are estimated using the SU(2)
version of the local NJL model. For the model parameters we take the values m0 = 5.419 MeV,
Λ = 6.395 MeV and GΛ2 = 2.136, which (for vanishing external magnetic field) correspond to
a quark effective mass M = 350 MeV and a quark-antiquark condensate ⟨f¯f ⟩ = (−243.3 MeV)3 .
This parametrization, denoted as S350, properly reproduces the empirical values of mπ0 and fπ0 in
vacuum, namely mπ0 = 135 MeV and fπ = 92.4 MeV. To test the sensitivity of our results to the
model parametrization we have also considered two alternative parameter sets, denoted as S320
and S380, which also reproduce the phenomenological values of mπ0 and fπ in vacuum, and lead to
quark effective masses M = 320 and 380 MeV, respectively. The corresponding model parameters
are listed in Table I together with the predictions for gπ0 qq and Γπ0 γγ at B = 0. As one can see,
the values of Γπ0 γγ are all compatible with the experimental range Γexp
π 0 γγ
= (7.72 ± 0.11) eV within
less than 1% accuracy.

−⟨q q̄⟩1/3 m0 GΛ2 Λ M gπ0 qq Γπ0 γγ

MeV MeV MeV MeV eV
S320 246.9 5.185 2.138 639.5 320 3.396 7.67
S350 243.3 5.418 2.252 613.4 350 3.720 7.65
S380 241.4 5.543 2.366 596.1 380 4.046 7.65

TABLE I. Model parameters and predictions for gπ0 qq and Γπ0 γγ at B = 0.

In addition, since local NJL models fail to describe the IMC effect observed in LQCD at finite
 14

temperature, we also consider a NJL Lagrangian with a magnetic field-dependent four-quark inter-
action coupling G(B), so as to incorporate the sea effect produced by the backreaction of gluons
to magnetized quark loops. For definiteness, we adopt for G(B) the form proposed in Ref. [21],
namely

G(B) = G F(B) , (53)

where

2
F(B) = κ1 + (1 − κ1 ) e−κ2 (eB) , (54)

with κ1 = 0.321 and κ2 = 1.31 GeV−2 . Even though G(B) does not appear explicitly in our
expressions for the decay width Γπ0 γγ , the B dependence affects the values of M , mπ0 and gπ0 qq ,
which implicitly depend on the four-quark coupling strength.
In Fig. 2 we show our results for the normalized magnetic decay width for the three parameter
sets of Table I, considering both a constant coupling G (left panel) and a magnetic-field dependent
coupling G(B) (right panel). As seen, the NJL model predicts a rather strong suppression of the
width, which is even more striking in the case of G(B). We can also observe that this feature
is basically independent of the model parametrization. Before discussing the possible origin of
this suppression, it is convenient to compare the above results with the predictions arising from
Eq. (52), obtained in the chiral limit. This comparison is shown in Fig. 3 for the parameter set
S350. As seen from the graphs, our calculations show that for both G = constant and G(B) the
result in Eq. (52) is actually a good approximation to the full width. Indeed, the accuracy of this
approximation is found to be less than 4% at B = 0, becoming even smaller as the magnetic field
increases.
The similarity between the numerical results for Γπ0 γγ (B) and Γch
π 0 γγ (B) can be used as a hint

to trace the origin of the width suppression with the external magnetic field. In this sense, we can
take into account the fact that in the chiral limit there is a simple relation (at leading order in mπ0 )
between the π 0 → γγ width, the pion mass and the pion decay constant. The behavior of these
three quantities as functions of B is shown in Fig. 4 for set S350. Looking at these curves, and
considering Eq. (52), it becomes clear that the width suppression originates from the joined effect
of the decrease of the neutral pion mass and the increase of fπch0 with the magnetic field. In fact, it is
seen that both effects turn out to be numerically important. We note in this sense that in Ref. [60],
where the width calculation was performed within the framework of Chiral Perturbation Theory,
only the magnetic behavior of the pion mass was taken into account. It is interesting to observe
 15

7
S 3 2 0
S 3 5 0
S 3 8 0
6

5
Γπ0γγ [ e V ]

1 G = c o n sta n t G (B )

0
0 .0 0 .2 0 .4 0 .6 0 .8 1 .0 0 .0 0 .2 0 .4 0 .6 0 .8 1 .0

e B [G e V 2
] e B [G e V 2
]

FIG. 2. π 0 → γγ decay width as a function of eB. Results for G = constant and G(B) are shown in left
and right panels, respectively. Solid lines correspond to parameter set S350, while band limits correspond
to S320 (dashed lines) and S380 (dotted lines).

that the magnetic behavior of the pion decay constant can be qualitatively understood taking into
account the generalized Gell-Mann-Oakes-Renner relation [29] (mπ0 fπch
0 )
¯ ch /2
2 = −m ⟨ūu + dd⟩
0

(valid at lowest order in the current quark mass m0 ), which explains the enhancement of fπch
0 with

B in terms of both the Magnetic Catalysis effect and the decrease of mπ0 as B increases. Regarding
the use of G(B), it is seen from Fig. 4 that magnetic effects on mπ0 and fπch
0 are more pronounced

when considering G(B) in comparison with the case of a constant G. This leads to a stronger
suppression of the width for G(B) that also holds away from the chiral limit, as shown in Fig. 2.

Finally, the angular dependence of the decaying photon rate can be analyzed from Eq. (47).
Though in principle the differential width dΓ(B)/d cos θ depends on the polar angle θ between the
magnetic field and one of the photon 3-momenta, it is seen that the θ-dependent contribution is
suppressed by powers of the ratio (mπ0 /M )2 . Indeed, from our numerical calculations we find that
for values of eB up to 1 GeV2 the differential width varies less than 0.1% within the full range of
values of θ.
 16

G = c o n st. G (B )
7
G = c o n s t. (c h ir a l) G (B ) (c h ir a l)
6

5
Γπ0γγ [ e V ]

0
0 .0 0 .2 0 .4 0 .6 0 .8 1 .0 0 .0 0 .2 0 .4 0 .6 0 .8 1 .0

e B [G e V 2
] e B [G e V 2
]

FIG. 3. π 0 → γγ decay width as a function of eB for parameter set S350. Left and right panels correspond
to G = constant and G(B), respectively. Solid lines correspond to our full result, Eq. (43), while dotted
lines correspond to the lowest-order chiral expression in Eq. (52).

V. SUMMARY AND CONCLUSIONS

In this work we have studied the anomalous π 0 → γγ decay in the presence of a static uniform
external magnetic field. Our analysis has been performed in the framework of a two-flavor Nambu-
Jona-Lasinio effective model for low energy QCD dynamics, in which pions are treated as quantum
fluctuations in the random phase approximation. The ultraviolet divergences associated to the
nonrenormalizability of the model have been regularized using the magnetic field independent
regularization method, which has been shown to be free from unphysical oscillations and to reduce
the dependence of the results on the model parameters [66]. Additionally, we have explored the
possibility of using a magnetic field-dependent coupling constant G(B) to account for the effect of
the magnetic field on sea quarks.

Our calculations indicate that the π 0 → γγ decay width gets strongly reduced with the external
field, a fact that can be more easily understood by analyzing the process in the chiral limit. At
leading order in mπ0 , we find that the result for the full width reduces to a simple expression,
 17

1 .0
G = c o n sta n t G (B )

0 .8

0 .6

0 .4

[m π0( B ) / m π0( 0 )] 3

0 .2
[f c h
π0 (B ) / f c h
π0 (0 )] −2

Γπc 0h γγ( B ) / Γπc 0h γγ( 0 )

0 .0
0 .0 0 .2 0 .4 0 .6 0 .8 1 .0 0 .0 0 .2 0 .4 0 .6 0 .8 1 .0

e B [G e V 2
] e B [G e V 2
]

FIG. 4. Cubed π 0 mass (dashed lines), inverse squared pion decay constant (dashed-dotted lines) and
π 0 → γγ decay width (solid lines) at lowest order in Chiral Perturbation Theory as functions of eB. All
three quantities are normalized with respect to their values at B = 0. The cases of a constant coupling G
and a B-dependent coupling G(B) are shown in left and right panels, respectively. All values correspond to
the parameter set S350.

which turns out to be just an extension to finite magnetic field of the well-known B = 0 result
that follows from the anomalous gauged Wess-Zumino-Witten action. That expression linearly
relates Γch
π 0 γγ with the cube of the pion mass and the inverse of the squared pion decay constant,

(fπch −2 , see Eq. (52). Since m ch

0 ) π 0 decreases while fπ 0 increases with B, the joint effect implies a

strong suppression of Γch

π 0 γγ with the external field. Our analysis also shows that the effects of both

the behavior of mπ0 and fπch

0 with B contribute significantly to the width reduction. Moreover,

it is seen that this qualitative analysis can be extrapolated away from the chiral limit: from
our numerical results, it is found that the calculation of Γπ0 γγ in the chiral limit is actually a
good approximation to the full calculation, the differences lying below 4% for values of eB up to
1 GeV2 . The full analysis is found to be basically independent of the model parametrization, within
phenomenologically acceptable parameter ranges. Regarding the use of a B-dependent four-quark
coupling, we observe that magnetic effects on mπ0 and fπ0 are more pronounced when considering
G(B) with respect to the case of constant G. This leads to an even stronger suppression of the
 18

π 0 → γγ width.
As for the angular dependence of the decaying photons, an anisotropic angular distribution could
be in principle expected, since rotational symmetry is broken by the magnetic field. Nevertheless,
we find that angular-dependent contributions to the differential decay width dΓ(B)/d cos θ get
suppressed by powers of the ratio (mπ0 /M )2 , leading to nearly perfect isotropy: for magnetic fields
as large as eB = 1 GeV2 , dΓ(B)/d cos θ varies less than 0.1% within the full range of values of θ.
In this work we have considered only leading order dynamical electromagnetic contributions,
taking vacuum dispersion relations for the outgoing photons and dismissing possible effects arising
from π 0 − γ mixing. We leave the analysis of the effect of the corresponding corrections for future
studies.

ACKNOWLEDGMENTS

This work has been partially funded by CONICET (Argentina) under Grant No. PIP 2022-
2024 GI-11220210100150CO, by ANPCyT (Argentina) under Grants No. PICT20-01847 and
No. PICT22-03-00799, and by UNLP (Argentina) under Project No. X960.

Appendix A: π 0 polarization function

The regularized form of the π 0 polarization function, Eq. (18), in the MFIR scheme can be
written as

Πreg
π0
(q 2|| , q⊥
2
) = Ππ0,reg
0 (q 2 ) + Πmag
π0
(q 2|| , q⊥
2
). (A1)

Here, the regularized “B = 0” contribution is given by

h i
Π0,reg
π0
(q 2
) = 2Nc I 0,reg
1 − q 2 0,reg 2
I2 (q ) , (A2)

0,reg
where the expressions of I1,2 , obtained using a 3D momentum cutoff, read [29, 74]
 p  
1 M
I10,reg = 2 2
Λ Λ + M + M ln 2
√ , (A3)
2π 2 Λ + Λ2 + M 2
Z 1 " p
2 − y(1 − y)q 2
#
1 Λ M
I20,reg (q 2 ) = dy p + ln . (A4)
4π 2 0
p
Λ2 + M 2 − y(1 − y)q 2 Λ + Λ2 + M 2 − y(1 − y)q 2
 19

On the other hand, the magnetic contribution can be written as [29]

Z ∞ Z 1
Nc X 2 2
mag 2 2
Ππ0 (q || , q⊥ ) = 2
dz dy e−z [M −y(1−y)q∥ ]
4π 0 0
f =u,d
  
2 1 Bf
× e−ωf (y,z) q⃗⊥ M 2 + + y(1 − y)q 2||
z tanh(zBf )

Bf2 e−zy(1−y) q⃗⊥

2 
 
2

2 2 2
+ 1 − ωf (y, z) ⃗q⊥ − M + + y(1 − y)q ,
sinh2 (zBf ) z z
(A5)

where
sinh(yzBf ) sinh[(1 − y)zBf ]
ωf (y, z) = . (A6)
Bf sinh(zBf )
For a pion at rest this expression can be rewritten in terms of the digamma function, see Ref. [34]
for details.

Appendix B: Computation of the π 0 → γγ decay amplitude

To evaluate the decay amplitude we need to calculate the tensor Rfµν given by Eq. (37). We
take into account the Schwinger form of quark propagators in Eqs. (11-12) and perform the usual
deformation of the proper time integration paths which effectively reduces to the substitution
σ → −iτ [75]. After integration over the “parallel” coordinates s || and t || we obtain
qf2
Z ∞ Z ∞ Z ∞
µν 2
2
Rf = i 16π Nc gπ0 qq 2 dτ1 dτ2 dτ3 e−(τ1 +τ2 +τ3 )M
Bf 0 0 0

d2 r || τ1 (r∥ −p∥ )2 +τ2 r∥2 +τ3 (r∥ +k∥ )2 X

Z
µν
× 2
e Nεf , (B1)
(2π)
ε=±1

where
d2 r⊥ d2 s⊥ d2 t⊥
Z  
µν 1 2 2 ⃗ 2


Nεf = exp − t1 ⃗s⊥ + t2 ⃗r⊥ + t3 t⊥
(2π)2 (2π)2 (2π)2 Bf
 
2εsf  1 1 1 2 2 2 1 1 1 2 2 2
 µν
× exp i (p + s − r )(k + r − t ) − (k + r − t )(p + s − r ) Tεf , (B2)
Bf
µν
with ti = tanh(τi Bf ). Here, Tεf stands for the Dirac trace

µν
= TrD iγ 5 iCεf (r || − p || , s⊥ , t1 ) γ µ iCεf (r || , r⊥ , t2 ) γ ν iCεf (r || + k || , t⊥ , t3 ) ,
 
Tεf (B3)

where

Cεf (u || , v⊥ , ti ) = (u || · γ || + M ) 1 + iεsf γ 1 γ 2 ti − ⃗v⊥ · ⃗γ (1 − t2i ) .

 
(B4)
 20

From explicit computation of the traces in Eq. (B3) we obtain

T 00 = − 8 εsf M Tf p0 p3 ,

T 11 = T 22 = 8 εsf M [t1 + t3 − t2 (1 + t1 t3 )] p0 p3 ,

T 33 = − 8 εsf M Tf p0 (2r3 − p3 ) ,

T 01 ± iT 02 = ± 4 M f1± f2± f3∓ f1∓ (s1 ± is2 ) − f3± (t1 ± it2 ) p3 ,

 

T 10 ± iT 20 = ∓ 4 M f1± f2∓ f3∓ f1∓ (s1 ± is2 ) − f3± (t1 ± it2 ) p3 ,

 

T 03 + T 30 = − 16 εsf M Tf r0 p0 ,
n
T 03 − T 30 = 8M εsf Tf [M 2 + p2|| − r2|| + 2p3 (p3 + r3 )] + t1 (1 − t22 )(1 − t23 )(⃗r⊥ · ⃗t⊥ )
o
+ t2 (1 − t21 )(1 − t23 )(⃗s⊥ · ⃗t⊥ ) + t3 (1 − t21 )(1 − t22 )(⃗r⊥ · ⃗s⊥ )
n o
+ 8 iM ϵij3 ri tj (1 − t22 )(1 − t23 ) + ti sj (1 − t21 )(1 − t23 ) + si rj (1 − t21 )(1 − t22 ) ,

T 12 = − T 21 = 8 iM [1 + t1 t3 − t2 (t1 + t3 )] p0 p3 ,
n  o
T 13 ± iT 23 = ± 4 M f1± f2∓ f3∓ f1∓ (s1 ± is2 ) + f3± (t1 ± it2 ) − 2f2± f3± (r1 ± ir2 ) p0 ,


n  o
T 31 ± iT 32 = ∓ 4 M f2± f3∓ f1± f1∓ (s1 ± is2 ) + f3± (t1 ± it2 ) − 2f2∓ f1∓ (r1 ± ir2 ) p0 ,

(B5)

where we have defined fi± = 1 ± εsf ti and Tf = t1 + t2 + t3 + t1 t2 t3 . We have also considered

the usual photon on-shell conditions k 2 = p2 = 0 and taken into account the fact that in the rest
frame of the pion one has k 0 = p0 , k 3 = −p3 .
To calculate the integrals over perpendicular momenta in Eq. (B2), notice that the traces include
up to quadratic powers of r⊥ , s⊥ and t⊥ coordinates. In fact, we are left with Gaussian integrals
that can be straightforwardly performed. Let us define

d2 r⊥ d2 s⊥ d2 t⊥
Z  
1
I⊥ (un± ) = 1 2 n 2 2 ⃗ 2


(u ± iu ) exp − t ⃗
s
1 ⊥ + t ⃗
r
2 ⊥ + t t
3 ⊥
(2π)2 (2π)2 (2π)2 Bf
 
2εsf  1
(p + s1 − r1 )(k 2 + r2 − t2 ) − (k 1 + r1 − t1 )(p2 + s2 − r2 )

× exp i , (B6)
Bf

where n = 0, 1 and u = r, s, t . Taking ⃗k⊥ = −⃗

p⊥ , for n = 0 we obtain

Bf3
 
t2 (t1 + t3 ) 2
I⊥ (1) = exp − p⃗⊥ , (B7)
64π 3 Tf Bf Tf
 21

while for n = 1 we have

I⊥ (1)
I⊥ (r± ) = (t1 + t3 ) (p1 ± ip2 ) , (B8)
Tf
I⊥ (1)
I⊥ (s± ) = − t2 f3± (p1 ± ip2 ) , (B9)
Tf
I⊥ (1)
I⊥ (t± ) = − t2 f1∓ (p1 ± ip2 ) . (B10)
Tf
µν
From the above expressions it can be see that Nεf = 0 for µν = 01, 02, 10 and 20. On the other
µν
hand, notice that for µ = ν the expressions of Nεf depend linearly on ε. Thus, the corresponding
30 and N 03 , one is faced with terms that are linear
sums over ε = ±1 also vanish. In the case of Nεf εf

in ε and with integrals of terms that are quadratic in perpendicular coordinates. For the latter, in
′ = r − p and choose the 1-axis in the direction of p
the integrals we can shift r⊥ → r⊥ ⃗⊥ (i.e., we
⊥ ⊥

can take p2 = 0). Then, changing r′ 2 → −r′ 2 , s2 → −s2 and t2 → −t2 , it is seen that the integrals
30 = 03
P P
change sign by changing ε → −ε, which leads to ε Nεf ε Nεf = 0. The only nonvanishing

contributions to the amplitude can be obtained from the sums

X X
13 23 31 32



Nεf ± iNεf = − Nεf ± iNεf
ε=± ε=±

= ∓16 M I⊥ (1) [1 + t1 t3 − t2 (t1 + t3 )] p0 (p1 ± ip2 ) , (B11)

X X
12 21
Nεf = − Nεf
ε=± ε=±

= 16 iM I⊥ (1) [1 + t1 t3 − t2 (t1 + t3 )] p0 p3 . (B12)

P µν
In fact, the above results for ε Nεf can be more compactly written as

µν
X
Nεf = −16 iM I⊥ (1) [1 + t1 t3 − t2 (t1 + t3 )] p0 ϵ0µνλ pλ . (B13)
ε=±

Turning back to the expression of Rfµν in Eq. (B1), we are left with the integral over parallel
components of r
d2 r || τ1 (r∥ −p∥ )2 +τ2 r∥2 +τ3 (r∥ +k∥ )2
Z
I || = e , (B14)
(2π)2
P µν
(notice that the result for ε Nεf does not depend on r || ). Performing a Wick rotation in the
variable r0 we get, in the pion rest frame,
" #
i 4τ1 τ3 (p0 )2 + (τ1 + τ3 )τ2 p2||
I || = exp . (B15)
4π (τ1 + τ2 + τ3 ) τ1 + τ2 + τ3
 22

In this way, we end up with

qf2
Rfµν 2
= 256 π Nc gπ0 qq M p ϵ 0 0µνλ
pλ
Bf2
Z ∞ Z ∞ Z ∞
2
× dτ1 dτ2 dτ3 e−(τ1 +τ2 +τ3 )M I⊥ (1) I || [1 + t1 t3 − t2 (t1 + t3 )] , (B16)
0 0 0

which, after replacing the integrals I⊥ (1) and I || by the corresponding expressions in Eqs. (B7)
and (B15), leads to the result in Eq. (39).

Appendix C: Polarizations of outgoing photons

The polarization vectors ϵµ (p, ±1) for an on-shell photon with momentum p⃗ are given by

ϵ0 (p, ±1) = 0 ,

e± iϕ
ip̂2 ∓ p̂1 p̂3 , −ip̂1 ∓ p̂2 p̂3 , ± |p̂⊥ |2 ,


⃗ϵ (p, ±1) = √ (C1)
2 |p̂⊥ |
p |, p̂⊥ = p⃗⊥ /|⃗
where p̂ = p⃗/|⃗ p |, and ϕ is an arbitrary phase. They satisfy

ϵ(p, ±1) · ϵ(p, ±1)∗ = −1 , ϵ(p, +1) · ϵ(p, −1)∗ = 0 , ϵ(p, ±1) · p = 0 , ⃗ϵ (p, ±1) · p⃗ = 0 , (C2)

as well as
X pµ p̃ν + p̃µ pν
ϵµ (p, λp ) ϵν (p, λp )∗ = − g µν + , (C3)
p · p̃
λp =±

where p̃µ = (p0 , −⃗

p ).

[1] D. E. Kharzeev, K. Landsteiner, A. Schmitt, and H.-U. Yee, ‘Strongly interacting matter in magnetic
fields’: an overview, Lect. Notes Phys. 871, 1 (2013), arXiv:1211.6245 [hep-ph].
[2] J. Andersen, W. Naylor, and A. Tranberg, Phase diagram of QCD in a magnetic field: A review, Rev.
Mod. Phys. 88, 025001 (2016), arXiv:1411.7176 [hep-ph].
[3] V. Miransky and I. Shovkovy, Quantum field theory in a magnetic field: From quantum chromody-
namics to graphene and Dirac semimetals, Phys. Rept. 576, 1 (2015), arXiv:1503.00732 [hep-ph].
[4] P. Adhikari et al., Strongly interacting matter in extreme magnetic fields (2024), arXiv:2412.18632
[nucl-th].
[5] T. Vachaspati, Magnetic fields from cosmological phase transitions, Phys. Lett. B 265, 258 (1991).
[6] D. Grasso and H. Rubinstein, Magnetic fields in the early universe, Phys. Rept. 348, 163 (2001),
arXiv:astro-ph/0009061.
 23

[7] V. Skokov, A. Y. Illarionov, and V. Toneev, Estimate of the magnetic field strength in heavy-ion
collisions, Int. J. Mod. Phys. A 24, 5925 (2009), arXiv:0907.1396 [nucl-th].
[8] V. Voronyuk, V. D. Toneev, W. Cassing, E. L. Bratkovskaya, V. P. Konchakovski, and S. A. Voloshin,
(Electro-)Magnetic field evolution in relativistic heavy-ion collisions, Phys. Rev. C 83, 054911 (2011),
arXiv:1103.4239 [nucl-th].
[9] R. C. Duncan and C. Thompson, Formation of very strongly magnetized neutron stars - implications
for gamma-ray bursts, Astrophys. J. Lett. 392, L9 (1992).
[10] C. Kouveliotou et al., An X-ray pulsar with a superstrong magnetic field in the soft gamma-ray repeater
SGR 1806-20., Nature 393, 235 (1998).
[11] D. E. Kharzeev, L. D. McLerran, and H. J. Warringa, The Effects of topological charge change in heavy
ion collisions: ’Event by event P and CP violation’, Nucl. Phys. A 803, 227 (2008), arXiv:0711.0950
[hep-ph].
[12] K. Fukushima, D. E. Kharzeev, and H. J. Warringa, The Chiral Magnetic Effect, Phys. Rev. D 78,
074033 (2008), arXiv:0808.3382 [hep-ph].
[13] D. E. Kharzeev, J. Liao, S. A. Voloshin, and G. Wang, Chiral magnetic and vortical effects in high-
energy nuclear collisions—A status report, Prog. Part. Nucl. Phys. 88, 1 (2016), arXiv:1511.04050
[hep-ph].
[14] V. P. Gusynin, V. A. Miransky, and I. A. Shovkovy, Dimensional reduction and catalysis of dynamical
symmetry breaking by a magnetic field, Nucl. Phys. B 462, 249 (1996), arXiv:hep-ph/9509320.
[15] G. Bali, F. Bruckmann, G. Endrődi, Z. Fodor, S. Katz, S. Krieg, A. Schafer, and K. Szabo, The QCD
phase diagram for external magnetic fields, JHEP 02, 044 (2012), arXiv:1111.4956 [hep-lat].
[16] S. Fayazbakhsh, S. Sadeghian, and N. Sadooghi, Properties of neutral mesons in a hot and magnetized
quark matter, Phys. Rev. D 86, 085042 (2012), arXiv:1206.6051 [hep-ph].
[17] S. Fayazbakhsh and N. Sadooghi, Weak decay constant of neutral pions in a hot and magnetized quark
matter, Phys. Rev. D 88, 065030 (2013), arXiv:1306.2098 [hep-ph].
[18] H. Liu, L. Yu, and M. Huang, Charged and neutral vector ρ mesons in a magnetic field, Phys. Rev. D
91, 014017 (2015), arXiv:1408.1318 [hep-ph].
[19] S. Avancini, W. Tavares, and M. Pinto, Properties of magnetized neutral mesons within a full RPA
evaluation, Phys. Rev. D 93, 014010 (2016), arXiv:1511.06261 [hep-ph].
[20] R. Zhang, W.-j. Fu, and Y.-x. Liu, Properties of Mesons in a Strong Magnetic Field, Eur. Phys. J. C
76, 307 (2016), arXiv:1604.08888 [hep-ph].
[21] S. Avancini, R. Farias, M. Pinto, W. Tavares, and V. Timóteo, π0 pole mass calculation in a strong
magnetic field and lattice constraints, Phys. Lett. B 767, 247 (2017), arXiv:1606.05754 [hep-ph].
[22] S. Mao and Y. Wang, Effect of discrete quark momenta on the Goldstone mode in a magnetic field,
Phys. Rev. D 96, 034004 (2017), arXiv:1702.04868 [hep-ph].
[23] D. Gomez Dumm, M. F. Izzo Villafañe, and N. N. Scoccola, Neutral meson properties under an external
magnetic field in nonlocal chiral quark models, Phys. Rev. D 97, 034025 (2018), arXiv:1710.08950 [hep-
 24

ph].
[24] Z. Wang and P. Zhuang, Meson properties in magnetized quark matter, Phys. Rev. D 97, 034026
(2018), arXiv:1712.00554 [hep-ph].
[25] H. Liu, X. Wang, L. Yu, and M. Huang, Neutral and charged scalar mesons, pseudoscalar mesons, and
diquarks in magnetic fields, Phys. Rev. D 97, 076008 (2018), arXiv:1801.02174 [hep-ph].
[26] M. Coppola, D. Gomez Dumm, and N. Scoccola, Charged pion masses under strong magnetic fields in
the NJL model, Phys. Lett. B 782, 155 (2018), arXiv:1802.08041 [hep-ph].
[27] S. Mao, Pions in magnetic field at finite temperature, Phys. Rev. D 99, 056005 (2019), arXiv:1808.10242
[nucl-th].
[28] S. Avancini, R. Farias, and W. Tavares, Neutral meson properties in hot and magnetized quark matter:
a new magnetic field independent regularization scheme applied to NJL-type model, Phys. Rev. D 99,
056009 (2019), arXiv:1812.00945 [hep-ph].
[29] M. Coppola, D. Gomez Dumm, S. Noguera, and N. Scoccola, Neutral and charged pion properties
under strong magnetic fields in the NJL model, Phys. Rev. D 100, 054014 (2019), arXiv:1907.05840
[hep-ph].
[30] G. Cao, Magnetic catalysis effect prevents vacuum superconductivity in strong magnetic fields, Phys.
Rev. D 100, 074024 (2019), arXiv:1906.01398 [nucl-th].
[31] G. Cao, Recent progresses on QCD phases in a strong magnetic field: views from Nambu–Jona-Lasinio
model, Eur. Phys. J. A 57, 264 (2021), arXiv:2103.00456 [hep-ph].
[32] B.-k. Sheng, X. Wang, and L. Yu, Impacts of inverse magnetic catalysis on screening masses of neutral
pions and sigma mesons in hot and magnetized quark matter, Phys. Rev. D 105, 034003 (2022),
arXiv:2110.12811 [hep-ph].
[33] S. S. Avancini, M. Coppola, N. N. Scoccola, and J. C. Sodré, Light pseudoscalar meson masses under
strong magnetic fields within the SU(3) Nambu–Jona-Lasinio model, Phys. Rev. D 104, 094040 (2021),
arXiv:2109.01911 [hep-ph].
[34] M. Coppola, D. Gomez Dumm, S. Noguera, and N. N. Scoccola, Masses of magnetized pseudoscalar
and vector mesons in an extended NJL model: The role of axial vector mesons, Phys. Rev. D 109,
054014 (2024), arXiv:2312.16675 [hep-ph].
[35] K. Kamikado and T. Kanazawa, Chiral dynamics in a magnetic field from the functional renormalization
group, JHEP 03, 009 (2014), arXiv:1312.3124 [hep-ph].
[36] A. Ayala, R. Farias, S. Hernández-Ortiz, L. Hernández, D. Manreza Paret, and R. Zamora, Magnetic
field-dependence of the neutral pion mass in the linear sigma model coupled to quarks: The weak field
case, Phys. Rev. D 98, 114008 (2018), arXiv:1809.08312 [hep-ph].
[37] A. Ayala, J. Hernández, L. A. Hernández, R. Farias, and R. Zamora, Magnetic field dependence of the
neutral pion mass in the linear sigma model with quarks: The strong field case, Phys. Rev. D 103,
054038 (2021), arXiv:2011.03673 [hep-ph].
 25

[38] A. Ayala, R. L. S. Farias, L. A. Hernández, A. J. Mizher, J. Rendón, C. Villavicencio, and R. Zamora,

Magnetic field dependence of the neutral pion longitudinal screening mass in the linear sigma model
with quarks, Phys. Rev. D 109, 074019 (2024), arXiv:2311.13068 [hep-ph].
[39] N. Agasian and I. Shushpanov, Gell-Mann-Oakes-Renner relation in a magnetic field at finite temper-
ature, JHEP 10, 006 (2001), arXiv:hep-ph/0107128.
[40] J. Andersen, Chiral perturbation theory in a magnetic background - finite-temperature effects, JHEP
10, 005 (2012), arXiv:1205.6978 [hep-ph].
[41] G. Colucci, E. Fraga, and A. Sedrakian, Chiral pions in a magnetic background, Phys. Lett. B 728, 19
(2014), arXiv:1310.3742 [nucl-th].
[42] V. Orlovsky and Y. Simonov, Nambu-Goldstone mesons in strong magnetic field, JHEP 09, 136 (2013),
arXiv:1306.2232 [hep-ph].
[43] M. Andreichikov, B. Kerbikov, E. Luschevskaya, Y. Simonov, and O. Solovjeva, The Evolution of Meson
Masses in a Strong Magnetic Field, JHEP 05, 007 (2017), arXiv:1610.06887 [hep-ph].
[44] Y. Simonov, Pion decay constants in a strong magnetic field, Phys. Atom. Nucl. 79, 455 (2016),
arXiv:1503.06616 [hep-ph].
[45] M. Andreichikov and Y. Simonov, Chiral physics in the magnetic field with quark confinement contri-
bution, Eur. Phys. J. C 78, 902 (2018), arXiv:1805.11896 [hep-ph].
[46] C. A. Dominguez, M. Loewe, and C. Villavicencio, QCD determination of the magnetic field dependence
of QCD and hadronic parameters, Phys. Rev. D 98, 034015 (2018), arXiv:1806.10088 [hep-ph].
[47] E. V. Luschevskaya, O. A. Kochetkov, O. V. Teryaev, and O. E. Solovjeva, π ± and ρ0,± mesons in a
strong magnetic field on the lattice, JETP Lett. 101, 674 (2015).
[48] E. V. Luschevskaya, O. E. Solovjeva, O. A. Kochetkov, and O. V. Teryaev, Magnetic polarizabilities of
light mesons in SU (3) lattice gauge theory, Nucl. Phys. B 898, 627 (2015), arXiv:1411.4284 [hep-lat].
[49] G. Bali, B. Brandt, G. Endrődi, and B. Gläßle, Meson masses in electromagnetic fields with Wilson
fermions, Phys. Rev. D 97, 034505 (2018), arXiv:1707.05600 [hep-lat].
[50] H. T. Ding, S. T. Li, A. Tomiya, X. D. Wang, and Y. Zhang, Chiral properties of (2+1)-flavor QCD
in strong magnetic fields at zero temperature, Phys. Rev. D 104, 014505 (2021), arXiv:2008.00493
[hep-lat].
[51] H. T. Ding, S. T. Li, J. H. Liu, and X. D. Wang, Chiral condensates and screening masses of neutral pseu-
doscalar mesons in thermomagnetic QCD medium, Phys. Rev. D 105, 034514 (2022), arXiv:2201.02349
[hep-lat].
[52] G. Endrodi, QCD with background electromagnetic fields on the lattice: A review, Prog. Part. Nucl.
Phys. 141, 104153 (2025), arXiv:2406.19780 [hep-lat].
[53] A. I. Nikishov and V. I. Ritus, Quantum Processes in the Field of a Plane Electromagnetic Wave and
in a Constant Field 1, Sov. Phys. JETP 19, 529 (1964).
[54] A. I. Nikishov and V. I. Ritus, Quantum Processes in the Field of a Plane Electromagnetic Wave and
in a Constant Field 2, Sov. Phys. JETP 19, 1191 (1964).
 26

[55] G. S. Bali, B. B. Brandt, G. Endrődi, and B. Gläßle, Weak decay of magnetized pions, Phys. Rev. Lett.
121, 072001 (2018), arXiv:1805.10971 [hep-lat].
[56] M. Coppola, D. Gomez Dumm, S. Noguera, and N. N. Scoccola, Pion-to-vacuum vector and axial
vector amplitudes and weak decays of pions in a magnetic field, Phys. Rev. D 99, 054031 (2019),
arXiv:1810.08110 [hep-ph].
[57] M. Coppola, D. Gomez Dumm, S. Noguera, and N. N. Scoccola, Magnetic field driven enhancement of
the weak decay width of charged pions, JHEP 09, 058, arXiv:1908.10765 [hep-ph].
[58] M. Coppola, D. Gomez Dumm, S. Noguera, and N. N. Scoccola, Weak decays of magnetized charged
pions in the symmetric gauge, Phys. Rev. D 101, 034003 (2020), arXiv:1910.10814 [hep-ph].
[59] T. Brauner and S. V. Kadam, Anomalous low-temperature thermodynamics of QCD in strong magnetic
fields, JHEP 11, 103, arXiv:1706.04514 [hep-ph].
[60] P. Adhikari and B. C. Tiburzi, Chiral Symmetry Breaking and Pion Decay in a Magnetic Field (2024),
arXiv:2406.00818 [hep-ph].
[61] S. L. Adler, Axial vector vertex in spinor electrodynamics, Phys. Rev. 177, 2426 (1969).
[62] J. S. Bell and R. Jackiw, A PCAC puzzle: π 0 → γγ in the σ model, Nuovo Cim. A 60, 47 (1969).
[63] D. Ebert, K. G. Klimenko, M. A. Vdovichenko, and A. S. Vshivtsev, Magnetic oscillations in dense cold
quark matter with four fermion interactions, Phys. Rev. D 61, 025005 (2000), arXiv:hep-ph/9905253.
[64] D. P. Menezes, M. Benghi Pinto, S. S. Avancini, A. Perez Martinez, and C. Providencia, Quark matter
under strong magnetic fields in the Nambu-Jona-Lasinio Model, Phys. Rev. C 79, 035807 (2009),
arXiv:0811.3361 [nucl-th].
[65] P. G. Allen, A. G. Grunfeld, and N. N. Scoccola, Magnetized color superconducting cold quark mat-
ter within the SU(2)f NJL model: A novel regularization scheme, Phys. Rev. D 92, 074041 (2015),
arXiv:1508.04724 [hep-ph].
[66] S. Avancini, R. Farias, N. Scoccola, and W. Tavares, NJL-type models in the presence of in-
tense magnetic fields: the role of the regularization prescription, Phys. Rev. D 99, 116002 (2019),
arXiv:1904.02730 [hep-ph].
[67] A. Ayala, M. Loewe, A. J. Mizher, and R. Zamora, Inverse magnetic catalysis for the chiral transi-
tion induced by thermo-magnetic effects on the coupling constant, Phys. Rev. D 90, 036001 (2014),
arXiv:1406.3885 [hep-ph].
[68] R. L. S. Farias, K. P. Gomes, G. I. Krein, and M. B. Pinto, Importance of asymptotic freedom
for the pseudocritical temperature in magnetized quark matter, Phys. Rev. C 90, 025203 (2014),
arXiv:1404.3931 [hep-ph].
[69] M. Ferreira, P. Costa, O. Lourenço, T. Frederico, and C. Providência, Inverse magnetic catalysis in the
(2+1)-flavor Nambu-Jona-Lasinio and Polyakov-Nambu-Jona-Lasinio models, Phys. Rev. D 89, 116011
(2014), arXiv:1404.5577 [hep-ph].
[70] S. P. Klevansky, Neutral pion photodecay and weak decays of charged pions at finite temperature, Nucl.
Phys. A 575, 605 (1994).
 27

[71] J. Wess and B. Zumino, Consequences of anomalous Ward identities, Phys. Lett. B 37, 95 (1971).
[72] E. Witten, Global Aspects of Current Algebra, Nucl. Phys. B 223, 422 (1983).
[73] C. Itzykson and J. B. Zuber, Quantum Field Theory, International Series In Pure and Applied Physics
(McGraw-Hill, New York, 1980).
[74] S. P. Klevansky, The Nambu-Jona-Lasinio model of quantum chromodynamics, Rev. Mod. Phys. 64,
649 (1992).
[75] J. Schwinger, On gauge invariance and vacuum polarization, Phys. Rev. 82, 664 (1951).