2
ticular the spontaneous chiral-symmetry-breaking phase
transition at finite temperature and density, the use of
the NJL model furthermore ties this study to the develop-
ment of a general framework of PT-symmetric field the-
ories containing four-point contact interactions and the
possibility of non-Hermitian physics beyond the Standard
Model, see for example [23–27]. Here, analyzing the ef-
fects of finite temperature and density on non-Hermitian
and in particular PT-symmetric theories marks a cru-
cial step towards developing feasible non-Hermitian ap-
proaches and PT quantum field theories applicable to
experimental realizations, such as heavy-ion collisions
and astrophysical models of compact stars. The crucial
task of course is to identify characteristics beyond the
existence and generation of real effective fermion mass,
that may both differentiate between Hermitian and non-
Hermitian field theories, and examine whether new fea-
tures of quantum field theories may arise, when the un-
derlying system is generally non-Hermitian, and specifi-
cally when it is PT symmetric.
This paper is structured as follows. In Sec. II the
standard SU (2) NJL model is reviewed. This discus-
sion serves as the baseline for the examination and
the modified approach used in the study of the non-
Hermitian extensions of the NJL model. The gap
equation for the effective fermion mass is presented
in a self-consistent Hartree approximation and within
the Matsubara-formalism for finite temperature Tand
at finite chemical potential µ, introducing a three-
dimensional regulator Λ. The thermodynamic (grand)
potential Ωis obtained following the coupling-constant
integration method. Based on this, the phase diagram of
the physical fermion mass is determined and the behav-
ior of the thermodynamic observables – quark number,
pressure, entropy, and energy density, as well as the in-
teraction measure – is established.
Section III adapts the NJL formalism for the study
of the model which is modified through the inclusion
of a non-Hermitian non-PT-symmetric bilinear extension
based on the pseudoscalar term γ5, introduced with the
coupling constant g. The qualitative results obtained at
zero temperature and chemical potential are seen to ver-
ify the behavior found previously with an Euclidean four-
momentum cutoff under similar constraints [20]. The
behavior of the effective fermion mass, in particular the
spontaneous chiral-symmetry-breaking phase transition
and its critical end-point (CEP), as well as the effect
on the thermodynamic observables is analyzed in depen-
dence of the temperature Tand chemical potential µ, as
well as Tand the quark number density nfor illustrative
values of the coupling g. Despite a dynamical gener-
ation of fermion mass within the spontaneously broken
approximate chiral symmetry regime, the behavior of the
thermodynamic observables is demonstrated to coincide
with the standard NJL model behavior at low temper-
ature and small chemical potential. In the vicinity of
the phase transition and throughout the restored sym-
metry region, however, the pseudoscalar extension drives
a fermion excess compared to the standard NJL model
and exhibits interaction measures I=ε−3p < 0.
In Sec. IV the NJL model is extended through the in-
clusion of the non-Hermitian but PT-symmetric pseu-
dovector bilinear igBν¯
ψγ5γνψ. The influence of this
modification on the effective fermion mass at finite T
and µis analyzed for a spacelike, a lightlike and a time-
like background field Bν. We confirm that the results
for the spacelike background field obtained in the zero-
temperature and vanishing chemical potential limit coin-
cide qualitatively with those previously found using an
Euclidean cutoff, demonstrating the robustness of the
regularization procedure in this limit. The effect on the
position of the chiral phase transition, the CEP, and on
the behavior of the thermodynamic observables within
the T-µ–plane is investigated, finding a notable deviation
from the standard NJL model behavior and an empha-
sis on the antifermionic component of the system. This
contrasts with the findings within the pseudoscalar ex-
tension.
We conclude and summarize our results in Sec. V.
II. THE NJL MODEL
In the grand canonical ensemble, the two-flavor version
of the standard NJL model [21] is characterized by the
Hamiltonian density
HNJL−µN =¯
ψ(−iγk∂k+m0−µ)ψ−G[( ¯
ψψ)2
+( ¯
ψiγ5~τψ)2],
(1)
where Nis the quark number density operator, µis the
baryon chemical potential, Gis the two-body coupling
strength, and m0is a bare fermion mass term. ~τ denotes
the isospin SU (2) matrices. The Dirac matrices γin 3+1
dimensional spacetime have the form
γ0=10
0−1, γk=0σk
−σk0, γ5=iγ0γ1γ2γ3,
(2)
where σkwith k∈[1,3] are the Pauli matrices. In
the limit of vanishing bare mass m0, this model can be
used to study the spontaneous chiral symmetry breaking,
which occurs through fermion-antifermion pair produc-
tion, parallelling the Bardeen-Cooper-Schrieffer mecha-
nism of superconductivity [28]. It is therefore a com-
monly used effective model for the study of QCD in the
low-energy regime. Due to the non-renormalizability in-
troduced by the contact interaction, a cutoff length of
Λ = 653 MeV and the coupling strength GΛ2= 2.14,
determined traditionally through the quark condensate
density per flavor and the pion-decay constant, are fixed
within a three-momentum cutoff regularization scheme
in this context, cf. [22].
Following the Feynman-Dyson perturbation theory,
the gap equation for the effective fermion mass mis de-
termined in a self-consistent Hartree approximation to