Journal of High Energy Astrophysics 26 (2020) 21–44

Contents lists available at ScienceDirect

Journal of High Energy Astrophysics

www.elsevier.com/locate/jheap

Dark matter stars from beyond General Relativity models:

Magnetosphere dynamics and torsion fields
D.J. Cirilo-Lombardo a,b,c,∗
a
CONICET-Universidad de Buenos Aires, Instituto de Fisica del Plasma (INFIP), Buenos Aires, Argentina
b
Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Departamento de Fisica, Buenos Aires, Argentina
c
Bogoliubov Laboratory of Theoretical Physics, Joint Institute for Nuclear Research, 141980, Dubna (Moscow Region), Russian Federation

a r t i c l e i n f o a b s t r a c t

Article history: The magnetosphere structure of compact objects is considered in the context of a theory of gravity with
Received 20 December 2019 dynamical torsion field beyond standard General Relativity (GR) introduced previously in Cirilo-Lombardo
Received in revised form 5 February 2020 and Minotti (2019). To this end, a new spherically symmetric solution is obtained. A particular case
Accepted 5 February 2020
physically representing a compact object of pseudoscalar fields (for example, axion field) is analyzed
Available online xxxx
from the magnetosphere dynamics viewpoint using the axially symmetric version of the Grad-Shafranov
equation (GSE) previously derived. We explicitly show that, in sharp contrast with the dipole case, the
quadrupole case (where we have two well defined toroidal lobes) the energy stored in the magnetosphere
increases more than in the pure dipole case, fact that is important to explain the problem of excess
energy in mechanics of the pulsars and in the structure of magnetars. In this way we briefly discuss with
the same argument as the one presented in Wang et al. (2019), how in our theoretical frame our model
can help reinterpret phenomenological and observational results in the case of the X-ray emission of XTE
J1810-197.
© 2020 Elsevier B.V. All rights reserved.

Contents

1. Introduction: axisymmetric force-free magnetosphere beyond general relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2. Axion vs boson stars: qualitative aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1. Galactic center and supermassive pseudoscalar star . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2. Cerenkov radiation and pseudoscalar star . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3. Gravitational collapse and magnetosphere structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3. The model of gravity: geometrical Lagrangians and affine geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4. SSS solution: statement of the problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.1. Equations of motion with f = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2. Case ρ = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.3. Case II: ρ = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5. Simplest solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.1. Schwarzchild-Reissner-Nordstrom-(A)de Sitter-like solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.2. Wormhole-type solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6. The spherically symmetric solution: the axion star background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
7. Magnetosphere dynamics and GS equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7.1. Dipole case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8. Quadrupole and quadrupole-dipole case (Figs. 11–12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
9. Helicity and energy: the mechanism (Figs. 5–7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
9.1. About the behavior of helicity cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
10. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

*
Correspondence to: CONICET-Universidad de Buenos Aires, Instituto de Fisica del Plasma (INFIP), Intendente Güiraldes 2160 Pabellón I PB Ciudad Universitaria C1428EGA,
Ciudad de Buenos Aires, Argentina.
E-mail address: diego777jcl@gmail.com.

https://doi.org/10.1016/j.jheap.2020.02.001
2214-4048/© 2020 Elsevier B.V. All rights reserved.
22 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

10.1. Self-linking of magnetic lines: no reconnection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

10.2. Energy threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.3. Seed magnetic field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.4. Axion, cosmic magnetic field and magnetosphere dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.5. Remaining open questions from previous work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.6. Charge distribution and pair generation beyond GR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
10.7. Energy, topology and magnetar-pulsar mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Appendix A. Generalized Lorentz force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Appendix B. Generalized Grad-Shafranov equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
B.1. Magnetic fields and currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Appendix C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

1. Introduction: axisymmetric force-free magnetosphere beyond general relativity

As it is well known, a huge amount of the baryonic component of the Universe is found in the form of plasma, e.g. the standard
conditions of the interstellar medium. Generally astrophysical plasmas can be (at the macroscopical level) treated as an electric conductive
fluid, locally neutral, where currents and magnetic fields play a central role (magnetohydrodynamics, or MHD). This approach can be either
employed in its classical version (non-relativistic), or extended to a relativistic fluid if the physical scenario requires it. In the latter case the
theory is the relativistic MHD and it is necessary when the presence of strong gravity requires the use of general relativity (e.g. for compact
objects like neutron stars and black holes). The magnetic fields involved in astrophysical plasmas may be extremely strong and they are
responsible for many steady and variable emission processes, such as the coronal activity and the solar flares driven by subphotospheric
motions and flux emergence (Shibata and Magara, 2011) or the electromagnetic spindown emission from pulsars Pacini (1968), etc.
Ideal MHD alone fails at explaining the process of magnetization of a fluid initially with B = 0. Moreover, one of the most important
problems for astrophysical plasmas is the origin of the magnetic fields (primordial or not), in the various physical scenarios as that
described above, and their growth through the cosmic time.
For a long time, attempts have been made to give concrete answers to various astrophysical and cosmological mechanisms. In particular
the origin, both of primordial fields of different types, as well as of the stellar and cosmological dynamics. Given that both general
relativity (GR) and the standard model of elementary particles (SM) do not give fully conclusive explanations to these questions, the idea
of a reformulation of a unified theory beyond RG and MS seems very attractive. Now we give (referring textually to some introductory
comments of Kojima, 2017 and Cirilo-Lombardo, 2019c) the main motivations of this research.
In previous references, one of the authors has introduced a unified model based on a non-Riemannian geometry containing a dynamic
antisymmetric torsion that admits the same results of GR and SM already proven, but also satisfactorily solving problems for which GR and
SM present difficulties or inconsistencies. Some of those problems that were successfully treated in the context of this new formulation
as the determination of the mass of the axion (Alvarez-Castillo and Cirilo-Lombardo, 2017), the violation of CP of the neutrino (Cirilo-
Lombardo, 2013), primordial magnetogenesis (Cirilo-Lombardo, 2017c,b), etc. In this work we extend a previous research (Cirilo-Lombardo
and Minotti, 2019), calculating the equations controlling the stellar magnetospheres of compact objects, in particular beyond the dipole
approximation to include the quadrupole modes. To this end, force free conditions are adopted by deriving the equilibrium conditions
depending on a flow function with a pseudoscalar part coming from the torsion field though its dual, the torsion vector h μ .
Actually, a typical example is the axisymmetric force-free magnetosphere in the exterior of a neutron star. As we have been pointed
out before, two possibilities are proposed for the energy storage prior to magnetar outbursts to explain the relevant phenomena: storage
in the magnetar crust or in the magnetosphere (Thompson et al., 2002) (for the case of twisted magnetospheres in the context of standard
GR see Kojima, 2017). The latter model is discussed in terms of similarity with solar flares (Del Zanna and Bucciantini, 2018; Yokoi, 2013;
Lyutikov et al., 2003; Beloborodov and Thompson, 2007). In the solar flare model (e.g., Aly, 1984, 1991), the energy is quasi-statically
stored by thermal motion at the surface, and is suddenly released as large-scale eruptive coronal mass ejections. The energy is dissipated
via a magnetic reconnection associated with the field reconfiguration. Analogous energy buildup and release processes may be relevant
to the magnetar giant flares, although the energy scale differs by many orders. In sum, one must entirely rethink the physics of neutrino
cooling, photon emission, and particle emission from a neutron star, when its magnetic field (instead of its rotation) is the main source
of free energy. This possibility is completely feasible in the context of the model previously presented in Cirilo-Lombardo (2017c) that is
based on a geometric (Lagrangian) action that can be considered the non-Riemannian generalization of the Born-Infeld model (see details
in Cirilo-Lombardo, 2017b)
     
Rs RA
L gs = det λ g α β 1+ + λ F αβ 1 + (1)
4λ 2λ
R s ≡ g α β R (α β) ; R A ≡ f α β R [α β ] (2)
  
, det F μν = 2F μν

∂ ln det F μν
with f α β ≡ ∂ F αβ F μν .
In this model, the torsion T βαγ has a dynamic character (contrary to other models in the literature; Hehl et al., 1976) and is totally
antisymmetric, which allows it to be related to its dual torsion vector hμ . The other important feature is that the energy-momentum
tensor and fundamental constants (really functions of the spacetime) are geometrically induced and not imposed “by hand”.
Field equations linking the dual vector with the electromagnetic field via the following expression

∇α T α β γ = −λ F β γ → ∇[β h γ ] = −λ∗ F β γ (3)

 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 23

in the case of interest here indicate that the magnetic field is related in this theoretical context to the dynamics of the torsion vector h μ .
At the same time we demonstrate, generalizing the Helmholtz theorem in 4 dimensions, that the torsion vector admits unique geometric
decomposition of the form
axial vector
   polar vector
 
βγ δ βγ δ
hα = ∇α a + εα ∇β A γ δ + γ1 εα M β γ δ + γ2 Pα (4)
where a, A γ δ , M β γ δ , P α fields can be associated to particles (matter) and physical observables (e.g. vorticity, helicity etc). Also, we have
found via Killing-Yano symmetries, the fields and possible physical observables associated to the 2 and 0-forms A γ δ and a (later we
will define it as π ), in equation (4). In the 3 + 1 decomposition of the spacetime, expression (4) with geometrically admissible fields
(Killing-Yano symmetries) takes the form
4π  
h0 = ∇0 a + h M + q s n s u s · B + γ1 h V + γ2 P 0 (5)
3   
4π      ·
h i = ∇i a + − A + q s ns u s × E i + ( + q s ns u 0s ) B i + γ1 u 0 ∇ × u + u × ∇ u 0 + u × u + γ2 P i (6)
3 i
Notice that in h0 we can recognize the magnetic and vortical helicities where the A μ is the vector potential and q s is the particle charge,
γ
ns is the number density (in the rest frame) and the four-velocity of species s is u s . Consequently the simplest mechanism to generate
the necessary amount of energy of magnetospheres (even without star rotation) can be described as follows:
1) The axion and other pseudoscalars and pseudovector particles (contained geometrically in hα ) plus all helicities increase the original
magnetic field B, e.g. due the induction (dynamo) linearized expression from Section 2, as

∇ × α g B = h × E − h0 B − ( E · ∇ ω) m (7)
(with α g the metric coefficient gtt in standard notation).
2) The increased B, increases the magnetic helicity H M defined as (g 3 determinant of the absolute space, see Section 4)

√
HM = · B g 3 d3 x
A (8)
hM

3) The H M in turn increases B even more via expressions (5), that contains h M , and (7) through the torsion vector hα .
4) Consequently, the total energy in the magnetosphere will be increased to a certain limit (see Section 4)

√
EM = α g B 2 g 3 d3 x (9)

5) After some limit to be determined, the excess energy in the magnetosphere is ejected and the process is repeated. This limit, as
we saw before (Cirilo-Lombardo and Minotti, 2019), is subject to several topological bounds between the total energy of the system and
the helicity. Precisely, in flat space the well-known inequality between energy and helicity namely E (ξ )  C . |H (ξ )| (C being a positive
constant depending on the shape and size of a compact domain M) is exactly fulfilled in the case treated here, were the metric is fully
regular without any singularity. However, the problem of energy conservation in the case of theories based on a non-Riemannian structure
will not be discussed here and shall be treated in a separate work, due its importance.
This paper, unlike the previous paper that only focused on solutions of the Grad-Shafranov equation for the dipole case, now we will
see and analyze the pure quadrupole and dipolar-quadrupolar field cases also. The reasons are given below.
From the accretion mechanism viewpoint, as it is well known, interstellar magnetic fields play certainly a distinct role in star formation
and in galactic molecular disks. It is also believed that these magnetic fields are, however, also a key ingredient for accretion disks around
young stellar objects and in the nuclei of galaxies. In these objects Panchenko and Postnov (1994), Long et al. (2007) find a strong
observational link between accretion disks and bipolar outflows in the form of winds and jets. Was pointed out before (Cirilo-Lombardo
and Minotti, 2019 and references therein) that in turbulent disks, the magnetic fields evolve according to the induction equation of
mean field electrodynamics. In case of axisymmetric configurations, this leads to a coupled system of partial differential equations for
the poloidal magnetic flux and the toroidal magnetic field including the effects of advection, spatial diffusion, differential rotation and
helicity of the turbulence. In geometrically thin accretion disks, in the context of standard GR the lowest axisymmetric dynamo mode which
can grow exponentially until equipartition with turbulence, is the quadrupolar field. Dipolar fields are found to decay exponentially. Is
known that in the magnetar case, with surface field ∼ 1015 G with standard structure in the context of GR, the internal field can reach
∼ 1017 G (Thompson and Duncan, 1993, 1995, 1996). Since the poloidal and toroidal components are expected to be (in GR context) in
rough equipartition (see Thompson et al., 2002 and references therein), the huge toroidal field stresses the crust, producing a deformation
of the surface layers. This, in turn, induces a rotation of the external field lines which are anchored to the star crust and leads to the
appearance of an external toroidal component. The properties of such a twisted magnetosphere have been investigated by Thompson et al.
(2002) by means of a model analogous to that for the solar magnetic field (e.g. Low and Lou, 1990; Wolfson, 1995) under the assumptions
of a static, dipolar, globally twisted field and enforcing, as in the solar models, the force–free condition. The fact of not considering strong
gravitational fields, and even more, theories beyond general relativity make the study, analysis and possibilities of verifying the new
physics inherent in the structure and mechanisms of magnetars (through the existence of dark matter, dark energy and the associated
processes) practically close to zero.
In our case, we know that the torsion in our model has a dynamic character and that there exists an analogy between the induction
equations in our case with that of the standard midfield described above. Therefore, we must carefully analyze whether the scenario of
Kojima (2018) in the context of general relativity is also fulfilled in our case. In addition, the problems that appear in the formulation of
general relativity could be solved in the context of our proposal, as some of those discussed below.
24 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

Recently on September 14, 2015, the Laser Interferometer Gravitational-Wave Observatory (LIGO) detected the first gravitational wave
(GW) event, GW 150914, apparently produced by the merger of two black holes. This fact is the signature of the beginning of a new
era of gravitational wave astronomy. Two years later on August 17, 2017, LIGO and Virgo collaboration reported the detection of the first
binary neutron star merger event GW170817, which was spatially and temporarily associated with a short-duration gamma-ray burst GRB
170817A with a temporal delay of ∼1.7 s. What is the explanation of this delay? Zhang (2019) gives a review of this problem with
various scopes in the context of the GR. This is a possible case where a theory beyond GR like ours could help. Another recent problem
is the interpretation of the variable radio pulsation behaviors and the evolution of the X-ray hot spot of the radio-transient magnetar XTE
J1810-197. This problem is discussed in the last Section of this paper where we put the model suggested in Wang et al. (2019) in the
context of our proposal, showing how the dark matter represented by the axion together with the magnetic helicity for example, they can
modify in the same way as the twist angle, the luminosity.
With the above motivation, in this paper we will work out the problem of the free magnetosphere for the dipole quadrupole and pure
quadrupole case. To this end we revisiting the explicit computation of the Grad-Shafranov (Cirilo-Lombardo, 2019c) equation in the case
of a new spherically symmetric solution in the context of the theory proposed by us (that it is treated in full detail through the paper)
and finally seeing what are the physical consequences in the context of the numerous problems listed. For this purpose in Section 4
the new spherically symmetric solution (considering the dual tensor torsion hμ from the gravitational theory based on affine geometry
given in Cirilo-Lombardo, 2017b,c) representing a spherical configuration of pseudoscalars (the axion star) is obtained and the analysis of
the metric and the fields involved, carried out. Section 7 is devoted to the application of the generalized Grad-Shafranov equation (see
Cirilo-Lombardo, 2019c and Appendix) in the case of the solution obtained in Section 4: the dynamics and structure of the magnetosphere
is determined. In section 9 we describe the interplay between the helicity and energy in this theoretical description.
Finally in Section 10, we will present a concise discussion on the problem of the physics of magnetospheres based on the expressions
obtained, topology of the magnetic fields and the current knowledge regarding the intervention of high energy processes in these scenarios.

2. Axion vs boson stars: qualitative aspects

We will now briefly discuss some qualitative aspects regarding similarities and differences between axion and boson stars in the light
of the previous work (Cirilo-Lombardo and Minotti, 2019).

2.1. Galactic center and supermassive pseudoscalar star

The solution presented in this paper meets partially the conditions described in Torres et al. (2000a) for the stars of scalars. As we
see, although our solution have not singularity, the existence of horizons is evident and the interaction with the electromagnetic field is
clear. However, at least with respect to the horizons, the problem would be solved for the extreme solution, or that the horizons would
be hidden behind the stellar surface. Consequently one could roughly infer in a purely qualitative way, similar arguments to consider a
pseudoscalar star as a galactic center. A detailed discussion conveniently with phenomenological and quantitative aspects of this model
are given in a subsequent paper (Cirilo-Lombardo, in progress).

2.2. Cerenkov radiation and pseudoscalar star

In Torres et al. (2000b) the authors explore the possibility that a charged particle moving in the gravitational field generated by a
scalar star could radiate energy via a recently proposed gravitational Čerenkov mechanism. Due that they numerically prove that this is
not possible for stable boson stars. From the theoretical viewpoint, the emission of Cerenkov radiation is a kinematically allowed process
if one invokes the condition that the spatial variation of the scalar field ψ is greater than three half its potential (eq. (21) of Torres et al.,
2000b). However, that in our case from (59) we see that the energy-momentum tensor takes that same form for the axion field a that
for the scalar field ψ of Torres et al. (2000b) but in our case the potential is zero. Then, we have simply: |hr |2 = |∂r a|2 > 0. Consequently
from explicit expression for hr (eq. (69)) this Cerenkov process is possible allowing in our particular case, a concrete way of distinguishing
a star from axions from one of bosons or from another type of star composition.

2.3. Gravitational collapse and magnetosphere structure

From the purely qualitative point of view and in the context of recent work of Capozziello et al. (2019) in which it is shown that it is
possible to extend Einstein’s theory by considering f ( R ) gravity and inquiring if the further degrees of freedom in the gravitational action
can be modeled as perfect fluids sourcing the field equations. In our particular case with dynamic torsion, a similar treatment could be
done considering that one has a energy-momentum tensor that allows an interpretation of fluid (exotic or not).
The serious analysis of gravitational collapse in our context with an analogous treatment to the work of Astashenok et al. (2019) and
its relationship with the structure of the magnetosphere requires to go beyond a merely qualitative description that is left out of this
work.

3. The model of gravity: geometrical Lagrangians and affine geometry

Our research here is based on a new model of theory of gravitation that uses a pure affine geometrical construction with the Lagrangian
given by

Lg = det Raμ Raν , (10)

reminiscent of a nonlinear Sigma model, where the following geometrical objects are determined by the curvature
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 25

 
Raμ = λ eaμ + f aμ + R aμ , (11)

that corresponds to the breaking of the underlying SU (2, 2) symmetry of a group manifold in higher dimensions (generally belonging to
S O (2, 2) group). In (11), eaμ is the tetrad field

g μν ≡ eaμ e ν a , ηab ≡ eaν eν b (12)

a
and f μ is antisymmetric with respect to index permutation

eaμ f aν = f μν = − f νμ , (13)
which is associated with a central tensor extension of the original SU (2, 2) group. R μν is the Ricci contraction of the curvature tensor
in the manifold with torsion, M (e.g. U 4 ) and λ an arbitrary constant. Notice, that the Ricci tensor has symmetric and antisymmetric
parts corresponding to the Christoffel and torsion contributions to the connection.1 By using the simplified form of the basic blocks in the
Lagrangian density (11) the action becomes
     
Rs RA
L gs = det λ g α β 1+ + λ F αβ 1 + (14)
4λ λ
R s ≡ g α β R (α β) ; R A ≡ f α β R [α β ] (15)
where

R μν = R (μν ) + R [μν ]
   
S ymmetric antisymmetric

(see Shafranov, 1958 for details).

The basis of the theory considered is an hypercomplex construction of the metric compatible) space-time manifold M (Shafranov, 1958),
where for each point pertinent to M there exists a local affine space A. The connection over A,
, defines a generalized affine connection
 on M specified by ∇ and K , where K is an invertible (1, 1) tensor over M. We will demand that the connection is compatible and
rectilinear
α , ∇ g
∇μ K ρσ = K ρα T μσ μ μν = 0, (16)
α is the torsion and g
where T μσ μν is preserved under parallel transport. This generalized compatibility condition ensures that the gener-
alized affine connection  maps autoparallels of  on M in straight lines over the affine space A (locally). The first equation in (16) is the
condition determining the connection  in terms of the fundamental tensor K .
As is well known, the Palatini principle determines the connection required for the spacetime symmetry as well as the field equations.
By construction, the action (1) yields the G-invariant conditions (namely, the intersection of the 4 dimensional Lorentz group L 4 , the
symplectic Sp (4) and the almost complex group K (4)) without prior assumptions. As consequence, the gravitational, Dirac and Maxwell
equations arise from the action S g as a causally connected closed system. The selfconsistency is given by

1
f μν ≡ εμνρσ ϕ ρσ = ∗ϕμν (17)
2
 ρ
where ϕν λ is related to the torsion by 31! ∂μ ϕν λ + ∂ν ϕλμ + ∂λ ϕμν = T νμ ϕρ λ and f μν is assumed to be proportional to the electromag-
netic field.
Starting from the trace free gravitational equation from Cirilo-Lombardo (2017b) that is not assumed but arises from the model, it is
rewritten as
◦ gαβ ◦  gαβ  g
αβ
R αβ − R = 6 −hα hβ + hγ hγ + R s + T αF β + 2λρα β (18)
  2   2 2
≡G α β ≡ T αh β

where
  
1 F μν

F μν F ηρ

F ηρ
ρα β ≡ gα β + F α λ F βλ + gαβ − F α λ

F βλ (19)
W 2b2 8
      2 
Rs R (α β) 1 F ηρ

F ηρ F ηρ

F ηρ R (α β)
F
T αβ ≡ F α λ F βλ − F μν F μν + F μν

F μν gαβ − F α λ

F βλ + (20)
2W Rs 2b2 8 2 Rs
◦
the LHS of (18) is the Einstein tensor. The GR divergence ∇ of G α β is zero (geometrical identity) and, in an analog manner
◦α 
∇ T αh β + T αF β = 0 because both tensors have the same symmetry that the corresponding GR energy momentum tensors of a vector
field and electromagnetic field respectively:

1
Here and below, one has μ, ν = 0, 1, 2, 3, a = 1, 2, 3 and M aμ ≡ eaν M νμ .
26 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

◦α ◦α 
∇ G αβ = ∇ T αh β + T αF β = 0

consequently the remaining part must be a covariantly constant tensor that we assume proportional to g α β :
g 
αβ
∇α R s + 2λρα β = 0
g2 
αβ
⇒ R s + 2λρα β =  g α β → R s = 2 (21)
2
Coming back to the original trace free expressions we have the expected formula
◦ gαβ ◦  gαβ 
R αβ − R = 6 −hα hβ + hγ hγ + T αF β +  g α β (22)
  2   2
≡G α β ≡ T αh β

◦
Tracing the first expression in (21) we have R s = 2 = R + 6hμ hμ linking the value of the curvature and the square norm of the
torsion vector field. This fact is very important because, if the dual of the torsion field has the role of the energy-matter carrier, the
meaning of lambda as the vacuum energy is immediately established. Notice that the LHS in expression (21) instead to be proportional to
the metric tensor it can be proportional to the square of a Killing-Yano tensor containing the symmetries that are missing in the standard
formulations in the context of the General Relativity (GR), as the vorticity, spin and other skew-symmetric observables. We remark here
that the above equations (similar in form to the GR ones) were obtained from the trace free gravitational equation of Cirilo-Lombardo
(2017c) by tracing and it is extremely important in order to compare with other approaches due the fact that it is written in the LHS
as the GR equations with in the RHS a geometrically induced “energy-momentum like” term containing the torsion vector h and the
−term.

4. SSS solution: statement of the problem

Now we are going to look for the simplest solutions that are spherically symmetrical in our model beyond general relativity. Note that
in this Section we have λ from the geometric Lagrangian (3), the constant of integration from the equations “trace free”  g and  which
appears in the temporal exponential coefficient in the line element (28). To do this we will proceed as follows:

4.1. Equations of motion with f = 0

Our starting point are the gravitational plus field equation in their final form, namely
◦ gαβ ◦  gαβ 
R αβ − R = 6 −hα hβ + hγ hγ +  g gα β (23)
  2   2
≡G α β ≡ T αh β
R s = 2 g
◦
R [μν ] = ∇ α T αμν = −λ f μν (24)
= ∇α T αμν ,
here,  g is the cosmological constant that appears from integration of the trace free equations, T μνρ = εμνρσ hσ and the divergence of
the general connection coincides with the Christoffel one due the fully antisymmetry of the torsion tensor. The exact solution that we are
interested in, in principle can be obtained considering no EM field (e.g. f μν → 0) in equation (24):
∗
∇μ hν − ∇ν hμ = −λ f μν =0 (25)
then seeing (17) the possible choice admissible in (25) is
◦
hα = ±∇ α π (26)
Taking into account (26), equation (23) becomes to
 
◦ gαβ ◦ ◦ ◦ gαβ ◦ ◦γ
R αβ − R = 6 −∇ α π ∇ β π + ∇ γ π ∇ π + g g α β (27)
  2  
2

≡G α β
≡ T αh β

with π a pseudoscalar.
ii) The next important step is to propose the following line element

ds2 = −e 2 dt 2 + e 2 dr 2 + e 2F (r ) dθ 2 + e 2G (r ) sin2 θ dϕ 2 (28)

where the components of the metric tensor are
gtt = −e 2 g tt = −e −2
grr = e 2 g rr = e −2
g θ θ = e 2F g θ θ = e −2F (29)
2 −2G
g ϕϕ = sin θ e 2G
g ϕϕ = e 2
sin θ
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 27

To obtain the system of equations we use the Cartan’s structure equations method in the standard way. The line element (28) in the
1-forms basis takes the following form
 2  2  2  2
ds2 = − ω0 + ω1 + ω2 + ω3 (30)

were the 1-forms are

ω0 = e dt⇒ dt = e − ω0
ω1 = e dr⇒ dr = e − ω1
(31)
ω2 = e F (r )dθ
⇒ dθ = e − F (r ) ω2
ω = e sin θ dϕ ⇒ dϕ = e−G (r ) (sin θ)−1 ω3
3 G (r )

with the 1-forms at hand we can explicitly write, using the tetrad frame, the equation (25) and to solve it,
◦ 
∇ μ hμ = ∇μ hμ → er1 ∂r + a1a h1 = 0 (32)
 a 
ω b , ec = acb → a1a = e− ∂r ( + 2F ) (33)
From (32) and (33) we obtain

h1 = ρ e −(+2F ) = e − hr (34)
Consequently, the gravitational dynamic equations take the form
◦0 ◦1  
G 0 − G 1 = 2 (hr )2 e −2 = 2e −2 ∂r ∂r F − ∂r F ∂r ( + ) + (∂r F )2 (35)
2 −4F 2
−ρ e = ∂r ∂r F − ∂r F ∂r ( + ) + (∂r F ) (36)
then, taking into account (34)

−ρ 2 e −3F = ∂r ∂r Y − ∂r Y ∂r ( + ) (37)
where we have been defined e ≡ Y and the explicit form of ∂r ∂r e
F F
was used. This equation is simplified straightforwardly taking the
simplest form
 
1 ρ2 2
∂r + (∂r Y ) = (∂r Y )2 ∂r ( + ) (38)
2 Y2

4.2. Case ρ = 0

ρ = 0 that takes the role of vev of the torsion vector field, consequently eq. (38) becomes
This case is setting
1  
∂r (∂r Y )2 = (∂r Y )2 ∂r ( + ) (39)
2
that is easily solved by reducing the order obtaining (V 0 = const )

∂r Y = V 0 e + (40)
Notice that this case implies that π = const. By symmetry arguments and in order to compare the results with the well known solutions,
we can make  = − thus

Y = c 0 r + r0 = e F ⇒ F = ln |c 0 r + r0 | (41)
◦1
We now must go to G 1
 
◦1 ∂r (2) c 0
2 c 02 1
G1 = e + = 2 g + (42)
(c 0 r + r0 ) (c 0 r + r0 )2 (c 0 r + r0 )2
 
= c 0 ∂r e 2 (c 0 r + r0 ) = 2 g (c 0 r + r0 )2 + 1
without loss of generality we put c 0 = 1 and integrate
 
2 1
e 2 = r + r
0 +  g (r + r0 )3 (43)
3 r + r0
with r
0 the new arbitrary integration constant, that we assume as r
0 ≡ r0 ± 2m, immediately we obtain the following important result
2m 2
e 2 = 1 ± +  g (r + r0 )2 (44)
r + r0 3
This equation is identical to the well known Schwarzchild-(anti)de Sitter solution with an important difference: the r 0 constant that
apparently avoid the singularity problem. This derivation carry the same conclusion that the original work of Schwarzchild that the
singularity is a very particular case: the called Hilbert-Schwarzchild solution where the spacetime singularity appears.
28 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

4.3. Case II: ρ = 0

Again by symmetry arguments and in order to compare the results with the standard solutions, we make  = − as in usual spheri-
cally symmetric electrovac spacetimes, consequently eq. (38) becomes
 
1 ρ2
∂r − 2 − (∂r Y )2 = 0 ⇒ ρ 2 − Y 2 (∂r Y )2 = c 1 Y 2 (45)
2 Y
1  2
= ρ2 + ∂r Y 2 = c1 Y 2
4
as before, it is easily solved by reducing the order making Y 2 = ω then

1
ρ2 + (∂r ω)2 = c 1 ω ⇒ ∂r ω = ±2 −ρ 2 + c 1 ω (46)
4
consequently

(r − r0 )2 + ρ 2
e 2F = (47)
c1
◦1
Introducing it in G 1 , namely
◦1    
G 1 = ∂r F ∂r e 2 + (∂r F )2 − ρ 2 e −4F e 2 − e −2F = 2 g (48)

we obtain the main equation to solve

 
A (r − r0 )2 − 1 X A (r − r0 )2 + 1 1
∂r X ± = ±2 g + (49)
A (r − r0 )2 + 1 (r − r0 ) A (r − r0 ) c 1 (r − r0 )

where we define: X ≡ e 2 and A = c 12 /ρ 2 .

4.3.1. Homogeneous solution

From (49) we obtain
 
A (r − r0 )2 − 1 XH
∂r X H = ∓ (50)
A (r − r0 )2 + 1 (r − r0 )
then, there are two possible homogeneous solutions, namely

A (r − r0 )2 + 1
X+
H = (51)
(r − r0 )
(r − r0 )
X−
H = (52)
A (r − r0 )2 + 1
 −1
e.g. X +
H = X−
H.

4.3.2. General solution

With the homogeneous solutions (51), (52), we need to find a particular solution from the standard procedure obtaining:

 −1 2λ + e −2F
X±
P = X±
H dr + c 4 ⇒ X G± = X ± ±
H XP (53)
∂r F

5. Simplest solutions

Starting from the developments given in the previous Section, there appear two important and interesting cases for the metrics: a
Schwarzchild-Reissner-Nordstrom-(A)de Sitter-like solution and a wormhole-type one that will not be treated in this paper.

5.1. Schwarzchild-Reissner-Nordstrom-(A)de Sitter-like solution

The equation (53), after explicit substitution of F , can be exactly integrated obtaining the following result
  
− 2
c 12 + 4 g ρ 2 (r + r0 )2 c 2 + c 3 ρ 2 (r + r0 ) 2 (r + r0 )4 ρ 2 c12 + 2 g ρ 2
XG → e = ± + g 2 − (54)
c 12 (r + r0 )2 + ρ 2 c 12 (r + r0 )2 + ρ 2 3 c 1 (r + r0 )2 + ρ 2 (r + r0 )2 c 12 + ρ 2
when we see clearly the role of each term above looking their asymptotic behavior
  
2
c 12 + 4 g ρ 2 (r + r0 )2 c 2 + c 3 ρ 2 (r + r0 ) 2 (r + r0 )4 ρ 2 c12 + 2 g ρ 2
e = ± + g − (55)
c 12 (r + r0 )2 + ρ 2 c 12 (r + r0 )2 + ρ 2 3 c 12 (r + r0 )2 + ρ 2 (r + r0 )2 c 12 + ρ 2
       
deformation flat limit Newtonian type (mass) cosmological type charge (Reissner-Nordstrom type)
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 29

Notice that ρ can be interpreted, from the metric coefficient viewpoint, as “angular-momentum-like”. For r0 = 0 the value of c 12 = −(1 +
2 g ρ 2 ) in order to fulfill the flat boundary condition at r → 0.
When ρ → 0 we obtain
c2 2
e 2 = 1 ± +  g (r + r0 )2 (56)
(r + r0 ) 3
as we expected from the previous analysis.

5.2. Wormhole-type solution

The following expression will be analyzed in a next paper, which we present for completeness:
 
c 12 /ρ 2 (r − r0 )2 + 1   ρ2
2
e = c4 c3 + 2 g c 12 /ρ 2 (r − r0 )2 + 1 2 + (57)
(r − r0 ) c1
ρ2    
+ c 12 /ρ 2 (r − r0 )2 + 1 F 1 1/2, 1, 3/2, −c 12 /ρ 2 (r − r0 )2
c 12 2

This solution will be studied in detail soon.

6. The spherically symmetric solution: the axion star background

In this section we use the previous spherically symmetric solution that describes, in principle, an axion star in the theory with torsion
totally antisymmetric of Cirilo-Lombardo (2017b,c).
The starting point to determine the simple solution that we will give here for the axion star is based on the Lagrangian of Cirilo-
Lombardo (2017b) and with the following ansatz of the general metric of a spherically symmetric configuration:

ds2 = −e 2 dt 2 + e 2 dr 2 + e 2F (r ) dθ 2 + e 2G (r ) sin2 θ dϕ 2 (58)

The gravitational equations (Einstein-like equations) in the case of null cosmological constant  g , take the form
◦ gαβ ◦  gαβ 
R αβ − R = 6 −hα hβ + hγ hγ (59)
  2   2
≡G α β ≡ T αh β
 
◦ gαβ ◦ ◦ ◦ gαβ ◦ ◦γ
R αβ − R = 6 −∇ α π ∇ β π + ∇γ π ∇ π (60)
  2 2
≡G α β

where the overcircle on operators and tensors indicates that they are defined as in GR and π is a pseudoscalar field: e.g. axion. Note
that the fact of having zero cosmological constant in this type of theory implies a constraint of geometric character between energy and
geometrically induced matter. In equation (61) the following equation of motion for the torsion field was used, namely
◦
∇ α T αμν = λ∗ F μν = 0 (61)
◦
∇μ hν − ∇ν hμ = 0 ⇒ hα = ±∇ α π (62)
(however T μνρ = εμνρσ hσ and the divergence of the G-connection coincides with the Christoffel one due to the full antisymmetry of the
torsion tensor).
The conservation law is equivalent to the divergence of the hμ that allows to determine the radial component of the torsion vector

h1 = ρ e −(+2F ) = e − hr (63)
The above result is analog to the standard conservation law for a (pseudo) scalar field in standard GR due the RHS of (60). The remaining
components hμ , on the other hand, are determined by equations (62) and are related via Hodge operation to the totally antisymmetric
component of the connection, via the formula
ρ
∂μ hν = ∂ν hμ − 2[νμ] hρ (64)
ρ
= ∂ν hμ − νμσ hσ hρ (65)
h0 being determined by
  
 
∂r e  ∂r e − ht = 0 → ht = e  c dre − + d (66)

with c , d = const. and h2 = h3 = 0 in the simplest case.

Consequently, after solving the full system, the metric coefficients and the torsion vector components are given by the following
expressions (considering cosmological constant  g = 0 in (55)):
30 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

Fig. 1. Metric solution, full line M = 1, ρ = 1 (normalized) and dotted line with M = 1 and ρ = 0: e.g. without torsion field (that is precisely the Schwarzchild solution).

2(ρ 2 c 2 + M )r 2(ρ 2 + M )r
e 2 = 1 − =1−  2 (67)
(r − r0 )2 + ρ 2 ρ2
r − (ρ 2 c + M ) + ρ2
2
 2
ρ2
e 2F = (r − r0 )2 + ρ 2 = r − + ρ2 (68)
(ρ 2 c 2 + M )
In the case of SSS configurations the simplest admissible case is F = G and  = −. Consequently
ρ
hr = − (69)
(r − r0 )2 + ρ 2
  
−
ht = e 
a dre +b (70)

Through the next part of this work, the ρ constant plays the role of v.e.v. of the axion field. Another important feature of the solution is
that there exists a minimal radius, namely

ρ 2 c12 + 1
r0 = 2 , (71)
2c 1 (ρ 2 c 2 + M )
coming from the asymptotical conditions of the metric (67), c 1 is a constant that we will put equal to one without loss of generality
and c 2 a constant with units of inverse of mass depending on the torsion field through ρ , that makes spacetime completely regular as
described by Fig. 1 (e.g. free of singularities). Note that when the torsion field goes to zero, the metric goes to the Schwarzchild one.
While the SSS metric depends completely on h1 , with respect to the zero component of the torsion vector, the only nontrivial admissible
solution that does not modify the geometrically induced energy-momentum tensor is h0 = e − ht = constant.

7. Magnetosphere dynamics and GS equation

7.1. Dipole case

Now we review the pure dipole case treated in our previous work (Cirilo-Lombardo and Minotti, 2019). In this case, due the lack of
rotation of the compact object whose metric is given by (67) with the conditions we consider, the general equation (143) from Appendix B
(Cirilo-Lombardo, 2019c) will be simplified according to the following considerations:
The field hμ does not generate electromagnetic field directly obeying the equation (3). Consequently, ξ ≡ ψ ∈ C.
The additional term in the GSE in spherical coordinates below, that contains |h0 |, comes from the equation of induction (dynamo) and
with corresponds to the chiral anomaly in the generalized current (alpha effect) namely
   
∂ ∂ψ α 2g sin θ ∂ 1 ∂ψ dI (ψ)
α 2g α 2g + = − ( I (ψ) + |h0 | ψ/2) (72)
∂r ∂r r2 ∂θ sin θ ∂θ dψ
The simplest way to reduce the above equation to an homogeneous one is with I (ψ) = const or

I (ψ) = −h0 ψ/2 (73)

With this condition the solution for (72) is
 d P l (θ)
ψ (r , θ) = f l (r ) sin θ (74)
dθ
l ≥1

For the dipole field l = 1 the radial function is

 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 31

Fig. 2. Metric solution gtt (dotted line) and the corresponding radial (full line) component of the torsion vector hr ( M = 1, ρ = 2).

 
 2M




r2 2M 2M
f 1 (r ) = −3μ∞ ln 1 − + 1+ (75)

3
8M

r
with the following definitions in order to be compact:

= (ρ 2 c 2 + M )
M (76)
(r − r0 )2 + ρ 2

r= (77)
r


2M
α 2g = 1 + (78)

r
Consequently
 
 2M




r 2 sin2 θ 2M 2M
ψ (r , θ) = −3μ∞ ln 1 − + 1+ (79)
8M
3

r
which in the weak gravity regime is given by


2 
 ∗

r 2 sin2 θ 3M 12 M
= B 30

r 1+ + + ... (80)

r 2

r 5

r2
As can be seen, we have taken the usual boundary conditions for ψ (r , θ), i.e.

ψ (r , 0) = ψ (r , π ) = 0 (81)
−1
ψ (r → ∞, θ) ∼ r (82)

whereas at the stellar surface r = r ∗ the called magnetic function ψ (r , θ) (Fig. 4) is assumed to be dipolar field (e.g. l = 1)
 
ψ r ∗ , θ = f 1 r ∗ sin2 θ (83)

The components of the magnetic field can be obtained as

 
 ψ,θ α ψ,r I
Br , Bθ , Bφ = ,− ,
r  α
then
32 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

Fig. 3. Magnetic function ψ as function of θ and r. Note that the peaks coincide with the horizons of the compact star.

Fig. 4. Magnetic function ψ as function of θ and r near the first horizon of the compact object.

 2  
(r − r0 )2 + ρ 2 cos θ 2r (ρ 2 c 2 + M )
B r = −3μ∞ ln α + α (84)
8(ρ 2 c 2 + M )3 r 6 (r − r0 )2 + ρ 2
  
3 r 2 − r02 (r − r0 )2 + ρ 2 2 1
B θ = − μ∞ α sin θ 3 2 2 ln α + (ρ c 2 + M ) 1 + (85)
8 r (ρ c 2 + M )3 r α
 2  
|h0 | (r − r0 )2 + ρ 2 sin θ 2(ρ 2 c 2 + M )r
B φ = 3μ∞ 3 ln α + α (86)
2r 8(ρ 2 c 2 + M )3 α (r − r0 )2 + ρ 2

Note that in this case μ∞ ∼ B 0 r ∗3 . Relating F μν to the locally measured magnetic field B and defining x ≡ r

2M
we have

2μ cos θ    
Br = −3x3 ln 1 − x−1 − 3x2 1 + x−1 /2
r3
  
  6x2
1 + x−1 /2
2μ sin θ 1 /2
Bθ = 6x3 1 − x−1 ln 1 − x−1 +  1/2
r3 1 − x−1
|h0 | x2 sin θ    
B φ = B φ = 3μ∞  ln 1 − x−1 + x−1 1 − x−1

2r 2 1 − x−1 1/2 M

8. Quadrupole and quadrupole-dipole case (Figs. 11–12)

Analogously to the dipole case, from the general expression (74) for the magnetic function ψ (r , θ) we take l = 1 and l = 2 consequently
the dipole quadrupole expression takes the following form:
 
d P 1 (θ) d P 2 (θ)
ψ (r , θ) = − f 1 (r ) + f 2 (r ) sin θ (87)
dθ dθ
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 33

Fig. 5. Helicity density H in the spacetime geometry of the axion star for a h0 (normalized value) of 0.051.

Fig. 6. Helicity density H in the spacetime geometry of the axion star for a h0 (normalized value) of 0.51.


 
 M




r3 3M 2M M M
f 2 (r ) = −5μ Q 1+ ln 1 − + 2− − (88)

5
2M 2

r 3

r


2 
μQ 8M 40 M
 2 1+ + +···

r 3

r 7

r2

the second line in the equation above is given in the weak gravity regime. Note that in this case μ Q ∼ B 0 r ∗4 .
As can be seen, we have taken the usual boundary conditions for ψ (r , θ), i.e.

ψ (r , 0) = ψ (r , π ) = 0 (89)

ψ (r → ∞, θ) ∼ r −1 (90)

whereas at the stellar surface r = r ∗ the called magnetic function ψ (r , θ) is assumed now to be dipolar-quadrupolar field (e.g. l = 1)

   
ψ r ∗ , θ = f 1 r ∗ − 3a2 f 2 r ∗ cos θ sin2 θ, (91)

a2 > 0 North pole

a2 < 0 Inversion
34 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

Fig. 7. Helicity density H in the spacetime geometry of the axion star for a h0 (normalized value) of 1.51.

Fig. 8. Toroidal field (initial configuration) with no twist (lower magnetosphere energy).

9. Helicity and energy: the mechanism (Figs. 5–7)

The analysis below is in analogy with our previous work: the formulas are the same, only the magnetic function changes and can also
be quadrupole or dipole-quadrupole beside the pure dipole case analyzed before in Cirilo-Lombardo and Minotti (2019). The helicity in
function of I and the function ψ is
 
√ ψI
H≡ g 3 A i B i d 3 x = 4π drdθ (92)
α 2 sin θ
Looking at the geometrically induced energy-momentum tensor from (59) and taking into account the results of the previous Section, we
have the energy as
 
√ α Bi Bi 3 1
E EM ≡ g3 d x= B 2 r 2 sin θ drdθ (93)
8π 4
   2  2 
1 1 I drdθ
= (α ∂r ψ)2 + ∂θ ψ + (94)
4 r α sin θ
√
In the above expressions g 3 is the determinant of the 3-space and generally the integrals are performed in the region r  r ∗ . As we see
from the expression, energy and helicity are linked by h0 (our alpha term).
At this point we will see the role of the torsion in relation to the energy contained in the magnetospheric region. Since the constant λ
was considered null, there is no a priori direct relationship between B ρ and hρ via expression (61). Consequently the relationship will be
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 35

Fig. 9. One twist configuration with the magnetic lines in a helical bundle (Möbius strip isomorphic).

Fig. 10. Two twist configuration (higher energy stored in the magnetosphere) with the magnetic lines in a helical bundle (Möbius strip isomorphic).

via h0 that takes the place of alpha in the dynamo expression (7) and the metric coefficients that form a fundamental part in the stream
function ψ (also called magnetic function) that contains h1 . Therefore, h0 is relevant in the electrodynamics of the pulsar mechanism and
h1 is relevant in the gravitodynamic effects because it is part of the compact object itself by means of gtt (metric solution of the spherical
configuration).
As we are going to see now, the torsion is a special ingredient in the increase of the energy contained in the magnetosphere. The
helicity and the toroidal component of the energy ( E E M | T ) are directly related as follows

 
−h0 ψ 2
H ≡ 2π drd θ ≡ Hdrdθ (95)
α 2 sin θ
  
1 − |h0 | ψ 2 drdθ −h0
E E M |T = = H (96)
8 α sin θ 4π
 
|h0 |
= E E M | T drdθ = |H| drdθ (97)
4π

Consequently, the inequality suggested by Arnold and Khesin (1995) is fulfilled.

36 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

Fig. 11. Dipole-quadrupole case.

Fig. 12. Quadrupole case.

9.1. About the behavior of helicity cases

The helicity in the quadrupole case completely bifurcates into 2 well-defined toroidal lobes, as we can see in Fig. 14. Obviously this
case presents greater stability than the dipolar-quadrupole of Fig. 13. The origin of the dipolar-quadrupolar instability is the impossibility
to establish a complete match between the dipole and quadrupole terms of the solution (87) for the magnetic or stream function. As for
the magnetic function, the lobes are practically located above the horizons.

10. Concluding remarks

As we have shown in other contexts (Cirilo-Lombardo, 2012) and in Cirilo-Lombardo and Minotti (2019), if we initially suppose the
fields in a pure toroidal (axial) configuration (possibly confined), the field bundle allows twisted solutions. These configurations appear
through the geometrical constraint between the angles of the twist in question, reducing the 3-geometry of the torus to a unilateral helical
geometry (surface) isomorphic to the Möbius strip. Fig. 8 would be the basic configuration, then Fig. 9 with one and Fig. 10 with two
twists respectively. Geometrically and topologically speaking, we have here that the energy stored in the magnetosphere of the compact
object increases proportional to T w (twisting number), being the twist number h0 precisely in our case, fact that is corroborated in well
known topological theorems (e.g. Arnold and Khesin, 1995 in flat space). Note that here the configurations are not necessarily closed on
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 37

Fig. 13. Helicity for dipole-quadrupole case.

Fig. 14. Helicity for quadrupole case.

the surface of the star or in the horizons as it is usually assumed, but they close exactly on surfaces of toroidal geometry making the twist
(as in the Möbius strip case). The topological/geometric invariant is the cylinder that contains the central ring of the torus. The number of
twists (turns) will describe an increase in the energy contained in the configuration (the magnetosphere in our case). Consequently for the
quadrupole case (where we have two well defined toroidal lobes) the energy increase more than in the pure dipole case, as is described
clearly in the figures.
The important remark is that, because in flat space helicity is proportional to the linking number times the square of the flux of
the vector field across any section of the flux tubes (Arnold and Khesin, 1995), in our case this translates into the formula (95) namely
H ∝ h0 .ψ 2 .

10.1. Self-linking of magnetic lines: no reconnection

As we saw in the previous discussion, the number of twist in the toroidal fields increases the helicity of the system. As a result of
the expressions (96) and (95) we see that the toroidal component of the energy differs from the helicity by the constant pseudoscalar
quantity h0 (containing the macroscopical observables plus the particle currents, see (5)), therefore the increase in the number of turns
(increase of the linking number) increases the energy of the configuration. The number of turns in which the helical field lines close on
themselves does not produce the effect of usual reconnection, and for fixed energy of the magnetosphere this number has fixed values as
well influencing the conservation of the physical quantities contained in h0 .
38 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

10.2. Energy threshold

For the helicity/toroidal energy relationship discussed in the previous paragraph, the limit value in the force-free magnetosphere will be
defined by the topological/geometrical conditions of the helicity and higher invariants. Since in the case of dynamically torsion theories the
torsion vector contains the currents with the fields of the theory plus the macroscopic physical observables, the conditions of conservation
of energy, etc. are also subject by the same topological/geometrical conditions. Physically, the excess of energy will be emitted in the
vicinity of the horizons of events of the regular black hole as we see in Fig. 3 and also Fig. 2, since the helicity peaks are at those points.

10.3. Seed magnetic field

As we have been seen before (Cirilo-Lombardo, 2017c), considering only the terms of interest without the diffusive and advective term
in the induction equation (only time-dependence for the magnetic field is preserved) we had the following formula
⎡⎛ # #⎞ ⎤
# #
cos ξ # j gen #
η ∇ j × ⎣⎝ ⎠ ek ⎦ = −∂t B i (98)
cos ς

being {ek } our vector basis in g3 (usually in the orthonormal system) giving cos ξ and cos ς the relative directions between h and j and
between h and the generated B respectively (Cirilo-Lombardo, 2017b,c). We can see from expression (98) that the currents given by the
fields (related to the geometry via hα ) originate the magnetic field. Considering, for example, fermion currents for different species (e.g.
neutrino, electron, positron, etc.) (98) takes the form
⎡⎛ # #⎞ ⎤
#* #
# #
⎢⎜ cos ξ # q ψ γ μ f #⎟ ⎥
ψ
⎢⎜ # f f f #⎟ ⎥
η∇ j × ⎢
⎢⎜
⎜ ⎟ ek ⎥ = −∂t B i
⎟ ⎥
⎣⎝ cos ς ⎠ ⎦

As a simple example, one could consider a neutron star with a neutrino dynamics in its interior that under certain conditions (roughly
considered in (98)) would generate the initial magnetic field (seed) required.

10.4. Axion, cosmic magnetic field and magnetosphere dynamics

One point to consider is the possibility of a cosmological origin of the h0 that would appear as constant in our GS equations due to the
differences of temporal scales in the formation and dynamics of the magnetic fields in stars and pulsars and the global cosmological time.
Indeed, we have seen in previous papers (Cirilo-Lombardo, 2017c,b), that h0 appears as a cosmological solution describing an expanding
universe where the h0 falls as the scale factor (1/a (τ ))2 being h0 in that case proportional to the temporal derivative of the axion: e.g.
θ̇ introducing, in this way, the dynamics of dark matter candidates to large-scale astrophysical phenomena. This fact clearly justifies (in
addition to the purely mathematical reasons to obtain a solution) the choice h0 = constant.

10.5. Remaining open questions from previous work

There are many things still to be investigated (Cirilo-Lombardo and Minotti, 2019), one of which is the equilibrium condition beyond
the force free condition (Cirilo-Lombardo and Minotti, in progress) which also includes a turbulent part in the dynamo mechanism with
an original simulation treatment that was used successfully for the solar case. The other interesting topic that we will be working on has
its origin in the enormous velocities of the pulsars that can not be explained only in the asymmetric explosions of supernovas during the
birth of the pulsar, probably the existence of the dynamic torsion in the model will be the right mechanism to give the pulsar the required
“kick” that would take place in magnetic and vortical chiral effects depending also on matter and dark energy. Another very important
point to research is the higher order solutions and the self-similar cases in our theoretical framework.

10.6. Charge distribution and pair generation beyond GR

As we made mention in a previous research (Cirilo-Lombardo, 2019b), from the beginning of radio pulsar studies, three main parame-
ters determining the key electrodynamic processes were defined: from the calculations above, we will demonstrate in a simple way that
the quantities defined from the density are all altered. Said alteration comes from the dynamics of h, being able to accentuate or even
annul the effect that the rotation has on that density. The first was the electric charge density that is needed to screen the longitudinal
electric field near the neutron star surface, namely ρG J = − 2π· Bc . This quantity, introduced by Goldreich and Julian (GJ) in 1969 (Goldreich
# #
#ρG J #
and Julian, 1969) was used to determine the characteristic particle number density n G J = |e| (of the order of 10−12 cm3 near the neu-
tron star surface). Here, as h must be considered from the equation of motion (we concentrate on the linearized version to simplify the
analysis), the corresponding charge density to that of GJ is

( + h ) · B
ρU F T = − ≡ ρG J + ρh (99)
2π c
(subindices GJ indicate here the corresponding GJ quantity) consequently the characteristic charge density can only be determined through
the exact knowledge of hμ , and the corresponding characteristic number density that will be
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 39

# #
# ( + h ) · B #
nU F T = ##− # / |e |
# (100)
2π c
Also the characteristic current density, is modified as

j U F T = c ρU F T (101)

which is much more important as indicated in several references in the literature (e.g. Beskin, 2009 and references therein) because in
such approaches it is the longitudinal electric current circulating in the magnetosphere that will play the key role. Note that we consider
these quantities near the stellar surface, otherwise ρG J should be slightly modified (with the derived quantities) as − 2π· Bc 1
2
2 (1−(r /c ) sin θ)
etc.
The second parameter is the particle multiplication defined currently as λG J = ne /n G J , which shows how much the secondary particle
number density exceeds the critical number density n G J . Also this parameter is affected according to our work, as

λU F J = ne /nU F T (102)

that is evidently greater than the same GJ quantity. As inside of the above expression we have nh , the secondary particle number density
must be greater than in the GJ case to exceeds the new critical number density nU F T . Finally, the third relevant quantity is the hydrody-
namic particle flow that now is Ṅ U F T me c 2  ( here and below denotes the hydrodynamic Lorentz factor of the outflowing plasma) with
the electron-positron pair injection rate

Ṅ U F T = c π λU F T R 20 nU F T = Ṅ (103)

that, with these definitions, it is not modified.

10.7. Energy, topology and magnetar-pulsar mechanism

Our argument regarding the problem of the energy of pulsars and magnetars in the context of our model beyond the GR, is that if we
see, topologically, speaking the expressions of helicity and energy (95) and (96), it has been demonstrated:
i) that Arnold’s conjecture on the energy-helicity relationship is fulfilled,
ii) that the helicity is contained (together with other macroscopic physical
 observables and some particle fields) in the h 0 component
of the torsion vector, as it is easy to see from (5), namely h0 = ∇0 π + 43π H M + q s ns u s · B + γ1 H V + γ2 P 0 . As we also see the dynamics
of a pseudo scalar (here we use π ) is included as well.
iii) in the case of toroidally confined fields (typical cases in MHD in the context of GR, e.g. see Akgun et al. (2016), and beyond GR), we
show in previous sections that h0 is related to the linking and self-linking number (in particular, as it is also the case in the number of
closed twists in Figs. 8–10). This fact does not surprise us, since the relationship linking number helicity is well known both in geometry
and topology and in theoretical physics.
Concerning about the importance of the statements above, it is clear that we should know as clearly as possible, the influence of
the various factors that regulate the energetic regimes of the pulsars. In particular, the topological conditions of the magnetic fields and
their relationship with high energy processes both within compact objects and in its magnetospheres, which in the simplest case can be
considered force free. In Wang et al. (2019) for example, the twisted magnetosphere is considered the relevant condition. In this case
in general, the complete twist of the magnetosphere favors the formation of pairs by the self-induced electric field and the currents
associated with the movement of said pairs. We see that in Wang et al. (2019) (and references therein) the twist angle is associated with
the alpha term in the equation of induction, that is

ψ → α − term (104)

where, as is well known (Zhang, 2019)

 
B φ dl
ψ = dφ = (105)
Br sin θ
in such a way that assuming the sustained twist current only (in the Zhang, 2019 notation)

c cB
j= ∇ ×B sin2 θ ψ (106)
4π 4π r
where the current flows come from the acceleration of the pairs created by the self induced electric field from the whole twisted mag-
netosphere. In our model the term alpha is given as a function of the torsion vector hμ , the physical currents j ph and the magnetic field
as
  
h · j ph
α ∼ − h0 + (107)
(h · B )

Looking the exact expression (127) from the set of equations of the Appendix A, we retain the anomalous term of interest
  
 h · j ph
∇ × α g B  − h0 + B (108)
(h · B )
40 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

then compare expressions (106) from Zhang (2019) and (108) in our case. As we saw for only the toroidal field (which is the most relevant
for the phenomenology of interest), h0 contains the linking (self-linking) number in the case of considering only the toroidal field that
obviously contributes to total helicity. If we also consider the rest of the fields we must effectively include in the term α , namely eq. (107),
the twist angle (105). This extra term must be modulated by h0 just considering the twist equation (105) with the fields computed in our
model, namely (86), (84), (85). Consequently if we consider expressions of the components of magnetic field (86), (84), (85) at the lowest
order, in eq. (108) we have
⎡  ⎤
r
*
⎢ Ah q f ψ f γr ψ f ⎥
⎢ f ⎥
∇ × αg B  − ⎢ 2
⎢ht (1 − sin θ ψ) +
⎥B
⎥ (109)
⎣ (h · B) ⎦

1/ 2
It is important to remember: α g ≡ gtt and that in general we take in principle [ht ] ∼ [ f (t ) /r] in order to give the correct meaning to
expression (108). In (109) we also consider a current of different fermions (f-subindex) as the j ph , accompanied by a constant A in order
to give the correct dimensions. This current could be of neutrinos in interaction with the axion: if we see the decomposition (5), (6)
and consider our previous work (Alvarez-Castillo and Cirilo-Lombardo, 2017) and also Cirilo-Lombardo (2019a) we see that these exotic
interactions with neutrinos in the context of compact objects play a preponderant role in the energetic emission of jets, GRB, and pulsar
kicks in particular. These interactions sterilize neutrinos and also favor the chiral magnetic effect, which is crucial to explain the extreme
energetic conditions in magnetars.
A possible example of how our model can help reinterpret phenomenological and observational results is the case of the X-ray emission
of XTE J1810-197. In Wang et al. (2019) a mechanism based on a twisted magnetosphere model in the force free approach has been
described in order to explain the luminosity problem of XTE J1810-197. The fundamental expressions for the calculation of luminosity are
in our proposal as in Zhang (2019) but modified as follows.
The time of evolution takes the form
   2   −1  
B R 2 ψ 0 ht ψ 0 B R V ht
t ev =  15 yr, (110)
cV ht |re f 0.1 1015 G 10 km 109 V ht |re f
with h0 |re f a reference value. Consequently in the twisted magnetosphere, in the same theoretical context, the stored free energy is
approximately
 2  2  3  2  2
B R3 2 ht 44 B R ψ 0 ht
Et w ≈ (ψ0 ) ∼ 4 × 10 erg. (111)
24 ht |re f 1015 G 10 km 0.1 ht |re f
An average luminosity can be calculated
     
Et w B R ψ 0 V ht
LX ∼ = 2.6 × 1036 erg/s. (112)
t ev 1015 G 10 km 0.1 109 V ht |re f
as we see the three fundamental quantities (110), (111), (112) are modified by the temporal component of the torsion vector and this
is not trivial since from the point of view of the phenomenology of high energies. Well, if we see the decomposition of h 0 taking into
account the gradient of the axion (dark matter candidate) and the magnetic helicity (proportional to the chiral charges) and considering
the admissible solution (70) h0 = e − ht = constant, we arrive at the following expression
· 4π
ht = α g h 0 = π + HM (113)
3
where we took into account definitions (76), (77) and (78). In (113) we see how Lx will also depend on other factors (in addition to
the twist angle) associated with the dynamics of the axion, geometry and indirectly to the generation of pairs through the different
contributions to the total helicity of the system. This can help to understand and correct, complementing the mechanism proposed in
Wang et al. (2019), the anomaly in the observed luminosity one year after the 2003 outburst, with respect to the estimated luminosity in
models in the context of the GR.

Acknowledgments

DJCL is grateful to CONICET-Argentina for financial support. DJCL thanks the BLTP-JINR for hospitality where part of this work was
developed.

Appendix A. Generalized Lorentz force

As we showed before in Cirilo-Lombardo (2017b,c), the geometrically induced Lorentz force that we have obtained from the model in
the linear limit was
 
h · B + ρe E + J × B = ( E · B ) h (114)

consequently, the case of force free condition with non-vanishing torsion field implies: ( E · B ) = 0. General assumptions for 3+1 splitting
in axisymmetrical spacetimes can be introduced in standard form (e.g. j, E and B can be treated as 3-vectors in spacelike hypersurfaces).
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 41

In terms of these 3-vectors the nonlinear equations of the original model can be linearized and consequently written in a Maxwellian
form as

∇ · E = −h · B ≡ 4πρe (115)

∇·B =0 (116)

∇ × (α E ) = ( B · ∇ ω) m (117)

∇ × (α B) = h × E − h0 B − (E · ∇ ω) m (118)

The derivatives in these equations are covariant derivatives with respect to the metric of the absolute space γi j being α , β : lapse and
shift functions respectively. Notice that we need to replace h × E in order to introduce the physical currents as follows. From the above
equations in exact form, the geometrical current induced by the non-Riemannian framework is

J ≡ +h × E − h0 B (119)

consequently

J × B ⇒ (h · B) E = J × B + (B · E) h (120)

then h × E is
 
(B · E) h+ J × B (h · J ) B
h×E=h× = J− (121)
(h · B) (h · B)
consequently, the relation with the physical scenario can be implemented as follows:
  
(h · J ) B h · j ph B
J− → αg j ph − (122)
(h · B) (h · B)

where we define the lapse function α g (metric coefficient gtt ) in order to differentiate it from the alpha-term, transforming the set
((115)–(118)) at the linear level, namely E → E and B → B, to

∇ · E = −h · B ≡ 4πρe (123)

∇·B =0 (124)

∇ × α g E = ( B · ∇ ω) m (125)
  
 h · j ph B
∇ × α g B = α g j ph − − h0 B − ( E · ∇ ω) m (126)
(h · B )
  
h · j ph
= α g j ph − h0 + B − ( E · ∇ ω) m (127)
(h · B )

Similar equations, without considering a fundamental theory, were recently obtained in Zakharov and Shafranov (1986).

Appendix B. Generalized Grad-Shafranov equation

The consistent theoretical description of gravitational magnetohydrodynamics (MHD) equilibria is of fundamental importance for un-
derstanding the phenomenology of accretion disks (AD) around compact objects (black holes, neutron stars, etc.). The very existence of
these equilibria is actually suggested by observations, which not only show evidence of quiescent, and essentially non-relativistic, AD
plasmas close to compact stars, but also the dynamical interplay with high energy processes involving the magnetospheres of compact
objects, in particular pulsars, quasars and magnetars. The electromagnetic (EM) fields involved, in particular the electric field, may lo-
cally be extremely intense (Goldreich and Julian, 1969), so several standard processes such as electron positron pair creation occur, but
several exotic interactions involving neutrinos with axions and other dark matter candidates must also be taken into account. This sug-
gests therefore that such equilibrium (if it certainly exists) should be described in the framework of unified field theory beyond general
relativity (GR) (Cirilo-Lombardo, 2017a,b,c; Alvarez-Castillo and Cirilo-Lombardo, 2017) and beyond the standard model (SM). Extending
previous approaches, holding for compact objects/black hole axisymmetric geometries having into account effect of space-time curvature,
the purpose of this work is the formulation of a generalized Grad-Shafranov (GGS) (Grad and Rubin, 1958; Shafranov, 1958; Solov’ev, 1968;
Zakharov and Shafranov, 1986) equation based in a non-Riemannian geometry with dynamical torsion field suitable for the investigation
of accretion, jets and winds and other astrophysical effects when high energy effects (exotic or not) are present. Now we will calculate
the GSE with axisymmetry. Arguments and procedures for calculating GSE in this model are similar in form to works well known in the
context of GR (Macdonald and Thorne, 1982; Okamoto, 1975; Lovelace et al., 1986) (we use throughout the notation of Okamoto, 1975)
to have a reasonable comparison parameter with those results.
42 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

B.1. Magnetic fields and currents

From the electrodynamic equations in 3 + 1 formulation of curved spacetimes, under axisymmetry, the magnetic field is split as
B = B p + B t where

1  1   ∇ψ × m
Bp = ∇ψ + m × ∇ h0 × m = ∇ψ × e φ- +  ∇ h0 = + ∇χ (128)
2π 2 2π 2 2π
2I (ψ, χ )
Bt = − e φ-
α c
with e φ- unitary toroidal vector, h0 pseudoscalar field that we redefine as χ (zero component of the dual of the antisymmetric torsion
field) and m · m = g φφ =  2 . The expression for the toroidal magnetic field coming from the Ampere law and the currents enclosed by the
surface A, namely I (depending on ψ and χ ), are obtained similarly to the magnetic flux assuming the form: ∇ I = ∇ I (ψ) + m × ∇ I (χ ).
Consequently, since ( E · B ) = 0 eqs. (128) bring the force free condition as

1   ∇ I (ψ) × m ∇ I (χ )
jp = e φ- × ∇ I = − + (129)
2πα 2πα 2 2π
1 dI
=− Bp
α dζ
 
where we have defined the multi-vector ζ ≡ ψ + mχ m ≡  e φ- (and consequently under action of exterior derivative: dζ = dψ + m × dχ
and ∇ζ = ∇ψ + m × ∇ χ ). Thus:

B p · d A = dζ (130)
. . ζ
dI dI
α jp · dA = − Bp · dA = − Bp · dA (131)
dζ dζ
0

I (0) = I (ζ ) (132)
⎡ ⎛ ⎞ ⎤
1 ⎢c ⎜ α  ⎟  ( F − ω)  ⎥
jt = − ⎣ ∇ · ⎝ 2 ∇ψ + m × ∇ χ ⎠ + ∇ψ + m × ∇ χ · ∇ ω⎦ (133)
8π α    2
α c
≡∇ζ
1
vF = ( F − ω)  (134)
α
vF   1
Ep = − e φ- × B p = − ( F − ω) ∇ζ (force free) (135)
c 2π c α
1   1
ρe = ∇ · Ep + h · B = ∇ · Ep (136)
4π 4π
then
 α  /  0
 2 ( F − ω )
1 α  2 ( F − ω )
ρe = − 2 ∇ · ∇ζ + ∇ζ · ∇ ln (137)
8π α2c 2 2 α2c
  
As usual we can eliminate the first factor: ∇ · α2 ∇  + m × ∇ χ between expressions (133) and (137), consequently
 2  
 ( F − ω) jt 1  ( F − ω) cα2  2 ( F − ω)
ρe − =− ∇ζ · ∇ ω − 2 ∇ζ · ∇ (138)
αc c 8απ 2 αc  α2c
In this case with dynamical torsion field, the transfield component of the momentum equation for the force free case, namely
 
h · B + ρe E + J × B = 0 (139)

becomes
jt  ( F − ω) 1 IdI
− ρe = 2 2 (140)
c αc 2α  c d ζ
Solving for ρe and jt
 (  F −ω )   
αc 8π 2 IdI  ( F − ω ) c  2 ( F − ω)
8π 2 ρe =  2 + ∇ζ · ∇ ω − ∇ζ · ∇ ln (141)
 (  F −ω ) 2α 2  c d ζ 2
α c  α2c
1− αc
 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44 43

  
1 8π 2 IdI  ( F − ω) c  2 ( F − ω)
8π 2 j t =   + ∇ζ · ∇ ω − ∇ζ · ∇ ln (142)
 ( F −ω) 2 2α 2  c d ζ 2
α c  α2c
1− αc
(it is due to Ampere equation (fourth Maxwell equation above) projected toroidally)
Consequently, from (141) and (142) we obtain
 
 α  ( F − ω )2 ( F − ω) d F 8π 2 IdI
∇· 2
∇ζ = ∇ · ∇ζ − |∇ζ | 2
− (143)
 αc2 αc 2 dζ 2
2α (  c ) d ζ

Eqs. (141), (142) and (143) are the Grad Shafranov ones, notice that they can be seen as a 2-dimensional Poisson type equation
(axisymmetry) of a complex variable ζ ≡ ψ + mχ → ψ + i χ with a source term ∝ IdI dζ
.
Here we derive the expression for the toroidal magnetic field in the case described in previous sections. Starting from the equation of
the current (Ampere), namely

1 2 3
          
(∇ × (α B)) · ds = 4π α j · ds − (E · ∇ ω) m · ds − h0 B · ds + (h × E) · ds (144)
A A A A A
  
h· j B
in the linearized case and taking into account that: i) h0 = const, ii) (h × E ) ∝ α g j ph −  ph , iii) ∇μ hν − ∇ν hμ = 0 (hμ does not
h·B

generate field directly because λ = 0) and iv) 2 →∝ 2π m dr × m = 0 we have finally

1
   
 
(∇ × (α B )) · ds = 4π α j · ds − h0 B · ds (145)
A A A

Consequently we obtain
 
2I (ψ) + |h0 | ψ
BT = − e φ- (146)

αω
Notice that the h0 enter into the induction equation as alpha term, also it produces the twist into the flux lines (due its pseudoscalar
character) as explained in the last part of this paper.

Appendix C

The standard procedure of E. Cartan has its startpoint in the following equations

dωα = −ωα β ∧ ωβ (147)

Rα β = d ωα β + ωα λ ∧ω λ
β (148)
These are denominated the structure equations. The procedure for to obtain the Einstein equations is by mean the following steps:
i. Making the exterior derivatives of ωα we computing the connection 1-forms ωα β :

ω0 1 = ω1 0 = e− ∂r  ω0
ω2 1 = −ω1 2 = e− ∂r F (r ) ω2
ω3 1 = −ω1 3 = e− ∂r G (r ) ω3
cos θ − F (r ) 3
ω3 2 = −ω2 3 = e ω (149)
senθ
ii. Making the exterior derivatives of ωα β we computing the curvature 2-forms Rα β :
 
R0 1 = e −2 ∂r ∂r  − ∂r  ∂r  + (∂r )2 ω1 ∧ ω0
 
R2 1 = e −2 ∂r ∂r F − ∂r  ∂r F + (∂r F )2 ω1 ∧ ω2
cos θ 1  
R3 2 = e −( F +) ∂r (G − F ) ω ∧ ω3 + e−2 ∂r G ∂r F − e−2F ω2 ∧ ω3
senθ
  cos θ 2
−2 
3
R 1=e ∂r ∂r G − ∂r  ∂r G + (∂r G )2 ω1 ∧ ω3 + e −( F +) ∂r (G − F ) ω ∧ ω3
senθ
R0 2 = −e −2 ∂r  ∂r F ω0 ∧ ω2
R0 3 = −e −2 ∂r  ∂r G ω0 ∧ ω3 (150)
44 D.J. Cirilo-Lombardo / Journal of High Energy Astrophysics 26 (2020) 21–44

iii. The components of the Riemann tensor are easily obtained from the well known geometrical relation of Cartan:

Rα β = R α β ρσ ωρ ∧ ωσ
where we obtain explicitly
 
R 0 110 = e −2 ∂r ∂r  − ∂r  ∂r  + (∂r )2
 
R 2 112 = e −2 ∂r ∂r F − ∂r  ∂r F + (∂r F )2
 
R 3 113 = e −2 ∂r ∂r G − ∂r  ∂r G + (∂r G )2
cos θ
R 3 213 = e −( F +) ∂r (G − F )
sin θ
cos θ
R 3 123 = e −( F +) ∂r (G − F )
sin θ
R 3 223 = e −2 ∂r G ∂r F − e −2F
R 0 330 = e −2 ∂r  ∂r G
R 0 220 = e −2 ∂r  ∂r F (151)
from which we can construct the Einstein equations of the usual manner.

References

Akgun, T., Miralles, J.A., Pons, J.A., Cerda-Duran, P., 2016. Mon. Not. R. Astron. Soc. 462 (2), 1894–1909.
Alvarez-Castillo, David, Cirilo-Lombardo, D.J., Zamora-Saa, Jilberto, 2017. J. High Energy Astrophys. 13–14, 10–16.
Aly, J.J., 1984. Astrophys. J. 283, 349.
Aly, J.J., 1991. Astrophys. J. 375, L61.
Arnold, Vladimir I., Khesin, Boris A., 1995. Topological Methods in Hydrodynamics. Applied Mathematical Sciences. Springer-Verlag, New York.
Astashenok, Artyom V., Mosani, Karim, Odintsov, Sergey D., Samanta, Gauranga C., 2019. Int. J. Geom. Methods Mod. Phys. https://doi.org/10.1142/S021988781950035X.
Beloborodov, Andrei M., Thompson, Christopher, 2007. Astrophys. J. 657, 967.
Beskin, V.S., 2009. MHD Flows in Compact Astrophysical Objects: Accretion, Winds and Jets. Astronomy and Astrophysics Library. Springer, Heidelberg.
Capozziello, Salvatore, Mantica, Carlo Alberto, Molinari, Luca Guido, 2019. Int. J. Geom. Methods Mod. Phys. 16 (01), 1950008.
Cirilo-Lombardo, D.J., 2012. J. Phys. A, Math. Theor. 45, 244026.
Cirilo-Lombardo, D.J., 2013. Astropart. Phys. 50–52, 51–56.
Cirilo-Lombardo, D.J., 2017a. Phys. Part. Nucl. Lett. 14 (6), 799–810.
Cirilo-Lombardo, D.J., 2017b. J. High Energy Astrophys. 16, 1–14.
Cirilo-Lombardo, D.J., 2017c. Int. J. Geom. Methods Mod. Phys. 14 (07), 1750108.
Cirilo-Lombardo, D.J., 2019a. Europhys. Lett. 127, 10002.
Cirilo-Lombardo, D.J., 2019b. Phys. Part. Nucl. Lett. 16 (6), 586–592.
Cirilo-Lombardo, D.J., 2019c. Int. J. Geom. Methods Mod. Phys. https://doi.org/10.1142/S0219887819500130. arXiv:1812.04481.
Cirilo-Lombardo, D.J. Work in progress.
Cirilo-Lombardo, D.J., Minotti, F.O., 2019. Int. J. Geom. Methods Mod. Phys. 16 (04), 1950064.
Cirilo-Lombardo, D.J., Minotti, F.O. Work in progress.
Del Zanna, L., Bucciantini, N., 2018. Mon. Not. R. Astron. Soc. 479 (1), 657.
Goldreich, P., Julian, W.H., 1969. Astrophys. J. 157, 869.
Grad, H., Rubin, H., 1958. In: Proceedings of the Second United Nations Conference on the Peaceful Uses of Atomic Energy, vol. 21. United Nations, Geneva, p. 190.
Hehl, Friedrich W., von der Heyde, Paul, Kerlick, G. David, Nester, James M., 1976. Rev. Mod. Phys. 48, 393.
Kojima, Yasufumi, 2017. Mon. Not. R. Astron. Soc. 468 (2), 2011–2016.
Kojima, Yasufumi, 2018. Mon. Not. R. Astron. Soc. 477 (3), 3530–3538.
Long, M., Romanova, M.M., Lovelace, R.V.E., 2007. Mon. Not. R. Astron. Soc. 374, 436–444.
Lovelace, R.V.E., Mehanian, C., Mobarry, C.M., Sulkanen, M.E., 1986. Astrophys. J. Suppl. Ser. 62, 1.
Low, B.C., Lou, Y.Q., 1990. Astrophys. J. 352, 343.
Lyutikov, Maxim, Pariev, V.I., Blandford, Roger, 2003. Astrophys. J. 597, 998–1009.
Macdonald, D.A., Thorne, K.S., 1982. Mon. Not. R. Astron. Soc. 198, 345.
Okamoto, I., 1975. Mon. Not. R. Astron. Soc. 173, 357.
Pacini, F., 1968. Nature 219, 145–146.
Panchenko, I.E., Postnov, K.A., 1994. A&A 286, 497.
Shafranov, V.D., 1958. Sov. Phys. JETP 6, 545.
Shibata, K., Magara, T., 2011. Living Rev. Solar Phys. 8 (1), 6.
Solov’ev, L.S., 1968. Sov. Phys. JETP 26, 400.
Thompson, C., Duncan, R., 1993. Astrophys. J. 408, 194.
Thompson, C., Duncan, R., 1995. Mon. Not. R. Astron. Soc. 275, 255.
Thompson, C., Duncan, R., 1996. Astrophys. J. 473, 322.
Thompson, C., Lyutikov, M., Kulkarni, S.R., 2002. Astrophys. J. 574, 332.
Torres, Diego F., Capozziello, S., Lambiase, G., 2000a. Phys. Rev. D 62, 104012.
Torres, Diego F., Capozziello, S., Lambiase, G., 2000b. Class. Quantum Gravity 17, 3171–3182.
Wang, Weiyang, et al., 2019. Astrophys. J. 875, 84.
Wolfson, R., 1995. Astrophys. J. 443, 810.
Yokoi, N., 2013. Cross helicity and related dynamo. Geophys. Astrophys. Fluid Dyn. 107 (1–2), 114–184.
Zakharov, L.E., Shafranov, V.D., 1986. Rev. Plasma Phys. 11, 153.
Zhang, Bing, 2019. Front. Phys. 14 (6), 64402.