betaFeedback

WorksPhysics and Astronomy

Torsional oscillations of magnetized neutron stars: Impacts of Landau-Rabi quantization of electron motion

Abstract

Torsional oscillations of magnetized neutron stars have been well studied since they may be relevant to the physical interpretation of some of the observed quasiperiodic oscillations in the magnetar giant flares. In the crustal region of a magnetar, the strong magnetic field can alter the equation of state and composition due to the Landau-Rabi quantization of electron motion. In this paper, we study this effect on the crust-confined, torsional oscillation modes of neutron stars with mixed poloidal-toroidal magnetic fields in general relativity under the Cowling approximation. Furthermore, the inner and outer crusts are treated consistently based on the nuclear-energy density functional theory. Depending on the magnetic-field configurations, we find that the Landau-Rabi quantization of electrons can change the frequencies of the fundamental torsional oscillation mode of 1.4M⊙1.4 M_\odot neutron star models with a normal fluid core by about 10% when the magnetic field strength at the pole reaches the order of 101610^{16} G. The shift can even approach 20% at a field strength of 101510^{15} G for neutron stars with a simple model of superconducting core where the magnetic field is assumed to be expelled completely.

Introduction

Neutron stars, which are born from the gravitational core-collapse of massive stars during supernova explosions, contain a superdense core that can reach several times the nuclear matter density [1]. It is now well established that highly magnetized neutron stars, magnetars, with magnetic field strengths reaching over 1014−101510^{14}-10^{15} G can exist, making them the most magnetic objects in the universe [2]. They are responsible for the observed soft-gamma ray repeaters (SGRs) and anomalous X-ray pulsars [3]. High energy sudden X-ray outbursts with luminosities up to ∼1036 erg s−1\sim10^{36}\,\mathrm{erg}\,\mathrm{s}^{-1} have been observed for some of these objects [4]. More energetic giant flares with peak luminosities 1044−1047 erg s−110^{44}-10^{47}\,\mathrm{erg}\,\mathrm{s}^{-1} have also been observed [5, 6, 7].

Quasiperiodic oscillations (QPOs) with a large range of frequencies have been detected in the late-time tail phase of giant flares. Examples for such frequencies are 18, 26, 29, 92.5, 150, 626.5 and 1837 Hz for SGR 1806-20 [8, 9, 10]; 28, 54, 84 and 155 Hz for SGR 1900+14 [11] and 43.5 Hz for SGR 0526-66 [12]. More recently, very high-frequency QPOs at 2132 and 4250 Hz have also been identified in the main peak of a giant flare [13].

While the origin of the QPOs is still an open question, they are commonly associated to the magneto-elastic oscillations of the star. The fundamental theory for torsional oscillations of elastic solid crusts in the absence of magnetic fields was developed in [14, 15, 16]. Models focusing on the torsional oscillations of the solid crust with dipole magnetic fields were investigated in [17], or even with mixed poloidal-toroidal fields in [18, 19]. Torsional oscillations might explain some but not all observed QPOs. Complications due to the coupling of the crust to the Alfvén modes in the fluid core are still an unsettled issue [20, 21]. Nevertheless, oscillations of highly magnetized neutron stars have been extensively investigated in numerous publications. Some studies have taken into account the intricate coupling to the magnetic field and have discussed the qualitative changes of the QPOs from shear-like to Alfvén-like behavior (e.g., [17, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29]). Among these studies, the global crust-core oscillations of highly magnetized neutron stars was discussed by [20, 23]. The continuum nature of QPOs spectrum was pointed out in [22] using a toy model (see also [25, 26, 30] for further discussions). However, it was later found that the axial spectrum loses its continuum character when the coupling between the polar and axial oscillations is taken into account [27]. Moreover, it was demonstrated that the presence of the neutron star crust has a significant impact on the continuum spectrum. Specifically, it gives rise to the existence of discrete Alfvén modes near the edges of the continua, which correspond to the turning points of the continuum [21, 28, 31, 32]. Despite extensive efforts, the explanation for the high-frequency QPOs at approximately 625 Hz has proven challenging for neutron stars with nonsuperfluid core [29]. This is due to the fact that at high frequencies, the continua dominate and overlap with each other, effectively absorbing the discrete frequencies. Consequently, it becomes very difficult to identify the edges of the continuum, further complicating the interpretation of the observed QPOs. Nevertheless, it has been suggested that the high-frequency QPOs can be interpreted in terms of global magneto-elastic oscillations when the effects of superfluidity are taken into account [33, 34, 35]. An attempt was also made to investigate the polar oscillations of magnetars (without crust) in an effort to understand the high-frequency QPOs [36] as the polar Alfvén oscillations was found to be discrete, in contrast to the axial Alfvén oscillations. The frequencies of the fundamental polar Alfvén oscillations were found to be on the order of a few hundred Hz for typical magnetars. On the other hand, the low-frequency QPOs may be relevant to torsional Alfvén modes [30]. The study of the magneto-elastic oscillations of magnetars touches upon various areas of physics, ranging from the underlying magnetic field configuration to the physics of dense matter, and so far there is no self-consistent model that can incorporate all the important physics components in the analysis.

The intense magnetic field also affects the properties of the crust. In particular, one important effect is due to the quantization of the electron motion perpendicular to the field into Landau-Rabi levels [37, 38]. As a consequence, the composition and equation of state (EOS) of the crustal region can be modified significantly in a high magnetic field [39, 40, 41]. The role of Landau-Rabi quantization of electron motion has recently been studied by treating the inner and outer crusts consistently based on the nuclear-energy density functional theory [41]. More specifically, the outer crust is described using the experimental atomic mass data [42] supplemented with the Brussels-Montreal atomic mass model HFB-24 [43], while the inner crust is calculated based on the same functional BSk24 that underlies the HFB-24 model. The EOS of the inner and outer crusts are thus treated in a unified way. As a result, the composition and EOS of the magnetar crust are found to vary with the magnetic field due to the Landau-Rabi quantization of electron motion. The EOS of the crust matches smoothly that calculated in [44] for the core using the same functional. It is worth noting that the application of a thermodynamically consistent and unified EOS throughout the whole star is important as it has been recently shown that an ad hoc matching of different EOSs for the crust and core can lead to significant errors on the neutron-star structure [45, 46]. The stellar radius can differ by a few percent while the thickness of the crust can differ by up to 30%. This can have significant effects on the crustal torsional oscillation modes considered in this work.

Previous studies (e.g., [17, 47, 48]) have shown that torsional oscillation mode frequencies depend sensitively on the EOS and properties of the crust, and hence one may try to constrain the EOS with observational data if these modes contribute to the QPOs observed in the giant flares of magnetars [49]. In this work, we extend the study of torsional oscillation modes of magnetars by incorporating the effects of Landau-Rabi quantization of electron motion. We shall in particular focus on the errors that will be made if one ignores the effects of magnetic field on the EOS in the oscillation mode calculation. Moreover, it has been proven that both purely dipole and purely toroidal magnetic field configurations are generally unstable for non-convective star (e.g., [50, 51]). Therefore, we shall consider mixed poloidal-toroidal magnetic fields in this work.

The paper is structured as follows. The equations governing a hydrostatic equilibrium star in general relativity with mixed poloidal-toroidal magnetic fields are summarized in Sec. II. In Sec. III, the magnetic-field-dependent EOS model employed in this work is described. § 1.3 discusses the assumptions and methods that we introduce to construct our magnetized stellar models and magnetic field configurations. In § 1.4, the perturbation equations for calculating the torsional oscillation modes of magnetized neutron stars are described. The structures of neutron stars calculated from our magnetized stellar models are then presented in Sec. VI. Our main numerical results for the torsional oscillation modes are presented in Sec. VII. The effects of different stellar masses on the mode frequencies are discussed in Sec. VIII. Finally, we summarize and discuss our results in Sec. IX. Unless stated otherwise, we assume geometric units with G=c=1G=c=1 and the metric signature is (−,+,+,+)(-,+,+,+).

STATIC STELLAR BACKGROUNDS

Hydrostatic equilibrium

In all the models presented in this work, neutrons stars are assumed to be spherical irrespective of the magnetic field configuration. That is, the stellar deformations due to the tensions of magnetic fields are neglected. This treatment is expected to be a good approximation for the magnetic field strengths B∼1015−1016B\sim10^{15}-10^{16} G considered in this work since the energy of the magnetic field is considerably much smaller than the gravitational energy [17].

magnetic energygravitational energy∼B2R3GM2/R∼10−4(B1016 G)2,(1)\frac{\text{magnetic energy}}{\text{gravitational energy}}\sim\frac{B^{2}R^{3}}{GM^{2}/R}\sim10^{-4}\left(\frac{B}{10^{16}\,\mathrm{G}}\right)^{2}, \tag*{(1)}

Detailed numerical solutions of Einstein-Maxwell equations [52] have shown that the stellar structure remains almost unchanged for magnetic fields B≲1017B\lesssim10^{17} G.

In addition, magnetars are in general slowly rotating as most of the angular momentum is extracted via enhanced magnetic braking due to their strong magnetic fields comparing to ordinary neutron stars. For this reason, we ignore any rotational deformations for simplicity. The resulting equilibrium stellar models can be considered as spherically symmetric stars. So, we consider non-rotating and non-accreting spherically symmetric neutron stars composed of cold matter. The spacetime for the unperturbed equilibrium stellar model is described by the static and spherically symmetric metric

ds2=−e2νdt2+e2λdr2+r2(dθ2+sin⁡2θ dϕ2),(2)ds^{2}=-e^{2\nu}dt^{2}+e^{2\lambda}dr^{2}+r^{2}\left(d\theta^{2}+\sin^{2}\theta\,d\phi^{2}\right), \tag*{(2)}

where λ\lambda and ν\nu are two metric functions of radial coordinate rr. The 4-velocity of the fluid inside an unperturbed static background star is thus given by

uμ=(e−ν,0,0,0).(3)u^{\mu}=\left(e^{-\nu},0,0,0\right). \tag*{(3)}

The hydrostatic equilibrium stellar model is determined by the standard Tolman-Oppenheimer-Volkoff (TOV) equation (see, e.g., [1])

dP(r)dr=−[ρ(r)+P(r)][m(r)+4πr3P(r)]r2[1−2m(r)r],(4)\frac{dP(r)}{dr}=-\frac{[\rho(r)+P(r)]\left[m(r)+4\pi r^{3}P(r)\right]}{r^{2}\left[1-\frac{2m(r)}{r}\right]}, \tag*{(4)}
dm(r)dr=4πr2ρ(r),(5)\frac{dm(r)}{dr}=4\pi r^{2}\rho(r), \tag*{(5)}

where ρ(r)\rho(r) and P(r)P(r) are the energy density and pressure at coordinate radius rr, respectively. The enclosed mass m(r)m(r) is defined by

eλ(r)=11−2m(r)r.(6)e^{\lambda(r)}=\frac{1}{\sqrt{1-\frac{2m(r)}{r}}}. \tag*{(6)}

On the other hand, the metric function ν(r)\nu(r) is determined by

dνdr=−1ρdPdr(1+Pρ)−1.(7)\frac{d\nu}{dr}=-\frac{1}{\rho}\frac{dP}{dr}\left(1+\frac{P}{\rho}\right)^{-1}. \tag*{(7)}

The system is closed by providing an EOS P(ρ)P(\rho), which will be discussed in Sec. III. The above system of equations is integrated from the center to the stellar surface with radius RR, which is defined by P(R)=0P(R)=0. The total mass of the star is given by M=m(R)M=m(R). Note also that the metric functions satisfy the following boundary conditions at the surface

eν(R)=e−λ(R)=1−2MR.(8)e^{\nu(R)}=e^{-\lambda(R)}=\sqrt{1-\frac{2M}{R}}. \tag*{(8)}

We will describe how we construct magnetized stellar models taking into account the effects of magnetic field on the EOS in § 1.3.

Mixed poloidal-toroidal magnetic fields

To model the mixed poloidal-toroidal magnetic fields, we employ the linearized relativistic Grad-Shafranov equation [53], which is derived from the Maxwell equations solved for static spherically symmetric stars described in Sec. II A. The dipolar (ℓ=1\ell=1) component of the vector potential is related to a radial function a1(r)a_1(r) which is determined by [54]

e−2λd2a1dr2+(dνdr−dλdr)e−2λda1dr+(ξ2e−2ν−2r2)a1=−4πf0(ρ+P)r2.(9)e^{-2\lambda}\frac{d^2a_1}{dr^2}+\left(\frac{d\nu}{dr}-\frac{d\lambda}{dr}\right)e^{-2\lambda}\frac{da_1}{dr}+\left(\xi^2e^{-2\nu}-\frac{2}{r^2}\right)a_1=-4\pi f_0(\rho+P)r^2. \tag*{(9)}

where f0f_0 is an arbitrary constant and the parameter ξ\xi represents the ratio between the toroidal and poloidal components of the magnetic field. The components of the magnetic field can be expressed as

Bμ=(0,−eνr2a1cos⁡θ,e−λda1drsin⁡θ,ξe−νa1sin⁡2θ).(10)B^\mu=\left(0,-\frac{e^\nu}{r^2}a_1\cos\theta,e^{-\lambda}\frac{da_1}{dr}\sin\theta,\xi e^{-\nu}a_1\sin^2\theta\right). \tag*{(10)}

For convenience, we define Hμ≡Bμ/4πH^\mu\equiv B^\mu/\sqrt{4\pi} which will be used in Eqs. (40) and (42).

Outside the star (vacuum), Eq. (9) can be solved analytically. The general solution is given by [54]

a1(ex)=−3μb8M3r2[ln⁡(1−2Mr)+2Mr+2M2r2].(11)a_1^{(\mathrm{ex})}=-\frac{3\mu_b}{8M^3}r^2\left[\ln\left(1-\frac{2M}{r}\right)+\frac{2M}{r}+\frac{2M^2}{r^2}\right]. \tag*{(11)}

where μb\mu_b is the magnetic dipole moment observed by an observer at infinity and the model becomes purely dipole outside the star (ξ=0\xi=0). We require the interior solutions a1a_1 and da1/drda_1/dr to match with Eq. (11) at the stellar surface continuously. In the interior, a1a_1 is determined by solving Eq. (9) numerically.

One may see from Eq. (10) that there is a discontinuity in BϕB_\phi at the surface when ξ>0\xi>0. In fact, this is due to the simplified treatment used in the Grad-Shafranov equation2. In reality, the drop in BϕB_\phi will more likely spread out in the magnetosphere so as to match the vacuum solution continuously (see [18] for a more detailed discussion).

As pointed out in [54], there may exist some points inside the star r=rˉ<Rr=\bar{r}<R such that a1(rˉ)=0a_1(\bar{r})=0, and hence Br(rˉ)=0B_r(\bar{r})=0; conversely such points never exist for r>Rr>R. The locations of these points will depend on the value of ξ\xi. If such a point exists, the magnetic flux would be confined inside the spherical surface with radius rˉ\bar{r}. In this case, the field lines are said to be in disjoint domains. For such magnetic field configuration, a physical interpretation is still missing and we treat it as unphysical case in this work. We shall not study those configurations.

In this work, two different types of cores inside neutron stars will be considered: normal fluid cores and superconducting cores. As they have different properties, their corresponding magnetic field models will also be different.

For a normal fluid core, the magnetic field extends throughout the star. This situation has been well discussed in the literature (e.g., [53]). The regularity condition at the center of the star implies that a1a_1 should have the form

a1=αcr2+O(r4),(12)a_1=\alpha_c r^2+O\left(r^4\right), \tag*{(12)}

where αc\alpha_c is a constant determined by the junction condition with a1(ex)a_1^{(\mathrm{ex})} at the stellar surface. Here, we require a1(r)a_1(r) to have no node for r<Rr<R for the following ranges:

range I: 0≤ξ≤ξa,\text{range I}: \ 0\leq\xi\leq\xi_a,
range II: ξb≤ξ≤ξc.(13)\text{range II}: \ \xi_b\leq\xi\leq\xi_c. \tag*{(13)}

It should be noted that the values of (ξa,ξb,ξc)(\xi_a,\xi_b,\xi_c) all depend on the chosen EOS and the stellar model. If ξ\xi lies inside these ranges, it is defined to be physical. We shall not consider parameters outside these ranges.

For a superconducting core, we only consider the simplest model for which the matter inside the core behaves as a type I superconductor. The magnetic field is expelled completely from the core due to the Meissner effect and is confined in the crust. In such case, we set a1a_1 to be zero at the core-crust interface. The regular behavior near the interface implies that a1a_1 should have the form [19]

a1=αI[r−(R−δr)]+O([r−(R−δr)]2),(14)a_1=\alpha_I\left[r-(R-\delta r)\right]+O\left(\left[r-(R-\delta r)\right]^2\right), \tag*{(14)}

where the constant αI\alpha_I is determined by the junction condition with a1(ex)a_1^{(\mathrm{ex})} at the stellar surface and δr\delta r is the thickness of the crust. Here we require a1(r)a_1(r) to have no node for R−δr<r<RR-\delta r<r<R in the following range:

0≤ξ≤ξγ.(15)0\leq\xi\leq\xi_\gamma. \tag*{(15)}

Similar to the case of normal fluid core, the value of ξγ\xi_\gamma depends on the EOS and the stellar model. If ξ\xi lies inside this range, it is defined to be physical.

EQUATIONS OF STATE

In this section, we present the EOS employed in our calculations. Our discussion here serves as a summary of the main results reported in [41, 55]. We refer the reader to these original works and references therein for a more detailed discussion. The outer crust, the inner crust, and the core of the star are discussed as follows.

Outer crust

The EOS for arbitrary magnetic field strength is constructed using the procedure described in Sec. II B combining the data given in Table II and 3 of [55]. In short, the experimental atomic masses from the 2016 Atomic Mass Evaluation [42] with the microscopic atomic mass table HFB-24 [43] from the BRUSLIB database [56] are used. We will briefly review the microphysics inputs here [55]. For the crustal region at density below the neutron-drip point and above the ionization threshold, each crustal layer is assumed to be made of fully ionized atomic nuclei (A,Z)(A,Z) with mass number AA and proton number ZZ embedded in a relativistic electron gas.

M/M⊙M/M_{\odot}ξa\xi_aξb\xi_bξc\xi_cξγ\xi_{\gamma}
1.40.2980.3540.4621.4
2.00.2550.2650.3652.1

Table II. The values of ξa\xi_a, ξb\xi_b, ξc\xi_c and ξγ\xi_{\gamma} for 1.4M⊙1.4M_{\odot} and 2.0M⊙2.0M_{\odot} neutron star models in the weak field limit (B≪1015B \ll10^{15} G).

While the pressure due to nuclei is negligible, they contribute to the energy density

ρN=nNM′(A,Z)c2,(16)\rho_N=n_N M'(A,Z)c^2, \tag*{(16)}

where nNn_N is the number density of nuclei and M′(A,Z)M'(A,Z) is the mass of nuclei including the rest mass of ZZ electrons. In principle, M′(A,Z)M'(A,Z) might also depend on the magnetic field. However, the correction is very small for the magnetic fields considered here and will thus be neglected [57]. In the following, the magnetic field will be expressed using the dimensionless ratio B∗≡B/BrelB_* \equiv B/B_{\mathrm{rel}} with

Brel=me2c3eℏ≈4.41×1013 G,(17)B_{\mathrm{rel}}=\frac{m_e^2c^3}{e\hbar}\approx4.41\times10^{13}\ {\rm G}, \tag*{(17)}

where mem_e is the electron mass, ℏ\hbar is the Planck-Dirac constant and ee is the elementary electric charge. Electrons are well approximated by an ideal Fermi gas. However, the presence of a magnetic field leads to the quantization of the electron motion perpendicular to the field into Landau-Rabi levels [37, 38]. By ignoring thermal effects and the small electron anomalous magnetic moment, the electron energy density (with rest-mass excluded) and pressure are

ρe=B∗mec2(2π)2λe3∑ν~=0ν~max⁡gν~(1+2ν~B∗)ψ+[xe(ν~)1+2ν~B∗]−nemec2,(18)\rho_e=\frac{B_*m_ec^2}{(2\pi)^2\lambda_e^3}\sum_{\widetilde{\nu}=0}^{\widetilde{\nu}_{\max}}g_{\widetilde{\nu}}(1+2\widetilde{\nu}B_*)\psi_+\left[\frac{x_e(\widetilde{\nu})}{\sqrt{1+2\widetilde{\nu}B_*}}\right]-n_em_ec^2, \tag*{(18)}
Pe=B∗mec2(2π)2λe3∑ν~=0ν~max⁡gν~(1+2ν~B∗)ψ−[xe(ν~)1+2ν~B∗],(19)P_e=\frac{B_*m_ec^2}{(2\pi)^2\lambda_e^3}\sum_{\widetilde{\nu}=0}^{\widetilde{\nu}_{\max}}g_{\widetilde{\nu}}(1+2\widetilde{\nu}B_*)\psi_-\left[\frac{x_e(\widetilde{\nu})}{\sqrt{1+2\widetilde{\nu}B_*}}\right], \tag*{(19)}

respectively, where λe=ℏ/(mec)\lambda_e=\hbar/(m_ec) is the electron Compton wavelength, gν~=1g_{\widetilde{\nu}}=1 for ν~=0\widetilde{\nu}=0 and gν~=2g_{\widetilde{\nu}}=2 for ν~≥1\widetilde{\nu}\geq1,

ψ±(x)=x1+x2±ln⁡(x+1+x2),(20)\psi_\pm(x)=x\sqrt{1+x^2}\pm\ln\left(x+\sqrt{1+x^2}\right), \tag*{(20)}
xe(ν~)=γe2−1−2ν~B∗,(21)x_e(\widetilde{\nu})=\sqrt{\gamma_e^2-1-2\widetilde{\nu}B_*}, \tag*{(21)}

where γe\gamma_e is the electron Fermi energy in the units of mec2m_ec^2. ν~max⁡\widetilde{\nu}_{\max} is related to the electron number density nen_e by [55]

ne=2B∗(2π)2λe3∑ν~=0ν~max⁡gν~xe(ν~).(22)n_e=\frac{2B_*}{(2\pi)^2\lambda_e^3}\sum_{\widetilde{\nu}=0}^{\widetilde{\nu}_{\max}}g_{\widetilde{\nu}}x_e(\widetilde{\nu}). \tag*{(22)}

The index ν~max⁡\widetilde{\nu}_{\max} is the highest integer for which the combination inside the square root of Eq. (21) is larger than or equal to zero, i.e.,

ν~max⁡=⌊γe2−12B∗⌋,(23)\widetilde{\nu}_{\max}=\left\lfloor\frac{\gamma_e^2-1}{2B_*}\right\rfloor, \tag*{(23)}

where ⌊⋅⌋\lfloor\cdot\rfloor denotes the integer part. According to the condition of charge neutrality, the average baryon number density is given by

nˉ=AZne=AnN.(24)\bar{n}=\frac{A}{Z}n_e=An_N. \tag*{(24)}

For pointlike ions embedded in a uniform electron gas, the energy density associated to the electron-ion interactions is given by (see, e.g., Chapter 2 of [58])

ρL=CM(4π3)1/3e2ne4/3Z2/3,(25)\rho_L=C_M\left(\frac{4\pi}{3}\right)^{1/3}e^2n_e^{4/3}Z^{2/3}, \tag*{(25)}

where CMC_M is the Madelung constant. The corresponding contribution to the pressure is therefore given by

PL=ne2d(ρL/ne)dne=ρL3.(26)P_L=n_e^2\frac{d(\rho_L/n_e)}{dn_e}=\frac{\rho_L}{3}. \tag*{(26)}

The total pressure of the Coulomb plasma is thus given by P=Pe+PLP=P_e+P_L, and the corresponding energy density is ρ=ρN+ρe+ρL\rho=\rho_N+\rho_e+\rho_L. In this work, the Wigner-Seitz estimate CM=−9/10C_M=-9/10 will be adopted for the Madelung constant [59]. The composition of the crust in thermodynamical equilibrium under the presence of magnetic field is determined by minimizing the Gibbs free energy per nucleon, which coincides with the baryon chemical potential (see, e.g., Appendix A in [60]):

μc=ρ+Pnˉ=M′(A,Z)c2A+ZAmec2[γe−1+43CM(4π3)1/3αλene1/3Z2/3](27)\mu_c = \frac{\rho+P}{\bar{n}} = \frac{M'(A,Z)c^2}{A} + \frac{Z}{A}m_e c^2\left[\gamma_e-1+\frac{4}{3}C_M\left(\frac{4\pi}{3}\right)^{1/3}\alpha\lambda_e n_e^{1/3}Z^{2/3}\right] \tag*{(27)}

where α\alpha is the fine-structure constant. When descending in the crust region from a layer made of nuclei (A1,Z1)(A_1,Z_1) to a denser layer made of nuclei (A2,Z2)(A_2,Z_2), the pressure P1→2P_{1\to2} associated to the transition is determined by the equilibrium condition

μc(A1,Z1,P1→2)=μc(A2,Z2,P1→2),(28)\mu_c(A_1,Z_1,P_{1\to2})=\mu_c(A_2,Z_2,P_{1\to2}), \tag*{(28)}

which can be approximately written as [61]

γe+CMαλeF(Z1,A1;Z2,A2)(4πne3)1/3=γe1→2,(29)\gamma_e+C_M\alpha\lambda_e F(Z_1,A_1;Z_2,A_2)\left(\frac{4\pi n_e}{3}\right)^{1/3}=\gamma_e^{1\to2}, \tag*{(29)}
F(Z1,A1;Z2,A2)≡(43Z12/3Z1A1−13Z12/3Z2A2−Z22/3Z2A2)(Z1A1−Z2A2)−1,(30)F(Z_1,A_1;Z_2,A_2)\equiv\left(\frac{4}{3}\frac{Z_1^{2/3}Z_1}{A_1}-\frac{1}{3}\frac{Z_1^{2/3}Z_2}{A_2}-\frac{Z_2^{2/3}Z_2}{A_2}\right)\left(\frac{Z_1}{A_1}-\frac{Z_2}{A_2}\right)^{-1}, \tag*{(30)}
γe1→2≡1+[M′(A2,Z2)A2me−M′(A1,Z1)A1me](Z1A1−Z2A2)−1.(31)\gamma_e^{1\to2}\equiv1+\left[\frac{M'(A_2,Z_2)}{A_2m_e}-\frac{M'(A_1,Z_1)}{A_1m_e}\right]\left(\frac{Z_1}{A_1}-\frac{Z_2}{A_2}\right)^{-1}. \tag*{(31)}

For the detailed composition and nuclear parameters, we refer the reader to Tables 2 and 3 in [55].

Inner crust

The effects of Landau-Rabi quantization in the inner crust have been implemented within the fourth-order extended Thomas Fermi method with proton shell corrections added consistently via the Strutinsky integral [41]. By ignoring the neutron band-structure effects, the boundary between the inner and outer crustal region is determined by the condition μc=mnc2\mu_c=m_n c^2, where mnm_n is the mass of neutron [40, 62]. Nuclear clusters3 are assumed to be spherical and unaffected by the magnetic field. The Coulomb lattice is considered using the Wigner-Seitz approximation and the nucleon density distributions in the Wigner-Seitz cell are parameterized as

nq(r′)=nB,q+nΛ,qfq(r′),(32)n_q(r')=n_{B,q}+n_{\Lambda,q}f_q(r'), \tag*{(32)}

where q=p(n)q=p(n) for protons (neutrons), nB,qn_{B,q} are the background nucleon number densities (typically protons are bound inside clusters therefore nB,p≈0n_{B,p}\approx0), nΛ,qn_{\Lambda,q} are the nucleon number densities characterizing the clusters, r′r' is the radial coordinate for the Wigner-Seitz cell and fq(r′)f_q(r') describes the spatial inhomogeneities. This function is given by

fq(r′)={1+exp⁡[(Cq−air′−ai)2−1]exp⁡(r′−Cqaq)}−1,(33)f_q(r')=\left\{1+\exp\left[\left(\frac{C_q-a_i}{r'-a_i}\right)^2-1\right]\exp\left(\frac{r'-C_q}{a_q}\right)\right\}^{-1}, \tag*{(33)}

in which

ai=(34πnN)1/3,(34)a_i=\left(\frac{3}{4\pi n_N}\right)^{1/3}, \tag*{(34)}

where aia_i is the ion-sphere radius, CqC_q is the cluster radius defined as the half width at half maximum and aqa_q is the diffuseness of the cluster surface. The numbers of neutrons and protons in the Wigner-Seitz cell are respectively determined by

N=4π∫0air′2nn(r′) dr′,(35)N=4\pi\int_0^{a_i}r'^2n_n(r')\,\mathrm{d}r', \tag*{(35)}
Z=4π∫0air′2np(r′) dr′.(36)Z=4\pi\int_0^{a_i}r'^2n_p(r')\,\mathrm{d}r'. \tag*{(36)}

The nuclear energy density functional BSk24 [43], which underlies the nuclear mass model HFB-24 used for the outer crust, is employed to determine the EOS of nuclear clusters and free neutrons. The energy per nucleon E/AE/A is minimized at fixed average baryon density nˉ\bar{n}, which is given by

EA=4πA∫0air′2ε(r′) dr′,(37)\frac{E}{A}=\frac{4\pi}{A}\int_0^{a_i}r'^2\varepsilon(r')\,\mathrm{d}r', \tag*{(37)}

where ε(r′)\varepsilon(r') is the energy density and A=N+ZA=N+Z is the total number of nucleons in the Wigner-Seitz cell [44], taking into account the contributions of magnetized electron Fermi gas as discussed in [41].

Core

The zero temperature unified EOS in beta equilibrium based on the same Brussels-Montreal energy-density functional BSk24 [63] is employed for the dense fluid core. This is the same EOS for unmagnetized neutron stars and we refer the reader to [44, 64] for details. This EOS was shown to be consistent with astrophysical observations, including the gravitational-wave signal from GW 170817 and its electromagnetic counterpart [65]. As an illustration, Figure 1 plots the pressure PP against energy density ρ\rho for the complete EOS model at different normalized magnetic field strength B∗B_* for comparison. It should be a good approximation as the magnetic field has negligible effects on the EOS in the high density core for the range of field strength considered in this work (see § 1.3 for further discussion). Indeed, as can be seen in Figure 1, the EOS of the crust matches smoothly that of the core.

Pressure $P$ plotted against energy density $\rho$ for different normalized magnetic-field strengths

Figure 1. The pressure PP is plotted against the energy density ρ\rho for different values of normalized magnetic-field strength B∗≡B/BrelB_* \equiv B/B_{\rm rel}, where BrelB_{\rm rel} is defined in Eq. (17).

Shear modulus of the crust

The elastic property of the crystallized matter formed by the ions in the neutron star crust treated as an isotropic polycrystal can be characterized by an effective shear modulus μ\mu. Since the shear modulus in the crust is responsible for the restoring force of torsional oscillations, the frequencies of torsional oscillation modes strongly depend on the shear modulus. To compute the torsional modes, an estimation for the shear modulus μ\mu is required. For practical applications in the neutron star crust, the model of isotropic Coulomb solid [66] is considered and the shear modulus (in CGS system of units) is given by

μ=0.1194 nNZ2e2ai(38)\mu= 0.1194 \, n_N \frac{Z^2 e^2}{a_i} \tag*{(38)}
No. of cycle iterationM/M⊙M/M_\odotRR (km)δr\delta r (km)
01.4012.43600.8722
11.4012.54260.9765
21.4012.54480.9786
31.4012.54490.9787

Table 1. The values of MM, RR and δr\delta r for a 1.4M⊙1.4M_\odot neutron star model with a normal fluid core during the iterative calculation of a magnetized stellar model. The magnetic-field configuration is purely dipole and the field strength at the pole of the surface is 4×10164 \times10^{16} G.

In Figure 2, we plot μ\mu against ρ\rho for our EOS with different values of B∗B_*.

Shear modulus $\mu$ as a function of energy density $\rho$ for different normalized magnetic-field strengths

Figure 2. The shear modulus μ\mu as a function of the energy density ρ\rho for different values of normalized magnetic-field strength B∗≡B/BrelB_* \equiv B/B_{\rm rel}, where BrelB_{\rm rel} is defined in Eq. (17).

MAGNETIZED STELLAR MODELS

In standard calculations of the torsional modes of magnetized neutron stars (e.g., [17]), one needs to first obtain the spherically symmetric background star by solving the TOV equation. A magnetic field configuration is then imposed and solved for the equilibrium stellar model as discussed in Sec. II B. However, when the EOS of the crust depends on the magnetic field, so will the structure of the crust. The standard approach is thus not applicable here.

In order to construct a magnetized equilibrium stellar model that includes the effects of magnetic field on the EOS, we first construct a stellar model without magnetic field and then impose a magnetic field configuration on it just like the standard approach outlined above. Next, with this magnetic field configuration, a new stellar background model is calculated by solving the TOV equation with a magnetic-field-dependent EOS. After that, a new magnetic field configuration is constructed using the new stellar background. This process is repeated until the properties of the star (e.g., the radius RR and the crust thickness δr\delta r) converge to some values. When this is achieved, we declare that a magnetized stellar model is obtained. Our method is summarized by the iteration loop presented in Figure 3.

Graphical illustration of the iterative method for the construction of a magnetized stellar model

Figure 3. Graphical illustration of the iterative method for the construction of a magnetized stellar model starting from the upper-left picture. The words in black indicate the information obtained from the previous iteration while those in pink indicate the information to be determined in the current iteration (see text for more details).

As an illustration, Table I presents the convergence of RR and δr\delta r after each iteration for a 1.4 M⊙1.4\,M_\odot magnetized neutron star with a normal fluid core and a dipole magnetic field having the magnetic field strength of 4×10164 \times10^{16} G at the pole of the surface. The convergence is achieved when the new δr\delta r only differs from the previous δr\delta r by less than 0.1%. It is seen that the values of RR and δr\delta r converge rapidly with a few number of iteration cycles only. However, the number of iterations will be greater for stronger magnetic field strengths and more complex magnetic field configurations.

In our study, two approximations are made to simplify the problem. Firstly, we assume that the magnetic field has negligible effect on the structure of the fluid core, and hence the fluid core of the star is fixed during the iteration. This should be a good approximation as the EOS at the high-density range relevant to the core is insensitive to the magnetic field for the field strength being considered in this work [41]. It has also been pointed out [52] that the EOS in the core may be affected by the magnetic field only for field strength of the order of 101810^{18}ગી GG, which is beyond our range of consideration.

Secondly, when we solve for the spherically symmetric model using the TOV equation with a magnetic-field-dependent EOS, we make use of the root mean square (r.m.s.) values of the magnetic field B(r,θ)B(r,\theta) on any shell of radius rr. For a given magnetic field configuration, the magnitude of magnetic field in general depends on both the radial and angular positions. In principle, the stellar structure will no longer be spherically symmetric, though the deviation is expected to be very small for the field strength being considered in this work. By using the r.m.s. values of B(r,θ)B(r,\theta), the background stellar model is kept spherically symmetric, but its radius and crust thickness will in general be different from those obtained without considering the effects of the magnetic field on the EOS.

According to our simplified magnetic field models, BϕB_\phi is discontinuous at the stellar surface when ξ>0\xi> 0, which is unphysical as already noted in Sec. II B. During the iterative process of constructing a new equilibrium model that incorporates the effects of magnetic field on the EOS, the stellar radius increases slightly. As a result, a thin vacuum layer just outside the star in the previous iteration step becomes filled with matter at the current step. To account for this effect, we extrapolate the magnetic field within the matter but not outside the star for the next iteration. The continuity of the magnetic field inside the star is maintained throughout this process. The method of extrapolating BϕB_\phi is to keep the corresponding value of ξ\xi when the magnitude of BϕB_\phi is being calculated using the expression shown in Eq. (10). At the end of the iteration when a new equilibrium model is obtained, the magnetic field still exhibits the same characteristic, where BϕB_\phi remains discontinuous at the stellar surface for ξ>0\xi> 0, as discussed in the standard treatment outlined in Sec. II B.

As an illustration, Figure 4 shows different magnetic field configurations on the meridional plane for 1.4 M⊙M_\odot neutron stars with a normal fluid core. The upper (lower) panel shows the case for ξ=0\xi= 0 (0.45). In each panel, the red and blue circles indicate the stellar surface and the crust-core interface, respectively. One can see that the complexity of the magnetic field configuration increases with ξ\xi.

Magnetic field lines on the meridional plane for 1.4 $M_\odot$ magnetized neutron stars with a normal fluid core

Figure 4. Magnetic field lines on the meridional plane for 1.4 M⊙M_\odot magnetized neutron stars with a normal fluid core. The upper (lower) panel shows the case for ξ=0\xi= 0 (0.45). In each panel, the red and blue circles indicate the stellar surface and the crust-core interface, respectively.

In the following, the variable BB will be used to represent the strength of magnetic field at the pole of the stellar surface.

TORSIONAL OSCILLATIONS

General-relativistic pulsation equations

For the torsional oscillation modes, the only perturbed fluid variable is the azimuthal component of the perturbed 4-velocity [15]

δuϕ=e−νsin⁡θ∂Y(r,t)∂tdPℓ(cos⁡θ)dθ(39)\delta u^{\phi}=\frac{e^{-\nu}}{\sin\theta}\frac{\partial Y(r,t)}{\partial t}\frac{dP_{\ell}(\cos\theta)}{d\theta} \tag*{(39)}

where Pℓ(cos⁡θ)P_{\ell}(\cos\theta) is the Legendre polynomial of order ℓ\ell. A harmonic time-dependence is assumed such that the perturbation function Y(r,t)=Y(r)eiσtY(r,t)=Y(r)e^{i\sigma t}, where σ\sigma is the angular mode frequency.

In the relativistic Cowling approximation where the metric perturbations are neglected, the equations for determining the torsional oscillation modes can be obtained (see [17, 67] for details) by linearizing the fluid equation and the magnetic induction equation, which is derived from the homogeneous Maxwell’s equations with the ideal magnetohydrodynamical approximations such that the electric field 4-vector vanishes.

In the following, we briefly summarize the main perturbation equations studied in [17] relevant to our discussion. The perturbed magnetic induction equation is given by

δ{(uαHβ−Hαuβ);β}=0,(40)\delta\left\{\left(u^{\alpha}H^{\beta}-H^{\alpha}u^{\beta}\right)_{;\beta}\right\}=0, \tag*{(40)}

and the perturbed fluid equation is

δ([δαj+ujuα]T ;βαβ)=0,(41)\delta\left(\left[\delta^{j}_{\alpha}+u^{j}u_{\alpha}\right]T^{\alpha\beta}_{\ ;\beta}\right)=0, \tag*{(41)}

where the spatial index j=(r,θ,ϕ)j=(r,\theta,\phi); TαβT^{\alpha\beta} is the stress-energy tensor for the background magnetized neutron star

Tαβ=(P+ρ+H2)uαuβ+(P+H22)gαβ−HαHβ,(42)T^{\alpha\beta}=(P+\rho+H^{2})u^{\alpha}u^{\beta}+\left(P+\frac{H^{2}}{2}\right)g^{\alpha\beta}-H^{\alpha}H^{\beta}, \tag*{(42)}

where gαβg^{\alpha\beta} is the (inverse) spacetime metric. We assume the background equilibrium star is described by a perfect fluid under zero strain, and thus the stress is given by the isotropic pressure. The shear stress due to elasticity only contributes at the perturbative level in Eqs. Eqs. (40) and (41) after perturbing TαβT^{\alpha\beta} (see [17] for details).

When attempting to separate variables of the system of equations using Eq. (39), it has been suggested by [17] that the presence of a mixed poloidal-toroidal magnetic field causes the eigenfunction of a single mode, which would initially correspond to a single value of ℓ=ℓ0\ell=\ell_{0} in the non-magnetized case, to couple with eigenfunctions of other values of ℓ\ell, namely ℓ0−2\ell_{0}-2 and ℓ0+2\ell_{0}+2. However, it has been found that this coupling is unimportant for calculating the lowest-order mode. We refer the reader to Figure 3 in [19] for a comparison of the frequencies of fundamental torsional modes of ℓ=2\ell=2 with and without taking into account the ℓ±2\ell\pm2 couplings. To simplify the problem and since we will only focus here on the fundamental torsional modes of ℓ=2\ell=2, the ℓ±2\ell\pm2 couplings will be ignored in the following. The required final pulsation equation obtained from combining Eqs. (40) and (41) is given by [17]

d2Ydr2+[(1+2λ1)a12μπr4]d2Ydr2+{(4r+dνdr−dλdr+1μdμdr)+(1+2λ1)a1μπr4[(dνdr−dλdr)a1+2da1dr]}dYdr+e2λ−2ν[(ρ+P)σ2μ−e2ν(ℓ−1)(ℓ+2)r2]Y+{(1+2λ1)a12μπr4e2λ−2νσ2−λ12μπr4(da1dr)2e−2ν+(ℓ−1)(ℓ+2)λ12μπr4(da1dr)2+(2+5λ1)a12μπr4[(dνdr−dλdr)da1dr+d2a1dr2]}Y=0,(43)\begin{aligned} \frac{d^{2}Y}{dr^{2}}&+\left[(1+2\lambda_{1})\frac{a_{1}^{2}}{\mu\pi r^{4}}\right]\frac{d^{2}Y}{dr^{2}}+\left\{\left(\frac{4}{r}+\frac{d\nu}{dr}-\frac{d\lambda}{dr}+\frac{1}{\mu}\frac{d\mu}{dr}\right)+(1+2\lambda_{1})\frac{a_{1}}{\mu\pi r^{4}}\left[\left(\frac{d\nu}{dr}-\frac{d\lambda}{dr}\right)a_{1}+2\frac{da_{1}}{dr}\right]\right\}\frac{dY}{dr}\\ &+e^{2\lambda-2\nu}\left[(\rho+P)\frac{\sigma^{2}}{\mu}-e^{2\nu}\frac{(\ell-1)(\ell+2)}{r^{2}}\right]Y+\left\{(1+2\lambda_{1})\frac{a_{1}^{2}}{\mu\pi r^{4}}e^{2\lambda-2\nu}\sigma^{2}-\frac{\lambda_{1}}{2\mu\pi r^{4}}\left(\frac{da_{1}}{dr}\right)^{2}e^{-2\nu}\right.\\ &\left.\quad+(\ell-1)(\ell+2)\frac{\lambda_{1}}{2\mu\pi r^{4}}\left(\frac{da_{1}}{dr}\right)^{2}+(2+5\lambda_{1})\frac{a_{1}}{2\mu\pi r^{4}}\left[\left(\frac{d\nu}{dr}-\frac{d\lambda}{dr}\right)\frac{da_{1}}{dr}+\frac{d^{2}a_{1}}{dr^{2}}\right]\right\}Y=0, \tag*{(43)} \end{aligned}

where λ1=−ℓ(ℓ+1)(2ℓ−1)(2ℓ+3)\lambda_{1}=-\frac{\ell(\ell+1)}{(2\ell-1)(2\ell+3)}. Using Eq. (9) and the variables

Y1≡Yr1−ℓ,Y2≡[μ+(1+2λ1)a12πr4]eν−λr2−ℓdYdr,(44)Y_{1}\equiv Yr^{1-\ell},\qquad Y_{2}\equiv\left[\mu+(1+2\lambda_{1})\frac{a_{1}^{2}}{\pi r^{4}}\right]e^{\nu-\lambda}r^{2-\ell}\frac{dY}{dr}, \tag*{(44)}

Eq. (43) then becomes two equations in terms of Y1Y_{1} and Y2Y_{2}:

dY1dr=−ℓ−1rY1+πr3πr4μ+(1+2λ1)a12e−ν+λY2,(45)\frac{dY_{1}}{dr}=-\frac{\ell-1}{r}Y_{1}+\frac{\pi r^{3}}{\pi r^{4}\mu+(1+2\lambda_{1})a_{1}^{2}}e^{-\nu+\lambda}Y_{2}, \tag*{(45)}
dY2dr=−{[ρ+P+(1+2λ1)a12πr4−λ1e−2λ2πr2(da1dr)2]σ2r2e2(λ−ν)−(ℓ−1)(ℓ+2)[μe2λr−λ12πr3(da1dr)2]\frac{dY_{2}}{dr}=-\left\{\left[\rho+P+(1+2\lambda_{1})\frac{a_{1}^{2}}{\pi r^{4}}-\frac{\lambda_{1}e^{-2\lambda}}{2\pi r^{2}}\left(\frac{da_{1}}{dr}\right)^{2}\right]\sigma^{2}r^{2}e^{2(\lambda-\nu)}-(\ell-1)(\ell+2)\left[\frac{\mu e^{2\lambda}}{r}-\frac{\lambda_{1}}{2\pi r^{3}}\left(\frac{da_{1}}{dr}\right)^{2}\right]\right.
+(2+5λ1)a1e2λπr3[a1r2−ξ2e−2νa12−2πf0(ρ+P)r2]}eν−λY1−ℓ+2rY2.(46)\left.\quad+(2+5\lambda_{1})\frac{a_{1}e^{2\lambda}}{\pi r^{3}}\left[\frac{a_{1}}{r^{2}}-\frac{\xi^{2}e^{-2\nu}a_{1}}{2}-2\pi f_{0}(\rho+P)r^{2}\right]\right\}e^{\nu-\lambda}Y_{1}-\frac{\ell+2}{r}Y_{2}. \tag*{(46)}

This is the first-order system of equations we solved for the real eigenvalues σ\sigma.

Boundary conditions

Boundary conditions are needed to be imposed to solve the above system of equations for the mode frequencies. As torsional modes are confined to the crust region, a zero traction condition at the core-crust interface and a zero torque condition at the stellar surface implies that Y2=0Y_2=0 at both r=R−δrr=R-\delta r and r=Rr=R [17]. However, it should be pointed out that in reality one should also consider the coupling of the crust to the fluid core [19]. In this work, we ignore the coupling with the fluid core to demonstrate the effects of magnetic-field-dependent EOS on the spectrum of torsional modes.

Code tests

Before studying the effects of Landau-Rabi quantization, we first test the torsional mode frequencies ℓfn_\ell f_n produced by our code with the results presented in [17, 19, 68], which do not consider the effects of magnetic field on the EOS.

Firstly, the mode frequencies in the non-magnetized limit ℓfn_\ell f_n are tested. Our results were compared with those in [17] for ℓfnˉ\bar{_\ell f_n} with n=0,1n=0,1 (which is the number of nodes along the coordinate rr, from the center to stellar surface) and ℓ=2,3,4,…,10\ell=2,3,4,\ldots,10 for their stellar models A+DH1.4/1.6_{1.4/1.6}, WFF3+DH1.4/1.6_{1.4/1.6}, A+NV1.4/1.6_{1.4/1.6} and WFF3+NV1.4/1.6_{1.4/1.6}. For all the cases, we find that our mode frequencies differ from those reported in [17] by only about 1%, which is much better than similar code tests reported in [18]. The deviations could be due to different numerical treatments of the tabulated EOS data in different codes.

The coefficients ℓαn_\ell\alpha_n appeared in the following fitting formula proposed in [17]

ℓfnℓfnˉ=1+ℓαn(BBμ)2(47)\frac{_\ell f_n}{\bar{_\ell f_n}}=\sqrt{1+{}_\ell\alpha_n\left(\frac{B}{B_\mu}\right)^2} \tag*{(47)}

for purely dipolar magnetic configurations are also tested with ℓ=2\ell=2 and n=0,1n=0,1, where Bμ=4×1015B_\mu=4\times10^{15} G. We again find that the values of ℓαn_\ell\alpha_n obtained for the same stellar models tested above differ from the results reported in [17] by only about 1%.

We then tested our mode frequencies 2f0_2f_0 with mixed poloidal-toroidal magnetic fields. In this part, no numerical data are given in [19]. So we compared our data with the results extracted digitally from the figures presented in [19]. The results are presented in Appendix A.

STRUCTURES OF NEUTRON STARS

As mentioned above, having the magnetic-field-dependent EOS, the structure of a neutron star in general varies with the strength and configuration of magnetic field. In this section, we will compare the differences in the stellar structures

between models constructed with and without the effects of magnetic field on the EOS. In particular, we focus on the relative differences in δr\delta r defined by

Δδr=δr′−δr∘δr∘,(48)\Delta_{\delta r}=\frac{\delta r'-\delta r^\circ}{\delta r^\circ}, \tag*{(48)}

where δr′\delta r' is the thickness of the crust after considering the effects of magnetic-field-dependent EOS; δr∘\delta r^\circ is the original thickness without the effects of magnetic field.

We shall focus on 1.4 M⊙1.4\,M_\odot neutron stars as our canonical models in the study. For our chosen BSk24 EOS, a 1.4M⊙1.4M_\odot neutron star has crust thickness δr∘=0.8722\delta r^\circ=0.8722 km and radius R∘=12.4360R^\circ=12.4360 km in the non-magnetized limit. A more massive 2.0M⊙2.0M_\odot model will be considered later in this work.

Normal fluid cores

We first consider the case for normal fluid core and plot Δδr\Delta_{\delta r} against the magnetic-field strength BB at the pole in the upper panel of Figure 5. In the figure, three different field configurations ξ=0\xi=0, 0.28, and 0.45 are considered, where ξ=0\xi=0 represents the purely dipole case. It can be seen that Δδr\Delta_{\delta r} in general increases with BB and ξ\xi. Even for a purely dipole field, Δδr\Delta_{\delta r} can exceed 10% when the strength of magnetic field is of the order of 101610^{16} G. For more complex magnetic field configurations with larger values of ξ\xi, Δδr\Delta_{\delta r} can even exceed 20% when BB reaches over 101610^{16} G. The strength of the magnetic field in the crustal region generally increases with ξ\xi (i.e., a stronger toroidal component) and hence leads to greater effects on the stellar structure via the EOS.

Three-panel plots showing magnetic-field-dependent crust thickness and parameters

Figure 5. Upper panel: Δδr\Delta_{\delta r} as a function of BB for ξ=0\xi=0, 0.28 and 0.45. Middle panel: ξb\xi_b and ξc\xi_c as a function of BB. Lower panel: ξa\xi_a as a function of BB. Here BB is the magnetic-field strength at the pole of a magnetized neutron star, which is assumed to have a normal fluid core.

Increasing the crust thickness δr\delta r also affects the allowed parameter space for physical magnetic-field configurations mentioned in Sec. II B. In the middle and bottom panels of Figure 5, we show how the parameters ξa\xi_a, ξb\xi_b, and ξc\xi_c defined in Eq. (13) change with the field strength BB. It is seen that ξb\xi_b and ξc\xi_c remain essentially unchanged. However, ξa\xi_a decreases more significantly as BB increases. As a result, the gap between the two physical ranges I and II defined in Eq. (13) increases with BB. It should be noted that as the parameter ξ\xi increases towards the limiting value ξa\xi_a, the field configuration in general would become more complex. In other words, for physical magnetic-field configurations defined in the physical range I, increasing the field strength BB would effectively increase the complexity of the field configuration for a fixed value of ξ\xi. Similar effect also exists in the physical range II as it can be seen that its upper limit ξc\xi_c also decreases, though not as significant as ξa\xi_a, when BB increases towards 4×10164 \times10^{16} G. The consequences of this important effect will be discussed in Sec. VII.

In order to understand why Δδr\Delta_{\delta r} is positive and increases with the strength of the magnetic field, we first note that, for the same value of pressure PP, increasing the field strength will result in a smaller value of energy density ρ\rho near the transition between the inner and outer crustal regions, which happens at the density range between about 5×10115 \times10^{11} g cm−3^{-3} and 101210^{12} g cm−3^{-3} for different field strengths shown in Figure 1. According to Eq. (4), a smaller value of ρ\rho leads to a smaller value of ∣dP/dr∣|dP/dr|, and hence a flatten density profile which can extend further away until it reaches the surface.

To illustrate this phenomenon, we show the normalized energy-density profiles (blue lines) and pressure profiles (yellow lines) in the crustal region for three different star models in Figure 6. The solid lines represent the profiles for a unmagnetized neutron star. The dotted lines are the profiles for a magnetized star with a purely dipole field (ξ=0\xi= 0) and B=4×1016B = 4 \times10^{16} G, which branch off from the solid lines leading to a thicker crust. However, the dotted lines in fact drop more rapidly near the stellar surface compared to the solid lines. This can be understood from Figure 1 that increasing the field strength for a fixed pressure leads to a higher density at the low density region, and hence a steeper ∣dP/dr∣|dP/dr| near the surface. One can also see from Figure 6 that, for the same field strength BB at the pole, increasing the value of ξ\xi from 0 to 0.28 (dashed lines) can further increase the thickness of the crust.

Normalized energy-density and pressure profiles in the crustal region

Figure 6. The profiles of energy density ρ\rho and pressure PP normalized by their own values ρcrust\rho_{\rm crust} and PcrustP_{\rm crust}, respectively, at the base of the crust. The solid lines represent the profiles for an unmagnetized neutron star. The dotted and dashed lines are the profiles for magnetized stars with the same magnetic field strength B=4×1016B = 4 \times10^{16} G, but with different field configurations ξ=0\xi= 0 and 0.28.

Superconducting cores

For neutron stars with a superconducting core, Δδr\Delta_{\delta r} also increases with BB and ξ\xi as shown in the upper panel of Figure 7. However, Δδr\Delta_{\delta r} now becomes more sensitive to the field strength comparing to the case of normal fluid core. For neutron stars with a normal fluid core, Δδr\Delta_{\delta r} increases to about 5% for ξ=0.45\xi= 0.45 and B=4×1015B = 4 \times10^{15} G as shown in Figure 5. For the same value of BB, superconducting-core models can reach up to 10% when ξ≈1\xi\approx1. The bottom panel of Figure 7 shows that the parameter ξy\xi_y, which defines the physical range of parameter in Eq. (15), also decreases with increasing BB.

Fractional crust-thickness change and superconducting-core parameter

Figure 7. Upper panel: Δδr\Delta_{\delta r} as a function of BB for ξ=0\xi= 0, 0.9, 1.1 and 1.2. Lower panel: ξy\xi_y as a function of BB. The neutron star is assumed to have a superconducting core and the magnetic field is confined in the crust.

MODE FREQUENCIES

As mentioned above, the torsional oscillation modes ℓfn\ell f_n for neutron stars with mixed poloidal-toroidal magnetic fields, but without considering the effect of Landau-Rabi quantization, have been studied [18]. In this work, we only focus on the ℓ=2\ell= 2 fundamental torsional oscillation modes and investigate the changes in the mode frequencies when the effect of magnetic field on the EOS is considered. We define the relative differences in mode frequencies by

Δf=f′−f∘f∘,(49)\Delta_f = \frac{f' - f^\circ}{f^\circ}, \tag*{(49)}

where f′f' and f∘f^\circ are the frequencies calculated with and without the effect of magnetic field on the EOS, respectively.

Normal fluid cores

In the upper panel of Figure 8, we plot f′f' (dotted lines) and f∘f^\circ (solid lines) against BB for neutron stars with a normal fluid core and magnetic field configurations defined by ξ=0\xi= 0, 0.280.28 and 0.450.45. One can observe that f∘f^\circ increases rapidly with BB and ξ\xi when the field strength is larger than about 101510^{15} G. This phenomenon has been discussed in [17, 18]. A semi-analytical understanding is provided in Appendix B. Including the effect of magnetic field on the EOS (dotted lines), the mode frequency further increases for fixed values of BB and ξ\xi. The difference between f′f' and f∘f^\circ also increases with the complexity (i.e., the value of ξ\xi) of the field configuration. The difference Δf\Delta_f is plotted in the lower panel of Figure 8. For B=4×1015B = 4 \times10^{15} G, Δf\Delta_f can reach only up to a few percent even for the case ξ=0.45\xi= 0.45, which is already near the upper limit of the physical range II defined in Eq. (13). However, Δf\Delta_f can rise up to about 10%10\% when the field strength B≈2×1016B \approx2 \times10^{16} G for ξ=0.28\xi= 0.28 and 0.450.45. For the case of purely dipole field (ξ=0\xi= 0), Δf\Delta_f can reach only up to about 5%5\% even when B=4×1016B = 4 \times10^{16} G. As discussed in Sec. VI, for a fixed value of ξ\xi, including the effect of magnetic field on the EOS would effectively increase the complexity of the field configuration, and hence leading to the increase of Δf\Delta_f as BB increases.

Mode frequency and relative frequency difference as functions of magnetic field strength for different magnetic field configurations

Figure 8. Upper panel: Mode frequency ff as a function of BB for ξ=0\xi= 0, 0.280.28 and 0.450.45. The solid lines correspond to f∘f^\circ, while the dotted lines correspond to f′f'. Lower panel: Δf\Delta_f as a function of BB. The neutron star is assumed to have a normal fluid core and the magnetic field extends throughout the star.

Superconducting cores

For neutron stars with a superconducting core, the results differ from the above results qualitatively. Similar to Figure 8, we plot f∘f^\circ and f′f' in the upper panel of Figure 9 to illustrate how the mode frequencies change with BB and ξ\xi. In this case, it is seen that f∘f^\circ increases with BB but decreases with ξ\xi. Hence, in contrast to neutron stars with a normal fluid core, increasing the complexity of the field configuration (i.e., the value of ξ\xi) would decrease the mode frequency. A semi-analytical understanding is provided in Appendix B. Again, as discussed in Sec. VI, including the effect of magnetic field on the EOS effectively increases the complexity of the field configuration for a fixed value of ξ\xi, since the upper limit ξγ\xi_\gamma of the physical range defined in Eq. (15) decreases with increasing BB. This is why the frequency f′f' is smaller than f∘f^\circ for given values of BB and ξ\xi. The difference Δf\Delta_f between the two frequencies increases with BB and ξ\xi as shown in

Upper and lower panels showing $f$ and $\Delta f$ as functions of $B$

Figure 9. Upper panel: ff as a function of BB for ξ=0\xi= 0, 0.90.9, 1.11.1 and 1.21.2. The solid and dotted lines correspond to f∘f^{\circ} and f′f^{\prime}, respectively. Lower panel: Δf\Delta f as a function of BB. The neutron star is assumed to have a superconducting core and the magnetic field is confined in the crust.

the lower panel of Figure 9. In particular, the magnitude of Δf\Delta_f can reach up to about 20%20\% when B=4×1015B = 4 \times10^{15} G for the case ξ=1.2\xi= 1.2, which is close to the upper limit of ξγ\xi_\gamma as shown in Figure 7.

Magnetic field complexities

In the above discussion, we characterize the complexity of a magnetic field configuration by the parameter ξ\xi, which is compared to a purely dipole configuration (ξ=0\xi= 0) constructed with a TOV background without the effect of Landau-Rabi quantization (i.e., an unmagnetized background). We see that the value of ξ\xi, and hence the complexity of the field configuration, plays an important role in the oscillation mode frequencies. However, ξ\xi is not intuitive and it may be better to compare a physical quantity associated with the magnetic energy. In order to better capture the difference between a generic field configuration and a purely dipole field on an unmagnetized background, we define the quantity

⟨ΔE⟩=(ΔEr)2+(ΔEθ)2+(ΔEϕ)2,(50)\langle\Delta E \rangle= \sqrt{(\Delta E_r)^2 + (\Delta E_\theta)^2 + (\Delta E_\phi)^2}, \tag*{(50)}

where ΔEi\Delta E_i (i=r,θ,ϕi = r, \theta, \phi) stands for the difference of the magnitude of EiE_i between the two field configurations at the same value of BB, in which EiE_i is the mean magnetic energy density of the ii-component of the magnetic field stored in the crust. In Figure 10, we plot ⟨ΔE⟩/EBμ\langle\Delta E \rangle/ E_{B_\mu} against BB, where EBμE_{B_\mu} is the energy density for the field strength of Bμ=4×1015B_\mu= 4 \times10^{15} G (which is introduced in Eq. (47)) in the crust. For a fixed value of BB, Figure 10 shows that ⟨ΔE⟩\langle\Delta E\rangle increases with ξ\xi for both the normal fluid (upper panel) and superconducting cores (lower panel). Hence, ξ\xi can serve as a proxy for the physical quantity ⟨ΔE⟩\langle\Delta E\rangle which quantifies the complexity of the field configuration relative to a purely dipole field on an unmagnetized background.

Upper and lower panels showing normalized $\langle\Delta E\rangle$ as a function of $B$

Figure 10. ⟨ΔE⟩\langle\Delta E\rangle as a function of BB for 1.4M⊙1.4M_{\odot} neutron stars with a normal fluid (upper panel) and superconducting core (lower panel). Here we normalize ⟨ΔE⟩\langle\Delta E\rangle by EBμE_{B_{\mu}} for convenience, where EBμE_{B_{\mu}} is the magnetic energy density for the magnetic field strength Bμ=4×1015B_{\mu} = 4 \times10^{15} G.

Effects of stellar mass

So far we have focused on a 1.4M⊙1.4M_{\odot} neutron star. Let us now present the results for a more massive 2.0M⊙2.0M_{\odot} model with δr∘=0.4615\delta r^{\circ} = 0.4615 km and R∘=12.2298R^{\circ} = 12.2298 km in the non-magnetized limit. We shall see how Δδr\Delta\delta r and Δf\Delta f change for neutron stars with different masses. It is worth noticing that the mode frequency f∘f^{\circ} for a 2.0M⊙2.0M_{\odot} neutron star model is not the same as that of the 1.4M⊙1.4M_{\odot} model. For comparison, we also present the corresponding values of ξa\xi_a, ξb\xi_b, ξc\xi_c and ξγ\xi_{\gamma} for 1.4M⊙1.4M_{\odot} and 2.0M⊙2.0M_{\odot} neutron star models in the weak field limit (B≪1015B \ll10^{15} G) in Table II.

Normal fluid cores

We first compare how the differences in the thickness of the crust Δδr\Delta\delta r defined in Eq. (48), relative to the nonmagnetic limit, depend on the stellar mass in the upper panel of Figure 11. In the figure, the blue lines correspond to the purely dipole field configuration (ξ=0\xi= 0) for star models with 1.4M⊙1.4M_{\odot} (solid line)

Upper and lower panels showing $\Delta_{\delta r}$ and $\Delta_f$ as functions of $B$

Figure 11. Upper panel: Δδr\Delta_{\delta r} as a function of BB for ξ=0\xi= 0 (blue lines) and 0.28 (red lines). Lower panel: Δf\Delta_f as a function of BB. The solid and dotted-dashed lines correspond to neutron star models with 1.4M⊙1.4M_\odot and 2.0M⊙2.0M_\odot, respectively. The neutron stars are assumed to have a normal fluid core, and the magnetic field extends throughout the stars.

and 2.0M⊙2.0M_{\odot} (dotted-dashed line). It is seen that the two lines agree quite well with each other even up to B=4×1016B = 4 \times10^{16} G, and hence Δδr\Delta\delta r is insensitive to the stellar mass for ξ=0\xi= 0. For a more complex field configuration defined by ξ=0.28\xi= 0.28 (red lines), the results are slightly more sensitive to the stellar mass. Similarly, we plot the corresponding differences in the mode frequency Δf\Delta f in the lower panel of Figure 11 and notice that Δf\Delta f is somewhat less sensitive to the stellar mass for ξ=0.28\xi= 0.28 comparing to Δδr\Delta\delta r.

Superconducting cores

For neutron stars with a superconducting core, the effects of stellar mass are different from those discussed above. Similar to Figure 11, we consider how Δδr\Delta_{\delta r} depends on stellar mass in the upper panel of Figure 12. The yellow (blue) lines indicate the results for stellar models constructed by the field configuration ξ=0.9\xi= 0.9 (ξ=1.1\xi= 1.1). The 1.4M⊙1.4M_\odot and 2.0M⊙2.0M_\odot models are represented by the solid and dotted-dashed lines, respectively. Unlike the situation for a normal fluid core, it is seen that Δδr\Delta_{\delta r} is now more sensitive to the stellar mass as the blue solid and dotted-dashed lines, and similarly for the yellow lines, already deviate from each other significantly when B=4×1015 GB = 4 \times10^{15}\,\mathrm{G}. Nevertheless, Δδr\Delta_{\delta r} in general increases with the stellar mass for both normal fluid and superconducting cores. In the lower panel of Figure 12, we plot the corresponding Δf\Delta_f for neutron stars with a superconducting core. In contrast to the case for a normal fluid core, where Δf\Delta_f is weakly dependent on the stellar mass, it is seen that Δf\Delta_f now depends more sensitively on the stellar mass. In particular, the magnitude of Δf\Delta_f increases with decreasing stellar mass for a given magnetic field strength.

Upper and lower panels showing $\Delta_{\delta r}$ and $\Delta_f$ as functions of $B$

Figure 12. Upper panel: Δδr\Delta_{\delta r} as a function of BB for ξ=0.9\xi= 0.9 (yellow lines) and 1.1 (blue lines). Lower panel: Δf\Delta_f as a function of BB. The solid and dotted-dashed lines correspond to neutron star models with 1.4M⊙1.4M_\odot and 2.0M⊙2.0M_\odot, respectively. The neutron stars are assumed to have a superconducting core, and the magnetic field is confined in the crust.

Discussion

The intense magnetic field of a magnetar can affect the structure and EOS of the crust of the star. In this paper, we have studied the effects of Landau-Rabi quantization of electron motion on the crustal torsional oscillations of magnetars in general relativity under the Cowling approximation. For the EOS model, we employ the study of [41] which incorporates the effects of Landau-Rabi quantization by treating the inner and outer crusts consistently based on the nuclear-energy density functional theory. We also consider mixed poloidal-toroidal magnetic field configurations for our stellar models. In our study, the magnetic field configuration first contributes to the calculation of the torsional oscillation modes through its effect on the magnetic-field-dependent EOS due to the Landau-Rabi quantization, and hence the resulting magnetized stellar model as discussed in § 1.3. The magnetic field then contributes at the level of the perturbation equations that determine the oscillation modes. We assume spherically symmetric stellar models and ignore the crust deformation [69] and the modification of the energy-momentum tensor due to the strong magnetic field [52]. These effects are expected to be non-negligible only when the strength of the magnetic field is of the order of 1018 G10^{18}\,\mathrm{G}, which is far beyond our range of study here and the typical strength of magnetic field of magnetars.

The magnetic field configurations are characterized by the strength of the magnetic field BB at the pole of the stellar surface and the parameter ξ\xi, which represents the ratio between the toroidal and poloidal components of the magnetic field. Our numerical results show that the crust thickness δr\delta r increases with both BB and ξ\xi when the effect of magnetic field is considered in the EOS. For a more complex magnetic field configuration with a nonzero ξ\xi, δr\delta r can increase by more than 20% when B∼1016 GB \sim10^{16}\,\mathrm{G} for neutron stars with a normal fluid core. On the other hand, for neutron stars with a superconducting core, δr\delta r can already increase by 10% when B∼1015B \sim10^{15} G.

We focus on the fundamental ℓ=2\ell= 2 torsional oscillation modes and study the changes in the mode frequencies when the effect of the magnetic field on the EOS is incorporated. It is found that the magnitude of the fractional change in mode frequency ∣Δf∣|\Delta_f| increases with BB and ξ\xi. For the models that we have considered, ∣Δf∣|\Delta_f| can exceed 10% when B∼1016B \sim10^{16} G for neutron stars with a normal fluid core. For neutron stars with a superconducting core, however, ∣Δf∣|\Delta_f| can even approach 20% at a smaller field strength B∼1015B \sim10^{15} G, which lies at the top end of the observed range of magnetic field for magnetars. It is also worth to note that Δf\Delta_f is positive (negative) for neutron stars with a normal fluid (superconducting) core, and hence the effect of Landau-Rabi quantization would increase (decrease) the mode frequency correspondingly. We have also studied the effect of stellar mass and found that Δf\Delta_f is not sensitive to the stellar mass for models with a normal fluid core. However, for a superconducting core, Δf\Delta_f depends more sensitively on the stellar mass and increases with decreasing mass, especially for field strength larger than B∼1015B \sim10^{15} G.

Let us now put together the results of all the stellar models studied in this work and discuss the implication from an observational perspective. While the mechanism for the magnetar giant flares and the nature of the observed QPOs in the late-time tail phase are still not well understood, it is believed that the magneto-elastic oscillations of the star must play a role. Although the role of torsional oscillation modes of magnetars in explaining (at least) some of the observed QPOs is far from clear, let us adopt this simple scenario as an example to illustrate the implication of our numerical results. In particular, we want to clarify the situation under which we need to worry

about the effect of Landau-Rabi quantization and the level of deviations one could have if this effect is not considered in the theoretical calculation of the oscillation modes, when comparing to the observations.

In Figure 13, we plot ∣Δf∣|\Delta_f| as a function of the mode frequencies f∘f^\circ without considering the effect of magnetic field on the EOS for different models studied in this work, where the subscripts nn and ss in the parameter ξ\xi indicate that the neutron star models consist of a normal fluid and superconducting core, respectively. The solid and dashed lines with the same color represent 1.4M⊙1.4M_\odot and 2.0M⊙2.0M_\odot models for the same value of ξ\xi, respectively. In addition, the vertical lines denote the frequencies of some of the observed QPO as mentioned in Sec. I. Figure 13 shows that the effect of Landau-Rabi quantization is not important when one considers the observed QPOs with smaller frequencies (≲80\lesssim80 Hz), but becomes relevant for higher frequencies modes. It can be seen that ∣Δf∣|\Delta_f| is always less than 1% for f∘f^\circ being around 40 Hz. However, ∣Δf∣|\Delta_f| exceeds 3% for f∘f^\circ being around 90 Hz and approaching 10% for ff being around 150 Hz.

Plot of $|\Delta_f|$ as a function of $f^\circ$ for different models

Figure 13. ∣Δf∣|\Delta_f| as a function of f∘f^\circ for different models. The subscripts nn and ss in the parameter ξ\xi indicate that the neutron star models consists of a normal fluid and superconducting core, respectively. The solid lines and dashed lines represent the results for 1.4 M⊙1.4\,M_\odot and 2.0 M⊙2.0\,M_\odot neutron star models, respectively. The vertical lines denote some of the observed QPO frequencies in magnetar giant flares.

In addition, one can see that different models show similar trends in Figure 13, which can be fitted approximately by

∣Δf∣=K(f∘)1.12±0.12,(51)|\Delta_f| = K\left(f^\circ\right)^{1.12\pm0.12}, \tag*{(51)}

where KK is a constant depending on the background star model. This fit applies to the range for ∣Δf∣|\Delta_f| being between 1% and 10%. On the other hand, we plot ∣Δf∣|\Delta_f| against Δδr\Delta_{\delta r} in Figure 14. As one may expected, there should be some correlation between the two quantities. Different models also show similar trends in Figure 14, which can be fitted for Δδr\Delta_{\delta r} in the range between 2% and 10% approximately by

Plot of $|\Delta_f|$ as a function of $\Delta_{\delta r}$ for different models

Figure 14. ∣Δf∣|\Delta_f| as a function of Δδr\Delta_{\delta r} for different models, similar to Figure 13.

∣Δf∣=K′(Δδr)1.11±0.04,(52)|\Delta_f| = K^\prime\left(\Delta_{\delta r}\right)^{1.11\pm0.04}, \tag*{(52)}

where K′K^\prime is a constant depending on the background star model.

In conclusion, depending on the neutron star models and magnetic field configurations, we have shown that the effect of Landau-Rabi quantization of electron motion in the crust can change the frequencies of the fundamental torsional oscillation modes by up to 10% to 20% for magnetic field strength B∼1015−1016B \sim10^{15} - 10^{16} G, which is within the measured surface field strength of magnetars. However, it should be noted that all the results presented here only hold in cases where the important coupling to the core and a more realistic treatment of the boundary conditions are not considered. Our work provides a benchmark on the parameter space of stellar models and observation precision within which one needs to consider the Landau-Rabi quantization. In the future, we plan to extend our study to other oscillation modes such as the interfacial and shear modes, which also depend sensitively on the properties of the solid crust and may be excited in magnetar starquakes.

Acknowledgements

The work of NC was financially supported by the Fonds de la Recherche Scientifique (Belgium) under Grant No. PDR T.004320. LML is supported by a grant from the Research Grants Council of the Hong Kong SAR, China [Project No: CUHK 14304322].

CODE TESTS WITH MIXED POLOIDAL-TOROIDAL MAGNETIC FIELDS

In this appendix, we verify our code by comparing our numerical results for torsional oscillation modes with mixed poloidal-toroidal magnetic field configurations with the data published in [19]. We consider their 1.4 M⊙M_\odot stellar model described by the following EOS:

P=0.00936n0mb(nˉn0)2.46,(53)P = 0.00936 n_0 m_b \left(\frac{\bar{n}}{n_0}\right)^{2.46}, \tag*{(53)}
ρ=nˉmb+P1.46,(54)\rho= \bar{n}m_b + \frac{P}{1.46}, \tag*{(54)}
μ=ρvs2,(55)\mu= \rho v_s^2, \tag*{(55)}

where n0=0.1 fm−3n_0 = 0.1\ \mathrm{fm}^{-3}, mb=1.66×10−24 gm_b = 1.66 \times10^{-24}\,\mathrm{g} and vs=μ/(ρ+P)v_s = \sqrt{\mu/(\rho+ P)} is the velocity of the shear wave as measured by a local observer which is assumed here to be fixed at its typical value of 108 cm s−110^8\,\mathrm{cm\,s}^{-1} [15]. In the upper panel of Figure 15, we plot the fundamental torsional oscillation mode frequencies ff against the normalized magnetic field strength B/BμB/B_\mu (Bμ=4×1015B_\mu= 4 \times10^{15} G) for this background star with a normal fluid core and magnetic field configurations defined by ξ=0.1\xi= 0.1 and 0.31. The solid lines are the results extracted from the figures in [19] and the dots are our numerical results. Similarly, the lower panel of Figure 15 shows the comparison for neutron stars with a superconducting core and magnetic field configurations defined by ξ=0.8\xi= 0.8 and 1.9. We see that our results match those presented in [19] very well.

Figure 15. Comparison between our numerical results (data points) and the results (solid lines) presented in [19] for a 1.4M⊙M_\odot neutron star model with different magnetic field configurations (see text for more details). The upper (lower) panel plots the fundamental torsional oscillation mode frequency ff against the normalized field strength B/BμB/B_\mu for the case of a normal fluid (superconducting) core.

SEMI ANALYTICAL UNDERSTANDING OF TORSIONAL OSCILLATION MODES WITH MIXED POLOIDAL-TOROIDAL MAGNETIC FIELDS

To understand the behaviours of the torsional oscillation mode frequency ff with mixed poloidal-toroidal magnetic fields, we can apply approximations on Eq. (43). In this appendix, we do not consider the effects of magnetic field on the EOS.

Normal fluid core

Let us first consider neutron stars with a normal fluid core. To simplify the problem, we only focus on the Newtonian limit. As the torsional oscillations are confined in the solid crust, we only need to analyze the crustal region. In addition, thin crust will be assumed in the following. For a thin crust, we have d2Y/dr2≈dY/dr≈0d^{2}Y/dr^{2}\approx dY/dr\approx0 in our analysis due to the boundary conditions at the base of the crust and the stellar surface, which require Y2=dY/dr=0Y_{2}=dY/dr=0.

As we are only interested in how the mode angular frequency σ\sigma scales with the magnetic field strength BB and the parameter ξ\xi qualitatively, we can perform an order-of-magnitude approximation for Eq. (46) and obtain the following result:

σ2≈(ℓ−1)(ℓ+2)R2⟨vs2⟩+2+5λ12π⟨ρ⟩R4[⟨4πf0ρR2a1⟩+ξ2⟨a12⟩](56)\sigma^{2}\approx\frac{(\ell-1)(\ell+2)}{R^{2}}\left\langle v_{s}^{2}\right\rangle+\frac{2+5\lambda_{1}}{2\pi\langle\rho\rangle R^{4}}\left[\left\langle4\pi f_{0}\rho R^{2}a_{1}\right\rangle+\xi^{2}\left\langle a_{1}^{2}\right\rangle\right] \tag*{(56)}

where ⟨x⟩\langle x\rangle is the average value of xx throughout the crust. In the absence of magnetic field (i.e., a1=0a_{1}=0), we obtain the non-magnetic eigenvalue σ2≈σˉ2≡(ℓ−1)(ℓ+2)⟨vs2⟩/R2\sigma^{2}\approx\bar{\sigma}^{2}\equiv(\ell-1)(\ell+2)\langle v_{s}^{2}\rangle/R^{2}, which agrees with the result obtained in [70] except that a metric function of order unity is ignored in our analysis. We thus have

σ2σˉ2≈1+2+5λ12πσˉ2⟨ρ⟩R4[⟨4πf0ρR2a1⟩+ξ2⟨a12⟩].(57)\frac{\sigma^{2}}{\bar{\sigma}^{2}}\approx1+\frac{2+5\lambda_{1}}{2\pi\bar{\sigma}^{2}\langle\rho\rangle R^{4}}\left[\left\langle4\pi f_{0}\rho R^{2}a_{1}\right\rangle+\xi^{2}\left\langle a_{1}^{2}\right\rangle\right]. \tag*{(57)}

On the other hand, the source term f0(ρ+P)R2f_{0}(\rho+P)R^{2} in Eq. (9) scales like BB, as that is the current distribution producing the magnetic field. As a1∝Ba_{1}\propto B at the surface, we can approximate ⟨a12⟩∝B2\left\langle a_{1}^{2}\right\rangle\propto B^{2} and obtain the final relation

σ2σˉ2=1+CℓB2+DℓB2ξ2,(58)\frac{\sigma^{2}}{\bar{\sigma}^{2}}=1+C_{\ell}B^{2}+D_{\ell}B^{2}\xi^{2}, \tag*{(58)}

where CℓC_{\ell} and DℓD_{\ell} are some positive constants depending on ℓ\ell and the stellar models. Rewriting the terms give

σσˉ=[1+Cℓ(1+Eℓξ2)B2]1/2,(59)\frac{\sigma}{\bar{\sigma}}=\left[1+C_{\ell}(1+E_{\ell}\xi^{2})B^{2}\right]^{1/2}, \tag*{(59)}

where Eℓ=Dℓ/CℓE_{\ell}=D_{\ell}/C_{\ell}. This is exactly the results provided numerically in [18]. For purely dipole field (ξ=0\xi=0), it reduces to the fitting formula proposed in [17] (i.e., Eq. (47)).

Superconducting core

As stated in [19], the θ\theta-component of magnetic field will be much stronger than the other components when the magnetic field is confined only in the crustal region. So, here we repeat the same analysis above but keep the terms with (da1/dr)2(da_{1}/dr)^{2} instead of a12a_{1}^{2} (see Eq. (10)). The final result can be written as

σ2σˉ2≈1+CℓB2+Fℓ⟨(da1dr)2⟩,(60)\frac{\sigma^{2}}{\bar{\sigma}^{2}}\approx1+C_{\ell}B^{2}+F_{\ell}\left\langle\left(\frac{da_{1}}{dr}\right)^{2}\right\rangle, \tag*{(60)}

where Fℓ=(ℓ−1)(ℓ+2)∣λ1∣2πσˉ2⟨ρ⟩R4F_{\ell}=(\ell-1)(\ell+2)\frac{|\lambda_{1}|}{2\pi\bar{\sigma}^{2}\langle\rho\rangle R^{4}}. Applying similar approximations as above and only considering the cases with ξ∼1\xi\sim1 for our numerical results, Eq. (9) can be approximated by

d2a1dr2+2MR2da1dr+ξ2a1≈ΓB,(61)\frac{d^{2}a_{1}}{dr^{2}}+\frac{2M}{R^{2}}\frac{da_{1}}{dr}+\xi^{2}a_{1}\approx\Gamma B, \tag*{(61)}

where Γ\Gamma is some constant depending on the stellar model. We have also neglected the term 2a1/r22a_{1}/r^{2} in Eq. (9) in the analysis as here it is much smaller than the other terms. In the following, M/R2M/R^{2} will be taken as 0.01 in our order-of-magnitude analysis. Solving this second-order ordinary differential equation, we obtain approximately

a1(r)≈χ1e−0.01rsin⁡(ξr)+χ2e−0.01rcos⁡(ξr)+ΓBξ2,(62)a_{1}(r)\approx\chi_{1}e^{-0.01r}\sin(\xi r)+\chi_{2}e^{-0.01r}\cos(\xi r)+\frac{\Gamma B}{\xi^{2}}, \tag*{(62)}

where χ1\chi_{1} and χ2\chi_{2} are some constants. By requiring a1(R−δr)=d2a1dr2(R−δr)=0a_{1}(R-\delta r)=\frac{d^{2}a_{1}}{dr^{2}}(R-\delta r)=0 (see Sec. II B for the boundary conditions), both χ1\chi_{1} and χ2\chi_{2} scale like B/ξ2B/\xi^{2}. As a result,

⟨da1dr⟩∝Bξ.(63)\left\langle\frac{da_{1}}{dr}\right\rangle\propto\frac{B}{\xi}. \tag*{(63)}

Eq. (60) can then be expressed as

σ2σˉ2=1+CℓB2+GℓB2ξ2,(64)\frac{\sigma^{2}}{\bar{\sigma}^{2}}=1+C_{\ell}B^{2}+G_{\ell}\frac{B^{2}}{\xi^{2}}, \tag*{(64)}

where GℓG_{\ell} is some constant depending on ℓ\ell and the stellar model. We have tested this relation with our numerical results to verify that the ξ\xi-dependence obtained is valid and the error is less than 10%10\% for the range of magnetic field strength considered. Therefore, given a fixed value of BB, increasing the value of ξ\xi will decrease σ\sigma.

For the cases with ξ=0\xi=0, the solution of Eq. (61) is given approximately by

a1(r)≈χ1e−0.01r+χ2+100ΓBr.(65)a_{1}(r)\approx\chi_{1}e^{-0.01r}+\chi_{2}+100\Gamma Br. \tag*{(65)}

Again, having a1(R−δr)=d2a1dr2(R−δr)=0a_{1}(R-\delta r)=\frac{d^{2}a_{1}}{dr^{2}}(R-\delta r)=0, we determine that χ1=0\chi_{1}=0 and χ2\chi_{2} is just some constant. Therefore, we have ⟨da1/dr⟩∝B\left\langle da_{1}/dr\right\rangle\propto B and

σ2σˉ2=1+JℓB2,(66)\frac{\sigma^{2}}{\bar{\sigma}^{2}}=1+J_{\ell}B^{2}, \tag*{(66)}

where JℓJ_{\ell} is some constant depending on ℓ\ell and the stellar model. This relation has been tested with our numerical results and the error is around 1%1\% only.

References

References

  1. [3]V. M. Kaspi and A. M. Beloborodov, Magnetars, Annual Review of Astronomy and Astrophysics 55, 261 (2017), arXiv:1703.00068 [astro-ph.HE].DOI
  2. [4]N. Rea and P. Esposito, Magnetar outbursts: an observational review, in High-Energy Emission from Pulsars and their Systems, Astrophysics and Space Science Proceedings, Vol. 21 (2011) p. 247, arXiv:1101.4472 [astro-ph.GA].DOI
  3. [5]K. Hurley, T. Cline, E. Mazets, S. Barthelmy, P. Butterworth, F. Marshall, D. Palmer, R. Aptekar, S. Golenetskii, V. Il’Inskii, D. Frederiks, J. McTiernan, R. Gold, and J. Trompka, A giant periodic flare from the soft γ-ray repeater SGR1900+14, Nature (London) 397, 41 (1999), arXiv:astro-ph/9811443 [astro-ph].DOI
  4. [6]K. Hurley, S. E. Boggs, D. M. Smith, R. C. Duncan, R. Lin, A. Zogler, S. Krucker, G. Hurford, H. Hudson, C. Wigger, W. Hajdas, C. Thompson, I. Mitrofanov, A. Sanin, W. Boynton, C. Fellows, A. von Kienlin, G. Lichti, A. Rau, and T. Cline, An exceptionally bright flare from SGR 1806-20 and the origins of short-duration γ-ray bursts, Nature (London) 434, 1098 (2005), arXiv:astro-ph/0502329 [astro-ph].
  5. [7]S. Mereghetti, D. Götz, A. von Kienlin, A. Rau, G. Lichti, G. Weidenspointner, and P. Jean, The First Giant Flare from SGR 1806-20: Observations Using the Anticoicidence Shield of the Spectrometer on INTEGRAL, Astrophys. J. Lett. 624, L105 (2005), arXiv:astro-ph/0502577 [astro-ph].
  6. [8]G. L. Israel, T. Belloni, L. Stella, Y. Rephaeli, D. E. Gruber, P. Casella, S. Dall’Osso, N. Rea, M. Persic, and R. E. Rothschild, The Discovery of Rapid X-Ray Oscillations in the Tail of the SGR 1806-20 Hyperflare, Astrophys. J. Lett. 628, L53 (2005), arXiv:astro-ph/0505255 [astro-ph].DOI
  7. [9]A. L. Watts and T. E. Strohmayer, Detection with RHESSI of High-Frequency X-Ray Oscillations in the Tailof the 2004 Hyperflare from SGR 1806-20, Astrophys. J. Lett. 637, L117 (2006), arXiv:astro-ph/0512630 [astro-ph].DOI
  8. [10]T. E. Strohmayer and A. L. Watts, The 2004 Hyperflare from SGR 1806-20: Further Evidence for Global Torsional Vibrations, Astrophys. J. 653, 593 (2006), arXiv:astro-ph/0608463 [astro-ph].
  9. [11]T. E. Strohmayer and A. L. Watts, Discovery of Fast X-Ray Oscillations during the 1998 Giant Flare from SGR 1900+14, Astrophys. J. Lett. 632, L111 (2005), arXiv:astro-ph/0508206 [astro-ph].DOI
  10. [12]C. Barat, R. I. Hayles, K. Hurley, M. Niel, G. Vedrenne, U. Desai, W. G. Kurt, V. M. Zenchenko, and I. V. Estulin, Fine time structure in the 1979 March 5 gamma ray burst, Astron. Astrophys. 126, 400 (1983).
  11. [13]A. J. Castro-Tirado, N. Østgaard, E. GӧÇüş, C. Sánchez-Gil, J. Pascual-Granado, V. Reglero, A. Menzemtsev, M. Gabler, M. Marisaldі, T. Neubert, C. Budtz-Jørgensen, A. Lindanger, D. Sarria, I. Kuvvetli, P. Cerdá-Durán, J. Navarro-González, J. A. Font, B. Zhang, N. Lund, C. A. Oxborrow, S. Brandt, M. D. Caballero-García, I. M. Carrasco-García, A. Castellón, M. A. Castro Tirado, F. Christiansen, C. J. Eyles, E. Fernández-García, G. Genov, S. Guzy, Y. D. Hu, A. Nicuesa Guelbenzu, S. B. Pandey, Z. K. Peng, C. Pérez del Pulgar, A. J. Reina Terol, E. Rodríguez, R. Sánchez-Ramírez, T. Sun, K. Ullaland, and S. Yang, Very-high-frequency oscillations in the main peak of a magnetar giant flare, Nature (London) 600, 621 (2021).DOI
  12. [14]C. J. Hansen and D. F. Cioffi, Torsional oscillations in neutron star crusts, Astrophys. J. 238, 740 (1980).
  13. [15]B. L. Schumaker and K. S. Thorne, Torsional oscillations of neutron stars, Mon. Not. R. Astron. Soc. 203, 457 (1983), https://academic.oup.com/mnras/article-pdf/203/2/457/3878163/mnras203-0457.pdf.
  14. [16]P. N. McDermott, H. M. van Horn, and C. J. Hansen, Nonradial
  15. [17]H. Sotani, K. D. Kokkotas, and N. Stergioulas, Torsional oscillations of relativistic stars with dipole magnetic fields, Mon. Not. R. Astron. Soc. 375, 261 (2007), arXiv:astro-ph/0608626 [astro-ph].DOI
  16. [18]G. H. de Souza and C. Chirenti, Torsional oscillations of magnetized neutron stars with mixed poloidal-toroidal fields, Phys. Rev. D 100, 043017 (2019), arXiv:1810.06628 [astro-ph.HE].
  17. [19]H. Sotani, A. Colaiuda, and K. D. Kokkotas, Constraints on the magnetic field geometry of magnetars, Mon. Not. R. Astron. Soc. 385, 2161 (2008).DOI
  18. [20]Y. Levin, QPOs during magnetar flares are not driven by mechanical normal modes of the crust, Mon. Not. R. Astron. Soc. 368, L35 (2006), arXiv:astro-ph/0601020 [astro-ph].DOI
  19. [21]M. Gabler, P. Cerdá Durán, J. A. Font, E. Müller, and N. Stergioulas, Magneto-elastic oscillations and the damping of crustal shear modes in magnetars, Mon. Not. R. Astron. Soc. 410, L37 (2011), arXiv:1007.0856 [astro-ph.HE].DOI
  20. [22]Y. Levin, On the theory of magnetar QPOs, Mon. Not. R. Astron. Soc. 377, 159 (2007), arXiv:hep-ph/0612725 [astro-ph].DOI
  21. [23]K. Glampedakis, L. Samuelsson, and N. Andersson, Elastic or magnetic? A toy model for global magnetar oscillations with implications for quasi-periodic oscillations during flares, Mon. Not. R. Astron. Soc. 371, L74 (2006), arXiv:astro-ph/0605461 [astro-ph].DOI
  22. [24]P. Cerdá-Durán, N. Stergioulas, and J. A. Font, Alfvén QPOs in magnetars in the anelastic approximation, Mon. Not. R. Astron. Soc. 397, 1607 (2009), arXiv:0902.1472 [astro-ph.HE].
  23. [25]A. Colaiuda, H. Beyer, and K. D. Kokkotas, On the quasi-periodic oscillations in magnetars, Mon. Not. R. Astron. Soc. 396, 1441 (2009), arXiv:0902.1401 [astro-ph.HE].DOI
  24. [26]A. Colaiuda and K. D. Kokkotas, Magnetar oscillations in the presence of a crust, Mon. Not. R. Astron. Soc. 414, 3014 (2011), arXiv:1012.3103 [gr-qc].DOI
  25. [27]A. Colaiuda and K. D. Kokkotas, Coupled polar-axial magnetar oscillations, Mon. Not. R. Astron. Soc. 423, 811 (2012), arXiv:1112.3561 [astro-ph.HE].DOI
  26. [28]M. van Hoven and Y. Levin, Magnetar oscillations - I. Strongly coupled dynamics of the crust and the core, Mon. Not. R. Astron. Soc. 410, 1036 (2011), arXiv:1006.0348 [astro-ph.HE].DOI
  27. [29]M. van Hoven and Y. Levin, Magnetar oscillations - II. Spectral method, Mon. Not. R. Astron. Soc. 420, 3055 (2012), arXiv:1110.2107 [astro-ph.HE].DOI
  28. [30]H. Sotani, K. D. Kokkotas, and N. Stergioulas, Alfvén quasi-periodic oscillations in magnetars, Mon. Not. R. Astron. Soc. 385, L5 (2008), arXiv:0710.1113 [astro-ph].
  29. [31]K. D. Kokkotas, E. Gaertig, and A. Colaiuda, Neutron star dynamics and gravitational waves, in Journal of Physics Conference Series, Vol. 222 (2010) p. 012031.
  30. [32]M. Gabler, P. Cerdá-Durán, N. Stergioulas, J. A. Font, and E. Müller, Magnetoelastic oscillations of neutron stars with dipolar magnetic fields, Mon. Not. R. Astron. Soc. 421, 2054 (2012), arXiv:1109.6233 [astro-ph.HE].DOI
  31. [33]M. Gabler, P. Cerdá-Durán, N. Stergioulas, J. A. Font, and E. Müller, Imprints of Superfluidity on Magnetoelastic Quasiperiodic Oscillations of Soft Gamma-Ray Repeaters, Phys. Rev. Lett. 111, 211102 (2013), arXiv:1304.3566 [astro-ph.HE].
  32. [34]A. Passamonti and S. K. Lander, Quasi-periodic oscillations in superfluid magnetars, Mon. Not. R. Astron. Soc. 438, 156 (2014), arXiv:1307.3210 [astro-ph.SR].DOI
  33. [35]M. Gabler, P. Cerdá-Durán, N. Stergioulas, J. A. Font, and E. Müller, Constraining properties of high-density matter in neutron stars with magneto-elastic oscillations, Mon. Not. R. Astron. Soc. tron. Soc. 476, 4199 (2018), arXiv:1710.02334 [astro-ph.HE].
  34. [36]H. Sotani and K. D. Kokkotas, Alfvén polar oscillations of relativistic stars, Mon. Not. R. Astron. Soc. 395, 1163 (2009), arXiv:0902.1490 [astro-ph.HE].DOI
  35. [37]I. v. Rabi, Das freie elektron im homogenen magnetfeld nach der diracschen theorie, Zeitschrift für Physik 49, 507 (1928).DOI
  36. [38]L. Landau, Diamagnetismus der metalle, Zeitschrift für Physik 64, 629 (1930).
  37. [39]N. Chamel, R. L. Pavlov, L. M. Mihailov, C. J. Velchev, Z. K. Stoyanov, Y. D. Mutafchieva, M. D. Ivanovich, J. M. Pearson, and S. Goriely, Properties of the outer crust of strongly magnetized neutron stars from hartree-fock-bogoliubov atomic mass models, Phys. Rev. C 86, 055804 (2012).DOI
  38. [40]N. Chamel, Z. K. Stoyanov, L. M. Mihailov, Y. D. Mutafchieva, R. L. Pavlov, and C. J. Velchev, Role of landau quantization on the neutron-drip transition in magnetar crusts, Phys. Rev. C 91, 065801 (2015).DOI
  39. [41]Y. D. Mutafchieva, N. Chamel, Z. K. Stoyanov, J. M. Pearson, and L. M. Mihailov, Role of Landau-Rabi quantization of electron motion on the crust of magnetars within the nuclear energy density functional theory, Phys. Rev. C 99, 055805 (2019), arXiv:1904.05045 [astro-ph.HE].DOI
  40. [42]M. Wang, G. Audi, F. G. Kondev, W. J. Huang, S. Naimi, and X. Xu, The AME2016 atomic mass evaluation (II). Tables, graphs and references, Chinese Physics C 41, 030003 (2017).
  41. [43]S. Goriely, N. Chamel, and J. M. Pearson, Further explorations of Skyrme-Hartree-Fock-Bogoliubov mass formulas. XIII. The 2012 atomic mass evaluation and the symmetry coefficient, Phys. Rev. C 88, 024308 (2013).DOI
  42. [44]J. M. Pearson, N. Chamel, A. Y. Potekhin, A. F. Fantina, C. Ducoin, A. K. Dutta, and S. Goriely, Unified equations of state for cold non-accreting neutron stars with Brussels-Montreal functionals - I. Role of symmetry energy, Mon. Not. R. Astron. Soc. 481, 2994 (2018), arXiv:1903.04981 [astro-ph.HE].arxiv.org/abs/1903.04981
  43. [45]M. Ferreira and C. Providência, Neutron star properties: Quantifying the effect of the crust–core matching procedure, Universe 6, 220 (2020).DOI
  44. [46]L. Suleiman, M. Fortin, J. L. Zdunik, and P. Haensel, Influence of the crust on the neutron star macrophysical quantities and universal relations, Phys. Rev. C 104, 015801 (2021).DOI
  45. [47]H. Sotani, K. Iida, and K. Oyamatsu, Constraints on the nuclear equation of state and the neutron star structure from crustal torsional oscillations, Mon. Not. R. Astron. Soc. 479, 4735 (2018), arXiv:1807.00528 [nucl-th].DOI
  46. [48]H. Sotani, H. Togashi, and M. Takano, Effects of finite sizes of atomic nuclei on shear modulus and torsional oscillations in neutron stars, Mon. Not. R. Astron. Soc. 516, 5440 (2022), arXiv:2209.05416 [nucl-th].DOI
  47. [49]H. Sotani, K. D. Kokkotas, and N. Stergioulas, Neutron star mass-radius constraints using the high-frequency quasi-periodic oscillations of GRB 200415A, Astron. Astrophys. 676, A65 (2023), arXiv:2303.03150 [astro-ph.HE].DOI
  48. [50]E. Flowers and M. A. Ruderman, Evolution of pulsar magnetic fields., Astrophys. J. 215, 302 (1977).
  49. [51]J. Frieben and L. Rezzolla, Equilibrium models of relativistic stars with a toroidal magnetic field, Mon. Not. R. Astron. Soc. 427, 3406–3426 (2012).DOI
  50. [52]D. Chatterjee, J. Novak, and M. Oertel, Structure of ultra-magnetised neutron stars, European Physical Journal A 57, 249 (2021), arXiv:2108.13733 [nucl-th].DOI
  51. [53]K. Ioka and M. Sasaki, Relativistic Stars with Poloidal and Toroidal Magnetic Fields and Meridional Flow, Astrophys. J. 600, 296 (2004), arXiv:astro-ph/0305352 [astro-ph].
  52. [54]A. Colaiuda, V. Ferrari, L. Gualtieri, and J. A. Pons, Relativistic models of magnetars: structure and deformations, Mon. Not. R. Astron. Soc. 385, 2080 (2008).DOI
  53. [55]N. Chamel and A. F. Fantina, Onset of electron captures and shallow heating in magnetars, Universe 8, 328 (2022).DOI
  54. [56]Y. Xu, S. Goriely, A. Jorissen, G. L. Chen, and M. Arnould, Databases and tools for nuclear astrophysics applications. BRUSsels Nuclear Library (BRUSLIB), Nuclear Astrophysics Compilation of REactions II (NACRE II) and Nuclear NETwork GENerator (NETGEN), Astron. Astrophys. 549, A106 (2013), arXiv:1212.06268 [nucl-th].DOI
  55. [57]D. P. Arteaga, M. Grasso, E. Khan, and P. Ring, Nuclear structure in strong magnetic fields: Nuclei in the crust of a magnetar, Phys. Rev. C 84, 045806 (2011), arXiv:1107.5243 [nucl-th].DOI
  56. [58]P. Haensel, A. Y. Potekhin, and D. G. Yakovlev, Neutron stars 1: Equation of state and structure, Vol. 326 (Springer, New York, USA, 2007).
  57. [59]E. Salpeter, Electron screening and thermonuclear reactions, Australian Journal of Physics 7, 373 (1954).
  58. [60]N. Chamel and A. F. Fantina, Electron capture instability in magnetic and nonmagnetic white dwarfs, Phys. Rev. D 92, 023008 (2015).DOI
  59. [61]N. Chamel and Z. K. Stoyanov, Analytical determination of the structure of the outer crust of a cold nonaccreted neutron star: Extension to strongly quantizing magnetic fields, Phys. Rev. C 101, 065802 (2020).DOI
  60. [62]N. Chamel, A. F. Fantina, J. L. Zdunik, and P. Haensel, Neutron drip transition in accreting and nonaccreting neutron star crusts, Phys. Rev. C 91, 055803 (2015).DOI
  61. [63]S. Goriely, N. Chamel, and J. M. Pearson, Further explorations of skyrme-hartree-fock-bogoliubov mass formulas. xiii. the 2012 atomic mass evaluation and the symmetry coefficient, Phys. Rev. C 88, 024308 (2013).DOI
  62. [64]J. M. Pearson, N. Chamel, A. Y. Potekhin, A. F. Fantina, C. Ducoin, A. K. Dutta, and S. Goriely, Erratum: Unified equations of state for cold non-accreting neutron stars with Brussels-Montreal functionals. I. Role of symmetry energy, Mon. Not. R. Astron. Soc. 486, 768 (2019).DOI
  63. [65]L. Perot, N. Chamel, and A. Sourie, Role of the symmetry energy and the neutron-matter stiffness on the tidal deformability of a neutron star with unified equations of state, Phys. Rev. C 100, 035801 (2019).DOI
  64. [66]S. Ogata and S. Ichimaru, First-principles calculations of shear moduli for monte carlo–simulated coulomb solids, Phys. Rev. A 42, 4867 (1990).
  65. [67]N. Messios, D. B. Papadopoulos, and N. Stergioulas, Torsional oscillations of magnetized relativistic stars, Mon. Not. R. Astron. Soc. 328, 1161 (2001).
  66. [68]H. Sotani, K. D. Kokkotas, N. Stergioulas, and M. Vavoulidis, Torsional Oscillations of Relativistic Stars with Dipole Magnetic Fields II. Global Alfvén Modes, arXiv e-prints , astro-ph/0611666 (2006), arXiv:astro-ph/0611666 [astro-ph].
  67. [69]B. Franzon, R. Negreiros, and S. Schramm, Magnetic field effects on the crust structure of neutron stars, Phys. Rev. D 96, 123005 (2017).
  68. [70]L. Samuelsson and N. Andersson, Neutron star asteroseismology. Axial crust oscillations in the Cowling approximation, Mon. Not. R. Astron. Soc. 374, 256 (2007).

Paper details

vlnc:W_sz3jgp9t
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents