betaFeedback

WorksPhysics and Astronomy

Superconducting multi-vortices and a novel BPS bound in chiral perturbation theory

Abstract

We derive a novel BPS bound from chiral perturbation theory minimally coupled to electrodynamics at finite isospin chemical potential. At a critical value of the isospin chemical potential, a system of three first-order differential field equations (which implies the second-order field equations) for the gauge field and the hadronic profile can be derived from the requirement to saturate the bound. These BPS configurations represent magnetic multi-vortices with quantized flux supported by a superconducting current. The corresponding topological charge density is related to the magnetic flux density, but is screened by the hadronic profile. Such a screening effect allows to derive the maximal value of the magnetic field generated by these BPS magnetic vortices, being Bmax=2,04×1014GB_{max}=2,04 \times 10^{14}G. The solution for a single BPS vortex is discussed in detail, and some physical consequences, together with the comparison with the magnetic vortices in the Ginzburg-Landau theory at critical coupling, are described.

Introduction

The phase diagram of the low energy limit of quantum chromodynamics (QCD) under non-trivial external conditions (such as finite density, low temperatures and in presence of strong magnetic fields) is a very hard nut to crack (see [1]-[4], and references therein). Not only analytic perturbative methods fail, but also lattice QCD may experience some problems (see, for instance, Refs. [5]-[9]).

An extremely important region of the phase diagram corresponds to low temperatures, vanishing baryonic chemical potential (μB=0\mu_B = 0) and finite isospin chemical potential (μI≠0\mu_I \ne0) [10]-[21]. There are many reasons behind the relevance of this sector. First, when μI≠0\mu_I \ne0 but μB=0\mu_B = 0, there is no sign problem, so that detailed results on the phase diagram are available in this region. Second, in the seminal paper by Son and Stephanov [10], it has been shown that a transition occurs from the hadronic phase to a superfluid phase and, eventually, to a superconducting phase. It is expected that at low isospin density the ground state is a pion condensate [10], [22], [23], while a Fermi liquid with Cooper pairing should appear at higher density. These results have been confirmed by lattice studies and still offer many surprises, as we will see.

In particular, the analysis of the interplay between strong interactions and electromagnetic fields exhibit difficult unsolved problems at low energies and temperatures. At these scales, the main role is played by the topological solitons of QCD [24, 25, 26, 27]. Topological solitons are characterized by a non-vanishing topological charge, which prevents such classical configurations from decaying into the trivial vacuum. Especially taking into account the Cooper pairing appearing at finite μI\mu_I, it is clear that the most natural topological solitons that should appear are superconducting vortices with quantized magnetic flux. Indeed, vortex-like solitons are expected both in superfluids and in superconductors. At finite density, many vortex-like solutions have been constructed numerically, and many of the following references have been important sources of inspiration as far as the present work is concerned; [28, 29, 30, 31, 32, 33, 34, 35, 36, 37].

Analytical results have also been obtained in Refs. [38, 39, 40, 41, 42, 43]. However, these analytical solutions (which describe inhomogeneous baryonic condensates with the shapes of tubes and layers) have been obtained at finite baryon density but not at finite isospin density1 and, moreover, do not possess the required quantized magnetic flux characteristic of vortices in superconductors. One of the most important open theoretical questions in this area appears when we consider the situation in the Ginzburg–Landau (GL) theory. In such a theory, depending on the value of the Higgs coupling, one can be in a type-I or in a type-II superconductor and, exactly at the transition, a remarkable phenomenon appears. One can saturate a Bogomol’nyi–Prasad–Sommerfield bound (BPS bound in what follows) for the free energy per unit of length of the vortices, where the magnetic flux plays the role of the topological charge, and multi-vortices solutions (where the vortices can be placed at arbitrary locations) appear [44].

Since μI\mu_I is responsible for the Cooper pairing [10] (and so μI\mu_I plays a similar role to the Higgs coupling), one would expect that at a special value of μI\mu_I it should be possible to saturate a suitable BPS bound (where the magnetic flux is expected to be the corresponding topological charge) in such a way that multi-vortices solutions should appear. Such a special value of μI\mu_I should be a signal of a change of behavior from type-I to type-II (in which case flux tubes can penetrate the superconductor). Despite all the efforts, such a BPS completion has not been found. In this paper, we will fill this gap.

We will work with chiral perturbation theory (χ\chiPT), which describes the low energy limit of QCD; see [45, 46, 47, 48, 49, 50, 51, 52], and references therein. As it will be clarified in the following sections, the present novel results are well within the range of validity of χ\chiPT. Hence, our starting point is the non-linear sigma model (NLSM) coupled to the Maxwell theory in the case of SU(2)SU(2) isospin global symmetry, which is one of the most relevant effective field theories [24], [53]. The analysis of soliton crystals in χ\chiPT in the presence of magnetic fields and/or at finite isospin chemical potential is a very active topic (see [30, 31, 32, 33, 34, 35, 36, 37], and references therein). The conditions allowing the appearance of topologically stable non-homogeneous condensates, besides its intrinsic interest, are also relevant for the thermodynamics of dense nuclear/quark matter [1, 2, 3]. Moreover, when the SU(2)SU(2) valued scalar field is assumed to be homogeneous, and the corresponding spatial fluctuations are neglected, many important physical effects are missed (see [54], [55], and references therein).

Until very recently, the more common techniques (which allow describing stable topological solitons such as instantons, monopoles and vortices in Yang-Mills-Higgs and the abelian-Higgs models) did not work in the case of the gauged χ\chiPT. The reason is that, at the level of the gauged NLSM, there is no obvious BPS bound to saturate for static configurations. Indeed, the obvious bounds (where one would put, on the right-hand side, the baryonic charge density or the magnetic flux density) cannot be saturated. On the other hand, the recent results in Ref. [56] have shown that when the obvious BPS bound cannot be saturated, a novel less obvious BPS bound exist that can actually be saturated, and it has all the expected properties that a BPS bound should have. In the present paper, we will see a quite remarkable manifestation of this phenomenon; we will find a novel BPS bounds for a special value of μI\mu_I in the gauged χ\chiPT which can actually be saturated by superconducting multi-vortices with quantized magnetic flux.

The reason why this bound has not been found before is very tricky. The topological charge density, which is needed for the BPS completion, is not the magnetic flux density (as it happens in the GL case), but rather a magnetic flux density dressed by the hadronic profile itself. Such a dressing screens the magnetic field BzB_z providing with a maximal possible value for BzB_z. It is worth emphasizing that magnetized topological solitons are extremely relevant in many situations, such as heavy ions collisions and neutron stars (see Refs. [5, 6, 7, 8, 9], and [57, 58, 59]). A further by-product of the present results is that the Landau-Peierls criteria is avoided.

The techniques introduced in Refs. [38, 39, 40, 41, 42, 43], allowed the construction of analytical solutions describing multi-solitons at finite density for both in the NLSM as well as in the Skyrme model minimally coupled to the Maxwell theory (see, for details, Refs. [60, 61, 62, 63, 64, 65, 66]). However, in such an approach, the electromagnetic field (although self-consistently generated by the non-homogeneous condensates) has necessarily the electric components of the same order as the magnetic components. Therefore, in all the situations in which the magnetic field dominates, this framework must be modified. This is the aim of the present paper, in which the technique introduced in Ref. [56] will be used to derive a novel BPS bound that can be saturated for magnetized configurations.

This paper is organized as follows: In § 2 the gauged χ\chiPT will be introduced together with the Ansatz for the matter fields. In § 3 we derive a novel BPS bound from the gauged χ\chiPT. In § 4 we construct magnetized vortex solutions from the BPS equations and we discuss its main physical properties. In § 5 we show in detail the case of a single vortex. § 6 is devoted to the conclusions. In the Appendix we will provide with mathematical arguments supporting the existence of such BPS multi-vortices configurations.

Preliminaries

The gauged χPT\chi\mathrm{PT} (up to order O(p2)\mathcal{O}(p^{2})) [51], is described by the action

I[U,A]=14e2∫Md4x−g(KTr⁡[LμLμ]−FμνFμν)(1)I[U,A]=\frac{1}{4e^{2}}\int_{\mathcal{M}}d^{4}x\sqrt{-g}\left(K\operatorname{Tr}[L^{\mu}L_{\mu}]-F_{\mu\nu}F^{\mu\nu}\right) \tag*{(1)}
Lμ=U−1DμU=Lμjtj,Fμν=∂μAν−∂νAμ,tj=iσj.L_{\mu}=U^{-1}D_{\mu}U=L_{\mu}^{j}t_{j},\qquad F_{\mu\nu}=\partial_{\mu}A_{\nu}-\partial_{\nu}A_{\mu},\qquad t_{j}=i\sigma_{j}.

Here U(x)∈SU(2)U(x)\in SU(2) is the pionic field, AμA_{\mu} is the Maxwell potential and σi\sigma_{i} are the Pauli matrices. The coupling KK is a positive constant given by K=(efπ2)2K=\left(\frac{ef_{\pi}}{2}\right)^{2}, being fπf_{\pi} the pions decay constant and ee the electric charge. Here we will use fπ=93 MeVf_{\pi}=93\,\mathrm{MeV} and e=4πα=0.303e=\sqrt{4\pi\alpha}=0.303. In our convention c=1c=1 and ℏ=1\hbar=1. In Eq. (1), DμD_{\mu} denotes the covariant derivative, defined as

DμU=∇μU+AμUO^,O^=U−1[t3,U],(2)D_{\mu}U=\nabla_{\mu}U+A_{\mu}U\hat{O},\qquad\hat{O}=U^{-1}[t_{3},U], \tag*{(2)}

being ∇μ\nabla_{\mu} the partial derivative.2 The isospin chemical potential can be introduced to the model through the covariant derivative in the following form

DμU→DˉμU=DμU+μI[t3,U]gμt,(3)D_{\mu}U\rightarrow\bar{D}_{\mu}U=D_{\mu}U+\mu_{I}[t_{3},U]g_{\mu t}, \tag*{(3)}

where μI\mu_{I} is the value of isospin chemical potential. It is important to point out that critical values for the isospin chemical potential have been derived from effective models (such as the Skyrme model or ladder-QCD) showing different phase transitions (see, for instance, Refs. [67]-[70]). In the next section, we will compute a critical value for μI\mu_{I} in χPT\chi\mathrm{PT} that allows the formation of a topologically stable inhomogeneous condensate made of magnetized multi-vortices.3

The electromagnetic current obtained from Eq. (1) is given by

Jμ=−K2Tr⁡(O^Lμ).(4)J_{\mu}=-\frac{K}{2}\operatorname{Tr}(\hat{O}L_{\mu}). \tag*{(4)}

The pionic field in the exponential representation is written as

U=cos⁡(α)1+sin⁡(α)niti,(5)U=\cos(\alpha)\mathbf{1}+\sin(\alpha)n_{i}t^{i}, \tag*{(5)}
ni={sin⁡Θcos⁡Φ,sin⁡Θsin⁡Φ,cos⁡Θ},n_{i}=\{\sin\Theta\cos\Phi,\sin\Theta\sin\Phi,\cos\Theta\},

where α=α(xμ)\alpha=\alpha(x^{\mu}), Θ=Θ(xμ)\Theta=\Theta(x^{\mu}), Φ=Φ(xμ)\Phi=\Phi(x^{\mu}) are the three degrees of freedom of the UU field, and 1\mathbf{1} denotes the 2×22\times2 identity matrix. In terms of this parametrization, the covariant derivative DμUD_{\mu}U in Eq. (2) reads

Dμα=∂μα,DμΘ=∂μΘ,DμΦ=∂μΦ−2Aμ.(6)D_{\mu}\alpha=\partial_{\mu}\alpha,\qquad D_{\mu}\Theta=\partial_{\mu}\Theta,\qquad D_{\mu}\Phi=\partial_{\mu}\Phi-2A_{\mu}. \tag*{(6)}

Hence, one can see that the scalar degree of freedom Φ(xμ)\Phi(x^{\mu}) plays the role of the phase of the complex Higgs field in the GL theory (while α\alpha and Θ\Theta are not affected by the covariant derivative).

Since we are interested in multi-vortices (with quantized magnetic field along the third spatial direction) in χ\chiPT minimally coupled to the Maxwell theory, the natural Ansatz is

α=α(x1,x2),Φ=Φ(x1,x2),Θ=π2,(7)\alpha=\alpha(x_1,x_2),\qquad\Phi=\Phi(x_1,x_2),\qquad\Theta=\frac{\pi}{2}, \tag*{(7)}
Aμdxμ=A1 dx1+A2 dx2,Ai=Ai(x1,x2),(8)A_\mu dx^\mu=A_1\,dx_1+A_2\,dx_2,\qquad A_i=A_i(x_1,x_2), \tag*{(8)}

which leads to a magnetic field in the x3x_3 direction. This choice is very convenient since, as it will be clear from the following computations, the combination sin⁡(α)niti\sin(\alpha)n_i t^i, plays the role of the complex Higgs field ϕ\phi in the GL free energy:

sin⁡(α)niti∼ϕ,\sin(\alpha)n_i t^i\sim\phi,

while the choice of the gauge potential is forced by the requirement to have a magnetic field along x3x_3.

A novel BPS bound from χ\chiPT

In this section, for simplicity, we will assume a flat metric in Cartesian coordinates, but our results can be obtained in the same way using arbitrary coordinate systems, as we will see in the next section.

When one replaces in Eq. (1) the Ansatz defined in Eqs. (5), (7) and (8), (taking into account the non-trivial isospin chemical potential in Eq. (3)) one obtains the lagrangian density L\mathcal{L}, which will play the role of free energy density FF, that is

F=−L=1e2[K2{(∇⃗α)2+sin⁡2(α)(D⃗Φ)2}+12(B⃗ 2)−2Kμ12sin⁡2(α)],(9)F=-\mathcal{L}=\frac{1}{e^2}\left[\frac{K}{2}\left\{\left(\vec{\nabla}\alpha\right)^2+\sin^2(\alpha)\left(\vec{D}\Phi\right)^2\right\}+\frac{1}{2}\left(\vec{B}^{\,2}\right)-2K\mu_1^2\sin^2(\alpha)\right], \tag*{(9)}

where we have denoted

(∇⃗α)2=∑i=12(∂iα)2,(D⃗Φ)2=∑i=12(DiΦ)2.\left(\vec{\nabla}\alpha\right)^2=\sum_{i=1}^{2}(\partial_i\alpha)^2,\qquad\left(\vec{D}\Phi\right)^2=\sum_{i=1}^{2}(D_i\Phi)^2.

In order to derive a BPS bound from Eq. (9), it is convenient to rewrite the free energy density as follows:

e2F=K2{(∇⃗α)2+sin⁡2(α)(D⃗Φ)2}+12(B⃗ 2)+2Kμ12cos⁡2(α)−2Kμ12.(10)e^2F=\frac{K}{2}\left\{\left(\vec{\nabla}\alpha\right)^2+\sin^2(\alpha)\left(\vec{D}\Phi\right)^2\right\}+\frac{1}{2}\left(\vec{B}^{\,2}\right)+2K\mu_1^2\cos^2(\alpha)-2K\mu_1^2. \tag*{(10)}

One can see that the above expression would be positive definite if one could eliminate the last constant term. From the viewpoint of the differential equations, the constant term is certainly not important (and one could simply drop it). However, such a term is telling us something physical: in the present case, the appropriate thermodynamic potential is not the free energy FF, but rather the Gibbs free energy GG

G=F+PV,F=∫d3x F,G=∫d3x G,(11)G=F+PV,\qquad F=\int d^3x\,F,\qquad G=\int d^3x\,G, \tag*{(11)}

being VV the volume and PP the pressure fixed by the chemical potential, namely P=2KμI2P = 2K\mu_I^2, according to Eq. (10). Thus, the Gibbs free energy density GG reads

e2G=K2{(∇⃗α)2+sin⁡2(α)(D⃗Φ)2}+12(B⃗2)+2KμI2cos⁡2(α)(12)e^2G = \frac{K}{2}\left\{\left(\vec{\nabla}\alpha\right)^2 + \sin^2(\alpha)\left(\vec{D}\Phi\right)^2\right\} + \frac{1}{2}\left(\vec{B}^2\right) + 2K\mu_I^2\cos^2(\alpha) \tag*{(12)}

so that

GL=∫dx1dx2G.(13)\frac{G}{L} = \int dx_1dx_2G . \tag*{(13)}

Eq. (13) has to do with the fact that we are interested in superconducting vortices whose Gibbs free energy does not depend on the third direction. From now on, we will work with GG since it is a positive definite quantity, and it simplifies the discussion of the boundary conditions.

Boundary conditions and flux quantization

The requirement to have a finite Gibbs free energy per unit of length gives rise to boundary conditions that are close analogs of the boundary conditions used for vortices in the GL theory.

Let us introduce the usual coordinate system to discuss boundary conditions for a single vortex at the origin of the cylindrical coordinates:

ds2=−dt2+dr2+r2dθ2+(dx3)2.ds^2 = -dt^2 + dr^2 + r^2d\theta^2 + (dx_3)^2 .

In order for the integral in Eq. (13) to be convergent for large values of the two-dimensional radius, we must require

α→r→∞π2+c1π,c1∈N,D⃗Φ→r→∞0⟺A⃗→r→∞12∂⃗Φ.(14)\alpha\xrightarrow{r\to\infty} \frac{\pi}{2} + c_1\pi,\quad c_1\in\mathbb{N},\qquad\vec{D}\Phi\xrightarrow{r\to\infty} 0 \Longleftrightarrow\vec{A} \xrightarrow{r\to\infty} \frac{1}{2}\vec{\partial}\Phi. \tag*{(14)}

On the other hand, close to the origin, regularity requires

α→r→0c2π,c2∈N,A⃗→r→00.(15)\alpha\xrightarrow{r\to0} c_2\pi,\quad c_2\in\mathbb{N},\qquad\vec{A} \xrightarrow{r\to0} 0 . \tag*{(15)}

The condition in Eq. (15), from the viewpoint of the chiral field UU in Eqs. (5) and (7), reads:

U→r→0±1.(16)U \xrightarrow{r\to0} \pm1 . \tag*{(16)}

In fact, Eqs. (15) and (16) must be satisfied at the locations of all the vortices as one can always choose a local coordinates system where the position of the jj−th vortex is the origin:

U→r→rj±1,∀j=1,…,N,(17)U \xrightarrow{r\to r_j} \pm1,\quad\forall j = 1,\ldots,N , \tag*{(17)}

where NN is the total number of vortices and rjr_j is the position of the jj−th vortex.

On the other hand, one can derive the quantization of the magnetic flux in the usual way. The chiral field UU in Eq. (5) must be single valued:

U(r,θ)=U(r,θ+2π),(18)U(r,\theta) = U(r,\theta+ 2\pi) , \tag*{(18)}

and then, taking into account Eq. (7), one arrives at the conclusion that

Φ(r,θ)=Φ(r,θ+2π)+2nπ,n∈N,(19)\Phi(r,\theta) = \Phi(r,\theta+ 2\pi) + 2n\pi,\quad n \in\mathbb{N}, \tag*{(19)}

which is natural taking into account that Φ\Phi plays the role of the phase of the complex Higgs field in the GL theory (see Eq. (6)). This implies that the following integral is an integer

12π∮Γ∞∂⃗Φ⋅dl⃗=n,(20)\frac{1}{2\pi}\oint_{\Gamma_\infty} \vec{\partial}\Phi\cdot d\vec{l} = n, \tag*{(20)}

where Γ∞\Gamma_\infty is the circle at spatial infinity. It is worth emphasizing here that the above integral would be an integer for any closed curve Γ\Gamma enclosing all the points {rj}j=1,…,N\{r_j\}_{j=1,\ldots,N}, where sin⁡(α)\sin(\alpha) vanishes. On the other hand, the finiteness of the Gibbs free energy requires Eq. (14). Consequently, due to the Stokes theorem, as in the case of the GL theory, one obtains that nn is exactly the magnetic flux, and it is quantized.

Improper gauge transformations

In order to derive a BPS bound, we rewrite Eq. (12) as

e2G=K2((∂1α)2+(∂2α)2+sin⁡2(α)[(D1Φ)2+(D2Φ)2])+12Bz2+2KμI2cos⁡2(α)=K2((∂1α+sin⁡(α)D2Φ)2+(∂2α−sin⁡(α)D1Φ)2)+12(Bz−2μIKcos⁡(α))2−Ksin⁡(α)(∂1αD2Φ−∂2αD1Φ)+2μIKcos⁡(α)Bz.\begin{aligned} e^2G={}&\frac{K}{2}\left((\partial_1\alpha)^2+(\partial_2\alpha)^2+\sin^2(\alpha)\left[(D_1\Phi)^2+(D_2\Phi)^2\right]\right)+\frac{1}{2}B_z^2+2K\mu_I^2\cos^2(\alpha)\\ ={}&\frac{K}{2}\left((\partial_1\alpha+\sin(\alpha)D_2\Phi)^2+(\partial_2\alpha-\sin(\alpha)D_1\Phi)^2\right)+\frac{1}{2}\left(B_z-2\mu_I\sqrt{K}\cos(\alpha)\right)^2\\ &{}-K\sin(\alpha)(\partial_1\alpha D_2\Phi-\partial_2\alpha D_1\Phi)+2\mu_I\sqrt{K}\cos(\alpha)B_z. \end{aligned}

Note that, locally, we can gauge away the phase Φ\Phi as follows

Aμ→A^μ=Aμ+12∂μΦ,(21)A_\mu\to\widehat{A}_\mu=A_\mu+\frac{1}{2}\partial_\mu\Phi, \tag*{(21)}
Φ→Φ^=Φ−Φ=0.(22)\Phi\to\widehat{\Phi}=\Phi-\Phi=0. \tag*{(22)}

However, it is very important to emphasize that Φ\Phi itself is a multi-valued function precisely due to the non-vanishing magnetic flux (see Eqs. (3.11) and (3.12)). Therefore, Eqs. (3.13) and (3.14) do not define a proper gauge transformation. In order to clarify which is the price to be paid (in particular, which extra terms do appear) when one performs a gauge transformation with a multi-valued gauge parameter, let us consider the case of a single vortex (which will be analyzed in more detail in § 5). The corresponding Ansatz is

α=α(r),Ar=Az=0,Aθ=A(r),Φ=nθ,(23)\alpha=\alpha(r),\quad A_r=A_z=0,\quad A_\theta=A(r),\quad\Phi=n\theta, \tag*{(23)}

where the cylindrical coordinates are

ds2=−dt2+dr2+r2dθ2+dz2.(24)ds^2=-dt^2+dr^2+r^2d\theta^2+dz^2. \tag*{(24)}

The spatial coordinates {r,θ,z}\{r,\theta,z\}, have the ranges

0<r<∞,0≤θ<2π,−∞<z<∞.0<r<\infty,\quad0\leq\theta<2\pi,\quad-\infty<z<\infty.

As one can see from Eq. (6), the interaction between the matter field and the gauge field appears through the combination ∂μΦ−2Aμ\partial_\mu\Phi-2A_\mu. It is clear that with the Ansatz in Eq. (23), the conditions of regularity at the origin in Eq. (15) reduce to

Aθ=A(r)→r→00,sin⁡α(r)→r→00,A_\theta=A(r)\xrightarrow{r\to0}0,\qquad\sin\alpha(r)\xrightarrow{r\to0}0,

which means, of course, that the origin is the position of the vortex. If one now performs the improper gauge transformation in Eqs. (21) and (22) one would get

A^θ=A(r)+12n→r→012n,sin⁡α(r)→r→00.\widehat{A}_\theta=A(r)+\frac{1}{2}n\xrightarrow{r\to0}\frac{1}{2}n,\qquad\sin\alpha(r)\xrightarrow{r\to0}0.

This is not an innocuous shift of the gauge potential since it introduces a singularity. This is most easily seen in Cartesian coordinates, where

Φ=nθ=narctan⁡(yx).\Phi=n\theta=n\arctan\left(\frac{y}{x}\right).

A gauge potential AμA_\mu proportional to ∂μΦ\partial_\mu\Phi, with the Φ\Phi here above, is nothing but the gauge potential of an Aharonov-Bohm gauge field. As it is well known, such a gauge potential, despite being locally a pure gauge, it corresponds to the magnetic field of a very thin solenoid along the zz axis. Therefore, the magnetic field, BzABB_z^{AB}, associated with the Aharonov-Bohm gauge field is

BzAB=nδ(x,y),(25)B_z^{AB}=n\delta(x,y), \tag*{(25)}

where δ(x,y)\delta(x,y) is the Dirac delta at the position of the thin solenoid. This argument is true actually at the positions of all the vortices. The reason is that, close to the position of any of the vortices present in the configuration, one can choose a coordinate system with the origin at the position of such vortex and repeat this argument. Consequently, the price to pay if one decides to “gauge away" the function Φ\Phi (which is convenient from the computational point of view) in a configuration with NN vortices is that one has to include explicitly in the magnetic field the “Aharonov-Bohm contributions" from all the vortices:

BzAB=∑j=1Nnjδ(x−xj,y−yj),(26)B_z^{AB}=\sum_{j=1}^{N}n_j\delta\left(x-x_j,y-y_j\right), \tag*{(26)}

(note that we have absorbed the factors of 4π4\pi in the njn_j in order to simplify the notation) which corresponds to a Φ\Phi of the following form

Φ=∑j=1Nnjarg⁡(z−zj),\Phi=\sum_{j=1}^{N}n_j\arg\left(z-z_j\right),

where z=x+iyz=x+iy, and zjz_j is the position of the jj–th vortex in complex coordinates (compare, for instance, with the expression below Eq. (10) of Ref. [71]). Of course, this same result has been obtained in a more rigorous way in the classic references [72] and [73]. In particular, the above expression corresponds to the typical sum over the Dirac deltas, which appear in the mathematical literature (see, for instance, the right hand side of Eq. (11) in Ref. [71]). Although the present argument is less rigorous than the one in the above mentioned references, it offers a very intuitive physical picture: if one eliminates Φ\Phi with an improper gauge transformation, then the “Aharonov-Bohm contributions" from all the vortices in Eq. (26) must be included.

BPS completion

The elimination of Φ\Phi simplifies the derivation of the BPS bound. One can rewrite the free energy density as

e2G=K2((∂1α+sin⁡(α)A^2)2+(∂2α−sin⁡(α)A^1)2)+12(B^z−2μIKcos⁡(α))2−Ksin⁡α(∂1αA^2−∂2αA^1)+2μIKcos⁡(α)B^z.(27)\begin{aligned} e^{2}G={}&\frac{K}{2}\left((\partial_{1}\alpha+\sin(\alpha)\widehat{A}_{2})^{2}+(\partial_{2}\alpha-\sin(\alpha)\widehat{A}_{1})^{2}\right)+\frac{1}{2}\left(\widehat{B}_{z}-2\mu_{I}\sqrt{K}\cos(\alpha)\right)^{2}\\ &-K\sin\alpha(\partial_{1}\alpha\widehat{A}_{2}-\partial_{2}\alpha\widehat{A}_{1})+2\mu_{I}\sqrt{K}\cos(\alpha)\widehat{B}_{z}. \tag*{(27)} \end{aligned}

In order to obtain the sought BPS completion, we need to write the last two terms in the above equation as a total derivative. This is only possible at the following special critical value for the isospin chemical potential

μIc=K=efπ2=14,1 MeV.(28)\mu_{I}^{c}=\sqrt{K}=\frac{ef_{\pi}}{2}=14,1\ \mathrm{MeV}. \tag*{(28)}

Quite remarkably, this particular value is well within the range of validity of χPT [16]. In fact, this value allows that

−K∫dx1dx2[sin⁡(α)(∂1αA^2−∂2αA^1)−cos⁡(α)B^z]=∫dω,ω=cos⁡(α)(A^1dx1+A^2dx2),(29)\begin{aligned} -K\int dx^{1}dx^{2}\left[\sin(\alpha)(\partial_{1}\alpha\widehat{A}_{2}-\partial_{2}\alpha\widehat{A}_{1})-\cos(\alpha)\widehat{B}_{z}\right]&=\int d\omega,\\ \omega&=\cos(\alpha)(\widehat{A}_{1}dx^{1}+\widehat{A}_{2}dx^{2}), \tag*{(29)} \end{aligned}

where we have used that B^z=F12=∂1A^2−∂2A^1\widehat{B}_{z}=F_{12}=\partial_{1}\widehat{A}_{2}-\partial_{2}\widehat{A}_{1}. Thus, the Gibbs free energy density is minimized when the following first order equations are satisfied

∂1α+sin⁡(α)A^2=0,(30)\partial_{1}\alpha+\sin(\alpha)\widehat{A}_{2}=0, \tag*{(30)}
∂2α−sin⁡(α)A^1=0,(31)\partial_{2}\alpha-\sin(\alpha)\widehat{A}_{1}=0, \tag*{(31)}
B^z−2Kcos⁡(α)=0,(32)\widehat{B}_{z}-2K\cos(\alpha)=0, \tag*{(32)}

and the following BPS bound emerges

G≥∫∂Σdω.(33)G\geq\int_{\partial\Sigma}d\omega. \tag*{(33)}

where ∂Σ\partial\Sigma is usually taken as the S1S^{1} circle at spatial infinity.

It is important to note that the value of the free energy depends on the topological invariant that is fixed by the boundary conditions on the sample. Thus, even when the pressure and chemical potential are fixed, the BPS bound allows the existence of configurations with different GG determined by the choice of the boundary conditions. In particular, for the BPS vortex phase and the homogeneous phase with the same (B,P,μI)(B,P,\mu_{I}) and different free energies: if the boundary conditions are such that the topological invariant vanishes, then the homogeneous phase has lower free energy than the vortex phase, while if the boundary conditions are such that the topological invariant is non-trivial, then the vortex phase wins. This also happens for normal superconductors at a special critical coupling for Ginzburg-Landau theory, which allows BPS vortices. One has, at a special point in the parameter space, both the homogeneous solution and the BPS crystal, and which solution is favored is decided by the boundary conditions.

The relation between the topological charge on the right hand side of the above equation and the magnetic flux will be discussed in the following sections.

It is interesting to note that, for small α\alpha, the above BPS system reduces exactly to the BPS system of multi-vortices in critical superconductors. In fact, considering

∣α∣≪1,sin⁡(α)∼α,cos⁡(α)∼1−α22,|\alpha| \ll1,\quad\sin(\alpha) \sim\alpha,\quad\cos(\alpha) \sim1-\frac{\alpha^{2}}{2},

we obtain the well-known system

∂1α+αA^2=0,\partial_{1}\alpha+\alpha\widehat{A}_{2}=0,
∂2α−αA^1=0,\partial_{2}\alpha-\alpha\widehat{A}_{1}=0,
B^z−K(1−α22)=0.\widehat{B}_{z}-K\left(1-\frac{\alpha^{2}}{2}\right)=0.

This fact explains why the boundary conditions at the positions of the vortices (where α\alpha is small) are the same as in the GL case.

In order to solve the BPS system in Eqs. (30), (31) and (32), it is convenient to define the function H(α)=ln⁡(tan⁡(α/2))H(\alpha)=\ln(\tan(\alpha/2)). It follows that

A^1=1sin⁡α∂2α=∂2H(α),A^2=−1sin⁡α∂1α=−∂1H(α).\widehat{A}_{1}=\frac{1}{\sin\alpha}\partial_{2}\alpha=\partial_{2}H(\alpha),\qquad\widehat{A}_{2}=-\frac{1}{\sin\alpha}\partial_{1}\alpha=-\partial_{1}H(\alpha).

Using the above expression for the Maxwell potential, the first order equations in Eqs. (30), (31) and (32), can be written as a second order differential equation for H(α)H(\alpha), namely,

∇2H=Ktanh⁡(H)+∑j=1Nnjδ(x−xj,y−yj),(34)\nabla^{2}H=K\tanh(H)+\sum_{j=1}^{N}n_{j}\delta(x-x_{j},y-y_{j}), \tag*{(34)}

where the second term on the right-hand side corresponds to the “Aharonov-Bohm contributions” from all the vortices.4 This equation (without the δ\delta-like terms) can be obtained from the variation of the following effective action:

Leff(H)=12(∇⃗H)2+Kln⁡(2cosh⁡H).(35)\mathcal{L}_{\mathrm{eff}}(H)=\frac{1}{2}(\vec{\nabla}H)^{2}+K\ln(2\cosh H). \tag*{(35)}

Advantages of the availability of BPS bounds

The appearance of a BPS configuration arises from a delicate balance between opposite forces (typically, when attractive and repulsive interactions of the theory of interest are of the same order), one of the most important examples being BPS vortices in superconductors. Despite the fact that these configurations appear at special points in parameters space, they have fundamental importance for several reasons (see [24]-[27] and references therein). Here we will mention just a few of them.

  1. These configurations signal a transition from one type of behavior to a different type of behavior (as in the transition from type-I to type-II superconductors).

  1. The physical effects generated by BPS configurations are non-perturbative in nature, namely that it cannot be captured by perturbation theory as the dependence of the relevant physical quantities on the coupling constants is manifestly non-analytic.

  1. There is a very powerful tool to describe the low-energy dynamics of these BPS configurations (which is not available for non-BPS configurations) in terms of geodesics on moduli space. Such a tool allows for the description of dynamical processes (such as vortex-vortex scattering), which are very difficult to analyze with different methods.

  1. BPS configurations are topologically stable (if one starts from a positive definite action functional), then, these solutions cannot be destroyed by quantum/thermal fluctuations.

  1. It is possible to derive many exact results about the number of fermionic zero modes which BPS solitons are able to support just looking at the topological charge: such results are unavailable for generic configurations [25], [74].

These configurations are known to appear in supersymmetric theories as well as in Yang-Mills-Higgs theory [75] (where non-abelian monopoles, instantons and so on are found in the non-perturbative sectors). On the other hand, to the best of author’s knowledge, here we constructed the first BPS vortices in χ\chiPT, which is very far from being supersymmetric. It is also worth noticing that even in the case of BPS vortices in BCS superconductors, the corresponding BPS equations must be solved numerically (unlike what happens for instantons and non-abelian monopoles). In fact, as it has already been stressed, the importance of BPS vortices in superconductors lies in their existence and not in the possibility of solving analytically the BPS equations.

At this point it is important to note relevant differences between the present results and the ones reported in Ref. [29]. In our case, in order to take full advantage of the BPS bound, we need to start from a free energy functional which is positive definite from the very beginning. Here, such a functional is the Gibbs free energy at the pressure in Eq. (11). Indeed, if we would use the free energy at zero external pressure, our BPS configuration would be saddle points but not necessarily minima in the corresponding topological sector. This means that the transition between type-I and type-II behavior, disclosed by the presence of the BPS vortices, appears at the value of the pressure given above. This is not surprising since external pressure affects heavily the properties of superconductors [76]. On the other hand, the results in Ref. [29] hold at zero external pressure.

Also, it is important to note that here GG is simply the free energy with the inclusion of a vacuum subtraction at fixed μI\mu_I, i.e., the free energy at vanishing fields is subtracted. On the other hand, such a subtraction is not necessary if, instead of the free energy one uses the Gibbs free energy, with the PVPV term included. Consequently, if one uses the Gibbs free energy instead of the free energy, not only any subtraction is avoided, but also one gets directly a positive definite functional, which ensures the property 4 in the list here above.

The interpretation of this fact is that a region of type-I behavior opens up at this value of the pressure. It is also worth emphasizing that in our results there is no approximation, as we have found the BPS point analytically, and due to the properties of BPS configurations, our results are robust against perturbations (something that cannot be said about generic vortices far from the BPS point).

Regarding property 5, let us shortly describe Weinberg’s computations in Ref. [74] to obtain the number of fermionic zero modes captured by our BPS vortices.

In order to compute the index, which is defined as

I=dim⁡(Ker⁡D)−dim⁡(Ker⁡D†) ,(36)\mathcal{I} = \dim(\operatorname{Ker}\mathcal{D}) - \dim(\operatorname{Ker}\mathcal{D}^{\dagger}) \, , \tag*{(36)}

one can follow Weinberg’s technique, considering the limit of M2→0M^2 \to0 of the quantity

I(M2)=Tr⁡(M2D†D+M2)−Tr⁡(M2DD†+M2).(37)\mathcal{I}(M^2) = \operatorname{Tr}\left(\frac{M^2}{\mathcal{D}^{\dagger}\mathcal{D} + M^2}\right) - \operatorname{Tr}\left(\frac{M^2}{\mathcal{D}\mathcal{D}^{\dagger} + M^2}\right) . \tag*{(37)}

The expression in Eq. (37) is extremely important because it counts the fermionic zero modes supported by the BPS configuration of interest. We would like to choose a suitable Dirac operator D\mathcal{D} in such a way that Weinberg’s procedure applied to our case produces, as a result, the topological invariant appearing on the right-hand side of the BPS bound in Eq. (33). This reverse engineering process is worthless unless the resulting Dirac operator turns out to be physically interesting. The comparison with the Dirac equation analyzed in Eq. (11) of Ref. [74] suggests that we should take D\mathcal{D} as

D=γμ(∂μ−QAμ)+Uγ5 ,(38)\mathcal{D} = \gamma^{\mu}(\partial_{\mu} - Q A_{\mu}) + U\gamma^{5} \, , \tag*{(38)}

where γ5\gamma_{5} is needed for chirality reasons and QQ is the charge matrix of the quarks. Indeed, in order to repeat the steps of Ref. [74] (see, in particular, Eqs. (9), (10) and (11) of Ref. [74]), the gauge potential and the chiral field UU should appear on the same footing (similarly as the gauge potential and the Higgs field appear in the case of vortices in BCS superconductors at critical coupling). On the other hand, Uγ5U^{\gamma_{5}} is necessary in the same way, as the Higgs field couples with fermions while its conjugate couples to anti-fermions. Quite interestingly, the above Dirac operator in Eq. (38) (whose form has been guessed with the “naive” requirement to reproduce the topological invariant in Eq. BPS upon using Weinberg’s procedure) is compatible with the Dirac operator proposed in Refs. [77, 78, 79, 80], to describe the coupling of quarks with the chiral field UU. The present arguments strongly suggest that the coupling of quarks with the BPS vortices introduced in the previous section could share many properties with the coupling of electrons with BPS vortices in critical superconductors. The analysis of the coupling of quarks with the present BPS vortices will be studied in detail in a future publication.

Multi-vortices at critical values

The novel BPS bound presented in the previous section presents many peculiar properties. In this section, we detail the most important ones.

Strong magnetic field and finite isospin chemical potential

  1. First, the topological charge density, which appears naturally in the BPS bound is not simply the magnetic flux density, as one would naively expect; instead, it is a “dressed magnetic flux" modulated by the hadronic profile (this possibility has been pointed out for the first time in Ref. [56]). The presence of the pions generates a screening effect on the magnetic flux, tending to suppress the magnetic field of the vortices in the condensate.

  1. The inhomogeneous condensate is stabilized by the topological charge as it happens for the usual GL vortices, and, as in the GL theory, the possibility to have multi-vortices configurations at arbitrary locations appears at a special value of the relevant parameter (which is the Higgs coupling in the GL case, while the isospin chemical potential in the present case). From the viewpoint of differential equations, this could be enough. However, from the physical viewpoint, the fact that these are topologically non-trivial minima of the Gibbs free energy (and not of the free energy) means that they can be realized in actual experiments keeping the pressure fixed at the value P=2K(μIc)2=858.474  (MeV)4P = 2K(\mu_I^c)^2 = 858.474\;(\mathrm{MeV})^4.

  1. Unlike what happens in the GL theory, where the derivation of a bound on the maximal magnetic field requires quite some work (see, for instance, Refs. [71, 72, 73]), the present novel BPS bound allows to find very easily the maximum Bmax⁡B_{\max} value for the magnetic field beyond which the condensate ceases to exist. Indeed, looking at Eq. (32), one gets

Bmax⁡=2K=2(efπ2)2=397,03  (MeV)2=2,04×1014  G.(39)B_{\max}=2K=2\left(\frac{e f_\pi}{2}\right)^2=397,03\;(\mathrm{MeV})^2=2,04\times10^{14}\;\mathrm{G}. \tag*{(39)}

The above value for the maximal magnetic field is valid at the BPS point. Even more, it is still a good estimate for Bmax⁡B_{\max} when the isospin chemical potential and pressure are close to corresponding values at the BPS bound. It is worth emphasizing that it is a novelty of our approach to be able to compute exactly (thanks to the BPS properties of the vortices) the maximal magnetic field even if just at a special point (as there are very few exact results in the low energy limit of QCD).

  1. The behavior of the electromagnetic field, the current and the energy density can be easily described thanks to availability of the BPS equations. First, as we have point out above, the magnetic field only possesses component in the zz direction. This magnetic field is generated by a self-sustained current, given by

Jμ=2Ksin⁡2(α)DμΦ.(40)J_\mu=2K\sin^2(\alpha)D_\mu\Phi. \tag*{(40)}

From Eq. (40) one can see that the current is not-null even when the electromagnetic field is suppressed. In fact, there is a persistent current generated by the coupling with pions, given by Jμ(0)=2Ksin⁡2(α)∂μΦJ_\mu^{(0)}=2K\sin^2(\alpha)\partial_\mu\Phi.

  1. The stationary point of the free energy density F\mathcal{F} can be determined through the variation of Eq. (10) with respect to the field α\alpha, Φ\Phi and AμA_\mu, leading to the following set of second order field equations

Δα−sin⁡(α)cos⁡(α)((D⃗Φ)2−4μI2)=0 ,\Delta\alpha-\sin(\alpha)\cos(\alpha)\left((\vec{D}\Phi)^2-4\mu_I^2\right)=0 \,,
ΔΦ−2∇⃗⋅A⃗+2cot⁡(α)∇⃗α⋅D⃗Φ=0 ,\Delta\Phi-2\vec{\nabla}\cdot\vec{A}+2\cot(\alpha)\vec{\nabla}\alpha\cdot\vec{D}\Phi=0 \,,
∂jFij−2Ksin⁡2(α)DiΦ=0 ,(41)\partial_jF^{ij}-2K\sin^2(\alpha)D^i\Phi=0 \,, \tag*{(41)}

where Δ\Delta is the Laplacian operator. As expected, the first order BPS equations in Eqs. (30), (31) and (32) imply the second order system in Eqs. (41). Moreover, one can check that the BPS equations also imply the field equations obtained directly varying the action in Eq. (1) with respect to the UU field and the spatial components of the Maxwell potential A⃗\vec{A}.

Inclusion of the pions mass

The present formalism can be applied even when the gauged χPT includes a pions mass term. As it has been discussed many times in the literature, it is possible to introduce a mass term for the pions in many different ways; see, for instance, Refs. [81, 82, 83, 84, 85]. The reason is that, if one write the chiral field as

U=σ1+πiti ,σ2+∑i=13(πi)2=1 ,U=\sigma1+\pi_i t^i \,,\qquad\sigma^2+\sum_{i=1}^{3}(\pi_i)^2=1 \,,

the mass term for the pions in the lagrangian can be seen as an interaction potential for σ\sigma, namely, Lmass=V(σ)\mathcal{L}_{\mathrm{mass}}=V(\sigma), and the mass of the pions mπm_\pi reads

mπ2=−∂2V∂σ2∣σ=1 .m_\pi^2=-\left.\frac{\partial^2V}{\partial\sigma^2}\right|_{\sigma=1} \,.

It is worth emphasizing that it is impossible to distinguish between the potentials by just knowing the pion mass (as it has been discussed in details in the above references). Here we will consider a mass term proposed in Refs. [81] and [82],

Lmass=−2Kmπ2e2(1−14Tr⁡[U]2) ,(42)\mathcal{L}_{\mathrm{mass}}=-\frac{2Km_\pi^2}{e^2}\left(1-\frac{1}{4}\operatorname{Tr}[U]^2\right) \,, \tag*{(42)}

which has several benefits. Such a mass term, at the same time, supports domain walls solutions (hadronic layers) and (for small pions field) gives the usual pions mass term. The appearance of hadronic layers at finite density is confirmed both by numerical simulations and by astronomical observations (see [86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97], and references therein). For the above reasons, there are very strong experimental evidence in favor of the mass term in Eq. (42).

Using the parametrization in Eqs. (5) and (7), the mass term in Eq. (42) reads

Lmass=2Kmπ2e2sin⁡2(α) .(43)\mathcal{L}_{\mathrm{mass}}=\frac{2Km_\pi^2}{e^2}\sin^2(\alpha) \,. \tag*{(43)}

and the free energy density in Eq. (10) becomes

e2F=K2{(∇⃗α)2+sin⁡2(α)(D⃗Φ)2}+12(B⃗2)+2KμI2cos⁡2(α)+2Kmπ2sin⁡2(α)−2KμI2=K2{(∇⃗α)2+sin⁡2(α)(D⃗Φ)2}+12(B⃗2)+2K(μI2−mπ2)cos⁡2(α)−2K(μI2−mπ2).(44)\begin{aligned} e^2 F = \frac{K}{2}\left\{(\vec{\nabla}\alpha)^2+\sin^2(\alpha)(\vec{D}\Phi)^2\right\}+\frac{1}{2}(\vec{B}^2)+2K\mu_I^2\cos^2(\alpha)+2Km_\pi^2\sin^2(\alpha)-2K\mu_I^2 \\ &= \frac{K}{2}\left\{(\vec{\nabla}\alpha)^2+\sin^2(\alpha)(\vec{D}\Phi)^2\right\}+\frac{1}{2}(\vec{B}^2)+2K(\mu_I^2-m_\pi^2)\cos^2(\alpha)-2K(\mu_I^2-m_\pi^2). \tag*{(44)} \end{aligned}

Consequently, the inclusion of the mass term for the pions manifests itself in a negative shift5 of the μI2\mu_I^2:

μI2→μI2−mπ2.\mu_I^2 \rightarrow\mu_I^2-m_\pi^2.

Hence, all the results in the previous sections still hold provided the shift in the above equation is taken into account. The critical value for the isospin chemical potential when the pions mass is taken into account reads

μIc=K+mπ2<1.1mπ.(45)\mu_I^c=\sqrt{K+m_\pi^2}<1.1m_\pi. \tag*{(45)}

It is important to emphasize that, even with the inclusion of the mass term, our value for the critical isospin chemical potential μIc\mu_I^c are within the range of validity of χ\chiPT (namely, 0.1mπ≲μIc<1.1mπ0.1m_\pi\lesssim\mu_I^c<1.1m_\pi). In fact, it is known (see [16], and references therein) that χ\chiPT needs improvement starting from μI∼1.3mπ\mu_I\sim1.3m_\pi.

A single vortex

In this section, we will study the case of a single vortex. This is great interest because it allows to explicitly show many of the most important properties of our solution.

The starting point is the pionic field in Eq. (23) together with the metric in Eq. (24). The Gibbs free energy density for a single vortex reduces to

e2G=K2r2{r2(α′)2+sin⁡2(α)(n−2A)2}+12r2(A′)2+2KμI2cos⁡2(α)=K2r2{rα′+sin⁡(α)(n−2A)}2+12r2{A′+μˉrcos⁡(α)}2−K1rsin⁡(α)(n−2A)α′−μˉrcos⁡(α)A′,\begin{aligned} e^2G=\frac{K}{2r^2}\{r^2(\alpha')^2+\sin^2(\alpha)(n-2A)^2\}+\frac{1}{2r^2}(A')^2+2K\mu_I^2\cos^2(\alpha) \\ &=\frac{K}{2r^2}\{r\alpha'+\sin(\alpha)(n-2A)\}^2+\frac{1}{2r^2}\{A'+\bar{\mu}r\cos(\alpha)\}^2-K\frac{1}{r}\sin(\alpha)(n-2A)\alpha'-\frac{\bar{\mu}}{r}\cos(\alpha)A', \end{aligned}

where we have defined μˉ2=4KμI2\bar{\mu}^2=4K\mu_I^2. Now, we can follow the same steps presented in the previous sections in a straightforward way. Integrating the last expression, we obtain the Gibbs free energy

GL=2π∫12e2r2{K(rα′+sin⁡(α)(n−2A))2+(A′+μˉrcos⁡(α))2} dr\frac{G}{L}=2\pi\int\frac{1}{2e^2r^2}\{K(r\alpha'+\sin(\alpha)(n-2A))^2+(A'+\bar{\mu}r\cos(\alpha))^2\}\,dr
−2πe2∫{Ksin⁡(α)(n−2A)α′+μˉcos⁡(α)A′} dr.(46)-\frac{2\pi}{e^2}\int\{K\sin(\alpha)(n-2A)\alpha'+\bar{\mu}\cos(\alpha)A'\}\,dr. \tag*{(46)}

The second term in Eq. (46) is a total derivative for the particular value of the isospin chemical potential presented in the previous section, that is μˉ=2K⇒μIc=K\bar{\mu}=2K \Rightarrow\mu_I^c=\sqrt{K}. In fact, for this value

−Ksin⁡(α)(n−2A)α′−μˉcos⁡(α)A′={Kcos⁡(α)(n−2A)}′.-K\sin(\alpha)(n-2A)\alpha' - \bar{\mu}\cos(\alpha)A' = \{K\cos(\alpha)(n-2A)\}'.

It follows from Eq. (46) that the Gibbs free energy is minimized when the following BPS equations are satisfied

rα′+sin⁡(α)(n−2A)=0,(47)r\alpha' + \sin(\alpha)(n-2A) = 0, \tag*{(47)}
A′+2Krcos⁡(α)=0,(48)A' + 2Kr\cos(\alpha) = 0, \tag*{(48)}

where, also in this case, one can check that the above first order BPS system implies the second-order field equations corresponding to the stationary points of the Gibbs free energy in Eq. (46).

The topological charge of this configuration reads

ω=2πe2∫{Kcos⁡(α)(n−2A)}′dr=2πKe2n.(49)\omega= \frac{2\pi}{e^2}\int\{K\cos(\alpha)(n-2A)\}'dr = \frac{2\pi K}{e^2}n. \tag*{(49)}

Note that, as we have pointed out in the previous section, the topological charge density of the vortices here presented is different from the magnetic flux density (due to the presence of the term cos⁡(α)\cos(\alpha) in Eq. (49)). However, once the boundary conditions are implemented, the topological charge is in fact proportional to the magnetic flux (times a dimensional constant that depends on the couplings of the theory), as can be seen from Eqs. (20) and (49). Now, if the solitons are confined in a finite volume, the topological charge and the magnetic flux may be different.6 This more general case will be studied in a forthcoming paper.

Figure 1 shows the behavior of the pionic profile α\alpha as well as the Maxwell potential characterized by the function AA for the single vortex with n=1n=1. To integrate the system in Eqs. (47) and (48) we have used the boundary conditions in Eqs. (14) and (15), with c1=0c_1=0 and c2=1c_2=1. To obtain clearer plots, we have fixed the value of the coupling constants to K=17K=\frac{1}{7} and fπ=2f_\pi=2. Figure 2 shows the Gibbs free energy density of a single vortex for different values of nn.

Figure 1. Profiles α\alpha and AA of the single vortex with n=1n=1.

Figure 2. Gibbs free energy density for a single vortex with n=1n=1, n=2n=2 and n=3n=3.

Regarding the electromagnetic field self-sustained by the vortex, the magnetic field is in the zz direction, given by

Bz=−A′r,(50)B_z=-\frac{A'}{r}, \tag*{(50)}

while the current in the θ\theta direction is

Jθ=2Ksin⁡2(α)(n−2A).(51)J_\theta=2K\sin^2(\alpha)(n-2A). \tag*{(51)}

Figure 3 shows the magnitude of the magnetic field of the single vortex for different values of nn.

Figure 3. Magnetic field of the single vortex for n=1n=1, n=2n=2 and n=3n=3.

Final remarks

In the present manuscript, we have shown that, at a critical value of the isospin chemical potential, superconducting multi-vortices configurations appear in gauged χ\chiPT. This has been achieved through the derivation of a novel BPS bound in χ\chiPT minimally coupled to electrodynamics, which is saturated by these multi-vortices configurations. This bound has not been discovered before because the topological charge density is not the magnetic flux density but rather is a magnetic flux density “dressed” by the hadronic profile. Such a dressing discloses the screening effects of the hadrons on the magnetic field, allowing us to derive an upper bound on the intensity of the magnetic field itself. These configurations are supported by a persistent current, which does not vanishes when the Maxwell field is suppressed, showing that the condensate has superconducting characteristics. These BPS vortices exist for a particular value of the isospin chemical potential which is within the range of validity of χ\chiPT. The present results open many interesting questions on which we hope to come back in a future publication. First of all, in the pioneering paper [44], E. Weinberg, using in a very clever way the index theorem, was able to determine the dimension of the moduli space of BPS multi-vortices in the GL theory at critical coupling. Then, there are two natural questions. The first is whether the dimension in the present case is the same as in the GL theory (the arguments presented in the Appendix suggest a positive answer). The second is whether the index appearing in the present case is related to the magnetic flux density, or the dressed magnetic flux density. Another closely related direction of research is the remarkable idea proposed by N. Manton (see [98]-[101], and references therein), according to which one can describe the dynamics of BPS solitons in terms of the geodesics of the geometry of the moduli space. Hence, it is an extremely interesting issue to analyze the geometry of the moduli space in the present case. On a more phenomenological side, the analysis of the coupling of up and down quarks to these BPS multi-vortices is very relevant. We think that all the above open questions deserve further investigation.

Acknowledgments

F. C. has been funded by FONDECYT Grant No. 1240048. M. L. has been funded by FONDECYT Iniciación No. 11241079.

On the existence of multi-vortices configurations

In this Appendix, we outline two mathematical strategies (which can be promoted to rigorous arguments, as it will be discussed in a future publication) leading to the proof of the existence of multi-vortices configurations in the χ\chiPT coupled to Maxwell at the critical value of the isospin chemical potential defined in the previous sections. Besides the intrinsic importance to prove rigorously the existence of multi-vortices solutions, the first strategy presented in this Appendix shows in a very clear way the difference between the present case, in which the hadronic profile screens substantially the magnetic field, and the usual GL case.

First Strategy

As it has been explained in the previous sections, the BPS equations in Eqs. (30), (31) and (32) are equivalent to the master equation in Eq. (34) which is a semi-linear elliptic equation in two coordinates. Moreover, in the case of multi-vortices in the GL theory at critical coupling, the BPS equations are equivalent to a single master equation, which is also a semi-linear elliptic equation in two coordinates (see [72], [73], and references therein). The differences between the two master equations only lie in the non-linear terms. Thus, let us write the two master equations in a unified fashion:

∇2H=ΨI(H)+∑j=1Nnjδ(x−xj,y−yj),I=1,2(52)\nabla^2 H = \Psi_I(H) + \sum_{j=1}^{N} n_j \delta(x-x_j,y-y_j), \quad I=1,2 \tag*{(52)}
Ψ1(H)=Ktanh⁡(H),Ψ2(H)=−1+exp⁡H\Psi_1(H)=K\tanh(H), \qquad\Psi_2(H)=-1+\exp H

where Ψ1(H)\Psi_1(H) corresponds to the BPS vortices in χ\chiPT presented in this paper while Ψ2(H)\Psi_2(H) corresponds to the BPS vortices in GL theory at critical coupling. Now, we will describe a technique (which is quite standard in the theory of non-linear elliptic equations; see Ref. [102]) based on the Banach-Caccioppoli theorem (see Ref. [103] and references therein). The idea to use a strategy based on the Banach-Caccioppoli theorem is that it highlights a very important difference between Ψ1(H)\Psi_1(H) and Ψ2(H)\Psi_2(H) which, physically, is related to the screening of the magnetic field from the hadronic profile, which is stronger in χ\chiPT than in GL theory at critical coupling. The mathematical consequence is that in the case of Ψ1(H)\Psi_1(H) one can apply directly the Banach-Caccioppoli theorem, while in the case of Ψ2(H)\Psi_2(H) the existence and uniqueness parts require far more refined techniques (see [72], [73], and references therein).

Let us begin with the definition of an auxiliary linear problem:

∇2u=ΨI(V)+∑j=1NnjPj,nj∈N,u∣∂Ω=0,V∈X,N<∞(53)\nabla^{2}u = \Psi_{I}(V) + \sum_{j=1}^{N} n_{j}P_{j},\quad n_{j} \in\mathbb{N},\quad u|_{\partial\Omega}=0,\quad V \in X,\quad N < \infty \tag*{(53)}
Ψ1(V)=Ktanh⁡V,Ψ2(V)=−1+exp⁡V,∇2=∂2∂x2+∂2∂y2(54)\Psi_{1}(V)=K\tanh V,\quad\Psi_{2}(V)=-1+\exp V,\quad\nabla^{2}=\frac{\partial^{2}}{\partial x^{2}}+\frac{\partial^{2}}{\partial y^{2}} \tag*{(54)}

where the functions PjP_{j} are defined as follows:

Pj(x,y)=Λ,(x,y)∈Ωj,Pj(x,y)=0,(x,y)∈Ω∖Ωj(55)P_{j}(x,y)=\Lambda,\quad(x,y)\in\Omega_{j},\quad P_{j}(x,y)=0,\quad(x,y)\in\Omega\setminus\Omega_{j} \tag*{(55)}
measure⁡(Ωj)=1Λ2,Ωj⊂Ω ∀j=1,…,N,Ωj∩Ωk=0 for j≠k(56)\operatorname{measure}(\Omega_{j})=\frac{1}{\Lambda^{2}},\quad\Omega_{j}\subset\Omega\ \forall j=1,\ldots,N,\quad\Omega_{j}\cap\Omega_{k}=0\ \text{for }j\ne k \tag*{(56)}
Pj∈L2(Ω)∀j=1,…,N(57)P_{j}\in L^{2}(\Omega)\quad\forall j=1,\ldots,N \tag*{(57)}
Ω=(0,a)×(0,a)(58)\Omega=(0,a)\times(0,a) \tag*{(58)}

The conditions in Eqs. (A.4), (A.5) and (A.6) describe the functions PjP_{j} (where Ωj\Omega_{j} is a small circle of area 1Λ2\frac{1}{\Lambda^{2}} around the point of coordinates (xj,yj)(x_{j},y_{j})). It is clear that when Λ\Lambda is a large number, the PjP_{j} tend to be very peaked (xj,yj)(x_{j},y_{j}): the idea to introduce the functions PjP_{j} is to define a limit in which, as Λ\Lambda becomes bigger and bigger, PjP_{j} becomes closer and closer to a Dirac delta. There are actually many different ways to represent the Dirac delta as the limit of bounded (even smooth) functions. This issue will be discussed in more detail in a future publication.

In the linear elliptic equation in Eq. (A.2), u=u(x,y)u=u(x,y) is the unknown, while V=V(x,y)V=V(x,y) is supposed to be known. For the moment, we will take as domain Ω\Omega a square of side a<K−1/2a<K^{-1/2} in Eq. (A.7) (then we will explain how to obtain similar results in bigger squares).

The first technical step is to choose the function VV (the “source term” of the auxiliary problem) in a suitable functional space XX in such a way that the solution u(x,y)u(x,y) of the auxiliary problem belongs to the same functional space of the function VV.

The second technical step is to define a mapping Φ\Phi such that for any Vk∈XV_{k}\in X the mapping Φ\Phi sends VkV_{k} into the solution uku_{k} of the auxiliary problem in Eq. (A.2) with V=VkV=V_{k}:

uk=Φ(Vk)∣ ∇2uk=ΨI(Vk)+∑j=1NnjPj,nj∈N,uk∣∂Ω=0u_{k}=\Phi(V_{k})\mid\ \nabla^{2}u_{k}=\Psi_{I}(V_{k})+\sum_{j=1}^{N}n_{j}P_{j},\quad n_{j}\in\mathbb{N},\quad u_{k}|_{\partial\Omega}=0

The third technical step is to show that such a mapping is a contraction so that, due to the Banach-Caccioppoli theorem, it possesses a unique fixed point. Then, it is trivial to see that such a fixed point is a solution of the original non-linear problem in Eq. (A.1).

First step

Here we need to put the basis to define such a mapping. The easiest choice is to assume that the functional space XX (to which the source VV in Eq. (A.2) belongs) is L2(Ω)L^{2}(\Omega):

X=L2(Ω)X=L^{2}(\Omega)

where L2(Ω)L^2(\Omega) is the space of square integrable function on Ω\Omega. Moreover, one observes the following. If

∇2u=f,u∣∂Ω=0,f∈L2(Ω)⇒u∈L2(Ω),(59)\nabla^2u=f,\quad u|_{\partial\Omega}=0,\quad f\in L^2(\Omega)\Rightarrow u\in L^2(\Omega), \tag*{(59)}

then (due to u∣∂Ω=0u|_{\partial\Omega}=0) one gets

∫Ωuf=∫Ωu∇2u=−[∫Ω(∂u∂x)2+∫Ω(∂u∂y)2]⇒[∫Ω(∂u∂x)2+∫Ω(∂u∂y)2]≤∥f∥L2(Ω)∥u∥L2(Ω),(60)\begin{aligned} \int_{\Omega}uf=\int_{\Omega}u\nabla^2u=-\left[\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2+\int_{\Omega}\left(\frac{\partial u}{\partial y}\right)^2\right] \\ &\Rightarrow\left[\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2+\int_{\Omega}\left(\frac{\partial u}{\partial y}\right)^2\right]\leq\|f\|_{L^2(\Omega)}\|u\|_{L^2(\Omega)}, \tag*{(60)} \end{aligned}

where we have used the Schwarz inequality (which will be used often in the following computations). Using the obvious identity

∫0x∂u∂σ(σ,y) dσ=u(x,y),\int_{0}^{x}\frac{\partial u}{\partial\sigma}(\sigma,y)\,d\sigma=u(x,y),

it follows that for any u(x,y)u(x,y) which solves the problem in Eq. (A.8) one gets

(∥u∥L2(Ω))2=∫Ωu2 dxdy=∫Ω[∫0x∂u∂σ(σ,y) dσ]2dxdy⇒(∥u∥L2(Ω))2≤∫Ω[(∫0x(∂u∂σ(σ,y))2dσ)1/2(∫0x12dσ)1/2]2dxdy=defΓ.(61)\begin{aligned} \left(\|u\|_{L^2(\Omega)}\right)^2=\int_{\Omega}u^2\,dxdy=\int_{\Omega}\left[\int_{0}^{x}\frac{\partial u}{\partial\sigma}(\sigma,y)\,d\sigma\right]^2dxdy \\ &\Rightarrow\left(\|u\|_{L^2(\Omega)}\right)^2\leq\int_{\Omega}\left[\left(\int_{0}^{x}\left(\frac{\partial u}{\partial\sigma}(\sigma,y)\right)^2d\sigma\right)^{1/2}\left(\int_{0}^{x}1^2d\sigma\right)^{1/2}\right]^2dxdy\mathrel{\overset{\mathrm{def}}{=}}\Gamma. \tag*{(61)} \end{aligned}

It is possible to further bound from above Γ\Gamma as follows:

Γ≤a∫Ω(∫0a(∂u∂σ(σ,y))2dσ)dxdy=a∫0adx∫0ady[∫0a(∂u∂σ(σ,y))2dσ]=a∫0adx[∫Ω(∂u∂x)2dxdy]dx=a2∫Ω(∂u∂x)2dxdy,⇒(∥u∥L2(Ω))2≤a2∫Ω(∂u∂x)2dxdy.(62)\begin{aligned} \Gamma\leq a\int_{\Omega}\left(\int_{0}^{a}\left(\frac{\partial u}{\partial\sigma}(\sigma,y)\right)^2d\sigma\right)dxdy=a\int_{0}^{a}dx\int_{0}^{a}dy\left[\int_{0}^{a}\left(\frac{\partial u}{\partial\sigma}(\sigma,y)\right)^2d\sigma\right] \\ &=a\int_{0}^{a}dx\left[\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2dxdy\right]dx=a^2\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2dxdy, \\ &\Rightarrow\left(\|u\|_{L^2(\Omega)}\right)^2\leq a^2\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2dxdy. \tag*{(62)} \end{aligned}

Hence, from Eqs. (A.9) and (A.11) it follows that

∫Ω(∂u∂x)2≤a(∫Ω(∂u∂x)2dxdy)1/2∥f∥L2(Ω)⇒(∫Ω(∂u∂x)2dxdy)1/2≤a∥f∥L2(Ω),(63)\begin{aligned} \int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2\leq a\left(\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2dxdy\right)^{1/2}\|f\|_{L^2(\Omega)} \\ &\Rightarrow\left(\int_{\Omega}\left(\frac{\partial u}{\partial x}\right)^2dxdy\right)^{1/2}\leq a\|f\|_{L^2(\Omega)}, \tag*{(63)} \end{aligned}

and this is true ∀\forall uu which is a solution of the problem in Eq. (A.8). Similarly, we get

(∫Ω(∂u∂y)2dxdy)1/2≤a∥f∥L2(Ω).(64)\left(\int_{\Omega}\left(\frac{\partial u}{\partial y}\right)^2dxdy\right)^{1/2}\leq a\|f\|_{L^2(\Omega)}. \tag*{(64)}

Moreover, from Eqs. (A.11) and (A.12) together with (A.13), we also get

∥u∥L2(Ω)≤a2∥f∥L2(Ω).(65)\lVert u\rVert_{L^2(\Omega)} \le a^2 \lVert f\rVert_{L^2(\Omega)} . \tag*{(65)}

As it is well known (see, for instance, [102]) the problem in Eq. (A.8) possesses an existence and uniqueness theorem: ∀f∈L2(Ω)\forall f \in L^2(\Omega) there is a unique uu which solves the problem in Eq. (A.8). One can also say that, actually, the solution uu actually belongs to a well-behaved subspace of L2(Ω)L^2(\Omega):

u∈W,W=def{u∈W2,2(Ω), u∣∂Ω=0},u \in\mathrm{W}, \qquad\mathrm{W} \mathrel{\mathop{=}\limits^{\mathrm{def}}} \left\{u \in W^{2,2}(\Omega),\ u|_{\partial\Omega}=0\right\},

where W2,2(Ω)W^{2,2}(\Omega) is the (2,2) Sobolev space on the domain Ω\Omega (for a definition of Sobolev space, see [102] and references therein).

Second step

Here we define such a mapping. As a particular case of the problem in Eq. (A.8), let us consider the problem in Eq. (A.2) with Ψ1(V)=Ktanh⁡V\Psi_1(V)=K\tanh V, where

V∈L2(Ω).V \in L^2(\Omega) .

Thus, in this case,

f=Ktanh⁡V+∑j=1NnjPj∈L2(Ω).f=K\tanh V+\sum_{j=1}^{N}n_jP_j \in L^2(\Omega) .

The existence and uniqueness result for the problem in Eq. (A.8) provides with the following:

∀V∈L2(Ω), ∃!u∈W⊂L2(Ω),\forall V \in L^2(\Omega),\ \exists! u \in\mathrm{W} \subset L^2(\Omega),

which solves the problem in Eq. (A.2) with Ψ1(V)=Ktanh⁡V\Psi_1(V)=K\tanh V. We are in a position to define a mapping Φ\Phi such that for any V∈L2(Ω)V \in L^2(\Omega) the mapping Φ\Phi sends VV into the unique solution uu of the auxiliary problem in Eq. (A.2):

u=Φ(V)(66)u=\Phi(V) \tag*{(66)}
∇2u=Ψ1(V)+∑j=1NnjPj,nj∈N,u∣∂Ω=0,Ψ1(V)=Ktanh⁡V.(67)\nabla^2u=\Psi_1(V)+\sum_{j=1}^{N}n_jP_j,\quad n_j\in\mathbb{N},\quad u|_{\partial\Omega}=0,\quad\Psi_1(V)=K\tanh V . \tag*{(67)}

Let us consider now two different “sources” V1V_1 and V2V_2 together with the corresponding images under the mappings u1=Φ(V1)u_1=\Phi(V_1) and u2=Φ(V2)u_2=\Phi(V_2). From the above definitions it follows that

∇2(u1−u2)=K(tanh⁡V1−tanh⁡V2)∈L2(Ω),(u1−u2)∣∂Ω=0,u1−u2∈W.\nabla^2(u_1-u_2)=K(\tanh V_1-\tanh V_2)\in L^2(\Omega),\quad(u_1-u_2)|_{\partial\Omega}=0,\quad u_1-u_2\in\mathrm{W}.

Then, from Eq. (A.14), it follows that

∥u1−u2∥L2(Ω)≤a2∥K(tanh⁡V1−tanh⁡V2)∥L2(Ω)⇒∥u1−u2∥L2(Ω)≤Ka2∥tanh⁡V1−tanh⁡V2∥L2(Ω)=Ka2(∫Ω∣tanh⁡V1−tanh⁡V2∣2 dx dy)1/2.(68)\begin{aligned} \lVert u_1-u_2\rVert_{L^2(\Omega)}\le a^2\lVert K(\tanh V_1-\tanh V_2)\rVert_{L^2(\Omega)} \\ &\Rightarrow\lVert u_1-u_2\rVert_{L^2(\Omega)}\le Ka^2\lVert\tanh V_1-\tanh V_2\rVert_{L^2(\Omega)}=Ka^2\left(\int_{\Omega}|\tanh V_1-\tanh V_2|^2\,dx\,dy\right)^{1/2}. \tag*{(68)} \end{aligned}

Then, using the elementary result from calculus (namely, the mean value theorem) that for any continuous differentiable function h(x)h(x) we have

h(x)−h(y)=h′(z)(x−y),h′(z)=dhdz,z∈(x,y)⇒∣tanh⁡V1−tanh⁡V2∣=1cosh⁡2(z)∣V1−V2∣,1cosh⁡2(z)≤1 ∀z∈R,(69)\begin{aligned} h(x)-h(y)=h'(z)(x-y),\quad h'(z)=\frac{dh}{dz},\quad z\in(x,y) \\ &\Rightarrow|\tanh V_1-\tanh V_2|=\frac{1}{\cosh^2(z)}|V_1-V_2|,\quad\frac{1}{\cosh^2(z)}\leq1\ \forall z\in\mathbb{R}, \tag*{(69)} \end{aligned}

for some zz. Consequently, from Eq. (A.17) together with the mean value theorem (and taking into account that the derivative of the hyperbolic tangent is always less than or equal to 1 in absolute value) we get

∥u1−u2∥L2(Ω)≤Ka2(∫Ω∣V1−V2∣2 dxdy)1/2.(70)\|u_1-u_2\|_{L^2(\Omega)}\leq Ka^2\left(\int_{\Omega}|V_1-V_2|^2\,dxdy\right)^{1/2}. \tag*{(70)}

Third step

Here we prove that the mapping defined in Eqs. (A.15) and (A.16) is a contraction thanks to Eq. (A.19). Indeed, from the definition in Eqs. (A.15) and (A.16) together with Eq. (A.19) we obtain that

∀V1, V2∈L2(Ω),∥Φ(V1)−Φ(V2)∥L2(Ω)≤Ka2∥V1−V2∥L2(Ω).(71)\forall V_1,\ V_2\in L^2(\Omega),\qquad\|\Phi(V_1)-\Phi(V_2)\|_{L^2(\Omega)}\leq Ka^2\|V_1-V_2\|_{L^2(\Omega)}. \tag*{(71)}

Consequently, if we take Ka2<1Ka^2<1, we see that the mapping defined here above is a contraction so that, due to the Banach-Caccioppoli theorem, it possesses a unique fixed point:

∃V∗ such that Φ(V∗)=V∗.(72)\exists V^*\ \text{such that}\ \Phi(V^*)=V^*. \tag*{(72)}

If we explicitly write the condition to be a fixed point of the mapping defined in Eqs. (A.15) and (A.16), we see that V∗V^* satisfies the following equation:

∇2V∗=Ψ1(V∗)+∑j=1NnjPj,nj∈N,V∗∣∂Ω=0,Ψ1(V∗)=Ktanh⁡V∗\nabla^2V^*=\Psi_1(V^*)+\sum_{j=1}^{N}n_jP_j,\quad n_j\in\mathbb{N},\quad V^*|_{\partial\Omega}=0,\quad\Psi_1(V^*)=K\tanh V^*

namely, the fixed point V∗V^* of the mapping defined here above is the solution of the sought non-linear elliptic differential equation. Note that such simple argument would not work in the case of the BPS vortices in the GL theory. The reason is that, in the GL case, Ψ2(V)=−1+exp⁡V\Psi_2(V)=-1+\exp V, and if we follow the same procedure, when we apply the mean value theorem we would get

∣exp⁡V1−exp⁡V2∣=(exp⁡z)∣V1−V2∣,|\exp V_1-\exp V_2|=(\exp z)|V_1-V_2|,

for some zz. However, unlike what happens with the Ψ1(V)=Ktanh⁡V\Psi_1(V)=K\tanh V, which characterizes the BPS vortices in χ\chiPT introduced in the present work, ∣exp⁡z∣|\exp z| is not less than or equal to one. This is the reason why the proof of existence and uniqueness of BPS multi-vortices in the GL case requires very refined techniques (see Refs. [72] and [73]). This technical difference is related to the fact that, in the case of χ\chiPT, the hadronic profile screens considerably the magnetic field.

It is worth emphasizing that the above argument is not complete. First of all, one should define properly the limit of Λ→∞\Lambda\to\infty in order to obtain Pj→δ(x−xj,y−yj)P_j \to\delta(x-x_j,y-y_j). This issue appears to be within the reach of usual techniques in functional analysis (see Refs. [102] and [103]). Secondly, the smoothness of the solution must be improved, as we would like V∗V^* to be smooth, far from the positions of the vortices. The usual “bootstrap arguments” or “elliptic regularity theory” in Sobolev spaces (which have been already used in the GL case: see [71, 72, 73]) appear to work in the present case as well without much change. Thirdly, we would like to eliminate the constraint Ka2<1Ka^2 < 1 on the size of the domain Ω\Omega allowing domains of generic sizes. This can be certainly be achieved using more refined inequalities in functional analysis (see [104] and references therein).

It is important to remark here that the idea of the present paper is not to give a fully rigorous proof of the existence and uniqueness (for this a separate more mathematical paper is necessary) but to provide with sound evidences that such BPS multi-vortices configurations do exist. In the present work we have discussed (in this Appendix) not only strong mathematical arguments in favor of the existence of such BPS configurations but also (in the main text) we have constructed numerical solutions for the single vortex case.

Second Strategy

The first strategy (which has been presented here above) is interesting since it emphasizes how the physical properties, which has been discussed in the main text (such as the hadronic screening of the magnetic field) manifest themselves in the possibility of using the contraction theorem directly. However, the easiest and fastest strategy to prove rigorously that such multi-vortices exist is actually to follow exactly the same steps of Ref. [72]. A direct analysis reveals that many of the steps can be repeated without change in the present case. In particular, we can start from our main equation

∇2H=Ktanh⁡(H)+∑j=1Nnjδ(x−xj,y−yj).(73)\nabla^2 H = K \tanh(H) + \sum_{j=1}^{N} n_j \delta(x-x_j,y-y_j). \tag*{(73)}

Then, the functions u0u_0 and g0g_0 in Eqs. (10) and (12) of Ref. [72] can be introduced in the present case as well in order to take care of the Dirac deltas. More precisely, one can choose (the constant parameters appearing in) u0u_0 in such a way that its Laplacian cancels exactly the sum over the Dirac deltas, leaving the smooth function g0g_0 only. Then, (as in Eq. (13) of Ref. [72]) the new unknown vv can be defined as follows:

H=u0+v.H = u_0 + v.

If the above expression is replaced into Eq. (A.22) one arrives at

∇2v=g0+Ktanh⁡(u0+v).\nabla^2 v = g_0 + K \tanh(u_0+v).

The above equation for vv can be derived by the following action

Seff(v)=12(∇⃗v)2+g0v+Kln⁡(2cosh⁡(u0+v)),(74)\mathcal{S}_{\mathrm{eff}}(v) = \frac{1}{2}(\vec{\nabla}v)^2 + g_0v + K\ln\left(2\cosh(u_0+v)\right), \tag*{(74)}

to be compared with Eq. (15) of Ref. [72]. The most important step in the analysis of Ref. [72] is the proof that the action in his Eq. (15) is a convex functional. In particular, in Eq. (4.28) of [72] a fundamental role is played by the strict convexity of the non-linear part of the action (the exponential in that case). On the other hand, one can observe that the term Kln⁡(2cosh⁡(u0+v))K \ln(2\cosh(u_0+v)), in Eq. (A.23), is a strict convex function of its argument as well. Moreover, there are many technical steps of Ref. [72] which actually do not depend on the explicit form of the non-linear term in the action in his Eq. (15): such steps can be imported directly to the present situations. At last, there are steps (such as the inequalities in Eqs. (44) and (45) of [72]) which can be easily generalized to the present case thanks both to Eq. (A.18) discussed in the first part of the Appendix and to the fact that ln⁡(2cosh⁡X)\ln(2\cosh X) is well approximated by ∣X∣|X|.

The above sketchy arguments show that the strategy of Ref. [72] can be adapted to the present case, leading to a rigorous proof of the existence of multi-vortex solutions. We will come back to this issue in a future publication.

References

  1. [1]P. de Forcrand, *Simulating QCD at finite density*, PoS(LAT2009)010 [arXiv:1005.0539] [INSPIRE].
  2. [2]J. B. Kogut, M. A. Stephanov, The phases of Quantum Chromodynamics, Cambridge University Press (2004).
  3. [3]N. Brambilla et al., QCD and Strongly Coupled Gauge Theories: Challenges and Perspectives, Eur. Phys. J. C 74 (2014) 2981 [arXiv:1404.3723] [INSPIRE].DOI
  4. [4]R. Pisarski, Lect. Notes Phys. 999, 89-145 (2022).
  5. [5]A. Bzdak, S. Esumi, V. Koch, J. Liao, M. Stephanov et al., Phys. Rept. 853 (2020) 1-87.
  6. [6]K. Nagata, Prog. Part. Nucl. Phys. 127 (2022) 103991.
  7. [7]N. Astrakhantsev, V.V. Braguta, N.V. Kolomoyets, A.Yu. Kotov, D.D. Kuznedelev et al., Phys.Part.Nucl. 52 (2021) 4, 536-541; N. Astrakhantsev, V. Braguta, M. Cardinali, M. D’Elia, L. Maio et al., PoS LATTICE2021 (2022) 119; B. B. Brandt, F. Cuterí, G. Endrődi, G. Markó, L. Sandbote, A. D. M. Valois, arXiv:2305.19029.
  8. [8]W. Busza, K. Rajagopal, W. van der Schee, Annu. Rev. Nucl. Part. Sci. 2018. 68:339–76.
  9. [9]K. Yagi, T. Hatsuda, Y. Miake, *Quark–Gluons Plasma*, Cambridge University Press (2005).
  10. [10]D.T. Son, M. A. Stephanov, Phys. Rev. Lett. 86 (2001) 592-595.
  11. [16]T. Kojo, D. Suenaga, Phys. Rev. D 105 (2022) 7, 076001.
  12. [21]B. B. Brandt, G. Endrodi and S. Schmalzbauer, EPJ Web Conf. 175, 07020 (2018).
  13. [22]S. Carignano, L. Lepori, A. Mammentarella, M. Mannarelli and G. Pagliaroli, Eur. Phys. J. A 53, no.2, 35 (2017).
  14. [23]M. Mannarelli, Particles 2, no.3, 411-443 (2019).
  15. [24]N. Manton and P. Sutcliffe, Topological Solitons (Cambridge University Press, Cambridge, 2007).
  16. [25]E. Shuryak, Nonperturbative Topological Phenomena in QCD and Related Theories (Lecture Notes in Physics, 977, 2021 edition).DOI
  17. [26]M. Shifman, Advanced Topics in Quantum Field Theory: A Lecture Course, Cambridge University Press 2022.
  18. [27]E. J. Weinberg, “Classical solutions in quantum field theory: Solitons and Instantons in High Energy Physics,” Cambridge University Press, 2012, ISBN 978-0-521-11463-9, 978-1-139-57461-7, 978-0-521-11463-9, 978-1-107-43805-7.DOI
  19. [28]P. Adhikari and J. Choi, Int. J. Mod. Phys. A 33, no.36, 1850215 (2019).
  20. [29]P. Adhikari, Phys. Lett. B 790, 211-217 (2019).
  21. [30]G. W. Evans and A. Schmitt, JHEP 09, 192 (2022).
  22. [31]G. W. Evans and A. Schmitt, JHEP 2024, no.02, 041 (2024).
  23. [32]A. Edery, JHEP 01, 166 (2021).
  24. [33]A. Edery, Phys. Rev. D 106, no.6, 065017 (2022).
  25. [34]Z. Qiu and M. Nitta, JHEP 06, 139 (2024).
  26. [35]M. Eto, K. Nishimura and M. Nitta, JHEP 12, 032 (2023).
  27. [36]M. Eto, K. Nishimura and M. Nitta, [arXiv:2304.02940 [hep-ph]].arxiv.org/abs/2304.02940
  28. [37]M. Eto and M. Nitta, Phys. Rev. D 104, no.9, 094052 (2021).
  29. [38]F. Canfora, Phys. Rev. D 88, (2013), 065028; Eur. Phys. J. C78, no. 11, 929 (2018).
  30. [39]F. Canfora, S. H. Oh and A. Vera, Eur. Phys. J. C 79, no.6, 485 (2019).
  31. [40]F. Canfora, M. Lagos and A. Vera, Eur. Phys. J. C 80, no.8, 697 (2020).
  32. [41]F. Canfora, S. Carignano, M. Lagos, M. Mannarelli, A. Vera, Phys. Rev. D 103 (2021) 7, 076003.
  33. [42]P. D. Alvarez, F. Canfora, N. Dimakis and A. Paliathanasis, Phys. Lett. B 773, (2017) 401-407; F. Canfora, N. Dimakis, A. Paliathanasis, Eur. Phys. J. C79 (2019) no.2, 139.
  34. [43]S. L. Cacciatori, F. Canfora, M. Lagos, F. Muscolino, A. Vera, JHEP 12 (2021) 150; Nucl. Phys. B 976 (2022) 115693.
  35. [44]E. J. Weinberg, Phys. Rev. D 19 (1979) 3008.
  36. [45]S. Scherer, M. R. Schindler, A Primer for Chiral Perturbation Theory, Lecture Notes in Physics. Berlin Heidelberg: Springer-Verlag (2012). ISBN 978-3-642-19253-1.DOI
  37. [46]J. Donoghue, E. Golowich, B. Holstein, Dynamics of the Standard Model, (Cambridge University Press, 1994).
  38. [47]R. Machleidt, D. R. Entem, Physics Reports 503 (1): 1–75 (2011).
  39. [48]J. Gasser and H. Leutwyler, Ann. Phys., vol. 158, p. 142, 1984.
  40. [49]H. Leutwyler, Ann. Phys., vol. 235, pp. 165–203, 1994.
  41. [50]G. Ecker, Prog. Part. Nucl. Phys., vol. 35, 1-80, 1995.
  42. [51]S. Scherer, Adv. Nucl. Phys., vol. 27, p. 277, 2003.
  43. [52]S. Carignano, A. Mammarella and M. Mannarelli, Phys. Rev. D 93, no.5, 051503 (2016).
  44. [53]A. Balachandran, G. Marmo, B. Skagerstam, A. Stern, Classical Topology and Quantum States, World Scientific (1991).
  45. [54]G. Baym, B. L. Friman and G. Grinstein, Nucl. Phys. B210 (1982) 193.
  46. [55]Y. Hidaka, K. Kamikado, T. Kanazawa and T. Noumi, Phys. Rev. D92 (2015) 034003.
  47. [56]F. Canfora, JHEP 11 (2023) 007.
  48. [57]K. Hattori, T. Kojo and N. Su, Nucl. Phys. A 951, 1-30 (2016).
  49. [58]K. Hattori and X. G. Huang, Nucl. Sci. Tech. 28, no.2, 26 (2017).
  50. [59]H. T. Ding, S. T. Li, A. Tomiya, X. D. Wang and Y. Zhang, Phys. Rev. D 104, no.1, 014505 (2021).
  51. [60]F. Canfora, H. Maeda, Phys. Rev. D 87, 084049 (2013).
  52. [61]F. Canfora, F. Correa, J. Zanelli, Phys. Rev. D 90, 085002 (2014).
  53. [62]F. Canfora, M. Di Mauro, M. A. Kurkov, A. Naddeo, Eur. Phys. J C75 (2015) 9, 443.
  54. [63]E. Ayon-Beato, F. Canfora, J. Zanelli, Phys. Lett. B 752, (2016) 201-205; E. Ayón-Beato, F. Canfora, M. Lagos, J. Oliva and A. Vera, Eur. Phys. J. C 80, no.5, 384 (2020).
  55. [64]L. Aviles, F. Canfora, N. Dimakis, D. Hidalgo, Phys. Rev. D 96 (2017), 125005; F. Canfora, M. Lagos, S. H. Oh, J. Oliva and A. Vera, Phys. Rev. D 98, no. 8, 085003 (2018).
  56. [65]F. Canfora, D. Hidalgo, M. Lagos, E. Meneses and A. Vera, Phys. Rev. D 106, no.10, 105016 (2022).
  57. [66]S. L. Cacciatori, F. Canfora, F. Muscolino, Nucl. Phys. B 1000 (2024) 116477.
  58. [70]X. Cao, H. Liu, D. Li and G. Ou, Chin. Phys. C 44, no.8, 083106 (2020).
  59. [71]S. Wang, Y. Yang, Siam Journal of Mathematical Analysis 23 (1992), 1125-1140.
  60. [72]C. H. Taubes, Comm. Math. Phys. 72, 277-292 (1980).
  61. [73]A. Jaffe, C. H. Taubes, Vortices and Monopoles, Birkhauser, Boston (1980).
  62. [74]E. J. Weinberg, Phys. Rev. D 24, 2669 (1981).
  63. [75]M. Mariño, Instantons and Large N: An Introduction to Non-Perturbative Methods in Quantum Field Theory, Cambridge University Press, 2015. ISBN 978-1-107-06852-0, 978-1-316-37154-1.DOI
  64. [76]Lilia Boeri et al 2022 J. Phys.: Condens. Matter 34 183002.
  65. [77]R. Jackiw and P. Rossi, Nucl. Phys. B 190 (1981), 681-691.
  66. [78]A. P. Balachandran, V. P. Nair, S. G. Rajeev and A. Stern, Phys. Rev. D 27 (1983), 1153 [erratum: Phys. Rev. D 27 (1983), 2772].
  67. [79]A. P. Balachandran, V. P. Nair, S. G. Rajeev and A. Stern, Phys. Rev. Lett. 49 (1982), 1124 [erratum: Phys. Rev. Lett. 50 (1983), 1630].
  68. [80]S. Kahana, R. Perry and G. Ripka, Phys. Lett. B 163 (1985), 37-40.
  69. [81]S. B. Gudnason and M. Nitta, Phys. Rev. D 94, no.6, 065018 (2016).
  70. [82]S. B. Gudnason, Phys. Rev. D 93, no.6, 065048 (2016).
  71. [83]G. S. Adkins and C. R. Nappi, Nucl. Phys. B 233, 109-115 (1984).
  72. [84]V. B. Kopeliovich, B. Piette and W. J. Zakrzewski, Phys. Rev. D 73, 014006 (2006).
  73. [85]L. Marleau, Phys. Rev. D 43, 885-890 (1991).
  74. [86]J. A. Lopez, C. O. Dorso, G. A. Frank, Front.Phys. (Beijing) 16 (2021) 2, 24301.
  75. [87]C. J. Horowitz, D. K. Berry, C.M. Briggs, M. E. Caplan, A. Cumming, A. S. Schneider, Phys. Rev. Lett. 114, 031102 (2015).
  76. [88]D. K. Berry, M. E. Caplan, C. J. Horowitz, G. Huber, A. S. Schneider, Phys. Rev. C 94, 055801 (2016).
  77. [89]C. O. Dorso, G. A. Frank, J. A. López, Nucl. Phys. A978, 35 (2018).
  78. [90]A. da Silva Schneider, M. E. Caplan, D. K. Berry, C. J. Horowitz, Phys. Rev. C 98, 055801 (2018).
  79. [91]M. E. Caplan, A. S. Schneider, and C. J. Horowitz, Phys. Rev. Lett. 121, 132701 (2018).
  80. [92]R. Nandi and S. Schramm, J. Astrophys. Astron. 39, 40 (2018).
  81. [93]Z. Lin, M. E. Caplan, C. J. Horowitz, C. Lunardini, Phys. Rev. C 102 (2020) 4, 045801.
  82. [94]C.O. Dorso, A. Strachan, G.A. Frank, Nucl. Phys. A 1002 (2020) 122004.
  83. [95]C.J. Pethick, Z. Zhang, D.N. Kobyakov, Phys. Rev. C 101 (2020) 5, 055802.
  84. [96]G. Watanabe and T. Maruyama, [arXiv:1109.3511 [nucl-th]].
  85. [97]A. Schmitt, Phys. Rev. D 101, no.7, 074007 (2020).
  86. [101]T.M. Samols, Commun.Math.Phys. 145 (1992) 149-180.
  87. [102]D. Gilbarg, N. Trudinger, Elliptic Partial Differential Equations of Second Order, Springer (2013).
  88. [103]D. Kinderlehrer, G. Stampacchia, Variational Inequalities in R^N: An Introduction to Variational Inequalities and Their Applications, New York: Academic Press (1980), pp. 7–22.
  89. [104]P. Pedregal, Functional Analysis, Sobolev Spaces, and Calculus of Variations, Springer 2024.DOI

Paper details

vlnc:W_d5dnq3yk
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