betaFeedback

WorksPhysics and Astronomy

Metastable vortex string decay in a brane picture

Abstract

We describe the decay of a metastable string as a bounce in a brane configuration. We consider a particular N=1\mathcal N=1 supersymmetric U(2)\mathrm{U}(2) gauge theory with cascade Higgsing U(2)→U(1)2→U(1)\mathrm{U}(2)\to \mathrm{U}(1)^2 \to \mathrm{U}(1), where the vortex string can break by creating a pair of 't Hooft-Polyakov monopole and anti-monopole. The theory is engineered in a Type IIA brane setup, where the metastable string is realized as a D2-brane stretched between NS5- and D4-branes. The string breaks when the endpoints of the D2-brane, moving along the NS5- and D4-branes, meet so that the D2-brane shrinks to nothing. We check that the tension of the D2-brane reproduces that of the vortex string obtained numerically in the field theory, with the ratio approaching one in the limit where the vortex string becomes BPS. Within the thin-wall approximation, we derive the bounce action from the Dirac-Born-Infeld action of the D2-brane and find that it agrees with that of the field theory, with the monopole mass given by the area swept by the D2-brane. As a consistency check, we numerically construct a bounce solution of the D2-brane equation of motion under simplifying assumptions, whose action agrees with the thin-wall result in the regime where our description is reliable.

Introduction

Solitons play important roles in high-energy physics, both from theoretical and phenomenological points of view. For the theoretical point of view, one of the most important solitons is branes in string theory [1]. Branes are non-perturbative objects in string theory, and they are crucial for understanding not only the non-perturbative dynamics of string theory but also the non-perturbative dynamics of gauge theories realized on the branes [2, 3].

For the phenomenological point of view, topological defects in field theories, such as cosmic strings can be produced in the early universe and can have various cosmological implications [4]. In particular, cosmic strings can produce gravitational waves, and the recent evidence of gravitational wave background by multiple pulsar timing array (PTA) collaborations [5, 6, 7, 8, 9, 10, 11] has triggered a lot of interest in metastable cosmic strings as a possible source of the observed gravitational wave background [12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39].

Interestingly, some topological defects in field theories can be described by branes in string theory [40, 41, 42, 43, 44]. For instance, in N=2\mathcal{N} = 2 supersymmetric gauge theories with U(N)\mathrm{U}(N) gauge group with NN fundamental hypermultiplets, the gauge symmetry is completely broken if we introduce the Fayet-Iliopoulos (FI) term, and vortex strings can exist in the theory as π1(U(N))=Z\pi_{1}(\mathrm{U}(N)) = \mathbb{Z} topological defects. These vortex strings preserve half of the supersymmetry and are Bogomol’nyi-Prasad-Sommerfield (BPS) states. In the brane picture, the N=2\mathcal{N} = 2 supersymmetric gauge theory with U(N)\mathrm{U}(N) gauge group with NN fundamental hypermultiplets and the FI term is realized by the Hanany-Witten setup [2, 3] in Type IIA string theory, which consists of NS5-branes, D4-branes and D6-branes. In this brane configuration, the vortex strings correspond to D2-branes stretched between the NS5-brane and the D4-branes. The tension of the vortex strings can be calculated from the tension of the D2-branes, and it matches with that calculated in the field theory picture. The brane picture also successfully describes the structure of the moduli space of the vortex strings and the dynamics on the string worldsheet.

Then, a natural question arises: can we describe non-BPS topological defects in field theories using the brane picture as well? There has been several attempts for that. Of the most well-known examples is Sen’s conjecture on the tachyon condensation of non-BPS D-branes [45], where, for example, the unstable pair of a D(p+2)(p+2)-brane and an anti-D(p+2)(p+2)-brane can annihilate into a Dpp-brane, or a kink on an unstable D(p+1)(p+1)-brane can be identified with a DpD p-brane. Furthermore, several works have focused on the non-BPS properties of topological defects in field theories and their brane realizations [40].

Of particular interest are metastable cosmic strings, which can be produced in relatively simple field theories; suppose, for example, we have a (non-supersymmetric) SU(2) gauge theory with an adjoint scalar field and a fundamental scalar field in the UV theory and the gauge symmetry is broken hierarchically as SU(2)→U(1)→\mathrm{SU}(2) \to\mathrm{U}(1) \to nothing. Then, from the second symmetry breaking, we have cosmic strings as π1(U(1))=Z\pi_{1}(\mathrm{U}(1)) = \mathbb{Z}, but the cosmic strings are not stable since the vacuum manifold is simply connected: π1(SU(2))=0\pi_{1}(\mathrm{SU}(2)) = 0. Therefore, the cosmic strings are metastable with endpoints [46, 47], which correspond to monopoles for π2(SU(2)/U(1))=Z\pi_{2}(\mathrm{SU}(2)/\mathrm{U}(1)) = \mathbb{Z}.

For hierarchical symmetry breaking, the decay rate of the metastable strings is analytically calculable using the thin-wall approximation [46, 47]. However, it becomes unreliable in the parameter region of interest for the PTA observations, and one must resort to numerical computations in the field theory [48, 49, 50].

These metastable strings and their monopole endpoints have also been realized in the brane picture. In Ref. [51], monopoles confined by vortices are realized in MQCD as D2D2-branes bounded by D4D4-branes and NS5-branes, and the vortex is identified with a strip of the D2D2-brane continuing into the spacetime directions, whose width is proportional to the vortex tension. In Ref. [52], an electroweak-like string and the Nambu monopole at its end are realized by a single D2D2-brane, and the monopole masses are obtained from the minimal surface area of the D2D2-brane in the internal space. There, it is also pointed out that the string can break by bending the D2D2-brane into the internal space, which creates a monopole-anti-monopole pair. However, the tunneling process of this string breaking itself is not described there in the brane picture; for the decay rate, Ref. [52] refers to the thin-wall formula in the field theory [47], which is not derived from the dynamics of the D2D2-brane.

On the other hand, describing tunneling processes by a Euclidean string worldsheet or brane worldvolume that ends on other branes has a long history. The nucleation of brane bubbles bounded by brane junctions was studied in Ref. [53], and the Schwinger-type creation of charge and monopole pairs was interpreted in terms of brane configurations in Ref. [54]. The pair creation of W-bosons and of open strings has been computed with Euclidean worldsheet instantons, both in flat space [55, 56] and holographically [56, 57]. The breaking of a confining string by quark pair creation, which describes meson decays, has been studied with a worldsheet instanton that splits the string on a flavour brane [58, 59]. Decay rates have also been computed in brane configurations for the decay of metastable vacua seeded by solitons [60, 61], where in Ref. [61] the seed is a D2-brane string identified in Ref. [62]. However, the breaking of a string by monopole-anti-monopole pair creation, where the string itself is a D2-brane ending on other branes, has not been described in this way.

In this paper, we describe this tunneling process in the brane picture and derive its bounce action from the dynamics of the D2-brane. We focus on the Hanany-Witten setup [2, 3] in Type IIA string theory. There, the N=1\mathcal{N}=1 supersymmetric gauge theory with U(2) gauge group is realized with an adjoint chiral multiplet and two fundamental chiral multiplets. The theory undergoes two-step symmetry breaking, U(2) →\to U(1) ×\times U(1) →\to U(1), so that metastable vortex strings can exist associated to the second symmetry breaking. In the brane picture, the metastable vortex strings correspond to non-BPS D2-branes stretched between the NS5-brane and the D4-branes. We first identify the parameters of the field theory with the geometrical parameters of the brane configuration, including the numerical coefficients, and check the identification by comparing the tension of the D2-brane with that of the vortex string obtained by explicitly solving the field equations.

We then describe the string breaking as the tunneling of the D2-brane, and formulate its bounce using the Dirac-Born-Infeld (DBI) action of the Euclidean D2-brane, together with the boundary condition given by the NS5-branes and the D4-branes in the Hanany-Witten setup. The “stable” vacuum in this process corresponds to the absence of the metastable string, i.e., the D2-brane itself, where the DBI action is no more applicable. We point out that special care is needed for the bounce there, unlike the tunneling process in field theory [63, 64]: the D2-brane suddenly starts to exist at a finite radius in the Euclidean spacetime. Within the thin-wall approximation, the DBI action of the bounce takes the same form as the field-theory bounce action, 2πmR−πTR22\pi mR-\pi T R^{2}: the D2-brane tension gives the string tension TT and the area swept by the D2-brane in the internal space gives the monopole mass mm, with the string length and the string coupling entering only through the gauge coupling. Finally, we construct the bounce solution explicitly by numerically solving the Euclidean equation of motion of the D2-brane. We find that such a solution exists and that its action agrees with the thin-wall result, which provides a consistency check of our picture of the tunneling. Since the D2-brane is described by the low-energy DBI action as a thin object, it does not capture the massive modes that form the cores of the monopole and the vortex string; our description is therefore reliable where the thin-wall approximation is valid, i.e., where the bubble is much larger than these cores.

The rest of this paper is organized as follows. In Sec. II, we introduce the brane construction of the N=1\mathcal{N}=1 supersymmetric gauge theory with U(2)\mathrm{U}(2) gauge group with an adjoint chiral multiplet and two fundamental chiral multiplets, identify the parameters in the field theory picture with the geometrical parameters in the brane setup, and compare the tension of the D2-brane with that of the vortex string obtained in Appendix § 1; we also discuss the conditions under which the identification is valid. In Sec. III, we describe the string breaking as the tunneling of the D2-brane, formulate its bounce with the DBI action of the Euclidean D2-brane, and show that the bounce action reproduces that of the field theory within the thin-wall approximation. In Sec. IV, we present our numerical algorithm to construct the bounce solution of the Euclidean D2-brane, compare the resulting bounce action with the thin-wall result, and discuss the validity of the thin-wall approximation in the brane and field theory pictures. Sec. V is devoted to the conclusion and discussion.

Brane construction and parameters

In this section, we introduce the brane construction of the N=1\mathcal{N}=1 supersymmetric gauge theory with U(2)\mathrm{U}(2) gauge group with an adjoint chiral multiplet and two fundamental chiral multiplets. As we discuss later, this theory admits a cascade pattern of symmetry breaking, U(2)→U(1)×U(1)→U(1)\mathrm{U}(2) \to\mathrm{U}(1) \times\mathrm{U}(1) \to\mathrm{U}(1) in the field theory picture, so that metastable (non-BPS) vortex strings can exist in the intermediate phase with U(1)×U(1)\mathrm{U}(1) \times\mathrm{U}(1) gauge symmetry whereas the metastable vortex strings correspond to non-BPS D2-branes in the brane picture. To perform quantitative analysis of the metastable vortex strings in the brane picture, we need to identify the parameters in the field theory picture with the geometrical parameters in the brane setup with correct numerical coefficients. In this section, we perform this identification and discuss the conditions for the validity of our identification.

The brane configuration we consider here is the Hanany-Witten setup [2, 3] in Type IIA string theory, which consists of NS5-brane, NS5′-brane, D41_1-branes and D42_2-brane, as shown in Figure 1.1 The NS5-branes are extended along the x1x^1, x2x^2, x3x^3, x4x^4, x5x^5 directions.

Figure 1. The brane configuration in Type IIA string theory realizing the N=1\mathcal{N}=1 supersymmetric gauge theory with U(2) gauge group with an adjoint chiral multiplet and two fundamental chiral multiplets. The two NS5-branes are separated in the x6x^{6} direction by a distance L6L_{6}. One of the NS5-branes, which we call NS5′-brane, is rotated in the (v,w)(v,w) plane by an angle θ\theta to break the supersymmetry down to N=1\mathcal{N}=1. The D41_1-brane is hanging between the two NS5-branes, whereas the D42_2-brane is semi-infinite, ending on the NS5-brane on one side. The distance between the D41_1-brane and the D42_2-brane in the (x4,x5)(x^{4},x^{5}) plane is ∣v′∣|v'|. We also show a (non-BPS) D2-brane, which corresponds to a (non-BPS) vortex string in the field theory picture.

One of the NS5-branes, which we call NS5′-brane, is rotated to break the supersymmetry down to N=1\mathcal{N}=1. The D41_1-branes are suspended between the two NS5-branes along the x6x^6 direction, and the D42_2-brane is semi-infinite, ending on one of the NS5-branes and extending to x6→+∞x^6 \to+\infty. The D4-branes are extended along the x1,x2,x3,x6x^1,x^2,x^3,x^6 directions. The 3+13+1 dimensional gauge theory lives on the x1,x2,x3x^1,x^2,x^3 directions of the branes. In the language of the field theory picture, this brane configuration realizes the N=1\mathcal{N}=1 supersymmetric gauge theory with U(2) gauge group with an adjoint chiral multiplet, Φ\Phi, and two fundamental chiral multiplets, QQ and Q~\widetilde{Q}. When the NS5′-brane is not rotated, the brane configuration preserves the N=2\mathcal{N}=2 supersymmetry, so that the adjoint chiral multiplet Φ\Phi is massless and forms the N=2\mathcal{N}=2 vector multiplet together with the gauge field and the chiral multiplets QQ and Q~\widetilde{Q} form the N=2\mathcal{N}=2 hypermultiplet. The superpotential of the theory is given by

W=yQ~ΦQ−mQ~Q−μTr⁡(Φ2).(1)W=y\widetilde{Q}\Phi Q-m\widetilde{Q}Q-\mu\operatorname{Tr}(\Phi^2). \tag*{(1)}

We need to express the parameters yy, mm, and μ\mu in the superpotential, along with the gauge coupling constant gg in terms of the string theory parameters, which are string scale ℓs≡α′\ell_s\equiv\sqrt{\alpha'}, the string coupling constant gsg_s, and the geometrical parameters of the brane configuration. Since we aim to perform quantitative analysis of the field theory, we need to identify these parameters with correct numerical coefficients, not merely their parametric dependences. However, as we will see, our identification is consistent only when certain conditions are satisfied in addition to the weak string coupling condition gs≪1g_s\ll1. We discuss these conditions at the end of this section.

The distance between the two NS5-branes in the x6x^6 direction, L6L_6, controls the gauge coupling constant of the gauge theory. The gauge coupling constant gg is calculated from the DBI action of the D4-branes. The gauge coupling constant in the DBI action, g5g_5 is given by

1g52=τ4(2πℓs)22gs,(2)\frac{1}{g_5^2}=\frac{\tau_4(2\pi\ell_s)^2}{2g_s}, \tag*{(2)}

where τp=2π(2πℓs)p+1\tau_p=\frac{2\pi}{(2\pi\ell_s)^{p+1}} is the Dp-brane tension in the string frame, ℓs\ell_s is the string length [65]. The 3+13+1 dimensional gauge coupling constant is given by compactifying the x6x^6 direction with length L6L_6:

1g2=L6g52=L68π2gsℓs.(3)\frac{1}{g^2}=\frac{L_6}{g_5^2}=\frac{L_6}{8\pi^2g_s\ell_s}. \tag*{(3)}

on the Nambu monopole. Here, instead, we consider the cascade breaking U(2) →\to U(1) ×\times U(1) →\to U(1) with a rotated NS5-brane, so that the endpoint of the metastable string is the ’t Hooft-Polyakov monopole.

Throughout this paper, we take the limit gs≪1g_s \ll1 with fixed gg, so that the stringy effects are suppressed and the classical description of the brane configuration is valid.

Next, we consider v′v', the distance between the D41_1-brane and the D42_2-brane in the (x4,x5)(x^{4},x^{5}) plane, where we define the complex coordinate v=x4+ix5v=x^{4}+ix^{5}. In the following discussion, we assume gs≪1g_s \ll1 and all the energy scales we consider are much smaller than the string scale ℓs−1\ell_s^{-1}, so that the kinetic terms of the fields on the branes are well approximated by the DBI action and assumed to be canonical. Let us denote the positions of D41_1-brane and D42_2-brane in the (x4,x5)(x^{4},x^{5}) plane as 00 and v′v', respectively. The mass of the WW-boson, which corresponds to the open string stretched between the two D4-branes, is given by mW=∣v′∣/(2πℓs2)m_W=|v'|/(2\pi\ell_s^2). In the field theory picture, the mass of the WW-boson is given by mW=2g∣Φ3∣m_W=\sqrt{2}g|\Phi^3|. Here, we use the convention that the generators of the fundamental representation of U(2) are normalized as Tr⁡(TaTb)=12δab\operatorname{Tr}(T^aT^b)=\frac{1}{2}\delta^{ab} and Φ\Phi is expanded as Φ=ΦaTa\Phi=\Phi^aT^a. Then, we have

Φ=v′22πℓs2g(1000).(4)\Phi=\frac{v'}{2\sqrt{2}\pi\ell_s^2g}\begin{pmatrix}1 & 0 \\ 0 & 0\end{pmatrix}. \tag*{(4)}

In other words, we identify Φ3=v′22πℓs2g\Phi^3=\frac{v'}{2\sqrt{2}\pi\ell_s^2g} with Φ0=Φ3\Phi^0=\Phi^3. Here and after, we take x5=0x^5=0 for simplicity, so that v′v' is real.

With this identification of the adjoint scalar VEV and v′v', we can write the correspondence between the angle of the NS5′-brane rotation and the adjoint mass parameter in the field theory picture. Let us define another complex coordinate w=x8+ix9w=x^8+ix^9 of the (x8,x9)(x^8,x^9) plane. Before the rotation, the NS5′-brane is at w′=0w'=0 plane, extending along the x1x^1, x2x^2, x3x^3 and vv directions. We rotate the NS5′-brane in the (v,w)(v,w) plane, so that the NS5′-brane is at

w′=R(ϕ,θ)v′(5)w'=R(\phi,\theta)v' \tag*{(5)}

plane with a rotation parameter R(θ,ϕ)=eiϕtan⁡θR(\theta,\phi)=e^{i\phi}\tan\theta, where θ\theta and ϕ\phi are the rotation angles. The rotation breaks the supersymmetry down to N=1\mathcal{N}=1 by introducing the mass of the adjoint scalar field, μ\mu. To identify the relation between μ\mu and R(θ,ϕ)R(\theta,\phi), let us consider the energy density when one of the D4-branes is at v′=0v'=0 and the other is at v′=ϵ≪ℓsv'=\epsilon\ll\ell_s. In the field theory picture, the potential is given by V=∣μ∣2∣Φa∣2V=|\mu|^2|\Phi^a|^2. Using Eq. (4), we have V=∣μ∣2ϵ2/(4π2ℓs4g2)V=|\mu|^2\epsilon^2/(4\pi^2\ell_s^4g^2), where we add both Φ0\Phi^0 and Φ3\Phi^3. In the brane picture, the length of the D4-brane in the x6x^6 and ww directions is elongated by ΔL≃∣R(θ,ϕ)∣2ϵ2/(2L6)\Delta L\simeq|R(\theta,\phi)|^2\epsilon^2/(2L_6) for small ϵ\epsilon. The additional energy density is given by ΔE=τ4gsΔL=2π∣R(θ,ϕ)∣2ϵ2(2πℓs)52L6gs\Delta E=\frac{\tau_4}{g_s}\Delta L=\frac{2\pi|R(\theta,\phi)|^2\epsilon^2}{(2\pi\ell_s)^5 2L_6g_s}. Matching the two energy densities, we have

μ=R(θ,ϕ)L6.(6)\mu=\frac{R(\theta,\phi)}{L_6}. \tag*{(6)}

In the following, we take ϕ=0\phi=0 as well for simplicity, so that R(θ,ϕ)=tan⁡θR(\theta,\phi)=\tan\theta.

Next, we identify the mass parameter of the fundamental chiral multiplets. To discuss this, let us first consider the vacua of the theory we have introduced so far. In the brane picture shown in Figure 1, the D42_2-brane is semi-infinite, ending on the NS5-brane on one side and extending to x6→+∞x^6\to+\infty. This brane configuration allows another vacuum, as shown in Figure 2. There, the D42_2-brane is split into two segments, D4′2_2-brane and D4″2_2-brane. The D4′2_2-brane is hanging between the NS5-brane and the NS5′-brane, being coincident with the D41_1-brane so that the gauge symmetry is U(2) and unbroken. The D4″2_2-brane is hanging between the NS5′-brane and extending to x6→+∞x^6 \to+\infty. One can check the total length of the D42_2-brane in the x6x^6 direction is the same in both configurations, so that they are degenerate vacua of the theory. Note that one may consider D6-brane extending along the x1x^1, x2x^2, x3x^3, x7x^7, x8x^8, x9x^9 directions at some x6x^6 position much larger than the NS5′-brane position. In this case, the D42_2-brane can end on the D6-brane and one can easily see that the two configurations shown in Figure 1 and Figure 2 are indeed degenerate vacua. As long as the D6-brane is far away from the other branes, the presence of the D6-brane does not affect the dynamics of the other branes, and we can ignore it in the following discussion.

Figure 2. Another vacuum of the brane configuration shown in Figure 1, where D42_2-brane is split into two segments, D42′_2'-brane and D42′′_2''-brane. The D42′_2'-brane is hanging between the NS5-brane and the NS5′-brane, whereas the D42′′_2''-brane is hanging between the NS5′-brane and extending to x6→+∞x^6 \to+\infty.

In the unbroken vacuum shown in Figure 2, the fundamental chiral multiplets corresponds to the excitations of the open strings stretched between the stack of D4-branes (D41_1-brane and D42′_2'-brane) and the D42′′_2''-brane. Since the two segments of the D42_2-brane are at different positions in vv coordinate, the fundamental chiral multiplets are massive. Around the endpoint of the D4-branes on the NS5′-brane, the brane configuration is locally the same as that of Figure 3, where the two NS5-branes are completely parallel, and the distance between the D41_1-brane (and D42′_2'-brane) and the D42′′_2''-brane in the vv coordinate is v′/cos⁡θv'/\cos\theta. The physical mass of the hypermultiplet in this brane configuration is given by

Figure 3. Brane configuration realizing the N=2\mathcal{N}=2 supersymmetric gauge theory with U(2)U(2) gauge group with a fundamental hypermultiplet. The two NS5-branes are completely parallel, and the D42′′_2^{\prime\prime}-brane is at v=v′/cos⁡θv=v^{\prime}/\cos\theta, whereas the D41_1-brane and D42′_2^{\prime}-brane are at v=0v=0.

mhyperphys=v′2πℓs2cos⁡θ.(7)m_{\mathrm{hyper}}^{\mathrm{phys}}=\frac{v^{\prime}}{2\pi\ell_s^2\cos\theta}. \tag*{(7)}

Therefore, the physical mass of the fundamental chiral multiplets in the unbroken vacuum of Figure 2 is mhyperphysm_{\mathrm{hyper}}^{\mathrm{phys}} as well.

However, this is not the mass parameter mm appearing in the superpotential of the field theory picture, as it does not have the correct quantum number under the RR symmetry. In the N=2\mathcal{N}=2 theory shown in Figure 3, the RR-symmetry is SU(2)R×U(1)R\mathrm{SU}(2)_R \times\mathrm{U}(1)_R, where the SU(2)R\mathrm{SU}(2)_R symmetry corresponds to the rotation in the (x7,x8,x9)(x^7,x^8,x^9) directions and the U(1)R\mathrm{U}(1)_R symmetry does to the rotation in the (x4,x5)(x^4,x^5) directions. With the rotation of the NS5′-brane shown in Eq. (5), the SU(2)R×U(1)R\mathrm{SU}(2)_R \times\mathrm{U}(1)_R symmetry is broken down to U(1)R\mathrm{U}(1)_R symmetry, which is the simultaneous rotation in the vv and ww planes with the same angle, v→eiαvv \to e^{i\alpha}v and w→eiαww \to e^{i\alpha}w. Another U(1)\mathrm{U}(1) symmetry is the simultaneous rotation in the vv and ww planes with R(θ,ϕ)R(\theta,\phi), v→eiβvv \to e^{i\beta}v, w→e−iβww \to e^{-i\beta}w and R(θ,ϕ)→e−2iβR(θ,ϕ)R(\theta,\phi) \to e^{-2i\beta}R(\theta,\phi). Obviously, this U(1)\mathrm{U}(1) symmetry is broken by the definite value of R(θ,ϕ)R(\theta,\phi), but we can formally assume R(θ,ϕ)R(\theta,\phi) is a spurion field with the U(1)\mathrm{U}(1) charge to keep the U(1)\mathrm{U}(1) symmetry. Indeed, the adjoint mass in the superpotential, W⊃−μTr⁡(Φ2)W \supset-\mu\operatorname{Tr}(\Phi^2), is invariant under both the U(1)R\mathrm{U}(1)_R and U(1)\mathrm{U}(1) symmetries since Φ\Phi is the fluctuation of the D4-brane position in the vv direction and has the same quantum numbers as vv.

To identify the mass parameter mm in the superpotential, let us first consider the N=2\mathcal{N}=2 limit, μ→0\mu\to0. In this limit, the SU(2)R×U(1)R\mathrm{SU}(2)_{R} \times\mathrm{U}(1)_{R} symmetry is restored and the Yukawa coupling in the superpotential, W⊃Q~ΦQW \supset\widetilde{Q}\Phi Q, is invariant under the SU(2)R×U(1)R\mathrm{SU}(2)_{R} \times\mathrm{U}(1)_{R} symmetry. Therefore, the Yukawa coupling is also invariant under both the U(1)R\mathrm{U}(1)_{R} and U(1)\mathrm{U}(1) symmetries. The mass term of the fundamental chiral multiplets in the superpotential, W⊃−mQ~QW \supset-m\widetilde{Q}Q, should also be invariant under both the U(1)R\mathrm{U}(1)_{R} and U(1)\mathrm{U}(1) symmetries; in other words, the mass parameter mm should have the same quantum numbers as Φ\Phi, or vv. Therefore, the mass term such as Eq. (7), or v′/R(θ,ϕ)v'/R(\theta,\phi), to maintain the holomorphicity, cannot enter the superpotential as it has different quantum numbers under the U(1)\mathrm{U}(1) symmetries. Instead, the mass parameter mm in the superpotential should be proportional to v′v', which has the same quantum numbers as vv. Matching the dimensions, we have

m=v′2πℓs2.(8)m = \frac{v'}{2\pi\ell_s^2}. \tag*{(8)}

Then, to have the correct physical mass, we assume that the Kähler potential of the fundamental chiral multiplets, KQ\mathcal{K}_{Q}, is non-trivial and given by

KQ=cos⁡θ(Q†eVQ+Q~†e−VQ~),(9)\mathcal{K}_{Q} = \cos\theta\left(Q^\dagger e^{V}Q + \widetilde{Q}^\dagger e^{-V}\widetilde{Q}\right), \tag*{(9)}

where VV is the vector superfield. With this Kähler potential, the physical mass of the fundamental chiral multiplets is given by mhyperphys=∣m/cos⁡θ∣m_{\mathrm{hyper}}^{\mathrm{phys}} = \lvert m/\cos\theta\rvert, which matches Eq. (7).

Finally, let us identify the Yukawa coupling constant between the adjoint chiral multiplet and the fundamental chiral multiplets. Let the Yukawa coupling constant be yy. With the full superpotential shown in Eq. (1), the FF-flatness condition for Φ\Phi, QQ and Q~\widetilde{Q} requires

ycos⁡θQQ~−2μΦ=0,yΦQ−mQ=0,yQ~Φ−mQ~=0.(10)\frac{y}{\cos\theta}Q\widetilde{Q} - 2\mu\Phi= 0,\qquad y\Phi Q - mQ = 0,\qquad y\widetilde{Q}\Phi- m\widetilde{Q} = 0. \tag*{(10)}

Together with the DD-flatness condition, we can fix yy and the VEVs of the chiral multiplets:

y=2g,Q=(q0),Q~=(q0),(11)y = \sqrt{2}g,\qquad Q = \begin{pmatrix} q \\ 0 \end{pmatrix},\qquad\widetilde{Q} = \begin{pmatrix} q & 0 \end{pmatrix}, \tag*{(11)}

with

q2=v′μcos⁡θ2πℓs2g2.(12)q^2 = \frac{v'\mu\cos\theta}{2\pi\ell_s^2g^2}. \tag*{(12)}

Let us then discuss the conditions for the validity of the above identification of the parameters. First, we need to require that the string coupling constant is weak, so that the stringy effects are suppressed and the classical description of the brane configuration is valid. Since we would like to fix the gauge coupling constant gg in the field theory picture, we need to take the limit

gs→0,L6/ℓs→0,(13)g_s \to0,\qquad L_6/\ell_s \to0, \tag*{(13)}

simultaneously, with fixed gg. In the numerical analysis below, we take finite values, gs=10−3g_s = 10^{-3} and L6/ℓs=0.3L_6/\ell_s = 0.3 for the reference parameters (up to 0.6 in the scaling checks of Sec. IV), which give g≃0.5g \simeq0.5; L6/ℓsL_6/\ell_s is thus small but not parametrically small. Next, with the above identification of the parameters, we first require that the physical masses of the fields on the branes in the unbroken vacuum are much smaller than heavier modes such as the stringy modes and the Kaluza-Klein modes in the x6x^6 direction, so that the low-energy effective field theory on the branes is well approximated by the field theory picture we have introduced. This requires

tan⁡θ≪π,v′2πℓs≪1,(14)\tan\theta\ll\pi,\qquad\frac{v'}{2\pi\ell_s} \ll1, \tag*{(14)}

given that we take the limit L6/ℓs→0L_6/\ell_s \to0.

With the above conditions, both the brane picture and the field theory picture are valid descriptions of the same physics. Since we are interested in the metastable vortex strings, the scale of the first symmetry breaking U(2)→U(1)×U(1)\mathrm{U}(2) \to\mathrm{U}(1) \times\mathrm{U}(1) should be larger than that of the second symmetry breaking U(1)×U(1)→U(1)\mathrm{U}(1) \times\mathrm{U}(1) \to\mathrm{U}(1). Therefore, we need to require that

(v′22πℓs2g)2>(q2)2,(15)\left(\frac{v'}{2\sqrt{2}\pi\ell_s^2 g}\right)^2 > \left(\frac{q}{2}\right)^2, \tag*{(15)}

where the factor of 1/21/2 comes from the normalization of the generators of the fundamental representation of SU(2)\mathrm{SU}(2). Equivalently, we have

v′πℓs2≫μ.(16)\frac{v'}{\pi\ell_s^2} \gg\mu. \tag*{(16)}

With this condition, the metastable vortex strings appear in the field theory picture.

In the field theory picture, metastable vortex strings and monopoles appear as the result of the hierarchical symmetry breaking, U(2)→U(1)×U(1)→U(1)\mathrm{U}(2) \to\mathrm{U}(1) \times\mathrm{U}(1) \to\mathrm{U}(1). The mass of the monopole and the tension of the vortex string are [66] given by

mmonopole∼2v′g2ℓs2,Tvortex∼4πq2,(17)m_{\mathrm{monopole}} \sim\frac{2v'}{g^{2}\ell_{s}^{2}},\qquad T_{\mathrm{vortex}} \sim4\pi q^{2}, \tag*{(17)}

respectively, up to non-BPS corrections.

Since both the brane picture and the field theory picture are valid descriptions of the same physics, the metastable vortex strings and the monopoles should appear in the brane picture as well. We will discuss the monopole in the brane picture in the next section, and here we focus on the metastable vortex strings. Unlike the BPS vortex strings in the N=2\mathcal{N}=2 supersymmetric gauge theories [41, 42], the metastable vortex strings are non-BPS, so that they do not correspond to BPS configuration of D-branes in the brane picture. However, based on the observation in Ref. [51], we can identify the metastable vortex strings with the non-BPS configuration of D2-branes stretched between the NS5′-brane and the D42_2-brane and infinite in the, for example, x3x^{3} direction, as shown in Figure 1. Then, the tension of the metastable vortex string in the brane picture is

T3=τ2gs×v′sin⁡θ≃4πq2.(18)T_{3}=\frac{\tau_{2}}{g_{s}}\times v'\sin\theta\simeq4\pi q^{2}. \tag*{(18)}

Here, however, we have ignored the bending of the D42_2-brane due to the presence of the D2-brane. As L6<ℓsL_{6}<\ell_{s}, the bending is not negligible [67, 68], and the tension of the D2-brane is smaller than that in Eq. (18). This bending is the brane counterpart of the classical profile of the fields around the vortex string, just as a D-string ending on a D3-brane bends it into the profile of the BPS monopole [67, 68]; neglecting it amounts to neglecting the core structure of the vortex string. The bending is not controlled by the conditions in Eqs. (14) and (16), and we come back to its possible effects in § 0.5.

Now, let us compare the tension of the metastable vortex in the brane picture with that in the field theory picture. The order of magnitude of the tension matches, as in Ref. [51], where the tension is identified up to a numerical factor. The question is whether the numerical coefficient matches as well. To check this, we explicitly construct the string solution in the field theory picture and calculate the tension of the vortex string. The details of the construction of the vortex string solution and the calculation of its tension are shown in Appendix A. In Figure 4, we show the ratio of the tension of the vortex string in the field theory picture, TvortexT_{\mathrm{vortex}}, to that in the brane picture, T3T_{3}, as a function of θ\theta. Here, we take L6=6L_{6}=6, v′=25v'=25, ℓs=20\ell_{s}=20 and gs=10−3g_{s}=10^{-3}, which satisfy Eqs. (14) and (16). We can see that the ratio approaches 1 as θ→0\theta\to0 and stays within about 15% for θ≤0.1\theta\le0.1, although we have ignored the bending of the D42D4_2-brane in the brane picture. This is a non-trivial check of the identification of the metastable vortex string with the non-BPS D2-brane in the brane picture. The ratio is in fact bounded from above. Setting η=1\eta= 1 in the tension functional of Appendix A freezes the adjoint field at its vacuum value, and the remaining Abelian-Higgs model has the BPS potential, whose vortex tension is 4πq2=T34\pi q^2 = T_3 up to O(θ2)O(\theta^2) corrections from the Kähler potential. Since the actual solution minimizes the functional, Tvortex≤T3T_{\mathrm{vortex}} \le T_3. The deviation at finite θ\theta comes from the relaxation of the adjoint field inside the core of the vortex string, η<1\eta< 1, which is suppressed as θ→0\theta\to0, where the adjoint field becomes much heavier than the fields of the second symmetry breaking. In the brane picture, this relaxation is a displacement of the D42D4_2-brane in the vv direction near the D2-brane. Heuristically, the D2-brane is perpendicular to the NS5′-brane and pulls the D42D4_2-brane with a component along vv proportional to sin⁡θ\sin\theta, which vanishes as θ→0\theta\to0. This part of the bending, neglected in Eq. (18), lowers the tension, consistent with the sign of the deviation.

Figure 4. The ratio of the tension of the vortex string in the field theory picture, TvortexT_{\mathrm{vortex}}, to that in the brane picture, T3T_3, as a function of θ\theta. Here, we take L6=6L_6 = 6, v′=25v' = 25, ℓs=20\ell_s = 20 and gs=10−3g_s = 10^{-3}, which satisfy Eqs. (14) and (16).

String breaking processes in the brane picture

With the setup we have introduced in the previous section, we can analyze the string breaking processes in the brane picture. In the field theory picture, the metastable vortex string can break by creating a monopole-anti-monopole pair; the homotopy class of the second symmetry breaking U(1)×U(1)→U(1)\mathrm{U}(1)\times\mathrm{U}(1) \to\mathrm{U}(1) is non-trivial, π1(U(1)×U(1)/U(1))=Z\pi_1(\mathrm{U}(1)\times\mathrm{U}(1)/\mathrm{U}(1)) = \mathbb{Z}, but the homotopy class of the total symmetry breaking U(2)→U(1)\mathrm{U}(2) \to\mathrm{U}(1) is trivial, π1(U(2)/U(1))=0\pi_1(\mathrm{U}(2)/\mathrm{U}(1)) = 0. Correspondingly, in the brane picture, the non-BPS D2-brane can shrink to nothing. Then, what is the string breaking process in the brane picture?

First, we show the schematic picture of the end of the broken string straight in the x3x^3 direction in the brane picture in Figure 5. The D2-brane first hangs between the NS5′-brane and the D42_2-brane, which corresponds to the metastable vortex string. However, the energy of the D2-brane can be as small as we want; the smallest energy configuration is when the D2-brane shrinks to nothing. The path from the D2-brane configuration to the configuration with no D2-brane is shown by the gray arrows in the left figure, which corresponds to the path along the direction of the string to its end, as shown in the right figure. The length of the D2-brane first increases as the path goes from the initial configuration, which corresponds to the energy of the string end, the monopole [51, 52]. With this identification of the monopole in the brane picture, we may calculate the mass of the monopole in the brane picture. Obviously, it is impossible to separate the monopole end and the string itself, since they are continuously connected, but we can approximately regard the monopole as the configuration of the D2-brane apart from the vortex string, the D2-brane stretched between the NS5′-brane and the D42_2-brane. Then, when θ\theta is small enough, the mass of the monopole is approximately given by the energy of the D2-brane stretched between the NS5-branes and the D42_2-brane, mmonopole≃τ2v′L6/gsm_{\mathrm{monopole}} \simeq\tau_2 v^\prime L_6/g_s. This indeed matches with the mass of the monopole in the field theory picture in Eq. (17).

Figure 5. A schematic picture of the string end in the brane picture. The configuration of the D4-branes and the NS5-branes is the same as that in Figure 1. The D2-brane, which is shown in magenta, first hangs between the NS5′-brane and the D42_2-brane, which may shrink to nothing as we go along the string direction to the monopole end. In the left figure, we show the configuration of the branes in the x4,…,9x^{4,\ldots,9} directions, where the D2-brane is shown in magenta. The gray arrows show the path along the direction of the string. In the right figure, we project the brane configuration in the (x4,x6,8)(x^4,x^{6,8}) plane and show the x3x^3 direction as well. (t,x3)(t,x^3)-space and show the configuration of the branes at those points. At t=t1t=t_1, the D2-brane is hanging between the NS5′-brane and the D42_2-brane for any x3x^3 position, which corresponds to the straight vortex string. At t=t2t=t_2, the D2-brane shrinks to nothing at some x3x^3 position in the brane picture, which corresponds to the instanton process in the field theory picture. For the D2-brane to shrink to nothing, we need to tunnel through the energy barrier, so that the process in the brane picture is also the instanton process. After the tunneling process, at t=t3t=t_3, the straight string is broken into two pieces, and the monopole-anti-monopole pair is created at the end of both pieces.

With this picture of the string end, the string breaking process can be understood as the tunneling process of the D2-brane from the initial configuration to the configuration with no D2-brane. As an example, we show the schematic figure of the breaking process of a straight string in x3x^3 direction in Figure 6. In the left figure, we show the configuration of the vortex string in the brane picture whereas in the right figure, we show the corresponding configuration of the vortex string in the field theory picture. We take some points in the

Figure 6. A schematic picture of the tunneling process of the string breaking in the brane picture (left) and in the field theory picture (right). In the brane picture, we show the configuration of the D4-branes and the NS5-branes, which is the same as that in Figure 1, together with the D2-brane, which is shown in magenta, for some points in the (t,x3)(t, x^3)-space. The corresponding configuration of the vortex string in the field theory picture is shown on the right, where the vortex string is shown in magenta. In both pictures, at t=t2t=t_2, the instanton process occurs, and the string breaks by creating a monopole-anti-monopole pair. In the right picture, both the instanton process and the monopole-anti-monopole pair are shown by the magenta blobs.

Before discussing how to calculate the bounce action of the metastable vortex string in the brane picture, let us first mention analytical estimates of the decay rate in terms of mmonopolem_{\mathrm{monopole}} and TvortexT_{\mathrm{vortex}}. The decay rate of the metastable vortex string is given by Γ∼e−Sinst\Gamma\sim e^{-S_{\mathrm{inst}}} with SinstS_{\mathrm{inst}} being the action of the instanton corresponding to the tunneling process, which we also call the bounce action. The action of the instanton can be analytically estimated by using the thin-wall approximation [46, 47]. In the thin-wall approximation, we approximate the total action of the system as the sum of the action of the string and the action of the monopole-anti-monopole pair:

S=mmonopole∫ds+Tvortex∫dA,(19)S = m_{\mathrm{monopole}} \int ds + T_{\mathrm{vortex}} \int dA, \tag*{(19)}

where dsds is the line element along the worldline of the monopole, and dAdA is the area element of the worldsheet of the string. For the straight string in the x3x^3 direction, we assume SO(2) symmetry in the (t,x3)(t, x^3)-space. Then, the tunneling action is given as a function of the radius of the “bubble” on the string worldsheet, RR, as

Sinst(R)=2πmmonopoleR−πTvortexR2.(20)S_{\mathrm{inst}}(R) = 2\pi m_{\mathrm{monopole}}R - \pi T_{\mathrm{vortex}}R^2. \tag*{(20)}

Here, we subtract the action of the metastable string. By minimizing Sinst(R)S_{\mathrm{inst}}(R) with respect to RR, we have

Rinst=mmonopoleTvortex,Sinst=πmmonopole2Tvortex.(21)R_{\mathrm{inst}} = \frac{m_{\mathrm{monopole}}}{T_{\mathrm{vortex}}}, \qquad S_{\mathrm{inst}} = \pi\frac{m_{\mathrm{monopole}}^2}{T_{\mathrm{vortex}}}. \tag*{(21)}

Then, the next question is how to calculate SinstS_{\mathrm{inst}} in the brane picture. In the brane picture, SinstS_{\mathrm{inst}} is given by the action of the instanton corresponding to the tunneling process of the D2-brane. The action of the instanton should be calculated from the DBI action of the D2-brane if the motion of the D2-brane corresponding to the tunneling process is given. The DBI action of the D2-brane is given by

SDBI=τ2gs∫d3ξdet⁡∂αXμ∂βXμ(22)S_{\mathrm{DBI}}=\frac{\tau_2}{g_s}\int d^3\xi\sqrt{\det\partial_\alpha X^\mu\partial_\beta X_\mu} \tag*{(22)}

where ξα\xi^\alpha are the worldvolume coordinates of the D2-brane and XμX^\mu are the spacetime coordinates. We have ignored the contribution from the gauge field on the D2-brane and the NS-NS BB-field. We also assume that the spacetime metric is flat and ignore the backreaction of the branes. Note that the DBI action has a gauge symmetry corresponding to the diffeomorphism of the worldvolume coordinates, so that we can choose any gauge to calculate the action. We will take the static gauge later. We then need to find the tunneling path of the D2-brane, whose ends are fixed on the NS5-branes and the D4-branes shown in Figure 6, and calculate the DBI action along the tunneling path.

To see how to find the tunneling path, let us first review the calculation of the decay rate of the false vacuum in quantum field theory [63, 64]. Suppose we have a scalar field ϕ\phi with a potential V(ϕ)V(\phi), which has a local minimum at ϕ=ϕfalse\phi=\phi_{\mathrm{false}} and a global minimum at ϕ=ϕtrue\phi=\phi_{\mathrm{true}}. The instanton corresponding to the tunneling process from the false vacuum to the true vacuum is given by the bounce solution ϕbounce\phi_{\mathrm{bounce}}, which is a solution of the Euclidean equation of motion with appropriate boundary conditions, ϕbounce→ϕfalse\phi_{\mathrm{bounce}}\to\phi_{\mathrm{false}} at ∣x∣→∞\lvert x\rvert\to\infty and ∂μϕbounce=0\partial_\mu\phi_{\mathrm{bounce}}=0 at x=0x=0. Here, we have assumed that the bounce solution is O(4) symmetric, so that the solution depends only on the radial coordinate r=xμxμr=\sqrt{x_\mu x^\mu}. The decay rate of the false vacuum is given by Γ∼e−Sbounce\Gamma\sim e^{-S_{\mathrm{bounce}}} with SbounceS_{\mathrm{bounce}} being the Euclidean action of the bounce solution. The simplest way to find the bounce solution is so-called overshoot/undershoot method. We first choose some initial value of the field at r=0r=0, ϕbounce(0)=ϕ0\phi_{\mathrm{bounce}}(0)=\phi_0, and solve the Euclidean equation of motion with the boundary condition ∂μϕbounce=0\partial_\mu\phi_{\mathrm{bounce}}=0 at r=0r=0. If ϕ0\phi_0 is sufficiently close to ϕtrue\phi_{\mathrm{true}}, the solution overshoots the false vacuum at some finite rr, whereas if ϕ0\phi_0 is sufficiently far from ϕtrue\phi_{\mathrm{true}}, the solution undershoots the false vacuum and starts to oscillate at some finite rr. By tuning ϕ0\phi_0, we can find the bounce solution, which approaches the false vacuum at r→∞r\to\infty. We schematically depict this solution in Figure 7.

Figure 7. A schematic picture of the bounce solution for the false vacuum decay of a scalar field ϕ\phi. In the left figure, we show the O(4)-symmetric bounce configuration on the Euclidean spacetime. The field takes the value ϕ=ϕ0\phi= \phi_0, close to the true vacuum, at the center (black dot), and approaches the false vacuum, ϕ≈ϕfalse\phi\approx\phi_{\mathrm{false}}, far from the center, through the wall region where the field varies rapidly. In the right figure, we show the profile of the bounce solution (thick curve), which starts at ϕ0\phi_0 with ∂rϕ=0\partial_r\phi= 0 and asymptotes to ϕfalse\phi_{\mathrm{false}} (dashed line). The thin curves show solutions with slightly different initial values, which overshoot or undershoot the false vacuum.

We try to apply these methods to find the tunneling path of the D2-brane. First, we perform the Wick rotation to the DBI action, t→−iτt\to-i\tau, to rewrite the action in the Euclidean signature. Then, just like the O(4) symmetry in the bounce solution, we may assume that the tunneling path of the D2-brane corresponding to the straight vortex string in the x3x^3 direction is O(2) symmetric in the (τ,x3)(\tau, x^3)-space. We rewrite the coordinates as τ=ρcos⁡α\tau= \rho\cos\alpha and x3=ρsin⁡αx^3 = \rho\sin\alpha with ρ=τ2+(x3)2\rho= \sqrt{\tau^2 + (x^3)^2} and α=arctan⁡(x3/τ)\alpha= \arctan(x^3/\tau), and assume that the tunneling path of the D2-brane is independent of α\alpha. We next take the static gauge, ξ0=τ\xi^0 = \tau and ξ1=x3\xi^1 = x^3. The final coordinate, ξ2≡s\xi^2 \equiv s, is defined as the coordinate along the direction of the string; without loss of generality, we can take s∈[0,1]s \in[0,1] and s=0s = 0 and s=1s = 1 correspond to the ends of the D2-brane. With these simplifications, the DBI action is reduced to

SDBI=2πτ2gs∫0∞dρ∫01ds[1+(∂X⃗∂ρ)2](∂X⃗∂s)2−(∂X⃗∂ρ⋅∂X⃗∂s)2(23)S_{\mathrm{DBI}} = \frac{2\pi\tau_2}{g_s} \int_0^\infty d\rho\int_0^1 ds \sqrt{\left[1+\left(\frac{\partial\vec{X}}{\partial\rho}\right)^2\right]\left(\frac{\partial\vec{X}}{\partial s}\right)^2-\left(\frac{\partial\vec{X}}{\partial\rho}\cdot\frac{\partial\vec{X}}{\partial s}\right)^2} \tag*{(23)}

where X⃗(ρ,s)\vec{X}(\rho,s) are the coordinates of the D2-brane in the directions transverse to the (τ,x3)(\tau,x^3)-space.

One may then try to find the bounce solution of the DBI action, which approaches the configuration of the D2-brane corresponding to the straight vortex string at ρ→∞\rho\to\infty, starting from some initial configuration sufficiently close to the true vacuum at ρ=0\rho= 0. However, this does not work, as the true vacuum configuration corresponds to the configuration with no D2-brane and cannot be described by the DBI action of the D2-brane. The true stable vacuum should consist of some stringy excitations in the brane background. Therefore, instead of changing the initial field value at r=0r=0 in the vacuum decay calculation, we claim that the bounce solution suddenly starts to exist at some finite ρ=ρ0\rho=\rho_0, with a short enough D2-brane. The change of the initial field value at r=0r=0 corresponds to not only the change of the initial configuration of the D2-brane such as its velocity and its shape, but also the change of ρ0\rho_0, the time when the D2-brane suddenly appears (See Figure 8 left panel). The overshoot/undershoot method in the brane picture is then to tune the initial configuration of the D2-brane and ρ0\rho_0 to find the solution of the Euclidean equation of motion of the DBI action, which approaches the configuration of the D2-brane corresponding to the straight vortex string at ρ→∞\rho\to\infty.

Schematic picture of the conjectured bounce solution for the decay of the metastable vortex string

Figure 8. A schematic picture of the conjectured bounce solution for the decay of the metastable vortex string, to be compared with Figure 7. Left: the (τ,x3)(\tau, x^3)-plane. In contrast to the scalar field case, no D2-brane exists for ρ<ρ0\rho< \rho_0; the D2-brane suddenly starts to exist at ρ=ρ0\rho= \rho_0 (dashed circle), sweeps the configurations in the lower left inset across the wall region, and approaches the metastable vortex string configuration (lower right inset) as ρ→∞\rho\to\infty. Right: a schematic profile of the endpoint parameters u0(ρ)u_0(\rho) (blue) and u1(ρ)u_1(\rho) (red), which asymptote to the false vacuum values (dashed lines); see Figure 10 for the numerical solutions.

Before constructing the bounce solution explicitly, we derive its action within the thin-wall approximation in this brane framework, by assuming that in the bounce solution the variation of D2-brane configuration sharply concentrates around ρ∼ρ0\rho\sim\rho_0 for some ρ0\rho_0. This means that the bounce action behaves as

SDBI≃τ2gs(2πρ0Asweep−πρ02v′sin⁡θ),(24)S_{\mathrm{DBI}}\simeq\frac{\tau_2}{g_s}\left(2\pi\rho_0 A_{\mathrm{sweep}}-\pi\rho_0^2v'\sin\theta\right), \tag*{(24)}

where AsweepA_{\mathrm{sweep}} is the area of the surface the D2 brane sweeps in the space spanned by x4,5,6,7,8,9x^{4,5,6,7,8,9} coordinates during the decay. The first term is the worldvolume swept by the D2-brane along the circle ρ=ρ0\rho=\rho_0 in the (τ,x3)(\tau,x^3)-space, whose circumference is 2πρ02\pi\rho_0. The second term is from the loss of action due to the absence of the D2-brane in the region ρ≤ρ0\rho\leq\rho_0, whose area is πρ02\pi\rho_0^2, the D2-brane there being of width v′sin⁡θv'\sin\theta. While the area AsweepA_{\mathrm{sweep}} should be obtained by minimizing the area among the potential trajectory of the D2 brane (in [52], such minimal surfaces give the monopole masses in a different brane configuration), when θ\theta is small we can approximate the minimal area surface by the rectangle formed by NS5 and D4 branes projected onto x4,5,6x^{4,5,6} directions, whose area is L6v′L_6v'. Thus in this approximation we have

SDBI(ρ0)≃τ2gs(2πρ0L6v′−πρ02v′sin⁡θ).(25)S_{\mathrm{DBI}}(\rho_0)\simeq\frac{\tau_2}{g_s}\left(2\pi\rho_0L_6v'-\pi\rho_0^2v'\sin\theta\right). \tag*{(25)}

Comparing Eq. (24) with the field theory result of Eq. (20), we see that the two expressions agree term by term, once we identify

mmonopole=τ2gsAsweep,Tvortex=τ2gsv′sin⁡θ=T3.(26)m_{\mathrm{monopole}}=\frac{\tau_2}{g_s}A_{\mathrm{sweep}},\qquad T_{\mathrm{vortex}}=\frac{\tau_2}{g_s}v'\sin\theta=T_3. \tag*{(26)}

The identification of the string tension is nothing but Eq. (18), which we have already checked against the explicit vortex string solution in Figure 4. As for the monopole mass, using the small θ\theta approximation Asweep≃L6v′A_{\mathrm{sweep}} \simeq L_{6}v' together with Eq. (3), we obtain

mmonopole≃τ2v′L6gs=2v′g2ℓs2=4πmWg2,(27)m_{\mathrm{monopole}} \simeq\frac{\tau_{2}v'L_{6}}{g_{s}} = \frac{2v'}{g^{2}\ell_{s}^{2}} = \frac{4\pi m_{W}}{g^{2}}, \tag*{(27)}

which is the mass of the monopole in the field theory picture quoted in Eq. (17). In other words, the thin-wall approximation in the brane picture reproduces the field theory one of Eqs. (20) and (21).

The radius of the bubble and the bounce action in this approximation are then given by

ρ0=Asweepv′sin⁡θ≃L6sin⁡θ,Sinst≃τ2L62v′gssin⁡θ=2πv′L6g2ℓs2sin⁡θ.(28)\rho_{0} = \frac{A_{\mathrm{sweep}}}{v'\sin\theta} \simeq\frac{L_{6}}{\sin\theta}, \qquad S_{\mathrm{inst}} \simeq\frac{\tau_{2}L_{6}^{2}v'}{g_{s}\sin\theta} = \frac{2\pi v'L_{6}}{g^{2}\ell_{s}^{2}\sin\theta}. \tag*{(28)}

Note that the string scale ℓs\ell_{s} and the string coupling constant gsg_{s} enter these expressions only through the gauge coupling constant gg. Indeed, using μ=tan⁡θ/L6\mu= \tan\theta/L_{6} and taking θ\theta small, we have ρ0≃1/μ\rho_{0} \simeq1/\mu and Sinst≃2πv′/(g2ℓs2μ)S_{\mathrm{inst}} \simeq2\pi v'/(g^{2}\ell_{s}^{2}\mu), which are written purely in terms of the field theory quantities. This is the analytic counterpart of the invariance of the bounce action under the scalings of the brane parameters, which we check numerically in Sec. IV. The validity of the thin-wall approximation itself is discussed in Sec. IV, where the numerical solutions let us estimate the thickness of the “wall”.

Numerical calculation and results

As a consistency check of the thin-wall result, we now construct the bounce solution numerically, based on the conjecture of the bounce solution in the previous section. To simplify the calculation, we make one key assumption: the D2-brane remains straight in the space of x4,5,6,7,8,9x^{4,5,6,7,8,9}, with its shape entirely determined by the positions of its endpoints. In principle, the D2-brane can be curved, and its shape is determined by solving the equation of motion from the DBI action. However, we expect this straight-brane approximation to be valid in the limit of small θ\theta. The endpoints of the D2-brane are fixed on the NS5-branes and the D4-branes, so that they can be described by one-parameter functions: X⃗=X⃗E(u)\vec{X} = \vec{X}_{E}(u). We normalize the parameter uu such that dX⃗E/dud\vec{X}_{E}/du is a unit vector along the branes, with X⃗E(0)\vec{X}_{E}(0) at the intersection of the NS5-brane and D41_1-brane, X⃗E(v′)\vec{X}_{E}(v') at the intersection of the NS5-brane and D42_2-brane, X⃗E(−L6)\vec{X}_{E}(-L_{6}) at the intersection of the NS5′-brane and D41_1-brane, and so forth. The kinematic degrees of freedom of the D2-brane are then the endpoint parameters as functions of ρ\rho, which we denote as u0(ρ)u_0(\rho) and −u1(ρ)-u_1(\rho), with both u0u_0 and u1u_1 non-negative at the initial configuration. For given (ρ,s)(\rho,s), the field X⃗(ρ,s)\vec{X}(\rho,s) is given by the linear interpolation of the endpoints, X⃗(ρ,s)=(1−s)X⃗E(u0(ρ))+sX⃗E(−u1(ρ))\vec{X}(\rho,s)=(1-s)\vec{X}_{E}(u_0(\rho))+s\vec{X}_{E}(-u_1(\rho)). We note that this parameterization of the D2-brane is non-differentiable at the edges, the intersection of a NS5-brane and a D4-brane. In the numerical calculation, we smooth the edge by a small smoothing parameter ϵ\epsilon; instead of rectangular corners, we use quarter circles with radius ϵ\epsilon at the edges. ϵ\epsilon is taken to be much smaller than any other length scale in the problem, and we check that the results do not depend on ϵ\epsilon.

With this straight-brane approximation, the DBI action is reduced to a functional of u0(ρ)u_0(\rho) and u1(ρ)u_1(\rho) as

SDBI=2πτ2gs∫0∞dρ L,L=ρL(u0,u1)I(u0,u1,u˙0,u˙1),(29)S_{\mathrm{DBI}}=\frac{2\pi\tau_2}{g_s}\int_0^\infty d\rho\,\mathcal{L},\qquad\mathcal{L}=\rho L(u_0,u_1)I(u_0,u_1,\dot{u}_0,\dot{u}_1), \tag*{(29)}

where

L⃗≡X⃗E(−u1)−X⃗E(u0),\vec{L}\equiv\vec{X}_{E}(-u_1)-\vec{X}_{E}(u_0),
I≡∫01ds1+M2−(L^⋅M⃗)2,I\equiv\int_0^1 ds\sqrt{1+M^2-(\hat{L}\cdot\vec{M})^2},
M⃗≡∂X⃗∂ρ=u˙0(1−s)v⃗0+u˙1sv⃗1\vec{M}\equiv\frac{\partial\vec{X}}{\partial\rho}=\dot{u}_0(1-s)\vec{v}_0+\dot{u}_1s\vec{v}_1
v⃗0≡∂X⃗E(u0)∂u0,v⃗1≡∂X⃗E(−u1)∂u1,(30)\vec{v}_0\equiv\frac{\partial\vec{X}_{E}(u_0)}{\partial u_0},\qquad\vec{v}_1\equiv\frac{\partial\vec{X}_{E}(-u_1)}{\partial u_1}, \tag*{(30)}

with the dot denoting the derivative with respect to ρ\rho, A^≡A⃗/A\hat{A}\equiv\vec{A}/A and A=∣A⃗∣A=|\vec{A}| for any vector A⃗\vec{A}. The equations of motion for u0(ρ)u_0(\rho) and u1(ρ)u_1(\rho) are given by

I(u¨0u¨1)=(F0F1),(31)\mathcal{I} \begin{pmatrix} \ddot{u}_0\\ \ddot{u}_1 \end{pmatrix} = \begin{pmatrix} \mathcal{F}_0\\ \mathcal{F}_1 \end{pmatrix}, \tag*{(31)}

where

I≡(∂2∂u˙02∂2∂u˙0∂u˙1∂2∂u˙1∂u˙0∂2∂u˙12)I,\mathcal{I}\equiv \begin{pmatrix} \frac{\partial^2}{\partial\dot{u}_0^2} & \frac{\partial^2}{\partial\dot{u}_0\partial\dot{u}_1}\\ \frac{\partial^2}{\partial\dot{u}_1\partial\dot{u}_0} & \frac{\partial^2}{\partial\dot{u}_1^2} \end{pmatrix}I,
Fi≡1L∂L∂uiI+∂I∂ui−∂2I∂uj∂u˙iu˙j−1ρ∂I∂u˙i−∂L∂uj∂I∂u˙iu˙jL,(32)\mathcal{F}_i\equiv\frac{1}{L}\frac{\partial L}{\partial u_i}I+\frac{\partial I}{\partial u_i}-\frac{\partial^2 I}{\partial u_j\partial\dot{u}_i}\dot{u}_j-\frac{1}{\rho}\frac{\partial I}{\partial\dot{u}_i}-\frac{\partial L}{\partial u_j}\frac{\partial I}{\partial\dot{u}_i}\frac{\dot{u}_j}{L}, \tag*{(32)}

with i,j=0,1i,j=0,1 and the summation over jj is implied. From this equation of motion, we can explicitly check the observation in the previous section that the DBI action fails to describe the motion of the D2-brane near the true vacuum, as I\mathcal{I} vanishes when both endpoints of the D2-brane are on the same brane, e.g., u0<v′u_0 < v' and u1=0u_1 = 0. In the actual numerical calculation, we may avoid this singularity due to the smoothing of the edge.

Now, we would like to find the bounce solution of the above equation of motion with the overshoot/undershoot method. We need to tune the initial configuration of the D2-brane and ρ0\rho_0, the time when the D2-brane suddenly appears, to find the solution which approaches the configuration of the D2-brane corresponding to the straight vortex string at ρ→∞\rho\to\infty. Then, what is the initial configuration of the D2-brane and which parameters describe the initial configuration? As we mentioned above, the two endpoints of the D2-brane cannot be on the same brane at the initial configuration in our approximation, and we assume that this reflects the fact that the state there is so close to the true vacuum that the DBI action cannot describe it. Therefore, we assume that the initial configuration of the D2-brane is a straight line with length LinitialL_{\mathrm{initial}} whose endpoints are both on the NS5-brane or both on the D41_1-brane; i.e., u0=Linitialu_0 = L_{\mathrm{initial}} and u1=0u_1 = 0 or u0=0u_0 = 0 and u1=Linitialu_1 = L_{\mathrm{initial}}, as schematically shown in Figure 8. The initial configuration of the D2-brane is then described by four parameters, LinitialL_{\mathrm{initial}}, u˙0\dot{u}_0 and u˙1\dot{u}_1 at ρ=ρ0\rho=\rho_0.

To further simplify the calculation, we assume that the motion of the D2-brane experiences a point where F0=F1=0\mathcal{F}_0=\mathcal{F}_1=0 at some ρ=ρtop\rho=\rho_{\mathrm{top}}. We also assume that at ρ=ρtop\rho=\rho_{\mathrm{top}}, the endpoints of the D2-brane, XE(u0)X_{\mathrm{E}}(u_0) and XE(−u1)X_{\mathrm{E}}(-u_1), are on the NS5-brane and the D41_1-brane, respectively. This is motivated by the observation that the bounce solution of the vacuum decay in quantum field theory experiences a point where the potential gradient vanishes, ∂V/∂ϕ=0\partial V/\partial\phi=0, at some rr when ϕbounce\phi_{\mathrm{bounce}} is between ϕtrue\phi_{\mathrm{true}} and ϕfalse\phi_{\mathrm{false}}. We, however, admit that this is just an assumption and we do not have a strong reason to believe that this is the case; for example, we ignore the friction term that may well exist in the equation of motion of the D2-brane just like the friction term in the field theory case, although we expect that the friction term is not important for large ρtop\rho_{\mathrm{top}}. Also, even if our first assumption is the case, the endpoints of the D2-brane may be on the different branes at ρ=ρtop\rho=\rho_{\mathrm{top}}, e.g., on NS5-brane and NS5′-brane, instead of on the NS5-brane and the D41_1-brane. This second assumption is just for numerical convenience as we discuss later, and indeed, we observe that u0(ρtop)→vu_0(\rho_{\mathrm{top}})\to v or u1(ρtop)→L6u_1(\rho_{\mathrm{top}})\to L_6 as we take smaller θ\theta, and thus the second assumption might be violated in the small θ\theta limit. Having said that, as we will see later, we can find the bounce solution with the above assumptions, and we also find that without the above assumptions, we cannot find the bounce solution, which supports the validity of the first assumption to some extent.

With the above assumptions, the initial condition is reduced to three parameters: u0(ρtop)u_0(\rho_{\mathrm{top}}), u1(ρtop)u_1(\rho_{\mathrm{top}}), and ρtop\rho_{\mathrm{top}}. The velocities u˙0(ρtop)\dot{u}_0(\rho_{\mathrm{top}}) and u˙1(ρtop)\dot{u}_1(\rho_{\mathrm{top}}) are determined by the condition F0=F1=0\mathcal{F}_0=\mathcal{F}_1=0. We compute them numerically using the following procedure.

  1. Symmetric case: For simplicity, we first consider the symmetric case, where u0(ρtop)=u1(ρtop)≡usymu_0(\rho_{\mathrm{top}})=u_1(\rho_{\mathrm{top}})\equiv u_{\mathrm{sym}} and we fix usymu_{\mathrm{sym}}. By symmetry, we have u˙0=u˙1≡u˙sym\dot{u}_0=\dot{u}_1\equiv\dot{u}_{\mathrm{sym}} and F0=F1≡Fsym\mathcal{F}_0=\mathcal{F}_1\equiv\mathcal{F}_{\mathrm{sym}}. Since Fsym\mathcal{F}_{\mathrm{sym}} depends only on u˙sym\dot{u}_{\mathrm{sym}}, we find the value satisfying Fsym=0\mathcal{F}_{\mathrm{sym}}=0 using the bisection method.

  1. General case: Next, we extend to the general case u0(ρtop)≠u1(ρtop)u_0(\rho_{\mathrm{top}})\ne u_1(\rho_{\mathrm{top}}). In this case, we parametrize the configuration by uoffset=min⁡(u0,u1)u_{\mathrm{offset}}=\min(u_0,u_1) and udiff=u0−u1u_{\mathrm{diff}}=u_0-u_1. For fixed uoffsetu_{\mathrm{offset}}, we discretize udiffu_{\mathrm{diff}} into NdiffN_{\mathrm{diff}} steps and compute u˙0(ρtop)\dot{u}_0(\rho_{\mathrm{top}}) and u˙1(ρtop)\dot{u}_1(\rho_{\mathrm{top}}) for each step. At each step nn, we minimize ∣I−1F∣2\lvert\mathcal{I}^{-1}\mathcal{F}\rvert^2 using the Nelder-Mead method.

For the initial simplex in the Nelder-Mead algorithm, we use:

(u˙0∣n=k−1+δ0,u˙1∣n=k−1),(u˙0∣n=k−1,u˙1∣n=k−1+δ1),(u˙0∣n=k−1+δ0,u˙1∣n=k−1+δ1),(33)\begin{aligned} &(\dot{u}_0|_{n=k-1}+\delta_0,\dot{u}_1|_{n=k-1}),\\ &(\dot{u}_0|_{n=k-1},\dot{u}_1|_{n=k-1}+\delta_1),\\ &(\dot{u}_0|_{n=k-1}+\delta_0,\dot{u}_1|_{n=k-1}+\delta_1), \tag*{(33)} \end{aligned}

where δi=u˙i∣n=k−1−u˙i∣n=k−2\delta_i=\dot{u}_i|_{n=k-1}-\dot{u}_i|_{n=k-2} for k>2k>2. For the first step (k=1k=1), we use:

(u˙sym,u˙sym),(1.00025u˙sym,u˙sym),(u˙sym,1.00025u˙sym),(34)\begin{aligned} &(\dot{u}_{\mathrm{sym}},\dot{u}_{\mathrm{sym}}),\\ &(1.00025\dot{u}_{\mathrm{sym}},\dot{u}_{\mathrm{sym}}),\\ &(\dot{u}_{\mathrm{sym}},1.00025\dot{u}_{\mathrm{sym}}), \tag*{(34)} \end{aligned}

where u˙sym\dot{u}_{\mathrm{sym}} is obtained from the symmetric case 22. We verify that ∣I−1F∣2|\mathcal{I}^{-1}\mathcal{F}|^2 is sufficiently small for each step. Finally, we interpolate the computed values to obtain u˙0\dot{u}_0 and u˙1\dot{u}_1 for arbitrary u0(ρtop)u_0(\rho_{\mathrm{top}}) and u1(ρtop)u_1(\rho_{\mathrm{top}}).

Our task is then to tune uoffsetu_{\mathrm{offset}}, udiffu_{\mathrm{diff}}, and ρtop\rho_{\mathrm{top}} to find the bounce solution using a nested bisection approach. The algorithm proceeds as follows:

  1. Bisection over udiffu_{\mathrm{diff}} for fixed ρtop\rho_{\mathrm{top}} and uoffsetu_{\mathrm{offset}}: We first fix uoffsetu_{\mathrm{offset}} and ρtop≡ρtrial\rho_{\mathrm{top}}\equiv\rho_{\mathrm{trial}}. The fate of the D2-brane depends on udiff=u0−u1u_{\mathrm{diff}}=u_0-u_1:

  • When udiffu_{\mathrm{diff}} is very large and positive, the D2-brane is nearly parallel to the NS5-brane, and the solution overshoots the false vacuum as XE(u0)X_{\mathrm{E}}(u_0) goes beyond the endpoint of the target false vacuum configuration.

  • When udiffu_{\mathrm{diff}} is very large and negative, the D2-brane is nearly parallel to the D4-brane, and again overshoots as XE(−u1)X_{\mathrm{E}}(-u_1) goes beyond the endpoint.

  • For intermediate values of udiffu_{\mathrm{diff}}, the solution may either simultaneously overshoot (both endpoints together approach and pass the false vacuum endpoints) or undershoot (one endpoint turns back before reaching the false vacuum endpoint).

By bisecting over udiffu_{\mathrm{diff}}, we find the value where the solution transitions between one-sided overshooting and simultaneous overshooting/undershooting regimes, which corresponds to the most-bounce-like solution for that particular ρtrial\rho_{\mathrm{trial}}. Note that the optimal udiffu_{\mathrm{diff}} for the bounce solution is dependent on ρtrial\rho_{\mathrm{trial}} in general, since the initial velocities and the equation of motion depend explicitly on ρ\rho.

  1. Bisection over ρtop\rho_{\mathrm{top}}: Once we have found the most-bounce-like solution for a given ρtrial\rho_{\mathrm{trial}} by tuning udiffu_{\mathrm{diff}}, we determine whether this solution simultaneously overshoots or undershoots the false vacuum asymptotically. We assume that small ρtop\rho_{\mathrm{top}} (weak friction) leads to undershooting, while large ρtop\rho_{\mathrm{top}} (strong friction) leads to overshooting, analogous to the vacuum decay problem in field theory. By bisecting over ρtop\rho_{\mathrm{top}}, we find the value where the solution transitions from undershooting to overshooting, which identifies the bounce solution. Since the optimal udiffu_{\mathrm{diff}} depends on ρtrial\rho_{\mathrm{trial}}, we repeat the bisection over udiffu_{\mathrm{diff}} for each ρtrial\rho_{\mathrm{trial}} during this step.

  1. Search over uoffsetu_{\mathrm{offset}}: Finally, we repeat the above nested bisection for different values of uoffsetu_{\mathrm{offset}}. We observe that solutions exist and approach the false vacuum configuration for a wide range of uoffsetu_{\mathrm{offset}} values, but the action varies with uoffsetu_{\mathrm{offset}}. We interpret this as a numerical artifact due to finite precision: ideally, only one specific value of uoffsetu_{\mathrm{offset}} corresponds to the true bounce solution, because if we took the initial configuration of the D2-brane at ρ→∞\rho\to\infty, we would find that only one specific value of uoffsetu_{\mathrm{offset}} leads to the solution that approaches the false vacuum configuration at ρ→∞\rho\to\infty, but simply our numerics cannot resolve this. Therefore, we assume that the bounce solution is characterized by the maximum bounce action value across all uoffsetu_{\mathrm{offset}} choices, as this most closely resembles the vacuum decay instanton in field theory. Note that, to calculate the bounce action, we solve the equation of motion backwards from ρ=ρtop\rho= \rho_{\mathrm{top}} to ρ=ρ0\rho= \rho_0, where either u0u_0 or u1u_1 reaches zero, and then compute the action by integrating from ρ=ρ0\rho= \rho_0 to such ρ\rho that the solution sufficiently approaches the false vacuum configuration.

To illustrate how the above algorithm works, we perform the numerical calculation for some parameters, L6=6L_6 = 6, v′=25v' = 25, ℓs=20\ell_s = 20, θ=0.1\theta= 0.1. This choice of parameters satisfies the conditions, Eq. (14), and Eq. (16). Also, we may take, for example, gs=10−3g_s = 10^{-3} so that g∼0.5g \sim0.5. According to our algorithm, we find uoffset=1.14u_{\mathrm{offset}} = 1.14, udiff=−0.626u_{\mathrm{diff}} = -0.626, and ρtop=57.0\rho_{\mathrm{top}} = 57.0 gives the bounce solution. The bounce action for this solution is Sinst=90.2S_{\mathrm{inst}} = 90.2 for gs=10−3g_s = 10^{-3}. We note that this value is close to the thin-wall approximation, which gives Rinst∼60R_{\mathrm{inst}} \sim60 and Sinst∼89.5S_{\mathrm{inst}} \sim89.5 for the above parameters.

In Figure 9, we plot Δ\Delta (to be defined shortly) as a function of ρtop\rho_{\mathrm{top}} for different values of udiffu_{\mathrm{diff}} with uoffset=1.14u_{\mathrm{offset}} = 1.14. The red, green, orange, and blue curves correspond to udiff=−0.55u_{\mathrm{diff}} = -0.55, −0.626-0.626, −0.7-0.7, and −0.8-0.8, respectively. The value Δ\Delta is defined as the difference between either u0u_0 or u1u_1 and the corresponding endpoint of the false vacuum configuration when the other endpoint approaches the false vacuum endpoint. The sign of Δ\Delta is defined such that Δ\Delta is positive when u0u_0 is at the false vacuum endpoint and Δ\Delta is negative when u1u_1 is at the false vacuum endpoint. For example, if u0u_0 is at the intersection of the NS5-brane and the D424_2-brane when u1u_1 is at the false vacuum endpoint, Δ\Delta is −L6-L_6. Another example is that when the simultaneous overshooting occurs, Δ=0\Delta= 0. If the solution is in the undershooting regime, Δ\Delta is defined not to have any value.

Figure 9. Δ\Delta as a function of ρtop\rho_{\mathrm{top}} for different values of udiffu_{\mathrm{diff}} with uoffset=1.14u_{\mathrm{offset}} = 1.14. The red, green, orange, and blue curves correspond to udiff=−0.55u_{\mathrm{diff}} = -0.55, −0.626-0.626, −0.7-0.7, and −0.8-0.8, respectively. When the solution is in the undershooting regime, Δ\Delta does not have any value and the curve is therefore interrupted.

From Figure 9, we can see how Step 1 and Step 2 of our algorithm works. First, let us fix ρtop\rho_{\mathrm{top}} to some smaller value, e.g., ρtop=53.6\rho_{\mathrm{top}} = 53.6. For this ρtop\rho_{\mathrm{top}}, we perform Step 1, the bisection over udiffu_{\mathrm{diff}}. For a smaller udiffu_{\mathrm{diff}}, e.g., udiff=−0.8u_{\mathrm{diff}} = -0.8, Δ\Delta is negative. On the other hand, for a larger udiffu_{\mathrm{diff}}, e.g., udiff=−0.55u_{\mathrm{diff}} = -0.55, Δ\Delta is positive. We can find the value of udiffu_{\mathrm{diff}} where Δ\Delta changes its sign by bisection. However, as the orange curve for udiff=−0.7u_{\mathrm{diff}} = -0.7 shows, in the intermediate value of udiffu_{\mathrm{diff}}, the solution undershoots the false vacuum. We therefore conclude that the solution with ρtop=53.6\rho_{\mathrm{top}} = 53.6 is in the undershooting regime in Step 2. Next, we are back to Step 1 and perform the bisection over udiffu_{\mathrm{diff}} for a larger ρtop\rho_{\mathrm{top}}, e.g., ρtop=57.0\rho_{\mathrm{top}} = 57.0. For this ρtop\rho_{\mathrm{top}}, we similarly find the value of udiffu_{\mathrm{diff}} where Δ\Delta changes its sign by bisection. Then, as the green curve for udiff=−0.626u_{\mathrm{diff}} = -0.626 shows, the solution with ρtop=57.0\rho_{\mathrm{top}} = 57.0 is in the simultaneous overshooting regime in Step 2. We repeat the above procedure to find the value of ρtop\rho_{\mathrm{top}} where the solution transitions from undershooting to overshooting, which gives the bounce solution. With the above parameters, we find ρtop=57.0\rho_{\mathrm{top}} = 57.0 gives the bounce solution.

To show that Step 3 is necessary, we plot the most-bounce-like solutions for uoffset=1.14u_{\mathrm{offset}} = 1.14 and uoffset=0.3u_{\mathrm{offset}} = 0.3 in Figure 10. The left panel shows the solution for uoffset=1.14u_{\mathrm{offset}} = 1.14 and the right panel shows the solution for uoffset=0.3u_{\mathrm{offset}} = 0.3. We identify the former as the bounce solution, as the bounce action for the former, Sinst=90.2S_{\mathrm{inst}} = 90.2, is larger than that for the latter, Sinst=86.4S_{\mathrm{inst}} = 86.4. However, as we can see from Figure 10, both solutions approach the false vacuum configuration at large ρ\rho, seemingly asymptotically. Within our numerical precision, we cannot determine which solution is the true bounce solution, and thus we assume that the solution with the larger bounce action is the true bounce solution.

Figure 10. The most-bounce-like solutions for uoffset=1.14u_{\mathrm{offset}} = 1.14 (left) and uoffset=0.30u_{\mathrm{offset}} = 0.30 (right). Solid curves show the endpoint positions: blue for ∣u0(ρ)∣|u_0(\rho)| and red for ∣u1(ρ)∣|u_1(\rho)|. Dashed curves show the endpoint velocities: cyan for ∣u˙0(ρ)∣/∣u˙0(ρtop)∣|\dot{u}_0(\rho)|/|\dot{u}_0(\rho_{\mathrm{top}})| and pink for ∣u˙1(ρ)∣/∣u˙1(ρtop)∣|\dot{u}_1(\rho)|/|\dot{u}_1(\rho_{\mathrm{top}})|.

We admit that some numbers in the above calculation are close to or even smaller than the string length ℓs\ell_s; heavier degrees of freedom such as the stringy excitations may be relevant in this case. Nonetheless, the resulting bounce action is close to the thin-wall result, which suggests that our algorithm captures the decay process at least where the thin-wall approximation is valid.

To perform further non-trivial checks of our algorithm, we have also checked the non-trivial scaling of the bounce action with respect to the parameters in the brane picture. That is, different choices of parameters in the brane picture may give the same parameters in the field theory, and thus they should give the same bounce action. In addition to the trivial scaling, which scales all the length scales in the brane picture by the same factor, we have two types of scaling:

  1. Scaling of gsg_s:

L6→aL6,gs→ags,θ→aθL_6 \to aL_6,\quad g_s \to ag_s,\quad\theta\to a\theta

with aa being a scaling factor.

  1. Simultaneous scaling of gsg_s and ℓs\ell_s:

ℓs→bℓs,gs→b−1gs,v′→b2v′\ell_s \to b\ell_s,\quad g_s \to b^{-1}g_s,\quad v' \to b^2v'

with bb being a scaling factor.

For small θ\theta limit, these scalings does not change the parameters in the field theory, and thus the bounce action should be invariant under these scalings. In Figure 11, we plot the bounce action in terms of θ\theta for different values of the scaling factors aa and bb. Again, we take L6=6L_6 = 6, v′=25v' = 25, ℓs=20\ell_s = 20, gs=10−3g_s = 10^{-3} as the reference parameters for a=b=1a = b = 1. For the scaling of aa, we scale back θ\theta by aa to show the bounce action as a function of the original θ\theta for different values of aa. As a reference, we also plot the bounce action from the thin-wall approximation; for this calculation, we have used the string tension and monopole mass from the brane picture, Tvortex=τ2v′sin⁡θ/gsT_{\mathrm{vortex}} = \tau_2v'\sin\theta/g_s and mmonopole=τ2v′L6/gsm_{\mathrm{monopole}} = \tau_2v'L_6/g_s, respectively. We see that all the bounce action curves are sufficiently close to the thin-wall approximation, and the bounce action is indeed invariant under the above scalings.

Figure 11. The bounce action for different values of the scaling factors aa (left) and bb (right). We take L6=6L_6 = 6, v′=25v' = 25, ℓs=20\ell_s = 20, gs=10−3g_s = 10^{-3} as the reference parameters for a=b=1a = b = 1 and plot the bounce action with varying θ\theta for different values of aa and bb. The correspondence between the scaling factors and the style of the curves are shown in the legends. The bounce action from the thin-wall approximation is also shown as the orange dot for some θ\theta values for reference. For the left figure, we scale back θ\theta by aa to show the bounce action as a function of the original θ\theta for different values of aa.

Before turning to a wider range of θ\theta, let us examine the validity of the thin-wall approximation itself, which the solutions of Figure 10 allow us to check directly. In the brane picture, the approximation requires the thickness of the “wall”, the range Δρ\Delta\rho over which the D2-brane configuration deviates appreciably from the metastable one, to be much smaller than the radius of the bubble. This thickness is not fixed by the thin-wall approximation. itself; in the wall region the integrand of Eq. (23) reduces to ρ∣∂ρX⃗∧∂sX⃗∣\rho\lvert\partial_\rho\vec{X} \wedge\partial_s \vec{X} \rvert, the density of the area swept by the D2-brane, which is invariant under reparameterizations of ρ\rho, so that Δρ\Delta\rho is fixed only by the subleading terms. In both panels of Figure 10 the D2-brane reaches the metastable configuration within Δρ≃6\Delta\rho\simeq6 of its appearance, while the two solutions have different bubble radii, ρtop=57.0\rho_{\mathrm{top}} = 57.0 and 46.446.4. The wall is thus thinner than the bubble by an order of magnitude, and the thin-wall approximation is self-consistent in the brane picture at θ=0.1\theta= 0.1.

The same is not true in the field theory picture, where the approximation further requires the bubble to be larger than the core sizes of the monopole and of the vortex string, 1/mW=2πℓs2/v′≃1001/m_W = 2\pi\ell_s^2/v' \simeq100 and 1/(gq)≃781/(gq) \simeq78, respectively. Both exceed the bubble radius of Eq. (28), Rinst=L6/sin⁡θ≃60R_{\mathrm{inst}} = L_6/\sin\theta\simeq60, for the parameters used here. Equivalently, the monopole core gives the stronger condition, sin⁡θ≪L6v′/(2πℓs2)≃0.06\sin\theta\ll L_6v'/(2\pi\ell_s^2) \simeq0.06. The thin-wall approximation in the field theory picture is therefore already marginal at θ=0.1\theta= 0.1, whereas the one in the brane picture is not.

Finally, in Figure 12, we plot the bounce action for a wider range of θ\theta to see if the numerical bounce action deviates from the thin-wall approximation for larger θ\theta. We take L6=6L_6 = 6, v′=25v' = 25, ℓs=20\ell_s = 20, gs=10−3g_s = 10^{-3} and plot the bounce action by the blue solid line. For the comparison, we plot the bounce actions from the thin-wall approximation with the string tension from the corresponding field theory, which is shown in Figure 4, (orange dash-dot line) and with the string tension from the brane picture (blue dash-dot line). For large θ\theta, field theoretic calculation should be unreliable and thus we show the bounce action from the thin-wall approximation with the string tension from the field theory only for θ≲0.1\theta\lesssim0.1. In the figure, we see that the bounce action from our numerical calculation is close to the thin-wall approximation with the string tension from the brane picture for a wide range of θ\theta.

Figure 12. The bounce action for a wider range of θ\theta. We take L6=6L_6 = 6, v′=25v' = 25, ℓs=20\ell_s = 20, gs=10−3g_s = 10^{-3} and plot the bounce action by the blue solid line. For the comparison, we also plot the bounce actions from the thin-wall approximation; the orange dash-dot line shows the thin-wall approximation with the string tension from the corresponding field theory and the blue dash-dot line shows the thin-wall approximation with the string tension from the brane picture.

For θ≳0.1\theta\gtrsim0.1, the thin-wall approximation is no longer justified in the field theory picture, as discussed above, while in the brane picture it still is: if Δρ\Delta\rho stays of order L6L_6, the ratio of the wall thickness to the bubble radius is Δρ/ρ0≃sin⁡θ\Delta\rho/\rho_0 \simeq\sin\theta, and the approximation remains reasonable up to θ∼0.3\theta\sim0.3. The agreement with the thin-wall curve in that range is therefore expected from the brane point of view, but it also means that the brane picture does not capture the deviation from the thin-wall result that the field theory bounce should exhibit there. This is the limitation stated in the Introduction: the D2-brane is described as a thin object and does not capture the cores of the monopole and the vortex string, so our description cannot reproduce the deviation of the field theory bounce action from the thin-wall result in this regime. For θ≳0.3\theta\gtrsim0.3, on the other hand, the wall and the bubble become comparable even in the brane picture, and yet the bounce action from our numerical calculation keeps following the thin-wall curve up to θ∼1\theta\sim1. We do not have a strong reason for this; numerically, our resolution on, in particular, Step 3 might not be good enough, where the spacing of uoffsetu_{\mathrm{offset}} is taken to be ∼0.14\sim0.14 in our calculation, although we have tried ten times smaller spacing and the bounce action does not change significantly, or it might be because our straight-brane approximation is not good for θ≳0.3\theta\gtrsim0.3 and the deviation from the thin-wall approximation is not captured in our numerical calculation. It is of particular interest to understand the origin of this behavior, which may be a limitation of the straight-brane approximation or of our numerical resolution in uoffsetu_{\mathrm{offset}}. We leave this as an open question for future work.

Conclusion and Discussion

In this paper, we have studied the metastable vortex string in the brane picture and its decay process. We have calculated and compared the tension of the vortex string in the brane picture and in the field theory, and we have found that they are consistent with each other despite the fact that the D2-brane is a non-BPS object and we have ignored the bending of the branes in our calculation. We have then derived the bounce action from the DBI action of the D2-brane within the thin-wall approximation, and found that it takes the same form as the field theory one, with the monopole mass given by the area swept by the D2-brane. As a consistency check, we have developed an algorithm to calculate the bounce action for the decay of the metastable vortex string in the brane picture and performed the numerical calculation based on our algorithm and found that our results successfully reproduce the thin-wall approximation where the thin-wall approximation is expected to be valid. On the other hand, even in the regime where the thin-wall approximation is not expected to be a good approximation in the field theory, our numerical results are still close to the thin-wall approximation. This is partially expected, as the D2D2-brane is described as a thin object and does not capture the cores of the monopole and the vortex string, which are responsible for the deviation from the thin-wall result in the field theory. We have also checked the non-trivial scaling of the bounce action with respect to the parameters in the brane picture, which provides a non-trivial check of our algorithm. In the brane picture, the scaling of gsg_s and ℓs\ell_s with appropriate scaling of other parameters represents non-trivial changes of the parameters, but they do not change the parameters in the field theory, and thus the bounce action should be invariant under these scalings. We have confirmed this invariance of the bounce action under these scalings in our numerical calculation.

The first thing to be discussed is the effect of the non-BPS nature of the D2D2-brane on the decay process. Indeed, the non-BPS nature shows up in the tension: Figure 4 shows that Tvortex/T3T_{\mathrm{vortex}}/T_3 approaches 1 as θ→0\theta\to0, where the vortex string approaches a BPS one, and deviates from it at finite θ\theta, by about 15% at θ=0.1\theta= 0.1. As discussed in Sec. II, the deviation comes from the relaxation of the adjoint field in the core of the vortex string, which corresponds to the bending of the D42D4_2-brane in the vv direction. It is interesting to reproduce this deviation in the brane picture by taking into account the bending of the branes, and to see whether higher derivative terms in the effective action of the field theory from our brane configuration, which are not captured in our Kähler potential and superpotential, also contribute.

Another way to study the effect of the bending of the branes is to consider a different brane configuration where the BPS limit of the whole brane configuration may be taken; possibly a brane configuration including orientifold planes may be useful for this purpose. One may also consider non-BPS configurations of D4-branes and NS5-branes, such as Ref. [52], but then it might well be unclear if the correspondence between the brane configuration in Type IIA string theory and the field theory is valid in such a non-BPS configuration.

Once the brane bending effect and/or higher derivative terms in the effective action of field theory from our brane configuration are taken into account, it is interesting to see if the brane picture can reproduce the field theory result for the bounce action even in the regime where the thin-wall approximation is not a good approximation. In the field theory picture, several studies have been done to go beyond the thin-wall approximation [48, 49, 50] and it is worth studying if brane picture can reproduce at least some of the results in these studies.

Finally, the description of the bounce by a Euclidean D-brane ending on other branes is not specific to the configuration considered in this paper. It would be interesting to apply it to other brane/field-theory configurations, including those describing the decay of other objects such as domain walls, although it might be more difficult to find the corresponding brane configuration.

Acknowledgments

This work was supported by JSPS KAKENHI Grant Nos. 24K17042 [HF], 25H00638 [HF], 26H00403 [HF], 26K17131 [HF] and 24K00522 [KO]. This work was also supported by Grant-in-Aid for JSPS Research Fellow 24KJ0832 [AC] and by FoPM, WINGS Program, the University of Tokyo [AC]. This work was also supported by JST PRESTO Grant Nos. JPMJPR26AB [HF] and JPMJPR26A3 [KO]. This work was also supported by RIKEN ECL program [KO]. Finally, this work was also supported by the Simons Foundation Grant #888984 [KO]. In this research work, we used the UTokyo Azure (https://utelecon.adm.u-tokyo.ac.jp/en/research_computing/utokyo_azure/).

String solution in the field theory picture

In this appendix, we introduce the vortex solution in the field theory picture for our superpotential, Eq. (1), and Kähler potential, Eq. (9). First, the FF-term potential is given by

VF=12Tr⁡(∣2gcos⁡θQ~Q−2μΦ∣2)+∣Q~(2gcos⁡θΦ−mcos⁡θ1)∣2+∣(2gcos⁡θΦ−mcos⁡θ1)Q∣2.(35)V_F = \frac{1}{2}\operatorname{Tr}\left(\left|\frac{\sqrt{2}g}{\cos\theta}\widetilde{Q}Q-2\mu\Phi\right|^2\right)+\left|\widetilde{Q}\left(\frac{\sqrt{2}g}{\cos\theta}\Phi-\frac{m}{\cos\theta}\mathbf{1}\right)\right|^2+\left|\left(\frac{\sqrt{2}g}{\cos\theta}\Phi-\frac{m}{\cos\theta}\mathbf{1}\right)Q\right|^2 . \tag*{(35)}

For the vortex solution, we take the following ansatz:

Q=Q~†=(qξ(r)eiφ000),Φ=v′22πℓs2g(η(r)000),Aφ3=Aφ=1−f(r)r(36)Q=\widetilde{Q}^{\dagger}=\begin{pmatrix}q\xi(r)e^{i\varphi} & 0 \\ 0 & 0\end{pmatrix},\qquad\Phi=\frac{v'}{2\sqrt{2}\pi\ell_s^2g}\begin{pmatrix}\eta(r)&0\\0&0\end{pmatrix},\qquad A_{\varphi}^{3}=A_{\varphi}=\frac{1-f(r)}{r} \tag*{(36)}

with the following boundary conditions:

ξ(0)=0,ξ(∞)=1,η′(0)=0,η(∞)=1,f(0)=1,f(∞)=0.(37)\begin{aligned} \xi(0)&=0,\qquad\xi(\infty)=1,\\ \eta'(0)&=0,\qquad\eta(\infty)=1,\\ f(0)&=1,\qquad f(\infty)=0. \tag*{(37)} \end{aligned}

With this ansatz, the DD-term potential is automatically vanishing, and the tension functional is given by

T=2π∫0∞dr r[2q2(ξ′)2+2q2ξ2f2r2+1g2(f′)2r2+(v′2πℓg2)2(η′)2+g2q4cos⁡2θ(ξ2−η)2+2m2q2cos⁡2θ(η−1)2ξ2].(38)T = 2\pi\int_{0}^{\infty} dr\, r \left[ 2q^{2}(\xi')^{2} + 2\frac{q^{2}\xi^{2}f^{2}}{r^{2}} + \frac{1}{g^{2}}\frac{(f')^{2}}{r^{2}} + \left(\frac{v'}{2\pi\ell_{g}^{2}}\right)^{2}(\eta')^{2} + \frac{g^{2}q^{4}}{\cos^{2}\theta}(\xi^{2}-\eta)^{2} + 2\frac{m^{2}q^{2}}{\cos^{2}\theta}(\eta-1)^{2}\xi^{2} \right]. \tag*{(38)}

Here, rr is the radial coordinate in the x1x^{1}-x2x^{2} plane. The profile functions ξ(r)\xi(r), η(r)\eta(r), and f(r)f(r) are determined by minimizing the above tension functional. To derive Figure 4, we numerically minimize the above tension functional by discretizing the profile functions in rr direction.

References

References

  1. [1]J. Polchinski, Tasi lectures on D-branes, in Theoretical Advanced Study Institute in Elementary Particle Physics (TASI 96): Fields, Strings, and Duality (1996) pp. 293–356, arXiv:hep-th/9611050.arxiv.org/abs/hep-th/9611050
  2. [2]A. Hanany and E. Witten, Type IIB superstrings, BPS monopoles, and three-dimensional gauge dynamics, Nucl. Phys. B 492, 152 (1997), arXiv:hep-th/9611230.DOI
  3. [3]A. Giveon and D. Kutasov, Brane Dynamics and Gauge Theory, Rev. Mod. Phys. 71, 983 (1999), arXiv:hep-th/9802067.DOI
  4. [4]A. Vilenkin, Cosmic Strings and Domain Walls, Phys. Rept. 121, 263 (1985).DOI
  5. [5]H. Xu et al., Searching for the Nano-Hertz Stochastic Gravitational Wave Background with the Chinese Pulsar Timing Array Data Release I, Res. Astron. Astrophys. 23, 075024 (2023), arXiv:2306.16216 [astro-ph.HE].DOI
  6. [6]G. Agazie et al. (NANOGrav), The NANOGrav 15 yr Data Set: Evidence for a Gravitational-wave Background, Astrophys. J. Lett. 951, L8 (2023), arXiv:2306.16213 [astro-ph.HE].DOI
  7. [7]G. Agazie et al. (NANOGrav), The NANOGrav 15 yr Data Set: Observations and Timing of 68 Millisecond Pulsars, Astrophys. J. Lett. 951, L9 (2023), arXiv:2306.16217 [astro-ph.HE].DOI
  8. [8]A. Afzal et al. (NANOGrav), The NANOGrav 15 yr Data Set: Search for Signals from New Physics, Astrophys. J. Lett. 951, L11 (2023), [Erratum: Astrophys.J.Lett. 971, L27 (2024), Erratum: Astrophys.J. 971, L27 (2024)], arXiv:2306.16219 [astro-ph.HE].DOI
  9. [9]J. Antoniadis et al. (EPTA, InPTA;), The second data release from the European Pulsar Timing Array - III. Search for gravitational wave signals, Astron. Astrophys. 678, A50 (2023), arXiv:2306.16214 [astro-ph.HE].DOI
  10. [10]M. T. Miles et al., The MeerKAT Pulsar Timing Array: the first search for gravitational waves with the MeerKAT radio telescope, Mon. Not. Roy. Astron. Soc. 536, 1489 (2024), arXiv:2412.01153 [astro-ph.HE].DOI
  11. [11]D. J. Reardon et al., Search for an Isotropic Gravitational-wave Background with the Parkes Pulsar Timing Array, Astrophys. J. Lett. 951, L6 (2023), arXiv:2306.16215 [astro-ph.HE].DOI
  12. [12]W. Buchmuller, V. Domcke, and K. Schmitz, Metastable cosmic strings, JCAP 11, 020, arXiv:2307.04691 [hep-ph].arxiv.org/abs/2307.04691
  13. [13]A. Afzal, Q. Shafi, and A. Tiwari, Gravitational wave emission from metastable current-carrying strings in E6, Phys. Lett. B 850, 138516 (2024), arXiv:2311.05564 [hep-ph].
  14. [14]W. Ahmed, T. A. Chowdhury, S. Nasri, and S. Saad, Gravitational waves from metastable cosmic strings in the Pati-Salam model in light of new pulsar timing array data, Phys. Rev. D 109, 015008 (2024), arXiv:2308.13248 [hep-ph].
  15. [15]A. Afzal, M. Mehmood, M. U. Rehman, and Q. Shafi, Supersymmetric hybrid inflation and current-carrying metastable cosmic strings in SU(4)×SU(2)L×U(1)R, Phys. Rev. D 112, 083545 (2025), arXiv:2308.11410 [hep-ph].
  16. [16]W. Ahmed, M. U. Rehman, and U. Zubair, Probing stochastic gravitational wave background from SU(5) × U(1)χ strings in light of NANOGrav 15-year data, JCAP 01, 049, arXiv:2308.09125 [hep-ph].
  17. [17]G. Lazarides, R. Maji, A. Moursy, and Q. Shafi, Inflation, superheavy metastable strings and gravitational waves in non-supersymmetric flipped SU(5), JCAP 03, 006, arXiv:2308.07094 [hep-ph].
  18. [18]G. Lazarides, R. Maji, and Q. Shafi, Superheavy quasistable strings and walls bounded by strings in the light of NANOGrav 15 year data, Phys. Rev. D 108, 095041 (2023), arXiv:2306.17788 [hep-ph].
  19. [19]S. Antusch, K. Hinze, S. Saad, and J. Steiner, Probing SUSY at gravitational wave observatories, Phys. Lett. B 856, 138924 (2024), arXiv:2405.03746 [hep-ph].arxiv.org/abs/2405.03746
  20. [20]A. Chitose, M. Ibe, S. Neda, and S. Shirai, Gravitational waves from metastable cosmic strings in supersymmetric new inflation model, JCAP 04, 010, arXiv:2411.13299 [hep-ph].
  21. [21]C. Pallis, T-model Higgs inflation and metastable cosmic strings, JHEP 01, 178, arXiv:2409.14338 [hep-ph].
  22. [22]R. Maji, A. Moursy, and Q. Shafi, Induced gravitational waves, metastable cosmic strings and primordial black holes in GUTs, JCAP 01, 106, arXiv:2409.13584 [hep-ph].arxiv.org/abs/2409.13584
  23. [23]S. Datta and R. Samanta, Cosmic superstrings, metastable strings and ultralight primordial black holes: from NANOGrav to LIGO and beyond, JHEP 02, 095, arXiv:2409.03498 [gr-qc].arxiv.org/abs/2409.03498
  24. [24]S. Antusch, K. Hinze, and S. Saad, Explaining PTA results by metastable cosmic strings from SO(10) GUT, JCAP 10, 007, arXiv:2406.17014 [hep-ph].arxiv.org/abs/2406.17014
  25. [25]S. Antusch and K. Trailović, Cosmic strings from Tribird inflation, JCAP 10, 095, arXiv:2406.12521 [hep-ph].DOI
  26. [26]C. Pallis, PeV-Scale SUSY and Cosmic Strings from F-Term Hybrid Inflation, Universe 10, 10.3390/universe10050211 (2024), arXiv:2403.09385 [hep-ph].DOI
  27. [27]J. Ingoldby, V. V. Khoze, and J. Turner, Metastable strings and gravitational waves in one-scale models, JHEP 04, 094, arXiv:2511.08546 [hep-ph].arxiv.org/abs/2511.08546
  28. [28]A. Chitose, M. Ibe, S. Neda, and S. Shirai, Do Cosmic String Segments Emit Gravitational Waves?, (2025), arXiv:2507.12386 [hep-ph].arxiv.org/abs/2507.12386
  29. [29]R. Maji and Q. Shafi, Superheavy metastable strings in SO(10), JHEP 06, 217, arXiv:2504.09055 [hep-ph].DOI
  30. [30]S. Antusch, K. Hinze, and S. Saad, Metastable cosmic strings and gravitational waves from flavor symmetry breaking, Phys. Rev. D 112, 035043 (2025), arXiv:2503.05868 [hep-ph].arxiv.org/abs/2503.05868
  31. [31]Y. Hu and K. Kamada, Gravitational waves from metastable cosmic strings in the delayed scaling scenario, JCAP 04, 044, arXiv:2501.18380 [astro-ph.CO].arxiv.org/abs/2501.18380
  32. [32]M. N. Ahmad, M. Mehmood, M. U. Rehman, and Q. Shafi, μ-hybrid inflation and metastable cosmic strings in SU(3)c×SU(2)L×SU(2)R×U(1)B-L, Phys. Rev. D 111, 083526 (2025), arXiv:2501.06307 [hep-ph].
  33. [33]S. Blasi, M. Grandjean, and A. Mariotti, Metastable strings at PTAs: classical stability analysis, (2026), arXiv:2605.03003 [hep-ph].arxiv.org/abs/2605.03003
  34. [34]D. H. Asl and K. Schmitz, New gravitational-wave templates for metastable cosmic strings: Loop breaking versus network collapse, (2026), arXiv:2604.28097 [hep-ph].arxiv.org/abs/2604.28097
  35. [35]L. Tranchedone, E. Carragher, E. Hardy, and N. Koscelanská van IJcken, Metastable cosmic strings are broken at the start, (2026), arXiv:2601.04320 [hep-ph].DOI
  36. [36]A. de Giorgi, J. Ingoldby, V. V. Khoze, and J. Turner, Thermal Metastable Strings in One-Scale Models and Gravitational Waves, (2026), arXiv:2606.02689 [hep-ph].arxiv.org/abs/2606.02689
  37. [37]R. Maji and Q. Shafi, Monopoles, Strings, Walls and Gravitational waves, (2026), arXiv:2607.20340 [hep-ph].arxiv.org/abs/2607.20340
  38. [38]S. Datta and R. Samanta, LISA Reconstruction Landscape for Metastable Cosmic Strings, (2026), arXiv:2607.28592 [hep-ph].arxiv.org/abs/2607.28592
  39. [39]I. Rybak, R. Maji, and Q. Shafi, Gravitational Wave Background from Metastable Current-Carrying Cosmic Strings, (2026), arXiv:2610.38140 [astro-ph.CO].arxiv.org/abs/2610.38140
  40. [40]A. Hanany, M. J. Strassler, and A. Zaffaroni, Confinement and strings in MQCD, Nucl. Phys. B 513, 87 (1998), arXiv:hep-th/9707244.DOI
  41. [41]A. Hanany and D. Tong, Vortices, instantons and branes, JHEP 07, 037, arXiv:hep-th/0306150.
  42. [42]A. Hanany and D. Tong, Vortex strings and four-dimensional gauge dynamics, JHEP 04, 066, arXiv:hep-th/0403158.DOI
  43. [43]D. Tong, Monopoles in the higgs phase, Phys. Rev. D 69, 065003 (2004), arXiv:hep-th/0307302.DOI
  44. [44]D. Tong, TASI lectures on solitons: Instantons, monopoles, vortices and kinks, in Theoretical Advanced Study Institute in Elementary Particle Physics: Many Dimensions of String Theory (2005) arXiv:hep-th/0509216.DOI
  45. [45]A. Sen, NonBPS states and Branes in string theory, in Advanced School on Supersymmetry in the Theories of Fields, Strings and Branes (1999) pp. 187–234, arXiv:hep-th/9904207.DOI
  46. [46]A. Vilenkin, COSMOLOGICAL EVOLUTION OF MONOPOLES CONNECTED BY STRINGS, Nucl. Phys. B 196, 240 (1982).DOI
  47. [47]J. Preskill and A. Vilenkin, Decay of metastable topological defects, Phys. Rev. D 47, 2324 (1993), arXiv:hep-ph/9209210.DOI
  48. [48]M. Shifman and A. Yung, Metastable strings in Abelian Higgs models embedded in nonAbelian theories: Calculating the decay rate, Phys. Rev. D 66, 045012 (2002), arXiv:hep-th/0205025.DOI
  49. [49]A. Chitose, M. Ibe, Y. Nakayama, S. Shirai, and K. Watanabe, Revisiting metastable cosmic string breaking, JHEP 04, 068, arXiv:2312.15662 [hep-ph].DOI
  50. [50]V. Domcke and Y. Hamada, The decay rate of metastable cosmic strings beyond the thin-string approximation, (2026), arXiv:2606.03008 [hep-ph].arxiv.org/abs/2606.03008
  51. [51]R. Auzzi, S. Bolognesi, and J. Evslin, Monopoles can be confined by 0, 1 or 2 vortices, JHEP 02, 046, arXiv:hep-th/0411074.DOI
  52. [52]M. Eto, K. Konishi, M. Nitta, and Y. Ookouchi, Brane Realization of Nambu Monopoles and Electroweak Strings, Phys. Rev. D 87, 045006 (2013), arXiv:1211.2971 [hep-th].DOI
  53. [53]A. Gorsky and K. Selivanov, Junctions and the fate of branes in external fields, Nucl. Phys. B 571, 120 (2000), arXiv:hep-th/9904041.DOI
  54. [54]A. S. Gorsky, K. A. Saraikin, and K. G. Selivanov, Schwinger type processes via branes and their gravity duals, Nucl. Phys. B 628, 270 (2002), arXiv:hep-th/0110178.DOI
  55. [55]C. Schubert and A. Torrielli, Open string pair creation from worldsheet instantons, J. Phys. A 43, 402003 (2010), arXiv:1008.2068 [hep-th].DOI
  56. [56]S. Bolognesi, E. Rabinovici, and G. Tallarita, String pair production in non homogeneous backgrounds, JHEP 04, 174, arXiv:1601.04758 [hep-th].DOI
  57. [57]G. W. Semenoff and K. Zarembo, Holographic Schwinger Effect, Phys. Rev. Lett. 107, 171601 (2011), arXiv:1109.2920 [hep-th].DOI
  58. [58]K. Peeters, J. Sonnenschein, and M. Zamaklar, Holographic decays of large-spin mesons, JHEP 02, 009, arXiv:hep-th/0511044.DOI
  59. [59]K. Peeters, M. Matuszewski, and M. Zamaklar, Holographic meson decays via worldsheet instantons, JHEP 06, 083, arXiv:1803.06318 [hep-th].DOI
  60. [60]A. Kasai and Y. Ookouchi, Decay of False Vacuum via Fuzzy Monopole in String Theory, Phys. Rev. D 91, 126002 (2015), arXiv:1502.01544 [hep-th].DOI
  61. [61]A. Kasai and Y. Ookouchi, Gravitational Correction to Fuzzy String in Metastable Brane Configuration, JHEP 06, 098, arXiv:1504.00479 [hep-th].DOI
  62. [62]M. Eto, K. Hashimoto, and S. Terashima, Solitons in Supersymmetry Breaking Meta-Stable Vacua, JHEP 03, 061, arXiv:hep-th/0610042.DOI
  63. [63]S. R. Coleman, The Fate of the False Vacuum. 1. Semiclassical Theory, Phys. Rev. D 15, 2929 (1977), [Erratum: Phys.Rev.D 16, 1248 (1977)].
  64. [64]C. G. Callan, Jr. and S. R. Coleman, The Fate of the False Vacuum. 2. First Quantum Corrections, Phys. Rev. D 16, 1762 (1977).DOI
  65. [65]J. Polchinski, String theory. Vol. 2: Superstring theory and beyond, Cambridge Monographs on Mathematical Physics (Cambridge University Press, 2007).
  66. [66]M. Shifman, Advanced topics in quantum field theory: A lecture course (Cambridge Univ. Press, Cambridge, UK, 2012).
  67. [67]C. G. Callan and J. M. Maldacena, Brane death and dynamics from the Born-Infeld action, Nucl. Phys. B 513, 198 (1998), arXiv:hep-th/9708147.DOI
  68. [68]G. W. Gibbons, Born-Infeld particles and Dirichlet p-branes, Nucl. Phys. B 514, 603 (1998), arXiv:hep-th/9709027.DOI
  69. [69]P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, E. Burovski, P. Peterson, W. Weckesser, J. Bright, S. J. van der Walt, M. Brett, J. Wilson, K. J. Millman, N. Mayorov, A. R. J. Nelson, E. Jones, R. Kern, E. Larson, C. J. Carey, İ. Polat, Y. Feng, E. W. Moore, J. VanderPlas, D. Laxalde, J. Perktold, R. Cimmrman, I. Henriksen, E. A. Quintero, C. R. Harris, A. M. Archibald, A. H. Ribeiro, F. Pedregosa, P. van Mulbregt, and SciPy 1.0 Contributors, SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python, Nature Methods 17, 261 (2020).

Paper details

vlnc:W_rtm5xfr5
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

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