hep-th/0007078 CERN-TH/2000-193 MRI-P-000701 TIFR/TH/00-36 Nielsen-Olesen Vortices in Noncommutative Abelian Higgs Model Dileep P. Jatkara,b, Gautam Mandalb,c and Spenta R. Wadiac a Mehta Research Institute of Mathematics and Mathematical Physics Chhatnag Road, Jhusi, Allahabad 211 019, India b Theory Division, CERN CH-1211, Geneva, 23, Switzerland b Department of Theoretical Physics, Tata Institute of Fundamental Research Homi Bhabha Road, Mumbai 400 005, India Abstract We construct Nielsen-Olesen vortex solution in the noncommutative abelian Higgs model. We derive the quantized topological flux of the vortex solution. We (cid:12)nd that the flux is integral by explicit computation in the large θ limit as well as in the small θ limit. In the context of a tachyon vortex on the brane- antibrane system we demonstrate that it is this topological charge that gives rise to the RR charge of the resulting BPS D-brane. We also consider the left- right-symmetric gauge theory which does not have a commutative limit and construct an exact vortex solution in it. July 2000 e-mail: [email protected], [email protected], [email protected] 1 Introduction and Summary Noncommutative (cid:12)eldtheories appear naturallyin thelowenergy description ofstring theory in a constant Neveu-Schwarz antisymmetric tensor background [1]-[11]. They have also appeared previously in the study of c = 1 matrix model (two-dimensional string theory) [12] and of two-dimensional QCD [13]. In [2] it was observed that D- branes compacti(cid:12)ed on a torus with constant Neveu-Schwarz B (cid:12)eld background gives rise to an e(cid:11)ective noncommutative (cid:12)eld theory on the compacti(cid:12)ed world-volume. There has been some study of perturbative dynamics of these theories[14]. However, their utility in understanding nonperturbative aspects of (cid:12)eld theories has attracted more attention lately[15]. Presence of constant Neveu-Schwarz antisymmetric tensor (cid:12)eld background tangential to the D-brane leads to noncommutative gauge theory on the D-brane world volume. In case of the abelian gauge theory it has been shown that the abelian noncommutative gauge theory is related by (cid:12)eld rede(cid:12)nition to the Born- Infeld electrodynamics action on the D-brane[4]. Noncommutative gauge theories have been further discussed in [16, 17]. Solitons in noncommutative scalar (cid:12)eld theories were studied in [15]. All the soli- tons studied in [15] are non-topological, and are stable at large θ. These solutions lack stability in the commutative limit as they violate Derrick’s theorem. However, coupling of these scalars to, say, gauge (cid:12)elds could add stability to these solutions. Simplicity of the noncommutative (cid:12)eld theory formalism in the large θ limit was exploited in [18, 19] to gain better understanding of the tachyon condensation phe- nomenon in the non-BPS D-branes[20] using the physics of noncommutative solitons. It is important to generalize the class of noncommutative solitons to accommodate solitonic solutions which have topological conserved charge. It is clear from the work of [18, 19, 20, 21, 22] that the topological solitons are relevant to the study of brane descent relations in the non-BPS brane dynamics. It has been conjectured that the topological charge of the tachyonic soliton on the brane worldvolume of the non-BPS Dp brane is the RR charge of the D(p−k) brane[20, 22], where k < p and for k = 1 the soliton is a kink, for k = 2, it is a vortex and so on. Study of topology in noncommutative (cid:12)eld theories, however, is also important in its own right. Here we will address this question in the abelian Higgs model. We will show that the noncommutative abelian Higgs model supports topological vortex solution with quantized flux, which can be any arbitrary integer. The model that we will study (cid:12)rst is a left module on the Hilbert space, that is, in the complex Higgs (cid:12)eld notation, 1 (cid:22) the gauge (cid:12)eld multiplies the Higgs (cid:12)eld φ from the left whereas for φ it multiplies from right. In section 2, we will briefly review the Nielsen-Olesen vortex solution in the ordinary abelian Higgs model, highlighting the (cid:12)rst order,i.e., Bogomolnyi formulation. We generalise the Bogomolnyi equations to the noncommutative abelian Higgs model in section 3 and obtain exact solutions in the large θ limit. In this limit it is possible to do a systematic 1/θ expansion and we obtain corrections to the solution of the leading order equations of motion. We (cid:12)nd that these corrections converge quite rapidly. In the large distance limit this solution matches with the large distance Nielsen-Olesen ansatzforvortexsolutionintheordinaryabelianHiggsmodel. Wealso derivetheSeiberg-Witten(SW)mapforthenoncommutativeabelianHiggsmodel. We (cid:12)nd that to order θ, the SW map for the gauge (cid:12)eld A and the gauge transformation parameter λ is unaltered whereas the SW map for the Higgs φ is linear in φ. It has been argued in the pure gauge theory case[4], that the SW map in the zero slope limit gives a (cid:12)eld rede(cid:12)nition from an ordinary (cid:12)eld theory with the Born-Infeld action to the noncommutative gauge theory action which is quadratic in the gauge (cid:12)eld strength. The SW map for A is a nonlinear function of A, whereas as mentioned above SW map for φ is linear in φ. Therefore it is tempting to conjecture that the corresponding action for the Higgs (cid:12)eld in the ordinary (cid:12)eld theory will retain its form with each term in the Higgs Lagrangian multiplied by some function of the gauge (cid:12)eld A. We also determine the pro(cid:12)le of the magnetic (cid:12)eld B due to this vortex solution. In the original coordinates it is just a δ-function. This is as expected since in the large θ limit, the dispersion generated by the derivative terms in the kinetic energy is totally suppressed. In the scaled coordinates, however, the magnetic (cid:12)eld pro(cid:12)le is proportional to the ground state wavefunction of the harmonic oscillator. The coe(cid:14)cient in front of this wavefunction encodes the topological charge of the solution. Insection4, wediscuss theissue ofthetopologyofthissolution. Weshow thatthis solution carries the topological conserved charge (magnetic flux) which is determined by the behaviour of the Higgs (cid:12)eld. This topological charge is conserved, quantized and takes integer values. We establish this result in both small as well as in large θ limit. In the small θ limit the leading result reproduces the integer vortex charge of the commutative limit; we show that higher order terms in θ expansion are total derivatives and fall too rapidly at large distances to contribute to the charge. In the large θ limit too, the leading term itself gives the entire charge and subleading 2 terms in the 1/θ expansion do not contribute. We discuss the connection with this topological charge with various other quantized charges and indices. We also discuss the \semiclassical" limit of the (cid:12)eld con(cid:12)guration in which the topology reduces to that of the commutative limit. In section 5, we address the same question but for theleft-right symmetric module. In this case we get an interesting vortex solution which at the face value seems to have charge 1 (using Witten’s identi(cid:12)cation [21] of the vortex charge with Atiyah- Singer index of the Higgs (cid:12)eld). The solution is not square integrable, however; in the operator language the trace of the topological charge over the one particle Hilbertspacediverges andrequires regularisation. Thequestion of(cid:12)nding aconsistent regularisation scheme to compute this charge is left for the future. 2 Ordinary Abelian Higgs Model In this section we will briefly review the Nielsen-Olesen vortex solution to the abelian Higgs model[23], emphasising the Bogomolnyi limit of this model[24]. In the next section we will generalise these equations to the noncommutative abelian Higgs model and work only in the Bogomolnyi limit. Let us start with the usual Abelian Higgs model. The Lagrangian is 1 1 L = − F Fµν − j(∂ +ieV )Ψj2 −λ(jΨj2−jΨ j2)2. (2.1) µν µ µ 0 4 2 As mentioned earlier we will concentrate on the Bogomolnyi limit. Bogomolnyi limit and the corresponding (cid:12)rst order equations of motion are obtained from the energy functionalby writing it in terms ofstrictly positive quantities anda topological charge density and then minimising the energy functional. We will take the static ansatz. That means we will set V = 0 and ∂ V = 0 = ∂ φ. Energy functional of the 0 t m t abelian Higgs model for such an ansatz is (cid:90) (cid:20) (cid:21) 1 E = d2x F Fmn +jD Ψj2 +λ(jΨj2 −Ψ2)2 (2.2) mn m 0 4 Let us rescale coordinates and (cid:12)elds and write them down in terms of the following dimensionless variables 1 1 Ψ = Ψ φ, V = Ψ A , x = y , z = p (y +iy ) (cid:17) r eiϕ and 0 m 0 m m m 1 2 eΨ0 2 E = 2πΨ2E (2.3) 0 3 The Nielsen-Olesen ansatz for the asymptotic form of a vortex solution with winding number n is zn (cid:22) (cid:3) φ = exp(inϕ) = , φ = φ n (zz(cid:22))2 n A = −i , A = A(cid:3) (2.4) z z¯ z 2z The second equation is equivalent to m A dx = n dϕ (2.5) m With the de(cid:12)nitions (2.3) the energy functional becomes (cid:34) (cid:35) (cid:90) 1 1 β E = d2z B2 +jD φj2 + (φφ(cid:22)−1)2 (2.6) m 2π 2 2 where β = 2λ/e2, D is a covariant derivative with gauge (cid:12)eld A and B = ∂ A −∂ A = −i(∂A(cid:22)−∂(cid:22)A) = [D,D(cid:22)], (2.7) 1 2 2 1 It is easy to rewrite this as (see Appendix B for the derivation in the noncommutative case) (cid:34) (cid:35) (cid:90) 1 1 β −1 E = d2z (B +(φφ(cid:22)−1))2 +D φD φ(cid:22)+ (φφ(cid:22)−1)2 +T (2.8) z¯ z 2π 2 2 where T = ∂ Sm +B (2.9) m 1 Sm = (cid:15)mn(iφD φ(cid:22)−iD φφ(cid:22)) (2.10) n n 2 Our convention for (cid:15)mn is that (cid:15) = 1. 12 We will argue below that T is a topological density, which generalises naturally to the noncommutative case as well. It is easy to see in the commutative case (the same (cid:82) will be true for the noncommutative generalisation) that 1 d2zT gives the magnetic 2π flux of the vortex and is an integer n. The energy functional for a vortex with winding number n, then, is (cid:90) 1 1 β −1 E = n+ d2z[ (B +(φφ(cid:22)−1))2 +D φD φ(cid:22)+ (φφ(cid:22)−1)2] (2.11) z¯ z 2π 2 2 Notice E is a sum of absolute square terms except the last term. However, when β = 1, this term drops out and minimum of E can be obtained if the Bogomolnyi equations are satis(cid:12)ed: D φ = 0, D φ(cid:22)= 0, B = 1−φφ(cid:22) (2.12) z¯ z 4 It is interesting to notethat the Euclidean action(2.8) can bewritten in anelegant form using Quillen’s superconnection [25] A de(cid:12)ned below (cid:90) (cid:16) (cid:17) E = d2x jjF −1jj2 +Strexp[F] (2.13) where in the exponential only the term proportional to the volume form contributes, namely (cid:90) d2xStrexp[F] = F+ +D(cid:22)φ^Dφ(cid:22)−fF+,φφ(cid:22)g. (2.14) Here F is the curvature of the superconnection [25] A (cid:32) (cid:33) d+A+ φ A = (2.15) (cid:22) φ d and F+ is the curvature of A+. The Bogomolnyi equations (2.12) can be written in a very suggestive form F −1 = 0 (2.16) 3 Noncommutative Abelian Higgs Model Nowletusconsidernoncommutative generalisationoftheseequations. Wepresent our conventions and de(cid:12)nitions for noncommutative (cid:12)eld theory following [16] (see also appendix A). The energy functional for the noncommutative abelian Higgs model is given by [16] (cid:20) (cid:21) 1 1 E = Tr B2 +D φD φ(cid:22)+D φD φ(cid:22)+ (φφ(cid:22)−1)2 . (3.1) z¯ z z z¯ 2 2 where B is now de(cid:12)ned as B = [D,D(cid:22)] = −i(∂ A −∂ A )−[A ,A ] (3.2) z z¯ z¯ z z z¯ Following the steps described in Appendix B, we can recast the energy functional as (cid:20) (cid:21) 1 E = Tr (B +(φφ(cid:22)−1))2 +D φD φ(cid:22)+T) , (3.3) z¯ z 2 where the topological term is now T = D Sm +B (3.4) m with B de(cid:12)ned as in (3.2) and D Sm = ∂ Sm −i[A ,Sm] (3.5) m m m 5 Sm is de(cid:12)ned as before (2.10) with due attention to operator ordering now. The use of the covariant derivative on Sm in (3.4) is necessary since Sm is now gauge-covariant. We will argue in the next section that I = Tr T, (3.6) corresponds to a topological charge. In particular, for the noncommutative Nielsen- Olesen vortex of charge n constructed below, I evaluates to n for any value of the noncommutativity parameter θ. Bogomolnyi equations It is clear from (3.3) that the Bogomolnyi equations remain the same as in (2.12), namely D φ = 0, D φ(cid:22)= 0, B = 1−φφ(cid:22) (3.7) z¯ z which are now to be interpreted as operator equations (conventions described in Ap- pendix A). It is di(cid:14)cult to solve the Bogomolnyi equations exactly. We will (cid:12)nd the solution in various limits. First we will solve the large θ limit (θ ! 1 and p1 corrections). θ 3.1 Large θ We will now consider the limit when the the noncommutativity parameter θ, de(cid:12)ned by 1 2 [X ,X ] = iθ (3.8) is large. Let us de(cid:12)ne the following scaled operators p Xi = θX(cid:102)i, i = 1,2 (3.9) We de(cid:12)ne the annihilation and creation operators as 1 1 a = p (X(cid:103)1 +iX(cid:103)2), ay = p (X(cid:103)1 −iX(cid:103)2) (3.10) 2 2 The scaled complex coordinates w,w(cid:22) are de(cid:12)ned as p p z = θw, z(cid:22)= θw(cid:22) (3.11) In accordance with the fact that the gauge potential A is a 1-form and the magnetic (cid:12)eld is a 2-form, the scaled gauge potential A~ and scaled magnetic (cid:12)eld B~ are given 6 by (cid:101) (cid:101) A B p A = , B = . (3.12) θ θ With the above rescalings the energy functional (3.3) becomes  (cid:32) (cid:33)  1 B(cid:101) 2 1 B(cid:101) D Sm E = θ Tr +(φφ(cid:22)−1) + D φD φ(cid:22)+ + m ). (3.13) w¯ w 2 θ θ θ θ Here ∂ = −Aday,∂ = Ada (see Appendix A). w w¯ Let us now solve the Bogomolnyi equations (3.7) in these rescaled variables order by order in 1/θ. We write the following 1/θ-expansion of the Higgs (cid:12)eld and the gauge (cid:12)eld. 1 φ = φ + φ +... 1 −1 θ 1 1 1 p (cid:101) p A = A = (A + A +...) (3.14) 1 −1 θ θ θ The expansion of A(cid:22) is identical to that of A except that A(cid:22) are hermitian conjugates i of A . The large θ expansion of the magnetic (cid:12)eld (3.2) is given by i 1 1 1 (cid:101) B = B = B + B +... (3.15) θ θ 1 θ2 −1 where B = −i(∂ A(cid:22) −∂ A )−[A ,A(cid:22) ] = i([ay,A(cid:22) ]+[a,A ])−[A ,A(cid:22) ], (3.16) 1 w 1 w¯ 1 1 1 1 1 1 1 B = −i(∂A(cid:22) −∂(cid:22)A )−[A ,A(cid:22) ]−[A ,A(cid:22) ] −1 −1 −1 1 −1 −1 1 = i([ay,A(cid:22) ]+[a,A ])−[A ,A(cid:22) ]−[A ,A(cid:22) ]. (3.17) −1 −1 1 −1 −1 1 Let us substitute the expansions (3.14), (3.15) in (3.13) or in the Bogomolnyi equations and solve these equations order by order in 1/θ. o(θ) Bogomolnyi equations The relative orders of the Bogomolnyi equations are easiest to (cid:12)gure out from (3.13). The o(θ) term in (3.13) gives the leading Bogomolnyi equation give (cid:22) φ φ = 1. (3.18) 1 1 This equation is solved by Witten[21] and the solution is 1 1 p n (cid:22) ynp φ = a , φ = a . (3.19) 1 1 anayn anayn 7 It is easy to see by simple substitution that this ansatz indeed solves the leading order Bogomolnyi equation (3.18). o(1) Bogomolnyi equations At o(1) in (3.13) we get the Bogomolnyi equations involving the covariant deriva- tive of the Higgs (cid:12)eld, namely ∂ φ −iA(cid:22) φ = 0, ∂ φ(cid:22) +iφ(cid:22) A = 0 (3.20) w¯ 1 1 1 w 1 1 1 (cid:22) For the sake of simplicity we will work only with the φ equation of motion (φ, being hermitian conjugate of φ, can be determined from the solution to the φ equation of motion). The φ-equation can be written in the operator form as [a,φ ] = iA(cid:22) φ . (3.21) 1 1 1 Let us recall at this point the action of creation and annihilation operators of the harmonic oscillator on the one particle Hilbert space. p p ajmi = mjm−1i, ayjmi = m+1jm+1i. (3.22) For future purposes, it is useful to write down the action of the Higgs (cid:12)eld on the one particle Hilbert space. The Higgs (cid:12)eld con(cid:12)guration for the vortex solution is written in terms of harmonic oscillator creation and annihilation operators. Using (3.22), it is easy to see that the action of the Higgs (cid:12)eld on the Hilbert space is φjmi = jm−ni. (3.23) Now let us look at the subleading equation, that is (3.21). Using the above results it is easy to derive the gauge (cid:12)eld A(cid:22) (A is given by its hermitian conjugate). Thus 1 1 p p 1 A(cid:22) = −ip a( N − N +n) 1 N +1 p p 1 A = i( N − N +n)ayp , (3.24) 1 N +1 where N = aya is the number operator. In the large w limit or equivalently in the large hNi limit, the vortex solution becomes φ = exp(inϕ), 1 n n A(cid:22) = i , A = −i (3.25) 1 1 2w(cid:22) 2w 8 which is the same as the large distance behaviour of the usual Nielsen-Olesen vortex. In a way, this result is expected because in the large θ limit we have ignored the derivative terms and then in the large hNi limit the asymptotic behaviour of the vortex solution becomes exact. Behaviour of the vortex solution in the (cid:12)nite domain of the w plane, that is, for (cid:12)nite values of hNi depends on the competition between the kinetic energy terms and the potential energy terms in the energy functional. Exact solution to the equations in the large θ expansion, which essentially ignores the kinetic energy e(cid:11)ect, of the Bogomolnyi equations, reduces to the potential energy minimisation in the large hNi limit. We will show below that (3.6) evaluates to n for this solution. Therefore, this solution carries topological charge, which is determined by the quantized magnetic flux through the vortex solution. The magnetic field From (3.16) we see that B is given entirely in terms of A . Using the solution 1 1 (3.24) in (3.16) (details in Appendix C) we get B = nj0ih0j, (3.26) 1 where j0ih0j is a projection operator onto the vacuum state. It is curious that the leading term in the large θ expansion of the magnetic (cid:12)eld has such a remarkably simple form. It is also interesting to note that the trace of B in the one particle 1 Hilbert space gives us exactly the integer n as desired. We will elaborate more on this when we will discuss the topology of our solution. In terms of the original unscaled coordinates this vacuum projection operator is essentially a δ-function. However, in the scaled variables the vacuum projection operatoris represented by the groundstate wavefunction, i.e., by the Gaussian. Signi(cid:12)cance of this will be discussed in the next section. o(1/θ) and higher orders So far we have looked at the leading and the (cid:12)rst subleading term in the large θ expansion of the Bogomolnyi equations. Here we will look at the higher corrections to the Higgs (cid:12)eld φ as well as the gauge (cid:12)eld A. It is interesting to see that the large θ correction are quite small and the convergence of the solution is remarkably fast. The o(1/θ) equation of motion is given by B = −φ φ(cid:22) −φ φ(cid:22) . (3.27) 1 1 −1 −1 1 9