FTPI-MINN-08/04, UMN-TH-2534/08 ITEP-TH-09/08 March 2, 2008 N Heterotic Flux Tubes in = 2 SQCD N with = 1 Preserving Deformations 9 M. Shifmana and A. Yunga,b,c 0 0 2 n a aWilliam I. Fine Theoretical Physics Institute, University of Minnesota, J Minneapolis, MN 55455, USA 0 3 bPetersburg Nuclear Physics Institute, Gatchina, St. Petersburg 188300, Russia cInstitute of Theoretical and Experimental Physics, Moscow 117259, Russia ] h t - Abstract p e Weconsidernon-AbelianBPS-saturatedfluxtubes(strings)in = 2 super- h N [ symmetric QCD deformed by superpotential terms of a special type breaking 2 = 2 supersymmetry down to = 1. Previously it was believed that v N N worldsheet supersymmetry is “accidentally” enhanced due to the facts that 8 5 = (1,1) SUSY is automatically elevated up to = (2,2) on CP(N 1) 1 N N − and, at the same time, there are no = (0,2) generalizations of the bosonic 0 N . CP(N 1) model. Edalati and Tong noted that the target space is in fact 3 − 0 CP(N 1) C rather than CP(N 1). This allowed them to suggest a “het- 8 − × − erotic” = (0,2) sigma model, with the CP(N 1) target space for bosonic 0 N − : fields and an extra right-handed fermion which couples to the fermion fields of v i the = (2,2) CP(N 1)model. Wederivetheheterotic = (0,2) worldsheet X N − N model directly from the bulk theory. The relation between the bulk and world- r a sheet deformation parameters we obtain does not coincide with that suggested by Edalati and Tong at large values of the deformation parameter. For polyno- mial deformation superpotentials in the bulk we find nonpolynomial response in the worldsheet model. We find a geometric representation for the heterotic model. Supersymmetry is proven to be spontaneously broken for small defor- mations (atthequantumlevel). ThisconfirmsTong’s conjecture. Aproofvalid for large deformations will be presented in the subsequent publication. Contents 1 Introduction 2 2 Bulk theory 6 3 Non-Abelian strings 11 3.1 Fermion zero modes: N = 2 limit . . . . . . . . . . . . . . . . . . . . . . . 14 3.2 CP(1)×C model on the string worldsheet: direct calculation in the N = 2 limit . . . . . . . . . . . . . . . . . . . . . . . . . 17 4 Digression: Edalati–Tong’s suggestion 20 5 Heterotic CP(1) 22 6 Geometric formulation of the N = (0,2) heterotic model 25 7 From the bulk N = 1 theory to the heterotic deformation of the CP(1) model on the worldsheet 30 7.1 Fermion zero modes in N = 1 theory . . . . . . . . . . . . . . . . . . . . . 31 7.2 Small-µ limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 7.3 Large-µ limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 8 Parameters of the heterotic CP(1) model from the bulk theory 37 8.1 Small-µ limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 8.2 Large-µ limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 9 Twisted mass in the worldsheet theory 40 10 Adding the polynomial deformation superpotential 43 10.1 Small-µ limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 10.2 Large-µ limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 11 Conclusions 48 Appendices 50 References 52 1 1 Introduction Non-Abelian BPS-saturated flux tubes were discovered and studied in = 2 super- N symmetric QCD [1, 2, 3, 4, 5]. The simplest model supporting such flux tubes, to be referred to as the basic model, has the gauge group U(N), with the U(1) Fayet–Iliopoulos (FI) term, and N flavors (N hypermultiplets in the fundamental representation). Multiple developments in supersymmetric solitons and ideas about confinement ensued (for reviews see [6, 7, 8]). A crucial feature of non-Abelian strings is the presence of orientational (and superorientational) moduli associated with ro- tations of their color fluxes inside a non-Abelian group, in addition to “standard” translational and supertranslational moduli. The low-energy theory on the string worldsheet is split into two disconnected parts: a free theory for (super)translational moduli and a nontrivial part, a theory of interacting (super)orientational moduli, CP(N 1) model. The latter is completely fixed by the fact that the basic bulk − theory has eight supercharges, and the string under consideration is 1/2 BPS. As well-known (e.g. [9, 10]), the only supergeneralization of the bosonic CP(N 1) is − the = (2,2) supersymmetric CP(N 1) model with four supercharges. N − In a bid to decrease the level of supersymmetry (SUSY) in the bulk theory an = 2 breaking deformation of the type N = µ 2 (1.1) deform W A was introduced [11] where is the adjoint chiral superfield. The above deformation A preserves = 1 in the bulk. As µ increases, the adjoint fields become heavier and N eventually decouple from the spectrum at µ . → ∞ With = 1 preserving deformation of the basic model, there are four con- N served supercharges in the bulk rather than eight. At the same time, the description of the orientational moduli is the same as in the = 2 basic model; the bosonic N part of the worldsheet theory is CP(N 1). Since the string solution remains 1/2 − BPS, the worldsheet theory must have two conserved supercharges. Endowing the bosonic CP(N 1) model with two supercharges automatically endows it with four − supercharges [9, 10]. A conclusion was made [11] that in the problem at hand, un- expectedly, the worldsheet supersymmetry enhances up to = (2,2). If it were the N case, the situation would be similar to supersymmetry enhancement on domain walls [12]. Recently Edalati and Tong noted [13] that the bosonic part of the worldsheet sigma model on the string is, in fact, CP(N 1) C rather than CP(N 1), and − × − endowing CP(N 1) C with two supercharges need not necessarily lead to = − × N (2,2) supersymmetry on the worldsheet. They built an = (0,2) heterotic model N which supergeneralizes the bosonic model with the above target space. Moreover, 2 basing on a number of indirect checks they concluded that the Edalati–Tong heterotic modelemerges onthestringworldsheet inthe = 1 bulktheoryandsuggested arule N of converting the bulk = 2 breaking superpotential into an = (2,2) breaking N N superpotential on the string worldsheet. To be more exact, the Edalati–Tong model is designed as follows. Consider for example the U(2) model in the bulk with CP(1) on the worldsheet. If = 2 in the N bulk is unbroken, the 1/2 BPS flux tube has two translational moduli associated with its center x , and four supertranslational moduli. The above set is totally decoupled 0 from two orientational moduli parameterizing the coset SU(2)/U(1) accompanied by four superorientational moduli. When = 2 is broken by an = 1-preserving deformation, the number of the N N moduli fields remains intact, but their grouping changes. The four supertranslational moduli split into two plus two. Two left-handed fermion fields combine with x to 0 forman = (0,2) supermultiplet. These fields aredescribed by a free theory andare N decoupled fromtherest oftheworldsheet theory. (Generalaspectsoftwo-dimensional = (0,2) sigma models were discussed in [14].) N ¯ The right-handed fermion fields ζ and ζ , which used to be “two other” su- R R pertranslational moduli, “mix” with two right-handed superorientational moduli tan- gential to the coset SU(2)/U(1). Together with two orientational moduli of CP(1) and four superorientational moduli they form the = (0,2) extension of the CP(1) N model. For brevity sometimes we will refer to it as the heterotic CP(1) (or heterotic ¯ CP(N 1) for generic N). The fermion fields ζ and ζ lie outside the target space R R − SU(2)/U(1). They are remnants of C. With respect to = (0,2) supersymmetry N they transform through F terms which are expressible, via equations of motion, in ζ terms of the fermion fields of the conventional CP(1) model. In this paper we present a direct derivation of the string worldsheet theory for a generic superpotential in the bulk theory breaking = 2 while preserving = N N 1 and the 1/2-BPS nature of the flux tube solution at the classical level. The µ 2 A superpotential mentioned above is a particular case. Generally speaking, the minimal choice one can consider is a cubic in superpotential (in the U(2) bulk theory) A with coefficients rigidly fixed by the quark mass terms. In the U(N) bulk theory with N = N flavors the minimal admissible = 1 -preserving deformation is a f N polynomial of the (N+1)-th order whose coefficients are unambiguously fixed. These more general superpotentials will be considered as well. Focusing on the simplest example of U(2) in the bulk we prove that an = N (0,2) extension of the CP(1) model `a la Edalati–Tong does indeed emerge on the string worldsheet in the low-energy limit. While gross features of the emergent het- erotic worldsheet theory are those predicted by Edalati and Tong, details do not quite 3 coincide. In particular, for polynomial deformations in the bulk we find, generally speaking, a non-polynomial response in the worldsheet theory. Our direct derivation of the heterotic string model relies, in addition to already known results, on explicit form of the fermion zero modes on the BPS flux tubes in = 1 bulk theories. To N obtain the fermion zero modes we had to extend previous analyses [11, 15]. Thus, a large part of this paper bears a technical nature. It is based, however, on an ob- servation of conceptual nature (Sect. 5) which is responsible for the very possibility of direct derivation of the heterotic CP(1) model on the string worldsheet. Indeed, in the Edalati–Tong formulation the difference between the = (2,2) and heterotic N models reveals itself in four-fermion terms. It is very hard, if possible at all, to derive these terms starting directly from the bulk theory. In our formulation the most straightforward distinction between two models occurs in the kinetic part of the Lagrangian, in the term bilinear in the fermion fields, of the type ζ†χa ∂ Sa, (1.2) R R L (cid:16) (cid:17) where Sa is the bosonic field of the O(3) model subject to the constraint S~2 = 1, while χa is its fermionic superpartner, S~χ~ = 0. Since the term in (1.2) is bilinear R in the fermion fields, the knowledge of the fermion zero modes allows one to get this term from the bulk Lagrangian in a very explicit and direct way. Other additional terms transforming = (2,2) model into = (0,2) unambiguously follow from N N (1.2) by virtue of = (0,2) supersymmetry. N The basic features of the heterotic CP(1) model we obtain are as follows. The term (1.2) entails the occurrence of the four-fermion interaction of the type ζ†ζ χaχb Scε , (1.3) R R L L abc (cid:16) (cid:17)(cid:16) (cid:17) and a suppression of the coefficient in front of the conventional four-fermion term 1 (χaχa)2 . (1.4) 2 L R ¯ The addition of seemingly rather insignificant ζ ,ζ terms to the = (2,2) R R N CP(N 1) model drastically changes its dynamical behavior. In particular, Witten’s − index I = N for CP(N 1) [16] changes and becomes zero. Supersymmetry on − the worldsheet is no longer protected by Witten’s index. In fact, we will prove, at small µ, that spontaneous SUSY breaking does take place. The fields ζ , ζ† play R R the role of Goldstinos. In the accompanying paper [17] we will solve the heterotic CP(N 1) model at large N and prove that supersymmetry is spontaneously broken − at the quantum level for any value of the deformation parameter, as was anticipated 4 by Tong [18]. This result seems to be intuitively clear given that small variations of the deformation superpotential ruin the BPS nature of the flux-tube solutions already at the classical level. We will derive a long-sought geometric representation of the heterotic = N (0,2) model, in terms of the metric and curvature tensor of the CP(N 1) space. − Organizationofthepaperisasfollows. InSect.2wereviewourbasicbulktheory witheight superchargesanddiscuss possibledeformationsofthisbulktheorybreaking = 2 down to = 1 without destroying the BPS nature of the flux-tube solution. N N In Sect. 3 we review construction of non-Abelian strings in the = (2,2) limit. N Moreover, we perform derivation of those fermion zero modes which had not been explicitly derived in the literature previously. Section 4 summarizes general aspects of the Edalati–Tong model. In Sect. 5 we present our formulation of the heterotic CP(1) model. Section 6 is devoted to yet another, geometric, formulation of the heterotic CP(1) model. Here we also show that at small µ the vacuum energy density of the heterotic model is proportional to the square of the chiral condensate. In Sect. 7 we begin our direct derivation of the worldsheet model from the bulk theory deformed by the superpotential (1.1). Section 7 is devoted to the fermion zero modes. Section 8establishes the relationbetween the parametersof theworldsheet modeland those of the bulk theory. In Sects. 9 and 10 we proceed to a more general case of a polynomial deformation superpotential replacing the simplest superpotential (1.1). Here we calculate the worldsheet superpotential in two limits, µ 0 and µ . → → ∞ While the first result agrees with the Edalati–Tong conjecture, the large-µ limit defies it. We show that in this case the main effect of = 2 breaking deformation in the N µ limit is that the potential of the worldsheet theory gets enhanced. It forces → ∞ the string orientational vector to point towards the north or south poles of the sphere S = SU(2)/U(1). The string becomes exceedingly more “Abelian” as we increase 2 the deformation superpotential in the bulk. Section 11 summarizes our findings. Remark: In Sects. 2–5 and 7–10 we use Euclidean notation most suitable for consideration of static solitons. This is explained in Appendix A. Section 6 which bears a general nature is presented in Minkowski notation. This is explained in Appendix B. In Appendix C we briefly discuss the Witten index for the heterotic = (0,2) CP(N 1) models. In Appendix D we collect for convenience various N − definitions of the deformation parameters. 5 2 Bulk theory The gauge symmetry of the basic bulk model is SU(N) U(1). We will focus on the × SU(2) U(1) case, which presents the simplest example. Besides the gauge bosons, × gauginos and their superpartners, the model has the matter sector consisting of N = f N = 2 “quark” hypermultiplets. In addition, we will introduce the Fayet–Iliopoulos D-term for the U(1) gauge field which triggers the quark condensation. Let us first discuss the undeformed theory with = 2. The superpotential has N the form 1 2 = q˜ qA +q˜ aτaqA , (2.1) WN=2 √2 AA AA AX=1(cid:16) (cid:17) where and a are chiral superfields, the = 2 superpartners of the gauge bosons A A N of U(1) and SU(2), respectively. Furthermore, q and q˜ (A = 1,2) represent two A A matter (quark) hypermultiplets. The flavor index is denoted by A. Thus, in our model the number of colors coincides with the number of flavors. The qA mass terms are denoted by m . A Next, we add a superpotential which breaks supersymmetry down to = 1. In N this paper we will consider two types of = 1 preserving deformation superpoten- N tials. The first superpotential is the mass term for the adjoint fields, µ = 2 +( a)2 , (2.2) 3+1 W 2 A A h i where µ is a common mass parameter for the chiral superfields in = 2 gauge N supermultiplets, U(1) and SU(2), respectively. The subscript 3+1 tells us that the deformation superpotential (2.2) refers to the bulk four-dimensional theory. Clearly, the mass term (2.2) splits = 2 supermultiplets, breaking = 2 supersymmetry N N down to = 1 . N For the deformation (2.2), in order to preserve the BPS nature of the flux-tube solutions, it is necessary to set the quark mass terms at zero, m = m = 0. (2.3) 1 2 As was shown in [11] and [13] (see also the review paper [8]), in this case the deformed theory supports 1/2 BPS -saturated flux-tube solutions at the classical level. The second (more general) deformation we will consider in this paper is a poly- nomial superpotential of the form N=2 c = Tr k ˆk+1, (2.4) 3+1 W k +1 A k=1 X 6 where we introduce the adjoint matrix superfield 1 τa ˆ= + a; (2.5) A 2A 2 A τa are the SU(2) Pauli matrices. The hat over will remind us that is a matrix A A from U(2) rather than SU(2). The coefficients c are not arbitrary. As explained at k the end of this section, they are unambiguously fixed by the bulk theory parameters. The bosonic part of our SU(2) U(1) theory has the form × S = d4x 1 Fa 2 + 1 (F )2 + 1 D aa 2 + 1 ∂ a 2 Z "4g22 (cid:16) µν(cid:17) 4g12 µν g22 | µ | g12 | µ | + qA 2 + ¯q˜A 2 +V(qA,q˜ ,aa,a) . (2.6) µ µ A ∇ ∇ (cid:12) (cid:12) (cid:12) (cid:12) (cid:21) (cid:12) (cid:12) (cid:12) (cid:12) Here D is the cov(cid:12)ariant(cid:12) der(cid:12)ivative(cid:12) in the adjoint representation of SU(2), while µ i τa = ∂ A iAa . (2.7) ∇µ µ − 2 µ − µ 2 The coupling constants g and g correspond to the U(1) and SU(2) sectors respec- 1 2 tively. With our conventions, the U(1) charges of the fundamental matter fields are 1/2. ± The potential V(qA,q˜ ,aa,a) in the Lagrangian (2.6) is a sum of various D and A F terms, g2 1 τa τa 2 V(qA,q˜ ,aa,a) = 2 εabca¯bac +q¯ qA q˜ ¯q˜A A 2 g22 A 2 − A 2 ! + g12 q¯ qA q˜ ¯q˜A 2ξ 2 A A 8 − − (cid:16) (cid:17) g2 ∂ 2 g2 ∂ 2 + 2 q˜ τaqA +√2 W3+1 + 1 q˜ qA +√2 W3+1 2 (cid:12) A ∂aa (cid:12) 2 (cid:12) A ∂a (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 1 (cid:12)2 (cid:12) 2(cid:12) (cid:12) + (a+τaaa +√2mA)qA 2 AX=1(cid:26)(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 2 + (a+τaaa +√2mA)¯q˜ , (2.8) A (cid:12) (cid:12) (cid:27) (cid:12) (cid:12) where the sum over repeat(cid:12)ed flavor indices A is(cid:12)implied. The first and second lines here represent D terms, the third line the F terms, while the fourth and the fifth A lines represent the squark F terms. We also introduced the Fayet–Iliopoulos D-term 7 for the U(1) field, with the FI parameter ξ in (2.8). Note that the Fayet–Iliopoulos term does not break = 2 supersymmetry [19, 15]. The parameters which do break N = 2 down to = 1 are µ or c in (2.2) or (2.4). k N N The vacuum structure and the mass spectrum of perturbative excitations in this theory were studied in [11] for the case of mass-type deformation (2.2). Here we briefly review relevant results for convenience. The Fayet–Iliopoulos term triggers the spontaneous breaking of the gauge sym- metry. The vacuum expectation values (VEV’s) of the squark fields can be chosen as 1 0 qkA = ξ , ¯q˜kA = 0, h i 0 1 ! h i q k = 1,2, A = 1,2, (2.9) while the VEV’s of the adjoint fields vanish aa = 0, a = 0. (2.10) h i h i Here we write down q as a 2 2 matrix, the first superscript (k = 1,2) refers to SU(2) × color, while the second (A = 1,2) to flavor. We keep the quark masses m = m = 0 1 2 in conjunction with (2.2). The color-flavor locked form of the quark VEV’s in Eq. (2.9) and the absence of VEV of the adjoint scalar aa in Eq. (2.10) results in the fact that, while the theory is fully Higgsed, a diagonal SU(2) survives as a global symmetry. The presence C+F of this symmetry leads to the emergence of orientational zero modes of Z strings in 2 the model (2.6) [2]. With two matter hypermultiplets, the SU(2) part of the gauge group is asymp- totically free, implying generation of a dynamical scale Λ. In order to stay at weak coupling we assume that √ξ Λ, so that the SU(2) coupling running is frozen by ≫ the squark condensation at a small value. Since both U(1) and SU(2) gauge groups are broken by the squark condensation, all gauge bosons become massive. From (2.6) we get for the U(1) gauge boson m = g ξ, (2.11) γ 1 q while three gauge bosons of the SU(2) group acquire the same mass m = g ξ. (2.12) W 2 q To get the masses of the scalar bosons we expand the potential (2.8) near the vacuum (2.9), (2.10) and diagonalize the corresponding mass matrix. The four com- ponents of the eight-component1 scalar qkA are eaten by the Higgs mechanism for 1We mean here eight real components. 8 U(1) and SU(2) gauge groups. Another four components are split as follows: one component acquires the mass (2.11). It becomes a scalar component of a massive = 1 vector U(1) gauge multiplet. Other three components acquire masses (2.12) N and become scalar superpartners of the SU(2) gauge boson in = 1 massive gauge N supermultiplet. Other 16 real scalar components of the fields q˜ , aa and a produce the following Ak states: two states acquire mass m+ = g ξλ+, (2.13) U(1) 1 1 q while the mass of other two states is given by m− = g ξλ−, (2.14) U(1) 1 1 q where λ± are two roots of the quadratic equation 1 λ2 λ (2+ω2)+1 = 0, (2.15) i − i i for i = 1. Here we introduced two = 2 supersymmetry breaking parameters N associated with the U(1) and SU(2) gauge groups, respectively, g2µ g2µ ω = 1 , ω = 2 . (2.16) 1 2 m m γ W Furthermore, other 2 3=6 states acquire mass × m+ = g ξλ+, (2.17) SU(2) 2 2 q while the remaining 2 3=6 states also become massive. Their mass is × m− = g ξλ−. (2.18) SU(2) 2 2 q Here λ± are two roots of the quadratic equation (2.15) for i = 2. Note that all states 2 come either as singlets or triplets of unbroken SU(2) . C+F In the large-µ limit the larger masses m+ and m+ become U(1) SU(2) m+ = m ω = g2µ, m+ = m ω = g2µ. (2.19) U(1) U(1) 1 1 SU(2) SU(2) 2 2 Clearly, in the limit µ these are the masses of the heavy adjoint scalars a and → ∞ aa. At ω 1 these fields decouple and can be integrated out. i ≫ Thelow-energybulktheoryinthislimitcontainsmassivegauge = 1 multiplets N and chiral multiplets with lower masses m− . Equation (2.15) gives for these U(1),SU(2) masses m ξ m ξ m− = U(1) = , m− = SU(2) = . (2.20) U(1) ω µ SU(2) ω µ 1 2 9