Elsevier

Nuclear Physics B

Volume 873, Issue 3, 21 August 2013, Pages 682-727
Nuclear Physics B

On string theory on AdS3×S3×T4 with mixed 3-form flux: Tree-level S-matrix

https://doi.org/10.1016/j.nuclphysb.2013.05.005Get rights and content

Abstract

We consider superstring theory on AdS3×S3×T4 supported by a combination of RR and NSNS 3-form fluxes (with parameter of the NSNS 3-form q). This theory interpolates between the pure RR flux model (q=0) whose spectrum is expected to be described by a (thermodynamic) Bethe ansatz and the pure NSNS flux model (q=1) which is described by the supersymmetric extension of the SL(2,R)×SU(2) WZW model. As a first step towards the solution of this integrable theory for generic value of q we compute the corresponding tree-level S-matrix for massive BMN-type excitations. We find that this S-matrix has a surprisingly simple dependence on q: the diagonal amplitudes have exactly the same structure as in the q=0 case but with the BMN dispersion relation e2=p2+1 replaced by the one with shifted momentum and mass, e2=(p±q)2+1q2. The off-diagonal amplitudes are then determined from the classical Yang–Baxter equation. We also construct the Pohlmeyer-reduced model corresponding to this superstring theory and find that it depends on q only through the rescaled mass parameter, μ1q2μ, implying that its relativistic S-matrix is q-independent.

Introduction

The subject of this paper is the superstring theory on AdS3×S3×T4 space–time supported by a combination of RR and NSNS 3-form fluxes. The corresponding type IIB supergravity background is the near-horizon limit of the mixed NS5–NS1 + D5–D1 solution that is the basis for an interesting example of AdS/CFT duality [1], [2]. In the near-horizon limit the dilaton is constant, the 3-form fluxes through AdS3 and S3 have related coefficients and the radii of the two spaces are equal.

The S-duality symmetry of type IIB supergravity transforms the NSNS 3-form into the RR 3-form, so that if the coefficients of the NSNS and RR fluxes are chosen as q and q, respectively, then they enter symmetrically into the supergravity equations, e.g., as q2+q2=1 (we set the curvature radius to 1). The perturbative fundamental superstring theory is not invariant under the S-duality and should thus depend non-trivially on the parameter q.2

We shall assume that 0q1, with q=0 corresponding to the AdS3×S3×T4 theory with pure RR flux and q=1 – to the AdS3×S3×T4 theory with pure NSNS flux. The NSNS flux theory is given by the superstring generalization of the SL(2)×SU(2) WZW model while the RR flux theory has a Green–Schwarz (GS) formulation [3], [4] similar to the one [5] in the AdS5×S5 case. The “mixed” theory for generic value of q (first discussed in GS formulation in [3]) has a nice PSU(1,1|2)×PSU(1,1|2)SU(1,1)×SU(2) supercoset formulation [6], exposing its classical integrability and UV finiteness.3

The free-string spectrum of the NSNS (q=1) theory can be found using the chiral decomposition property of the WZW model [7] while the apparently more complicated spectral problem of the RR (q=0) theory is expected to be solved, as in the AdS5×S5 case [8], by a thermodynamic Bethe ansatz (for recent progress towards its construction see [9], [10], [11], [12], [13], [14], [15], [16], [17]).

Solving the “interpolating” theory with 0<q<1 is thus a very interesting problem as that may help to understand the relation between the more standard CFT approach in the q=1 case and the integrability-based TBA approach in the q1 case.4

The first step towards constructing the q0 generalization of the Bethe ansatz for the string spectrum is to find the corresponding S-matrix for the elementary (BMN-like) massive excitations. Our aim here will be to determine the string tree-level term in this S-matrix following the approach used in the AdS5×S5 case in [23], [24], [25], [26], [27], i.e. computing it directly from the gauge-fixed string action.

We shall start in Section 2 with an explicit description of the bosonic part of the string action in the sector corresponding to the string moving on R×S3. In conformal gauge it is described by the SU(2) principal chiral model with a WZ term (with coefficient proportional to q). Fixing a gauge corresponding the BMN vacuum in which the center of mass of the string moves along a circle in S3 we expand the action to quartic order in the fields, sufficient to compute the two-particle tree-level S-matrix.

The computation of this S-matrix for the bosonic string on S3 with B-field flux is the subject of Section 3. This S-matrix has a direct generalization to the full bosonic AdS3×S3 sector found by using the expression for the relevant gauge-fixed action given in Appendix A.1.

The simplicity of the bosonic result and the requirements of integrability (symmetry factorization and the Yang–Baxter equation) suggest a natural generalization to the fermionic sector, i.e. leading to the full tree-level S-matrix of the GS superstring theory on AdS3×S3×T4 with mixed RR–NSNS flux, which we present in the Section 4.

In Appendix A.2 we explain how the quadratic fermionic action that reproduces the non-trivial BBFF part of this S-matrix should follow from the AdS3×S3×T4 superstring action upon light-cone gauge fixing and expansion in powers of bosons. A candidate for the symmetry algebra of this S-matrix is discussed in Appendix B where we also comment on the symmetry of the S-matrix in the q=0 case. Section 5 contains some concluding remarks.

Imposing the conformal gauge, one may solve the Virasoro conditions explicitly and reformulate the classical string theory in terms of current field variables. One way to do this is the Pohlmeyer reduction and another is the Faddeev–Reshetikhin construction (that applies in the bosonic SU(2) case). In Appendix C we construct the Faddeev–Reshetikhin model corresponding to the bosonic string on R×S3 with B-flux, generalizing the discussion in [28], and present the q0 expression for the corresponding tree-level S-matrix (which is different from the bosonic string sigma model one).

In Appendix D we give a detailed construction of the Pohlmeyer-reduced model corresponding to the superstring theory on AdS3×S3×T4 with mixed 3-form flux. Somewhat unexpectedly, we find that the reduced action is the same as in the q=0 case [29], [30] but has the rescaled mass scale parameter, μ1q2μ. As a result, the corresponding relativistic S-matrix does not depend on q.

Section snippets

Bosonic string action on R×S3 with B-flux

Let us start with some basic definitions. In general, the bosonic string sigma model action is5S=12παdτdσL,L=12[ggabGmn(x)+ϵabBmn(x)]axmbxn. The action of the AdS3×S3×T4 superstring theory with mixed 3-form flux has the following structure:Stot=SAdS+SS+fermionic terms, where the first two terms are given by the principal chiral models with an extra WZ term for the groups SL(2,R) and SU(2) respectively. The WZ term represents the NSNS 3-form flux

Tree-level S-matrix of bosonic string on R×S3 with B-flux

Starting with the Lagrangian (2.30), (2.31) or (2.36), (2.37) it is straightforward (following [24], [25], [26]) to compute the corresponding tree-level 2-particle S-matrix for the elementary massive excitations of the bosonic string on S3 with non-zero B-field flux.

Despite the apparently complicated dependence of L4 on q (cf. Eq. (2.37)) we shall find that the resulting S-matrix has a very simple structure: its expression for q0 can be found from its q=0 limit by replacing e(p)=p2+1 with the

Tree-level S-matrix of AdS3×S3×T4 superstring with mixed flux

The bosonic-sector results of the previous section suggest a natural generalization to the full tree-level world-sheet S-matrix for the massive BMN modes of superstring theory on AdS3×S3×T4 with a mixed RR–NSNS flux.

Concluding remarks

In this paper we have found the generalization of the tree-level S-matrix for massive BMN-type excitations of the AdS3×S3×T4 superstring theory in the case of non-zero NSNS 3-form flux (parametrized by q(0,1)). We have directly computed the S-matrix in the bosonic sector discovering its very simple dependence on q via the modified dispersion relation. Using the requirements of integrability (factorization and Yang–Baxter properties of the S-matrix) we then suggested its generalization to the

Acknowledgements

We are grateful to A. Babichenko for many useful discussions and collaboration in the initial stages of this project. We would like to thank T. Klose, M. Kruczenski and R. Roiban for important explanations and comments and also S. Frolov, O. Ohlsson Sax, R. Roiban, A. Torrielli, L. Wulff and K. Zarembo for useful discussions and comments on the draft. We thank A. Sfondrini for pointing out misprints in Eq. (4.1) in the original version of this paper. B.H. is supported by the Emmy Noether

References (55)

  • B. Hoare et al.

    Towards the quantum S-matrix of the Pohlmeyer reduced version of AdS5×S5 superstring theory

    Nucl. Phys. B

    (2011)
  • V. Forini et al.

    Generalized cusp in AdS4×CP3 and more one-loop results from semiclassical strings

  • J.M. Maldacena

    The large N limit of superconformal field theories and supergravity

    Adv. Theor. Math. Phys.

    (1998)
  • A. Giveon et al.

    Comments on string theory on AdS3

    Adv. Theor. Math. Phys.

    (1998)
  • I. Pesando

    The GS type IIB superstring action on AdS3×S3×T4

    JHEP

    (1999)
  • J. Rahmfeld et al.

    The GS string action on AdS3×S3 with Ramond–Ramond charge

    Phys. Rev. D

    (1999)
    J. Park et al.

    Green–Schwarz superstring on AdS3×S3

    JHEP

    (1999)
    R.R. Metsaev et al.

    Superparticle and superstring in AdS3×S3 Ramond–Ramond background in light cone gauge

    J. Math. Phys.

    (2001)
  • R.R. Metsaev et al.

    Type IIB superstring action in AdS5×S5 background

    Nucl. Phys. B

    (1998)
  • A. Cagnazzo et al.

    B-field in AdS3/CFT2 correspondence and integrability

    JHEP

    (2012)
  • N. Beisert et al.

    Review of AdS/CFT integrability: An overview

    Lett. Math. Phys.

    (2012)
  • A. Babichenko et al.

    Integrability and the AdS3/CFT2 correspondence

    JHEP

    (2010)
  • J.R. David et al.

    S-matrix for magnons in the D1–D5 system

    JHEP

    (2010)
    J.R. David et al.

    Giant magnons in the D1–D5 system

    JHEP

    (2008)
  • O. Ohlsson Sax et al.

    Integrability, spin-chains and the AdS3/CFT2 correspondence

    JHEP

    (2011)
    O.O. Sax et al.

    On the massless modes of the AdS3/CFT2 integrable systems

  • N. Rughoonauth et al.

    Near BMN dynamics of the AdS3×S3×S3×S1 superstring

    JHEP

    (2012)
    P. Sundin et al.

    Classical integrability and quantum aspects of the AdS3×S3×S3×S1 superstring

    JHEP

    (2012)
  • C. Ahn et al.

    Exact S-matrices for AdS3/CFT2

  • R. Borsato et al.

    All-loop Bethe ansatz equations for AdS3/CFT2

    R. Borsato et al.

    A dynamic su(1|1)2 S-matrix for AdS3/CFT2

  • M.C. Abbott

    Comment on strings in AdS3×S3×S3×S1 at one loop

    JHEP

    (2013)
  • M. Beccaria et al.

    Quantum corrections to spinning superstrings in AdS3×S3×M4: Determining the dressing phase

  • Cited by (146)

    View all citing articles on Scopus
    1

    Also at Lebedev Institute, Moscow.

    View full text