In two-dimensional integrable sigma-models expanded around the trivial vacuum, all excitations are massless and the tree-level S-matrix generically exhibits particle production and fails to factorize, an effect attributed to infrared (IR) ambiguities of massless tree-level amplitudes. Georgiou (JHEP 01 (2025) 189) proposed a prescription that repairs this: expand the Lagrangian around a non-trivial vacuum that turns on an isometry coordinate linearly in worldsheet time with slope omega, compute scattering amplitudes with the resulting (massive, Lorentz-breaking) Feynman rules, and send omega to zero only at the end. At tree level, for the SU(2) principal chiral model, this yields an elastic, factorized, Lorentz-invariant massless S-matrix. The open problem is whether this prescription is valid and well-defined beyond the setting where it was demonstrated: at the quantum (loop) level, where new IR divergences appear and the diagrammatic cancellations must persist; for models with a B-field (an antisymmetric tensor coupling in addition to the target-space metric), several isometry parameters, or no continuous isometries; and in exceptional cases where the induced mass matrix degenerates. A resolution would establish the massless S-matrix as a well-founded probe of integrability and provide the foundation for a proposed classification of massless integrable sigma-models by order-by-order absence of particle production.

Background

A two-dimensional sigma-model is a field theory whose fields map a two-dimensional spacetime (the worldsheet) into a target space equipped with a metric; the principal chiral model (PCM), the paradigmatic example, takes the fields to be group-valued functions g(τ,σ)g(\tau,\sigma) with action proportional to ∫dτ dσ Tr(∂μg−1∂μg)\int d\tau\, d\sigma\, \mathrm{Tr}(\partial_\mu g^{-1}\partial^\mu g). A sigma-model is classically integrable when its equations of motion can be rewritten as the flatness condition of a Lax connection, a gauge connection depending on an auxiliary spectral parameter; flatness yields infinitely many conserved charges. For theories with massive excitations, quantum integrability implies very strong constraints on scattering: there is no particle production (the number of particles and the set of momenta are conserved, so only elastic processes like 2→22\to2 occur) and the S-matrix factorizes into a product of two-body S-matrices. This was established by Luscher (1978) and Parke (1980) under the essential assumption that the scattered particles are massive, so that wave packets can be spatially separated.

Sigma-models such as the PCM are usually expanded around the trivial vacuum g=const.g=\mathrm{const.}, around which all excitations are massless. Surprisingly, for such massless excitations the link between integrability and scattering breaks down: since Nappi (1980) it has been known that massless tree-level amplitudes can show particle production, and Hoare, Levine and Tseytlin (2019) demonstrated that in well-known integrable sigma-models the tree-level massless S-matrix exhibits particle production and fails to factorize. The accepted diagnosis is that massless tree-level amplitudes suffer from infrared (IR) ambiguities: for special kinematic configurations, internal propagators go on-shell at the same time as the vertices connected to them vanish, producing indeterminate 0/00/0 expressions whose value depends on the regularization. Crucially, the regularization schemes tried previously (the iϵi\epsilon prescription on the propagators, or artificially giving all particles a common mass) resolve the 0/00/0 ambiguities but are incompatible with the absence of particle production: production survives these regulators. This failure blocks a natural program — using absence of particle production as an order-by-order criterion to constrain the couplings and classify integrable sigma-models, in the spirit of Wulff's classification of integrable symmetric-space strings (strings whose target space is a symmetric coset G/HG/H) via factorized scattering. (Non-perturbatively, massless factorized scattering is not forbidden: Zamolodchikov and Zamolodchikov (1992) constructed such S-matrices for sigma-models with topological terms.)

Georgiou (JHEP 01 (2025) 189) proposed a definition of the massless S-matrix that reestablishes the link with integrability. The prescription: choose a coordinate xix^i along an isometry (a continuous symmetry direction of the target-space metric) and expand the action around the non-trivial vacuum solution xi=ωτx^i = \omega\tau, where τ\tau is worldsheet time and ω\omega a parameter, instead of around the trivial vacuum xi=0x^i=0. The excitations in the non-isometric directions then acquire masses proportional to ω\omega, the IR ambiguities are regulated, and Lorentz symmetry — broken intermediately by the moving vacuum — is restored in the final amplitudes after ω→0\omega\to0 is taken at the end of the calculation. For the SU(2)SU(2) PCM, Georgiou verified at tree level that all 2→32\to3 and 2→42\to4 amplitudes vanish, that the sets of incoming and outgoing momenta are equal, and that 3→33\to3 amplitudes factorize into products of 2→22\to2 amplitudes; the new vertices generated by the deformed vacuum are precisely what cancel the would-be production terms. The construction extends verbatim to any sigma-model whose target space has at least one isometry, and it is conjectured to apply even without isometries whenever a suitable one-parameter family of classical solutions exists; natural next targets include the lambda-deformed interpolations between exact conformal field theories and non-Abelian T-dual sigma-models (T-duality — the equivalence of string backgrounds under inversion of a compactification radius — has a non-Abelian version acting on group-valued fields) constructed by Sfetsos, whose integrability is likewise believed to persist at the quantum level. Earlier S-matrix computations around omega-type vacua exist (e.g. Georgiou and Sfetsos on pp-wave (plane-wave) backgrounds), but there ω\omega was never sent to zero and Lorentz symmetry remained explicitly broken.

Problem Statement

Establish whether Georgiou's omega-regulated definition of the massless S-matrix is valid beyond the setting in which it has been demonstrated (tree level, a single isometry turned on, no B-field, non-degenerate mass matrix), by resolving the following linked questions about the same construction.

  1. Loop level: does the prescription define an integrable massless S-matrix at the quantum level? Concretely, for the SU(2)SU(2) PCM — the model in which the prescription was demonstrated at tree level — compute the one-loop (and, if possible, higher-loop) amplitudes using the Feynman rules of the Lagrangian expanded around the vacuum xi=ωτx^i=\omega\tau, with ω→0\omega\to0 taken only at the end: do the resulting amplitudes remain free of particle production, admit only equal sets of incoming and outgoing momenta, factorize into two-body S-matrices, and come out Lorentz invariant — and if so, by what diagrammatic mechanism do the new omega-induced vertices cancel the would-be production terms loop by loop? A demonstration of failure at any loop order (for instance a one-loop amplitude that produces particles after ω→0\omega\to0, or a loop IR divergence not cured by the regulator) refutes the prescription at the quantum level.

  2. Generality of the model class: does the prescription extend to (a) integrable sigma-models with a non-zero BB-field, i.e. a target-space antisymmetric tensor BμνB_{\mu\nu} entering the action in addition to the metric, where the quadratic part of the omega-deformed Lagrangian mixes isometric and non-isometric modes and the amplitudes must be computed in the appropriate diagonal basis; (b) models with several isometries turned on with independent parameters ωi\omega^i, where it must be determined whether the amplitudes depend on ratios ωi/ωj\omega^i/\omega^j and whether the massless limit is unique and path-independent (including independent of the choice among multiple vacuum solutions); and (c) models without continuous isometries, where applicability requires constructing a one-parameter family of non-trivial classical solutions that degenerates continuously to the trivial vacuum and verifying the prescription on it?

  3. Exceptional cases: how does the prescription behave when the mass matrix induced on the non-isometric directions is degenerate or not positive definite — as in the Nappi-type model, where some modes stay massless even around the non-trivial vacuum — and must the prescription be modified or declared inapplicable in such cases?

The common objective is a single one: determine whether the omega-regulated massless S-matrix is a well-founded, model-independent definition of the scattering of integrable sigma-models with massless excitations (each component addresses one facet of validity or generality of the same construction). Completion is the joint resolution of these facets — the loop-level question (1), each part of the general model class 2(a)-(c), and the exceptional cases (3) — and answering only one facet, however completely, is a partial result rather than a completion of the task. The motivating application, stated by the source but not part of the completion boundary here, is to then use the regulated massless S-matrix to constrain the couplings of massless integrable sigma-models order by order in perturbation theory, as the first step of a classification program by absence of particle production.

References

  1. G. Georgiou, "The massless S-matrix of integrable sigma-models", JHEP 01 (2025) 189. DOI: 10.1007/JHEP01(2025)189; arXiv:2408.03673; https://arxiv.org/abs/2408.03673
  2. B. Hoare, N. Levine and A. A. Tseytlin, "On the massless tree-level S-matrix in 2d sigma models", J. Phys. A 52 (2019) 144005. DOI: 10.1088/1751-8121/ab0b79; arXiv:1812.02549; https://arxiv.org/abs/1812.02549
  3. M. Lüscher, "Quantum non-local charges and absence of particle production in the two-dimensional non-linear sigma-model", Nucl. Phys. B 135 (1978) 1. DOI: 10.1016/0550-3213(78)90211-0; https://doi.org/10.1016/0550-3213(78)90211-0
  4. S. J. Parke, "Absence of particle production and factorization of the S-matrix in 1+1 dimensional models", Nucl. Phys. B 174 (1980) 166. DOI: 10.1016/0550-3213(80)90196-0; https://doi.org/10.1016/0550-3213(80)90196-0
  5. C. R. Nappi, "Some properties of an analog of the chiral model", Phys. Rev. D 21 (1980) 418. DOI: 10.1103/PhysRevD.21.418; https://doi.org/10.1103/PhysRevD.21.418
  6. A. B. Zamolodchikov and Al. B. Zamolodchikov, "Massless factorized scattering and sigma models with topological terms", Nucl. Phys. B 379 (1992) 602. DOI: 10.1016/0550-3213(92)90136-Y; https://doi.org/10.1016/0550-3213(92)90136-Y
  7. L. Wulff, "Classifying integrable symmetric space strings via factorized scattering", JHEP 02 (2018) 106. DOI: 10.1007/JHEP02(2018)106; arXiv:1711.00296; https://arxiv.org/abs/1711.00296
  8. G. Georgiou and K. Sfetsos, "Scattering in integrable pp-wave backgrounds: S-matrix and absence of particle production", Nucl. Phys. B 987 (2023) 116096. DOI: 10.1016/j.nuclphysb.2023.116096; arXiv:2208.01072; https://arxiv.org/abs/2208.01072
  9. K. Sfetsos, "Integrable interpolations: From exact CFTs to non-Abelian T-duals", Nucl. Phys. B 880 (2014) 225. DOI: 10.1016/j.nuclphysb.2014.01.004; arXiv:1312.4560; https://arxiv.org/abs/1312.4560
  10. M. Fabri and D. Polvara, "One-loop elastic amplitudes from tree-level elasticity in 2d", JHEP 06 (2024) 104. DOI: 10.1007/JHEP06(2024)104; arXiv:2402.12087; https://arxiv.org/abs/2402.12087
  11. R. Borsato and S. Driezen, "Particle production in a light-cone gauge fixed Jordanian deformation of AdS5×S5AdS_5\times S^5", Phys. Rev. D 111 (2025) 086010. DOI: 10.1103/PhysRevD.111.086010; arXiv:2412.08411; https://arxiv.org/abs/2412.08411
  12. M. de Leeuw, A. Fontanella and J. M. Nieto García, "An integrable deformed Landau-Lifshitz model with particle production?", SciPost Phys. 20 (2026) 041. DOI: 10.21468/SciPostPhys.20.2.041; arXiv:2506.13598; https://arxiv.org/abs/2506.13598
  13. A. Antunes, N. Levine and M. Meineri, "Demystifying integrable QFTs in AdS: No-go theorems for higher-spin charges", SciPost Phys. 20 (2026) 088. DOI: 10.21468/SciPostPhys.20.3.088; arXiv:2502.06937; https://arxiv.org/abs/2502.06937