Completing the classification of two-dimensional Lagrangians with no tree-level particle production: multi-field seeds, non-Toda solutions, and quantum persistence

The 2018 work of Gabai, Mazac, Shieber, Vieira and Zhou showed that imposing absence of tree-level particle production on two-dimensional scalar Lagrangians, analyzed in a multi-Regge high-energy limit, yields recursion relations that fix all higher-point couplings in terms of the cubic and quartic ones; a seed-and-induction existence analysis recovered sine-Gordon, Bullough-Dodd, the rank-two affine Toda theories, and the $U(N)$ non-linear sigma model, but was completed only for a single field and for two fields under a genericity assumption. Subsequent literature — necessary-and-sufficient tree-level conditions and the area rule by Dorey and Polvara, one-loop elasticity theorems by Polvara and by Fabri and Polvara, the single-superfield supersymmetric classification by Bercini and Trancanelli, and a rigorous single-field proof by Bering — sharpens these gaps without closing them; the surviving open problem is to complete the seed-and-induction classification of no-production Lagrangians for arbitrary numbers of fields including the degenerate quartic-coupling kernels, prove the planar jet-amplitude induction for the colour-ordered two-derivative matrix ansatz, and determine whether tree-level elasticity persists beyond the one-loop order now established; in particular, whether any genuinely non-Toda production-free Lagrangian exists.

Background

In a generic relativistic quantum field theory, collisions at high enough energy produce additional particles. Two-dimensional integrable quantum field theories are the known exception: they possess an infinite tower of conserved charges of arbitrarily high spin, which forces the scattering to be purely elastic (every scattering process factorizes into two-body amplitudes) and forbids particle production, meaning that amplitudes such as $\mathcal{M}_{2\to4}$ vanish identically and $\mathcal{M}_{3\to3}$ is supported only on measure-zero factorization kinematics. In $d\geq3$ space-time dimensions, analyticity and unitarity make particle production unavoidable in any interacting theory, so two dimensions is the unique setting in which an interacting, gapped theory can be exactly production-free. Interest in how rigid this property is was sharpened by the two-dimensional S-matrix bootstrap of Paulos, Penedones, Toledo, van Rees and Vieira, which found that the theories maximizing couplings at a fixed mass spectrum are precisely the no-production theories, alongside numerical evidence for strikingly small (but nonzero) production in higher-dimensional bootstrap searches.

The tree-level version of the question has a long history: Arefeva and Korepin observed in 1974 that the Feynman diagrams contributing to production in the sine-Gordon model cancel order by order once the couplings are tuned; Dorey's review and the affine Toda perturbation theory of Braden and Sasaki publicized the recursive program of tuning each successive interaction vertex to cancel the next production amplitude. (Affine Toda field theories are scalar theories whose exponential interactions are labelled by an affine Dynkin diagram, the relativistic cousins of the Toda lattice, and form the main known class of massive integrable Lagrangians; sine-Gordon, equivalently sinh-Gordon, is the $a_1^{(1)}$ case with a $\cosh$ potential, and Bullough-Dodd is the $a_2^{(2)}$ case, whose potential combines two exponentials.) What was missing until 2018 was a complete analysis of this recursive procedure. Gabai, Mazac, Shieber, Vieira and Zhou introduced a multi-Regge high-energy limit — one incoming particle at rest, the others taken to infinite rapidity in geometric progression, $p_j=-m(x^{j-2},x^{-(j-2)})$ with $x\to\infty$ — which isolates the linear-chain exchange diagrams and converts the cancellation conditions into explicit recursion relations determining all higher-point vertices in terms of the cubic and quartic ones. For a single real massive scalar, the recursion supplemented by a seed analysis (proofs that particular sums of diagrams in the low-point amplitudes add up to kinematic constants, so that a contact term can cancel them) and an induction step resums the Lagrangian uniquely to sine-Gordon or Bullough-Dodd.

For $N$ massive real scalars with non-derivative interactions,

$$\mathcal{L}=\sum_{\alpha=1}^{N}\frac{1}{2}\,\partial_\mu\phi_\alpha\,\partial^\mu\phi_\alpha-\frac{1}{2}m_\alpha^2\phi_\alpha^2-\sum_{n=3}^{\infty}\frac{v_n^{\alpha_1\ldots\alpha_n}}{n!}\,\phi_{\alpha_1}\cdots\phi_{\alpha_n},$$

the same limit yields a recursion fixing all $n\geq5$ vertices from the cubic tensor $v_3$ and quartic tensor $v_4$, and imposing nontrivial constraints on those. The seed analysis was completed only for $N=2$ and only under a genericity assumption, namely that the linear system constraining the quartic couplings has full rank: the outcome is a set of six discrete solutions, which are exactly the affine Toda theories with two particles ($A_2$ with equal masses; $B_2=C_2$, $G_2$, $A_3^{(2)}$, $A_4^{(2)}$, $D_4^{(3)}$ with distinct mass ratios). The degenerate lower-rank cases found at $N=2$ correspond to decoupled pairs of sine-Gordon and Bullough-Dodd theories. For $N>2$ only the recursion was verified (for example on $B_4$ Toda); the seed problem was left open, as was the question of whether any genuinely non-Toda solution exists. For colour-ordered theories of massless Hermitian matrix-valued fields with two-derivative interactions,

$$\mathcal{L}(h)=\frac{1}{2}\mathrm{Tr}(\partial_\mu h\,\partial^\mu h)+\sum_{n=3}^{\infty}\sum_{k=1}^{\lfloor n/2\rfloor}g_{n,k}\,\mathrm{Tr}(h^{k-1}\,\partial_\mu h\,h^{\,n-k-1}\,\partial^\mu h),$$

where the physical particles are massless right- and left-movers that cluster into jets (groups of collinear particles of a single chirality), a recursion was derived in the planar (single-trace, large-$N$) limit, and it admits a single solution: the $U(N)$ non-linear sigma model (the theory of unitary matrix-valued fields with two-derivative interactions) in its Cayley parametrization (a rational parametrization of the unitary group-valued field in terms of the Hermitian field $h$); but the seed and induction proof that all higher planar jet amplitudes vanish was not given. (Single-scalar two-derivative theories, by contrast, are free in disguise, since their interactions can be removed by a field redefinition.)

Subsequent work sharpened several pieces of this program without completing it. Bering gave a rigorous proof, via factorization of the generating functional of connected tree diagrams as the Legendre transform of the action, that a single massive scalar without derivative couplings or tadpoles is production-free only for sinh-/sine-Gordon or Bullough-Dodd. Bercini and Trancanelli carried out the supersymmetric analysis announced in the source paper: for a single two-dimensional $\mathcal{N}=1$ scalar superfield with a $\mathbb{Z}_2$ symmetry, the only no-production superpotential is that of supersymmetric sinh-Gordon, while without the $\mathbb{Z}_2$ no supersymmetric Bullough-Dodd exists. Dorey and Polvara, for massive two-dimensional bosonic theories with polynomial interactions satisfying their simply-laced scattering conditions (poles in inelastic $2\to2$ processes cancel in pairs, and singularities in allowed processes arise from a single on-shell particle at a time), proved that vanishing of inelastic amplitudes implies the area rule — the three-point coupling $C^{(3)}_{abc}$ is a universal function of the area of the triangle with sides $(m_a,m_b,m_c)$ — together with a set of tree-level bootstrap relations, showed these conditions are necessary and sufficient, and verified them universally for affine Toda theories via root-system geometry; they explicitly left the full classification and the existence of non-Toda solutions open. In the loop direction, Polvara and Fabri established that for two-dimensional massive theories with polynomial-like interactions any one-loop inelastic amplitude equals its tree-level counterpart evaluated with one-loop shifted masses, so that tree-level elasticity implies one-loop elasticity once the physical masses are renormalized — verified on simply-laced and then on all non-simply-laced affine Toda theories. Hoare, Levine and Tseytlin added a caveat delimiting the massless sector: classically integrable massless sigma models expanded about the trivial vacuum exhibit tree-level production and infrared ambiguities.

Problem Statement

Determine the complete set of two-dimensional Lagrangians with no tree-level particle production, with existence proofs, and decide whether the absence of production survives quantization beyond the loop order already established. The problem concerns the two ansaetze fixed by the source work.

Part 1 (multi-field polynomial ansatz). Consider $N$ massive real scalar fields in $1+1$ dimensions with the most general non-derivative interaction Lagrangian

$$\mathcal{L}=\sum_{\alpha=1}^{N}\frac{1}{2}\,\partial_\mu\phi_\alpha\,\partial^\mu\phi_\alpha-\frac{1}{2}m_\alpha^2\phi_\alpha^2-\sum_{n=3}^{\infty}\frac{v_n^{\alpha_1\ldots\alpha_n}}{n!}\,\phi_{\alpha_1}\cdots\phi_{\alpha_n}.$$

Complete the seed-and-induction analysis for arbitrary $N\geq2$: determine all mass spectra and cubic and quartic couplings for which every tree-level production amplitude (every amplitude not supported purely on factorized elastic kinematics) vanishes. The analysis must cover the non-generic cases in which the linear system constraining the quartic couplings degenerates and admits a nonzero kernel (already of rank 0 and rank 3 at $N=2$), which the source could interpret only as decoupled sine-Gordon/Bullough-Dodd pairs. Either prove that every solution is a decoupled sum of free theories and affine Toda field theories — completing the classification and subsuming the six known $N=2$ Toda solutions, the $B_4$ recursion check, the exclusion of $O(N)$-symmetric theories, and the decoupled cases already known — or exhibit a genuinely new interacting Lagrangian together with a complete proof that all its tree-level production amplitudes vanish.

Part 2 (colour-ordered matrix ansatz). For massless Hermitian matrix-valued fields with two-derivative interactions

$$\mathcal{L}(h)=\frac{1}{2}\mathrm{Tr}(\partial_\mu h\,\partial^\mu h)+\sum_{n=3}^{\infty}\sum_{k=1}^{\lfloor n/2\rfloor}g_{n,k}\,\mathrm{Tr}(h^{k-1}\,\partial_\mu h\,h^{\,n-k-1}\,\partial^\mu h),$$

complete the missing seed and induction proof that all planar colour-ordered jet amplitudes with more than four jets — the production amplitudes, excluding the two-jet-to-two-jet elastic amplitudes, which remain nonzero even in an integrable theory — vanish, so that the Cayley-parametrized $U(N)$ non-linear sigma model is established, not merely conjectured from the necessary recursion, as the unique solution of the ansatz with this amplitude-vanishing pattern — or find other solutions.

Part 3 (quantum persistence). It is now known that tree-level elasticity implies one-loop elasticity once the physical masses are renormalized, for two-dimensional massive theories with polynomial-like interactions. Determine whether absence of particle production persists to all loop orders: prove the all-orders statement for the massive polynomial ansatz class (the Part 1 ansatz), or exhibit a tree-elastic theory in which a specific higher-loop inelastic amplitude is nonzero.

These parts share one objective: to convert the necessary conditions of the multi-Regge recursion program into a complete, existence-backed classification of classically production-free Lagrangians, and to fix whether that classical property determines the quantum theory. A further associated direction raised by the source — whether analogous high-energy constraints can explain the very small particle production observed numerically in higher-dimensional S-matrix bootstrap studies — is untouched by the literature and is treated here as motivation rather than as part of the completion boundary.

References

1. Barak Gabai, Dalimil Mazac, Andrei Shieber, Pedro Vieira, Yehao Zhou, "No Particle Production in Two Dimensions: Recursion Relations and Multi-Regge Limit", JHEP 02 (2019) 094, DOI 10.1007/JHEP02(2019)094, arXiv:1803.03578, https://arxiv.org/abs/1803.03578
2. Irina Arefeva, Vladimir Korepin, "Scattering in two-dimensional model with Lagrangian (1/gamma) ((d(mu)u)2/2 + m2 cos(u-1))", Pisma Zh. Eksp. Teor. Fiz. 20 (1974) 680, https://inspirehep.net/literature/95457
3. Patrick Dorey, "Exact S matrices", arXiv:hep-th/9810026, https://arxiv.org/abs/hep-th/9810026
4. H. W. Braden, R. Sasaki, "Affine Toda perturbation theory", Nuclear Physics B 379 (1992) 377, DOI 10.1016/0550-3213(92)90601-7, https://doi.org/10.1016/0550-3213(92)90601-7
5. Miguel F. Paulos, Joao Penedones, Jonathan Toledo, Balt C. van Rees, Pedro Vieira, "The S-matrix bootstrap II: two dimensional amplitudes", JHEP 11 (2017) 143, DOI 10.1007/JHEP11(2017)143, arXiv:1607.06110, https://arxiv.org/abs/1607.06110
6. Carlos Bercini, Diego Trancanelli, "Supersymmetric integrable theories without particle production", Physical Review D 97 (2018) 105013, DOI 10.1103/PhysRevD.97.105013, arXiv:1803.03612, https://arxiv.org/abs/1803.03612
7. Klaus Bering, "A note on tree factorization and no particle production", Journal of Physics A 56 (2023) 195401, DOI 10.1088/1751-8121/accb1e, arXiv:2207.13003, https://arxiv.org/abs/2207.13003
8. Patrick Dorey, Davide Polvara, "Tree level integrability in 2d quantum field theories and affine Toda models", JHEP 02 (2022) 199, DOI 10.1007/JHEP02(2022)199, arXiv:2111.02210, https://arxiv.org/abs/2111.02210
9. Patrick Dorey, Davide Polvara, "From tree- to loop-simplicity in affine Toda theories I: Landau singularities and their subleading coefficients", JHEP 09 (2022) 220, DOI 10.1007/JHEP09(2022)220, arXiv:2206.09368, https://arxiv.org/abs/2206.09368
10. Patrick Dorey, Davide Polvara, "From tree- to loop-simplicity in affine Toda theories II: higher-order poles and cut decompositions", JHEP 10 (2023) 177, DOI 10.1007/JHEP10(2023)177, arXiv:2307.15498, https://arxiv.org/abs/2307.15498
11. Davide Polvara, "One-loop inelastic amplitudes from tree-level elasticity in 2d", JHEP 04 (2023) 020, DOI 10.1007/JHEP04(2023)020, arXiv:2302.04709, https://arxiv.org/abs/2302.04709
12. Matheus Fabri, Davide 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
13. Matheus Fabri, Davide Polvara, "One-loop integrability with shifting masses", JHEP 07 (2025) 165, DOI 10.1007/JHEP07(2025)165, arXiv:2411.15080, https://arxiv.org/abs/2411.15080
14. Piotr Tourkine, Alexander Zhiboedov, "Scattering from production in 2d", JHEP 07 (2021) 228, DOI 10.1007/JHEP07(2021)228, arXiv:2101.05211, https://arxiv.org/abs/2101.05211
15. Ben Hoare, Nat Levine, Arkady A. Tseytlin, "On the massless tree-level S-matrix in 2d sigma models", Journal of Physics A 52 (2019) 144005, DOI 10.1088/1751-8121/ab0b79, arXiv:1812.02549, https://arxiv.org/abs/1812.02549
16. Clifford Cheung, James Mangan, Julio Parra-Martinez, Nabha Shah, "Nonperturbative Double Copy in Flatland", Physical Review Letters 129 (2022) 221602, DOI 10.1103/PhysRevLett.129.221602, arXiv:2204.07130, https://arxiv.org/abs/2204.07130