Caetano and Komatsu (2020) reformulated the exact g-function (Affleck-Ludwig boundary entropy) of the sinh-Gordon model at the self-dual coupling, on a cylinder of circumference L, as a coupled system of nonlinear integral equations of the thermodynamic Bethe ansatz (TBA) type -- the Tracy-Widom TBA -- whose input is the ordinary sinh-Gordon TBA pseudo-energy and whose output is the ratio of Fredholm determinants that represents the g-function. The equations are numerically efficient, but the ultraviolet limit L -> 0, where the theory degenerates to boundary Liouville conformal field theory, is understood only numerically and qualitatively: the difference of g-functions for two boundary parameters agrees only qualitatively, near zero Liouville momentum, with a combination of Liouville bulk one-point functions, the momentum being fixed by an approximate quantization condition and the state normalization being unresolved. The open problem is to solve the Tracy-Widom TBA analytically in the L -> 0 limit -- in the spirit of the classic analytic derivation of the sinh-Gordon ultraviolet central charge and of the match between the sinh-Gordon boundary TBA ground-state energy and the boundary Liouville reflection amplitude -- and thereby derive controlled ultraviolet asymptotics of the g-function that make the comparison with Liouville boundary data quantitative, including the pole-subtraction prescription at zero momentum and the discrete-to-continuum normalization. No existing work performs this L -> 0 analysis; the problem is open.

Background

Two-dimensional quantum field theories with a boundary carry a boundary contribution to the entropy: the g-function (Affleck-Ludwig boundary entropy), a number gg counting the degrees of freedom residing on the boundary. It is exactly known at conformal fixed points but flows away from them, so in a massive theory its finite-size value must be extracted by exact nonperturbative methods.

The sinh-Gordon model is the simplest interacting integrable quantum field theory in 1+1 dimensions: a single neutral particle of mass mm whose two-body scattering matrix S(u,v)S(u,v) depends on a coupling bb (throughout, at the self-dual point b=1b=1). Integrable boundary conditions preserve factorized scattering and are characterized by a reflection factor RR -- the amplitude for the particle to bounce off the wall -- depending on parameters (η,ϑ)(\eta,\vartheta); the reparametrization s=bη/πs=b\eta/\pi is the convenient one. On a cylinder of circumference LL (the closed-string channel; the open-string channel swaps space and Euclidean time), the exact ground-state g-function admits a Fredholm-determinant representation: a ratio of determinants of integral operators whose kernels are built from ∂ulog⁡S\partial_u\log S and ∂ulog⁡R\partial_u\log R, and whose entries are dressed by a pseudo-energy ε(u)\varepsilon(u) that itself solves the standard thermodynamic Bethe ansatz (TBA) equation of the sinh-Gordon model (TBA being the standard exact method that computes finite-size energies from infinite-volume scattering data). This representation, conjectured in unpublished work and developed by Kostov-Serban-Vu and by Jiang-Komatsu-Vescovi, is exact but numerically awkward because it requires multiple integrals.

Caetano and Komatsu reformulated these determinants as the solution of a system of functional equations -- equivalently nonlinear integral equations -- which they named the Tracy-Widom TBA, because Tracy and Widom proved the underlying observation (first made in the context of two-dimensional polymers, and much used in topological string theory) that Fredholm determinants whose kernels have the special form E(u)E(v)/(M(u)+M(v))E(u)E(v)/(M(u)+M(v)) obey TBA-like nonlinear integral equations. The g-function system is double-layer: one first solves the ordinary sinh-Gordon TBA for ε(u)\varepsilon(u), then solves a second layer of integral equations (for auxiliary functions η⋆(u)\eta_\star(u) with ⋆∈{s,+}\star\in\{s,+\}, or equivalently even/odd resolvent traces of the dressed kernels) that feed on it. This is far more efficient numerically than the determinants, but it is a new and analytically uncontrolled object.

The ultraviolet (UV) regime is L→0L\to 0. There the bulk interaction degenerates and the theory is expected to reduce to boundary Liouville conformal field theory -- the exactly solved theory of a scalar field with an exponential potential, whose boundary one-point functions of bulk primary fields are known in closed form (Fateev-Zamolodchikov-Zamolodchikov). Caetano-Komatsu analyzed this limit in the minisuperspace approximation (the quantum mechanics of the zero mode of the field): the Liouville momentum PP is quantized by an approximate classical condition and tends to zero only logarithmically, P∼−1/log⁡LP\sim-1/\log L, for the ground state; and they proposed to compare the difference of g-functions for two boundary parameters with the combination Us1,s0(P)U_{s_1,s_0}(P) of Liouville bulk one-point functions, with an explicit pole subtraction at P=0P=0 (their eq. (71)). Numerically the agreement is only qualitative near P=0P=0, and the conversion between the discrete finite-volume momentum normalization and the continuum Liouville normalization is unresolved (their footnote 10).

Two analytic prototypes exist for exactly this kind of question: the classic derivation of the sinh-Gordon UV central charge from the TBA, and the analysis of Bajnok-Rim-Zamolodchikov, who extracted the UV behavior of the sinh-Gordon boundary TBA ground-state energy and matched it to the boundary Liouville reflection amplitude in perfect quantitative agreement. What does not exist is the analogous analytic control for the Tracy-Widom (double-layer) TBA and the g-function.

Problem Statement

Starting from the Tracy-Widom TBA system of Caetano-Komatsu for the boundary sinh-Gordon g-function at the self-dual coupling -- i.e. the standard sinh-Gordon TBA equation for the pseudo-energy ε(u)\varepsilon(u) together with the second-layer nonlinear integral/functional equations for the auxiliary functions η⋆(u)\eta_\star(u), ⋆∈{s,+}\star\in\{s,+\} (equivalently the even/odd resolvent traces of the dressed kernels), exactly as pinned by equations (11), (44)-(46) and the kernels (21)-(22) of arXiv:2004.05071 -- derive its ultraviolet (small-circumference) limit L→0L\to 0 analytically, and from it obtain controlled asymptotics of the ground-state g-function: in particular the L→0L\to 0 expansion of log⁡g\log g and of its differences log⁡gs1−log⁡gs0\log g_{s_1}-\log g_{s_0} between boundary parameters, including the leading logarithmic dependence on the Liouville momentum P(L)∼−1/log⁡(L/2π)→0P(L)\sim-1/\log(L/2\pi)\to 0.

Use these asymptotics to make the proposed identification with boundary Liouville conformal field theory quantitative, in place of the currently qualitative numerical match near P=0P=0: determine whether the combination of Liouville bulk one-point functions Us1,s0(P)U_{s_1,s_0}(P) (eq. (71) of the source) reproduces the ultraviolet g-function differences with explicit coefficients, and fix (i) the pole-subtraction prescription for the Liouville one-point function at P=0P=0 and (ii) the discrete-to-continuum normalization of the bulk states, i.e. whether a Gaudin-like factor built from the momentum quantization condition must be included.

A complete answer consists of: the degenerate form of the coupled integral equations at L→0L\to 0 and its solution; the resulting explicit ultraviolet asymptotics of log⁡g\log g with error control; and the quantitative statement of the match -- or of a precisely located discrepancy -- with the Liouville prediction. This problem is the ultraviolet core of the open directions listed by Caetano-Komatsu; the companion directions (excited-state g-functions via analytic continuation of the Tracy-Widom TBA; generalization to theories with bound states or non-diagonal scattering; the normalization constant N\mathcal{N} of the separation-of-variables conjecture) are deliberately outside its completion boundary.

References

  1. J. Caetano and S. Komatsu, Functional Equations and Separation of Variables for Exact g-Function, JHEP 09 (2020) 180, arXiv:2004.05071, DOI 10.1007/JHEP09(2020)180, https://arxiv.org/abs/2004.05071
  2. Z. Bajnok, C. Rim and A. Zamolodchikov, Sinh-Gordon Boundary TBA and Boundary Liouville Reflection Amplitude, Nucl. Phys. B796 (2008) 622, arXiv:0710.4789, DOI 10.1016/j.nuclphysb.2007.12.023, https://arxiv.org/abs/0710.4789
  3. C. A. Tracy and H. Widom, Proofs of Two Conjectures Related to the Thermodynamic Bethe Ansatz, Commun. Math. Phys. 179 (1996) 667, arXiv:solv-int/9509003, DOI 10.1007/BF02100102, https://arxiv.org/abs/solv-int/9509003
  4. V. Fateev, A. Zamolodchikov and Al. Zamolodchikov, Boundary Liouville Field Theory I. Boundary State and Boundary Two-point Function, arXiv:hep-th/0001012, https://arxiv.org/abs/hep-th/0001012
  5. I. Affleck and A. W. W. Ludwig, Critical theory of overscreened Kondo fixed points, Nucl. Phys. B360 (1991) 641, DOI 10.1016/0550-3213(91)90419-X, https://doi.org/10.1016/0550-3213(91)90419-X
  6. I. Kostov, D. Serban and D.-L. Vu, Boundary TBA, trees and loops, Nucl. Phys. B949 (2019) 114817, arXiv:1809.05705, DOI 10.1016/j.nuclphysb.2019.114817, https://arxiv.org/abs/1809.05705
  7. Y. Jiang, S. Komatsu and E. Vescovi, Structure Constants in N=4 SYM at Finite Coupling as Worldsheet g-Function, JHEP 07 (2020) 037, arXiv:1906.07733, DOI 10.1007/JHEP07(2020)037, https://arxiv.org/abs/1906.07733
  8. J. Caetano and S. Komatsu, Crosscap States in Integrable Field Theories and Spin Chains, J. Stat. Phys. 187 (2022) 30, arXiv:2111.09901, DOI 10.1007/s10955-022-02914-6, https://arxiv.org/abs/2111.09901
  9. Z. Bajnok and T. L. Tompa, TCSA and the finite volume boundary state, Nucl. Phys. B964 (2021) 115330, arXiv:2008.01979, DOI 10.1016/j.nuclphysb.2021.115330, https://arxiv.org/abs/2008.01979
  10. Y.-J. He and Y. Jiang, Exact g-function without strings, arXiv:2412.12869 (2024), https://arxiv.org/abs/2412.12869
  11. N. Brizio, T. Morone and R. Tateo, A Note on T Tbar Deformations and Boundaries, arXiv:2410.20947 (2024), https://arxiv.org/abs/2410.20947
  12. O. A. Castro-Alvaredo, S. Negro and F. Sailis, Boundary Quantum Field Theories Perturbed by T Tbar: Towards a Form Factor Program, Nucl. Phys. B1017 (2025) 116924, arXiv:2501.11647, DOI 10.1016/j.nuclphysb.2025.116924, https://arxiv.org/abs/2501.11647
  13. S. Ekhammar, N. Gromov and P. Ryan, Boundary Overlaps from Functional Separation of Variables, JHEP 05 (2024) 268, arXiv:2312.11612, DOI 10.1007/JHEP05(2024)268, https://arxiv.org/abs/2312.11612
  14. C. Bercini, A. Homrich and P. Vieira, Structure Constants in N=4 Supersymmetric Yang-Mills Theory and Separation of Variables, Phys. Rev. D 110 (2024) L121901, arXiv:2210.04923, DOI 10.1103/PhysRevD.110.L121901, https://arxiv.org/abs/2210.04923
  15. T. Bargheer, C. Bercini, A. Cavaglià, D. Lai and P. Ryan, Orthogonality of Q-Functions up to Wrapping in Planar N=4 Super Yang-Mills Theory, SciPost Phys. 20 (2026) 064, arXiv:2507.00110, DOI 10.21468/SciPostPhys.20.2.064, https://arxiv.org/abs/2507.00110
  16. T. Bargheer, C. Bercini, G. Lefundes and P. Ryan, Structure Constants from Q-Systems and Separation of Variables, arXiv:2603.26617 (2026), https://arxiv.org/abs/2603.26617