LeClair (2007) reformulated the equilibrium statistical mechanics of quantum gases so that the two-body scattering amplitude, rather than the Hamiltonian, is the input: a pseudo-energy satisfies an integral equation whose kernel is built from the logarithm of the two-body forward amplitude, and the free energy is a functional of the resulting filling fraction. The paper states that a complete derivation of the exact thermodynamic Bethe ansatz (TBA) equations of integrable (1+1)-dimensional models from this formalism is not known. The open problem is to close this gap: derive the exact TBA equations (whose kernel acts on the dressed density) from the Dashen–Ma–Bernstein trace formula, cluster decomposition, and filling-fraction saddle point without assuming the Bethe ansatz or integrability. A 2024–2026 revival of S-matrix thermodynamics proves only second-order-in-fugacity agreement with TBA and documents multi-particle obstructions, so the exact bridge remains open.

Background

Equilibrium thermodynamics from scattering data has a long history. The Beth–Uhlenbeck formula gives the leading density correction to the free energy of a dilute quantum gas in terms of the derivative of the two-body scattering phase shifts, and Dashen, Ma and Bernstein (DMB) wrote the grand-canonical partition function of a relativistic gas as a trace over multi-particle scattering states, organizable as a cluster (fugacity) expansion in powers of eβμe^{\beta\mu}, where β\beta is the inverse temperature and μ\mu the chemical potential. In this framework the two-body S-matrix — the operator relating incoming to outgoing two-particle momentum eigenstates — is the only dynamical input, through its forward amplitude (evaluated at equal incoming and outgoing momenta).

LeClair (2007) turned this into a self-contained "S-matrix statistics" for both relativistic and non-relativistic gases. One introduces a pseudo-energy ε(k)\varepsilon(\mathbf{k}) obeying a closed integral equation

ε(k)=ω(k)−μ−∫ddk′(2π)d K(k,k′) f(k′),\varepsilon(\mathbf{k}) = \omega(\mathbf{k}) - \mu - \int \frac{d^d k'}{(2\pi)^d}\,\mathbf{K}(\mathbf{k},\mathbf{k}')\, f(\mathbf{k}'),

where ω\omega is the single-particle dispersion, f(k)=1/(eβε(k)−s)f(\mathbf{k}) = 1/(e^{\beta\varepsilon(\mathbf{k})} - s) is the filling fraction (s=+1s=+1 for bosons, s=−1s=-1 for fermions), and the kernel K\mathbf{K} is built from the logarithm of the two-body forward scattering amplitude. This is the two-body-level form of the general variational saddle point of the formalism, in which ε=ω−μ−1β ∂U/∂f\varepsilon = \omega - \mu - \tfrac{1}{\beta}\,\partial U/\partial f for a two-particle-irreducible functional UU. The grand-canonical free energy is then a variational functional of ff, exact for constant kernels and approximate (two-body level) in general, with no assumption of integrability. The generic non-constant-kernel case must be solved numerically or perturbatively.

Separately, for integrable (1+1)-dimensional models — systems such as the Lieb–Liniger Bose gas (bosons on a line with contact repulsion) or the sinh-Gordon model (an integrable relativistic field theory of a single scalar particle) — every multi-particle scattering event factorizes into two-body scatterings depending only on rapidity differences θ\theta (the parametrization p=m(cosh⁡θ,sinh⁡θ)p = m(\cosh\theta, \sinh\theta)), and the exact finite-density thermodynamics is governed by the thermodynamic Bethe ansatz (TBA) introduced by Yang and Yang: a pseudo-energy ε(θ)\varepsilon(\theta), taken here to be energy-valued (so that the dressed density is log⁡(1+e−βε)\log(1+e^{-\beta\varepsilon}) and T=1/βT = 1/\beta), satisfies the nonlinear integral equation

ε(θ)=mcosh⁡θ−μ−T∫dθ′2π [−i ∂θlog⁡S(θ−θ′)] log⁡ ⁣(1+e−βε(θ′)),\varepsilon(\theta) = m\cosh\theta - \mu - T\int \frac{d\theta'}{2\pi}\,\left[-i\,\partial_\theta \log S(\theta-\theta')\right]\,\log\!\left(1 + e^{-\beta\varepsilon(\theta')}\right),

and the free energy density is −m2πβ∫dθ cosh⁡θ log⁡(1+e−βε(θ))-\tfrac{m}{2\pi\beta}\int d\theta\,\cosh\theta\,\log(1+e^{-\beta\varepsilon(\theta)}) for bosons. For the non-relativistic Lieb–Liniger gas the same structure holds in momentum space: with exact two-body S-matrix S(p)=(p−ic)/(p+ic)S(p) = (p-ic)/(p+ic), where pp is the relative momentum and c>0c>0 the coupling,

ε(p)=p22m−μ−T∫dp′ K(p−p′) log⁡ ⁣(1+e−βε(p′)),K(p)=12πi ddplog⁡S(p)=1π cp2+c2,\varepsilon(p) = \frac{p^2}{2m} - \mu - T\int dp'\,K(p-p')\,\log\!\left(1 + e^{-\beta\varepsilon(p')}\right), \qquad K(p) = \frac{1}{2\pi i}\,\frac{d}{dp}\log S(p) = \frac{1}{\pi}\,\frac{c}{p^2+c^2},

and free energy density −T∫dp2π log⁡(1+e−βε(p))-T\int\frac{dp}{2\pi}\,\log(1+e^{-\beta\varepsilon(p)}); for sinh-Gordon the kernel is K(θ)=12πi ∂θlog⁡S(msinh⁡θ)=mcosh⁡θ KLL(msinh⁡θ)K(\theta) = \tfrac{1}{2\pi i}\,\partial_\theta \log S(m\sinh\theta) = m\cosh\theta\,K_{\mathrm{LL}}(m\sinh\theta), with pp in S(p)S(p) understood as msinh⁡θm\sinh\theta. TBA equations are normally derived from the Bethe ansatz solution of the many-body problem (the quantization conditions, called Bethe roots, that fix the allowed many-particle momenta on a large circle), or, in relativistic models, by self-consistently quantizing a gas on a circle (Zamolodchikov's construction).

The structural mismatch is the crux. In TBA the kernel −i∂θlog⁡S-i\partial_\theta\log S acts on the dressed density log⁡(1+e−βε)\log(1+e^{-\beta\varepsilon}), which is an infinite resummation of the bare two-body interaction, while LeClair's saddle-point equation applies the two-body kernel directly to the bare filling fraction; bridging them requires the three- and higher-particle connected clusters that the naive two-body DMB formula does not generate. Thacker (1977) took early steps relating S-matrix methods to finite-density thermodynamics in 1+1D. LeClair's paper states (footnote [26]): "To our knowledge, a complete derivation of the thermodynamic Bethe ansatz equations from this formalism is not known." Whether this pseudo-energy-to-TBA bridge has been closed in the two decades since is the subject of this problem.

Problem Statement

Starting from the S-matrix/filling-fraction formalism of LeClair (2007) — the Dashen–Ma–Bernstein trace formula for the grand-canonical partition function, its cluster (fugacity) decomposition, and the filling-fraction saddle-point equation ε(k)=ω(k)−μ−∫ddk′(2π)d K(k,k′) f(k′)\varepsilon(\mathbf{k}) = \omega(\mathbf{k}) - \mu - \int \frac{d^d k'}{(2\pi)^d}\,\mathbf{K}(\mathbf{k},\mathbf{k}')\, f(\mathbf{k}') (the two-body-level form of the general variational saddle point) with f(k)=1/(eβε(k)−s)f(\mathbf{k}) = 1/(e^{\beta\varepsilon(\mathbf{k})}-s), s=+1s=+1 for bosons and s=−1s=-1 for fermions, and K\mathbf{K} built from the two-body forward scattering amplitude — derive the exact thermodynamic Bethe ansatz (TBA) equations of integrable (1+1)-dimensional models, without assuming the Bethe ansatz or factorized (integrable) scattering as an input. Supplying the exact two-body S-matrix of an integrable model as scattering data is part of the formalism's input, as it is of the TBA equations themselves; what may not be used as input are Bethe-ansatz quantization conditions or the factorization of multi-particle scattering. The target is a derivation for integrable (1+1)-dimensional models generally, not a model-by-model result. In the diagonal scattering case it must, at minimum for the Lieb–Liniger and sinh-Gordon models, reproduce the exact TBA pseudo-energy equation with kernel −i ∂θlog⁡S-i\,\partial_\theta \log S — in the sign and normalization conventions fixed in the Background, and with the corresponding momentum-space kernel K(p)=12πi ddplog⁡S(p)K(p) = \tfrac{1}{2\pi i}\,\tfrac{d}{dp}\log S(p) for Lieb–Liniger — acting on the dressed density log⁡(1+e−βε)\log(1+e^{-\beta\varepsilon}), together with the standard TBA free energy; and for non-diagonal scattering, where the two-body S-matrix mixes particle species so that several mutually dressed pseudo-energies are needed, it must reach the nested TBA equations, the coupled systems of pseudo-energy equations required in that case.

References

  1. A. LeClair, "Quantum statistical mechanics of gases in terms of dynamical filling fractions and scattering amplitudes", J. Phys. A 40 (2007) 9655, DOI 10.1088/1751-8113/40/31/033, arXiv:hep-th/0611187, https://arxiv.org/abs/hep-th/0611187
  2. R. Dashen, S.-K. Ma and H. J. Bernstein, "S-Matrix Formulation of Statistical Mechanics", Phys. Rev. 187 (1969) 345, DOI 10.1103/PhysRev.187.345, https://doi.org/10.1103/PhysRev.187.345
  3. C. N. Yang and C. P. Yang, "Thermodynamics of a one-dimensional system of bosons with repulsive delta-function interaction", J. Math. Phys. 10 (1969) 1115, DOI 10.1063/1.1664947, https://doi.org/10.1063/1.1664947
  4. Al. B. Zamolodchikov, "Thermodynamic Bethe ansatz in relativistic models: Scaling 3-state Potts and Lee-Yang models", Nucl. Phys. B 342 (1990) 695, DOI 10.1016/0550-3213(90)90333-9, https://doi.org/10.1016/0550-3213(90)90333-9
  5. H. B. Thacker, "Statistical mechanics of a (1+1)-dimensional quantum field theory at finite density and temperature", Phys. Rev. D 16 (1977) 2515, DOI 10.1103/PhysRevD.16.2515, https://doi.org/10.1103/PhysRevD.16.2515
  6. A. LeClair, "Interacting Bose and Fermi gases in low dimensions and the Riemann hypothesis", Int. J. Mod. Phys. A 23 (2008) 1371, DOI 10.1142/S0217751X08039451, arXiv:math-ph/0611043, https://arxiv.org/abs/math-ph/0611043
  7. Pye-Ton How and Andre LeClair, "Critical point of the two-dimensional Bose gas: An S-matrix approach", Nucl. Phys. B 824 (2010) 415, DOI 10.1016/j.nuclphysb.2009.07.022, arXiv:0906.0333, https://arxiv.org/abs/0906.0333
  8. Pye-Ton How and Andre LeClair, "S-matrix approach to quantum gases in the unitary limit: I. The two-dimensional case", J. Stat. Mech. (2010) P03025, DOI 10.1088/1742-5468/2010/03/P03025, arXiv:1001.1121, https://arxiv.org/abs/1001.1121
  9. Pye-Ton How and Andre LeClair, "S-matrix approach to quantum gases in the unitary limit: II. The three-dimensional case", J. Stat. Mech. (2010) P07001, DOI 10.1088/1742-5468/2010/07/P07001, arXiv:1004.5390, https://arxiv.org/abs/1004.5390
  10. Andre LeClair, "Thermodynamics of the Two-Dimensional Hubbard Model Based on the Exact Two-Body S-Matrix", Braz. J. Phys. 42 (2012) 28, DOI 10.1007/s13538-012-0068-z, arXiv:1007.1195, https://arxiv.org/abs/1007.1195
  11. Dinh-Long Vu, "A diagrammatic approach towards the thermodynamics of integrable systems", arXiv:2008.06901 (2020), https://arxiv.org/abs/2008.06901
  12. Daniel Schubring, "Free energy from forward scattering in 1+1d", JHEP 03 (2026) 030, DOI 10.1007/JHEP03(2026)030, arXiv:2408.00729, https://arxiv.org/abs/2408.00729
  13. Pietro Baratella, Joan Elias Miró and Emanuele Gendy, "Thermodynamics from the S-matrix reloaded, with applications to QCD and the confining Flux Tube", JHEP 12 (2024) 001, DOI 10.1007/JHEP12(2024)001, arXiv:2408.06729, https://arxiv.org/abs/2408.06729
  14. Pietro Baratella and Joan Elias Miró, "Thermodynamics from the S-matrix reloaded: emergent thermal mass", JHEP 03 (2026) 229, DOI 10.1007/JHEP03(2026)229, arXiv:2512.16466, https://arxiv.org/abs/2512.16466
  15. J. Balog and A. Hegedűs, "Virial expansion and TBA in O(N) sigma-models", Phys. Lett. B 523 (2001) 211, DOI 10.1016/S0370-2693(01)01307-7, https://doi.org/10.1016/S0370-2693(01)01307-7
  16. Pok Man Lo and Francesco Giacosa, "Thermal contribution of unstable states", Eur. Phys. J. C 79 (2019) 336, DOI 10.1140/epjc/s10052-019-6844-y, arXiv:1902.03203, https://arxiv.org/abs/1902.03203