Continuum-limit identification of the staggered six-vertex model with integrable open boundaries: Euclidean or Lorentzian black-hole sigma model, and the emergence of the non-compact direction

The critical antiferromagnetic Potts model — equivalently the $\mathbb{Z}_2$-staggered six-vertex model or staggered XXZ spin chain — is the flagship example of a lattice model with finitely many states per site whose continuum limit is a non-compact conformal field theory (CFT) with a continuous spectrum of conformal dimensions. With periodic boundaries the continuum theory was identified with the $SL(2,\mathbb{R})/U(1)$ Euclidean black-hole sigma model, but later work indicates the lattice Hilbert space rather matches the Lorentzian black-hole sigma model. Robertson–Jacobsen–Saleur (JHEP 2021) extended the program to open integrable ($U_q(\mathfrak{sl}(2))$-invariant) boundaries, found both discrete and continuous parts of the boundary spectrum, and left open the density of states of the continuous part, the mechanism by which the non-compact direction emerges from compact lattice variables, and the identification of the continuum theory. The density of states and the full scaling-limit partition function have since been computed exactly by Frahm–Gehrmann–Kotousov (2024), but they do not appear to match any of the published brane partition functions of the Euclidean black-hole CFT, and the conformal boundary conditions of the Lorentzian model have never been constructed. The surviving problem is to construct the boundary CFT data (branes) of the candidate continuum theories and match them against — or exhaustively rule them out by comparison with — the exact lattice data, thereby settling the Euclidean-versus-Lorentzian identification and completing the microscopic explanation of how non-compactness emerges.

Background

Two-dimensional lattice models at criticality flow in the continuum limit to conformal field theories (CFTs) — quantum field theories invariant under angle-preserving maps whose spectra are organised by conformal dimensions $h$, the scaling exponents labelling energy levels at lattice size $L$ through $E\propto h-c/24$, with $c$ the central charge. For a lattice model with finitely many states per site the limiting CFT is normally 'compact': its conformal dimensions form a discrete set. The critical antiferromagnetic (AF) Potts model on the square lattice is a striking exception. It maps to the staggered six-vertex model — Baxter's inhomogeneous six-vertex model (an exactly solvable model of arrows on lattice edges) with spectral parameters arranged in a checkerboard pattern — whose Hamiltonian limit is the $\mathbb{Z}_2$-staggered spin-$\tfrac12$ XXZ chain, solvable by the Bethe ansatz (an algebraic method reducing eigenvalue problems to equations for auxiliary rapidities, called Bethe roots). In this model a subset of levels approaches its scaling value only logarithmically in $L$, so that as $L\to\infty$ the conformal dimensions fill a continuum: the continuum theory is non-compact.

Ikhlef–Jacobsen–Saleur identified this continuum theory as the two-dimensional Euclidean black-hole sigma model, the $SL(2,\mathbb{R})/U(1)$ gauged Wess–Zumino–Witten (WZW) model. A WZW model is a CFT whose field takes values in a group manifold; gauging a $U(1)$ subgroup of $SL(2,\mathbb{R})$ yields a sigma model whose target space is an infinitely long 'cigar' with a non-compact radial direction — the Euclidean black hole of two-dimensional string theory, whose CFT was worked out by Maldacena–Ooguri–Son. The lattice identification went beyond the continuous spectrum: the density of states agreed with the CFT formulas, providing a lattice 'measurement' of the reflection amplitude for a closed string scattering off the black hole.

A major refinement came from Bazhanov–Kotousov–Lukyanov and Bazhanov–Kotousov–Koval–Lukyanov. Using the ODE/IQFT correspondence — a technique relating the scaling limit of an integrable lattice model to the spectral problem of an ordinary differential equation — together with large-scale Bethe-ansatz numerics, they showed for (quasi-)periodic boundary conditions that one half of the lattice scaling-limit partition function coincides with the torus partition function of the Euclidean black-hole CFT, while the structure of the Hilbert space (global symmetries and the algebra of extended conformal symmetry, the $W_\infty$ algebra of infinitely many commuting charges) matches instead the Lorentzian black-hole sigma model: the gauged $SL(2,\mathbb{R})$ WZW model with certain reality and boundary conditions on the fields, whose space of states is a pseudo-Hilbert space with indefinite metric. The lattice model thus appeared to realise a Lorentzian-signature theory — something no other lattice regularisation had achieved.

All of this concerned periodic boundary conditions. Robertson–Jacobsen–Saleur (RJS) extended the program to boundaries: they constructed integrable open boundary conditions preserving the quantum group $U_q(\mathfrak{sl}(2))$ (the $q$-deformation of the $\mathfrak{sl}(2)$ symmetry algebra, with $q=e^{i\pi/k}$), built from the $D_2^{(2)}$ vertex-model transfer matrix (a six-vertex-type model associated with the twisted affine Lie algebra $D_2^{(2)}$ that is equivalent to the staggered six-vertex model). With these quantum-group-invariant boundaries the finite-size spectrum splits into a discrete part and a continuous part, and the loop-model partition function in the boundary sector labelled by $j$ takes the form $Z^{\rm cont}_{2j}=\mathfrak{q}^{l(l+2)/4k}\,\mathfrak{q}^{-c_{\rm PF}/24}\int_0^\infty{\rm d}s\,\rho_l(s)\,\mathfrak{q}^{s^2/(k-2)}$ with $l=k-2j-2$ and $c_{\rm PF}$ the central charge of the $\mathbb{Z}_{k-2}$ parafermion theory, where $\mathfrak{q}=e^{-\pi\tau}$ denotes the modular nome ($\tau$ the annulus modular parameter; it is a different symbol from the deformation parameter $q=e^{i\pi/k}$ above), and $\rho_l(s)$ is a density of states the authors explicitly left undetermined. They also exhibited a boundary renormalisation-group flow from the non-compact theory towards the compact $\mathbb{Z}_{k-2}$ parafermionic theory, whose generating functions are $su(2)/u(1)$ string functions (the characters of the $\mathbb{Z}_{k-2}$ parafermion CFT). Their conclusion posed the open questions addressed in this task: understand how a non-compact direction emerges from the compact lattice variables; determine $\rho_l(s)$; and settle whether the continuum limit is the Euclidean or the Lorentzian black-hole sigma model — 'hopefully, an analysis of the conformal boundary conditions in the latter model could help settle this question'.

Substantial progress followed on the lattice side. Frahm–Gehrmann showed the continuous component of the spectrum is present for a range of the staggering parameter and related the finite-size amplitudes of both continuous and discrete levels to the eigenvalues of a commuting quasi-momentum operator — a conserved charge that parametrises the emerging non-compact degree of freedom directly on the lattice; they also classified which staggerings admit local integrable open boundaries. The decisive step is due to Frahm–Gehrmann–Kotousov (FGK): combining the ODE/IQFT correspondence with an explicit new formula for the matrix elements of the Baxter $Q$ operator (an operator commuting with the transfer matrix whose eigenvalues generate the Bethe equations), they derived the quantisation condition $4s\log(L/L_0)+\delta(s)=2\pi m$ for the continuum levels — level spacing $\sim 1/\log L$ — the explicit density of states $\rho^{(\mathtt d)}_p(s)$, the discrete set of remaining levels, and the full scaling-limit partition function $Z^{\rm(scl)}=Z^{\rm(cont)}+Z^{\rm(disc)}$. On the string-theory side the conformal boundary conditions of the Euclidean black-hole CFT are known: they are branes (boundary states describing D-branes, the extended objects on which strings can end), classified by Ribault–Schomerus. But FGK found that their $Z^{\rm(scl)}$ 'does not seem to correspond to the published results in the literature concerning branes in the 2D Euclidean black hole CFT', and that relating it to the black-hole CFTs in the presence of boundaries 'would require a separate and detailed study'. The conformal boundary conditions of the Lorentzian black-hole sigma model have never been constructed. This is the surviving open problem.

Problem Statement

For the $\mathbb{Z}_2$-staggered six-vertex model at its self-dual point (the special checkerboard staggering, equivalent to the critical antiferromagnetic Potts model and to the $\mathbb{Z}_2$-staggered spin-$\tfrac12$ XXZ chain, at which the model gains an extra $\mathbb{Z}_2$ symmetry) with anisotropy $q=e^{i\pi/k}$ in the critical regime $0<\pi/k<\pi/2$, with integrable open boundary conditions — in the first instance the $U_q(\mathfrak{sl}(2))$-invariant family for which the scaling limit contains a continuous component — determine precisely which boundary conformal field theory the scaling limit realises, and thereby explain the microscopic emergence of the non-compact direction from the compact lattice variables. Concretely, an answer should:

1. Construct the conformal boundary conditions (brane boundary states) of the candidate continuum theories — the Euclidean black-hole sigma model $SL(2,\mathbb{R})/U(1)$, whose branes are known, and the Lorentzian black-hole sigma model (the gauged $SL(2,\mathbb{R})$ WZW model with reality conditions on the fields), for which they have not been constructed — and compute the corresponding boundary partition functions.

2. Match these continuum data against the lattice scaling-limit data that are now known exactly — the density of states $\rho^{(\mathtt d)}_p(s)$ as a function of the continuous-spectrum parameter $s\in\mathbb{R}$, the discrete spectrum, the degeneracies, and the partition function $Z^{\rm(scl)}=Z^{\rm(cont)}+Z^{\rm(disc)}$ of Frahm–Gehrmann–Kotousov — where $n=k-2$ (so that $q=e^{i\pi/(n+2)}$), $p$ is the boundary-sector label, $p=S-\tfrac{n+1}{2}$ with $S$ the $U_q(\mathfrak{sl}(2))$ spin of the sector (the same sector as the through-line label $j=S$ of the Background, hence $l=k-2j-2=-2p-1$, and $\rho_l(s)$ is the $\mathtt d=0$ case of $\rho^{(\mathtt d)}_p(s)$), and $\mathtt d=0,1,2,\dots$ counts the $W_\infty$-descendant tower — including the treatment of its $\log L$-divergent normalisation. A match must be operational: state which lattice and continuum quantities are compared and the normalisation that defines agreement, in particular how the $\log L$-divergent normalisation is handled. The primary completion standard is the identification of the realised boundary CFT, as stated in item 3, together with a demonstrated match of that theory's boundary partition function to the lattice data; the Euclidean and Lorentzian black-hole sigma models are the two candidates to be constructed and compared first, and if the lattice data exclude both, the answer must still name and match the boundary CFT that does describe the scaling limit. Exhaustively excluding the two candidates is therefore a step in the comparison, not a completion in itself.

3. State the resulting identification — Euclidean, Lorentzian, or another CFT — together with the dictionary between the lattice boundary data of the $U_q(\mathfrak{sl}(2))$-invariant family (its discrete labels) and the continuum branes, and the account of how the quasi-momentum quantisation $4s\log(L/L_0)+\delta(s)=2\pi m$ (level spacing $\sim 1/\log L$) turns compact, finite-dimensional lattice degrees of freedom into a non-compact continuum direction.

Two further questions raised in the same source conclusion — the explicit Hamiltonian of the apparently integrable free-boundary AF Potts model at the isotropic point, and the interpretation of the boundary renormalisation-group flow of string-function combinations for general $l$ — are separate side problems and are not part of the objective above.

References

1. N. F. Robertson, J. L. Jacobsen and H. Saleur, 'Lattice regularisation of a non-compact boundary conformal field theory', J. High Energy Phys. 02 (2021) 180, doi:10.1007/JHEP02(2021)180, arXiv:2012.07757, https://arxiv.org/abs/2012.07757
2. Y. Ikhlef, J. L. Jacobsen and H. Saleur, 'An integrable spin chain for the SL(2,R)/U(1) black hole sigma model', Phys. Rev. Lett. 108 (2012) 081601, doi:10.1103/PhysRevLett.108.081601, arXiv:1109.1119, https://arxiv.org/abs/1109.1119
3. J. Maldacena, H. Ooguri and J. Son, 'Strings in AdS_3 and the SL(2,R) WZW model. Part 2: Euclidean black hole', J. Math. Phys. 42 (2001) 2961, doi:10.1063/1.1377039, arXiv:hep-th/0005183, https://arxiv.org/abs/hep-th/0005183
4. S. Ribault and V. Schomerus, 'Branes in the 2D black hole', J. High Energy Phys. 02 (2004) 019, doi:10.1088/1126-6708/2004/02/019, arXiv:hep-th/0310024, https://arxiv.org/abs/hep-th/0310024
5. V. V. Bazhanov, G. A. Kotousov and S. L. Lukyanov, 'Equilibrium density matrices for the 2D black hole sigma models from an integrable spin chain', J. High Energy Phys. 03 (2021) 169, doi:10.1007/JHEP03(2021)169, arXiv:2010.10603, https://arxiv.org/abs/2010.10603
6. V. V. Bazhanov, G. A. Kotousov, S. M. Koval and S. L. Lukyanov, 'Scaling limit of the $\mathbb{Z}_2$ invariant inhomogeneous six-vertex model', Nucl. Phys. B 965 (2021) 115337, doi:10.1016/j.nuclphysb.2021.115337, arXiv:2010.10613, https://arxiv.org/abs/2010.10613
7. H. Frahm and S. Gehrmann, 'Finite size spectrum of the staggered six-vertex model with $U_q(\mathfrak{sl}(2))$-invariant boundary conditions', J. High Energy Phys. 01 (2022) 070, doi:10.1007/JHEP01(2022)070, arXiv:2111.00850, https://arxiv.org/abs/2111.00850
8. H. Frahm and S. Gehrmann, 'Integrable boundary conditions for staggered vertex models', J. Phys. A 56 (2023) 025001, doi:10.1088/1751-8121/acb29f, arXiv:2209.06182, https://arxiv.org/abs/2209.06182
9. H. Frahm, S. Gehrmann and G. A. Kotousov, 'Scaling limit of the staggered six-vertex model with $U_q(\mathfrak{sl}(2))$ invariant boundary conditions', SciPost Phys. 16 (2024) 149, doi:10.21468/SciPostPhys.16.6.149, arXiv:2312.11238, https://arxiv.org/abs/2312.11238
10. H. Frahm and S. Gehrmann, 'Finite-size spectrum of the staggered six-vertex model with antidiagonal boundary conditions', Nucl. Phys. B 1006 (2024) 116655, doi:10.1016/j.nuclphysb.2024.116655, arXiv:2405.20919, https://arxiv.org/abs/2405.20919
11. H. Frahm, S. Gehrmann, R. I. Nepomechie and A. L. Retore, 'Quantum-group-invariant $D^{(2)}_{n+1}$ models: Bethe ansatz and finite-size spectrum', J. High Energy Phys. 12 (2025) 117, doi:10.1007/JHEP12(2025)117, arXiv:2509.00610, https://arxiv.org/abs/2509.00610
12. M. Azhari and A. Klümper, 'Managing singular kernels and logarithmic corrections in the staggered six-vertex model', J. High Energy Phys. 12 (2024) 040, doi:10.1007/JHEP12(2024)040, arXiv:2406.09889, https://arxiv.org/abs/2406.09889
13. S. Gehrmann, 'The influence of boundary conditions on the emergence of non-compact degrees of freedom in the scaling limit of integrable lattice models', PhD thesis, Leibniz Universität Hannover (2024), doi:10.15488/17537, https://www.repo.uni-hannover.de/handle/123456789/17667
14. D. Chernyak, A. M. Gainutdinov and H. Saleur, '$U_q\mathfrak{sl}_2$-invariant non-compact boundary conditions for the XXZ spin chain', J. High Energy Phys. 11 (2022) 016, doi:10.1007/JHEP11(2022)016, arXiv:2207.12772, https://arxiv.org/abs/2207.12772