Completing the lattice-to-continuum correspondence for integrable seams in unitary and nonunitary A-D-E minimal models: r-type seams, noncommutative $D_{2\ell}$ fusion algebras, and nonunitary dilogarithm identities

Pearce, Heymann and Quella (arXiv:2512.21808) argued that the coset graph $A\otimes G/\mathbb{Z}_2$ of the unitary and nonunitary A-D-E minimal models $\mathcal{M}(m,m')=(A_{m-1},G)$ universally encodes the coset graph fusion algebra, the Affleck-Ludwig boundary and defect $g$-factors, the relative symmetry-resolved entanglement entropies, and the effective central charges and conformal weights via dilogarithm identities in the quantum dimensions, and that these structures are realized on the lattice by Yang-Baxter-integrable seams of the critical A-D-E RSOS models. Completing this correspondence in full generality requires closing four gaps left open or deferred in that work: (1) a direct lattice-to-continuum correspondence for the r-type (bulk) integrable seams, which unlike the s-type (braid) seams are not topological on the finite lattice and only acquire the label $r$ after a nontrivial two-step scaling limit; (2) the noncommutative enlarged fusion algebra obtained by adjoining the $\mathbb{Z}_2$ diagram automorphism in the $D_{2\ell}$ models, and its consequences for defect operator constructions and lattice seams; (3) the detailed nonunitary TBA and dilogarithm derivations of the effective central charges and conformal weights; (4) the polynomial-ring presentations of the coset graph fusion algebras explicitly deferred for the $D_4$-based 3-state Potts coset examples. None of the four gaps has been resolved in the literature to date.

Background

Two-dimensional conformal field theories (CFTs) with Virasoro symmetry and finitely many primary fields — the Virasoro minimal models $\mathcal{M}(m,m')$ — have central charge $c=1-6/(mm')$ and primaries labelled by Kac labels $(r,s)$ with $1\le r\le m-1$, $1\le s\le m'-1$. They are unitary precisely when $m'=m\pm 1$; otherwise (for example the Lee-Yang model $\mathcal{M}(2,5)$) they are nonunitary, meaning the Hilbert space is not positive definite, the ground state is not the vacuum $(1,1)$, and the physically meaningful data are the effective central charge $c_{\rm eff}$ and effective conformal weights measured relative to the true ground state. The modular-invariant partition functions of these theories are classified by pairs of simply-laced Dynkin diagrams (the A-D-E classification), written $(A_{m-1},G)$ with $G=A$ (diagonal theories) or $G=D,E$ (non-diagonal, type I and type II theories).

A topological defect line in a 2D CFT is a line operator that can be freely deformed without changing correlators, as long as it does not cross other lines or operator insertions; such lines implement generalized (generally non-invertible) global symmetries and fuse according to a fusion algebra. For minimal models the elementary lines are the Verlinde lines $\mathcal{L}_{(1,s)}$, the Pasquier lines, and, for non-diagonal theories, additional Ocneanu lines; the fusion of compound lines is encoded by graph fusion algebras — algebras of nonnegative-integer matrices (nimreps) whose structure constants count fusion multiplicities, attached to suitable graphs.

Pearce, Heymann and Quella define the coset graph $A\otimes G/\mathbb{Z}_2$ — the tensor product of the $A_{m-1}$ and $G$ Dynkin diagrams quotiented by the Kac-table $\mathbb{Z}_2$ symmetry $(r,a)\equiv(m-r,a)$ — and argue that it is a universal object which always factorizes into tadpole graphs tensor Dynkin graphs and which encodes: (i) the coset graph fusion algebra; (ii) the Affleck-Ludwig boundary $g$-factors (universal non-integer ground-state degeneracies of conformal boundary conditions) and the defect $g$-factors given by quantum dimensions (ratios of eigenvalues of fusion matrices, giving the quantum dimension of a defect line); (iii) the relative symmetry-resolved entanglement entropies (SREEs — the decomposition of an interval's entanglement entropy into symmetry sectors); and (iv) the effective central charges and conformal weights through analytically continued Rogers dilogarithm functions of the quantum dimensions.

On the lattice side, the critical A-D-E restricted solid-on-solid (RSOS) models are two-dimensional Yang-Baxter-integrable lattice models (their local face weights satisfy the Yang-Baxter equation, so that transfer matrices built with different spectral parameters commute) whose heights live on the nodes of a Dynkin diagram, with crossing parameter $\lambda=\pi(m'-m)/m'$ where $m'-m$ is a Coxeter exponent of $G$; their continuum scaling limit is the corresponding minimal CFT. Defects are implemented as integrable seams — vertical columns of faces with a modified spectral parameter $\xi$, built as commuting column transfer matrices. Two families arise. The s-type (braid) seams come from the braid limits $\xi\to\pm i\infty$ of the face weights and are exactly topological on the finite lattice, following the Temperley-Lieb centralizer construction of Aasen, Fendley and Mong and of Bellettête et al. The r-type (bulk) seams come from the fused transfer matrices $\boldsymbol{T}^{(r,1)}$. The transfer matrices satisfy T-system and Y-system functional equations (the usual fusion-hierarchy functional equations of integrable models), whose braid and bulk asymptotics reproduce the fusion algebra of the defect lines; the thermodynamic Bethe ansatz (TBA) plateaux of these Y-systems yield the dilogarithm identities expressing $c_{\rm eff}$ and $\Delta_{r,s}$ in terms of the quantum dimensions.

The paper leaves four items open or deferred. First, the detailed analytic derivations for nonunitary models — the TBA functional equations and the analytic continuation of the dilogarithms, where negative components of the Perron-Frobenius eigenvector and principal-branch square roots require care — are deferred to a separate paper, with only the Lee-Yang model treated in the earlier literature. Second, whereas s-type seams are topological on the lattice, the r-type seams are not: at finite size $r$ is not a good quantum number, and the label $r$ emerges only after a two-step limit (scaling limit with $x\sim\mp\log M$, then $x\to\pm\infty$), so the lattice-to-continuum correspondence for r-type seams is indirect and a complete general proof is not given in the paper (the detailed arguments are referred to the work of Sinha, Tavares, Roy and Saleur). Third, the polynomial-ring presentations of the coset graph fusion algebras — presentations of the fusion ring as a quotient of a polynomial ring via a fusion potential, as known for the $A$-series — are not worked out for the $D$- and $E$-type coset examples: for the $D_4$-based 3-state Potts coset examples $T_2\otimes D_4$ and $T_3\otimes D_4$ the paper states 'The polynomial ring for this case will be discussed elsewhere' even though the fusion Cayley tables are given, and the $(A_4,E_6)$ example is likewise presented without a polynomial-ring presentation. Fourth, among the A-D-E minimal models only the $D_{2\ell}$ models require the $\mathbb{Z}_2$ diagram automorphism to be added separately to the fusion algebra, and adding it makes the enlarged algebra noncommutative (in the $D_4$-based Potts examples, because it contains within it the noncommutative symmetric group $S_3$); the authors explicitly do not construct this algebra ('we do not do this here') and leave its implications for defect operators and lattice seams open. The only pre-existing treatment of the seam algebra beyond the $A$-series is the Pearce-Rasmussen analysis of the critical unitary $E_6$ RSOS model.

Problem Statement

The common objective is to complete, in full generality, the lattice-to-continuum correspondence for the unitary and nonunitary A-D-E coset minimal models: to show that the integrable-seam construction of the critical A-D-E RSOS lattice models realizes all topological defect lines of the coset CFTs $(A_{m-1},G)$ with their fusion algebras, and that the associated TBA/dilogarithm machinery verifies the conformal data, uniformly for all Coxeter-number pairs $(m,m')$, all $G=A,D,E$, and both the unitary and nonunitary cases. The source paper leaves four items open or deferred, each addressing one axis of this correspondence:

1. (r-type seams) Establish a direct lattice-to-continuum correspondence for the r-type (bulk) integrable seams: give a complete and fully general proof that the two-step scaling limit of the seam transfer matrices $\boldsymbol{T}^{(r,1)}(\xi)$ converges to the CFT defect lines $\mathcal{L}_{(r,1)}$ and that the resulting lines satisfy the coset graph fusion algebra, for all $(m,m')$ and all $(A_{m-1},G)$ families, including the associated r-type boundary conditions of the double-row transfer matrices.

2. (noncommutative $D_{2\ell}$ algebras) Construct the enlarged coset graph fusion algebra for the $D_{2\ell}$ models obtained by adjoining the $\mathbb{Z}_2$ diagram automorphism $\sigma$ — an algebra which becomes noncommutative — determine its structure constants and representation content, and work out its implications for defect operator constructions and lattice seams in these models.

3. (nonunitary derivations) Provide the detailed analytic treatment of the nonunitary models deferred by the source to a separate paper — the TBA functional equations, the analytic continuation of the dilogarithms, and the consistent handling of negative Perron-Frobenius components and principal-branch square roots in the face weights — deriving in detail the formulas, stated in the source as valid for all $(m,m')$, that express the effective central charges $c^{m,m'}_{\rm eff}$ and conformal weights $\Delta^{m,m'}_{r,s}$ as Rogers dilogarithm sums of the quantum dimensions, and generalizing the existing Lee-Yang-model computation to the other nonunitary A-D-E cases.

4. (polynomial rings) Give the polynomial-ring presentations (fusion potentials) explicitly deferred by the source for the $D_4$-based 3-state Potts coset examples ($T_2\otimes D_4$ and $T_3\otimes D_4$), carrying out the polynomial-reduction steps.

All four parts are mandatory: completing any one of them, or any proper subset of them, is a partial result and does not by itself resolve the task. In part 2, the representation content means a complete list of irreducible representations, and the implications for defect operator constructions and lattice seams mean the induced seam and boundary conditions.

References

1. Paul A. Pearce, Jared Heymann, Thomas Quella, 'Unitary and Nonunitary A-D-E minimal models: Coset graph fusion algebras, defects, entropies, SREEs and dilogarithm identities', arXiv:2512.21808 (2025), DOI: 10.48550/arXiv.2512.21808, https://arxiv.org/abs/2512.21808
2. Madhav Sinha, Thiago Silva Tavares, Ananda Roy, Hubert Saleur, 'Integrability and lattice discretizations of all topological defect lines in minimal CFTs', Nucl. Phys. B 1023 (2026) 117308, arXiv:2509.04257, DOI: 10.1016/j.nuclphysb.2026.117308, https://arxiv.org/abs/2509.04257
3. Madhav Sinha, Thiago Silva Tavares, Hubert Saleur, Ananda Roy, 'Lattice topological defects in nonunitary conformal field theories', Phys. Rev. B 114 (2026) 084107, arXiv:2604.25999, DOI: 10.1103/1sqr-9n7d, https://arxiv.org/abs/2604.25999
4. David Aasen, Paul Fendley, Roger S. K. Mong, 'Topological Defects on the Lattice: Dualities and Degeneracies', arXiv:2008.08598 (2020), https://arxiv.org/abs/2008.08598
5. Paul A. Pearce, Jorgen Rasmussen, 'Ocneanu Algebra of Seams: Critical Unitary $E_6$ RSOS Lattice Model', 2024 MATRIX Annals, Part II (MATRIX Book Series, 2026) pp. 113-135, arXiv:2409.06236, DOI: 10.1007/978-3-032-16206-9_8, https://arxiv.org/abs/2409.06236
6. Zoltan Bajnok, Omar el Deeb, Paul A. Pearce, 'Finite-Volume Spectra of the Lee-Yang Model', JHEP 04 (2015) 073, arXiv:1412.8494, DOI: 10.1007/JHEP04(2015)073, https://arxiv.org/abs/1412.8494
7. Jared Heymann, Thomas Quella, 'Revisiting the symmetry-resolved entanglement for non-invertible symmetries in $1{+}1$d conformal field theories', Phys. Rev. D 112 (2025) 025004, arXiv:2409.02315, DOI: 10.1103/lr47-yv3j, https://arxiv.org/abs/2409.02315
8. J. Belletête, A. M. Gainutdinov, J. L. Jacobsen, H. Saleur, T. S. Tavares, 'Topological defects in periodic RSOS models and anyonic chains', arXiv:2003.11293 (2020), https://arxiv.org/abs/2003.11293
9. Yoshiki Fukusumi, Yuma Furuta, 'Organizing transitions and their cascades: Generalized symmetry enforcement in massless flows or Higgs transitions', arXiv:2608.08428 (2026), https://arxiv.org/abs/2608.08428
10. Yoshiki Fukusumi, Shuma Nakashiba, 'Characterizing gapped phases by smeared boundary conformal field theories: Duality in unusual ordering with spontaneously broken generalized symmetries', arXiv:2605.07734 (2026), https://arxiv.org/abs/2605.07734