Completing the exact solution of the $U_q\mathfrak{sl}_2$-invariant non-compact boundary XXZ chain: two-boundary Temperley–Lieb representation theory at roots of unity, indecomposable boundary weights, critical scaling limits, and fusion

Chernyak–Gainutdinov–Saleur (JHEP 2022) introduced open XXZ spin chains whose quantum-group symmetry is preserved by coupling the chain ends to infinite-dimensional Verma modules ('non-compact' boundary conditions), and posed a coordinated set of open problems: full two-boundary Schur–Weyl duality with the two-boundary Temperley–Lieb (2B) algebra, identification of the multiplicity modules with 2B standard modules, the root-of-unity and indecomposable regimes, the critical continuum-limit (Virasoro) correspondence, and a fusion construction. The 2023 sequel by the same authors with Jacobsen has resolved the generic-$q$ half — proving Conjecture 1 (mutual maximal centralisers of $U_q\mathfrak{sl}_2$ and the universal two-boundary TL algebra on the two-boundary Hilbert space), identifying the multiplicity modules, and deriving the two-boundary Bethe equations by algebraic Bethe ansatz — while the remaining core is open: the representation theory of 2B at the non-generic parameters at roots of unity (Conjecture 2, the structure of the vacuum module, Bethe-ansatz completeness there), the indecomposable regime of non-generic boundary weights $q^\alpha\in\pm q^{\mathbb{Z}}$ with non-diagonalisable Hamiltonians, the rigorous sector-to-Virasoro correspondence for $|q|=1$, and the fusion interpretation. This surviving core is one problem: completing the exact solution of these chains at all parameter values.

Background

The open XXZ spin chain (the anisotropic spin-1/2 Heisenberg chain) with quantum-group-invariant boundary conditions is the Hamiltonian $H_{\mathrm{XXZ}}=-\sum_{i=1}^{N-1} e_i$ acting on $(\mathbb{C}^2)^{\otimes N}$, where the $e_i$ are Temperley–Lieb (TL) generators — diagrammatic operators satisfying $e_i^2=\delta e_i$, $e_i e_{i\pm1}e_i=e_i$ and $e_i e_j=e_j e_i$ for $|i-j|\ge 2$, with loop weight $\delta=q+q^{-1}$ (each closed loop evaluates to $\delta$). The TL algebra is in Schur–Weyl duality with the quantum group $U_q\mathfrak{sl}_2$ (the $q$-deformation of the universal enveloping algebra of the Lie algebra $\mathfrak{sl}_2$): the two actions on $(\mathbb{C}^2)^{\otimes N}$ commute and each is the maximal centraliser of the other — the centraliser of a set of operators being the algebra of all operators commuting with it — the lattice analogue of the classical duality between symmetric-group and general-linear-algebra representations.

Chernyak–Gainutdinov–Saleur (2022) constructed new boundary conditions preserving this symmetry by coupling 'non-compact' boundary degrees of freedom: one or both ends of the chain are tensored with a Verma module $\mathcal{V}_\alpha$, the infinite-dimensional highest-weight $U_q\mathfrak{sl}_2$-module of highest weight $q^\alpha$, where $\alpha\in\mathbb{C}$ plays the role of a continuous boundary spin. The fusion rule $\mathcal{V}_\alpha\otimes\mathbb{C}^2\cong\mathcal{V}_{\alpha+1}\oplus\mathcal{V}_{\alpha-1}$ (valid whenever $q^\alpha\neq\pm1$, even when the summands $\mathcal{V}_{\alpha\pm1}$ are non-irreducible) yields boundary projectors $b_\pm$ onto the two summands, and the one-boundary Hamiltonian $H_b=-\mu b-\sum_i e_i$ then represents the blob algebra — the TL algebra extended by an idempotent $b$ with $e_1 b e_1=y e_1$ — faithfully and Schur–Weyl-dually to $U_q\mathfrak{sl}_2$ for $q^\alpha\notin\pm q^{\mathbb{Z}}$. At roots of unity $q=e^{i\pi/p}$ the Verma modules are replaced by $p$-dimensional analogues and the relevant symmetry becomes the unrolled quantum group $\overline{U}^H_q\mathfrak{sl}_2$, a non-semisimple modification of $U_q\mathfrak{sl}_2$ which admits a non-trivial braiding and underlies logarithmic conformal field theories.

With a Verma module on each end, $H_{2b}=-\mu_l b_l-\sum_i e_i-\mu_r b_r$ on $\mathcal{H}_{2b}=\mathcal{V}_{\alpha_l}\otimes(\mathbb{C}^2)^{\otimes N}\otimes\mathcal{V}_{\alpha_r}$ represents the universal two-boundary TL algebra $2\mathsf{B}^{\mathrm{uni}}_{\delta,y_l,y_r,N}$. The ordinary two-boundary TL algebra $2\mathsf{B}_{\delta,y_l,y_r,Y,N}$ of de Gier–Nichols additionally assigns a weight $Y\in\mathbb{C}$ to loops touching both boundaries; in the universal version $Y$ is a central generator. On the spin chain, $Y$ is realised as an explicit non-local central element built from the Casimir (the standard quadratic central element) of $U_q\mathfrak{sl}_2$, and decomposing $\mathcal{H}_{2b}$ into $Y$-eigenspaces yields, on each eigenspace with eigenvalue $Y_m$, a representation of $2\mathsf{B}_{\delta,y_l,y_r,Y_m,N}$. The lattice $Y$ operator has the diagrammatic form of a 'hoop' (a loop encircling the whole chain) and is conjecturally related to the topological defect operators known from periodic (affine-TL) models. Through the Schur–Weyl framework these chains encode the spectra of the integrable open XXZ chains with general non-diagonal boundary terms restricted by the Nepomechie condition — a constraint on the six boundary parameters under which the Bethe-ansatz solution of the non-diagonal open chain is complete; in this realisation the constraint selects precisely the discrete set of Casimir-allowed $Y$ values.

At the free-fermion point $q=i$ the Hamiltonian reduces to a modified XX model, diagonalised exactly by a Jordan–Wigner transformation (the standard mapping of spin operators to fermionic ones); the continuum scaling limit is the $(\eta,\xi)$ ghost (symplectic-fermion) conformal field theory of central charge $c=-2$ on the upper half-plane, with a continuum of conformally invariant boundary conditions labelled by $\alpha$, and each Hamiltonian sector matches an irreducible representation of the Virasoro algebra (the infinite-dimensional symmetry algebra of two-dimensional conformal field theory). This proves, at this single point, the correspondence conjectured on Coulomb-gas (free-field) and numerical grounds by Jacobsen–Saleur and Dubail–Jacobsen–Saleur between the scaling limits of boundary and two-boundary loop-model sectors and Virasoro modules. What makes the general problem hard is that the Casimir-allowed values $Y_m$ are non-generic: at these values the representation theory of $2\mathsf{B}$ — in particular the structure of its vacuum module $W_0$ (the standard module with no through-lines, i.e. no diagram strands connecting the two boundaries) — is non-semisimple and was not classified. The source paper therefore established the decomposition of $\mathcal{H}_{2b}$ only into $Y$-eigenspaces (its Propositions 3–6) and conjectured the full Schur–Weyl duality and module identifications: Conjecture 1 for generic $q$ and Conjecture 2 for $q$ a root of unity, the latter stating that the multiplicity modules $\tilde{\underline{\mathcal{H}}}_m$ are isomorphic to the semi-simplified vacuum module (the direct sum of the composition factors of $W_0$) with all summands $\tilde{\underline{\mathcal{H}}}_m^{(s)}$ irreducible, and with $2\mathsf{B}^{\mathrm{uni}}_{\delta,y_l,y_r,N}$ and $\overline{U}^H_q\mathfrak{sl}_2$ mutual maximal centralisers on $\mathcal{H}_{2b}$. A 2023 sequel (Chernyak–Gainutdinov–Jacobsen–Saleur) proved Conjecture 1 and the generic-$q$ module identification, and derived the two-boundary Bethe equations by algebraic Bethe ansatz (ABA) for all Nepomechie-allowed boundary parameters at generic $q$; the root-of-unity representation theory, the indecomposable boundary-weight regime, the rigorous critical scaling limits, and the fusion construction remain open.

Problem Statement

Complete the exact solution of the $U_q\mathfrak{sl}_2$-invariant XXZ spin chain with non-compact boundary conditions — the program announced by Chernyak–Gainutdinov–Saleur (spectra and scaling limits, lattice-algebra representation content, and fusion, at all values of the parameters) — in the regimes not settled by the free-fermion analysis at $q=i$ or by the generic-$q$ sequel:

(1) Roots of unity (Conjecture 2 of the source paper). For $q=e^{i\pi/p}$, $p\ge 2$, with $\alpha_l,\alpha_r\notin\mathbb{Z}$ and $\alpha_l+\alpha_r\notin\mathbb{Z}$, three sub-claims are to be settled independently. (1a) Determine the structure of the vacuum module $W_0$ of $2\mathsf{B}_{\delta,y_l,y_r,Y,N}$ at the non-generic values $Y=Y_m$ produced by the Casimir, and classify the irreducible $2\mathsf{B}$-modules at these parameters; this sub-claim is required regardless of the outcome of (1b). (1b) Prove or refute that each multiplicity module $\tilde{\underline{\mathcal{H}}}_m$ is isomorphic to the $2^N$-dimensional semi-simplified vacuum module, with its summands $\tilde{\underline{\mathcal{H}}}_m^{(s)}$ irreducible — so that $2\mathsf{B}_{\delta,y_l,y_r,Y_m,N}$ and $\overline{U}^H_q\mathfrak{sl}_2$ are mutual maximal centralisers on each $\underline{\mathcal{H}}_m$, and $2\mathsf{B}^{\mathrm{uni}}_{\delta,y_l,y_r,N}$ and $\overline{U}^H_q\mathfrak{sl}_2$ on $\mathcal{H}_{2b}$; a refutation of (1b) settles (1b) alone and leaves (1a) open. (1c) Establish or refute the conjectured completeness and counting of solutions of the two-boundary Bethe equations at roots of unity, expected to differ from the generic-$q$ case; a refutation of (1c) settles (1c) alone.

(2) Non-generic boundary weights. Within $q^\alpha\in\pm q^{\mathbb{Z}}$ two distinct mechanisms must be separated. When $q^\alpha=\pm 1$ the fusion rule fails and $\mathcal{V}_\alpha\otimes\mathbb{C}^2$ is reducible but indecomposable, so the boundary projectors $b_\pm$ degenerate and $H_b$ is recovered only by the rescaling limit; when $q^\alpha=\pm q^n$ with $n\neq0$ the projectors and $H_b$ remain well defined, but reducible-and-indecomposable modules appear in sufficiently large chains and make the Hamiltonians non-diagonalisable with non-trivial Jordan blocks. In both cases determine the spectra and the Jordan-block structure of $H_b$ and $H_{2b}$, and determine the composition factors of the corresponding modules of the one-boundary (blob) and two-boundary ($2\mathsf{B}$) algebras, for both chains.

(3) Critical scaling limits. Prove, for all $|q|=1$, the conjectured sector-to-Virasoro correspondence: identify the scaling limit of each sector of $H_b$ and $H_{2b}$ (equivalently, of the boundary and two-boundary loop models) with explicit Virasoro representations, as established at $q=i$ and conjectured in general by Jacobsen–Saleur and Dubail–Jacobsen–Saleur.

(4) Fusion. Construct the fusion procedure for these chains, making precise the statement that the two-boundary chain is the fusion of two one-boundary systems: define the operation at the level of the lattice algebras $2\mathsf{B}$ and their modules (extending the fusion bifunctor known for the periodic Temperley–Lieb algebra), show how lattice fusion rules map to Virasoro fusion in the continuum, and determine whether the additive combination $\alpha_l+\alpha_r$ of boundary weights found at $q=i$ is universal. This includes clarifying the relation between the lattice $Y$ operator and the topological defect operators of periodic models.

Items (1)–(4) are jointly the completion of the stated program: they share one model, one lattice-algebra framework, and one continuum limit, and each supplies structural input that the others use (the module classification underlies the sector decomposition of the spectrum and its scaling limit; the fusion construction organises the same modules). Each of items (1)–(4) is an independent mandatory part of that program: none of the four items, and no proper sub-part of any of them, resolves the task on its own. The generic-$q$ Schur–Weyl duality, the generic-$q$ module identification, and the algebraic Bethe ansatz at generic $q$ are settled and are not part of this problem.

References

1. Dmitry Chernyak, Azat M. Gainutdinov, Hubert Saleur, '$U_q\mathfrak{sl}_2$-invariant non-compact boundary conditions for the XXZ spin chain', JHEP 11 (2022) 016, DOI 10.1007/JHEP11(2022)016, arXiv:2207.12772, https://arxiv.org/abs/2207.12772
2. Dmitry Chernyak, Azat M. Gainutdinov, Jesper Lykke Jacobsen, Hubert Saleur, 'Algebraic Bethe Ansatz for the Open XXZ Spin Chain with Non-Diagonal Boundary Terms via $U_q\mathfrak{sl}_2$ Symmetry', SIGMA 19 (2023) 046, DOI 10.3842/SIGMA.2023.046, arXiv:2212.09696, https://arxiv.org/abs/2212.09696
3. Jan de Gier, Alexander Nichols, 'The two-boundary Temperley–Lieb algebra', Journal of Algebra 321 (2009) 1132–1167, DOI 10.1016/j.jalgebra.2008.10.023, arXiv:math/0703338, https://arxiv.org/abs/math/0703338
4. Jérôme Dubail, Jesper Lykke Jacobsen, Hubert Saleur, 'Conformal two-boundary loop model on the annulus', Nuclear Physics B 813 (2009) 430–459, DOI 10.1016/j.nuclphysb.2008.12.023, arXiv:0812.2746, https://arxiv.org/abs/0812.2746
5. Jesper Lykke Jacobsen, Hubert Saleur, 'Conformal boundary loop models', Nuclear Physics B 788 (2008) 137–166, DOI 10.1016/j.nuclphysb.2007.06.029, arXiv:math-ph/0611078, https://arxiv.org/abs/math-ph/0611078
6. Zajj Daugherty, Arun Ram, 'Calibrated representations of two boundary Temperley–Lieb algebras', 2020, arXiv:2009.02812, https://arxiv.org/abs/2009.02812
7. Chris Bowman, Zajj Daugherty, Maud De Visscher, Rob Muth, Loïc Poulain d'Andecy, 'The orientifold Temperley–Lieb algebra', 2026, arXiv:2601.04012, https://arxiv.org/abs/2601.04012
8. Yacine Ikhlef, Alexi Morin-Duchesne, 'Fusion in the periodic Temperley–Lieb algebra: general definition of a bifunctor', 2025, arXiv:2509.11756, https://arxiv.org/abs/2509.11756
9. Azat M. Gainutdinov, Jesper Lykke Jacobsen, Hubert Saleur, 'A fusion for the periodic Temperley–Lieb algebra and its continuum limit', JHEP 11 (2018) 117, DOI 10.1007/JHEP11(2018)117, arXiv:1712.07076, https://arxiv.org/abs/1712.07076
10. J. Belletête, A. M. Gainutdinov, J. L. Jacobsen, H. Saleur, T. S. Tavares, 'Topological defects in lattice models and affine Temperley–Lieb algebra', Communications in Mathematical Physics 400 (2023) 1203–1254, DOI 10.1007/s00220-022-04618-0, arXiv:1811.02551, https://arxiv.org/abs/1811.02551
11. Guillaume Lemarthe, Pascal Baseilhac, Azat M. Gainutdinov, 'Fused K-Operators and the $q$-Onsager Algebra', SIGMA 22 (2026) 026, DOI 10.3842/SIGMA.2026.026, arXiv:2301.00781, https://arxiv.org/abs/2301.00781