Boundary crossing relations, dressing factors, bound states, and boundary thermodynamics for integrable open spin chains in AdS3/CFT2

For type IIB open superstrings ending on D-branes in the pure R-R flux backgrounds AdS3 x S3 x S3 x S1 and AdS3 x S3 x T4, the matrix parts of the integrable reflection (K-)matrices for the $d(2,1;\alpha)^2$ and $psu(1,1|2)^2$ open spin chains were derived in 2015, but their scalar boundary factors are constrained only by unitarity: the boundary crossing symmetry relations have never been derived or solved. Determining these boundary dressing factors — and through them the boundary bound-state spectrum and the boundary thermodynamics (boundary Bethe equations at momentum-carrying level and a boundary thermodynamic Bethe ansatz) — is the surviving open core of the AdS3/CFT2 boundary integrability program. Later work (2024–2026) resolved the massless-sector reflection matrices and the boundary algebraic Bethe ansatz for the massless and massive sectors, and classified classical mixed-flux integrable boundaries, but explicitly identifies the crossing and dressing-factor determination as a major outstanding task that gates all boundary thermodynamics.

Background

The AdS3/CFT2 correspondences relate type IIB superstring theory on AdS3 x S3 x T4 (dual to symmetric-orbifold CFTs such as the D1-D5 system) and on AdS3 x S3 x S3 x S1 to two-dimensional conformal field theories. In these backgrounds with pure Ramond–Ramond flux, the worldsheet scattering of the two fundamental massive magnon excitations is governed by an exact two-particle S-matrix (the R-matrix) invariant under a centrally extended Lie superalgebra, two copies of $psu(1,1|2)$ for AdS3 x S3 x T4 and of $d(2,1;\alpha)$ for AdS3 x S3 x S3 x S1. The exact bulk S-matrix consists of a fixed matrix part times a scalar dressing phase, which encodes the coupling dependence and is fixed by solving the bulk crossing relations; for AdS3/CFT2 these bulk dressing phases were determined by Borsato, Ohlsson Sax, Sfondrini, Stefański and Torrielli (2013, 2017). Open strings ending on D-branes (realized as D1- and D5-brane maximal giant gravitons — branes wrapped on cycles of the internal spheres and carrying the maximal allowed angular momentum) instead define open spin chains: the magnons reflect off the boundary, and reflection is described by a reflection (K-)matrix K(p) that solves the boundary Yang–Baxter equation (BYBE), the reflection analogue of the bulk Yang–Baxter equation. In the double-row transfer-matrix formalism for integrable open chains introduced by Sklyanin (1988), a K-matrix must intertwine a boundary coideal subalgebra B of the bulk Hopf (super)algebra A — a (super)algebra equipped with a coproduct, the rule prescribing how a symmetry generator acts on two-particle tensor-product states, with B a subalgebra whose coproduct closes on B ⊗ A — which encodes the residual symmetry preserved by the boundary. The bulk symmetry algebra is moreover completed to an infinite-dimensional extension, the Yangian, whose level-0 generators reproduce the finite (super)algebra and whose higher-level generators encode multi-magnon scattering; the boundary analogue, the twisted Yangian, is a coideal subalgebra of the bulk Yangian, and constructing these boundary extensions for the AdS3 boundaries is itself unfinished. Prinsloo, Regelskis and Torrielli (2015) derived the K-matrices for vector and singlet boundaries of the $d(2,1;\alpha)^2$ and $psu(1,1|2)^2$ open chains, covering totally supersymmetric, half-supersymmetric, quarter-supersymmetric and non-supersymmetric boundary coideal subalgebras, and matched many of them to D-brane maximal giant gravitons; they also uncovered hidden diagonally-supersymmetric coideal subalgebras with no AdS5/CFT4 analogue. However, the BYBE and the intertwining conditions fix only the matrix part of each K-matrix: every solution carries an undetermined scalar boundary factor (the scale factors $k^a(p)$ for chiral reflections, which preserve the magnon's left/right chirality, and the achiral prefactors $a^a(p)$ for achiral reflections, which flip it), constrained so far only by unitarity. Completing the exact boundary S-matrix therefore requires the boundary analogue of the bulk dressing-phase problem: deriving the boundary crossing symmetry relations (the consistency conditions obtained by analytically continuing a magnon through the boundary, as a double-row transfer matrix contains both R(p) and R(-p)) and solving them for the scalar factors, relating them to the known bulk dressing phases. This is the entry point to the full boundary program familiar from AdS5/CFT4 boundary integrability: boundary bound states (via the boundary bootstrap, i.e. the analysis of poles of the dressed K-matrices), the boundary spectrum, and boundary thermodynamics (the boundary thermodynamic Bethe ansatz, an integral-equation method extracting finite-volume energies from exact scattering data, which here would yield boundary ground-state energies). Substantial progress on adjacent parts of this program has appeared recently: the reflection matrices for the massless sector of AdS3 x S3 x T4 were classified exhaustively (Bielli, Gautam, Moustakis, Prinsloo, Torrielli 2024), the boundary algebraic Bethe ansatz was constructed for the massless sector (Bielli, Moustakis, Torrielli 2025) and comprehensively for the massive sector (Bielli, Moustakis, Torrielli 2025), and a classical classification of integrable boundary conditions for mixed NS-NS/R-R flux was obtained from the divisor structure of the Lax connection (the gauge connection whose flatness encodes classical integrability of the sigma model; Cabello Gil and Driezen 2026). None of these fixed the scalar factors: the crossing and dressing-factor determination was explicitly deferred to future work in each of these papers.

Problem Statement

For the integrable open spin chains describing magnon scattering off the boundaries of type IIB open superstrings on the pure R-R flux backgrounds AdS3 x S3 x S3 x S1 (chain algebra $d(2,1;\alpha)^2$) and AdS3 x S3 x T4 (chain algebra $psu(1,1|2)^2$), complete the exact boundary scattering theory and its thermodynamic consequences, whose components are chained by the requirement that each feeds the next: (i) derive the boundary crossing symmetry relations for the reflection K-matrices — which constrain the chiral scale factors $k^a(p)$ and achiral prefactors $a^a(p)$, and their massless-sector counterparts — obtained by analytically continuing magnons through the boundary in the double-row transfer matrix, and solve them to determine all scalar boundary dressing factors, establishing their relation to the known bulk AdS3 dressing phases (Borsato et al.) so that the boundary S-matrices are completely specified beyond the unitarity constraints; (ii) construct the boundary (twisted) Yangian coideal structures extending the level-0 boundary coideal subalgebras to the infinite-dimensional level, including the hidden diagonally-supersymmetric boundary algebras, with their Hopf-algebra properties; (iii) classify the boundary bound states through the boundary bootstrap — the pole structure of the fully dressed K-matrices — determining which boundaries support boundary bound states, their masses, and their role in the boundary excitation spectrum; and (iv) building on (i)–(iii) and on the existing massless and massive boundary algebraic Bethe ansatz results, formulate the momentum-carrying boundary Bethe equations and construct the boundary thermodynamic Bethe ansatz for these open chains, including boundary ground-state energies (the AdS3 analogue of the boundary g-function, i.e. the boundary contribution to the ground-state energy). All four components (i)–(iv) are jointly required for the task to be complete, with none of them resolving it on its own. The common objective is to bring the boundary sector of AdS3/CFT2 integrability to the same level of completeness as its bulk counterpart, where the dressing phases, Bethe equations, and thermodynamics are known; item (i) is the required entry point ('an absolute necessary step to proceed with the thermodynamics' per the 2025 state of the art), (ii) characterizes the symmetry structure the K-matrices intertwine, (iii) supplies the boundary spectrum entering (iv), and (iv) is the thermodynamic payoff.

References

1. A. Prinsloo, V. Regelskis, A. Torrielli, Integrable open spin-chains in AdS3/CFT2 correspondences, Phys. Rev. D 92 (2015) 106006, arXiv:1505.06767, doi:10.1103/PhysRevD.92.106006, https://arxiv.org/abs/1505.06767
2. R. Borsato, O. Ohlsson Sax, A. Sfondrini, B. Stefański, A. Torrielli, Dressing phases of AdS3/CFT2, Phys. Rev. D 88 (2013) 066004, arXiv:1306.2512, doi:10.1103/PhysRevD.88.066004, https://arxiv.org/abs/1306.2512
3. R. Borsato, O. Ohlsson Sax, A. Sfondrini, B. Stefański, A. Torrielli, On the dressing factors, Bethe equations and Yangian symmetry of strings on AdS3 x S3 x T4, J. Phys. A 50 (2017) 024004, arXiv:1607.00914, doi:10.1088/1751-8121/50/2/024004, https://arxiv.org/abs/1607.00914
4. D. Bielli, V. Gautam, V. Moustakis, A. Prinsloo, A. Torrielli, Boundary scattering in massless AdS3, JHEP 07 (2024) 266, arXiv:2403.18594, doi:10.1007/JHEP07(2024)266, https://arxiv.org/abs/2403.18594
5. D. Bielli, V. Moustakis, A. Torrielli, Boundary Bethe ansatz in massless AdS3, J. Phys. A 58 (2025) 045402, arXiv:2408.15674, doi:10.1088/1751-8121/ada7ae, https://arxiv.org/abs/2408.15674
6. D. Bielli, V. Moustakis, A. Torrielli, Boundary Bethe ansatz in massive AdS3, JHEP 10 (2025) 154, arXiv:2506.20133, doi:10.1007/JHEP10(2025)154, https://arxiv.org/abs/2506.20133
7. J. Cabello Gil, S. Driezen, Analytic approach to boundary integrability with application to mixed-flux AdS3 x S3, arXiv:2601.09935, https://arxiv.org/abs/2601.09935
8. E. K. Sklyanin, Boundary conditions for integrable quantum systems, J. Phys. A 21 (1988) 2375, doi:10.1088/0305-4470/21/10/015, https://doi.org/10.1088/0305-4470/21/10/015