The one-dimensional quantum Ashkin–Teller model — two transverse-field Ising chains coupled by a four-spin interaction of strength λ\lambda — is critical along a line 0≤λ≤10\le\lambda\le1 described by the Z2\mathbb{Z}_2 orbifold of the c=1c=1 compactified free boson. Liu, Chepiga, Fukusumi and Oshikawa (SciPost Phys. 20, 130 (2026)) identified which conformal boundary states of this theory are reached by microscopic boundary terms — the field hστzσzτzh^z_{\sigma\tau}\sigma^z\tau^z for the continuous Dirichlet family and the marginal operator hθNσzτxh^N_\theta\sigma^z\tau^x for the Neumann family — and verified the identification by density-matrix-renormalization-group (DMRG) finite-size spectroscopy, but left the quantitative lattice-to-boundary-CFT dictionary open. Three linked issues survive: the boundary value θ\theta labeling each family is a non-universal and so-far-unknown function of the microscopic coupling; the level crossing observed at hθN=1h^N_\theta=1, conjectured to realize the self-dual Neumann state θ=π/2\theta=\pi/2 for every λ\lambda under a protecting 'semi-duality', is proven only at the decoupled Ising point λ=0\lambda=0; and at the four-state Potts endpoint λ=1\lambda=1 the 1/ln⁡L1/\ln L decay of marginally irrelevant boundary couplings prevents standard extrapolations from demonstrating the asymptotic flow to the free boundary fixed point, leaving a persistent spectral deviation quantitatively unexplained. The problem is to complete this dictionary: determine the coupling-to-θ\theta maps for both boundary families, prove or refute the duality-protected crossing at hθN=1h^N_\theta=1 and its mechanism, and give a controlled account of the boundary asymptotics at the Potts point.

Background

The Ashkin–Teller (AT) model couples two Ising spin systems. In its one-dimensional quantum form it is a pair of transverse-field Ising chains — chains of spins-1/2 whose Hamiltonian combines a field flipping each spin with an Ising interaction aligning neighbors — here described by Pauli operators σx,z\sigma^{x,z} and τx,z\tau^{x,z}, with the two copies coupled both through the transverse field, by a term λσxτx\lambda\sigma^x\tau^x, and through the exchange, by a four-spin term λσjzτjzσj+1zτj+1z\lambda\sigma^z_j\tau^z_j\sigma^z_{j+1}\tau^z_{j+1} (Eq. (1) of the source paper). At the critical transverse field h=Jh=J the model is critical for every 0≤λ≤10\le\lambda\le1, and the whole critical line is described by a conformal field theory (a scale-invariant fixed-point quantum field theory, here with central charge c=1c=1, i.e. one effective gapless degree of freedom): the Z2\mathbb{Z}_2 orbifold of a compactified free boson — the free boson quotiented by identifying its compactification variable with its negative — with a compactification radius r(λ)r(\lambda) running from the decoupled-Ising pair at λ=0\lambda=0 to the four-state Potts point at λ=1\lambda=1, where the microscopic symmetry enlarges from the dihedral group D4D_4 (the symmetry of a square) to the full permutation group S4S_4.

For an open critical chain, boundary conformal field theory (BCFT) — the conformal field theory restricted to a half-space with a conformally invariant boundary — classifies the allowed boundary conditions: each is a superposition of Ishibashi states, coherent states built on a single bulk primary field that solve the gluing condition identifying left- and right-moving modes at the boundary (Cardy's construction). For the Z2\mathbb{Z}_2-orbifold theory the conformal boundary states form two one-parameter continuous families, Dirichlet ∣DO(θ)⟩|D_O(\theta)\rangle and Neumann ∣NO(θ)⟩|N_O(\theta)\rangle (the compact boson is pinned to a fixed value at the boundary, or left with a free slope, respectively, in the orbifold sense), labeled by a boundary value θ∈[0,π]\theta\in[0,\pi], together with eight discrete endpoint states. The boundary value fixes the entire boundary spectrum: the open-chain energy levels organize into conformal towers (families of levels generated by the descendants of a single boundary primary operator) whose scaling dimensions (eigenvalue spacings, up to 1/L1/L finite-size corrections) include a set of 'floating' boundary operators whose dimensions move continuously with θ\theta. The Affleck–Ludwig boundary entropy gg — the universal degeneracy the boundary contributes to the ground state — also depends on the boundary state. Kramers–Wannier (KW) duality, the exact self-duality of the AT model exchanging order and disorder variables, acts on these families as a defect operator: DKW∣NO(θ)⟩=2∣NO(π−θ)⟩D_{\mathrm{KW}}|N_O(\theta)\rangle=2|N_O(\pi-\theta)\rangle (Eq. (67) of the source), so θ=π/2\theta=\pi/2 is the self-dual point of the Neumann family, where floating operators become pairwise degenerate and the spectrum shows characteristic level crossings.

Liu, Chepiga, Fukusumi and Oshikawa (SciPost Phys. 20, 130 (2026)) constructed microscopic boundary terms that renormalize to these conformal boundary states and verified the identification by finite-size spectroscopy using the density matrix renormalization group (DMRG), a variational method for one-dimensional quantum chains. A boundary field hστzσzτzh^z_{\sigma\tau}\sigma^z\tau^z realizes the continuous Dirichlet family, and the boundary operator hθNσzτxh^N_\theta\sigma^z\tau^x — a truly marginal boundary perturbation, i.e. one of scaling dimension exactly 1 that deforms rather than destabilizes the fixed point (their Eq. (130)) — moves the system continuously along the Neumann family ∣NO(θ)⟩|N_O(\theta)\rangle. The catch is that the boundary value reached at a given microscopic coupling is a non-universal function — θ(hστz)\theta(h^z_{\sigma\tau}) for the Dirichlet family, θ(hθN)\theta(h^N_\theta) for the Neumann family — depending on microscopic details rather than universal CFT data. Only limiting points are pinned: hστz→±∞h^z_{\sigma\tau}\to\pm\infty projects the boundary onto the antipodal mixed states ∣AC⟩=∣DO(θc)⟩|AC\rangle=|D_O(\theta_c)\rangle and ∣BD⟩=∣DO(π−θc)⟩|BD\rangle=|D_O(\pi-\theta_c)\rangle with a known θc(λ)\theta_c(\lambda) (Eqs. (32)–(34) of the source).

At the decoupling point λ=0\lambda=0 the model is two free-fermion Ising chains, and the exact solution of the critical Ising model with a defect line by Oshikawa and Affleck gives θ=π/2\theta=\pi/2 at hθN=1h^N_\theta=1: this boundary is the KW transformation applied to half of an infinite critical Ising chain. For λ≠0\lambda\neq0 no exact statement exists. Numerically, the DMRG spectra of Liu et al. show the degenerate-level crossings characteristic of the self-dual point θ=π/2\theta=\pi/2 occurring at hθN=1h^N_\theta=1 for every λ\lambda sampled, leading them to speculate that the crossing is 'protected by a semi-duality' of the lattice model — their term for a duality acting on only half of the system — while explicitly stating that they have no proof.

At the four-state Potts endpoint λ=1\lambda=1, boundary operators that are relevant (of dimension below 1, hence destabilizing) for generic λ\lambda become marginal, and the flows they generate are only marginally irrelevant: the effective boundary coupling decays as geff∼1/ln⁡Lg_{\mathrm{eff}}\sim1/\ln L with chain length LL. Liu et al. analyzed the three-state mixed boundary perturbation ('BCD' — a boundary term mixing three of the discrete fixed-point boundary states, denoted B, C and D in the source), predicted by BCFT to flow back to the free boundary fixed point. But at the field strengths needed to reach the crossover plateau the spectra remain persistently displaced from the free-boundary conformal towers even after the standard logarithmic finite-size extrapolation (system sizes L≤40L\le40 for the Neumann spectroscopy, up to L=75L=75 for the mixed-boundary flows). The authors state that this discrepancy 'remains unresolved', interpreting it as a finite-size artifact of the slow crossover rather than a new fixed point — a physical picture, but not a controlled quantitative demonstration of the asymptotic approach to the fixed point.

Problem Statement

Complete the quantitative dictionary between the microscopic boundary couplings of the critical quantum Ashkin–Teller chain and the conformal boundary states of its Z2\mathbb{Z}_2-orbifold c=1c=1 BCFT, along the critical line h=Jh=J, 0≤λ≤10\le\lambda\le1 (lattice Hamiltonian and boundary terms as in Eqs. (1), (125) and (130) of Liu–Chepiga–Fukusumi–Oshikawa). Three linked deliverables constitute completion:

(1) The coupling-to-boundary-value maps. Determine θ(hστz)\theta(h^z_{\sigma\tau}) for the Dirichlet family and θ(hθN)\theta(h^N_\theta) for the Neumann family — exactly in closed form, or with controlled analytic error bounds, or by certified numerical determination — beyond the already-known limiting points (hστz→±∞h^z_{\sigma\tau}\to\pm\infty giving θc\theta_c and π−θc\pi-\theta_c, and the λ=0\lambda=0 Ising results).

(2) The duality-protected crossing. Prove or disprove that the self-dual Neumann boundary state θ=π/2\theta=\pi/2 is realized exactly at hθN=1h^N_\theta=1 for every λ∈[0,1]\lambda\in[0,1] — equivalently, that the level crossing of the floating boundary operators occurs at hθN=1h^N_\theta=1 along the whole critical line — and, if true, establish the protecting mechanism (the speculated 'semi-duality', i.e. a Kramers–Wannier duality acting on a half-infinite chain carrying the boundary operator σzτx\sigma^z\tau^x at its end).

(3) The Potts-point asymptotics. Give a controlled account — an analytic logarithmic-correction treatment of the marginally irrelevant boundary (and bulk) operators, or a certified extrapolation protocol — of the asymptotic approach to the free boundary fixed point at λ=1\lambda=1 under the three-state mixed (BCD) boundary perturbation, resolving the persistent plateau deviation between the finite-size spectra and the free-boundary conformal towers.

The source also proposes extending the simple-current method (the use of a chiral primary field whose fusion product with any primary field is again a single primary field, by which the chiral symmetry algebra of the theory is enlarged) and the boundary-state analysis to wider classes of rational and irrational conformal field theories, including generic Narain CFTs (the conformal field theories of free compact bosons on a torus, i.e. toroidal compactifications) and their deformations; that extension is a research program without a definite completion boundary and is not part of this problem window.

References

  1. Yifan Liu, Natalia Chepiga, Yoshiki Fukusumi, Masaki Oshikawa, Boundary critical phenomena in the quantum Ashkin-Teller model, SciPost Physics 20(5), 130 (2026), doi:10.21468/SciPostPhys.20.5.130, arXiv:2601.16951, https://arxiv.org/abs/2601.16951
  2. Masaki Oshikawa, Ian Affleck, Boundary conformal field theory approach to the critical two-dimensional Ising model with a defect line, Nuclear Physics B 495, 533 (1997), doi:10.1016/S0550-3213(97)00219-8, arXiv:cond-mat/9612187, https://arxiv.org/abs/cond-mat/9612187
  3. Davide Rossini, Ettore Vicari, Stacked quantum Ising systems and quantum Ashkin-Teller model, Physical Review B 113, 155143 (2026), doi:10.1103/y9n7-d6vr, arXiv:2601.18922, https://arxiv.org/abs/2601.18922
  4. Yifan Liu, Haruki Shimizu, Dongchang Liu, Kohei Kawabata, Extracting Boundary Conformal Data from Periodic Non-Hermitian Critical Chains, arXiv:2606.16785 (2026), https://arxiv.org/abs/2606.16785
  5. Qing-Yuan Zuo, Shuo Geng, Shan-Wen Tsai, Jin Zhang, Ground-state phase diagram of Rydberg atoms in a triangular-prism array, arXiv:2606.01116 (2026), https://arxiv.org/abs/2606.01116
  6. 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
  7. 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
  8. Christophe Chatelain, Tensor Renormalization-Group study of the surface critical behavior of a frustrated two-layer Ising model, arXiv:2510.21269 (2025), https://arxiv.org/abs/2510.21269
  9. Yihua Sun, Yuchen Fan, Relevant–marginal field mixing resolves the four-state Potts endpoint in the square-lattice J1–J2 Ising model, arXiv:2605.25946 (2026), https://arxiv.org/abs/2605.25946