Stochastic-geometry description of critical Ashkin–Teller interfaces: the boundary-operator map, the four fractal dimensions, SLE$_4$ traces, and FK–spin duality in the $c=1$ extended conformal field theory

The Ashkin–Teller (AT) model has a self-dual critical line described by the $\mathbb{Z}_2$ orbifold of the compactified free boson, a $c=1$ extended conformal field theory (CFT). Picco and Santachiara [Phys. Rev. E 83, 061124 (2011)] measured the fractal dimensions of spin interfaces and Fortuin–Kasteleyn (FK) cluster boundaries along this line and found: (i) at a generic point the spin-interface fractal dimension takes only one of four values; (ii) the $(13|24)$ interface has $d_f=3/2$ all along the line, with transfer-matrix evidence for a boundary operator of dimension $x_b=1/4$, suggesting an SLE$_4$ description; and (iii) FK-cluster boundaries satisfy the Duplantier duality $(1-d_s)(1-d_{FK})=1/4$ with the associated spin interfaces. None of these facts is derived. The problem is to give a theoretical (boundary-CFT/Coulomb-gas or rigorous-probability) account of this geometry: construct the map from lattice spin boundary conditions to the continuum boundary operators of the orbifold theory at general compactification radius, use it to derive the four fractal dimensions and the SLE$_4$ identification of the $x_b=1/4$ interface, extend the FK–spin duality beyond the specific one-Ising-component construction, and determine the asymptotic fractal dimensions near the four-state Potts endpoint with controlled logarithmic corrections. These questions remain open: existing work supplies a conjectural fractal-dimension Ansatz, discrete-holomorphic arguments for an SLE$(4,\rho,\rho)$ description, non-convergent numerics that even complicate the four-value picture, a boundary-BCFT construction for the quantum chain only, and rigorous level-line results only at the decoupled-Ising point with the AT line stated as a conjecture.

Background

The Ashkin–Teller (AT) model places on each site $i$ of a square lattice a pair of Ising spins $(\sigma_i,\tau_i)$ interacting through the Hamiltonian $H=-\sum_{\langle ij\rangle}(K(\sigma_i\sigma_j+\tau_i\tau_j)+K_4\,\sigma_i\sigma_j\tau_i\tau_j)$, so that it interpolates between two decoupled Ising models ($K_4=0$) and, at a special coupling, the four-state Potts model. Its ferromagnetic phase diagram contains a self-dual critical line along which the critical exponents vary continuously: in the continuum limit the line is described by the compactified free boson (a Gaussian model with central charge $c=1$) modulo a $\mathbb{Z}_2$ reflection, i.e. the $\mathbb{Z}_2$ orbifold of the compactified boson, with the compactification radius $r$ parametrizing the line. Landmark points on the line are the four-state Potts endpoint, the Fateev–Zamolodchikov (FZ) point where the model coincides with the $\mathbb{Z}_4$-parafermionic spin chain, and the decoupled-Ising point (two independent Ising models, equivalent to the XOR-Ising model obtained by multiplying the two spin fields).

A spin interface is a domain wall forced by boundary conditions: writing the four local states as $1=(++),2=(+-),3=(--),4=(-+)$, a condition such as $(13|24)$ forces the states $\{1,3\}$ on one side of the domain and $\{2,4\}$ on the other, and for each configuration the interface is the curve on the dual lattice separating the two basins. The symmetries of the AT model leave essentially four non-equivalent non-branching interfaces of this type, conventionally labeled $(1|234)$, $(1|2)$, $(12|34)$ and $(13|24)$. At criticality such a curve is a random fractal whose length scales with system size $L$ as $L^{d_f}$, defining the fractal dimension $d_f$. The Schramm–Loewner evolution SLE$_\kappa$ is the family of conformally invariant random curves parametrized by $\kappa$; an SLE$_\kappa$ trace has fractal dimension $1+\kappa/8$, and the SLE/CFT dictionary relates $\kappa$ to the conformal dimension of the boundary-condition-changing operator (the operator that, inserted where the interface meets the boundary, changes one boundary condition into the other). For a single Ising model the interface is SLE$_3$; level lines of the Gaussian free field are SLE$_4$, of fractal dimension $3/2$.

For models described by minimal CFTs (Ising, Potts, $O(n)$ loops) the geometric content is largely understood: the high-temperature polygons become the level lines of the Coulomb-gas (free-boson) representation, their fractal dimensions follow from Coulomb-gas scaling dimensions, and convergence to SLE curves is in several cases proven. The AT line is different in two ways. First, it is an extended CFT: besides conformal symmetry it carries the internal $\mathbb{Z}_4$ symmetry, and its bulk theory is the orbifold compactified boson. Second, the natural geometrical objects here are spin interfaces, curves coexisting with fluctuating non-local degrees of freedom that are not fixed locally by the curve; the lattice interface then loses the Markov property that underlies the standard SLE construction. What is known is fragmentary: Oshikawa and Affleck [Phys. Rev. Lett. 77, 2604 (1996)] classified boundary states on orbifolds and their defect lines; Santachiara [Nucl. Phys. B 793, 396 (2008)] derived CFT predictions for interface fractal dimensions at the FZ point; and at the FZ point the $\mathbb{Z}_4$ boundary-condition classification of Picco and Santachiara [J. Stat. Mech. (2010) P07027] explains some measured values. For general compactification radius no such identification exists.

Fortuin–Kasteleyn (FK) clusters are the random-cluster representation of a spin model: bonds between like spins are occupied with a percolation-like probability, producing clusters whose boundaries are the critical loops of the associated loop model. In the Potts model the Duplantier duality $(1-d_s)(1-d_{FK})=1/4$ relates the fractal dimension $d_s$ of spin-cluster boundaries to that $d_{FK}$ of FK-cluster boundaries; in Coulomb-gas language it is an electric–magnetic (T-duality) transformation. For the AT model Picco and Santachiara constructed FK clusters from one Ising component with wired boundary conditions and found the duality satisfied numerically along the whole line — evidence that the duality extends to an extended CFT, but only for that particular construction. Related facts: generalized FK clusters attempted for $\mathbb{Z}_4$ and $\mathbb{Z}_5$ models were found not to percolate at criticality, and a duality had earlier been observed at the random-bond Potts fixed point [Jacobsen et al., Phys. Rev. Lett. 102, 070601 (2009)].

Picco and Santachiara [Phys. Rev. E 83, 061124 (2011)] simulated all four interface types along the critical line and found that, at a generic point, the spin-interface fractal dimension takes only one of four values, one of which is $d_f=3/2$ all along the line (the $(13|24)$ interface, consistent with $x_b=1/4$ and an SLE$_4$ expectation), and that the measured FK boundary dimensions satisfy the Duplantier relation with the associated spin interfaces along the entire line. Near the four-state Potts endpoint the measured values degrade because that endpoint has a marginally irrelevant operator — a perturbation that fades only logarithmically slowly under coarse-graining — producing strong logarithmic corrections to finite-size scaling.

Problem Statement

Establish a theoretical — boundary-CFT/Coulomb-gas or rigorous-probability — description of the critical interfaces of the two-dimensional Ashkin–Teller model that derives the geometric facts discovered numerically by Picco and Santachiara along the entire self-dual critical line, which is the $\mathbb{Z}_2$ orbifold of the compactified free boson. The common objective is the extension of the SLE/CFT correspondence from minimal CFTs to this extended $c=1$ theory, with the following four jointly supporting deliverables, all of which a full resolution must address and any single one of which, if refuted, is independently reportable: (1) construct the map from the lattice spin boundary conditions $(1|234)$, $(1|2)$, $(12|34)$, $(13|24)$ to the complete set of continuum boundary operators (boundary states) of the orbifold theory at general compactification radius, and from it determine whether, and if so why, the spin-interface fractal dimension collapses to only four values along the whole line, including at the decoupled-Ising point; (2) identify analytically the continuum boundary operator of dimension $x_b=1/4$ generating the $(13|24)$ condition, and derive that the resulting interface is described by SLE$_4$ (or the appropriate SLE$(4,\rho,\rho)$) traces at every point of the AT family, not only at special points; (3) derive the Duplantier duality $(1-d_s)(1-d_{FK})=1/4$ between FK-cluster boundaries and spin interfaces in this extended CFT, and determine whether it holds beyond the one-Ising-component wired cluster definition for the natural FK-type constructions — namely the same cluster definition applied to the other Ising component and the $\mathbb{Z}_4$-symmetric cluster definitions built from the Coulomb-gas weights — toward a general FK–spin duality framework for extended CFTs; and (4) determine the asymptotic fractal dimensions at and near the four-state Potts endpoint, where marginally irrelevant operators generate logarithmic corrections, with these corrections controlled — that is, obtained from a finite-size extrapolation whose leading correction exponent is either fixed analytically or fitted and shown to be stable against the choice of fitting window and consistent with the quoted statistical errors.

References

1. M. Picco and R. Santachiara, "Critical interfaces and duality in the Ashkin Teller model", Phys. Rev. E 83, 061124 (2011). DOI: 10.1103/PhysRevE.83.061124; arXiv:1011.1159. https://arxiv.org/abs/1011.1159
2. Y. Ikhlef and M. A. Rajabpour, "Discrete holomorphic parafermions in the Ashkin–Teller model and SLE", J. Phys. A: Math. Theor. 44, 042001 (2011). DOI: 10.1088/1751-8113/44/4/042001; arXiv:1009.3374. https://arxiv.org/abs/1009.3374
3. Y. Ikhlef and M. A. Rajabpour, "Spin interfaces in the Ashkin–Teller model and SLE", J. Stat. Mech. (2012) P01012. DOI: 10.1088/1742-5468/2012/01/P01012; arXiv:1111.3197. https://arxiv.org/abs/1111.3197
4. M. Caselle, S. Lottini and M. A. Rajabpour, "Critical domain walls in the Ashkin–Teller model", J. Stat. Mech. (2011) P02039. DOI: 10.1088/1742-5468/2011/02/P02039; arXiv:0907.5094. https://arxiv.org/abs/0907.5094
5. Y. Fukusumi, M. Picco and R. Santachiara, "Spin interfaces and crossing probabilities of spin clusters in parafermionic models", arXiv:2006.09850 (2020, preprint). https://arxiv.org/abs/2006.09850
6. Y. Liu, N. Chepiga, Y. Fukusumi and M. Oshikawa, "Boundary critical phenomena in the quantum Ashkin–Teller model", SciPost Phys. 20, 130 (2026). DOI: 10.21468/SciPostPhys.20.5.130; arXiv:2601.16951. https://arxiv.org/abs/2601.16951
7. T. Alcalde López and A. Sepúlveda, "Excursion decomposition of the XOR-Ising model", arXiv:2602.06011 (2026, preprint). https://arxiv.org/abs/2602.06011
8. N. Lamsen, Y. Diouane and G. Delfino, "Critical points in coupled Potts models and correlated percolation", J. Stat. Mech. (2023) 013203. DOI: 10.1088/1742-5468/aca901; arXiv:2208.14844. https://arxiv.org/abs/2208.14844
9. M. Oshikawa and I. Affleck, "Defect Lines in the Ising Model and Boundary States on Orbifolds", Phys. Rev. Lett. 77, 2604 (1996). DOI: 10.1103/PhysRevLett.77.2604. https://doi.org/10.1103/PhysRevLett.77.2604
10. R. Santachiara, "SLE in self-dual critical Z(N) spin systems: CFT predictions", Nucl. Phys. B 793, 396 (2008). DOI: 10.1016/j.nuclphysb.2007.09.029; arXiv:0705.2749. https://arxiv.org/abs/0705.2749
11. J. L. Jacobsen, P. Le Doussal, M. Picco, R. Santachiara and K. J. Wiese, "Critical Interfaces in the Random-Bond Potts Model", Phys. Rev. Lett. 102, 070601 (2009). DOI: 10.1103/PhysRevLett.102.070601. https://doi.org/10.1103/PhysRevLett.102.070601
12. M. Picco and R. Santachiara, "Critical interfaces of the Ashkin–Teller model at the parafermionic point", J. Stat. Mech. (2010) P07027. DOI: 10.1088/1742-5468/2010/07/P07027; arXiv:1005.0493. https://arxiv.org/abs/1005.0493
13. A. Banerjee, P. Jana and P. K. Mohanty, "Geometric percolation of spins and spin-dipoles in Ashkin-Teller model", Phys. Rev. B 111, 014403 (2025). DOI: 10.1103/PhysRevB.111.014403; arXiv:2411.11644. https://arxiv.org/abs/2411.11644
14. N. Nemkov and S. Ribault, "Analytic conformal bootstrap and Virasoro primary fields in the Ashkin-Teller model", SciPost Phys. 11, 089 (2021). DOI: 10.21468/SciPostPhys.11.5.089; arXiv:2106.15132. https://arxiv.org/abs/2106.15132