Critical manifold of the Potts model: Exact results and homogeneity approximation

Abstract

The q-state Potts model has stood at the frontier of research in statistical mechanics for many years. In the absence of a closed-form solution, much of the past efforts have focused on locating its critical manifold, trajectory in the parameter \q, eJ\ space where J is the reduced interaction, along which the free energy is singular. However, except in isolated cases, antiferromagnetic (AF) models with J<0 have been largely neglected. In this paper we consider the Potts model with AF interactions focusing on deducing its critical manifold in exact and/or closed-form expressions. We first re-examine the known critical frontiers in light of AF interactions. For the square lattice we confirm the Potts self-dual point to be the sole critical point for J>0. We also locate its critical frontier for J<0 and find it to coincide with a solvability condition observed by Baxter in 1982. For the honeycomb lattice we show that the known critical point holds for all J, and determine its critical qc = 1 2 (3+ 5) = 2.61803 beyond which there is no transition. For the triangular lattice we confirm the known critical point to hold only for J>0. More generally we consider the centered-triangle (CT) and Union-Jack (UJ) lattices consisting of mixed J and K interactions, and deduce critical manifolds under homogeneity hypotheses. For K=0 the CT lattice is the diced lattice, and we determine its critical manifold for all J and find qc = 3.32472. For K=0 the UJ lattice is the square lattice and from this we deduce both the J>0 and J<0 critical manifolds and find qc=3 for the square lattice. Our theoretical predictions are compared with recent tensor-based numerical results and Monte Carlo simulations.

0

Turn this paper into a lesson

ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.

Discussion (0)

Sign in to join the discussion.

Loading comments…