First-order phase transition in Potts models with finite-range interactions (Q878368)

This paper deals with \(Q\)-states Potts model with Kac potential on \({\mathbb Z}^d\) \((Q \geq 3\), \(d \geq 2)\). In this model, the Hamiltonian in a finite region \( \Lambda \) with boundary conditions \(\sigma_{\Lambda^c}\) is given by \[ H_{\gamma,\Lambda}(\sigma_{\Lambda}| \sigma_{\Lambda^c}) = -\frac{1}{2}\sum_{i,j\in\Lambda}\gamma^dJ(\gamma(i-j)){\mathbf 1}_{\{\sigma(i)=\sigma(j)\}} - \sum_{i\in\Lambda, j\in \Lambda^c}\gamma^dJ(\gamma(i-j)){\mathbf 1}_{\{\sigma(i)=\sigma(j)\}}, \] where \( J(x) \) is supposed to be a spherically symmetric nonnegative function supported by the unit ball with bounded derivative. The result is: There exists \( \widetilde\gamma > 0\) such that for all \(\gamma \in (0, \widetilde\gamma)\), there exists \(\beta_c(\gamma)\) at which there are \(Q+1\) mutually independent Gibbs states. It indicates the first order phase transition of the model for sufficient large but finite range of interaction. This is contrast to the case of \( Q=3\), \(d=2 \) nearest neighbor interaction, for which available results show the second order phase transition. The method of the paper is based on the idea of perturbing around the mean field model (corresponding the limit \(\gamma \to 0\)) [\textit{J. L. Lebowitz} and \textit{O. Penrose}, J. Math. Phys. 7, 98--113 (1966; Zbl 0938.82520)] and the Pirogov-Sinai approach.