Scaling limit for a class of gradient fields with nonconvex potentials (Q624661): Difference between revisions

The authors consider the scaling limit of gradient Gibbs measures (GGM) with non-convex potentials. While for stricly convex potentials, the translation-invariant, ergodic GGMs are completely characterized by their tilt, a non-convex potential may lead to several ergodic gradient measures with zero tilt [the first author and \textit{R. Kotecký}, Probab. Theory Related Fields 139, 1--39 (2007; Zbl 1120.82003)]. Here, the tilt represents the average incline of typical configurations. The main result of the paper is that every translation-invariant, ergodic GGM with zero tilt and a possibly non-convex potential of the form \[ V(\eta) = -\log \int \varrho(d\kappa) \exp\left[ -\frac{1}{2} \kappa \eta^2 \right] \] scales to a Gaussian free field. In the final section of the paper the authors discuss the open question of whether this result also holds for GGMs with non-zero tilt.