Large deviations of the finite cluster shape for two-dimensional percolation in the Hausdorff and \(L^1\) metric (Q1582130)

This paper provides two large deviation principles for the shapes of random sets arising in percolation theory. Consider Bernoulli bond percolation on the two-dimensional square lattice in the percolating phase \(p>1/2\). Conditional on the event that the open connected component \(C\) of the origin is finite it is shown that -- as \(N\to\infty\) the laws of the random sets \(C/N\) satisfy a large deviation principle with respect to the Hausdorff metric, -- let \(f:{\mathbf N}\to{\mathbf R}\) be such that \(f(N)/N\to 0\) and \(f(N)/\log N\to\infty\). The laws of \(\{x\in{\mathbb{R}}^2: d(x,C)\leq f(N)\}/N\) satisfy a large deviation principle with respect to the \(L^1\) metric associated with planar Lebesgue measure. The two rate functions are suitable surface energies and coincide on regular sets. The second large deviation principle is applied to estimate the asymptotic behaviour of the probability that the random set \(\{x\in{\mathbb{R}}^2: d(x,C)\leq f(N)\}\) has a finite area larger than \(N^2\). This provides a link to the Wulff construction.