No label defined (Q4497729)

Consider a Bernoulli bond percolation process on a lattice \(Z^d\), i.e. place i.i.d. random variables taking the values \(1\) and \(0\) with probability \(p\) resp. \(1-p\) on the nearest neighbor bonds of \(Z^d\). It is well known that there exists a critical value \(p_c\) such that for \(p\) below \(p_c\), all clusters of bonds carrying a \(1\) (``open bonds'') are finite, while for \(p\) above \(p_c\) there exists a unique infinite cluster of open bonds. This book studies the properties of the finite cluster of open bonds containing a given point, say the origin, when \(p>p_c\). It has been known for a long time that the probability that such a cluster is larger than \(n\) is bounded from above and below by \(\exp(-c_\pm n^{(d-1)/d})\). It is shown that this probability behaves exactly like \(\exp(-c n^{(d-1)/d})\), and the constant \(c\) is identified as the minimum of a rate function, more precisely the so called Wulff-functional on the set of shape functions, describing the macroscopic shape of the droplet. Moreover, a stability result is proven that ensures that the macroscopic shape of a cluster of prescribed volume \(n\), in the limit \(n\uparrow \infty\), will be given with probability one, by the minimizer of this variational problem, the so-called Wulff shape. The problem considered here is an old one. The Wulff construction to determine the equilibrium shape of a crystal from the knowledge of the surface tension was introduced a century ago by Wulff on purely thermodynamic grounds. A derivation of this construction on the basis of microscopic statistical mechanics was first obtained for the two-dimensional Ising model only in the work of \textit{R. L. Dobrushin, R. Kotecký} and \textit{S. Shlosman} [``Wulff construction. A global shape from local interaction'' (1992; Zbl 0917.60103)]. The problem has ever since then drawn considerable attention, and numerous improvements and extensions have been provided [see in particular \textit{D. Ioffe} and \textit{R. H. Schonmann}, Commun. Math. Phys. 199, No. 1, 117-167 (1998; Zbl 0929.60076)]. All these results, however, concerned the case of two-dimensional models, and their extension to three dimensions was a desideratum that encountered considerable obstacles. The result coming closest to a solution of the three-dimensional problem concerned a particular class of so called Kac-Ising models with long range interactions [\textit{O. Benois, T. Bodineau} and \textit{E. Presutti}, Stochastic Processes Appl. 75, No. 1, 89-104 (1998; Zbl 0927.60080)], but a gap remained between the upper and lower bound that would vanish only in the limit of infinite interaction range. The contribution of the present book is thus the first full solution of a Wulff problem for a three-dimensional model which constitutes a major breakthrough in the field. Although its basic approach, in particular the coarse graining procedure and the topology used are quite similar to that of Benois et al., it brought an important clarification through the identification of the surface tension in the rate function. While certain aspects in the present work were simplified through independence properties particular to the percolation model, it was soon recognized that these were of minor importance, and in several papers the present work is already extended to interacting models such as the three-dimensional Ising model [\textit{T. Bodineau}, Commun. Math. Phys. 207, No. 1, 197-229 (1999) and \textit{R. Cerf} and \textit{A. Pisztora}, ``On the Wulff crystal in the Ising model'' (Preprint 1999)]. For an extensive review on the problem see \textit{T. Bodineau, D Ioffe} and \textit{Y. Velenik} [J. Math. Phys. 41, No. 3, 1033-1098 (2000)].