Critical exponents of planar gradient percolation (Q948744)

The author considers a gradient percolation model on the triangular lattice. Percolation theory was introduced in order to explain the diffusion of a fluid in a porous media. Since then, it has been applied to a variety of problems in physics and several other branches of science. In particular, the percolation problem posseses many challenges in mathematical physics and probability theory. Such richness of so many applications may be explained by the universality character presented by the percolation problem, in spite of its simplicity. Because of this, the main results do not depend on the details of a particular model or system. In the simplest case the model is defined on a regular lattice and we have only a parameter \(p\) which controls the system. On the square lattice, for instance, a \textit{site percolation} model is defined by the probability \(p\) that a site is occupied, while \(1-p\) is the probability that the site is empty. The main problem is to find the critical value \(p_c\): for \(p < p_c\) the system develops some finite non-percolating clusters; for \(p > p_c\) there are clusters of occupied sites which percolate the whole lattice, i.e., one can connect sites belonging to opposite boundaries through a walk along occupied sites. About 20 years ago a gradient percolation model was introduced. In this case, the parameter \(p\) is no longer uniform but varies along a given direction and we are faced with an inhomogeneous site percolation model. In the paper under consideration the author considers the case where the probability for a given site to be occupied varies linearly with the distance along a fixed direction. The paper consists of 6 sections. In Section 1 we have a general introduction to the problem. In Section 2 we have a review on some known results for the homogeneous percolation problem which are later used in the work. In particular, the concept of arm is introduced and some results about the arm exponents are reviewed. In Section 3 we find the definition of the gradient percolation model on the triangular lattice. The occupancy probability for a site along the \(y\)-direction is given by \(p(y) = 1/2 - y/2N ~(-N < y < N)\), and do not depends on the \(x\)-coordinate. The definition of a front is given and the localization of the front is discussed. The author proves a theorem ensuring that the front remains close to \(y = 0\) within a strip of width which goes as \( N^{4/7}\). He also proves the uniqueness of the front. In Section 4 the author presents his analysis on the length of the front. This is achieved by the introduction of two-arm probability estimates as an extension of previous results presented in Sec. 2. It is proved that the defined length beheaves as \(N^{3/7}\) and a bound on the length variance is also obtained. In Section 5 we have the definition of the outer (\textit{upper} and \textit{lower}) boundaries and it is showed that their respective lengths has the same behaviour, both scaling as \(N^{4/21}\), and a bound on the corresponding variances are also given. Finally, in Section 6 we have a discussion on some remaining problems. The results presented in the paper are very important for this class of problem and can be used as a guide to more sofisticad models, usually treated by numerical simulations.