Large systems with Coulomb interactions: variational study and statistical mechanics (Q502142)

This article reviews a series of works by the author and her collaborators, on the asymptotics of minimizers of energy functionals on point sets \(x_1,\dots,x_n \in \mathbb{R}^d\) of the form \[ H_n(x_1,\dots,x_n) = \underset{i \neq j}{\sum} g(x_i - x_j) +n \underset{i}{\sum} V(x_i) \] where \(V\) is a confinement potential and \(g\) a repulsive interaction kernel with a singularity at the origin. An important particular case is when \(g\) is a multiple of the Coulomb interaction kernel in \(\mathbb{R}^d\). As detailed in the article, this problem arises in several areas of mathematics and natural sciences, such as interpolation theory, where they are known as Fekete point sets, and the study of vortices in condensed matter physics. Moreover, probability measures \(\mathbb{P}_{n,V,\beta}\) with density proportional to \(e^{-\beta H_n}\) play an important role in several areas of statistical physics, such as particle systems and random matrix theory (where for some some choices of potentials these measures describe the statistics of eigenvalues of large random matrices). The leading order of the asymptotics of \(H_n\) is well understood, in the sense of gamma-convergence. It describes the macroscopic behavior of the point sets, and is related to large deviations for the measures \(\mathbb{P}_{n,V,\beta}\). The author and her collaborators have investigated in depth the second-order term, which help to understand the microscopic behavior of point sets that minimize \(H_n\), and is given by a so-called renormalized energy. This article reviews the derivation of an expression for the renormalized energy, and applications to the description of minimizers of \(H_n\) at the microsocpic level (rigidity of the number of points in a set, equidistribution of energy, links with Abrikosov lattices) as well as to the asymptotic behavior of \(\mathbb{P}_{n,V,\beta}\) at the microscopic level (large deviations of particle configurations at a microscopic level and asymptotic expansion of the partition function).