Statistical mechanics

The group studies the statistical mechanics of many-body systems, with a focus on collective behavior that emerges when systems are driven out of thermal equilibrium or governed by interactions that do not decay quickly with distance. In equilibrium, the Boltzmann-Gibbs framework provides a complete and well-tested description of phases and phase transitions; far from equilibrium no such universal framework exists, and even steady states sustained by external driving or boundary currents can display ordering, long-range correlations, and phase transitions with no equilibrium counterpart. The work seeks the organizing principles behind these phenomena using exactly solvable models, large-deviation theory, and careful analysis of simple driven systems.

A second thread concerns systems with long-range interactions, such as gravitational, dipolar, and unscreened Coulomb forces, where energy is non-additive and the usual equivalence between statistical ensembles can break down. Such systems exhibit features that are impossible in short-range models, including negative specific heat, ensemble inequivalence, and anomalously slow relaxation through long-lived quasi-stationary states. The group also applies statistical-mechanical reasoning to the physical properties of biomolecules, for example the thermodynamics and kinetics of DNA denaturation, where collective transitions emerge from many coupled microscopic degrees of freedom.