A BREAKDOWN OF EQUILIBRIUM STATISTICAL-MECHANICS

We consider the recent paper "Ultrafast Dynamics of a Quasi-Dissociative Diatomic Molecule in Solution" by Zhu and Robinson, in which the results of molecular dynamics simulations and their interpretation by the authors seem to contradict basic principles of classical equilibrium statistical mechanics. We discuss both the results expected from equilibrium statistical mechanics and the results of tests by computer simulation. Two types of simulations are presented: (i) a "canonical" case with frequent connections to a heat bath, and (ii) a "microcanonical" case which repeats a critical portion of Zhu and Robinson's molecular dynamics. We pay particular attention to the question of whether the numerical data can justify the rejection of equilibrium statistical mechanical expectations. © 1990 American Chemical Society.