Science Abstract

Abstract


The Feynman path integral formulation of quantum statistical mechanics is reviewed in detail. The evaluation of path integrals by molecular dynamics methods is discussed, and specific algorithms for treating the quantum canonical (NVT) and isothermal-isobaric (NPT) ensembles are developed. These algorithms employ a variety of novel techniques that address the problem of ergodic phase space sampling and ensure rapid convergence of quantum equilibrium properties compared to more standard Monte Carlo approaches. Specifically, variable transformations, Nosé-Hoover chain thermostatting, and multiple time scale integration are combined to produce algorithms that are almost as efficient as the best Monte Carlo methods. The molecular dynamics approach is often preferable to Monte Carlo, in that it can be both vectorized and parallelized more efficiently. In addition, certain methods, such as the Car-Parrinello ab inito molecular dynamics or the recently introduced adiabatic centroid molecular dynamics methods are inextricably tied to a dynamical propagation scheme. In these lectures, the equations of motion that generate the different quantum ensembles are presented, tested and used in a variety of applications.