System dynamics revealed by recurrence quantification analysis: application to molecular dynamics simulations.

The present work examines the applicability and efficacy of recurrence plots and recurrence quantification analysis in interpreting statistical-mechanics-based simulations of classical fluids and solids. We analyze temperature time series obtained from molecular dynamics simulations of a Lennard-Jones system at various fluid and solid states. It turns out that the structure of the recurrence plots reflects the different regimes of atomic motion as well as the degree of atomic diffusivity as the system density and temperature are varied. Recurrence plots (RPs) can help to localize a region where a phase transition occurs, while recurrence quantitative analysis descriptors confirm in a more clear way the results of RPs. The trends identified in our results are in qualitative agreement with direct computation of Lyapunov exponents for liquid Lennard-Jones systems reported in the literature.