Connectivity in the potential energy landscape for binary Lennard-Jones systems.

Connectivity in the potential energy landscape of a binary Lennard-Jones system can be characterized at the level of cage-breaking. We calculate the number of cage-breaking routes from a given local minimum and determine the branching probabilities at different temperatures, along with correlation factors that represent the repeated reversals of cage-breaking events. The number of reversals increases at lower temperatures and for more fragile systems, while the number of accessible connections decreases. We therefore associate changes in connectivity with super-Arrhenius behavior. Reversals in minimum-to-minimum transitions are common, but often correspond to "non-cage-breaking" processes. We demonstrate that the average waiting time within a minimum shows simple exponential behavior with decreasing temperature. To describe the long-term behavior of the system, we consider reversals and connectivity in terms of the "cage-breaking" processes that are pertinent to diffusion [V. K. de Souza and D. J. Wales, J. Chem. Phys. 129, 164507 (2008)]. These cage-breaking events can be modeled by a correlated random walk. Thus, a full correlation factor can be calculated using short simulations that extend up to two cage-breaking events.