Global aspects of the energy landscape of metastable crystal structures inionic compounds

Understanding the properties of complex systems requires the analysis of th eir energy landscape. Using simple models for the crystalline compounds MgF 2 and CaF2 as example landscapes, we show that, in addition to the construc tion of tree-graph approximations of the landscape, the threshold algorithm can be employed to calculate local densities of states for regions of the landscape around local minima, and to determine the probabilities of transi tions between such minima. This information is then correlated with the res ults of global optimizations and relaxations.