Efficient global geometry optimization of clusters

The method of global geometry optimization of atomic and molecular clusters by evolutionary algorithms is briefly presented and reviewed. As an exemplary application of a parallelized implementation of such an algorithm, neutral pure water clusters are globally optimized. In contrast to previous studies, the sophisticated and quantitatively reliable TTM2-F potential is employed. Significant qualitative differences to the earlier results are found, implicating a breakdown of simple water models for water clusters of non-trivial size