The Monte Carlo threshold algorithm applied to crystal structure prediction landscapes of polycyclic aromatic hydrocarbons

Crystal structure prediction (CSP) is a valuable tool for predicting the crystalline arrangement that a molecular compound will adopt. Although powerful, the technique is limited to locating the local minima of the crystal potential energy surface (PES). The Monte Carlo threshold (MCT) algorithm can be used to extract more information from this PES, namely the energy barriers between local minima. We use this combined approach to analyze the crystal PESs of three polycyclic aromatic hydrocarbons (PAHs): phenanthrene, pyrene, and perylene. We also investigate the impact of the choice of potential energy model, comparing three atom- atom potentials: FIT, PAHAP and isoPAHAP. We find that the choice of potential does not have a large impact on the relative ranking of experimental polymorphs or the energy barriers that separate them. The data obtained from the MCT algorithm allows us to rationalize the differences in experimental conditions under which each crystal polymorph is observed, and find that the crystal packing PES of PAHs is highly frustrated. This study shows the value of using the MCT algorithm to enrich our knowledge of the crystal structure landscapes of small molecules.