Influences of electrostatic models on organic crystal structure prediction – a case study of pentacene

Abstract

Crystal structure prediction (CSP) is widely used to explore the molecular packing of organic semiconductors. The reliability of CSP is related to the accurate ranking of trial crystals, which is dependent on the description of the intermolecular interaction. The intermolecular interaction has several components including non-polarizable and polarizable electrostatic interactions and van der Waals interactions. Among them, the electrostatic interaction plays a key role in the identification of the molecular packing. However, the influence of electrostatic interaction models on the CSP of organic crystals is largely unknown. In this study, we use a pentacene model system to systematically evaluate the impact of four different electrostatic models coupled with MM2 standard Buckingham vdW interactions on the reliability of CSP. The four electrostatic models include the nonpolar point-charge model (Charge), atomic multipole (AMP) model, polarizable AMOEBA model, and QM-derived polarizable force field (QM-PFF) model. The evolution algorithm based USPEX is employed to predict the structure of pentacene. Our results demonstrate that the anisotropic electrostatic intermolecular interaction is a decisive factor to obtain the correct herringbone packing structures. And the accurate polarizable model has superior performance in the CSP of pentacene since it can provide more accurate crystal ranking.