Decoding Supramolecular Packing Patterns from Computed Anisotropic Deformability Maps of Molecular Crystals

The ability to encode and embed desired mechanical properties into active pharmaceutical ingredient solid forms would significantly advance drug development. In recent years, computational methods, particularly dispersion-corrected density functional theory (DFT), have come of age, opening the possibility of reliably predicting and rationally engineering the mechanical response of molecular crystals. Here, many-body dispersion and Tkatchenko-Scheffler dispersion-corrected DFT were used to calculate the elastic constants of a series of archetypal systems, including paracetamol and aspirin polymorphs and model hydrogen-bonded urea and π- π-bound benzene crystals, establishing their structure-mechanics relations. Both methods showed semiquantitative and excellent qualitative agreement with experiment. The calculations revealed that the plane of maximal Young's modulus generally coincides with extended H-bond or π-πnetworks, showing how programmable supramolecular packing dictates the mechanical behavior. In a pharmaceutical setting, these structure-mechanics relations can steer the molecular design of solid forms with improved physicochemical and compression properties.