Reducing overprediction of Molecular Crystal structures via threshold clustering Patrick Butler and Graeme M. Day School of Chemistry, University of Southampton, UK The structure of crystalline molecular materials is crucial to applications ranging from pharmaceuticals to organic semi-conductors. Polymorphs, crystals of the same compound with different structures, are consequently of great interest as they can significantly alter the chemophysical properties. Over the past decades there has been considerable effort to improve our understanding of polymorphism. However, determining the accessible polymorphs of a given compound still relies mostly on screening crystallisations under a wide range of conditions. 1 Crystal structure prediction (CSP) has shown potential to augment polymorph screens. The conventional CSP approach involves locating all minima on the potential energy surface for the target compound and then ranking them, typically by energy, to identify the most likely observed structures. 2 Despite significant advances, a consistent issue of this approach is it invariably overpredicts the number of polymorphs with no ability to distinguish which of the candidate structures are experimentally accessible. One of the causes for this overprediction is in neglecting thermal effects that cause potential energy minima, separated by relatively small energy barriers, to coalescence into a single basin at finite temperate. 3 While methods based on a series of molecular dynamics (MD) and enhanced sampling simulations have been proposed 4–6 to account for this such approaches have not become widely adopted due to the complexity in both the simulations and in the processing and analysis of the results, which adds significant cost on top of generating the initial CSP landscape. Moreover, these methods typically rely on common MD force fields rather than the more elaborate and accurate energy models typically required for CSP leading to ambiguity in the connection between the original CSP landscape and the reduced set of structures. Considering the above, we propose a simple method underpinned by the Monte Carlo threshold algorithm for clustering potential energy minima into basins. 7 We demonstrate this method, termed threshold clustering, on calculated CSP landscapes for benzene, acrylic acid and resorcinol—three systems of varying intermolecular interactions and conformational flexibility. The results show threshold clustering can significantly reduce the number of candidate structures on CSP landscapes while retaining matches to experimental structures. This is achieved without a complex pipeline and moreover on the same energy surface as the original CSP, therefore eliminating any ambiguity regarding the connection between the reduced structure set and the original landscape.
References 1. A. J. Cruz-Cabeza, S. M. Reutzel-Edens and J. Bernstein, Chem. Soc. Rev. , 2015, 44 , 8619–8635. 2. D. H. Bowskill, I. J. Sugden, S. Konstantinopoulos, C. S. Adjiman and C. C. Pantelides, Annu. Rev. Chem. Biomol. Eng. , 2021, 12 , 593–623. 3. E. C. Dybeck, D. P. McMahon, G. M. Day and M. R. Shirts, Cryst. Growth Des. , 2019, 19 , 5568–5580. 4. P. Raiteri, R. Martoňák and M. Parrinello, Angew. Chem. Int. Ed. , 2005, 44 , 3769–3773. 5. E. Schneider, L. Vogt and M. E. Tuckerman, Acta Crystallogr. Sect. B Struct. Sci. Cryst. Eng. Mater. , 2016, 72 , 542–550. 6. N. F. Francia, L. S. Price, J. Nyman, S. L. Price and M. Salvalaglio, Cryst. Growth Des. , 2020, 20 , 6847–6862. 7. P. W. V. Butler and G. M. Day, 2022, DOI: 10.26434/chemrxiv-2022-jmd78
P07
© The Author(s), 2023
Made with FlippingBook Learn more on our blog