One Size Fits All? Development of the CPOSS209 Data Set of Experimental and Hypothetical Polymorphs for Testing Computational Modeling Methods DOI Creative Commons
Louise S. Price, Matteo Paloni, Matteo Salvalaglio

et al.

Crystal Growth & Design, Journal Year: 2025, Volume and Issue: 25(9), P. 3186 - 3209

Published: April 28, 2025

Organic crystal structure prediction (CSP) studies have led to the rapid development of methods for predicting relative energies known and computer-generated structures. There is a compromise between level theoretical treatment, its reliability across different types organic systems, how accuracy depends on size shape unit cell, number structures that can be modeled at an affordable computational cost. We used our database studies, often performed as complement experimental screening, produce sets comprising 6 15 structures, covering polymorphs, observed packings closely related molecules, CSP-generated energetically competitive but distinct 20 molecules. These been chosen illustrate some issues need consideration in any lattice energy method, seeking generally applicable moderate-sized including small drug included crystallization reported polymorphs. In all examples, original CSP electronic calculations molecule give conformational anisotropic atom-atom model electrostatic intermolecular energy, combined with empirical "exp-6" repulsion dispersion energy. The are compared those obtained by reoptimizing periodic, plane-wave, dispersion-corrected density functional theory, specifically PBE TS correction, single point where many body (MBD) correction applied, example widely "workhorse" method. use this data set preliminary test modeling illustrated two Machine Learned Foundation Models, MACE-MP-0 MACE-OFF23. challenges putative polymorphs range their energies, possible agreement illustrated. Very similar molecules differ significantly observed, only partially reflecting polymorph screening experiments produced approaches based purely thermodynamic paradigm.

Language: Английский

Polymorphic ROYalty: The 14th ROY Polymorph Discovered via High-Throughput Crystallization DOI Creative Commons
Jake Weatherston, Michael R. Probert, Michael J. Hall

et al.

Journal of the American Chemical Society, Journal Year: 2025, Volume and Issue: unknown

Published: March 25, 2025

Polymorphism, when a substance can exist in more than one crystalline form yet return to the same liquid or solution phase, is characterized by differences packing molecular conformation. Polymorphs often exhibit differing physical properties, and are therefore particularly important development of materials pharmaceuticals. However, gaining thorough understanding solid-state landscape molecule requires exhaustive experimental screening crystallization conditions, particular challenge using classical methods. We show that high-throughput Encapsulated Nanodroplet Crystallization (ENaCt) enable rapid efficient exploration highly polymorphic molecules, through an in-depth study 5-methyl-2-((2-nitrophenyl)amino)thiophene-3-carbonitrile (ROY), most small known. An ENaCt screen encompassing 1536 individual experiments, spanning 320 unique resulted direct access single crystals, suitable for X-ray diffraction analysis, all six known polymorphs accessible from (Y, R, YN, ON, ORP R18). In addition, two (Y04 Y19) previously accessed only via melt heteroseeded new polymorph ROY (O22) were obtained. Furthermore, identification first solvate (ROY· methyl anthranilate) example dimer, formed situ oxidation. thus shown be impactful tool mapping molecules and, discovery O22, has ensured tetradecamorphic retains record molecule.

Language: Английский

Citations

1
One Size Fits All? Development of the CPOSS209 Data Set of Experimental and Hypothetical Polymorphs for Testing Computational Modeling Methods DOI Creative Commons
Louise S. Price, Matteo Paloni, Matteo Salvalaglio

et al.

Crystal Growth & Design, Journal Year: 2025, Volume and Issue: 25(9), P. 3186 - 3209

Published: April 28, 2025

Organic crystal structure prediction (CSP) studies have led to the rapid development of methods for predicting relative energies known and computer-generated structures. There is a compromise between level theoretical treatment, its reliability across different types organic systems, how accuracy depends on size shape unit cell, number structures that can be modeled at an affordable computational cost. We used our database studies, often performed as complement experimental screening, produce sets comprising 6 15 structures, covering polymorphs, observed packings closely related molecules, CSP-generated energetically competitive but distinct 20 molecules. These been chosen illustrate some issues need consideration in any lattice energy method, seeking generally applicable moderate-sized including small drug included crystallization reported polymorphs. In all examples, original CSP electronic calculations molecule give conformational anisotropic atom-atom model electrostatic intermolecular energy, combined with empirical "exp-6" repulsion dispersion energy. The are compared those obtained by reoptimizing periodic, plane-wave, dispersion-corrected density functional theory, specifically PBE TS correction, single point where many body (MBD) correction applied, example widely "workhorse" method. use this data set preliminary test modeling illustrated two Machine Learned Foundation Models, MACE-MP-0 MACE-OFF23. challenges putative polymorphs range their energies, possible agreement illustrated. Very similar molecules differ significantly observed, only partially reflecting polymorph screening experiments produced approaches based purely thermodynamic paradigm.

Language: Английский

Citations

0