University of North Texas
Computational Study of Molecular Crystal Polymorphism and Continuum Solvation Model
Abstract
dc:descriptionMolecular dynamics simulations play a crucial role in understanding the structural and thermodynamic properties of molecular systems. It can provide valuable insights into the stability, energetics, and enantiotropic transformations of polymorphic crystals. However, the accuracy of classical molecular dynamics simulations heavily depends on the choice of the force field, which defines the atomic and molecular interactions. In this thesis, we investigate the conformational polymorphic crystal ROY (5-methyl 2-((2-nitrophenyl)amino)-3-thiophene carbonitrile), a highly polymorphic molecule with fourteen known crystal forms exhibiting distinct color and conformational variations. Classical molecular dynamics simulations are performed using the Generalized Amber Force Field (GAFF) and the Open Force Field (OpenFF) to evaluate their ability to reproduce crystal properties. The validation involves comparing physical properties, torsional angles, hydrogen bonding distances, and SMAC order parameters. The SMAC descriptors are further employed to classify the polymorphs, providing insights into their polymorphism, stability, and packing characteristics. Overall, this work demonstrates the importance of force field validation for complex molecular crystals. Additionally, we have developed and parameterized a continuum solvation model applicable to periodic DFT calculations. Specifically, we parameterize the Soft-Sphere Continuum Solvation (SSCS) model for non-aqueous and ionic liquid solvents. The SSCS model employs atom-centered, smoothly varying spheres to define solute cavities, enabling a physically consistent and computationally efficient treatment of solvation effects. Using large partition-coefficient databases, dielectric constants, and solvent densities (particularly for ionic liquids), we parameterize the model for a broad range of solvents through a high-throughput automated workflow. The optimized SSCS model achieves an average root-mean-square error (RMSE) of 0.85 kcal/mol for non-aqueous and 0.89 kcal/mol for ionic liquid solvents, within the threshold of chemical accuracy, using only a few adjustable parameters. The results highlight the model’s robustness, transferability, and predictive capability for diverse solvent environments.
Degree
thesis:*- Grantor dc:publisher
- University of North Texas
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Si, Pradip
- Contributors dc:contributor
-
- Cundari, Thomas R.
- Ko, Hsin-Yu
- Valsson, Omar
- Marshall, Paul
Subjects
dc:subject × 3Rights
dc:rights- Statement dc:rights
-
- Public
- Si, Pradip
- Copyright
- Copyright is held by the author, unless otherwise noted. All rights Reserved.
- Language dc:language
- English
Identifiers
dc:identifier.*- Identifier
-
local-cont-no: submission_4677
ark: ark:/67531/metadc2588573