A non-empirical intermolecular force-field for trinitrobenzene and its application in crystal structure prediction

Area of Science:

• Computational Chemistry
• Materials Science
• Crystallography

Background:

• Developing accurate intermolecular force fields is crucial for molecular simulations.
• Existing empirical models for organic crystals often struggle with predicting accurate lattice energies and structures.
• Trinitrobenzene (TNB) presents a unique challenge due to its complex intermolecular interactions.

Purpose of the Study:

• To develop a novel, non-empirical anisotropic atom-atom distributed intermolecular force-field (DIFF) for rigid trinitrobenzene (TNB).
• To assess the performance of the developed force field against empirical models in predicting the solid-state properties of TNB polymorphs.
• To establish a methodology for creating accurate force fields for molecular dynamics simulations.

Main Methods:

• Derivation of distributed multipole moments, polarizabilities, and dispersion coefficients from isolated TNB charge density.
• Fitting short-range parameters using symmetry-adapted perturbation theory dimer interaction energies and the distributed density-overlap model.
• Parameterization of long-range damping coefficients and relaxation of isotropic coefficients using second-order calculations.
• Comparison with empirical force fields (FIT) fitted to crystal structures using iterated stockholder atoms (ISAs) and Gaussian Distributed Analysis (GDMA) multipoles.

Main Results:

• The non-empirical DIFF models provide more accurate relative lattice energies for TNB polymorphs compared to empirical models.
• The developed models propose more realistic hypothetical crystal structures than empirical force fields.
• The DIFF-srL2(rel) model correctly identifies the most stable structure within the coordination sphere of TNB form III.
• The models show improved prediction of solid-state properties, including relative lattice energies and structural preferences.

Conclusions:

• The developed anisotropic atom-atom DIFF force field offers a significant improvement over empirical models for TNB solid-state properties.
• This non-empirical approach provides a more reliable method for predicting crystal structures and lattice energies.
• The methodology represents a step towards developing more comprehensive force fields for molecular dynamics simulations, capable of modeling all molecular phases.