Can Coarse-Grained Molecular Dynamics Simulations Predict Pharmaceutical Crystal Growth?

Area of Science:

• Computational chemistry
• Materials science
• Crystallography

Background:

• Predicting crystal growth is crucial for pharmaceutical development.
• Coarse-grained (CG) molecular dynamics (MD) offers a computationally efficient alternative to atomistic simulations.
• Evaluating CG methods for predicting relative crystal facet growth rates is essential.

Purpose of the Study:

• To assess the ability of two CG MD strategies to predict pharmaceutical crystal facet growth rates.
• To compare CG simulation results with established theories like Bravais-Friedel-Donnay-Harker (BFDH) and attachment energy (AE).

Main Methods:

• Applied two CG strategies: MARTINI-level mapping with Particle Swarm Optimization (PSO) and a coarser-grained Iterative Boltzmann Inversion (IBI) method.
• Used CG force fields to simulate crystal growth from the melt for phenytoin and carbamazepine.
• Compared simulation predictions with BFDH and AE theories.

Main Results:

• Both CG strategies successfully stabilized crystal structures and predicted growth from the melt using modest computational resources.
• The finer-grained PSO-generated force field (FF) using MARTINI mapping demonstrated superior accuracy in predicting facet growth rates and crystal morphology for carbamazepine.
• While most methods performed well for phenytoin, the IBI-generated model showed limitations.

Conclusions:

• CG MD simulations are viable tools for predicting pharmaceutical crystal growth.
• The choice of CG strategy significantly impacts prediction accuracy, with finer-grained approaches like MARTINI-PSO showing better performance for complex morphologies.