Crystal Polymorph Search in the NPT Ensemble via a Deposition/Sublimation Alchemical Path

Area of Science:

• Computational chemistry
• Materials science
• Crystallography

Background:

• Active pharmaceutical ingredient formulation requires identifying stable polymorphs with specific properties.
• In silico crystal structure prediction (CSP) aids experimental screening but often lacks realistic condition modeling.
• Current CSP methods using force fields or quantum mechanics do not fully integrate temperature, pressure, or solution conditions.

Purpose of the Study:

• To develop an innovative alchemical path for crystal structure prediction using direct NPT ensemble sampling.
• To enhance sampling efficiency for crystal structure prediction by incorporating alchemical intermediates and advanced sampling techniques.
• To demonstrate the algorithm's capability in reproducing experimentally determined polymorphs.

Main Methods:

• Utilized an advanced polarizable atomic multipole force field (AMOEBA) for crystal structure prediction.
• Employed an alchemical path with nonphysical intermediates, a novel Monte Carlo barostat, and orthogonal space tempering.
• Applied the algorithm to 2-((4-(2-(3,4-dichlorophenyl)ethyl)phenyl)amino)benzoic acid (XAFPAY) using short NPT ensemble simulations.

Main Results:

• Successfully reproduced all experimentally determined polymorphs of XAFPAY with one molecule in the asymmetric unit.
• Achieved high sampling efficiency, generating over 300 ns/day on a dual CPU node using the Force Field X software.
• Demonstrated the feasibility of predicting crystal structures through direct NPT ensemble sampling.

Conclusions:

• The proposed alchemical path and NPT ensemble sampling represent a significant advancement in crystal structure prediction.
• This method enhances the accuracy and efficiency of CSP, aiding pharmaceutical development.
• Future work can integrate solution conditions into CSP algorithms using continuum solvation methods.