A Reliable and Inexpensive Flexible Molecule Crystal Structure Prediction Protocol Based on First Principles.

Area of Science:

• Materials Science
• Computational Chemistry
• Crystallography

Background:

• Crystal structure prediction (CSP) is crucial for designing new materials.
• Current CSP methods often use empirical force fields (empFFs) or periodic density functional theory (pDFT+D).
• EmpFFs are unreliable for crystals with flexible monomers.

Purpose of the Study:

• To develop a reliable CSP method for flexible molecular crystals.
• To design and implement novel intramonomer force fields (intra-aiFFs) fitted to ab initio calculations.
• To improve the efficiency and accuracy of CSP protocols.

Main Methods:

• Developed intra-aiFFs by fitting to ab initio calculations for monomers.
• Combined intra-aiFFs with intermonomer aiFFs (inter-aiFFs) for full-dimensional CSP.
• Screened 1000 low-energy polymorphs using aiFFs, followed by pDFT+D calculations.
• Optimized top 100 polymorphs using pDFT+D.

Main Results:

• The novel aiFF-based CSP method accurately predicted the experimental crystal structure of 2-acetamido-4,5-dinitrotoluene as the second-ranked structure.
• This protocol achieved high accuracy at significantly lower computational cost compared to other reliable methods.
• The developed intra-aiFFs show promise for biomolecular simulations.

Conclusions:

• A new, reliable, and cost-effective CSP method for flexible molecular crystals has been established.
• The developed intra-aiFFs are crucial for accurate CSP of flexible systems.
• This approach has broad implications for materials discovery and computational drug design.