Framework for Inverse Mapping Chemistry-Agnostic Coarse-Grained Simulation Models into Chemistry-Specific Models.

Area of Science:

• Computational chemistry
• Molecular modeling
• Materials science

Background:

• Coarse-grained (CG) models enable molecular simulations at larger scales.
• An inverse-design challenge exists in mapping specific chemistries to generic CG models.
• Current methods lack automated tools for optimizing this mapping.

Purpose of the Study:

• To develop and validate tools for automatically generating and refining the mapping of chemistry-specific (CS) molecules to CG models.
• To identify optimal CS-molecule candidates for CG model representation based on objective criteria.
• To assess the sensitivity of optimal CS-molecule selection to CG model parameters.

Main Methods:

• Developed automated tools for mapping CS-molecule candidates to CG model constraints.
• Utilized representative optimization criteria and an objective function to assess mapping fit.
• Applied the methodology to diverse CG models, from small molecules to block copolymers.

Main Results:

• Successfully identified optimal CS-molecule candidates for various CG models.
• Demonstrated the ability to uncover the underlying length scale of CG models.
• Validated the methodology by correctly identifying known molecular chemistries and selecting plausible candidates for unknown systems.
• Showcased the sensitivity of the optimal chemistry to CG model variations.

Conclusions:

• The developed tools provide an automated and effective solution for the inverse-design problem in CG modeling.
• This approach enhances the accuracy and applicability of CG simulations across different molecular systems.
• The methodology offers a robust way to select appropriate molecular chemistries for CG representations.