Modeling microsolvation clusters with electronic-structure calculations guided by analytical potentials and...

Area of Science:

• Computational Chemistry
• Physical Chemistry
• Materials Science

Background:

• Studying microsolvation clusters of alkali-metal ions with rare-gas atoms is crucial for understanding ion-solvent interactions.
• Traditional methods for characterizing these clusters can be computationally expensive and time-consuming.

Purpose of the Study:

• To develop and validate a novel computational methodology for efficiently studying alkali-metal ion microsolvation clusters.
• To identify an accurate density functional theory (DFT) approach for characterizing these systems.
• To leverage machine learning (ML) for predicting and optimizing cluster structures.

Main Methods:

• Global optimization using an analytical potential energy surface (PES) and an evolutionary algorithm (EA).
• Systematic benchmark study of DFT functionals and basis sets to determine the optimal approach.
• Application of ML classification algorithms to predict the mapping of low-energy PES minima to DFT minima.
• Re-optimization of selected low-energy minima at the DFT level.

Main Results:

• The B3LYP-D3/aug-pcseg-1 DFT approach was identified as the most suitable for Li+Krn (n = 2-14, 16) clusters.
• The ML classifier accurately predicted the majority of structures requiring DFT re-optimization.
• The methodology significantly enhanced computational efficiency, allowing for the study of larger clusters.

Conclusions:

• The proposed methodology provides an efficient and accurate approach for studying microsolvation clusters.
• Combining global optimization, DFT benchmarking, and ML prediction accelerates the characterization of complex chemical systems.
• This work paves the way for investigating larger and more complex microsolvation phenomena.