Exploration of Minimum Energy Paths Through Feature-Vector Force-Driven Reaction Coordinate Exploration Approach

Area of Science:

• Computational Chemistry
• Theoretical Chemistry
• Chemical Kinetics

Background:

• Predicting reaction minimum-energy paths (MEPs) and locating transition states (TSs) are computationally intensive tasks in chemistry.
• Current methods struggle with large molecular systems due to high computational costs of electronic structure calculations.

Purpose of the Study:

• To benchmark the Feature-vector Force-driven Reaction Coordinate Exploration (FFoRCE) approach for generating MEPs and TSs.
• To assess the accuracy and efficiency of FFoRCE compared to traditional methods.

Main Methods:

• The FFoRCE approach utilizes a two-step process: a reaction-coordinate navigator to generate an initial path, followed by local relaxations.
• The navigator transforms coordinate space into features, assigning unique values to structures along the reaction coordinate.
• Symmetry function-constrained geometry optimization is employed to generate MEP structures by minimizing energy along the TS.

Main Results:

• FFoRCE was benchmarked on 10 elementary gas-phase organic reactions.
• Geometries, activation energies, and vibrational frequencies from FFoRCE showed fair agreement with whole-molecule DFT calculations.
• The method successfully reproduced MEPs and located TSs accurately.

Conclusions:

• The FFoRCE approach provides an accurate and computationally efficient method for tracing MEPs and locating TSs.
• FFoRCE shows promise for studying complex reaction mechanisms where traditional methods are prohibitive.
• This method can accelerate the exploration of chemical reaction pathways.