Biased gradient squared descent saddle point finding method.

Area of Science:

• Computational Chemistry
• Chemical Dynamics
• Theoretical Chemistry

Background:

• Calculating chemical reaction rates is crucial for understanding chemical processes.
• Transition state theory (TST) is a key theoretical framework, often relying on identifying saddle points on potential energy surfaces.
• Existing methods for finding saddle points can be computationally intensive and may require prior knowledge of reaction pathways or product states.

Purpose of the Study:

• To develop a novel saddle point finding method for chemical reaction rate calculations.
• To present a method that does not require knowledge of specific product states.
• To offer a computationally efficient alternative to existing saddle point search algorithms.

Main Methods:

• The proposed method transforms the potential energy landscape into the square of the gradient.
• This transformation converts all critical points into global minima.
• A biasing term is incorporated to stabilize relevant low-energy saddle points near a minimum of interest.

Main Results:

• The method successfully identifies low energy saddle points without needing product state information.
• The approach converts critical points into global minima, simplifying the search.
• The method demonstrates competitiveness with the dimer min-mode following method in terms of force evaluations.

Conclusions:

• The presented saddle point finding method offers an efficient approach for calculating chemical reaction rates.
• Its ability to bypass the need for product state information simplifies the theoretical treatment of chemical reactions.
• This method provides a valuable tool for computational chemists studying reaction dynamics.