Empirical Valence-Bond Models for Reactive Potential Energy Surfaces Using Distributed Gaussians

Area of Science:

• Chemical Physics
• Computational Chemistry
• Theoretical Chemistry

Background:

• Accurate potential energy surfaces (PES) are crucial for understanding chemical reaction dynamics.
• Existing methods for constructing empirical valence bond (EVB) PES can suffer from divergence problems.
• The generalized Gaussian approach provides a foundation for developing improved PES construction techniques.

Purpose of the Study:

• To present a novel method for constructing empirical valence bond potential energy surfaces.
• To generalize the approach to accommodate an arbitrary number of data points.
• To address and overcome divergence issues inherent in some existing PES construction methods.

Main Methods:

• The V12(2)(q) method represents the PES using a Gaussian function multiplied by a polynomial at the transition state.
• The method builds upon the generalized Gaussian approach.
• Applications involve testing on two model surfaces and the HCN isomerization reaction.

Main Results:

• The V12(2)(q) method successfully constructed empirical valence bond potential energy surfaces.
• The applications demonstrated the method's ability to overcome divergence problems.
• The study discussed the implications of using different coordinate systems (Cartesian vs. internal/redundant internal).

Conclusions:

• The presented V12(2)(q) method offers a robust approach for constructing EVB potential energy surfaces.
• This method provides a viable alternative to existing techniques, mitigating common divergence issues.
• The findings contribute to more accurate simulations of chemical reactions.