Non-Hermitian Multiconfiguration Molecular Mechanics

Area of Science:

• Computational chemistry
• Theoretical chemistry

Background:

• Accurate potential energy surface (PES) fitting is crucial for simulating complex reactive systems.
• Existing multiconfiguration molecular mechanics (MCMM) algorithms have limitations in fitting accuracy over broad geometric ranges.

Purpose of the Study:

• To introduce a novel version of the MCMM algorithm with improved capabilities for fitting PES.
• To enhance the accuracy and efficiency of computational modeling for chemical reactions.

Main Methods:

• Developed a new MCMM algorithm allowing non-Hermitian valence bond configuration interaction matrices.
• Implemented simplified gradients and Hessians for faster computation.
• Evaluated algorithm performance using two model reactive systems.

Main Results:

• The non-Hermitian matrix approach broadens the range of geometries for accurate PES fitting.
• The new algorithm demonstrates faster execution due to simpler gradients and Hessians.
• Successful application to two model reactions validates the enhanced MCMM performance.

Conclusions:

• The improved MCMM algorithm offers enhanced accuracy and efficiency for reactive system simulations.
• This development advances computational chemistry methods for studying chemical dynamics.
• The algorithm is suitable for fitting complex potential energy surfaces in computational studies.