Nondirect-Product Local Diabatic Representation with Smolyak Sparse Grids

Area of Science:

• Chemical Physics
• Computational Chemistry
• Quantum Dynamics

Background:

• Nonadiabatic conical intersection dynamics are crucial for understanding photophysical, photochemical, and biological processes.
• Accurate modeling of these dynamics is computationally demanding, especially for larger molecular systems.

Purpose of the Study:

• To develop and validate a novel computational method for nonadiabatic conical intersection dynamics.
• To reduce the computational cost associated with simulating these dynamics.
• To apply the method to realistic molecular systems.

Main Methods:

• Development of a nonadiabatic conical intersection wave packet dynamics method.
• Implementation in the local diabatic representation utilizing Smolyak sparse grids.
• Comparison with traditional direct-product grids for accuracy and efficiency.

Main Results:

• Sparse grids significantly reduce computational costs compared to direct-product grids.
• The method achieves excellent agreement with direct-product grid results for pyrazine and phenol models.
• Accurate results were obtained for a 4D pyrazine model where direct-product grids are infeasible.
• The method successfully applied to the Shin-Metiu model without quasi-diabatization.

Conclusions:

• Smolyak sparse grids offer an efficient and accurate approach for modeling nonadiabatic conical intersection dynamics.
• This method alleviates the computational scaling issues of traditional grid-based methods.
• The developed technique is applicable to complex, realistic molecular systems and opens new avenues for theoretical studies.