Positioning of grid points for spanning potential energy surfaces-How much effort is really needed?

Area of Science:

• Computational Chemistry
• Theoretical Chemistry
• Quantum Chemistry

Background:

• Accurate potential energy surfaces (PESs) are essential for calculating molecular vibrational frequencies.
• The placement of grid points significantly influences PES accuracy and computational expense.
• Current methods often focus on reducing grid points without directly optimizing computational cost.

Purpose of the Study:

• To investigate the impact of different grid point positioning schemes on PES accuracy and computational cost.
• To compare static and adaptive (Gaussian process regression-based) schemes for PES generation.
• To evaluate these schemes for calculating anharmonic vibrational frequencies using vibrational configuration interaction theory.

Main Methods:

• Studied six grid point positioning schemes for n-mode expansions of PESs.
• Employed a static approach and five adaptive schemes utilizing Gaussian process regression.
• Assessed accuracy of fundamental vibrational modes and the number of required grid points for test molecules.

Main Results:

• Adaptive schemes show promise in optimizing PES generation compared to static methods.
• The number of grid points and computational cost are not directly correlated.
• Accuracy of anharmonic vibrational frequencies is sensitive to grid point placement.

Conclusions:

• Grid point positioning is a critical factor in the efficiency and accuracy of PES calculations.
• Adaptive schemes based on Gaussian process regression offer a more effective strategy for PES generation.
• The primary objective should be minimizing overall computational cost, not solely reducing grid points.