An adaptive potential energy surface generation method using curvilinear valence coordinates.

Area of Science:

• Computational Chemistry
• Molecular Spectroscopy

Background:

• Calculating vibrational spectra is crucial for understanding molecular properties.
• Accurate potential energy surfaces (PES) are essential for these calculations.
• Existing methods can be computationally expensive for large molecules.

Purpose of the Study:

• To present an automatic method, AGAPES, for generating potential energy surfaces (PES).
• To enable efficient calculation of vibrational spectra for large polyatomic molecules.
• To explore computational savings and scalability of the method.

Main Methods:

• Developed an automatic Born-Oppenheimer potential energy surface (PES) generation method (AGAPES).
• Employed an adaptive approach using intermediate vibrational calculations.
• Utilized a multi-mode expansion of the PES in internal valence coordinates.
• Tested on molecules: HNO, HClCO, and formaldoxime.

Main Results:

• Demonstrated significant computational savings in PES generation.
• Verified the versatility of AGAPES across different molecular structures.
• Examined the potential for linear scaling of sampling grid size with molecular size.
• Observed decreased correlation of remote coordinates in larger molecules.

Conclusions:

• AGAPES provides an efficient approach for computing vibrational spectra of large molecules.
• The method shows promise for scalability, reducing computational cost.
• Further improvements can enhance the applicability of AGAPES.