Approximating Periodic Potential Energy Surfaces with Sparse Trigonometric Interpolation.

Area of Science:

• Computational chemistry
• Theoretical chemistry
• Quantum chemistry

Background:

• Potential energy surfaces (PES) are crucial for understanding chemical systems.
• Constructing full-dimensional PESs for large molecules is computationally intensive.
• Previous methods using polynomial interpolation struggle with periodic properties of angular coordinates.

Purpose of the Study:

• To develop a novel method for constructing surrogate potential energy surfaces (PES).
• To address the limitations of polynomial interpolation for periodic properties in PES.
• To improve the accuracy and physical realism of PES for molecular systems.

Main Methods:

• Utilized sparse interpolation with trigonometric basis functions.
• Focused on systems where reaction coordinates are primarily torsion angles.
• Constructed reduced-dimensional PESs.

Main Results:

• The trigonometric interpolation basis guarantees the periodicity of the PES gradient for torsion angles.
• This approach resulted in a periodically repeating PES.
• Achieved slightly lower approximation errors compared to polynomial interpolation.

Conclusions:

• Trigonometric interpolation is a superior method for constructing PES with periodic angular coordinates.
• This technique enhances the physical accuracy of computational chemistry models.
• Offers a more reliable approach for studying molecular systems with torsional dynamics.