An Algorithm for Atom-Centered Lossy Compression of the Atomic Orbital Basis in Density Functional Theory...

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Large atomic orbital (AO) basis sets are crucial for accurate Kohn-Sham density functional theory (DFT) calculations.
• Increasing basis set size dramatically increases computational cost, posing a significant challenge.

Purpose of the Study:

• To develop and validate a threshold-based natural atomic orbital (NAO) scheme for AO basis set compression.
• To assess the trade-off between basis set compression and accuracy in DFT calculations.

Main Methods:

• Developed a threshold-based natural atomic orbital (NAO) scheme using eigenfunctions of the density matrix.
• Implemented a compression strategy by discarding NAOs below a specified occupation number threshold (10^-ϵ).
• Performed pilot test calculations using Hartree-Fock functional with quadruple-ζ (QZ) pc-3 basis sets.

Main Results:

• A threshold of 10^-5 achieved compression factors of 2.5–4.5 for QZ pc-3 basis sets, with relative energy errors below 0.1 kcal/mol.
• A threshold of 10^-7 yielded compression factors of 2–2.5 with significantly smaller errors (typically <0.01 kcal/mol).
• The developed scheme offers optimal basis set compression for a given threshold.

Conclusions:

• The threshold-based NAO scheme effectively compresses AO basis sets for DFT calculations.
• This method provides a computationally efficient approach to achieve high accuracy in electronic structure calculations.
• The technique allows for significant reduction in computational demand without compromising the reliability of results.