Low-Order Scaling G₀W₀ by Pair Atomic Density Fitting.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Accurate prediction of electronic properties is crucial for understanding molecular behavior.
• Traditional G0W0 calculations are computationally expensive, limiting their application to large systems.
• Developing scalable algorithms is essential for advancing computational chemistry.

Purpose of the Study:

• To develop and implement a low-scaling G0W0 algorithm for molecular electronic structure calculations.
• To demonstrate the scalability and efficiency of the new algorithm for large molecular systems.
• To assess the accuracy of the G0W0 implementation for predicting HOMO and LUMO energies.

Main Methods:

• Derivation of a low-scaling G0W0 algorithm using pair atomic density fitting (PADF).
• Implementation within the Slater type orbital (STO)-based Amsterdam density functional (ADF) code.
• Utilized an imaginary time representation of the Green's function for computational efficiency.

Main Results:

• Demonstrated asymptotic quadratic scaling for water clusters up to 432 atoms.
• Achieved significant computational speed-up, with large cluster calculations completed in 240 CPU hours.
• Reported average errors of 80 meV for large organic molecules on the augmented double-ζ level.

Conclusions:

• The developed low-scaling G0W0 algorithm offers a computationally efficient approach for large molecular systems.
• The method shows promise for studying single-particle properties of large organic systems like chromophores.
• Further improvements may be achieved by optimizing auxiliary fit sets for better representation of diffuse functions.