Efficient periodic resolution-of-the-identity Hartree-Fock exchange method with k-point sampling and Gaussian basis...

Area of Science:

• Condensed matter physics
• Computational chemistry
• Materials science

Background:

• Hybrid density functional theory (DFT) simulations of condensed matter systems are computationally expensive.
• The high cost is primarily due to the non-local Hartree-Fock exchange (HFX) and the need for thermodynamic limit approximations.
• Efficient calculation of HFX in periodic systems is crucial for advancing materials simulations.

Purpose of the Study:

• To develop an efficient method for calculating HFX in periodic systems using k-point sampling.
• To reduce the computational complexity of hybrid DFT simulations.
• To enable accurate and scalable simulations of condensed matter systems.

Main Methods:

• A local atom-specific resolution-of-the-identity (RI) scheme.
• Use of atom-centered Gaussian type orbitals.
• Truncation of Coulomb interaction and auxiliary density matrix method for efficiency.
• Real-space approach with linear scaling with k-points.
• Implementation in the CP2K software package.

Main Results:

• The method demonstrates excellent agreement with Γ-point supercell calculations for relative energies and nuclear gradients.
• Achieved good strong and weak scaling performance.
• Successfully implemented graphics processing unit (GPU) acceleration.

Conclusions:

• The developed method significantly enhances the efficiency of HFX calculations in periodic systems.
• This approach offers a promising solution for high-performance computing in condensed matter simulations.
• The implementation provides accurate and scalable computational tools for materials science research.