Deterministic/Fragmented-Stochastic Exchange for Large-Scale Hybrid DFT Calculations.

Area of Science:

• Computational Chemistry
• Quantum Mechanics
• Materials Science

Background:

• Accurate evaluation of exact exchange is computationally demanding for large systems in density functional theory.
• Grid- and plane-wave-based methods face challenges in efficiently handling long-range interactions.

Purpose of the Study:

• To develop an efficient computational approach for range-separated exact exchange within the GKS-DFT framework.
• To enable the calculation of long-range exchange for large molecular systems with high accuracy.

Main Methods:

• Fragmenting the Coulomb kernel in reciprocal space.
• Employing a mixed deterministic-stochastic representation for efficient computation.
• Projecting the Hamiltonian onto valence and conduction states from local-DFT calculations.

Main Results:

• The method efficiently calculates long-range exchange for large molecular systems (hundreds to thousands of states).
• Accurate convergence of HOMO and LUMO energies is achieved with a limited number of states.
• Facilitates excellent tuning of long-range separated hybrids (RSH) for complex systems.

Conclusions:

• The developed method offers an efficient and accurate way to compute exact exchange in large systems.
• This approach is suitable for studying electronic properties of complex molecular systems, including biological light-harvesting complexes.