Efficient and Robust Ab Initio Self-Consistent Field Acceleration Algorithm Based on a Semiempirical Model...

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Hartree-Fock (HF) and Density Functional Theory (DFT) calculations are fundamental in quantum chemistry.
• Accelerating the self-consistent field (SCF) convergence is crucial for computational efficiency.
• Existing SCF convergence techniques can be slow, especially for complex systems.

Purpose of the Study:

• To introduce a novel doubly iterative self-consistent field (SCF) approach, the SMH algorithm.
• To accelerate HF and DFT calculations by reducing the number of exact Fock matrix constructions.
• To improve the efficiency and robustness of SCF convergence.

Main Methods:

• Developed a semiempirical model Hamiltonian (SMH) algorithm for SCF calculations.
• Implemented a doubly iterative approach with exact Fock matrix construction and semiempirical microiterations.
• Combined SMH with standard convergence accelerators like DIIS and EDIIS.

Main Results:

• The SMH algorithm significantly reduces the number of exact Fock builds needed for SCF convergence.
• When integrated with DIIS, SMH improved convergence for organic molecules by ~10%.
• SMH achieved speedups of up to 60% for transition metal systems compared to EDIIS + DIIS.

Conclusions:

• The SMH algorithm is a reliable and robust SCF accelerator.
• This method offers substantial speedups for HF and DFT calculations, particularly for challenging systems.
• SMH seldom deteriorates convergence, making it a valuable tool in computational chemistry.