MCSCF optimization revisited. II. Combined first- and second-order orbital optimization for large molecules

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Theoretical Chemistry

Background:

• Multiconfiguration self-consistent field (MCSCF) methods are crucial for accurately describing electron correlation in molecules.
• Traditional optimization methods can suffer from slow convergence, limiting their application to larger systems.
• Efficient orbital optimization is key to advancing theoretical chemistry calculations.

Purpose of the Study:

• To introduce a new, efficient orbital optimization method for MCSCF.
• To improve convergence speed and robustness compared to existing techniques.
• To enable accurate electronic structure calculations for larger and more complex molecular systems.

Main Methods:

• The SO-SCI method combines second-order (SO) optimization for active orbitals with first-order super configuration interaction (SCI) optimization for remaining rotations.
• Density fitting is employed to efficiently reuse intermediates for calculating two-electron integrals.
• A limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) quasi-Newton method accounts for CI-orbital coupling.

Main Results:

• The SO-SCI method demonstrates significantly improved convergence compared to conventional SCI methods.
• Computational overhead is minimized by reusing gradient calculation intermediates for the active Hessian.
• The method shows enhanced efficiency and robustness in benchmark calculations on aromatic molecules and transition metal complexes.

Conclusions:

• The SO-SCI method offers a substantial advancement in MCSCF orbital optimization.
• Its efficiency and robustness make it suitable for large molecular systems.
• This method paves the way for more accurate and feasible quantum chemical calculations.