Stochastic Multiconfigurational Self-Consistent Field Theory.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Multiconfigurational self-consistent field (MCSCF) theory is foundational for strongly correlated molecular problems.
• Conventional MCSCF methods face computational cost limitations that grow exponentially with the number of correlated orbitals.

Purpose of the Study:

• To develop a novel MCSCF approach that overcomes the computational limitations of traditional methods.
• To enable the study of larger and more complex strongly correlated systems.

Main Methods:

• Optimization of determinant coefficients using stochastic Full Configuration Interaction Quantum Monte Carlo (FCIQMC).
• Variational update of nonlinear orbital rotation parameters based on sampled wave functions.
• Introduction of stochastic noise into determinant amplitudes and orbital rotation gradients/Hessians.

Main Results:

• The new stochastic MCSCF approach achieves robust convergence for orbital optimization.
• Stochastic noise aids in avoiding local minima, leading to improved energy solutions.
• The method successfully treats polycyclic aromatic hydrocarbons, including coronene, with large active spaces (24 electrons in 24 orbitals).

Conclusions:

• The developed stochastic MCSCF method significantly reduces computational cost for strongly correlated systems.
• This approach expands the scope of treatable molecular systems, allowing for accurate electronic structure calculations with modest resources.
• The findings demonstrate the potential of integrating stochastic techniques into ab initio electronic structure calculations.