Assessing orbital optimization in variational and diffusion Monte Carlo

Area of Science:

• Computational Physics
• Quantum Chemistry
• Materials Science

Background:

• Correlated magnetic systems present significant challenges for electronic structure calculations.
• Variational Monte Carlo (VMC) and diffusion Monte Carlo (DMC) are powerful quantum mechanical methods.
• The quality of orbitals significantly impacts the accuracy of DMC calculations.

Purpose of the Study:

• To investigate the impact of orbital optimization within VMC on DMC results for correlated magnetic systems.
• To assess whether orbital optimization systematically improves DMC energies and other observables.
• To use Chromium Sulfide Bromide (CrSBr) as a model system for these investigations.

Main Methods:

• Employed variational Monte Carlo (VMC) with orbital optimization.
• Performed diffusion Monte Carlo (DMC) calculations using optimized and standard density functional theory (DFT) orbitals.
• Incorporated short-range Jastrow factors and analyzed locality and fixed-node errors.
• Calculated both energies and other observables to assess bias.

Main Results:

• Short-range Jastrow factors are crucial for improving DMC, irrespective of orbital quality.
• Orbital optimization in VMC led to worse DMC energies compared to standard DFT orbitals, attributed to increased pseudopotential locality errors.
• Orbital optimization improved the fixed-node error in DMC and systematically reduced mixed-estimator bias for observables other than energy.

Conclusions:

• Orbital optimization in VMC is a reliable method for enhancing variational and pure fixed-node energies.
• It effectively reduces mixed-estimator bias for observables beyond the energy.
• Despite potential drawbacks in energy accuracy due to pseudopotentials, orbital optimization offers benefits for correlated magnetic systems.