FCIQMC-CASPT2 with Imaginary-Time-Averaged Wave Functions

Area of Science:

• Quantum chemistry
• Computational physics
• Theoretical chemistry

Background:

• Accurate electronic structure calculations are crucial for understanding molecular properties.
• Large active spaces in quantum chemistry present significant computational challenges.
• Full configuration interaction quantum Monte Carlo (FCIQMC) offers a path to treat large active spaces.

Purpose of the Study:

• To develop a novel method for performing second-order perturbation theory (CASPT2) on large active spaces.
• To address numerical instabilities and fermionic positivity violations in such calculations.
• To apply and benchmark the new method on relevant chemical systems.

Main Methods:

• Complete active space second-order perturbation theory (CASPT2).
• Optimization of large active spaces using full configuration quantum Monte Carlo (FCIQMC).
• Computation of density matrices from imaginary-time-averaged wave functions.

Main Results:

• The developed method resolves fermionic positivity violations, ensuring numerical stability.
• Successful application to complex systems like [NiFe]-hydrogenase, [Cu2O2]-oxidase, and Fe-porphyrin models.
• Calculations involving up to 26 electrons in 27 orbitals were performed and benchmarked.

Conclusions:

• The new CASPT2 protocol is robust and numerically stable for large active spaces.
• This method provides a reliable approach for electronic structure calculations of challenging molecular systems.
• The findings pave the way for more accurate theoretical studies in catalysis and bioinorganic chemistry.