Multireference configuration interaction theory using cumulant reconstruction with internal contraction of density...

Area of Science:

• Quantum Chemistry
• Computational Chemistry

Background:

• Accurately describing electron correlation is crucial in quantum chemistry.
• Existing methods struggle with large active spaces, limiting accuracy for complex systems.
• Density Matrix Renormalization Group (DMRG) excels at static correlation but needs dynamic correlation correction.

Purpose of the Study:

• To develop a novel multireference configuration interaction (MRCI) method.
• To integrate DMRG's static correlation capabilities with MRCI's dynamic correlation treatment.
• To enable accurate quantum chemical calculations in significantly larger active spaces.

Main Methods:

• Developed a multireference configuration interaction (MRCI) method scalable to large active spaces.
• Combined DMRG and MRCI (DMRG-MRCI) for a comprehensive treatment of static and dynamic correlation.
• Utilized commutators to eliminate five-particle rank reduced density matrices (RDMs) and cumulant reconstruction for four-particle rank RDMs.
• Employed computer-aided tensor contraction for efficient implementation and parallel computing.

Main Results:

• Successfully combined DMRG and MRCI, overcoming complexity associated with high-rank RDMs.
• The DMRG-MRCI method accurately includes high-level dynamic correlation.
• Demonstrated the method's efficacy through benchmark applications, including the singlet-triplet gap of free-base porphyrin with 24 active orbitals.
• Extended the approach to size-consistency-corrected variants like MRCI+Q, MR-ACPF, and MR-AQCC.

Conclusions:

• The developed DMRG-MRCI method significantly advances the ability to treat static and dynamic electron correlation.
• This computational approach allows for accurate quantum chemical calculations in much larger active spaces than previously possible.
• The method provides a robust framework for studying complex molecular systems with high accuracy.