Renormalized-Residue-Based Multireference Configuration Interaction Method for Strongly Correlated Systems.

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Theoretical Chemistry

Background:

• Multireference configuration interaction (MRCI) methods face challenges with large active spaces due to configuration basis expansion and complex reduced density matrix (RDM) handling.
• Existing methods struggle to efficiently capture entanglement between active and inactive orbitals in complex quantum systems.

Purpose of the Study:

• To introduce a novel spin-adapted renormalized-residue-based MRCI (RR-MRCI) approach for quantum systems.
• To enhance computational accuracy and efficiency in electronic structure calculations for large active spaces.
• To develop an efficient algorithm for deploying spin-adapted matrix product state MRCI (MPS-MRCI).

Main Methods:

• Developed a spin-adapted RR-MRCI framework leveraging renormalized residues to capture orbital entanglement.
• Implemented an efficient algorithm for spin-adapted MPS-MRCI.
• Utilized quantum information theory with small buffer environments to enhance computational efficiency.

Main Results:

• The RR-MRCI method demonstrates improved computational accuracy and efficiency compared to internally contracted (ic) MRCI by considering orbital entanglement and using compressed matrix product state (MPS) structures.
• The approach offers enhanced computational efficiency over standard MPS-MRCI, with potential for application to large molecular systems.
• The RR framework, when combined with ic-MRCI, bypasses the need for high-rank RDMs through the use of distinct renormalized residues.

Conclusions:

• The developed RR-MRCI method is versatile and effective, as demonstrated across nine diverse molecular systems, including challenging cases with large active spaces.
• This approach provides a significant advancement in handling complex quantum systems, offering a more accurate and efficient alternative to existing MRCI methods.
• The method's ability to handle orbital entanglement and its computational efficiency make it a promising tool for future electronic structure studies.