Taming the virtual space for incremental full configuration interaction.

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Electronic Structure Theory

Background:

• Incremental full configuration interaction (iFCI) offers accurate approximations to the full configuration interaction (FCI) limit with polynomial computational cost.
• Existing iFCI methods face limitations with larger basis sets, hindering broader applicability.
• The choice of virtual orbitals is crucial for the convergence and accuracy of iFCI calculations.

Purpose of the Study:

• Introduce a novel natural orbital (NO) screening approach, incremental NO full configuration interaction (iNO-FCI).
• To enhance the efficiency and scalability of iFCI methods by optimizing virtual orbital selection.
• To enable accurate electronic structure calculations with significantly larger basis sets.

Main Methods:

• Developed iNO-FCI by incorporating a natural orbital screening strategy into the iFCI framework.
• Maximized consistency in virtual orbital selection across different many-body terms.
• Employed a principle of error cancellation by using the same set of virtual NOs for interdependent terms.

Main Results:

• Achieved computational savings of up to 95% compared to traditional iFCI methods.
• Demonstrated the ability to utilize larger, previously computationally prohibitive basis sets.
• Validated iNO-FCI performance on challenging systems including double-bond dissociation, biradicals, conjugated systems, and a transition metal complex.

Conclusions:

• iNO-FCI significantly reduces computational cost while maintaining high precision in electronic structure calculations.
• The novel NO screening approach extends the applicability of iFCI to larger basis sets.
• iNO-FCI represents a substantial advancement for accurate quantum chemical modeling of complex molecular systems.