Stretching Bonds without Breaking Symmetries in Density Functional Theory.

Area of Science:

• Computational Chemistry
• Quantum Mechanics
• Electronic Structure Theory

Background:

• Kohn-Sham density functional theory (KS-DFT) offers a balance of accuracy and efficiency for electronic structure calculations.
• A persistent challenge in KS-DFT is the "symmetry dilemma," where achieving chemically accurate energies requires breaking fundamental symmetries.
• Standard density functional approximations often necessitate artificial symmetry breaking for accurate results.

Purpose of the Study:

• To present an embedding framework that addresses the symmetry dilemma in KS-DFT.
• To develop a method that improves the accuracy of energy calculations without compromising charge or spin symmetries.
• To explore a new computational approach using fragment spin densities as primary variables.

Main Methods:

• Developed an embedding framework utilizing fragment spin densities instead of total molecular densities.
• Constructed a novel functional approximation, termed the "overlap approximation," based on the spatial overlap of fragment densities.
• Applied the method to assess binding energies of molecules, including covalently bonded and strongly correlated systems.

Main Results:

• The proposed embedding framework partially resolves the symmetry dilemma inherent in KS-DFT.
• The "overlap approximation" significantly improves the accuracy of semilocal KS-DFT binding energies.
• The method successfully avoids artificial breaking of charge and spin symmetries in calculations.

Conclusions:

• The fragment spin density embedding approach offers a promising solution to the KS-DFT symmetry dilemma.
• This physically motivated "overlap approximation" enhances computational accuracy for molecular systems.
• The framework demonstrates applicability to diverse chemical systems, including challenging strongly correlated ones.