DFVB: A Density-Functional-Based Valence Bond Method.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Ab initio valence bond (VB) theory accurately describes static electron correlation but often neglects dynamic correlation.
• Existing methods to include dynamic correlation in VB theory can be computationally expensive.

Purpose of the Study:

• Introduce a new ab initio valence bond method incorporating density-functional theory (DFT) for correlation correction, termed DFVB.
• To provide a computationally efficient approach for improving the accuracy of ab initio VB calculations.

Main Methods:

• The DFVB method utilizes density correlation functionals to account for dynamic correlation energy.
• Static correlation energy is handled by the Valence Bond Self-Consistent Field (VBSCF) wave function.
• The method's performance is tested on spectroscopic parameters, dipole moments, singlet-triplet gaps, reaction barriers, and resonance energies.

Main Results:

• DFVB accurately calculates spectroscopic parameters for diatomic molecules.
• The method precisely determines dipole moments for the NF molecule across various electronic states.
• DFVB effectively computes singlet-triplet gaps, reaction barriers, and charge-shift resonance energies for diradical species.

Conclusions:

• The DFVB method offers a computationally economical route to enhance the accuracy of ab initio VB theory.
• DFVB achieves high accuracy comparable to existing post-VBSCF methods but with significantly lower computational cost.