Generalized Valence Bond Perfect-Pairing Made Versatile Through Electron-Pairs Embedding.

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Theoretical Chemistry

Background:

• Accurate description of electron correlation is crucial for understanding chemical phenomena.
• Generalized Valence Bond Perfect-Pairing (GVB-PP) offers a suitable reference but requires enhancement for complex correlation effects.
• Existing methods struggle with phenomena like symmetry breaking and phase transitions.

Purpose of the Study:

• To develop a uniformly accurate electron-pair-based method for diverse chemical systems.
• To improve upon the GVB-PP ansatz by incorporating electron pair correlation.
• To address limitations of uncorrelated electron-pair models, such as symmetry breaking.

Main Methods:

• Development of the Electron-Pair-based Extended Random Phase Approximation with Generalized Valence Bond Perfect-Pairing (EERPA-GVB) method.
• Embedding electron pairs in their environment to calculate intra- and inter-fragment correlation.
• Employing truncated extended random phase approximation (RPA) equations for computational efficiency.

Main Results:

• EERPA-GVB demonstrates high accuracy for systems with short-range (energy barriers) and long-range (molecular interactions) correlation.
• The method successfully resolves spatial symmetry breaking in benzene, a known issue for other models.
• Accurate prediction of the boron chain phase transition, where correlation character changes.

Conclusions:

• EERPA-GVB provides a versatile and accurate approach for electron correlation.
• The method overcomes limitations of previous electron-pair models.
• Its computational cost scales quadratically with the number of electron pairs, making it efficient.