Strong electron correlation from partition density functional theory.

Area of Science:

• Computational chemistry
• Quantum mechanics
• Materials science

Background:

• Standard approximations in Kohn-Sham density functional theory (KS-DFT) struggle with strongly correlated systems.
• Partition-DFT (PDFT) offers an exact reformulation of KS-DFT using fragment calculations and inter-fragment interactions.

Purpose of the Study:

• To suppress typical errors of the local density approximation (LDA) in KS-DFT for strongly correlated systems.
• To introduce and validate the multi-fragment overlap approximation (MFOA) for the partition energy in PDFT.

Main Methods:

• Applying the multi-fragment overlap approximation (MFOA) to the partition energy within PDFT.
• Utilizing the local density approximation (LDA) for fragment calculations.
• Testing the method on one-dimensional linear hydrogen chains.

Main Results:

• MFOA significantly suppresses LDA errors in strongly correlated systems.
• Improved LDA dissociation curves for hydrogen chains, comparable to spin-unrestricted LDA but preserving spin symmetry.
• MFOA corrects LDA electron density, capturing dimerization in hydrogen chains.
• Quantitative agreement with density matrix renormalization group (DMRG) for dissociation energies with an additional 1D LDA correction.

Conclusions:

• MFOA is a promising approximation for improving PDFT calculations of strongly correlated systems.
• The method offers an accurate and spin-symmetric alternative to existing approximations.
• Further refinements can lead to quantitative agreement with highly accurate methods.