Magnetizabilities at Self-Interaction-Corrected Density Functional Theory Level.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Magnetizability calculations are crucial for understanding magnetic properties of molecules.
• Accurate prediction requires reliable density functional approximations (DFAs).
• Previous studies show varying performance of DFAs for magnetic properties.

Purpose of the Study:

• To evaluate the performance of various density functionals for calculating molecular magnetizabilities.
• To assess the impact of self-interaction correction (SIC) on the accuracy of these calculations.
• To identify the most suitable DFAs for magnetizability predictions.

Main Methods:

• Utilized a high-quality ab initio coupled cluster benchmark set for magnetizabilities.
• Tested a range of density functionals: LDA, GGA, and meta-GGA.
• Investigated the effect of self-interaction correction (SIC) on DFA performance.

Main Results:

• The performance of density functionals varied significantly across different complexity levels.
• Self-interaction correction (SIC) led to a substantial reduction in errors for most functionals.
• Magnetizability calculations are highly sensitive to the physical realism of the exchange-correlation potential.

Conclusions:

• Standard density functionals often struggle with accurate magnetizability predictions.
• Self-interaction correction is essential for improving the reliability of density functional theory for magnetic properties.
• Future development should focus on exchange-correlation potentials that minimize self-interaction error.