Does the ionization potential condition employed in QTP functionals mitigate the self-interaction error?

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Materials Science

Background:

• Kohn-Sham eigenvalues are often interpreted as vertical ionization potentials (IPs).
• Density functional approximations (DFAs) are evaluated for fractionally charged systems to assess self-interaction errors (MSIEs).
• Satisfying the IP theorem in DFAs is linked to mitigating MSIEs.

Purpose of the Study:

• To demonstrate that optimizing DFAs for IPs also reduces MSIEs.
• To introduce and test IP-optimized (CAM-QTP) and MSIE-optimized (rCAM-B3LYP) functionals.
• To compare the performance of these new functionals against the parent CAM-B3LYP.

Main Methods:

• Formal and numerical justification for interpreting Kohn-Sham eigenvalues as IPs.
• Development and application of IP-optimized QTP functionals (reparameterized CAM-B3LYP).
• Testing functionals on He2+ potential curves and evaluating MSIEs.

Main Results:

• IP-optimized QTP functionals exhibit reduced MSIEs.
• MSIE-optimized rCAM-B3LYP functionals show accurate orbital eigenvalues.
• Both CAM-QTP and rCAM-B3LYP demonstrate improved dissociation limits, fundamental gaps, and thermochemical accuracy over CAM-B3LYP.

Conclusions:

• Optimizing functionals for vertical ionization potentials effectively mitigates self-interaction errors.
• The developed CAM-QTP and rCAM-B3LYP functionals offer enhanced accuracy for various chemical properties.
• This work unifies the perspectives of IP prediction and MSIE reduction in DFAs.