Intricacies of DFT+U, Not Only in a Numeric Atom Centered Orbital Framework

Area of Science:

• Computational Materials Science
• Condensed Matter Physics
• Quantum Chemistry

Background:

• Density-functional theory (DFT) often struggles with strongly correlated materials.
• The DFT+U method is a popular extension to address these limitations.
• Accurate implementation of DFT+U is essential for reliable material property predictions.

Purpose of the Study:

• To implement the spherically averaged DFT+U method within the FHI-aims all-electron code.
• To validate the implementation using benchmark systems like NiO and TiO2.
• To investigate the impact of localized numeric atomic orbital (NAO) basis sets and projector functions on DFT+U accuracy.

Main Methods:

• Implementation of spherically averaged DFT+U in FHI-aims using NAOs.
• Validation on bulk NiO to observe band gap opening.
• Study of polaron formation on rutile TiO2(110) surface and bulk.
• Analysis of projector function choice and basis set effects.

Main Results:

• Successful implementation of DFT+U in FHI-aims.
• Observed band gap opening in NiO, consistent with expectations.
• Accurate prediction of polaron formation on TiO2, matching experimental and hybrid functional results.
• Identified potential inaccuracies arising from multiple localized basis functions and provided remedies.
• Demonstrated the critical influence of projector function choice on DFT+U outcomes.

Conclusions:

• The DFT+U implementation in FHI-aims is effective for correlated materials.
• Careful selection of projector functions and basis sets is vital for accurate DFT+U calculations.
• DFT+U is not a black-box method and requires careful application.