Evaluation of DFT plus U and HSE Frameworks for Strongly Correlated Iron Oxide

For strongly correlated metal oxides, several density functional theories have been used to describe the electronic structure of bulk iron oxide (alpha-Fe2O3) and corresponding surfaces, including GGA+U and hybrid functionals (HSE), but the accuracy of these methods is unclear for these metal oxides. In the present study, first-principles simulations were carried out to examine the effects of a hybrid density functional on alpha-Fe2O3 bulk and surface slabs. Further, by using GGA+U, van der Waals corrections and O-2 over binding corrections were also examined. It has been found that HSE functionals with 17.5 % Fock exchange predict better properties than those with 12 % or 25 % exchange. Methodological studies indicate that GGA+U predicted zero-bandgap surface slabs become semiconducting slabs at HSE, and that calculated bandgaps are dependent on the actual exchange addition percentage. Our studies also show that the surface energy and relative stability of different Fe2O3 terminations are less sensitive to the inclusion of dispersion terms. However, when accounting for the GGA-error in O-2 over-binding, significant changes occur in the computed surface energies. Additionally, HSE06 functional was tested on MnO2(0001) and MnO2(110) surfaces. Both surfaces were identified as metallic by a PBE+U calculation, and an HSE06 analysis revealed a nonzero band gap.

Automated summary

Several density functional theories have been used to describe the electronic structure of strongly correlated metal oxides, but their accuracy for alpha-Fe2O3 is unclear. First-principles simulations were carried out in this study to investigate the effects of a hybrid density functional on alpha-Fe2O3. The results showed that HSE functionals with 17.5% Fock exchange predict better properties and that GGA+U predicted zero-bandgap surface slabs become semiconducting when using HSE.