Dear Quantum espresso community, It's well known that adapting the Hubbard model in DFT give a better insight for the electronic structure on correlated materials, here the systems with d and f shell localisation are corrected by applying a U potential, but I don't understand why do we apply U also for oxygen, unfortunately I can't find any reference (may be my research wasn't so deep) giving more explanation.
Finally I would like to discuss the effect of hybridisation on U, especially with mixed compounds where the oxygen electrons are shared on between both Hubbard and non Hubbard atoms, from my experience on double pervoskite the calculated value of U on Hubbard atoms including oxygen using the linear response approach by cococcioni et al overestimate the band gap . >From my point of view I think that the value of U must be proportional to the number of electron only shared between Hubbard atoms and the oxygen. Thanks in advance. Youssef Aharbil Faculty of sciencs ben msik. Morocco
_______________________________________________ Pw_forum mailing list [email protected] http://pwscf.org/mailman/listinfo/pw_forum
