Dear all, I am currently running calculations to determine the electronic structure of a 3d transition metal - containing phosphate and am facing an issue with evaluating the correct value for the Hubbard U parameter. I am trying to determine the value through a benchmarking method, attempting to reproduce the electronic bandgap reported in literature for bulk NiO (usually between 3.6 to 4.0 eV is reported). The .xyz file was set up in a manner that we handle a spin-polarized system with an equal number of UP and DOWN spin states for (A Antiferromagnet), within which there are ferromagnetic layers along the {111} direction. Firstly the PLUS_U_METHOD section was set to MULLIKEN, which resulted in negative DFT+U energies for any value of U above U=2.0 eV (-0.38 eV for U=3.2 eV for instance) , but the overal DFT+U occupancies of the d-orbitals were quite reasonable. Nevertheless, I switched to LOWDIN method in an attempt to obtain a positive DFT+U energy, which then in turn causes a lowering of the bandgap value (2.9 eV HOMO-LUMO gap for U=3.4 eV using LOWDIN, compared to e.g 3.3 eV HOMO-LUMO Gap for U=3.2 eV using MULLIKEN). My question would then be, should I stick to the LOWDIN method and ramp up to even higher U values until I reach a bandgap comparable to literature, or can the negative DFT+U energy contribution that comes out from the MULLIKEN method be ignored as long as the orbital occupancies are of a reasonable value? Thank you in advance for you assistance, Kind Regards, Lovro Šarić -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to <a href="mailto:cp2k+unsubscribe@googlegroups.com">cp2k+unsubscribe@googlegroups.com</a>. To view this discussion visit <a href="https://groups.google.com/d/msgid/cp2k/99d5fc8d-dd9a-4505-a5a8-c02af8e9109en%40googlegroups.com?utm_medium=email&utm_source=footer">https://groups.google.com/d/msgid/cp2k/99d5fc8d-dd9a-4505-a5a8-c02af8e9109en%40googlegroups.com</a>.