My Community

for transition metals with fully-occupied f-orbitals, i.e. Hf-Au, the provided pesuopotentials (GGA-PW91 or PBE) include the outmost s- and d- electrons only. Are they enough for conventional DFT calculations, or we should use DFT+U?
B)

Thanks in advance for your reply !

the included valence electrons depend on the POTCAR (standard, X_pv,...) you use. Please have a look into the
pseudoppdatabase.pdf file of the VASP-workshop (http://cms.mpi.univie.ac.at/vasp-workshop) or the respective V_RHFIN file if you are not sure about the electronic configuration).
Whether you have to use DFT+U depends on the strength of correlation of the electrons in your specific problem. Only if the
Coulomb splitting of occupied and unoccupied levels is substantially too small, DFT+U has to be applied.