To know the utility of the method " mbj+U" read the reply of Dr Peter Blaha

Re: [Wien] Regarding mBJ plus U

Peter Blaha Tue, 08 Mar 2016 21:29:03 -0800

 

 

mBJ works very well for bandgaps of s,p and even 3d electrons. HOwever, it is too weak to handle the atomic-like 4f electrons properly. 

xxx+U can handle the 4f electrons, but still very often the gap is too small (because it is then determined by the "other electrons" in 4f compounds), or one would need an unphysically large value of U. A combination of mBJ+U with a rather moderate U value looks quite good in many cases .... 

Reference: https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg13955.html