[Wien] SPLIT parameter for ELNES calculation

[Shu Miao]

Dear users,
For the whitelines of transition metals, L2 leaves a hole in 2p_1/2 and L3
leaves a hole in 2p_3/2 in the final state. Since we ignore the
possibility that two electrons are excited simulately within same atom,
one atom can contribute to either L2 or L3, but not both. So is the SPLIT
parameter in case.innes redundant?
To simulate the whitelines, should I always set SPLIT to 0 and prepare L2
and L3 separately, then overlay them manually? In this case, the splitting
can be taken from the core eigenvalues in case.scf as Dr Peter Blaha
mentioned in the "EELS of  L3 Ti edge" a few days ago. Does anyone have
experience on the accuracy of this method? Can these eigenvalues
reflect the energy change as small as chemical shift, eg, a few eV, in the
material?


Thanks

Shu Miao