[Wien] about SOC band structure

[Peter Blaha]

>    I have a question confusing me for a long time. For WIEN, it is possible to distinct the spin up and spin down channels for DOS plot in SOC calculation, while this is impossible for band(only one sort of line would be performed). I understand that it is the mixing effect for SOC. Then, what is the meaning for the two spins in DOS?

For any "partial"-DOS (and spin-up/dn are a "partial-DOS") you get the
contribution to this DOS from a projection of the corresponding
character of each state.

If a single eigenvalue has a mixture of s + p character, we calculate
its corresponding  weight to the s (or p) partial DOS from its s (p)
partial charge, i.e. the amount of its "s (p)-character".
Of course the same thing can be done for spin-up/dn in a SO calculation,
where a single eigenvalue may contain a certain fraction spin-up and dn,
respectively.

On the other hand you cannot make a "s-bandstructure" and a
"p-bandstructure" (only with fat-bands you can indicate its character),
and thus with SO you cannot plot spin-up/dn bandstructures (since they
are the same), but again, you can use "fat-bands" to indicate the
main-character of a band.

> 
> So I wonder how WIEN deals with this matter. If someone is familiar with VASP, it is known that neither DOS nor band plot can display two spins in SOC calculation.
> 
> Furthermore, in WIEN, a given state could be heavier plotted in band structure. So the other question is does it make any sense if one spin state (such as spin up) is specified for the d orbitals of an ion in SOC-included band?