[Wien] Questions regarding xspec and optic

[Peter Blaha]

>                1. For XANES: If my structure file contains atoms with
> multicplicity > 1 (in my case, 3), then after making the supercell
> (1x1x2), if I remove one 1s electron from Fe1, it actually creating 3
> core-holes instead of 1 (inside the supercell). So How should I create
> the case.struct file at the first place?

You have to break symmetry and split a position off, so that you have 
only one atom with a core hole.  PS: 18 atom supercells could be a bit 
small, but one can start with it...

>                2. For obtaining XANES without core-hole, { and also with
> core-hole } do I need to create k-points over the whole BZ ( in chapter
> 8.16 of UG (pp 156), it is written that for X-ray ABS spectra,
> eigenvalue must be calculated over whole BZ )?

For XANES you don't need a full k-mesh. For XMCD I do not know, but 
would be surprised ....

>                3. Is the Energy window in case.inxs depends on EMAX,
> NBANDS in case.in1c?
Of course. Without eigenvalues in this energy region, you cannot 
calculate XANES .....

>                4. For XMCD, I need only K-edge . Even for that should I
> do SO calculation? If yes, as far as I understood, the lineup of steps
> will look like this after creating the unit-cell,:: a) Init_lapw ; b)
> Initso_lapw (or, symmetso -c ?); c) runsp_lapw -so ; d) x kgen -so ; e)
> put FERMI in case.in2c ; f) runsp_lapw -so -c -s lapw1 -e lcore ; g) x
> opticc -so -up ; h) x joint -up ; i) x kram -up .

For XMCD you always need SO.

PS: don't worry about runsp_lapw in 11.1 or 12.1 and "complex". It 
should be ok.
-- 
Peter Blaha
Inst.Materials Chemistry
TU Vienna
Getreidemarkt 9
A-1060 Vienna
Austria
+43-1-5880115671