[Wien] txspec

[Joo Yull Rhee]

Thanks, Dr. P. Blaha. But I still have questions.

I am talking about the SO splitting of core level, not about the SO splitting of valence states. Pt 2p-level, for example, splits into 2p_1/2 and 2p_3/2 due to the SO coupling and the transitions from 2p_1/2 level to empty states above the Fermi level corresponds to the L2-edge spectrum and the transitions from 2p_3/2 level to the L3-edge spectrum. Therefore, for Pt the L2- and L3-edges are split by about 1730 eV.

 I do not think there is any 'relativistic selection rule.' Only the dipole selection rule for optical transition is used in both x-ray spectra and optical conductivity calculations. For this case the SO coupling has nothing to do with the optical transitions. The SO coupling plays, however, an important role when magneto-optical properties (off-diagonal components of the optical conductivity tensor) are calculated. Since for the Pt 2p level the angular momentum quantum number is already fixed (l=1), there are only two possible transitions: to l=2 (d-states) and to l=0 (s-states). I think that the program 'txspec' can split these two transitions, but not for the L2- and L3-edges. And I do not think 'SPLIT' can differentiate the L2- and L3-edges, either. I think the parameter 'SPLIT' can do distinguish the l -> l+1 tanstion and l -> l-1 transition only. So, if the program 'txspec' uses only 'l' qunatum number without differentiating the L2- and L3-edges, the resultant spectrum is the sum of L2- and L3-edge spectra. Am I correct?

Joo Yull Rhee

-----Original Message-----
From: wien-admin at zeus.theochem.tuwien.ac.at [mailto:wien-admin at zeus.theochem.tuwien.ac.at]On Behalf Of Peter Blaha
Sent: Saturday, August 16, 2003 4:46 PM
To: wien at zeus.theochem.tuwien.ac.at
Subject: Re: [Wien] txspec


> I am trying to calculate the Pt L2-edge x-ray absorption spectrum. I have several questions.
>
> 1. In UG, if I set appropriately nonzero 'SPLIT', then I can get a spectrum in which the L2- and L3-edge are split. However, I tried SPLIT=30, I got almost the identical L2- and L3-edge spectra. What is wrong? It shows some difference, but the difference is not due to the L2- and L3-edge. The difference is due to two transitions, l->l+1 and l->l-1.
>
> 2. In SRC_txspec directory, I checked the 'txspec.f' file. In the 'txspec.f' file on the line 95, it is said that 'just take the negative kappa.' I think the parameter 'kappa' is a quantum number that is determined by the total angular momentum (j=l+s). In order to calculate the L2-edge, I set n=2, l=1. Then the 'txspec.f' file autimatically select kappa=-3, which corresponds to j=3/2, i.e., L3-edge. Am I correct?
>
> 3. Since I think that the negative kappa corresponds to the L3-edge, I tried to change the line 96 of the 'txspec.f' file from 'kappa=-1-LC' to 'kappa=LC' to calculates the L2-edge, however, it does make almost no difference. Then I checked which program would use the paramete 'kappa.' It is used only in the 'hfsd.f' file. In the 'hfsd.f' file it is used to check only it the choices of NATT and NC are correct. I mean the parameter kappa is practically never used. Why?

XSPEC does NOT use any relativistic selection rules,.....

SPLIT is most useful for "lighter" elements, where the core state is
splitted by a few eV only and the L2 and L3 spectra "overlap".
SPLIT does not do anything else than taking the non-relativistic spectrum,
and overlaps two identical spectra (with specified intensities) shifted by
SPLIT.

You can add SO also for the valence states, work out the selection rules
yourself and choose the appropriate partial DOS (maybe using the QTL
program) for "fully relativistic spectra.

Regards
                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------

_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien