[Wien] valence energies

[Stefaan Cottenier]

Citeren Michael Gurnett <michael.gurnett at kau.se>:

> I know this topic was discussed previously, but I'm a little
> confused
> 
> I have a system containing Ge, Li and H atoms. In the case of Ge the
> cutoff energy is such that the 3d, 4s and 4p electrons are treated as
> valence electrons. Looking in the scf1 file I find
> 
> ATOMIC SPHERE DEPENDENT PARAMETERS FOR ATOM  Ge1       
>           OVERALL ENERGY PARAMETER IS    0.3000
>           OVERALL BASIS SET ON ATOM IS LAPW
>           E( 2)=   -1.7550   E(BOTTOM)=   -1.950   E(TOP)=   -1.560
>              APW+lo
>           E( 2)=    0.3000
>              LOCAL ORBITAL
>           E( 0)=    0.3000
>              APW+lo
>           E( 1)=    0.3000
>              APW+lo
> 
> if the 3d electrons have an energy of -1.7550, where are the energies
> for the 4s and 4p electrons.

4s and 4p are these ones:

>           E( 0)=    0.3000
>              APW+lo
>           E( 1)=    0.3000
>              APW+lo

As soon as bands are not too localized in energy, wien will use 
default linearization energies of 0.3 Ry (which for most systems falls 
a bit below E_F, in the middle of the occupied part of the valence 
band). For such broad bands, there is not much meaning in giving one 
single energy (look at the s-DOS and p-DOS: what is 'the' energy that 
characterizes the s- and p- 'peaks'...?)

> Also where is the split of the 3d electrons. 

You will see it in the DOS. For the linearization energy, a single 
value is chosen that does not care about the split. From the above 
data, you see only that the band limits of 3d are roughly -1.95 and -
1.56 Ry.

> With the chosen cutoff energy, analysis in w2web is reporting 0 ryd
> for all H and Li atoms, and in the scf1 file for H
> 
>           ATOMIC SPHERE DEPENDENT PARAMETERS FOR ATOM  H 1       
>           OVERALL ENERGY PARAMETER IS    0.3000
>           OVERALL BASIS SET ON ATOM IS LAPW
>           E( 0)=    0.3000
>              APW+lo
> 
> so where do I find the 1S energies

Same explanation as above: it forms a broad band in the valence region.
 
> Li is similar 0 ryd for core and
> 
>           ATOMIC SPHERE DEPENDENT PARAMETERS FOR ATOM  Li1       
>           OVERALL ENERGY PARAMETER IS    0.3000
>           OVERALL BASIS SET ON ATOM IS LAPW
>           E( 0)=    0.3000
>              APW+lo
>           E( 0)=   -3.5225   E(BOTTOM)=   -3.720   E(TOP)=   -3.325
>              LOCAL ORBITAL
> 
> for valence. Could someone explain this. 

The 1s is localized enough to form a sharp peak (make a DOS plot), so 
you see a specific energy. The 2s is a broad band with the default 
value.

If you don't like these default 0.3, you can improve a little on them 
by using the in1new switch (but most often this changes only tiny 
details).

> I want to do core-level
> calculations on the 3d Ge electrons and I'm guessing that treating
> these as core would be incorrect as the energies are so high.

Yes, for a meaningfull standard calculation, you need to include such 
states as valence states. I cannot comment on the core-level part, 
maybe it can be meaningfull still to play a trick and freeze the Ge 3d 
in the core (by the old 'open core' method, see faq ?), I don't know.

Stefaan