[Wien] Virtual Crystal Approximation (VCA): not neighboring atoms in periodic table

[Rocquefelte]

Many other posts have been sent more recently. Search the word virtual 
and not VCA.

The idea is that you replace an atom defined by a nucleus +Z surrounded 
by Z electrons, by an effective atom defined by a nucleus Zeff = Z+x  
surrounded by Z+x electrons. Such an effective treatment can only be 
done for neigboring atoms, which have the same number of core electrons 
for instance and similar radii...
Indeed, remember the Slater approach (effective nucleus). The screening 
is defined by the inner electrons Zeff = Z - Sigma, and the way the 
nucleus is screened defines all the properties of the atom. In your case 
the valence electrons of La and Nb feel a completely different effective 
nucleus (the screening is too much different) and a mixing of these two 
atoms cannot be simply treated by an effective nucleus. In a way VCA is 
a perturbative approach and could give results only if the atoms are not 
too far in the periodic table.

In addition, as already explained by Peter in previous posts it should 
be used only for non-active electrons in the valence band (like 
Sr2+/La3+), but not for active ones (like O/F) ...

In your case, I will use a supercell approach and to insure that you 
didn't create an artificial order you must try different models and 
compare them in terms of total energy, DOS, properties...

Best Regards

Xavier





  valence electrons/core charge


Le 3/9/2013 1:11 AM, Qingyun Mao a ?crit :
 Dear Prof. Blaha and wien users,

 Recently I am trying to do calculations on systems like 
 La(x)Nb(1-x)O(y). I searched in the mailing list and the earliest 
 comments showed up was posted around 2004. I saw comments like: VCA 
 in WINE2k is only possible between neighboring elements, like Na-Mg or 
 Ba-La. But I have not seen any explanation for that yet.

 May I ask the reason why we cannot use VCA for atoms which are not 
 neighbors (like La and Nb) in periodic table? Thanks a lot!

 Regards,
 Qingyun Mao

 School of Applied and Engineering Physics
 Cornell University




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

-- next part --
An HTML attachment was scrubbed...
URL: 
http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20130309/c3f0d16f/attachment.htm

[Wien] supercell warning

[prasenjit roy]

The problem is with the symmetry of the supercell you created. You must
have chosen 2x1x1 or 1x1x2 supercell here. So you need to remember that
after doubling the lattice parameters, the space group and point group
symmetries also changes. It means that while doing sgroup during
initialization the program will find new spacegroup (which is not 225
anymore), and the wyckoff positions of the atoms in your unitcell will be
classified under that spacegroup. That is why the multiplicity changes. Its
of no harm to accept it as  Prof Blaha suggested, or else you can rename
all atoms with different index and can continue with space group P1 ( your
multiplicity all other stuff will remain unchanged in that case).

Prasenjit Roy
Electronic Structures of Materials
Radboud University
Nijmegen
+31 (0) 24 36 52805
-- next part --
An HTML attachment was scrubbed...
URL: 
http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20130309/eab40c0f/attachment.htm