[Pw_forum] inequivalent atoms of the same type

[Xunlei Ding]

Dear Yu,
The "same" atom at different symmetry sites is just the same at initial 
time.
When scf calculation is done, they can be totally different.
When we set ntype=2 for the same kind (chemically) of atoms, usually in 
magnetic calculation,
we just set different initial parameters for them, such as 
starting_magnetization(i), to make the initial state close to the final 
result.

For example, we do a calculation on a meterial X (only one kine of atom 
X, with two different symmetry site X1, X2).
If the final spin state is X1: 0.5, X2:0.3, then you can set them as the 
same atom and just set starting_magnetization(1)=0.7,
and also, you can set them as two kind of atoms, and set 
starting_magnetization(1)=0.7, starting_magnetization(2)=0.9.
If the final spin state is X1: 0.5, X2:-0.3, then I think we must set 
them as two kind of atoms, and set starting_magnetization(1)=0.7, 
starting_magnetization(2)=-0.6.

Hope what I said is right.

Best,
Ding

W. YU wrote:

>Dear PW users,
>
>I have a question concerning the "ntype" keyword:
>
>In a non-magnetic case, if I have the same kind
>(chemically) of atoms which occupy different symmetry
>sites, say "4c" and "4e",and I made them belong to the
>same "type", how will the code find the symmetry for
>these two inequivalent kinds of atoms. If it cannot
>differentiate between these two symmetry sites, will
>this has some effect on the results?
>
>Furthermore, in a phonon calculation, the phonon modes
>are associated with different site symmetries. If the
>two atoms in different symmetry sites were treated as
>one kind, will the phonon modes be  different?  
>
>I would appreciate it if anybody could explain this.
>
>Wen Yu 
>
>
> 
>____________________________________________________________________________________
>Be a PS3 game guru.
>Get your game face on with the latest PS3 news and previews at Yahoo! Games.
>http://videogames.yahoo.com/platform?platform=120121
>_______________________________________________
>Pw_forum mailing list
>Pw_forum at pwscf.org
>http://www.democritos.it/mailman/listinfo/pw_forum
>
>  
>