[Pw_forum] constrained_magnetization

[chu-chun fu]

Dear Lorenzo,

Thanks again. I am aware of the calculation procedure you described in 
your previous email, i.e. to asign as different species that are not 
equivalent from electronic point of view bu equivalent as chemical 
species. However, concrning atomic magnetization constraint, it requires 
to asign one fixed value to each species, and what I asked was about the 
possibility to constrain certain species (atoms) while leaving other 
species (atoms) free from magnetic constraint. I just wonder I may 
probably do it by modify the subroutine add_bfield ? But just want to be 
sure.

Will appreciate if possible that you let me know which subroutines would 
be modified to do such a constraint ?

Best Regards

Chu Chun



Lorenzo Paulatto wrote:
> In data 03 giugno 2009 alle ore 11:35:06, chu-chun fu 
> <chuchun.fu at cea.fr> ha scritto:
>
>> Thank you for your response. Maybe I was not clear enough in my previous
>> email. My question is not to give occupation Z to atom X, but local M
>> (magnetization) to atom X. My question actually  concerns the 'atomic'
>> constrained magnetization option already implemented in PWscf (in
>> particular version 4.0.3) by adding a 'penalty functional' as can be
>> seen in the SUBROUTINE add_bfield,
>
> Well, yes I mistaken your question a bit, yet that's what constrained 
> DFT is about.
>
>> Just wonder if we can only constrain the local magnetization of only one
>> species instead of all the atoms to respective values as fixed in the 
>> input.
>
> You cannot, but you can assign a different (formally) specie to each 
> atom doing something like this:
>
> ATOMIC_SPECIES
>  Pd_1  106.40    Pd.pbe-paw_kj.UPF
>  Pd_2  106.40    Pd.pbe-paw_kj.UPF
>  Pd_3  106.40    Pd.pbe-paw_kj.UPF
> ATOMIC_POSITIONS
> Pd_1      0.000000         0.000000       0.000000
> Pd_2      0.250000        -0.144340       0.408250
> Pd_3      0.500000        -0.288680       0.816490
>
> the species will be equivalent in practice, but distinct from the code 
> point of view, hence you can assign different constraints to them. It 
> has to be done this way because of how the symmetry analysis works: 
> atoms of the same specie can in principle occupy equivalent positions, 
> in which case it would make no sense to assign different properties to 
> each of them.
>
> best regards
>
>
> --Lorenzo Paulatto
> SISSA  &  DEMOCRITOS (Trieste)
> phone: +39 040 3787 511
> skype: paulatz
> www:   http://people.sissa.it/~paulatto/
>
>     *** save italian brains ***
>  http://saveitalianbrains.wordpress.com/


-- 
Chu Chun FU			 	
DEN/DMN/SRMP CEA-Saclay				
91191 Gif-sur-yvette Cedex	
France 				
Tel: +33 1 69 08 29 32
Fax: +33 1 69 08 68 67
E-mail: chuchun.fu at cea.fr