[Octopus-users] Ground-state calculation of magnetic materials

[Micael Oliveira]

Dear Motohiko,

  There is a variable that allows you to set the initial magnetization 
density of your system: GuessMagnetDensity. So, if you want to start 
your calculation with an an antiferromagnetic density, you should set 
GuessMagnetDensity = user_defined and then assign a value for the 
magnetization around each atom using the AtomsMagnetDirection block.

  Regards,

Micael



Motohiko Tanaka wrote:
> Dear Octopus developpers,
> 
>    I am performing the ground-state calculation of 
> anti-ferromagnetic material which involves irons. 
> My platform for this purpose is a Linux-based Opteron 
> system.
>  
>    As I understand that Octopus 3.1 is not yet ready for 
> parallel computation of ground states in a periodic system, 
> I calculate for an isolated crystal with vacuum around it.
>  
>    For the crystal of pure irons which is expected to be 
> ferromagnetic, irons (electron spins) tend to orient in one
> direction, although there is an accuracy issue.
>  
>    However, with an anti-ferromagnetic crystal for which 
> half of irons should have up spins while the rest have 
> down spins depending on their site, calculated irons tend to 
> have up spins (none has down spins). 
> I think that the initial electrons should be assigned
> either up or down spins so that their distribution   
> is not very far from the anti-ferromagnetic state. 
>    Does Octopus have the controlling feature of initial
> up and down spins depending on their site ? How can I 
> control them?  If not at present, which routines and which 
> variables do I need to work on ?  Thank you for your help.  
> 
> Best regards,
> M.Tanaka
>        
> _______________________________________________
> Octopus-users mailing list
> Octopus-users at tddft.org
> http://www.tddft.org/mailman/listinfo/octopus-users
>