hanghui chen wrote:
>       However if I run a spin-polarized calculation (LSDA), I find 
> that a system could be paramagnetic at "high temperature" and 
> ferromagnetic at 'low temperature'. 

That's plausible, since the magnetization is more hard to converge wrt to 
smearing than 
the total energy. The former is much more sensitive to the details of the band 
structure 
around the Fermi energy (and hence on the value used for the BROADENING 
parameter) than 
the latter.

If you're sure that the ground state of your relaxed system will be ferro, why 
don't you 
try to constrain the total magnetization (search for constrained_magnetization 
in 
Doc/INPUT_PW.txt) and then relax the constrain at the end (though not sure that 
this last 
step can be done actually...).

Another, more simple, solution would be to start from_scratch the "small 
temperature" run 
using the atomic positions from the previous run at "high temperature".

GS

>  
> Hanghui Chen
> Department of Physics
> Yale University
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

-- 


o ------------------------------------------------ o
| Gabriele Sclauzero, PhD Student                  |
| c/o:   SISSA & CNR-INFM Democritos,              |
|        via Beirut 2-4, 34014 Trieste (Italy)     |
| email: sclauzer at sissa.it                         |
| phone: +39 040 3787 511                          |
| skype: gurlonotturno                             |
o ------------------------------------------------ o

Reply via email to