[Pw_forum] fixed magnetization to check stoner criterion

Arles V. Gil Rebaza arvifis at gmail.com
Thu Jul 24 00:00:20 CEST 2014


Hi Rene Wirnata, I think that you must to obtaind the total energy (Et) vs
total magnetic moment (MT) and fot with a curve of 4to order (Ax^4 +
Bx^2+C). The coeficient of the  cuadratic term (B) is related with the
stones parameter. So, to get this curve you must to do scf calculations for
each MT value using fixed spin calculations (use "total_magnetization" in
QE).
Maybe this paper could help you:
http://journals.aps.org/prb/abstract/10.1103/PhysRevB.74.014416

In your input file, if you are doing SCF calculation, why you use
 &cell
    press_conv_thr = 0.1
these tag is for vc-relax calculations

 why ...
&electrons
    mixing_mode='TF'
try with the default setting "plain"

nspin       = 4  .... are you performing non-collinear calculations.??

I hope I've helped

Best.

Dr. Arles V. Gil Rebaza
Instituto de Física La Plata
Argentina.





2014-07-23 11:44 GMT-03:00 Rene Wirnata <renewirnata at gmx.de>:

>  Hi @ all!
>
> Unfortunately I couldn't find anything on this topic with google, so I
> have to ask here:
>
> Currently, I try to test the stoner criterion for the Al5Fe2 system and
> therefore I set
> *"constrained_magnetization = 'total', fixed_magnetization(3) = i, lambda
> = 1.d0"*
> where i = {0.0 .. 2.2}. This setup will first of all tell me in the output
> file that I have to
> set *nspin = 4*, setting it to sth. else won't work, I already tested.
> Now the problem is
> none of these calculations will converge doing max. 100 iterations. Even
> if I set
> *lambda =** 0.1*, there's no big change as you can see in the output
> files.
>
> Further, the documentation tells me:
> a) for LDA potentials use *tot_magnetization*, --> mine is GGA, so stay
> with *contrained_mag**...*
> b) if I used *tot_magnetization* then I shouldn't use
> *starting_magnetization*, so maybe this
> could help? But to what number should I set it, because for every atom in
> my cell it will
> become different during the iterations.
>
> Unfortunately, no one in my institut ever did such calculations before and
> I am
> myself still a total beginner with pwscf, so pls excuse if the question is
> stupid :).
>
> Thanks a lot in advance!
>
> Best regards from Germany,
>
> Rene Wirnata
>
>
> --
> Attachements see here:
> https://www.dropbox.com/sh/zxr7phpxdngo0zl/AABWJEqMog7eawxvTec9FH7pa
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>



-- 
###--------->   Arles V.   <---------###
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20140723/6f59ea5e/attachment.html>


More information about the users mailing list