[Pw_forum] spin splitting

vtmtrinh at caltech.edu vtmtrinh at caltech.edu
Mon Aug 17 21:49:56 CEST 2009


Hi Gabriele,

I am sorry. Before, there are usually the words "spin up" and "spin down"
in the spin polarization cases.  I did not see it this time, so I thought
something was wrong here.  When I looked at the band data again, I saw the
number of eigen values are double there, which means one for each spin
state.  So I saw it now.  I am sorry for my bad.

Thanks,

Trinh

> Dear user,
>
> vtmtrinh at caltech.edu wrote:
>> Dear PWSCF Users,
>>
>> Does any of you know how to calculate spin splitting using QE?  More
>> clearly, if I have 2 structures (Si 8-atom cell), one at equilibrium and
>> the other one with atoms being displced according to phonon mode.  I
>> would
>> like calculate the energies of the spin up and down.  There are 2
>> concerns
>
> You did not clarify much, in my poor understanding. What do you want to
> calculate? The
> splittings between the spin up/spin down eigenvalues?
> That should be easy, you do a calculation with nspin=2 and a guess for the
> starting
> magnetization and, if your system stays magnetic, you can easily compute
> the splittings
> from the list of eigenvalues given on output at the end of the
> calculation.
>
> Or maybe you want to compute total energy difference between an AF and a
> FM configuration?
>
> Why do you mention two configurations that differ by atomic displacements
> and ask about
> computing spin up/spin down splittings? Does magnetism arise in the second
> case due to the
> displacement from equilibrium configuration?
>
>
>> that I have: (i) The difference in eneries for these 2 states is too
>> small.  Can we get to that level of accuracy by using QE.
>
> It depends what is the accuracy you need. In principle you can increase it
> by increasing
> PW/charge-density cutoffs, number of k-points, and decreasing degauss (you
> may also need
> to decrease conv_thr or other things that did not come in my mind in this
> moment...). Then
> you probably  have to face with the error due to the use of
> pseudo-potentials, but
> generally the error in energy differences can be believed to be lower than
> absolute error
> on the total energies.
>
> (ii) how to do
>> it (how to fix spin up and down).
>
> I don't understand exactly what you want to fix. You can for instance fix
> the total spin
> magnetization of your system using the tot_magnetization or multiplicity
> variables.
>
> HTH
>
> GS
>
>
>>
>> Any suggestion would be appreciated.
>>
>> Thank you,
>>
>> MyTrinh Vo
>>
>> Computation group
>> JPL/CalTech
>>
>> _______________________________________________
>> 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
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>




More information about the Pw_forum mailing list