[QE-users] Odd number of electrons yields even total magnetization

[Ayestaran Latorre, Carlos]

Dear Farah,

I stumbled upon that work and an older question of yours in the forum similar to mine for that very FG system.

Even if our systems are different, did you notice whether having the right magnetisation had a large effect on the final energies/electronic structure? I'm not concerned with magnetic properties, only energies and structures; I was just puzzled by the fact I was getting unphysical results and inconsistent between similar systems (e.g., if I use a Si dopant instead of B, everything is fine). As you suggest, I guess it's just simple DFT trying to do its best in a strongly correlated system.

Kind regards,
Carlos
________________________________
De: users <users-bounces at lists.quantum-espresso.org> en nombre de FARAH MARSUSI <marsusi at aut.ac.ir>
Enviado: martes, 16 de julio de 2019 15:43
Para: Quantum ESPRESSO users Forum
Asunto: Re: [QE-users] Odd number of electrons yields even total magnetization

​Dear Carlos,

If everything in your input file is OK, when DFT suffers so much from self interaction error, e.g. for a highly correlated system, this may happen. Please see : " Carbon 144, 615 (2019).

Best wishes,

Farah.

On Tue, 07/16/2019 09:34 AM, "Ayestaran Latorre, Carlos" <carlos.ayestaran-latorre17 at imperial.ac.uk> wrote:
Hi,

I have a system consistent of a diamond slab with a boron substitutional defect, an adsorbed hydroxyl and an adsorbed hydrogen on the surface. The system has an odd number of electrons (487) and I run a spin polarized (nspin=2) structural relaxation (see attached). One would expect an odd value for the total magnetization, but the it quickly converges to zero.

I assumed this would be a side effect of smearing allowing partial electronic occupations (I employed gaussian smearing with degauss=0.02). I tried lowering the smearing value down to 0.001, both for gaussian and mv smearing, but only managed to get total magnetization ~ 0.3 Bohr mag/cell in both cases. Fixing tot_magnetization=1, however, yields a noticeably higher final energy (~0.3 eV).

When I run a structural relaxation on a similar system (with an adsorbed H2O molecule instead of the hydroxyl+hydrogen fragments), the total magnetization satisfyingly reaches 1 Bohr mag/cell rather quickly. Other combinations (for example, without the H2O molecule at all) also give total magnetization values that don't match the even/odd number of electrons.

Regards,
Carlos Ayestaran Latorre




--
This email was Anti Virus checked by  Security Gateway.

_______________________________________________
Quantum ESPRESSO is supported by MaX (http://www.max-centre.eu/quantum-espresso)
users mailing list users at lists.quantum-espresso.org<mailto:users at lists.quantum-espresso.org>
https://lists.quantum-espresso.org/mailman/listinfo/users


--
This email was Anti Virus checked by  Security Gateway.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190716/39945aba/attachment.html>