[Pw_forum] magnetic moments in graphene with single vacancy :: discrepancy

[Martin Gmitra]

Dear all,

I did some calculations for induced magnetism in graphene due to single
vacancy (sometimes reffered as V1(5-9))
using USPP (C.pbe-van_ak.UPF<http://www.quantum-espresso.org/wp-content/uploads/upf_files/C.pbe-van_ak.UPF>;
ecutwfc=30, ecutrho=160) and NCPP
(C.pbe-mt_fhi.UPF<http://www.quantum-espresso.org/wp-content/uploads/upf_files/C.pbe-mt_fhi.UPF>,
ecutwfc=35, ecutrho=140).
Calculations of relaxed structure give magnetic moment in the supercell of
about 0.65 \mu_B in both the PP cases.
However, I did calculations of the same structure using full-potential
LAPWcode FLEUR where the magnetic moment
(spin_up - spin_down) drops to value of 0.12256. There are also studies
where similar structures have been studied
with SIESTA code:
depending of supercell size from 1.1 to 1.5 \mu_B [Yazyev et al, Phys Rev B
75, 125408]
Normconserving PPs give about 1.2 \mu_B [http://arxiv.org/abs/1006.0589]
GIPAW gives 1.53 \mu_B [http://www.researchgate
.net/publication/235566299_Non-collinear_magnetic_moments_in_graphene
_with_vacancies]

It seems that USPP and NCPP could not be the reason for the discrepancy, or?
Could you comment please on the origin of the discrepancies between QE and
LAPW?

Best regards,

Martin Gmitra
Uni Regensburg, Germany
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20130509/dd83a913/attachment.html>