[Wien] AFM NiO wrong magnetic moment

[Marcus Ekholm]

Dear Wien2k experts,

I use Wien2k 13.1. I am trying some basic LSDA/GGA (not +U)  calculations
for (111)-layered AFM NiO with the setup as suggested in the exercises
below:

http://www.wien2k.at/events/ws2008/talks/Exercises_08.pdf

However, I find that although it reaches an AFM solution, the local
magnetic moment at fixed volume is much too small compared to previous
LSDA/GGA calculations (e.g. Tran, Blaha, et al, Phys. Rev. B 74:155108). In
addition, relaxing the volume I find it to collapse to a much too small
volume and consequently a very small magnetic moment.

More specifically:
With plain PBE-GGA I get the local spin moment 1.1 bohr magnetons as
compared to 1.4. This is at the GGA (cubic) lattice constant 4.2 Å
corresponding to the hexagonal lattice constants a=5.61220 and c=27.49406
a.u.. With LSDA I get 0.9. I've tried increase the number of k-points,
rkmax etc, but this doesn't seem to be the problem. Relaxing the volume
with LSDA I get a=5.15171 and c=25.23811 a.u. corresponding to a=3.86 Å,
and with the magnetic moment 0.6 bohr magnetons. In the reference above you
got 4.07 Å and 1.2 bohr magnetons.

I use the attached struct file and initalize calculations with

init -b -sp -vxc 5 -numk 10000 -rkmax 9.0 -ecut -9.0

having switched up/down for Ni1 and Ni2 in the inst file, and I make O
nonmagnetic. I start it with runsp_lapw and I get no warnings or
convergence problems down to 1 microRy.

So I believe there must be something wrong with my struct file.
Could you please check if you see anything weird?

Best regards,
Marcus Ekholm
Linköping University
Sweden
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20131120/4e0d7225/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: NiO.struct
Type: application/octet-stream
Size: 1694 bytes
Desc: not available
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20131120/4e0d7225/attachment.dll>