wannier90 and spin projections

Question on input files/tags, interpreting output, etc.

Please check whether the answer to your question is given in the VASP online manual or has been discussed in this forum previously!

Moderators: Global Moderator, Moderator

Post Reply
Message
Author
michaela._fechner
Newbie
Newbie
Posts: 6
Joined: Tue Dec 15, 2020 7:41 am

wannier90 and spin projections

#1 Post by michaela._fechner » Fri Feb 05, 2021 1:23 pm

Dear all,

is there a way to use the spin projections of wannier90 in combination with vasp?
Currently, I just try to project out the spin up bands of MnO, but vasp seems
to ignore the (u) and (d) flag and crashes with the message that spinor=.false.
although I set it to true. Any idea what to fix ?

my win file reads

num_wann = 10 # number of target Wannier functions

spinors=.true.
exclude_bands=1-8
Begin Projections
c=0.00,0.00,0.00:l=2(u)
c=0.50,0.50,0.50:l=2(d)
End Projections

henrique_miranda
Global Moderator
Global Moderator
Posts: 414
Joined: Mon Nov 04, 2019 12:41 pm
Contact:

Re: wannier90 and spin projections

#2 Post by henrique_miranda » Fri Feb 05, 2021 3:32 pm

The wannier90 interface was recently updated in VASP 6.2.0 (the latest release) to correctly handle the spinor case.
Give it a try and let us know in case you find any problems :)

The spinor case was not implemented in previous versions of VASP. See this thread for some more details.
https://www.vasp.at/forum/viewtopic.php?f=3&t=17841

michaela._fechner
Newbie
Newbie
Posts: 6
Joined: Tue Dec 15, 2020 7:41 am

Re: wannier90 and spin projections

#3 Post by michaela._fechner » Fri Feb 05, 2021 3:43 pm

I realized this also today thanks for the swift reply,

just to be sure as I checked the code

the precompiled flag is now just

-DVASP2WANNIER90

so v2 has been dropped ?

best wishes
Michael

henrique_miranda
Global Moderator
Global Moderator
Posts: 414
Joined: Mon Nov 04, 2019 12:41 pm
Contact:

Re: wannier90 and spin projections

#4 Post by henrique_miranda » Fri Feb 05, 2021 4:06 pm

Yes, exactly.
With this change the support for Wannier90 1.0 is dropped so -DVASP2WANNIER90 means you can link with Wannier90 2.x or 3.x.

michaela._fechner
Newbie
Newbie
Posts: 6
Joined: Tue Dec 15, 2020 7:41 am

Re: wannier90 and spin projections

#5 Post by michaela._fechner » Mon Feb 08, 2021 11:28 am

Dear henrique,

sorry to bother again just to ask how to operate W90 with the new version, as I realized there are some changes:

1.) the library mode is the new default as I noticed

so for my test I use MnO as a testcase (pseudos are Mn_pv and O) and the following INCAR see below.
However, still vasp quits with

Calling wannier_setup of wannier90 in library mode
param_get_projections: spin is defined but spinors=.false.

any suggestions ?

Also what is the SCMDC method to define projections ?

INCAR:

Code: Select all

SYSTEM = MnO
ISTART  =  0
ICHARG  =  2
ISMEAR  =  0
SIGMA   =  0.02
PREC    =  A
EDIFF   =  1E-8
ENCUT   =  600
NELM    =  200
ISYM    =  2
ISPIN   =  2
LORBIT  =  10
LMAXMIX =  4
ADDGRID = .TRUE.

#Antiferromagnetic configuration
MAGMOM =  3 -3 0 0

NEDOS = 5000
EMIN = -10
EMAX =  15

#to get an insulating state
LDAU    =  .True.
LDAUTYPE=  1
LDAUL   =    2  -1
LDAUU   =  2.0 0.0
LDAUJ   =  0.0 0.0


LWANNIER90 =.TRUE. # generate wannier90 inputs (projections and overlap
LWRITE_UNK = .TRUE.
LWRITE_MMN_AMN  = .TRUE.

NUM_WANN=10

WANNIER90_WIN = "
spinors=.true.
exclude_bands=1-8
Begin Projections
Mn:d(u)
End Projections
"
POSCAR:

Code: Select all

MnO
   1.00000000000000
  4.2912822905   2.0698391900   2.0698391900
  2.0698391900   4.2912822905   2.0698391900
  2.0698391900   2.0698391900   4.2912822905
   Mn  O
    2  2
Direct
  0.0000000000000000  0.0000000000000000  0.0000000000000000
  0.5000000000000000  0.5000000000000000  0.5000000000000000
  0.2500000000000000  0.2500000000000000  0.2500000000000000
  0.7500000000000000  0.7500000000000000  0.7500000000000000

  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00

henrique_miranda
Global Moderator
Global Moderator
Posts: 414
Joined: Mon Nov 04, 2019 12:41 pm
Contact:

Re: wannier90 and spin projections

#6 Post by henrique_miranda » Mon Feb 08, 2021 5:35 pm

Don't worry about it :)
I am working on updating the wiki page regarding the new changes in the wannier90 interface with vasp 6.2.
I will update you here once that is finished.

In the meantime and to answer your question:
There are three different modes for VASP:
1. spin unpolarized calculation (ISPIN=1)
2. collinear spin calculation (ISPIN=2)
3. non-collinear spin calculation (ISPIN=1 and LNONCOLLINEAR=.TRUE. or LSORBIT=.TRUE.)

Wannier90 has two modes:
1. spin unpolarized calculation
2. same as case 1. for spin up and another for spin down
3. non-collinear spin calculation (spinors=.true.)

If you want to get a non-collinear calculation you should either specify LSORBIT=.TRUE. or LNONCOLLINEAR=.TRUE. in VASP.
The input file for wannier90 will be created in a consistent way i.e. with spinors=.true.

Post Reply