Error while running TB2J with VASP wit SOC

[Adam]

Hello,

Thank you very much for this software. It is so much helpful to the magnetic community. I have a problem running the program after wannerization  in other to obtain the DMI components. My wannier band structures almost matches with that of the nscf band. this is the error message:

wann2J.py --elements Fe --efermi 0  --spinor --prefix_spinor='wannier90' --kmesh 7 7 7  --nz 150 --groupby orbital

TB2J version 0.7.6.2
Copyright (C) 2018-2020  TB2J group.
This software is distributed with the 2-Clause BSD License, without any warranty. For more details, see the LICENSE file delivered with this software.


Reading atomic structure from file ./POSCAR.
Reading Wannier90 hamiltonian: non-colinear spin.
Starting to calculate exchange.
Traceback (most recent call last):
  File "/home/adam/anaconda3/bin/wann2J.py", line 186, in <module>
    run_wann2J()
  File "/home/adam/anaconda3/bin/wann2J.py", line 156, in run_wann2J
    gen_exchange(path=args.path,
  File "/home/adam/anaconda3/lib/python3.11/site-packages/TB2J/manager.py", line 212, in gen_exchange
    exchange = ExchangeNCL(tbmodels=tbmodel,
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/adam/anaconda3/lib/python3.11/site-packages/TB2J/exchange.py", line 77, in __init__
    self._prepare_orb_dict()
  File "/home/adam/anaconda3/lib/python3.11/site-packages/TB2J/exchange.py", line 206, in _prepare_orb_dict
    raise ValueError(
ValueError: The number of spin-orbitals for atom 6 is not even,
7 spin-orbitals are found near this atom.
which means the spin up/down does not have same number of orbitals.
This is often because the Wannier functions are wrongly defined,
or badly localized. Please check the Wannier centers in the Wannier90 output file.

[Xu He]

Hi, 

As the error message suggested, can you check the Wannier centers are close to the atoms? (also check the Wannier spread in the .wout file).  The matching of the band structure is not enough to tell a good Wannierization. 

And check if the " --groupby orbital" is indeed the right option for the DFT code you're using. 

 If you can send the .wout file I can have a look. 

Best wishes, 

HeXu