Help with Wannier90 band alignment issue

[alok_shukla1]

Dear Experts,

I am working on a Bi₂Se₃/MnTe bulk heterostructure and am trying to construct Wannier-interpolated bands using VASP + Wannier90. I am interested in the Bi-p and Se-p states. However, I am facing an issue where the Wannier-interpolated bands do not overlap well with the DFT bands, even after several trials. I am attaching wannier90.amn, wannier90.win, wannier90.eig, wannier90.wout, wannier90.mmn files that I am using for the final step and also the obtained wannier90_band.dat as well as my DFT band str. files.

https://drive.google.com/drive/folders/ ... -hero-goto

I am not able to get how can I improve the band overlap or if something is missing from wannier90.win file. Despite several attempts, I have not been able to obtain satisfactory agreement between the Wannier and DFT bands.
Thank you in advance for your help.

[martin.schlipf]

Did you try to do a simpler unit cell first? From the output, I can see that your unit cell has almost 100 Angstrom along the z direction. Ideally you should start from the bulk materials first and then move on to a supercell afterwards. You have many bands in your band structure so it is hard to see what is exactly going on.

For the Wannierization itself, are you sure that the bands in the window agree with the ones that you selected in the projections section of the wannier.win file? Did you look at the projected DOS and identified the main band character of these bands?

Also for future reference, please also include the VASP input files so that we can assess whether you made a mistake in the setup of the VASP calculation before you even reach the Wannier stage.

[alok_shukla1]

Dear Expert,
Thank you so much for your response.
I first performed Wannierization for the bulk Bi2Se3 primitive cell (without MnTe or supercell) and obtained a very good agreement between Wannier and DFT bands, so the Wannier procedure and projections for pure Bi2S3 appear to be working fine.
For the heterostructure, I have checked both the projected bands and PDOS and I tried several different outer/frozen windows (slightly increasing and decreasing the energy ranges), the result changes a bit but the overall mismatch between DFT and Wannier bands for the heterostructure still persists.
I am adding my system's SCF INCAR, KPOINTS, POTCAR and POSCAR files. I am also adding PDOS plotted results as well. Please let me know if I need to add some more files. I will be very thankfull for futher suggestions. Many thanks again.

https://drive.google.com/drive/folders/ ... -hero-goto

[martin.schlipf]

Looking at the PDOS, I can see that there is a significant contribution of Mn:d and Te:p near the Fermi energy (I assume that is what you indicate with the dashed line). In your Wannierization you only include Bi:p and Se:p which is probably not good enough.

If you already succeeded in Wannierization of bulk Bi2Se3, the next step would be a Wannierization of bulk MnTe. Afterwards I would create a small interface of say one layer each and try to wannierize that. Once that is working, I would try to increase the number of layers until you reach the size you are interested in.