EPW Forum

Dear users,

In the diamond example, the chosen dis_froz_max and dis_win_max are 7 and 15 eV, respectively.

Why do we choose the values? Why don't we choose 15eV for dis_froz_max since we are interested in the four valence band? If my nbnd=6, should I use dis_win_max=20?

What about n-type diamond? Should I choose 8 nbnd, and use dis_froz_max=19 and dis_win_max=30?

The band energy is shown below:
-7.95675 13.89078 13.89078 13.89078 19.55356 19.55356

Thank you!

Tianshi Wang
University of Delaware, DE, USA

Dear Tianshi,

You can plot the Wannier function and overlay them with the nscf bandstructure. The band should be quite close.
This is to show the disentanglement procedure. If you choose dis_froz_max 14 eV and have only 4 bands, then you Wannierized everything (so no disentanglement).

You could indeed impose dis_froz_max 14 eV for valence band only and dis_froz_max = 20 eV for conduction bands as well.

But in any case you have to make sure that the Wannier interpolation reproduces very closely the correct electronic and phonon bandstructure.

Best,

Samuel

Hi Samuel,

Thank you for your reply! It is very helpful to me.

I chose (0 for dis_froz_max, 10 for dis_win_max) for p-type and (10, 16) for n-type based on my understanding. The interpolated band structure doesn't have notable change. Do you think it is fine?

This material is 3C-SiC. The band gap is around 10 eV. For n-type, only care the lowest conduction band.


Best regards,

Tianshi

Dear Tianshi,

This look good for the wannierization of the valence band indeed.

Maybe you can also make a zoom close to the VBM to be sure that the curvature (effective mass) is correct.

Best,

Samuel