Dear Mr. Nishihara,
I don't know if this is the right place to ask this , But I didn't get the answers at the Pw_forum. So, I am posting this.
I have gone through the following steps . Please suggest me that if they are okay
1. First I have computed the ecutwfc convergence test.
2. Then I have performed a ionic relaxation calculation at the converged ecut values of my supercells ( only for a few k-points . i.e 1 1 3 )
3. Then I performed a scf calculation for a denser k-point with the relaxed structure ( 1 1 12)
4 Performed band-calculation.
Now my confusion is this Does relax calculation has anything to do with how much k-point I take in the relax calculation ? Because much k-point means a very long relax calculation
And One more thing I have doped In atom in my GaN structure . I intend to increase the number of In atom by one ( 1, 2, 3, 4) in the subsequent processes. I think k-point convergence depends on the bonding properties of the cell. Do I need to performance k-point convergence test every time I put an extra In atom ?
Thank you Mr. Nishihara.
> Relax calculation
I think your 4 steps of calculation is correct.
Small #k-point for relax is not bad, to reduce computational cost.
> In doped GaN
Once you determine #k-point, you do not change #k-point for all doped models.
Because it is allowed to compare results among models.