[QE-users] NEB : path length is increasing

[Tone Kokalj]

On Mon, 2020-10-26 at 14:38 +0000, Omer Mutasim wrote:
> Dear Dr. Michal 
> I have relaxed both initial and final structure with forc_conv_thr = 
> 1.0e-03  , 
> but after running NEB , it increased to the above-mentioned values
> (0.02 & 0.03).

Which is OK, because the unit of force in pw.x is Ry/Bohr, while in
neb.x it is eV/Angstrom; note that 0.001 Ry/Bohr = 0.026 eV/Angs.

This implies that your 1st and last images with the errors of about
0.02 eV/Angs are converged within the defaults of pw.x.

Best regards, 
Tone Kokalj
-- 
Jožef Stefan Institute, Ljubljana, Slovenia