No convergence for dipole-corrected structural relaxation
Posted: Wed May 15, 2013 1:28 pm
Dear VASPers,
I am working on graphene (84 C atoms) with 6 H atoms adsorbed on the surface. I performed a structural relaxation (IBRION=2, ISIF=0) without dipole correction. By using ALGO=Fast and real space projection (LREAL = Auto), the system converged rapidly, as I expected.
Now, I would like to optimize the new geometry (restart from CONTCAR) by switching on the dipole correction (LDIPOL=TRUE). However, when I use IALGO=38 and LREAL=FALSE (for very accurate calculation), the electronic minimization never converges.
My question is: Why it is so? Is it due to the change of electronic minimization algorithm or the real space projection?
Thanks for your help.
L.
I am working on graphene (84 C atoms) with 6 H atoms adsorbed on the surface. I performed a structural relaxation (IBRION=2, ISIF=0) without dipole correction. By using ALGO=Fast and real space projection (LREAL = Auto), the system converged rapidly, as I expected.
Now, I would like to optimize the new geometry (restart from CONTCAR) by switching on the dipole correction (LDIPOL=TRUE). However, when I use IALGO=38 and LREAL=FALSE (for very accurate calculation), the electronic minimization never converges.
My question is: Why it is so? Is it due to the change of electronic minimization algorithm or the real space projection?
Thanks for your help.
L.