Hello everyone ! I want to do a structure relaxation for a 2d Janus crystal In2SSe , since I am going to calculate phonon bands afterwards , I need to do a very accurate relaxation , my INCAR is as follows :
SYSTEM = ln2SSe
ISPIN = 2 # spin-polarized
ISMEAR = -5 #semidonductor
SIGMA = 0.1
ENCUT = 400
EDIFF = 1E-8
IBRION = 2
POTIM = 0.5
ISIF = 4 #2d material
NSW = 1000
EDIFFG = -1E-8
LVDW = .TRUE #turn on van der Waals interaction
LREAL = .FALSE
LWAVE = .FALSE
LCHARG = .FALSE
During my first relaxation , an error occurs as follows:
ZBRENT: fatal error in bracketing,please rerun with smaller EDIFF, or copy CONTCAR to POSCAR and continue
I think maybe i encounter a local minima,so I set IBRION = 1 and lower POTIM and continue.
BUT after 1000 ionic steps (I have set NSW = 1000) , it still doesn't echo ''reached required accuracy' and the total energy E0 changes little during almost all the steps.
What's the problem ? Really apreciate your comments !!
Dear Kong,
I have several observation.
1) If your system is bulk in nature then ISIF = 3 is okay. If not, then go for ISIF = 2. It will do ionic relaxation keeping the cell parameter fixed.
2) EDIFF = 10-6 is more than accurate.
3) Whenever you are doing ISMEAR = -5 , go for Sigma = 0.5, it usually stabilize the relaxation process.
4) Moreover, my suggestion is for relaxation purpose you can safely use , ISMEAR = -1 , it saves time.
5) Use NELM = 60, NSW = 500 tag for better convergence ( It's experience only).
6) And finally, ZBRENT: fatal error in bracketing,please rerun with smaller EDIFF, or copy CONTCAR to POSCAR and continue is a very common error when your EDIFF is very low. I have faced it several time (more than 100 !!!). You can safely consider that, the convergence criteria has been reached when you get this message.
I hope this will solve your problem Jian-Gang Kong
1-The issue may come from the way you constructed your POSCAR
2-Use ISIF=2 and reduce EDIFF to 10^-6.
- Badges
- Science topic
- Similar topics
- Philosophy
- Moral Philosophy
- Free Will