Convergence issue in slab calculations with Zero-Kpointss

To share experience including discussions about scientific questions.


Moderators: Global Moderator, Moderator

Post Reply
Message
Author
rishikanta_m
Jr. Member
Jr. Member
Posts: 72
Joined: Wed Jun 15, 2022 5:52 am

Convergence issue in slab calculations with Zero-Kpointss

#1 Post by rishikanta_m » Thu Jul 27, 2023 4:55 am

Hi,

I am running a PBE calculation with zero kpoints for a slab. The INCAR file is such:

Code: Select all

# Wave calculation 
   ISTART = 0
   ICHARG = 2

# general
   ENCUT  = 500
   LWAVE = .TRUE.

# SCF specific setting
   IBRION = -1
   NSW = 0
   ISMEAR = 0
   SIGMA = 0.02
   EDIFF = 1E-5
   NELM = 300

#   LORBIT = 11
   LSORBIT=.TRUE.                
   SAXIS= 0 0 1  
   ISYM=0

# For large systems
  LREAL=Auto
  ALGO=FAST
  LPLANE =.TRUE. 


However, in the OSZICAR, the EDIFF reaches 1E-05 but still has not converged and still running. What can be the issues? The oszicar file is attached.
You do not have the required permissions to view the files attached to this post.

pedro_melo
Global Moderator
Global Moderator
Posts: 138
Joined: Thu Nov 03, 2022 1:03 pm

Re: Convergence issue in slab calculations with Zero-Kpointss

#2 Post by pedro_melo » Thu Jul 27, 2023 8:02 am

Dear rishikanta_m,

When you wrote that you are running a calculation with zero k-points, what exactly did you mean? For a very large system or a molecule, all calculation will have one k-point, Gamma.

Another point: the total energy of the system is an extensive quantity, so it will increase with the number of atoms. Depending on that number it might be unreasonable to ask any code to converge the total energy to the same precision as a system with only 2 or 3 atoms per unit cell.

Finally, results with ALGO=FAST are often used for quick calculations, and can lead to inaccurate results. I suggest that you change it to NORMAL and see if the convergence improves.

Kind regards,
Pedro Melo

Post Reply