Troubleshooting electronic convergence: Difference between revisions

From VASP Wiki
No edit summary
Line 1: Line 1:
There can be many reasons why the SCF convergence is not reached.
There can be many reasons why convergence to the electronic ground state fails. Below you find some strategies to overcome convergence issues in the [[Category:Electronic minimization|electronic minimization]].
Naturally, there are also many different possible solutions.
== Step-by-step instructions ==


Try to create a minimal [[INCAR]] file with as few tags as possible.
'''Step 1:''' Simplify the calculation and reduce time-to-solution. Try to create a minimal [[INCAR]] file with as few tags as possible. If the calculation converges, then gradually add them back until you find which one was causing the problem. Try to reduce the time-to-solution as much as possible by lowering the '''k'''-point sampling (or using gamma-only calculations if applicable), lower {{TAG|ENCUT}}, use {{TAG|PREC}}=Normal.
If the calculation converges then gradually add them back until you find which one was causing the problem.
Try to reduce the time-to-solution as much as possible by lowering the '''k''' point sampling (or using gamma-only calculations is that is sensible), lower {{TAG|ENCUT}}, use {{TAG|PREC}}=Normal.


Check if you have enough bands. You can do this by looking at the [[OUTCAR]] file and checking that there are enough 'empty' states (i.e. states with zero occupation). When using an iterative solver the last states might not be accurately described, if these are occupied then convergence is likely to fail.
'''Step 2:''' Fixing the charge density (for cases where density mixing is used)
Often the VASP default setting for {{TAG|NBANDS}} is insufficient for systems with f-orbitals.


Systems with f-orbitals are in general difficult to converge.
'''Step 3:''' Increase {{TAG|NBANDS}}. Check if you have enough bands. You can do this by looking at the [[OUTCAR]] file and checking that there are enough empty states, i.e., states with zero occupation. When using an iterative solver, the last states might not be accurately described, if these are occupied, then convergence is likely to fail. Often, the VASP default setting for {{TAG|NBANDS}} is insufficient for systems with f-orbitals or calculations with meta-GGA's.
 
'''Step 4:''' Switch {{TAG|ALGO}}.
 
'''Step 5:''' For {{TAG|ALGO}}=5X or 4X change {{TAG|TIME}}.
 
== Method-specific recommendations ==


In the following, we will describe a few recipes that work for particular systems.
In the following, we will describe a few recipes that work for particular systems.
Some of these recipes might be transferable but it is hard to provide a recipe that is guaranteed to work for all cases.
Some of these recipes might be transferable even to other methods.
 
=== Magnetic calculation with LDA+U ===
== Magnetic calculation with LDA+U ==
Magnetic calculations present a lot of challenges, in particular when the energy differences between different magnetic configurations are small.
Magnetic calculations present a lot of challenges in particular when the energy differences between different magnetic configurations are not large.
Our recommendation is to split the calculation into multiple steps:
Our recommendation is to split the calculation into multiple steps:
# give initial magnetization only to the magnetic atoms
# give initial magnetization only to the magnetic atoms
Line 24: Line 26:
## step 3 add LDA+U tags keeping {{TAG|ALGO}}=All and small {{TAG|TIME}}
## step 3 add LDA+U tags keeping {{TAG|ALGO}}=All and small {{TAG|TIME}}


It might be useful to split step 1. in two by first running with a smaller {{TAG|ENCUT}} and then restarting the calculation from the {{FILE|WAVECAR}} with the desired {{TAG|ENCUT}}
It might be helpful to split step 1. in two by first running with a smaller {{TAG|ENCUT}} and then restarting the calculation from the {{FILE|WAVECAR}} with the desired {{TAG|ENCUT}}.
 
== MBJ calculation ==


=== MBJ calculation ===
This exchange-correlation functional is not particularly easy to converge in some systems.   
This exchange-correlation functional is not particularly easy to converge in some systems.   
For these systems we recommend that you split the calculation into multiple steps that successively bring you closer to the solution (always restarting from the WAVECAR of the previous step):
For these systems, we recommend that you split the calculation into multiple steps that successively bring you closer to the solution (always restarting from the WAVECAR of the previous step):
# Converge with the PBE functional
# Converge with the PBE functional
# Converge with the {{TAG|METAGGA}}=MBJ functional with the {{TAG|CMBJ}} parameter set to some value and {{TAG|ALGO}}=All and {{TAG|TIME}}=0.1
# Converge with the {{TAG|METAGGA}}=MBJ functional with the {{TAG|CMBJ}} parameter set to some value and {{TAG|ALGO}}=All and {{TAG|TIME}}=0.1
# Converge with the {{TAG|METAGGA}}=MBJ functional without {{TAG|CMBJ}} parameter set and {{TAG|ALGO}}=All and {{TAG|TIME}}=0.1
# Converge with the {{TAG|METAGGA}}=MBJ functional without {{TAG|CMBJ}} parameter set and {{TAG|ALGO}}=All and {{TAG|TIME}}=0.1


Similar to the [[Difficult_to_converge_systems#Magnetic calculation with LDA+U|previous recipe]] it might be useful to run steps 1. to 3. with a low {{TAG|ENCUT}} and then perform step 3. again with the desired {{TAG|ENCUT}}.
Similar to the [[#Magnetic calculation with LDA+U|recipe for magnetic calculation with LDA+U]], it might be helpful to run steps 1. to 3. with a low {{TAG|ENCUT}} and then perform step 3. again with the desired {{TAG|ENCUT}}.


[[Category:Common Pitfalls]]
[[Category:Howto]][[Category:Electronic minimization]]

Revision as of 10:52, 20 October 2023

There can be many reasons why convergence to the electronic ground state fails. Below you find some strategies to overcome convergence issues in the.

Step-by-step instructions

Step 1: Simplify the calculation and reduce time-to-solution. Try to create a minimal INCAR file with as few tags as possible. If the calculation converges, then gradually add them back until you find which one was causing the problem. Try to reduce the time-to-solution as much as possible by lowering the k-point sampling (or using gamma-only calculations if applicable), lower ENCUT, use PREC=Normal.

Step 2: Fixing the charge density (for cases where density mixing is used)

Step 3: Increase NBANDS. Check if you have enough bands. You can do this by looking at the OUTCAR file and checking that there are enough empty states, i.e., states with zero occupation. When using an iterative solver, the last states might not be accurately described, if these are occupied, then convergence is likely to fail. Often, the VASP default setting for NBANDS is insufficient for systems with f-orbitals or calculations with meta-GGA's.

Step 4: Switch ALGO.

Step 5: For ALGO=5X or 4X change TIME.

Method-specific recommendations

In the following, we will describe a few recipes that work for particular systems. Some of these recipes might be transferable even to other methods.

Magnetic calculation with LDA+U

Magnetic calculations present a lot of challenges, in particular when the energy differences between different magnetic configurations are small. Our recommendation is to split the calculation into multiple steps:

  1. give initial magnetization only to the magnetic atoms
  2. use spin-polarized calculation
  3. perform the calculation in 3 steps (always starting from the previous WAVECAR):
    1. step 1 with ICHARG=12 and ALGO=Normal without any LDA+U tags
    2. step 2 with ALGO=All (Conjugate gradient) and a small TIME step 0.05 instead of the default 0.4 (this is crucial)
    3. step 3 add LDA+U tags keeping ALGO=All and small TIME

It might be helpful to split step 1. in two by first running with a smaller ENCUT and then restarting the calculation from the WAVECAR with the desired ENCUT.

MBJ calculation

This exchange-correlation functional is not particularly easy to converge in some systems. For these systems, we recommend that you split the calculation into multiple steps that successively bring you closer to the solution (always restarting from the WAVECAR of the previous step):

  1. Converge with the PBE functional
  2. Converge with the METAGGA=MBJ functional with the CMBJ parameter set to some value and ALGO=All and TIME=0.1
  3. Converge with the METAGGA=MBJ functional without CMBJ parameter set and ALGO=All and TIME=0.1

Similar to the recipe for magnetic calculation with LDA+U, it might be helpful to run steps 1. to 3. with a low ENCUT and then perform step 3. again with the desired ENCUT.