[CP2K-user] [CP2K:21668] Re: SCF convergence

[Sihem AZIZI]

Thanks a lot

On Wednesday, July 16, 2025 at 12:21:27 PM UTC+1 Frederick Stein wrote:

> Dear Sihem,
> I would first try to reduce the mixing by adding
> &MIXING
>   # Lower values improve convergence but increase the number of SCF 
> iterations
>   # Difficult systems may require even lower values
>   ALPHA 0.01
> &END
> to your SCF section. You can also try different mixing schemes (see here: 
> https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/SCF/MIXING.html 
> ). Unfortunately, there is no standard recipe with difficult to converge 
> systems.
> HTH,
> Frederick
>
> Sihem AZIZI schrieb am Mittwoch, 16. Juli 2025 um 12:35:59 UTC+2:
>
>> Dear Johann,
>>
>> Thanks a lot for your response.
>> I started a new job with the file attached. I paid attention to your 
>> comments, but I still get the same warning.
>> *** WARNING in qs-scf.F:542 :: SCF run NOT converged ***
>> I reduced the number of Ru atoms to reduce the calculation time, and now 
>> I don't know whether to stop the calculation or let it continue. Can you 
>> tell me what's wrong? Thanks.  ,
>> Sihem
>>
>> On Thursday, July 10, 2025 at 5:38:21 PM UTC+1 Johann Pototschnig wrote:
>>
>>> Your SCF shows convergence, but doesn't reach the threshold. 
>>> It might be sufficient to increase the number of iterations: MAX_SCF 200
>>>
>>> Otherwise you could try different convergence accelerators. 
>>>
>>> https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/SCF.html#CP2K_INPUT.FORCE_EVAL.DFT.SCF
>>>
>>> For a geometry optimization you should increase the SCF 
>>> threshold: EPS_SCF 1.0E-6
>>> otherwise the gradients will be unreliable. 
>>>
>>> You have a lot of Ru atoms, but only add 10 orbitals in the Fermi 
>>> smearing, which might also cause problems in the SCF convergence. 
>>>
>>> VdW correction are described in:
>>>
>>> https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/XC/VDW_POTENTIAL.html
>>>
>>> On Wednesday, July 9, 2025 at 8:18:18 PM UTC+2 Sihem AZIZI wrote:
>>>
>>>> Dear all,
>>>>
>>>> I ran a geometry optimization calculation for a 3-layer Ru slab. The 
>>>> calculation ended with the following warning:
>>>>  *** WARNING in qs_scf.F:542 :: SCF run NOT converged ***
>>>> I have attached the input and the output files.
>>>> Also, I would like to know how to add the dispersion and contribution 
>>>> of the VDW forces.
>>>> Thanks a lot in advance for your help
>>>>
>>>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/7b76cde0-ebf1-4556-bfca-24b7ee220460n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250717/93cd2a0b/attachment-0001.htm>