System Explodes After Stable Initial Steps

[Parveen Rawal]

Dear all,

I am a novice to cp2k software and this is my first time performing an AIMD simulation with CP2K. 

I have been studying an isolated intermediate for which I have used a box of 25x25x25. The simulation runs smoothly up to a certain number of steps (see snap1). However, in the very next step, all the atoms suddenly scatter away, resulting in an unstable simulation (see snap2 for reference).

Could you please suggest how to address this issue? Is there anything wrong with my input file?

Note: snap1 shows the structure immediately before snap2.

Any suggestions would be greatly appreciated.

Thank you

Best,

Parveen  

 

[Stein, Dr. Frederick]

Dear Parveen,
I have the following recommendations:
1. Did you properly converge all relevant parameters to have consistent
gradients: CUTOFF as large as necessary and EPS_DEFAULT as low as necessary
(in
https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/QS.html#CP2K_INPUT.FORCE_EVAL.DFT.QS.EPS_DEFAULT))?
2. The D3BJ_SCALING is not necessary because you provided
REFERENCE_FUNCTIONAL. I am not sure which parameters are taken here but the
values you provided are probably wrong.
Best,
Frederick

[Parveen Rawal]

Dear Dr. Frederick

Thank you very much for your suggestions. 

Yes, I have converged all the relevant parameters cutoff, rel_cutoff, EPS_default. 

I also removed D3BJ_SCALING, but the problem still persists.

Below are the convergence results for CUTOFF and EPS_default:

Cut_off	Energy
100	-336.717771997
150	-336.659886509
200	-336.657577737
250	-336.655498918
300	-336.654830160
350	-336.654922067
400	-336.654516782
450	-336.654516777
500	-336.654498644
550	-336.654498679
600	-336.654524447
650	-336.654537025
700	-336.654529932
750	-336.654525174
800	-336.654525294
850	-336.654529077
900	-336.654525222

EPS_default	Energy
1.0E-5	-336.6542499233
1.0E-6	-336.6545261902
1.0E-7	-336.6545296232
1.0E-8	-336.6545167817
1.0E-9	-336.6545164402
1.0E-10	-336.6545159953
1.0E-11	-336.6545158358
1.0E-12	-336.6545157974

Best,

Parveen

--
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+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/ximss-52900851%40cg4.fz-rossendorf.de.

--

Best Regards

Parveen Rawal

Research Fellow 

Department of Chemistry IIT Roorkee

[Stein, Dr. Frederick]

Dear Parveen,
Are the forces also converged (see
https://manual.cp2k.org/cp2k-2026_1-branch/CP2K_INPUT/FORCE_EVAL/PRINT/FORCES.html
for information on how to print them)? What happens if you decrease EPS_SCF
(While using an EPS_DEFAULT of EPS_SCF^2 for a reasonable accuracy of the
calculation)?
Because you did not provide the geometry, we cannot reproduce the problem.
Can you provide an output file or at least the last one or two MD steps and
the header of the output file with the general information on CP2K (version,
how compiled etc.)?
Best,
Frederick

Am Thu, 25 Jun 2026 10:12:25 +0200 schrieb Parveen Rawal
<rawa...@gmail.com>:

> *Parveen Rawal*

> Research Fellow
> Department of Chemistry IIT Roorkee
>

> --
> 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+uns...@googlegroups.com.
> To view this discussion visit

>https://groups.google.com/d/msgid/cp2k/CAERtc%2B0oR%3DM6%3DOMBH8uokMC2BGUZvNU3rzGbaVdcpzED4arW_w%40mail.gmail.com.

Dr. Frederick Stein
Scientific Computing Core
CASUS - Center for Advanced Systems Understanding, Görlitz
Helmholtz-Zentrum Dresden-Rossendorf e.V. (HZDR)

[Parveen Rawal]

 NVT_Ru_CBr_complex-pos-1.pdb

 NVT_Ru_CBr_complex.inp

 NVT_Ru_CBr_complex.out

Dear Dr. Frederick,

I have attached the input, output and trajectory files for your reference.

In the meantime, I will try your suggestions. 

Best,

Parveen

To view this discussion visit https://groups.google.com/d/msgid/cp2k/ximss-52832036%40cg6.fz-rossendorf.de.

--

Best Regards

Parveen Rawal

[Stein, Dr. Frederick]

Dear Parveen,
After MD step 1991, the SCF iteration did not converge with the warning that
the density is non-zero at the boundary. This suggests that you box size is
too small. You may try a larger box of 40-50 A. If this kind of warning is
shown again, you should increase the box size further. In recent versions of
CP2K, the failing SCF convergence lets CP2K abort.
Best,
Frederick

Am Thu, 25 Jun 2026 12:19:34 +0200 schrieb Parveen Rawal
<rawa...@gmail.com>:
> NVT_Ru_CBr_complex-pos-1.pdb
> <https://drive.google.com/file/d/1YPNQmQSzvWBgTCkL9e0-w4kUr-O99uCT/view?usp=drive_web>
> NVT_Ru_CBr_complex.inp
> <https://drive.google.com/file/d/1sFR1bQrAhrI2AcvFU8BoWybMlaHX94zW/view?usp=drive_web>
> NVT_Ru_CBr_complex.out
> <https://drive.google.com/file/d/1toOehFqmPSQe69GpCHFNXgwQuEMe_Rr4/view?usp=drive_web>

>https://groups.google.com/d/msgid/cp2k/CAERtc%2B2cJ1%2BxBNnkJZN4A7QzgF5BPFivTh43fKPRbmmBBrci2g%40mail.gmail.com.