[CP2K-user] [CP2K:21320] SCF Convergence in High Impact Energy QM/MM Collisions

[Nicholas Laws]

Hi CP2K Community,

I hope this message finds you well. I am trying to perform a mixed QM/MM 
high energy collision of a fast incident neutral (EMI-BF4) system (composed 
of a cation, EMI+, and an anion, BF4-) with an Au surface. In my model, the 
Au surface is handled at the MM level and initially placed about 20 Å away 
from the EMI-BF4 (QM region). The incoming EMI-BF4 travels toward the Au 
surface at approximately 100 eV, and my goal is to observe the 
post-collision fragmentation and charge redistribution while preserving SCF 
convergence.

However, I’m encountering SCF convergence issues. As bonds begin to break 
from collision, the first few OUTER_SCF loops of the final time step prior 
to failure converge seemingly well, but the SCF begins to diverge. I can 
see the total electronic density on the regular grids creeping upward each 
subsequent OUTER_SCF iteration, eventually leading to failed convergence. 
Given this is a high-energy collision and the system is partly in vacuum 
(non-periodic boundary conditions), I suspect that something about the 
setup or the transition between the vacuum and the Au surface region might 
be causing the SCF solver to struggle.

I’ve attached my input script (which shows the QM/MM method, the basis 
sets, potentials (DZVP-GTH/PBE), and my MD parameters), output file, and 
truncated trajectory file. I’m wondering if the community could suggest 
best practices for the MGRID section (for example, choices of NGRIDS, 
CUTOFF, REL_CUTOFF), or advice on how the cell size for the QM region might 
need adjusting (e.g., wrapping/fragments drifting out of the defined 
region). I’d also appreciate any tips on SCF convergence strategies (e.g., 
different guess methods, more robust smoothing for the density, or 
alternative Poisson solver approaches) that might stabilize the calculation 
in a high energy collision scenario.

Thank you very much for your time, and please let me know if you need any 
further information.

All my best, 
Nick

-- 
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/ba15268d-5199-4361-97b2-91b9cb3ba3f6n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250331/6f11dabf/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 100eV_EMIBF4_Surface_NVE-pos-1.xyz
Type: chemical/x-xyz
Size: 1295796 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250331/6f11dabf/attachment-0001.xyz>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: EMIBF4_Surface_100eV.inp
Type: chemical/x-gamess-input
Size: 234247 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250331/6f11dabf/attachment-0001.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: EMIBF4_Surface_100eV.out
Type: application/octet-stream
Size: 342108 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250331/6f11dabf/attachment-0001.obj>