[CP2K-user] [CP2K:22181] Re: qm_mm/image_charge: optical spectrum

[jgh]

Hi
you are testing your luck here. The combination of this many non-standard 
options is prone to errors.
QM/MM image charges
hexagonal grid
2d PBC
TDDFT
XC grid smoothing
...

When I remove the grid smoothing and run the input with 3d PBC it works 
(latest trunk version of CP2K), see attached input file.
If the results can be trusted is another question.

Using the SAOP optin also run through but had negative excitation energies.
I wouldn't use SAOP in a production run without first do extensive tests on 
small systems.

The new TDDFT options to use a restricted occupied orbital set for 
excitation works for the few systems we have tested up to now.
If you want to try it, please keep in mind that it is a new option.

regards
JH

On Friday, April 3, 2026 at 10:47:09 AM UTC+2 volk... at gmail.com wrote:

> Good day.
>
> I wish to bring under attention 
> of the developers 
> (of this approach
> https://manual.cp2k.org/trunk/methods/qm_mm/image_charges.html)
> two examples, 
> where 
> their code, apparently, fails to compute
> optical electronic transitions.
>
> Specifically, following the online examples,
> I condition OH neutral molecule at four gold layers (facet 111),
> where gold atoms are fixed under hexagonal symmetry
> to provide a stable, well-defined structural reference.
>
> The neutral OH molecule has one unpaired electron.
> Therefore multiplicity of the system is 2.
>
> My coworker and me, we run time-dependent DFT to compute three transitions
> under regular setting
> and under SAOP setting.
>
> Here, I attach inputs and outputs for both cases 
> and the coordinate file.
>
> If you wish to try to run the jobs yourself,
> you would need Au.pot file, which is present on the CP2K website.
> and you would need to comment out the restart wavefunction 
> and reset charges to ATOMIC. 
>
> My co-worker and me tested the jobs 
> under linux cp2k of different installations,
> including the most recent one.
>
> The results are the same:
> 1) SAOP job dies with the same
>  *** WARNING in qs_tddfpt2_utils.F:634 :: Orbital energy correction ***
>  *** potential is an experimental feature. Use it with extreme care ***
> 2) regular TDDFT job dies without any notice.
>
> If possible, 
> would you be kind to look at the case and comment,
> if the failure(s) can be amended.
>
> If the CP2K QMMM-IMAGE theory cannot solve such cases,
> would you instruct kindly
> if the recent CP2K version got an update for TD-DFT
> to compute optical transitions 
> to include only a narrow sub-set of orbitals, which may be selected
> to be specific to HOMO-LUMO properties of the adsorbed molecule
> rather than to the metal continuum sp-d transitions.   
>
> If such TD-DFT is possible, would you be kind to hint
> how to instruct this in input file.
>
> Thank you for your time and attention.
>
> Sincerely, 
> I wish you a wonderful and peaceful time 
> during the spring festival of life revival.
>
> Victor
>
>

-- 
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/05bf4e0a-fc52-4223-907e-41b7c3139747n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20260414/7467f1cc/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Au4_OH.inp
Type: chemical/x-gamess-input
Size: 3443 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20260414/7467f1cc/attachment.inp>