Partial charges in CP2K - simulation fails

[Hasan Al-Mahayni]

Hello everyone,

I have a simple (100) copper slab (100 atoms, 4 layer==> see input.xyz); and I am trying to make the overall charge of my system equal to +1.39. Since the CHARGE tag only takes integers, I am trying to give partial charges to the atoms in the top layer to go from charge +1 to charge +1.39. I used a cp2k tutorial as reference:  https://www.cp2k.org/howto:resp 

I attach my input file ("cp2k.inp") and my output file "RESULTS".  The simulation initially runs fine, the first SCF loop converges, then the simulation crashes and I get this error (See "error.txt"):

 

srun: error: cdr1443: tasks 3,5,7-8,14-15,20,23-25,27,30: Segmentation fault (core dumped)

This is the tags I added:

&PROPERTIES
  &RESP
    &RESTRAINT
    ATOM_LIST 76..100
    TARGET 0.015477832
    STRENGTH 0.0001
    &END RESTRAINT
  &END RESP
&END PROPERTIES

Does anyone know why I am getting this error? Or how can I include partial charges correctly in my system? The goal here is to have a certain Fermi Energy for my system,  so if there is a better way to do this, I would love to know that.

Thanks for your time!

Hasan.

[Marcella Iannuzzi]

Dear Hasan 

The RESP feature is meant to assign charges that reproduce the DFT electrostatic potential as post processing.

The electronic density is not modified and the SCF calculation is not affected.

DFT in CP2K works only with a total integer charge. The constant Fermi energy method cannot be applied.

Regards

Marcella

[Hasan Al-Mahayni]

Hello Marcella,

Thank you for the helpful information! On that note, I would like to ask 2 final  questions:

-is there a way to perform GCE-DFT (Grand Canonical Ensemble) in CP2K? I am asking because my goal here is to include the potential of the electrode (like in experiments). I read the following work, https://www.sciencedirect.com/science/article/pii/S2451910321000636, and I wanted to know if something similar is implemented in CP2K.

- I am also trying to include the solvent effects of the electrolyte with the implicit SCCS model (Andreussi) that is available on CP2K. However, it does not work with my stress tensor tag, which I had to remove. I also saw that in previous threads here, SCCS does not work with kpoints? Any insight on this matter would be immensely helpful.

Thank you so much for your time.

Hasan.

Hasan Al-Mahayni

--
You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cp2k/08xyElGa3Jg/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/b159b64d-d9ce-434c-b445-10fa4d536d81n%40googlegroups.com.