Is it possible to set charge on a specific atom in a system?

[Jingyun Ye]

Hi, all

If I want to calculate metal complex, is there any parameter I can set charges on the metal? I don't want to the charge shared in the whole system.

Thanks!

[Florian Schiffmann]

Hi,

I guess you are doing DFT simulations. However I am not quite sure what you attempt without more information. The only thing I can imagine you acually want to do is to run the system with a broken symmetry guess which can be acchieved with the BS section in the kinds section.

Flo

[Zhendong Guo]

Hi ,

 I guess the meaning of jingyun's question is like that: For example, he wants to put a Cu2+ into his neutral system, two extra positive charges is needed. How to set parameters to make the extra charges located on the Cu atom rather than being shared by the whole system.

Best regards
zhendong

 

--
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 post to this group, send email to cp...@googlegroups.com.
Visit this group at http://groups.google.com/group/cp2k.
For more options, visit https://groups.google.com/d/optout.

--

Zhendong Guo 

Phone: +41786943316

Email:  zhendon...@gmail.com

            zhendo...@epfl.ch-------------------------------------------------------------------------------------------------

[Jingyun Ye]

Hi, Zhendong

That's what I mean. Thanks!

[Samuel Andermatt]

As fa as I know it is only possible to charge the whole system (but maybe somebody knows something more). The charge will then be placed on the energetically most favorable spot. I think under periodic boundary conditions a uniform background charge will be applied to keep the whole system neutral (as otherwise the energy would diverge).

What you could do is to use the Kim-Gorden method implemented in linear scaling DFT in CP2K. It allows you to subdivide the system into multiple subsystems. Therefore electrons could at least not be shared between subsystems and you could ensure your metal complex has the charge of a whole number (although you could still not specify what this number will be).

[Florian Schiffmann]

With the BS section in the KIND section where you specify the Basis set, you have total control of the elecron configuration in the initial guess. However, I expect the SCF will find back the delocalized solution. Most of the time this stragtegy  is only meaningful in combination with open shell calculations, where you can have symmetry breaking.
Using this keyword you will have to have a look here
http://manual.cp2k.org/cp2k-2_3-branch/CP2K_INPUT/FORCE_EVAL/SUBSYS/KIND/BS.html
and then start to modify the electron configuration from what you have in your pseudo

Flo

[Marcella Iannuzzi]

Hi 

Indeed the BS initialisation might work to localise the charge, even if not in combination with open shell calculations. 

This is true if the electronic structure with the localised charge is at least a metastable minimum. 

In this case, BS helps the convergence to this state. If not, then the electrons will be redistributed by the optimisation. 

regards

Marcella