Point charges to simulate an electric field?

[Бранислав Миловановић]

Dear cp2k users,

Is it possible to introduce ghost atoms with no basis set and with finite non-integer point charge in AIMD calculations? Something like MM point charges.

I want to try to simulate behaviour of some small molecules in the electric field which originates from a particular sequence of the point charges.

Thanks

Branislav

[Matt Watkins]

Multiple options - most directly:

https://manual.cp2k.org/cp2k-9_1-branch/CP2K_INPUT/FORCE_EVAL/SUBSYS/KIND.html#list_CORE_CORRECTION

lets you change the core charge on a pseudopotential. If you make a fake hydrogen and adjust charge appropriately you'd get close to what you were after (won't be a point charge but gaussian smeared charge).

There are other routines that could load an external potential on a grid or make use of the QM/MM code in some form.

Matt

[Бранислав Миловановић]

Thanks, just want to check if this is right way to do this:

I defined atoms labeled X in the coord file and then I added this section into the SUBSYS:

    &KIND X
      ELEMENT H
      GHOST .TRUE.
      BASIS_SET NONE
      POTENTIAL GTH-PBE-q1
      CORE_CORRECTION 0.12   # 3/25
    &END KIND

Is this the right way to divide for example 3e charge over 25 fake hydrogen atoms?

Branislav

[Бранислав Миловановић]

Additionally, should I in such cases change total charge od the system?

[Matt Watkins]

There will be the core charge of the original atom as well - so you need neutralise the atom core (-1 in your case) then add your charge back.

Yes - the charge keyword will alter the number of electrons in the system so if you want these to be virtual protons not hydrogen atoms then you do need to remove the electron as well with the charge keyword.

If your overall system is charged then normal issues of charged periodic systems apply.. 

Matt

[Бранислав Миловановић]

Thanks.
OK, I think I get it now. 

So I have set CHARGE keyword to 25, since I want to have 25 fake H atoms and each to represent smeared charge of +0.28e, and the rest of my system in neutral.

My KIND section now:

    &KIND X
      ELEMENT H
      BASIS_SET NONE
      POTENTIAL GTH-PBE-q1
      CORE_CORRECTION -0.88
    &END KIND

However, when I perform population analysis I find finite population on these fake atoms:

                           Hirshfeld Charges

  #Atom  Element  Kind  Ref Charge     Population                    Net charge
      1       H      1       0.280          0.046                         0.234
      2       H      1       0.280          0.047                         0.233
      3       H      1       0.280          0.052                         0.228
      4       H      1       0.280          0.055                         0.225
      5       H      1       0.280          0.046                         0.234
      6       H      1       0.280          0.051                         0.229
.

.

.

How is this possible when they are described with no basis at all?

Best,

BM

[Matt Watkins]

Hi, 

those you list are Hirshfeld charges that are defined geometrically without reference to a basis set (intergrated density in some radius or similar). If you look at Mulliken charges those should be as you expect.

At a glance the setup looks right to me.

Matt

[Бранислав Миловановић]

Hi,

Ok, thank once again for clarifying.

Mulliken population analysis gives expected values for these fake atoms.

Best
BM