Wannier function calculation

[SS]

Dear cp2k users

I want to run wannier function calculations of my system  (Metal aqua complex) using cp2k. I followed the tutorials given in the cp2k forum related to wannier function calculations. When I follow the same procedure in my system it's writing cube files. (A large number of cube files as my system has near 300 atoms including metal). 

My queries are 

1) How to find out which cube file has the information corresponding to the metal.

2) How to identify the wannier functions which are associated with the metal and to visualize it.

Thank You 

Shyam

[Patrick Gono]

Dear Shyam,

You can select which states to localize either through their index (keyword LIST in CP2K_INPUT / FORCE_EVAL / DFT / LOCALIZE), or by selecting an energy range. Likewise, you can select which orbitals are to be printed in the form of a cube file by selecting their indexes via the keyword CUBES_LIST in the subsection CP2K_INPUT / FORCE_EVAL / DFT / LOCALIZE / PRINT / WANNIER_CUBES.

The localized orbitals are ordered by their respective energy.

You can print the Wannier centers as an .xyz file via CP2K_INPUT / FORCE_EVAL / DFT / LOCALIZE / PRINT / WANNIER_CENTER. I prefer to print the positions of the ions as well, makes it easier to inspect the whole thing. You can do that by switching on the keyword IONS+CENTERS in that same subsection.

You can then use VMD (https://www.ks.uiuc.edu/Research/vmd/), or a program of your choice, to plot the Wannier centers (+ions, optionally). This allows you to locate in space the centers of interest. You then look at their indexes and find the appropriate cube files.

So, my general workflow is to:

1) localize all orbitals but only print out centers + ions

2) inspect using a script, or visually, and identify which wannier orbitals are of interest

3) rerun and compute the cubes corresponding to these centers.

Alternatively, if you have only a few hundred of them (and are not limited by memory), you can print out all cube files at the start and then retain only those that interest you.

Yours sincerely,

Patrick Gono

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/c74af534-e1bb-4e5e-9d9d-e4760dccef13%40googlegroups.com.

[SS]

Thank you Patrick Gono for your help.

I have some doubts regarding the CP2K_INPUT / FORCE_EVAL / DFT / LOCALIZE  section you mentioned. Here I could only find the keywords LIST and LIST_UNOCCUPIED for indices of occupied and unoccupied states. I think you were mentioning about this. If not please correct me.

As you mentioned it is better to localize all the orbitals and select the needed one later. 

Is there any way to do  the calculation particularly for an atom in the system? Like just for the metal atom in the system.

Thank you 

Shyam

To unsubscribe from this group and stop receiving emails from it, send an email to cp...@googlegroups.com.