Mulliken Analysis for Individual Orbitals

[Leili Rassouli]

Dear all, 

I want to print the Mulliken analysis for all occupied and unoccupied orbitals. I used PRINT_ALL and PRINT_GOP under CP2K_INPUT / FORCE_EVAL / DFT / PRINT / MULLIKEN directory, but it doesn't print the Mulliken analysis for individual orbitals. 

I want to check which atomic orbital has the highest population of alpha and beta in each orbital (for the range of HOMO-10 - LUMO+10). Is there any way to print these values?

In addition, I want to do the Mulliken analysis for Natural Transition Orbitals (NTOs). Can I use cp2k to do that? 

 Best regards,

Lili

[Martin Konôpka]

Dear Leili,

As for the 2nd part of your question, I am not able to give an advice. As for the 1st part, I am also not sure but you might perhaps try to use PDOS section, something like the one shown below.

If you want to do a geometry relaxation or molecular dynamics as well, you should somehow prevent calculating and printing all that information at each ionic step. (It would be waste of resources.) I personally do PDOS/LDOS calculations in separate runs after geometry relaxation was completed.

&DFT

..............

..............

..............

  &PRINT

    &MO_CUBES
      NLUMO  10
      STRIDE  2 2 2
       WRITE_CUBE  .FALSE.
    &END MO_CUBES

    &PDOS

       FILENAME  PDOS

       NLUMO  10

       COMPONENTS    ! optional

    &END PDOS

    &LDOS

      LIST   1         ! the atom number 1

      COMPONENTS

    &END LDOS

.....................

.....................

    &LDOS

       LIST   8         ! the atom number 8 say

       COMPONENTS

    &END LDOS

    &EACH
      QS_SCF  0
    &END EACH

  &END PRINT

&END DFT

Best regards

Martin.

st 7. 9. 2022 o 15:33 Leili Rassouli <rassoul...@gmail.com> napísal(a):

--
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/4144231f-6138-4f54-a893-cbadca574c1en%40googlegroups.com.

[Leili Rassouli]

Dear Martin,
Thank you for sharing this information with me. I really appreciate it.

Regarding the Mulliken analysis of NTO, I use  Multiwfn program. I enter  .molden file from cp2k as an input file in  Multiwfn program for doing this calculation. However, I still have some problems to do this.
 
I did  PDOS/LDOS calculation and I have some questions:
1- I am working on Fe2O3 supercell with 120 atoms. Do I need to write the LDOS part for each atom (writing 120 times or writing a for loop in Linux to do that) as I want to check the relative contribution of each atom to the total DOS population?
2- Do you know how PDOS is calculated in cp2k? Is the relative contribution based on  Mulliken analysis?
Best,
Leili

 &LDOS

       LIST   1         ! the atom number 1

       COMPONENTS

    &END LDOS   

 &LDOS

       LIST   2         ! the atom number 2 

       COMPONENTS

    &END LDOS

 . 

 . 

 .

&LDOS

       LIST   120         ! the atom number 120 

       COMPONENTS

    &END LDOS

[Martin Konôpka]

Dear Leili,

To your 1st question: I am aware that CP2K knows some loops syntax in input files. I am used to VIM editor which makes me easy to edit text files even if they are very large (and I try to keep order in structure of my files and use the diff tool to check them). Anyway, to keep your main input file short, you can use the "@include" statement (which works similarly but not identically as the #include in C codes). Simply put the long part with the LDOS sections to an external file, say "ldos.inc" and then, inside the CP2K input file, type '@include ldos.inc' (if I am not mistaken).

To your 2nd question - I was not able to find a reference to the subject but I believe the CP2K's way to calculate PDOS and LDOS is a pretty standard one, based on calculating overlaps of eigenstates with some properly defined localised orbitals.

Best
Martin.

št 8. 9. 2022 o 19:08 Leili Rassouli <rassoul...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/a65acfd0-c670-40ea-918c-5513ebe9531an%40googlegroups.com.

[Martin Konôpka]

PS Sorry for my mistype: I am NOT aware of such a syntax.

št 8. 9. 2022 o 22:27 Martin Konôpka <konop...@gmail.com> napísal(a):