Questions on the use of the WFN_MIX function to obtain the open-shell singlet state

[John Zhang]

Hi

I want to obtain an open-shell singlet state wave function for a C2 molecule using the WFN_MIX function. I first performed a quintet state calculation (converged smoothly) and tried to use this quintet state wave function as the initial guess in the following open-shell singlet state calculation. In the WFN_MIX section, I tried to flip the spin of the original HOMO and HOMO-1 orbitals (Also see the attached input)

    &WFN_MIX

      &UPDATE

        RESULT_MO_INDEX    -1

        RESULT_SPIN_INDEX  BETA

        RESULT_SCALE       0.00000

        ORIG_MO_INDEX      2

        ORIG_SPIN_INDEX    ALPHA

        ORIG_SCALE         1.0000

        ORIG_IS_VIRTUAL    .FALSE.

      &END UPDATE

      &UPDATE

        RESULT_MO_INDEX    -2

        RESULT_SPIN_INDEX  BETA

        RESULT_SCALE       0.00000

        ORIG_MO_INDEX      1

        ORIG_SPIN_INDEX    ALPHA

        ORIG_SCALE         1.0000

        ORIG_IS_VIRTUAL    .FALSE.

      &END UPDATE

    &END WFN_MIX

but the calculation failed with an error message "If ORIG_IS_VIRTUAL is set, the array unoccupied_orbs must be associated!"

I think my usage of WFN_MIX is not correct. Could someone please help on this issue? Thanks!

Best 

John

[Marcella Iannuzzi]

Dear John

I think that your input might work if you remove the key ADDED_MOS from SCF.

Anyway, the input structure for wfn_mix has slightly changed in most recent version of the code. 

Regards

Marcella

[John Zhang]

Dear Dr Lannuzzi

Thank your for your response!

Following you suggestions, I performed a few test calculations (without the ADDED_MOS)

1. Starting from the ATOMIC guess with singlet spin multiplicity. The calculation stopped at the beginning of "Performing wfn mixing" and the error message looks like (I pasted only a few lines) 

"PBLAS ERROR 'Array subscript out of bounds: JX = 7, DESCX[N_] = 4'

from {0,7}, pnum=7, Contxt=0, in routine 'PDCOPY'.

PBLAS ERROR 'Parameter number 4 had an illegal value'

from {0,7}, pnum=7, Contxt=0, in routine 'PDCOPY'.

{0,7}, pnum=7, Contxt=0, killed other procs, exiting with error #-4.

"

2. if the calculation started using the quintet wave function, it immediately stopped and the error message is "Number of occupied MOs on restart unit larger than allocated MOs"

I also checked the examples in the tests, but I cannot figure out what the problem is. Any insight? Thank you again!

Best wishes

John

[Marcella Iannuzzi]

Dear John, 

The restart does not work if the previous calculation had ADDED_MOS, but the ATOMIC guess should work. 

You have to keep the MO_CUBES in the PRINT section of DFT

It works for me. 

Regards

Marcella

[John Zhang]

Dear Prof. Lannuzzi

The previous quintet calculation did not have the "ADDED_MOS". Using ATOMIC guess, I have MO_CUBES in the PRINT section of DFT, but got the same error. Please see attached outputs (I did not include the generated cube files). I am using 2023.2. Maybe my compilation have some issues? or should I try other version of cp2k? Thanks!

Best wishes

Peng

[Marcella Iannuzzi]

Dear Peng

But why have you set ORIG_IS_VIRTUAL as FALSE and given negative indexes to the MOs?

Regards

Marcella

[John Zhang]

Dear Prof. Lannuzzi

The quintet state has 4 singly occupied alpha spin orbitals. I tried to flip two of these alpha spin orbitals to beta spin. therefore, I set the "RESULT_MO_INDEX" as negtive to represent the unoccupied MO space. I actually also tries the opposite way: setting ORIG_MO_INDEX as the virtual MO and RESULT_MO_INDEX as HOMO and HOMO-1. This run was "successful", however, I am not sure I understand the results. The output file shows that "Performing wfn mixing and Writing new MOs to file" at the end of the calculation. In this specific calculation, the spin multiplicity is 5 and I used the ATOMIC guess, the program seems to perform wfn mixing after SCF for the initial quintet state (although not converge), How do i know the wfn mixing is correctly performed? Does the saved wfn file contains the spin-fliped singlet state? Could you please provide some suggestions? Thank you! Attached please find the input and output.

 

Best wishes

 

John