Wannier centers - post-processing AIMD equilibrium trajectories

[Emma Rossi]

Dear all,

I writing since I would simulate the IR spectrum of my system using equilibrium trajectories obtained through ab initio molecular dynamics. I need the transition dipole moment to get the correct intensities of peaks.

I was wondering if there is a post-processing procedure to compute Wannier centers with CP2K using my ab initio trajectory. I've found only tutorials to print Wannier centers on-the-fly during the simulation. 

If this is not possible, are there different strategies to extract dipole information from AIMD trajectories without performing again the simulations?

I would be very grateful if you could help me. 

Thank you very much in advance for your support.

Best regards,

Emma Rossi

[Marcella Iannuzzi]

Dear Emma

For the calculation of the Wannier centres   the localisation of the molecular orbitals has to be performed on the optimised electronic structure. 

Hence for each configuration for which the WC are needed, also the MOS are needed, the trajectory is not enough. 

If it is enough to have the WC every N MD time step, one possibility would be to run REFTRAJ on the generated trajectory, with a stride N, where for each selected configuration a new wave function optimization is performed. Choosing this option you can select  the settings also differently from what was used in MD. 

Though for the IR spectrum  it is recommended to compute the dipole quite frequently to resolve properly the high frequencies.

Kind regards

Marcella

[Emma Rossi]

Dear Marcella,

I solved the problem computing the total dipole of the simulation box with the Berry phase approach, since I didn't need the molecular dipole of each molecule.

However, I have a doubt. I used the TOTAL_DIPOLE keyword as follows

&DFT

 &LOCALIZE  T
       &PRINT
         &TOTAL_DIPOLE  ON
           FILENAME =totdipole
           PERIODIC  T
           &EACH
             MD  1
           &END EACH
         &END TOTAL_DIPOLE
       &END PRINT
     &END LOCALIZE
   &END DFT

, but I found also  this possibility on the manual:  FORCE_EVAL / DFT / PRINT / MOMENTS

What is the difference between these two approaches to compute the dipole moment of the box?

Thank you very much in advance for your help and support.

Best regards,

Emma Rossi