Apparently not possible to get Wannier centers and Lowdin charges during Ehrenfest MD ?

Matt W mattwa... at gmail.com
Tue Apr 10 18:42:05 CEST 2018


The wavefunctions generated during Ehrenfest are quite different to 
standard ground state DFT (for instance they have complex coefficients), so 
it is very likely that the standard routines do not work.

I am not sure what is available.

Matt

On Friday, April 6, 2018 at 6:59:28 PM UTC+1, Dan_M wrote:
>
>
> Dear all, 
>
> I am trying to perform an Ehrenfest MD in such a way that I get also the 
> evolution of the Lowdin and Mulliken charges during the EMD, together with 
> the Wannier centers and cubes every N steps.
>
> The issue is that while I do get in the output the Mulliken charges, I 
> don't get neither the Lowdin charges nor the Wannier centers/cubes. Since 
> the EMD itself runs fine, and I am able to get all that output in the 
> previous single point calculation in which I compute the WFN that I am 
> using to start the EMD, I guess that some flag must be activated when using 
> real_time_propagation or ehrenfest_dyn that prevents the calculation of 
> Lowdins/Wanniers (however I don't see any warning at all even when using 
> PRINT_LEVEL DEBUG).
>
> I looked a bit at the code (5.1 is the version I'm using) and I found that 
> in the file qs_scf_post_gpw.F, in SUBROUTINE write_mo_dependent_results, 
> there is a sentence IF (.NOT. qs_env%run_rtp) THEN ...
>
> While I'm not sure that this is what it is preventing the code to compute 
> that info, if that is the case I guess that this was put there for the case 
> in which one uses DENSITY_PROPAGATION in the &real_time_propagation section 
> (since then the MO's would not be computed?), but I am not using that. 
> Could this be a bug then? Or am I missing something else?
>
> Thanks and best,
> D.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180410/43de764b/attachment.html>


More information about the CP2K-user mailing list