[Pw_forum] born effective charges

Vivek Ranjan vranjan at ncsu.edu
Mon Oct 9 16:40:41 CEST 2006


Hi !

I am trying to calculate born effective charges for a system consisting of
more than hundred atoms. I am doing it using linear response, i.e. I relax
the structure, then I do a scf calculation and then I run a gamma point
phonon calculation. After the programme prints the dielectric constants,
it must calculate the born effective charges. With the given computer
resources at my disposal, it cannot complete this part. So, I must restart
the calculation from the previous step. I use "recover.true." flag. But it
never seems to go beyond the dielectric constant calculation part. Is
there a way I can do this calculation ?

Thanks,
Vivek

PS : I know that I can calculate polarization using berry phase
calculation. Then I can move atoms one-by-one and calculate effective
charges by finite difference. But that will be doing many calculations, so
I want to avoid this method.







More information about the users mailing list