ERROR: should be no new maxima in edge refinement

[juganta]

Thanks a lot ! Here is the link for CHGCAR and CHGCAR_sum.

https://www.dropbox.com/s/6cajmi7keojnd ... ar.gz?dl=0

Regards,

[graeme]

There is something wrong with your CHGCAR_sum file. I've seen this before (perhaps even earlier in this thread) but I'm not sure what causes is. Anyway, you can't do a sensible Bader analysis on a CHGCAR_sum file which has the density that you see in the figure (and negative values).

[juganta]

Thanks a lot for your quick reply. Do you have any suggestion how to overcome it? To generate CHGCAR_sum file I follow as mentioned in the bader analysis website. chgsum.pl AECCAR0 AECCAR2 (fft grid was 240*240*960 which was checked for the conversed atomic charge).

You have mentioned "negative values" I did not get it.

Do you have any suggestions to resolve this problem?
Regards

[graeme]

I don't know why some charge density files are like this. By negative values I mean that the file contains negative charge densities, which is not physical.

[juganta]

Thanks for your reply.

[juganta]

[quote=graeme post_id=8695 time=1487287132 user_id=2]
I don't know why some charge density files are like this. By negative values I mean that the file contains negative charge densities, which is not physical.
[/quote]

I have checked CHGCAR_sum and the negative value is coming from core charge containing AECCAR0 file. Is their any intervention in the code can help to get well defined CHGCAR_sum file?

In my some calculations, I am getting bad CHGCAR_sum and bad isosurface like the previous one without any error (no new maxima in edge refinement) by issuing : bader CHGCAR -ref CHGCAR_sum.

Thanks,

[graeme]

Assuming you are using vasp to generate these charge density files, this is a question for the vasp team. The Bader analysis requires a physically reasonable charge density file for sensible results. So no, you can not expect the Bader analysis code to deal with charge density files that are messed up. We need to figure out why vasp is generating them and how to fix the problem.

[juganta]

Thanks for your reply.