About getting force constant in VASP.

Queries about input and output files, running specific calculations, etc.


Moderators: Global Moderator, Moderator

Locked
Message
Author
hszhao.cn@gmail.com
Full Member
Full Member
Posts: 191
Joined: Tue Oct 13, 2020 11:32 pm

About getting force constant in VASP.

#1 Post by hszhao.cn@gmail.com » Sat Sep 02, 2023 12:31 pm

I found that the force constant matrix in OUTCAR differed by a negative sign from the force constant matrix extracted from vasprun.xml. Any hints for this will be helpful. See here for the related discussion.

Regards,
Zhao

ferenc_karsai
Global Moderator
Global Moderator
Posts: 460
Joined: Mon Nov 04, 2019 12:44 pm

Re: About getting force constant in VASP.

#2 Post by ferenc_karsai » Sun Sep 03, 2023 4:54 pm

Thank you very much for informing us. I will forward this to the persons who work on this or related topics.

manuel_engel1
Global Moderator
Global Moderator
Posts: 126
Joined: Mon May 08, 2023 4:08 pm

Re: About getting force constant in VASP.

#3 Post by manuel_engel1 » Tue Sep 05, 2023 10:06 am

Hi,

Sorry for the confusion. The negative sign is due to compatibility reasons only, i.e., to preserve backwards compatibility with earlier versions of VASP. In VASP, the force constants are calculated directly from the forces, but a negative sign is not added at this step. Therefore, the output of Hessian and force-constant matrix is generally with a negative sign as compared to standard literature.

However, there is no output of the force constants to the OUTCAR file in recent versions of VASP. Therefore, I'm not 100% certain if what you mean is related to this minus sign. Could you please specify your VASP version and clarify which OUTCAR section you are talking about?

EDIT: There is still OUTCAR output for certain parts of the code. Could you please also upload your INCAR and OUTCAR files so I know which part of the code produces the output.
Manuel
VASP developer

hszhao.cn@gmail.com
Full Member
Full Member
Posts: 191
Joined: Tue Oct 13, 2020 11:32 pm

Re: About getting force constant in VASP.

#4 Post by hszhao.cn@gmail.com » Thu Sep 07, 2023 11:18 am

Thank you for your comment and explanation. In fact, currently, I'm not doing any concise computation related to this problem, instead, I completely restated based on the original description on phonopy-users@lists.sourceforge.net, as shown below:
[Phonopy-users] Questions about getting force constant in VASP
Dear developer,

When I used VASP's DFPT to calculate phonons, I found that the force constant matrix in OUTCAR differed by a negative sign from the force constant matrix extracted from vasprun.xml by phonopy. By the source code, it seems that Phonopy artificially added a negative sign when calculating the force constant through the Hessian matrix in vasprun.xml. I wonder about the meaning of this negative sign, whether it means that the Heissen matrix output by VASP differs by one sign from the real force constant. Hope for help from you. Thanks a lot.


Best wishes,

Longju
If I come across specific relevant examples in the future, we can continue the discussion.

Regards,
Zhao

Locked