VASP 5.4.4 Performing Volume + Geometry Optimisations

Question on input files/tags, interpreting output, etc.

Please check whether the answer to your question is given in the VASP online manual or has been discussed in this forum previously!

Moderators: Global Moderator, Moderator

Post Reply
Message
Author
charlie_adams
Newbie
Newbie
Posts: 2
Joined: Mon Jun 28, 2021 9:21 am

VASP 5.4.4 Performing Volume + Geometry Optimisations

#1 Post by charlie_adams » Wed Sep 15, 2021 10:12 am

I'd like to compare Mp2 adsorption energies with a few DFT functionals. I saw a forum post a few years back saying that volume relaxations aren't implemented in VASP5.3.5. Is this the case with VASP 5.4.4?

If ionic optimizations are not implemented, is it valid to take a DFT optimized geometry (PBE specifically) and perform MP2 electronic optimization, and then compare against the energies generated purely from DFT?

Thanks,
Charlie

ferenc_karsai
Global Moderator
Global Moderator
Posts: 128
Joined: Mon Nov 04, 2019 12:44 pm

Re: VASP 5.4.4 Performing Volume + Geometry Optimisations

#2 Post by ferenc_karsai » Wed Sep 15, 2021 6:29 pm

Did you mean you want to do geometry optimizations using forces and stress within MP2? That is not available. For LDA, PBE and other (semi-)local functionals the geometry optimizations (ionic and volume optimizations) are available in VASP 5.4.4 and higher.

The minimum energy PBE structure is not necessarily the minimum energy structure for MP2. In principle you would have to search the energy landscape, but the more parameters (lattice constants, angles for volume optimizations) you have the more cumbersome it gets. Usually cubic and tetragonal cells work, but everything above is too cumbersome.

charlie_adams
Newbie
Newbie
Posts: 2
Joined: Mon Jun 28, 2021 9:21 am

Re: VASP 5.4.4 Performing Volume + Geometry Optimisations

#3 Post by charlie_adams » Thu Sep 16, 2021 7:48 am

Forces and stresses yes. I will have to rethink what I plan to do. I was hoping to use the MP2 calculation as a sort of benchmark to my adsorption energies calculated with OptPBE, in the absence of experimental data. I have seen other papers benchmark against meta-GGA functionals, to justify their choice of GGA functional. However, there is no really way of knowing whether the meta-GGA functional is systematically obtaining more accurate results (which was why I was originally going for MP2). Nonetheless, I might try the HSE06 of SCAN functionals as a comparison.

Thanks for your help.

Post Reply