Adsorption energy calculation without D3 correction

[julien_steffen]

Dear VASP community,

I calculated the adsorption energy of chemisorbed acetone on a Pt(111) surface, thus a quite straightforward job. Since the calculated adsorption energy (1.16 eV) was roughly twice as large as a value reported in literature, also calculated with VASP (https://doi.org/10.1016/j.cattod.2005.04.013), I checked in which settings my calculation differed from theirs.
Main differences are the used DFT functional (PW91 there, PBE in my calculations) and that no D3 VDW correction has been used in the paper (since its 20 years old).
I then, however, noticed that if I do not use the VDW correction as well, the adsorption energy significantly shrinks to only 0.029 eV (PBE) or 0.032 eV (PW91). The DFT functional thus seems to have only a small influence.
The structure, however, is still a chemisorbed one, with the sp3 hybridization of the central C atom clearly visible. I would expect a much larger adsorption energy (similar to the cited paper) even without VDW corrections.

I already checked my settings but I found no obvious reason why the adsorption energy without D3 correction is almost zero.
Could you maybe look over the settings and the output if I forgot some crucial setting?
The input and output files for the calculations of the clean surface, the isolated acetone as well as the adsorbate system with and without applied VDW correction are added as attachment. In the case of adsorbate calculations, the OUTPUT files were cutted such that only the last two full geometry optimization cycles are included in order to keep the size of the attachment reasonable.

Best wishes,
Julien

[michael_wolloch]

Dear Julien,

thanks for reaching out with your adsorption question. Unfortunately, it is not possible for us at the VASP company to analyze and answer questions regarding specific quality of calculations or detailed settings, due to our limited time.

I thus moved your post to the "from users for users" section and want to invite the community to discuss the problem here. We are working on adding some surface science exercises to our tutorials page, so keep a lookout for them!

A quick note I want to make anyway: For accurate total energies, we recommend static follow-up calculations to relaxations with ISMEAR=-5 .

Good luck with your question and happy computing,
Michael