Pt(111) surface --- Oxygen molecule geometry optimization

[neogien]

Hello evryone.

I had posted some questions on the geometry optimization of Pt(111) surface and its energy calculation.

Now, I'm doing some practices on the platinum-oxygen molecule adsorption at Top-Bridge-Top position.

I used several settings(different INCAR files) with various POTIM values, different algorithms, smearing factors, and positions of oxygen molecule.

However, in all cases, optimized structure were same and they were slightly different with reported results (for example, Hafner's famous result published on the PRL 1997).

The axis of molecular oxygen was slightly rotated from T-B-T line (Hafner reported that they were parallel with each other) about 15 degree.

I'm wondering to believe my reslt as itself because it is so different with reported ones.

I want to ask to you is there any physical or calculation factors I missed or should consider more carefully?

If you need more informations on my calculation please see below (INCAR file).


System = Pt 3x3x4
ENMAX = 400

ISMEAR = 1
SIGMA = 0.15
GGA = RP
ISPIN = 2

NSW = 300
IBRION = 1
ISIF = 2

ALGO = fast
LREAL = A
NPAR = 8
POTIM = 0.2

NSIM = 4
LPLANE = .TRUE
LSCalu = .FALSE


------------------------------------------------

I also have another question.

I prepared 4-layer Pt(111) slab and fully optimized the upper 2 layers and oxygen molecule placed on the TBT site.
I also calculated total energy of Pt-O adsorbed system and free O2 molecule.

For oxygen-Pt binding energy calculation, which slab structure should I choose for reference Pt slab energy calculation?

1) Erase the oxygen molecule from the fully optimized Pt-O2 structure. Fix all positions and calculate total energy.

2) Choose 4-layer bare Pt slab which structure optimization was not performed.

Please don't hesitate to give me your advices.
Any kind of your comments will be very helpful to me.

<span class='smallblacktext'>[ Edited ]</span>

[aarondesk]

Have you verified that all simulation parameters are sufficiently converged? 400 eV seems like it might be a low cut-off energy for O.

Also, symmetry issues may be at play. Have you tried turning off (or on) symmetry and putting the O2 parallel and seeing what happens?

Lastly, the correct reference state should be the relaxed 4-layer slab with no O2 present.