Adsorption simulations

[fsimeski]

Hi VASP users!

I want to set up some simulations to study adsorption of a gas molecule on a solid substrate. Is there a VASP command that I can use to incrementally vary distance between the gas molecule and the surface?

Previously, I ran ten separate single-point energy calculations where I had used a bash script to vary the distance between the gas molecule and the substrate. These calculations produced results as I expected, but there is noise and oscillations in the energy levels, instead of a smoothly varying curve. I am looking for a way to remove these unphysical oscillations.

Thank you,
Filip

[fabien_tran1]

Hi,

Using a script to run a series of calculations is the convenient method to use. Where do these oscillations come from is another thing, but without knowing more details it is difficult to give specific advice. Oscillations can be due to various reasons like parameters (e.g., EDIFF or ENCUT) which are not set properly. Also, which properties are you looking at and what do you mean exactly with "energy levels"? Have you also considered methods related to molecular dynamics?

[fsimeski]

Hi Fabien,

Here is my INCAR file:

Code: Select all

SYSTEM = Quartz adsorption calculation

# start parameters for this run
ISTART = 0
ICHARG = 2

# electronic optimization
ALGO = Normal
IVDW = 11         # DFT-D3 correction
NELM = 60
EDIFF = 1E-06
ENCUT = 400       # from POTCAR file
AMIN = 0.01       # recommended setting for long systems

SIGMA = 0.05
ISMEAR = 0

EDIFFG = -1E-02
IBRION = 2
POTIM = 0.5

KPAR = 8
NCORE = 8
LREAL = A

Do you notice anything out of place here? Any red flags about what you would expect from typical adsorption calculations?

I am simply looking at binding energies to confirm the physisorption that I observed through MD simulations in LAMMPS is the expected behavior and not an artifact of the model. To calculate the binding energy, I use:

E_b = E_system - E_slab - E_mol

where E_system is the energy of the system with the slab and the adsorbed molecule, E_slab is the energy of the slab, and E_mol is the energy of the molecule. These are single molecule adsorption simulations. The reason I called them energy levels is because I am plotting the binding energy vs. distance. The oscillations that I report are shown in the plot below:

adsorption_quartz.jpg

Please let me know if you have any insight on where these artifacts may originate from. Thank you very much for your help.

Happy Holidays,
Filip

[fabien_tran1]

Dear Filip,

Without trying it is a bit difficult to make a specific comment about your chosen paramaters. Have you already tested the influence of EDIFF and EDIFFG? ENCUT may also be of importance. Actually, ICONST to monitor the geometry may be helpful.

Happy Holidays, too.