2-D/Thin film/Slab Convergence and Calculations

[rishikanta_m]

Hi

I am really interested in calculating the opto-electronic properties of thin film.
The procedure I followed is:
1. Optimize the bulk.
2. Create thin film along a preferred plane.
3. Include vacuum and optimize vacuum thickness layer by convergence w.r.t. total energy.
4. Check Convergence of K point sampling.
5. Increase atomic layers and optimize the no. of layers by convergence w.r.t. total energy/atom.
6. Relaxation of surface atoms.
7. Calculation of properties.

Q1. Is my procedure correct?

Q2. I started calculations. All steps till step 4, its alright, I could achieve convergence criteria of 1 meV. But, in step 5, I got the following values.
No. of layers Total energy/ atom
2 -4.6519
3 -4.7378
4 -4.7839
5 -4.7487
6 -4.8302
7 -4.8432
8 -4.8530
Should I keep on increasing the number of layers, or is the convergence already achieved (on increasing no. of layers from 6 to 7, the total energy/atom differs by 0.01 eV) ?

[andreas.singraber]

Hello!

Sorry for the late reply! To be honest, I have little experience with slab calculations.. nevertheless, I think your overall procedure is correct but there is one important note: I believe in steps 3 and 5 you are not considering the correct quantity to converge. Instead of the total energy per atom you should rather converge the surface energy, see e.g. this definition (the surface area can be neglected, if always the same). If you compute this and create corresponding plots you should see convergence (maybe accompanied by oscillatory behavior). When to stop and accept the residual error is a question one cannot answer in general because it depends on your desired accuracy and the actual system you are studying. Can you please analyze the surface energy convergence and show the plots? Thank you!

All the best,

Andreas Singraber

[rishikanta_m]

Thank you for your reply. I will try out your suggestions.

[rishikanta_m]

Hi again,

In the optimization of thin films, I got the fact that only ionic relaxations should be done.
However, I have seen three different types of relaxation.

1. Relaxation of all atoms in the slab.
2. Relaxation of outermost layers of atoms (both two sides e.g. upper and lower ) while fixing inner atoms located near the center
3. Relaxation of only the upper layers of atoms at the top and fixing the atoms at the middle and lower layers of atoms.

Which one is the most reasonable for the calculation of electronic and optical properties?

[henrique_miranda]

I would think option 3 is the most reasonable but you should probably look in the literature and of course do your own testing.