VASP Forum

Dear all,

I hope you're well.

I'm doing structure optimization of Co@monolayer MoS2 with single-S vacancy.
The Co atom replaces the central upper S atom of the 5x5x1-sized MoS2, which is obtained by optimizing the structure and getting the supercell.

But here are some troubles. The energy has serious vibration and cannot be converged.
I'm not sure if it's due to an unreasonable initial structure or an issue with the calculation settings.

The calculation files are attached below. Please check the details about setting and OUTCAR.
I'd appreciated if you could give me some advices.

Best,
Xin

Dear Xin,

This calculation is not converging. The first electronic minimization didn't converge after 120 iterations and that leads to a bad prediction for the next step in the geometry optimization.

To solve this problem you should first make sure that the first electronic minimization is reasonably converged. To do that follow this guide on the wiki.
Furthermore, considering that your initial geometry is far from the equilibrium you should try using a smaller step, for example set POTIM=0.2.

Best,
Alexey

Dear Alexey,

Thanks so much for your reply and advice!

But I'm not so sure if my initial structure is rational or not.
And I'm not quite sure about the importance of the initial guess structure.

I will have a try.
Thanks!

Best,
Xin

It is important that atoms are not placed too close too each other as this would lead to really large forces and likely to result in an unreasonable configuration. However, in your OUTCAR file I don't see any atoms that are too close. You can analyze the nearest neighbor table in OUTCAR to check that. At the same time the Co atom is not too far, so I think that it is a reasonable initial guess.

Furthermore, I see that you are using the dipole corrections and didn't not provide DIPOL. Setting the coordinate in the center of the cell should improve the convergence in this case.