Hard time on Pt 111 slab

[Ednilsom Orestes]

Dear all,

I'm absolutely newbie on CP2K and having hard time on Pt111 slab.

After convergence of cutoff/rel_cutoff to the Pt unit cell, should the k-points be converged to the slab or... the unit cell of the bulk has to be optimized?

[Lucas Lodeiro]

In my experience, slabs need a denser k-point grid in the periodic directions to converge the surface states. So, if you need surface properties, maybe it is better to do it directly for the slab. If bulk properties are the objective, the bulk kpoint grid it is sufficient. For example in my research, for a perovskite a 3x3x2 kpoint grid was sufficient for bulk properties, but when surface energy was computed, a 4x4x1 kpoint grid was better.

Regards - Lucas

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/2e932a1c-34e2-4f48-9969-e6883ae579fbn%40googlegroups.com.

[Ednilsom Orestes]

Completely agree about k-points density of bulk vs slab!

But about the steps: 

cutoff/rel_cutoff (slab) --> k-points (slab) --> Geo_opt (slab) ?

Thanks in advance for your time and fast response!

Ednilsom