[Wien] bulk-like slab

[Stefaan Cottenier]

>         In order to simulate the interface between two bulk-like slabs, we
must ensure the presence of a bulk-like slab. So I want to know is there a
criterion for  the presence of a bulk-like slab?

It depends on the property you are looking at. For very local properties (a
hyperfine field, for instance), 5 to 7 layers are enough. More extended
properties (a strain field near an interface) you need more. In my
experience, you should use at least 7 layers (for 5 layers usually there are
clear signs of communication between the two surfaces/interfaces of the
slab). A possible strategy: calculate the property you are interested in in
a test system with 5,7,9,... layers, and see where it saturates.

Stefaan