The ABINIT Users Mailing List ( CLOSED )

[Steven Homolya <steven.homolya@spme.monash.edu.au>]

zeng zhenhua wrote:
> The reason I want to fix the bottom three layers is that I want to make 
> these  layers like bulk layres to the top three layers.Then the top 
> three layeres exhibit characteristics of a real surface layers. But if I 
> have full  relaxation ,the top thre layers will lose real surface layers 
> condition. So i have one surface relaxation. Is there has other methods? 
> your suggestion will be greatly appreciated!

This method wil not give you a "real" surface layer. If there is
significant relaxation at the centre of your slab (when you don't fix
any layers) then your slab is not thick enough, i.e., for a "real"
surface you can probably expect to see significant strain more than
three layers from the surface.

You do need to fix all lattice constants though. Those parallel to the
surface should be fixed at their bulk values (as obtained from a bulk
calculation).

Steve