[Pw_forum] Relaxation of impurity in bulk materials

[giuseppe.mattioli at mlib.ism.cnr.it]

Dear Izaak
You (often) suppose to investigate an "isolated" dopant or defect, but  
in fact you are always simulating an array (a superlattice) of  
impurities in a semiconductor (SiC) host. In principle you should  
verify that your results are reasonably insensitive to the supercell  
volume. In such a framework you can quite safely fix the equilibrium  
lattice parameters of the 2x2x2 bulk SiC and relax all the atoms, but  
if you want to study modifications to the lattice parameters induced  
by defects, which may depend on defect concentration, codoping, etc.  
You may be interested in some details: I've often investigated defects  
in zincblend GaAs by using cubic 64-atom supercells (and a 4x4x4  
k-points mesh) at fixed lattice parameters. A 64-atoms simple cubic  
SiC supercell may represent a good starting point to study defects  
(point defects...) in SiC. You may want to read the following Zunger's  
paper, which (as usual...:-)) addresses some of your questions better  
than me.

S. Lany and A. Zunger Phys. Rev. B 78, 235104 (2008)

HTH

Giuseppe> Hello everyone,
>
> I am trying to study impurities in SiC bulk material. For this, I use a big
> unitcell of 2x2x2. I do not understand how to do a relaxation in this system,
> ie, which atoms should be kept fixed? I was thinking that the   
> unitcell boundary
> atoms would automatically be kept fixed but that does not seem to be  
>  the case.
> Any help will be much appreciated.
>
> --
> Izaak Williamson
> Research Assistant
> Physics Department
> Boise State University

GiuseppeMattioli
ISM-CNR
Italy