[Pw_forum] vc-relax

[Lorenzo Paulatto]

In data 08 giugno 2009 alle ore 14:47:08, ali kazempour  
<kazempoor2000 at yahoo.com> ha scritto:

>
> Dear all
> I want to study defect in TiO2. For a supercell with 2*2*3(72 atoms)  
> size , Do I relax   the atomic position and lattice parameter  
> simultaneously (vc relax)or not?what about bigger supercell?

Dear Ali,
it depends on the system you want to simulate!
In other words, if you assume the density of defects to be small, you'll  
expect the defective supercell to be forced, by the surrounding flawless  
crystal, to have the same lattice parameter as the unperturbed TiO2. E.g.  
if you remove a single Oxygen from the middle of a big TiO2 crystal it  
does not provide enough energy to force the entire crystal to change its  
lattice parameter, there may be a local strain which you account by taking  
a big enough supercell. As periodic-boundary condition always imply a  
relatively high density of defects (one every 12 unit cells, in your case)  
you have to use the lattice parameter of unpertubed TiO2, and only relax  
the atoms positions (which will move to compensate the local strain). The  
system will actually feel a stress, but that's exactly what's happening in  
the experiment.

On the other hand, if you want to study a system with a high density of  
defects, in your case about 1 every 12 unit cells, than you have to relax  
the cell parameters together with the ionic positions.

regards


-- 
Lorenzo Paulatto
SISSA  &  DEMOCRITOS (Trieste)
phone: +39 040 3787 511
skype: paulatz
www:   http://people.sissa.it/~paulatto/

     *** save italian brains ***
  http://saveitalianbrains.wordpress.com/