[Pw_forum] Constraints

m.g.p.krishna at chem.leidenuniv.nl m.g.p.krishna at chem.leidenuniv.nl
Tue Jun 14 04:49:13 CEST 2011

yuva: there are some quick answer...

you can use FLAGS in atoms to relax/not relax in a particular  
Cartesian direction.
Al     11.039040    5.580000   24.296861   0 0 1,
means that Aluminium is fixed EXCEPT in the Z direction. So it will  
relax on Z direction. In your case I suggest to use this kind of  
relaxation for every atom with a fixed Constraint on bond lengths (see  
previous posts of mine).

KPoints: I, generally, use 'automatic' K points which is a kind of  
Monkhorst-Pack grid (more genereic). Suppose you have orthorhombic  
cell (a,b,c) where the magnitudes are (a=b < c), then I will use more  
Kpoints in a and b direction and less kpoints in c direction. Since  
Kpoints are defined in the reciprocal space (say, a'b'c' where  
magnitudes are a'b' > c'), you should use more k points to span the  
necessery irrep. areas in that reciproal space of a'b'c'.
simply You can start with (see manual)
  K_POINTS automatic
   nx ny nz shiftx shifty shifty

It is a good idea to visualize your supercell prior to put some values  
in nx,...

By the way why you are using a BASH script? It was in QE package just  
to make tests quite smoothly/interactively. So pls use a simple *.in  
file which contains the input-data BETWEEN cat > alpen148.rx.in << EOF  
to EOF (bottom, near to Kpoints...)

good luck

Krishna Mohan GP
Center for Superfunctional Materials
Sotuh Korea.

More information about the users mailing list