[gpaw-users] : gpaw.KohnShamConvergenceError: Did not converge!

[Mathias Ljungberg]

Hi Yuelin,

I assume that it is a solid and not a cluster in gas phase you want to  
calculate? In that case you need periodic boundary conditions.  For a  
small unit cell you generally need to sample the Brillouin zone (=more  
kpoints) to get meaningful results; equivalently you might duplicate  
the cell. To duplicate the cell twice in each direction is the same as  
having 2x2x2 kpoints. If you have a core hole you want to duplicate  
the cell to minimize the interactions of the core holes. Do not change  
the cell, instead write

slab *= (2,2,2)

to create a supercell

Mathias



On Dec 3, 2010, at 12:32 AM, Yuelin Li wrote:

> <CMO.out>