[CPMD-list] Surface calculations

["J. Schüle"]

Hello,
I tried to get some meaningfull surface calculations. Therefor I put 48 
atoms in  a cell and set the cell-size to in a way half the interatomic 
distances to the "next cell" according to the periodic boundary 
conditions. After a Wavefunction calculation I performed a geometry 
optimization which I expected to be lower in energy - but the optimzed 
geometry is instead as much as 20.2eV higher in energy ??
Then I optimized the wavefunction again for this optimized geometry and 
expected no great change in energy - but the energy dropped by 1.96eV.

Am I wrong, or is this method not suited at all to be compared to 
surface experiments with a resolution of 0.1eV and beyond?

Thanks for any comments,

  Josef Schüle