[CPMD-list] molecular crystal geometries

[mkosmows]

Dear CPMD community:

I'm working with some molecular crystal systems, and find that when I optimize 
a geometry within the literature cell parameters that the optimized bonds get 
a bit long.  And when I optimize the cell, the cell shrinks, but I still get 
long bonds.

My advisor suspects this is due to geometry optimization occurring at 0K, 
while the crystal structures were obtained at a higher temperature and that 
the cell at 0K may well actually be smaller than the literature cell.  This 
still doesn't explain the longer bonds.  I think that this is the failure of 
LDA and GGA DFT methods to handle van der Waals forces properly, as described 
in Physical Review Letters, Volume 80, Number 19, Page 4153, Year 1998.

Can anyone shed some light on this subject?  And are there any methods to get 
better bond lengths?

Thank you,

Mark Kosmowski

Chemistry Department
Syracuse University
mkosmows at syr.edu