Possible bug in DFT-D3

[Leopold Talirz]

Dear Flo,

thanks a lot for your reply and the link. Nice table with the cutoff values 
for the elements by the way :)

Concerning our question - I completely agree with your point, where it 
concerns quantities that are computed on the grids. Our question, however, 
concerns only the *dispersion energy*. As far as I know (please correct me, 
if I am wrong) both DFT-D2 and DFT-D3 compute the dispersion energy based 
only on atomic coordinates. 
If the interaction between the periodic replicas of the slab across the 
vacuum region are set to zero by R_CUTOFF, then the dispersion energy 
should be completely independent of the amount of vacuum. If you check the 
values for the DFT-D2, you will find that the dispersion energies for the 
two different amounts of vacuum are identical up to the last digit.

Best,
Leo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20141006/34b0fac1/attachment.htm>