hybrid functionals with periodic boundary conditions

[Florian Schiffmann]

Hi,

this warning is about the operator when you calculate the exchange 
potential. The normal exchange uses <ab¦1/r¦cd>. The decay of the 1/r 
operator is very slow which causes a problem in PBC. In CP2K there are 
other operators for this term which are more suitable with PBC. One is that 
the integration is only performed up to a given cutoff radius instead of 
the whole space and the other uses a HSE like approach in which (if I am 
not completely mistaken) an erf is used.
Have a look in the HF section INTERACTION_POTENTIAL. For truncation you 
will have to use the t_c_g.dat file in addtion. And maybe a long range 
correction. 

Cheers
Flo
  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150322/0ed62c0d/attachment.htm>