Cutoff distance question

[Yiyang]

Hello,

While reading the page demonstrating the Hartree-Fock exchange:

https://www.cp2k.org/exercises:2015_pitt:hfx

and reading the 2009 paper of Guidon 
(http://pubs.acs.org/doi/abs/10.1021/ct900494g),
I came up with a few questions:


1. How can we determine the CUTOFF_RADIUS of the truncated electrostatic 
potential?
    In [2009 Guidon] it is suggested that CUTOFF_RADIUS < 0.5*Cell_size
    guarantees no spurious self-exchange parts enter the gamma-point 
calculation.
    Does this mean we should always use 0.5*Cell_size when we are not using 
long range correction?
    Because [2009 Guidon] also reported that accurate HF energy is obtained 
more easily without long range correction.


2. For the dispersion correction, the VDW_POTENTIAL section, if we use type 
PAIR_POTENTIAL, then
    we need to specify the R_CUTOFF for the pair potential. 
    As shown in the example script (
https://www.cp2k.org/exercises:2015_pitt:hfx):
    R_CUTOFF [angstrom] 16
    so the cutoff used is 16*2 =32 A.
    However the cell size is
    ABC [angstrom] 12.42 12.42 12.42
    I am not sure why we can use the cutoff much bigger than the cell size.
    Is there any rule about setting this cutoff distance?


3. Also in the VDW_POTENTIAL section, using type PAIR_POTENTIAL we have 
option LONG_RANGE_CORRECTION.
    However, using type NON_LOCAL there is no option for long range 
correction.
    Is the correction built in the KERNAL_FILE? 
    

Thank you for your time.

Regard,
Yiyang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20161017/d01644fc/attachment.htm>