Choosing parameters in CP2K

huan... at huan... at
Tue Oct 1 14:43:48 CEST 2013

Dear All,
I am a newbie to CP2K. I have a project on benzene-water cluster and I 
would like to run MD simulation using DFT PBE functional. 

I have some questions on choosing basis set and parameters for MD 
simulation. However, after I checked the examples or CP2K reference, I did 
not get any answers.

In the example file of /opt/cp2k/test/QS/regtest-dft-vdw-corr/dftd3_t1.inp, 
it used two different basis set for one system, like,
    &KIND H
    &KIND C

Is that OK? I mean if there is an argon atom in my cluster, as basis set 
not always including all the atoms after the 3rd period, can I choose 
different basis set for calculation?

In the section of  DISPERSION_FUNCTIONAL, I saw another example like,
            TYPE DFTD3
            PARAMETER_FILE_NAME ../dftd3.dat
            CALCULATE_C9_TERM T
            R_CUTOFF 7.

the R_CUTOFF parameter is set to 7. The description in the CP2K reference 
is too brief to understand.
*       R_CUTOFF 24.0*
*       Range of potential. The cutoff will be 2 times this value  [Edit]*
*       This optional keyword cannot be repeated and it expects precisely 
one real.*
*       Default value: 1.05835442E+01*
*       Default unit: [angstrom]*
Is there any rules to set the value of R_CUTOFF? Is there any relationships 
with what kinds of potential I chose?  Or for what kind of system 
(contained what kinds of atoms), the R_CUTOFF value should be increased or 

Another question is what's the difference between DFTD3 and DFTD2? Which 
one is more suitable for the weak interaction such like pi-hydrogen bond? I 
did not find any description in the CP2K input reference.
*    TYPE DFTD2*
*   Type of potential:*
*     DFTD2 → Grimme D2 method*
*     DFTD3 → Grimme D3 method*
*   [Edit]*
*   This optional keyword cannot be repeated and it expects precisely one 
*   Default value: DFTD3*
*   List of valid keywords:*
*   DFTD2*
*    Without description, yet.  [Edit]*
*   DFTD3*
*    Without description, yet.  [Edit]*

I appreciated it very much for any of you can give me explanations or 
suggestions on choosing those parameters. Thank you.

Best regards,

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the CP2K-user mailing list