Choosing parameters in CP2K

[huan...@mail.huji.ac.il]

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

      BASIS_SET DZV-GTH

      POTENTIAL GTH-PBE-q1

    &END

    &KIND C

      BASIS_SET DZVP-GTH

      POTENTIAL GTH-PBE-q4

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,

      &vdW_POTENTIAL

         DISPERSION_FUNCTIONAL PAIR_POTENTIAL

         &PAIR_POTENTIAL

            TYPE DFTD3

            PARAMETER_FILE_NAME ../dftd3.dat

            REFERENCE_FUNCTIONAL PBE

            CALCULATE_C9_TERM T

            R_CUTOFF 7.

         &END PAIR_POTENTIAL

      &END vdW_POTENTIAL

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 reduced? 

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 keyword.

   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,

Huan

[Andreas Funk]

Dear Huan,

I can give you reply to your second question. You should use the D3 method. It is less empirical than the D2 method and improved some other issues, i.e. it takes the molecular geometry better into consideration. My own experience is that the D3 method gives more reliable results with weak interactions. I suggest you also compute the C9 term. Maybe you should also read the paper of Grimme: J. Chem. Phys. 132, 154104 (2010); doi: 10.1063/1.3382344. It explains the D3 method in detail.

Best regards,
Andreas

[huan...@mail.huji.ac.il]

Dear Andreas,

Thanks very much for your quick response and suggestion. I will read the paper you recommended.

Best wishes,

Huan

[simin pahlavi]

Hello everybody,

I have same the problem, " Is there any rules to set the value of R_CUTOFF?"Could we set it to  1/2 (2.5 sigma) in Lenard johns interactions?

Thank you in advance