Invoking B97-3C in CP2K

[awer...@gmail.com]

Dear CP2K Developers and Users,

I would like to kindly ask you about the way of using B97-3C to get close to original definition.

I use the following sections along with TZVP-MOLOPT-GTH basis set, but I am not sure if I am correct:

&XC

        &XC_FUNCTIONAL  
                 &BECKE97
                    PARAMETRIZATION B97-3C
                 &END BECKE97
        &END XC_FUNCTIONAL

        &vdW_POTENTIAL
         DISPERSION_FUNCTIONAL PAIR_POTENTIAL
        &PAIR_POTENTIAL
         TYPE DFTD3(BJ)
          CALCULATE_C9_TERM .TRUE.
          REFERENCE_C9_TERM .TRUE.
          LONG_RANGE_CORRECTION .TRUE.
          PARAMETER_FILE_NAME  ./dftd3.dat
          VERBOSE_OUTPUT .TRUE.
          REFERENCE_FUNCTIONAL BECKE97
       &END PAIR_POTENTIAL
       &END vdW_POTENTIAL

...

Should I define DFTD3(BJ) with BECKE97 parametrization?

Should I use the pseudopotentials defined for PBE?

Am I doing anything wrong here? Could anybody comment on this, please?

Any additional tricks are needed here?

With best wishes,

Ana

[Thomas Kühne]

Dear Ana, 

the correct usage of B97-3c within CP2K is best described in Perlt, Ray, Hansen, Malberg, Grimme & Kirchner, J. Chem. Phys. 148, 193835 (2018). 

Even though not fully identical with the receipt of Grimme the following section with the standard TZVP basis set is the closest: 

&FORCE EVAL 

&DFT

&XC

&XC FUNCTIONAL

&BECKE97
PARAMETRIZATION B97-3c 

SCALE C 1.0
SCALE X 1.0

&END BECKE97

&END XC FUNCTIONAL 

&vdW POTENTIAL

DISPERSION FUNCTIONAL PAIR POTENTIAL 

&PAIR POTENTIAL

TYPE DFTD3(BJ)
PARAMETER FILE NAME dftd3.dat 

REFERENCE FUNCTIONAL B97-3c

R CUTOFF 7.93766
CALCULATE C9 TERM
SHORT RANGE CORRECTION

&END PAIR POTENTIAL 

&END vdW POTENTIAL

&END XC 

&END DFT

&END FORCE EVAL 

In conjunction with the MOLOPT basis sets you are using in my group we use DZVP-MOLOPT-SR as default … 

Best, 

Thomas Kühne

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/311c5e76-b7ab-4475-ad6f-c54690993fa3n%40googlegroups.com.

==============================

Thomas D. Kühne

Dynamics of Condensed Matter

Chair of Theoretical Chemistry

University of Paderborn

Warburger Str. 100

D-33098 Paderborn

Germany

thomas...@upb.de

+49/(0)5251/60-5726

[awer...@gmail.com]

Dear Thomas, Thank you so much for your invaluable help. I do confirm it does work properly. The quoted paper is brilliant. Thank you! Ana

[flytomoo...@gmail.com]

Dear Prof. Kühne,

May I ask that if your group members specially optimize the GTH pseudopotentials for the B97-D functional (for the usage of B97-3c), or directly use the existing GTH pseudopotentials optimized for other functionals (e.g., BLYP, BP, or PBE)? Thanks in advance !

Sincerely,
Li

[Thomas Kühne]

Dear Li, 

yes we are typically using the BLYP GTH pseudopotentials. 

Greetings, 

Thomas Kühne

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/df4bc97b-f5a6-49ee-b097-d46bd623d3cbn%40googlegroups.com.

[Daniel Count]

Has the above input by Thomas K. ever been validated, e.g. for the water dimer binding energy, against the results of the reference implementation of "B97-3c" by Grimme provided in QC codes like ORCA?

[Thomas Kühne]

In the paper by Perlt, Ray, Hansen, Malberg, Grimme & Kirchner, J. Chem. Phys. 148, 193835 (2018), 

the implementation of B97-3c was compared to ORCA, since one the corrections is recommended to 

be used in conjunction with the mTZVP that differs from the ones employed in CP2K. 

The summary of the comparison, which you will find in the SI of the aforementioned paper, reads as: 

"While the mean signed deviation is similar for both basis sets (ORCA: 0.06 kcal mol−1, 

CP2K: −0.06 kcal mol−1), the mean absolute deviation is significantly larger if the 

TZVP-GTH basis set in CP2K is used (ORCA: 0.28 kcal mol−1, CP2K: 0.44 kcal mol−1)."

Greetings, Thomas Kühne  

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/6bed0cc3-08f7-4724-8e54-e6d73da8abc5n%40googlegroups.com.