[Pw_forum] Huge dispersion energy

[Ari P Seitsonen]

Dear venkataramana,

   I would guess that the problem is that you have extracted the 
"dispersion energy" - please be aware that this can only be "defined" in 
the case of DFT-D2, not with vdW-density fucntionals, as there also the 
density is changed etc - for the whole system (that includes also the 
dispersion between the atoms of the substrate and intra-molecular 
contributions of methane) whereas Henkelman et alia only consider the 
difference in the binding, or adsorption, energy. Please calculate the 
adsorption energies and see if then the value makes more sense.

     Greetings from Zurich,

        apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
   Ecole Normale Supérieure (ENS), Département de Chimie, Paris
   Mobile (F) : +33 789 37 24 25    (CH) : +41 79 71 90 935


On Sat, 10 Oct 2015, Venkataramana Imandi wrote:

> Dear QE users,
> 
> The geometry of methane molecule on the top of Iridium slab is relaxed.
> In that relaxed geometry, methane molecule is sitting 4.32 angstroms on the top of Iridium atom, which is exactly same as Henkelman results (prl, 86,664-667,2001).
> The total energy of relaxed geometry is -5861698.6 kJ/mol. (val1)
> I took that relaxed geometry coordinates and then I applied dispersion interactions of that configuration only. (input file is attached, Dispersion interactions with
> non-local functional)
> The total energy with dispersion interactions is -5893444.5 kJ/mol. (val2)
> The difference of val2 and val1 is -31745.9 kJ/mol.
> Moreover, I used grimme-d2 dispersion interactions (commenting out input_dft='vdw-df') , then total energy is -5876733.8 kJ/mol. (val3)
> The difference of val3 and val1 is -15035.2 kJ/mol.
> 
> I want to reproduce the results of Henkelman. In these results the total energy difference between with dispersion and without dispersion is roughly 10 kJ/mol.
> 
> I am unable to find out where the mistake is. are input parameters correct in the input file (input file is attached).
> Thanks in advance for any suggestions and help.
> 
> 
> ...
>  
> venkataramana
> PhD student
> IIT Bombay
>