[Wien] Help needed in DFTD3 of layered structure

[tran at theochem.tuwien.ac.at]

Hi,

We have looked at this problem and it seems that there is a bug in the
DFTD3 package. The forces are wrong for the Becke-Johnson (bj) damping
function when they are calculated analytically. Unfortunately, this is
just what corresponds to the default choice in case.indftd3.

We will mention this problem to the authors of the DFTD3 package, but
at the moment there are several alternatives to get correct forces:

1) choose "yes" for "num" (last line of case.indftd3) such that the
forces are calculated numerically. Note that the numerical calculation
of forces for the 3-body-term abc may take some time. If this is really a
problem, then switch off the 3-body term (choose "no" for "abc").

2) choose the "zero" (instead of bj) damping function (1st line of
case.indftd3).

Thanks for having pointed out the problem.

F. Tran


> Dear users and developers,
>
> i am using DFT-D3 for calculation of structure optimization of bulk MoS2. 
> For the minimum energy structure (calculated by including dftd3) there is 
> huge a net attractive force on the S atom (-2810 mRy/bohr). When I am 
> trying to minimize the forces, the energy of system goes high making it 
> unstable. How to find the stable atomic positions?
>
> -- 
> Dileep Krishnan,
> Int. Ph. D Student,
> International Centre for Materials Science (ICMS),
> Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR),
> Jakkur, Bangalore-560064,
> INDIA.