[gmx-users] Regargding simulation of periodic silica surface

[Mirco Wahab]

On 12.09.2014 13:30, Sathish Kumar wrote:
> Yes i created a block of SiO2 according to the length of the lipid
> membrane and kept in the center of the box and i kept the distance from
> edges of the box is 2 nm. Here i have not used the periodic-molecules
> option in the mdp file.  I hydroxylated the surface of SiO2. The below
> force field parameters i have used in the simulation.
>
> atom types
> Si     28.0855  ; Si
>   BO     15.999400 ; O
>   SO     15.999400 ; O
>   HS     1.00800 ;   H
>
> bond types
>
>    Si   BO       1    0.1680    125500.00;
>    Si   SO       1    0.1680    125500.00;
>    SO   HS       1    0.0945    207100.00;
>
> Non bonded terms
>
> Si      14      28.085   1.10000  A  0.000000E+00   0.0000E+00
> BO       8      15.9994 -0.55000  A  0.2700E+00     1.91107E+00
> SO       8      15.9994 -0.67500  A  0.3000E+00     1.91107E+00
> HS       1       1.008   0.4000   A  0.00000E+00    0.00000E+00
>
> And here i attached the md.log file and em.gro file. Please once check.

The SiO2 block in your attached file is periodic in x/y.

According to your coordinate file, you additionally have 4 "CL" Atoms
dispersed in the SOL water. Is this for charge compensation? Then, there
are about 6 to 8 SOL waters placed *within* the SiO2 crystal, which
could give rise to high force gradients due to overlaps (only guessed).

The values and esp. charges in your "non-bonded" atom parameter list
are somehow non-standard. Which force field do you use with these
parameters?

Furthermore, you have irregular patterns of atom species in your
  .gro coordinate file starting from atom 8782. This should be mirrored
in your topology file. Your "bond types" would imply you have one
large SiO2 super-molecule - whose topology and connectivity should
then be provided to gromacs. Can you post your topology .top/.itp
file that describes *this* structure?

Regards

Mirco