[gmx-users] Re: Water spreading on graphene!

[David van der Spoel]

On 2013-04-21 12:39, Dr. Vitaly Chaban wrote:
> Yes, the water droplet should remain a droplet on graphene.
>
> Since you have a droplet, I assume you have much vacuum in the system. I
> would not use PME for such system, but would rather "compensate" it by
> using larger cutoffs.
>
It could be an artifact if the box dimension is small compared to the 
droplet (less than double the droplet size, make it at least three times 
the droplet). In principle PME should work though. Check the energy 
before and after the droplet spreads to see whether the energy goes down.


> If you need a simple advice, it is -- enlarge the box to avoid any
> interactions, which you would not have had in real experiment.
Indeed.

>
> I also expect a specific behavior, if your droplet is very small.
>
> Maybe, I would simulate this system without PBC at all.
>
>
> Dr. Vitaly Chaban
>
>
>
>
>
>
>
>> I have a graphene like surface (carbons on a hexagonal lattice with zero
>> partial charges and some LJ parameters) with a drop of water placed on it.
>> In principle, this is a hydrophobic surface and the drop should remain a
>> droplet on it. However, surprisingly I am seeing that the drop is spreading
>> within a 1 ns of the simulation. In the past I have done some similar
>> simulations (with a different structure but basically zero partial charge
>> surface and LJ) on which the drop remains a drop.
>>
>> Does any body have any ideas of what I may be doing wrong? Pasted below is
>> my mdp and top file. Briefly, I am doing NVT simulation, PME for
>> electrostatics, TIP3P water model, the surface is frozen and intra-surface
>> interactions are excluded, the cut-off distances are 1 nm, velocity rescale
>> thermostat (I tried both with the entire system coupled to the same
>> thermostat and with sheet and water being coupled to different
>> thermostats).
>> Any suggestions are welcome.
>>
>> My analysis suggests this comes from the long range electrostatics because
>> we have tested the same thing in LAMMPS. When we turn off PPPM and use
>> cut-off based Coulomb (~5 nm) in LAMMPS (yes now we have changed the
>> software) we don't see the drop spreading.
>>
>> I have done so many simulations and to have this problem stumps me!! Thanks
>> for being helpful always.
>>


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se