[gmx-users] Re:Re: Re: water genbox 3x3x0.8 EM not successful T not high

[Mark Abraham]

On 12/17/10, gromacs  <ptf1242 at 163.com> wrote:
> Yes. I should first run the 3x3x0.8 water to make the water 
> 
> 
> in equilibrium.
> 
> 

No. Any box dimension below about 2nm is nonsense with a normal force field.

You want a film in vacuum, so use a box that is large enough to contain both the film and enough vacuum that the films cannot feel each other. So your box dimension in the direction normal to the film must be at least the thickness of the film plus twice rcoulomb.

NVT will also be important for equilibration, per Vitaly's comments.

You may also need to use a very short MD timestep - films are intrinsically less stable than bulk solvent, and the "normal" values are all for bulk, condensed media.

Mark


> 
> And then i use editconf to put the water at the centre of 
> 
> 
> 3x3x9 box, so the film was 0.8 nm, the Z dimension is 9 nm, so there will be film.
> 
> 
> I want to simulate the rupture of the thickness of water film. 
> 
> 
>  
> 
> 
> If i directly put the  3x3x0.8 water in the 3x3x9 box, i will try to see the result.
> 
> 
>  
> 
> 
> I
> 
> 
>  
> 
> 
> 
> 
> 
> At 2010-12-17 13:12:43，"Vitaly Chaban" <vvchaban at gmail.com> wrote:
> 
> >I still do not understand what physical system you want to create in
> >the MD system or what phenomenon you want to reproduce.
> >
> >If you need a water film, then the only solution is to follow my
> >former suggestion.
> >
> >In no case should you use "3x3x0.8 NPT" MD box! If somebody suggested
> >you to do such thing, then this person simply does not understand
> >classical MD simulations.
> >
> >--
> >Dr.Vitaly V. Chaban
> >
> >
> >
> >
> >
> >
> >On Thu, Dec 16, 2010 at 11:48 PM, gromacs <ptf1242 at 163.com> wrote:
> >> OK. I will use your method. But i am afraid there will be something wrong. Because generally, i should run the 3x3x0.8 NPT water, and then i use editconf to add the water to 3x3x3. If i donot run 3x3x0.8, i may not be able to run the final interface box. Because the water in the box is not at proper situation or position. Anyway, i will have a try first.
> >>
> >> If i have problem or success, i will give an answer to all.
> >> At 2010-12-17 09:51:57，"Vitaly Chaban" <vvchaban at gmail.com> wrote:
> >>>>>Hey, gromacs -
> >>>>>
> >>>>>Nice to hear from you here... It is a bad idea to simulate something
> >>>>>with any cartesian dimension lower than a couple of nanometers using
> >>>>>classical FFs.  All your problems are generated by 0.8 nm of the
> >>>>>z-side.
> >>>>>
> >>>>>--
> >>>>>Dr. Vitaly V. Chaban
> >>>>>Rochester, U.S.A.
> >>>>>
> >>>>
> >>>> Thanks very much. I use the cutoff=0.38, because the cutoff should be smaller than the half box size. I guess it may be the Z dimension is too small, Plus the PBC. But when i use 1.8x1.8x1.8nm bulk water, all results are OK.
> >>>>
> >>>> My teacher wants me to simulate 3x3x0.8, 3x3x0.9 nm water film's rupture. So i have to run bulk >3x3x0.8 water and then add box to 3x3x9 to form interface.
> >>>>
> >>>> So if i want to run run bulk 3x3x0.8 water, how could i do? which FF should i use? the high T may be >due to EM not success??
> >>>
> >>>
> >>>I believe you use the term "bulk water" in another sense as compared
> >>>to all other guys here. When people say "bulk liquid" they imply the
> >>>endless ocean of this liquid. It is not important what particular side
> >>>lengths you have, since due to PBC you get infinitively large system.
> >>>
> >>>I suspect your teacher wants you to obtain a film of water that would
> >>>be about 3 molecular layers wide. That's easy. Take your current box
> >>>of water and apply editconf -
> >>>
> >>>editconf -f <current_conf> -o <water_film_conf> -box 3 3 3
> >>>
> >>>Finally, just start MD run with a water_film_conf. You'll have the
> >>>same periodic box , but water molecules will be present only at its
> >>>center surrounded by vacuum in one (Z) direction.
> >>>
> >>>Happy GROMACSing,
> >>>
> >>>Dr. Vitaly V. Chaban
> >>>Rochester, U.S.A.
> 
> 
> 
> 
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20101217/9debd5fa/attachment.html>