Fwd: [gmx-users] Re: Frozen N2 in a very long tube

[Vitaly Chaban]

Shuai:

In addition, this is generally a bad idea to fix anything in your
system unless you really need it to be fixed. In your system, I would
not have fixed the nanotube in such a way. You may also want to
consider the center-of-mass motion removal keywords. Please see the
keywords below.

comm-mode
nstcomm
comm-grps


Vitaly



---------- Forwarded message ----------
From: Vitaly Chaban <vvchaban at gmail.com>
Date: Sun, Jul 11, 2010 at 9:05 PM
Subject: Re: [gmx-users] Re: Frozen N2 in a very long tube
To: Shuai Ban <banshuai at gmail.com>
Cc: gmx-users at gromacs.org


1. You may use any thermostat you like.

2. The particle fixation procedure and the temperature coupling
procedure are located in the different branches of the code and are
independent. When one procedure applies thermostat, it does not know
that another kills all the velocities of the nanotube carbons. In
fact, the completely fixed particle will have a temperature of ZERO,
but formally you write two separate groups to couple anyway. Your
"ref_t" for the CNT is insignificant here.

Good luck!

Vitaly


On Sun, Jul 11, 2010 at 8:39 PM, Shuai Ban <banshuai at gmail.com> wrote:
> Dear all,
>
> Many thanks for your answers. First, the nitrogen is model by 2 LJ bead with
> negative partial charges and one dummy bead in the center to keep the
> neutrality.
>
> One question to the temperature coupling. Even though the carbon nanotube is
> fixed (all atoms do not move at all) in the simulation box, do you think it
> is still possible to couple it to temperature? And is it possible to use
> noose-hover coupling instead of V-rescale? Thank in advance.
>
> shuai
>
>
> 于 7/10/2010 11:25 下午, Vitaly Chaban 写道:
>>
>> Dear shuai:
>>
>> CNT is also to be coupled in spite it is rigid. Otherwise, the program
>> will give you an error. Please use something like:
>>
>> Tcoupl   = V-rescale
>> tc-grps  = CNT N2
>> tau_t      = 0.1  0.1
>> ref_t       = 298  298
>>
>> Cheers,
>> Dr. Vitaly Chaban
>>
>>
>> On Sun, Jul 11, 2010 at 12:53 AM, shuai shuai<banshuai at gmail.com>  wrote:
>>
>>>
>>> Dear all,
>>>
>>> thanks for your suggestions. During this NVT simulation, I keep
>>> nanotube rigid. So the temperature should only couple to Nitrogen
>>> instead of the whole system as my original input file. I will try to
>>> couple temperature only to nitrogen to see whether this is the reason.
>>> Many thanks again, especially Dr. Vitaly Chaban.
>>>
>>> shuai
>>>
>>> 2010/7/10 Vitaly Chaban<vvchaban at gmail.com>:
>>>
>>>>
>>>> Dear Shuai:
>>>>
>>>> Please try to couple temperature for nanotube and nitrogen separately.
>>>>
>>>> BTW, what is your partial charge and how did you construct it?
>>>>
>>>> Good luck!
>>>>
>>>> Dr. Vitaly Chaban
>>>>
>>>>
>>>>
>>>>
>>>>>
>>>>> Recently I met a strong error from gromacs. Therefore I would like to
>>>>> know
>>>>> whether you have experienced the same thing or what is the reason for
>>>>> that.
>>>>> The system I simulated is nvt MD for N2 (with partial charge) diffusion
>>>>> in a
>>>>> long tube (100x2.5x2.5 nm^3) at 300k. I get molecules freezed in a
>>>>> small cluster after a few ns. The weird thing is that I do not have
>>>>> such problem for either longer domain or short domain. �The input
>>>>> parameters are:
>