[gmx-users] pure methane condensation in NVT simulation

[Alex]

It isn't possible for two identical setups to produce two different 
results. That is, if they are actually identical, i.e. what Justin asked 
earlier. You are only sharing a small portion of your input.

It also isn't clear what "normal box" and "the result will be correct as 
we think" mean, i.e. whether anything has been tested already. If the 
forcefield is known to reproduce methane's properties, one needs to make 
sure the inputs to Gromacs are correct, and that includes the 
forcefield, the cutoffs, the timing constants for anything dynamic, etc, 
etc. I am unable to understand how the simulation box size was selected 
with respect to how many molecules are in it at the target temperature. 
Have you looked at the equation of state for methane at the given 
conditions?

All those things can lead to condensation, seeming phase separation, and 
a number of other effects that are very unlikely to have anything to do 
with Gromacs specifically. From personal experience, small organics are 
a pretty serious topic in terms of setup, especially when you are using 
united-atom forcefields that promote local electrical neutrality, etc.

Alex


On 5/8/2018 4:34 AM, Ming Geng wrote:
> Thanks Alex,
>      "common NVT" I mean is the simulation is only using the T coupling and
> no P coupling. The T P coupling part was set as following
>
> *    ;; TEMPERATURE COUPLING*
> *     ; Temperature coupling methods *
> *     Tcoupl                   = nose-hoover*
> *     ; frequency for coupling temperature*
> *     nsttcouple               = -1*
> *     ; Groups to couple separately*
> *     tc-grps                  = SYSTEM*
> *     ; Time constant (ps)*
> *     tau_t                    =  2*
> *     ; reference temperature (k)*
> *      ref_t                    = 400      *
> *   ;; PRESSURE COUPLING     *
> *   ; pressure coupling methods*
> *   pcoupl                   = no  *
> *   ; way of pressure coupling *
> *   pcoupltype               = isotropic*
> *   ; frequency for coupling pressure*
> *   nstpcouple               = -1*
> *   ; Time constant (ps)*
> *   tau_p                    = 2*
> *   ; compressibility in bar^(-1)*
> *   compressibility          = 4.5e-5 *
> *   ; reference pressure*
> *   ref_p                    = 1*
>
>       The settings of the whole simulation are simple, but we get a weird
> result. That is why I am writing here to get help. In our simulation, the
> only unusual thing is the box size(4x4x180nm). In our test with normal box,
> like 4 x 4x 4 nm, the result will be correct as we think. Do you think this
> could be the reason we get a strange result? Thanks a lot.
>
> Ming Geng
>
>