[gmx-users] Why does my box change dimensions?

[Chris O'Brien]

Dear gmx-users,

When performing NPT simulations (2ps tauP) on my system of 32 infinite 
polymer chains, something unexpected occurred near the 6-ns mark.  The 
system density, which had been creeping steadily upward near a value of 
about 1250 (the experimental density), suddenly plummeted to around 850 
over the course of 1.5 ns.  Although the run has not finished yet, the 
density is showing no indications that it will return to its previous value.

When this first occurred, I was experimenting with a 3-fs time step (I have 
only C and O atoms) and thought that maybe I had been unlucky and some 
atoms overlapped.  However, I saw the same behavior in an identical 
simulation using a 2-fs time step.  The chain dimension Z shrinks about 20% 
while the X- and Y- dimensions increase noticeably.  None of the energy 
components show any indication to account for this behavior, nor do I see 
anything useful in the log file except for:

Grid: 7 x 5 x 25 cells
Grid: 7 x 4 x 25 cells
Grid: 7 x 5 x 25 cells
Grid: 7 x 4 x 25 cells
Grid: 7 x 5 x 25 cells
Grid: 7 x 4 x 25 cells
Grid: 7 x 5 x 25 cells


Related to this, when I plot the box vectors, all of the curves show an 
increase at this time period even though (as noted above) the box is 
clearly getting smaller in the z-direction when looking at the trajectory 
in VMD.

Relevant sections of my mdp file:

integrator            = md
dt                       = 0.002 ;ps = 1 fs
nsteps                = 5000000 ; 10.0 ns
nstlist                 = 10
ns_type              = grid
pbc                    = full
rlist                    = 1.0
coulombtype       = cut-off
rcoulomb            = 1.2; 2*1.4>~2.8
vdwtype              = cut-off
rvdw                   = 1.0
tcoupl                 = berendsen
tc-grps               = System
tau_t                  = 0.1
ref_t                   = 300; 300 300
Pcoupl               = berendsen
pcoupltype          = isotropic
tau_p                  = 2.0
ref_p                   = 1.0
gen_vel               = yes
gen_temp            = 300
gen_seed             = 214567


Any ideas or suggestions?

Thank you,
Chris O'Brien

Department of Chemical Engineering
Clemson University
Clemson, SC  29634-0909