[gmx-users] Pressure negative after NVT simulation

[Mark Abraham]

> Dear GROMACS users and developers,
>
> I am running MD simulation of lipid self-assembly with total of 53
> molecules in a 1000 nm3 cubic simulation box with spc216 as the solvent.
> After minimization i ran constant NPT MD equilibration up to 1 ns (just to
> make sure) for the system to have suitable volume and density before
> running with NVT. After the equilibration, i observed the potential energy
> and it looked fine. So i took the volume of the equilibration trajectory
> which then was 1025.51 nm3 as my new box volume. Then i ran 5ns of
> constant NVT MD simulation. After simulation i checked the pressure of the
> system throughout the simulation and found that the average pressure of
> the system from g_energy was -12.7856 bar. The xvg file demonstrated that
> the pressure fluctuate very high from about 100 to -100 troughout 5 ns. I
> am still not familiar with this situation but i believe that negative
> value of pressure is not a good thing. Can anyone kindly help me tell
> what's wrong with my simulation?

For the fluctuations, see http://wiki.gromacs.org/index.php/Pressure_Coupling

The negative pressure indicates the system wants to contract, so if you
did a restart that preserved atomic velocities
(http://wiki.gromacs.org/index.php/Doing_Restarts), then you have probably
determined that the NPT was not yet equilibrated. This might also be
apparent from looking at NPT observables other than PE.

Mark