[lammps-users] negative pressure

hi all,
Can someone take a look at my input to check that does not have any error. Is to simulate an ionic glasses system (Li2SiO3). I have attached the output file and input file. The problem is if I dont really equilibrate the system. As shown in the output file, the energy, temperature and volume remain constant, but the pressure is large and negative values, then the last cooling cycle. What they really mean ALL these negative values in the pressure?
On the one hand makes me think that the system really is in equilibrium before cooling, but it seems that instantly disarms the system during cooling. If I compute the MSD gives very large values without physical meaning.
Well annex the input file and the thermo file for if someone wants to make a suggestion about my problem.
Thanks
Cristian.

termalizacion-432.sh (2.99 KB)

termalizacion-432.salida (276 KB)

A large negative pressure means that your box is too big for the system you’re trying to study. Therefore, as soon as you turn on the fix npt, the system is going to start to shrink. However, since your box size is only slightly larger than twice your system size to begin with, when the system starts shrinking, you will quickly reach a point where particles and their images are both in an atom’s neighbor list, at which point bad things are almost guaranteed to happen.

–AEI

Hi Ahmed !
<However, since your box size is only slightly larger than twice your system size to begin with, when the system <starts shrinking…

Do you mean by this that the (Rc + skin) > (1 / 2 L) when the system reduces its volume, then there the problems begin, because you can add to the list of neighboring real particle and its image when ppp apply.
But the cycle where the problem happens I carried out in a NVT ensemble,

fix 15 all nvt 1000.0 800.0 100
thermo 1000
run 500000
unfix 15

In anyway in this case, would be a size problem, which I can prove it easily by increasing the size of my system.Homework!

An important consideration to keep in mind is that if I do cooling cycle in a NPT ensemble, instead of NVT,

fix 15 all npt 1000.0 800.0 100 xyz 1.0 1.0 10.0 drag 1.5
thermo 1000
run 500000
unfix 15

the pressure oscillates around 1atm correctly, but at a given temperature pressures become all positive values and large, the 1E4 order, contrary to what happens in the NVT.
In both cases the MSD gives me completely illogical results, extremely large values.
thank you very much again for the suggestions!
Cristian