hi,
I have been confused by the large pressure disturbation in NPT ensemble these days.
In your last reply, you said that For only 150 atoms, large fluctuations in pressure (including negative) are normal for MD.But there are about 5000 atoms in my system,large pressure fluctuations in pressure still exists.
Some friends using lammps also found the question.when we use atmosphere as unit,the fluctuations will be several atmospheres or so.
I want to know is this the flaw of MD method or just in lammps? How should we measure the affect of large pressure fluctuation when we use 10 atmosphere as our result but the real situation is 1 atmosphere?
Thank you for your reply
Hi Herofather,
The large pressure distribution is not a flaw of lammps. It is not a flaw of MD either. As others have said, it is a consequence of the small system sizes that we all use to model systems. So, it is natural to have a mean pressure of 1 atmosphere with a RMS deviation that is much much larger (say 10 or 20 atmospheres or more). Is this a problem. Not if you are studying a system which is not-so-compressible. Most people who do MD look at condensed phase systems which are weakly compressible relative to gases. So, even if the average pressure of the simulated system turns out to 50 atmospheres, it is ok. As an example, please check out the density of liquid water as a function of pressure at 300K. You would find it hardly changing. The marginal changes that one would find in the experimental densities would be subsumed within the fluctuations of density that one sees in a MD trajectory of liquid water at 300K and 0 atm. So, it is ok.
Cheers,
Bala
http;//www.jncasr.ac.in/bala