I am working with N2 + CH4 + H2O mixtures. Since N2 is a three-site model which contains a zero mass point charge at the N2 center of mass. I set a tiny value 0.0000001 on this point charge/massless atom. CH4 is a single point model. H2O is TIP4P . I need to simulate a mixture of rigid and non rigid bodies. I tried to carry out NPT simulation following the options reported in fix rigid documentation:
1- fix rigid/nvt for N2 and npt with dilate all for H2O and CH4.
2- rigid/npt with dilate all for N2 and nvt for H2O and CH4.
However, I found using VMD to visualise the dump file that the centre of mass drift over timescale of 500 ps. The timestep is 1 fs. Is the drift in COM part of fix rigid? Anybody had this problem before? Thanks in advance for any support.
I am working with N2 + CH4 + H2O mixtures. Since N2 is a three-site model
which contains a zero mass point charge at the N2 center of mass. I set a
tiny value 0.0000001 on this point charge/massless atom. CH4 is a single
point model. H2O is TIP4P . I need to simulate a mixture of rigid and non
rigid bodies. I tried to carry out NPT simulation following the options
reported in fix rigid documentation:
1- fix rigid/nvt for N2 and npt with dilate all for H2O and CH4.
2- rigid/npt with dilate all for N2 and nvt for H2O and CH4.
However, I found using VMD to visualise the dump file that the centre of
mass drift over timescale of 500 ps. The timestep is 1 fs. Is the drift in
COM part of fix rigid? Anybody had this problem before? Thanks in advance
1 fs seems like a large timestep for your system. rigid body
integrators usually require rather shorter timesteps for cleanly
integrating the rotational motions.
however, most of the time a C.O.M. momentum is induced into a system
by its original configuration and initial velocity assignment and -
once present - will gradually increase over the course of the
simulation. using an aggressively large time step typically
accelerates this effect.