Is it possible to use the Langevin thermostat and have a temperature gradient since fix Langevin specifies the temperature (Tstart and Tstop) for a region? In other words, to look at the behavior of Brownian particles (with an implicit solvent) in a system with a temperature gradient.
Would cutting the box into thin slices with a different fix-ID at different temps for each group work?
Looking for some examples.
Thanks,
Fred
Is it possible to use the Langevin thermostat and have a temperature
gradient since fix Langevin specifies the temperature (Tstart and Tstop) for
a region? In other words, to look at the behavior of Brownian particles
(with an implicit solvent) in a system with a temperature gradient.
outside of programming something like that yourself,
i don't see how that would be possible with any of the
facilities in LAMMPS that i know. without having given
it much thought i would also worry about potential
side effects of such an setup.
Would cutting the box into thin slices with a different fix-ID at different
temps for each group work?
no. since fixes are applied to groups of atoms, not regions.
those would only coincide at the very beginning of the calculation.
axel.
You can use fix langevin (or other thermostats) with regions
of atoms, if you use fix_modify temp to assign a compute temp/region
to it. Then it will thermostat atoms in a region. You can use 2 such
tstats, one at each end of your box to assign one to a hi Temp, the
other to a lo Temp, and get an effective gradient in between.
If you mean to tstat every atom, but to a different spatially-dependent
Temp, depending on its position, that would have to be coded into
fix langevin. Don't think that's very physical however.
Steve