Force field in molecular dynamics

[thinhgt18]

Hi everyone,
I have a question about molecular dynamics force field.
Do the force field parameters depend on the equilibrium temperature?
For example in the case of polymers, a material that is highly dependent on the temperature used?
Thanks

[studiot]

The force field depends upon the source of the force, for example gravity or electrostatic attraction/repulsion.

 

Do you know any that depend upon temperature and if so the temperature of what?

[timo]

A "force field" in molecular dynamics is the interaction rules between the simulated objects, which is only vaguely related to a force-valued function over space (which is what the meaning of "force field" would be in most physics).

The answer to the question asked depends on the actual force field. Force fields of atomistic simulations should not depend on the temperature. The effect of temperature on the system comes from thermodynamics, which is still captured in MD simulations even if the force field has no temperature dependence. Essentially, the temperature comes via thermostats, that give the particles suitable random kicks to keep the whole system at the target temperature. That said, force field parameters can depend on temperature, at least in principle (at least that is what I think - I am not a super expert on MD but I have some background in scientific computer simulations). For example, consider a so-called coarse-grained force field, where 3-4 atoms are combined into a single simulation point for performance reasons. Force fields for such simulation points may well try to catch some of the internal dynamics of the multi-atom point which, in turn, depends on the temperature (at least in principle). I am not sure if such theoretically imaginable dependencies play a role in practice, though. I'd certainly be careful to perform simulations near absolute zero with a force field that has been designed and tuned for biological applications. Or even more radically: A force field for beads of 3-4 bound atoms certainly won't be applicable for simulations of plasmas, where not only the force field's parameters but even underlying assumptions (bound states) break down.

 

EDIT: Thinking of it: Even atomistic force fields make assumptions and approximations that hold only within some temperature range. So the simpler answer is: The applicability of a force field and its parameters depends on the temperature in principle. For a given problem at hand, the temperature effect may or may not be irrelevant (especially compared to other reasons for getting results that are not 100% exact).

Edited October 16, 2013 by timo