[gmx-users] Re: NPT simulation

[ms]

On 29/08/10 19:39, Justin A. Lemkul wrote:

> When dividing the energy of a system by the number of molecules (in a
> homogeneous system), you are extracting what I believe is commonly
> referred to as "configurational energy" which, for relatively simple
> systems, should converge fairly quickly. The reason your total energy
> values increase with system size is a simple matter of potential energy.
> More interactions mean that the magnitude of the potential will
> increase, and likewise with the kinetic energy, more particles that have
> velocity imply a greater sum.
>
> You can convince yourself of this fact by running relatively short
> simulations of water boxes, using, i.e. spc216.gro and then a larger
> construct from it. Within a few hundred ps, you should get reasonably
> converged configurational energies, though of course the total energy
> will be larger simply by virtue of the system size.
>
> Energies are in kJ/mol, no question. The extrapolation would be that a
> mole of a given species in the given configuration would have this
> energy in kJ. For a single molecule, the energy is whatever value you
> obtain divided by Avogadro's number, which is a relatively unimportant
> quantity if you are interested in bulk dynamics.

Wait, now I am confused too.
I always understood that energy was meant as kJ/mol for a mol of my 
*system*, that is, if I have a box with 10 identical molecules, then I 
g_energy gives me out the energies of that *box*.

Is it wrong?

m.