Re: Second law in the microscopic scale

[Jack Uretsky <jlu@HEP.ANL.GOV>]

        For dimensional reasons, [xi] has to be something like the
correlation length over the particle (or lattice) spacing.  The point I
intended, however, is that (barring phase transitions, which I don't know
how to envision for few-particle systems) is that when N is Avagodro's
number, fluctations are ignorable.


On Sat, 27 Dec 2003, David Bowman wrote:

> Regarding Jack U's comment:
>
> >    Essentially yes.  If N is the number of particles in the system,
> > then the random-fluctuation probability is like 1/sqrt(N) - See
> > Landau and Lifshitz introduction.
> >                             Jack
> =20
> If we include the effects of correlations between the particles I
> believe the relative fluctuations go more like sqrt([xi]^d / N) where
> [xi] is the correlation length and d is the spatial dimensionality
> of the system (assuming the interactions between the particles are
> finite-ranged).
>
> David Bowman

--
"Don't push the river, it flows by itself"
                                Frederick Perls