\( \newcommand{\be}{\begin{equation}}
\newcommand{\ee}{\end{equation}} \newcommand{\ba}{\begin{array}}
\newcommand{\ea}{\end{array}} \newcommand{\bea}{\begin{eqnarray}}
\newcommand{\eea}{\end{eqnarray}}
\newcommand{\bean}{\begin{eqnarray*}}
\newcommand{\eean}{\end{eqnarray*}} \newcommand{\la}{\label}
\newcommand{\nn}{\nonumber} \newcommand{\half}{{\scriptstyle
\frac{1}{2}}} \)
On the macroscopic level, the state of a many-particle system is
described by the conserved charge densities $N_{\rm
A}(x)$, the energy density $E(x)$, and the momentum
density $\vecG(x)$. Furthermore, one may define a space- and
time-dependent entropy density $S(x)$. Since entropy is
not conserved, $S(x)$ satisfies a balance equation of the form
$$ \be {\ptt}S(x) + \nabla
\cdot {J_{S}}(x) = \sigma (x) {\rm{ }}{\rm{,}} \la{1.1} \ee $$
where $J_{S}$ is the entropy flow, and $\sg$ the local entropy
production (per unit volume and unit time) which is never
negative to be in accord with the second law of thermodynamics:
$$ \be \sigma (x) \ge 0
{\rm{ }}{\rm{.}} \la{1.2} \ee $$
Exercise 1
- Show by integrating over all space that (2.1) implies that
the total entropy never decreases.
A non-equilibrium state of a system is characterized by a
non-zero entropy production. However, at this stage we do not yet
know how the entropy production is related to the irreversible
processes which may occur in the system. Therefore, we shall use
for now a more heuristic criterion for non-equilibrium behaviour:
we shall declare a system to be outside equilibrium if the distribution
function
$$ \be f_k (x,p) = {n_k}(p) +
\delta {f_k}(x,p) \la{1.3} \ee $$
differs by a non-zero amount $\dl f_k$ from the equilibrium Fermi
or Bose distribution function $n_k (p)$.
Since, in general, the properties of a non-equilibrium system
are non-stationary and non-uniform, the distribution function is
space and time dependent. To keep the discussion general we assume
that there are a number $k = 1,2,...,$ of different particle
species in the system. However, we shall confine ourselves to
systems only slightly out of equilibrium which means that $\dl
f_k$ is small, and that quadratic terms may be neglected.
Let us now write down the particle density of component
$k$ :
$$\be N_k (x) = {\int
{d\omega }_k}\,{f_k}(x,p){\rm{ }}{\rm{,}} \la{1.4} \ee $$
where we introduced the notation
$$ \be d{\omega _k}: =
{g_k}\frac{{{d^3}p}}{{{{(2\pi )}^3}}}{\rm{ }}{\rm{.}} \la{1.5} \ee
$$
with $g_k$ the spin—weight factor, i.e. the number of
independent spin states of the particles. It is customary to
regard $f_k$ as giving the distribution of the particles in
$\mu$-space:
$\triangleright\,\, f_k(x,p) \Dl^3 p \Dl^3 x =$ the
number of particles of species $k$ which at time $t$ are
situated in the small volume $\Dl^3 x$ at the point $\vecx$ with
momenta lying in the range $(\vecp - \half \Dl \vecp , \vecp +
\half \Dl \vecp)$.
However, it must be stressed that this definition is not really
necessary. It is sufficient to assume that the density can be
expressed in terms of an ancillary quantity $f_k (x,p)$ which in
equilibrium reduces to the Fermi or Bose distribution function
$n_k (p)$. In terms of this ancillary quantity the energy
density and entropy density of species $k$ take the
forms:
$$ \be {E_k}\left( x \right)
= {\rm{ }}\int {d{\omega _k}} \,{\varepsilon _k}\, {f_k}\left( x,p
\right){\rm{ }}{\rm{,}} \la{1.6} \ee $$
$$ \be {S_k}\left( x
\right)= {\rm{ }}- \int {d{\omega _k}}\, [{f_k}\log{f_k}-\eta (1 +
\eta {f_k})\log\left( {1 + \eta {f_k}} \right)]{\rm{ }}{\rm{,}}
\la{1.7} \ee $$
- These expressions are the direct generalizations of the energy
and entropy densities of a quantumgas
in equilibrium by means of the substitution $n_k
(p)\rightarrow f_k (x,p)$; $\et = +1$ for bosons, and $\et= —1$
for fermions.
- Because of the absence of any interaction terms, the
expressions are only valid for weakly coupled or dilute particle
systems. Nevertheless, this theory of non—equilibrium "ideal"
gases is very important as a model and guideline for the
construction of more complicated theories.