Statistical Mechanics - Physics at Oregon State University
Statistical Mechanics - Physics at Oregon State University
Statistical Mechanics - Physics at Oregon State University
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
144 CHAPTER 7. CLASSICAL STATISTICAL MECHANICS.<br />
7.3 Ergodic theorem.<br />
Connection with experiment.<br />
Most experiments involve systems th<strong>at</strong> involve many degrees of freedom. In<br />
all those cases we are only interested in average quantities. All average quantities<br />
follow from the quantum mechanical formula, and can be expressed as classical<br />
averages if we change sums to integrals:<br />
< A >=<br />
1<br />
N!h 3N<br />
<br />
d Xρ( X)A( X) (7.21)<br />
The factor N! is needed for indistinguishable particles, and should be omitted<br />
for distinguishable particles.<br />
The previous formula does, of course, not describe the way we measure<br />
quantities! In a measurement we often consider time averaged quantities or<br />
space averaged quantities. In the first case, we start the system <strong>at</strong> an initial<br />
st<strong>at</strong>e <br />
∂pi<br />
∂H<br />
Xin <strong>at</strong> t = 0 and use Hamilton’s equ<strong>at</strong>ions ∂t = − and ∂xi<br />
∂xi<br />
∂t =<br />
<br />
+ to calcul<strong>at</strong>e X(t). We then evalu<strong>at</strong>e<br />
∂H<br />
∂pi<br />
< A >τ (t) = 1<br />
t+τ<br />
A(<br />
τ t<br />
X(t ′ ))dt ′<br />
(7.22)<br />
where X(t) is the st<strong>at</strong>e of the system <strong>at</strong> time t. The value of this average<br />
depends on the measurement time τ and on the initial st<strong>at</strong>e. Birkhoff showed<br />
th<strong>at</strong> in general limτ→∞ < A >τ (t) is independent of the initial conditions. So<br />
if we take τ large enough it does not m<strong>at</strong>ter where we start. Note th<strong>at</strong> in cases<br />
where we have external potentials th<strong>at</strong> vary with time we cannot make τ too<br />
large, it has to be small compared to the time scale of the external potential. In<br />
different words, the effects of external potentials th<strong>at</strong> vary too rapidly are not<br />
described by thermodynamic equilibrium processes anymore.<br />
Equivalency of averages.<br />
Now we ask the more important question. Under wh<strong>at</strong> conditions is this<br />
measured quantity equal to the thermodynamical ensemble average? In the<br />
quantum-mechanical case we argued th<strong>at</strong> all accessible quantum st<strong>at</strong>es are<br />
equally probable if the total energy U of the system is fixed. We needed to<br />
invoke some random external small perturb<strong>at</strong>ion to allow the system to switch<br />
quantum st<strong>at</strong>es. In the classical case we need a similar argument.<br />
The time evolution of a classical system is determined by Hamilton’s equ<strong>at</strong>ions:<br />
∂pi<br />
∂t<br />
= −∂H<br />
∂xi<br />
, ∂xi<br />
∂t<br />
= ∂H<br />
∂pi<br />
(7.23)