04.08.2013 Views

Statistical Mechanics - Physics at Oregon State University

Statistical Mechanics - Physics at Oregon State University

Statistical Mechanics - Physics at Oregon State University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

82 CHAPTER 4. STATISTICS OF INDEPENDENT PARTICLES.<br />

we need absolute convergence, th<strong>at</strong> is we need to be able to find a minimum value<br />

of N for which the partial sum is a good approxim<strong>at</strong>ion for all temper<strong>at</strong>ures.<br />

Hence we need for every ɛ > 0 we a value Nɛ such th<strong>at</strong><br />

N > Nɛ ⇒ |SN(T ) − S(T )| < ɛ (4.58)<br />

In th<strong>at</strong> case we can interchange summ<strong>at</strong>ion and limits, and can integr<strong>at</strong>e the<br />

sum of the series by integr<strong>at</strong>ing each term and sum the resulting integrals.<br />

In our case, for large values of the quantum numbers the energy is very large<br />

and the distribution function can be approxim<strong>at</strong>ed by<br />

lim<br />

ɛ→∞ fF D(ɛ) ≈ e µ−ɛ<br />

kB T (4.59)<br />

Hence the terms in the series for N decay very rapidly. If we consider a temper<strong>at</strong>ure<br />

interval [0, Tmax] we always have th<strong>at</strong> e −ɛ −ɛ<br />

kB T k e B Tmax . Therefore we<br />

can always use the value <strong>at</strong> Tmax to condition the convergence of the series.<br />

Hence we find th<strong>at</strong> the series converges uniformly on the interval [0, Tmax], with<br />

endpoints included. We may interchange the limit T → 0 and the sum and get<br />

N = <br />

Θ(ɛF − ɛo) (4.60)<br />

orb<br />

If we use free particle energies, which are always positive, we find immedi<strong>at</strong>ely<br />

th<strong>at</strong> ɛF > 0 because N > 0. Note th<strong>at</strong> the only problem is when we take<br />

T → ∞. Th<strong>at</strong> limit needs to be analyzed separ<strong>at</strong>ely, since we now have an<br />

infinite number of infinitesimally small terms to sum. We will see l<strong>at</strong>er how<br />

to deal with th<strong>at</strong> limit. We already know the answer, though, from a previous<br />

section. At large temper<strong>at</strong>ures we can replace the sum by an integral, if the<br />

orbital energies are free particle energies. This is always true for large quantum<br />

numbers, due to Bohr’s correspondence principle.<br />

Grand partition function.<br />

The grand partition function for a Fermi gas involves a sum over st<strong>at</strong>es.<br />

If we enumer<strong>at</strong>e the orbitals for the independent particles, a st<strong>at</strong>e of the total<br />

system can be specified by the occup<strong>at</strong>ion of each orbital, or by a set of numbers<br />

{n1, n2, . . .}. Here ni denotes the number of particles in orbital i. Since we are<br />

dealing with fermions ni is zero or one only. Examples of st<strong>at</strong>es are {0, 0, 0, . . .}<br />

for a st<strong>at</strong>e with no particles or {1, 1, 0, . . .} for a st<strong>at</strong>e with one particle in orbit<br />

1 and in orbit 2.<br />

The grand partition function is<br />

Z(T, µ, V ) = <br />

{n1,n2,...}<br />

The total number of particles is easy to find:<br />

e 1<br />

k B T (µN({n1,n2,...})−E({n1,n2,...}))<br />

(4.61)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!