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

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

1.3. STATES OF A SYSTEM. 9<br />

which also is much smaller than one when N is very large. In other words, for<br />

all values where the Gaussian approxim<strong>at</strong>ion to the multiplicity function has a<br />

large value it is close to the original function! In th<strong>at</strong> sense we can replace the<br />

multiplicity function by a Gaussian form for all values of x. The rel<strong>at</strong>ive errors<br />

are large when x is not small, but so is the value of g(N,x). In other words,<br />

when the rel<strong>at</strong>ive difference between the actual multiplicity function and the<br />

Gaussian approxim<strong>at</strong>ion is large, the values of the multiplicity function are very<br />

small and the absolute value of the error is very small. This result is general for<br />

all multiplicity functions. In the thermodynamic limit the devi<strong>at</strong>ions from the<br />

maximum value are always of a Gaussian n<strong>at</strong>ure in the area where the values<br />

are large. This is one of the reasons th<strong>at</strong> st<strong>at</strong>istical mechanics works!<br />

It also reinforces the importance of the thermodynamic limit! It is<br />

directly rel<strong>at</strong>ed to the law of large numbers in m<strong>at</strong>hem<strong>at</strong>ics.<br />

In order to keep the normaliz<strong>at</strong>ion correct we replace the factor N − 1 in the<br />

exponent by N and will use l<strong>at</strong>er<br />

Because we know th<strong>at</strong><br />

g(N, x) ≈<br />

<br />

2<br />

πN 2N 1 −<br />

e 2 x2N <br />

g(N, x) = 2 N<br />

x<br />

(1.24)<br />

(1.25)<br />

and th<strong>at</strong> the stepsize for x is 2<br />

N (which includes a factor of two since M can only<br />

change in units of two!), we can write<br />

N 2<br />

2<br />

N<br />

<br />

= g(N, x)∆x →<br />

For the Gaussian approxim<strong>at</strong>ion we find<br />

1<br />

−1<br />

x<br />

1<br />

−1<br />

<br />

2<br />

πN 2N 1 −<br />

e 2 x2 <br />

N<br />

dx =<br />

√ N<br />

2<br />

− √ <br />

2<br />

N<br />

2<br />

∞<br />

−∞<br />

<br />

2<br />

πN 2N e −t2<br />

<br />

2<br />

dt<br />

N<br />

which is the correct normaliz<strong>at</strong>ion indeed.<br />

Wh<strong>at</strong> is the physical meaning?<br />

g(N, x)dx (1.26)<br />

πN 2N e −t2<br />

dt<br />

= 2N 2<br />

N<br />

<br />

N<br />

2 ≈<br />

(1.27)<br />

At this point we start to wonder wh<strong>at</strong> the multiplicity function means. Suppose<br />

the parameter x in the previous example is an internal parameter of the<br />

system. This would be the case in a system in zero magnetic field, where there<br />

is no work involved when the magnetic moment changes. In thermodynamics<br />

we have seen th<strong>at</strong> the equilibrium st<strong>at</strong>e is the st<strong>at</strong>e of maximal entropy. Here<br />

we see th<strong>at</strong> the st<strong>at</strong>e with x = 0, which we expect to be the equilibrium st<strong>at</strong>e,

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

Saved successfully!

Ooh no, something went wrong!