Scribd
Upload
60 views

"Derivation" of Canonical Distribution

The document summarizes Schrödinger's derivation of the canonical distribution. It considers a system with states indexed by l, each with energy ǫl . The occupation number al represents the number of systems in state l. The total energy E and number of systems N are fixed. Maximizing the probability P of the system configuration subject to these constraints using Lagrange multipliers yields the canonical distribution, where the probability of a state is proportional to e−βǫl , where β is determined by the total energy. The partition function Z emerges, from which thermodynamic quantities like the mean energy per system U can be calculated.

Uploaded by

ANDRÉS FELIPE SERRANO BARRIOS
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
60 views

"Derivation" of Canonical Distribution

The document summarizes Schrödinger's derivation of the canonical distribution. It considers a system with states indexed by l, each with energy ǫl . The occupation number al represents the number of systems in state l. The total energy E and number of systems N are fixed. Maximizing the probability P of the system configuration subject to these constraints using Lagrange multipliers yields the canonical distribution, where the probability of a state is proportional to e−βǫl , where β is determined by the total energy. The partition function Z emerges, from which thermodynamic quantities like the mean energy per system U can be calculated.

Uploaded by

ANDRÉS FELIPE SERRANO BARRIOS
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Chapter 2

“Derivation” of Canonical
Distribution

By way of illustration, we now give Schrödinger’s “derivation” of the canonical


distribution. The argument (roughly) applies to any system, be it classical or
quantum. Let the states be enumerated as follows

State No. 1 2 3 ... l ...


Energy ǫ1 ǫ2 ǫ3 . . . ǫl . . . (2.1)
Occupation No. a1 a2 a3 . . . al . . . .

The occupation number means the number of systems in the state. Recall that
there are N systems altogether in the ensemble, so

X
∞
N= al . (2.2)
l=1

We also assume that the total energy E in the ensemble is fixed, so

X
∞
E= al ǫ l . (2.3)
l=1

(These sums are actually finite sums, since only a finite number of the al s can be
different from zero.) The number of states of this class for the whole ensemble
is the number of ways of getting this configuration:

N!
P = . (2.4)
(a1 )!(a2 )! . . . (al )! . . .

We seek to find the most probable distribution; that is, to maximize P ,


subject to the constraints (2.2) and (2.3). Equivalently, we seek to maximize
ln P . The usual way of solving such a constrained maximization problem is to

7 Version of January 25, 2010


8 Version of January 25, 2010CHAPTER 2. “DERIVATION” OF CANONICAL DISTRIBUTION

introduce Lagrange multipliers. That is, we seek the unconstrained maximum


of X X
ln P − α al − β al ǫ l . (2.5)
l l

where α and β are parameters to be determined. Now in so doing, recall Stir-


ling’s formula,
ln n! ∼ n(ln n − 1), n → ∞. (2.6)
Varying the occupation numbers (assumed large), we find then the extremum
condition
X X X
0 = − [δal (ln al − 1) + δal ] − α δal − β ǫl δal
l l l
X
= − δal [ln al + α + βal ] . (2.7)
l

This must be true for all δal , so

ln al + α + βǫl = 0, (2.8)

or
al = e−α e−βǫl . (2.9)
Thus, from constraint (2.2),
X
N= e−α−βǫl , (2.10)
l

and from constraint (2.3),


X
E= ǫl e−α−βǫl . (2.11)
l

The average energy per system is


P
E ǫl e−βǫl d
U≡ = Pl −βǫ = − ln Z, (2.12)
N l e l dβ
where the partition function is
X
Z= e−βǫl . (2.13)
l

The mean fraction of systems in the l state is

al e−βǫl
= . (2.14)
N Z
Thus, the canonical distribution has emerged as the most probable (overwhelm-
ingly so as N → ∞) distribution of systems within the ensemble of fixed total
energy.

You might also like