Microcanonical Ensemble

Prof. Sami Mahmood

The University of Jordan

1 2/28/2023
 Course Outline
I. Macroscopic and microscopic description of a system

II. Classical postulate of statistical mechanics

III. Microcanonical ensemble

IV. Ideal gas statistics: classical gas, bosons, fermions

2 2/28/2023
 Course Outline
V. The Canonical ensemble

VI. The grand canonical ensemble

VII. Thermodynamic functions

VIII. The phonon gas; black body radiation; B-E

condensation

3 2/28/2023
 Introduction

An equilibrium macrostate of a system is

identifiable by a set of macroscopic
properties that are reproducible upon
repetition of the measurements. The
temperature, volume and pressure of a
gas, and the magnetic moment of a piece
of iron, etc., all should be reproducible at
equilibrium.

4 2/28/2023
 Introduction

A macroscopic system in a non-

equilibrium state proceeds with
time toward equilibrium, passing
through all possible microstates.
We can describe the equilibrium
properties of the system by two
ways.

5 2/28/2023
 Introduction

Firstly, by monitoring the state of

the system with time, and take the
average of the properties as its
microstate changes with time

6 2/28/2023
 Introduction

Alternatively, we can represent our

system by an ensemble consisting of
a large number (N ) of replicas
(identical copies) of the system. The
ensemble is isolated and has a fixed
energy (E).

7 2/28/2023
 Introduction

For each accessible state of our system

there is at least one corresponding
replica; each possible microstate of the
system is represented in the ensemble

The study of the ensemble is equivalent

to studying the system as its state varies
with time passing through all the
possible microstates.

8 2/28/2023
 Introduction

The average properties of the systems in

the ensemble will therefore give the
same result as if this average were taken
for a single system as its state varies
with time

The time average of a system property is

equal to the ensemble average of that
property

9 2/28/2023
 Introduction

Depending on the constraints imposed

on the system, three different ensembles
are used to study the equilibrium
properties of a system

The microcanonical ensemble, the

canonical ensemble, and the grand
canonical ensemble.

10 2/28/2023
 Entropy
We now show that the ensemble average
of the entropy is quite generally given
by:

𝑺 = −𝒌 ෍ 𝒑𝒓 𝐥𝐧 𝒑𝒓
𝒓

Here 𝒑𝒓 = 𝒏𝒓 ΤN is the probability that the

system is in the microstate 𝒓, and 𝒏𝒓 is
the number or replicas in that state.

11 2/28/2023
 Entropy
Consider an ensemble of (N ) replicas of
a system sharing a total energy (E). Let
𝒏𝒓 be the number of systems in the
microstate 𝒓. The constraints imposed on
the ensemble are such that:

෍ 𝒏𝒓 = N
𝒓

෍ 𝒏𝒓 𝑬 𝒓 = E
𝒓

12 2/28/2023
 Entropy
The number of ways in which we can
distribute the N systems over the set 𝒏𝒓
is then given by:

N!
𝛀 𝒏𝒓 =
𝒏𝟏 ! 𝒏𝟐 ! 𝒏𝟑 ! …

𝒏𝒓 = N 𝒑𝒓

13 2/28/2023
 Entropy
The entropy of the ensemble is:

𝑺N = 𝒌 𝐥𝐧𝛀

𝑺N = 𝒌 𝐥𝐧 N ! − ෍ 𝐥𝐧 (𝒏𝒓 !)
𝒓

14 2/28/2023
 Entropy
Stirling formula for large n

𝒏! ≈ 𝒏𝒏 𝒆−𝒏 𝟐𝝅𝒏 → 𝐥𝐧 𝒏! ≈ 𝒏 𝐥𝐧 𝒏 − 𝒏

Notice that 𝐥𝐧 𝟐𝝅 × 𝟏𝟎𝟐𝟑 ≈ 𝟐𝟕 ≪ 𝟏𝟎𝟐𝟑

At macroscopic scales, the natural

logarithm of the term under the square-
root is ignorable.

15 2/28/2023
 Entropy
Using Stirling formula

𝑺N = 𝒌 N 𝐥𝐧N − N − ෍ 𝒏𝒓 𝐥𝐧 𝒏𝒓 − 𝒏𝒓
𝒓

𝑺N = 𝒌 N 𝐥𝐧N − ෍ 𝒏𝒓 𝐥𝐧 𝒏𝒓
𝒓

Now substitute 𝒏𝒓 = N 𝒑𝒓

16 3/13/2023
 Entropy

𝑺N = 𝒌 N 𝐥𝐧N − ෍ N 𝒑𝒓 (𝐥𝐧N + 𝐥𝐧 𝒑𝒓 )
𝒓

𝑺N = −N 𝒌 ෍ 𝒑𝒓 𝐥𝐧 𝒑𝒓
𝒓

𝑺N
𝑺= = −𝒌 ෍ 𝒑𝒓 𝐥𝐧 𝒑𝒓
N
𝒓
17 2/28/2023
 Entropy

In the derivation of this formula, notice

that no constraints on our system were
specified. Therefore, this formula is a
general.

In reaching the final result, extensivity

of the entropy was assumed as we will
show in more details later.

18 2/28/2023
 Microcanonical Ensemble

The macrostate of an isolated

system is defined by fixed 𝑵, 𝑽,
𝑬 (practically 𝑬 𝒕𝒐 𝑬 + 𝜹𝑬). The
ensemble representing the system
in this case is known as the
microcanonical ensemble

19 2/28/2023
 Microcanonical Ensemble

20 3/7/2023
 Microcanonical Ensemble

Any system in any accessible

microstate should be a member of
the microcanonical ensemble.

N = 𝛀(𝑬, 𝑵, 𝑽)

21 2/28/2023
 Microcanonical Ensemble
In the microcanonical ensemble, the
number of replicas in every possible
microstate is unity (equal a priori
probabilities)

𝟏
𝒏𝒓 = 𝟏 → 𝒑𝒓 =
𝛀(𝑬, 𝑽, 𝑵)

𝛀
𝟏 𝟏
𝑺(𝑬, 𝑽, 𝑵) = −𝒌 ෍ 𝐥𝐧 = 𝒌 𝐥𝐧 𝛀(𝑬, 𝑽, 𝑵)
𝛀 𝛀
𝒓=𝟏
22 2/28/2023
 Ideal Classical gas
Consider a classical, ideal gas composed
on 𝑵 monatomic particles in a volume 𝑽.
The number of microstates with energy
equal or smaller than 𝑬 is:

𝒑𝟐𝒊
σ𝒊 =𝑬
𝟐𝒎
𝟏
𝚽 𝑬 = 𝟑𝑵 න … න … න 𝒅𝟑𝑵 𝒒𝒅𝟑𝑵 𝒑
𝒉
𝟎

23 2/28/2023
 Ideal Classical gas
The integral over 𝒒’𝒔 is straightforward:

𝒑𝟐𝒊
σ𝒊 =𝑬
𝟐𝒎
𝑽𝑵
𝚽 𝑬 = 𝟑𝑵 න … න … න 𝒅𝟑𝑵 𝒑
𝒉
𝟎

The integral is the volume of a 𝟑𝑵 − 𝑫

sphere of radius given by 𝑹𝟐 = 𝟐𝒎𝑬.

24 2/28/2023
 Ideal Classical gas
𝒑𝟐𝒊
σ𝒊 =𝑬
𝟐𝒎
න… න … න 𝒅𝟑𝑵 𝒑 = 𝑪𝟑𝑵 𝟐𝒎𝑬 𝟑𝑵/𝟐

𝝅𝟑𝑵/𝟐
𝑪𝟑𝑵 =
𝟑𝑵
𝟐 !

25 2/28/2023
 Ideal Classical gas
Accordingly:

𝑽𝑵 𝝅𝟑𝑵/𝟐 𝟑𝑵/𝟐
𝚽 𝑬 = 𝟑𝑵 𝟐𝒎𝑬
𝒉 𝟑𝑵
𝟐 !

In the microcanonical ensemble, the

accessible states are in the energy range
from 𝑬 𝐭𝐨 𝑬 + 𝜹𝑬

26 2/28/2023
 Ideal Classical gas
Therefore:

𝑽𝑵 𝟐𝝅𝒎 𝟑𝑵/𝟐 𝟑𝑵/𝟐

𝛀 𝑬 = 𝟑𝑵 𝑬 + 𝜹𝑬 − 𝑬𝟑𝑵/𝟐
𝒉 𝟑𝑵
!
𝟐

𝒅𝚽 𝑬
𝛀 𝑬 = 𝜹𝑬
𝒅𝑬
𝑽𝑵 𝟐𝝅𝒎 𝟑𝑵/𝟐 𝟑𝑵/𝟐 𝟑𝑵 𝜹𝑬
𝛀 𝑬 = 𝟑𝑵 𝑬
𝒉 𝟑𝑵 𝟐 𝑬
!
𝟐
27 3/13/2023
 Ideal Classical gas
Final result:

𝑽𝑵 𝟐𝝅𝒎 𝟑𝑵/𝟐 𝟑𝑵/𝟐 𝜹𝑬

𝛀 𝑬 = 𝟑𝑵 𝑬
𝒉 𝟑𝑵 𝑬
𝟐 − 𝟏 !

𝑺 𝑬, 𝑵, 𝑽 = 𝒌 𝐥𝐧 𝛀(𝑬, 𝑵, 𝑽)

28 2/28/2023
 Ideal Classical gas
Now the stage set for thermodynamics:

𝑺 𝑬, 𝑵, 𝑽
𝑽 𝟑 𝟑𝑵 𝟑𝑵
= 𝒌𝑵 𝐥𝐧 𝟑 𝟐𝝅𝒎𝑬 𝟐 − 𝒌 − 𝟏 𝐥𝐧 −𝟏
𝒉 𝟐 𝟐
𝟑𝑵 𝜹𝑬
+𝒌 − 𝟏 + 𝒌 𝐥𝐧
𝟐 𝑬

29 2/28/2023
 Ideal Classical gas
Entropy:

𝑺 𝑬, 𝑵, 𝑽
𝑽 𝟑 𝟑𝑵 𝟑𝑵 𝟑𝑵
= 𝒌𝑵 𝐥𝐧 𝟑 𝟐𝝅𝒎𝑬 𝟐 −𝒌 𝐥𝐧 + 𝒌
𝒉 𝟐 𝟐 𝟐

𝟑Τ𝟐
𝑽 𝟒𝝅𝒎𝑬 𝟑
𝑺 𝑬, 𝑵, 𝑽 = 𝒌𝑵 𝐥𝐧 𝟑 + 𝑵𝒌
𝒉 𝟑𝑵 𝟐

30 2/28/2023
 Ideal Classical gas
Energy:

𝟑 𝟒𝝅𝒎𝑽𝟐/𝟑 𝟑
𝑺 𝑬, 𝑵, 𝑽 = 𝑵𝒌 𝐥𝐧 𝟐
𝑬 + 𝑵𝒌
𝟐 𝟑𝑵𝒉 𝟐

𝟑𝑵𝒉𝟐 𝟐𝑺
−𝟏
𝑬 𝑺, 𝑵, 𝑽 = 𝟐/𝟑
𝒆 𝟑𝑵𝒌
𝟒𝝅𝒎𝑽
Is this entropy extensive?
31 3/13/2023
 Ideal Classical gas
Entropy of Mixing and Gibbs Paradox

Check if the derived entropy extensive?

𝟑 𝟒𝝅𝒎𝑽𝟐/𝟑 𝟑
𝑺 𝑬, 𝑵, 𝑽 = 𝑵𝒌 𝐥𝐧 𝟐
𝑬 + 𝑵𝒌
𝟐 𝟑𝑵𝒉 𝟐

𝟑 𝟒𝝅𝒎𝑬
𝑺 𝑻, 𝑵, 𝑽 = 𝑵𝒌 𝐥𝐧 𝑽 + 𝑵𝒌 𝐥𝐧 𝟐
+𝟏
𝟐 𝟑𝑵𝒉
32 3/13/2023
 Ideal Classical gas
Notice that S here is not an extensive
property of the system as it should be!

(Paradoxical)

33 2/28/2023
 Ideal Classical gas
Consider mixing two gases. In case (I),
the initial system is more ordered, and
the microstates form a subset of the final
microstates after removing the partition.

𝑺𝑻 > 𝑺𝑳 + 𝑺𝑹 is understood

L R T

Irreversible Process
34 2/28/2023
 Ideal Classical gas
In case (II), the removal of the partition
microstates are the same as the should
not lead to new microstates (𝛀𝑻 = 𝛀𝑳 ∙ 𝛀𝑹 )

𝑺𝑻 = 𝑺𝑳 + 𝑺𝑹 (extensivity of S)

L R T

Rreversible Process
35 3/13/2023
 Ideal Classical gas
Examine our expression for the entropy
of two identical systems at equilibrium
𝟑 𝟒𝝅𝒎𝑬
𝑺 = 𝑵𝒌 𝐥𝐧 𝑽 + 𝑵𝒌 𝐥𝐧 𝟐
+𝟏
𝟐 𝟑𝑵𝒉

𝟑 𝟒𝝅𝒎𝑬
𝑺𝑳 = 𝑺𝑹 = 𝑵𝒌 𝐥𝐧 𝑽 + 𝑵𝒌 𝐥𝐧 𝟐
+𝟏
𝟐 𝟑𝑵𝒉

𝟑 𝟒𝝅𝒎 ∙ 𝟐𝑬
𝑺𝑻 = 𝟐𝑵𝒌 𝐥𝐧 𝟐𝑽 + 𝟐𝑵𝒌 𝐥𝐧 𝟐
+𝟏
𝟐 𝟑𝒉 ∙ 𝟐𝑵
36 2/28/2023
 Ideal Classical gas
This leads to higher entropy after the
removal of the partition

𝑺𝑻 = 𝑺𝑳 + 𝑺𝑹 + 𝟐𝑵𝒌 𝐥𝐧 𝟐

In direct contradiction with the second

law of thermodynamics (the entropy of a
system at equilibrium can be reduced by
inserting a partition!)

37 2/28/2023
 Ideal Classical gas
The paradox was a result of assuming
that the particles are distinguishable.

𝛀 ∝ 𝑽𝑵 L R
𝐥𝐧 𝛀𝑻 ∝ 𝟐𝑵 𝐥𝐧 𝟐𝑽
𝐥𝐧 𝛀𝑳 ∙ 𝛀𝑹 ∝ 𝟐𝑵 𝐥𝐧 𝑽

𝑺 𝜶𝑬, 𝜶𝑵, 𝜶𝑽 > 𝜶𝑺 𝑬, 𝑵, 𝑽 (paradoxical

result)
38 2/28/2023
 Ideal Classical gas
Gibbs correction factor

Gibbs proposed that the paradox can be

avoided by correctly enumerating the
microstates of the system (assuming
that the particles are indistinguishable)

L R T

39 2/28/2023
 Ideal Classical gas
The corrected number of states leads to
the correct Sackur-Tetrode Equation:
𝛀 𝑬, 𝑵, 𝑽
𝛀𝒄 𝑬, 𝑵, 𝑽 =
𝑵!

𝑺𝒄 𝑬, 𝑵, 𝑽 = 𝑺 − 𝑵𝒌 𝐥𝐧 𝑵 + 𝑵𝒌

𝑽 𝟑 𝟒𝝅𝒎𝑬 𝟓
𝑺 = 𝑵𝒌 𝐥𝐧 + 𝑵𝒌 𝐥𝐧 𝟐
+
𝑵 𝟐 𝟑𝑵𝒉 𝟑
40 2/28/2023
 Ideal Classical gas
Sackur-Tetrode Equation is consistent
with the extensivity of the entropy

𝑽 𝟑 𝟒𝝅𝒎𝑬 𝟓
𝑺 = 𝑵𝒌 𝐥𝐧 + 𝑵𝒌 𝐥𝐧 𝟐
+
𝑵 𝟐 𝟑𝑵𝒉 𝟑

𝑺(𝜶𝑵, 𝜶𝑽, 𝜶𝑬) = 𝜶𝑺(𝑵, 𝑽, 𝑬)

41 2/28/2023
 Ideal Classical gas
The thermodynamics of the system
follows immediately.

𝑽 𝟑 𝟒𝝅𝒎𝑬 𝟓
𝑺 = 𝑵𝒌 𝐥𝐧 + 𝑵𝒌 𝐥𝐧 𝟐
+
𝑵 𝟐 𝟑𝑵𝒉 𝟑

𝟐/𝟑
𝟑 𝑽 𝟑 𝟒𝝅𝒎𝑬 𝟓
𝑺 = 𝑵𝒌 𝐥𝐧 + 𝑵𝒌 𝐥𝐧 𝟐
+
𝟐 𝑵 𝟐 𝟑𝑵𝒉 𝟑

42 2/28/2023
 Ideal Classical gas
Simplify to determine the energy.

𝟑 𝟒𝝅𝒎𝑽𝟐/𝟑 𝟓
𝑺 = 𝑵𝒌 𝐥𝐧 𝟓/𝟑 𝟐
𝑬 +
𝟐 𝟑𝑵 𝒉 𝟑

𝟑𝑵𝟓/𝟑 𝒉𝟐 𝟐𝑺 𝟓
−
𝑬= 𝒆 𝟑𝑵𝒌 𝟑
𝟒𝝅𝒎𝑽 𝟐/𝟑

43 2/28/2023
 Ideal Classical gas
Now we determine the intensive
parameters and the thermodynamic
potentials. The general form of the first
law of thermodynamics allows exchange
of particles between the system and the
surroundings. Thus

𝒅𝑬 = 𝑻𝒅𝑺 − 𝑷𝒅𝑽 + 𝝁𝒅𝑵

44 2/28/2023
 Ideal Classical gas
Accordingly, the intensive parameters
(temperature T, pressure P, and the
Chemical potential 𝝁 are given by:

𝒅𝑬 = 𝑻𝒅𝑺 − 𝑷𝒅𝑽 + 𝝁𝒅𝑵

𝝏𝑬 𝝏𝑬 𝝏𝑬
→ = 𝑻; = −𝑷; =𝝁
𝝏𝑺 𝑵,𝑽
𝝏𝑽 𝑵,𝑺 𝝏𝑵 𝑺,𝑽

45 2/28/2023
 Ideal Classical gas
𝟑𝑵𝟓/𝟑 𝒉𝟐 𝟐𝑺 𝟓
−
𝑬= 𝒆 𝟑𝑵𝒌 𝟑
𝟒𝝅𝒎𝑽 𝟐/𝟑

𝝏𝑬 𝟐
𝑻= = 𝑬
𝝏𝑺 𝑵,𝑽
𝟑𝑵𝒌
𝟑
→ 𝑬 = 𝑵𝒌𝑻
𝟐

46 2/28/2023
 Ideal Classical gas
𝟑𝑵𝟓/𝟑 𝒉𝟐 𝟐𝑺 𝟓
−
𝑬= 𝒆 𝟑𝑵𝒌 𝟑
𝟒𝝅𝒎𝑽 𝟐/𝟑

𝝏𝑬 𝟐
−𝑷 = =− 𝑬
𝝏𝑽 𝑵,𝑺
𝟑𝑽
𝟐
→ 𝑷𝑽 = 𝑬 = 𝑵𝒌𝑻
𝟑

47 2/28/2023
 Ideal Classical gas
𝟑𝑵𝟓/𝟑 𝒉𝟐 𝟐𝑺 𝟓
−
𝑬= 𝒆 𝟑𝑵𝒌 𝟑
𝟒𝝅𝒎𝑽 𝟐/𝟑

𝝏𝑬 𝟓 𝟐𝑺
𝝁= = 𝑬− 𝟐
𝑬
𝝏𝑵 𝑽,𝑺
𝟑𝑵 𝟑𝑵 𝒌
𝟑 𝟓 𝟐𝑺
→ 𝝁 = 𝒌𝑻 −
𝟐 𝟑 𝟑𝑵𝒌
𝟑 𝟓 𝟐 𝟑 𝟒𝝅𝒎𝑽𝟐/𝟑 𝟓
= 𝒌𝑻 − 𝑵𝒌 𝐥𝐧 𝟓/𝟑 𝟐
𝑬 +
𝟐 𝟑 𝟑𝑵𝒌 𝟐 𝟑𝑵 𝒉 𝟑
48 3/13/2023
 Ideal Classical gas
Consequently, the chemical potential is:

𝟑 𝟒𝝅𝒎𝑽𝟐/𝟑
𝝁 = − 𝒌𝑻 𝐥𝐧 𝟓/𝟑 𝟐
𝑬
𝟐 𝟑𝑵 𝒉

𝟑 𝑵𝟐/𝟑 𝒉𝟐
𝝁 = 𝒌𝑻 𝐥𝐧
𝟐 𝟐𝝅𝒎𝑽𝟐/𝟑 𝒌𝑻
𝟑/𝟐
𝑵 𝒉𝟐
= 𝒌𝑻 𝐥𝐧
𝑽 𝟐𝝅𝒎𝒌𝑻
49 2/28/2023
 Ideal Classical gas
The rest of the thermodynamics is
straightforward. Thermodynamic
potentials (Helmholtz free energy F,
Gibbs free energy G, and the enthalpy H:

𝒅𝑬 = 𝑻𝒅𝑺 − 𝑷𝒅𝑽 + 𝝁𝒅𝑵

𝑭 = 𝑬 − 𝑻𝑺 → 𝒅𝑭 = −𝑷𝒅𝑽 − 𝑺𝒅𝑻 + 𝝁𝒅𝑵
𝑮 = 𝑭 + 𝑷𝑽 = 𝝁𝑵
𝑯 = 𝑬 + 𝑷𝑽 = 𝑮 + 𝑻𝑺

50 2/28/2023
 Ideal Classical gas
The heat capacity:

𝝏𝑺 𝝏𝑬
𝑪𝑽 = 𝑻 =
𝝏𝑻 𝑵.𝑽
𝝏𝑻 𝑵,𝑽
𝝏𝑺 𝝏 𝑬 + 𝑷𝑽 𝝏𝑯
𝑪𝑷 = 𝑻 = =
𝝏𝑻 𝑵,𝑷
𝝏𝑻 𝑵,𝑷
𝝏𝑻 𝑵,𝑷

51 2/28/2023
 Ideal Classical gas
The heat capacity:

𝝏𝑬 𝟑
𝑪𝑽 = = 𝑵𝒌
𝝏𝑻 𝑵,𝑽
𝟐

𝝏 𝑬 + 𝑷𝑽 𝝏 𝟓𝑵𝒌𝑻/𝟐 𝟓
𝑪𝑷 = = = 𝑵𝒌
𝝏𝑻 𝑵,𝑷
𝝏𝑻 𝑵,𝑷
𝟐

52 2/28/2023
 Ideal Classical gas
Ratio of the two specific heats is:

𝑪𝑷 𝟓
𝜸= =
𝑪𝑽 𝟑

53 2/28/2023
 Microcanonical Ensemble

54 Sat. Mech 2/28/2023