Chapter 0.

Review of thermodynamics

Without use of any microscopic properties, TDs

interrelates macroscopic properties, such as
pressure, volume, mass, and etc.

The basic notions of TDs inherit in:

0th law - new TDs quantity - temperature, 𝑇𝑇
1st law - new TDs quantity - internal energy, 𝑈𝑈
2nd law - new TDs quantity - entropy, 𝑆𝑆

1
Macroscopic properties are represented by TDs
variables, e.g.

• pressure, 𝑃𝑃, internal energy, 𝑈𝑈, temperature,

𝑇𝑇, ... , intensive: independent of mass, or
"local",

• mass, 𝑀𝑀, volume, 𝑉𝑉, ..., extensive: depend on

mass, or "whole system"

TDs equilibrium: state of system, when no

changes happen w.r.t time.
2
Reversible process: can be reverse by
infinitesimal changes (quasi-static, no hysteresis)
in conditions (e.g. 𝑃𝑃, 𝑉𝑉, 𝑇𝑇)

0th law: if system 1 and system 2 are separately

in equilibrium with system 3, system 1 and
system 2 are also in equilibrium.

1st law: 𝑑𝑑𝑑𝑑 = 𝛿𝛿𝛿𝛿 + 𝛿𝛿𝛿𝛿

𝛿𝛿𝛿𝛿 = heat transfer to the system
𝛿𝛿𝛿𝛿 = work done on the system

3
2nd law:
• Kelvin-Planck: No process is possible whose
sole result is the absorption of heat from a
reservoir and the complete conversion of this
heat into work.
• Clausius: No process is possible whose sole
result is the transfer of heat from a cooler to
a hotter body.

4
Chapter 1. Introduction
1.1 Introduction to statistical mechanics,
macroscopic and microscopic systems

To obtain detailed information concerning the

thermodynamics properties of systems without
experimental measurements, we required
calculations based on one of the two theories:
(i) kinetic theory
(ii) statistical mechanics (SM)

5
SM - study physical properties of macroscopic
system based on averaged effects of its
microscopic constituents, such as atoms.
Main purpose - formulate bulk properties of
matter from QM laws that governed individual
molecules.
A macroscopic system has dimension ≥ 1𝜇𝜇𝜇𝜇, e.g.
a dust particle, water droplet.
A microscopic system has dimensions of atoms.
E.g. a free Helium atom.

6
Quantum-state: state of atom described by its
energy (energy levels), angular momentum and
charge, which are quantized or discrete.

Microstate: state of macroscopic system, with

complete description on atomic scale.

Macrostate: specified by macroscopic observable

quantities: energy E, temperature T, volume V,
pressure P, magnetization M, and polarization P.

7
1.2 Basic probability concepts
Eg1: Ideal gas in boxes (4 atoms, distinguishable)

1 2 3 4

A B

Possible configurations or microstates:

8
1234 24 1 3
123 4 23 1 4
12 4 3 3 4 1 2
1 34 2 1 2 3 4
234 1 2 1 3 4
12 3 4 3 1 2 4
1 4 23 4 1 2 3
1 3 2 4 12 3 4

9
In summary:
A B Configurations Probabilities
or microstates
(i) 4 0 1 1/16
(ii) 3 1 4 4/16
(iii) 2 2 6 6/16
(iv) 1 3 4 4/16
(v) 0 4 1 1/16

Ω = 24 = 16 ( 𝑁𝑁 atoms, 𝑀𝑀 boxes: Ω = 𝑀𝑀𝑁𝑁 )

10
Combinations:

𝑛𝑛 atoms in A, �𝑁𝑁 – 𝑛𝑛� atoms in B:

𝑁𝑁!
Ω𝑛𝑛 = 𝐶𝐶 (𝑛𝑛) =
𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
Ω𝑛𝑛 𝐶𝐶 (𝑛𝑛) 1 𝑁𝑁!
Prob. : 𝑃𝑃𝑛𝑛 = = 𝑁𝑁 = 𝑁𝑁
Ω 2 2 𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!

States in (iii): 2atoms in A , 2atoms in B,

4! 𝐶𝐶 (2) 3
𝐶𝐶 (2) = = 6 and 𝑃𝑃2 = =
2!(4−2)! 24 8
12
 𝐶𝐶 (𝑛𝑛) = 𝐶𝐶 (𝑁𝑁 − 𝑛𝑛)

𝐶𝐶 (0) = 𝐶𝐶 (𝑁𝑁) = 1, 1! = 1, 0! = 1
𝐶𝐶 (𝑛𝑛 + 1) 𝑁𝑁! 𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
=
𝐶𝐶 (𝑛𝑛) (𝑛𝑛 + 1)! (𝑁𝑁 − 𝑛𝑛 − 1)! 𝑁𝑁!
𝑁𝑁 − 𝑛𝑛
=
𝑛𝑛 + 1

13
Eg2: Ideal gas in boxes (4 atoms, indistinguishable)

A B
Ω=5
11
1.3 Statistical ensemble
Consider collection of very large number of
exactly identical sub-systems.
This is macroscopic system or statistical
ensemble.

Eg. bar-magnet, water.

The sub-systems are still macroscopic, i.e. consist
of large number of molecules:

14
Suppose N measurements, say magnetization,
made on system, 𝑁𝑁𝑟𝑟 measurements reveal results
𝜇𝜇.

- - - -A - - - -A--A- - - A

N measurements

Prob. of 𝜇𝜇 is
𝑁𝑁𝑟𝑟
𝑃𝑃𝜇𝜇 =
𝑁𝑁

15
A statistical ensemble is independent of time or in
equilibrium if 𝑃𝑃𝜇𝜇 is obtained from repeated N
measurements.
N measurements → 𝑃𝑃𝜇𝜇
N measurements → 𝑃𝑃𝜇𝜇
⋮
An isolated macroscopic system is in statistical
equilibrium if its macroscopic sub-systems reveal
macroscopic quantities equal to their mean value.
Mean magnetization,

16
 𝑁𝑁
1 1
𝜇𝜇 = � 𝜇𝜇𝑖𝑖 = (𝜇𝜇1 + 𝜇𝜇2 + ⋯ + 𝜇𝜇𝑁𝑁 )
𝑁𝑁 𝑁𝑁
𝑖𝑖=1

If the system is in equilibrium, most of µi ≈ µ .

1.4 Basic probability theory

Events 1 2 3 ⋯ 𝑀𝑀
↕ ↕ ↕ ↕ ↕
Variables 𝑥𝑥1 𝑥𝑥2 𝑥𝑥3 ⋯ 𝑥𝑥𝑀𝑀

17
 ↕ ↕ ↕ ↕ ↕
Probabilities 𝑃𝑃1 𝑃𝑃2 𝑃𝑃3 ⋯ 𝑃𝑃𝑀𝑀

Variables 𝑥𝑥𝑖𝑖 = physical quantity for each event.

Normalization condition: ∑𝑖𝑖 𝑃𝑃𝑖𝑖 = 1

Mean value of 𝑥𝑥𝑖𝑖 :

𝑀𝑀
∑𝑀𝑀
𝑖𝑖=1 𝑃𝑃𝑖𝑖 𝑥𝑥𝑖𝑖
𝑥𝑥 = 〈𝑥𝑥 〉 = = � 𝑃𝑃𝑖𝑖 𝑥𝑥𝑖𝑖
∑𝑀𝑀
𝑖𝑖=1 𝑃𝑃𝑖𝑖 𝑖𝑖=1
18
Variance for 𝑥𝑥 is
(∆𝑥𝑥 )2 = (𝑥𝑥𝑖𝑖 − 𝑥𝑥 )2 = � 𝑃𝑃𝑖𝑖 (𝑥𝑥𝑖𝑖 − 𝑥𝑥 )2
𝑖𝑖
2 2
= � 𝑃𝑃𝑖𝑖 �𝑥𝑥𝑖𝑖 − 2𝑥𝑥𝑖𝑖 𝑥𝑥 + 𝑥𝑥 �
𝑖𝑖
2 2
= � 𝑃𝑃𝑖𝑖 𝑥𝑥𝑖𝑖 − 2𝑥𝑥 � 𝑃𝑃𝑖𝑖 𝑥𝑥𝑖𝑖 + 𝑥𝑥 � 𝑃𝑃𝑖𝑖
𝑖𝑖 𝑖𝑖 𝑖𝑖
2 2 2 2
= 𝑥𝑥 − 2𝑥𝑥𝑥𝑥 + 𝑥𝑥 = 𝑥𝑥 − 𝑥𝑥

19
1.5 Binomial distribution

Large number of independent measurements,

each has only 2 possible outcomes, prob.
distribution of outcomes =binomial distribution.

Eg.: tossing 50 cents coin, each trial can be flower

or kite:

𝑃𝑃𝐹𝐹 (flower) = p , 𝑃𝑃𝐾𝐾 (kite) =𝑞𝑞 = 1 − 𝑝𝑝

20
For N trials, a possible permutation of
𝑁𝑁𝐹𝐹 “ flower ” and 𝑁𝑁𝐾𝐾 “ kite" (*)

has probability

𝑝𝑝𝑁𝑁𝐹𝐹 𝑞𝑞 𝑁𝑁𝐾𝐾 = 𝑝𝑝𝑁𝑁𝐹𝐹 𝑞𝑞 𝑁𝑁−𝑁𝑁𝐹𝐹 = 𝑝𝑝𝑁𝑁−𝑁𝑁𝐾𝐾 𝑞𝑞 𝑁𝑁𝐾𝐾

Prob. for any combination of (*)is

21
 𝑃𝑃𝑁𝑁 (𝑁𝑁𝐾𝐾 ) = 𝐶𝐶 (𝑁𝑁𝐾𝐾 )𝑝𝑝𝑁𝑁−𝑁𝑁𝐾𝐾 𝑞𝑞 𝑁𝑁𝐾𝐾
𝑁𝑁!
= 𝑝𝑝𝑁𝑁−𝑁𝑁𝐾𝐾 𝑞𝑞 𝑁𝑁𝐾𝐾
(𝑁𝑁 − 𝑁𝑁𝐾𝐾 )! 𝑁𝑁𝐾𝐾 !
𝑁𝑁!
= 𝑝𝑝𝑁𝑁𝐹𝐹 𝑞𝑞 𝑁𝑁𝐾𝐾
𝑁𝑁𝐹𝐹 ! 𝑁𝑁𝐾𝐾 !

E.g. For 8 trials of tossing 50 cents coin, if 𝑝𝑝 = 0.3

, 𝑞𝑞 = 0.7 , the probability distribution is

8! 8( 0
( )
𝑃𝑃8 0 = ( ) )
0.3 0.7 = 0.00006561
8! 0!
22
 8!
𝑃𝑃8 (1) = (0.3)7 (0.7)1 = 0.00122472
7! 1!
8!
𝑃𝑃8 (2) = (0.3)6 (0.7)2 = 0.01000188
6! 2!
8!
𝑃𝑃8 (3) = (0.3)5 (0.7)3 = 0.04667544
5! 3!
8!
𝑃𝑃8 (4) = (0.3)4 (0.7)4 = 0.1361367
4! 4!
8!
𝑃𝑃8 (5) = (0.3)3 (0.7)5 = 0.25412184
3! 5!
8!
𝑃𝑃8 (6) = (0.3)2 (0.7)6 = 0.29647548
2! 6!
23
 8!
𝑃𝑃8 (7) = (0.3)1 (0.7)7 = 0.19765032
1! 7!
8!
𝑃𝑃8 (8) = (0.3)0 (0.7)8 = 0.05764801
0! 8!

24
25
 𝑁𝑁 𝑁𝑁
𝑁𝑁!
� 𝑃𝑃𝑁𝑁 (𝑁𝑁𝐾𝐾 ) = � 𝑝𝑝𝑁𝑁−𝑁𝑁𝐾𝐾 𝑞𝑞 𝑁𝑁𝐾𝐾
(𝑁𝑁 − 𝑁𝑁𝐾𝐾 )! 𝑁𝑁𝐾𝐾 !
𝑁𝑁𝐾𝐾 =0 ���������������������
𝑁𝑁 𝐾𝐾 =0
Binomial expansion
= (𝑝𝑝 + 𝑞𝑞 )𝑁𝑁 = 1

Challenge: For 3 trials of rolling a fair dice, work

out the probability distribution.

26
Answer: Let
𝑁𝑁𝑎𝑎 , 𝑎𝑎 = number and probability of 1 dot occurs,
𝑁𝑁𝑏𝑏 , 𝑏𝑏 =number and probability of 2 dots occurs,
𝑁𝑁𝑐𝑐 , 𝑐𝑐 =number and probability of 3 dots occurs,
𝑁𝑁𝑑𝑑 , 𝑑𝑑 = number and probability of 4 dots occurs,
𝑁𝑁𝑒𝑒 , 𝑒𝑒 = number and probability of 5 dots occurs,
𝑁𝑁𝑓𝑓 , 𝑓𝑓 = number and probability of six dots occurs,

𝑃𝑃3 (𝑁𝑁𝑎𝑎 ) = 𝐶𝐶 (𝑁𝑁𝑎𝑎 )𝑎𝑎𝑁𝑁𝑎𝑎 𝑏𝑏𝑁𝑁𝑏𝑏 𝑐𝑐 𝑁𝑁𝑐𝑐 𝑑𝑑 𝑁𝑁𝑑𝑑 𝑒𝑒 𝑁𝑁𝑒𝑒 𝑓𝑓 𝑁𝑁𝑓𝑓

3!
= Ω𝑎𝑎 𝑎𝑎𝑁𝑁𝑎𝑎 𝑏𝑏𝑁𝑁𝑏𝑏 𝑐𝑐 𝑁𝑁𝑐𝑐 𝑑𝑑 𝑁𝑁𝑑𝑑 𝑒𝑒 𝑁𝑁𝑒𝑒 𝑓𝑓 𝑁𝑁𝑓𝑓
𝑁𝑁𝑎𝑎 ! 𝑁𝑁𝑏𝑏 ! 𝑁𝑁𝑐𝑐 ! 𝑁𝑁𝑑𝑑 ! 𝑁𝑁𝑒𝑒 ! 𝑁𝑁𝑓𝑓 !

27
 1
𝑎𝑎 = 𝑏𝑏 = 𝑐𝑐 = 𝑑𝑑 = 𝑒𝑒 = 𝑓𝑓 =
6
3
3! 1
∴ 𝑃𝑃3 (𝑁𝑁𝑎𝑎 ) = Ω𝑎𝑎 � �
𝑁𝑁𝑎𝑎 ! 𝑁𝑁𝑏𝑏 ! 𝑁𝑁𝑐𝑐 ! 𝑁𝑁𝑑𝑑 ! 𝑁𝑁𝑒𝑒 ! 𝑁𝑁𝑓𝑓 ! 6

For 𝑵𝑵𝒂𝒂 = 𝟎𝟎 , 𝑁𝑁𝑏𝑏 + 𝑁𝑁𝑐𝑐 + 𝑁𝑁𝑑𝑑 + 𝑁𝑁𝑒𝑒 + 𝑁𝑁𝑓𝑓 = 3

5!
0+0+0+0+3=3 , Ω01 =
4! 1!
5!
0+0+1+1+1=3 , Ω02 =
2! 3!
5!
0+0+0+1+2=3 , Ω03 =
3! 1! 1!
28
 𝑃𝑃3 (𝑁𝑁𝑎𝑎 = 0)
3
5! 3! 1
=� � � �
4! 1! 0! 0! 0! 0! 0! 3! 6
3
5! 3! 1
+� � � �
2! 3! 0! 0! 0! 1! 1! 1! 6
3
5! 3! 1 125
+� � � � =
3! 1! 1! 0! 0! 0! 0! 1! 2! 6 216

For 𝑵𝑵𝒂𝒂 = 𝟏𝟏 , 𝑁𝑁𝑏𝑏 + 𝑁𝑁𝑐𝑐 + 𝑁𝑁𝑑𝑑 + 𝑁𝑁𝑒𝑒 + 𝑁𝑁𝑓𝑓 = 2

5!
0+0+0+0+2=2 , Ω11 =
4! 1!
29
 5!
0+0+0+1+1=2 , Ω12 =
3! 2!

𝑃𝑃3 (𝑁𝑁𝑎𝑎 = 1)
3
5! 3! 1
=� � � �
4! 1! 1! 0! 0! 0! 0! 2! 6
3
5! 3! 1 75
+� � � � =
3! 2! 1! 0! 0! 0! 1! 1! 6 216

For 𝑵𝑵𝒂𝒂 = 𝟐𝟐 , 𝑁𝑁𝑏𝑏 + 𝑁𝑁𝑐𝑐 + 𝑁𝑁𝑑𝑑 + 𝑁𝑁𝑒𝑒 + 𝑁𝑁𝑓𝑓 = 1

5!
0+0+0+0+1=1 , Ω21 =
4! 1!
30
 3
5! 3! 1 15
𝑃𝑃3 (𝑁𝑁𝑎𝑎 = 2) = � � � � =
4! 1! 2! 0! 0! 0! 0! 1! 6 216

For 𝑵𝑵𝒂𝒂 = 𝟑𝟑 , 𝑁𝑁𝑏𝑏 + 𝑁𝑁𝑐𝑐 + 𝑁𝑁𝑑𝑑 + 𝑁𝑁𝑒𝑒 + 𝑁𝑁𝑓𝑓 = 0

5!
0+0+0+0+0=0 , Ω31 =
5!
3
5! 3! 1 1
𝑃𝑃3 (𝑁𝑁𝑎𝑎 = 3) = � � � � =
5! 3! 0! 0! 0! 0! 0! 6 216

𝑃𝑃3 (0) + 𝑃𝑃3 (1) + 𝑃𝑃3 (2) + 𝑃𝑃3 (3) = 1

31
1.6 Stirling’s approximation
𝑁𝑁! = 𝑁𝑁 ∙ (𝑁𝑁 − 1) ∙ (𝑁𝑁 − 2) ∙∙∙ 3 ∙ 2 ∙ 1
ln(𝑁𝑁!)
= ln N + ln(𝑁𝑁 − 1) + ln(𝑁𝑁 − 2) + ⋯ + ln 3
𝑁𝑁

+ ln 2 + ln 1 = � ln(𝑚𝑚)
𝑚𝑚 =1

For large N,

32
 𝑁𝑁 𝑁𝑁
ln(𝑁𝑁!) = � ln(𝑚𝑚) ≈ � ln 𝑥𝑥 𝑑𝑑𝑑𝑑
𝑚𝑚 =1 1
[ ]𝑁𝑁
= 𝑥𝑥 ln 𝑥𝑥 − 𝑥𝑥 1 = 𝑁𝑁 ln 𝑁𝑁 − 𝑁𝑁 + 1

𝑁𝑁 ≫ 1,1 is neglected:

∴ ln(𝑁𝑁!) ≈ 𝑁𝑁 ln 𝑁𝑁 − 𝑁𝑁

𝑁𝑁! ≈ 𝑁𝑁 𝑁𝑁 𝑒𝑒 −𝑁𝑁

33
Graph ln(m)

34
1.7 Mean values and standard deviations
Eg: Ideal spin 12 paramagnet
1
Consider sys. of N localized spin magnetic
2
particle.
Magnetic moment of each particle:
𝑚𝑚𝑖𝑖 = +𝜇𝜇for spin up, ,prob. is p, or
𝑚𝑚𝑖𝑖 = −𝜇𝜇 for spin down, prob. is q

35
Find mean, variance, standard deviation, and
relative standard deviation of a particle and sys.
magnetic moment.
Solution:
For particle:
𝑁𝑁

𝑚𝑚 = 〈𝑚𝑚〉 = � 𝑚𝑚𝑖𝑖 𝑃𝑃(𝑚𝑚𝑖𝑖 )

𝑖𝑖=1
𝑁𝑁
𝑃𝑃(𝑚𝑚𝑖𝑖 ) = prob. of𝑚𝑚𝑖𝑖 , ∑𝑖𝑖=1 𝑃𝑃(𝑚𝑚𝑖𝑖 ) =1

Variance (dispersion) with respect to the mean is,

36
 2
(∆𝑚𝑚)2 = 𝑚𝑚2 − 𝑚𝑚
2
(∆𝑚𝑚)2 or 〈(∆𝑚𝑚 )2 〉 ≥ 0, ∴ 𝑚𝑚2 ≥ 𝑚𝑚

For system:
𝑁𝑁 𝑁𝑁

𝑀𝑀 = 〈𝑀𝑀〉 = � 𝑚𝑚𝑖𝑖 = � 𝑚𝑚𝑖𝑖 = 𝑁𝑁𝑚𝑚

𝑖𝑖=1 𝑖𝑖=1

𝑀𝑀 = 𝑚𝑚1 + 𝑚𝑚2 + ⋯ + 𝑚𝑚𝑁𝑁 = � 𝑚𝑚𝑖𝑖

𝑖𝑖

37
 2
Variance is (∆𝑀𝑀)2 = 𝑀𝑀2 − 𝑀𝑀
In terms of𝑚𝑚𝑖𝑖 ,

𝑀𝑀2 = (𝑚𝑚1 + ⋯ 𝑚𝑚𝑖𝑖 + ⋯ 𝑚𝑚𝑁𝑁 )�𝑚𝑚1 + ⋯ 𝑚𝑚𝑗𝑗 + ⋯ 𝑚𝑚𝑁𝑁 �

= � 𝑚𝑚𝑖𝑖 � 𝑚𝑚𝑗𝑗 = � � 𝑚𝑚𝑖𝑖 𝑚𝑚𝑗𝑗

𝑖𝑖 𝑗𝑗 𝑖𝑖 𝑗𝑗
𝑁𝑁 𝑁𝑁 𝑁𝑁

= � 𝑚𝑚𝑖𝑖2 + � � 𝑚𝑚𝑖𝑖 𝑚𝑚𝑗𝑗

𝑖𝑖=𝑗𝑗 =1 𝑖𝑖=1 𝑗𝑗 =1 (𝑖𝑖≠𝑗𝑗 )

38
 𝑁𝑁 𝑁𝑁 𝑁𝑁 𝑁𝑁
2
𝑀𝑀 = � 𝑚𝑚𝑖𝑖 � 𝑚𝑚𝑗𝑗 = � 𝑚𝑚𝑖𝑖 � 𝑚𝑚𝑗𝑗
𝑖𝑖=1 𝑗𝑗 =1 𝑖𝑖=1 𝑗𝑗 =1
𝑁𝑁 𝑁𝑁 𝑁𝑁

( ) 2
= � 𝑚𝑚𝑖𝑖 + � � 𝑚𝑚𝑖𝑖 𝑚𝑚𝑗𝑗
𝑖𝑖=𝑗𝑗 =1 𝑖𝑖=1 𝑗𝑗 =1 (𝑖𝑖≠𝑗𝑗 )
𝑁𝑁 𝑁𝑁
2
∴ ∆𝑀𝑀 = � 𝑚𝑚𝑖𝑖 − � (𝑚𝑚𝑖𝑖 )2 = 𝑁𝑁𝑚𝑚2 − 𝑁𝑁(𝑚𝑚)2
( )2

𝑖𝑖=𝑗𝑗 =1 𝑖𝑖=𝑗𝑗 =1
= 𝑁𝑁 �𝑚𝑚2 − (𝑚𝑚)2 � = 𝑁𝑁(∆𝑚𝑚)2

39
Standard deviation, ∆∗ 𝑥𝑥, is defined as square root
of variance.

∆∗ 𝑚𝑚 = �(∆𝑚𝑚)2

∆∗ 𝑀𝑀 = �(∆𝑀𝑀)2 = √𝑁𝑁∆∗ 𝑚𝑚

Comparison between ∆∗ 𝑥𝑥 and𝑥𝑥 gives prob.

distribution.

Small ∆∗ 𝑥𝑥: prob. distribution displays sharp peak

about𝑥𝑥.
40
Large ∆∗ 𝑥𝑥: prob. distribution displays broad peak
about𝑥𝑥.

Relative standard deviation:

∆ 𝑚𝑚∗ �(∆𝑚𝑚)2
=
𝑚𝑚 𝑚𝑚
∗
∆ 𝑀𝑀 �(∆𝑀𝑀)2 1 ∆∗ 𝑚𝑚
= =
𝑀𝑀 𝑀𝑀 √𝑁𝑁 𝑚𝑚

41
 1
System of N localized spin magnetic particles,
2
with P(spin-up) = p, and P(spin-down) = q,
𝑚𝑚 = 𝑝𝑝𝜇𝜇0 − 𝑞𝑞𝜇𝜇0 = 𝑝𝑝𝜇𝜇0 − (1 − 𝑝𝑝)𝜇𝜇0 = (2𝑝𝑝 − 1)𝜇𝜇0
2 2 2
𝑚𝑚2 = 𝑝𝑝𝜇𝜇0 + 𝑞𝑞𝜇𝜇0 = 𝜇𝜇0

2 2 2 2
(∆𝑚𝑚)2 = 𝑚𝑚2 − 𝑚𝑚 = 𝜇𝜇0 ( )
− 2𝑝𝑝 − 1 𝜇𝜇0
= 4𝑝𝑝(1 − 𝑝𝑝)𝜇𝜇02 = 4𝑝𝑝𝑝𝑝 𝜇𝜇02

(∆𝑀𝑀)2 = 𝑁𝑁(∆𝑚𝑚)2 = 4𝑁𝑁𝑁𝑁𝑁𝑁 𝜇𝜇02

42
 ∗ 2
2�𝑝𝑝𝑝𝑝
∆ 𝑚𝑚 1 �4𝑝𝑝𝑝𝑝 𝜇𝜇0
= �(∆𝑚𝑚)2 = =
𝑚𝑚 𝑚𝑚 (2𝑝𝑝 − 1)𝜇𝜇0 (2𝑝𝑝 − 1)

If 𝑝𝑝 ≠ 𝑞𝑞,
∆∗ 𝑀𝑀 1 ∆∗ 𝑚𝑚 1 2�𝑝𝑝𝑝𝑝
= =
𝑀𝑀 √𝑁𝑁 𝑚𝑚 √𝑁𝑁 (2𝑝𝑝 − 1)

43
Eg: Distribution of atoms in ideal gas
Consider an ideal gas with 𝑁𝑁 atoms in container
with volume 𝑉𝑉0 .
If container is divided into two parts, with
′ ( ′
volumes 𝑉𝑉 and 𝑉𝑉 𝑉𝑉 + 𝑉𝑉 = 𝑉𝑉0 ).
Prob. of atoms in𝑉𝑉 is 𝑝𝑝 ,
Prob. of atoms in𝑉𝑉 ′ is𝑞𝑞, 𝑝𝑝 + 𝑞𝑞 = 1.

Find relative standard deviation for atoms

distribution.
44
 ′
𝑉𝑉

(𝑁𝑁 − 𝑛𝑛)atoms
𝑉𝑉

𝑛𝑛 atoms

45
Solution:
𝑁𝑁!
𝑃𝑃(𝑛𝑛) = 𝑝𝑝𝑛𝑛 𝑞𝑞 (𝑁𝑁−𝑛𝑛 )
𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
𝑛𝑛=𝑁𝑁 𝑛𝑛=𝑁𝑁
𝑛𝑛 𝑁𝑁!
𝑛𝑛� = � 𝑛𝑛 𝑃𝑃(𝑛𝑛) = � 𝑝𝑝𝑛𝑛 𝑞𝑞 (𝑁𝑁−𝑛𝑛 )
𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
𝑛𝑛=0 𝑛𝑛=0
𝑛𝑛=𝑁𝑁
𝜕𝜕 𝑁𝑁! 𝑛𝑛 (𝑁𝑁−𝑛𝑛 )
= 𝑝𝑝 � � 𝑝𝑝 𝑞𝑞 �
𝜕𝜕𝜕𝜕 𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
𝑛𝑛=0
𝜕𝜕 𝑁𝑁 𝑁𝑁−1
= 𝑝𝑝 ( ) (
𝑝𝑝 + 𝑞𝑞 = 𝑝𝑝 𝑁𝑁 𝑝𝑝 + 𝑞𝑞 ) = 𝑁𝑁𝑁𝑁
𝜕𝜕𝜕𝜕
46
 𝑛𝑛=𝑁𝑁 𝑛𝑛=𝑁𝑁
2
��� 𝑛𝑛 𝑁𝑁!
2
𝑛𝑛2 = � 𝑛𝑛 𝑃𝑃(𝑛𝑛) = � 𝑝𝑝𝑛𝑛 𝑞𝑞 (𝑁𝑁−𝑛𝑛 )
𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
𝑛𝑛=0 𝑛𝑛=0
𝑛𝑛=𝑁𝑁
𝜕𝜕 𝜕𝜕 𝑁𝑁! 𝑛𝑛 (𝑁𝑁−𝑛𝑛 )
= 𝑝𝑝 �𝑝𝑝 � � 𝑝𝑝 𝑞𝑞 ��
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!
𝑛𝑛=0
𝜕𝜕 𝜕𝜕
= 𝑝𝑝 �𝑝𝑝 (𝑝𝑝 + 𝑞𝑞 )𝑁𝑁 �
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
𝜕𝜕
= 𝑝𝑝 [𝑝𝑝 𝑁𝑁 (𝑝𝑝 + 𝑞𝑞 )𝑁𝑁−1 ]
𝜕𝜕𝜕𝜕
= 𝑝𝑝𝑝𝑝[(𝑝𝑝 + 𝑞𝑞 )𝑁𝑁−1 + 𝑝𝑝(𝑁𝑁 − 1)(𝑝𝑝 + 𝑞𝑞 )𝑁𝑁−2 ]
2 2 2 2 2
= 𝑁𝑁𝑁𝑁 + 𝑁𝑁 𝑃𝑃 − 𝑁𝑁𝑃𝑃 = 𝑁𝑁 𝑃𝑃 + 𝑁𝑁𝑁𝑁(1 − 𝑝𝑝)
47
 ��������
(∆𝑛𝑛)2 = �����������
(𝑛𝑛 − 𝑛𝑛�)2 = ���
𝑛𝑛2 − (𝑛𝑛�)2 = 𝑁𝑁𝑁𝑁(1 − 𝑝𝑝)
= 𝑁𝑁𝑁𝑁𝑁𝑁

∆ 𝑛𝑛 = �(��������
∗
∆𝑛𝑛)2 = �𝑁𝑁𝑁𝑁𝑁𝑁

Relative standard deviation for 𝑛𝑛:

∆∗ 𝑛𝑛 �𝑁𝑁𝑁𝑁𝑁𝑁 1 𝑞𝑞
= = �
𝑛𝑛� 𝑁𝑁𝑁𝑁 √𝑁𝑁 𝑝𝑝
′ 1 1 ∆∗ 𝑛𝑛 1
Notes: If 𝑉𝑉 = 𝑉𝑉 = 𝑉𝑉0 , 𝑝𝑝 = 𝑞𝑞 = , =
2 2 𝑛𝑛� √𝑁𝑁

48
For large 𝑁𝑁, e.g.𝑁𝑁 = 1024 ,
∆∗ 𝑛𝑛 1
= = 1012 , 𝑛𝑛� = 𝑁𝑁𝑁𝑁 = 0.5 × 1024
𝑛𝑛� √𝑁𝑁

For small𝑁𝑁, e.g.𝑁𝑁 = 18,

∆∗ 𝑛𝑛 1 1
= ≈ ≈ 0.24 , 𝑛𝑛� = 𝑁𝑁𝑁𝑁 = 9
𝑛𝑛� √𝑁𝑁 4.243
For 𝑁𝑁 = 36,
∗
∆ 𝑛𝑛 1 1
= = ≈ 0.17 , 𝑛𝑛� = 𝑁𝑁𝑁𝑁 = 18
𝑛𝑛� √36 6
49
For𝑁𝑁 = 170,
∗
∆ 𝑛𝑛 1 1
= ≈ ≈ 0.08 , 𝑛𝑛� = 𝑁𝑁𝑁𝑁 = 85
𝑛𝑛� √170 13.04

50
 Probability for n atoms in V, P(n), N = 18

0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

51
 Probability for n atoms in V, P(n), N = 36

0.14

0.12

0.1

0.08

0.06

0.04

0.02

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37

52
 Probability of n atoms in V, P(n), N = 170

0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
1 11 21 31 41 51 61 71 81 91 10111112113114115116117

53
Chapter 2 Determination of the quantum states
2.1 Quantum states
In QM, microstate (= quantum state) of atom
(microscopic system) is described by
wavefunction 𝜙𝜙(𝑥𝑥, 𝑡𝑡).
Alternatively, quantum state can be described by
energy and momentum.
In QM, energy and momentum are represented by
operators (Hamiltonian act on wavefunction).

1
Isolated microscopic system (e.g. atom), in space
(e.g. box), spatial and temporal behaviours are
governed by Schrodinger eqn and BCs:

𝜕𝜕𝜕𝜕(𝑥𝑥, 𝑡𝑡)
𝑖𝑖ℏ = 𝐻𝐻𝐻𝐻(𝑥𝑥, 𝑡𝑡)
𝜕𝜕𝜕𝜕
ℏ2 2
𝐻𝐻 = − ∇ + 𝑉𝑉
2𝑚𝑚
𝑉𝑉 is zero if no interaction and potential energy.

E.g an atom in a box, BC is 𝜙𝜙(𝑥𝑥𝑏𝑏 , 𝑡𝑡) = 0.

2
System in equilibrium, quantum state
independent of time (stationary quantum state):
Schrodinger eqn reduce to

𝐻𝐻𝐻𝐻(𝑥𝑥 ) = 𝐸𝐸𝐸𝐸(𝑥𝑥 )

Usually, more than one possible wavefunctions:

𝜓𝜓1 , 𝜓𝜓2 , 𝜓𝜓3 , ⋯ , 𝜓𝜓𝑗𝑗
with energies
𝐸𝐸1 , 𝐸𝐸2 , 𝐸𝐸3 , ⋯ , 𝐸𝐸𝑗𝑗
satisfy Schrodinger eqn.
3
𝜓𝜓1 , 𝜓𝜓2 , 𝜓𝜓3 , ⋯ , 𝜓𝜓𝑗𝑗 : eigenfunctions,
𝐸𝐸1 , 𝐸𝐸2 , 𝐸𝐸3 , ⋯ , 𝐸𝐸𝑗𝑗 : eigenvalues (eigen energies)
System can be in any possible states, represented
by 𝜓𝜓𝑗𝑗 and 𝐸𝐸𝑗𝑗 .
𝐸𝐸𝑗𝑗 are discrete energy levels.

If 𝑔𝑔𝑗𝑗 quantum states with same energy 𝐸𝐸𝑗𝑗 , this

energy level has 𝑔𝑔𝑗𝑗 degeneracy.

4
Energy

𝐸𝐸𝑗𝑗 𝜓𝜓𝑗𝑗 1 𝜓𝜓𝑗𝑗 2 𝜓𝜓𝑗𝑗 3 

⋮
𝐸𝐸2 𝜓𝜓21 𝜓𝜓22 𝜓𝜓23 𝜓𝜓24

𝐸𝐸1 𝜓𝜓11 𝜓𝜓12

𝐸𝐸0 𝜓𝜓0
degenerate
energy levels
Ground state
5
Eg 1: Consider three spin ½ magnetic particles
B

µ0 µ0 − µ0
− µ0 B − µ0 B µ0 B
magnetic moment Energy for each
for each particle particle =−μ ⋅ B

th 1
State of 𝑖𝑖 particle is 𝑚𝑚𝑖𝑖 = ± .
2
State of system is: 𝑚𝑚1 𝑚𝑚2 𝑚𝑚3 .
6
 1 1
Let 𝑚𝑚𝑖𝑖 = + → + , 𝑚𝑚𝑖𝑖 = − → −
2 2
All possible states (microstates) of system:
m1 m2 m3 ∑mag. mom. ∑energy
1 +++ 3𝜇𝜇0 −3𝜇𝜇0 𝐵𝐵

degenerate
2 ++− +𝜇𝜇0 −𝜇𝜇0 𝐵𝐵
3 +−+ +𝜇𝜇0 −𝜇𝜇0 𝐵𝐵
4 −++ +𝜇𝜇0 −𝜇𝜇0 𝐵𝐵
5 +−− −𝜇𝜇0 +𝜇𝜇0 𝐵𝐵

degenerate
6 −+− −𝜇𝜇0 +𝜇𝜇0 𝐵𝐵
7 −−+ −𝜇𝜇0 +𝜇𝜇0 𝐵𝐵
8 −−− −3𝜇𝜇0 +3𝜇𝜇0 𝐵𝐵
7
Eg 2: A particle in 1-D box

V ( x)

0 , 0 < x < L
V ( x) = 
∞ , x ≤ 0 or x ≥ L

0 L x
Schrodinger eqn. of particle for 0 < 𝑥𝑥 < 𝐿𝐿 :
8
 𝐻𝐻𝐻𝐻(𝑥𝑥 ) = 𝐸𝐸𝐸𝐸(𝑥𝑥 )
or
d ψ ( x)
2
2mE
2
=− 2 ψ ( x) =
−k ψ ( x)
2
(1)
dx 
2mE
k = 2
2
(2)

ψ ( x) A sin kx + B cos kx
General solution of (1): =

Applying BCs :
ψ (= = 0 ⇒ B=0
x 0)
9
 ψ (=
x L=
) 0 ⇒ A sin kL = 0
For nontrivial solution, A ≠ 0, sin kL = 0,
nπ , n = 1, 2, ( discrete )
∴ kL =

nπ
=
k → kn , A → An
L
k2 2
nπ 2 2 2
(4)
E → En= =
n
2
2m 2mL
Possible states of particle :
10
 ψ n ( x) A=
n sin k n x , n 1, 2, (5)
Superposition of all possible states:
∞
ψ ( x) = ∑ An sin kn x general solution of (1).
n =1

An : determined by normalization condition:

∞

∫ ψ ψ = (11)
*
( x ) ( x ) dx 1
−∞

 ∞
 ∞

L
∴ ∫  ∑ Anψ n ( x)   ∑ Asψ s ( x)  dx =
* *
1
0  n 1= s 1 
11
 ∞ ∞
 nπ x   sπ x 
L
⇒ ∑∑ A As ∫ sin 
*
n  sin   dx =
1
=
n 1 =s 1  L   L 

0
I
L
For n ≠ s , I = zero; for n = s , I =
2
∞ ∞
L 2
⇒ ∑ An An*
= 1, ∑ An An =
*

n =1 2 n =1 L
or A A + A A2 +  =
*
1 1
*
2 2 L
A single particle can occupy only one possible
state at any instant, so An = 2 L .
12
Possible quantum states is

2 nπ x 2
k nπ
2 2 2 2
ψ n ( x) = sin , =
En = n
2
L L 2m 2mL

13
Eg 3: Single particle in 3-D box
z

Lz

0 y
Ly

x Lx

14
V (r ) = 0 inside 3D box
V (r ) = ∞ on the boundary and outside box
2

Schrodinger eqn.: − ∇ ψ (r ) =
2
Eψ (r ) (1)
2m
∂ ψ (r ) ∂ ψ (r ) ∂ ψ (r )
2 2 2
+ + = − k ψ (r )
2
(2)
∂x 2
∂y 2
∂z 2

2mE
k = 2 . BCs ψ (r ) = 0 at surface of the box.
2

15
(2) is 2nd order PDE, solved by separation of
variables: assumed
ψ (r ) = P( x)Q( y ) R( z ) (4)
Substitute (4) into (2),
𝜕𝜕 2 𝑃𝑃(𝑥𝑥 ) 𝜕𝜕 2 𝑄𝑄(𝑦𝑦)
𝑄𝑄(𝑦𝑦)𝑅𝑅(𝑧𝑧) + 𝑃𝑃 ( 𝑥𝑥 )𝑅𝑅 ( 𝑧𝑧 )
𝜕𝜕𝑥𝑥 2 𝜕𝜕𝑦𝑦 2
𝜕𝜕 2 𝑅𝑅(𝑧𝑧) 2 ( ) ( ) ( )
+ 𝑃𝑃(𝑥𝑥 )𝑄𝑄(𝑦𝑦) 2
= −𝑘𝑘 𝑃𝑃 𝑥𝑥 𝑄𝑄 𝑦𝑦 𝑅𝑅 𝑧𝑧
𝜕𝜕𝑥𝑥
Divide by 𝑃𝑃(𝑥𝑥 )𝑄𝑄(𝑦𝑦)𝑅𝑅(𝑧𝑧):
16
 1 𝜕𝜕 2 𝑃𝑃(𝑥𝑥 ) 1 𝜕𝜕 2 𝑄𝑄(𝑦𝑦) 1 𝜕𝜕 2 𝑅𝑅(𝑧𝑧)
+ +
𝑃𝑃(𝑥𝑥 ) 𝜕𝜕𝑥𝑥 2 𝑄𝑄 (𝑦𝑦) 𝜕𝜕𝑦𝑦 2 𝑅𝑅 (𝑧𝑧) 𝜕𝜕𝑥𝑥 2
2
= −𝑘𝑘
2
1 𝜕𝜕 𝑃𝑃(𝑥𝑥 ) 2
= −𝑘𝑘𝑥𝑥
𝑃𝑃(𝑥𝑥 ) 𝜕𝜕𝑥𝑥 2

1 𝜕𝜕 2 𝑄𝑄(𝑦𝑦) 2
= −𝑘𝑘 𝑦𝑦
𝑄𝑄 (𝑦𝑦) 𝜕𝜕𝑦𝑦 2

1 𝜕𝜕 2 𝑅𝑅(𝑧𝑧) 2
= −𝑘𝑘𝑧𝑧
𝑅𝑅(𝑧𝑧) 𝜕𝜕𝑥𝑥 2
17
 𝑘𝑘𝑥𝑥2 + 𝑘𝑘𝑦𝑦2 + 𝑘𝑘𝑧𝑧2 = 𝑘𝑘 2
General solutions are

𝑃𝑃(𝑥𝑥 ) = 𝐴𝐴𝑥𝑥 cos 𝑘𝑘𝑥𝑥 𝑥𝑥 + 𝐵𝐵𝑥𝑥 sin 𝑘𝑘𝑥𝑥 𝑥𝑥

𝑄𝑄 (𝑦𝑦) = 𝐶𝐶𝑦𝑦 cos 𝑘𝑘𝑦𝑦 𝑦𝑦 + 𝐷𝐷𝑦𝑦 sin 𝑘𝑘𝑦𝑦 𝑦𝑦
𝑅𝑅(𝑧𝑧) = 𝐹𝐹𝑧𝑧 cos 𝑘𝑘𝑧𝑧 𝑧𝑧 + 𝐺𝐺𝑧𝑧 sin 𝑘𝑘𝑧𝑧 𝑧𝑧

Apply BCs on the results yields

∞

𝑃𝑃(𝑥𝑥 ) = � 𝐵𝐵𝑥𝑥𝑥𝑥 sin 𝑘𝑘𝑥𝑥𝑥𝑥 𝑥𝑥

𝑛𝑛=1
18
 ∞

𝑄𝑄 (𝑦𝑦) = � 𝐷𝐷𝑦𝑦𝑦𝑦 sin 𝑘𝑘𝑦𝑦𝑦𝑦 𝑦𝑦

𝑠𝑠=1
∞

𝑅𝑅 (𝑧𝑧) = � 𝐺𝐺𝑧𝑧𝑧𝑧 sin 𝑘𝑘𝑧𝑧𝑧𝑧 𝑧𝑧

𝑣𝑣=1

nπ sπ vπ
=
where k xn = , k ys = , k zv
Lx Ly Lz
n = 1, 2,, s = 1, 2, and v = 1, 2, ,
19
 k →k
2 2
nsv = k +k +k
2
xn
2
ys
2
zv


π n 2 2
s 2 2
v 2 
and E →=
Ensv  2 + 2 + 2  discrete
2m  Lx Ly Lz ,
Possible states :
nπ x sπ y vπ z
ψ nsv ( x, y, z ) = Fnsv sin sin sin
Lx Ly Lz
Superposition of all states :

20
 ∞ ∞
nπ x
∞
sπ y vπ z
ψ (r ) = ∑∑∑ Fnsv sin sin sin
=
n 1 =s 1 =
v 1 Lx Ly Lz
n, s , v ≠ 0 .
Fnsv is determined by normalization condition:
∗( ) ( ) 3
� 𝜓𝜓 𝐫𝐫 𝜓𝜓 𝐫𝐫 𝑑𝑑 𝑟𝑟 = 1
over all space
or
∞ ∞ ∞

∫ ∫ ∫ ψ ( x, y, z ) ψ ( x, y, z ) dxdydz = 1
*

x = −∞ y = −∞ z = −∞

21
∴ Fnsv =8V
8 nπ x sπ y vπ z
∴ψ nsv ( x, y, z ) = sin sin sin
V Lx Ly Lz

π n 22 2
s 2
v 2 
=
Ensv  2 + 2 + 2 
2m  Lx Ly Lz 

If L=
x L=y L=
z L (cube),
3 π
2 2
for ψ 1,1,1 , =
E1,1,1 = 2
E1
2mL
22
 3ℏ2 𝜋𝜋 2
for ψ 1,1,2 , ψ 1,2,1 , ψ 2,1,1 , 𝐸𝐸1,1,2 =
𝑚𝑚 𝐿𝐿2
For E1 , g1 = 1 ; E2 , g 2 = 3
For N non-interacting atoms in 3-D box, states of
system is determined by sets of quantum
numbers:

𝑛𝑛1 , 𝑠𝑠1 , 𝑣𝑣1 , 𝑛𝑛2 , 𝑠𝑠2 , 𝑣𝑣2 , ⋯ , 𝑛𝑛𝑁𝑁 , 𝑠𝑠𝑁𝑁 , 𝑣𝑣𝑁𝑁

23
2.2 Phase space and classical states
Consider a mac. system with s degrees of
freedom:
i.e. position of molecules in space are described
by s coordinates, 𝑞𝑞𝑖𝑖 , i = 1, 2,, s .
Note: 1-D system with N molecules, s = N ,
3-D system with M molecules, s = 3M .
Classical state is defined by s space coordinates 𝑞𝑞𝑖𝑖
and 𝑠𝑠 corresponding momenta 𝑝𝑝𝑖𝑖 .

24
Every system has its own phase space, with
dimensions equal to 2s.
Any set ( q1 , q2 ,  , qs , p1 , p2 ,  , ps ) , or point in
phase space (phase point), represents a particular
state of system.
State of system changes with time, the
corresponding phase point moves along a curve
called phase trajectory.

25
Consider an isolated mac. system. A very small
part of the system but still mac., is called a
subsystem.
Subsystem interacts with other parts of system.
These interactions are very complex an intricate.
Complicated manner of subsystem’s state
variation allows statistical approach to the
solution.

26
Let ∆p∆q denote a small “volume” of phase space
of subsystem in between of
(𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 , 𝑝𝑝1 , 𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 )
and
(𝑞𝑞1 + ∆𝑞𝑞1 , 𝑞𝑞2 + ∆𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 + ∆𝑞𝑞𝑠𝑠 ;
𝑝𝑝1 + ∆𝑝𝑝1 , 𝑝𝑝2 + ∆𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 + ∆𝑝𝑝𝑠𝑠 )

with
∆𝑝𝑝∆𝑞𝑞 = ∆𝑞𝑞1 ∆𝑞𝑞2 ⋯ ∆𝑞𝑞𝑠𝑠 ∆𝑝𝑝1 ∆𝑝𝑝2 ⋯ ∆𝑝𝑝𝑠𝑠
= (∆𝑝𝑝1 ∆𝑞𝑞1 )(∆𝑝𝑝2 ∆𝑞𝑞2 ) ⋯ (∆𝑝𝑝𝑠𝑠 ∆𝑞𝑞𝑠𝑠 )

27
28
This “volume” represents all possible states or
phase points inside it.
During a sufficiently long time 𝑡𝑡, subsystem will
be many times in every possible state.
Let ∆t be duration of subsystem in ∆p∆q . As t
increases indefinitely: =
w lim ∆t t
t →∞

𝑤𝑤 = prob. for subsystem in ∆𝑝𝑝∆𝑞𝑞.

Taking infinitesimal limit of phase volume:

29
 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑞𝑞1 𝑑𝑑𝑞𝑞2 ⋯ 𝑑𝑑𝑞𝑞𝑠𝑠 𝑑𝑑𝑝𝑝1 𝑑𝑑𝑝𝑝2 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
= (𝑑𝑑𝑝𝑝1 𝑑𝑑𝑞𝑞1 )(𝑑𝑑𝑝𝑝2 𝑑𝑑𝑞𝑞2 ) ⋯ (𝑑𝑑𝑝𝑝𝑠𝑠 𝑑𝑑𝑞𝑞𝑠𝑠 ) = lim ℎ0𝑠𝑠
ℎ 0 →0

Prob. of states in between

(𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 , 𝑝𝑝1 , 𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 )
and
(𝑞𝑞1 + 𝑑𝑑𝑞𝑞1 , 𝑞𝑞2 + 𝑑𝑑𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 + 𝑑𝑑𝑞𝑞𝑠𝑠 ;
𝑝𝑝1 + 𝑑𝑑𝑝𝑝1 , 𝑝𝑝2 + 𝑑𝑑𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 + 𝑑𝑑𝑝𝑝𝑠𝑠 )
is
𝑑𝑑𝑑𝑑 = 𝜌𝜌(𝑝𝑝1 , 𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 , 𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 )𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑

30
 𝜌𝜌(𝑝𝑝1 , 𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 , 𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 )
is statistical distribution function (probability
density) in phase space.

This function satisfy normalization condition

31
 � 𝜌𝜌(𝑝𝑝1 , 𝑝𝑝2 , ⋯ 𝑝𝑝𝑠𝑠 , 𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 )𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑
all phase
space

= � 𝑑𝑑𝑑𝑑 = 1
all phase
space

Over a sufficiently long time, effects of initial

conditions of subsystems will be entirely
overweighed by the effects of much larger
remaining parts of the system.
32
Statistical distribution of subsystem does not
depend on initial conditions of any subsystems.
Without solving mechanical problems
(Schrodinger eqns), statistical distribution for
small parts of system can be found.
Probabilities of various physical quantities which
depend on the states (𝑝𝑝𝑖𝑖 , 𝑞𝑞𝑖𝑖 ), 𝑖𝑖 = 1,2, ⋯ , 𝑠𝑠 of
subsystem can be calculated.
Note: mac. system considered is also a subsystem
of a larger closed system.
33
Mean value of physical quantities are obtained by
multiplying each possible values by the
corresponding probability and integrating over all
states:

𝑓𝑓 ̅ = 〈𝑓𝑓 〉 = � 𝑓𝑓(𝑝𝑝, 𝑞𝑞 )𝜌𝜌(𝑝𝑝, 𝑞𝑞 )𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑

all phasespace
or all possible states
or

34
=
f =
f ∫
all possible states
f ( p1 ,, qs ) ρ ( p1 ,, qs )dp1  dqs
or all phase space

35
2.3 Isolated system in Equilibrium:
“Postulate of equal a priori probabilities (PPP)”

Consider an isolated system. In equilibrium it is

specified by macroscopic variables (𝐸𝐸, 𝑉𝑉, 𝑁𝑁).

For non-equilibrium macrostates, additional α

must be specified.

Each 𝛼𝛼 corresponds a Ω(𝐸𝐸, 𝑉𝑉, 𝑁𝑁, 𝛼𝛼 ), number of

microstates comprising the macrostate.

34
Eg. in a gas, due to collisions, microstate and
hence 𝛼𝛼 and Ω(𝐸𝐸, 𝑉𝑉, 𝑁𝑁, 𝛼𝛼 ) change continually.

No preference of one microstate to another, so

must assign equal probabilities to them:

PPP: for an isolated system, all microstates

compatible with system’s constraints (𝐸𝐸, 𝑉𝑉, 𝑁𝑁)
are equally likely to occur.

Prob. of system in macrostate specified by

(𝐸𝐸, 𝑉𝑉, 𝑁𝑁, 𝛼𝛼) is proportional to Ω(𝐸𝐸, 𝑉𝑉, 𝑁𝑁, 𝛼𝛼 ).
35
Equilibrium state of system, with (𝐸𝐸, 𝑉𝑉, 𝑁𝑁) fixed,
is corresponding to maximum Ω(𝐸𝐸, 𝑉𝑉, 𝑁𝑁, 𝛼𝛼).

Thus equilibrium state is the state of maximum

probability.

Eg 4: system with three spin ½ magnetic

particles.
Assume system is isolated. If 𝐸𝐸 = −𝜇𝜇0 𝐵𝐵, system
can be in any of these states:
(+ + −) , (+ − +) , (− + +)
36
 �⃗
𝐵𝐵

𝜇𝜇0 𝜇𝜇0 −𝜇𝜇0

−𝜇𝜇0 𝐵𝐵 −𝜇𝜇0 𝐵𝐵 +𝜇𝜇0 𝐵𝐵

Magnetic moment Energy for each
for each particle partticle
B

37
PPP presses that when system is in equilibrium,
probabilities for finding it in
(+ + −) , (+ − +) , or (− + +)
are equal.

2.4 Calculation of probability and density of states

Consider isolated system in equilibrium, energy

of system within 𝐸𝐸 to 𝐸𝐸 + 𝛿𝛿𝛿𝛿.

38
 𝐸𝐸 + 𝛿𝛿𝛿𝛿
δE
𝐸𝐸

Ω = total number of attainable states.

Ω : function of E , defined as

Ω(𝐸𝐸 ) = 𝑊𝑊 (𝐸𝐸 )𝛿𝛿𝛿𝛿 (*)

39
𝑊𝑊 (𝐸𝐸 ) = density of states (DOS).

= characteristic function of system

= number of attainable states per unit

energy.

𝜙𝜙(𝐸𝐸 ) = total number of states with energy < 𝐸𝐸.

𝑑𝑑𝑑𝑑(𝐸𝐸 )
Ω(𝐸𝐸 ) = 𝜙𝜙(𝐸𝐸 + 𝛿𝛿𝛿𝛿 ) − 𝜙𝜙(𝐸𝐸 ) = 𝛿𝛿𝛿𝛿
𝑑𝑑𝑑𝑑
(**)
40
Compare (*) and (**) :
𝑑𝑑𝑑𝑑(𝐸𝐸 )
𝑊𝑊 (𝐸𝐸 ) =
𝑑𝑑𝑑𝑑

Eg: Single particle in a cube, 𝑉𝑉 = 𝐿𝐿3 . Eigen

energies of particle are:

ℏ2 𝜋𝜋 2 2 2 2)
𝐸𝐸𝑛𝑛𝑛𝑛𝑛𝑛 = ( 𝑛𝑛 + 𝑠𝑠 + 𝑣𝑣
2𝑚𝑚𝐿𝐿2
2
2𝑚𝑚𝐿𝐿
⟹ 𝑛𝑛2 + 𝑠𝑠 2 + 𝑣𝑣 2 = 2 2 𝐸𝐸𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑅𝑅2
ℏ 𝜋𝜋
41
 𝐿𝐿
∴ 𝑅𝑅 = �2𝑚𝑚𝐸𝐸𝑛𝑛𝑛𝑛𝑛𝑛 , 𝑛𝑛, 𝑠𝑠, 𝑣𝑣 = 1,2, ⋯
𝜋𝜋ℏ

In first octant with radius 𝑅𝑅 ,

3 3
1 4𝜋𝜋𝑅𝑅 1 4𝜋𝜋 𝐿𝐿
𝜙𝜙(𝐸𝐸 ) = � � = � � � � (2𝑚𝑚𝑚𝑚 )3/2
8 3 8 3 𝜋𝜋ℏ
3
𝜋𝜋 𝐿𝐿
= � � � � (2𝑚𝑚𝑚𝑚 )3⁄2
6 𝜋𝜋ℏ

42
 𝑑𝑑𝑑𝑑(𝐸𝐸 )
Ω(𝐸𝐸 ) = 𝛿𝛿𝛿𝛿
𝑑𝑑𝑑𝑑
3
𝜋𝜋 𝐿𝐿 ⁄ 3
= � � � � (2𝑚𝑚)3 2 � � √𝐸𝐸 𝛿𝛿𝛿𝛿
6 𝜋𝜋ℏ 2
𝑉𝑉
= � 2 3 � (2𝑚𝑚)3⁄2 √𝐸𝐸 𝛿𝛿𝛿𝛿 , 𝑉𝑉 = 𝐿𝐿3
4𝜋𝜋 ℏ

𝑉𝑉
𝑊𝑊 (𝐸𝐸 ) = � 2 3 � (2𝑚𝑚)3⁄2 √𝐸𝐸
4𝜋𝜋 ℏ

43
Chapter 3 Thermal interactions of
macroscopic systems

3.1 Thermal interactions of macroscopic systems

Consider 2 isolated macroscopic systems being

combined, and in weak thermal interactions:

A + A' = A = A + A′
(0)

E E′ E (0)
≈ E + E′ Figure 3.1
1
Before thermal contact:

Ω(𝐸𝐸 ) = number of attainable microstates of 𝐴𝐴

with energy 𝐸𝐸 to 𝐸𝐸 + 𝛿𝛿𝛿𝛿 .

Ω′ (𝐸𝐸 ′ ) = number of attainable microstates of 𝐴𝐴′

with energy 𝐸𝐸 ′ to 𝐸𝐸 ′ + 𝛿𝛿𝐸𝐸 ′ .

With weak thermal interactions, they form a

(0)
combine isolated system 𝐴𝐴 , total energy
𝐸𝐸 (0) ≈ 𝐸𝐸 + 𝐸𝐸 ′ = constant.

2
 ′ ′ (0)
Energy of 𝐴𝐴 : 𝐸𝐸 = 𝐸𝐸 − 𝐸𝐸 (1)

Ω(0) (𝐸𝐸 ) = number of attainable microstates of

(0)
𝐴𝐴 when 𝐴𝐴 is within energy range 𝐸𝐸 to 𝐸𝐸 + 𝛿𝛿𝛿𝛿.

Note: In thermal interactions, energy of A is not

restricted to 𝐸𝐸 to 𝐸𝐸 + 𝛿𝛿𝛿𝛿.

PPP presses that: when 𝐴𝐴(0) in equilibrium

(0) (0)
(𝐸𝐸 constant), 𝐴𝐴 has equal probabilities in any
of its attainable microstates.
3
 (0)
Probability of 𝐴𝐴 when A within energy 𝐸𝐸 to
𝐸𝐸 + 𝛿𝛿𝛿𝛿 :

𝑃𝑃(𝐸𝐸 ) = 𝐶𝐶 Ω(0) (𝐸𝐸 ) (2)

1
𝐶𝐶 = (0)
Ωtotal
(0) (0)
Ωtotal
= total attainable microstates for 𝐴𝐴 (with
(0)
energy 𝐸𝐸 ).

4
Each microstate of 𝐴𝐴 combined with each
′ (0)
microstate of 𝐴𝐴 to form microstates of 𝐴𝐴 ,
Ω(0) (𝐸𝐸 ) = Ω(𝐸𝐸 )Ω′ (𝐸𝐸 ′ ) = Ω(𝐸𝐸 )Ω′ �𝐸𝐸 (0) − 𝐸𝐸�

′ (0)
∴ 𝑃𝑃(𝐸𝐸 ) = 𝐶𝐶 Ω(𝐸𝐸 )Ω �𝐸𝐸 − 𝐸𝐸� (3)

( ) ( ) ′ (0)
ln 𝑃𝑃 𝐸𝐸 = ln 𝐶𝐶 + ln Ω 𝐸𝐸 + ln Ω �𝐸𝐸 − 𝐸𝐸� (4)

First order derivative of (4) is

5
 ′ (0)
𝜕𝜕 ln 𝑃𝑃(𝐸𝐸 ) 𝜕𝜕 ln 𝐶𝐶 𝜕𝜕 ln Ω(𝐸𝐸 ) 𝜕𝜕 ln Ω �𝐸𝐸 − 𝐸𝐸�
= + +
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
1 𝜕𝜕𝜕𝜕(𝐸𝐸 ) 1 𝜕𝜕Ω(𝐸𝐸 ) 1 𝜕𝜕Ω′ (𝐸𝐸 ′ )
⟹ = + ′ ′
𝑃𝑃(𝐸𝐸 ) 𝜕𝜕𝜕𝜕 Ω(𝐸𝐸 ) 𝜕𝜕𝜕𝜕 Ω (𝐸𝐸 ) 𝜕𝜕𝜕𝜕

(0)
In equilibrium, 𝐴𝐴 is in the microstates with
maximum probability.

When 𝑃𝑃(𝐸𝐸 ) attain its maximum at 𝐸𝐸 = 𝐸𝐸� ,

6
 𝜕𝜕𝜕𝜕(𝐸𝐸 )
� =0
𝜕𝜕𝜕𝜕 𝐸𝐸=𝐸𝐸�
or
1 𝜕𝜕Ω(𝐸𝐸 ) 1 𝜕𝜕Ω′ (𝐸𝐸 ′ )
� + ′ ′ � = 0 (5a)
Ω(𝐸𝐸 ) 𝜕𝜕𝜕𝜕 𝐸𝐸=𝐸𝐸� Ω (𝐸𝐸 ) 𝜕𝜕𝜕𝜕 𝐸𝐸=𝐸𝐸�
or
1 𝜕𝜕Ω(𝐸𝐸 ) 1 𝜕𝜕Ω′ (𝐸𝐸 ′ ) 𝜕𝜕𝐸𝐸 ′
� + ′ ′ �
Ω(𝐸𝐸 ) 𝜕𝜕𝜕𝜕 𝐸𝐸=𝐸𝐸� Ω (𝐸𝐸 ) 𝜕𝜕𝐸𝐸 ′ 𝜕𝜕𝜕𝜕 𝐸𝐸 ′ =𝐸𝐸� ′ (5b)
=0

7
From (1),
𝜕𝜕𝐸𝐸 ′
= −1
𝜕𝜕𝜕𝜕
(5b) becomes
1 𝜕𝜕Ω(𝐸𝐸 ) 1 𝜕𝜕Ω′ (𝐸𝐸 ′ )
� − ′ ′ � =0 (6)
Ω(𝐸𝐸 ) 𝜕𝜕𝜕𝜕 𝐸𝐸=𝐸𝐸� Ω (𝐸𝐸 ) 𝜕𝜕𝐸𝐸 ′
𝐸𝐸 ′ =𝐸𝐸� ′

where 𝐸𝐸 ′ = 𝐸𝐸 (0) − 𝐸𝐸 , 𝐸𝐸� ′ = 𝐸𝐸 (0) − 𝐸𝐸� .

Let

8
 1 𝜕𝜕Ω(𝐸𝐸 )
𝛽𝛽 (𝐸𝐸 ) = (7)
Ω(𝐸𝐸 ) 𝜕𝜕𝜕𝜕

′( ′)
′( ′)
1 𝜕𝜕Ω 𝐸𝐸
𝛽𝛽 𝐸𝐸 = ′ ′ (8)
Ω (𝐸𝐸 ) 𝜕𝜕𝐸𝐸 ′

� ′ � ′
When 𝐸𝐸 = 𝐸𝐸 and 𝐸𝐸 = 𝐸𝐸 , numbers of microstates
for 𝐴𝐴 and 𝐴𝐴′ are maximum, and hence
� ′ �′
𝛽𝛽�𝐸𝐸 � = 𝛽𝛽 �𝐸𝐸 � (9)

9
Note: 𝛽𝛽 has dimension of inverse energy.

3.2 Absolute temperature and entropy

Let 𝑇𝑇 and 𝑇𝑇 ′ defined as

1 ′
1
𝑘𝑘𝐵𝐵 𝑇𝑇 = , 𝑘𝑘𝐵𝐵 𝑇𝑇 = ′
𝛽𝛽 𝛽𝛽

From (7) and (8),

10
 1 1 𝜕𝜕Ω(𝐸𝐸 ) 𝜕𝜕 ln Ω(𝐸𝐸 )
= =
𝑘𝑘𝐵𝐵 𝑇𝑇 Ω(𝐸𝐸 ) 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

1 1 𝜕𝜕Ω′ (𝐸𝐸 ′ ) 𝜕𝜕 ln Ω′ (𝐸𝐸 ′ )

= ′ ′ =
𝑘𝑘𝐵𝐵 𝑇𝑇 ′ Ω (𝐸𝐸 ) 𝜕𝜕𝐸𝐸 ′ 𝜕𝜕𝜕𝜕′
Then
1 𝜕𝜕{𝑘𝑘𝐵𝐵 ln Ω(𝐸𝐸 )} 𝜕𝜕𝜕𝜕
= =
𝑇𝑇 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

1 𝜕𝜕{𝑘𝑘𝐵𝐵 ln Ω′(𝐸𝐸′)} 𝜕𝜕𝜕𝜕′

= =
𝑇𝑇′ 𝜕𝜕𝜕𝜕′ 𝜕𝜕𝜕𝜕′
11
where 𝑆𝑆 = 𝑘𝑘𝐵𝐵 ln Ω(𝐸𝐸 ) , 𝑆𝑆 ′ = 𝑘𝑘𝐵𝐵 ln Ω′(𝐸𝐸′),

𝑘𝑘𝐵𝐵 = Boltzmann constant.

𝑇𝑇 and 𝑇𝑇 ′ are temperatures ;

𝑆𝑆 and 𝑆𝑆 ′ are entropies

From (4),
( ) ( ) ′( ′)
𝑘𝑘𝐵𝐵 ln 𝑃𝑃 𝐸𝐸 = 𝑘𝑘𝐵𝐵 ln 𝐶𝐶 + 𝑘𝑘𝐵𝐵 ln Ω 𝐸𝐸 + 𝑘𝑘𝐵𝐵 ln Ω 𝐸𝐸
= 𝑘𝑘𝐵𝐵 ln 𝐶𝐶 + 𝑆𝑆 + 𝑆𝑆 ′

12
 � ′ � ′
When 𝐸𝐸 = 𝐸𝐸 and 𝐸𝐸 = 𝐸𝐸 , 𝑃𝑃(𝐸𝐸 ) attains its
maximum, i.e. 𝑃𝑃�𝐸𝐸� � and 𝑆𝑆 + 𝑆𝑆 ′ are maximum.
From (9),
𝛽𝛽�𝐸𝐸� � = 𝛽𝛽 ′ �𝐸𝐸� ′ �
or
1 1 ′
= or 𝑇𝑇 = 𝑇𝑇
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇 ′

In other words, when isolated system is in

equilibrium, probability of its macrostate 𝑃𝑃(𝐸𝐸 ) is
maximum with absolute temperature 𝑇𝑇.
13
3.3 Second law of thermodynamics
In thermodynamics,
𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑 = = increment of entropy
𝑇𝑇
𝑑𝑑𝑑𝑑 = total heat absorbed at 𝑇𝑇 .

2nd law of thermodynamic:

(i) Entropy 𝑆𝑆 is a state function, its value is

unique for every macrostate, and was
14
 independent of the previous states of the
system
(ii) Entropy of an isolated system cannot be
decreased, i.e. ∆𝑆𝑆 > 0 .

3.4 Properties of entropy

Consider 2 systems (same temperatures &
pressures), are combined, with weak thermal
interaction.
S1 S2 S12= S1 + S 2
Ω1 + Ω2 = Ω = Ω1Ω 2
15
 Figure 3.2
Entropy of combined system:
𝑆𝑆 ∝ ln Ω , ln Ω = ln Ω1 + ln Ω2
3rd law of thermodynamics: 𝑇𝑇 → 0 , 𝑆𝑆 → 0
𝑆𝑆 = 𝑘𝑘𝐵𝐵 ln Ω
If Ω = 1 , ln Ω = 0 , 𝑆𝑆 = 0 .

Entropy is referred as disorder of system.

3.5 Properties of micro-canonical ensemble

16
Micro-canonical ensemble: consist of isolated
systems with same number of molecules 𝑁𝑁,
energy 𝐸𝐸, and volume 𝑉𝑉 .
All thermodynamics relations can be derived from
SM based on micro-canonical ensemble :

(i) Calculate number of attainable microstates, Ω,

(ii) Determine entropy of system: 𝑆𝑆 = 𝑘𝑘𝐵𝐵 ln Ω .

(iii) Derive 𝐸𝐸 and physical quantities in terms of 𝑆𝑆
and 𝑉𝑉 .
17
From
1 𝑃𝑃
𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑑𝑑 + 𝑑𝑑𝑑𝑑
𝑇𝑇 𝑇𝑇
1 𝜕𝜕𝜕𝜕
=� �
𝑇𝑇 𝜕𝜕𝜕𝜕 𝑉𝑉
𝑃𝑃 𝜕𝜕𝜕𝜕
=� �
𝑇𝑇 𝜕𝜕𝜕𝜕 𝐸𝐸

𝑆𝑆 = 𝑘𝑘𝐵𝐵 ln Ω

18
 1 𝜕𝜕{𝑘𝑘𝐵𝐵 ln Ω} 1 𝜕𝜕 ln Ω
=� � ⟹ =� �
𝑇𝑇 𝜕𝜕𝜕𝜕 𝑉𝑉
𝑘𝑘 𝐵𝐵 𝑇𝑇 𝜕𝜕𝜕𝜕 𝑉𝑉

𝑃𝑃 𝜕𝜕{𝑘𝑘𝐵𝐵 ln Ω} 𝑃𝑃 𝜕𝜕 ln Ω
=� � ⟹ =� �
𝑇𝑇 𝜕𝜕𝜕𝜕 𝐸𝐸
𝑘𝑘 𝐵𝐵 𝑇𝑇 𝜕𝜕𝜕𝜕 𝐸𝐸

Helmholtz free energy, 𝐹𝐹 = 𝐸𝐸 − 𝑇𝑇𝑇𝑇 ,

𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑑𝑑 − 𝑇𝑇𝑇𝑇𝑇𝑇 − 𝑆𝑆𝑆𝑆𝑆𝑆

= 𝑑𝑑𝑑𝑑 − (𝑑𝑑𝑑𝑑 + 𝑃𝑃𝑃𝑃𝑃𝑃 ) − 𝑆𝑆𝑆𝑆𝑆𝑆
= −𝑃𝑃𝑃𝑃𝑃𝑃 − 𝑆𝑆𝑆𝑆𝑆𝑆

19
 𝜕𝜕𝜕𝜕
⟹ 𝑃𝑃 = − � � = pressure
𝜕𝜕𝜕𝜕 𝑇𝑇
𝜕𝜕𝜕𝜕
⟹ 𝑆𝑆 = − � � = entropy
𝜕𝜕𝜕𝜕 𝑉𝑉

Eg: Consider system of 𝑁𝑁 spin ½ magnetic

�⃗.
particles, with applied field 𝐵𝐵
��⃗ of the system in terms of 𝑇𝑇 and 𝐵𝐵
Derive 𝐸𝐸 and 𝑀𝑀 �⃗.

�⃗
Solution: energy = −𝜇𝜇⃗𝑖𝑖 ∙ 𝐵𝐵
�⃗
𝐵𝐵
20
 𝑛𝑛 𝑁𝑁 − 𝑛𝑛
𝜇𝜇 −𝜇𝜇
−𝜇𝜇𝜇𝜇 +𝜇𝜇𝜇𝜇
Magnetic moment Energy for each
for each particle partticle
𝑁𝑁 𝑁𝑁
�⃗ = −𝐵𝐵 � 𝜇𝜇𝑖𝑖
𝐸𝐸 = − � 𝜇𝜇⃗𝑖𝑖 ∙ 𝐵𝐵
(1)
𝑖𝑖=1 𝑖𝑖=1
= −𝐵𝐵{𝑛𝑛𝑛𝑛 + (𝑁𝑁 − 𝑛𝑛)(−𝜇𝜇)} = (𝑁𝑁 − 2𝑛𝑛)𝜇𝜇𝜇𝜇

21
 1 𝐸𝐸
∴ 𝑛𝑛 = �𝑁𝑁 − � (2a)
2 𝜇𝜇𝜇𝜇
1 𝐸𝐸
𝑁𝑁 − 𝑛𝑛 = �𝑁𝑁 + � (2b)
2 𝜇𝜇𝜇𝜇

𝑁𝑁!
Ω= (3)
𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!

22
 ln Ω = ln 𝑁𝑁! − ln 𝑛𝑛! − ln(𝑁𝑁 − 𝑛𝑛)! (4)

With Stirling’s approximation,

ln Ω = (𝑁𝑁 ln 𝑁𝑁 − 𝑁𝑁) − (𝑛𝑛 ln 𝑛𝑛 − n)

− {(𝑁𝑁 − 𝑛𝑛) ln(𝑁𝑁 − 𝑛𝑛) − (𝑁𝑁 − 𝑛𝑛)} (5)
= 𝑁𝑁 ln 𝑁𝑁 − 𝑛𝑛 ln 𝑛𝑛 − (𝑁𝑁 − 𝑛𝑛) ln(𝑁𝑁 − 𝑛𝑛)

Substitute (2) into (5),

23
 1 𝐸𝐸 1 𝐸𝐸
ln Ω = 𝑁𝑁 ln 𝑁𝑁 − �𝑁𝑁 − � ln � �𝑁𝑁 − ��
2 𝜇𝜇𝜇𝜇 2 𝜇𝜇𝜇𝜇
1 𝐸𝐸 1 𝐸𝐸 (6)
− �𝑁𝑁 + � ln � �𝑁𝑁 + ��
2 𝜇𝜇𝜇𝜇 2 𝜇𝜇𝜇𝜇

𝐸𝐸
1 𝜕𝜕 ln Ω 1 𝑁𝑁 −
𝜇𝜇𝜇𝜇
=� � = ln � �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝜕𝜕𝜕𝜕 𝑉𝑉 2𝜇𝜇𝜇𝜇 𝐸𝐸
𝑁𝑁 +
𝜇𝜇𝜇𝜇

24
 𝐸𝐸
2𝜇𝜇𝜇𝜇 𝑁𝑁 −
𝜇𝜇𝜇𝜇
= ln � �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝐸𝐸
𝑁𝑁 +
𝜇𝜇𝜇𝜇
𝐸𝐸
2𝜇𝜇𝜇𝜇 𝑁𝑁 −
𝜇𝜇𝜇𝜇
exp � �=
𝑘𝑘𝐵𝐵 𝑇𝑇 𝐸𝐸
𝑁𝑁 +
𝜇𝜇𝜇𝜇
2𝜇𝜇𝜇𝜇
1 − exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇
𝐸𝐸 = 𝑁𝑁𝑁𝑁𝑁𝑁
2𝜇𝜇𝜇𝜇
1 + exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇
25
 2𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
1 − exp � � exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
𝐸𝐸 = 𝑁𝑁𝑁𝑁𝑁𝑁 � �
2𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
1 + exp � � exp �− 𝑘𝑘 𝑇𝑇�
𝑘𝑘𝐵𝐵 𝑇𝑇 𝐵𝐵
𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
exp �− � − exp � � (7)
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
= 𝑁𝑁𝑁𝑁𝑁𝑁
𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
exp �− � + exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
𝜇𝜇𝜇𝜇
= −𝑁𝑁𝑁𝑁𝑁𝑁 tanh � �
𝑘𝑘𝐵𝐵 𝑇𝑇

From (1),
26
 𝑁𝑁 𝑁𝑁
�⃗ = −𝐵𝐵 � 𝜇𝜇𝑖𝑖 = −𝐵𝐵𝐵𝐵
𝐸𝐸 = − � 𝜇𝜇⃗𝑖𝑖 ∙ 𝐵𝐵
𝑖𝑖=1 𝑖𝑖=1

𝐸𝐸 𝜇𝜇𝜇𝜇
∴ 𝑀𝑀 = − = 𝑁𝑁𝑁𝑁 tanh � � (8)
𝐵𝐵 𝑘𝑘𝐵𝐵 𝑇𝑇

27
𝑀𝑀

𝜇𝜇𝜇𝜇
𝑘𝑘 𝐵𝐵 𝑇𝑇

28
Eg: Consider a system with two energy levels:

ε2 𝑛𝑛 particles
ε
ε1 𝑁𝑁 – 𝑛𝑛 particles

Derive energy 𝐸𝐸 and heat capacity 𝐶𝐶𝑉𝑉 in terms of

𝑇𝑇 when the system is in equilibrium.

29
Answer:

𝐸𝐸 = 𝑛𝑛𝜀𝜀2 + (𝑁𝑁 − 𝑛𝑛)𝜀𝜀1 = 𝑛𝑛𝜀𝜀2 + 𝑁𝑁𝜀𝜀1 − 𝑛𝑛𝜀𝜀1

= 𝑁𝑁𝜀𝜀1 + 𝑛𝑛𝑛𝑛 (1)

𝜀𝜀 = 𝜀𝜀2 − 𝜀𝜀1 = energy gap

𝑁𝑁!
Ω= , 𝑆𝑆 = 𝑘𝑘𝐵𝐵 ln Ω
𝑛𝑛! (𝑁𝑁 − 𝑛𝑛)!

30
Use Stirling’s approximation:

ln Ω = 𝑁𝑁 ln 𝑁𝑁 − 𝑛𝑛 ln 𝑛𝑛 − (𝑁𝑁 − 𝑛𝑛) ln(𝑁𝑁 − 𝑛𝑛)

∴ 𝑆𝑆 = 𝑘𝑘𝐵𝐵 {𝑁𝑁 ln 𝑁𝑁 − 𝑛𝑛 ln 𝑛𝑛 − (𝑁𝑁 − 𝑛𝑛) ln(𝑁𝑁 − 𝑛𝑛)}

Helmholtz free energy of system :

𝐹𝐹 = 𝐸𝐸 − 𝑇𝑇𝑇𝑇
= 𝑁𝑁𝜀𝜀1 + 𝑛𝑛𝑛𝑛
− 𝑘𝑘𝐵𝐵 𝑇𝑇 {𝑁𝑁 ln 𝑁𝑁 − 𝑛𝑛 ln 𝑛𝑛 − (𝑁𝑁 − 𝑛𝑛) ln(𝑁𝑁 − 𝑛𝑛)}

31
In equilibrium, 𝐹𝐹 is minimum:
𝜕𝜕𝜕𝜕
= 0 ⟹ 𝜀𝜀 − 𝑘𝑘𝐵𝐵 𝑇𝑇 {ln(𝑁𝑁 − 𝑛𝑛) − ln 𝑛𝑛} = 0
𝜕𝜕𝜕𝜕
𝑁𝑁 − 𝑛𝑛 𝜀𝜀
⟹ ln � �=
𝑛𝑛 𝑘𝑘𝐵𝐵 𝑇𝑇

𝑁𝑁
⟹ 𝑛𝑛 = 𝜀𝜀 (2)
1 + exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇

32
(2) in (1) gives
𝑁𝑁𝑁𝑁
𝐸𝐸 = 𝑁𝑁𝜀𝜀1 + 𝜀𝜀
1 + exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇

𝜀𝜀
𝜕𝜕𝜕𝜕 𝑁𝑁𝜀𝜀 2 exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇
𝐶𝐶𝑉𝑉 = � � = � 2
� 2
𝜕𝜕𝜕𝜕 𝑉𝑉 𝑘𝑘𝐵𝐵 𝑇𝑇 𝜀𝜀
�1 + exp � ��
𝑘𝑘𝐵𝐵 𝑇𝑇
𝑁𝑁 𝑘𝑘𝐵𝐵 𝜃𝜃 2 𝑒𝑒 𝜃𝜃 𝜀𝜀
= , 𝜃𝜃 =
(1 + 𝑒𝑒 𝜃𝜃 )2 𝑘𝑘𝐵𝐵 𝑇𝑇

33
𝐶𝐶𝑉𝑉 ⁄𝑁𝑁𝑘𝑘𝐵𝐵
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0 2 4 6 8 10

θ 34
Chapter 4 Equilibrium of system in heat
bath: Canonical ensemble

4.1 Canonical ensemble

Micro-canonical ensemble: ensemble of sub-

systems, all with the same 𝑁𝑁, 𝐸𝐸, 𝑉𝑉, 𝑀𝑀, 𝑃𝑃, shape
and etc.
Note: 𝑀𝑀 = magnetization, 𝑃𝑃 = polarization.

1
Canonical ensemble: ensemble of sub-systems, all
with the same N, V, M, P, shape and etc, but
energy of each sub-system is not constant.

In Canonical ensemble, each sub-system is in

thermal contact with the remainder sub-systems
which acts as heat bath.

Energy pass from sub-systems to neighbours:

energy of sub-systems fluctuate.

2
Any system in contact with a heat bath: it is said
to be in canonical state (becomes sub-system to
combined system).

Heat bath has large heat capacity: its

temperature does not change when heat is added
to, or extracted from it.
The system and heat bath eventually reach
thermal equilibrium at the temperature of heat
bath.

3
In other words, canonical ensemble has constant
absolute 𝑇𝑇, but 𝐸𝐸 of its sub-systems fluctuates.

Consider a small system placed in heat bath at 𝑇𝑇,

combined system is isolated with 𝑁𝑁 and 𝑉𝑉 fixed.
𝐴𝐴 = combined system = 𝐴𝐴1 + 𝐴𝐴2
Energy of 𝐴𝐴, 𝐸𝐸0 = 𝐸𝐸𝑟𝑟 + 𝐸𝐸2 = constant .
𝐴𝐴1 = small system, 𝐴𝐴2 = heat bath, ∴ 𝐴𝐴 ≈ 𝐴𝐴2 .
𝐸𝐸2 = energy of heat bath 𝐴𝐴2 , 𝐸𝐸2 ≈ 𝐸𝐸0 .

4
 𝑇𝑇
𝐴𝐴1

𝐸𝐸𝑟𝑟
𝐸𝐸2 𝐴𝐴2

In equilibrium, macrostate of 𝐴𝐴1 is specified by

𝑇𝑇, 𝑉𝑉1 , 𝑁𝑁1 .
𝐴𝐴1 possess discrete microstates 1,2, ⋯ , 𝑟𝑟, ⋯,
with energies 𝐸𝐸1 , 𝐸𝐸2 , ⋯ , 𝐸𝐸𝑟𝑟 , ⋯.
5
Energy will not suffice to specify a unique
microstate, because many microstates possess
the same energy (degenerate).
When 𝐴𝐴1 has energy 𝐸𝐸𝑟𝑟 , 𝐸𝐸2 ’s energy is
𝐸𝐸2 = 𝐸𝐸0 − 𝐸𝐸𝑟𝑟 .
Total attainable microstates for 𝐴𝐴1 is Ω1 (𝐸𝐸𝑟𝑟 ) .
Total attainable microstates for 𝐴𝐴2 is
Ω2 (𝐸𝐸2 ) = Ω2 (𝐸𝐸0 − 𝐸𝐸𝑟𝑟 )

Total attainable microstates for 𝐴𝐴 is Ω(𝐸𝐸0 ) :

6
 Ω(𝐸𝐸0 ) = Ω1 (𝐸𝐸𝑟𝑟 ) × Ω2 (𝐸𝐸0 − 𝐸𝐸𝑟𝑟 )

According to PPP, when 𝐴𝐴 is in equilibrium,

probability for 𝐴𝐴1 in a definite state (with energy
𝐸𝐸𝑟𝑟 ) is

𝑃𝑃(𝐸𝐸𝑟𝑟 ) ∝ 1 × Ω2 (𝐸𝐸0 − 𝐸𝐸𝑟𝑟 ) (1)

𝐴𝐴1 ≪ 𝐴𝐴2 ⟹ 𝐸𝐸𝑟𝑟 ≪ 𝐸𝐸2 ⟹ 𝐸𝐸𝑟𝑟 ≪ 𝐸𝐸0

⟹ Taylor expansion of ln Ω2 :
7
ln Ω2 (𝐸𝐸0 − 𝐸𝐸𝑟𝑟 )
𝜕𝜕 ln Ω2
= ln Ω2 (𝐸𝐸0 ) + (−𝐸𝐸𝑟𝑟 ) � +⋯ (2)
𝜕𝜕𝜕𝜕 𝐸𝐸𝑟𝑟 =0
≈ ln Ω2 (𝐸𝐸0 ) + (−𝐸𝐸𝑟𝑟 )𝛽𝛽

∴ Ω2 (𝐸𝐸0 − 𝐸𝐸𝑟𝑟 ) = Ω2 (𝐸𝐸0 ) exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) (2')

or
𝑃𝑃(𝐸𝐸𝑟𝑟 ) = 𝐶𝐶 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) (3)

8
exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) = Boltzmann factor.

𝐶𝐶 is determined by:
� 𝑃𝑃(𝐸𝐸𝑟𝑟 ) = 1 (4)
𝑟𝑟

(3) in (4):
𝐶𝐶 � exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) = 1
𝑟𝑟

9
 1
∴ 𝐶𝐶 =
∑𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )

exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
∴ 𝑃𝑃(𝐸𝐸𝑟𝑟 ) = (5)
∑𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )

(5) = Boltzmann probability distribution

⟹ gives prob. of microstate.

Microstates of heat bath dropped out completely;

heat bath defines T of system but has no other
effects.
10
Distribution of 𝑃𝑃(𝐸𝐸𝑟𝑟 ) changes with 𝑇𝑇.

At low 𝑇𝑇 , only microstates of low energy are

occupied.

As 𝑇𝑇 increase, microstates of higher energy start

to have significant prob. of being occupied.

As 𝑇𝑇 → ∞, all energy levels have the same

probabilities.

11
𝑃𝑃(𝐸𝐸𝑟𝑟 ) in (5) applicable to system in contact with
heat bath, with 𝑁𝑁 constant.

Eg. systems: electron in an atom, or a piece of

iron.

4.2 The partition function

Denominator in (5) = partition function, denoted

by 𝑍𝑍:
𝑍𝑍 = � exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) (6a)
𝑟𝑟
12
The sum is over all distinct microstates of system
(𝑟𝑟 refer to microstates, some may have same
energy).

If 𝑔𝑔𝑟𝑟 microstates with same energy 𝐸𝐸𝑟𝑟 , Z can be

rewritten as (𝑟𝑟 refer to distinct energies)
𝑍𝑍 = � 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) (6b)
𝑟𝑟

If physical quantity 𝑦𝑦 reveals value 𝑦𝑦𝑟𝑟 when

system in energy level 𝐸𝐸𝑟𝑟 ,
13
 ∑𝑟𝑟 𝑦𝑦𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
𝑦𝑦� = � 𝑃𝑃(𝐸𝐸𝑟𝑟 ) 𝑦𝑦𝑟𝑟 =
𝑍𝑍
𝑟𝑟

4.3 Energy fluctuations in canonical ensemble

System's mean energy:

14
 ∑𝑟𝑟 𝐸𝐸𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
𝐸𝐸� = � 𝑃𝑃(𝐸𝐸𝑟𝑟 ) 𝐸𝐸𝑟𝑟 =
∑𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
𝑟𝑟
𝜕𝜕 ln 𝑍𝑍
=−
𝜕𝜕𝜕𝜕

�
𝜕𝜕𝐸𝐸 ∑ 2
𝑟𝑟 𝐸𝐸𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
=−
𝜕𝜕𝜕𝜕 ∑𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
2
∑𝑟𝑟 𝐸𝐸𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
+� ������
( 2 )
� = − 𝐸𝐸 + 𝐸𝐸( � ) 2
∑𝑟𝑟 𝑔𝑔𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
(*)
15
 𝜕𝜕𝐸𝐸� 𝜕𝜕𝐸𝐸� 𝜕𝜕𝜕𝜕 �
𝜕𝜕𝐸𝐸
= = −𝑘𝑘𝐵𝐵 𝑇𝑇 2 = −𝑘𝑘𝐵𝐵 𝑇𝑇 2 𝐶𝐶𝑉𝑉
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
(**)

Compare (*) and (**):

𝐸𝐸 2 ) − (𝐸𝐸� )2 = (������������
𝑘𝑘𝐵𝐵 𝑇𝑇 2 𝐶𝐶𝑉𝑉 = (������ 𝐸𝐸 − 𝐸𝐸� )2 = (��������
∆𝐸𝐸 )2

Relative standard deviation:

16
 ∆∗ 𝐸𝐸 ���������
(∆𝐸𝐸 )2 �𝑘𝑘𝐵𝐵 𝑇𝑇 2 𝐶𝐶𝑉𝑉
= =
𝐸𝐸� 𝐸𝐸� 𝐸𝐸�

Ideal monatomic gas:

3 3
𝐶𝐶𝑉𝑉 = 𝑁𝑁𝑘𝑘𝐵𝐵 , 𝐸𝐸� = 𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇
2 2
� 2 3
∗
∆ 𝐸𝐸 𝑘𝑘 𝑇𝑇 𝑁𝑁𝑘𝑘 1
𝐵𝐵 2 𝐵𝐵 �2 1
= = ∝
𝐸𝐸� 3
𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 3 √𝑁𝑁 √𝑁𝑁
2

17
 ∆∗ 𝐸𝐸
⟶0 if 𝑁𝑁 ⟶ ∞
𝐸𝐸�

⟹Relative energy fluctuation of system in heat

bath: inversely proportional to square root of
number of particle.
−11
System with N ≈ 10 , ∆ E E ≈ 10 : extremely
23 *

small fluctuations.

Energy of mac. body in heat bath can be said

completely determined.
18
Note: Above statements not applicable when
physical system is undergoing phase transition.

4.4 Entropy in canonical ensemble

Generalize isolated-system’s entropy to canonical

ensemble:

Let 𝐴𝐴 (heat bath) composed of 𝑀𝑀 small system 𝐴𝐴1

(𝑀𝑀 is large).
19
Let

𝑛𝑛1 small systems in microstates with energy 𝐸𝐸1 ,

𝑛𝑛2 small systems in microstates with energy 𝐸𝐸2 ,


𝑛𝑛𝑟𝑟 small systems in microstates with energy 𝐸𝐸𝑟𝑟 ,

Number of attainable microstates for 𝐴𝐴 :

𝑀𝑀!
Ω=
𝑛𝑛1 ! 𝑛𝑛2 ! ⋯ 𝑛𝑛𝑟𝑟 ! ⋯
20
Entropy (Stirling’s appr.):
𝑆𝑆𝑀𝑀 = 𝑘𝑘𝐵𝐵 ln Ω
= 𝑘𝑘𝐵𝐵 {(𝑀𝑀 ln 𝑀𝑀 − 𝑀𝑀) − (𝑛𝑛1 ln 𝑛𝑛1 − 𝑛𝑛1 )
− (𝑛𝑛2 ln 𝑛𝑛2 − 𝑛𝑛2 ) − ⋯ − (𝑛𝑛𝑟𝑟 ln 𝑛𝑛𝑟𝑟 − 𝑛𝑛𝑟𝑟 )

− ⋯ } = 𝑘𝑘𝐵𝐵 �𝑀𝑀 ln 𝑀𝑀 − � 𝑛𝑛𝑟𝑟 ln 𝑛𝑛𝑟𝑟 �

𝑟𝑟

𝑀𝑀 = � 𝑛𝑛𝑟𝑟
𝑟𝑟

21
∴ 𝑆𝑆𝑀𝑀 = 𝑘𝑘𝐵𝐵 �� 𝑛𝑛𝑟𝑟 ln 𝑀𝑀 − � 𝑛𝑛𝑟𝑟 ln 𝑛𝑛𝑟𝑟 �
𝑟𝑟 𝑟𝑟

= 𝑘𝑘𝐵𝐵 �� 𝑛𝑛𝑟𝑟 (ln 𝑀𝑀 − ln 𝑛𝑛𝑟𝑟 ) �

𝑟𝑟
𝑛𝑛𝑟𝑟
= −𝑘𝑘𝐵𝐵 𝑀𝑀 �� (ln 𝑛𝑛𝑟𝑟 − ln 𝑀𝑀) �
𝑀𝑀
𝑟𝑟
𝑛𝑛𝑟𝑟 𝑛𝑛𝑟𝑟
= −𝑘𝑘𝐵𝐵 𝑀𝑀 � ln
𝑀𝑀 𝑀𝑀
𝑟𝑟

22
When 𝑀𝑀 ⟶ ∞ ,

𝑛𝑛𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
⟶ 𝑃𝑃𝑟𝑟 =
𝑀𝑀 𝑍𝑍

where 𝑍𝑍 = ∑𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )

Therefore entropy of a small system is

23
 𝑆𝑆𝑀𝑀
𝑆𝑆 = = −𝑘𝑘𝐵𝐵 � 𝑃𝑃𝑟𝑟 ln 𝑃𝑃𝑟𝑟
𝑀𝑀
𝑟𝑟
exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
= −𝑘𝑘𝐵𝐵 � 𝑃𝑃𝑟𝑟 ln � �
𝑍𝑍
𝑟𝑟
= −𝑘𝑘𝐵𝐵 � 𝑃𝑃𝑟𝑟 (−𝛽𝛽𝐸𝐸𝑟𝑟 − ln 𝑍𝑍)
𝑟𝑟
#
𝜕𝜕𝜕𝜕 2
𝜕𝜕𝜕𝜕
= −𝑘𝑘𝐵𝐵 𝑇𝑇 = − � 𝐸𝐸𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
𝑟𝑟

24
 𝜕𝜕𝜕𝜕
∴ � 𝐸𝐸𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) = 𝑘𝑘𝐵𝐵 𝑇𝑇 2
𝜕𝜕𝜕𝜕
𝑟𝑟

Mean energy is
∑𝑟𝑟 𝐸𝐸𝑟𝑟 exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) 1 𝜕𝜕𝜕𝜕
𝐸𝐸� = � 𝑃𝑃𝑟𝑟 𝐸𝐸𝑟𝑟 = = 𝑘𝑘𝐵𝐵 𝑇𝑇 2
𝑍𝑍 𝑍𝑍 𝜕𝜕𝜕𝜕
𝑟𝑟
2
𝜕𝜕 ln 𝑍𝑍 𝜕𝜕 ln 𝑍𝑍
= 𝑘𝑘𝐵𝐵 𝑇𝑇 =−
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

25
Helmholtz free energy is
𝐹𝐹 = 𝐸𝐸� − 𝑇𝑇𝑇𝑇 = 𝐸𝐸� + 𝑘𝑘𝐵𝐵 𝑇𝑇 � 𝑃𝑃𝑟𝑟 (−𝛽𝛽𝐸𝐸𝑟𝑟 − ln 𝑍𝑍)
𝑟𝑟
= 𝐸𝐸� − ���
𝑘𝑘𝐵𝐵 𝑇𝑇𝑇𝑇 � 𝑃𝑃𝑟𝑟 𝐸𝐸𝑟𝑟 − 𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝑍𝑍 � 𝑃𝑃𝑟𝑟
=1 �����
𝑟𝑟 ���
𝑟𝑟
=𝐸𝐸� =1
= −𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝑍𝑍

Conclusion for a system in heat bath:

𝜕𝜕 ln 𝑍𝑍
𝐸𝐸� = − and 𝐹𝐹 = −𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝑍𝑍
𝜕𝜕𝜕𝜕
26
4.5 Example: application of canonical
distribution in paramagnetic system
1
Consider a piece of paramagnet (𝑁𝑁 spin
2
particles ) in applied magnetic field B.

𝝁𝝁 −𝝁𝝁 ��⃗
𝑩𝑩

𝒏𝒏 𝑵𝑵 − 𝒏𝒏
27
 1
A spin particle act as small system, remainding
2
as heat bath.

If a particle is spin up (or spin down), its

magnetic moment is 𝜇𝜇 (or −𝜇𝜇) , its energy is
−𝜇𝜇𝜇𝜇 (or +𝜇𝜇𝜇𝜇).

𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
𝑍𝑍 = � exp(−𝛽𝛽𝐸𝐸𝑟𝑟 ) = exp � � + exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
𝑟𝑟

28
The spin up probability:

𝜇𝜇𝜇𝜇
exp �+ �
𝑘𝑘𝐵𝐵 𝑇𝑇
𝑃𝑃+ =
𝑍𝑍

The spin down probability:

𝜇𝜇𝜇𝜇
exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇
𝑃𝑃− =
𝑍𝑍

29
Mean magnetic moment for particle:

𝜇𝜇̅ = 𝜇𝜇𝑃𝑃+ + (−𝜇𝜇)𝑃𝑃−

𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
exp � � exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
= 𝜇𝜇 + (−𝜇𝜇)
𝑍𝑍 𝑍𝑍
𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
exp � � − exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
= 𝜇𝜇 � �
𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇
exp � � + exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
𝜇𝜇𝜇𝜇
= 𝜇𝜇 tanh � �
𝑘𝑘𝐵𝐵 𝑇𝑇
30
Energy for particle :

𝜇𝜇𝜇𝜇
𝜀𝜀̅ = (−𝜇𝜇𝜇𝜇)𝑃𝑃+ + (𝜇𝜇𝜇𝜇)𝑃𝑃− = −𝜇𝜇𝜇𝜇 tanh � �
𝑘𝑘𝐵𝐵 𝑇𝑇

For system, total energy and magnetization :

𝜇𝜇𝜇𝜇
𝐸𝐸 = 𝑁𝑁𝜀𝜀̅ = −𝑁𝑁𝑁𝑁𝑁𝑁 tanh � �
𝑘𝑘𝐵𝐵 𝑇𝑇

𝜇𝜇𝜇𝜇
𝑀𝑀 = 𝑁𝑁𝜇𝜇̅ = 𝑁𝑁𝑁𝑁 tanh � �
𝑘𝑘𝐵𝐵 𝑇𝑇
31
For high 𝑇𝑇 and weak 𝐵𝐵,

𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇

≪1 , tanh � �≈
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇

𝑁𝑁𝜇𝜇2 𝐵𝐵
∴ 𝑀𝑀 ≈
𝑘𝑘𝐵𝐵 𝑇𝑇

For low 𝑇𝑇 and strong 𝐵𝐵,

𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇 𝜇𝜇𝜇𝜇

≫1 , exp � � ≫ exp �− �
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇
32
 𝜇𝜇𝜇𝜇
tanh � �≈1 , ∴ 𝑀𝑀 ≈ 𝑁𝑁𝑁𝑁
𝑘𝑘𝐵𝐵 𝑇𝑇

33
1.2
M
Nµ
1

0.8

0.6

0.4

0.2

0
µB
0 0.5 1 1.5 2 2.5 3 kT
34
Chapter 5 Perfect classical gas
(Ideal classical gas)

5.1 Maxwell-Boltzmann statistics

Perfect gas: potential energy or interaction

between molecules is negligible .

Actual gas at low density is approximately a

perfect gas.

Total 𝐸𝐸 is sum of energy of every molecule.

1
Molecule’s energy: determine quantum
mechanically: 𝜀𝜀1 , 𝜀𝜀2 , ⋯ , 𝜀𝜀𝑟𝑟 , ⋯,corresponding to
quantum states 1,2, ⋯ , 𝑟𝑟, ⋯, or single particle
states.

For classical gas, molecules are distinguishable.

When molecules exchange positions, states before

and after positions exchange are different
microstates.

2
Concept of symmetry (of wavefunction) is not
applied on perfect classical gas.

Energies of attainable microstates are determined

by occupation numbers 𝑛𝑛1 , 𝑛𝑛2 , ⋯ , 𝑛𝑛𝑟𝑟 , ⋯, of single
particle states.

Eg: A classical perfect gas consisting of 2

molecules, each has 3 single particle states, as
shown in figure.

3
Determine all possible microstates and the
partition function.

Energy
ε 3 = 3ε
ε 2 = 2ε
ε1 = ε

4
Solution: All particles are distinguishable:
microstates are determined by MB statistics: each
energy level can be occupied by any number of
particles:
Ω = 32 = 9
𝑍𝑍 = � exp(−𝛽𝛽𝐸𝐸𝑟𝑟 )
𝑟𝑟
−2𝛽𝛽𝛽𝛽 −4𝛽𝛽𝛽𝛽 −6𝛽𝛽𝛽𝛽 −3𝛽𝛽𝛽𝛽 −3𝛽𝛽𝛽𝛽
= 𝑒𝑒 + 𝑒𝑒 + 𝑒𝑒 + 𝑒𝑒 + 𝑒𝑒
+ 𝑒𝑒 −4𝛽𝛽𝛽𝛽 + 𝑒𝑒 −4𝛽𝛽𝛽𝛽 + 𝑒𝑒 −5𝛽𝛽𝛽𝛽 + 𝑒𝑒 −5𝛽𝛽𝛽𝛽
= 𝑒𝑒 −2𝛽𝛽𝛽𝛽 + 2𝑒𝑒 −3𝛽𝛽𝛽𝛽 + 3𝑒𝑒 −4𝛽𝛽𝛽𝛽 + 2𝑒𝑒 −5𝛽𝛽𝛽𝛽
+ 𝑒𝑒 −6𝛽𝛽𝛽𝛽
5
 energy ε 2ε 3ε Er
2ε
4ε
6ε
3ε
9 possible
3ε
microstates
4ε
4ε
5ε
5ε

6
In terms of occupation number of distinct energy
levels, i.e. for a given set of occupation number:

{𝑛𝑛1 , 𝑛𝑛2 , ⋯ , 𝑛𝑛𝑟𝑟 , ⋯ }

𝑁𝑁!
Ω{𝑛𝑛 1 ,𝑛𝑛 2 ,⋯,𝑛𝑛 𝑟𝑟 ,⋯ } =
𝑛𝑛1 ! 𝑛𝑛2 ! ⋯ 𝑛𝑛𝑟𝑟 ! ⋯

The partition function can be rewritten as

7
 𝑁𝑁! −𝛽𝛽 {𝑛𝑛 1 𝜀𝜀 1 +𝑛𝑛 2 𝜀𝜀 2 +⋯ }
𝑍𝑍 = � 𝑒𝑒
𝑛𝑛1 ! 𝑛𝑛2 ! ⋯ 𝑛𝑛𝑟𝑟 ! ⋯
{𝑛𝑛 1 ,𝑛𝑛 2 ,⋯ }
(*)
or
𝑁𝑁! −𝛽𝛽 𝜀𝜀 1 𝑛𝑛 1 −𝛽𝛽 𝜀𝜀 2 𝑛𝑛 2
𝑍𝑍 = � �𝑒𝑒 � �𝑒𝑒 � ⋯
𝑛𝑛1 ! 𝑛𝑛2 ! ⋯
{𝑛𝑛 1 ,𝑛𝑛 2 ,⋯ }
−𝛽𝛽 𝜀𝜀 1 −𝛽𝛽 𝜀𝜀 2 𝑁𝑁
= �𝑒𝑒 + 𝑒𝑒 + ⋯�
𝑁𝑁

= �� exp(−𝛽𝛽𝜀𝜀𝑟𝑟 )� = 𝑞𝑞 𝑁𝑁
𝑟𝑟
(**)
8
Note:
• 𝜀𝜀𝑟𝑟 is distinct single particle state's energy.
• Also, different single particle states may have
same energy.
Summation over all {𝑛𝑛1 , 𝑛𝑛2 , ⋯ , 𝑛𝑛𝑟𝑟 , ⋯ },

� 𝑛𝑛𝑟𝑟 = 𝑁𝑁
𝑟𝑟

Note: 𝑞𝑞=single particle partition function.

9
 𝑁𝑁
�𝑥𝑥1 + 𝑥𝑥2 + ⋯ + 𝑥𝑥𝑝𝑝 �
𝑁𝑁! 𝑛𝑛 1 𝑛𝑛 2 𝑛𝑛 𝑝𝑝
= � 𝑥𝑥1 𝑥𝑥2 ⋯ 𝑥𝑥𝑝𝑝
𝑛𝑛1 ! 𝑛𝑛2 ! ⋯ 𝑛𝑛𝑝𝑝 !
𝑛𝑛 1 ,𝑛𝑛 2 ,⋯,𝑛𝑛 𝑝𝑝
(multinomial)
From (*), mean occupation number 𝑛𝑛�Λ in a
particular single particle state Λ is

10
 𝑁𝑁! −𝛽𝛽 {𝑛𝑛 1 𝜀𝜀 1 +𝑛𝑛 2 𝜀𝜀 2 +⋯𝑛𝑛 Λ 𝜀𝜀 Λ +⋯ }
∑𝑛𝑛 1 ,𝑛𝑛 2 ,⋯ 𝑛𝑛Λ 𝑒𝑒
𝑛𝑛 1 !⋯𝑛𝑛 Λ !⋯
𝑛𝑛�Λ = 𝑁𝑁!
∑𝑛𝑛 1 ,𝑛𝑛 2 ,⋯ 𝑒𝑒 −𝛽𝛽 {𝑛𝑛 1 𝜀𝜀 1 +𝑛𝑛 2 𝜀𝜀 2 +⋯𝑛𝑛 Λ 𝜀𝜀 Λ +⋯ }
𝑛𝑛 !⋯𝑛𝑛 !⋯
1 Λ

1 𝜕𝜕𝜕𝜕 1 𝜕𝜕 ln 𝑍𝑍
=− =−
𝛽𝛽𝛽𝛽 𝜕𝜕𝜀𝜀Λ 𝛽𝛽 𝜕𝜕𝜀𝜀Λ

By using (**),

ln 𝑍𝑍 = 𝑁𝑁 ln �� exp(−𝛽𝛽𝜀𝜀𝑟𝑟 )�
𝑟𝑟

11
 1 𝜕𝜕{𝑁𝑁 ln(∑𝑟𝑟 exp(−𝛽𝛽𝜀𝜀𝑟𝑟 ))}
𝑛𝑛�Λ = −
𝛽𝛽 𝜕𝜕𝜀𝜀Λ
𝑁𝑁 (−𝛽𝛽) exp(−𝛽𝛽𝜀𝜀Λ ) 𝑁𝑁 exp(−𝛽𝛽𝜀𝜀Λ )
=− =
𝛽𝛽 ∑𝑟𝑟 exp(−𝛽𝛽𝜀𝜀𝑟𝑟 ) ∑𝑟𝑟 exp(−𝛽𝛽𝜀𝜀𝑟𝑟 )
= Maxwell-Boltzmann distribution function.
(#)
Note: In (#), subscript Λ refer to single particle
state.
In terms of energy level,

12
 𝑔𝑔𝑖𝑖 exp(−𝛽𝛽𝜀𝜀𝑖𝑖 )
𝑛𝑛�𝑖𝑖 = 𝑁𝑁
∑𝑗𝑗 𝑔𝑔𝑗𝑗 exp�−𝛽𝛽𝜀𝜀𝑗𝑗 �

5.2 Equation of state

For classical perfect gas,

𝑁𝑁
𝑝𝑝𝑖𝑖2
𝐸𝐸𝐾𝐾 = �
2𝑚𝑚
𝑖𝑖=1

13
 2 2 2 2
where 𝑝𝑝𝑖𝑖 = 𝑝𝑝𝑖𝑖𝑖𝑖 + 𝑝𝑝𝑖𝑖𝑖𝑖 + 𝑝𝑝𝑖𝑖𝑖𝑖 .

Energy depends on position and momentum of

each molecule:
𝐸𝐸 = 𝐸𝐸 (𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 , 𝑝𝑝1 , 𝑝𝑝2 , ⋯ , 𝑝𝑝𝑠𝑠 )
𝑠𝑠 = 3𝑁𝑁 = degrees of freedom
.

Consider a perfect gas (𝑁𝑁 molecules) in enclosure

of volume 𝑉𝑉 (immerse in heat bath at 𝑇𝑇 )

14
 Perfect gas
V
Heat hath

Figure 1

“Volume” element in phase space is

15
 (�� ) ( ) ( ) 𝑠𝑠
∆𝑝𝑝
���∆𝑞𝑞
1 ��1 �� ∆𝑝𝑝
���∆𝑞𝑞
2 ��2 ⋯ ��∆𝑝𝑝
��� ∆𝑞𝑞
𝑠𝑠 ��𝑠𝑠 = ℎ0
ℎ0 ℎ0 ℎ0

Differential “volume” element of phase space is

𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑞𝑞1 𝑑𝑑𝑞𝑞2 ⋯ 𝑑𝑑𝑞𝑞𝑠𝑠 𝑑𝑑𝑝𝑝1 𝑑𝑑𝑝𝑝2 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠 (#)

Number of states in (#) is approximately

1
𝑠𝑠 (𝑑𝑑𝑞𝑞1 𝑑𝑑𝑞𝑞2 ⋯ 𝑑𝑑𝑞𝑞𝑠𝑠 𝑑𝑑𝑝𝑝1 𝑑𝑑𝑝𝑝2 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠 ) (i)
ℎ0
16
Energy of these states is approximately

𝐸𝐸 (𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 , 𝑝𝑝1 , 𝑝𝑝2 , ⋯ , 𝑝𝑝𝑠𝑠 ) (ii)

Partition function of perfect gas: multiple

integrations over all degrees of freedom:

1
𝑍𝑍 = 𝑠𝑠 � ⋯ � exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 ⋯ 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑞𝑞1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
ℎ0
(iii)
𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 , 𝑝𝑝1 , 𝑝𝑝2 , ⋯ , 𝑝𝑝𝑠𝑠
17
to
𝑞𝑞1 + 𝑑𝑑𝑞𝑞1 , 𝑞𝑞2 + 𝑑𝑑𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 + 𝑑𝑑𝑞𝑞𝑆𝑆 ,
𝑝𝑝1 + 𝑑𝑑𝑝𝑝1 , 𝑝𝑝2 + 𝑑𝑑𝑝𝑝2 , ⋯ , 𝑝𝑝𝑠𝑠 + 𝑑𝑑𝑝𝑝𝑆𝑆

Note: In a perfect gas, potential energy is zero, so

𝑁𝑁 2
𝑝𝑝𝑖𝑖
𝐸𝐸 = 𝐸𝐸𝐾𝐾 = �
2𝑚𝑚
𝑖𝑖=1
By using notations

𝑑𝑑 3 𝑟𝑟1 = 𝑑𝑑𝑥𝑥1 𝑑𝑑𝑦𝑦1 𝑑𝑑𝑧𝑧1 = 𝑑𝑑𝑞𝑞1 𝑑𝑑𝑞𝑞2 𝑑𝑑𝑞𝑞3

18
 𝑑𝑑 3 𝑟𝑟2 = 𝑑𝑑𝑥𝑥2 𝑑𝑑𝑦𝑦2 𝑑𝑑𝑧𝑧2 = 𝑑𝑑𝑞𝑞4 𝑑𝑑𝑞𝑞5 𝑑𝑑𝑞𝑞6
⋯
3
𝑑𝑑 𝑝𝑝1 = 𝑑𝑑𝑝𝑝1𝑥𝑥 𝑑𝑑𝑝𝑝1𝑦𝑦 𝑑𝑑𝑝𝑝1𝑧𝑧 = 𝑑𝑑𝑝𝑝1 𝑑𝑑𝑝𝑝2 𝑑𝑑𝑝𝑝3

𝑑𝑑 3 𝑝𝑝2 = 𝑑𝑑𝑝𝑝2𝑥𝑥 𝑑𝑑𝑝𝑝2𝑦𝑦 𝑑𝑑𝑝𝑝2𝑧𝑧 = 𝑑𝑑𝑝𝑝4 𝑑𝑑𝑝𝑝5 𝑑𝑑𝑝𝑝6

Kinetic part of 𝑍𝑍 in (iii) is

𝑁𝑁
1 𝑝𝑝𝑖𝑖2
𝑍𝑍 = 3𝑁𝑁 � ⋯ � exp �−𝛽𝛽 � �
ℎ0 2𝑚𝑚
𝑖𝑖=1
3 3 3 3
× 𝑑𝑑 𝑟𝑟1 ⋯ 𝑑𝑑 𝑟𝑟𝑁𝑁 𝑑𝑑 𝑝𝑝1 ⋯ 𝑑𝑑 𝑝𝑝𝑁𝑁
19
Integrand is independent of 𝑟𝑟𝑖𝑖 ,

3 3 3
∴ � 𝑑𝑑 𝑟𝑟1 = � 𝑑𝑑 𝑟𝑟2 = ⋯ = � 𝑑𝑑 𝑟𝑟𝑁𝑁 = 𝑉𝑉

∴ � ⋯ � 𝑑𝑑 3 𝑟𝑟1 ⋯ 𝑑𝑑 3 𝑟𝑟𝑁𝑁 = 𝑉𝑉 𝑁𝑁
and

20
 𝑁𝑁 𝑁𝑁 2
𝑉𝑉 𝑝𝑝𝑖𝑖
𝑍𝑍 = 3𝑁𝑁 � ⋯ � exp �−𝛽𝛽 � � 𝑑𝑑 3 𝑝𝑝1 ⋯ 𝑑𝑑 3 𝑝𝑝𝑁𝑁
ℎ0 2𝑚𝑚
𝑖𝑖=1
𝑁𝑁 2 2
𝑉𝑉 𝛽𝛽𝑝𝑝1
3
𝛽𝛽𝑝𝑝𝑁𝑁
3
= 3𝑁𝑁 � exp �− � 𝑑𝑑 𝑝𝑝1 ⋯ � exp �− � 𝑑𝑑 𝑝𝑝𝑁𝑁
2𝑚𝑚
ℎ0 ������������� 2𝑚𝑚
�������������
𝑄𝑄 𝑄𝑄
𝑉𝑉 𝑁𝑁 𝑄𝑄 𝑁𝑁
= 3𝑁𝑁
ℎ0

Evaluation of Q :

21
 2
𝛽𝛽𝑝𝑝𝑖𝑖
𝑄𝑄 = � exp �− � 𝑑𝑑 3 𝑝𝑝𝑖𝑖
2𝑚𝑚
∞ ∞ ∞
2 2 2
𝛽𝛽�𝑝𝑝𝑖𝑖𝑖𝑖 + 𝑝𝑝𝑖𝑖𝑖𝑖 + 𝑝𝑝𝑖𝑖𝑖𝑖 �
= � � � exp �− � 𝑑𝑑𝑝𝑝𝑖𝑖𝑖𝑖 𝑑𝑑𝑝𝑝𝑖𝑖𝑖𝑖 𝑑𝑑𝑝𝑝𝑖𝑖𝑖𝑖
2𝑚𝑚
−∞ −∞ −∞
∞ ∞
2 2
𝛽𝛽𝑝𝑝𝑖𝑖𝑖𝑖 𝛽𝛽𝑝𝑝𝑖𝑖𝑖𝑖
= � exp �− � 𝑑𝑑𝑝𝑝𝑖𝑖𝑖𝑖 × � exp �− � 𝑑𝑑𝑝𝑝𝑖𝑖𝑖𝑖
2𝑚𝑚 2𝑚𝑚
−∞ −∞
∞
2
𝛽𝛽𝑝𝑝𝑖𝑖𝑖𝑖
× � exp �− � 𝑑𝑑𝑝𝑝𝑖𝑖𝑖𝑖
2𝑚𝑚
−∞
22
Table of integrals:
∞
𝜋𝜋
� exp{−𝑎𝑎𝑥𝑥 2 } 𝑑𝑑𝑑𝑑 = �
𝑎𝑎
−∞
3⁄2
2𝑚𝑚𝑚𝑚
∴ 𝑄𝑄 = � �
𝛽𝛽

𝑁𝑁 3𝑁𝑁 ⁄2
𝑉𝑉 2𝑚𝑚𝑚𝑚
𝑍𝑍 = 3𝑁𝑁 � �
ℎ0 𝛽𝛽

23
 3𝑁𝑁 3𝑁𝑁
ln 𝑍𝑍 = 𝑁𝑁 ln 𝑉𝑉 + ln(2𝑚𝑚𝑚𝑚) − ln 𝛽𝛽 − 3𝑁𝑁 ln ℎ0
2 2

Some physical quantities of perfect gas calculated

from 𝑍𝑍:

𝐹𝐹 = −𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝑍𝑍 (Helmholtz free energy)

𝜕𝜕 ln 𝑍𝑍 3𝑁𝑁 3
𝐸𝐸� = − = = 𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 (mean energy)
𝜕𝜕𝜕𝜕 2𝛽𝛽 2

24
 𝜕𝜕𝐸𝐸� 3
𝐶𝐶𝑉𝑉 = � � = 𝑁𝑁𝑘𝑘𝐵𝐵 (heat capacity)
𝜕𝜕𝜕𝜕 𝑉𝑉 2

𝜕𝜕𝜕𝜕 𝜕𝜕 ln 𝑉𝑉 𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇
𝑝𝑝 = − � � = 𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 = (pressure)
𝜕𝜕𝜕𝜕 𝑇𝑇 𝜕𝜕𝜕𝜕 𝑉𝑉

Rearranging the last equation

𝑝𝑝𝑝𝑝 = 𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 = equation of state

25
5.2 Theorem of equipartition of energy (TEE)

Energy of perfect classical gas is

𝐸𝐸 = 𝐸𝐸 (𝑞𝑞1 , 𝑞𝑞2 , ⋯ , 𝑞𝑞𝑠𝑠 , 𝑝𝑝1 , 𝑝𝑝2 , ⋯ , 𝑝𝑝𝑠𝑠 )

Frequently occurred situations are:

(a) Total energy can be written as a sum of two
parts:

26
 3𝑁𝑁
′(
𝐸𝐸 = � 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 ) + 𝐸𝐸 ��� 𝑞𝑞� , 𝑞𝑞2 , ⋯
1 ��� 𝑞𝑞𝑠𝑠 )
��, ��
�� ���
𝑖𝑖=1 �� 𝑎𝑎𝑞𝑞 𝑖𝑖2 , independent of 𝑝𝑝 𝑖𝑖
kinetic energy

(b) Kinetic energy of each molecule ∝ square of

momentum:
𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 ) = 𝑏𝑏𝑝𝑝𝑖𝑖2 , 𝑏𝑏 = constant

In thermal equilibrium at 𝑇𝑇, mean kinetic energy

is:
27
������������
� 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 ) =
𝑖𝑖
𝑑𝑑𝑑𝑑 1 ⋯𝑑𝑑𝑝𝑝 𝑠𝑠
∫𝔸𝔸ℙ𝕊𝕊 ∫Ω𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 [∑𝑖𝑖 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 )] exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} ℎ 0𝑠𝑠
𝑑𝑑𝑑𝑑 1 ⋯𝑑𝑑𝑝𝑝 𝑠𝑠
∫𝔸𝔸ℙ𝕊𝕊 ∫Ω𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} ℎ 0𝑠𝑠

where

𝔸𝔸ℙ𝕊𝕊 = all phase space

Ω𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = all possible states in 𝑑𝑑𝑞𝑞1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
28
Size of 𝑑𝑑𝑞𝑞1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠 is chosen ≈ ℎ0𝑠𝑠 , number of
states in 𝑑𝑑𝑞𝑞1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠 ≈ 1.

������������
� 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 ) =
𝑖𝑖
𝑑𝑑𝑑𝑑 1 ⋯𝑑𝑑𝑝𝑝 𝑠𝑠
∫𝔸𝔸ℙ𝕊𝕊[∑𝑖𝑖 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 )] exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} ℎ 0𝑠𝑠
𝑑𝑑𝑑𝑑 1 ⋯𝑑𝑑𝑝𝑝 𝑠𝑠
∫𝔸𝔸ℙ𝕊𝕊 exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} ℎ 0𝑠𝑠

29
 ∫𝔸𝔸ℙ𝕊𝕊 𝜀𝜀1 (𝑝𝑝1 ) exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
=
∫𝔸𝔸ℙ𝕊𝕊 exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠

∫𝔸𝔸ℙ𝕊𝕊 𝜀𝜀2 (𝑝𝑝2 ) exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠

+
∫𝔸𝔸ℙ𝕊𝕊 exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
+⋯
∫𝔸𝔸ℙ𝕊𝕊 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 ) exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
+
∫𝔸𝔸ℙ𝕊𝕊 exp{−𝛽𝛽𝛽𝛽 (𝑞𝑞1 , ⋯ , 𝑝𝑝𝑠𝑠 )} 𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠

+⋯
30
 −𝛽𝛽 𝜀𝜀 1 (𝑝𝑝 1 ) −𝛽𝛽𝛽𝛽 (𝑞𝑞 1 ,⋯,𝑝𝑝 𝑠𝑠 )
∫ 𝜀𝜀1 (𝑝𝑝1 )𝑒𝑒 𝑑𝑑𝑝𝑝1 ∫(𝑝𝑝 ) 𝑒𝑒 𝑑𝑑𝑞𝑞2 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
1
=
∫ 𝑒𝑒 −𝛽𝛽 𝜀𝜀 1 (𝑝𝑝 1 ) 𝑑𝑑𝑝𝑝1 ∫(𝑝𝑝 ) 𝑒𝑒 −𝛽𝛽𝛽𝛽 (𝑞𝑞 1 ,⋯,𝑝𝑝𝑠𝑠 ) 𝑑𝑑𝑑𝑑2 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
1

+⋯
∫ 𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 ) exp{−𝛽𝛽𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 )} 𝑑𝑑𝑝𝑝𝑖𝑖 ∙ ∫(𝑝𝑝 ){⋯ }𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
𝑖𝑖
+
∫ exp{−𝛽𝛽𝜀𝜀𝑖𝑖 (𝑝𝑝𝑖𝑖 )} 𝑑𝑑𝑝𝑝𝑖𝑖 ∙ ∫(𝑝𝑝 ){⋯ }𝑑𝑑𝑑𝑑1 ⋯ 𝑑𝑑𝑝𝑝𝑠𝑠
𝑖𝑖
+⋯

(𝑝𝑝𝑖𝑖 ) in lower limit of ∫ means "exclude 𝑝𝑝𝑖𝑖 ".

31
Mean kinetic energy of any molecule :
2 2
∫ 𝑏𝑏𝑝𝑝𝑖𝑖exp�−𝛽𝛽𝑏𝑏𝑝𝑝𝑖𝑖 � 𝑑𝑑𝑝𝑝𝑖𝑖
�����
𝜀𝜀𝑖𝑖 = 𝑏𝑏𝑝𝑝 2 =
𝑖𝑖
∫ exp{−𝛽𝛽𝛽𝛽𝑝𝑝𝑖𝑖2 } 𝑑𝑑𝑝𝑝𝑖𝑖
𝜕𝜕 2
=− �ln � exp�−𝛽𝛽𝑏𝑏𝑝𝑝𝑖𝑖 � 𝑑𝑑𝑝𝑝𝑖𝑖 �
𝜕𝜕𝜕𝜕
(*)
Let 𝑦𝑦 = �𝛽𝛽 𝑝𝑝𝑖𝑖 , then 𝑑𝑑𝑑𝑑 = �𝛽𝛽 𝑑𝑑𝑝𝑝𝑖𝑖
∞ ∞
1
� exp�−𝛽𝛽𝛽𝛽𝑝𝑝𝑖𝑖2 � 𝑑𝑑𝑝𝑝𝑖𝑖 = � exp{−𝑏𝑏𝑦𝑦 2 } 𝑑𝑑𝑑𝑑
−∞ �𝛽𝛽 −∞
32
 ∞
ln � exp�−𝛽𝛽𝛽𝛽𝑝𝑝𝑖𝑖2 � 𝑑𝑑𝑝𝑝𝑖𝑖
−∞
∞
1 2}
{
= − ln 𝛽𝛽 + ln � exp −𝑏𝑏𝑦𝑦 𝑑𝑑𝑑𝑑
2 �������������
−∞
independent of 𝛽𝛽

Substitute the last result into (*),

1 1
𝜀𝜀𝑖𝑖 = = 𝑘𝑘𝐵𝐵 𝑇𝑇
2𝛽𝛽 2

33
 �����
2 1
By using same procedures, 𝑏𝑏𝑞𝑞𝑖𝑖 = 𝑘𝑘𝐵𝐵 𝑇𝑇
2

The last two results verify TEE:

For a system in thermal equilibrium at T, a

generalized position or momentum coordinate,
occurs in the Hamiltonian as a square term,
1
contributes an energy 𝑘𝑘𝐵𝐵 𝑇𝑇 to mean energy of
2
the system.

34
5.3 Application for TEE

Eg: Calculate the mean kinetic energy and specific

heat for a perfect gas at temperature T.

Solution: The translational kinetic energy of a

molecule:
1 2 2 2
𝐾𝐾 = �𝑝𝑝𝑥𝑥 + 𝑝𝑝𝑦𝑦 + 𝑝𝑝𝑧𝑧 �
2𝑚𝑚
1 ���2 1
According to TEE, 𝑝𝑝𝑖𝑖 = 𝑘𝑘𝐵𝐵 𝑇𝑇
2𝑚𝑚 2

35
 1 ���2 1 �������������������
2 2
3
𝑝𝑝 = �𝑝𝑝𝑥𝑥 + 𝑝𝑝𝑦𝑦 + 𝑝𝑝𝑧𝑧2 � = 𝑘𝑘𝐵𝐵 𝑇𝑇
2𝑚𝑚 2𝑚𝑚 2

For monatomic perfect gas, mean energy per mole

is
3 3
𝐸𝐸� = 𝑁𝑁𝐴𝐴 � 𝑘𝑘𝐵𝐵 𝑇𝑇� = 𝑅𝑅𝑅𝑅
2 2

Specific heat per mole is

𝜕𝜕𝐸𝐸� 3
𝐶𝐶𝑉𝑉 = � � = 𝑅𝑅
𝜕𝜕𝜕𝜕 𝑉𝑉 2

36
Eg: Consider a 1-D simple harmonic oscillator
(SHO) in equilibrium at T. Determine its mean
energy.

Solution: Energy of SHO is composed of kinetic

2
p 1 2
energy , and potential energy k ′x ,
2m 2
2
p 1 2
∴ E= + k ′x
2m 2

According to TEE,

37
 1 ���
2 1 1 ′ ���
2 1
𝑝𝑝 = 𝑘𝑘𝐵𝐵 𝑇𝑇 and 𝑘𝑘 𝑥𝑥 = 𝑘𝑘𝐵𝐵 𝑇𝑇
2𝑚𝑚 2 2 2

Therefore mean energy of this SHO is

1 1
𝐸𝐸� = 𝑘𝑘𝐵𝐵 𝑇𝑇 + 𝑘𝑘𝐵𝐵 𝑇𝑇 = 𝑘𝑘𝐵𝐵 𝑇𝑇
2 2

38
Chapter 6 Perfect quantal gas
6.1 Quantum statistics
Consider a gas of N non-interacting identical
particles, with discrete single particle states
1,2,, r ,and energies ε1 , ε 2 ,, ε r ,.
A microstate of the gas is specified by a set of
occupation numbers n1 , n2 ,, nr , of the single
particle states.
n1 , n2 ,, nr , are subjected to QM restrictions.

1
A perfect quantal gas with identical particles can be
one of two cases:
[1] Identical particles are bosons:
(a) Occupation numbers can be any positive
integer: nr = 0, 1, 2,, r = 1,2,.
(b) Spin ( intrinsic angular momentum) of boson
is any of {0, , 2, 3,}.
(c) Wavefunction of boson gas is symmetric
under exchange of spatial coordinates of any
two particles,
2
 ψ (Q j Qi ) = ψ (Qi Q j )
(d) Egs: photon (spin  ), π meson (spin 0).

[2] Identical particles are fermions:

(a) Occupation numbers nr are restricted to 1 or
0: nr = 0, 1 , r = 1,2,.
(b) Spin of fermion is any of { 12 , 32 , 52 ,}.
(c) Wavefunction of fermion gas is anti-
symmetric under exchange of spatial
coordinates of any two particles,
3
 ψ (Q j Qi ) = −ψ (Qi Q j )
(d) Egs: electron, positron, proton, neutron (spin
1
2
 ).
Note: (a) and (c) state Pauli exclusion principle:
two fermions cannot be in the same single particle
state.
Eg: A gas with 2 identical particles, each has 3
single particle states. Determine all microstates
and partition function if theparticles are(a)
bosons, (b) fermions.
4
 Energy
ε 3 = 3ε
ε 2 = 2ε
ε1 = ε

Solution: (a) By using BE statistics, all possible

microstates are:

5
Energy ε 2ε 3ε Er
2ε
4ε
6ε
6 possible 3ε
microstates 4ε
5ε

The partition function is

Z = ∑e − β Er −2 βε −3 βε −4 βε
= e + e + 2e + e + e −5 βε −6 βε

r
6
(b) By using FD statistics, all possible microstates
are:
energy ε 2ε 3ε Er
3ε
3 possible
4ε
microstates
5ε
The partition function is
Z = ∑e − β Er
=e −3 βε
+e −4 βε
+e −5 βε

7
6.2 The perfect quantal gas
Total energy of perfect quantal gas = kinetic
energy.
Hamiltonian of perfect quantal gas :
𝑁𝑁 𝑁𝑁
𝑝𝑝𝑖𝑖2
ℏ 2
2
𝐻𝐻 = � =− � ∇𝑖𝑖
2𝑚𝑚 2𝑚𝑚
𝑖𝑖=1 𝑖𝑖=1

8
If perfect quantal gas is in a rectangular box with
volume V = Lx Ly Lz , eigen functions, eigen energies
of any particle in the box is

8 𝑛𝑛𝑛𝑛𝑛𝑛 𝑠𝑠𝑠𝑠𝑠𝑠 𝑣𝑣𝑣𝑣𝑣𝑣

𝜙𝜙𝑛𝑛𝑛𝑛𝑛𝑛 (𝑥𝑥, 𝑦𝑦, 𝑧𝑧) = � sin sin sin
𝑉𝑉 𝐿𝐿𝑥𝑥 𝐿𝐿𝑦𝑦 𝐿𝐿𝑧𝑧

ℏ2 𝜋𝜋 2 𝑛𝑛2 𝑠𝑠 2 𝑣𝑣 2
𝜀𝜀𝑛𝑛𝑛𝑛𝑛𝑛 = � 2 + 2 + 2�
2𝑚𝑚 𝐿𝐿𝑥𝑥 𝐿𝐿𝑦𝑦 𝐿𝐿𝑧𝑧

n = 1, 2,, s = 1, 2, and v = 1, 2, .

9
Single particle states 𝑟𝑟 are determined by quantum
numbers 𝑛𝑛𝑛𝑛𝑛𝑛.
𝑟𝑟 = 1, 𝑛𝑛 = 1, 𝑠𝑠 = 1, 𝑣𝑣 = 1,
𝑟𝑟 = 2, 𝑛𝑛 = 1, 𝑠𝑠 = 1, 𝑣𝑣 = 2,
𝑟𝑟 = 3, 𝑛𝑛 = 1, 𝑠𝑠 = 2, 𝑣𝑣 = 1,
𝑟𝑟 = 4, 𝑛𝑛 = 2, 𝑠𝑠 = 1, 𝑣𝑣 = 1, …
Microstates of perfect quantal gas with N identical
particles are specified by
𝜀𝜀𝑟𝑟 = single particle states energies,

10
𝑛𝑛𝑟𝑟 = occupation numbers = number of particles
in 𝜙𝜙𝑟𝑟 , and
𝑁𝑁 = � 𝑛𝑛𝑟𝑟
𝑟𝑟

Energy of a possible microstate, 𝐸𝐸𝑆𝑆 , is

corresponding to a set of occupation numbers of
single particle states, {𝑛𝑛𝑟𝑟 } = 𝑛𝑛1 , 𝑛𝑛2 , ⋯ , 𝑛𝑛Λ , ⋯

𝐸𝐸𝑆𝑆 = � 𝑛𝑛𝑟𝑟 𝜀𝜀𝑟𝑟

{𝑛𝑛 𝑟𝑟 }

11
6.3 Quantum distribution function of perfect gas

Consider perfect gas with N particles and in

equilibrium at T, with

𝑁𝑁 = � 𝑛𝑛𝑟𝑟 = 𝑛𝑛1 + 𝑛𝑛2 + ⋯ + 𝑛𝑛Λ + ⋯

𝑟𝑟

Note:
(i) Subscript 𝑆𝑆 refer to microstates of perfect gas
(ii) Subscript 𝑟𝑟 refer to single particle states
12
  
∑S  ∑r nr  exp [ − β ES ]
N ∑
= n
r
r
∑ exp [ − β ES ]
S

∑ ( n +  n ) exp − β {n ε +  n ε + }

1 Λ 1 1 Λ Λ
= S

∑ exp − β {n ε +  n ε + }

S
1 1 Λ Λ

∑ ( n +  n ) exp − β {n ε +  n ε + }

1 Λ 1 1 Λ Λ

=
n1 ,nΛ 

∑ exp − β {n ε +  n ε + }

n1 ,nΛ 
1 1 Λ Λ

13
  +
( nΛ )

∑n
nΛ
Λ exp ( − β nΛε Λ ) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }
( nΛ )

∑ exp ( − β n ε ) ∑
nΛ
Λ Λ
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }

+ (*)
( nΛ )

∑
n1 , n2 
is excluding the single particle state Λ .

From (*), mean occupation number of single

particle state is
14
nΛ =
( nΛ )

∑n
nΛ
Λ exp ( − β nΛε Λ ) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }
( nΛ )

∑ exp ( − β n ε ) ∑ exp − β {n ε

nΛ
Λ Λ
n1 , n2 
1 1 + n2ε 2 + }

(1)
6.4 Photons statistics
All objects emit e.m. radiations (or photons) by
virtue of their temperature, usually these
radiations are not in thermal equilibrium.
15
A photon gas with fixed volume and in
equilibrium: opaque enclosure with its wall
maintained at a uniform temperature T.
Thermal equilibrium between photons in the
cavity and the enclosure's wall is achieved by
continual emission and absorption of photons by
atoms of the enclosure maintained at T.
* first property of photon gas: total number of
photons in a cavity maintained at temperature T is
not constant.

16
Photon gas do not obey condition

𝑁𝑁 = � 𝑛𝑛𝑟𝑟 = 𝑛𝑛1 + 𝑛𝑛2 + ⋯ + 𝑛𝑛Λ + ⋯

𝑟𝑟
because occupation number of single particle
states can be any positive integer numbers:
𝑛𝑛Λ = 0,1,2,3, ⋯ ∞, for all 𝑟𝑟.
Total number of photons in cavity at 𝑇𝑇 fluctuates
about mean value (determined by equilibrium
conditions).

17
**2nd property: photons are spin 1 particles,
hence photons obey BE statistics.

***Third property , photons do not interact with

each other and hence photon gas is perfect gas.
Calculation of nΛ is a specific case of BE statistics:
(i) undetermined total number of particles N (or
photons), and
(ii) undetermined occupation number in every
single particle =
states, nΛ 0, 1, 2, 3, ∞.
18
 ( nΛ )
With these properties, ∑ ( ) in (1) are cancelled,
n1 , n2 

∑ n exp ( − β n ε )
Λ Λ Λ

∴ nΛ =
nΛ

∑ exp ( − β n ε )
nΛ
Λ Λ

1 ∂  
=
− ln  ∑ exp ( − β nΛε Λ ) 
β ∂ (ε Λ )  nΛ 

19
 1 ∂
ln 1 − exp ( − βε Λ ) 
β ∂ (ε Λ )
exp ( − βε Λ ) 1
=
1 − exp ( − βε Λ ) exp ( βε Λ ) − 1
1
nΛ = = photons distribution function
exp ( βε Λ ) − 1
=Planck distribution function.

6.5 Fermi-Dirac Statistics for electron gas

For electron gas, 𝑛𝑛𝑟𝑟 = 0 or 1 .
20
If one electron occupy single particle state Λ ,
( nΛ )
terms ∑()
n1 , n2 
in (1) will cover only remaining

( N − 1) electrons.
Let

21
nΛ =
( nΛ )

∑n
nΛ
Λ exp ( − β nΛε Λ ) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }
( nΛ )

∑ exp ( − β n ε ) ∑ exp − β {n ε

nΛ
Λ Λ
n1 , n2 
1 1 + n2ε 2 + }

A
=
B

Since nr = 0 or 1,

22
 ( nΛ )
A (0) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }

N electrons
( nΛ )
+ (1)exp ( − βε Λ ) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }

( N −1) electrons
( nΛ )
Let Z Λ (=
N) ∑ exp − β {n ε
n1 , n2 
+ n2ε 2 + }1 1

N electrons

∴ A= exp ( − βε Λ ) Z Λ ( N − 1)
23
 Similarly,
( nΛ )
B exp ( 0 ) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }

N electrons
( nΛ )
+ exp ( − βε Λ ) ∑
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }

( N −1) electrons

= Z Λ ( N ) + exp ( − βε Λ ) Z Λ ( N − 1)

24
 − βε Λ
e Z Λ ( N − 1) 1
∴ nΛ = − βε Λ βε Λ
ZΛ ( N ) + e Z Λ ( N − 1) Z Λ ( N )e
+1
Z Λ ( N − 1)

For ∆N << N ,

∂ {ln Z Λ ( N )}
ln Z Λ ( N=
− ∆N ) ln Z Λ ( N ) − ∆N + 
∂N
ln Z Λ ( N ) − α Λ ∆N + 

( N − ∆N ) Z Λ ( N ) exp ( −α Λ ∆N )
Z Λ=
25
 ∂ {ln Z Λ ( N )}
where αΛ =
∂N
∂ {ln Z ( N )}
α Λ : independent of Λ , ∴α = (2)
∂N

Let ∆N =
1, then Z Λ ( N − 1) =
ZΛ ( N ) e −α

ZΛ ( N )
or = exp (α )
Z Λ ( N − 1)

26
 1
Then nΛ =
exp (α + βε Λ ) + 1
=Fermi-Dirac distribution function.

α is determined by ∑ nΛ =
N ⇒ ∑n Λ =
N
Λ Λ
1
or ∑n
Λ
Λ ∑Λ=
exp (α + βε Λ ) + 1
N

Helmholtz free energy, F = − kT ln Z .

27
 ∂ ln Z 1 ∂F
From (2), α= = −
∂N kT ∂N
∂F
= µ= "chemical potential",
∂N
µ 1
∴α =− =− βµ ∴ nΛ =
kT exp {β ( ε Λ − µ )} + 1

6.6 Bose-Einstein Statistics

Particles are indistinguishable, nr = 0, 1, 2, 3, and
∑ nΛ = N .
Λ
28
From (1),
( nΛ )

∑n Λ exp ( − β nΛε Λ ) ∑ exp  − β {n1ε1 + n2ε 2 + }

nΛ =
nΛ n1 , n2 
( nΛ )

∑ exp ( − β n ε ) ∑
nΛ
Λ Λ
n1 , n2 
exp  − β {n1ε1 + n2ε 2 + }

− βε Λ −2 βε Λ
0+e Z Λ ( N − 1) + 2e Z Λ ( N − 2) + 
= − βε Λ −2 βε Λ
ZΛ ( N ) + e Z Λ ( N − 1) + e Z Λ ( N − 2) + 

29
 − βε ΛZ Λ ( N − 1) −2 βε Λ Z Λ ( N − 2)
e + 2e +
ZΛ ( N ) ZΛ ( N )
=
− βε Λ Z Λ ( N − 1) −2 βε Λ Z Λ ( N − 2)
1+ e +e +
ZΛ ( N ) ZΛ ( N )
Z Λ ( N − ∆N ) = Z Λ ( N ) e −α∆N
, ∆N =
1, 2, 3,
Z Λ ( N − 1) −α ZΛ ( N − 2) −2α
∴ = e , = e , 
ZΛ ( N ) ZΛ ( N )

30
 exp ( − βε Λ ) exp ( −α ) + 2exp ( −2 βε Λ ) exp ( −2α )
∴ nΛ =
1 + exp ( − βε Λ ) exp ( −α ) + exp ( −2 βε Λ ) exp ( −2α ) + 
N

∑n
nΛ = 0
Λ exp  − nΛ {α + βε Λ }
= N

∑ exp −n {α + βε }

nΛ = 0
Λ Λ

24
𝑁𝑁~10 , let 𝑁𝑁 → ∞

31
 ∑∞
𝑛𝑛Λ =0 𝑛𝑛Λ exp[−𝑛𝑛Λ {𝛼𝛼 + 𝛽𝛽𝜀𝜀Λ }]
𝑛𝑛�Λ ≈
∑∞ 𝑛𝑛Λ =0 exp[−𝑛𝑛Λ {𝛼𝛼 + 𝛽𝛽𝜀𝜀Λ }]
∞
𝜕𝜕
=− ln � exp[−𝑛𝑛Λ {𝛼𝛼 + 𝛽𝛽𝜀𝜀Λ }]
𝜕𝜕𝜕𝜕
𝑛𝑛Λ =0
∞

� exp[−𝑛𝑛Λ {𝛼𝛼 + 𝛽𝛽𝜀𝜀Λ }]

𝑛𝑛 Λ =0
= 1 + 𝑒𝑒 −(𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ ) + 𝑒𝑒 −2(𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ ) + ⋯
2
1 1
= 1 + 𝑎𝑎 + 𝑎𝑎 + ⋯ = =
1 − 𝑎𝑎 1 − 𝑒𝑒 −(𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ )

32
 ∞
𝜕𝜕
∴ 𝑛𝑛�Λ = − ln � exp[−𝑛𝑛Λ {𝛼𝛼 + 𝛽𝛽𝜀𝜀Λ }]
𝜕𝜕𝜕𝜕
𝑛𝑛Λ =0
𝜕𝜕 1
=− ln � �
𝜕𝜕𝜕𝜕 1 − 𝑒𝑒−(𝛼𝛼+𝛽𝛽𝜀𝜀Λ )
𝜕𝜕
=− �ln(1) − ln�1 − 𝑒𝑒−(𝛼𝛼+𝛽𝛽𝜀𝜀Λ ) ��
𝜕𝜕𝜕𝜕
𝜕𝜕 −(𝛼𝛼+𝛽𝛽𝜀𝜀Λ )
=− �0 − ln�1 − 𝑒𝑒 ��
𝜕𝜕𝜕𝜕
𝜕𝜕
= ln�1 − 𝑒𝑒−(𝛼𝛼+𝛽𝛽𝜀𝜀Λ ) �
𝜕𝜕𝜕𝜕

33
 1 𝜕𝜕 −(𝛼𝛼+𝛽𝛽𝜀𝜀Λ )
= ( )
� 1 − 𝑒𝑒 �
1 − 𝑒𝑒− 𝛼𝛼+𝛽𝛽𝜀𝜀Λ 𝜕𝜕𝜕𝜕
−(𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ )
𝑒𝑒
=
1 − 𝑒𝑒 −(𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ )
𝑒𝑒 −(𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ ) 𝑒𝑒 (𝛼𝛼+𝛽𝛽 𝜀𝜀 Λ )
= −( 𝛼𝛼+𝛽𝛽 𝜀𝜀 )
× (𝛼𝛼+𝛽𝛽 𝜀𝜀 )
1 − 𝑒𝑒 Λ 𝑒𝑒 Λ

1
𝑛𝑛�Λ = (𝛼𝛼+𝛽𝛽𝜀𝜀 )
𝑒𝑒 Λ − 1

= Bose-Einstein distribution function

34
and 𝛼𝛼 = −𝛽𝛽𝛽𝛽

For photons, condition ∑n

Λ
Λ = N is unnecessary,
∂ ln Z ( N )
=
Z Λ ( N ) independent of N, α = 0.
∂N
Bose-Einstein distribution function reduced to
photons distribution function
1
nΛ = βε Λ
e −1

35
Chapter 7 Applications of quantum statistics
7.1 Derivation of heat capacities of solids
7.1.1 Classical theory
Molecules of solid are assumed as independent
SHO, at rest at T = 0.
At low T, molecules execute small vibrations about
their equilibrium positions.
In 1D, 𝐸𝐸𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 = 𝐸𝐸𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘𝑘 + 𝐸𝐸𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
𝑝𝑝 𝑥𝑥2 1
or 𝜀𝜀 = + 𝑚𝑚𝜔𝜔2 𝑥𝑥 2
2𝑚𝑚 2
1
 𝟏𝟏 𝟏𝟏
According to TEE , 𝜺𝜺 = 𝑘𝑘𝐵𝐵 𝑇𝑇 + 𝑘𝑘𝐵𝐵 𝑇𝑇 = 𝑘𝑘𝐵𝐵 𝑇𝑇.
𝟐𝟐 𝟐𝟐

In 3D, a solid with N molecules has 3N degrees of

freedom.
For a molecule 𝜺𝜺 = 𝟑𝟑𝑘𝑘𝐵𝐵 𝑇𝑇.
Mean energy of solid :
𝐸𝐸� = 3𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 = 3𝑅𝑅𝑅𝑅
Specific heat of solid:
𝜕𝜕𝐸𝐸�
𝐶𝐶𝑉𝑉 = � � = 3𝑁𝑁𝑘𝑘𝐵𝐵 = 3𝑅𝑅
𝜕𝜕𝜕𝜕 𝑉𝑉
2
⟹ Dulong and Petit's law, valid only in high T
region.

CV
3R

At low T, this law violates 3rd law of

thermodynamics and experimental data:
CV → 0 when T → 0 .
3
7.1.2 Einstein’s model of vibrations in solid
Einstein's assumptions :
(i) molecules of solid are independent oscillators,
they execute small vibrations with the same
frequency ω about their equilibrium positions.

4
(ii) According to quantum theory, possible energy
levels for an oscillator is
( 2 ) ω , n =
εn =n + 1
0, 1, 2, 

∞
−𝛽𝛽 𝜀𝜀 𝑛𝑛
𝑍𝑍 = � 𝑒𝑒
𝑛𝑛=0

Mean energy:

5
 ∑∞
𝑛𝑛=0 𝜀𝜀𝑛𝑛 exp(−𝛽𝛽𝜀𝜀𝑛𝑛 ) 𝜕𝜕
𝜀𝜀̅ = ∞ =⋯=− ln 𝑍𝑍
∑𝑛𝑛=0 exp(−𝛽𝛽𝜀𝜀𝑛𝑛 ) 𝜕𝜕𝜕𝜕
Partition function of an oscillator :
∞
1
𝑍𝑍 = � exp �− �𝑛𝑛 + � 𝛽𝛽ℏ𝜔𝜔�
2
𝑛𝑛=0
∞
1
= exp �− 𝛽𝛽ℏ𝜔𝜔� � exp(−𝑛𝑛𝑛𝑛ℏ𝜔𝜔)
2
𝑛𝑛=0
1
− 𝛽𝛽ℏ𝜔𝜔 −𝛽𝛽ℏ𝜔𝜔 −2𝛽𝛽ℏ𝜔𝜔
= 𝑒𝑒 2 �1 + 𝑒𝑒 + 𝑒𝑒 + ⋯�

6
 1
− 𝛽𝛽ℏ𝜔𝜔
𝑒𝑒 2
∴ 𝑍𝑍 =
1 − 𝑒𝑒 −𝛽𝛽ℏ𝜔𝜔

1
ln 𝑍𝑍 = − 𝛽𝛽ℏ𝜔𝜔 − ln�1 − 𝑒𝑒 −𝛽𝛽ℏ𝜔𝜔 �
2
𝜕𝜕 1 ℏ𝜔𝜔 𝑒𝑒 −𝛽𝛽ℏ𝜔𝜔
∴ 𝜀𝜀̅ = − ln 𝑍𝑍 = ℏ𝜔𝜔 +
𝜕𝜕𝜕𝜕 2 1 − 𝑒𝑒 −𝛽𝛽ℏ𝜔𝜔
1 ℏ𝜔𝜔
= ℏ𝜔𝜔 + 𝛽𝛽ℏ𝜔𝜔
2 𝑒𝑒 −1

7
1
ℏ𝜔𝜔 = zero point (T = 0) energy of oscillator.
2

Total energy and specific heat of solid:

1 1
𝐸𝐸 = 3𝑁𝑁𝜀𝜀̅ = 3𝑁𝑁ℏ𝜔𝜔 � + �
2 exp(𝛽𝛽ℏ𝜔𝜔) − 1

1 1
= 3𝑁𝑁ℏ𝜔𝜔 � + �
2 exp � ℏ𝜔𝜔 � − 1
𝑘𝑘𝐵𝐵 𝑇𝑇

8
 ℏ𝜔𝜔
𝜕𝜕𝜕𝜕 ℏ𝜔𝜔 2 exp � �
𝑘𝑘𝐵𝐵 𝑇𝑇
𝐶𝐶𝑉𝑉 = � � = 3𝑁𝑁𝑘𝑘𝐵𝐵 � � 2
𝜕𝜕𝜕𝜕 𝑉𝑉 𝑘𝑘𝐵𝐵 𝑇𝑇 ℏ𝜔𝜔
�exp � � − 1�
𝑘𝑘𝐵𝐵 𝑇𝑇

Description of 𝐸𝐸 and 𝐶𝐶𝑉𝑉 feature: introduce

“Einstein’s temperature”:

ℏ𝜔𝜔 ℏ𝜔𝜔 𝜃𝜃𝐸𝐸

𝜃𝜃𝐸𝐸 = ⟶ =
𝑘𝑘𝐵𝐵 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑇𝑇

9
 𝜃𝜃𝐸𝐸
𝜃𝜃𝐸𝐸 2 exp � �
∴ 𝐶𝐶𝑉𝑉 = 3𝑅𝑅 � � 𝑇𝑇
𝑇𝑇 2
𝜃𝜃𝐸𝐸
�exp � � − 1�
𝑇𝑇
𝜃𝜃 𝐸𝐸 𝜃𝜃 𝐸𝐸 𝜃𝜃 𝐸𝐸
At high T , ≪ 1 , exp � � ~1 + +⋯
𝑇𝑇 𝑇𝑇 𝑇𝑇

10
 2 𝜃𝜃𝐸𝐸
𝜃𝜃𝐸𝐸 1+ +⋯
∴ 𝐶𝐶𝑉𝑉 = 3𝑅𝑅 � � 𝑇𝑇
𝑇𝑇 2
𝜃𝜃𝐸𝐸
�1 + + ⋯ − 1�
𝑇𝑇
2
𝜃𝜃𝐸𝐸 1+⋯
≈ 3𝑅𝑅 � � 2 ≈ 3𝑅𝑅
𝑇𝑇 𝜃𝜃𝐸𝐸
� + ⋯�
𝑇𝑇

𝜃𝜃 𝐸𝐸 𝜃𝜃 𝐸𝐸
At low T , ≫1 , exp � � ≫ 1
𝑇𝑇 𝑇𝑇

11
 𝜃𝜃𝐸𝐸
𝜃𝜃𝐸𝐸2 exp � �
∴ 𝐶𝐶𝑉𝑉 ≈ 3𝑅𝑅 � � 𝑇𝑇
𝑇𝑇 2
𝜃𝜃𝐸𝐸
�exp � ��
𝑇𝑇
2
𝜃𝜃𝐸𝐸 𝜃𝜃𝐸𝐸
= 3𝑅𝑅 � � exp �− �
𝑇𝑇 𝑇𝑇

⟶ 0 as 𝑇𝑇 ⟶ 0

Conclusion: 𝑇𝑇 depends on 𝐶𝐶𝑉𝑉 :

𝐶𝐶𝑉𝑉 ⟶ 3𝑅𝑅 at high 𝑇𝑇, 𝐶𝐶𝑉𝑉 ⟶ 0 at low 𝑇𝑇

12
CV 3R
1

0.8

0.6

0.4

0.2

0.2 0.4 0.6 0.8 1 1.2 1.4 T θE

13
7.1.3 Debye’s model of vibrations in a solid
A solid with N atoms has 3N degrees of freedom.
Classical analysis: normal modes of solid, i.e.
propagation of sound waves, are elastic, have low
frequency and with long wavelength (λ >> a , a =
inter-atomic spacing),

ω c=
= Sk cS k

ω = angular frequency, k = wave vector
cS = speed of sound waves.
14
Based on classical results, Debye ignored discrete
atomic structure of solid, assumed solid as a
homogeneous and isotropic elastic medium.
Calculation of number of states (or normal modes)
is based on solutions of SHO with appropriate BCs:
2π 2π 2π
kx = nx , k y = ny , k z nz
Lx Ly Lz
nx , n y , nz = 0, 1, 2, 

“Volume” of one mode in k space :
15
 ∆=
k x ∆k y ∆k z
( 2π=
)( 2π )( 2π ) 8π 3
Lx Ly Lz V
V
DOS in k-space =
8π 3

Number of modes (NOM) in a spherical shell is

2

3 (
4π k dk ) =
V 2 Vk
dk
8π 2π 2

16
 
Each k has 3 independent modes (3 directions of
polarizations): one longitudinal waves
(compression) and two transverse waves (shear).
“ volume ” of spherical shell: 4𝜋𝜋𝑘𝑘 2 𝑑𝑑𝑑𝑑

𝑘𝑘

𝜔𝜔

𝑘𝑘 + 𝑑𝑑𝑑𝑑
𝜔𝜔 + 𝑑𝑑𝑑𝑑
17
Longitudinal and transverse waves have different
velocities in solid, cL and cT .
NOM with frequency between ω and ω + dω is
Vk2
Vω 2
1 2
dk →  3 + 3  dω
2π 2
2π 2
 cL cT 
𝜔𝜔 𝑑𝑑𝑑𝑑 𝜔𝜔 2
Note: 𝑘𝑘 = , 𝑑𝑑𝑑𝑑 = , 𝑘𝑘 2 =
𝑐𝑐 𝑐𝑐 𝑐𝑐 2

For low ω , assume c=

L c=
T cS , then NOM within
(ω , ω + dω ) :

18
 3ω 2
D(ω )dω = V dω (*)
2π cS
2 3

D(ω ) = density of states (DOS).

From classical SHO, a solid of N atoms possesses

3N normal modes.
Debye assumed (*) apply to all normal modes.
He introduced maximum frequency ωD ( Debye
ωD
frequency ) such that ∫ D(ω )dω = 3N .
0

19
 𝜔𝜔 𝐷𝐷
3𝑉𝑉
∴ 2 3 � 𝜔𝜔2 𝑑𝑑𝑑𝑑 = 3𝑁𝑁
2𝜋𝜋 𝑐𝑐𝑆𝑆 0
3 𝜔𝜔 𝐷𝐷
3𝑉𝑉 𝜔𝜔
⟹ 3 � � = 3𝑁𝑁
2𝜋𝜋 2 𝑐𝑐𝑆𝑆 3 0
𝑉𝑉𝜔𝜔𝐷𝐷3
⟹ 3 = 3𝑁𝑁
2𝜋𝜋 2 𝑐𝑐𝑆𝑆
2 3
3
6𝑁𝑁𝜋𝜋 𝑐𝑐𝑆𝑆
⟹ 𝜔𝜔𝐷𝐷 =
𝑉𝑉
2 1⁄3
6𝑁𝑁𝜋𝜋
⟹ 𝜔𝜔𝐷𝐷 = 𝑐𝑐𝑆𝑆 � �
𝑉𝑉
20
Also
3
3
𝑉𝑉𝜔𝜔 𝐷𝐷
𝑐𝑐𝑆𝑆 =
6𝑁𝑁𝜋𝜋 2

3V ω 9 Nω 2 2
and D(ω ) =
=
2π cS
2 3
ωD
3

Mean energy of one mode :

1 1 
ε (ω=
)  +  ω
 2 exp ( β ω ) − 1 

21
Combine D (ω ) and ε , mean energy of solid is
ωD
1 1 
E= ∑ ω  +
2 exp ( β ω )  0
− 1 ∫
 = D (ω ) ε (ω ) d ω
all modes 
ωD
9 Nω 2 1 1 
∫0 ωD3 ω  2 + exp ( ω / kBT ) − 1  dω
 
ωD
9 N ωD 9 N  ω dω
3

8
+ 3
ωD ∫
0
exp ( ω / k BT ) − 1

22
 ωD
 ∂E  9N  ω 4
exp(ω / k BT )
CV = 
 ∂T V ω 3
D
∫
0
k BT [ exp(ω / k BT ) − 1]
2 2
dω

Note:
𝑑𝑑 ℏ𝜔𝜔 ℏ𝜔𝜔 ℏ𝜔𝜔
�exp � �� = − exp � �
𝑑𝑑𝑑𝑑 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇 2 𝑘𝑘𝐵𝐵 𝑇𝑇

Illustration of E and CV features:

change variable 𝑥𝑥 ⟶ 𝜔𝜔 by

23
 ω ω D θ D ωD
letting x = ,=
xD = with θ D =
k BT k BT T kB

ℏ 𝑘𝑘𝐵𝐵 𝑇𝑇 𝜔𝜔 𝑥𝑥
𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑑𝑑 , 𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑑𝑑 , =
𝑘𝑘𝐵𝐵 𝑇𝑇 ℏ 𝜔𝜔𝐷𝐷 𝑥𝑥𝐷𝐷

9 Nk Bθ D
xD 3
9 Nk BT x dx
=
∴E
8
+ 3
xD ∫
0
e −1
x

24
 xD 4 x
9 Nk B x e
CV = 3 ∫ dx
( e − 1)
2
xD 0
x

At high T, T >> θ D or x << 1 , e ≈ 1 + x +

x

9 Nk Bθ D
xD 3
9 Nk BT x dx
E
8
+ 3
xD ∫0 1 + x +  − 1
9 Nk Bθ D
≈ + 3 Nk BT
8

25
 9 Nk B
xD
x (1 + x + )
4
9 Nk B
xD

∫0 (1 + x + ) − 1 2 dx ≈ xD3 ∫ x dx
2
CV 3
xD
  0

=
 ≈ 3 Nk B = 3R

At low T , T << θ D or xD → ∞ .
∞ 4
x e x
4π 4
By using ∫0 e x − 1 2 dx = 15
( )
3
9 Nk B 4π 4
12π 4
T 
CV = Nk B   ∝ T 3

 θD 
3
xD 15 5
26
CV 3R
1

0.8

0.6

0.4

0.2

0.2 0.4 0.6 0.8 1 1.2 1.4 T θD

27
7.2 Black body radiation
All objects emit photons at a rate varies with T.
At thermal equilibrium, all objects emit and
absorb radiation at exactly the same rate.
A black body will absorb all radiation incident
upon it, none is reflected.
Artificial black body: a very small hole on the wall
of an opaque enclosure with a large cavity inside.
This hole absorbs all radiation incident upon it
and none is reflected.
1
2
Radiation in cavity hit the wall and rattles around
and comes to thermal equilibrium with the wall
before emerging from the hole.
Experiment: enclosure is heated up to T,
(i) radiation from the hole depends on T and size
of the hole (area A),
(ii) independent of wall’s materials.
Radiation inside the cavity is assumed as
ensemble of perfect quantal gas (photon gas),
with mean occupation number of single particle
state
3
 1
𝑛𝑛�Λ =
exp(𝛽𝛽𝜀𝜀Λ ) − 1

7.2.1 Derivation of Planck’s radiation law

If the enclosure is a cube, 𝐿𝐿𝑥𝑥 = 𝐿𝐿𝑦𝑦 = 𝐿𝐿𝑧𝑧 = 𝐿𝐿 , then
2 1⁄2
ℏ𝑐𝑐�𝑘𝑘𝑥𝑥2
𝜀𝜀𝑟𝑟 = ℏ𝜔𝜔𝑟𝑟 = ℏ𝑐𝑐𝑘𝑘𝑟𝑟 = + 𝑘𝑘𝑦𝑦2
+ 𝑘𝑘𝑧𝑧 �
2𝜋𝜋ℏ𝑐𝑐 2 2 1⁄2
= �𝑛𝑛𝑥𝑥 + 𝑛𝑛𝑦𝑦2 +
𝑛𝑛𝑧𝑧 �
𝐿𝐿
2𝜋𝜋ℏ𝑐𝑐 2 2 2 1⁄2
= 1⁄3 �𝑛𝑛𝑥𝑥 + 𝑛𝑛𝑦𝑦 + 𝑛𝑛𝑧𝑧 �
𝑉𝑉
(i)
4
 r = ( nx , n y , nz ) , nx , n y , nz = 1, 2,.
Mean pressure on the wall due to a photon with
energy 𝜀𝜀𝑟𝑟 is
 ∂ε r  1 εr
pr =
− ==
 ∂V  3V
Mean pressure due to all photons is
1 1
= P ∑ = nr pr ∑
= nrε r E
r 3V r 3V
Calculation of photons energy density in terms of
𝝎𝝎 and T :
5
(A) Calculation of number of states (NOM):
“Volume” of one state in k space is

∆=
k x ∆k y ∆k z
( 2π
=
)( 2π )( 2π ) 8π 3

Lx Ly Lz V
V
DOS in 𝑘𝑘-space is
8π 3

NOM in spherical shell, k to k + dk , is

2 2

3 (
4π k dk ) = 2 dk ⇒ f (k )d k = 2 dk
V Vk Vk
2 3
(ii)
8π 2π 2π
6
(B) A photon with momentum p has 2 mutually
perpendicular linear polarizations.
2
Vk
(ii) becomes f (k )d k =
3
dk
π 2

Number of photon states within [ω , ω + dω ] is

V  ω  dω Vπ
2

f (ω )dω =
2   ω dω2
(iii)
π c c (π c )
3

1
Combine (iii), nΛ = & ε Λ = ω :
exp ( βε Λ ) − 1
7
yield number of photons within [ω , ω + dω ] :
Vπ ω dω
2
dNω =
(π c) exp( β ω ) − 1
3

Energy density of photon radiation within [ω ,

ω + dω ] :
dNω  ω dω 3
ω , T )dω =
U (= ω (iv)
V π c exp( β ω ) − 1
2 3

→ Planck’s radiation law, and

8
  ω3
U (ω , T ) = 2 3 →Planck’s formula
π c exp( β ω ) − 1
(v)
U (ω , T ) gives energy density distribution as a
function of radiation’s frequency ω in thermal
equilibrium at T.
(iv) or (v) give complete description of black-
body radiation, in excellent agreement with
observations.

#Limiting features for Planck’s radiation law:

9
(a) In high ω , β ω >> 1, ∴ exp( β ω ) >> 1
 ω dω
3
U (ω , T )dω ≈ 2 3 → Wein’s law
π c exp( β ω )

(b) In low ω , β ω << 1, ∴ exp( β ω ) ≈ 1 + β ω + 

 ω dω
3
U (ω , T )dω ≈ 2 3
π c 1 + β ω +  − 1
 ω dω k BT 2
3
= ω dω → Rayleigh-Jeans formula
π c β ω π c
2 3 2 3

10
At low ω , energy distribution match equipartition
of energy: mean energy of a SHO is k BT .
Integration of (iv) over all ω gives total energy
density of black-body radiation at T:
∞ ∞
 ω dω
3
U (T ) ∫0=
U (ω , T )dω 2 3 ∫
π c 0 exp( β ω ) − 1
(vi)

ω k BT k BT
Let=η β= ω , ∴ ω = η , dω = dη
k BT  
4 ∞ 3
  k BT  η dη π kB 4
2 4
→ U (T ) 2 3 
=  ∫ η T
π c    0 e − 1 15 c 3 3

11
→ U (T ) ∝ T 4
→ Stefan-Boltzmann law
Kinetic theory of gases: radiation energy per unit
area incident on walls of black-body:
c
U (T ) = σ T 4

4
Area of black-body = A .
Rate of radiation energy emitted by black-body :
dQ c
= = U (T ) A σ AT 4

dt 4

12
 π k
2 4
σ
= =
3 2
B −8
5.67 × 10 Jm S K -2 -1 -4

60 c

= Stefan’s constant
Find ωmax corresponds to maximum U (ω , T ) : apply
 ∂U (ω , T ) 
  =0
 ∂ω T
 ∂  ω 3

∴ 2 3   =0
π c ∂ω  exp ( β ω ) − 1 T

13
 exp ( β ω ) − 1 3ω − ω β  exp ( β ω )
2 3

⇒ =
0
exp ( β ω ) − 1
2

⇒ 3 exp ( β ω ) − 3 − β ω exp ( β ω ) =0
⇒ exp(η ) (3 − η ) =
3
3
∴ exp(η ) = (vii)
3 −η
solve (vii) graphically or numerically:

14
η 3
e , 3 −η

20

15

η 3
10 e 3 −η

0.5 1 1.5 2 2.5

15
 ωmax
= η=
2 2.82 (**)
k BT
η2 k BT 2.82 k BT
∴ωmax= = (viii)
 
ωmax = maximum frequency of black-body at T.
Two black-bodies at T1 and T2 , from (**),
ωmax ωmax
2.82 k B
= = 1 2

T1 T2 
 
Wein's displacement law
Surface T of sun (black-body) = 6000K.
16
U (ω ,6000 K ) :
8E+44
6000K
7E+44

6E+44

5E+44

4E+44

3E+44

2E+44 4000K

1E+44

0
2000K ω
0 2E+15 4E+15 6E+15 8E+15 1E+16

17
 −23
2.82 k BT 2.82 × (1.38066 × 10 ) × 6000
ωmax = −34
 1.05456 × 10
≈ 2.2163 × 1015
radian / s
Frequency:
=
f max 3.5273 × 10 14
Hertz

Value of ωmax match the maximum of U (ω ,6000 K ) .

18
7.3 Application of Fermi-Dirac distribution:
conduction electrons in metals
Properties of conduction electrons in metals,
simplest approximation: free electron model (FEM):
Valence electrons that weakly bound to atomic ion
cores: detached in metal and roam around freely.
Apply potential difference: these electrons produce
a current ⟶ conduction electron of metals.
In FEM, conduction electrons ⟶ perfect quantal gas
obeying Fermi-Dirac (FD) statistics:
1
 1
nΛ = (i)
exp ( βε Λ − βµ ) + 1

7.3.1 Properties of Fermi’s function

In metal, density of single particle states is large,
discrete energies are very closed to each other.
nΛ ⟶ continuous function of energy ⟶ Fermi’s
function:
1
𝐹𝐹 (𝜀𝜀 ) =
exp(𝛽𝛽𝛽𝛽 − 𝛽𝛽𝛽𝛽) + 1
(iii)
2
(a) If 𝛽𝛽𝛽𝛽 ≪ 0, exp(𝛽𝛽𝛽𝛽 − 𝛽𝛽𝛽𝛽) ≫ 1
𝐹𝐹 (𝜀𝜀 ) → exp{−(𝛽𝛽𝛽𝛽 − 𝛽𝛽𝛽𝛽)} → MB distribution
(b) (i) If βµ >> 1 and ε << µ , β (ε − µ ) << 0

and exp {β (ε − µ )} << 1 , F (ε ) → 1

(ii) If βµ >> 1 and ε >> µ , β (ε − µ ) >> 1
and exp {β (ε − µ )} >> 1
F (ε ) → exp {− β (ε − µ )} → MB distribution

3
 =
(c) If ε µ=
, F (ε ) 0.5
(i) T = 0K , µ = 8 eV
F (Fε )

0.8

0.6

0.4 εF
8eV
0.2

-19 -18 -18 -18 ε

510 110 1.5 10 210
4
(ii) T = 5800K , µ = 8 eV
F

0.8

0.6 1 ε F = 8eV
2
0.4

0.2

-19 -18 -18 -18


510 110 1.5 10 210

2k BT
Figure A
5
7.3.2 Fermi-Dirac energy distribution
Number of states in k-space :
4π V
f (k )dk = 2
k dk (iv)
( 2π )
3

Electrons: 12  particles → 2 independent spin states

for each k,
8𝜋𝜋𝜋𝜋 2
→ 𝑓𝑓 (𝑘𝑘 )𝑑𝑑𝑑𝑑 = 𝑘𝑘 𝑑𝑑𝑑𝑑
(2𝜋𝜋) 3
(v)

6
Convert (v) into energy DOS by using 𝜀𝜀 = ℏ2 𝑘𝑘 2 ⁄2𝑚𝑚 ,
𝑚𝑚 = electron mass
2
2𝑚𝑚𝑚𝑚 √2𝑚𝑚 √2𝑚𝑚 1 1
𝑘𝑘 = 2 , 𝑘𝑘 = √𝜀𝜀 , 𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑑𝑑
ℏ ℏ ℏ 2 √𝜀𝜀

Number of states within energy interval [𝜀𝜀 , 𝜀𝜀 + 𝑑𝑑𝑑𝑑 ] :

8𝜋𝜋𝜋𝜋 2𝑚𝑚𝑚𝑚 √2𝑚𝑚 1
𝑓𝑓(𝜀𝜀 )𝑑𝑑𝑑𝑑 = � � � 𝑑𝑑𝑑𝑑�
(2𝜋𝜋)3 ℏ2 2ℏ √𝜀𝜀
4𝜋𝜋𝜋𝜋
= 3 (2𝑚𝑚)3⁄2 √𝜀𝜀𝑑𝑑𝑑𝑑
ℎ
(vi)
7
Combine eqns. (iii) and (vi)
( ) 3⁄2
4𝜋𝜋𝜋𝜋 2𝑚𝑚 √𝜀𝜀
𝑑𝑑𝑑𝑑(𝜀𝜀 ) = 𝐹𝐹 (𝜀𝜀 )𝑓𝑓(𝜀𝜀 )𝑑𝑑𝑑𝑑 = 3 𝑑𝑑𝑑𝑑
ℎ exp(𝛽𝛽𝛽𝛽 − 𝛽𝛽𝛽𝛽) + 1
(vii)
→ Fermi-Dirac energy distribution → number of
electrons within [𝜀𝜀 , 𝜀𝜀 + 𝑑𝑑𝑑𝑑 ].

8
 dN (ε ) d ε , µ = 8 eV
1.20E-09

1.00E-09

8.00E-10

6.00E-10

4.00E-10

2.00E-10

0.00E+00
0 0.4 0.8 1.2 1.6

Figure B
9
Total number of electrons :
∞ ( ) 3⁄2 ∞
4𝜋𝜋𝜋𝜋 2𝑚𝑚 √𝜀𝜀
𝑁𝑁 = � 𝑑𝑑𝑑𝑑(𝜀𝜀 ) = 3
� 𝛽𝛽𝛽𝛽 −𝛽𝛽𝛽𝛽 𝑑𝑑𝑑𝑑
0 ℎ 0 𝑒𝑒 +1
(viii)
Eqn.(viii) determines the chemical potential
𝜇𝜇 = 𝜇𝜇(𝑇𝑇, 𝑉𝑉, 𝑁𝑁).
Metal: V and N are fixed, 𝜇𝜇 = 𝜇𝜇(𝑇𝑇).
Fermi energy ε F = value of chemical potential at
T = 0:
10
 ε F µ=
= (T 0)

7.3.3 Fermi energy ε F for electron gas at T = 0

At T = 0, system in lowest energy state.
Due to Pauli’s exclusion principle, only one particle
per single particle state (2 independent states for
spin 12  particle had considered).
ε1 ≤ ε 2 ≤ ε ≤ , the first N lowest single particle
states are filled, and the rest are empty.

11
Fermi energy ε F is the topmost occupied energy
level at T = 0.
At T = 0, β → ∞ . If ε − ε F < 0, then
exp ( βε − βε F ) → 0 , (viii) becomes

4π V ( 2m ) εF
8π V ( 2m )
32 32

= ∫0 ε ε ε 32
N 3
d 3 F
h 3h
(*)
23
h  3 N
2
Rearranging (*): ε F =   (ix)
2m  8π V 
12
The effect on the conduction electrons of raising the
temperature from 0K to 𝑇𝑇K is shown in Figs. A or B.
For such a temperature change, we expect the
energy of an electron to increase by about 𝑘𝑘𝐵𝐵 𝑇𝑇.

But for most electrons, with the exception of those

whose energy is close to 𝜀𝜀𝐹𝐹 , the states to which they
would be excited are already occupied.

So because of the Pauli principle they can’t be

excited into these states.

13
It is only a small fraction of electrons with energy
near 𝜀𝜀𝐹𝐹 which can be excited, namely those which lie
approximately within an energy interval kT of the
Fermi energy 𝜀𝜀𝐹𝐹 .

This number of excited electrons 𝑁𝑁𝑒𝑒𝑒𝑒 is given by

𝑁𝑁𝑒𝑒𝑒𝑒 ~ 𝑓𝑓(𝜀𝜀𝐹𝐹 )𝑘𝑘𝐵𝐵 𝑇𝑇

(@)
From (vi)
4𝜋𝜋𝜋𝜋
𝑓𝑓(𝜀𝜀𝐹𝐹 ) = 3 (2𝑚𝑚)3⁄2 �𝜀𝜀𝐹𝐹
ℎ
14
 (**)
and from (*),
8𝜋𝜋𝜋𝜋 ⁄
3 2 3
𝑁𝑁 = 3
( 2𝑚𝑚 ) ��𝜀𝜀𝐹𝐹 �
3ℎ
(***)

Divide (**) by (***), and rearrange,

3𝑁𝑁
𝑓𝑓 (𝜀𝜀𝐹𝐹 ) =
2𝜀𝜀𝐹𝐹
(#)
Substitute (#) in (@),
15
 3𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 3𝑁𝑁 𝑇𝑇
𝑁𝑁𝑒𝑒𝑒𝑒 ~ =
2𝜀𝜀𝐹𝐹 2 𝑇𝑇𝐹𝐹
where 𝜀𝜀𝐹𝐹 = 𝑘𝑘𝐵𝐵 𝑇𝑇𝐹𝐹 .

⁄ 4
(Note: for 𝜀𝜀𝐹𝐹 = 8 𝑒𝑒𝑒𝑒 , 𝑇𝑇𝐹𝐹 = 𝜀𝜀𝐹𝐹 𝑘𝑘𝐵𝐵 ~ 9.3 × 10 𝐾𝐾)

The excitation energy per electron is of the order of

𝑘𝑘𝐵𝐵 𝑇𝑇.

Hence the electronic energy (relative to the ground

state at T = 0 K) is given by

16
 3𝑁𝑁𝑘𝑘𝐵𝐵 𝑇𝑇 2
𝐸𝐸 (𝑇𝑇) ~ 𝑁𝑁𝑒𝑒𝑒𝑒 𝑘𝑘𝐵𝐵 𝑇𝑇 =
2 𝑇𝑇𝐹𝐹
and the electronic heat capacity becomes
𝜕𝜕𝜕𝜕 𝑇𝑇
𝐶𝐶𝑉𝑉 (𝑇𝑇) = � � ~ 3𝑁𝑁𝑘𝑘𝐵𝐵
𝜕𝜕𝜕𝜕 𝑉𝑉 𝑇𝑇𝐹𝐹
At room temperature the electronic heat capacity is
very small compared with the lattice heat capacity
of about 3𝑁𝑁𝑘𝑘𝐵𝐵 and will give a contribution of only a
few per cent to the observed heat capacity of metals.

At very low temperatures the situation is different.

17
We saw in Chapter 7a that at low temperatures the
lattice heat capacity is proportional to 𝑇𝑇 3 whereas
the electronic heat capacity varies linearly with 𝑇𝑇.
At sufficiently low temperatures the electronic
contribution dominates. This behavior is fully borne
out by experiment.

Note: Alternative method to achieve eqn. (ix)

without integration: each electron’s energy is 𝑘𝑘𝐵𝐵
2 2 2
p k
=εk =
2m 2m
18
 2 2 2
p k
∴ε F = = F F
2m 2m
At T = 0, all the states with k < k F are filled, and with
k > k F empty.
“Volume” of sphere with radius k F is 4π k 3
F 3 .
There are V /(2π ) states per unit “volume in k-
3

space”.
Fermi sphere with radius k F contains

19
 V 4 3
π kF states (*)
(2π ) 3
3

Electrons: spin 1
2
 particles, each k has 2 possible
spin states :
V 4 3 V 3
= N 2= π kF k
(2π ) 3
3
3π 2 F

or
1/ 3
 3π N 
2
kF =  
 V 
2/3
k   3π N 
2 2 2 2
∴ ε F= = 
F

2m 2m  V  20
Chapter 8 The grand canonical ensemble

8.1 The grand canonical ensemble

An open system can exchange both heat and matter
with the surroundings and, therefore, both energy
and particle number will fluctuate.
For an open system in equilibrium, both average
energy 〈𝐸𝐸 〉, and average particle number 〈𝑁𝑁〉 are
fixed.
The system exchanges energy with a heat reservoir
at temperature 𝑇𝑇, which determines the average
1
energy 〈𝐸𝐸 〉, and it exchanges particles with a
"particle reservoir" of chemical potential 𝜇𝜇, which
determines the average number of particles 〈𝑁𝑁〉.

heat reservoir 𝑇𝑇

particle reservoir 𝜇𝜇

2
The grand canonical ensemble is a more realistic
representation of physical systems than the
canonical ensemble, for we can rarely fix the total
number of particles in a macroscopic system.

Eg 1. A volume of air in the atmosphere: the particle

reservoir is the rest of the atmosphere. Number of
air molecules in the volume fluctuates about a mean
value determined by the rest of the atmosphere.∎

Grand canonical ensemble can also be considered as

an ensemble of many identical systems, with both
3
energy and particles flowing from each system to its
neighbours.

Mean value of any physical quantities is calculated

through averaging over the microstates of the
ensemble.

Footnote: To obtain the equilibrium probability density in such a

fluctuating environment, one can use the method of Lagrange
multipliers to extremize the Gibbs entropy subject to three
constraints.

4
8.2 The particle reservoir

The grand canonical ensemble is built upon the

canonical ensemble by relaxing the restriction to a
definite number of particles.

In canonical ensemble, the relative probability of

finding the system in an accessible state with fixed
𝑁𝑁 particles at temperature 𝑇𝑇 , in a volume 𝑉𝑉 , is

5
 exp(−𝛽𝛽𝐸𝐸𝑆𝑆 )
𝑃𝑃(𝐸𝐸𝑆𝑆 , 𝑉𝑉, 𝑇𝑇) = 𝐶𝐶 exp(−𝛽𝛽𝐸𝐸𝑆𝑆 ) =
∑𝑆𝑆 exp(−𝛽𝛽𝐸𝐸𝑆𝑆 )
exp(−𝛽𝛽𝐸𝐸𝑆𝑆 )
=
𝑍𝑍𝑁𝑁

𝑍𝑍𝑁𝑁 = � exp(−𝛽𝛽𝐸𝐸𝑆𝑆 )
𝑆𝑆

In grand canonical ensemble, the relative probability

of finding the system with 𝑁𝑁 particles at
temperature 𝑇𝑇 , in a volume 𝑉𝑉 , is
6
 𝑁𝑁
𝜌𝜌 𝑁𝑁, 𝐸𝐸𝑆𝑆 , 𝑉𝑉, 𝑇𝑇 = 𝒞𝒞 𝓯𝓯 𝑍𝑍𝑁𝑁 = 𝒞𝒞 exp(𝑁𝑁𝑁𝑁𝑁𝑁) 𝑍𝑍𝑁𝑁
( )
∑𝑆𝑆 exp{−𝛽𝛽(𝐸𝐸𝑆𝑆 − 𝜇𝜇𝜇𝜇)} 𝑁𝑁
𝓯𝓯 𝑍𝑍𝑁𝑁
= =
∑𝑁𝑁 ∑𝑆𝑆 exp{−𝛽𝛽(𝐸𝐸𝑆𝑆 − 𝜇𝜇𝜇𝜇)} ∑𝑁𝑁 𝓯𝓯𝑁𝑁 𝑍𝑍𝑁𝑁
= 𝐠𝐠𝐠𝐠𝐠𝐠𝐠𝐠𝐠𝐠 𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜𝐜 𝐝𝐝𝐝𝐝𝐝𝐝𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭𝐭 𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟𝐟

where 𝓯𝓯 = exp(𝛽𝛽𝛽𝛽) is the fugacity, and 𝜇𝜇 is the

chemical potential.

It is useful to introduce the grand partition function:

7
 { ( )} 𝑁𝑁
𝓩𝓩 = � � exp −𝛽𝛽 𝐸𝐸𝑆𝑆 − 𝜇𝜇𝜇𝜇 = � 𝓯𝓯 𝑍𝑍𝑁𝑁
𝑁𝑁 𝑆𝑆 𝑁𝑁

= normalization constant

Eg2. Suppose there is a surface on which there are

sites where atoms can be absorbed; each site can be
empty or it can be occupied by only one atom.

Find the mean occupation number of particles on

site-𝑖𝑖 , i.e. 𝑛𝑛�𝑖𝑖 .

8
Answer: For site-𝑖𝑖 , when it is vacant, its energy is
zero:
𝑛𝑛𝑖𝑖 = 0 , 𝐸𝐸𝑖𝑖 = 0
0 𝛽𝛽𝛽𝛽 0 −𝛽𝛽0
Boltzmann factor = 𝓯𝓯 𝑍𝑍0 = �𝑒𝑒 � 𝑒𝑒 =1

When it is occupied, its energy is 𝜀𝜀𝑖𝑖 :

𝑛𝑛𝑖𝑖 = 1 , 𝐸𝐸𝑖𝑖 = 𝜀𝜀𝑖𝑖

Boltzmann factor
1 𝛽𝛽𝛽𝛽 1 −𝛽𝛽 𝜀𝜀 𝑖𝑖 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 )
= 𝓯𝓯 𝑍𝑍1 = �𝑒𝑒 � 𝑒𝑒 = 𝑒𝑒
9
The grand partition function of site-𝑖𝑖 is the sum of
all Boltzmann factors of all its accessible
microstates:

𝓩𝓩𝒊𝒊 = � 𝓯𝓯𝑛𝑛 𝑖𝑖 𝑍𝑍𝑛𝑛 𝑖𝑖 = 𝓯𝓯0 𝑍𝑍0 + 𝓯𝓯1 𝑍𝑍1 = 1 + 𝑒𝑒 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 )

𝑛𝑛 𝑖𝑖

∑𝑛𝑛 𝑖𝑖 𝑛𝑛𝑖𝑖 𝓯𝓯𝑛𝑛 𝑖𝑖 𝑍𝑍𝑛𝑛 𝑖𝑖 0(1) + 1�𝑒𝑒 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 ) �

𝑛𝑛�𝑖𝑖 = =
∑𝑛𝑛 𝑖𝑖 𝓯𝓯𝑛𝑛 𝑖𝑖 𝑍𝑍𝑛𝑛 𝑖𝑖 𝓩𝓩𝒊𝒊
𝑒𝑒 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 ) 1
= ( )
=
1 + 𝑒𝑒𝛽𝛽 𝜇𝜇 −𝜀𝜀 𝑖𝑖 𝑒𝑒 𝛽𝛽 (𝜀𝜀 𝑖𝑖 −𝜇𝜇 ) + 1
10
8.3 The grand potential

In the grand canonical ensemble, there is a relation

between the grand potential, Φ𝐺𝐺 (a new
thermodynamic energy) and the grand partition
function, 𝓩𝓩 .

For a system in contact with a heat reservoir, its

entropy is
𝑆𝑆 = −𝑘𝑘𝐵𝐵 � 𝑝𝑝𝑆𝑆 ln(𝑝𝑝𝑆𝑆 )
𝑆𝑆
(same as canonical ensemble)
11
In grand canonical ensemble, the probability that
the system is in any microstate 𝜓𝜓𝑆𝑆 (with energy 𝐸𝐸𝑆𝑆 )
is

−𝛽𝛽 (𝐸𝐸𝑆𝑆 −𝜇𝜇𝑁𝑁 )

𝑒𝑒
𝜌𝜌𝑆𝑆 =
𝓩𝓩

Therefore for system in contact with both heat and

particle reservoirs ,

12
𝑆𝑆 = −𝑘𝑘𝐵𝐵 � � 𝜌𝜌𝑆𝑆 ln(𝜌𝜌𝑆𝑆 )
𝑁𝑁 𝑆𝑆
−𝛽𝛽 (𝐸𝐸𝑆𝑆 −𝜇𝜇𝑁𝑁 )
𝑒𝑒
= −𝑘𝑘𝐵𝐵 � � 𝜌𝜌𝑆𝑆 ln � �
𝓩𝓩
𝑁𝑁 𝑆𝑆
= −𝑘𝑘𝐵𝐵 � � 𝜌𝜌𝑆𝑆 �ln 𝑒𝑒 −𝛽𝛽 (𝐸𝐸𝑆𝑆 −𝜇𝜇𝑁𝑁 ) − ln 𝓩𝓩�
𝑁𝑁 𝑆𝑆

𝜌𝜌𝑆𝑆 𝐸𝐸𝑆𝑆 − 𝜇𝜇 𝜌𝜌𝑆𝑆 𝑁𝑁

= � �� + (𝑘𝑘𝐵𝐵 ln 𝓩𝓩)𝜌𝜌𝑆𝑆 �
𝑇𝑇
𝑁𝑁 𝑆𝑆

13
 1 𝜇𝜇
= � � 𝜌𝜌𝑆𝑆 𝐸𝐸𝑆𝑆 − � � 𝜌𝜌𝑆𝑆 𝑁𝑁
𝑇𝑇 𝑇𝑇
��
𝑁𝑁 ����
𝑆𝑆 ��� ��
𝑁𝑁 �
���
𝑆𝑆 �
��
=𝐸𝐸� �
=𝑁𝑁
+ 𝑘𝑘𝐵𝐵 ln 𝓩𝓩 � � 𝜌𝜌𝑆𝑆
��
𝑁𝑁 ���
𝑆𝑆 ��
=1

𝐸𝐸� 𝜇𝜇𝑁𝑁
�
∴ 𝑆𝑆 = − + 𝑘𝑘𝐵𝐵 ln 𝓩𝓩
𝑇𝑇 𝑇𝑇
where
𝐸𝐸� = average energy ,
14
 � = average number of particles.
𝑁𝑁
Multiply every term with 𝑇𝑇 ,

𝑇𝑇𝑇𝑇 = 𝐸𝐸� − 𝜇𝜇𝑁𝑁

� + 𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝓩𝓩
Rearrange,
−𝑘𝑘���
�� 𝐵𝐵 𝑇𝑇 ln
��𝓩𝓩 = �
𝐸𝐸 − �
𝜇𝜇𝑁𝑁 − 𝑇𝑇𝑇𝑇
Φ 𝐺𝐺

Φ𝐺𝐺 = −𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝓩𝓩 = 𝐸𝐸� − 𝜇𝜇𝑁𝑁

� − 𝑇𝑇𝑇𝑇

is called the grand potential.

15
All the thermodynamics for systems with variable
numbers of particles can be derived from the grand
potential Φ𝐺𝐺 .

In grand canonical ensemble, the entropy is a

function of 𝐸𝐸� , 𝑉𝑉 , 𝑁𝑁
� , i.e. 𝑆𝑆 = 𝑆𝑆(𝐸𝐸� , 𝑉𝑉, 𝑁𝑁
� ).

Thus the exact differential of 𝑆𝑆 is

𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

𝑑𝑑𝑑𝑑 = � � 𝑑𝑑𝐸𝐸� − � � 𝑑𝑑𝑑𝑑 + � � 𝑑𝑑𝑁𝑁 �
𝜕𝜕𝐸𝐸� V,𝑁𝑁� 𝜕𝜕𝜕𝜕 𝐸𝐸� ,𝑁𝑁� � 𝐸𝐸� ,V
𝜕𝜕𝑁𝑁

16
 1 𝑃𝑃 𝜇𝜇
= � � 𝑑𝑑𝐸𝐸� + � � 𝑑𝑑𝑑𝑑 − � � 𝑑𝑑𝑁𝑁
�
𝑇𝑇 𝑇𝑇 𝑇𝑇
Hence
𝑑𝑑𝐸𝐸� = −𝑃𝑃𝑃𝑃𝑃𝑃 + 𝜇𝜇𝜇𝜇𝑁𝑁
� + 𝑇𝑇𝑇𝑇𝑇𝑇

Thus the exact differential of Φ𝐺𝐺 can be derived:

Φ𝐺𝐺 = 𝐸𝐸� − 𝜇𝜇𝑁𝑁
� − 𝑇𝑇𝑇𝑇
𝑑𝑑Φ𝐺𝐺 = 𝑑𝑑𝐸𝐸� − 𝑑𝑑 (𝜇𝜇𝑁𝑁
� ) − 𝑑𝑑(𝑇𝑇𝑇𝑇)
⟹ 𝑑𝑑Φ𝐺𝐺
� + 𝑇𝑇𝑇𝑇𝑇𝑇) + (−𝜇𝜇 𝑑𝑑𝑁𝑁
= (−𝑃𝑃𝑃𝑃𝑃𝑃 + 𝜇𝜇𝜇𝜇𝑁𝑁 � − 𝑁𝑁
� 𝑑𝑑𝑑𝑑)
+ (−𝑇𝑇𝑇𝑇𝑇𝑇 − 𝑆𝑆𝑆𝑆𝑆𝑆)
17
 � 𝑑𝑑𝑑𝑑 − 𝑆𝑆𝑆𝑆𝑆𝑆
⟹ 𝑑𝑑Φ𝐺𝐺 = −𝑃𝑃𝑃𝑃𝑃𝑃 − 𝑁𝑁
The last equation can be used to obtain the entropy,
the pressure and the average number of particles:

𝜕𝜕Φ𝐺𝐺 𝜕𝜕(𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝓩𝓩)

𝑆𝑆 = − � � =� �
𝜕𝜕𝜕𝜕 𝑉𝑉,𝜇𝜇 𝜕𝜕𝜕𝜕 𝑉𝑉,𝜇𝜇

𝜕𝜕Φ𝐺𝐺 𝜕𝜕(𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝓩𝓩)

𝑃𝑃 = − � � =� �
𝜕𝜕𝜕𝜕 𝑇𝑇,𝜇𝜇 𝜕𝜕𝜕𝜕 𝑇𝑇,𝜇𝜇

18
 𝜕𝜕Φ𝐺𝐺 𝜕𝜕(𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝓩𝓩)
� = −�
𝑁𝑁 � =� �
𝜕𝜕𝜕𝜕 𝑉𝑉,𝑇𝑇 𝜕𝜕𝜕𝜕 𝑉𝑉,𝑇𝑇

Consequently, once 𝓩𝓩 is obtained, all the

thermodynamic quantities can be calculated.

As an example, in Eg2, the grand partition function

of site-𝑖𝑖 is

𝑛𝑛 𝑖𝑖 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 )
𝓩𝓩𝒊𝒊 = � 𝓯𝓯 𝑍𝑍𝑛𝑛 𝑖𝑖 = 1 + 𝑒𝑒
𝑛𝑛 𝑖𝑖
19
 Φ𝐺𝐺𝐺𝐺 = −𝑘𝑘𝐵𝐵 𝑇𝑇 ln 𝓩𝓩𝒊𝒊 = −𝑘𝑘𝐵𝐵 𝑇𝑇 ln�1 + 𝑒𝑒 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 ) �

𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 )
𝜕𝜕Φ𝐺𝐺𝐺𝐺 𝜕𝜕�ln�1 + 𝑒𝑒 ��
𝑛𝑛�𝑖𝑖 = − = 𝑘𝑘𝐵𝐵 𝑇𝑇
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
𝛽𝛽𝑒𝑒 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 ) 𝑒𝑒 𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 )
= 𝑘𝑘𝐵𝐵 𝑇𝑇 )
=
(
1 + 𝑒𝑒𝛽𝛽 𝜇𝜇 −𝜀𝜀 𝑖𝑖 1 + 𝑒𝑒𝛽𝛽 (𝜇𝜇 −𝜀𝜀 𝑖𝑖 )
1
= 𝛽𝛽 (𝜀𝜀 −𝜇𝜇 )
𝑒𝑒 𝑖𝑖 +1

The result is the same as eg2.

20