Statistical Mechanics

Macroscopic system: It is a system that is large enough to be observed in ordinary

sense.
Microscopic system: It is a system of roughly of the order of atomic dimension or
smaller.
So, a macroscopic system consists of a large number of microscopic systems such
as molecule, atoms, ions etc. Consider, for example, a mole of gas. The number of
molecules in a mole of gas is 6.023x1023. Suppose we get on constructing and integrating
the equation of motions of the molecules in a mole of gas. For describing the motion of
each molecule we must know its initial position and velocity. The position is determined
by the co-ordinates x,y,z and the velocity by the three components Vx,Vy,Vz. We
therefore, have to fix 6x6.023x1023 numbers, i.e., more than 3x1024 numbers in order to
describe the position and velocities of all the molecules at a particular instant. Suppose,
we have an instrument which can integrate can register numbers at the rate of 106 per
second. Even at this rate it would take 3x1018 s i.e., about 11 years! This is a hopeless
undertaking! But it is not only the case for which dynamic consideration impractical. The
main trouble is that the solution of the equation of motion requires initial values of all the
variables and can not record data on the initial positions and momenta of all atoms at a
particular instant. So we can not apply simply classical mechanics to solve such
problems.
So, the alternative approach is the thermodynamic approach. In thermodynamics,
macroscopic system in equilibrium (equilibrium means, when parameters of an isolated
system do not change in time) is studied from macroscopic constituent of the system. But
the thermodynamic parameters (temperature, pressure, volume etc.) can not at all be
independent of the micro-constituents that make the system.
Hence alternative approach is the statistical mechanics, where macroscopic
parameters of a system in equilibrium is studied from the knowledge of microscopic
properties of its constituents particles by applying statistical techniques and laws of
mechanics. So, in statistical technique, we get the average value without going into
details of the constituents particles. In this technique the greatest probability can be found
out but not the exact value. The deviation from the exact value decreases when the size of
system increases. Statistical predictions are nearly correct for a system dealing with a
large number of particles.

Some basic concepts of statistical mechanics

Energy levels and energy states: While describing the atomic structure of the atom
Bohr assumed that electrons are revolving around the nucleus. The specific orbits are
known as energy levels. Different energy levels can be represented by principal quantum
number n having integer values 1,2,3………The concept of energy levels explains the
stability of the atom.
When the electrons are present in the lowest energy level, the electron is said to
be in the ground state. If electron absorbs or emits energy it gets shifted into one to
another energy levels, known as the 1st excited state, 2nd excited states and so on.

_________________________________________________________________________________ 1
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
Phase space: For complete dynamical specification of a point in a system from
molecular viewpoint, 6 co-ordinates are requires, 3 positions (x,y,z) and 3 corresponding
momentum co-ordinates (Px,Py,Pz). The six dimensional space with the combination of
position space and momentum space is known as phase space.
The phase space may be subdivided into 6-dimensional volume element, called
phase cells with each sides δx, δy, δz, δPx, δPy, δPz and volume H= δxδyδzδPxδPyδPz.

Microstate and Macrostate: A complete specification of the set of 6 co-ordinates (δx,

δy, δz, δPx, δPy, δPz) of each particle of a system is called microstate of the system.
Each distinct arrangement of the particle as specified by their identity,
corresponding to a given microstate is known as the microstate of the system.
Suppose we have 4 phase points a,b,c,d (particles) and 2 cells in phase space.
Cell Distribution1 Distribution2 Distribution3 Distribution4 Distribution5
1 0 1 2 3 4
2 4 3 2 1 0
So, only 5 distributions (0,4),(1,3),(2,2),(3,1),(4,0) is possible. Each cell wise
distributions is known as macrostate.

Macrostate Possible arrangement No. of microstate

Cell 1 Cell 2
0,4 0 a,b,c,d 1
1,3 a bcd 4
b cda
c abd
d abc
2,2 ab cd 6
ac bd
ad bc
bc ad
bd ac
cd ab
3,1 bcd a 4
cda b
abd c
abc d
4,0 abcd 0 1
So, no. microstates=16
No. of macrostates=5
Thermodynamic probability: The number of possible microstates corresponding to any
given macrostate is known as thermodynamic probability. It is denoted by W. In the
above example thermodynamic probability (W) for (1,3) macrostate is 4.

Equilibrium state: If there is one such particular macrostate for which there are
maximum microstates, then that macrostate is called most probable state or equilibrium
state. It is denoted by Wmax. In above example the macrostate (2,2) is most probable state.
The observable property of a system is due to this state.
_________________________________________________________________________________ 2
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
Number of quantum states within energy range E and E+dE in phase space
or Density of states.

Let us consider a phase space of volume containing a system of particles.

A small volume element in the phase space, dτ=dxdydzdpxdyydpz (x,y,z,px,py,py, are the
co-ordinates of a point in phase space)
    dxdydzdp x dp y dp z   dxdydz  dp x dp y dp z
From Heisenberg’s uncertainty relation, dx.dpx≈h h=Plank’s constant
So, dx.dpx dy.dpy dz.dpz ≈h3
The number of cells in phase space corresponding to momentum p and p+dp is

g(p)dp=
 dxdydz  dp x dp y dp z
h3
Now, the volume of system of particles V=  dxdydz
V  dp x dp y dp z
So, g(p)dp=
h3
Now dpxdpydpz represents a volume element in momentum space and thus the volume
enclosed between two concentric spheres of radii p and p+dp, which is 4πp2dp
So, which is the no. of energy states between P and p+dP
V
g ( p)dp  3 4p 2 dp P+dP
h

Again, p2=2mE , E=energy of the particle of mass m

So, 2pdp=2mdE P
mdE
So, dp=mdE/P=
2mE
V 4 (2mE)mdE 2V
So, g(E)dE= 3  3 (2m) 3 / 2 E 1 / 2 dE
h 2mE h
So,
4 2Vm 3 / 2 E 1 / 2
g(E)dE= dE
h3

g(E)dE gives the number of quantum states per unit energy range at given energy E and is
known as the density of states.

Note: Since an electron has two independent direction of spin orientation,

8 2Vm 3 / 2 E 1 / 2
g(E)dE= dE
h3

_________________________________________________________________________________ 3
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
 Maxwell-Boltzmann (MB) statistics

Basic postulates:
(i) The particles of the system is distinguishable.
(ii) The particles are spin less.
(iii) Heisenberg’s uncertainty relation is not applicable, since the particles are
classical.
(iv) Since the particles are spin less, Pauli’s exclusion principle is not applicable.
(v) Any number of particles can be accommodated in any energy state.

MB statistical count: Thermodynamic probability

 Let Ni particles are to be arranged in gi states. So, the number of ways in which
Ni particles could be distributed in gi states is (gi)Ni.

 Let N be the total number of distinguishable particles in a system;

N1,N2,N3……Nn are number of particles with energies E1,E2,E3…..En respectively
and gi be the number of quantum states for the energy level Ei. n=number of
energy levels. Then the total number of ways, in which the particles are
N!
distributed, is W  n ( g i ) Ni
 i 1 N i !
W is known as thermodynamic probability.

Most probable distribution

For most probable distribution Wmax, entropy S=klnW must be maximum. It can
be shown S will be maximum when,
N i  g i e   Ei
Ni
Take f(Ei)=  e   Ei
gi
f(Ei) is called distribution function which is the ratio of the number of particles Ni to the
number of states gi. Ni/gi is also known as occupation index.

MB distribution function= e   Ei

It can be shown β=1/KT , K=Boltzmann’s constant=1.38x10-23 JK-1

Ni 1
So, distribution function=f(Ei)= =  Ei / KT
gi e e

Limitation of MB statistics
In MB statistics it is considered that the particles of a system (i.e., ideal gas) are
distinguishable. So, it can explain the energy distribution of the molecules of ideal gas to
a fair degree of accurate. But this statistics failed to explain the energy distribution of
indistinguishable particles like electrons and photons.
_________________________________________________________________________________ 4
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
 Bose-Einstein’s statistics
Basic postulates:
i) The particles are indistinguishable and identical.
ii) They do not obey Pauli’s exclusion principle. So, each quantum state can
accommodate any number of particles.
iii) These particles have zero or integral spin (i.e., 0, h,2h,3h etc).

Bosons: The particles those are indistinguishable, identical having integral spin (i.e.,
0, h,2h,3h etc) and obeys Bose-Einstein’s statistics are called bosons. Examples-photons,
phonons, Π-mesons, K-mesons , η-mesons, α-particles.

BE statistical count: Thermodynamic probability

 If Ni number of identical, indistinguishable particles can be distributed among gi
( N  g i  1)!
quantum states of energy, then possible number of arrangement = i
N i !( g i  1)!

 If Ni number of identical, indistinguishable particles can be distributed among gi

quantum states of energy Ei, n=number of energy levels. Then thermodynamic
( N  g i  1)!
probability= W   in1 i
N i !( g i  1)!
Most probable distribution
N 1
f(Ei)= i   i Ei / KT Ni=number of particles,
gi e e 1
gi=number of states, Ei=energy levels
Bose condensation
It is seen that, as temperature approaches to zero, α→0 and distribution function takes the
1
form; f(Ei)= Ei / KT
e 1
This indicates that at absolute zero, the particles tend to occupy the lowest energy states.
This phenomenon is known as Bose condensation.

Plank’s radiation law from BE statistics

Suppose that a blackbody of volume V at temperature T is filled with energy that can be
considered as a ‘gas’ whose particles are photons. Now photons have unit angular
momentum hence they are bosons. Therefore BE distribution is used to derive Plank’s
law of blackbody radiation.
N 1 N
From B-E distribution law, f(Ei)= i  Ei
, i is called occupation number.
gi gi
e e KT  1
So, the number of photons with energy between E and E+dE is N(E)dE=f(E)g(E)dE
g (E)
Or, N(E)dE= Ei
dE , …..(i) g(E)= no. of quantum states of photons of energy
e  e KT  1
between P and P+dP
_________________________________________________________________________________ 5
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
We know the no. of quantum states correspondence to momentum between P and P+dP is
V
g ( p)dp  3 4p 2 dp … … (ii)
h
Since each photon is endowed with unit spin angular momentum, there are two allowed
momentum states for photon. So, for photons,
V
g ( p)dp  3 8p 2 dp
h
Now, energy of a photon, E=hγ
So, its momentum, p= hγ/c
So, dp= hdγ/c … … (iii)
Put (iii) in (ii), we obtain number of quantum states with frequency between γ and γ+dγ
8V (h / c 2 )h / c 8V
as: g ( )d  3
d  3  2 d … (iv)
h c
In a black body chamber at constant temperature, photons of different energies are
absorbed and reemitted by the walls of the container.
So,  N i  0 , which can be setting by Lagrange multiplier, -α=0.
So, expressing (i) in terms of γ and setting α=0,
g ( ) 8V 2
N ( )d  h / KT d  3 h / KT d
e 1 c e 1
Since the energy of a photon of frequency γ is hγ, the energy per unit volume of the
enclosure or the energy density is
h h 8V 2
u ( )d  N ( )d  . 3 h / KT d
V V c e 1
Or,

8h 3
u ( )d  d
c 3 e h / KT  1

This is Plank’s law of radiation in terms of frequency.

c
[sin ce   c, d   d ,can be omitted because frequence can' t negative]
2
8h c 3 c
( ) ( 2 d )
c 3
  8hc 1
U  d  hc / KT
 5 hc / KT d
e 1  e 1

8hc 1
U  d  hc / KT
d
 5
e 1
This relation is Plank’s law in black body radiation in terms of wavelength.

_________________________________________________________________________________ 6
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
 Fermi Dirac statistics (FD)
Basic postulates:
(i) The particles are indistinguishable and identical.
(ii) They obey Pauli’s exclusion principle. So, each quantum state can
accommodate either no particle or only one particle.
(iii) These particles have 1/2 integral spin (i.e., h / 2,3h / 2,5h / 2 etc).

Fermions: The particles those are indistinguishable, identical having half integral spin
(i.e., h / 2,3h / 2,5h / 2 etc) and obeys F-D statistics are called fermions. Examples-
electrons, protons, neutrons, positrons, μ-mesons.

FD statistical count: Thermodynamic probability

If Ni number of identical, indistinguishable particles can be distributed among gi quantum
states of energy Ei, ni=number of energy levels. then,
gi !
W   in1
N i !( g i  1)!

Most probable distribution

N 1
f(Ei)= i   i Ei / KT Ni=number of particles,
gi e e 1
gi=number of states, Ei=energy levels

Fermi energy or Fermi level

We got the FD distribution function as
1
f(Ei)= Ei

e e KT  1 T=0
1.0
Now, α=-Ef/KT, Ef=Fermi energy,
K=Boltzmann’s
F(Ei)
constant,
T=Temperature in
absolute scale.
0
1 Ei Ef
So, f(Ei)= ( Ei  E F ) / KT
e 1
Now at T=0K, we have
F(Ei)=1 when Ei<Ef
=0 when Ei>Ef
Thus at absolute zero temperature f(Ei) is a step function, where f(Ei) has been plotted
against Ei. The probability of occupation of all states with energy less than Ef is unity and
that of states with energy higher than Ef is zero. Thus at absolute zero, the Fermi energy
or Fermi level represents the highest occupied energy level.
As temperature increases, the occupation index f(Ei) changes from 1 to 0 more
and more gradually. At all non zero temperatures i.e., T>0 when Ei=Ef,
_________________________________________________________________________________ 7
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
 1 1
F(Ei)= 
e 1 2
0

Thus Fermi level is that energy level for T>0

which the probility of occupation at T>0 is 1.0
1/2, i.e., 50% of the quantum states are
occupied and 50% are empty. F(Ei)
0.5

0
Fermi temperature Ei Ef
It is defined as the ratio of Fermi
energy (Ef) at absolute tempersture to Boltzmann’s constant K and its corresponding
f
absolute temperature. So, Fermi temperature,  f 
KT

Determination of Fermi level: Electron gas (electrons in metal)

Electrons are fermions. So, FD statistics can apply to an electron gas to determine Fermi
level of the system.
1
From FD distribution law, f(Ei)=Ni/gi= ( Ei  EF ) / KT
e 1
The number of electrons with energy between E and E+dE is
N(E)dE=f(E)g(E)dE
8 2Vm 3 / 2 E 1 / 2 . f ( Ei )
= dE
h3
Total number of electrons is
  
8 2Vm 3 / 2
N ( E )   N ( E )dE   f ( E ) g ( E )dE  3  E 1 / 2 . f ( Ei )dE
0 0 h 0

At T≠0, the above integral can not be evaluated analytically. One should take recourse to
numerical methods.
At T=0, the highest occupied energy is Ef(0) and f(Ei)=1 for E<Ef(0).
8 2Vm 3 / 2 F 1 / 2
E ( 0)

So, N 
h3 0 E dE
16 2Vm 3 / 2
So, N  3
[ E F (0)]3 / 2
3h
This is the expression of total number of electrons in metal at absolute zero.
h 2 3N 2 / 3
From this expression, E F (0)  ( )
8m V
This is the expression for Fermi energy of electrons in metal at absolute zero.

_________________________________________________________________________________ 8
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
Average electron energy at absolute zero.

Total energy U0 at T=0 of an electron gas,

8 2Vm 3 / 2 F 8 2Vm 3 / 2 2
EF ( 0) E ( 0)
3
U 0   EN ( E )dE  3  E 3/ 2
f ( E ) dE  3
. [ E F (0)]5 / 2  NE F (0)
0 h 0 h 5 5

[ At T=0, the highest occupied energy is Ef(0) and f(Ei)=1 for E<Ef(0)]

So, average electron energy 3

U0  E F (0)
5

Comparison of MB, BE and FD statistics

Features MB BE FD
Particles Distinguishable Indistinguishable Indistinguishable
Eg. Gas particles at Eg. Photons, Eg. Electrons,
ordinary phonons, α-particles protons, neutrons
temperature
Relation Uncertainty and Obey uncertainty Obey Uncertainty
Pauli’s exclusion principle but do not principle and Pauli’s
principle do not obey Pauli’s exclusion principle.
apply exclusion principle
Particle spin Spinless 0,1,2,3, ….. 1/2,3/2,5/2,……..
No. of particles No upper limit No upper limit Only one per
per energy state quantum state is
allowed
Distribution 1 1 1
F(Ei)= 
F(Ei)= 
F(Ei)=  Ei  Ef / KT
function e e Ei / KT
e e Ei / KT
1 e e 1

MB statistics is a limiting case of BE and FD statistics

N 1
MB: f(Ei) = i   Ei / KT … … … (i)`
gi e e
N 1
BE: f(Ei) = i   Ei / KT … … … (ii)
gi e e 1
N 1
FD: f(Ei) = i   Ei / KT … … … (iii)
gi e e 1
The quantum statistical distribution (ii) and (iii) can be combined as
N 1
f(Ei) = i   Ei / KT … … … (iv)
gi e e 1
N
When i <<1, i.e, occupation index is very small,
gi

_________________________________________________________________________________ 9
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
 1

 1
e e Ei / KT
1
Or, e e Ei / KT  1  1
Hence 1 can be ignored, i.e.,
1
f(Ei)   Ei / KT , which is MB distribution.
e e

Ni
Note: When <<1, i.e, occupation index is very small, T is high and density is low the
gi
BE and FD statistics approaches to MB statistics.

Ex: There are 2 particles in 3 quantum states. Distribute the particles according to MB,
BE and FD statistics.

N
N! g i i 2!(3) 2 2 x1x(3) 2
MB: W    9
N! 2! 2 x1
( N  g i  1)! (2  3  1)! 4!
BE: W  i   6
N i !( g i  1)! 2!(3  1)! 2!2!
gi ! 3!
FD: W   3
N i !( g i  N i )! 2!(3  2)!

Ex: Consider 2 particles each of which can exists in energy states ε1, ε2 and ε3. What are
the possible states if the particles are (i) bosons (ii) fermions.
Bosons
Microstates ε1 ε2 ε3
1 aa X X
2 X aa X
3 X X aa
4 a a X
5 X a a
6 a X a
So, there are 6 possible distributions.
( N  g i  1)! (2  3  1)! 4!
Alternatively, W  i   6
N i !( g i  1)! 2!(3  1)! 2!2!
Fermions
Microstates ε1 ε2 ε3
1 a a X
2 X a a
3 a X a
So, there are 3 possible distributions
gi ! 3!
Alternatively, W   3
N i !( g i  N i )! 2!(3  2)!
_________________________________________________________________________________ 10
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
Ex: 3 distinguishable particles of which can be in one of the ε, 2ε, 3ε, 4ε energy states
having total energy 6ε. Find all possible number of distributions of all particles in energy
states. Find the number of macrostates in each state.

ε 2ε 3ε 4ε Total energy
2 0 0 1 6ε
0 3 0 0 6ε
1 1 1 0 6ε

So, number of possible macrostates=3 i.e., (2,0,0,1), (0,3,0,0), (1,1,1,0)

Macrostates ε 2ε 3ε 4ε No. of
microstates
(2,0,0,1) ab 0 0 c 3
ac 0 0 b
bc 0 0 a
(0,3,0,0) 0 abc 0 0 1
(1,1,1,0) A b c 0 6
B c a 0
C a b 0
B a c 0
A c b 0
C b a 0
Total number of microstates=10

Ex: A system has non-degenerate single particle states with 0,1,2,3 energy units. Three
particles ae to be distributed in these states such that total energy of the system is 3 units.
Find the number of microstates if the particle obey (a) MB (b) BE c) FD statistics.

MB:
Macrostates 0ε 1ε 2ε 3ε Total No. of
energy microstates
3ε
(0,3,0,0) 0 abc 0 0 3 1
(2,0,0,1) ab 0 0 c 3
ac 0 0 b
bc 0 0 a
(1,1,1,0) a b c 0 6
b c a 0
c a b 0
b a c 0
a c b 0
c b a 0
Total microstates=10

_________________________________________________________________________________ 11
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur
BE:
Macrostates 0ε 1ε 2ε 3ε Total No. of
energy microstates
3ε
(0,3,0,0) 0 aaa 0 0 3 1
(2,0,0,1) aa 0 0 a 1
(1,1,1,0) a a a 0 1
Total no. of microstates=3

FD:
Macrostates 0ε 1ε 2ε 3ε Total No. of
energy microstates
3ε
(1,1,1,0) a a a 0 1
Total no. of microstates=1

_________________________________________________________________________________ 12
Lecture notes prepared by Dr. I. Sadhukhan, Techno Engineering College, Banipur