Lectures on Thermodynamics and Statistical Physics 1

Chapter: 9
QUANTUM STATISTICAL PHYSICS

9.1 INTRODUCTION

Three kinds of particles

The assemblies consist of, in general, three kinds of particles :

1. Identical but distinguishable particles without any spin : The molecules of a gas are the particles
of this kind. In quantum terms, the wave functions of the particles overlap to a negligible extent.
The Maxwell-Boltzmann distribution function holds such particles. We shall use Maxwell-
Boltzmann statistics.

2. Identical and indistinguishable particles of zero or integral spin : They cannot be distinguished
one from another because their wave functions overlap. Such particles are called bosons, and do
not obey the Pauli exclusion principle. The Bose- Einstein distribution function holds for
them. The helium atoms at low temperature and the photons are the particles of this kind and we
shall use Bose-Einstein statistics to account for the spectrum of radiation from a black body.

3. Identical and indistinguishable particles with odd half- integral spin (1/2, 3/2, 5/2, …): Their
wave functions overlap on each other. Such particles are called fermions or fermi particles,
obeying the Pauli exclusion principle. The Fermi-Dirac distribution function holds for them.
Electrons, protons etc. are the examples of this kinds. We shall use F-D statistics to study the
behaviour of the free electrons in a metal that are responsible for its ability to conduct electric
current.

FERMI LEVEL AND FERMI ENERGY

The topmost filled energy level at absolute zero temperature is called Fermi level and corresponding
energy at topmost level (Fermi level) is called Fermi energy. i.e., the energy values upto which all the
states are full at 0 K and above which all the energy states are empty is known as Fermi energy.

The topmost filled energy level at absolute zero temperature is called Fermi level and corresponding
energy at topmost level (Fermi level) is called Fermi energy. i.e., the energy values upto which all the
states are full at 0 K and above which all the energy states are empty is known as Fermi energy.

Condition for Bose-Einstein statistics

1. The particles of the system are are identical and indistinguishable.

2. Particles don't obey Pauli exclusion principle.
3. The sum of particles in each quantum state is the total number of particles.
4. The sum of energy of each particle in the quantum state is the total energy.
5. The number of phase space cells is comparable with the number of particles.
h
6. Bose-Einstein statistics is applicable to particles with integral spin angular momentum in units of .
2
Particles which obey Bose-Einstein statistics are called Bosons.
2 Chapterer-9, Vijay Kumar Jha, Pigauna (Jaleshwar -11), Madhesh Province, Mithila
9.2 BOSE-EINSTEIN STATISTICS

Bose-Einstein distribution Law

Consider N particles in system. Let particles are in first, second, …ith cell with average energy
1, 2, …. i for each particle. Let g be the degeneracy factor.

Let ni particles be arranged in a row and distributed among gi quantum states with (gi – 1) partitions in
between. The total number of possible arrangements of particles and partitions is equal to the total
number of permutations of (n i + gi – 1) objects in a row. Therefore, the total possible ways of arranging n i
particles with (gi – 1) partitions = (n i + gi – 1)

As the particles are identical and indistinguishable,

(ni + gi – 1)
the possible number of distinct arrangements =
ni (gi – 1)

The total number of different and distinguishable ways of arranging N particles in all the available energy
states is given by:

(n1 + g1 – 1) (n2 + g2 – 1)

=   .........
n1 (g1 – 1) n2 (g2 – 1)

(ni + gi – 1)
 = (1)
i ni (gi – 1)

Also, log x = x log x – x, for large x

 From (1),

log  =  [log (ni + gi – 1) – log ni  – log (gi – 1)]

i

=  (ni + gi – 1) log (ni + gi – 1) – (ni + gi – 1) – [(ni log ni – ni) – (gi – 1) log (gi – 1) – (gi – 1)]
i

 log  =  (ni + gi – 1) log (ni + gi – 1) – ni log ni – (gi – 1) log (gi – 1) (2)

i

  log  =  [ log (ni + gi – 1) ni – log ni ni] = 0 (for maxima)

i

  [– log (ni + gi – 1) + log ni] ni = 0 (3)

i

Also,  ni = 0 (4)

i
Lectures on Thermodynamics and Statistical Physics 3
And  i ni = 0 (5)
i

Multiplying (4) by  and (5) by  and adding to (3), we get,

 [– log (ni + gi – 1) + log ni +  + i) ni = 0

i

or, – loge (ni + gi – 1) + log ni +  + i) = 0

 – loge (ni + gi) + log ni +  + i) = 0 [1 can be neglected compared with (n i + gi)]

log 
ni 

ni + gi = –  – i
n i + gi
or, = e  + i
ni

gi
or, 1+ = e  + i
ni

gi
or, = (e  + i – 1)
ni

gi
 ni =  + i
(e – 1)

This is Bose-Einstein distribution law.

9.3 FERMI-DIRAC STATISTICS

The condition for Fermi-Derac statistics are

(i) Particles are identical and indistinguishable.

(ii) The particles obey Pauli exclusion principle. The number of cells must be much greater than the
number of particles. i. e. gi >> ni.

Let ni be the number of particles in the ith compartmentfor energy i. Also let g be the degeneracy factor.

Then the first particle can be placed in any one of the available gi states. In other words, the first particle
can be assigned to any of the gi sets of quantum numbers. Thus the first particle can be distributed in gi
different ways and the remaining (gi  1) will remain vacant. The second partical can be arranged in(gi 
1) different ways and process continues. Thus, the total number of different ways of arranging n i particles
gi
among the gi states with energy level i = gi (gi – 1) …….[gi – (ni – 1)] =
(gi – ni)
4 Chapterer-9, Vijay Kumar Jha, Pigauna (Jaleshwar -11), Madhesh Province, Mithila
Further, if the particles are taken to be indistinguishable, it will be possible to detect any difference when
ni particles are reshuffled into different states occupied by them in the energy level i.

gi
The total number of different and distinguishable ways is
ni (gi – ni)

 The total number of different and distinguishable ways of getting the distribution n 1, n2, n3 etc. among
the various energy levels 1, 2, ..... i etc. can be calculated by multiplying the various factors. The
thermodynamic probability () for the microstate (n 1, n2, ……, ni) of the system is given by

g1 g2  gi gi

 =   …… =
n1 (g1 – n1) n2 (g2 – n2) ni (gi – ni) i n i (gi – n i)

Where  denotes multiplication of terms for vaious values of i.

i

Taking natural logarithm on both sides we get

 ln  =  [ ln gi  – log ni  – log (gi – ni)  ]

i

∵ ni and gi are very large numbers, we can use Sterling approximation log n  = n log n – n.

 ln  =  [(gi ln gi – gi) – (ni ln ni – ni) – (gi – ni) ln (gi – ni) – (gi – ni)]
i

ln  =  [(gi – ln gi) – ni ln ni – (gi – ni) ln (gi – ni)]

i

  ln  =  [ln ni – ln (gi – ni)]ni

i

For maximum,  (ln ) = 0

  [ln ni – ln (gi – ni)] ni = 0 (1)

i

But  ni = 0 (2)

i

And  i ni = 0 (3)

i
Multiplying equation (2) by  and equation (3) by  and then adding to equation (1), we get,

 [ ln ni – ln (gi – ni) +  +  I ] ni = 0

i
Lectures on Thermodynamics and Statistical Physics 5
or, ln ni – ln (gi – ni) +  + i = 0

gi – ni
 ln 
 ni  = –  – i
ni
or, gi – = e  + i
gi

gi
or, – 1 = e  + i
ni

gi
or, = 1 + e  + i
ni

gi
 ni =  + i
(e + 1)

This is Fermi Dirac distribution law.

9.4 ELECTRON GAS IN METALS

Electrons in metals are fermions because they obey Pauli’s exclusion principle.

From Fermi Dirac distribution law,

gi
ni =  + i
(e + 1)

where, ni is the number of particles in ith cell and gi is the degeneracy factor for ith cell.

As the energy of electron in the conduction band is continuous, the degeneracy factor gi is replaced by
g(E) dE and ni by n (E) dE,

g(E)dE
i.e., n (E) dE =
(e  e E / KT + 1)

Here, n(E) refers to number of electrons, g(E) refers to number of phase cells.

But g(E) dE in terms of momentum is:

8v 2 8v
g(p) dp = P dP = 3 P.P dP
h3 h

1
But E = mv2 and P = mv
2

P = (2mE)1/2

 P2 = 2mE  2P dP = 2m dE
6 Chapterer-9, Vijay Kumar Jha, Pigauna (Jaleshwar -11), Madhesh Province, Mithila
 P dp = m dE

8v 8 2 v 3/2 1/2

 g(E) dE = 3 (2mE)
1/2
m dE = m E dE
h h3

8 2 v m3/2 E1/2dE
 n(E) dE =
h3 e  e E/KT + 1

EF
But for electron gas  = – ; E is Fermi energy.
KT F

8 2 v m3/2 E1/2dE
 n(E) dE =
h3 e (E – EF) / KT + 1

This is the form of Fermi-Dirac law of energy distribution of electrons. At the temperature T = 0 K,
number of electrons = total number of energy states occupied by the electrons from zero to E F.

EF E
8 2 v 3/2 F 1/2 8 2 v m3/2 2 3/2 EF m3/2
 n =  g(E) dE = m  E dE = E = 16 2 v 3 EF3/2
h 3
h 3
3  0 3h
0 0

3nh3
or, EF3/2 =
16 2vm3/2

2/3
h2  3n 
or, EF =
2m 8v

9.3 PLANCK’S RADIATION LAW

Statement:
According to the law of radiation of wavelength lying between  and  + d emitted per unit volume by a
perfectly black body at temperature is given by
E d = 5  hc/kT
8hc d 
(1)
 e  1
Derivation

Suppose that photons do not interact among themselves but they interact only with the atoms of
the walls of the cavity. Let photons be indistinguishable and many photons can have the same energy.
Photons are considered as bosons and they obey Bose-Einstein statistics.
The most probable distribution of the particles among various energy levels for a system obeying Bose-
Einstein’s distribution law is
gi gi 1
ni = ( + Ei) =  ; where  =
e 1 e . e Ei / kT  1 kT
It is to be noted that photon gas differ from the ideal gas in one respect. Although the total energy of
photons inside the hollow enclosure at a particular temperature T remains constant, the total number of
photons may be completely absorbed on striking the wall or the hot wall may emit a new photon of energy
Lectures on Thermodynamics and Statistical Physics 7

h, i.e., the photons may be created or destroyed. In other words, for a photon gas system,  ni= 0 is no
i
longer valid. Hence, n =  ni  0.
i
For this reason, the value of  is equal to zero in the Bose- Einstein distribution law. Thus
gi
ni = Ei / kT (2) [∵  = 0]
e 1

In this case, the cavity is large as compared to the wavelength of radiation and the energy spectrum of
photon is taken to be continuous. The energy difference between successive allowed energy values is very
small. Thus replacing gi by g(E) dE and n i by n(E) dE, we get
g(E) dE
n(E) dE = Ei / kT (3)
e 1
Since E = h, the value of g(E) dE can be written as g() d. The quantity g() d corresponds to the
number of oscillatory modes in the frequency range  and ( + d) in the energy range E and (E + dE) .
The number of states in a black body radiation in the frequency range  and ( + d) can be
h( + d) h
obtained by calculating the spherical volume bounded by two spheres of ratio and .
c c
4 h3 4 h3
The volume of the spherical shell =  3 ( + d)3   3 3
3 c 3 c
4 h3 3 4 h3
=  3 [ + 32 d + ….  3 ] =  3  32 d
3 c 3 c
h3 2
= 4 3  d (4)
c
3
The phase space has volume h and there are two states of polarization (left hand and right hand) for the
radiation. The number of states in the black body radiation in the frequency range  and ( + d) is given
by
V
g() d = 2  4 3 2 d
c
V
or g() d = 8 3 2 d (5)
c
Substituting in equation (3), we get the number of photons in the frequency range  and ( + d) given by
8V2 d 1
n() d =  h / kT
c3 e 1
Since each photon has energy h, the energy density E() d defined as the amount of energy per unit
volume lying in the frequency range  and ( + d) is given by
h n() d
E() d =
V
8h3 d 1
 E() d =  h / kT (6)
c3 e 1
c c
Note: To express the law in terms of wavelength (), substituting  = and d =  2 In equation (6), we
 
get (in magnitude):
8h c3 c d 1 8hc d
E d = 3 3  = 5  hc /  kT
c  2 e hc /  kT  1  e 1
E d = 5  hc /  kT 
8hc d
 (7)
 e  1
8 Chapterer-9, Vijay Kumar Jha, Pigauna (Jaleshwar -11), Madhesh Province, Mithila

Extra
Comparison

Difference between Classical and Quantum Statistics

::O::