Statistical Physics

Lecture - 4

Dr. Amit Jain

Associate Professor
Department of Physics
Institute for Excellence in Higher Education,
Bhopal, M.P. (INDIA)
 General process to study a system
To study a system: (Various Macro and Micro-
states)
• Do experiments
• Take repeated set of observations
• Drive result in terms of statistical probability
• Difficult to keep the same initial condition of the system
every time (or observation)
Suppose we want to find the mean value of
property x of a system
• Take system of N particles,
• Take n1, n2, n3 …. are the no. of values corr. to special
property x1, x2, x3… .
 Mean value of x
n1 x1 + n2 x2 + ... + nr xr + ...
x =
n1 + n2 + ... + nr ...
n1 x1 + n2 x2 + ... + nr xr + ...
x =
N
n1 n2 nr
x = x1 + x2 + ... + xr + ...
N N N
x = P1 x1 + P2 x2 + ... + Pr xr + ...
nr
x =  Pr xr (as Pr = )
N
Where P1, P2,…. are the value of prob. of special
property x1, x2,… of property x.
 Gibbs concept of Ensemble
To make the study of the system
simpler and dependable scientist Gibbs
presented the concept of ensemble or
community.
• Group of N particles is called System.
• According to Gibbs it will be more suitable to
take observations on the group of identical
systems with different microstates which is
equivalent of the given system at the same
time.
 Gibbs concept of Ensemble
Such Group of identical systems is called
ensemble/community.

Every system(member) of the ensemble is

called element.

It’s elements are distributed in all accessible

states instead of in a particular state.
In this situation ensemble’s maximum
probability represents the mean state of the
system.
 Properties of Element
Elements are imaginary yet they are
independent of each other i.e. they do not
interact with each other.
Macroscopic properties of these elements
are always the same i.e. they are exact copy
of each other but their position, velocities
are randomly distributed, i.e. their micro-
states may differ. (Distributed in all accessible
states).
All elements follow the same constraints.
 Terminology
E: Energy

V: Volume

N: Number of Particles

T: Temperature

μ: Chemical potential: chemical potential of a

species is energy that can be absorbed or released
due to a change of the particle number of the given
species.
 Ensembles
Types of ensembles

Grand Canonical:
Micro-canonical: Canonical: The
The systems can
The systems are systems can
exchange heat
completely exchange heat
and particles
closed (a with the universe
with the universe
thermos) (a sealed beaker)
(an open beaker)

Micro Canonical Canonical Grand Canonical

 Micro-canonical Ensemble
Systems are E E E
V V V
Thermally Insulated N N N

E E E
V V V
Separated by N N N
Impermeable wall
E E E
V V V
N N N
E, V, N are constant
 Canonical Ensemble

Systems are separated by: T T T

V V V
Thermally Conducting and N N N
T T T
Impermeable wall V V V
N N N
T T T
V V V
N N N
T, V, N are constant
 Grand Canonical Ensemble

System are separated by:

T, V, μ T, V, μ T, V, μ
Thermally Conducting and
Permeable wall
T, V, μ T, V, μ T, V, μ

T, V, μ T, V, μ T, V, μ
T, V, μ are constant
 Gibbs Canonical Ensemble
Micro-canonical Canonical Grand Canonical

E, V, N are T, V, N are T, V, μ are

constant constant constant

Systems are Separated by

thermally Separated by thermally
insulated thermally conducting
conducting, permeable
impermeable wall wall
Separated by
impermeable
wall
 Equilibrium between two systems in
Thermal contact
Two systems in which there can be an exchange of
heat, are said to be in thermal contact.

Energy of system A and B are

A B
E1 & E2 respectively.
E = E1 + E2 = constant E1 E2
or E2 = E − E1...........(1)
If system A and B are in equilibrium then prob. for
the system A & B to have the energy E1 & E2 resp.
will be maximum.
Let the minimum measurable energy of system A &
B are dE1 and dE2 respectively.
Let : w(E1) and w(E2): No. of microstate in energy
interval dE1 and dE2 corresponding to energy E1
and E2 of the system A and B respectively.
From the postulate of equal a priori prob.
P( E1 ) = C.w( E1 ).......(2) Where C and C’ are
P( E2 ) = C .w( E2 )......(3) constant of proportionality.

The composite prob. For sys. A to have ene. E1 and

the sys. B to have energy E2 is:
P( E1 , E2 ) = P( E1 ).P( E2 ) = C.w( E1 ).C .w( E2 ).......(4)
P( E1 , E2 ) = P( E1 ).P( E2 ) = C.w( E1 ).C .w( E2 ).......(4)
Taking Log, we get.
ln( E1 , E2 ) = ln C + ln C  + ln w( E1 ) + ln w( E2 )......(5)
But, for P(E1, E2) or Log P(E1, E2) to be maximum, the
condition is that:
 
ln P( E1 , E2 ) = 0 or ln P( E1 , E2 ) = 0
E1 E2
Differentiating equation (5) w.r.t. E1, we get:
 
ln w( E1 ) + ln w( E2 ) = 0
E1 E1

Or,     E2
ln w( E1 ) +  ln w( E2 )  = 0 .....(6)
E1  E2  E1
 since, E1 + E2 = E
E2
 E1 + E2 = 0 or = −1
E1
Hence from equation (6), we have
 
ln w( E1 ) − ln w( E2 ) = 0
E1 E2
 
ln w( E1 ) = ln w( E2 ).......(7)
E1 E2

or ,  ( E1 ) =  ( E2 ) where  = ln w( E )
E
Obviously, dimension of β is same as that of (energy)-1
 
 ( E1 ) =  ( E2 ) where  = ln w( E )
E
But T.D. two system are said to be in thermal
equilibrium if their temperatures are same,
therefore, the function β and temp. T are related as:
1 1
 = or 
kT T
1 1
=
kT1 kT2
Or, T1 = T2
This is the necessary condition for thermal
equilibrium according to the thermodynamics.
 Determination of β parameter
Thermodynamically
From 2nd law of T.D. dQ = TdS

Replacing dQ with dE dE = TdS

Or, E = T .S ........(1)

We have defined β as: = ln w( E )
E
 ln w( E )  ln w( E )
Using eq. (1) = =
E T .S
S 1
= .......(2)
 ln w( E )  T
Now, from Boltzmann Entropy-Prob. relation:
S = k log e w
Hence, S = k(log e w)
S
Or, = k .......(3)
(log e w)
Comparing eq. (2) and (3), we get,
1 1
k= Or, =
T kT
Where k is Boltzmann constant
 Boltzmann Canonical Distribution Law
Statement: Prob. of a system to be in a specific
micro-state or energy Er in thermal equilibrium
with heat source at absolute temp. T is proportional
to exp(-Er/kT), where k is Boltzmann’s const., i.e.
P r  e − Er / kT Or, P r = Ae − Er / kT .....(1) A – const.
Proof:
Let us assume a system A is in thermal equilibrium
with Source B at temperature T. Also the total energy
Using eq. (1) + source) remains constant.
(E = system
Assume that system A is in microstate Er
Then the energy of Source will be: E s = E − Er
No. of accessible micro-state for the sys. in ene. state Er is:
= w( Er )
No. of accessible micro-state for the Source will be :
w( Es ) = w( E − Er )
Assume that the sys. A is << Source, then no. of total
accessible micro-states of the total ensemble (source
and system) will be taken to the equal to the total no.
of accessible micro-states of the source, i.e., w(E-Er).
From the postulate of equal a priori prob.

Pr  w( E − Er ) Or, Pr = Cw( E − Er )....(1)

Where, C is a constant.
 Boltzmann Canonical Distribution Law
Since Er << Es and Es ≈ E, so,
  
ln w( E − Er ) = ln( E ) −  ln w( E ) Er .....(2)
 E 
We know that the condition for the two systems in
equilibrium is that their β values must be equal.
 1
Since, = ln w( E ) = ......(3)
E kT
So, from eq. (2) and (3)
ln w( E − Er )
ln w( E − Er ) = ln( E ) − Er OR = −  Er
ln( E )
OR w( E − Er ) = w( E )e − Er ......(4)
 w( E − Er ) = w( E )e − Er ......(4)
Here , w(E) is constant because it does not depend on r.
Therefore, w( E − Er ) = C e − Er ......(5)
From eq. (1) and (5) Pr = CC e − Er

OR Pr = Ae − Er = Ae − Er kT
.....(6)
Where A = CC’ is
another constant. This is Boltzmann Canonical Law.
1
P = 1 or A e
Calculation of A: − βEr
= 1 or A =
 − Er
r kT
e

e − Er kT
OR Pr = .....(7)
e − Er kT
 Boltzmann Canonical Distribution Law
− Er kT
e
Pr = .....(7)
e − Er kT

From above equation it is clear that:

At any given temp. the prob. of finding the system

at higher energy state is less than that at lower
energy state and prob. decreases exponentially
with increasing energy.

At any energy state, the prob. of a system is more

at higher temperature than at lower temperature.
 Partition Function
− E r kT
e
Pr = .....(7)
e − E r kT

From above equation the function  e − E kT is called

r

partition function and is represented by Z, i.e.

Z =  e − Er kT

It is mostly used in determining average value of

different properties of canonical ensemble and this
function gives information about different energy
states of a micro-state in an ensemble.
Partition Function for a Canonical Ensemble
• Consider ideal gas is in thermal equilibrium at
absolute temp. T
• No. of particles of gas in the system is n
• Distribution of molecules is such that in energy
range Er and Er+dEr the no. of molecules are nr.
• No. of micro-states corresponding to energy Er is gr
Now, from the law of equal a priori prob.:
nr  g r
and from Boltzmann canonical law:
− Er kT
nr  e − Er kT
OR nr = Ag r e .....(1)
Since the no. of particles of gas in the system is n, therefore,

n =  nr = A g r e − Er kT

n
=  g r e − Er kT
= Z ( Assumed )
A
Where, Z is called partition function of Boltzmann partition function.

This function express the distribution of molecules in

diff. ene. states in thermal equilibrium at absolute
temp. T of the system.

Clearly, this energy expression includes not only

transport energy, but other energies associated with
rotational and virbrational energies are also included.
That’s all for Today