Professional Documents
Culture Documents
Lec SM
Lec SM
net/publication/360729702
CITATIONS READS
0 1,177
10 authors, including:
All content following this page was uploaded by Sabyasachi Ghosh on 20 May 2022.
Preface
This lecture note on Statistical Mechanics is made by BTech and MSc stu-
dents of PH506 course, taught by instructor Dr. Sabyasachi Ghosh in Winter
semester of 2022 with the help of teaching assistant Mr. Cho Win Aung.
Contents
1 Revisit Thermodynamical Relations for ideal gas 2
3 Canonical Ensemble 13
3.1 Micro to Macro connecting relation for CE . . . . . . . . . . . . 17
3.2 CE Partition function for non-interacting or ideal gas . . . . . . 19
3.3 CE Partition function for Classical Harmonic Oscillator (CHO) . 22
3.4 CE Partition function for quantum Harmonic Oscillator (QHO) . 24
4 Grand-Canonical Ensemble 26
4.1 Micro to Macro connecting relations for GCE . . . . . . . . . . . 27
1
1 Revisit Thermodynamical Relations for ideal
gas
Here, we will quickly revisit our old Thermodynamical understanding. Let us
start with Euler thermodynamical relation:
T S = U + P V − µN
=⇒ U = T S − P V + µN
(1)
U = Internal energy
T = Temperature
S = Entropy
P = Pressure
V = Volume
µ = Chemical potential
N = Number of particles
T dS = dU + P dV − µdN (3)
2
From eqn.(1), we get :
U P µ
S= +V −N
T T T
1 P µ 1 P µ
=⇒ dS = U d( ) + V d( ) − N d( ) + ( )dU + ( )dV − ( )dN (Product Rule)
T T T T T T
1 P µ 1 P µ
=⇒ dS = U d( ) + V d( ) − N d( ) + dS (Using dS = ( )dU + ( )dV − ( )dN from (3))
T T T T T T
1 P µ
=⇒ U d( ) + V d( ) − N d( ) = 0
T T T
(4)
3N KB
We know that U = 23 N KB T (Internal energy of ideal gas) i.e. 1
T = 2U and
P N KB
T = V (from Ideal Gas equation).
3
Again, from eqn.(1), we obtain :
U P µ
S= +V −N
T T T
3
N K B T µ
=⇒ S = 2 + N KB − N ( )(Using Ideal Gas Equation)
T " T #
5
5 N2
=⇒ S = N KB − N KB ln 3 (Using eq.(5))
2 U 2V C
" 3
! #
U 2V C 5
=⇒ S = N KB ln 5 +
N2 2
H =T +V
3N N
X p2i X
H(qi , pi ) = + ϕ(xi ) (6)
i=1
2m i=1
Thus the expression for Hamiltonian inside the volume of the container is given
by,
3N
X p2i
H= (8)
i=1
2m
4
But for our system of particles we know,
H(qi , pi ) = E (9)
3N
X p2i
=E
i=1
2m
3N
X
p2i = 2mE = R (10)
i=1
Eqn.(9) is the constant energy equation which imposes constraint on the sys-
tem limiting the available the momentum space to the surface of sphere. This
geometric symmetry is evident from the close resemblance of the eqn.(10) with
the equation of sphere.
On comparing the eqn.(12) and eqn.(11) we can deduce the value of the radius
of the analogous sphere in the 3-N dimensional space.
√
R = 2mE (13)
Using symmetry and applying constraints to our system we have reduced the
problem from abstract integral over 3-N dimensional space to integration over
a surface of 3-N dimensional hyper-sphere.
For a micro-canonical ensemble, the equilibrium density distribution represented
by ρeq is assumed to be constant in between two constant energy surfaces sep-
arated by a very small distance in the phase space.The equilibrium density
distribution is analogous to to the probability distribution function from the
quantum mechanics. In order to understand how ρeq evolves with time we need
to first revisit a few properties of the phase space. A given space is represented
by a trajectory in the phase space. Some of the important properties of the
trajectories in the phase space are as follows,
• Flow lines are deterministic, i.e. there is no intersections between the
trajectories which may give rise to the uncertainty in the sate of the system
in phase space.
• Liouvilles’ theorem states that there can be no points of convergence in
the phase space,i.e.
dρ
=0 (14)
dt
If a small volume with constant density is allowed to evolve with time every
point in the system will have moved a certain distance. However from liouville’s
theorem we van say that the density of the system will stay constant, but the
5
Figure 1: Intersection of flow lines is strictly not permitted and it gives rise to
uncertainty in the state of the system
Figure 2: Such points where phase space trajectories converge are not
permitted according to Liouvilles’ theorem implying absence of source or sink
new topology of the system will be quite complicated due to the constraint on
the system. Thus for our micro-canonical ensemble density will be constant
between two constant energy surfaces. The equilibrium density distribution can
be used to obtain the following properties of the system,
6
Figure 3: Figure depicting the area of interest between two constant energy
surfaces bounded by E and E + δE
7
Substituting the value of eqn.(15) in eqn.(16)
Z
ρeq (µ)d6N µ = 1
Z
C ′ d6N µ = 1
E≤H(µ)≤E+δ
Z
C′ d6N µ = 1
E≤H(µ)≤E+δE
h i
Ω̃(E + δE) − Ω̃(E) × C ′ = 1
1
C′ = (17)
Ω̃(E + δE) − Ω̃(E)
Eq.(18) represents the generalised expression for finding the ensemble average.
The macroscopic quantity to be measured is represented by the A(µ)and mul-
tiplied by the equilibrium density distribution to obtain the ensemble average.
The procedure followed is analogous to the method of finding expectation values
in the quantum mechanics using the probability distribution function.
Z
⟨A⟩eq = A(µ)ρeq (µ)d6N µ (18)
8
being 6N which represents the 6N dimensional phase space through which our
system of N-particles is defined.
Integration of position coordinates gives us the volume of the closed system. For
a single particle integral over a volume is represented by V . As all the particles
are identical, using symmetry the value of the integral over position will give
same results for all the other particles. Thus the total value of the integral is
V N. Z
dx1 dx2 dx3 ...dx3N = V N (21)
As energy is not dependent on the value of the position of the particle, hence the
integral of the position dependent terms can be treated separately. The value
of this separated integral was already obtained above in the eq.(21). Thus to
obtain the value of the integral we need to obtain the value of the momentum
integral. Z
Ω̃(E) = V N P dp1 dp2 dp3 ....dp3N (22)
3N
i=1 p2i ≤2mE
From eqn.(24) and eqn.(23) we get the following simplified expression for the
value of integral in the momentum space.
3N
π 2 R3N √
Vsp (R, 3N ) = where R= 2mE (25)
Γ( 3N
2 + 1)
In the phase space the momentum and position cannot take up any small value
indiscriminately. The uncertainty principle puts a limiting condition on the
minimum values of the momentum and position of the molecules in the phase
9
space. This is obtained by dividing our final expression for the no. of micro-
states by h3N as shown below.
1 ∂ Ω̃(E)
Ω= × × δE (27)
h3N ∂E
Thus eqn.(27) represents the final form for obtaining the number of micro-states
in the micros-canonical ensemble. The value of the entropy can be obtained
using the following procedure. From eqn.(27) and eqn.(26) we get,
3N 3N
−1
dΩ̃ (2πm) 2 E 2
=VN (28)
dE ( 3N
2 − 1)!
Applying Stirling’s approximation on the above equation, and keeping only the
lower order terms of N,
3 3 3N 3N
log(Ω) ≈ −N log(h ) + N log(2πmE) − 2 log
2 2
3N
+N log(V ) + (31)
2
10
2.1 Gibb’s paradox
There are two identical boxes of volume V1 and V2 and the number of molecules
in it are N1 and N2 respectively.
After mixing two system at same temperature(T),
V1 + V2 = Vf and N1 + N2 = Nf
Entropy Change:-
3 3
Sf −(S1 +S2 ) = KNf (ln[V1 ( 23 KT ) 2 ]+ 32 )+ 32 Nf Kln( 4πm 3 2
3h2 )−KN1 (ln[V1 ( 2 KT ) ]+
3
3
2 + 32 ln( 4πm 3
3h2 )) − KN2 (ln[V2 ( 2 KT ) ] +
2 3
2 + 32 ln( 4πm
3h2 ))
We know that, Nf − N1 − N3 = 0
Therefore,
V V
∆S = KN1 ln[ Vf1 ] + KN2 ln[ Vf2 ] ——– eqn.[A]
∆S = 0 (expected)
V V1 V2
∆S = KNf ln[ Nff ]−KN1 ln[ N 1
]−KN2 ln[ N 2
]+K( 32 ln[ 32 KT ]+ 25 + 32 ln[ 4πm
3h2 ])(Nf −
N1 − N2 )
We know that, Nf − N1 − N3 = 0
V N1 V N2
Therefore, ∆S = KN1 ln[ Vf1 Nf ] + KN2 ln[ Vf2 Nf ]
11
Now for V1 = V2 = V and Vf = 2V, N1 = N2 = N and Nf = 2N
∆S = KN ln[ 2V
V
N
2N ] + KN ln[ 2V
V
N
2N ] =0 ∵ ln1 = 0
V 3 3
So, to get Scarnot = N K[ln[ N ( 2 KT ) 2 ] + 52 ] + C
Ω 3
N! = ΩGibbs ←− Ω = (V ( 2πmυ 2 N
3 ∆ 1
h2 ) ) 2 N U Γ 3N ! 2
3
SB ↑= N K[ln(V ( 32 KT ) 2 ) + 3
2 + 3 4πm
2 ln[ 3h2 ]]
V 3 3 5
SB − KlnN ! = N K[ln( N ( 2 KT ) 2 ) + 2 + 32 ln[ 4πm
3h2 ]]
...................
12
3 Canonical Ensemble
In the above sections, we have discussed the basic thermodynamic relations
and derived an expression of entropy as a function of number of particles (N ),
volume (V ) and total internal energy of the system (U ) for a micro-canonical
ensemble. In the micro-canonical ensemble, the thermodynamic properties U ,
V , N remain constant throughout the process. In the canonical ensemble how-
ever, the internal energy of the system ceases to be a constant parameter, as
the system is allowed to exchange energy with the surroundings while keeping
the number of particles and volume of the system constant.
In the canonical ensemble, the system is in thermal equilibrium with the sur-
roundings (heat reservoir). Both the system and heat reservoir attain the same
temperature at the equilibrium condition. Hence, to derive all the thermody-
namic properties of this type of system configuration, we need to define a state
function at the macroscopic level which is a function of T , V and N . But before
that, it would be a good idea to define a function at the microscopic level, and
then derive a relationship between the microscopic function and its macroscopic
counterpart.
Let us consider a system immersed into a very large heat reservoir. At equi-
librium, the system and the reservoir attain the same temperature T , but their
energies would be variable at any instant. The energy values of system and
reservoir will lie between 0 and E0 , where E0 is the combined total energy of
the system and reservoir. At any instant, the system and reservoir will have the
energy values E and ER such that
E + ER = E0
where E0 is constant at any instant. Now, since the reservoir is much larger
than the system, ER is much greater than E. It also implies that
E ≪ E0
i.e., the energy of the system E is a very small fraction of the combined total
energy E0 .
Now let us focus on the energy configuration of the reservoir. The probability
of the reservoir assuming a particular energy value ER is directly proportional
to the number of micro states having the energy value ER .
p ∝ ΩR (ER )
p ∝ ΩR (E0 − E)
13
Now, if we take the ratio of the probabilities of the reservoir assuming the energy
states Ej and Ek respectively, then we can write it as follows
p(Ej ) ΩR (E0 − Ej )
=
p(Ek ) ΩR (E0 − Ek )
Using the famous Boltzmann expression for entropy, we can write the number
of micro states of a reservoir having the energy value E as a function of entropy
of the reservoir at that particular energy value.
SR = kB ln ΩR (E0 − E)
SR (E0 − E)
⇒ ΩR (E0 − E) = exp
kB
Hence we can write the probability ratios at Ej and Ek as
p(Ej ) (SR (E0 − Ej ) − SR (E0 − Ek ))
= exp
p(Ek ) kB
We have already proved above that E ≪ E0 for all practical purposes. In this
case, we can expand the entropy term as
∂SR
SR (E0 − E) = SR (E0 ) − E
∂E V,N
1
where β =
kB T
We have discussed the relationship between probability at a particular energy
14
value p(E) and energy of the system E. Now, it would be great if could ob-
tain a complete expression of p(E) by normalizing it. We can write p(E) at a
particular energy value E as
e−βE
p(E) =
Z
where Z is the normalizing constant.
To compute the value of Z, we add up all the probabilities at all possible energy
values. This can be represented mathematically as given below
X
p(Ej ) = 1
j
e−βEj
P
j
⇒ =1
Z
X
⇒ Z= e−βEj
j
In the above equation, Z signifies the sum of all the possible energy states in
the system at equilibrium, and is also known as the ”Canonical Partition Func-
tion”. To avoid confusion with the Grand Canonical Partition Function ZGCE ,
we shall represent canonical partition function by ZCE hereafter.
Ej e−βEj
P
j
⇒ ⟨E⟩ = P
j e−βEj
Now we already know that
X
ZCE = e−βEj
j
15
or
∂ (ln ZCE )
⟨E⟩ = −
∂β V,N
The calculations done above were based on the assumption that each state has
its own distinct energy values, which is not true in reality. We always have a
certain degeneracy in the number of states Ω(E) corresponding to an energy
value. So, to account for all these degenerate states, we need to modify the
expression of canonical partition function ZCE by multiplying the number of
states Ω(E) with the exponential term in the former expression of ZCE .
A = U − TS
16
Now, by substituting the above relation in the expression for ZCE and taking
log on both sides, we get
A = U − TS (36)
A = −P V + µN (37)
17
From the second law of thermodynamics, we know that :
T dS = dU + P dV − µdN (38)
dA = dU − T dS − SdT
∂A ∂(T lnZ)
⇒S=− = KB (39)
∂T V,N ∂T V,N
∂A ∂(lnZ)
⇒P =− = KB T (40)
∂V T,N ∂V N,T
∂A ∂(lnZ)
⇒µ= = −KB T (41)
∂N V,T ∂N V,T
Now,
U = A + TS
∂(T lnZ)
U = −KB T lnZ + T KB
∂T V,N
2 ∂(lnZ)
U = T KB (42)
∂T V,N
18
3.2 CE Partition function for non-interacting or ideal gas
Now we are going to calculate the partition function for the Canonical En-
semble and derive the physical macro quantities like Helmholtz free energy,
Entropy,Pressure and internal energy form that partition function of canonical
ensemble.
X
Z= Ω(Ev )e−βEv (44)
v
X d3N xd3N p
Z= e−βEv (45)
N !h3N v
d3N xd3N p p2
Z
Z= e−βEv whereE = (46)
N !h3N v 2m
N
∞
−βp2
VN
Z
−
Z= 4πp2 dpe 2m (47)
N !h3N
0
βp2 2m
But = z =⇒ pdp = dz;
2m β
N
1 1
N Z ∞
V 2m 2 m
Z= 4π z 2 dz e−z (48)
N !h3N β β
0
N
3 Z ∞ 1
N
√ m 2
V
Z= 4π 2 z 2 dze−z (49)
N !h3N β 0
Z ∞ √ √
π
since zdze−z = 2
0
19
N
3
N
V 2πm 2
Z= (50)
N !h3N β
3 N
1 V 1
Z= [2πmkT ] 2 as = kT (51)
N ! h3 β
3 N
1 V
Z= [2πmkT ] 2 (52)
N ! h3
V
A(N, V, T ) = −N KT ln 3 + ln N − 1 (54)
λ
ln(N !) = N ln N − N
h
and Here λ = √ which is ”Thermal de-Broglie wavelength”
2mπkT
N λ3
A(N, V, T ) = N KT ln −1 (55)
V
N λ3
A = N KT ln −1 (56)
V
20
Entropy:
N λ3
3
S = −N k ln[ − 1 + N KT (58)
V 2T
V 5
S = N K ln + (59)
N λ3 2
V 5
S = N K ln + (60)
N λ3 2
Pressure :
∂A
P =− (61)
∂V N,T
1
P = N KT (62)
V
N KT
P = (63)
V
P V = N KT (64)
Internal energy :
21
U = A + TS (65)
N λ3
V 5
U = N KT ln − 1 + N KT ln + (66)
v N λ3 2
3N KT
U= (67)
2
3
U= N KT (68)
2
Thus, we know that Then Z = Z1N where Z1 is the partition function of one of
them and is given by the following equation:
Z
dxdP
Partition function, Z1 = e−βH (70)
h
22
" Z #" Z #
∞ −1 −1 ∞ m 12 −1
1 −Z
2 2 −Z
=⇒ Z1 = 2 e 2βmω z 2 dz 2 e z 2 dz
h 0 0 2β
This is two Gaussian integrals:
" #" #
1 1 12 √ m 21 √
Z1 = 2 π 2 π
h 2βmω 2 2β
" #" #
1 2πKT 12 12
=⇒ Z1 = 2πKT m
h mω 2
Thus, we get:
2πKT K T
B
Z1 = =
hω ℏω
Note that in the last step, we re-wrote the answer in terms of the natural fre-
quency of the oscillator.(Note that this is a ratio of energies.)
Now consider many oscillators (like positions of atoms in a crystal), and let’s
extract the thermodynamics:
K T N
B
Z = Z1N =
ℏω
also,
∴ A = −KT lnZN (71)
ℏω
=⇒ A = N KT ln (72)
KT
Also P, S, V and U can be calculated from the above equation as follows:
∂A
P =− = 0, Since A(V ) = constant (73)
∂V
Above equation holds true since, A(V ) = constant.
∂A
S=− (74)
∂T N,V
ℏω 1 KT
S = −N K ln( ) + N KT = N K ln +1 (75)
KT T ℏω
U = A + T S = N KT (76)
23
hold:
CP is heat capacity at constant pressure
∂ ∂ 3 5
CP = (U + P V ) ⇒ CP = N KT + N KT ⇒ CP = N K
∂T P ∂T 2 2
Similarly,
CV is heat capacity at constant volume
∂U ∂ 3 3
CV = ⇒ CV = N KT ⇒ CV = N K
∂T V ∂T 2 2
This implies:
CP − CV = N K = R (77)
∂U
CP = CV = = NK = R (78)
∂T
where n = 0, 1, 2, 3, ..........
Partition function of single particle :
∞
e−βkω/2 1
e−β (
n+ 21 )ℏω
X
z1 = = = (80)
1 − e−βℏω
n=0 2 sinh βℏω
2
24
Total partition function:
N
1 N
e−( 2 βℏω) 1 − e−βℏω
N N
zn = (z1 ) = (81)
βℏω
2 sinh 2
ℏω −βℏω
A = −KT ln 2 = N KT β + ln 1 − e (82)
2
N KT e−βℏω (ℏω/kT 2 )
∂A
= −N K ln 1 − e−βℏω +
S=−
∂T V,N 1 − e−βℏω
" #
ℏω
KT −βℏω
S = NK − ln 1 − e (84)
eβℏω − 1
ℏω ℏω
U = TS + A = N + βℏω (85)
2 e −1
25
4 Grand-Canonical Ensemble
According to grand canonical ensemble,
E + ER = E0 = Constant
N + NR = N0 = Constant
26
∵ T
dS = dU + P dV − µdN
∂S 1 ∂S µ
⇒ = and =−
∂U V,N T ∂N V,U T
Ej µNj
∴ SR (E0 − Ej , N0 − Nj ) = SR (E0 , N0 ) − +
T T
ρ(Ej , Nj ) SR Ej µNj SR Ek µNk
∴ = exp − + − − +
ρ(EK , NK ) KB KB T KB T KB KB T KB T
(E − µN )
−
⇒ ρ(E, N ) ∝ e kT
∝ e−β(E−µN )
e−β(Ej −µNj ) X
⇒ ρj = P −β(E −µN ) ⇒ Z = e−β(Ej −µNj )
je
j j
j
Partition Function
Now,P
Z = j e−β(Ej −µNj )
Z = γ ω(Eγ )e−β(Eγ −µNγ ) (This is Density of states concept)
P
P S(Eγ ) Eγ µNγ
Z = γ exp − +
K KT KT
S γ µN
Z ≃ exp − +
K KT KT
=⇒ T S − U + µN = KT lnZ = P V = Φ which is Grand Canonical poten-
tial.
27
Figure 5: Grand Canonical Ensemble
Φ = −P V (87)
Φ = U − T S − µN (88)
T dS = dU + P dV − µdN (89)
From above,
dΦ = dU − T dS − SdT − µdN − N dµ
dΦ = −P dV − SdT − N dµ
∂Φ ∂(T lnZ)
⇒S=− = KB (90)
∂T µ,V ∂T µ,V
28
∂Φ ∂(lnZ)
⇒P =− = KB T (91)
∂V µ,T ∂V µ,T
∂Φ ∂(lnZ)
⇒N =− = KB T (92)
∂µ V,T ∂µ V,T
Now,
U = Φ + T S + µN
∂(T lnZ) ∂(lnZ)
U = −KB T lnZ + T KB + µKB T
∂T µ,V ∂µ V,T
∂(lnZ) ∂(lnZ)
U = T 2 KB + µT KB
∂T µ,V ∂µ V,T
Here Finally,
∂(lnZ) ∂(lnZ)
U = KB T T +µ (93)
∂T ∂µ
29
volume, and number of particles.
X
CE
ZN = e−(βEi ) (94)
i=0
X d3N xd3N p
= 3N
e(−βEν ) (Eν = N ϵ)
ν
N !h
VN
Z
= [ d3 pe−βϵ ]N
N !h3N
CE
ZN
= (95)
N!
For Grand Canonical Ensemble :
The grand canonical partition function applies to a grand canonical ensemble,
in which the system can exchange both heat and particles with the environment,
at fixed temperature, volume, and chemical potential.
X
Z GCE = e−β(E−µ) (96)
X d3N xd3N p
= ( 3N
)Eν ,Nα e−β(Eν −µα )
ν,α
N α !h
X VN Z
= 3N
[ d3 pe−β(E−µ) ]N (∵ µα = N µ)
α
N !h
X (eβµ ZCE )Nα
=
α
Nα !
p2
Z
V V
where Z1CE = d3 pe−βϵ = for ϵ = (97)
h3 λ3 2m
is one particle partition function of CE system with
h
λ= √ (thermal de-Broglie wavelength) (98)
2πmKT
So if we take all Nα from 0 to α
∞
X (eβµ ZCE )Nα
Z GCE (µ, T, V ) = (99)
Nα !
Nα =0
= exp{eβµ ZCE }
⇒ −P V = Φ = −KT lnZ GCE = −KT {eβµ ZCE }
V d3 p −β(ϵ−µ)
Z
= −KT e
h3
V d3 p −β(ϵ−µ)
Z
where N = e (100)
h3
GCE (Classical Ideal Gas) :
30
1. For one-particle: Partition function is written as:
X
Z1 = e−β(ϵ−µ) (101)
Z 3 3
d xd p −β(ϵ−µ)
= e
h3
d3 xd3 p
X Z
Remember, =
h3
ZN = Z1 × Z2 × Z3 ... (102)
Y
= Zi
i
Y (ϵi −µi )
= e−β
i
Z = exp{e−β(ϵ−µ) }
α
X {e−β(ϵ−µ) }Nα
=
Nα !
Nα
This imples,
d3 p −β(ϵ−µ)
Z
KT N
P = KT 3
e = (104)
h V
V d3 p −β(ϵ−µ)
Z
Φ = −P V = −KT e
h3
Z
KT V
⇒P = N, where N = 3 d3 pe−β(ϵ−µ)
V h
Z
V
= d3 pfo (ϵ)
h3
31
where fo (ϵ) is Maxwell distribution function
Now,
dΦ
N = −
dµ V,T
V d3 p −β(ϵ−µ)
Z
= −(−KT )β e
h3
p2
Z
V 1
N= 3 d3 pfo (ϵ) = V for ϵ = (105)
h λ3 (T ) 2m
(Thermal distribution of function of particle with energy ϵ)
Since,
dΦ
S = − (106)
dT µ,V
V d3 p −β(ϵ−µ) (µ − ϵ) V d3 p −β(ϵ−µ)
Z Z
= K e +KT e
h3 −KT 2 h3
| {z } | {z }
N N
V d3 p
Z
T S = KT N − µN + ϵfo (ϵ)
h3
| {z }
U
=⇒ T S = P V − µN + U (107)
Internal Energy:
V d3 p
Z
U = [ϵ]fo (108)
h3
3
= N KT
2
p2
3 V
= KT f or ϵ= (109)
2 λ3 (T ) 2m
U + P V − µN
S =
T
5 µN
= NK − (110)
2 T
32
6 Quantum mechanical version of partition func-
tion
According to classical physics, the partition function for a single particle is
defined mathematically as :
Z = e−β(ϵ−µ) (111)
In quantum mechanics, an energy level is degenerate if it corresponds to two
or more different measurable states of a quantum system. Conversely, two or
more different states of a quantum mechanical system are said to be degenerate
if they give the same value of energy upon measurement.
If quantum degeneracy is n for energy state ϵ, then its partition function is:
X
Z= e−βn(ϵ−µ) (112)
n
Subatomic particles fall into two classes, based on their statistical behaviour.
Those particles which follow the Pauli exclusion principle are called fermions
while those that do not obey this principle are called bosons.
According to Pauli Exclusion Principle, n = 1 for fermion i.e only one particle
is allowed in a particular energy level. But, for Boson, the quantum degeneracy
of energy levels,n can take any value from 1, 2, 3, ...∞. The classical definition
also misses the possibility of n = 0. But this is a valid quantum degeneracy
value for both boson and fermion. Hence, the partition function for each energy
level for fermions and bosons is given by :
(i) Fermion -
1
X
Z= e−βn(ϵ−µ) = [1 + e−β(ϵ−µ) ] (113)
n=0
(ii) Boson -
∞
X 1
Z= e−βn(ϵ−µ) = (114)
n=0
1 − e−β(ϵ−µ)
(Using the formula for sum of infinite G.P.)
The total partition function, Zt for all particles is given by the expression below :
Y Y X
Zt = Zi = Z1 × Z2 × Z3 × ... =⇒ ln Zt = ln { Zi } = ln Zi (115)
i i i
33
The above calculation of ln Zt will be useful for the calculation of Grand Canon-
ical Potential i.e Φ since we know that Φ = −KT ln Zt
Bosons are particles which have integer spin and which therefore are not con-
strained by the Pauli exclusion principle like the half-integer spin fermions.
Fermions are particles which have half-integer spin and therefore are constrained
by the Pauli exclusion principle.
For Fermion :
X
ϕf ermion = −KT ln [1 + e−β(ϵ−µ) ] (117)
For Boson :
X
ϕboson = −KT ln [1 − e−β(ϵ−µ) ]−1 (118)
1 X ηe−β(ϵ−µ) V d3 p
Z
∂Φ X 1 1
N =− = −β(ϵ−µ)
= β(ϵ−µ)
= 3 β(ϵ−µ)
∂µ T,V η 1 + ηe e +η h e +η
(121)
34
The Fermi-Dirac distribution applies to fermions, particles with half-integer
spin which must obey the Pauli exclusion principle. It is a type of quantum
statistics that applies to the physics of a system consisting of many identical
particles.
35
exactly zero chance of two particles being in the same state. Thus, the difference
among Fermi-Dirac, Bose-Einstein and Maxwell-Boltzmann statistics arises en-
tirely from situations were a single state has a nonzero change of being multiply
occupied.
V dp
Z Z
KT V 1
PV = ln{1 + ηeβ(µ−ϵ) } = 3 d3 p[?] (132)
η h3 h eβ(ϵ−µ)+η
36
Now we can cross check through the value of PV deduced above if it is same
as that of -KTln z and calculate pressure for fermi-dirac, bose-einstein and
maxwell-boltzmann accordingly:
P ∂
−1 X β(µ−ϵ)
− i ∂η ln[1 + ηeβ(µ−ϵ) ]
ln z = ln[1 + ηe ] = lim ∂η
(133)
η i η→0
∂η
It will give us
X eβ(µ−ϵ) X
ln z = − lim =− eβ(µ−ϵ) (134)
η→0 1 + ηeβ(µ−ϵ)
i i
Hence:
d3 xd3 p −β(ϵ−µ)
X Z
−β(ϵ−µ)
P V = Φ = −KT ln z = KT e = KT e (135)
i
h3
d3 xd3 P −β(ϵ−µ)
R
where N = h3 e
Now, similarly we apply the three distributions individually and find the Pres-
sure for respective distributions.
Pressure for Fermi-Dirac(PF D ):
Z 3
d p
PF D = KT V (ln{1 + e−β(ϵ−µ) }) (137)
h3
d3 p
Z
PBE = −KT V ln{1 − e−β(ϵ−µ) } (138)
h3
37
6.2 Photon Gas : example Boson system
In the previous sections, we have explored the realms of the Grand Canonical
Ensemble, and have ventured into the quantum mechanical aspects of statistical
mechanics. We are also familiar with some of the statistical distributions used
to calculate the macroscopic variables from microscopic variables.
Before proceeding to the next topic, it would be great if we go through the sta-
tistical distributions part once again as a primer to the next topic. In the Grand
Canonical Ensemble, Maxwell-Boltzmann distribution is used for classical sys-
tems (systems under the domain of classical mechanics), while Fermi-Dirac and
Bose-Einstein distributions are used for quantum systems. In addition, Fermi-
Dirac distribution is applicable on Fermions (electrons, protons, muons etc.),
while Bose-Einstein distribution is applicable on Bosons (photons, gluons etc.)
Now that we have gone through the statistical distributions used in the Grand
Canonical Ensemble, let us proceed to the application of the theories that we
have studied above. One of the most prominent applications of the Bose-Einstein
statistical distribution is the study of photon gas and calculation of macroscopic
variables for the same. As it is already known, light constitutes of energy packets
known as photons, which are basically Bosons with spin 1. Hence we can con-
sider light to be a gas of photons, with each photon having energy-momentum
relation as
ϵ = pc
We can use the above energy-momentum relation to calculate the macroscopic
parameters such as number of particles (N ), total internal energy (U ) and pres-
sure (P ). Let us start with the calculation of the number of particles in the
system. Z 3 3
d xd p 1
N =2 3 βϵ
(141)
h e −1
In eqn.(141), the general expression for number of particles is multiplied by 2
to account for two spin states (+1 and −1) of the photons. We already know
that Z
d3 x = V
38
Let βpc = x
kT
=⇒ dp = dx
c
On substituting these values in eqn.(142), we get
8πV (KT )3 ∞ x2
Z
N= dx (143)
h3 c3 0 ex − 1
In the above equation, the integral term is a special integral also known as
Riemann Zeta function, and it is denoted by ζ (n). The Riemann Zeta-function
is defined as Z ∞ n−1
1 x
ζ(n) = dx
Γ (n) 0 ex − 1
On comparing the above equation with the integral term in eqn.(143), we can
clearly see that n=3.
Let us now proceed to the calculation of internal energy of the photon gas.
The microscopic variable in this case is the energy of a single photon. The total
internal energy of the photon gas system is defined as
Z 3 3
d xd p ϵ
U =2
h3 eβϵ − 1
Z
2V ϵ
=⇒ U = 3 4πp2 dp βϵ
h e −1
Dividing both sides by V , we get
p3
Z Z
U 8πc
= 3 dp = uV dν (144)
V h eβpc − 1
where
8πc p3
uν dν = dp
h3 eβpc − 1
hν hdν
Substituting p = c and dp = c in the above equation, we get
8πν 3 h dν
uν dν = (145)
c3 eβhν−1
39
In eqn.(145), the term uν represents the spectral distribution of energy in the
black body radiation. It is now very clear that we can easily calculate the total
energy density by integrating the spectral distribution from 0 to infinity.
8πc kT 4 α x4−1
Z
U
= 3 ( ) x
dx
V h c 0 e −1
8π
= 3 3 (kT )4 ζ(4) ∝ T 4 (147)
h c
8π
= 3 3 (kT )4 ζ(4)
h c
Here
π4
ζ(4) =
90
Hence
U 8(π)5
= (kT )4
V 15h3 c3
(148)
π2
= (kT )4 ∝ T 4
15(ℏc)3
Hence we are able to get a relationship between energy density and temperature
in the above equation. We can clearly see from eqn.(148) that the energy den-
sity is directly proportional to the fourth power of temperature. This relation
goes a long way in deriving the Stefan-Boltzmann law of radiation.
Now that we have already ventured into the total internal energy of the photon
gas, let us try to calculate the value of Stefan constant. For a blackbody, the
intensity is given by
cU
I= (149)
4V
By Stefan-Boltzmann’s law, we already know that the intensity of a blackbody
can be written as
I(T ) = σT 4 (150)
40
On comparing eqns. (149) and (150), we get
π2 k4
σ= (151)
60h3 c2
Now, let us try to calculate the pressure of the photon gas.
Φ
Pressure, P = −
V
2d3 pd3 x
1 −KT
Z
−β ϵ
=− ln 1 − e
V −1 h3
Z
2KT
d3 p ln 1 − e−β ϵ
=− 3
h
8π ∞ 2
Z
P
P dP ln 1 − e−βpc ,
=− 3 ϵ= pc
KT h 0
∞ Z ∞ 3
8π p3 p βce−βpc
−βpc
=− 3 ln 1 − e − dp
h 3 0 0 3 1 − e−βpc
∞
p3
Here,| ln 1 − e−βpc =0
3 0
8π ∞ βp3
Z
c KT
= 3 βpc
dp, βpc = x and dp = dx
h 0 3 e −1 c
4 Z ∞ 4−1
8π βc KT x
= 3 x−1
dx
h 3 c 0 e
Then, we multiplied and divided by β 3 c3 to convert the integral into Γ(4)ζ(4) .Solved
using Riemann zeta function
8π 1 1U 1
P = 3 3
(KT )4 Γ(4)ζ(4) = , PV = U
3 h c 3V 3
Now, lets substitute the value of Γ(4)ζ(4) in the above equation
4
16π π
= 3 3 K 4T 4
h c 90
5 4
8π K
= T4
45h3 c3
41
This is the value of pressure of photon gas wich is directly proportional to T 4
where ρ is the density,M is the mass of the object, R is its radius. Here we
have considered a small sphere inside a big sphere,as its density is constant
we can integrate over the radius from 0 to R to get the average mass density.
Calculating the Gravitational potential energy of the sun. Ω
Z R
GM (r)
Ω=− 4πr2 ρ(r)dr
0 r·
Z R
4
≈− G πr3 ρ(r)4πrρ(r)dr
0 3
5
4π R
= −G 4π⟨ρ⟩2
3 5
2
3 4π 3
=− R ⟨ρ⟩
5R 3
Here G is the Gravitational constant. Using M (r) = 43 πr3 ρ(r) and ρ(r) ≈ ⟨ρ⟩
2
Ω = − 35 GM
R (152)
⟨ρ⟩
P= KBT
m
42
where m = mass of gas Particles
3K B R
Z
U= T (r)⟨ρ⟩4πr2 dr
2m 0
3 KB
U= M ⟨T ⟩
2 m
1
RR
⟨T ⟩ = M 0
4πr2 dr⟨ρ⟩T (r) (153)
The virial theorem relates the total kinetic energy of a self-gravitating body due
to the motions of its constituent parts, U to the gravitational potential energy,
Ω of the body.
Here the Virial theorem 2U + Ω = 0 become −2U = Ω
3 M 3 GM 2
⇒ −2( K B ⟨T ⟩ ) = −
2 m 5 R
1 GM m
⟨T ⟩ =
5 KBR
3M
Where ⟨ρ⟩ 43 πR3 = M ⇒ R3 =
4π⟨ρ⟩
1 Gm 4π⟨ρ⟩ 13
⟨T ⟩ = M
5 kB 3M
mH = 1.6 × 10−27 kg
23
G = 6.67 × 10−11 m3 /kg/s2 , k B = 1.38 × 10− m2 kgs−2 k −1
43
Now, the Values of the SUN:
4 ×106 ◦ K
∆x∆p ≥ h (154)
where h = 6.63 X 10− 27 erg s is Planck’s constant. In 3D, we can write this as
∆V ∆3 p ≥ h3 (155)
To get the pressure in this fully degenerate state, we need to know the momen-
tum distribution d3 p.
Since the shell has volume 4πp2 dp, and each grid point takes up a volume
∆3 p = h3 /∆V , the number of electrons inside the shell is
4πp2 dp 4πp2 dp
Ne = = .∆V (156)
∆3 p h3
To figure out the momentum pF where this distribution stops, we simply set
44
it by the condition that, when we integrate over all momenta, we get the right
number of particles. Z pF
4πp2
n=g dp (158)
0 h3
4π 3
n=g .p (159)
3h3 F
4π 3
n=g .[2mEF ] 2 (160)
3h3
p2F
As EF = 2m for non relativistic gases
32
4π EF2
2 2
n = g 3. −m c (161)
3h c2
p
As EF = p2F c2 − m2 c4 for relativistic gases.
13
3h3 n
pF = (162)
4πg
ϵ 3
< E >= = EF (167)
n 5
45
Z pF
g p.p
P = 4πp2 dp (169)
3h3 0 m
here p=mv momentum of gas
g 4π p5F
P = (170)
3h3 m 5
1
15P h3 m 5
pF = (171)
g.4π
5
g4π 3h3 n 3
P = (172)
15h3 4πg
3 53
g4π 3h n
P = (173)
15h3 m 4πg
2
1 3h3 3 5
P = .n 3 (174)
5m 4πg
Hydro dynamical Equilibrium :
1 dpe GM
. = 2 (175)
ρe dr R
4 3
AS M= πR ρe (176)
3
M
ρe = 4 3
(177)
3 πR
M
ρe ∝ (178)
R3
Z R Z R
dpe GM M
.dr = . .dr (179)
0 dr 0 R2 43 πR3
GM 2 R−5+1
P = 4 (180)
3π
−5 + 1
M2
Pe ∝ (181)
R4
5
But degenerate pressure Pe ∝ ρe3
5
M3
Pe ∝ (182)
R5
−1
M 3 Pe
=> ∝ (183)
R Pe
46
1
1
=> M 3 ∝ (184)
R
M
=> ρe ∝ 3 (185)
R
=> ρe ∝ M 2 (186)
gp3F
n(kF ) =
6π 2 ℏ3
gcp4F
u(kF ) =
8π 2 ℏ3
µ = pF c
1
P = u
3
dP 1 1
= pF c = u
dn 3 3
4
γ=
3
Note that the relativistic gas is more compressible than the non-relativistic one
dn dn
>
dP REL dP N R
47
[Note that for p2F << m2 this reduces to the non-relativistic result including
the rest mass energy.]
4πgm4 xF 2 p 2 gm4
Z
ū = 3
x dx x + 1 ≡ f (xF )
(2π) 0 2π 2
kF
where xF ≡ m . f (xF ) can be evaluated by elementary means,
1 3 1 1 1 p
f (x) = x(1 + x2 ) 2 − x(1 + x2 ) 2 − ln(x + 1 + x2 )
4 8 8
but we can evaluate the compressibility without knowing this integral
dP 1 p2
= p 2F (187)
dn 3 pF + m2
p2
which interpolates between the non-relativistic result, 3m F
, and the relativistic
pF
result, 3 . Not surprisingly, the adiabatic index, γ, slowly changes from 35 to 43
as xF goes from 0 to ∞.
48
Academic Section
Indian Institute of Technology Bhilai
IIT Bhilai Government Engineering College Campus Old Dhamtari Road,
Chhattisgarh, India