Solutions to Pathria’s Statistical Mechanics

Chapter 1
SM-at-THU
January 20, 2019

Problem 1.1
ln Ω(0) (E (0) , E1 ) = ln Ω(E1 ) = ln(Ω1 (E1 )Ω2 (E2 )) (1)
We can expand ln Ω1 (E1 ) and ln Ω2 (E1 ) into series near E¯1 , and the leading terms are

∂ ln Ω1 (E1 ) 1 ∂ 2 ln Ω1 (E1 )
ln Ω1 (E1 ) = ln Ω1 (E¯1 ) + (E1 − E¯1 ) + (E1 − E¯1 )2 + · · · (2)
∂E1 E1 =E¯1 2 ∂E12 E1 =E¯1
∂ ln Ω2 (E1 ) 1 ∂ 2 ln Ω2 (E1 )
ln Ω2 (E1 ) = ln Ω2 (E¯1 ) + (E1 − E¯1 ) + (E1 − E¯1 )2 + · · · (3)
∂E1 E1 =E¯1 2 ∂E12 E1 =E¯1

We can set the complicated derivatives to simple symbols

∂ ln Ω1 (E1 )
= a1 (4)
∂E1 E1 =E¯1
∂ ln Ω2 (E1 )
= a2 (5)
∂E1 E1 =E¯1
∂ 2 ln Ω1 (E1 )
= b1 (6)
∂E12 E1 =E¯1
∂ 2 ln Ω2 (E1 )
= b2 (7)
∂E12 E1 =E¯1

b1 + b2
ln Ω(E1 ) = ln(Ω1 (E¯1 )Ω2 (E¯1 )) + (a1 + a2 )(E1 − E¯1 ) + (E1 − E¯1 )2 + · · · (8)
2
This fuction reach its maximum under thermodynamic equilibrium condition at E1 = E¯1 . Thus the linear term must
vanish.
b1 + b2
ln Ω(E1 ) = ln(Ω1 (E¯1 )Ω2 (E¯1 )) + (E1 − E¯1 )2 + · · · (9)
2
b1 + b2
≈ (E1 − E¯1 )2 + ln(Ω(E¯1 )) (10)
2

¯ b1 +b2
2
+ln(Ω(E¯1 )) (E1 −E¯1 )2
Ω(E1 ) = e(b1 +b2 )(E1 −E1 ) = Ae 2 (11)
ln(Ω(E¯1 ))
A=e (12)

1
Which is obviously a gaussian function.
Gaussian RMS can be easily determined to be
1 1 1 1 1
= ∂β ∂β2
= 1 1 (13)
2(b1 + b2 ) 2 ( ∂E +
1
∂E2 )
2 kT1 2 Cv1
+ kT2 2 Cv2
1

For the example of ideal classical gases, we can subtitute Cv i = 23 Ni k and obtain 32 k 2 T 2 NN11+N
N2
2

Problem 1.2
Utilizing the additive characteristic of S = f (Ω) and get

S = S1 + S2 = f (Ω1 ) + f (Ω2 ) (14)

dS
( )Ω = f 0 (Ω1 ) (15)
dΩ1 2
dS
( )Ω = f 0 (Ω2 ) (16)
dΩ2 1

Inspect a small pertubation near the equilibrium state using the fact that S = f (Ω) = f (Ω1 Ω2 )

dS f ((Ω1 + ∆)Ω2 ) − f (Ω1 Ω2 )

( )Ω2 = lim (17)
dΩ1 ∆→0 ∆
Assume that δ = ∆Ω2
dS f (Ω1 Ω2 + ∆Ω2 ) − f (Ω1 Ω2 ) f (Ω + δ) − f (Ω)
( )Ω = lim Ω2 = lim Ω2 = Ω2 f 0 (Ω) (18)
dΩ1 2 ∆→0 ∆Ω2 δ→0 δ
dS
Apply to ( dΩ 2
)Ω1 , we can get similar result.
dS
( )Ω = Ω1 f 0 (Ω) (19)
dΩ2 1
Finally,
Ω2 0
f 0 (Ω1 ) = Ω2 f 0 (Ω) = f (Ω2 ) (20)
Ω1
Ω1 f 0 (Ω1 ) = Ω2 f 0 (Ω2 ) (21)

It is obvious that this equation holds for all Ω. Set the value of the equation constant k.

df (Ω)
Ω =k (22)
dΩ
f (Ω) = klnΩ + C (23)

Using a special value Ω = 1

f (Ω ∗ 1) = f (Ω) + f (1) (24)

C =f (1) = 0 (25)

And get the result

S = f (Ω) = k ln Ω (26)

2
Problem 1.3
When the two systems are brought together, they can form an isolated system. Energy and particle number are constant
while entropy will not not decreasing in such a system.

EA + EB = E0 (27)
NA + NB = N0 (28)
dSA + dSB ≥ 0 (29)

Apply derivation and get

dEA + dEB = 0 (30)

dNA + dNB = 0 (31)
dSA + dSB ≥ 0 (32)

Subtitute these relations into equation

dEA = TA dSA − pA dVA + µA dNA (33)

dEB = TB dSB − pB dVB + µB dNB (34)
dVA = 0 (35)
dVB = 0 (36)

and get
dEA µA TB − µB TA
≥ (37)
dNA TB − TA

Problem 1.4
Suppose N is the number of particles, v0 is the volume occupied by one particle and therefore the total number of
microstates Ω is
1 V V
Ω= ( )...( − N + 1) (38)
N ! 8v0 8v0
The “8” comes from the geometrical configurations that two particle cannot be too close, i.e the distance between two
particles is larger than 2r. Following (1.4.2), we have
 
P ∂ ln Ω
= k (39)
T ∂V N,E
∂Ω
= k (40)
Ω∂V
 
N 8(N − 1)v0
= k 1+ + ... (41)
V 2V

Considering only the first two terms, it corresponds to P (V − b) = N kT with b = 4N v0 .

3
Problem 1.5
√
Using equation (A.11), and setting K = π ε/L, it is straight forward to achieve
π 3/2 3π
Σ1 (ε) = ε ± ε (42)
6 8
where the first term is the volume term (V = L3 ) and the next one is the surface correction (S = 6L2 ).

Problem 1.6
Use the formula for ideal gas P V = N kT .
π
N k × 300 = 105 × (43)
10
Thus ∆T = 104 /N k ∼ 955K.

Problem 1.7
If the energy have follow extreme relativistic [Link] can learn from 1.4.
2
S(N, E, V ) = S(N, 2 V 3 )
Hence,for the constancy of S and N,which defines a reversible adiabatic process.
2
2 V 3 = const
Pressure could be inferred by
∂E
P = −( )N,S
∂V
So we can get the adiabatic equation.
4
PV 3 = const
Cp 4
So the adiabatic exponent Cv = 3

Problem 1.8
We use ni to show the number of quasiparticles in the nth energy state,and we can derive:
+∞
X
ni = N
i=0
+∞
X
ni i = E
i=0
We use the partition function of canonical system.
+∞
X i
Z= gi e KT
i=0
i
e KT
Pi =
Z
finally you can get the T
hν
T =
kln(1 + NEhν )

4
Problem 1.9
According that S is an extensive [Link] S is a linear function.

S = aN + bV + cE
∂S ∂S ∂S
a=( )V,E , b = ( )N,E , c = ( )V,N
∂N ∂V ∂E
So
∂S ∂S ∂S
N( )V,E + V ( )N,E + E( )V,N = S
∂N ∂V ∂E

Problem 1.10
Just use equation (1.4.21) and (1.4.23), we have:
"  3/2 #
2πmkT 3
S(N, V, E) = N k ln V 2
+ Nk (44)
h 2

Since He and Ar have the same N,V. We can get the T that He and Ar have the same entropy:

T = 0K(?)

Problem 1.11
As N2 and O2 are mixed together at the same pressure and temperature, we can know that the volume of mixed gas is:
V = V1 + V2 . And we can get the entropy of mixing by utilizing equation (1.5.3):
 
V V
∆S = k N1 ln + N2 ln (45)
V1 V2

for per mole of the air formed:  

V V
∆Sn = k N1 ln + N2 ln /(n1 + n2 )
V1 V2
 
V V (46)
= R n1 ln + n2 ln /(n1 + n2 )
V1 V2
= 4.16 J · mol−1 · K −1

Problem 1.12
(a) Equation (1.5.3a) can be written as:

(V1 + V2 )N1 (V1 + V2 )N2

(∆S)1≡2 = N1 ln + N2 ln
V1 (N1 + N2 ) V2 (N1 + N2 )
  (47)
y 1−y
= (N1 + N2 ) y ln + (1 − y) ln
x 1−x

Here x = V1 /(V1 + V2 ), y = N1 /(N1 + N2 ).

5
 Consider the function f (x, y) = y ln xy , we can get the second derivatives:

y/x2
 
2 −1/x
D f (x, y) = (48)
−1/x 1/y

Since D2 f (x, y) is a positive-semidefinite, f (x, y) is a convex function. Then we can know that:
1 1
f (x, y) + f (1 − x, 1 − y) ≥ f (1/2, 1/2) = 0 (49)
2 2
This means (∆S)1≡2 ≥ 0 and the equality holding only when N1 /V1 = N2 /V2
∗
(b) Suppose that N = N1 + N2 . And we have (∆S) by utilizing equation (1.5.4) :
 
∗ N N
(∆S) = k N1 ln + N2 ln
N1 N2 (50)
= k [N ln N − N1 ln N1 − N2 ln N2 ]
∗
Then we have the derivative of (∆S) with respect to N1 :
∗
d (∆S) ∂N2
= − ln N1 − ln N2
dN1 ∂N1 (51)
= − (ln N1 − ln N2 )
d(∆S)∗
It shows that dN1 satisfies:

<0 ∗
N1 > N2
d (∆S) 
=0 N1 = N2 (52)
dN1 
>0 N1 < N2
∗
So we can know that (∆S) have the only maximum value at N1 = N2 :
∗
max (∆S) = (N1 + N2 ln 2) (53)

Then we get:
∗
max (∆S) ≤ (N1 + N2 ln 2) (54)
The equality holding when and only when N2 = N2

Problem 1.13
In both cases the entropy will increase, but the increment of different gases mixing is larger. You can check that with the
equations in Section 1.5 and I don’t want to type it again, for entropy-saving’s sake:)

Problem 1.14
3N k T2
(∆S)V = ln( ) (55)
2 T1
While fixing p, V2 /V1 = T2 /T1 .
V2 3N k T2 5N k T2
(∆S)p = N k ln( )+ ln( ) = ln( ) (56)
V1 2 T1 2 T1
Clealy, (∆S)p = 5/3(∆S)V .

6
Problem 1.15
From V1 → V2 , we have

p2a V γ1
= 1γ1 (57)
p1a V2
p2b V1γ2
= γ2 (58)
p1b V2
p2 Vγ
= 1γ (59)
p1 V2

where p1a = f1 p1 and p2a = f2 p1 . After some algebra it’s straightforward to see
1 f1 f2
= + (60)
γ−1 γ1 − 1 γ2 − 1

Problem 1.16
Theorem:
If f (x, y, z) = 0,then we have    
∂x ∂y
=1
∂y z ∂x z
     
∂x ∂y ∂z
= −1
∂y z ∂z x ∂z y

(a)  
S ∂µ
=−
N ∂T P
 
V ∂µ
=
N ∂P T
 
∂µ  
S ∂T 1 ∂P
= −  P = −     =
V ∂µ ∂T ∂µ ∂T µ
∂P ∂µ ∂P
T P T

(b)  
V ∂µ
=
N ∂P T
 
∂P
V =N
∂µ T