Home-Work 1

Himanshu Sahu himanshusah1@iisc.ac.in

Problem 1 ............... 1

Problem 2 ............... 4

P 01- Filament
Part (a)
. . . . . . . . . . . .
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ... 6
6
Part (b) 7
Part (c)
Part (d)
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 7
7

P 02-Hard core gas ........... 8

P 03- Superconducting transition ...... 10

P 04- Photon gas Carnot cycle ....... 12

P 05- Irreversible Processes ........ 14

P 06- Constant Heat Capacities ....... 15

P 07- Non-Carnot Engine ......... 16

Problem 1
Answer: The basic stability condition given by
𝛿𝑆𝛿𝑇 + 𝛿𝐽 𝛿𝑋 + 𝛿𝜇𝛿𝑁 ≥ 0
We will set throughout in this answer 𝛿𝑁 = 0 without mention again. We will not use
subscript either to show this.
Furthermore, Let 𝛿𝐽 𝛿𝑋 = −𝛿𝑝𝛿𝑉 so that our basic stability condition now becomes
𝛿𝑆𝛿𝑇 − 𝛿𝑝𝛿𝑉 ≥ 0

To make the proof sort, Let’s write this in form of general variable and functions
𝛿𝑓 𝛿𝑔 − 𝛿𝜙𝛿𝜓 ≥ 0
Note that here any one from 𝑓 , 𝑔 and any one from 𝜙, 𝜓 are function of any other two
variables 𝑓 , 𝑔𝜙, 𝜓 . For example, We may have 𝑓 (𝜙, 𝑔) and 𝜓 (𝜙, 𝑔). In this case, We can
write

1
 𝛿𝑓 = 𝜕𝜙 𝑓 |𝑔 𝛿𝜙 + 𝜕𝑔 𝑓 |𝜙 𝛿𝑔
𝛿𝜓 = 𝜕𝜙 𝜓 |𝑔 𝛿𝜙 + 𝜕𝑔 𝜓 |𝜙 𝛿𝑔

We are using a sort notation for partials so that

𝜕𝑓
𝜕𝜙 𝑓 |𝑔 ≡ |
𝜕𝜙 𝑔

Putting the partial differential into stability condition, we have

𝛿𝑔(𝜕𝜙 𝑓 |𝑔 𝛿𝜙 + 𝜕𝑔 𝑓 |𝜙 𝛿𝑔) − 𝛿𝜙(𝜕𝜙 𝜓 |𝑔 𝛿𝜙 + 𝜕𝑔 𝜓 |𝜙 𝛿𝑔) ≥ 0

𝜕𝑔 𝑓 |𝜙 (𝛿𝑔)2 + 𝜕𝜙 𝜓 |𝑔 (𝛿𝜙)2 + (𝜕𝜙 𝑓 |𝑔 − 𝜕𝑔 𝜓 |𝜙 )(𝛿𝑔)(𝛿𝜙) ≥ 0

The above inequality is of the form:

𝑎(Δ𝑥)2 + 𝑏(Δ𝑦)2 + 𝑐(Δ𝑥)(Δ𝑦) ≥ 0

and is valid for all Δ𝑥 and Δ𝑦. We want to impose condition on 𝑎 and 𝑏. Consider a
variation such that Δ𝑦 = 0, In this case, We get

𝑎≥0

Similarly If we set Δ𝑥, We get

𝑏≥0
Hence we must have 𝑎 ≥ 0, 𝑏 ≥ 0 for all variations. Of course, this is a weak condition
and we can impose further strong bound on 𝑎, 𝑏 but these are not needed to solve the
problem. See for further discussion on this. We conclude that

𝜕𝑔 𝑓 |𝜙 ≥ 0, 𝜕𝜙 𝜓 | 𝑔 ≥ 0

1. Internal Energy condition

We have build all the machinery. Now let (𝑓 , 𝑔, 𝜙, 𝜓 ) = (𝑇 , 𝑆, 𝑝, 𝑉 ) since the internal
energy is taken as function of 𝑆, 𝑉 then

𝜕𝑆 𝑇 |𝑝 ≥ 0, 𝜕 𝑝 𝑇 |𝑆 ≥ 0

Now using
𝑑𝑈 = 𝑇 𝑑𝑆 − 𝑝𝑑𝑉
We have
𝑇 = 𝜕𝑆 𝐸|𝑉 , 𝑝 = −𝜕𝑉 𝑈 |𝑆
Putting these into our resultant inequalities, we get

𝜕𝑆2 𝐸|𝑉 ≥ 0, 𝜕𝑝2 𝐸|𝑆 ≤ 0

Hence proved!

2
2. Enthalpy condition
Now let (𝑓 , 𝑔, 𝜙, 𝜓 ) = (𝑇 , 𝑆, 𝑝, 𝑉 ) since the enthalpy is taken as function of 𝑆, 𝑝 then

𝜕𝑆 𝑇 |𝑝 ≥ 0, 𝜕𝑝 𝑉 |𝑆 ≥ 0

Now using
𝑑𝐻 = 𝑇 𝑑𝑆 + 𝑉 𝑑𝑝
We have
𝑇 = 𝜕𝑆 𝐻 |𝑝 , 𝑉 = −𝜕𝑝 𝐻 |𝑆
Putting these into our resultant inequalities, we get

𝜕𝑆2 𝐻 |𝑝 ≥ 0, 𝜕𝑝2 𝐻 |𝑆 ≤ 0

Hence proved!

3. Helmholtz function condition

Now let (𝑓 , 𝑔, 𝜙, 𝜓 ) = (𝑆, 𝑇 , 𝑉 , 𝑝) since the Helmholtz function is taken as function of
𝑇 , 𝑉 then
𝜕𝑇 𝑆|𝑉 ≥ 0, 𝜕𝑉 𝑝|𝑇 ≥ 0
Now using
𝑑𝐹 = −𝑆𝑑𝑇 − 𝑝𝑑𝑉
We have
𝑆 = −𝜕𝑇 𝐹 |𝑉 , 𝑝 = −𝜕𝑉 𝐹 |𝑇
Putting these into our resultant inequalities, we get

𝜕𝑇2 𝐹 |𝑉 ≤ 0, 𝜕𝑉2 𝐹 |𝑇 ≤ 0

Hence proved!

4. Gibbs function condition

Now let (𝑓 , 𝑔, 𝜙, 𝜓 ) = (𝑆, 𝑇 , 𝑝, 𝑉 ) since the Helmholtz function is taken as function of
𝑇 , 𝑉 then
𝜕𝑇 𝑆|𝑝 ≥ 0, 𝜕𝑝 𝑉 |𝑇 ≥ 0
Now using
𝑑𝐺 = −𝑆𝑑𝑇 + 𝑉 𝑑𝑝
We have
𝑆 = −𝜕𝑇 𝐺|𝑝 , 𝑉 = 𝜕𝑝 𝐺|𝑇
Putting these into our resultant inequalities, we get

𝜕𝑇2 𝐺|𝑝 ≤ 0, 𝜕𝑝2 𝐺|𝑇 ≥ 0

Hence proved!

3
The entropy of black hole given by

𝐴 𝜋𝑅2
𝑆= = = 4𝜋𝐺𝑁 𝑀 2
4𝐺𝑁 𝐺𝑁

and energy 𝐸 = 𝑀. Assuming that the work done is negligible, We have

𝑑𝐸 = 𝑇 𝑑𝑆 → 𝑑𝑀 = 𝑇 𝑑(4𝜋𝐺𝑁 𝑀 2 ) = 𝑇 4𝜋𝐺𝑁 (2𝑀𝑑𝑀)

1
→ 𝑇=
8𝜋𝐺𝑁 𝑀
Stability condition given by

𝛿𝑇 𝛿𝑆 + 𝛿𝐽 𝛿𝑋 + 𝛿𝜇𝛿𝑁 ≥ 0
Assuming 𝛿𝑁 = 0 = 𝛿𝐽 , we have
𝛿𝑇 𝛿𝑆 ≥ 0
In our case,
1
𝛿𝑇 𝛿𝑆 = 𝛿 ( ) 𝛿(4𝜋𝐺𝑁 𝑀 2 )
8𝜋𝐺𝑁 𝑀
1 1 1
= (− 2 ) (2𝑀)(𝛿𝑀)2 = − (𝛿𝑀)2 ≤ 0
2 𝑀 𝑀
Hence system is not thermodynamically stable.
The reason being that black hole is radiating it’s rest energy as heat until it run out of
mass and fully evaporated.
Here’s How?

𝑀′
𝛿𝑄 = 𝑇 𝑑𝑆 = 𝑑𝑀 → ∫ 𝛿𝑄 = 𝑄lost = ∫ 𝑑𝑀 = Δ𝑀 < 0
𝑀
The heat lost during a temperature change can also be written as 𝛿𝑄 = 𝑚𝐶𝑑𝑇 where 𝐶 is
specific heat (constant pressure). Here

1 −𝑑𝑀
𝑚𝐶𝑑𝑇 = 𝑑𝑀 = 𝑚𝐶 ( 2 ) → 𝐶 = −8𝜋𝐺𝑁 𝑀 < 0
8𝜋𝐺𝑁 𝑀

which again shows that system is unstable. Integrating 𝑇 𝑑𝑆 over the entire mass 𝑀 of a
black hole would give 𝑄 = −𝑀 which suggest that its total energy would eventually be
radiated away as heat.

Problem 2

4
Answer: We know that
𝐶𝑝 = 𝑇 𝜕𝑇 𝑆|𝑝 (1)
Using identity
𝜕𝑇 𝑆|𝑝 ⋅ 𝜕𝑝 𝑇 |𝑆 ⋅ 𝜕𝑆 𝑝|𝑇 = −1
𝐶𝑝
→
⋅ 𝜕 𝑇 | ⋅ 𝜕 𝑝| = −1
𝑇 𝑝 𝑆 𝑆 𝑇
Furthermore, the response functions given by
1 1
𝛽𝑝 = 𝜕𝑇 𝑉 |𝑝 & 𝛽𝑆 = 𝜕𝑇 𝑉 |𝑆
𝑉 𝑉
1 1
𝜅𝑇 = − 𝜕𝑝 𝑉 |𝑇 & 𝜅𝑆 = − 𝜕𝑝 𝑉 |𝑆
𝑉 𝑉

1 1 1
𝜕𝑆 𝑝|𝑇 = = =
𝜕𝑝 𝑆|𝑇 −𝜕𝑇 𝑉 |𝑝 𝛽𝑝 𝑉
where we used the Maxwell’s relation in going from second step to third. Furthermore,
1
𝜕𝑝 𝑇 |𝑆 = 𝜕𝑝 𝑉 |𝑆 ⋅ 𝜕𝑉 𝑇 |𝑆 = (−𝜅𝑆 𝑉 )
𝑉 𝛽𝑆
Putting this all we have

𝐶𝑝 1 1
⋅ ⋅ (−𝜅𝑆 𝑉 ) = −1
𝑇 𝛽𝑝 𝑉 𝑉 𝛽𝑆
𝛽𝑆 𝛽𝑝
𝐶𝑝 = ( )𝑉 𝑇
𝜅𝑆

Now consider
𝜕𝑇 𝑆|𝑉 ⋅ 𝜕𝑆 𝑉 |𝑇 ⋅ 𝜕𝑉 𝑇 |𝑆 = −1 (2)
Let’s see term by term
1 1
𝜕𝑉 𝑇 |𝑆 = =
𝜕𝑇 𝑉 |𝑆 𝑉 𝛽𝑆
1 1
𝜕𝑆 𝑉 |𝑇 = =
𝜕𝑉 𝑆|𝑇 𝜕𝑇 𝑝|𝑉
1 −𝜕𝑝 𝑉 |𝑇 𝑉 𝜅𝑇 𝜅𝑇
=− = = =
𝜕𝑉 𝑝|𝑇 ⋅ 𝜕𝑇 𝑉 |𝑝 𝜕𝑇 𝑉 |𝑝 𝑉 𝛽 𝑝 𝛽𝑝

Putting these, we have

𝜅𝑇 1 𝑉 𝛽 𝑝 𝛽𝑆
𝜕𝑇 𝑆|𝑉 ⋅ ( )( ) = −1 → 𝜕𝑇 𝑆|𝑉 = −
𝛽𝑝 𝑉 𝛽𝑆 𝜅𝑇

5
Hence
𝛽𝑝 𝛽𝑆
𝐶𝑉 = −𝑉 𝑇
𝜅𝑇

Consider 𝑆 = 𝑆(𝑇 , 𝑉 )
𝑑𝑆 = 𝜕𝑇 𝑆|𝑉 𝑑𝑇 + 𝜕𝑉 𝑆|𝑇 𝑑𝑉
differentiating this equation with respect to 𝑇 at constant 𝑝 yields

𝜕𝑇 𝑆|𝑝 = 𝜕𝑇 𝑆|𝑉 + 𝜕𝑉 𝑆|𝑇 ⋅ 𝜕𝑇 𝑉 |𝑝

𝜕𝑉 𝑆|𝑇 = 𝜕𝑇 𝑝|𝑉 = −𝜕𝑉 𝑝|𝑇 ⋅ 𝜕𝑇 𝑉 |𝑝

Hence
𝛽𝑝2
𝐶𝑝 − 𝐶𝑉 = 𝑉 𝑇
𝜅𝑇

The relation for 𝐶𝑝 and 𝐶𝑉 are not unique. For example, Consider

𝐶𝑝 𝜕𝑇 𝑉 |𝑝 ⋅ 𝜕𝑇 𝑝|𝑆 𝜕𝑇 𝑉 |𝑝 𝜕𝑇 𝑉 |𝑝
=− =− ⋅ 𝜕𝑉 𝑝|𝑆 =
𝐶𝑉 𝜕𝑇 𝑝|𝑉 ⋅ 𝜕𝑇 𝑉 |𝑆 𝜕𝑇 𝑝|𝑉 𝜕𝑝 𝑉 |𝑆
Hence
𝐶𝑝 𝜅𝑇
=
𝐶𝑉 𝜅𝑆
Hence using 𝐶𝑝 − 𝐶𝑉 , we can write

𝑇 𝑉 𝛽𝑝2 𝜅𝑆 𝑇 𝑉 𝛽𝑝2
𝐶𝑉 = & 𝐶𝑝 =
(𝜅𝑇 − 𝜅𝑆 )𝜅𝑇 𝜅𝑇 − 𝜅 𝑆

P 01- Filament
Answer: For an elastic filament it is found that, at a finite range in temperature, a dis-
placement 𝑥 requires a force

𝐽 = 𝑎𝑥 − 𝑏𝑇 + 𝑐𝑇 𝑥
where 𝑎, 𝑏 and 𝑐 are constants. Furthermore, its heat capacity at displacement is propor-
tional to temperature, that is, 𝐶𝑥 = 𝐴(𝑥)𝑇

Part (a)
From Maxwell’s equation

𝜕𝑥 𝑆|𝑇 = −𝜕𝑇 𝐽 |𝑥 = −𝜕𝑇 (𝑎𝑥 − 𝑏𝑇 + 𝑐𝑇 𝑥) = 𝑏 − 𝑐𝑥

6
Part (b)
Integrating the last result, we have

𝑎𝑥 2
𝑆(𝑇 , 𝑥) = 𝑏𝑥 − + 𝑓 (𝑇 )
2
By definition
𝐶𝑥 = 𝑇 𝜕𝑇 𝑆|𝑥
comparing with the form 𝐶𝑥 = 𝐴(𝑥)𝑇 , we have

𝐴(𝑥) = 𝜕𝑇 𝑆|𝑥

From the form of 𝑆(𝑇 , 𝑥), We have

𝜕𝑇 𝑆|𝑥 = 𝑓 ′ (𝑇 ) → 𝐴(𝑥) = 𝑓 ′ (𝑇 )

But since 𝐴(𝑥) is only function of 𝑥, we must have 𝑓 ′ (𝑇 ) = 𝜙0 where 𝜙0 is constant. There-
fore
𝑑 𝑑
𝐴(𝑥) = 𝜙0 = 0
𝑑𝑥 𝑑𝑥

Part (c)
We found in the last section that

𝑓 ′ (𝑇 ) = 𝜙0 → 𝑓 (𝑇 ) = 𝜙0 𝑇 + 𝜙1

hence
𝑎𝑥 2
𝑆(𝑇 , 𝑥) = 𝑏𝑥 − + 𝜙0 𝑇 + 𝜙 1
2
Assuming 𝑆(0, 0) = 𝑆0 , we have

𝑎𝑥 2
𝑆(0, 0) = 𝜙1 = 𝑆0 → 𝑆(𝑇 , 𝑥) = 𝑏𝑥 − + 𝜙0 𝑇
2
Note that 𝜙0 = 𝐴, we have
𝑎𝑥 2
𝑆(𝑇 , 𝑥) = 𝑏𝑥 − + 𝐴𝑇
2

Part (d)
Using 𝐽 (𝑥, 𝑇 ), we have
𝐽 + 𝑏𝑇
𝐽 = 𝑎𝑥 − 𝑏𝑇 + 𝑐𝑇 𝑥 → 𝑥 =
𝑎 + 𝑐𝑇
Now writing the entropy as 𝑆(𝑇 , 𝑥) = 𝑆(𝑇 , 𝑥(𝑇 , 𝐽 )),leads to

𝜕𝑇 𝑆|𝐽 = 𝜕𝑇 𝑆|𝑥 + 𝜕𝑥 𝑆|𝑇 𝜕𝑇 𝑥|𝐽

7
We already found
𝜕𝑥 𝑆|𝑇 = 𝑏 − 𝑐𝑥 & 𝜕𝑇 𝑆|𝑥 = 𝐴
Using 𝑥(𝑇 , 𝐽 ), we have
𝑎𝑏 − 𝐽 𝑐
𝜕𝑇 𝑥|𝐽 =
(𝑎 + 𝑐𝑇 )2
Using
𝑎𝑏 − 𝐽 𝑐 = 𝑎𝑏 − 𝑐(𝑎𝑥 − 𝑏𝑇 + 𝑐𝑇 𝑥) = 𝑎𝑏 + 𝑏𝑐𝑇 − 𝑎𝑐𝑥 − 𝑐 2 𝑇 𝑥 = (𝑎 + 𝑐𝑇 )(𝑏 − 𝑥𝑐)
Putting this we get
(𝑎 + 𝑐𝑇 )(𝑏 − 𝑥𝑐) 𝑏 − 𝑥𝑐
𝜕𝑇 𝑥|𝐽 = =
(𝑎 + 𝑐𝑇 )2 𝑎 + 𝑐𝑇

Thus
(𝑏 − 𝑐𝑥)2
𝐶𝐽 = 𝑇 𝜕𝑇 𝑆|𝐽 = 𝑇 (𝐴 + )
(𝑎 + 𝑐𝑇 )
We can write this as a function of 𝑇 and 𝐽 , as

1 𝐽 + 𝑏𝑇
𝐶𝐽 = 𝑇 [𝐴 + (𝑏 − 𝑐 )]
(𝑎 + 𝑐𝑇 ) 𝑎 + 𝑐𝑇

P 02-Hard core gas

Answer: A gas obeys the equation of state

𝑃(𝑉 − 𝑁 𝑏) = 𝑁 𝑘𝐵 𝑇 (3)

and has a heat capacity 𝐶𝑉 independent of temperature.

1. Part (a)
The Free energy potential given by

𝑑𝐹 = −𝑆𝑑𝑇 − 𝑝𝑑𝑉 (4)

Supposing 𝐹 = 𝐹 (𝑇 , 𝑉 , 𝑁 ), we have

𝑑𝐹 = 𝜕𝑇 𝐹 |𝑉 ,𝑁 𝑑𝑇 + 𝜕𝑉 𝐹 |𝑇 ,𝑁 𝑑𝑉 + 𝜕𝑁 𝐹 |𝑇 ,𝑉 𝑑𝑁

Comparing with Eq. (4), we have

−𝑆 = 𝜕𝑇 𝐹 |𝑉 ,𝑁 & − 𝑝 = 𝜕𝑉 𝐹 |𝑇 ,𝑁

Equating the mixed derivative, we have

𝜕 2𝐹 𝜕 2𝐹 𝜕𝑆 𝜕𝑃
| = | → | = |
𝜕𝑇 𝜕𝑉 𝑇 ,𝑁 𝜕𝑉 𝜕𝑇 𝑉 ,𝑁 𝜕𝑉 𝑇 ,𝑁 𝜕𝑇 𝑉 ,𝑁

8
2. Part (b) From now on, We consider 𝑑𝑁 = 0 and wouldn’t show as subscript to make
our life easy.

𝑑𝐸 = 𝑇 𝑑𝑆 − 𝑝𝑑𝑉 = 𝑇 (𝜕𝑇 𝑆|𝑉 𝑑𝑇 + 𝜕𝑉 𝑆|𝑇 𝑑𝑉 ) − 𝑝𝑑𝑉

Using the Maxwell’s equation, we have
𝑑𝐸 = 𝑇 𝜕𝑇 𝑆|𝑉 𝑑𝑇 + (𝑇 𝜕𝑇 𝑝|𝑉 − 𝑝)𝑑𝑉
Using the equation of state, we find
𝑁 𝑘𝐵 𝑇 𝑃
𝑃= → 𝜕𝑇 𝑃|𝑉 =
(𝑉 − 𝑁 𝑏) 𝑇
Putting this in, we get
𝑑𝐸 = 𝑇 𝜕𝑇 𝑆|𝑉 𝑑𝑇
which means that 𝐸 = 𝐸(𝑇 , 𝑁 ) and doesn’t depend on 𝑉 .
3. Part (c)
The heat capacity given by
𝐶𝑉 = 𝜕𝑇 𝐸|𝑉 = 𝜕𝑇 𝐸|𝑝 (5)
𝐶𝑝 = 𝜕𝑇 𝑄|𝑝 = 𝜕𝑇 𝐸|𝑝 + 𝑝𝜕𝑇 𝑉 |𝑝 (6)
𝐶𝑝 = 𝐶𝑉 + 𝑝𝜕𝑇 𝑉 |𝑝 (7)
Using the equation of state
𝑁 𝑘𝐵 𝑇 𝑁 𝑘𝐵
𝑉= + 𝑁 𝑏 → 𝜕𝑇 𝑉 |𝑃 =
𝑃 𝑃
Putthing this, we get
𝑁 𝑘𝐵
𝐶𝑝 = 𝐶𝑉 + 𝑁 𝑘𝐵 → 𝛾 = 1 +
𝐶𝑉
4. For adiabatic condition, we have
𝑑𝑆 = 0 → 𝑑𝐸 + 𝑝𝑑𝑉 = 0
Using the fact that 𝑑𝐸 = 𝐶𝑉 𝑑𝑇 and from equation of state
𝑝(𝑉 − 𝑁 𝑏) 1
𝑇= → 𝑑𝑇 = ((𝑉 − 𝑁 𝑏)𝑑𝑝 + 𝑝𝑑𝑉 )
𝑁 𝑘𝐵 𝑁 𝑘𝐵
Hence
1
𝐶𝑉 ((𝑉 − 𝑁 𝑏)𝑑𝑝 + 𝑝𝑑𝑉 ) + 𝑝𝑑𝑉 = 0
𝑁 𝑘𝐵
𝐶𝑉 𝐶𝑉 (𝑉 − 𝑁 𝑏) 𝑑𝑝 1 𝑑𝑉 𝐶𝑉 𝑑𝑝
→ 𝑑𝑉 [1 + ]+ = 0 → (1 + ) +
𝑁 𝑘𝐵 𝑁 𝑘𝐵 𝑝 𝛾 − 1 𝑉 − 𝑁 𝑏 𝑁 𝑘𝐵 𝑝
Or,
1 𝑑𝑉 1 𝑑𝑝 𝛾 𝑑𝑉 𝑑𝑝
→ (1 + ) + =0→ + =0
𝛾 − 1 (𝑉 − 𝑁 𝑏) 𝛾 − 1 𝑝 𝑉 − 𝑁𝑏 𝑝
Hence we have
𝑝(𝑉 − 𝑁 𝑏)𝛾 = constant

9
P 03- Superconducting transition

Answer: Many metals become superconductors at low temperatures 𝑇 , and magnetic

field 𝐵. The heat capacities of the two phase at zero magnetic field are approximately
given by

𝐶𝑠 (𝑇 ) = 𝑉 𝛼𝑇 3 In the superconducting phase (8)

𝐶𝑛 (𝑇 ) = 𝑉 [𝛽𝑇 3 + 𝛾 𝑇 ] In the normal phase (9)

1. Using
𝑑𝑆
𝐶 =𝑇
𝑑𝑇
we have
𝑑𝑆𝑠 𝑑𝑆
𝐶𝑠 = 𝑉 𝛼𝑇 3 = 𝑇 → 𝑠 = 𝑉 𝛼𝑇 2 (10)
𝑑𝑇 𝑑𝑇
𝑑𝑆 𝑑𝑆
𝐶𝑛 = 𝑉 [𝛽𝑇 3 + 𝛾 𝑇 ] = 𝑇 𝑛 → 𝑛 = 𝑉 [𝛽𝑇 2 + 𝛾 ] (11)
𝑑𝑇 𝑑𝑇
From third law of thermodynamics, we have 𝑆(𝑇 = 0) = 0. With integration, we have
3
𝑉 𝛼𝑇 3 𝛽𝑇
𝑆𝑠 = & 𝑆𝑛 = 𝑉 [ + 𝛾𝑇]
3 3

2. Part (b)
We know that the Latent heat define as

𝐿 = Δ𝑄rev = 𝑇𝑐 (𝑆2 − 𝑆1 )

where 𝑆1 and 𝑆2 are entropy of phase 1 and 2. For the superconducting phase
transition 𝐿 = 0, Hence
𝑉 𝛼𝑇𝑐3 𝛽𝑇𝑐3
𝑆1 = 𝑆2 → =𝑉[ + 𝛾 𝑇𝑐 ]
3 3
Hence
3𝛾
𝑇𝑐 =
√𝛼 − 𝛽
3. Part (c)
At zero temperature, the elctron in the superconductor form bound Cooper pairs.
As a result, the internal energy of the superconductor is reduced by an amount 𝑉 Δ

𝐸𝑛 (𝑇 = 0) = 𝐸0
𝐸𝑠 (𝑇 = 0) = 𝐸0 − 𝑉 Δ

10
 for metal and superconductor, respectively.
Using, First law of thermodynamics with 𝐽 𝑑𝑋 = 𝐵𝑑𝑀, we have

𝑑𝐸 = 𝑇 𝑑𝑆 + 𝐵𝑑𝑀
Since the magnetic field is zero (Note we have also assume that 𝑑𝑁 = 0), We have

𝑑𝐸 = 𝑇 𝑑𝑆 = 𝐶𝑑𝑇

Integrating this for the two states, we have

𝐸𝑠 (𝑇 ) 𝑇
𝛼
∫ 𝑑𝐸 = ∫ 𝑉 𝛼𝑇 3 𝑑𝑇 → 𝐸𝑠 (𝑇 ) = 𝐸0 + 𝑉 [−Δ + 𝑇 4 ]
𝐸0 −𝑉 Δ 0 4
𝐸𝑛 (𝑇 ) 𝑇
𝛽 𝛾
∫ 𝑑𝐸 = ∫ 𝑉 [𝛽𝑇 3 + 𝛾 𝑇 ]𝑑𝑇 → 𝐸𝑛 (𝑇 ) = 𝐸0 + 𝑉 [ 𝑇 4 + 𝑇 2 ]
𝐸0 0 4 2

4. Part (d) With zero magnetic field 𝐵, The Gibbs free energy 𝐺 given by

𝐺 = 𝐸 − 𝑇 𝑆 = 𝜇𝑁

Putting the values 𝐸, 𝑆 for the two states, we have

𝛼 𝛼 𝛼
𝐺𝑠 (𝑇 ) = 𝐸0 + 𝑉 [−Δ + 𝑇 4 ] − 𝑇 𝑉 𝑇 3 = 𝐸0 − 𝑉 [Δ + 𝑇 4 ]
4 3 4
𝛽 4 𝛾 2 𝛽 3 𝛽 𝛾
𝐺𝑛 (𝑇 ) = 𝐸0 + 𝑉 [ 𝑇 + 𝑇 ] − 𝑇 𝑉 [ 𝑇 + 𝛾 𝑇 ] = 𝐸0 − 𝑉 [ 𝑇 4 + 𝑇 2 ]
4 2 3 12 2
Using the fact that in phase equilibrium, each coexisting phase has the same chem-
ical potential or Gibbs free energy, we have

𝐺𝑠 (𝑇𝑐 ) = 𝐺𝑛 (𝑇 )
𝛽 4 𝛾 2 𝛼 𝛾 𝛼 −𝛽 4
→ 𝐸0 − 𝑉 [ 𝑇𝑐 + 𝑇𝑐 ] = 𝐸0 − 𝑉 [Δ + 𝑇𝑐4 ] → Δ = 𝑇𝑐2 − 𝑇
12 2 4 2 12 𝑐
Putting the value of 𝑇𝑐 obtain in part (b), we have

2
3 𝛾
Δ=
4𝛼 −𝛽

5. Part (e)
In the presence of a magnetic field 𝐵, inclusion of magnetic work results in

𝑑𝐸 = 𝑇 𝑑𝑆 + 𝐵𝑑𝑀 + 𝜇𝑑𝑁

11
 where 𝑀 is the magnetization. The superconducting phase is a perfect diamagnet,
expelling the magnetic field from its interior such that
𝑉𝐵
𝑀𝑠 =
4𝜋
in appropriate units. The normal metal can be regarded as approximately non-
magnetic with

𝑀𝑛 = 0

The Gibbs free energy will modify due to field 𝐵 in superconducting state but not
in normal state since 𝑀𝑛 = 0. For superconducting state, We have
𝐵
𝑉 𝐵2
𝐺𝑠 (𝑇 , 𝐵) = 𝐺𝑠 (𝑇 , 0) − ∫ 𝑀𝑠 𝑑𝐵 = 𝐺𝑠 (𝑇 , 0) + ∫ 𝐵𝑑𝐵 = 𝐺𝑠 (𝑇 , 0) + 𝑉
4𝜋 0 4𝜋

Equating the Gibbs free energies of two phases at critical temperature gives

𝐵𝑐2 𝛽 𝛾
𝐺𝑠 (𝑇𝑐 , 0) + 𝑉 = 𝐸0 − 𝑉 [ 𝑇𝑐4 + 𝑇𝑐2 ]
4𝜋 12 2
Therefore,
𝑉 𝐵𝑐2 𝛽 𝛾 𝛼
= 𝐸0 − 𝑉 [ 𝑇𝑐4 + 𝑇𝑐2 ] − 𝐸0 + 𝑉 [Δ + 𝑇𝑐4 ]
4𝜋 12 2 4
Solving this leads to
𝐵𝑐2 𝛼 − 𝛽 2
= (𝑇𝑐 − 𝑇 2 )2
8𝜋 12
Here we used the value of Δ and 𝑇𝑐 obtain in previous parts. We can shuffle the
equation
2𝜋(𝛼 − 𝛽) 2 𝑇2
𝐵𝑐 = 𝑇𝑐 (1 − 2 )
√ 3 𝑇𝑐
comparing this with the given form in problem, we have

2𝜋(𝛼 − 𝛽) 2
𝐵0 = 𝑇𝑐 = 𝑇𝑐 √2𝜋𝛾
√ 3
where in the last step, we put the value of 𝑇𝑐 to get rid of (𝛼 − 𝛽) term.

P 04- Photon gas Carnot cycle

12
 Figure 1: Photon Gas Carnot cycle

Answer: Consider the infinitesimal Carnot cycle on the photon gas shown in figure (1)

1. Part (a)
The work done is given by the area of the cycle 𝐴𝐵𝐶𝐷. This area is equal to area
of region 𝐴𝐵′ 𝐶𝐷, therefore
𝑊 = 𝑑𝑃𝑑𝑉

2. Part (b)
From the first law

𝑑 ̄𝑄 = 𝑑𝐸 + 𝑝𝑑𝑉 = [𝜕𝑇 𝐸|𝑉 𝑑𝑇 + 𝜕𝑉 𝐸|𝑇 𝑑𝑉 ) + 𝑝𝑑𝑉

For Isothermal process

𝑑 ̄𝑄 = 𝜕𝑉 𝐸|𝑇 𝑑𝑉 + 𝑝𝑑𝑉

3. Part (c)
By definition, The efficiency of the Carnot cycle

𝑊 𝑇 𝑇 𝑑𝑇 𝑑𝑇
𝜂= =1− 𝐶 =1− = ≈
𝑄 𝑇𝐻 𝑇 + 𝑑𝑇 𝑇 + 𝑑𝑇 𝑇
𝑑𝑃 𝑑𝑇
→ =
𝜕𝑉 𝐸|𝑇 + 𝑃 𝑇
4. Part (d)
Observations indicate that the pressure of the photon gas is given by

𝜋 2 𝑘𝐵4
𝑃 = 𝐴𝑇 4 , where 𝐴 =
45(ℎ ̄𝑐)3

13
 Using the result of part (c), we have

𝑇 𝑑𝑃 = (𝜕𝑉 𝐸|𝑇 + 𝑃)𝑑𝑇 → 𝑇 (4𝐴𝑇 3 𝑑𝑇 ) = (𝜕𝑉 𝐸|𝑇 + 𝐴𝑇 4 )𝑑𝑇

→ 𝜕𝑉 𝐸|𝑇 = 3𝐴𝑇 4 → 𝐸 = 3𝐴𝑉 𝑇 4

where we assumed 𝐸(𝑇 , 0) = 0.

5. Part (e) For adiabatic analysis, We set 𝑑 ̄𝑄 = 0 so that

0 = [𝜕𝑇 𝐸|𝑉 𝑑𝑇 + 𝜕𝑉 𝐸|𝑇 𝑑𝑉 ] + 𝑃𝑑𝑉

𝑑𝑃
→ 12𝐴𝑉 𝑇 3 𝑑𝑇 + 3𝐴𝑇 4 𝑑𝑉 + 𝐴𝑇 4 𝑑𝑉 = 12𝐴𝑉 𝑇 3 + 4𝐴𝑉 𝑇 4 𝑑𝑉 = 0
4𝐴𝑇 3
3𝑉 𝑑𝑃 + 4𝑃𝑑𝑉 = 0 → 𝑃𝑉 4/3 = constant

P 05- Irreversible Processes

Answer:
Part (a)
Consider two substances, initially at temperature 𝑇10 and 𝑇20 , coming to equilibrium at a
final temperature 𝑇𝑓 through heat exchange. Let 𝑑 ̄𝑄 is the amount of heat that flow from
substance 1 to substance 2. Then
𝑇𝑓 𝑇𝑓
𝑑 ̄𝑄1 𝑑 ̄𝑄2 𝑇 − 𝑇2
Δ𝑆 = Δ𝑆1 + Δ𝑆2 ≥ ∫ +∫ =∫ 1 𝑑 ̄𝑄 (12)
𝑇10 𝑇1 𝑇20 𝑇2 𝑇1 𝑇2

According to second law of thermodynamics, We know that the heat flows from hot to
cold so that if 𝑇10 > 𝑇20 then 𝑑 ̄𝑄 > 0 and if 𝑇10 < 𝑇20 then 𝑑 ̄𝑄 < 0. In both the cases, we have

𝑇1 − 𝑇 2
Δ𝑆 ≥ ∫ 𝑑 ̄𝑄 ≥ 0
𝑇1 𝑇2

Part (b)
Now consider a gas with adjustable volume 𝑉 , and diathermal walls, embedded in a
heat bath of constant temperature 𝑇 , and fixed pressure 𝑃. The change in the entropy
of the bath is given by

Δ𝑄bath Δ𝑄 1
Δ𝑆bath = =− = − (Δ𝐸gas + 𝑃Δ𝑉gas )
𝑇 𝑇 𝑇

14
From second law of thermodynamics, We know that Δ𝑆 ≥ 0.
1 1
Δ𝑆 = Δ𝑆bath + Δ𝑆gas = − (Δ𝐸gas + 𝑃Δ𝑉gas − 𝑇 Δ𝑆gas ) = − Δ𝐺gas
𝑇 𝑇
In the last step, we use

𝐺 = 𝐸 + 𝑃𝑉 − 𝑇 𝑆 → Δ𝐺 = Δ𝐸 − 𝑇 Δ𝑆 + 𝑃Δ𝑉 )

In the last step, we assume constant temperature and pressure.

Hence
1
Δ𝑆 ≥ 0 → − Δ𝐺gas ≥ 0 → Δ𝐺gas ≤ 0
𝑇
Therefore at constant temperature and pressure, we must minimize the Gibbs free en-
ergy to find the equilibrium state.

P 06- Constant Heat Capacities

Answer:

1. Part (a)
Since 𝑇1 > 𝑇2 , the heat flows from body 1 to body 2. The heat gained by body 1 is
equals to heat gained by body 2 and so

𝑄gained = 𝑄lost → 𝐶2 (𝑇𝑓 − 𝑇2 ) = 𝐶1 (𝑇1 − 𝑇𝑓 )

𝐶1 𝑇1 + 𝐶2 𝑇2
→ 𝑇𝑓 =
𝐶1 + 𝐶2

The change in entropy given by

𝑇𝑓 𝑇𝑓
𝑑 ̄𝑄1 𝑑 ̄𝑄2
Δ𝑆 = ∫ +∫
𝑇1 𝑇 𝑇2 𝑇
𝑇𝑓 𝑇𝑓
𝐶1 𝑑𝑇 𝐶2 𝑑𝑇
=∫ +∫
𝑇1 𝑇 𝑇2 𝑇
𝑇𝑓 𝑇𝑓
= 𝐶1 ln 𝑇 |𝑇1 + 𝐶2 ln 𝑇 |𝑇2
𝑇𝑓 𝑇𝑓
= 𝐶1 ln ( ) + 𝐶2 ln ( )
𝑇1 𝑇2
Therefore,
𝑇𝑓 𝑇𝑓
Δ𝑆 = 𝐶1 ln ( ) + 𝐶2 ln ( )
𝑇1 𝑇2

15
 2. Part (b)
In an infinitesimal change we have that

𝑑 ̄𝑄ℎ = −𝐶1 𝑑𝑇1

𝑑 ̄𝑄𝑙 = 𝐶2 𝑑𝑇2

and for a Carnot engine we have that

𝑑 ̄𝑄ℎ 𝑑 ̄𝑄𝑙
=
𝑇1 𝑇2

and integrating gives

𝑇𝑓 𝑇𝑓
𝑑 ̄𝑄𝑙 𝑑 ̄𝑄ℎ
∫ = −∫
𝑇2 𝑇2 𝑇 1 𝑇1
and hence
𝑇𝑓 𝑇𝑓
𝐶2 ln = −𝐶1 ln
𝑇2 𝑇1
and 𝑇𝑓 is the final temperature of each reservoir. Thus

𝐶 +𝐶2 𝐶 𝐶
𝑇𝑓 1 = 𝑇 1 1 𝑇2 2

The total heat extracted from each reservoir is Δ𝑄ℎ = 𝐶1 (𝑇1 − 𝑇2 ) and Δ𝑄𝑙 = 𝐶2 (𝑇𝑓 − 𝑇2 )
respectively and so that total work is

Δ𝑊 = Δ𝑄ℎ − Δ𝑄𝑙 = 𝐶1 𝑇1 + 𝐶2 𝑇2 − (𝐶1 + 𝐶2 )𝑇𝑓

P 07- Non-Carnot Engine

Consider an engine that operates between a set of temperature 𝑇max = 𝑇1 > 𝑇2 > … > 𝑇𝑛 =
𝑇min . For a subset of these {𝑇𝛽− } the engine takes in heat {𝑄𝛼+ }; while at the remaining
temperature {𝑇𝛽− }, heat {𝑄𝛽− } is released. The engine does a net work 𝑊 = 𝑄 + − 𝑄 − , and
its efficiency is given by
𝑊
𝜂= where 𝑄+ = ∑ 𝑄𝛼+ & 𝑄− = ∑ 𝑄𝛼−
𝑄+ 𝛼 𝛼

Show that this efficiency is less than that of a Carnot engine operating between 𝑇max
and 𝑇min .

Answer: Consider a reversible cycle EFGH. Any such cycle will operate between a max-
imum and a minimum temperature and a maximum and a minimum entropy. It can,
therefore, always be enclosed in a Carnot cycle, as shown in Fig. 2.

16
Figure 2: A reversible engine cycle and the enclosing Carnot cycle on a T-S diagram.
The arrow indicate the cycle direction

We will compare the efficiency of the enclosed cycle to that of the enclosing Carnot cycle.
The enclosing cycle has set of temperature lie between 𝑇1 and 𝑇𝑛 at which it absorb heat
and a set of temperature lie on the same range at which it releases the heat. Here we
made a arbitrary cycle.
The entropy is increasing along EFG, so heat is being absorbed. Along GHE, the entropy
is decreasing, so heat is being rejected. From

𝑑 ̄𝑄 = 𝑇 𝑑𝑆

the heat absorbed along BC is the area of the rectangle BCJI. The heat absorbed along
EFG is the area of EFGJI. Likewise, the heat rejected along DA is the area of the rectangle
DAIJ. The heat rejected along GHE is the are GHEIJ. Using

∫ 𝑑 ̄𝑄out
𝜂=1−
∫ 𝑑 ̄𝑄in

it’s clear that, any reversible engine operating between a maximum temperature 𝑇2 and
minimum temperature 𝑇1 must absorb less heat and reject more heat than the enclosing
Carnot cycle. It is, therefore, less efficient.
In the problem, we are given discrete step yet the it’s not very different from what we
proved here. To see consider the Fig. 3.

𝑄−
𝜂=1−
𝑄+

First, note that

𝑊 = Area inside the orange cycle = ∑ 𝑇𝑖 Δ𝑆𝑖
𝑖

17
Figure 3: The blue curve shows the Carnot cycle while orange curve shows the arbitrary
cycle that operate between set of temperature.

There are two set of temperatures. One at which the Δ𝑆 is positive that is the heat is
absorbed while other at which Δ𝑆 is negative that is the heat is released. So we can
write

𝑊 = ∑ 𝑇𝛼+ |Δ𝑆𝛼 | − ∑ 𝑇𝛽− |Δ𝑆𝛽 |

𝛼 𝛽

≤ 𝑇max ∑ |Δ𝑆𝛼 | − 𝑇min ∑ |Δ𝑆𝛽 |

𝛼 𝛽
≤ (𝑇max − 𝑇min )(Δ𝑆) ≡ 𝑊Carnot

Hence the work done in arbitrary cycle is less than the work done in Carnot cycle.
Next,
+
𝑄+ = ∑ 𝑇𝛼+ Δ𝑆𝛼 ≤ 𝑇max ∑ Δ𝑆𝛼 = 𝑄Carnot
𝛼 𝛼

𝑄− = ∑ 𝑇𝛽− Δ𝑆𝛽 ≥ 𝑇min ∑ Δ𝑆𝛽 = 𝑄Carnot

−
𝛽 𝛽

Dividing the two, we have

− −
𝑄− 𝑄Carnot 𝑄 𝑄Carnot
≥ + →1− − ≤1− +
𝑄+ 𝑄Carnot 𝑄+ 𝑄Carnot

Hence
𝜂non-Carnot ≤ 𝜂Carnot
The equality follows when non-Carnot follows when non-Carnot is identical to Carnot.

18