0% found this document useful (0 votes)
19 views7 pages

Thermodynamics Assignment 3 Solutions

Uploaded by

Garlic Bread
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
19 views7 pages

Thermodynamics Assignment 3 Solutions

Uploaded by

Garlic Bread
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

Solution to assignment 3 of Thermodynamics and Statistical Mechanics1

(Due on September 6 2024 at 2pm. Submit your solution to Canvas as a PDF file.)

Problem 1 (10 points). In the class, we have calculated the multiplicity of a magnet made of N two-state
systems, if it has a magnetization M > 0 (namely, the number of up-pointing dipoles minus the number of
down-pointing dipoles is M ). This multiplicity is given by N +M N! !N −M ! . In this problem, we will consider
( 2 )( 2 )
two interacting magnets A and B, which are made of NA two-state systems and NB two-state systems,
respectively, where both NA and NB are large. Denote the magnetizations of A and B by MA and MB ,
respectively. Suppose |MA | ≫ 1 and |MB | ≫ 1, as well as |NA − MA | ≫ 1 and |NB − MB | ≫ 1. Assume
M = MA + MB , the total magnetization of the combined system is fixed, and 1 ≪ |M | < NA + NB . We
will explore how the multiplicity of the combined system behaves as a function of MA .
(1) (2 points) Without making any approximation, show that the multiplicity of the combined system is
Ω =  NA +MA NA!NA −MA   NB +MB N
B!
NB −MB
 . What is the entropy of the combined system?
2 ! 2 ! 2 ! 2 !

(2) (4 points) Applying the Stirling formula (i.e., ln n! ≈ n ln n − n for n ≫ 1) to the entropy, show that the
entropy under this approximation is

NA + M A NA + M A NA − MA NA − MA
S = k NA ln NA − ln − ln
2 2 2 2
! (1)
NB + M B NB + M B NB − MB NB − MB
+NB ln NB − ln − ln
2 2 2 2

Fixing NA , NB and M = MA + MB , show that MA /NA = MB /NB for the most probable macrostate.
Remark: This result makes sense, because it means that in thermal equilibrium the system tends to be
uniform with the same magnetization density everywhere.
(3) (4 points) Denoting the most probable value of MA by MA∗ and δMA = MA − MA∗ . Expand Eq. (1) to
the second order of δMA and show that

k (NA + NB )3
S = S∗ − δMA2 + O(δMA3 ) (2)
2 NA NB (NA + NB + M )(NA + NB − M )

where S ∗ is the entropy for the most probable macrostate. Show that this result implies that near the most
probable macrostate where δMA is small,qthe multiplicity Ω as a function of δMA can be viewed as a normal
2
/σ 2
distribution with a width of the order of NA NB (N −A+N B +M )(NA +NB −M )
(NA +NB )3 . That is, Ω(δMA ) ∝ e−(δMA )
q
B +M )(NA +NB −M )
with σ of the order of NA NB (NA +N (NA +NB )3 .
Remark: Suppose NA and NB are of the same order, and typically M is an extensive quantity of the order
of NA , then the number of different macrostates is of the order of√NA . Therefore, the ratio between the
above width and the number of all macrostates is of the order of 1/ NA , which is quite small for large NA .
Namely, we will almost surely only observe the most probable macrostate.
1 You are welcome to get back with questions and clarifications if the wording of problems is ambiguous.

1
PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

Solution:
(1) In this question, we aim to write the multiplicity of the combined system as a function of MA . We
can firstly pick NA +M
2
A
up-pointing dipoles from NA total dipoles of A magnet, and then pick NB +M2
B

up-pointing dipoles from NB total dipoles of B magnet. Without any approximation about the interaction
between these two magnets, we think this two steps are independent to get the maximal multiplicity Ω:

NA +MA NB +MB
Ω = CNA 2 CNB 2 (3)
NA ! NB !
= NA +MA  NA −MA  NB +MB
 NB −MB  (4)
2 ! 2 ! 2 ! 2 !

Thus, the entropy S is


S = k ln(Ω) (5)
(2) In this question, we apply the Stirling formula to further simplify the entropy we get from the last
question (N ≫ 1):

S =k ln(Ω) (6)
        
NA + MA NA − M A NB + MB NB − M B
=k ln NA ! − ln ! − ln ! + ln NB ! − ln ! − ln !
2 2 2 2
(7)

NA + MA NA + MA NA + MA
=k NA ln NA − NA + NB ln NB − NB − ln +
2 2 2
NA − MA NA − MA NA − M A NB + M B NB + M B
− ln + − ln (8)
2 2 2 2  2
NB + MB NB − M B NB − M B NB − M B
+ − ln +
2 2 2 2
NA + MA NA + MA NA − MA NA − MA
=k NA ln NA − ln − ln
2 2 2 2
! (9)
NB + M B NB + M B NB − M B NB − M B
+ NB ln NB − ln − ln
2 2 2 2

∂S
With the most probable macrostate, the derivative of entropy should be zero, which is ∂MA ∗
= 0.
MA

 
∂S 1 NA + M A 1 1 NA − MA 1 1 NB + M − M A 1 1 NB − M + MA 1
=k − ln − + ln + + ln + − ln −
∂MA 2 2 2 2 2 2 2 2 2 2 2 2
(10)
k (NA − MA )(NB + MB )
= ln (11)
2 (NA + MA )(NB − MB )

2
PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

Thus, for the most probable macrostate, we have


NA − MA∗ NB − MB∗
∗ = (12)
NA + M A NB + MB∗
It is just
NA↑ NB↑
= (13)
NA↓ NB↓
∗ ∗
MA MB
Thus, we have NA = NB .
NA
(3) We perform a Taylor expansion for entropy at the most probable value MA∗ = NA +NB M .

∂S 1 ∂2S
S = S∗ + δMA + δ 2 MA + o(δ 3 MA ) (14)
∂MA ∗
MA 2 ∂MA2 ∗
MA

For the second derivative in Eq. (14),

∂2S
 
k 1 1 1 1
=− + + + (15)
∂MA2 M ∗ 2 NA − M A NA + M A NB + M − M A NB − M + M A ∗
MA
A

(NA + NB )3
=−k (16)
NA NB (NA + NB + M )(NA + NB − M )

∂S
We also have ∂MA ∗
= 0, so
MA

k (NA + NB )3
S =S ∗ − δMA2 + O(δMA3 ) (17)
2 NA NB (NA + NB + M )(NA + NB − M )

Then, the multiplicity Ω as a function of δMA can be viewed as a normal distribution

Ω =eS/k (18)
2 2
∝e−(δMA ) /σ
(19)
q q
2NA NB (NA +NB +M )(NA +NB −M ) NA NB (NA +NB +M )(NA +NB −M )
where σ = (NA +NB )3 , with the order (NA +NB )3 .

Problem 2 (10 points). In the class, we have learnt that the entropy of an ideal gas made of N identical
monatomic particles with total energy U and volume V is
"  3 ! #
V 4πmU 2 5
S = N k ln + (20)
N 3N h2 2

and the multiplicity is

Ω = f (N )V N U 3N/2 (21)

3
PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

Figure 1: Two ideal gases made of the same type of monatomic particles are separated by a plate (represented
by the red vertical line) that allows energy exchange but no particle exchange. The position of the plate can also
move horizontally without encountering any friction.

π 3N/2
where f (N ) = N !h3N (3N/2)!
(2m)3N/2 .
In this problem, we will examine the thermal equilibrium state of two ideal gases made of the same type
of monatomic particle with mass m. These two gases are separated by a plate that allows energy exchange
and can move without friction (see Fig. 1). Suppose the left side contains NA particles, has energy UA and
volume VA , and the right side contains NB particles, has energy UB and volume VB . Suppose there is no
particle exchange between the two containers, so that NA and NB are fixed. Also supposed the total energy
U = UA + UB and total volume V = VA + VB are both fixed.
(1) (2 points) Show that the total entropy of the two gases is
"  3 ! # "   23 ! #
VA 4πmUA 2 5 VB 4πmUB 5
S = NA k ln + + NB k ln + (22)
NA 3NA h2 2 NB 3NB h2 2

(2) (4 points) Take the derivatives of the total multiplicity with respect to UA and VA to show that UA /NA =
UB /NB and VA /NA = VB /NB in the most probable macrostate. Use the equipartition theorem and ideal
gas law to show that the temperatures on the two sides of the plate are the same, and the pressures on the
two sides of the plate are also the same.
Remark: Again, this result makes sense because it means that in thermal equilibrium the system tends to
be uniform.
(3) (4 points) For the most probable macrostate, calculate the second order derivatives of Eq. (22), the
∂2S ∂2S ∂2S
total entropy of the two gases, with respect to UA and VA , i.e., ∂U 2 , ∂U ∂V
A A
and ∂V 2 . You should find, for
A A
the most probable macrostate,
∂2S 3k(NA + NB )3 ∂2S ∂2S k(NA + NB )3
2 =− , = 0, 2 =− . (23)
∂UA 2NA NB U 2 ∂UA ∂VA ∂VA NA NB V 2
Show that these results imply that near the most probable macrostate, q the multiplicity Ω as a function of
NA NB U 2
UA and VA is a two-variable Gaussian function with a width of the order (N A +NB )
3 along the UA direction

(U −U ∗ )2 (V −V ∗ )2
− A 2A − A 2A
q
NA NB V 2
(NA +NB )3 along the VA direction. That is, Ω ∝ e
σ σ
and a width of the order U V , where
∗ ∗
U
qA and VA are the
q most probable values of UA and V A , respectively, and σ U and σ V are of the order of
NA NB U 2 NA NB V 2
(NA +NB )3 and (NA +NB )3 , respectively.
Remark: Again, consider the case where NA and NB are of the same order, typically the energy U and
volume V are extensive quantities, i.e., they are also of the same order of NA and NB . So the total number

4
PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

of all macrostates will be proportional to U V . Therefore, the above widths suggest that the fluctuations
around the most probable macrostates are concentrating in a narrow window. Namely, we will almost surely
only observe the most probable macrostate.
Solution:
(1) The total multiplicity Ω = ΩA ΩB , so the total entropy S = SA + SB
"  3 ! # "  3 ! #
VA 4πmUA 2 5 VB 4πmUB 2 5
S = NA k ln + + NB k ln + (24)
NA 3NA h2 2 NB 3NB h2 2

∂S ∂S
(2) For the most probable macrostate, we have ∂UA = 0 and ∂VA =0

∂S 3 3
= NA k − NB k (25)
∂UA 2UA 2(U − UA )

∂S 1 1
= NA k − NB k (26)
∂VA VA V − VA
Thus, we have UA /NA = UB /NB and VA /NA = VB /NB in the most probable macrostate. With U = 23 N kT ,
we have TA = TB . With the ideal gas law P V = N kT , we have PA = PB .
(3) In this question, we also perform a Taylor expansion for the entropy at the most probable macrostate.
!
∗ ∂S ∂S 1 ∂2S 2 ∂2S ∂2S
S=S + δUA + δVA + δ UA + 2 δUA δVA + δ VA +o3
2
∂UA U ∗ ∂VA V ∗ 2 ∂UA2 U ∗ ∂UA ∂VA U ∗ ,V ∗ ∂VA2 V ∗
A A A A A A
(27)
For the second derivatives, we have
∂2S
 
3k NA NB
=− + (28)
∂UA2 2 UA2 (U − UA )2

∂2S 3k (NA + NB )3
=− (29)
∂UA2 ∗
UA 2 NA NB U 2

∂2S
 
NA NB
= −k + (30)
∂VA2 VA2 (V − VA )2
∂2S (NA + NB )3
= −k (31)
∂VA2 VA∗ NA NB V 2

∂2S
=0 (32)
∂UA ∂VA
Then, near the most probable macrostate, the multiplicity Ω behaves like a two-variable Gaussian function
δUA2 δVA2
− −
σ2 σ2
Ω∝e U V (33)
q q
NA NB U 2 NA NB V 2
, where σU and σV are of the order of (NA +NB )3 and (NA +NB )3 , respectively.

5
PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

Figure 2: Initially, two ideal gases made of two different types of monatomic particles are separated by a plate
(represented by the red verticle line) that allows neither energy exchange nor particle exchange. There is also
no energy or particle exchange between the gases and the environment outside. After some time, the plate is
removed.

Problem 3 (10 points). In the class, we have mostly been discussing ideal gases made of the same type of
particles. In this problem, we consider ideal gases made of different types of monatomic particles A and B
with mass mA and mB , respectively (for example, A can be helium and B can be argon). Suppose the two
ideal gases are initially separated by a plate that does not allow either energy exchange or particle exchange
(see Fig. 2). There is also no energy or particle exchange between the gases and the environment outside
through the walls of the container. On the left side of the plate, the number of particles is NA , the energy
is UA and the volume is VA . On the right side of the plate, the number of particles is NB , the energy is UB
and the volume is VB . The two gases are in thermal equilibrium initially, and then the plate is removed.
(1) (2 points) According to the equipartition theorem, what are the temperatures of the two ideal gases
before the plate is removed? After the plate is removed and the system reaches thermal equilibrium again,
what is the final temperature according to the equipartition theorem? What is UA′ , the final energy of
particle type A after thermal equilibrium is reached after the plate is removed?
(2) (4 points) Use the result of part (1) to answer the following questions. What is the total entropy of the
combined system before and after the plate is removed? What is their difference, which is also known as
the entropy of mixing?
(3) (4 points) In the previous parts, we have used the fact that the entire system will share a single final
temperature after thermal equilibrium. In this part, we do not assume this, and we will derive it by using the
fundamental assumption of statistical mechanics and the equipartition theorem. Suppose after the plate
is removed, there is a macrostate where the energy of type-A particles is UA′ , and the energy of type-B
particles is UB′ . From the total multiplicity of the entire system, calculate the most probable value of UA′
and verify that it is identical to the answer obtained in part (1).
Solution: (1) Before we remove the plate, according to the equipartition theorem U = 32 N kT , we have

2UA
TA = (34)
3NA k
2UB
TB = (35)
3NB k
After the plate is removed, and the thermal equilibrium is reached

2(UA + UB )
T = (36)
3(NA + NB )k

6
PC2135 Lecturer: Liujun Zou TA: Jiulin Tang

The final energy of type A particle UA′ is

3
UA′ = NA kT (37)
2
NA
= (UA + UB ) (38)
NA + NB

(2) For the total entropy of the combined system before (S) and after (S ′ ) the plate is removed, we have
"  3 ! # "  3 ! #
VA 4πmA UA 2 5 VB 4πmB UB 2 5
S = NA k ln + + NB k ln + (39)
NA 3NA h2 2 NB 3NB h2 2
"  32 ! # "  32 ! #
4πmA UA′ 4πmB UB′
 
′ VA + VB 5 VA + VB 5
S = NA k ln + + NB k ln + (40)
NA 3NA h2 2 NB 3NB h2 2
The entropy of mixing ∆S is

∆S = S ′ − S (41)
  32 !   32 !
VA + VB (UA + UB )NA VA + VB (UA + UB )NB
= NA k ln + NB k ln (42)
VA UA (NA + NB ) VB UB (NA + NB )


∂S
(3) After the plate is removed, the most probable macrostate is featured by ∂U ′ = 0, which is similar to
A
′ ′
question 2. Thus, we also have UA /UB = NA /NB . Then, by using the equipartition theorem, we get the
thermal equilibrium TA′ = TB′ and UA′ = NAN+N
A
B
(UA + UB ).

Common questions

Powered by AI

Applying Stirling’s approximation to the entropy formula results in a simplification where ln(n!) is approximated as n ln(n) − n, enabling a more tractable entropy expression: S = k (NA ln NA − NA + (MA/2) ln(NA+MA/2) − (NA−MA/2) ln(NA−MA/2) + NB ln NB − NB + (MB/2) ln(NB+MB/2) − (NB−MB/2) ln(NB−MB/2)). This approximation facilitates analytical solutions to complex statistical systems and validates equilibrium properties, showing uniform magnetization density .

The total entropy of two ideal gases at equilibrium is the sum S = NAk[ln((VA/NA)(4πmUA/3NAh^2)^3/2)+5/2] + NBk[ln((VB/NB)(4πmUB/3NBh^2)^3/2)+5/2]. At equilibrium, conditions ∂S/∂UA = 0 and ∂S/∂VA = 0 imply UA/NA = UB/NB and VA/NA = VB/NB, leading to equality in temperature (TA = TB) and pressure (PA = PB). This reflects the principle that the distribution of energy and spatial arrangement of particles maximize entropy under equilibrium constraints .

For the system to tend to a uniform magnetization density in thermal equilibrium, the condition MA/NA = MB/NB must hold, indicating the most probable macrostate. This entails the same magnetization density across all parts of the system, achieved by equating the ratios of magnetization to number of dipoles in each subsystem. This equilibrates the per-site magnetizations and ensures that entropy maximization and energy equivalence guide the system to a uniform state .

The removal of a barrier between different types of gas particles leads to an increase in the system's entropy, reflecting the greater dispersal of particles and phase space accessibility. As the gases mix, energy and volume considerations drive the system towards a new equilibrium where entropy is maximized due to enhanced randomness. The statistical mechanics principles show that the entropy change, the entropy of mixing, quantifies the increased disorder as systems equilibrate energetically and spatially without restrictions .

The entropy of mixing ΔS when two gases made of different particle types mix is ΔS = S′ - S = NAk ln((VA + VB)/VA)((UA + UB)NA/(UA(NA + NB)))^3/2 + NBk ln((VA + VB)/VB)((UA + UB)NB)/(UB(NA + NB)))^3/2. This reflects increased randomness as the gases intermingle after removing separation barriers, which leads to a distribution of energy and space maximizing entropy beyond the initial compartmentalization .

The multiplicity of two interacting magnets behaves as a function of the magnetization of one system by considering the independent choices of up-pointing dipoles from each magnet. Specifically, the multiplicity Ω = (NA!/[(NA+MA)/2]![NA−MA/2]!) × (NB!/[(NB+MB)/2]![NB−MB/2]!) allows us to compute the entropy S = k ln(Ω). Applying Stirling's approximation simplifies the entropy expression to reflect the influence of each magnet's magnetization. This results in showing that in thermal equilibrium the system tends to exhibit uniform magnetization density, inferred from MA/NA = MB/NB .

Using a normal distribution to represent multiplicity near the most probable macrostate in magnet systems is justified by the statistical mechanics principle that macrostates close to equilibrium can be described by Gaussian functions. Near equilibrium, the probability of observing a macrostate is sharply peaked, making deviations quadratic in entropy changes, which implies a normal distribution for variations in magnetization δMA. This reflects how systems near equilibrium show minor fluctuations around their most probable macrostate due to constrained energy variance .

The equipartition theorem plays a crucial role in determining the temperatures of the gases before and after mixing. Initially, it gives TA = (2UA)/(3NAk) and TB = (2UB)/(3NBk). After the gases are allowed to mix, thermal equilibrium ensures a shared final temperature of T = (2(UA + UB))/(3(NA + NB)k). This highlights how energy redistribution leads to a single equilibrium temperature reflecting balanced kinetic energy per particle across the mixed system .

The normal distribution approximation of multiplicity near the most probable macrostate implies that the multiplicity Ω, as a function of δMA, can be viewed as a normal distribution with the form Ω(δMA) ∝ e^-(δMA)^2/σ^2, where σ represents the standard deviation. This suggests that near the equilibrium state, deviations in magnetization yield a Gaussian profile of multiplicity centered around the most probable macrostate. The width σ shows that fluctuations are small relative to the system size, ensuring that only small deviations from equilibrium are probable .

The mathematical derivation for energy and volume ratios that maximize entropy in ideal gas systems involves setting the derivatives of entropy with respect to UA and VA to zero: ∂S/∂UA = NAk(3/2UA) - NBk(3/2(U-UA)) = 0, and ∂S/∂VA = NAk(1/VA) - NBk(1/(V-VA)) = 0. Solving these reveals the conditions for equilibrium: UA/NA = UB/NB and VA/NA = VB/NB, which, along with the ideal gas law, ensure that temperature (TA = TB) and pressure (PA = PB) equalize, maximizing entropy .

You might also like