To focus your studying, I’ve identified the following textbook problems as ones that can be reasonably adapted to an exam question
4.14, 4.21 5.8, 5.12, 5.21, 5.23, 5.20 6.17, 6.18, 6.20, 6.22, 6.31, 6.41, 6.42
The additional problems are also exam friendly, and worth reviewing.
A thermodynamic cycle can be represented in particularly useful ways on a P-V diagram, or a T-S plane. These are just two slices through the somewhat higher dimensional state of macroscopic coordinates (S, V, N, T, P, \mu). The P-V plane is useful because total work done on a system W^{on \, sys} = - \int P dV is the area under the curve representing the process. On a T-S plane, we similarly have heat flow Q = \int T dS for any process represented by a curve on the T-S plane.
Only a quasi-static process can be represented by a curve on these planes. A thermodynamic cycle is just a closed contour in the state space (e.g. the P-V plane). And since a cycle returns to its starting point, all of the macroscopic state variables must return to their initial values. Therefore, for a thermodynamic cycle, the change in entropy of the system must be zero, \Delta S_{sys} = 0.
However, the cycle exchanges heat with its surroundings, and we have not accounted for how the entropy of the surroundings have changed. Typically we assume the system of interest is exchanging heat with some number of reservoirs.
A reversible cycle is one in which the net change in entropy of all of the reservoirs is also zero \Delta S_{R} = 0. An irreversible cycle is one in which \Delta S_{R} > 0.
In order to compute an engines efficiency, it is necessary to know the structure of the reservoirs.
The Carnot theorem states that if an engine operates between two reservoirs, the reversible engine has maximum efficiency which is given by the \eta = 1 - T_{c}/T_{h}.
A box of thermal radiation (a “photon gas”) has internal energy and pressure
U = b V T^{4}, \quad P = \frac{U}{3V} = \frac{b}{3} T^{4},
where b is a constant. Note that the pressure depends only on temperature. The gas is taken through a Carnot cycle: isothermal expansion from V_{1} to V_{2} at T_{h}, adiabatic expansion to T_{c}, isothermal compression at T_{c}, and adiabatic compression back to the start.
(a) Sketch the cycle on a P-V diagram. What do the isotherms look like, and how is this different from an ideal gas?
Solution: The important conceptual point of this problem is that a process on the P-V plane depends on the equation of state. For the ideal gas law, the equation of state tells us how pressure and volume are related to the other thermodynamics variables, PV = N k T. But we could easily have a different equation of state, which is what this problem highlights. The photon equation of state is
PV = \frac{1}{3} U So if we had a process at fixed energy, then we would see P \propto 1/V. However, for the photon gas, the energy does not solely depend on temperature. It depends on temperature and volume via U = b V T^{4}. Therefore, a constant temperature process means we fix U/V to be a constant. In that case, P = U/ 3V =constant. Therefore, for this system, an isotherm is a horizontal line, since it also implies pressure is constant.
The adiabatic process is different, but it can be obtained from the first law. Use the fact that the adiabatic process has \delta Q = 0, so that
dU = \delta W^{on \, gas} = - P dV Next, use the formula U = 3 P V to write
dU = 3 P dV + 3 V dP = \delta W^{on\, gas} = - P dV
Collecting terms proportional dV on the RHS, and dP on the RHS gives
3 V dP = - 4 P dV Which can be written
\frac{d P}{P} = - \frac{4}{3} \frac{d V}{V}
Integrating both sides
\int_{P_{0}}^{P_{f}} \frac{dP}{P} = - \frac{4}{3} \int_{V_{0}}^{V_{f}} \frac{d V}{V} Gives
\ln ( P_{f}/P_{0}) = - \frac{4}{3} \ln (V_{f}/V_{0}) Exponentiating then gives
P_{0} V_{0}^{4/3} = P_{f} V_{f}^{4/3}
Which means that for some initial point (P_{0}, V_{0}), the adiabatic curve looks like
P(V) = \frac{P_{0} V_{0}^{4/3}}{V^{4/3}}
An unknown working substance is taken reversibly around a triangle on the T-S plane: isothermal at T_{h} from S_{1} to S_{2}, then a diagonal line to the point (S_{1}, T_{c}) where T_{c} < T_{h}, then straight up at constant entropy back to the start.
(a) Find the heat absorbed, the net work, and the efficiency.
(b) Does your answer depend on what the substance is?
(c) Every step is reversible, but the efficiency is less than 1 - T_{c}/T_{h}. Why is this not a contradiction?
(This will not be tested on)
The important takeaways from this chapter are manipulating thermodynamic potentials to extract thermodynamic variables. The basic incredient is the thermodynamic identity
dU = TdS - P dV + \mu d N as well as the Euler identity
U = TS - PV + \mu N ***
For example, if we want the grand potential, it is
\Phi = U - TS - \mu N Which from the Euler relation implies \Phi = - P V.
The differential of \Phi (pronounced ‘fie’ or ‘fee’) is
d\Phi = dU - d (TS) - d ( \mu N) = \left( TdS - P dV + \mu dN\right) - T dS - S dT - \mu dN - N d \mu
Canceling terms gives
d\Phi = - S dT - P dV- N d\mu
We can read off some partial derivative relations from this, e.g.
\left( \partial \Phi/\partial T\right)_{V, \mu} = - S, \quad \left( \partial \Phi/\partial \mu\right)_{T, \mu} = - N,
The maxwell relations come by applying two derivatives and changing their order
\frac{\partial}{\partial \mu} \frac{\partial}{\partial T} \Phi = \frac{\partial}{\partial T} \frac{\partial}{\partial \mu}\Phi
\frac{\partial}{\partial \mu}\left( - S\right) = \frac{\partial}{\partial T} \left( - N \right)
Which implies
\left( \partial S/\partial \mu\right)_{T} = \left( \partial N/\partial T\right)_{\mu}
The Helmholtz free energy of a photon gas is F(T, V) = -\frac{b}{3} V T^{4}.
(a) Find the entropy, pressure, and internal energy.
Use the thermodynamic identity dF = - S dT - P dV to get
S= - \left(\frac{\partial F}{\partial T}\right)_{V} = \frac{4b}{3} V T^{3} P = - \left( \frac{\partial F}{\partial V}\right)_{T} = \frac{b}{3} T^{4}
And
U = F + TS = - \frac{b}{3} VT^{4} + \frac{4b}{3} V T^{4} = b V T^{4}
(b) Compute the Gibbs free energy G = F + PV. Using G = \mu N, what does this say about the chemical potential of photons? Why is this reasonable?
G = - \frac{b}{3} V T^{4} + \frac{b}{3} V T^{4} = 0 \Rightarrow \mu = 0

In this chapter, we learned about the Boltzmann distribution and the partition function.
Z = \sum_{s} e^{ - \beta E(s)} Where this is a sum over all configurations. For a classical system, we have the configurations labeled by position and momentum, and the sum replaced by a integral
Z = \int \prod_{i = 1}^{N}\frac{d^{D} {\bf r}_{i} d^{D} {\bf p}_{i}}{h^{D}} \exp \left( - \beta \sum_{i} E({\bf r}_{i}, {\bf p}_{i}) \right) E({\bf r}, {\bf p}) is just the total energy of a system, typically kinetic plus potential.
We saw that expectation value of energy are calculating using
\langle E \rangle = \sum_{s} E(s) \frac{e^{ - \beta E(s)}}{Z}
Which leads to a useful identity
U = \langle E \rangle = - \partial_{\beta} \ln Z
This led us to identify the log of the partition function with the Helmholtz free energy
F = - k_{B} T \ln Z
From which we can now get more thermodynamic properties by partial derivative relations from Chapter 6.
A system has two states with energies 0 and \epsilon.
(a) Find the average energy \langle E \rangle and the variance \sigma_{E}^{2} = \langle E^{2} \rangle - \langle E \rangle^{2} at temperature T. Since temperature is fixed, you must use the Boltzmann distribution.
(b) Use the formula:
C = \frac{1}{k T^{2}} \sigma_{E}^{2} To find C in the limits T \to 0 and T \to \infty. Give the physical reason in each limit. What does the relation in (b) say about the sign of C?
Z = 1 + e^{ - \beta \epsilon} So P(0) = \frac{1}{Z}, \quad P(\epsilon) = \frac{e^{ - \beta \epsilon}}{Z} \langle E \rangle = \epsilon P(\epsilon) = \frac{\epsilon e^{ - \beta \epsilon}}{1 + e^{ - \beta \epsilon}} \langle E^{2}\rangle = \sum_{i} \epsilon_{i}^{2} P(\epsilon_{i}) = \epsilon^{2} P(\epsilon) = \frac{\epsilon^{2} e^{ - \beta \epsilon}}{1 + e^{ - \beta \epsilon}} \sigma_{E}^{2} = \langle E^{2}\rangle - \langle E\rangle^{2} = \epsilon^{2} \frac{e^{\beta \epsilon}}{(1 + e^{ \beta \epsilon})^{2}}
C = \partial U/\partial T = - \frac{\epsilon}{(1 + e^{ \beta \epsilon})^{2}} \left( - \frac{\epsilon}{k T^{2}} e^{ \beta \epsilon}\right) = \frac{1}{k T^{2}} \sigma_{E}^{2} C vanishes in both extreme limits. As T \to 0, the system is stuck in its ground state, and cannot absorb heat due to the finite gap to drive the system into the excited state. The levels are discrete.
At high temperature, the the system is equally populating both states. Absorbing more heat will not change this
A particle moves in a two dimensional box of area A with energy E = c |{\bf p}|.
(a) Compute the average energy per particle.
The partition function is
Z(\beta) = \int \frac{d^{2} {\bf r} d^{2} {\bf p}}{h^{2}} e^{ - \beta c | {\bf p}|} In the exponent, we have
\beta \sqrt{ p_{x}^{2} + p_{y}^{2}} = \sqrt{ ( \beta p_{x})^{2} + (\beta p_{y})^{2}}
To eliminate the \beta dependence in the exponent, we just change variable \tilde{p}_{x} = \beta p_{x} and \tilde{p}_{y} = \beta p_{y}. In this case, we get
d^{2} {\bf p} = d p_{x} dp_{y} = \frac{d\tilde{p}_{x}}{\beta} \frac{d \tilde{p}_{y}}{\beta} = \frac{1}{\beta^{2}} d^{2} \tilde{{\bf p}}
Therefore
Z(\beta) = \frac{1}{\beta^{2}} Z(1) Which is sufficient to get the average energy
\langle E \rangle = - \partial_{\beta} \ln Z(\beta) = - \partial_{\beta} \left( - 2 \ln \beta + \ln Z(1)\right) = 2 \beta^{-1} = 2 k_{B} T
(b) A nonrelativistic particle in two dimensions has \langle E \rangle = kT, which is \frac{1}{2} kT per degree of freedom. Your answer to (a) is different. Why does the equipartition theorem not apply?
The equipartition theorem states that given f quadratic degrees of freedom, one has \langle E \rangle = f \frac{1}{2} k_{B} T. However, E(p) = c | {\bf p}| is clearly not quadratic, so we can’t expect the equipartition theorem to hold.