Statistical Physics A

winter term 2023-24

Prof. Krzysztof Meissner

Institute of Theoretical Physics, Faculty of Physics, University of Warsaw
Krzysztof.Meissner@fuw.edu.pl

Tutors:
dr Marta Waclawczyk, Marta.Waclawczyk@fuw.edu.pl, Wednesday 08:15-11:00
dr Jeffrey Everts, Jeffrey.Everts@fuw.edu.pl, Wednesday 14:15-17:00

Dates of colloquia and exams:

Colloquium I, 27th November 2023, room B0.14

Colloquium II, 22nd January 2024, room 1.40

Written exam, 6th February 2024, room 0.06

Tutorial
1 Week I
1. Equation of state - From an experiment it is known that
   
∂p 1 V
= − RT f ,
∂V T n n
 
∂p nR
= − 2nRT a,
∂T V V
where n - number of moles, a - a constant, f (V /n) - an arbitrary function.
Find the equation of state.
2. Exact and inexact differentials
(a) Show that dF ¯ (x, y) = xdx + xdy is inexact differential because the
integral from (0, 0) to (1, 1) depends on a path. Consider two paths
A: (0, 0) − (1, 0) − (1, 1) and B: (0, 0) − (0, 1) − (1, 1).
(b) Derive the condition for the differential dF (x, y) to be exact.
(c) Introduce an integrating factor dG(x, y) = r(x, y)dF
¯ (x, y). Consider
dF
¯ (x, y) = ydx + dy and find its integrating factor.

1
 3. Thermodynamic potentials and Maxwell relations - For a gas described by
the fundamental relation dU = T dS − pdV derive corresponding thermo-
dynamic potentials: internal energy U (S, V ), entalpy H(S, p), Helmholtz
free energy F (T, V ), and Gibbs free energy G(T, p) by choosing proper in-
dependent variables. Derive all possible Maxwell relations. Discuss equiv-
alence and usefulness of different thermodynamic potentials in description
a system.
4. Engine efficiency - Find efficiency of an engine working in
(a) the Carnot cycle,
(b) the Otto cycle with an ideal gas as the working substance.
Remind students the heat capacity and specific heat at constant V and p.

1.1 Homework
AS 3
1. A gas is described by the fundamental relation U = nV , where A = const.
Find the equation of state p(T, V, n).
p
Answer: p = nT 3 /27AV
RT
2. A gas fulfills the van der Waals equation of state: p(T, v) = v−b − va2 ,
where a, b, R are positive constants, and v = V /n. Demonstrate that the
specific heat cv of this gas does not depend on volume V .
3. Evaluate cp − cv for a gas fulfilling the Clapeyron equation of state p =
nRT /V .
4. The internal energy of a certain one-component gas involving n = 2 moles
fulfills the equation U = ApV 2 , where A is a known constant. Find
U (p, V, n) for any n.
Answer: U (p, V, n) = 2ApV 2 /n

2
2 Week II
1. Show that      
∂p ∂T ∂V
· · = −1
∂T V ∂V p ∂p T

´∞ 2
2. Gaussian integrals - Compute the integral I = −∞ e−αx dx by first com-
puting I 2 using polar coordinates.
´∞ By taking a derivative of I over α
2
determine the integral J = −∞ x2 e−αx dx.
´∞ 2
3. Gamma function - Gaussian integrals in the form Im = 2 0 xm e−αx dx,
with m > −1, can be formulated in terms of the gamma function Γ(z) =
´ ∞ z−1
0
t e−t dt. Show the recursion relation Γ(n + 1) = nΓ(n) for n > 0.
Compute explicitly Γ(1/2), Γ(l + 1/2), Γ(1), Γ(l + 1), for l = 1, 2, 3, ....
4. Ball in n-dimensions - Find the volume of a ball and the area of a sphere
with radius r in n dimensions. Express the results in terms of the Gamma
function (generalized factorial function).

5. Find the Legendre transform of the function g(x) = exp(x − 1).

Internal energy can be expressed in terms of its natural variables S and
∂U
V : U = U (S, V ). Find the Legendre transform Ũ (S, ∂V ). Which thermo-
dynamic potential does it define?

2.1 Homework
1. n-dimensional ball - Volume of n-dimensional ball is expressed by the
formula
π n/2 rn
Vn =
Γ(n/2 + 1)
(n > 0 is a natural number). Prove that the non-dimensional parameter
Vn /rn attains its maximum value for n = 5.
2. Gamma function - Express the following integral in terms of the Gamma
function ˆ 1  3
2 1
I= dxx ln .
0 x

Answer: Γ(4)/81
3. Find the Legendre transform of f (x) = x2 + x − 3.

4. Find the Legendre transform of the internal energy: Ũ ( ∂U

∂S , V ). Identify
which thermodynamic potential is defined in this way.

3
3 Week III
1. A gas expanded adiabatically from state 1, in which it had volume V1
and pressure p1 , to volume V2 and pressure p2 (state 2) passing on the
way through a sequence of states satisfying the relation pV γ = const with
some exponent γ > 1. The gas has been next isobarically and reversibly
compressed to the volume V1 (state 3). From the state 3 the gas has been
finally brought back to the initial state 1 by enclosing it adiathermally
and performing on it isochorically a work W (e.g. by placing in it a Joule
paddle-wheel device or by passing a controlled current through a resistor
immersed in it) which has been measured. How much heat has the gas
absorbed in the first two processes? Would it be possible to bring the gas to
the state 2 by performing on it some work while it remained adiathermally
isolated, if it were first isobarically (at the pressure p1 ) and reversibly
expanded to the volume V2 ?
2. A very large mass M of porous hot rock is to be utilized to generate
electricity by injecting water and utilizing the resulting hot steam to drive
a turbine. As a result of heat extraction, the temperature of the rock
drops, according to d̄Qh = −M CdTh where C is the specific heat of the
rock, assumed to be temperature independent. If the plant operates at the
Carnot limit, calculate the total amount W of electrical energy extractable
from the rock, if the temperature of the rock was initially Th = Ti , and
if the plant is to be shut down when the temperature has dropped to
Th = Tf . Assume that the lower reservoir temperature Tl stays constant.
3. Magnetic work - Derive the expression for work in magnetic systems

dW
¯ = µ0 HdM.

4. Electric work - Derive the expression for work in electric systems

dW
¯ = EdP.

5. (a) Let f be a homogeneous function of degree 1. That is

f (u1 , ..., un ) = λf (x1 , ..., xn ),

for all λ ∈ R and ui = λxi . Prove Euler’s theorem
n  
X ∂f
f (x1 , ..., xn ) = xi .
i=1
∂xi {xj }̸=xi

(b) Consider the internal energy U (S, V, N ) with dU = T dS−pdV +µdN .

Use the result in (a) to find the Euler form of the internal energy

U (S, V, N ) = T S − pV + µN.

4
 (c) Derive the Gibbs-Duhem relation

dµ = −sdT + vdp,

where s = S/N and v = V /N .

3.1 Homework
1. Using the First Law of Thermodynamics calculate the heat absorbed by an
ideal gas (with the equation of state pV = nRT and with cv = const) going
through a process from an equilibrium state 1 characterized by the pressure
p1 and volume V1 to the equilibrium state 2, characterized by p2 and V2 ,
where V2 < V1 and p2 > p1 . The gas was first heated isochorically (V =
const) and next isothermally (at T = const) and reversibly compressed.
Moreover, p1 V1γ = p2 V2γ , where γ = 5/3, that is the initial and final states
are on the same adiabath.
Answer: Q = 3/2nR(T2 − T1 ) − nRT2 ln(V1 /V2 )
2. Consider a magnetic system subject to a magnetic field H. Its fundamental
thermodynamic relation is given by

dU = T dS + µ0 HdM, (1)

where µ0 is the vacuum magnetic permeability and M is the total magnetic

moment of such a system. Derive the respective thermodynamic potentials
for such a system, as well as the heat capacities at constant magnetic
moment and constant magnetic field.
3. Consider an electric system subject to an electric field E. ts fundamental
thermodynamic relation is given by

dU = T dS + EdP, (2)

where P is the dipole moment. Redo problem 2 for such a system.

4. A magnetic compound behaves according to the Curie law, m = CH/T ,
where C is a constant, H is the applied magnetic field, m is the magne-
tization per particle, and T is temperature. In a quasi-static process, we
have
du = T ds + Hdm, (3)
where u = u(s, m) plays the role of an internal energy. For an infinitesimal
adiabatic process, show that we can write
CH
∆T = ∆H, (4)
cH T
where cH is the specific heat at constant magnetic field.

5
4 Week IV
1. Entropy change - A system with n mole of water at temperature T1 is in
thermal contact with a very large environment at temperature T2 . Find
changes of entropy: a) in the water, b) in the environment, and c) in the
whole system, when the temperature of both gets equal. Mole specific
heat is given cV . Discuss the importance of the assumption that the
environment must be infinitely large.
2. Heat capacity - Show that the heat capacities are expressed by the two
state functions    2 
∂S ∂ F
CV = T = −T
∂T V ∂T 2 V
and
∂2G
   
∂S
Cp = T = −T .
∂T p ∂T 2 p

3. Thermodynamics of a rubber band - Consider a rubber band of length L

at tension f . The total differential of the internal energy U is given by
dU = T dS +f dL+µdn. As usual, here T is temperature, S is the entropy,
µ is the chemical potential, and n is the mole number of the rubber band.
(a) An equation of state for the rubber band is given by either
"   #
2
1/2 1 L L0 3
S = L0 γ(θU/L0 ) − L0 γ + −
2 L0 L 2

or "   #
2
θnU/L0 1 L L0 3
S = L0 γe − L0 γ + − ,
2 L0 L 2
with L0 = nl0 . Here l0 > 0, θ > 0, and γ are constants. Explicitly
demonstrate which one of these entropy functions is physical. Can γ
be negative? For the proper choice of S determine the equation of
state for the tension f (T, l) with l = L/n.
(b) Compute for the correct equation of state (∂cl /∂l)T with cl being the
heat capacity per mole at constant length.
(c) Derive the analogue of the Gibbs-Duhem relation (see week III prob-
lem 5c) for a rubber band.
(d) Suppose you have two pieces of rubber band satisfying the equation
of state in (a). Each of these rubber bands is characterized by a
temperature T (α) , length per mole l(α) , and mole number n(α) with
α = 1, 2. Determine as a function of these state variables the final
energies and temperatures of the two rubber bands if they are held at
constant length and placed in thermal contact with one another. You
may neglect thermal convection and mass flow to the surroundings.

6
 4. Thermodynamic inequalities.
(a) Let Φ stand for a general thermodynamic potential, which is a natural
function of the extensive variables X1 , X2 , ..., Xr and the intensive
variables Ir+1 , ..., In . The total differential is given by
r
X n
X
dΦ = Ii dXi − Xj dIj .
i=1 j=r+1

Note that a pair Xi and Ii are thermodynamically conjugate with

each other. Prove the thermodynamic stability criteria
 
∂Ii
0≤ .
∂Xi X1 ,...,Xi−1 ,Xi+1 ,...,Xr ,Ir+1 ,...,In

(b) An experimentalist claims to have found that a certain gaseous ma-

terial obeys the conditions (i) (∂p/∂v)T < 0, (ii) (∂p/∂T )v > 0,
(iii) (∂µ/∂v)T < 0, (iv) (∂T /∂v)s > 0. Here v = V /n, and s = S/n.
Which one of these inequalities is determined by thermodynamic sta-
bility? Which pairs of inequalities are inconsistent with each other?

4.1 Homework
1. Entropy change in an ideal gas - Find the entropy change of an n mole of
ideal gas in a reversible process from a state A (pA , VA , TA ) to a state B
(pB ,VB ,TB ). Specific heats cV and cp are given.
2. Response functions - Show that

κT (Cp − CV ) = T V α2

and
Cp (κT − κS ) = T V α2 .
Here the compressibility is given by κx = −(1/V )(∂V /∂p)x and the coef-
ficient of thermal expansion is α = −(1/V )(∂V /∂T )p .
3. Thermodynamic stability of rubber bands and Gibbs-Duhem - One can
experimentally verify quite easily that a rubber band heats up when it is
stretched adiabatically.
(a) Given this fact, will a rubber band contract or expand when it is
cooled at constant tension?
(b) The same amount of heat flows into two identical rubber bands, but
one is held at constant tension, and the other one at constant length.
Which one has the largest increase in temperature?

7
 (c) Suppose the rubber band satisfies the equation of state U = θS 2 L/n2 ,
with θ being a constant. Determine the chemical potential of the
system µ(T, l) and show explicitly that the equation of state satisfies
the relevant Gibbs-Duhem relation from the tutorials.

4. Stability of magnetic systems - Consider a magnet with magnetization M

in the direction of the magnetic field strength H. In appropropriate units
the work is given by HdM .
(a) Determine the signs of the isothermal and adiabatic susceptibilities,
   
∂M ∂M
χT = , χS = .
∂H T,n,V ∂H S,n,V

Furthermore, determine the sign of χT − χS .

(b) Suppose a paramagnetic substance has a decreasing magnetization as
function of temperate at constant H. Will the temperature increase
or decrease when the material is adiabatically demagnetized?

5. Generalised Euler theorem - Given the generalised homogeneous function

f (λθ1 x1 , ..., λθn xn ) = λf (x1 , ..., xn ),

with θi a real number for i = 1, ..., n. Show that

   
∂f ∂f
θ 1 x1 + ... + θn xn = f (x1 , ..., xn ).
∂x1 x2 ,...,xn ∂xn x1 ,...,xn−1

This generalisation of Euler’s theorem plays an important role in the the-

ory of phase transitions close to the critical point.

8
5 Week V
1. Using the second law of thermodynamics in the form dQ ¯ = T dS and the
fact that the internal energy is a function of state show that

∂(T, S)
= 1.
∂(p, V )

2. Derive a formula for the internal energy and the entropy of the van der
Waals gas. The equation of state of the van der Waals gas reads
RT a
p(v, T ) = − 2,
v−b v
where a and b are constants. The specific heat at constant volume cV of
the van der Waals gas is given.

3. Consider an adiathermally isolated gas in a cylinder divided by a wall into

two parts. The gas initially occupies only one part of the cyllinder. After
removing the wall the gas expands by the volume ∆V . Which coefficients
characterize infinitesimal changes of the temperature and the entropy in
this process? Express these coefficients as functions of measurable quanti-
ties (heat capacities, pressure, etc...) Calculate the final temperature and
the change of entropy, if the gas had initially the temperature T1 and the
occupied the volume V1 . Consider a) an ideal gas, b) a van der Waals gas.
4. Find an expression for the internal energy per unit volume u(T ) of a gas,
if it is known that the pressure of this gas depend only on temperature
and that 3p = u(T ).

5. The Helmholtz free energy density f = F/V of a one-component system

of number density ϱ and temperature T is given within the van der Waals
approximation by

ϱΛ3
   
f = ϱkB T ln − 1 − Aϱ2 ,
1 − Bϱ
√
where A and B are positive constants and Λ = h/ 2πmkB T .
• Give the dimension and physical meaning of A and B.
• Calculate the pressure p and the chemical potential µ from f .
• Calculate the critical density ϱc and the critical temperature Tc . Is
it crucial that A, B > 0?
• Calculate the critical pressure pc and show that pc /(ϱc kTc ) = 3/8
independent from A, B.

9
 • Show that the van der Waals equation of state can be written in the
form  
T p ϱ
h , , = 0,
Tc pc ϱc
where the function h is independent of A and B. This is the law of
corresponding states.
• Sketch p(ϱ) and µ(ϱ) for T > Tc , T = Tc and T < Tc . Describe how
the condition for gas-liquid equilibrium at T < Tc can be determined
from p(ϱ, T ) and µ(ϱ, T ).

5.1 Homework
1. Prove the following properties of Jacobians
a)  
∂(u, v) ∂u
=
∂(x, v) ∂x v

b)
∂(u, v) ∂(u, v) ∂(a, b)
=
∂(x, y) ∂(a, b) ∂(x, y)

2. Calculate the work performed by n moles of the van der Waals gas during
isothermal expansion from V1 to 10V1 . The initial pressure of the gas was
equal to p1 .

3. Derive the entropy of the photon gas which follows the equation of state
p = αT 4 /3 and for which the internal energy U = αT 4 V . Express the work
done and heat exchanged with the surroundings when the gas undergoes
an isothermal compression from V1 to V2 . Derive a relationship between
the pressure and the volume during adiabatic expansion of the photon gas.

4. Calculate the chemical potential of the van der Waals gas a) by using the
expression for the Gibbs function per mole: g = u − T s + pv = µ b) by
integrating the Gibbs-Duhem relation dµ = −sdT + vdp. Show that the
two results are identical.

10
6 Week VI
1. Probability theory - A stone is falling down from a shelf at a height h. This
event is registered by a photo camera at discrete and random times. On
each photograph we measure the distance, which the stone has transversed.
Find a probability distribution function for these distances. Compute the
average distance ⟨x⟩ and the standard deviation σx . What is the proba-
bility that the stone was registered within a distance [⟨x⟩ − σx , ⟨x⟩ + σx ]?
(Using this example remind yourselves of: continuous random variables
and their probability distributions, averages and other moments of the
probability distribution).
2. Particles in a box - Consider N balls, each of which can be placed in the
left or the right half of a container. In the case of N = 4 write down
all possible microstates. Let the macrostate be defined by the number of
balls in the left half. What are the possible macrostates in this case? How
many microstates can realize a given macrostate? What is the probability
to find a given macrostate? In the case of N = 10 find all the possible
macrostates and probabilities of their occurrence. Which macrostate is
the most probable?
3. Stirling
√ approximation - Show that for n ≫ 1 we can approximate n! ≈
2
2πn nn e−n+1/(12n)+O(1/n ) . Write this approximation for ln n!. Discuss
the role of leading terms, 1/n corrections, and compare with the asymp-
totic form of the Gamma function.
4. Two-level system - It is a very popular model describing localized spins
1/2, photons with polarization or atoms in certain states. For spins 1/2
it is visualized that each particle has a magnetic moment pointing up
or down. Write down all possible microstates in the case of N = 2,
3, and 4 particles. For the spin system the natural variables describ-
ing different macrostates are the total number of particles N and the
”magnetization” M = N+ − N− , where N± is the number of moments
pointing up and down, respectively. Derive a formula for the number
of microstates Ω(N, M ), realizing a givenPmacrostate with N and M .
Find the total number of states Ω(N ) = M Ω(N, M ). In the limit of
2
large N showp that Ω(N, M ) ≈ Ω(N, 0)e−M /2N is Gaussian. Show that
Ω(N, 0) ≈ 2/πN 2N and estimate its value for N = 100. What are the
most probable macrostates?

6.1 Homework
1. Stirling’s formula - Use the Stirling’s formula to estimate the term:

1 · 3 · 5 · . . . · (2n + 1)

Use the Stirling’s formula to estimate the term:

11
 2 · 4 · 6 · . . . · (2n)

Use the two to calculate

∞
Y 2n 2n
n=1
2n − 1 2n + 1

2. Probability theory - We roll two dice. What is the probability that the
sum of the results is six? What is the average sum of the results and its
variance? What is the probability distribution function?
3. Distribution of particles in a container - Consider N identical, distin-
guishable particles occupying k cells. It is a model of a discrete space
with particles in it. Microstates are given by the distribution of particles
in each cell. Macrostates are characterized by providing numbers of par-
ticles ni in each cell i. Derive the formula for the number of microstates
for a given macro-state Ω(nP1 , n2 , ..., nk ) = N !/n1 !n2 ! . . . nk !. With a fixed
number of particles N = i ni show that the most probable distribution
of particles is such that in each cell there is ni = N/k of them. In other
words, the distribution is uniform.

4. Microcanonical ensemble - equilibrium conditions - An isolated system of

the energy U , volume V , and number of particles N (microcanonical en-
semble) is split with a partition that allows to exchange the energy and
particles and change the volumes of subsystems, i.e. Ui , Vi and Ni , with
i = 1, 2, are random variables which are constrained U1 + U2 = U ,
V1 + V2 = V and N1 + N2 = N . Let Ωi (Ui , Vi , Ni ) be the number of
microstates corresponding to a given macrostate (Ui , Vi , Ni ) of each sub-
system. Compute the total number of microstates for a given macrostate
of the whole system. Maximizing the corresponding probability, find the
equilibrium conditions for those subsystems.

12
7 Week VII
1. Connection with thermodynamics in the classical microcanonical ensemble.
Consider a closed system with fixed energy U , volume V , and number of
classical particles N . This is the microcanonical ensemble.
(a) Denote a point in phase space by Γ = (pN , rN ), with pN = (p1 , ..., pN )
and rN = (r1 , ..., rN ). What is the probability distribution fm (Γ) in
the microcanonical ensemble? Express your answer in terms of the
quantity ˆ
ω(U, V, N ) = dΓ δ(U − H(Γ)).

Can one derive this result for fm (Γ)? Give an explicit equation for
the ensemble average ⟨...⟩ in the microcanonical case.
(b) Suppose we change the volume of a microcanonical system from V
at t = t0 to V + dV at t = t0 + τ . Show that for an ergodic system
p = −⟨∂H/∂V ⟩.
(c) Introduce the so-called energy sphere
ˆ
Ω̃(U, V, N ) = dΓ Θ(U − H(Γ)),

with Θ the Heaviside step function. Derive expressions at fixed N

for ! !
∂ Ω̃(U, V, N ) ∂ Ω̃(U, V, N )
, ,
∂U ∂V
V,N U,N

Express your answers in terms of ω(E, V, N ).

(d) Use the first law of thermodynamics to show that
!−1
∂ Ω̃
T dS = dΩ̃.
∂U

Conclude that S = g(Ω̃) with a function g to be determined.

(e) Use extensivity to show that g(Ω) = k log[Ω̃(U, V, N )/(CN !)]. What
is the dimension of the constant C?
(f) Prove from the thermodynamic defintion of the temperature that
Ω = kT ω. Argue that in the thermodynamic limit (give a definition!),
we find the following equivalent expressions for the entropy
   
ω(U, V, N )kT ω(U, V, N )δU
S = k log = k log ,
CN ! CN !

with δU an arbitrary constant with dimension of energy.

13
 2. Simple classical harmonic oscillator. Consider the Hamiltonian of a simple
classical harmonic oscillator,
p2 1
H(p, q) = + kq 2 .
2m 2
Here m is the mass of the particle, k is the spring constant. Further-
more, p and q denote the canonical momentum and generalised position,
respectively.
(a) What is the phase space of the simple harmonic oscillator? How do
we call a subset Ω(U ) of the phase space fixed at constant energy U ?
(b) Write down the equations of motion from the Hamiltonian. Suppose
we fix the initial values p(t = 0) = p0 and q(t = 0) = q0 . Determine
q(t) and p(t) and give an explicit expression for the total energy U
for this specific phase space trajectory.
(c) Consider a phase space trajectory (p(t), q(t)) in Ω(U ). Show explic-
itly that ⟨p2 ⟩ = p2 (t), where ⟨...⟩ is the microcanonical ensemble
average and the overline denotes a time average over a period T of
(p(t), q(t)). What can you say about the ergodicity of the problem?
3. Number of microstates for a classical ideal gas in microcanonical ensem-
ble - For a classical ideal gas in three dimensions find Ω̃(U, V, N ) and
ω(U, V, N ). Find the corresponding expression for the entropy in case of
large N and show that the three definitions of the previous exercise are
equivalent in the thermodynamic limit. Give expressions for the pressure,
internal energy, and the chemical potential. Conclude that the constant k
from the previous exercise must be the Boltzmann constant.
4. The classical limit of a quantum ideal gas. Consider a particle in a cubic
box of volume V = L3 with energy ϵk = ℏ2 k2 /(2m). When we impose
closed boundary conditions, the wavevector k = kn is quantized and sat-
isfies ki = ni π/L with ni = 1, 2, 3, ... and i = x, y, z.
(a) What is the microcanonical partitition function in the case of discrete
energy levels? What is the probability distribution function? What
is the entropy in this case?
(b) In the classical limit we can neglect the discreteness of the energy
levels and we can take the continuum limit. Write down in this limit
an expression for the entropy.
(c) Identify the constant C from problem 2.

7.1 Homework
1. Harmonic oscillators and the Liouville equation.
(a) Argue that the Liouville equation is equivalent to the continuity equa-
tion for the probability density function ρ(pN , qN , t).

14
 (b) Consider again the simple harmonic oscillator from Problem 1 of the
tutorials. Write down the Liouville equation in this case. What
is the general solution ρ(p(t), q(t), t) to the Liouville equation when
ρ(p(0), q(0), 0) = ρ(p0 , q0 , 0) is given?
(c) Consider N independent classical harmonic oscillators with different
masses and force constants,
N 
p2i

X 1
H(pN , qN ) = + mi ωi2 q2i
i=1
2mi 2

Determine for this Hamiltonian Ω̃(U, V, N ). What is the internal

energy in this case? Interpret your answer.
2. Consider a microcanonical system with given number of particles N , inter-
nal energy U , and volume V . We divide this system into two subsystems
with fixed number of particles Ni and volume Vi for i = 1, 2. The total
energy is fixed U = U1 + U2 , however, the two subsystems can exchange
energy, i.e., U1 and U2 = U − U1 are not fixed.
(a) Let W (U1 ) be the probability that the energy of subsystem 1 is in
the regime between U1 and U1 + dU1 . Show that

W (U1 ) ∝ ω(U1 , V1 , N1 )ω(U − U1 , V2 , N2 ).

(b) Denote the most probable value of U1 by U1∗ , i.e. U1∗ maximises
W (U1 ). Write down the condition that determines U1∗ and show that
the result is consistent with the zeroth law of thermodynamics.
(c) Show that in the thermodynamic limit that
   
1 1 1 ∗ 2
W (U1 ) ∝ exp − + (U1 − U1 )
2kB T 2 CV,1 CV,2
and that
√ the standard deviation of this Gaussian distribution scales
as ∼ N . What do you conclude when N is large?
3. Consider the volume of a n-dimensional ball of radius R (see Week II).
Define
Vn (R) − Vn ((1 − α)R)
fn (α) =
Vn (R)
(a) Give a physical interpretation of fn (α).
(b) Show that fn (α) = 1 − exp(−αn) for 0 < α ≪ 1.
(c) In which regime is a Taylor expansion of fn (α) for small α appropri-
ate?
(d) Estimate numerical values for fn (10−6 ) for n = 1, 2, 3, 105 , 106 , 1020 .
(e) What is on the basis of these estimates, the validity of the Equations
derived for the entropy in Problem 2 of the tutorial?

15
8 Week VIII
Finish tasks from week VII and discuss tasks from the mid-term exam.

9 Week IX
1. Derivation of the Boltzmann factor. Consider a large system of fixed
energy U , volume V and the number of particles N . The system is divided
into two subsystems of energies Ui , volumes Vi and the number of particles
Ni , where i = 1, 2. A fixed wall between the subsystems allows exchange
of energy but not of particles, hence N1 = const, N2 = const, V1 = const,
V2 = const, and U = U1 + U2 . Moreover, U2 ≫ U1 , N2 ≫ N1 , V2 ≫ V1 .
Both subsystems are in thermal equilibrium at temperature T .
(a) Using microcanonical ensemble derive the distribution function of the
combined system.
(b) Using the assumption U2 ≫ U1 and the Taylor series expansion
around U = U2 derive the Boltzmann factor and the distribution
function in the canonical ensemble for the small subsystem.
2. Suppose an atom has only two available energy levels, which are separated
by 1.5 · 10−21 J.
(a) If the temperature is 1.5 K, what is the probability that the atom is
in the lower state? Plot this probability as a function of temperature.
(b) What is the temperature T ifpthe energy in lower state ϵ0 = 0 and
the standard deviation σE = ⟨E 2 ⟩ − ⟨E⟩2 = 0.5 · 10−21 J? How the
mean energy ⟨E⟩ and σE change as a function of temperature?
3. Classical ideal gas, the partition function in canonical ensemble - Derive
the partition function for the classical ideal gas starting from classical
kinetic energy for particles.
4. Thermodynamics of ideal gas derived microscopically in canonical ensem-
ble - From the partition function of the classical gas derive: internal energy,
specific heat, Helmholtz free energy, pressure and equation of states, in-
ternal energy, entropy. Derive the same expression for the entropy using
results from microcanonical ensemble.

9.1 Homework
1. Statistical sum satisfies the relation ln Z = aT α V , where a and α are pos-
itive constants. Calculate specific heat at constant volume of this system.
2. A system can be in states of energies 0, ϵ, ϵ, ϵ, 2ϵ. Calculate the internal
energy of this system and the specific heat at temperature T .

16
 3. Prove the Dalton’s law (which states that in a mixture of non-reacting
gases, the total pressure is equal to the sum of the partial pressures of
the individual gases) for a mixture of classical gasses using the canonical
ensemble. Hint: Calculate the statistical sum for the mixture of gasses
and make use of the definition of the pressure, as it was done for one-
component gas.
4. Using the classical canonical ensemble, justify the postulate about additiv-
ity of internal energies and entropies of two ideal gases (treated indepen-
dently) filling a container of volume V and having the same temperature
T.

10 Week X
1. Connection with thermodynamics in the classical canonical ensemble. Con-
sider the classical canonical partition function
ˆ
1
Z(N, V, T ) = dΓ exp[−βH(Γ)].
N !h3N
Here Γ is a shorthand notation for the phase space variables. We are
interested in the energy distribution W (E) in the canonical ensemble,
given by ˆ
W (E) = dΓ fc (Γ)δ(E − H(Γ)),

with fc (Γ) being the canonical probability distribution.

(a) Give an explicit expression for fc (Γ) and show that

exp(−βE)ω(E, V, N )
W (E) = .
N !h3N Z(N, V, T )

What is ω(E, V, N )?
(b) The distribution W (E) is maximal for E = E ∗ with E ∗ given by
β(E ∗ , V, N ) = β. In other words E ∗ is the energy for which the tem-
perature of the system equals that of the reservoir. Expand W (E)
around E ∗ and show that W (E) is in fact Gaussian in the thermo-
dynamic limit

(E − E ∗ )2
 
2 −1/2
W (E) = (2πkT CV ) exp − .
2kT 2 CV

Here, CV is the heat capacity at constant volume.

(c) Show that the relative energy fluctuations are characterised by
p
⟨(E − E ∗ )2 ⟩ 1
∝√
E∗ N

17
 (d) Show that in the thermodynamic limit
1/2
Z(N, V, T ) = exp[S ∗ /k] exp[−βE ∗ ](2πkT 2 CV ) ,
∗ ∗
with S = S(E , V, N ) =: ⟨S⟩ the entropy of the canonical ensemble.
(e) Show by ignoring O(log N ), it follows that
F (N, V, T ) := −kT log Z(N, V, T ) = E ∗ − T S ∗ = ⟨E⟩ − T ⟨S⟩.
Note that F is the Helmholtz free energy and it generates the full
thermodynamics of systems with fixed (N, V, T ).
2. For the Maxwell distribution of gas particle velocities in an ideal gas
mv 2
 m 3/2  
P (v) = exp − ,
2πkT 2kT
where T is the temperature and m is the mass of a gas particle
(a) find the mean speed ⟨|v|⟩ of a gas particle, and the standard deviation
of |v|,
(b) find the most probable speed of a gas particle,
(c) find the average relative velocity of two particles |v12 | = |v1 − v2 |.
3. Barometric formula - Find the formula describing how the pressure of the
ideal gas in a uniform gravitational field changes with the altitude and the
temperature.
4. Equipartition of energy Consider a system of 2s degrees Psof freedom with
2 2
the quadratic Hamiltonian H(p1 , ..., ps , q1 , ..., qs ) = i=1 (ai pi + bi qi ).
Here qi and pi are generalised coordinates and momenta, respectively.
Furthermore, ai and bi are constants (i = 1, ..., s). The internal energy U
is, when it is in thermal equilibrium with a heat bath at temperature T ,
given by
´ ´ ´ ´
dp1 ... dps dq1 ... dqs H exp(−βH)
U = ⟨H⟩ = ´ ´ ´ ´ .
dp1 ... dps dq1 ... dqs exp(−βH)
(a) Give an example of a physical model system described by H.
(b) Calculate U and show that each quadratic “mode” of H contributes
an energy amount kT /2 independent from ai and bi . This is called
the equipartition of energy.

10.1 Homework
1. Find the correlation function between two components of velocity
C(vx , vy ) = ⟨vx vy ⟩ − ⟨vx ⟩⟨vy ⟩
for an ideal gas assuming a Maxwell velocity distribution. Can you guess
the result before calculating?

18
 2. Find the center of gravity of an ideal gas in the cylinder in the uniform
gravity field g at the temperature T . An atomic mass m is given.
3. Ideal gas is composed of molecules with two atoms, each of mass m. Atoms
inside a molecule interact with each other with the potential of model
form V (r) = V (r0 ) + (1/2)mω0 ξ 2 , with ξ = r − r0 . For a single molecule
find the kinetic energy in relative and center of mass coordinates. Then
find the partition function and discuss each contributing terms. For the
translationally invariant part use the earlier results for a mono-atomic
ideal gas. For rotational and vibrational degrees of freedom show that
Z1rot = 2IkB T /ℏ2 and Z1vib = kB T /ℏω0 , where I = mr02 . Next, for a
gas of diatomic molecules find: internal energy, specific heat at constant
volume, Helmholtz free energy, entropy, pressure and equation of state.
Hints: Consider position of the center of mass and the relative distance
vector:
⃗ = (⃗x1 + ⃗x2 )/2, ⃗r = ⃗x1 − ⃗x2
R
with the corresponding momenta
P⃗ = p⃗1 + p⃗2 , p1 − p⃗2 )/2.
p⃗ = (⃗
To account for vibrational and rotational motion consider variables: qr =
r, qθ = θ, qϕ = ϕ and the corresponding generalised momenta: pr = mṙ,
pθ = mr2 θ̇, pϕ = mr2 ϕ̇ sin2 θ. Derive the kinetic energy:
!
p2r 1 2
p2ϕ
Ekin = + pθ +
2m 2mr2 sin2 θ

Consider small deviations ξ such that

! !
1 2
p2ϕ 1 2
p2ϕ 1
pθ + + V (r) ≈ pθ + + V (r0 ) + mω0 ξ 2 .
2mr2 sin2 θ 2mr02 sin2 θ 2

and pr = pξ .

11 Week XI
1. Grand canonical ensemble. Consider a large system of fixed energy U ,
volume V and the number of particles N . The system is divided into
two subsystems of energies Ui , volumes Vi , and number of particles Ni
where i = 1, 2. The subsystems can exchange energy and particles, hence
V1 = const, V2 = const, and U = U1 + U2 , N = N1 + N2 . Moreover, U2 ≫
U1 , N2 ≫ N1 , V2 ≫ V1 . Both subsystems are in thermal and chemical
equilibrium at temperature T = T1 = T2 and the chemical potential µ =
µ1 = µ2 .
(a) Using the microcanonical ensemble derive the distribution function
of the combined system.

19
 (b) Using the assumptions U2 ≫ U1 and N2 ≫ N1 and the Taylor series
expansion around U = U2 , N = N2 derive the distribution function
and the partition function in the grand canonical ensemble for the
small subsystem.

2. Summary of ensembles. Compare microcanonical, canonical and the grand

canonical ensembles. What are the constraints and what are the fluctu-
ating variables in the respective ensembles? Compare the distribution
functions and thermodynamic potentials derived from the respective par-
tition functions.

3. Ideal gas in the grand canonical ensemble. Derive the grand canonical
partition function for the ideal gas. Derive the mean energy, equation of
state and the entropy. Show that the probability distribution for number
of particles is a Poisson distribution.
4. Model of adsorption - As a simple model of adsorption consider an ideal
classical gas and a surface with M holes. In every hole there might be
maximally one particle and then its energy is −ϵ < 0. Find the covering
coefficient k = N̄ /M , where N̄ is the average number of particles adsorbed
on the surface.

11.1 Homework
1. Grand canonical ensemble - ultra-relativistic ideal gas - Consider ultra-
relativistic ideal classical gas in the grand canonical ensemble in three
space dimensions. The dispersion relation is ϵp = cp, where c is a given
velocity constant and p is a length of the momentum. Find: (a) the
partition function, (b) the grand canonical thermodynamic potential, (c)
the average number of particles, (d) the pressure, (e) the entropy, (f) the
internal energy.
2. Fluctuation of particle number i grand canonical ensemble - Derive the
2
formula for the variance of particle number σN = ⟨N 2 ⟩ − ⟨N ⟩2 in the
grand canonical ensemble andp show that the relative fluctuation of the
particle umber σN /⟨N ⟩ ∼ 1/ ⟨N ⟩ vanishes in the thermodynamic limit.
3. Show that ⟨N 2 ⟩ − ⟨N ⟩2 = ρkB T ⟨N ⟩κT , with κT the isothermal compress-
ibility and ρ the number density in the grand-canonical ensemble.

20