Solutions exam 2

Problem 1
(a) Which of those quantities defines a thermodynamic potential. Why? (2 points)
(i) T, p, N Gibbs free energy G
(ii) T, p, µ no thermodynamic potential, since T, p, µ are not independent of each other
(iii) S, p, N Enthalpy H

(b) Suppose we have a box with two chambers. In each of the two chambers we have the same ideal
gas. Which conditions need to be fulfilled by
(i) Thermal equilibrium? Heat exchange or T1 = T2
(ii) Chemical equilibrium? Particle exchange or µ1 = µ2
(iii) Mechanical equilibrium? Volume exchange or P1 = P2
(3 points)
(c) What is extremized in the microcanonical, the canonical, and the grand canonical ensemble? Is it
a maximum or a minimum? (3 points) microcanonical: entropy is maximal or partition function
maximal canonical: Helmholtz free energy is minimal, or entropy is maximal s.t. boundary con-
ditions, or partition function maximal grand-canonical: Grand-potential is minimal, or entropy
is maximal s.t. boundary conditions, or partition function maximal
(d) Define the Bose-Einstein condensation. What are the volume and chemical potential in the
condensed phase? Explain why. (4 points)
Below a critical temperature Tc a phase transition sets in, i.e. particles go from the gas phase
in the condensate. At low temperature most particles occupy the lowest energy level (at T = 0
all particles , see problem f) with
 3/2
hN0 i T
=1− (1)
N Tc

The volume of the condensate is Vcond ≈ 0 and the chemical potential is µ = 0 .

(e) Plot qualitatively the Carnot-cycle in a T -S and a p-V diagram. Show where the heat is going
in and out of the system. (4 points)

(shaded areas not to scale)

Heat goes into the system at step 1→2 (isothermal expansion) and out at step 3→¿4 (isothermal
compression). Steps 2→3 and 4→1 are adiabatic (dQ = 0) expansion and adiabatic (dQ = 0)
compression, respectively
(f) Plot qualitatively the Fermi-Dirac and Bose-Einstein as a function of β(E − µ) for T  0K and
T ≈ 0K. (4 points)

1
 left: Fermi distribution at differen temperatures, middle Fermi, Bose and Boltzmann distribution
at T =0K, right: Fermi, Bose and Boltzmann distribution at higher temperaure.
(g) What is the Maxwell construction for the van der Waals gas? Why is it used? (5 points)

Below a critical temperature Tc a gas can

condensate to a liquid. The equation for the van der Waals gas (that inclludes interaction such
that in principle the phase transition to liquid phase is possible) is cubic in the voume and
hence has three solutions, possible volumes for a given pressure. One of those can be dismissed
because of the positive slope of the vdW curve, corresponding to unphysical behaviour. The
other two solutions correspond to liquid phase (lower volume) and gas phase (higher volume).
The Maxwell construction finds the corresponding pressure form the horizontal line that equates
the areas between the line and the vdW isotherme (below and above). This is the pressure at
wchich, fo the given temperature. liquid and gas phase coexist

Problem 2 - Gas with two energy states

A classical gas in a volume V is composed of N non-interacting and indistinguishable particles. The
single particle Hamiltonian isH = p2 /2m + , with m the mass of the particle and p the absolute value
of the momentum. Moreover, for each particle, we find two internal energy levels: a ground state with
energy  = 0 and degeneracy g1 , and an excited state with energy  = 1 and degeneracy g2 .

(a) Determine the canonical and grand canonical partition function for N particles. Show that the
Helmholtz free energy has the form
VN
     
3N 2πm
A = −kB T ln + ln + N ln(g1 ) − N β ln(g2 )
h3N N ! 2 β
(10 points)
Solution:
When computing the canonical partition function, we have to consider the continuous part of
the Hamiltonian p2 /2m, plus the contribution coming from the internal degrees of freedom. The
partition function for the N particles is then the product of N single particle partition functions

∞ 3N
VN
Z
p2
ZN (T, V, N ) = 3N
e−β 2m dp (g1 + g2 e−β )N
h N! −∞
 3N
VN

2πm 2
= (g1 + g2 e−β )N
h3N N ! β
R∞ 2
Using −∞ e−ax dx = πa . The factorial N ! accounts for the classical indistiguishability of the
p
particles. From the partition function we compute the free energy
VN
     
3N 2πm
A = −kB T lnZN = −kB T ln + ln + N ln(g1 ) − N β ln(g2 )
h3N N ! 2 β

2
 (b) Compute the energy E of the total system as a function of the temperature T . (5 points)
Solution:

g2  e−/kB T
 
∂lnZN 3
E=− = N kB T + N
∂β V,N 2 g1 + g2 e−/kB T

(c) Compute specific heat CV as a function of T . (5 points)

Solution:
An the specific heat

N g2  e−/kB T N g1 g2 2 e/kB T
   
∂E 3 d 3
CV = = N kB + −/k T
= N kB +
∂T V,N 2 dT g1 + g2  e B 2 kB T 2 (g2 + g1 e/kB T )2

(d) Analyze the limit of low temperatures of the specific heat and explain this result. (5 points)
Solution:
For low temperatures, we find

N g1 g2 2 e/kB T g1 g2 2
≈ →0
kB T 2 (g2 + g1 e/kB T )2 kB T 2 g12  e2/kB T

That is the expected result because in such limit only the ground state (the one with energy 0)
is populated.

Problem 3 - Relativistic fermi gas

In high energy physics it is convenient to take into account relativistic effects. We now want to
calculate the energy and pressure of the non-interacting relativistic fermi gas. In this situation, the
energy of one particle is given by 2p = p2 c2 +m2 c4 . Assume that we are dealing with spin-1/2 particles,
further let be ρ() the density of states and Ω() the number of states with energy smaller than .

(a) The grand-canonical partition sum for fermions reads as

Y 
Ξ= 1 + e−β(p −µ)
p

Assuming we have a large system and small differences between the energy levels, we can trans-
form the sum over all states, that appears in the corresponding grand canonical potential, into
an integral:

X Z∞
−→ d ρ()
p 0

Perform this transformation and show that the grand canonical potential for a general system
of fermions can then be expressed as
Z∞
Ω()
Φ(T, V, µ) = −kB T d .
1 + eβ(−µ)
0

Explain all vanishing terms that appear during your calculation. (9 points)
Solution:

3
 The expression for the grand-canonical potential reads as
X  
Φ(T, V, µ) = −kB T ln Ξ = −kB T ln 1 + e−β(p −µ)
p

Z∞  
= −kB T d ρ() ln 1 + e−β(−µ)
0

Z∞

−β(−µ)
∞ −β e−β(−µ)
= −kB T Ω() ln 1 + e  + kB T d Ω()

0 1 + e−β(−µ)
0

Z∞
Ω()
=− d .
1 + eβ(−µ)
0

Here we performed a partial integration. The boundary term vanish since for  = 0 we get Ω = 0
from a).
(b) Calculate the number of states in 3 dimensions
Z
2V
Ω() = 3 d3 p
h
p <

with the fermion energy p being smaller than . (5 points)

Solution:
As usual:
Z Z
2V 8πV 8πV 3  √
Ω(E) = 3 d3 p = 3 p2 dp = p p= E 2 /c2 −m2 c2
h h 3h3
p <E p2 c2 <E 2 −m2 c4
3/2
E2

8πV
= − m2 c2
3h3 c2

(c) Go to the limit of highly relativistic particles (  mc2 ). Show that the density of states is then
given by ρ() = 3Ω()/. (4 points)
Solution:
In the highly relativistic limit we have E 2 /c2 − m2 c2 ≈ E 2 /c2 . We then just calculate the
derivative:
dΩ(E) 8πV d E 3 8πV E 2 3Ω(E)
ρ(E) = = 3 3
= = .
dE 3h dE c h3 c3 E

(d) Use your result from a) and the expression

R for the number of fermions per state hn()i = dN/dΩ
to show that Φ = −E/3, where E = dN (N ) is the total energy of the system (the integral
over all single fermion energies). (5 points)
Solution:
From the lecture we know that
1
hn()i =
1 + eβ(−µ)
It thus follows with part b) that
Z∞ Z∞ Z∞
1 1 dΩ() dN 1 1
Φ(T, V, µ) = − d ρ() hn()i = − d  =− dN (N ) = − E .
3 3 d dΩ 3 3
0 0 0

(e) Use the thermodynamic definition of the grand-canonical potential to obtain an expression for
the pressure p(E, V ). (2 points)
Solution:
We know that Φ = E − T S − µN = T S − pV + µN − T S − µN = −pV , hence p = E/3V .

4
Problem 4 - Gas with hard core interaction
A gas obeys the equation of state
 
α
P (V − N b) = N kB T exp − ,
V − Nb

where α and b are the parameters of system and N is kept fixed in the following.

(a) Find the Maxwell relation involving (∂S/∂V )T . (3 points)

Solution:
Using Helmholtz free energy we have

dA = d(E − T S) = −SdT − P dV ⇒ (∂S/∂V )T = (∂P/∂T )V

(b) Use a differential for S to calculate dE(T, V ). Show that for the upper equation of state E is a
function of T only. (9 points)
Solution:
We have
    
∂S ∂S
dE = T dS − P dV = T dT + dV − P dV
∂T V ∂V T
     
∂S ∂S
= T dT + T − P dV
∂T V ∂V T
     
∂S ∂P
= T dT + T − P dV
∂T V ∂T V

Using equation of state we have:

   
dP N kB α P
= exp − =
dT V (V − N b) V − Nb T

Then for dE we can write

     
∂S P ∂S
dE = T dT + T − P dV = T dT
∂T V T ∂T V

Then we can see that E is just a function of T , i.e E = E(T )

 
∂H
(c) Using CP = and the definition of the enthalpy show that:
∂T P
 
N kB (V − N b) α
γ = CP /CV = 1 + exp − .
CV (V − N b) − α V − Nb

(9 points)
Solution:
Using enthalpy definition H = E + P V we have
       
∂H ∂(E + P V ) ∂E ∂V
CP = = = +P .
∂T P ∂T P ∂T P ∂T P

Also because E = E(T ) for heat capacity CV at constant volume we have

   
∂E ∂E
CV = = .
∂T V ∂T P

Then for CP we can write

 
∂V
CP = CV + P .
∂T P

5
 
    
∂V ∂T  ∂V
For finding it is easiest to calculate =1 . Then for this reason we
∂T P ∂V P ∂T P
rewrite equation of state for T
 
P (V − N b) α
T = exp
N kB V − Nb

Then we have
     
∂T P α P (V − N b) α α
= exp − exp
∂V P N kB V − Nb N kB (V − N b)2 V − Nb
   
1 α α
= − P exp .
N kB N kB (V − N b) V − Nb

With substituting P as a state function in above relation we have

      
∂T V − Nb − α N kB T α α
= exp − exp
∂V P N kB (V − N b) V − N b V − Nb V − Nb
 
V − Nb − α
= T
(V − N b)2

Then

(V − N b)2
 
∂V 1
=  =
∂T P ∂T T (V − N b − α)
∂V P

Then for CP and substituting P as a state function we can write

P (V − N b)2
 
∂V
CP = CV + P = CV +
P ∂T T (V − N b − α)
2
 
(V − N b) N kB T α
= CV + exp −
T (V − N b − α)(V − N b) V − Nb
 
(V − N b) α
= CV + N kB exp −
V − Nb − α V − Nb

Then for γ we have

 
CP N kB (V − N b) α
γ= =1+ exp −
CV CV (V − N b − α) V − Nb

(d) In which limit does γ go to the value obtained from an ideal gas? (4 points)
Solution:
N kB
In the limit of α → 0 we have γ = 1 + . That is the same as the ideal gas relation for γ.
CV