Homework 3 Solutions
1. We first need to find an expression for the Fermi energy ϵF . We can do so by first
finding the Fermi momentum pF (maximum momentum of any electron in gas). The
number of particles with a momentum between p and p + dp is given by
gs × probability of a state with momentum between p and p + dp being occupied
× number of states with momentum between p and p + dp
V
= gs × f (ϵ) × 3 4πp2 dp
h
where ϵ is the kinetic energy of a particle with momentum between p and p + dp (we
ignore the rest mass energy), and f (ϵ) is the Fermi-Dirac distribution.
Integrating over all possible momentum, we should recover the total number of particles
over the volume V ,
Z ∞
V
N= gs f (ϵ) 3 4πp2 dp
h
Z0 ∞
1 V
= gs (ϵ−µ)/k T 3
4πp2 dp
0 e B + 1 h
where µ is the chemical potential. In the zero temperature limit, the Fermi-Dirac
distribution f (ϵ) becomes a step function:
(
1 if ϵ < µ
f (ϵ, T = 0) =
0 otherwise
and the Fermi energy is defined as ϵF = µ(T = 0). The Fermi momentum is given by
the non-relativistic kinetic energy- momentum relation: ϵF = p2F /2m. Thus, we only
need to integrate up to the Fermi momentum in our integral over p,
Z ∞
1 V
N= gs (ϵ−µ)/k T 3
4πp2 dp
e B + 1 h
Z0 pF
N 1
= gs 3 4πp2 dp (1)
V 0 h
8π
n = 3 p3F
3h
3 1/3
3h n
pF = (2)
8π
The Fermi energy is then
2/3
p2 h2
3n
ϵF = F = (3)
2m 8π 2m
1
√
Alternatively,
p one could have made the change of variables p = 2mϵ (which gives
dp = m/(2ϵ)dϵ ) and integrate
Z ϵF
1 p
n= gs 3 4π(2mϵ) m/(2ϵ)dϵ
0 h
8π (2m)3/2 ϵF 1/2
Z
= 3 ϵ dϵ
h 2 0
8π (2m)3/2 3/2
= 3 ϵF
h 3
2/3 2
3n h
=⇒ ϵF =
8π 2m
Now we set n = nQ = (2πme kB T /h2 )3/2 , then
2/3 2
3nQ h
ϵF =
8π 2me
2/3 2
3 2πme kB T h
= 2
8π h 2me
2/3 1/3
kB T 1 8π 64
= =
ϵF π 3 9π
2. Note: The textbook’s hint contains a typo in the number density relation between H2
and H. But points will be given if you followed that part of the hint.
Let’s say the initial number density of H2 is n0 . 50% of the H2 is dissociated, so the
number density of undissociated H2 is nH2 = 0.5n0 . One H2 gives two H atoms, so
number density of H2 is twice the number density of dissociated H2 , nH = 2 × 0.5n0 =
n0 . So nH2 = 0.5n0 = 0.5nH , or
2nH2 = nH (4)
At chemical equilibrium, the chemical potentials on both sides are equal. Since the
chemical potential of a photon is zero, we only need to consider
µ(H2 ) = 2µ(H) (5)
The chemical potential is given by
g n
s Q
µ = mc2 − kB T ln (6)
n
where gs is the statistical weight, nQ is the quantum concentration. Substituting
2
equation 6 into equation 5,
2 gH2 nQ,H2 2 gH nQ,H
mH2 c − kB T ln = 2mH c − 2kB T ln
nH 2 nH
2 2
gH nQ,H
2 2 gH2 nQ,H2
(mH2 c − 2mH c ) = kB T ln − ln
nH 2 n2H
!
gH2 nQ,H2 n2H
ϵB = kB T ln
gH2 n2Q,H nH2
where we use the fact that the rest mass energies of the two hydrogen atoms, 2mH c2
goes into rest mass energy of the hydrogen molecule plus the binding energy, mH2 c2 +ϵB ,
to write (mH2 c2 − 2mH c2 ) = ϵB . Note that ϵB is negative (ϵB = −4.48 eV); a hydrogen
molecule has a lower mass than two hydrogen atoms added together. Now, we can
exponentiate both sides of the equation,
gH2 nQ,H2 n2H
eϵB /kB T =
gH2 n2Q,H nH2
n2H g 2 n2Q,H ϵB /kB T
= H e
nH 2 gH2 nQ,H2
Now, we use the definition of the quantum concentration,
3/2
2πmkB T
nQ = (7)
h2
to write
3/2 3/2
n2H g2 (2πmH kB T )2 h2
= H eϵB /kB T
nH2 gH2 h4 2πmH2 kB T
3/2
gH2 2πkB T m2H
= eϵB /kB T
gH2 h2 mH2
3/2
gH2 2πkB T m2H
= eϵB /kB T
gH2 h2 2mH
3/2
n2H gH2 πmH kB T
= eϵB /kB T (8)
nH2 gH2 h2
Using the 50% dissociation condition, equation 4, we can write 2nH2 = nH , so
3/2
gH2 πmH kB T
2nH = eϵB /kB T (9)
gH 2 h2
We separately find nH as a function of temperature using the ideal gas law,
3 2 P
P = nkB T = (nH + nH2 )kB T = nH kB T =⇒ nH =
2 3 kB T
3
Substituting back, we get
3/2 3/2
gH2 πmH kB T h2
4 P 4 gH2 P
= eϵB /kB T =⇒ = eϵB /kB T
3 kB T gH2 h2 3 gH2 πmH (kB T )5/2
If you had followed the textbook’s hint on nH2 , you would get nH = P/(3kB T ), which
means n2H /nH2 = nH /2 = P/(6kB T ) and
2 3/2
1 gH2 h P
2
= eϵB /kB T
6 gH πmH (kB T )5/2
Either way, we approximate that gH2 = gH = 1 and numerically solve for T , given
P = 100 Pa, ϵB = −4.48 eV. With the book hint, T ≈ 2200 K; and without the typo
in the book hint, T ≈ 2400 K.
3. Consider equation 1, which integrates the occupation number f (ϵ) times density of
states g(p)dp over all momentum p to get the total number of particles. We now set
µ = 0, and use the relativistic expression ϵ = pc to write,
Z ∞
1 V
N= gs pc/k T 3
4πp2 dp
0 e B + 1 h
4πgs ∞
Z
1
n= 3 pc/k
p2 dp
h e BT + 1
0
We now make a change of variables x = pc/kB T (dx = cdp/kB T ),
3 Z ∞
x2
4πgs kB T
n= 3 dx
h c 0 ex + 1
One could solve this using an online integral calculator. However, here we follow the
book hint. We first write
3 Z ∞
x2
kB T
n = 4πgs dx
hc 0 ex [1 − (−e−x )]
For −1 < r < 1, the geometric series converges:
∞
1 X
= rn
1 − r n=0
so we can write
∞
3 Z !
∞
kB T X
n = 4πgs x2 e−x (−e−x )n dx
hc 0 n=0
∞
3 Z !
∞
kB T X
= 4πgs x2 (−1)n−1 e−nx ) dx
hc 0 n=1
∞
3 X Z ∞
kB T n−1 2 −nx
= 4πgs (−1) xe dx
hc n=1 0
4
Now consider
∞ ∞
e−nx
Z Z
a −nx a
x e dx = x d dx
0 0 −n
−nx ∞
a ∞ a−1 −nx
Z
a e
=x + x e dx
−n 0 n 0
a ∞ a−1 −nx
Z
= x e dx
n 0
We repeatedly integrate by parts for k times until a − k = 0. The resulting numerator
k+1
is then a · (a − 1) . . . 1 = a! and the denominator
R −nxis n = na+1 since after the k
integration by parts we still have to integrate e dx. Hence,
Z ∞
a!
xa e−nx dx = a+1 (10)
0 n
Here, a = 2, so
3 X∞
kB T n−1 2!
n = 4πgs (−1)
hc n=1
n3
"
3 X #
kB T 1 X 1
= 8πgs 3
−
hc odd n
n even n
n3
"
3 X #
kB T 1 X 1 X 1
= 8πgs + −2
hc odd n
n3 even n
n3 even n
n3
"
3 X ∞ ∞ #
kB T 1 X 1
= 8πgs −2
hc n=1
n3 k=1
(2k)3
3 "X ∞ ∞ #
kB T 1 1X 1
= 8πgs −
hc n=1
n3 4 k=1 k 3
3 " X ∞ #
kB T 3 1
= 6πgs
hc 4 n=1 n3
3
kB T
n = 6πgs ζ(3)
hc
We defined
8πk 3
b = 2ζ(3) ×
h3 c3
so
8πk 3 3
n = 2ζ(3) × × gs T 3
h3 c3 8
3 3
n= bT (11)
4
5
Similarly, we get the total energy by integrating ϵf (ϵ)g(p)dp, where ϵ = pc
Z ∞
1 V
E= pc · gs pc/k T 3
4πp2 dp
0 e B +1 h
Z ∞ 3
E 4πgs c p
= 3
dp
V h 0 e BT + 1
pc/k
Making a change of variables x = pc/kB T again, the energy density u = E/V is
4 Z ∞
x3
4πgs c kB T
u= dx
h3 c 0 ex + 1
Similarly to the previous part, we can write the integral as a sum of integrals,
4 4X ∞ Z ∞
kB T n−1 3 −nx
u = 4πgs (−1) x e dx
h3 c3 n=1 0
4 4X ∞
kB T n−1 3!
= 4πgs (−1)
h3 c3 n=1 n4
4 4" X X 1
#
kB T 1
= 24πgs −
h3 c3 odd n
n4 even n
n4
4 4 "X ∞ X 1
#
kB T 1
= 24πgs −2
h3 c3 n=1
n4 even n
n4
4 4 "X ∞ ∞ #
kB T 1 1 X 1
= 24πgs −2· 4
h3 c3 n=1
n4 2 n=1 n4
4 4
kB T
= 21πgs ζ(4)
h3 c3
4 4 4
kB T π
= 21πgs 3 3
·
hc 90
5 4 4
7 π kB T
u = gs
30 h3 c3
With
8π 5 kB4
a=
15h3 c3
we can write the energy density as
7
u= gs aT 4
16
Electrons and positions each have gs = 2, whereas the 3 types of neutrinos and anti-
neutrinos each have gs = 1. Then the sum of degeneracy factors for all 8 types of
6
fermions is Σgs = 10. The energy density of photons is aT 4 , so the total energy
density is
7 43
utot = (10)aT 4 + aT 4 = aT 4
16 8
and since the pressure for relativistic particles is one-third of the kinetic energy density,
1 43 4
P = utot = aT
3 24
4. The radius is given by
r
2 4 L
L = 4πσR T =⇒ R = 4
r 4πσT
0.01L⊙
=
4πσT 4
= 9.05 × 106 m
= 0.013 R⊙
The average density is
M 2 × 1030 kg
ρ̄ = 3
= 6 3
≈ 6.4 × 108 kg m−3 (12)
4πR /3 4π(9.05 × 10 m) /3
Using equation 3.43 of Phillips, and substituting L = 0.01 L⊙ and M = M⊙ , the
internal temperature of the white dwarf is
7/2
L TI M
= 7
=⇒ TI ≈ 1.88 × 107 K
L⊙ 7 × 10 K M⊙
or kTI ≈ 1.6 × 103 eV.
To show that the electrons are mostly degenerate, we can either compare kB TI with the
Fermi energy ϵF , or the electron number density ne with the quantum concentration
of electrons nQ,e . But first we have to find ne .
A one solar-mass white dwarf is mostly composed of carbon and oxygen. For simplicity
we assume that the white dwarf is composed of carbon. The number density of ions is
ρ̄ 6.4 × 108 kg m−3
nion ≈ = ≈ 3 × 1034 m−3 (13)
12mp 12 · 1.67 × 10−27 kg
Assuming complete ionization, each carbon contributes 6 electrons. Hence, the electron
number density is
6ρ̄ 6.4 × 108 kg m−3
ne ≈ 6 × nion = = −27
≈ 2 × 1035 m−3
12mp 2 · 1.67 × 10 kg
7
which gives a Fermi energy of
2/3
p2 h2
3ne
ϵF = F = ≈ 2 × 10−14 J = 1.2 × 105 eV
2me 2me 8π
Our mass density estimate is (6.4 × 108 )/(3.4 × 107 ) ≈ 18.8 times that of the textbook
2/3
hint. Since ϵF ∝ ne ∝ ρ2/3 , scaling our Fermi energy estimate by a factor of (18.8)−2/3
gives ϵF ≈ 1.7 × 104 eV which agrees with the book’s hint.
The ratio of thermal energy to the Fermi energy is then
kT 1.6 × 103 eV
≈ ≈ 0.01
ϵF 1.2 × 105 eV
which means the electrons are mostly degenerate.
Equivalently, the quantum concentration is
3/2
2πme kB TI
nQ,e = ≈ 1.93 × 1032 m−3
h2
and ne ≫ nQ,e again shows that the electrons are mostly degenerate.
The Coulomb coupling parameter is defined as the ratio of the electric potential energy
between two ions and the thermal energy:
UE
ΓC = (14)
kB T
which is 60 in our case. Note: literature usually quotes ΓC ≈ 175 as the criterion
for crystallization. To find UE , recall that the electric potential energy between two
charged particles Q1 and Q2 separated by a distance r is
Q1 Q2
UE =
4πϵ0 r
where ϵ0 is the permittivity of free space. Here, for an atom with atomic number
Z, there are Z protons each carrying a charge of +e. The charge of the ion is then
+Ze. Suppose that each ion occupies a volume V , the number density of ions is then
nion = 1/V = 3/(4πr3 ). This gives a distance of r = (4πnion /3)−1/3 between ions.
Gathering all these terms, the electrostatic energy between ions is
1/3
Z 2 e2 4π 1/3
UE = nion
4πϵ0 3
The Coulomb coupling parameter can then be expressed as
2 1/3
1/3
Z 2 e2 4π 1/3 −5 Z nion
ΓC = nion ≈ 2.69 × 10
4πϵ0 kB T 3 T
8
In terms of the mass density, let us approximate each ion as made up of equal number
of protons and neutrons. The atomic number is then A ≈ 2Z and the mass of each
ion is then mion ≈ 2Zmp . The mass density is then ρ = 2Zmp nion and the Coulomb
coupling parameter is
1/3 1/3 1/3
Z 2 e2 Z 5/3 e2
4π ρ 2π
ΓC = = ρ1/3
4πϵ0 kB T 3 2Zmp 4πϵ0 kB T 3mp
Solving for mass density, we find that crystallization happens when the density is
greater than
(ΓC 4πϵ0 kB T )3 3mp
ρ=
Z 5 e6 2π
3 3 −5
ΓC T Z
≈ 3 × 10 10
7
kg m−3
60 1 × 10 K 6
or when the ion number density is greater than nion ≈ 1.6 × 1036 m−3 .
Note that ΓC increases strongly with Z. This means that elements with higher atomic
number crystallize first. For example, oxygen would crystallize first in a carbon-oxygen
white dwarf, pushing liquid carbon out of the crystallized core. This process is known
as phase separation. See Kippenhahn Ch 37.
