MASSACHUSETTS INSTITUTE OF TECHNOLOGY

Physics Department

8.044 Statistical Physics I Spring Term 2013

Problem Set #11

Due in hand-in box by 4:00 PM, Friday, May 10

Problem 1: Ripplons

(k)

We have seen that the bulk motion of a solid or liquid can be described by harmonic normal
modes (that is, normal modes each having a harmonic oscillator Hamiltonian) known as
phonons. In a similar manner the two dimensional waves on an interface between a liquid
and its vapor can be described by harmonic normal modes known as “ripplons” each having a
single direction of polarization perpendicular to the interface. The dispersion curve
efor these
3/2
elementary excitations is isotropic and given by fω(k) ≡ f(k) = bk where k = kx + ky2 .2

For a rectangular sample with dimensions Lx and Ly , the wavevectors allowed by periodic
boundary conditions are kk = (2π/Lx )mx̂+(2π/Ly )nŷ where m and n can take on all positive
and negative integer values.

a) What is the density of allowed wavevectors D(kk) such that D(kk)dkx dky gives the num­
ber of allowed wavevectors in the area dkx dky around the point kk in k-space?

b) Find an expression for the density of states as a function of energy D(f) for the ripplons
in terms of the parameter b and the area A = Lx Ly . Sketch your result.

c) Find an expression for the ripplon contribution to the constant area heat capacity
CA (T ). Leave your result in terms of a dimensionless integral (do not try to evaluate
the integral). How does CA (T ) depend on T ? Sketch the result.

d) Does the system exhibit energy gap behavior? Explain your reasoning.

Problem 2: Two-Dimensional Metal

constant energy
ε contour

ky

ky
kx
kx

We have studied electrons moving in a box in which the potential energy was zero. Alter­
natively one could consider electrons moving in a box containing a periodic potential – a
simple model for the conduction electrons in a metal with a crystalline lattice. Under these
conditions the single particle states can still be indexed by a wavevector kk; however, the
energy of each state f(kk) need not be quadratic in kk nor even isotropic in space.

The figure at the left above shows an approximation to the dispersion relation, f(kk), in a
particular two-dimensional metal∗ . The energy has the form of an inverted square pyramid.
It has four-fold rotational symmetry. Along the kx direction the energy is given by f(kx ) =
γkx . The figure on the right shows a contour of constant energy on the kx , ky plane.

a) If one imposes periodic boundary conditions on the electron wavefunctions in a square

sample of side L, what are the allowed values of the wavevector kk?

b) Find D(kk), the density of allowed wavevectors as a function of kk.

c) Find D(f), the density of single particle states for the electrons as a function of their
energy f. Make a carefully labeled sketch of your result.

d) The metal contains N conduction electrons. Find the Fermi energy fF , the energy of
the last single particle state occupied at T = 0.

e) Find the total energy of the electrons at T = 0 in terms of N and fF .

f) Without doing any calculations, indicate how the electronic heat capacity depends on
the temperature for temperatures T << fF /kB .

g) What is the surface tension S (the negative of the spreading pressure) of the electron
gas at T = 0?

∗
Two dimensional planes of conduction electrons are not a fiction. They play an important role in
semiconductor electronics and in high temperature superconductivity.

Problem 3: Donor Impurity States in a Semiconductor

In lecture we studied the behavior of electrons in an intrinsic (pure) semiconductor. We

found the temperature dependence of the number of electrons in the conduction band and
found an expression for the temperature dependence of the chemical potential. Intrinsic
semiconductors are nice systems to study, but not very useful for semiconductor electronics
because of the strong dependence of the number of carriers on temperature. One would pre­
fer a semiconductor where the number of carriers depends only weakly on temperature, but
can be influenced by applied voltages through the position of the conduction band relative
to the chemical potential. This can be done by doping the semiconductor with impurities
that can easily donate an outer electron to the fluid of mobile electrons in the conduction
band. The density of states for such a system is shown below.

D(ε) doped semiconductor

εF ε

On the left one has the density of states associated with the top of the valence band. All
these states are occupied at T = 0. The delta function represents ND impurity states (one
electron per impurity) located somewhere within the energy gap of the intrinsic material.
They too are completely occupied at T = 0. On the right one has the density of states
associated with the bottom of the conduction band. These states are completely vacant at
T = 0. Considering the step-wise nature of the mean occupation number < n > at T = 0,
the Fermi energy fF (the chemical potential at T = 0) must lie above the energy of the
impurities and below the bottom of the conduction band.

As the temperature is raised from zero, electrons will first be thermally promoted from
impurity states into the valance band and then at a much higher temperature, after the
impurities have lost almost all of their electrons, electrons will be promoted from the valance
band to the conduction band. Analysis of this situation is complicated by the fact that
during this process the chemical potential will drop through and below the impurity energy.
One may be able to use approximations for < n > in the two bands, but one has to use
the full form for < n > applied to the impurity states. We will study some aspects of a
simplified model of this situation.

We will neglect the valance band entirely. Set f = 0 at the position of the impurity states
and let Δ be the position of the bottom of the conduction band.

D(ε)

0 ∆ ε
µ εF

The exact position of the Fermi energy fF is yet to be determined, and the location of the
chemical potential µ will change with temperature. Analytically we have
� �3/2
V 2me
D(f) = ND δ(f) + 2 (f − Δ)1/2
2π f2

where the second term is only present where f > Δ. In the following we will make the
approximation, realized in practice, that Δ » kB T . Before starting, you should review the
intrinsic semiconductor example presented at the end of Lecture 19.

a) A useful parameter
For a fixed value of the chemical potential, show that the number of electrons in the
conduction band, NC can be written in the form

NC = α(T )e−Δ/kB T eµ/kB T

where α(T ) is a useful dimensionless parameter, proportional to T 3/2 , with magnitude

comparable to the number of donors. Using this parameter will simplify subsequent
calculations.

b) A reference temperature
Find an equation, which when solved, gives the temperature at which exactly one half
of the donor sites are empty. Do not attempt to solve the equation.

c) Behavior of µ(T ) at very low temperatures

Assume that µ starts out for very low temperatures somewhere near the center of the
region between f = 0 and f = Δ so that we can use the approximate expressions for
< n > both at the donor energy f = 0, and inside the conduction band. In this case,
find an analytic expression for µ(T ). Use the fact that limx→0 (x ln x) = 0 to show that
the Fermi energy for this system is given by fF = Δ/2.

 d) Full equation for NC

First find the number of electrons on the donors using the exact expression for < n >.
This will involve e−µ/kB T . You can eliminate e−µ/kB T from the result by inverting the
expression you found in a) to isolate e−µ/kB T , then substituting this into your first
result. Next use the fact that the number of electrons on the donors plus NC must
equal the total number of electrons ND . Show that (after a fair amount of algebra)
this leads to the quadratic equation for NC /ND
�
 2
� �
 �

NC NC α −Δ/kB T α −Δ/kB T
+ e − e =0
ND ND ND ND

Check to see that this is consistent with the result you found in b). Without solving
the quadratic equation show that it gives the expected results limT →0 NC /ND → 0
and limkB T » Δ NC /ND → 1.

Problem 4: Spin Polarization

Consider a 3-dimensional non-interacting quantum gas of s = 1/2 Fermions. The two possible
spin states are ms = 1/2 and ms = −1/2. In a uniform magnetic field Hẑ the single particle
energies depend on the direction of the spin relative to the field: f = f2 k 2 /2m − 2µ0 Hms
where µ0 is a constant with the units of magnetic moment.

a) Find the density of states as a function of energy separately for the particles with spin
parallel and antiparallel to the external field, D1/2 (f) and D−1/2 (f). Make a careful
sketch of the total density of states as a function of energy D(f) = D1/2 (f) + D−1/2 (f).

b) For N particles in a volume V at absolute zero (T = 0), find an expression for the
minimum magnetic field H0 that will give rise to total polarization of the spins, that
is no particles left with the high energy spin orientation.

c) Evaluate H0 in Tesla (one Tesla = 104 Gauss) for the electrons in copper where the
electronic magnetic moment µ0 = −9.27 × 10−21 ergs-gauss−1 , the particle mass m =
9.11 × 10−28 grams and the number density of conduction electrons is n ≡ N/V =
8.45 × 1022 cm−3 .

d) Evaluate H0 in Tesla for liquid 3 He where the magnetic moment is that of the nucleus
with µ0 = 1.075 × 10−23 ergs-gauss−1 , m = 5.01 × 10−24 grams and the number density
of atoms in the liquid is n = 1.64 × 1022 cm−3 .

One can now buy commercial superconducting solenoids for laboratory research that go up
to about 15 Tesla, and at the National High Magnetic Field Laboratory one can obtain
continuous fields up to 45 Tesla and pulsed fields as high as 100 Tesla. As you can see, we
are a long way from being able to completely polarize either of these systems by brute force.

Problem 5: T = 0 solubility of 3 He in 4 He

Helium has two stable isotopes. The common isotope, 4 He, has nuclear spin 0 and is a
composite Boson. The rare lighter isotope, 3 He, has nuclear spin 1/2 and is a composite
Fermion. Helium atoms interact by a pairwise attractive Van der Waals interaction which is
the same for both isotopes.

At very low temperatures a liquid mixture of the two isotopes separates into a 3 He rich
phase and a 4 He rich phase as shown above. At T = 0 the 3 He rich phase becomes pure 3 He;
however, the 4 He rich phase can contain 3 He up to a critical concentration N3 /(N3 + N4 ) =
n3 /(n3 +n4 ) = xC . Here n3 and n4 represent the number densities, N/V , of the two isotopes.

The finite solubility of 3 He in 4 He at T = 0 can be understood in terms of a simple model.

Because the quantum mechanical zero point motion of the lighter isotope is greater than
that of the heavy isotope, the atoms in pure liquid 3 He are farther apart than the atoms in
liquid 4 He. As a result, the energy of a single 3 He atom dissolved in bulk 4 He is lower than
it would be if added to bulk 3 He due to the closer proximity of its neighbors. This can be
viewed as a potential energy difference of W between the two phases as shown above.

As more 3 He atoms are dissolved in the 4 He rich phase they behave as free particles with
an effective mass m∗ arising from their motion through the background fluid. A dissolved
3
He atom with momentum fk has an associated kinetic energy f(k) = f2 k 2 /2m∗ . 3 He will
dissolve in the 4 He until the total energy, potential plus kinetic, necessary to add another
atom exceeds that of having it join the bulk 3 He liquid.

a) Find an expression for the maximum number density of 3 He atoms that can be dissolved
in 4 He at T = 0.

b) Check to see if your answer in a) is reasonable by estimating xC using the following

values: f ≈ 1 × 10−27 erg-sec, kB ≈ 1 × 10−16 erg-o K −1 , m∗ ≈ 10 × 10−24 grams,
W ≈ 0.5 o K, and n4 ≈ 2 × 1022 cm−3 .

Problem 6: Melting Curve of 3 He

I II III

3
He is a composite Fermion (nuclear spin = 1/2) that remains a liquid under its own vapor
pressure down to T = 0. When compressed at low temperatures it solidifies at pressures of
the order of 30 atmospheres. A phase diagram showing part of the melting curve is shown
above. The phase diagram can be divided into three regions – I, II, and III – according to
the temperature.

In region III the temperature is too high to allow us to make simple models. In region II
the liquid can be treated as a degenerate spin 1/2 Fermi gas with temperature so low that
it can be considered to be at T = 0. The solid can be considered to be a collection of
non-interacting spin 1/2 paramagnets fixed on specific lattice sites (thus the 3 He atoms in
the solid phase are distinguishable). There is no magnetic field applied to the system.

a) What is the entropy of N atoms of liquid in region II?

b) What is the entropy of N atoms of solid in region II?

In region I the liquid still behaves as a degenerate spin 1/2 Fermi gas. The nuclear spins
in the solid become ordered in an antiferromagnetic state. In this state each nuclear spin
behaves as if it were in an effective magnetic field Hef f (this effective field is caused by the
alignment of all the other moments in the sample). We can assume that µHef f >> kT in
region I.

c) What is the entropy of the solid in region I (basically at T = 0)?

(continued on next page)

The slope of the melting curve is given by the expression

dPM ΔSs→l
=
dT ΔVs→l
where ΔSs→l is the difference between the entropy of the liquid and that of the solid and
ΔVs→l is the volume difference between the liquid and the solid. In 3 He, as in most well
behaved systems, ΔVs→l is positive and we can assume that it is a temperature independent
quantity N Δvo .

d) What is the slope of the melting curve in region II?

e) What is the slope of the melting curve in region I?

f) Assume that the system is on the melting curve in region II, that is, the system has
both liquid and solid coexisting in equilibrium. If heat is added to the system, will
some of the solid melt or will some of the liquid solidify? Explain why.

MIT OpenCourseWare
http://ocw.mit.edu

8.044 Statistical Physics I

Spring 2013

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.