Quantum Physics III (8.

06) — Spring 2016

Assignment 2

Readings

This week, and in general, you should consider the Griffiths reading as required and the
others as optional.

• Griffiths, Chapter 8
• Cohen-Tannoudji, Chapter XII
• Shankar, Section 16.2

Problem Set 2

1. Stark Effect (Adapted from Griffiths Problem 6.36, 35 points)

~ ext , the energy levels are
When an atom is placed in a uniform external electric field E
shifted, a phenomenon known as the Stark effect. This is the electrical analog to
the magnetic Zeeman effect. In this problem, you will analyze the Stark effect for the
n = 1 and n = 2 states of hydrogen. Let the electric field point in the ẑ direction, so
the electrostatic potential of the electron is

δHStark = eEext z (1)

Treat this as a perturbation on the Bohr Hamiltonian

p~ 2 e2
H0 = − , (2)
2m r
for now ignoring spin and fine structure effects.

(a) [This part is just math, but will make the later calculations much easier.] Suppose
that a, b, c are nonnegative integers and f (r) is a function. Prove that
Z
dx dy dz xa y b z c f (r) = 0 (3)
p
unless a, b, c are each even. Here r ≡ x2 + y 2 + z 2 and you may assume that
f (r) is a function such that the integral in (3) is always defined.
(b) Show that the ground state energy E1,0,0 is not affected by the perturbation in
(1), to first order in perturbation theory.
(c) The second-order shift to E1,0,0 is nonzero and is not so easy to calculate. In this
part you will compute a bound on the shift.

1
Quantum Physics III (8.06) — Spring 2016 Assignment 2

i. Calculate α |hα|z|1, 0, 0i|2 , where α runs over all states of the Hydrogen
P
atom, bound or unbound. [Hint: Dimensional analysis can be a good sanity
check of this result.]
ii. The quantity
1
0
(4)
E1,0,0 − Eα0
is always negative for all α 6= (1, 0, 0). What is the lowest (i.e. closest to −∞)
possible value for (4)?
iii. Conclude by arguing that the second order shift
2
E1,0,0 ≥ −Ca30 Eext
2
, (5)

for some C. What is C?

Discussion: This calculation shows an upper bound on the ground-state po-
larizability of the Hydrogen atom. Why? A system has polarizability α if
applying field E ~ induces dipole moment p~ = αE. ~ A dipole p~ in an electric
~ has energy −p~ · E;
field E ~ thus polarizability α corresponds to energy −αE 2
~ Conversely, if the term −αE 2 appears in a
in an electric field with E = |E|.
Hamiltonian, it corresponds to a system with polarizability α. Carrying out
2
the calculation of E1,0,0 exactly requires integrating over the unbound states,
which is doable using 8.06 methods but a lot of work. This leads to the value
Cexact = 49 . The value of C that you calculate should be consistent with this!
Calculating the polarizability is a step towards calculating the van der Waals
force.
(d) The first excited state of H0 is 4-fold degenerate, with |n, `, mi states |2, 0, 0i,
|2, 1, 1i, |2, 1, 0i, |2, 1, −1i. Using degenerate perturbation theory, determine the
first-order corrections to the energy. Into how many levels do these n = 2 states
split? (Ignore spin.)
(e) What are the “good” wave functions for part (d)? Find the expectation value of
the electric dipole moment (p~e ≡ −e~r) in each of these “good” states.
(f) Now we consider electron spin, and the fine-structure splitting between the 2S
and 2P levels. Calculate the electric field strength (in V/cm) at which the energy
of the Stark shift calculated above becomes equal to the fine-structure splitting
between the 2S and 2P energy levels.
A common household electric field strength is roughly 100V/cm. For example,
the live and neutral wires in an electrical socket have a voltage difference of 110V
and if they are 1.1cm apart then there will be a field of 100V/cm between them.
In this case, a hydrogen atom that happened to be between the two wires would
experience a field of strength 100V/cm. At this field strength would this atom
become polarized according to your result in part (d), or would the fine-structure
splitting be dominant? [Regardless of what you calculate, please do not try any
version of this at home.]

2. Quantum Mechanics of a Bouncing Ball (15 points)

2
Quantum Physics III (8.06) — Spring 2016 Assignment 2

The semiclassical approximation can also be used to estimate the energy eigenvalues
and eigenstates for potentials that cannot be treated exactly so easily. This problem
is loosely based on Griffiths 8.6. (See Griffiths 8.5 if you’d like to learn how to treat
this quantum mechanical problem exactly, using Airy functions.)
Consider the quantum mechanical analogue to the classical problem of a ball of mass
m bouncing elastically on the floor, under the influence of a gravitational potential
which gives it a constant acceleration g.

(a) Find the semiclassical approximation to the allowed energies En , in terms of m,

g, and ~.
(b) Estimate the zero point energy of a neutron “at rest” (i.e. in the quantum me-
chanical ground state) on a horizontal surface in the earth’s gravitational field.
Express your answer in eV. [This may sound artificial to you, but the experi-
ment has been done. See V. V. Nesvizhevsky et al., Nature 415, 297 (2002)
and arXiv:hep-ph/0306198 for an experimental measurement of the quantum me-
chanical ground state energy for neutrons bouncing on a horizontal surface in the
earth’s gravitational field. This experiment got a lot of press at the time, because
it involves both gravity and quantum mechanics, which made for an eye catching
press release. It of course has nothing to do with quantum gravity.]
(c) Now imagine dropping a ball of mass 1 gram from rest from a height of 1 meter,
and letting it bounce. Do the 8.01 “calculation” of the classical energy of the ball.
The quantum mechanical state corresponding to a ball following this classical
trajectory must be a coherent superposition of energy eigenstates, with mean
energy equal to the classical energy. How large is the mean value of the quantum
number n in this state?

3. Application of the Semiclassical Method to the Double Well Potential (25

points)
Do Griffiths Problem 8.15.
This is not as difficult a problem as its length would indicate. Griffiths leads you
through all the steps. This is an instructive problem in quantum dynamics. You
should recall that this is the potential that we used to describe the physics of the
ammonia molecule, early in 8.05. Back then, we had to wave our hands a little when
we talked about tunneling splitting the degeneracy between the even and odd states.
Now, you can do this calculation for real.
Hint for (a) and (b): The steps suggested by Griffiths are: work out the wave function
ψ1 in region (i); from ψ1 use the connection formulae at x2 to obtain the wave function
ψ2 in regions (ii); use ψ2 and the connection formulae at x1 to obtain the wave function
ψ3 in region (iii). Equation (8.59) can be found by requiring that ψ3 should satisfy
ψ3 (0) = 0 or ψ30 (0) = 0 at x = 0.
It is a bit easier (and more transparent) to use a slightly different approach from what
Griffiths suggests. Given that the wave function should be an even or odd function of

3
Quantum Physics III (8.06) — Spring 2016 Assignment 2

x, the wave function in region (iii) can be written down immediately. For example in
the even case,
 Z x 
C 1
ψ(x) = p cosh dy κ(y) , −x1 < x < x1 (6)
κ(x) ~ 0

using our standard notations. (6) is an example where by symmetry, the exponentially
small piece in a classically forbidden region is known exactly. The wave function ψ2 in
region (ii) then can be obtained using two ways: from ψ1 in region (i) via connection
formulae at x2 , or from ψ3 in region (iii) via connection formulae at x1 . The consistency
of two wave functions leads to equation (8.59) of Griffiths.

4. Hydrogen medley (25 points) Let me denote the mass of an electron and e its
charge. The 8.04 version of the Hydrogen Hamiltonian is

p2 e2
H0 = − .
2m r
In this problem we will consider the spin of the electron (whose corresponding operator
~ but we will ignore the spin of the proton.
we call S)

(a) Complete sets of commuting observables (CSCO). A CSCO is a set of

commuting operators whose simultaneous eigenspaces are each one dimensional.
Equivalently, specifying all the eigenvalues of all the operators in a CSCO uniquely
specifies a state (up to multiplication by a scalar). You may use without proof the
fact that H0 , L ~ 2 , Lz , Sz form a CSCO with eigenbasis {|n, l, ml , ms i}. For each of
the following sets of operators, either (a) explain why they are CSCOs and justify
this by expressing its eigenstates in terms of the {|n, l, ml , ms i} eigenstates, or
(b) explain why they are not CSCOs. For simplicity, consider only bound states.
~ 2, L
i. H0 , L ~ · S,
~ Jz .
~ , Lz , Sx .
ii. H0 , L 2

~ 2 , Jz , Sz .
iii. H0 , L
iv. H0 , J~2 , Jz , Sz .
v. H0 , J~2 , L~ · S,
~ Jz .
(b) Strong-field Zeeman effect. In the strong-field Zeeman effect, the unperturbed
eigenstates are |n, l, ml , ms i. The fine structure can be thought of as contributing
a term
m e c2 α 4
 
1 3
Hfs = − − ,
2n3 ĵ + 1/2 4n
where ĵ is an operator satisfying J 2 = ~2 ĵ(ĵ + 1). To compute the first-order
energy shifts here we need to evaluate
1
hn, l, ml , ms | |n, l, ml , ms i. (7)
ĵ + 1/2

4
Quantum Physics III (8.06) — Spring 2016 Assignment 2

i. Use the following strategy to evaluate (7). First compute the expectation
value of J 2 on the |n, l, ml , ms i state. Now imagine that we measure ĵ. Use
your calculation to find the probabilities of the two outcomes j = l + 1/2 and
j = l − 1/2. Finally use the fact that(7) equals

Pr[j = l + 1/2] Pr[j = l − 1/2]

+ (8)
l+1 l
to reproduce the known result for the strong-field Zeeman effect (see lecture
notes or Griffiths).
ii. The shifts in (7) used non-degenerate perturbation theory. But since the
unperturbed energies are a function of ml + 2ms there will be degeneracies.
Explain why, despite this, the above argument is still correct.

5
Quantum Physics III (8.06) — Spring 2016 Assignment 2

5. Tunneling and the Stark Effect (preview)

Do not turn in this problem with pset 2. It will appear on pset 3.
The Stark effect concerns the physics of an atom in an electric field. In this problem,
you will explore the possibility that in an electric field, the electron in an atom can
tunnel out of the atom, making the atomic bound states unstable. We consider this
effect in a simpler one-dimensional analog problem.
Suppose an electron is trapped in a one-dimensional square well of depth V0 and width
d:

V (x) = −V0 for |x| < d/2

= 0 for |x| ≥ d/2 .

Suppose a weak constant electric field in the x-direction with strength E is turned on.
That is V → (V − eEx). Assume throughout this problem that eEd  ~2 /2md2  V0 .

(a) Set E = 0 in this part of the problem. Estimate the ground state energy (i.e. the
amount by which the ground state energy is above the bottom of the potential
well) by pretending that the well is infinitely deep. (Because ~2 /2md2  V0 ,
this is a good approximation.) Use this estimate of the ground state energy in
subsequent parts of the problem. Note that the true ground state energy is lower
than what you’ve estimated, why?
(b) Sketch the potential with E = 6 0 and explain why the ground state of the E = 0
potential is no longer stable when E 6= 0.
(c) Use the semiclassical approximation to calculate the barrier penetration factor for
the ground state. [You should use the fact that eEd  ~2 /2md2 to simplify this
part of the problem.]
(d) Use classical arguments to convert the barrier penetration factor into an estimate
of the lifetime of the bound state.
(e) Now, lets put in numbers that are characteristic of an atomic system. Calculate
the lifetime for V0 = 20 eV, d = 2 × 10−8 cm and an electric field of 7 × 104 V/cm.
Compare the lifetime you estimate to the age of the universe.
(f) Show that the lifetime goes like exp(1/E), and explain why this result means that
this “instability” could not be obtained in any finite order of perturbation theory,
treating E as a perturbation to the Hamiltonian.

6
MIT OpenCourseWare
http://ocw.mit.edu

8.06 Quantum Physics III

Spring 2016

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