You are on page 1of 10

Quantum Physics III (8.

06) Spring 2003

Assignment 10

Read Cohen-Tannoudji, Chapter VIII and Griffiths Chapter 11.
Read Griffiths Chapter 9 on Time Dependent Perturbation Theory.

Problem Set 10 and Study Guide

The rst six problems are your problem set. The last ve
problems will not be graded, but should help you to study time dependent perturba
tion theory for the nal exam. Solutions to all eleven problems will be provided.

1. The Size of Nuclei (8 points)

In lecture we derived an expression for the scattering amplitude in the Born
approximation for the elastic scattering of a particle of mass m and charge -|e|
from a charge distribution |e|(r):


= 2 2
f (q) d3 reiqr


Recall that q = k k is the momentum transferred to the scattered particle

= 2|k| sin(/2). If the electrons
in the collision. For elastic scattering, q |q|
used in a scattering experiment are relativistic, k E/c.

(a) The charge distribution of a nucleus is not localized at a mathematical

point. f is therefore not exactly that for Rutherford scattering. The
charge distribution is roughly constant out to a radius R and then drops
rapidly to zero. A simple model is:
for r R

and = 0 for r > R. Calculate the cross section for electron scattering
from such a nucleus as a function of q 2 .
(b) The ratio of the actual amplitude for scattering from a point nucleus is
called the form factor. Sketch the form factor as a function of qR.
The form factor tells us about the shape of the charge distribution in a
nucleus, and thus tells us how the protons within a nucleus are arranged.
In our simple model, the form factor tells us the value of R. If nuclei
had precisely the shape we have used in our simple model, experimenters
would measure a form factor with precisely the functional form you have
calculated, and would then do a t to obtain a measurement of R, the
radius of the nucleus.
(c) For relativistic electrons with energy E, if you are able to count the scat
tered electrons at a variety of angles, ranging from close to zero to close
to , what range of q can you access? If you use electrons with E 1/R,
show that you will not be able to make an accurate determination of R.
You will not be able to resolve the fact that scattering o a nucleus
diers from Rutherford scattering.
The values of R for nuclei are around (27)1013 cm. Roughly how large
an electron energy do you need in order to do a reasonable measurement
of R?

First aside: The above problem uses a simple model, but it is not all that
far from the real thing. I am attaching copies of some gures of real data on
electron-nucleus scattering, along with the inferred nuclear charge distributions.
Second aside: The next step in the process of unveiling the structure of matter
on smaller and smaller length scales was the discovery that the protons and
neutrons that make up a nucleus have substructure. Electron beams with en
ergies appropriate for studying nuclear structure (ie the distribution of protons
within a nucleus, which youve been analyzing in this problem) cannot resolve
the substructure of a proton. Thus, the discovery of the quark structure of the
proton had to wait until the construction of the SLAC linear accelerator, which
began accelerating electrons to 18 GeV in the late 1960s. In 1967, Jerome
Friedman, Henry Kendall and Richard Taylor began the series of experiments
in which quarks were discovered. When an 18 GeV electron scatters at large

angles o a quark in a proton, the proton does not remain intact. This means
that the description of these experiments requires an understanding of inelastic
scattering. In an inelastic collision, the scattered electrons momentum changes
by q, and its energy also changes.

2. Partial Waves (10 points)

Suppose the scattering amplitude for a certain reaction is given by

1 k 2ik3 3
f () = + 3e sin 2k cos (1)
k k0 k ik

where , k0 , and are constants

characteristic of the potential which produces
the scattering. Of course k = 2mE/2 is the deBroglie wavenumber.

What partial waves are active (i.e. what values of )?

What are the phase shifts in the active partial waves? Do they have the
proper behavior as k 0?
What is the dierential cross section, d/d for general values of k?
What are the partial wave cross sections, ?
Assume k03 1. Give an approximation to the total cross section (k)
for k k0 .
What is the total cross section for general values of k? What is the imagi
nary part of the forward scattering amplitude? Do they satisfy the optical

3. Scattering from a -Shell (15 points)

Consider s-wave ( = 0) scattering from the potential

V (r) = (r R)


with a large positive constant. To nd the phase shift 0 (k) we have to solve

d2 u
+ k 2 u = (r R)u ,
dr R
with u = 0 at r = 0 and u = sin(kr + ) for r > R.

(a) What is u in r < R?

(b) By comparing u (r)/u(r) just inside and just outside r = R, nd a formula
to determine .
(c) Find the scattering length a, dened by limk0 0 = ka.

(d) Assume 1. Sketch (k). Show that for kR just below n, with n a
positive integer, (k) increases very rapidly by (as kR increases towards
n). Sketch the s-wave cross-section 0 . Show that the s-wave scattering
from this potential is the same as that from a hard sphere of radius R for
all values of kR except those such that kR is close to n. What is the
signicance of these values?
4. Ramsauer-Townsend Eect (8 points)
At very low energies only the s-wave contributes to scattering. If, for some
reason, the s-wave phase shift vanishes, then so does the scattering amplitude.
Under these circumstances a projectile can pass through material without any
scattering. This eect is known as the Ramsauer-Townsend Eect.
Consider a three dimensional square well,
V0 for r a
V (r) = (2)
0 for r > a
(a) Find the condition on 2 = 2mV0 a2 /2 such that the cross section for a
particle of mass m is zero at zero energy. Your answer should be in the
form of a set of values of 2 , specied graphically and with the rst few
numerical values given.
(b) As you can see from part (a), it is useful to think of the Ramsauer-
Townsend eect as a function of the depth of the potential. The existence
of bound states is also a function of the depth of the potential. Show that
if a square well which displays an exact Ramsauer-Townsend eect is made
a little deeper or shallower (you have to gure out which) it then has a
bound state at threshold.
5. Scattering in the Semiclassical Approximation (4 points)
The semiclassical approximation becomes better at high energies. For most
problems high energies means scattering as opposed to bound states. It is quite
straightforward to estimate the phase shift in the semiclassical approximation.
Consider scattering in the s-wave in three dimensions. The radial wavefunction
u (k, r) + 2 V (r)u(k, r) = k 2 u(k, r) (3)

and u(k, 0) = 0.
In this problem we will assume that V (r) is smooth and slowly varying and that
r2 V (r) 0 as r and that V (r) is negative at all r. (Well change the last
assumption in the next problem.)
Recall from 8.05 (and show for yourself if you like) that as r , u(k, r)
sin(kr + 0 (k)), where 0 (k) is the phase shift.

Show that in the semiclassical approximation

0 (k) = dr k 2 2 V (r) k (4)

6. A Semiclassical Analysis of Resonant Scattering (15 points)

Consider s-wave scattering for a particle of mass m o a potential V (r) which
vanishes at the origin, rises steadily as r increases from zero, reaches a maximum
at r = c, and then goes quickly to zero as r increases further.
For = 0, the radial wave function u(r) satises the same Schrodinger equation
as that for a particle in one dimension with potential V , subject to the boundary
condition u(0) = 0.
Consider scattering with energy E where 0 E V (c). The classical turning
points are at r = a and r = b with a < c < b.

(a) What is the semiclassical approximation to the wave function in the clas
sically allowed region, 0 r < a?
(b) What is the ratio of the amplitude of the wave function u(r) in the semi-
classical approximation in the region x > b compared to that in the region
x < a, for generic values of E?
(c) For some special values of E, there is a qualitative change in the ratio
of the amplitude for x > b to the amplitude for x < a, compared to its
generic value at other energies. What condition determines these special
values of E?
(d) Describe the qualitative behavior of the s-wave phase shift and s-wave
cross-section for energies in the vicinity of the special values of E.

7. Brick in a Square Well

Here is a simple enough time dependent perturbation of a simple enough system
that everything can be computed analytically.
Do Griths Problem 9.17.

8. Excitation of a hydrogen atom

A hydrogen atom is placed in an electric eld E(t) that is uniform and has the
time dependence,

E(t) =0 t<0
0 et
=E t>0 (5)

What is the probability that as t , the hydrogen atom, initially in the

ground state, makes a transition to the 2p state?

9. Decay of the three dimensional harmonic oscillator

The object of this problem is to calculate the lifetime of a charged particle
(charge q, mass m) in the rst p-state of the three dimensional harmonic oscil
lator (frequency ).

a) Write down an expression for the transition rate per unit time, (2p 1s),
for the particle to spontaneously emit electromagnetic radiation and make
a transition to the ground state. should depend on the frequency of the
emitted light and on the matrix element of the operator qr.

Note that the 1p state is three-fold degenerate: it has = 1 and can have
m = 1, 0, 1.

b) Show that the transition rate is independent of m .

c) Finally, give a formula for (2p 1s) in terms of m, , q, and fundamental
d) What is the relationship between the transition rate per unit time and the
lifetime of the 2p state?

10. Lifetime of Excited States of Hydrogen

(a) Do Griths Problem 9.10. You may use any results from Griths sec
tion 9.3.3 without proving them (even though Problem 9.10 comes before
section 9.3.3).
(b) Do Griths Problem 9.13, part (a) only.

11. A wave front crossing a bound particle

Consider a particle in one dimension moving under the inuence of some time-
independent potential, V (x). Assume that you know the energy levels and
corresponding eigenfunctions for this problem. We now subject the particle to
a traveling pulse represented by a space- and time-dependent potential,

V (t) = a(x ct)

where (x) is a Dirac -function.

(a) Suppose as t the particle is known to be in the ground state whose

wavefunction is x|i = ui (x). Find the probability for nding the system
in some excited state, with wavefunction x|f = uf (x) as t .
(b) (Note: you can skip part (b) and still do part (c). You may have to do
this, depending on exactly what I am able to cover in lecture.) Reinter
pret your result in part (a) as follows. Regard the -function pulse as a
superposition of harmonic perturbations, by recalling that the function
can be represented as a superposition of exponentials:

(x ct) = dei(x/ct) . (6)

Show that if you treat each frequency component of the function sep
arately, using for each the result we obtained in lecture for a harmonic
perturbation (namely that there is a transition if and only if = f i
and the amplitude of that transition is the matrix element of the operator
coecient of the harmonic time dependence between the initial and nal
states) then you get the same result as in part (a).
The lesson is that the analysis we did in lecture with a harmonic time
dependence can be applied to very dierent dime dependences via Fourier
(c) Apply the result of part (a) to the one dimensional (innite) square well,

V (x) = 0 for 0 < x < d,

= for x < 0 or x > d (7)

Express the probability to transition from the ground state to the rst
excited state as a function of the dimensionless parameters = c and
dE 2
3 2
= 2c , where E = 2md2 . Show that the transition probability has a
maximum for 1. Explain this in terms of the time it takes light to
cross the potential well and the natural timescale of the quantum system.