Kelley, J. Daniel - Leventhal, Jacob Joseph-Problems in Classical and Quantum Mechanics - Extracting The Underlying Concepts-Springer (2017)

J. Daniel Kelley · Jacob J.
Leventhal
Problems in
Classical and
Quantum
Mechanics
Extracting the Underlying Concepts
Problems in Classical and Quantum Mechanics
J. Daniel Kelley • Jacob J. Leventhal
Problems in Classical
and Quantum Mechanics
Extracting the Underlying Concepts
123
J. Daniel Kelley Jacob J. Leventhal
The Boeing Company (retired) Department of Physics and Astronomy
University of Missouri-St. Louis University of Missouri-St. Louis
St. Louis, MO, USA St. Louis, MO, USA
ISBN 978-3-319-46662-0 ISBN 978-3-319-46664-4 (eBook)

DOI 10.1007/978-3-319-46664-4
Library of Congress Control Number: 2016951927
© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, express or implied, with respect to the material contained herein or for any
errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
This book is dedicated to our families.
Preface
This book is not a textbook. It is a collection of problems intended to aid students in

their undergraduate and graduate level courses in physics. The book was, however,
formulated with students who are preparing for the PhD qualifying exam in mind.
Thus, the problems that are included are of the type that could be on this exam or
are problems that are meant to elucidate an important principle.
There are many compilations of physics problems available to students, so it is
reasonable to ask why this one is different. It is different because the aim is to place
the problems in the broader context of the subject. The book is meant to facilitate
the development of problem-solving skills to aid in the understanding of physics.
We state the problem and then present the solution in detail. Further, we note and
discuss the significance of the problem in the context of the subject under study.
We analyze the broader implications of the solution including limiting cases and the
relation to other problems. Many of the solutions are accompanied by a tutorial on
their meaning and the route to solution. We stress that the solution of the problem is
just the beginning of the learning process. As the subtitle infers, manipulation of the
solution and changing the associated parameters can provide a great deal of insight.
Our approach is to make the discussions of each problem seem as though the
student has come to one of our offices and asked for help solving it. It is our belief
that when students come for help, it is not enough to simply show them how to arrive
at the solution. We discuss with them the physics that they should be learning from
the problem. That is, after all, why problems are assigned. We want the students to
ask themselves “What physics can I learn from the problem?” not “How can I work
this problem and go on to the next one as quickly as possible?”
Another feature of this book is the inclusion of more mathematical detail in the
solutions than is usually provided. We have done this because the book is meant to
be an aid in learning physics. Thus, at the risk of lengthening the book, we have
attempted, when possible, to relieve the reader of the burden of spending a lot of
time on mathematical detail after the plan of attack has been formulated.
We could not include every aspect of the subjects that we treat. Our choice of
material was predicated on the notion that introductory concepts are the most vital.
Some will argue with our choices, but we have made them. Another consideration
vii
viii Preface
we used when choosing material was, as noted above, our perception of those
subjects most often included in PhD qualifying examinations. On occasion, we
have provided introductions to chapters, sometimes merely for the sake of notation.
None of these introductions should be considered substitutes for the appropriate
textbooks.
Advice for Students
Solving problems is an integral part of learning physics. Over the years, we have
heard many times “I understand the physics, but I can’t work the problems.” There
is no polite response to this beyond “Nonsense—you do not yet understand the
physics.” This does not mean that you should have been able to work every problem
the first time you encountered it. On the contrary, it is our experience that the
concepts retained best are those that you comprehend only after having struggled
with them. As in the weight room, “no pain, no gain.” A related excuse is “I
understand the physics, but I can’t do the math.” This is tantamount to saying that
you are a very good auto mechanic, but you really have trouble using wrenches.
Mathematics is the essential tool of physics.
We offer several hints for learning and retaining physics concepts. First, neatness
does count! When you arrive at a satisfactory solution to a problem, we urge you
to rewrite the solution in a clear comprehensible form. This permits you to review
the solution and to decipher it weeks later when you are studying for an exam.
Additionally, writing the solution in a coherent fashion is good practice for the exam.
After all, you want to transmit to the grader what you know. It does you no good to
know the material if the exam grader does not know that you know it.
Physics is not like some other disciplines. Rule: You cannot cram for a physics
test. At least the vast majority of people cannot cram for a physics test. You can
probably memorize the presidents and vice presidents of the United States or some
list of dignitaries or geographical locations the night before an examination (and
likely forget it after the exam), but learning a number of new concepts in a short
time is very difficult. As students we have tried it (unsuccessfully) and wish to pass
along our sad experiences. In short, it is worthwhile to stay as up to date as possible
in physics courses.
Finally, we offer some advice on the best way to use instructor-provided solutions
to assigned problems, a common aid in contemporary education. In our view, these
solutions are a two-edged sword because they can lull a student into a false sense
of security, thinking that they understand the material when, in fact, they do not.
Often students will simply read the problem, think it over for a short time, and then
peruse the solution. After a short digestion period, they think that they understand
the solution and go on to the next problem. They have ignored the “no pain, no
gain” rule. Glancing at the solution is like watching someone do push-ups. It does
not benefit the spectator. Proper use of provided solutions takes a great deal of
discipline. We recommend that the problem first be attempted without consulting
Preface ix
the solution. If the student reaches an impasse, after perhaps a half hour, the solution
should be consulted. Do not, however, look beyond the portion of the solution that
resolved the impasse. Discipline! The student should then continue this procedure
until the solution has been attained and (in all likelihood) understood. We contend
that this method will lead to a better understanding of physics and permanent
retention of the concepts that the problem was designed to illustrate. Yes, it takes
more time this way, but the rewards are worth it.
St. Louis, MO, USA J. Daniel Kelley

Jacob J. Leventhal
Contents
Part I Classical Mechanics

1 Newtonian Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1. Simple pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Free fall with drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3. Collisions—2 balls and a brick wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4. Collisions—two blocks and a spring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
5. Two unequal masses attached to a spring . . . . . . . . . . . . . . . . . . . . . . . . 11
6. Hole dug through a diameter of the earth . . . . . . . . . . . . . . . . . . . . . . . . 12
7. Hole dug through a chord of the earth . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
8. Sphere with a spherical (off center) hole . . . . . . . . . . . . . . . . . . . . . . . . 15
9. Moving inclined plane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
10. Particle moving in a cosine potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
11. Particle moving in inverted Gaussian potential . . . . . . . . . . . . . . . . . 21
12. Hard sphere scattering—classical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 Lagrangian and Hamiltonian Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1. Brachistochrone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2. Lagrange/Newton equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3. SHO - Lagrangian and Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4. Simple pendulum - Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5. Simple pendulum with vertically moving pivot point . . . . . . . . . . . 35
6. Sliding pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7. Atwood’s machine I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
8. Mass on table, pulley, hanging mass - Lagrangian . . . . . . . . . . . . . . . 41
9. Mass on table, pulley, hanging mass - Hamiltonian . . . . . . . . . . . . . 43
10. Projectile motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
11. Hanging disk - Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
12. Hanging disk - Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
13. Rod pivoted at the end . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
14. Moving inclined plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
15. Rotating massless rod - Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
xi
xii Contents
16. Rotating massless rod - Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

17. Atwood’s machine II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
18. Sphere rolling on inclined plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
19. Bead on a wire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
20. Two unequal masses attached to a spring . . . . . . . . . . . . . . . . . . . . . . . 64
3 Central Forces and Orbits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
1. Conservation of angular momentum I . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2. Conservation of angular momentum II . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3. Kepler’s laws. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4. Newton’s gravitational law deduced from Kepler’s laws . . . . . . . . 71
5. Total energy for a central potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6. Equation of the orbit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7. Total energy for a circular orbit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
8. Spiral orbit I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
9. Spiral orbit II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
10. Inverse cube force law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
11. Isotropic oscillator I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
12. Isotropic oscillator II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
13. Kepler effective potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
14. Kepler orbits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
15. Runge–Lenz vector I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
16. Runge–Lenz vector II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4 Normal Modes and Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
1. Two pendulums coupled by a spring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
2. Three springs, two masses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3. Double pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4. Pendulum with oscillating pivot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5. Two springs with thin rod attached . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
6. Three particles, two springs (CO2 ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Part II Quantum Mechanics

5 Introductory Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
1. Bohr atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
2. Gravitational bohr radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3. deBroglie wavelength of an electron in Bohr orbit . . . . . . . . . . . . . . 125
4. Uncertainty principle and the H-atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5. Classical radius of the electron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6. Magnetic dipole moment of spinning charged sphere . . . . . . . . . . 129
7. Uncertainty principle and the SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
8. Measurement, expectation values and probabilities . . . . . . . . . . . . . 131
9. Expectation values: SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
10. Expectation values: L-box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
11. Rectangular barrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Contents xiii
12. Transmission and reflection over a square well . . . . . . . . . . . . . . . . . 141

13. Delta function barrier I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
14. Delta function barrier II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
15. Semi-infinite barrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6 Bound States in One Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.1 Degeneracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.2 Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.3 Characteristics of the Eigenfunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.4 Superposition Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
1. Stationary states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
2. Characteristics of eigenfunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
3. Delta function potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
4. Half-well potential. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
5. Sudden approximation I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
6. Sudden approximation II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
7. Wave function, probabilities I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
8. Wave function, probabilities II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9. Sudden approximation III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
7 Ladder Operators for the Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . 173
1. Commutation of x and p . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
2. General relations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
3. Expectation values I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4. Expectation values II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
5. Expectation values III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
6. Matrix elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7. Expectation values IV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
8. Expectation values V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
8 Angular Momentum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
1. Operators and ladder operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
2. Angular momentum eigenvalues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
3. Measurements and expectation values. . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
4. Speed of a “spinning hard sphere” electron . . . . . . . . . . . . . . . . . . . . . . 199
5. Spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
6. Using Pauli matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
7. Spin ladder operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8. Electron in a B field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
9. Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
10. Construct a Clebsch–Gordan table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
11. More Clebsch–Gordan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
12. Hyperfine structure of H-atom I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
13. Hyperfine structure of H-atom II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
14. Hard sphere scattering - quantum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
xiv Contents
9 Indistinguishable Particles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

1. Bosons, fermions and the exchange force . . . . . . . . . . . . . . . . . . . . . . . . 218
2. Two fermions in an SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
10 Bound States in Three Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
1. 3D L-box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
2. Density of states for a 3D L-box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
3. Rigid rotor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
4. Infinitely deep spherical well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
5. Spherical shell well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
6. Classically forbidden region: H-atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
7. Expectation value for H-atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
8. Kramer’s relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
9. Bound state for unknown potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
10. Relation between H-atom quantum numbers . . . . . . . . . . . . . . . . . . . 235
11. H-atom degeneracy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
12. The quantum defect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
11 Approximation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
11.1 The WKB Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
1. Energies of an L-box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
2. Energies of an SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
3. Transmission through a parabolic barrier . . . . . . . . . . . . . . . . . . . . . . . . 246
11.2 The Variational Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
1. Estimate the ground state energy of an SHO . . . . . . . . . . . . . . . . . . . . . 249
2. Estimate the first excited state energy of an SHO . . . . . . . . . . . . . . . . 251
3. Linear potential well. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
4. Quartic potential well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
11.3 Non-degenerate Perturbation Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
1. Gravitational correction to H-atom energy . . . . . . . . . . . . . . . . . . . . . . . 259
2. L-box perturbed by a delta-function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
3. L-box perturbed by a linear function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
4. Relativistic correction to L-box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
5. Charged SHO perturbed by constant electric field . . . . . . . . . . . . . . . 266
6. H-atom perturbed by delta function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
7. Relativistic correction to SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
8. Relativistic correction to H-atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
9. Quartic perturbation of SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
10. Matrix eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
11.4 The Helium Atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
1. He-atom ground state energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
2. He-atom: perturbation/variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
11.5 Degenerate Perturbation Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
1. 2-D a-box with xy perturbation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
2. Particle on a ring with delta-function perturbation. . . . . . . . . . . . . . . 288
Contents xv
3. 2-D SHO with xy perturbation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292

4. Select set for 3x3 matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
11.6 Time Dependent Perturbation Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
1. Gaussian perturbation applied to particle in an L-box . . . . . . . . . . . 298
2. E-field pulse applied to to particle in an L-box . . . . . . . . . . . . . . . . . . 300
3. Exponential perturbation applied to SHO . . . . . . . . . . . . . . . . . . . . . . . . 302
4. Pulse applied to SHO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
5. Exponential E-field applied to H-atom . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
6. Fermi’s Golden Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
A Greek Alphabet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
B Acronyms, Descriptors and Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

B.1 Acronyms and Descriptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
B.2 Coordinate Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
C Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
D Conic Sections in Polar Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
E Useful Trigonometric Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
F Useful Vector Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
G Useful Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
H Useful Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329

H.1 Taylor Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
H.2 Binomial Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
I -Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
I.1 Integral -Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
I.2 Half-Integral -Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
J The Dirac Delta-Function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
K Hyperbolic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

K.1 Manipulations of Hyperbolic Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
K.2 Relationships Between Hyperbolic and Circular Functions . . . . . . . . 337
L Useful Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339

L.1 Classical Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
L.2 Quantum Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
L.2.1 One Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
L.2.2 Three Dimensions (Central Potentials) . . . . . . . . . . . . . . . . . . . . 341
xvi Contents
M The Infinite Square Well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343

M.1 The L-Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
M.2 The a-Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
N Operators, Eigenfunctions, and Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . 347

N.1 Eigenfunctions and Eigenvalues of Operators . . . . . . . . . . . . . . . . . . . . . . 347
N.2 Operator Algebra; Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
N.3 Commutator Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
N.4 Some Quantum Mechanical Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . 348
O The Quantum Mechanical Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . . . . . 349
P Legendre Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351

P.1 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
P.2 Legendre Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
P.3 The Function 1= jr1 r2 j . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
P.4 Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Q Orbital Angular Momentum Operators in Spherical Coordinates . . . 355
R Spherical Harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
S Clebsch–Gordan Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
T The Hydrogen Atom. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Part I
Classical Mechanics
Chapter 1
Newtonian Physics
Most of the problems in this chapter are the type assigned in introductory courses
on Classical Mechanics. Their inclusion in this volume serves two purposes. First,
they provide a “warm-up” for more advanced problems. More importantly, a
number of these problems are solved in subsequent chapters using more advanced
methods such as Lagrangian or Hamiltonian dynamics. We are of the opinion that
elementary problems solved using advanced techniques facilitate learning and better
understanding of these techniques. Correct solutions obtained using the advanced
methods also provide confidence in the use of these methods.
Problems
1. Solve the problem of a simple pendulum using Newtonian physics with the
coordinates shown. Assume that the bob of mass m is attached to a massless
rigid rod of length ` (Fig. 1.1).
Solution
Method 1
The force along the arc length s D ` is simply mg sin so according to

Newton’s second law, i.e. F D ma
mRs D mg sin (1.1)
© Springer International Publishing AG 2017 3

J.D. Kelley, J.J. Leventhal, Problems in Classical
and Quantum Mechanics, DOI 10.1007/978-3-319-46664-4_1
4 1 Newtonian Physics
Fig. 1.1 Problem 1
But, because s and ` are related according to s D ` we have

g
R C sin D 0 (1.2)
`
In the case of small oscillations sin Ð so that

g
R C D 0 (1.3)
`
p
and we have simple harmonic motion with frequency ! D g=`. This motion
is given by
.t/ D max sin .!t / (1.4)
The constant is the phase angle and determines the position of the
pendulum at t D 0. The constant max is the amplitude, which depends upon E,
the total mechanical energy (TME) as
E D mg .` ` cos max /
D mg` .1 cos max / (1.5)
For small oscillations

" !#
2
E D mg` 1 1 max (1.6)
2
so
s
2E
max D (1.7)
mg`
1 Newtonian Physics 5
Method 2
Again we use Newton’s second law, but this time in the form D I˛ where
torque, I is the moment of inertia of the bob about the pivot point and ˛ is the
angular acceleration. The torque is given by
D r F D `mg sin (1.8)
while the moment of inertia of the bob about the pivot point is
I D m`2 (1.9)
Therefore, we have

`mg sin D m`2 R (1.10)
which produces the same differential equation, Eq. (1.2).

This simple problem is the basis for many problems in classical mechanics
so it is worthwhile to work it using elementary methods.
2. A particle is dropped into a viscous fluid from rest at y D 0 and t D 0. Take

the force due to fluid resistance (the “drag”) to be proportional to the velocity
so that Fdrag D kmv, where k is a positive constant. Find the velocity v as a
function of the distance y. Take y to be positive downward. Manipulate your
answer to show that the answer is correct as k ! 0.
Fig. 1.2 Problem 2
Solution
Using Fig. 1.2 and F D ma we have
d2 y dv
m Dm D mg kmv (1.11)
dt2 dt
Separating this differential equation we have
1 dv 1 g
dt D H) t D ln v C C1 (1.12)
k .g=k/ v k k
Before proceeding we note that the units of k must be s1 . As we proceed

through the solution of his problem it is wise to check the units.
Using the initial condition to evaluate the constant of integration we arrive at
1 g
C1 D ln (1.13)
k k
and
1 g 1 g
t D ln v C ln
k k k k

g kv kv
kt D ln D ln 1 (1.14)
g g
dy
Now solve for v D
dt
kv g
ekt D 1 H) v D 1 ekt (1.15)
g k
so
g
dy D 1 ekt dt (1.16)
k
Integrating, we obtain

g 1
yD t C ekt C C2 (1.17)
k k
Using the initial conditions to evaluate the constant of integration we find

g 1 g
0D C C2 H) C2 D 2 (1.18)
k k k
so

g 1 g
yD t C ekt 2
k k k

g 1 1
D t C ekt (1.19)
k k k
Now we must eliminate t to obtain y D y .v/. To do this we solve Eq. (1.15)

for t. That is

g kt
1 kv
vD 1e H) t D ln 1 (1.20)
k k g
so that

kv
exp .kt/ 1 (1.21)
g
Inserting Eqs. (1.20) and (1.21) into Eq. (1.19) we have

g 1 kv 1 1 kv
yD ln 1 C 1
k k g k k g

1 g kv
D v C ln 1 (1.22)
k k g
If there were no fluid resistance the result is simple to calculate using

conservation of energy. Because the mass is dropped with no initial velocity
the initial potential energy must equal the kinetic energy at the distance y.
1 2
mv D mgy (1.23)
2
As it stands Eq. (1.22) is not very illuminating; it proves useful to look at the
limit as k ! 0. To facilitate this we use the Taylor series for ln .1 C x/ given in
Eq. (H.5).
x2 x3 x4
ln .1 C x/ D x C C (1.24)
2 3 4
Equation (1.22) then becomes
( " #)
1 g kv 1 kv 2 1 kv 3
yD vC
k k g 2 g 3 g

1 2 1 k 3
D v v
2g 3 g2

v2 1 k
D C v3 C (1.25)
2g 3 g2
The first term in Eq. (1.25) is clearly the correct answer as k ! 0. Moreover, the
positive sign on the term involving k shows that higher values of y are required
for higher values of v. Thus, with fluid resistance the particle must fall further
than without resistance to attain a given velocity.
3. A mass m2 is at rest a distance d from a brick wall. It is struck by a mass

m1 traveling with a velocity v1i as shown in Fig. 1.3. The collision is elastic
as is the subsequent collision of m2 with the wall. Find the conditions under
which a second collision will occur for the two cases m1 > m2 and m1 < m2 .
Discuss limits and find the smallest ratio for which a second collision will occur
(Fig. 1.4).
Fig. 1.3 Problem 3
Solution
Fig. 1.4 Problem 3—

solution
First, in terms of v1i , the initial velocity of m1 find the post-first-collision

velocities of m1 and m2 , v1f and v2f . All collisions are elastic so we recall an
important result from introductory physics:
In a one dimensional elastic collision the relative
velocity of approach before collision is equal to
the relative velocity of separation after collision.
This is a trivially derived result for the collision between m2 and the wall
(of infinite mass). Accordingly, the speed of m2 before it hits the wall will be
the same as it is after it hits the wall. Also, for the two masses
v1i D v2f v1f (1.26)
If the student does not remember this result, it is not necessary to panic
because it is readily derived, especially when m2 is initially at rest. We digress
to show this:
Conservation of momentum:

m1 v1i D m1 v1f C m2 v2f H) m1 v1i v1f D m2 v2f (1.27)
Conservation of TME:
1 2 1 2 1 2
m1 v1i D m1 v1f C m2 v2f (1.28)
2 2 2
Eq. (1.28) becomes

2
m1 v1i v1f v1i C v1f D m2 v2f (1.29)

Substituting Eq. (1.27) for m1 v1i v1f into Eq. (1.29) we have

v1i C v1f D v2f (1.30)
which is the same as Eq. (1.26).

Back to the problem at hand; using conservation of momentum with
Eq. (1.30) we have
m2
v1i D v1f C v2f
m1
m2
D v1f C v1i C v1f (1.31)
m1
Solving for v1f and v2f in terms of v1i we have

m1 m2
v1f D v1i (1.32)
m1 C m2
and
2m1
v2f D v1i (1.33)
.m1 C m2 /
Now, consider two different cases, m1 > m2 and m1 < m2 . Let x denote the
distance from the wall at which m1 and m2 collide for the second time. Note
that this distance can be greater than d, less than d or equal to d (for the case in
which m1 D m2 ).
Case I: m1 > m2 . From Eq. (1.32) we see that for m1 > m2 v1f is positive
(v1f is the velocity, not the speed) so x < d.
At the second of the collisions the masses will have traveled different
distances. Let
z1 D distance m1 travels between collisions D d x

z2 D distance m2 travels between collisions D d C x (1.34)
The travel time for each mass ttr is the same for both masses. As noted above,
the speed of m2 before it hits the wall will be the same as it is after it hits the
wall, i.e. v2f . We have
z1 z2 dx dCx
ttr D D D D (1.35)
v1f v2f m1 m2 2m1
v1i v1i
m1 C m2 .m1 C m2 /
Solving this equation for x we have

m1
1C
dx dCx m2
D )xD d (1.36)
m1 m2 2m1 m1
3 1
m2
Case II: m1 < m2 . Using Eq. (1.32) for this case we see that v1f is negative so
x > d. Because m1 backs up and moves to the right we can let
0
z1 D distance m1 travels between collisions D x d D z1
0
z2 D distance m2 travels between collisions D d C x D z2
We require the speed to calculate the time so we must reverse the sign of v1f
0
for m1 < m2 . But this reversal of sign is the same as reversing the sign of z1 to
make ttr > 0. This makes Eq. (1.36) valid for m1 < m2 .
Limits:
From Eq. (1.36) we have:
m1 D m2 H) x D d (this is sensible because from Eq. (1.32) v1f D 0)
m2 ! 1 H) x D d (no second collision occurs because m2 doesn’t
move)
From the denominator of Eq. (1.36) we see that if the ratio m1 =m2 D 1=3
then the second collision will occur at x D 1. Clearly, if m1 =m2 < 1=3 a
second collision cannot occur because x < 0 is forbidden so m1 =m2 D 1=3 is
the smallest ratio for which a second collision can occur.
4. A mass m1 with initial velocity v0 strikes a massless plate on a spring that is

attached to a mass m2 as shown in Fig. 1.5.
The spring constant is k and the table top is frictionless.
(a) What is x0 the maximum compression of the spring?
(b) What are the final velocities of m1 and m2 ?
(c) What criterion will insure that both masses are travelling in the same
direction after the collision?
Solution
(a) At maximum compression of the spring x0 the two masses have the same
velocity, call it v 0 . Therefore, using conservation of momentum and energy
we have
Fig. 1.5 Problem 4

m2 v0 m1

0 0 m1
m1 v0 D .m1 C m2 / v ) v D v0 (1.37)
m1 C m2
Conservation of energy gives

1 1 2 1
m1 v02 D .m1 C m2 / v 0 C kx02 (1.38)
2 2 2
Eliminating v 0 using Eq. (1.37) we have

1 m1 m2
x02 D v02
k m1 C m2
!
1 1
D v02 (1.39)
k m1 C m1
1 2
Clearly the maximum compression depends directly upon the square of

v0 and inversely on k. Because of the symmetry in m1 and m2 in the result,
Eq. (1.39), these masses play an equal role in determining x0 .
(b) Before the collision there is no potential energy because the spring is
neither compressed nor extended. Similarly, long after the collision it is
neither compressed nor extended. Therefore, the problem is the same as if
there were no spring and we have a perfectly elastic collision, the details of
which were worked out in Problem 3 of this chapter. The result is the same
as that of Eqs. (1.32) and (1.33)

m1 m2 2m1
v1f D v1i and v2f D v1i (1.40)
m1 C m2 .m1 C m2 /
(c) The final velocity of m2 , v2f , is manifestly positive (to the left in Fig. 1.5).
For both masses to be moving in the same direction after the collision v1f
and v2f must have the same sign so we must have v1f > 0. From Eq. (1.40)
it is clear that only for m1 > m2 will v1f > 0. This seems reasonable in the
limit if one considers m1 to be a bowling ball and m2 to be a ping pong ball.
5. Two unequal point masses m1 and m2 are attached to a spring having a spring
constant k and unextended length ` as shown in Fig. 1.6.
The system is free to oscillate on a frictionless table. Find the frequency of
oscillation for this system. It will be convenient to use the coordinates given.
Fig. 1.6 Problem 5

Solution
Using F D ma the equations of motion of each block are
m1 xR 1 D k Œx1 .x2 `/

m2 xR 2 D k Œx2 .x1 C `/ (1.41)
Because it is the mass separation, x2 x1 , that is important we re-write

Eq. (1.41) as
k
xR 1 D Œx1 x2 C `
m1
k
xR 2 D Œx2 x1 ` (1.42)
m2
Subtracting these equations we get

1 1
xR 2 xR 1 D k .x2 x1 `/ k .x2 x1 `/
m2 m1
.m1 C m2 /
D k .x2 x1 `/ (1.43)
m1 m2
Now define
& D .x2 x1 `/ (1.44)
so Eq. (1.43) becomes
.m1 C m2 /
&R C k & D0 (1.45)
m1 m2
which is the familiar equation that describes simple harmonic motion with
frequency ! given by
s
.m1 C m2 / k
!D (1.46)
m1 m2
6. Imagine a hole being dug through a diameter of the earth and an object of mass
m being dropped from rest at one surface. How long does it take for the object
to reach the other side of the earth? What is the nature of the motion of the
object? Take the radius of the earth to be R and ignore the rotation of the earth.
Assume that m traverses a frictionless path (Fig. 1.7).
Fig. 1.7 Problem 6
Solution
Because the gravitational force varies as 1=r2 , Gauss’s law applies. Therefore,
at any distance r from the center of the earth the attractive force toward the
center is proportional to the mass enclosed in a sphere of radius r. That is,
mM .r/
F .r/ D G (1.47)
r2
where M .r/ is the mass contained in the sphere of radius r. Taking to be the
density of the earth
4 3
M .r/ D r (1.48)
3
Thus, the force on the mass when it is situated at r is
m 4 3
F .r/ D G r
r2 3
4
D Gm r (1.49)
3
Using F .r/ D ma D mRr we have

4
mRr C m G r D 0 (1.50)
3
This equation can be simplified by noting that when r D R in Eq. (1.47)

4
F=m D g D G R (1.51)
3
where g is the gravitational acceleration at the surface of the earth. Equa-

tion (1.50) then becomes
g
rR C rD0 (1.52)
R
p This is the equation for simple harmonic motion with frequency ! D

g=R D 2=T where T is the period. Therefore, it takes one half T to arrive
at the other side of the earth, i.e.
s
1 R
TD (1.53)
2 g
Furthermore, if the mass m is not picked up when it arrives at the other side
of the earth it will return to its original drop-off point and continue to execute
simple harmonic motion indefinitely.
7. Imagine a hole that is dug through a chord of the earth, e.g. New York to Los
Angeles, and an object of mass m is dropped from rest at one end as depicted
in Fig. 1.8. How long does it take for the object to reach the other side? What
is the nature of the motion of the object? Take the radius of the earth to be R;
ignore friction and the rotation of the earth.
Solution
This problem is a generalization of Problem 6 of this chapter. Again Gauss’s

law applies so the attractive force toward the center of the earth at a radial
distance r from the center is proportional to the mass enclosed in a sphere of
Fig. 1.8 Problem 7

radius r. That is,
mM .r/
F .r/ D G (1.54)
r2
where M .r/ is the mass contained in the sphere of radius r and is the density
of the earth. Thus,
4 3
M .r/ D r (1.55)
3
As in Problem 6 of this chapter this equation can be simplified by noting that
the gravitational acceleration at the surface of the earth is given by
4
g D G R (1.56)
3
so that
r
F .r/ D mg (1.57)
R
In this case, however, only the component of F .r/ in the x -direction (see
Fig. 1.8) is the accelerating force. This component is
x x
Fx D F .r/ sin D F .r/ D mg (1.58)
r R
Using F D ma we have
mg
mRx C x D0 (1.59)
R
and that Eq. (1.50) becomes
g
xR C xD0 (1.60)
R
Not surprisingly (based on the result of Problem 6) the motion is simple
harmonic. Perhaps it is surprising though that the frequency of the oscillation
is exactly the same no matter where the hole is bored, through a diameter or a
chord.
8. A spherical cavity of diameter R is made in a sphere of radius R and uniform

density . The surfaces of the hole and the sphere coincide as shown in Fig. 1.9.
With what force will the large sphere with the hollowed out portion attract a
small sphere of mass m which lies a distance d > R from the center of the large
sphere on a line connecting the centers of both spheres?
Fig. 1.9 Problem 8
Solution
This problem demonstrates the magic of superposition. We treat the sphere as

if the hole doesn’t exist and treat the hole as a sphere having negative mass.
Assume the mass of the sphere prior to having cut the hole in it is M and the
mass of the hole is M 0 . These “masses” are then

4 3
MD R (1.61)
3
and
" 3 #
0 4 R M
M D D (1.62)
3 2 8
Therefore, using Gauss’s law and ignoring the hole, the force on m due to
M is
mM
FM D G (1.63)
d2
and the force on m from the (fictitious) “negative mass” M 0 is
mM 0
FM0 D G
.d R=2/2
G mM
D (1.64)
8 .d R=2/2
There are no vectors involved so the total force on the mass m is
Fm D FM C FM0

1 1
D GmM 2
d 8 .d R=2/2

GmM 1 1
D 1 (1.65)
d2 8 .1 R=2d/2
Note the trivial limit that as R ! 0, M ! 0 so Fm ! 0 as expected.

Suppose, however, that m is put at the center of the hole so d D R=2. The
force Fm in Eq. (1.65) blows up due to the denominator in the second term in
Eq. (1.65). This tells us that something is wrong. Upon reflection we recall that
the fields inside and outside a uniform distribution of mass (or charge) are not
the same (Gauss’s Law). In short, Eq. (1.65) is not valid if d < R.
We have already calculated the force on a mass m that is inside the sphere in
Problem 6 of this chapter. Using Gauss’s law for d < R , the force on m due to
the solid sphere (of radius d) without the hole is
4
FM D mG d (1.66)
3
and

4 R
FM0 D mG d (1.67)
3 2
where R=2 < d < R. Then
Fm D FM C FM0

4 4 R
D mG d mG d
3 3 2
4 R
D mG (1.68)
3 2
Replacing using Eq. (1.61) we have
1 mGM
Fm D (1.69)
2 R2
According to Eq. (1.69) the force on m is independent of d. In fact, the force
on m is not only independent of d, it is also independent of the location of m
within the hole.
Finally, let us compare the two results for Fm , one obtained for d > R,
Eq. (1.65), and the other with d < R, Eq. (1.69) at the only point they have in
common, d D R. Inserting d D R into Eq. (1.65) we have
1 mGM
Fm D (1.70)
2 R2
which is identical with Eq. (1.69). This is an interesting result because it shows
that the gravitational field is continuous at the common point.
9. The inclined plane of mass M shown in Fig. 1.10 rests on a frictionless surface.
The block of mass m is released from rest at the top of the frictionless
incline. The force on the block normal to the plane is N. Using the coordinates
in Fig. 1.10 find XR the acceleration of the inclined plane.
Fig. 1.10 Problem 9
Solution
Using Newton’s second law:
mRy D mg N cos
mRx D N sin
M XR D N sin (1.71)

Because xR and XR are antiparallel, the net horizontal of m is the sum xR C XR .
Thus, the ratio of the vertical to the horizontal acceleration is
yR
= tan
xR C XR
.mg N cos / =m
D (1.72)
1 1
N sin C
M m
Solving for N we have

g
ND (1.73)
1 1 1
tan sin C C cos
M m m
Putting Eq. (1.73) into the equation for XR in Eq. (1.71) and solving for XR we
have
g sin
XR D
M M
tan sin 1 C C cos
m m
g sin
D 2 2
sin M sin M cos
C C cos
cos m cos m cos
g sin
D 2
m sin M 1
C
m cos m cos
mg sin cos
D (to the left) (1.74)
m sin2 C M
Notice that if D 0 or if D =2, XR vanishes as it should. Moreover, as

M ! 1, XR ! 0 as it should.
This problem will be worked later in this volume in a tidier fashion using
Lagrangian dynamics (see Problem 14, Chap. 2).
10. A particle of mass m moves in one dimension without friction under the
influence of a potential energy function
U .x/ D U0 cos x (1.75)
The motion is restricted to the range =2 < x < 3=2.

(a) Sketch the potential energy function for the region of interest.
(b) What is the period T0 of the bounded motion for amplitudes small enough
so that the motion can be considered to be simple harmonic? Explain.
Solution
(a) The potential is a cosine function as shown in Fig. 1.11.
Fig. 1.11 p 3p
Problem 10—solution 2 2
U(x)
x
−U0
(b) For simple harmonic motion the potential energy is
1
Ushm .x/ D m! 2 x2 (1.76)
2
where m is the mass of the oscillating particle and ! is the angular
frequency of the motion. To find the conditions for which simple harmonic
motion is a good approximation to motion under U .x/ we must find the
coefficient of the quadratic term in a series expansion of the potential
energy, the well-known Taylor series given in Eq. (H.3).
U .x/ D U0 cos x

x2 x4
D U0 1 C (1.77)
2Š 4Š
Thus, U0 =2 is equivalent to 12 m! 2 and we have
1 U0 U0
m! 2 D ) !2 D
2 2 m
r
U0
!D (1.78)
m
Now, we know the period T0 in terms of !. It is
2 2
!D ) T0 D (1.79)
T0 !
Therefore
r
m
T0 D 2 (1.80)
U0
This harmonic approximation is valid for displacements in x small

enough that x4 =4Š << x2 =2Š.
This problem highlights the necessity of having the mathematical tools
to solve a problem readily accessible. In this case the Taylor series given in
Eq. (H.3) was a necessity.
11. A particle of mass m moves in one dimension without friction under the
influence of a potential energy function
2 2

U .x/ D U0 1 e˛ x (1.81)
where ˛ is a constant having suitable dimensions.

(a) Sketch the potential energy function. Draw lines of constant TME for
E0 < E1 < U0 and E2 > U0 .
(b) For a TME that gives bounded motion, what is the period T0 of the bounded
motion for amplitudes small enough for the motion to be considered simple
harmonic? Use the graph to discuss qualitatively the conditions under
which harmonic motion is a good approximation.
Solution
U(x)
E2
U0
E1
E0
x
Fig. 1.12 Problem 11—solution
(a) The heavy line is U .x/ and the lighter line represents the leading term of
the Taylor series expansion, the quadratic term, i.e. a parabola.
(b) To find the parameters for simple harmonic motion we must find the
coefficient of the quadratic term in the potential energy as in Problem 10
of this chapter. We must therefore expand U .x/ in a Taylor’ series. Using
Eq. (H.3) we have
2 2

U .x/ D U0 1 e˛ x

˛ 2 x2 ˛ 4 x4
D U0 1 1 C
1Š 2Š
U0 ˛ 2 x2 (1.82)
Thus, U0 ˛ 2 is equivalent to 12 m! 2 , the coefficient of x2 in the potential

energy function for simple harmonic motion. We have
1 2U0 ˛ 2
m! 2 D U0 ˛ 2 ) ! 2 D
2 m
s
2U0 ˛ 2
!D (1.83)
m
Now, we know the period T0 in terms of !. It is
2 2
!D ) T0 D (1.84)
T0 !
Therefore
r s
2 m 2m
T0 D D (1.85)
˛ 2U0 ˛ U0
Note that E1 (see Fig. 1.12) would not produce approximately harmonic
motion because the potential energy function is not very parabolic in
that region. The quadratic term in Eq. (1.82) is responsible for simple
harmonic motion so the region in which U .x/ most nearly approximates
a parabola is the region in which harmonic motion is expected to be a good
approximation, roughly that for which the TME is equal to or less than E0
as drawn on the graph.
12. Consider a particle that approaches a hard sphere of radius R, impacts the
surface and then bounces away elastically. The interaction potential is therefore
U .r/ D 1 r R
D0 r>R (1.86)
Fig. 1.13 Problem 12
The mass of the sphere is very much greater than that of the particle
(think of a small ball-bearing, a “BB,” impacting a bowling ball). The relevant
parameters are shown in Fig. 1.13.Take the incident velocity to be parallel to the
z-axis shown; the parameter b is known as the impact parameter and is given by
b D R sin ˛ for a hard sphere collision. The angles of incidence and reflection
are equal and also given by ˛. The “scattering angle” is .
This problem is three-dimensional, but it has circular symmetry in the
azimuthal angle around the z-axis. We define a “differential scattering cross-
section” D. / for particles incident with impact parameter b and scattered into
an infinitesimal solid angle d .
D. / D (1.87)
d
where
d
D bdbd
d D sin d d (1.88)
From Eq. (1.88) we see that

has units of area and the units of are
steradians (unitless). Now,
b db
D. / D (1.89)
sin d
(a) Find the relationship between the scattering angle and the angle of
incidence ˛, and use this to obtain the relationship between and b for
a hard sphere. Obtain D. / for a hard sphere. What happens when b > R?
(b) Evaluate the total elastic scattering cross section
by integrating D. / over
all space.
Solution
(a) From the diagram, the scattering angle is
D 2˛ (1.90)
so

b D R sin
2 2

D R cos (1.91)
2
The latter relation immediately leads to

1 b
D 2 cos if b R
R
D 0 if b > R (1.92)
Clearly, if b > R the particle misses the sphere, so there is no scattering.

Using Eq. (1.91) we can solve for D. /.
R cos 2 1
D. / D R sin
sin 2 2
R2
D (1.93)
4
(b) The total cross section is
Z

D D. /d

Z
R2
D d
4
D R2 (1.94)
This result is (or should be) intuitively obvious. The total scattering
cross section
in the hard sphere case is just the cross sectional area
of the sphere. If the particles were BBs and a large number were shot
horizontally with a wide range of impact parameters, a large sheet of
cardboard placed behind the sphere would be riddled with holes except for
an untouched circular region of radius R. When the hard sphere is replaced
by an arbitrary radial interaction potential U .r/ of longer range than a hard
sphere, Eq. (1.89) remains valid, but the relationship between b and may
be quite different.
Chapter 2
Lagrangian and Hamiltonian Dynamics
The Lagrangian and Hamiltonian formulations of mechanics contain no physics

beyond Newtonian physics. They are simply reformulations that provide recipes
to solve problems that are difficult to solve using elementary methods. Additionally,
these formulations, especially Hamiltonian dynamics, play an important role in the
development of quantum mechanics and in the elucidation of the deep relationships
between quantum and classical physics.
The Lagrangian equations of motion are obtained from Hamilton’s principle: The
actual path followed by a dynamical system is that which minimizes the time integral
of the Lagrangian L. The Lagrangian
LDT U (2.1)
is the difference between the kinetic energy T and potential energy U. The
minimizing path is found using the calculus of variations [3]. In equation form this
path is
Z t2
ı Ldt D 0 (2.2)
t1
This procedure leads to Lagrange’s equations of motion, which are

@L d @L
D0 (2.3)
@qi dt @Pqi
where qi and qP i represent the coordinates and their time derivatives. It is a simple
matter to remember where the d=dt must be in Eq. (2.3). We simply make sure that
the units of each term are the same.

26 2 Lagrangian and Hamiltonian Dynamics
Equation (2.3) does not explicitly contain momentum, but the generalized
momentum pi , also termed “conjugate momentum” (conjugate to the coordinate qi ),
is defined as
@L
pi D (2.4)
@Pqi
Thus, the Lagrangian equations of motion do contain the momentum, but it is

disguised. Note that if qi happens to be a linear coordinate then the generalized
momentum is the usual linear momentum. If qi is an angle as in cylindrical or
spherical coordinates, then the generalized momentum is an angular momentum.
Other pi ’s may not be recognizable as momenta, but they are nonetheless called
generalized momenta. The generalized coordinate qi and the generalized momentum
are termed “canonically conjugate.”
The Lagrangian L and the Hamiltonian H are related by
X
H .qi ; pi ; t/ D qP j pj L .qi ; pi ; t/ (2.5)
j
In Eq. (2.5) we have included the possibility that the time might appear explicitly in
the Lagrangian and the Hamiltonian. From Eq. (2.5) Hamilton’s equations of motion
are derived. They are
@H
qP i D
@pi
@H
Ppi D (2.6)
@qi
It is important to remember that to use these formulations the Lagrangian L

and the Hamiltonian H must be written in terms of their proper variables. For
generalized coordinates qi the Lagrangian must be written in terms of the qi , their
time derivatives qP i and possibly the time t. In contrast, the Hamiltonian H must be
written in terms of qi and pi . Equation (2.4) provides the link between pi and qP i .
If the Hamiltonian does not explicitly contain the time, then H is a conserved
quantity. To show this we write the total derivative of H with respect to time as
though it contains q, p, and t; replacing the partial derivatives with the Hamiltonian
equations of motion, Eq. (2.6). This leads to
dH @H @H @H
D qP i C pP i C
dt @qi @pi @t
@H
D Ppi qP i C qP i pP i C
@t
@H
(2.7)
@t
2 Lagrangian and Hamiltonian Dynamics 27
Thus, if H does not contain the time explicitly the partial time derivative is zero as
is the total derivative and H is a constant of the motion. This constant of the motion
is the TME if:
• The potential U contains no velocity dependence.
• The equations connecting Cartesian and generalized coordinates are independent
of time. This requires the kinetic energy to be a homogeneous quadratic function
of the qP i [3].
Having noted the conditions under which the Hamiltonian (and often the TME)
is a constant of the motion we turn to the generalized momenta pi as defined in
Eq. (2.4). From the second of the Hamiltonian equations of motion, Eq. (2.6), it
is clear that pP i vanishes if the Hamiltonian does not explicitly contain qi which
infers conservation of pi . In such a case the canonically conjugate coordinate qi is
referred to as a “cyclic coordinate.” It is also known as an “ignorable coordinate,”
an unfortunate appellation.
A good way to begin solving problems using Lagrangian or Hamiltonian
dynamics is to solve simple problems, problems for which you already know the
answers. This will give you confidence that these seemingly abstract formulations
are correct.
Problems
1. One of the first applications of the calculus of variations, upon which

Lagrangian dynamics is based, was to the classic problem known as the
brachistochrone. The goal is to find the path of a particle moving in a constant
force field, e.g. gravity, starting from rest at some point .x1 ; y1 / and ending at
a lower point .x2 ; y2 / such that the time to traverse this path is a minimum. It
is well known that this path is a cycloid. It is convenient to make the minimum
point of the cycloid at x D 2a. The parametric equations of the cycloid in this
form are
x D a .1 C cos /
y D a . C sin / (2.8)
where the coordinates axes, x and y, are shown in Fig. 2.1.

What is not so well-known about the brachistochrone problem is that the
time to traverse the cycloidal path is independent of the starting point .xi ; yi /.
To determine that the path is a cycloid the particle is always placed at x D 0 on
this coordinate system. It is a fact, however, that the time to reach the bottom
of the cycloid x D 2a is independent of the location of the starting point. Prove
this assertion.
Fig. 2.1 The coordinates (x1,y1) y

used to derive the parametric
equations of the trajectory of
the brachistochrone,
Eq. (2.8). These coordinates
are also used in Problem 1 of q = qi
this chapter (xi,yi)
(x2,y2)
x
Solution
The time dt it takes the particle to traverse a distance ds along the cycloid is
simply
ds
dt D (2.9)
v .x/
where v .x/ is the speed of the particle at any point .x; y/ along the cycloid
between .xi ; yi / and the minimum .x2 D 2a; y2 D 0/; this speed clearly depends
only upon x.
We define as the zero of potential energy to be at .xi ; yi /. Then, because the
particle starts from rest, the TME is zero and at any point .x; y/ we have
U .x/ C T .x/ D 0
1
mg .xi x/ C m Œv .x/2 D 0 (2.10)
2
At any point .x; y/ the speed of the particle v .x/ is given by
p
v .x/ D 2g .x xi / (2.11)
The increment of path ds is

p
ds D dx2 C dy2
p
D 1 C y02 dx (2.12)
where y0 D dy=dx. The time to reach the minimum is

Z 2a
ds
tD
xi v
Z 2a s
1 C y02
D dx (2.13)
xi 2g .x x1 /
To perform the integration we use the parametric equations for the cycloid
and find that
dy d 1
y0 D D a .1 C cos /
d dx a sin
.1 C cos /
D (2.14)
sin
and
dx D a sin d (2.15)
Also, when x D xi ; D i and when x D 2a ; D 0 so we have
Z " 1C2 cos Ccos2

#1=2
1 0 1C sin2
tD p .a sin d /
2g 1 a C a cos a a cos 1
r Z 1
a 2 C 2 cos 1=2
D d
2g 0 cos cos 1
r Z 1 1=2
a 1 C cos
D d (2.16)
g 0 cos cos 1
Using Eq. (E.9),

cos D 2 cos2 1 (2.17)
2
we have
2 31=2
r Z 1
a 2 cos2 2
tD 4 5 d
1
g 0 2 cos2 cos2 2 2
r Z 1
a cos 2
D 1=2 d
g 0
cos2 2 cos2 21
r Z 1
a cos 2 d
D r
g 0 1

1 sin2 2 1 sin2 2
r Z 1
a cos 2 d
D q (2.18)
g 0 sin2 21 sin2
2
To evaluate this integral we let
1
z D sin ) dz D cos d
2 2 2
1
D 0 ! z D 0, D 1 ! z D sin (2.19)
2
The time to reach the bottom of the cycloid, Eq. (2.18) becomes
r Z sin 1
a 2 2dz
tD q (2.20)
g 0 2 1
sin 2 z2
Applying the integral in Eq. (G.7)

Z
dx x
p D sin1 (2.21)
a2 x2 a
to Eq. (2.20) we have
r " !#sin 21

a 1 z
tD2 sin
g sin 21 0
r
a
1
D2 sin .1/ sin1 .0/
g
r
a
D (2.22)
g
It is surprising that no matter where the particle starts its path along the
trajectory, the time to reach the bottom of the cycloidal path is always the same.
This seems counterintuitive when comparing two different starting points. If,
however, the particle starts from .x1 ; y1 / as in Fig. 2.1, then it has gained speed
when it reaches .xi ; yi / so the path from .xi ; yi / to .x2 ; y2 / is covered in less time
than if the particle had started from rest at .xi ; yi /. This argument does not prove
the result, but it at least rationalizes it.
As stated above, this problem was included as an illustration of the calculus
of variations which is used as the tool applied to Hamilton’s principle to develop
Lagrangian dynamics.
2. Use a one-dimensional formulation to show that Lagrangian dynamics is

equivalent to Newton’s second law. Assume that the kinetic energy is a function
of only the velocity and the potential energy a function of only the position.
Solution
Because the Lagrangian is invariant under a transformation from any set of

generalized coordinates to another we may use Cartesian coordinates for con-
venience. We write Lagrange’s equation for a single particle in one dimension.
@L d @L
D0
@x dt @Px
@ ŒT .Px/ U .x/ d @ ŒT .Px/ U .x/
D0 (2.23)
@x dt @Px
where we have noted explicitly that the kinetic energy T is a function of only
the velocity xP and the potential energy U is a function of only the position x.
Carrying out the indicated differentiations we have
@U .x/ d @T .Px/
D
@x dt @Px

d @ 1 2
D mPx
dt @Px 2
d
D .mPx/
dt
D mRx (2.24)
For a conservative system the negative derivative of the potential energy is

the force so the last equation is
F D mRx (2.25)
which is Newton’s second law. This demonstrates that Lagrange’s equations

(and therefore Hamilton’s equations) are nothing more than Newton’s laws
recast into a form that permits solution of problems that are more complex
than those encountered in elementary physics courses.
3. Solve the problem of the one-dimensional frictionless simple harmonic oscil-

lator (SHO) of mass m and spring constant k initially at rest with initial
displacement x0 using
(a) Lagrangian dynamics
(b) Hamiltonian dynamics
Solution
(a) Designating x .D q/ as the displacement and defining ! as the angular

frequency (! 2 D k=m), the potential and kinetic energies are
1 2 1 1
U .x/ D kx D m! 2 x2 ; T D mPx2 (2.26)
2 2 2
which leads to a Lagrangian
1 2 1
LD mPx m! 2 x2 (2.27)
2 2
Inserting L into Eq. (2.3) produces
xR C ! 2 x2 D 0 (2.28)
which is one of the most often encountered equations in physics. The

general solution for the oscillator response x .t/ is well known. It is
x .t/ D A sin .!t/ C B cos .!t/ (2.29)
Upon inserting the boundary conditions given in the statement of the

problem we find that A D 0 and B D x0 so
x .t/ D x0 cos .!t/ (2.30)
(b) To solve using Hamiltonian dynamics we must first find the Hamiltonian as
given in Eq. (2.5). Everything in this equation is already known to us except
the generalized momentum p which, according to Eq. (2.4) is
@L
pD D mPx (2.31)
@Px
In this case p is the usual linear momentum because the lone coordinate
x has dimensions of length. The Hamiltonian is, according to Eq. (2.5)
p 1 1
H .x; p/ D p mPx2 C m! 2 x2
m 2 2
p2 1
D C m! 2 x2
2m 2
D T CU (2.32)
Note that we did not have to calculate H using L because the potential
energy is independent of the velocity and the kinetic energy is a homoge-
neous quadratic function of the xP . We could have simply written Eq. (2.32)
because under these conditions the Hamiltonian is the TME.
Inserting the Hamiltonian into Eq. (2.6) we obtain two linear first order
differential equations.
@H p @H
xP D D ; pP D D m! 2 x (2.33)
@p m @x
This is in contrast to the Lagrangian formulation which led to a single

second order differential equation. We can eliminate pP in the second
equation by differentiating the first equation with respect to time and
inserting the resulting expression for pP into the second. We obtain
mRx D m! 2 x (2.34)
which is the same as Eq. (2.28).
4. Use Lagrangian dynamics to find the frequency of the simple pendulum shown
in Fig. 2.2 for small vibrations.
Fig. 2.2 Problem 4
Solution
This problem is easily solved using elementary methods (see Problem 1 of

Chap. 1). Nevertheless, it provides a good opportunity to practice skills with
Lagrangian dynamics on a problem with a well-known solution. We begin by
writing the kinetic and potential energies in Cartesian coordinates and then
converting to the single coordinate using
x D ` sin
y D ` cos (2.35)
The kinetic and potential energies are
1 2
TD m xP C yP 2
2
1 2
D m`2 P (2.36)
2
and
U D mgy D mg` cos (2.37)
so
1 2 P2
LD m` C mg` cos (2.38)
2
The Lagrange equation of motion for is

d @L @L
D0
dt @P @
m`2 R C mg` sin D 0 (2.39)
which is the same equation derived using elementary methods in Problem 1 of

Chap. 1 (see Eq. (1.2)).
For small angles sin Ð so we have
g
R C D 0 (2.40)
`
whichp is the equation for undamped harmonic motion with frequency
! D g=`.
Note that in the small angle approximation the mass executes linear motion
in x with this same frequency because Ð x=`. In other words, the differential
equation in Eq. (2.40) could have just as well have been written as
g
xR C x D 0 (2.41)
`
under the small angle approximation.
There are a few things to note about this simple problem. First, let us
investigate the consequences of applying the small angle approximation to
the Lagrangian before obtaining the equations of motion. The small angle
approximation is
2
cos Ð 1 C (2.42)
2Š
When the approximation sin Ð is employed only the first term in

Eq. (2.42), i.e. unity, is taken. The Lagrangian then becomes
1 2 P2
LD m` C mg` (2.43)
2
and the Lagrangian equation of motion is
R D 0 (2.44)
This is clearly incorrect because we have eliminated the effect of the

potential. The lesson to be learned here is that we have approximated ourselves
out of a meaningful solution. Clearly we must retain (at least) the first two terms
in Eq. (2.42). Doing so makes the Lagrangian
1 2 P2 2
LD m` C mg` mg` (2.45)
2 2
so the Lagrangian equation of motion is

d @L @L
D0
dt @P @
m`2 R C mg` D 0 (2.46)
or
R C ! 2p D 0 (2.47)
where
r
g
!p D (2.48)
`
is the natural frequency of the simple pendulum.
5. The pivot point of a simple pendulum of mass m and length ` moves in the
y-direction according to y D f .t/. Using Lagrangian dynamics show that the
motion of the pendulum is that of a simple pendulum in a gravitational field
g C fR .t/.
Fig. 2.3 Problem 5
Solution
The x- and y-coordinates of the mass are
x D ` sin and y D f .t/ ` cos (2.49)
so the x- and y-coordinates of the velocities are
xP D `P cos and yP D fP .t/ C `P sin (2.50)
The kinetic energy is

1 2
TD m xP C yP 2
2
1 n 2 P2
2 2
o
D m ` cos2 C fP .t/ C `2 P sin2 C 2fP .t/ `P sin
2
1 n 2
2 o
D m `2 P C fP .t/ C 2fP .t/ `P sin (2.51)
2
The potential energy is
U D mgy D mg Œf .t/ ` cos (2.52)
The Lagrangian is
1 n 2
2 o
L D m `2 P C fP .t/ C 2fP .t/ `P sin mg Œf .t/ ` cos (2.53)
2
Then

d @L d h 2P P i
D m ` C f .t/ ` sin
dt @P dt
h i
D m `2 R C fR .t/ ` sin C fP .t/ `P cos (2.54)
and
@L
D mfP .t/ `P cos mg` sin (2.55)
@
so Lagrange’s equation is
h i
`2 R C fR .t/ ` sin C fP .t/ `P cos fP .t/ `P cos C g` sin D 0
`2 R C fR .t/ ` sin C g` sin D 0

fR .t/ C g
R C sin D 0 (2.56)
`
This is the same differential equation as that derived

in Problem 1, Chap. 1
(see Eq. (1.2)) except that g has been replaced by fR .t/ C g .
It is interesting to consider different forms of f .t/ and their effect on the
pendulum. Perhaps the most interesting is the case of free fall. In that case
fR .t/ D g and the coefficient of sin in Eq. (2.56) vanishes so that R D 0.
Integration yields
D P 0 t C 0 (2.57)
where 0 and P 0 are the initial angular displacement and angular velocity of the
bob of mass m. To be specific let us suppose that we have a simple pendulum in
an elevator and we pull the bob out to some initial angle 0 . We release the bob
from rest so that P 0 D 0 at the instant that the elevator cable is severed. The
elevator and pendulum are therefore in free fall. The equation that describes the
motion becomes
D 0 (2.58)
Thus, the bob floats in space at D 0 as the elevator drops. In effect, there
is no gravity to cause the periodic motion of a simple pendulum. Free fall has
negated the effect of gravity!
6. Describe the motion of the system shown in Fig. 2.4, a sliding pendulum
system, for small vibrations of the pendulum. The mass M can slide without
friction in either direction. It is permissible to make the small vibration
Fig. 2.4 Problem 6
approximation in the Lagrangian before obtaining the equations of motion. Use

the coordinates shown in Fig. 2.4; X is the distance of the sliding block of mass
M from the wall, x is the distance of the pendulum bob of mass m from the wall
and y is the vertical coordinate of the bob measured from the location of the
mass m when D 0.
Solution
There are four variables, but they are not independent. They are related
according to
x D X C ` sin and y D ` ` cos (2.59)
There are only two degrees of freedom and we choose to work the problem
in terms of the variables X and .
The kinetic energy is
1
P2
TD M X C m xP 2 C yP 2 (2.60)
2
Eliminating x and y using Eq. (2.59) we have
2
1 P2 1 P P 2 P2 2
T D M X C m X C ` cos C ` sin
2 2
1 1 P P cos C `2 P 2

D M XP 2 C m XP 2 C 2X` (2.61)
2 2
Now we wish to make the small angle approximation, but we must be
careful. In Problem 4 of this chapter we saw that making the approximation
cos Ð 1 for small angles was too much of an approximation and led to an
absurd result. In this case we do not lose any terms in the kinetic or potential
energies with the small-angle approximation, and it greatly simplifies the term
in the parentheses in Eq. (2.61). We have
1 P2 1 P 2
T' M X C m X C `P (2.62)
2 2
The potential energy taking U .y D 0/ D 0 is
U .y/ D mgy
D mg` .1 cos /
2
Ð mg` (2.63)
2
Notice that making the small angle approximation in Eq. (2.63) required
retention of the first two terms in the Taylor series (see Appendix H) of cos .
With the small angle approximation the Lagrangian becomes
1 P2 1 P 2 2
LD M X C m X C `P mg` (2.64)
2 2 2
The equations of motion are
XR C `R C g D 0
.M C m/ XR C m`R D 0 (2.65)
Solving the second of Eq. (2.65) for XR we have

m
XR D `R (2.66)
mCM
Now, substituting Eq. (2.66) for XR in the first of Eq. (2.65) eliminates X and
gives

M g
R C D 0 (2.67)
mCM `
which we recognize to be the equation for small-angle harmonic motion in a

simple pendulum with frequency

g MCm
!2 D (2.68)
` M
Integrating Eq. (2.66)

m
XP D `P C C (2.69)
mCM
we see that if XP ¤ 0, but the pendulum remains at rest with P D 0, then the
motion in X proceeds with a constant velocity C, the constant of integration
in Eq. (2.69). If C D 0 and P ¤ 0, then the block and pendulum oscillate
in opposite directions. In the general case the block slides along the support
at a constant average velocity with a superimposed oscillation at frequency !.
Suppose the mass M were very large compared with m. Then Eq. (2.68) gives
the frequency for a simple pendulum attached to a fixed point, and Eq. (2.66)
describes translation of the block undisturbed by the pendulum motion. If the
two masses were equal, ! 2 would be twice that of the simple pendulum because
the effective length of the pendulum is half that of the simple pendulum—the
block and pendulum oscillate in opposite direction about a point at y D `=2.
7. An Atwood’s machine consists of two point masses connected by an inexten-

sible massless string of length ` that passes over a frictionless peg as shown in
Fig. 2.5. Take y to be the vertical displacement of m2 . For convenience assume
that m2 > m1 and that the diameter of the peg is negligible.
(a) Find the Lagrangian in terms of the coordinates y and yP .
(b) Find the acceleration of the masses using Lagrangian dynamics.
Fig. 2.5 Problem 7
Solution
(a) The kinetic energy T is the sum of the kinetic energies of the masses m1
and m2 .
1
TD .m1 C m2 / yP 2 (2.70)
2
The potential energy measured from y D 0 is
U D m2 gy m1 g .` y/
D .m1 m2 / gy m1 g` (2.71)
We are confident that this is correct because when y D 0 (m2 at the top)
the potential energy is due solely to m1 and is m1 g`. When y D ` (m1 at
the top) the potential energy is entirely due to m2 and is m2 g`.
The Lagrangian is
1
LD .m1 C m2 / yP 2 .m1 m2 / gy C m1 g` (2.72)
2
(b) Lagrange’s equation in terms of y and yP is
@L d @L d
D 0 ) .m1 m2 / g Œ.m1 C m2 / yP D 0 (2.73)
@y dt @Py dt
Therefore
.m2 m1 /
yR D g (2.74)
.m2 C m1 /
Although this simple result is often obtained in introductory physics it

is satisfying to get the same results using a more advanced technique. Note
that because m2 > m1 , yR is manifestly positive because in this coordinate
system (y positive down) g is a positive number. Moreover, the answer is
correct in the limiting cases m2 D m1 and m1 D 0 (Ry D 0 and yR D g,
respectively). Thus, manipulating this solution leads us to believe that it is
correct because it has the correct units, the correct sign and the limiting
values are correct.
8. A system consists of a pulley of mass M, radius R, and moment of inertia

I D MR2 . Masses m1 and m2 are connected by an inextensible massless string
that passes over the pulley as shown in Fig. 2.6. When the pulley rotates the
string moves without slipping at the interface with the pulley. There is no
friction between m1 and the table. The cord is massless and m2 > m1 .
(a) Find the Lagrangian in terms of y and .
(b) Using Lagrangian dynamics find the acceleration of the masses assuming
that the system starts with m2 at y D 0 with D 0.
Solution
(a) The kinetic energy is
1 1 2 P 2
TD .m1 C m2 / yP 2 C MR (2.75)
2 2
Taking the potential energy to be zero at y D 0 we have
U D m2 gy (2.76)
so that
1 1 2
LD .m1 C m2 / yP 2 C MR2 P C m2 gy (2.77)
2 2
(b) Since we are asked for the acceleration of the masses, i.e. yR , we may
eliminate P using the relationship between y and (no slipping), which is
y
D (2.78)
R
In terms of y and yP the Lagrangian is
1
L .y; yP / D .m1 C m2 C M/ yP 2 C m2 gy (2.79)
2
Lagrange’s equation is
d
m2 g .m1 C m2 C M/ yP D 0 (2.80)
dt
from which we obtain
Fig. 2.6 Problem 8

m2
yR D g (2.81)
.m1 C m2 C M/
As usual it is advisable to examine the extreme cases. If m2 >> m1 and

m2 >> M, then m2 would essentially be in free fall so that yR Ð g. On the
other hand, if either m1 or the pulley mass M dominates, virtually nothing
happens or, at least, the motion is quite slow.
9. Solve Problem 8, part (b) of this chapter using Hamiltonian dynamics.
Solution
We could formally find the Hamiltonian using the relationship between it, the
Lagrangian and the generalized momentum. That is
XN

H qj ; pj D qP i pi L qj ; qP j (2.82)
iD1
where pj is the generalized momentum defined as
@L
pj D (2.83)
@Pqj
In the present case, in terms of the coordinate y this would be
H D py yP L .y; yP / (2.84)
We note, however, that the kinetic energy for this problem, Eq. (2.75), is a
homogeneous function of the coordinates y and . Moreover, as in most cases,
the potential energy is not a function of the velocity. These two conditions
assure us that the Hamiltonian is the TME. Because we already have both
the kinetic and potential energies it is simply a matter of calculating the
generalized momenta in order to convert the time derivative of the coordinates
in the kinetic energy into their momentum equivalents. This conversion is
necessary because the derivatives taken in Hamilton’s equations are those of
the generalized momenta and the generalized coordinates. In terms of the
generalized coordinates the Hamiltonian equations of motion are
@H @H
qP j D and pP j D (2.85)
@pj @qj
The upshot of all of this is that we need not do much in the way of
calculation other than to find the generalized momentum py . Because y is a
linear coordinate, a Cartesian coordinate, py is simply a linear momentum.

Using Eq. (2.79) we have
@L
py D
@Py
D .m1 C m2 C M/ yP (2.86)
In terms of this momentum the TME, and thus the Hamiltonian is

p2y
HD m2 gy (2.87)
2 .m1 C m2 C M/
Hamilton’s equations are then

py
yP D and pP y D m2 g (2.88)
.m1 C m2 C M/
Differentiating the first with respect to time and substituting the second for
pP y we obtain
m2 g
yR D (2.89)
.m1 C m2 C M/
which is identical with the answer obtained in part (b) of Problem 8 of this
chapter, Eq. (2.81).
10. A particle of mass m is projected upward with a velocity v0 at an angle ˛ to the

horizontal in the uniform gravitational field of the earth as shown in Fig. 2.7.
Ignore air resistance and take U .y D 0/ D 0.
Fig. 2.7 Problem 10 y parabola
v0
(a) Find the Lagrangian in terms of x and y and identify cyclic coordinates .
(b) Find the conjugate momenta, identify them and discuss which are con-
served and why.
(c) Find the x- and y-components of the velocity as functions of time.

(d) Find the Hamiltonian.
(e) Ignoring air resistance use Hamiltonian dynamics with the coordinates
shown to find the x- and y-components of the velocity as functions of time.
Solution
(a)
1 2 1
TD m xP C yP 2 I U D mgy H) L D m xP 2 C yP 2 mgy (2.90)
2 2
Only the x-coordinate is cyclic because it does not appear in the
Lagrangian.
(b) Conjugate momenta:
@L @L
px D D mPx; py D D mPy (2.91)
@Px @Py
Because x and y are Cartesian coordinates px and py are simply the x and
y components of linear momentum. From Lagrange’s equation

@L d @L
D0
@q dt @Pq
@L d
pq D 0 (2.92)
@q dt
it is clear that when the Lagrangian is cyclic in a coordinate then the

momentum conjugate to that coordinate is conserved. In the current
problem we have
dpx
D 0 ) px D constant
dt
dpy
D mg ) py ¤ constant (2.93)
dt
Because there is an outside force in the y-direction py is not a conserved
quantity. On the other hand, there is no outside force in the x-direction so
px is conserved.
(c) Lagrange’s equations for x and y are
0 mRx D 0 and mg mRy D 0 (2.94)
which lead to
xP .t/ D constant, and yP .t/ D gt C constant (2.95)

Using the initial conditions we have
xP .t/ D v0 cos ˛ and yP .t/ D gt C v0 sin ˛ (2.96)
(d) The Hamiltonian is given by Eq. (2.5).

X
H .qi ; pi ; t/ D qP j pj L .qi ; pi ; t/ (2.97)
j
where q1 D x and q2 D y. Using Eq. (2.91) we have
H D xP px C yP py L

px py 1 px 2 py 2
D px C py m C mgy
m m 2 m m
p2x p2y
D C C mgy D T C U (2.98)
2m 2m
The Hamiltonian is, as expected, the TME because the potential energy
is independent of the velocity.
(e) Hamilton equations of motion for x are
@H
pP x D D 0 H) px D constant D mv0 cos ˛ (2.99)
@x
and from Eq. (2.91)

@H px
xP D D (2.100)
@px m
Now for the y-direction. The Hamilton equations for y are
@H
pP y D D mg H) py D mgt C constant (2.101)
@y
Using the initial conditions the constant may be evaluated.
constant D mv0 sin ˛ (2.102)
so
py D mgt C mv0 sin ˛ (2.103)
and, again from Eq. (2.91)

@H py
yP D D (2.104)
@py m
The components of momentum are then
px D mPx D mv0 cos ˛; py D mPy D mgt C mv0 sin ˛ (2.105)
so the velocity components are
xP D v0 cos ˛; yP D gt C v0 sin ˛
This projectile problem is usually the first one encountered in introductory

physics courses. We see, again, that simple problems may be solved easily
using Lagrangian or Hamiltonian dynamics. (More importantly, we get the
same answers as were obtained using elementary techniques.)
11. A massless string is wound around a disk of mass m, radius R, and moment of
inertia I about the center. The disk unwinds from rest (like a yo-yo) with initial
conditions t D 0, y D 0, D 0 as shown in Fig. 2.8.
Fig. 2.8 Problem 11
(a) Find the Lagrangian in terms of y and yP only.

(b) Find py and the Hamiltonian and find the linear velocity of the disk in the
y-direction as a function of time using Hamiltonian dynamics.
Solution
(a)
1 2 1 P2 y
T D mPy C I and D
2 2 R
2
1 1 yP
) T D mPy2 C I (2.106)
2 2 R
and, taking U .y D 0/ D 0 we have
U D mgy (2.107)
Therefore
2
1 2 1 yP
LD mPy C I C mgy
2 2 R

1 I
D mC yP 2 C mgy (2.108)
2 R2
(b)

I py
py D m C 2 yP ) yP D (2.109)
R m C RI2
Then, because the equations of the transformation from Cartesian to

generalized coordinates do not contain the time and the potential energy
does not contain the velocity the Hamiltonian is the TME, so that

1 I
H D T CU D m C 2 yP 2 mgy
2 R

1 I p2y
D mC 2 2 mgy
2 R m C RI2
1 p2y
D mgy (2.110)
2 m C RI2
The Hamiltonian equations of motion for y and pP y are
@H
D Ppy D mg
@y
@H py
D yP D (2.111)
@py I
mC
R2
To solve these equations, we differentiate the second and substitute for
pP y in the first and obtain

I
m C 2 yR D mg (2.112)
R
Integrating Eq. (2.112) we get
m
yP D gt C K (2.113)
I
mC
R2
where K is a constant. But yP .t D 0/ D 0 ) K D 0 because the disk was

released from rest. Therefore
m
yP D gt (2.114)
I
mC 2
R
The moment of inertia of a cylindrically symmetric object may be
written I D ˇmR2 where ˇ depends upon the radial distribution of mass
about the axis of rotation. For example, for a flywheel (ring) ˇ D 1. Thus,
we may write Eq. (2.114) for any such disk as
1
yP D gt (2.115)
.1 C ˇ/
Note that the final speeds, both linear (Py) and rotational (Py=R), depend
upon neither the mass m nor the radius of the disk R. It is the distribution
of mass (ˇ) that determines these speeds.
12. Find the linear velocity yP in Problem 11 of this chapter using Lagrangian
dynamics.
Solution
The Lagrangian in terms of y and yP was found to be

1 I
LD m C 2 yP 2 C mgy (2.116)
2 R
so the Lagrange equation of motion for the coordinate y is

@L d @L I
D 0 ) mg m C 2 yR D 0 (2.117)
@y dt @Py R
Integrating and applying the boundary conditions we arrive at the same

answer as was obtained in Problem 11 of this chapter, Eq. (2.114).
m
yP D gt (2.118)
m C RI2
13. A uniform rod of mass m and length L is attached to a frictionless pivot as

shown. The rod is released from rest in the horizontal position.
Fig. 2.9 Problem 13
(a) Using the generalized coordinate shown in Fig. 2.9 write the Lagrangian.
Take the potential energy to be zero when the rod is horizontal ( D =2).
The Lagrangian should contain only the quantities given and the gravita-
tional acceleration.
(b) Use Lagrangian dynamics to find an expression for the angular velocity of
P
the rod as a function of and .
(c) Suppose that instead of being released from the horizontal position, the rod
were released from rest at a very small angular displacement from vertical
0 << =2. What would be the angular frequency in radians per second
of the small oscillations of the rod?
Solution
(a)
1 P2 L
TD Ie and U D mg cos (2.119)
2 2
where Ie is the moment of inertia of the rod about one end, which can be
quickly calculated. Letting x be the coordinate along the length of the rod
we have
Z L Z L
m
Ie D x2 dm D x2 dx
0 0 L
mL2
D (2.120)
3
Using this result we find that
L D T U
mL2 P 2 L
D C mg cos (2.121)
6 2
(b) The Lagrangian equation of motion is

@L d @L
D0
@ dt @P
L mL2 R
mg sin D0
2 3

3g
R C sin D 0 (2.122)
2L
Note that the coefficient of sin must have the same units as R , that is
1= s2 (“radians” is unitless). Indeed, g=L has the correct units.
Equation (2.122) is a nonlinear differential equation because sin is
nonlinear in . Solution of such equations is usually non-trivial. In our case,
however, we require only the solution for P . This can be obtained using
a standard “trick.” We multiply through by P and note that the resulting
equation may be written

3g P
P R C sin D 0
2L

d 1 P2 3g
cos D 0 (2.123)
dt 2 2L
Thus

1 P2 3g
cos D C (2.124)
2 2L
where C is a constant. But P D 0 when D =2 so C D 0 and we have

r
3g
P D cos (2.125)
L
This answer can be checked using elementary methods. The TME is

zero because the rod is released from rest and the zero of potential energy
is taken to be D =2. The TME is the Lagrangian, Eq. (2.121), with the
sign of the potential energy U reversed.
mL2 P 2 L
TME D mg cos D 0 (2.126)
6 2
Solving for P yields the same result as that obtained in Eq. (2.125).
(c) For an initial displacement 0 the TME is no longer zero, but this does not
change the Lagrangian in Eq. (2.121). Therefore, Eq. (2.122) still describes
the motion. Because 0 is small we may make the approximation sin '
(the first term in the Taylor series for sin ). We have then

3g
R C D0 (2.127)
2L
This equation describes simple harmonic motion about the coordinate

with angular frequency
r
3g
!D (2.128)
2L
We note that ! has the correct units, 1= s.
14. This problem is identical to Problem 9, Chap. 1 except now we wish to work it
using Lagrangian dynamics. It will be seen that the solution is much easier.
The inclined plane shown rests on a frictionless surface. The block of mass m
is released from rest at the top of the frictionless incline. Using the coordinates
in Fig. 2.10 use Lagrangian dynamics to find XR the acceleration of the inclined
plane. Note that the block and plane move in opposite directions, x and X.
Solution
The total kinetic energy is

1 P2 1 2
TD M X C mv (2.129)
2 2
where v is the velocity of the mass m along the coordinate s and is given,
according to Fig. 2.11, by the law of cosines.
Fig. 2.11
Problem 14—solution
v 2 D XP 2 C sP2 C 2XP
P s cos (2.130)
Using Eq. (E.2) we see that
cos . ˙ / D cos cos

D cos (2.131)
which accounts for the plus sign in the coefficient of cos in Eq. (2.130).
Only m contributes to the potential energy U so
U D mgs sin (2.132)
The Lagrangian is therefore
LDT U
1 P2 1 P2 1 2 P s cos C mgs sin
D M X C mX C mPs C mXP (2.133)
2 2 2
Lagrange’s equations are then

d @L @L
D 0 H) M XR C mXR C mRs cos D 0 (2.134)
dt @XP @X
or
.m C M/ XR D mRs cos (2.135)
and

d @L @L
D 0 H) mRs C mXP cos mg sin D 0 (2.136)
dt @Ps @s
or
mRs D mg sin mXR cos (2.137)

R
Now substitute mRs from Eq. (2.137) into Eq. (2.135) and solve for X.
.m C M/ XR D mg sin cos C mXR cos2

m sin cos
XR D g
m C M m cos2
m sin cos
D g (2.138)
M C m sin2
which is the same answer obtained using Newton’s second law directly in
Problem 9 of Chap. 1.
15. A point mass m is attached to one end of a massless rigid rod of length `. The
other end is rotated in a horizontal plane with a uniform angular speed ! in a
circle of radius R. Use Lagrangian dynamics to show that this configuration is
identical with that of a simple pendulum in a gravitational field provided the
gravitational acceleration is replaced by R! 2 .
Solution
There is no potential energy so the Lagrangian is the kinetic energy T. As

is frequently the case for rotating systems it is easiest to find the Cartesian
coordinates of the rotating mass and then convert to the appropriate generalized
coordinate, which in this case is . Using the axes shown we have
x D R cos !t C ` cos .!t C /

y D R sin !t C ` sin .!t C / (2.139)
where x and y are the Cartesian coordinates of the rotating mass. Differentiating,
we have

xP D R! sin !t ` ! C P sin .!t C /

yP D R! cos !t C ` ! C P cos .!t C / (2.140)
1 2
TDLD m xP C yP 2
2
1
D mŒR2 ! 2 sin2 !t C 2R!` ! C P sin !t sin .!t C /
2
2
C `2 ! C P sin2 .!t C / C R2 ! 2 cos2 !t

C 2R!` ! C P cos !t cos .!t C /
2
C `2 ! C P cos2 .!t C /
2
1
D m R2 ! 2 C 2R!` ! C P cos C `2 ! C P (2.141)
2
where we have used Eq. (E.2). Lagrange’s equation is

d @L @L
D0
dt @P @
d h i
2R!` cos C 2`2 ! C P C 2R!` ! C P sin D 0
dt
R!`P sin C `2 R C R! 2 ` sin C R!`P sin D 0
R! 2
R C sin D 0 (2.142)
`
This last equation reduces to that of a simple pendulum when is small

enough to use sin ' . This is derived in Problem 1, Chap. 1 (see Eq. (1.2)).
In Eq. (1.2) the coefficient of sin is g=`, the square of the frequency for small
displacements. Comparing this with the coefficient of sin in Eq. (2.142) we
see that the two equations are identical if g for the pendulum is replaced by R! 2 .
16. Work Problem 15 of this chapter using Hamiltonian dynamics.

Solution
We have already obtained the Lagrangian for this system in Problem 15

[Eq. (2.141)] using the coordinates given in Fig. 2.12. It is
2
1
L ; P D m R2 ! 2 C 2R!` ! C P cos C `2 ! C P (2.143)
2
We note that the equations of transformation connecting the Cartesian

coordinates .x; y/ and , Eq. (2.139), contain the time. Therefore, the kinetic
energy cannot be a quadratic homogeneous function of the P as may be seen
in Eq. (2.143) in which there are two linear terms in P . As a consequence
the Hamiltonian cannot be simply the TME so we must formally compute it
according to the relation

H . ; p / D P p L ; P (2.144)
First we must find the generalized momentum, which in this case is p and
is defined as
@L
p D
@P

D mR!` cos C m`2 ! C P
D mR!` cos C m`2 ! C m`2 P (2.145)
from which we find that

p R!
P D 2
cos ! (2.146)
m` `
Using Eq. (2.144) the Hamiltonian is

p R! 1
H .; p / D 2
cos ! p mR2 ! 2
m` ` 2

p R!
mR!` ! C cos ! cos
m`2 `
2
1 2 p R!
m` ! C cos ! (2.147)
2 m`2 `
which becomes
p2 R! 1
H .; p / D 2
p cos !p mR2 ! 2
m` ` 2
R!
p cos C mR2 ! 2 cos2
`
p2 R! 1
C p cos mR2 ! 2 cos2 (2.148)
2m`2 ` 2
Simplifying further we have
p2 R!
H . ; p / D 2
p cos !p
2m` `
1 1
mR2 ! 2 C mR2 ! 2 cos2 (2.149)
2 2
Hamilton’s equations are
@H @H
pP D and P D (2.150)
@ @p
so we have

R!
pP D p sin mR2 ! 2 cos sin
`
R!
D p sin C mR2 ! 2 cos sin (2.151)
`
and
p R!
P D cos ! (2.152)
m`2 `
Differentiating Eq. (2.152) with respect to time we have
pP R! P
R D C sin (2.153)
m`2 `
Substituting Eq. (2.151) to eliminate pP in Eq. (2.153) and using Eq. (2.145)
to eliminate P we have

1 R!
R D p sin C mR 2 2
! cos sin
m`2 `

R! p R!
C cos ! sin (2.154)
` m`2 `
which simplifies to
R! 2
R C sin D 0 (2.155)
`
Equation (2.155) is identical to the differential equation obtained using
Lagrangian dynamics, Eq. (2.142).
Because the Hamiltonian in this case is not the total energy there was quite
a bit more labor involved in obtaining Eq. (2.155) than that required to obtain
the Eq. (2.142) using Lagrangian dynamics. As always, it is comforting that we
obtained the same answer using both methods.
17. This problem is the same as Problem 7 of this chapter. That is, we have an
Atwood’s machine consisting of two point masses connected by an inextensible
massless string of length ` over a frictionless peg of negligible diameter.
Assume that m2 > m1 . In this problem we also ask for the tension in the string
and the acceleration of the masses obtained by using Lagrange’s method of
undetermined multipliers. This method of solving the problem, as contrasted
with the solution to Problem 7 of this chapter, illustrates the utility of the
method of undetermined multipliers and the added information that it yields.
Solution
The undetermined multipliers are related to the forces of constraint [3]. For
more detailed information the student is referred to any good textbook on
classical mechanics. To include the equation of constraint we modify the
generalized coordinates used in Problem 7 of this chapter (Fig. 2.13).
Fig. 2.13
The equation of constraint is the requirement that the length of the string
remain constant. Thus
y1 C y2 D ` ) f .y1 ; y2 / D y1 C y2 ` D 0 (2.156)
The kinetic and potential energies are
1 1
TD m1 yP 21 C m2 yP 22 ; U D m1 gy1 m2 gy2 (2.157)
2 2
The Lagrange equations of motion, with as the undetermined multiplier,

are
@L d @L @f .y1 ; y2 /
C D0 (2.158)
@yi dt @Pyi @yi
and the generalized forces of constraint are given by
@f .y1 ; y2 / @f .y1 ; y2 /
Q1 D ; Q2 D (2.159)
@y1 @y2
Because y1 and y2 are Cartesian coordinates the Qs are ordinary forces, i.e.
the tension in the string [3].
Inserting the Lagrangian and the equation of constraint into Eq. (2.158)
yields two uncoupled differential equations, one each in the coordinates y1
and y2 .
m1 yR 1 m1 g D 0 and m2 yR 2 m2 g D 0 (2.160)
Subtracting these equations and using the equation of constraint, Eq. (2.156),
to eliminate say yR 1 we immediately recover the result from Problem 7 in this
chapter, namely the linear acceleration of the system
.m2 m1 /
yR 2 D g (2.161)
.m2 C m1 /
Now, substituting this result into the second of Eq. (2.160) we may solve for
and thus the tension in the string.
2m1 m2
D g (2.162)
.m2 C m1 /
Therefore

2m1 m2
Q1 D g D Q2 (2.163)
.m2 C m1 /
which is the tension in the string. Notice that the tension forces are both in the
same direction, up.
18. A solid sphere of mass m and radius R having moment of inertia I D .2=5/ mR2
rolls down a stationary inclined plane without slipping. Using the method of
undetermined multipliers find the linear acceleration of the sphere down the
plane. Also find the force and torque required to maintain the condition “rolls
without slipping.” The method of undetermined multipliers is necessary in this
case in order to obtain the force and the torque.
Solution
Using the generalized coordinates shown in Fig. 2.14 the kinetic energy is
1 2 1 P2 1 1 2
TD mPs C I D mPs2 C mR2 P (2.164)
2 2 2 5
Taking the zero of potential energy to be at s D 0
U D mgs sin (2.165)
so the Lagrangian is
1 2 1 2 P2
LDT U D mPs C mR C mgs sin (2.166)
2 5
The condition that defines rolling without slipping is that the distance
traveled s is
s D R H) f .s; / D s R D 0 (2.167)
where f .s; / is the equation of constraint. We could eliminate from the

Lagrangian using Eq. (2.167), write Lagrange’s equation and solve for sR, thus
obtaining the linear acceleration of the disk down the plane. But, to obtain
the force and the torque it is necessary to use the “rolling without slipping”
constraint, Eq. (2.167), so the method of undetermined multipliers is required.
We proceed as in Problem 17 in this chapter.
Using the equivalent of Eq. (2.158), Lagrange’s equation with undetermined
multipliers, we obtain
2
mg sin mRs C D 0 and mR2 R R D 0 (2.168)
5
We may (temporarily) eliminate by solving the second of these equations
for and substituting it in the first. Thus,
2
D mRR (2.169)
5
so that
2
mg sin mRs mRR D 0 (2.170)
5
Now, using the equation of constraint in the form RR D sR, we eliminate R
and solve for sR the linear acceleration of the sphere down the plane
5g sin
sR D (2.171)
7
From the equation of constraint
sR 5g sin
R D D (2.172)
R 7R
and, reintroducing , Eq. (2.169), and inserting Eq. (2.172) we have
2mg sin
D (2.173)
7
Then, from Eq. (2.159) the generalized forces are
@f .s; / @f .s; /
Qs D ; Q D (2.174)
@s @
Note that in inasmuch as s is a linear coordinate Qs will be a force, but,
because is an angular coordinate Q is a torque. The units of these quantities,
once calculated, should verify this supposition. We find that
2mg sin 2mg sin

Qs D .1/ and Q D .R/ (2.175)
7 7
These signs are correct inasmuch as Qs , the force of friction that maintains
the “rolling without slipping” condition is in the s direction; Q is positive
because increases as the sphere rolls. Note that Qs and Q have the units of
force and torque, respectively.
19. A bead of mass m slides down a stiff straight wire without friction under the
influence of gravity as shown in Fig. 2.15.
The equation of the wire, considered to be a line, is y D kx where the

constant k > 0. Use Lagrange’s method of undetermined multipliers to find the
acceleration of the bead in both the x- and y-directions.
Solution
Taking the zero of potential energy to be at y D 0 the kinetic and potential

energies are
1 2 1 2
TD mPx C mPy ; U D mgy where g > 0 (2.176)
2 2
and the Lagrangian is
1 2 1 2
LDT U D mPx C mPy mgy (2.177)
2 2
The straight wire constrains the path of the bead to be
y D kx H) f .x; y/ D y C kx D 0 (2.178)
where f .x; y/ is the equation of constraint.

Lagrange’s equations of motion become
mRx C k D 0 (2.179)
and
mRy C mg C D 0 (2.180)
The terms in Eqs. (2.179) and (2.180) that do not contain are forces.
Therefore, must be a force (k is necessarily unitless).
To solve for xR and yR we first differentiate the equation of constraint,
Eq. (2.178), twice to obtain a relation between xR and yR .
yR D kRx (2.181)
Next, solve Eq. (2.179) for , insert this value and replace yR into Eq. (2.180)
to obtain

mRx
m .kRx/ C mg D0
k

1
kC xR D g (2.182)
k
or

k
xR D g (2.183)
k2 C 1
and

k2
yR D g (2.184)
k2 C 1
Now let us see if Eqs. (2.183) and (2.184) give sensible answers. First
examine the case for k D 0. This is a horizontal line so nothing happens; both
xR and yR vanish if k D 0. So far so good. Suppose the line is coincident with the
y-axis. In this case the slope of the line is infinite and we see that

k
lim xR D lim g
k!1 k!1 k2 C 1
D0 (2.185)
This is correct because if the line is vertical x 0. The bead never moves
left or right. On the other hand, how about yR ? We must examine

k2
lim yR D lim g
k!1 k!1 k2 C 1
D g (2.186)
This too is correct. The bead simply falls vertically with acceleration g.
Recall that g > 0 so the sign in Eq. (2.186) is correct.
The generalized forces are
@f .x; y/ @f .x; y/
Qx D ; Qy D (2.187)
@x @y
so that
Qx D k; Qy D (2.188)
First we solve for by inserting Eq. (2.183) into Eq. (2.179) to obtain

m 1
D xR D 2 mg (2.189)
k k C1
Therefore

1 1
Qx D 2 mgk; Qy D 2 mg (2.190)
k C1 k C1
Since k, the slope of the line on which the bead slides, is unitless it is clear
that both Qx and Qy have units of force. They are the x- and y-components of
the force imparted to the bead by the wire.
20. This problem is identical to Problem 5 of Chap. 1. Two unequal point masses
m1 and m2 are attached to a spring having spring constant k and unextended
length ` as shown in Fig. 2.16.
The masses are free to oscillate on a frictionless table. Use Lagrangian

dynamics to find the frequency of oscillation for this system.
Solution
The kinetic and potential energies are given by
1 1
TD m1 xP 12 C m2 xP 22 (2.191)
2 2
and the potential energy is
1
UD k Œx1 .x2 `/2
2
1
D k .x1 x2 C `/2 (2.192)
2
1 1 1
LD m1 x12 C m2 x22 k .x1 x2 C `/2 (2.193)
2 2 2
from which we obtain the equations of motion. They are
m1 xR 1 C k .x1 x2 C `/ D 0 (2.194)
and
m2 xR 2 k .x1 x2 C `/ D 0 (2.195)
Adding Eqs. (2.194) and (2.195) produces
d2
m1 xR 1 C m2 xR 2 D 0 ) .m1 x1 C m2 x2 / D 0 (2.196)
dt2
which, upon integration twice becomes
m1 x1 C m2 x2 D c1 t C c2 (2.197)
where c1 and c2 are constants of integration. This corresponds to translational

motion of the double mass system with no vibration of the spring, that
is ! 1 D 0.
To examine the oscillatory motion we divide Eqs. (2.194) and (2.195) by m1
and m2 , respectively, and subtract them to obtain

d2 1 1 1 1
.x2 x1 / C k C .x2 x1 / D k C ` (2.198)
dt2 m1 m2 m1 m2
which is simple harmonic motion in the coordinate .x2 x1 / with frequency

s
1 1
!2 D k C
m1 m2
s
.m1 C m2 / k
D (2.199)
m1 m2
The value of ! 2 in Eq. (2.199) is identical to the answer obtained in

Problem 5 of Chap. 1. Note that the grouping of masses in ! 2 is customarily
referred to as , the reduced mass. That is
m1 m2
D (2.200)
.m1 C m2 /
or, as it is conveniently remembered

1 1 1
D C (2.201)
m1 m2
Chapter 3
Central Forces and Orbits
Central forces are of considerable importance in both classical and quantum

mechanics so we briefly review their properties. In this chapter we deal with two
bodies of masses m1 and m2 . We ignore the motion of the center-of-mass so that only
the relative interparticle coordinates r and are treated, as will be discussed below.
The effective mass for motion in .r; / is the reduced mass D m1 m2 =(m1 +m2 ).
The central forces, F .r/, and the potentials that produce them, U .r/, depend only
upon the spherical coordinate r. The angular momentum vector L D r p, where
r is the position vector and p is the linear momentum vector, is conserved for all
central potentials. This will be proven in Problem 1 of this chapter. Because L is
conserved, the motion under the action of a central force takes place in a plane so
that only two dimensions are required to describe it. Plane polar coordinates .r; /
are almost always used in central force problems (see Appendix B.2). When written
in polar coordinates, the total energy of a system underthe influence of a central
potential contains an angular momentum term, `2 = 2r2 where ` is the magnitude
of the angular momentum L. For constant ` this term acts as an additional potential,
which is usually referred to as the centrifugal potential. Thus, we may define the
“effective potential” as
`2
Ueff .r/ D U .r/ C (3.1)
2r2
Note that ` D 0 corresponds to a linear trajectory passing through the force center.
Table 3.1 contains a few especially useful equations for central potential prob-
lems using plane polar coordinates.

68 3 Central Forces and Orbits
Table 3.1 Useful equations Quantity Equation

for central potential problems
Force F .r/ D @U .r/ =@r
Reduced mass D m1 m2 = .m1 C m2 /
2

Lagrangian L D 12 Pr2 C r2 P U .r/
Angular momentum ` D r2 P D constant

`2
Effective potential Ueff .r/ D U .r/ C 2r2
2
Total energy `
E D 12 Pr2 C 12 r 2 C U .r/
2 1
1 r2
Orbit equation d
d 2 r
C r D `2 F .r/
Problems
1. The angular momentum vector L is given by
LDrp (3.2)
where r is the position vector and p is the linear momentum vector. Show that
L is conserved for all central potentials.
Solution
If L is to be conserved, then its time derivative must vanish. Using Eq. (F.4) we
have
dL d
D .r p/
dt dt
dp dr
D r C p
dt dt
D r F .r/ C v p
D 0C0 (3.3)
The first term vanishes because F .r/ and r are colinear while the second
vanishes because p D v.
2. Starting with the Lagrangian in plane polar coordinates for a central potential
show that momentum conjugate to the coordinate is conserved for any central
potential. This momentum is, of course, the angular momentum.
3 Central Forces and Orbits 69
Solution
For any central potential the Lagrangian is
1 2 2

LD rP C r2 P U .r/ (3.4)
2
Lagrange’s equation for and P is
@L d @L
D0 (3.5)
@ dt @P
From Eq. (3.4) it is clear that for a central potential L does not contain
explicitly (although P is present) so the first term in Eq. (3.5) vanishes. This
means that (see Table 3.1)
@L
D r2 P D constant
P
@
D jLj D ` (3.6)
Note that , being absent from the Lagrangian, is a cyclic coordinate so that
the generalized momentum conjugate to this cyclic coordinate p D @L=@P is
conserved.
Theorem. When a coordinate is cyclic the generalized momentum conjugate
to this coordinate is conserved.
Remark. Sometimes this generalized momentum is not recognizable as a
momentum in the physical sense. For a central potential, however, it is the
angular momentum.
3. Kepler’s laws of planetary motion are

1. The planets travel about the sun in elliptical orbits with the sun at one focus
of the ellipse.
2. The radius vector from the sun to the planet sweeps out equal areas in equal
times.
3. The square of the period of the planet about the sun is proportional to the
cube of the semi-major axis of the elliptical orbit.
Show that Kepler’s second law is valid for any central force F.r/ between
the sun and the planet, or more generally between any two bodies.
Solution
The central force F .r/ is obtained from a central potential U .r/ by

differentiation
dU .r/
F .r/ D (3.7)
dr
The Lagrangian in spherical coordinates is
L D T U .r/
1 2 1 2 P2
D Pr C r U .r/ (3.8)
2 2
where is the reduced mass of the sun-planet system. This Lagrangian is
cyclic in the coordinate , i.e. it does not contain . Therefore, the momentum
conjugate to the cyclic coordinate is conserved.

d @L @L @L
D 0 H) D p D constant D r2 P D ` (3.9)
dt @P @ @P
This conserved momentum is p D `, the angular momentum.

If a particle is moving under the influence of the arbitrary central potential
margins U .r/, it sweeps out an area A in a time t2 t1 as shown in Fig. 3.1.
Fig. 3.1 Problem 3
This area dA is
1 1 dA 1
dA D r .rd / D r2 d ) D r2 (3.10)
2 2 d 2
Using Eq. (3.9) to obtain P the areal velocity dA=dt is

dA dA P
D
dt d

1 ` `
D r2 2
D
2 r 2
D constant (3.11)
thus proving the validity of Kepler’s second law for any central potential. In
contrast, the first and third laws depend upon F .r/ being an inverse square
force law.
4. Kepler deduced his laws from observations of the motion of the planets. His
observation that the planets travel in elliptical orbits about the sun (Kepler’s
first law) implies that the equation of the orbit in plane polar coordinates is
given by the equation of a conic section
˛
D 1 C cos (3.12)
r
where ˛ and are constants (see Appendix D). If the eccentricity is in the
range 0 < < 1, then the orbit described by Eq. (3.12) is a closed ellipse.
Newton had the advantage of knowing Kepler’s laws of planetary motion and
used them to obtain his famous 1=r2 law of gravitation.
Use the first two of Kepler’s laws to deduce Newton’s law of gravitation.
Solution
From Kepler’s second law Newton knew that the force must be a central force.
Recall that Kepler’s second law (equal areas in equal times) is the result of
conservation of angular momentum. The angular analog of Newton’s second
law F D pP (where p is the linear momentum) is a relation between the torque
T and the angular momentum L.
T D LP (3.13)
Conservation of angular momentum dictates that LP D 0 so there is no torque

on the particle to change its angular position. Therefore, the force is directed
along the radial coordinate, a central force.
Now that we have shown that the gravitational force is central, how do we
deduce the exact mathematical form of the law? We use the equation of the
ellipse in Eq. (3.12) together with the orbit equation for any central force. From
Table 3.1 we know the orbit equation is

d2 1 1 r2
2
C D 2 F .r/ (3.14)
d r r `
Solving Eq. (3.12) for 1=r and inserting it into Eq. (3.14) we have

d2 1 C cos 1 C cos r2
C D F .r/
d 2 ˛ ˛ `2
1 r2
cos C C cos D 2 F .r/ (3.15)
˛ ˛ ˛ `
so Newton’s law of gravitation is, indeed, the famous 1=r2 law
`2
F .r/ D (3.16)
˛r2
This problem shows the symbiotic relationship between theory and experi-
ment. In short, Newton didn’t guess the 1=r2 force law. He had observational
data at his disposal that led to his theory. Had Bertrand’s theorem [2] existed
in the latter part of the seventeenth century when Newton formulated his
theory of gravitation he might have correctly guessed 1=r2 from Kepler’s
first law. Bertrand’s theorem states that the only central forces that produce
closed repeating orbits are those that are proportional to 1=r2 and those that
are proportional to r2 . Unfortunately, this theorem was formulated by Joseph
Bertrand and published in 1873, nearly 200 years after Newton; fortunately,
Newton did not need it.
5. Express the total energy for a central potential in terms of only the radial
coordinate and its derivatives.
Solution
From Problem 3 of this chapter we know that the angular momentum is

conserved for any central potential so the motion is in a plane. In plane polar
coordinates the total energy is (see Table 3.1)
E D T C U .r/
1 2

D rP 2 C r2 P C U .r/ (3.17)
2
Also, the angular momentum is
` 1
` D r2 P H) P D 2 (3.18)
r
Substituting this in Eq. (3.17) we obtain the desired result
1 2 `2
ED Pr C C U .r/ (3.19)
2 2r2
6. The equation of the orbit for a central potential is given by (see Table 3.1)

d2 1 1 2 r 2
2
C D 2 F .r/ (3.20)
d r r `
This equation is a valuable tool when seeking the force F .r/ when the
equation of the orbit r ./ is known. It may also be used to determine r . /
when F .r/ is known. Starting from the Lagrangian for a central potential derive
Eq. (3.20) using the substitution u D 1=r so that

d2 u 1 1
CuD F (3.21)
d 2 `2 u2 u
Solution
The Lagrangian for a central potential is [see Eq. (3.8)]
1 2 1 2 P2
LD Pr C r U .r/ (3.22)
2 2
so the equation of motion for the coordinate r is
d @L @L
D0
dt @Pr @r
2 dU
Rr rP C D0 (3.23)
dr
Comparing Eq. (3.23) with Eq. (3.20) we see that we must eliminate P and
rR from Eq. (3.23). This is easily done because angular momentum is conserved
for any central potential. Thus
`
` D r2 P H) P D (3.24)
r2
To eliminate rR we use the chain rule on u D 1=r. We have
du du dr
D
d dr d
1 dr dt 1 rP
D 2 D 2 (3.25)
r dt d r P

D rP
`
where we have used conservation of angular momentum, Eq. (3.51). Differen-

tiating du=d with respect to we have
d2 u dPr dPr dt
D D
d 2 ` d ` dt d
rR
D
` P
2 2
D r rR (3.26)
`2
where we have again used conservation of angular momentum, Eq. (3.24), to
P Solving for rR we have
eliminate .
`2 1 d 2 u
rR D (3.27)
2 r2 d 2
Substituting Eq. (3.27) for rR , Eq. (3.24) for P and noting that F .r/ D
dU=dr Eq. (3.23) becomes

`2 1 d 2 u ` 2
r F .r/ D 0
r2 d 2 r2
`2 2 d2 u `2 3
u C u C F .r/ D 0 (3.28)
d 2
which is indeed the equation that we sought, Eq. (3.20), with r D 1=u. In the
form most often seen it is

d2 u 1 1
C u D F (3.29)
d 2 `2 u2 u
or

d2 1 1
C D 2 r2 F .r/ (3.30)
d 2 r r `
7. A particle of mass m is under the influence of a gravitational force of a mass

M located at the origin O. Assume M >> m so the center-of-mass is located at
the origin. The particle of mass m has a speed of vp at the point P, which is a
distance r0 from the origin (Fig. 3.2).
(a) What is E, the TME of the system for an arbitrary orbit having OP D r0 ?
(b) Define ˛ D vp =v0 where v0 would be the speed of the particle m in a
circular orbit of radius r0 . Show that
1
E D E0 C mv02 ˛ 2 1 (3.31)
2
Fig. 3.2 Problem 7
where E0 is the energy for a circular orbit. That is,

1 2 GmM
E0 D mv (3.32)
2 0 r0
(c) Show that the particle will be bound to M if ˛ 2 < 2.
Solution
(a) The TME is

1 2 GmM
ED mv (3.33)
2 p r0
(b) Adding and subtracting the kinetic energy for a circular orbit to Eq. (3.33)
we have

1 GmM 1 2 1 2
E D m˛ 2 v02 C mv0 mv0
2 r0 2 2
1 2 2 1
D m˛ v0 C E0 mv02
2 2
1 2 2
D E0 C mv0 ˛ 1 (3.34)
2
where E0 is the TME for a circular orbit.
(c) For a bound orbit the TME must be negative, E < 0. Eliminate r0 from the
total energy, Eq. (3.33), by finding it in terms of v0 . This is accomplished by
equating the centripetal force to the gravitational force for a circular orbit
mv02 GmM GM
D 2
) D v02 (3.35)
r0 r0 r0
Substituting this into Eq. (3.33) we have

1 2 2
ED m˛ v0 mv02
2
1
D mv02 ˛ 2 2 (3.36)
2
The condition for E < 0 is ˛ 2 < 2.
8. A particle moves in a spiral orbit given by the equation
r ./ D K 2 (3.37)
where K is a real constant.

(a) Use Eq. (3.19) derived in Problem 5 of this chapter to deduce the force law
F .r/ that produces this orbit.
(b) Use Eq. (3.20) derived in Problem 6 of this chapter to deduce the force law
F .r/ that produces this orbit.
Solution
(a) The total energy as given by Eq. (3.19) is
1 2 `2
ED Pr C C U .r/ (3.38)
2 2r2
Differentiating r . / and using conservation of angular momentum

` D r2 P
r 1=2 `
rP D 2 P K D 2K (3.39)
K r2
Inserting this result into Eq. (3.38) and solving for U .r/ we have
2K`2 `2
U .r/ D E
r3 2r2

`2 2K 1
DE C 2 (3.40)
r3 2r
Then the force law is
dU .r/
F .r/ D (3.41)
dr
2
` 6K 1
D C (3.42)
r4 r3
(b) The equation of the orbit as given by Eq. (3.20) is

d2 1 1
C D 2 r2 F .r/ (3.43)
d 2 r r `
Now
1 1
D 2 (3.44)
r K
so

d 1 2
D 3 (3.45)
d r K
and

d2 1 6
2
D 4
d r K
6K
D 2 (3.46)
r
Inserting the result of Eq. (3.46) into Eq. (3.43) and solving for F .r/ we
have
6K 1
C D 2 r2 F .r/
r2 r `

`2 6K 1
F .r/ D C 3 (3.47)
r4 r
We obtain the same answer using both methods. This is not surprising
because: (a) there is only one correct answer, and (b) the equation of
the orbit was derived using Lagrangian dynamics based on the kinetic
and potential energies so the methods are equivalent. Often, however, the
equation of the orbit is a more direct way of obtaining the force law when
the function r D r . / is known.
9. Using the equation of the orbit, see Eq. (3.20) and Table 3.1, show that there are
three different types of orbits that can result from a force law of the form
C
F .r/ D (3.48)
r3
where C is a real positive constant.
Solution
We use the equation of the orbit in the form
d2 u 1
2
CuD F .u/ (3.49)
d `2 u2
where u D 1=r and
F .u/ D Cu3 (3.50)
so we have
d2 u C
2
Cu D u
d `2
2
d u C
C 1 uD0 (3.51)
d 2 `2
This differential equation is the same as that describing simple harmonic
motion. By analogy, it is solved by assuming a solution of the form
u D em (3.52)
from which we arrive at the indicial equation

C
m2 C K 2 D 0 where K 2 D 1 2 (3.53)
`
The general solution is
1
uD D AeK C BeK (3.54)
r
where A and B are constants of integration to be determined by the boundary
conditions. The nature of the orbits will, however, differ greatly depending
upon the sign of K 2 . That is, the nature of the orbit will depend upon the
relative magnitudes of the constant C and the ratio =`2 . Examine the three
cases separately.
K 2 > 0: In this case the exponents are real and the orbit is a spiral.
K 2 D 0: The differential equation, Eq. (3.51), becomes
d2 u
D0 (3.55)
d 2
the solution to which is
1
u D A C B H) r D (3.56)
A C B
This too is a spiral as shown in Fig. 3.3.
Fig. 3.3 Problem 9
K 2 < 0: The exponents are imaginary and the solutions for u are therefore
sinusoidal. Thus, we may write
1
rD (3.57)
A cos . C ı/
which is also a spiral.

The three spirals that result from the attractive inverse cube central force law
are known as Cotes’ spirals. They are conveniently summarized by the three
different equations for their orbits r D r . /
A `2
rD C<
cos .K C ı/
A `2
D C>
cosh .K C ı/
1 `2
D CD (3.58)
A C B
The equation of the orbit was extremely useful in this problem because it
allowed us to expeditiously arrive at the solution.
10. A force law, the same as in Problem 9 of this chapter, is given by
C `2
F .r/ D where C D (3.59)
r3
The points .r; / D .1=4; 0/ and .r; / D .1=8; 4/ are known to lie on the
orbit of a particle subject to this force law. Also, it is known that at t D 0,
r D r0 . Find r as a function of time.
Solution
We know from Problem 9 of this chapter that, because C D `2 =, the orbit

must have the form
1
rD (3.60)
A C B
Using the two points on the orbit that are given we can evaluate A and B.
1 1
D H) B D 4 (3.61)
4 A0CB
and
1 1
D H) A D 1 (3.62)
8 A4C4
Therefore, the equation of the orbit r D r . / is
1
rD (3.63)
C4
At this point it is worth performing a consistency check on the units. From
Eq. (3.63) it might seem as though the units are not consistent because r has
units of length and everything on the right-hand side of this equation appears to
be unitless. Note, however, that A and B in Eq. (3.60) have units of r1 , so r has
the correct units.
Using conservation of angular momentum
` D r2 P (3.64)
we have
`
r2 d D dt (3.65)

We can eliminate d using Eq. (3.63)
1 1
D 4 H) d D 2 dr (3.66)
r r
So Eq. (3.65) becomes

`
dr D dt (3.67)

which, upon integration is
`
r .t/ D t C D (3.68)

where D is a constant of integration. At t D 0 the constant of integration D D r0

so the final answer is
`
r .t/ D t C r0 (3.69)

This contrived problem illustrates the manipulations that can yield the
time-dependent orbit r D r .t/ from the equation of the orbit in polar
coordinates r D r . /. As usual in central force problems, conservation of
angular momentum is central (no pun intended) to the solution.
11. A particle of mass having angular momentum ` is bound by a central force

F .r/ D kr, where k is a positive constant. This force causes the particle
to oscillate as an isotropic harmonic oscillator, “isotropic” because the spring
constant k is independent of direction.
(a) Sketch the potential energy U .r/ and the effective potential Ueff .r/.
(b) Find the radius rc of a circular orbit for this effective potential. Will the
circular orbit be stable or unstable? Explain.
Solution
(a) This force law for an isotropic harmonic oscillator is derived from the
corresponding potential energy, given by
dU .r/ 1
F .r/ D H) U .r/ D kr2 (3.70)
dr 2
Because angular momentum is conserved for central potentials, the
motion takes place in a plane and
`2
Ueff .r; `/ D U .r/ C (3.71)
2r2
the graph of which is shown in Fig. 3.4.

Fig. 3.4
Note that the curve that represents Ueff .r; ` D 0/ is half a parabola
(r cannot be negative) while Ueff .r; ` ¤ 0/ must approach the ordinate
asymptotically. Thus, the orbiting particle can never pass through the force
center at the origin.
(b) The radius rc of a stable circular orbit is the value of r at the minimum of
Ueff .r; ` ¤ 0/. Setting dUeff .r/ =dr D 0 to find rc we have
ˇ ˇ
dUeff .r/ ˇˇ `2 ˇˇ `2
ˇ D kr 3 ˇ D 0 H) rc4 D (3.72)
dr rDrc r rDrc k
ˇ
For rc to be a minimum d2 Ueff .r/ =dr2 ˇrDrc > 0 which is easily shown
to be the case. If the particle in this example were moving with constant
rc , a slight perturbation would produce small oscillations about rc . If rc had
been a maximum point, the particle would be at the top of the potential Ueff
and a perturbation would produce a large displacement in r.
12. Describe the possible orbits of the particle of mass m under the influence of an
isotropic harmonic oscillator potential (see Problem 11 of this chapter).
[Hint: It is recommended that Cartesian coordinates be used in conjunction
with F D ma rather than the equation of the orbit in polar coordinates.]
Solution
Using F D ma for Cartesian coordinates we have

F .x; y; z/ D mRr D m xR {O C yR |O C zRkO
D kxO{ ky|O kzkO (3.73)

Equation (3.73) represents three equations, one for each Cartesian compo-
nent. They are of the form
qR C ! 2 q D 0 where q D x; y or z and ! 2 D k=m (3.74)
Because the force is a central force we know that the motion takes place in
a plane so we may as well let z D 0. The solutions for x and y are

x D xm cos .!t x / and y D ym cos !t y (3.75)
where xm , ym , x and y are constants ( x and y are phase angles). The momenta
px and py are

px D m!xm sin .!t x / and py D m!ym sin !t y (3.76)
The coordinate equation (3.75) are the well-known

parametric equations
of an ellipse. The phase difference y x D ı determines the orientation
of the ellipse with respect to the xy-coordinate system. The phase difference
ı also determines the direction of motion of the particle on the ellipse. For
ı D 0; ˙, the ellipse is a straight line with particle oscillating back and forth;
for < ı < 0, the particle moves clockwise on the ellipse, and for 0 < ı <
it moves counterclockwise.
The total angular momentum, Lz is, from Eq. (3.2)
Lz D xpy ypx

D xm cos .!t x / m!ym sin !t y

C ym cos !t y m!xm sin .!t x /

D m! .xm ym / sin !t y cos .!t x /

cos !t y sin .!t x /

D m! .xm ym / sin !t y .!t x /
D m! .xm ym / sin ı (3.77)
Lz is positive or negative for counterclockwise or clockwise motion, respec-

tively, and it is zero for rectilinear motion, ı D 0; ˙. For a given value of the
product xm ym , jLz j is a maximum for ı D ˙=2. If xm D ym , the ellipse is a
circle.
Note that these elliptical orbits are closed paths that produce periodic
motion. In Problem 4 of this chapter it was noted that there are only two central
potentials for which the paths are closed, the isotropic oscillator and the Kepler
potential.
13. A particle of mass subjected to an attractive force law F .r/ D k=r2 where
k is a positive constant. This is also referred to as the “Kepler problem” because
planetary motion is derived from this force law, as we have already discussed.
(a) Sketch the effective potential Ueff .r; `/ for three different values of angular
momenta, `1 D 0 and 0 < `2 < `3 .
(b) Find rc the radius of a circular orbit for ` ¤ 0. Is the circular orbit stable?
Solution
(a) The potential is obtained from F .r/
F .r/ D k=r2 H) U .r/ D k=r (3.78)
so the effective potential is
`2
Ueff .r/ D k=r C (3.79)
2r2
the graph of which is (Fig. 3.5)
Fig. 3.5
Problem 13a—solution. The
radius of a circular orbit rc for
` ¤ 0 is shown for Ueff .r; `2 /
(b) Find the minimum of the effective potential Ueff .r; ` ¤ 0/ by setting the
derivative equal to zero.
dUeff .r/ k `2
D 2 3 D0 (3.80)
dr rc rc
so
`2 `2
krc D H) rc D (3.81)
k
The circular orbit is stable because rc occurs at a minimum in Ueff .r/.
14. For the gravitational force, F .r/ D k=r2 (k D positive constant), the possible
orbits are conic sections, the standard form of which is
˛
D 1 C cos (3.82)
r
See Appendix D for details. Choose one of the constants of integration
such that the polar angle is measured from the axis that coincides with the
minimum value of r.
(a) Prove that the equations of the orbits for an attractive inverse square force
(the Kepler problem) are conic sections (see Appendix D).
(b) Find the values of and ˛ in terms of the parameters `, , k, and E (the
TME).
(c) Show that the TME for a bound orbit is
k
ED (3.83)
2a
where a is the semi-major axis of the elliptical orbit.
Solution
(a) We begin with the equation of the orbit, Eq. (3.20) in the form
d2 u 1
CuD F .u/ (3.84)
d 2 `2 u2
with u D 1=r so
F .u/ D ku2 (3.85)
Thus, we have an inhomogeneous differential equation
d2 u k
2
CuD (3.86)
d `2
the solution of which is the sum of the solution to the homogeneous

equation and the particular solution. The solution to the homogeneous
equation is B cos . 0 / where B and 0 are constants of integration
[see the discussion surrounding Eq. (3.51)]. The particular solution is
u D constant D k=`2 . Thus
1 k
uD D 2 C B cos . 0 / (3.87)
r `
From Appendix D the standard form of a conic section with the origin at
one focus is
˛
D 1 C cos (3.88)
r
where ˛ and are constants given by
s
2E`2
D 1C D eccentricity
k2
`2
˛D D latus rectum (3.89)
k
The shape of the conic section depends upon the value of .
D 1 parabola
< 1 ellipse
> 1 hyperbola (3.90)
Equation (3.87) may be re-written

2 2
` =k B`
D1C cos . 0 / (3.91)
r k
from which it is clear that
`2
˛D (3.92)
k
and
D B˛ (3.93)
It is customary to take 0 D 0, a choice that forces r to have its minimum

value rmin at D 0. This convention conforms to the usual notation for the
Fig. 3.6
Problem 14—solution. The
energies E1 and E2 represent
the TME for a bound and
unbound orbit. The points
labelled TP are the turning
points for a particle having
these energies
“standard form” of the equation of the conic section, Eq. (3.88) and we are
left with
˛
D 1 C B˛ cos (3.94)
r
(b) To evaluate the constant of integration B so that we can find in terms of
the known parameters we use Fig. 3.6 as a guide. We note that whether the
TME is positive or negative the conic section must have at least one turning
point (TP in the figure) as long as ` ¤ 0. Only an ellipse has two turning
points, one at r1min and one at r1man (see Appendix D). From Eq. (3.94) we
see that rmin is given in terms of B by
˛
D 1 C B˛ (3.95)
rmin
We may write another expression for rmin by noting that the effective
potential must equal the TME at both r1min and r1max because the kinetic
energy at these values of r is zero for E < 0. For E > 0 the kinetic energy
is zero at the single turning point r2min , but we limit our discussion to the
elliptical orbits (E < 0). Setting Ueff .r/ D E we have
`2
Ueff .r/ D k=r C DE (3.96)
2r2
or
1 2k 1 2E
2
2 2 D0 (3.97)
r ` r `
which is a quadratic equation in the variable 1=r. The solutions to this

equation are
s
2
1 k k 2E
D 2 ˙ C
r ` `2 `2
s
1 1 1 2E
D ˙ 2
C (3.98)
˛ ˛ ˛ k
The plus sign yields r D rmin so we have

r
˛ 2E˛
D1C 1C (3.99)
rmin k
We now have two expressions for ˛=rmin , Eqs. (3.95) and (3.99), so we may
equate them to evaluate B and the eccentricity . Using Eq. (3.93) we have
r
2E˛
B˛ D D 1C (3.100)
k
s
2E`2
D 1C (3.101)
k2
Note that for the bound elliptical orbit the total energy E must be negative
and < 1 as prescribed by Eq. (3.90). Similarly, the eccentricities of the
parabolic and hyperbolic orbits follow from their properly inserted energies
(see the Fig. 3.6). According to Eq. (3.88) the minimum value of r D rmin
occurs when D 0 and for an elliptical orbit the maximum occurs when
D . Thus, the polar angle is measured from the “short side” of the
ellipse, the pericenter, as shown in Fig. 3.7. The long side, r D rmax , is
referred to as the apocenter.
Fig. 3.7
Problem 14—solution y
r
q
rmax rmin x
(c) From Eq. (3.95) it is clear that

˛ ˛
rmin D and rmax D (3.102)
1C 1
so the semi-major axis is
1 ˛
aD .rmin C rmax / D
2 1 2
2 2
` k
D
k 2 jEj `2
k
D (3.103)
2E
Solving for E we have
k
ED (3.104)
2a
where we have let jEj ! E because the energy must be negative for a
bound orbit.
Using the equation of the orbit we have derived the properties of the
conic sections (orbits) that result from an inverse square attractive force.
There are other ways to obtain this result, most of them equivalent to the
method used here. One of the most intriguing aspects of this problem is
that, in contrast to virtually all other central force orbits, the bound orbit
for the Kepler problem, the ellipse, closes on itself and is fixed in space
(Bertrand’s theorem again). This aspect of the problem is seldom stressed
at the introductory level. It is a consequence of a subtle symmetry of this
force law that produces an additional constant of the motion, a vector along
the semi-major axis that keeps the ellipse fixed in space. The perceptive
student is no doubt asking “what about the precession of the perihelion of
Mercury?” This motion of Mercury’s elliptical orbit is the result of several
perturbations to the pure Keplerian potential, the most famous being a
consequence of general relativity, which to first order adds a 1=r3 term to
the potential (see page 511 of [2]).
15. It was mentioned near the end of Problem 14 of this chapter that, in addition to
energy and angular momentum , the Kepler problem has an additional constant
of the motion [1]. Although it goes by other names [2], this constant of the
motion is commonly called the Runge–Lenz vector. It is designated by A and
defined as
r
A D p L k (3.105)
r
where the angular momentum vector is
LDrp (3.106)
and the other quantities in Eq. (3.105) have their customary meanings in this
chapter.
(a) Show that A is a constant of the motion.
(b) Show that A L D 0 so that A lies in the plane of the motion.
(c) By evaluating the scalar product A r and comparing the result with
Eq. (3.88) show that A points toward the pericenter (the short end) of the
semi-major axis of the elliptical orbit. Show also that the magnitude of A is
proportional to the eccentricity of the orbit and is given by
s
2E`2
A D k 1 C (3.107)
k2
Solution
(a) To show that A is a constant of the motion we must show that its time
derivative vanishes.
dA d r
D pP L C p LP k (3.108)
dt dt r
For a central potential angular momentum is conserved so LP D 0 and
dA d r
D pP L k (3.109)
dt dt r
For the Kepler problem
k k
pP D F .r/ D aO r D 3 r (3.110)
r2 r
Then, using the vector identity for the triple cross product, Eq. (F.2)
k
pP L D Œr .r Pr/
r3
k

D 3 r .r rP/ r2 rP (3.111)
r
Noting that
1d
r rP D .r r/ D rPr (3.112)
2 dt
we have
k
2

pP L D rP
r r r P
r
r3
k
D 2 ŒPrr rPr
r

rP rP r
D k
r r2
d r
D k (3.113)
dt r
P vanishes so A is a
Inserting this result into Eq. (3.109) shows that A
constant of the motion for the Kepler problem.
(b) We have
r
A L D p L k L
r
r
D k L
r
r
D k .r p/
r
D0 (3.114)
because L is perpendicular to .p L/ and r is perpendicular to .r p/.

(c) Taking the scalar product A r as directed we have
h ri
A r D r .p L/ k
r
D L .r p/ kr
D `2 kr (3.115)
where we have used the vector identity in Eq. (F.1).

By definition the scalar product A r is also given by
A r D Ar cos (3.116)
where is the angle between A and r. Equating these two expressions for
A r we see that the resulting equation is that of a conic section
Ar cos D `2 kr
`2
rD (3.117)
.A cos C k/
or

`2 =k A
D1C cos (3.118)
r k
Comparison of Eq. (3.118) with the equation of a conic section,
Eq. (3.88), shows that they are identical and D . Thus, the Runge–
Lenz vector points in the direction of the pericenter, that is, in the Cx
direction ( D 0/.
We can determine the magnitude of A in terms of the Keplerian
orbit parameters by comparing Eq. (3.118) with Eqs. (3.88) and (3.89).
Accordingly
A
D (3.119)
k
so that
s
2E`2
A D k 1 C (3.120)
k2
16. This problem shows why conservation of the Runge–Lenz vector demands
closed orbits for the Kepler problem.
(a) Express the Runge–Lenz vector A as defined in Eq. (3.105) in terms of r
and p only by eliminating L.
(b) Show that

k
A D rmax 2E C
rmax

k
D rmin 2E C (3.121)
rmin
where E is the total energy, r D rmin is the pericenter of the elliptical orbit,
and r D rmax is the apocenter.
(c) Show that A is parallel to rmin and antiparallel to rmax and that for a circular
orbit A D 0.
Solution
(a) A is defined as
r
A D p L k (3.122)
r
so, using the vector identity for the triple cross product, Eq. (F.2) and the
definition of angular momentum
r
A D p .r p/ k
r
r
D p2 r .p r/ p k (3.123)
r
(b) First eliminate p2 from Eq. (3.123) using the total energy
p2 k 2k
ED H) p2 D 2E C (3.124)
2 r r
Fig. 3.8 y
Problem 15b—solution
p
r
q
rmax rmin x
p
then
r
A D p2 r .p r/ p k
r
2k k
D 2Er C r .p r/ p r
r r

k
D 2E C r .p r/ p (3.125)
r
Now, when r D rmin or r D rmax , p ? r as illustrated in Fig. 3.8 so
p r D 0. Thus, at pericenter and apocenter we have

k k
A D rmin 2E C D rmax 2E C (3.126)
rmin rmax
(c) From Problem 14 of this chapter we know that the total energy of the
elliptical Kepler orbit is
k
ED (3.127)
2a
so Eq. (3.126) may be re-written

1 1
A D krmin C
a rmin

1 1
D krmax C (3.128)
a rmax
Moreover, we know that rmin < a < rmax so in the first of Eq. (3.128)
the quantity in parenthesis is positive; in the second, it is negative. This
establishes the directions of A:
A / rmin / rmax (3.129)
For a circular orbit rmin D rmax D a so, from Eq. (3.128), A D 0.

Although quantum mechanics has not yet been discussed here, it is
worthwhile to compare the nature of Keplerian orbits to the quantum
mechanical results for the hydrogen atom (H-atom) because both systems
are subject to the same 1=r2 force law. As listed in Table T.1 the energy
levels of the quantum mechanical H-atom are specified by a single quantum
number n, the principal quantum number.
E / 1=n2 (3.130)
These allowed energies do not depend upon the angular momentum

quantum number. For Keplerian elliptical orbits the energy depends only
upon the value of the semi-major axis a, Eq. (3.127), and not on the semi-
minor axis b which depends upon the angular momentum ` according to
r
` a
bD p D` (3.131)
2 jEj k
Thus, elliptical Keplerian orbits that have the same a, but different values
of b, all have the same energy. This is an example of a classical degeneracy.
Clearly, the classical degeneracy of the Keplerian orbits is related to those
of the H-atom.
Chapter 4
Normal Modes and Coordinates
Before we present the problems in this chapter we briefly review the properties of
normal modes and coordinates by stating three propositions about them:
Proposition 1. Using the normal coordinates to describe a multi-oscillator system
decouples the Lagrangian equations of motion. Each of the differential equations
of motion contains only a single variable, the normal coordinate, and its time
derivatives.
Proposition 2. Motion in the normal coordinates corresponds to sinusoidal oscil-
lation with a single well-defined frequency, a normal mode.
Proposition 3. When T and U are expressed in terms of the normal coordinates
they contain only quadratic functions of these coordinates and their time deriva-
tives; there are no cross terms.
Problems
1. The masses m of two identical pendulums of length ` are connected by a spring

having spring constant k. The motion is constrained to a plane and the angles
1 and 2 , each of which is measured counterclockwise from the vertical, are
assumed small. Assume also that the unstretched length of the spring is d, the
same as the separation between the supports.

96 4 Normal Modes and Coordinates
Fig. 4.1 Problem 1
(a) Set up Lagrange’s equations using the generalized coordinates 1 and 2 .

Use the small angle approximation.
(b) Find the normal mode frequencies and the normal coordinates. Describe the
motion for each normal mode.
Solution
(a) It is easier to find the kinetic energy T and potential energy U in terms of
Cartesian coordinates and then transform the result to 1 and 2 .
x1 D ` sin 1 ; y1 D ` cos 1
x2 D d C ` sin 2 ; y2 D ` cos 2 (4.1)
so
1 2
TD m xP 1 C yP 21 C xP 22 C yP 22
2
1 2 2 2 2

D m`2 P 1 sin2 1 C P 1 cos2 1 C P 2 sin2 2 C P 2 cos2 2
2
1 2 2

D m`2 P 1 C P 2 (4.2)
2
The potential energy due to gravity measured from the equilibrium
position of the mass is
Ugravity D mgy1 C mgy2

D mg` Œ.1 cos 1 / C .1 cos 2 /
4 Normal Modes and Coordinates 97
" ! !#
2 2
mg` 1 1 1 C1 1 2
2Š 2Š
1
D mg` 21 C 22 (4.3)
2
where we have, in the spirit of the small angle approximation, retained only
the first two terms of the Taylor expansion for cos .
The potential energy due to the spring (again using the small angle
approximation) is
1
Uspring D k Œ.x2 x1 / d2
2
1
D k`2 Œ.d C sin 2 sin 1 / d2
2
1
k`2 . 2 1 /2 (4.4)
2
so that
1 1
UD mg` 21 C 22 C k`2 . 2 1 /2 (4.5)
2 2
The Lagrangian is
LDT U
1 2 2
1 1
D m`2 P 1 C P 2 mg` 21 C 22 k`2 . 2 1 /2 (4.6)
2 2 2
and the Lagrangian equations of motion are

d @L @L
D0 (4.7)
dt @ P i @ i
so, using the notation
g
! 2p D D frequency of simple pendulum
`
k
! 2o D D frequency of SHO (4.8)
.m=2/
where m=2 is the reduced mass (see Problem 4 of Chap. 1 and Table 3.1).
The Lagrange equations of motion become
m`2 R 1 C mg` 1 k`2 . 2 1 / D 0 (4.9)
or
!2
R 1 C ! 2p 1 o . 2 1 / D 0 (4.10)
2
and
m`2 R 2 C mg` 2 C k`2 . 2 1 / D 0 (4.11)
or
!2
R 2 C ! 2p 2 C o . 2 1 / D 0 (4.12)
2
Notice that these equations of motion, Eqs. (4.10) and (4.12), are coupled
in the sense that each equation contains both 1 and 2 .
(b) Sometimes, as in this problem, it is possible to find the normal coordinate by
inspection. Let us add and subtract Eqs. (4.10) and (4.12) to cast them in a
different form. Adding them we have

R 1 C R 2 C ! 2p . 1 C 2 / D 0 (4.13)
Subtracting them gives

R 1 R 2 C ! 2p . 1 2 / C ! 2o . 1 2 / D 0 (4.14)
Inspection of Eqs. (4.13) and (4.14) shows that 1 and 2 occur only in the
forms . 1 C 2 / and . 1 2 /. This suggests that we make the substitutions
1 1
D . 1 C 2 / ; D . 1 2 / (4.15)
2 2
which lead to
R C ! 2p D 0

R C ! 2p C ! 2o D 0 (4.16)
In terms of the new variables, and , Eqs. (4.10) and (4.12) are now
“decoupled.” Moreover, Eq. (4.16) are the differential equations of simple
harmonic motion in the variables and so the system oscillates in and
at the frequencies
q
! D !p; ! D ! 2p C ! 2o (4.17)
which are the normal mode frequencies (see Proposition 2).

To check Proposition 3 we must solve Eq. (4.15) for 1 and 2 in terms of
and and then find T [Eq. (4.2)] and U [Eq. (4.5)] in terms of and . We
have
1 D C and 2 D (4.18)
From Eq. (4.2)
1 2 P 2 P 2
TD m` 1 C 2
2
1 h 2
2
i
D m`2 P 2 C 2PP C P C P 2 2PP C P
2
2

D m`2 P 2 C P (4.19)
and from Eq. (4.5)
1 1
UD mg` 21 C 22 C k`2 . 2 1 /2
2 2
1 h i
D mg` 2 C 2 C 2 C 2 2 C 2
2
1
C k`2 Œ. C / . /2
2
h i
D m`2 ! 2p 2 C ! 2p C ! 20 2 (4.20)
Thus, we see that the kinetic and potential energies, T and U are quadratic
functions of the normal coordinates and and their first time derivatives,
consistent with Proposition 3.
Examination of Eq. (4.17) provides further insight into the motion of the
system. In the symmetric mode, ! D ! p , we see that the characteristics
of the spring have no effect on the motion and the two pendulums sway
in unison with frequency ! p as in Fig. 4.1. In the antisymmetric mode the
parameters of both the pendulum and the spring determine the frequency of
the motion [see Eq. (4.17)]. In this mode the spring is alternately compressed
and stretched, as shown in Fig. 4.2, and the frequency of the motion depends
upon the spring constant.
Fig. 4.2
Problem 1b–solution
There is a more general way to obtain the normal coordinates than the
method of inspection employed above. We know that the solutions will be
sinusoidal so we try a solution of the form
i D Ai sin !t (4.21)
and attempt to solve for !. Inserting Eq. (4.21) into Eqs. (4.10) and (4.12)
we obtain
! 2o
A1 ! 2 C ! 2p A1 .A2 A1 / D 0
2
! 2o
A2 ! 2 C ! 2p A2 C .A2 A1 / D 0 (4.22)
2
Taking A1 and A2 as the independent variables we write two simultaneous
equations in the more familiar form

2 ! 2o 2 !2
!p C ! A1 o A2 D 0
2 2
2 2
!o 2 !o 2
A1 C ! p C ! A2 D 0 (4.23)
2 2
These simultaneous homogeneous equations for A1 and A2 may be written

in matrix form
!
!2 !2
! 2p C 2o ! 2 2o A1
! 2o ! 2o
D0 (4.24)
2 2
2
!p C ! 2
A2
The only way Eq. (4.23) can have a non-trivial solution is if the determi-
nant of the coefficients vanishes. This condition yields the secular equation,
which for Eq. (4.24) is
2
! 2o ! 4o
! 2p C !2 D0 (4.25)
2 4
or
! 4o !4
! 4p C ! 2p ! 2o 2! 2p ! 2 ! 20 ! 2 C C !4 o D 0 (4.26)
4 4
so that

! 4 2! 2p C ! 2o ! 2 C ! 4p C ! 2p ! 2o D 0 (4.27)
which is quadratic in ! 2 . Using the quadratic formula to solve for ! 2 we

have
q 2
2! 2 D 2! 2p C ! 2o ˙ 2! 2p C ! 2o 4 ! 4p C ! 2p ! 2o

D 2! 2p C ! 2o ˙ ! 2o (4.28)
so that

!2 !2
! 2 D ! 2p C o ˙ o (4.29)
2 2
These are the same two frequencies that we obtained using the less general
method [see qEq. (4.17)]. Clearly ! D ! p corresponds to the minus sign
while ! D ! 2p C ! 2o corresponds to the plus sign. To recover and we
first substitute ! into the first of Eq. (4.23) to obtain a relation between A1
and A2 for ! D !

! 2p C ! 2o ! 2p A1 ! 2o A2 D 0 H) A1 D A2 D A=2 (4.30)
so that for ! D !
1 D A sin ! t
2 D A sin ! t (4.31)
Thus, 1 D 2 and we have recovered the symmetric solution. The time

dependent motion .t/ is given by
.t/ D A sin ! t (4.32)
which is the solution to Eq. (4.13).

Substituting ! D ! into the second of Equations
2
! p C ! 2o ! 2p 2! 2o A1 ! 2o A2 D 0 H) A1 D A2 D A=2 (4.33)
so that for ! D !
1 D A sin ! t
2 D A sin ! t (4.34)
and 1 D 2 and we have recovered the antisymmetric solution. The time

dependent motion .t/ is given by
.t/ D A sin ! t (4.35)
We see again that the sum and difference of 1 and 2 as given in

Eq. (4.15) provide the sinusoidal response required of the solution to the
uncoupled differential equations in Eq. (4.13). They are the “eigenvectors”

corresponding to the eigenvalues ! and ! . They are, mathematically
speaking, analogous to the unit vectors {O and |O in Cartesian coordinates
in the sense that any motion of the system may be considered to be a
linear combination of and just as any two-dimensional vector may be
considered to be a linear combination of {O and |.
O
2. Find the frequencies and normal mode coordinates for the triple spring system
shown in Fig. 4.3.
Fig. 4.3 Problem 2
Solution
This problem is mathematically identical with Problem 1 of this chapter. For

small pendulum displacements the coupled pendulums execute linear harmonic
vibrations. We can immediately write the normal frequencies appropriate to this
p substituting the frequency ! of an SHO with spring
problem by simply p constant
, that is ! D =m, for the frequency of the pendulum, ! D g=` in the
symmetric and anti-symmetric frequencies of vibration ! s and ! a , Eq. (4.17).
Therefore
r r r r
g g k k
! D ! !s D ; ! D C 2 ! !a D C2 (4.36)
` m ` m m m
Now we justify the analogy and work out the details. We begin by writing
Largrange’s equations.
1 2 1 2
TD mPx C mPx (4.37)
2 1 2 2
and
1 2 1 1
UD x C k .x2 x1 /2 C x22 (4.38)
2 1 2 2
so
1 2 1 2 1 2 1 1
LD mPx1 C mPx2 x1 k .x2 x1 /2 x22 (4.39)
2 2 2 2 2
Lagrange’s equations are then
mRx1 C x1 k .x2 x1 / D 0

mRx2 C x2 C k .x2 x1 / D 0 (4.40)
As in Problem 1 of this chapter we could take the sum and difference of

Eq. (4.40), but here we take the more general approach letting
xi D Ai sin !t (4.41)
Since we are solving this problem using the general method we write the
simultaneous equations for A1 and A2 in matrix form.
0 1
k 2 k
Bm C ! C A1
@ m m A D0 (4.42)
k k A2
C !2
m m m
Solving the secular equation we obtain a quadratic equation in ! 2 the

solutions to which are
k
! 2C D ! 2s D and ! 2 D ! 2a D C 2 (4.43)
m m m
where the ˙ subscripts represent the solutions obtained using the ˙ signs in the
discriminant of the quadratic formula. We see that these frequencies are identical
with those obtained in Eq. (4.36) using the results of Problem 1. Not surprisingly,
the normal coordinates are D x1 C x2 and D x1 x2 as may be seen by adding
and subtracting or by solving Eq. (4.42) for A1 and A2 using the two different
values of !.
3. Find the frequencies and normal mode coordinates for small displacements
of the double pendulum system shown. Do not make the small displacement
approximation until after the Lagrange equations of motion have been derived.
Is it possible for the system to swing as a single pendulum?
Fig. 4.4 Problem 3
Check that the normal mode coordinates that you obtain are correct by
verifying Proposition 3. For this part of the problem care should be taken
in making the small angle approximation in the Lagrangian. The degree of
approximation is different than that required in the equations of motion.
Solution
To find the Lagrangian it is perhaps easiest to first write the Cartesian coordinates
in terms of the coordinates 1 and 2 and then find the kinetic and potential
energies. Letting 1 and 2 denote the coordinates of the upper and lower masses,
respectively, we have
x1 D ` sin 1 ; y1 D ` cos 1
xP 1 D `P 1 cos 1 ; yP 1 D `P 1 sin 1 (4.44)
For the lower mass we have
x2 D ` sin 1 C ` sin 2 ; y2 D ` cos 1 ` cos 2
xP 2 D `P 1 cos 1 C `P 2 cos 2 ; yP 2 D `P 1 sin 1 C `P 2 sin 2 (4.45)

Then, the kinetic energy is
1 2 1
TD m xP 1 C yP 21 C m xP 22 C yP 22
2 2

1 2 P2 1 2 P2 1 2 P2
D m` 1 C m` 1 C m` 2
2 2 2
Cm`2 P 1 P 2 .cos 1 cos 2 C sin 1 sin 2 / (4.46)
This may be re-written with the help of the trigonometric identity, Eq. (E.2):
cos .A B/ D cos A cos B C sin A sin B (4.47)
as

2 1 2
T D m`2 P 1 C P 2 C P 1 P 2 cos . 1 2 / (4.48)
2
U D mgy1 C mgy2
D mg` .2 cos 1 cos 2 / (4.49)

2 1 2
L D m`2 P 1 C P 2 C P 1 P 2 cos . 1 2 / C mg` .2 cos 1 C cos 2 /
2
(4.50)
Lagrange’s equations are
2
2`R 1 C `R 2 cos . 1 2 / C `P 2 sin . 1 2 / C 2g sin 1 D 0
2
`R 2 C `R 1 cos . 1 2 / `P 1 sin . 1 2 / C g sin 2 D 0 (4.51)
These are two coupled nonlinear differential equations the solutions to which
are not only difficult, but probably not very enlightening. As directed in the
statement of the problem we make the small angle approximation to simplify
and retain only the leading terms. Retaining only the leading terms eliminates
2 2
the nonlinearity of the equations because we ignore the P 1 and P 2 terms. Also,
letting
sin i Ð i and cos . 1 2 / Ð 1 (4.52)

we have
2`R 1 C `R 2 C 2g 1 D 0

`R 2 C `R 1 C g 2 D 0 (4.53)
Noting that the frequency of a single simple pendulum is given by

g
! 2p D (4.54)
`
we re-write Eq. (4.53) as
2R 1 C 2! 2p 1 C R 2 D 0
R 2 C R 1 C ! 2p 2 D 0 (4.55)
Letting
j D Aj sin !t (4.56)
we obtain
2A1 ! 2 C 2! 2p A1 A2 ! 2 D 0
A2 ! 2 A1 ! 2 C ! 2p A2 D 0 (4.57)
or

2 ! 2 2! 2p A1 ! 2 A2 D 0

! 2 A1 C ! 2 C ! 2p A2 D 0 (4.58)
The secular equation (see Problem 1 of this chapter) is therefore

ˇ ˇ
ˇ 2 ! 2 ! 2 ! 2 ˇ
ˇ ˇ
ˇ p
ˇD0 (4.59)
ˇ ! 2 ! 2 ! 2p ˇ
or
2
2 ! 2 ! 2p ! 4 D 0 (4.60)
which is a quadratic equation in ! 2 . The two frequencies are
1 p
! 2˙ D 4 ˙ 16 4 1 2 ! 2p
2
p
D 2 ˙ 2 ! 2p (4.61)
If we put the trial solution given in Eq. (4.56) into the first of Eq. (4.55), we
obtain

2! 2 C 2! 2p 1 ! 2 2 D 0 (4.62)
Substituting each of the values of the frequencies in Eq. (4.61) into Eq. (4.62)
leads to a relationship between 1 and 2 . We obtain
1
1 D p 2 ! D !C
2
1
1 D p 2 ! D ! (4.63)
2
We see that the second of the relations in Eq. (4.63), in which 1 and 2 are
in the same direction, is represented by Fig. 4.4. This is commonly known as the
symmetric mode so it is best to change the notation and let ! D ! s . The first of
the relations in Eq. (4.63) represents the antisymmetric mode so ! C D ! a and is
shown in Fig. 4.5.
Fig. 4.5 Problem 3–solution
The frequency of the antisymmetric mode is considerably higher than that

of the symmetric mode, roughly 2:4 times higher. For the system to swing as a
single pendulum we must have 1 D 2 , but Eq. (4.63) show that this is not the
case. Therefore, the double pendulum cannot swing as a single pendulum.
To find the normal coordinates, we seek combinations of 1 and 2 that vary
sinusoidally with each of the normal mode frequencies, ! s and ! a . To do this
we write 1 and 2 as (arbitrary) linear combinations of the sine functions in
Eq. (4.56). First we write
1 D B1 sin ! a t C B2 sin ! s t (4.64)

where, from Eq. (4.61),

p
! 2s D 2 2 ! 2p
p
! 2a D 2 C 2 ! 2p (4.65)
Using the relations between 1 and 2 in Eq. (4.63) we write 2 as a linear

combination of the two possible relations between it and 1 .
p p
2 D 2B1 sin ! a t C 2B2 sin ! s t (4.66)
p
Multiplying Eq. (4.64) by 2, adding and subtracting it to and from Eq. (4.66),
we arrive at

1 2
1 C p D B2 sin ! s t
2 2

1 2
1 p D B1 sin ! a t (4.67)
2 2
Thus, the normal coordinates are

1 2
D 1 C p (antisymmetric)
2 2

1 2
D 1 p (symmetric) (4.68)
2 2
because it is these combinations of 1 and 2 that oscillate with one of the single
normal mode frequencies, ! s () or ! a ().
4. This problem is similar to Problem 6 of Chap. 2 except that the mass at the pivot
point of the pendulum is attached to a spring. A plane pendulum of length ` and
mass m is attached to another equal mass m that is attached to a spring of constant
k that slides on a frictionless surface as shown in Fig. 4.6. Find the normal mode
frequencies using x and as the generalized coordinates.
Fig. 4.6 Problem 4

The coordinate x is the distance from the equilibrium position of the mass that
is attached to the spring. Assume small oscillations after obtaining the equations
of motion.
Solution
The kinetic energy and the potential energy (measured from y D 0) of the mass
on the end of the spring are
1 2 1
Ts D mPx and Us D kx2 (4.69)
2 2
The x and y positions of the pendulum bob are
xp D x C ` sin and yp D ` cos (4.70)
The kinetic energy of the pendulum bob is
1 2
Tp D m xP p C yP 2p
2
2
1
D m xP C 2Px ` cos C `P
2 P (4.71)
2
and the potential energy of the pendulum bob is
Up D mgyp
D mg` cos (4.72)
L D Ts C Tp Us Up
1 2
D mPx2 C mPxP ` cos C m `P
2
1
kx2 C mg` cos (4.73)
2
so the Lagrange equation for is

d @L @L
D0
dt @P @
xR cos xP P sin C `R C g sin D 0 (4.74)
For small oscillations we may drop the term xP P sin . Also, we take sin
and cos 1 and arrive at
xR C `R C g D 0 (4.75)
The Lagrange equation for x is obtained as follows.

2
2mRx m`P sin C m`R cos C kx D 0 (4.76)
2
For small oscillations we drop the P term and again take cos 1 obtaining
2mRx C m`R C kx D 0 (4.77)
Now we assume a sinusoidal solution and, because x and do not have the
same dimensions, we assume the solution to be of the form
A2
x D A1 sin !t and D sin !t (4.78)
`
as was done in Problem 6 of Chap. 2. Inserting these trial solutions into the
equations of motion, Eqs. (4.75) and (4.77) we obtain
g
! 2 A1 C ! 2 A2 D 0
`

k 2m! A1 m! 2 A2 D 0
2
(4.79)
which may be written in terms of the natural frequencies of the oscillator and the
simple pendulum.
k g
! 2o D and ! 2p D (4.80)
m `
Therefore

! 2 A1 C ! 2p ! 2 A2 D 0
2
! o 2! 2 A1 ! 2 A2 D 0 (4.81)
To solve for ! we solve the secular equation.

! 4 ! 2p ! 2 ! 2o 2! 2 D 0

! 4 2! 2p C ! 2o ! 2 C ! 2p ! 2o D 0 (4.82)
Solving the quadratic equation for ! 2 we obtain the normal mode frequencies
which we denote by ! ˙ .

1 2 q 2
! 2˙ D 2
2! p C ! o ˙ 2 2 2 2
2! p C ! o 4! p ! o (4.83)
2
Now we can examine special cases. If, for example, we set ! 2o D 2! 2p , then
the solution becomes

1
! 2 D ! 2o 1 ˙ p (4.84)
2
p
Replacing ! 2 with ! 20 .1 C 1= 2/ in the first of Eq. (4.81) and using
! 2p D ! 2o =2, we get
p
A1 .1=2 C 1= 2/
D p (4.85)
A2 .1 C 1= 2/
p
so the oscillator and pendulum move in opposite directions. Using ! 0 .1 1= 2/
in the first of Eq. (4.81) yields the ratio
p
A1 .1= 2 1=2/
D p (4.86)
A2 .1 1= 2/
which is positive so the pendulum and spring oscillate in the same direction.
Suppose we let the spring constant k ! 0 in Eq. (4.80). The two roots for this
case become
! 2 D 0; 2! 2p (4.87)
Setting k D 0 allows the block to become freely sliding, and the ! 2 D 0

root represents the constant velocity solution obtained in Problem 6 Chap. 2. The
! 2 D 2! 2p root represents the pendulum motion obtained in that problem for the
equal-mass case.
5. A very thin uniform rod of length ` and mass m hangs from two identical
massless springs with spring constant k (Fig. 4.7). The coordinate y is the vertical
displacement of the rod center from its equilibrium position. The displacements
of the rod ends are y1 D .`=2/ sin and y2 D .`=2/ sin . Consider only
vertical motion of the rod-spring system and assume small displacements.
Fig. 4.7 Problem 5

(a) Describe the system using coordinates .y1 ; y2 / where each represents the
coordinate of an end of the rod. Find the normal mode coordinates and
frequencies.
(b) Use the .y; / coordinates to find the normal mode coordinates and
frequencies.
Solution
(a) Using Newton’s second law the force is given by

m
k.y1 C y2 / mg D .Ry1 C yR 2 / (4.88)
2
The torque about the center of the rod, D I˛ (˛=angular acceleration,
I Dmoment of inertia), is

I k`
.Ry2 yR 1 / D .y2 y1 / (4.89)
` 2
where we have used
1
˛D .Ry2 yR 1 / (4.90)
`
consistent with the small-displacement assumption. The moment of inertia
of the thin rod about its center of mass is
1
ID m`2 (4.91)
12
Equations (4.88) and (4.89) are easily separable if we define new coordinates
mg
D y1 C y2 C
k
D .y2 y1 / (4.92)
In the new coordinates Eqs. (4.88) and (4.89) become

2k
R C D0 (4.93)
m
and

6k
R C D0 (4.94)
m
r
2k
! D (4.95)
m
Equations (4.93) and (4.94) describe harmonic motion in with frequencies

r r
2k 6k
! D and ! D (4.96)
m m
respectively.
The coordinates and are therefore the normal coordinates in the
.y1 ; y2 / system. (The constant term mg=k in the equation of transformation
for , Eq. (4.92), has no effect on the frequency.)
(b) The kinetic energy for the system in the .y; / system, after substituting for
I, is
1 P2 1 2
TD I C mPy
2 2
1 2 1
D m`2 P C mPy2 (4.97)
24 2
2 2
1 ` 1 `
U D mgy C k y C sin C k y sin
2 2 2 2
`2 2
Ð mgy C ky2 C k (4.98)
4
We note that T contains only the squares of the time derivatives of the
coordinates and that a simple translation of the y coordinate would make
U dependent upon only the squares of both coordinates. Thus, from Propo-
sition 3 we know that is a normal mode coordinate and that a simple
translation of y will produce the other normal coordinate.
The Lagrangian is
L D T U
1 2 1 `2
D m`2 P C mPy2 C mgy ky2 k 2 (4.99)
24 2 4
and the Lagrange equations of motion are
k
yR C 2 y g D 0
m
k
R C 6 D 0 (4.100)
m
Equation (4.100) are seen to represent harmonic motion in y and with

frequencies
r r
k k
!y D 6 and ! D 6 (4.101)
m m
As in part (a) the constant term in the y-equation does not affect the
calculated frequency. The frequencies are identical to those obtained in part
(a) as they must be. This result is not a surprise, because with the small-angle
approximation y and are proportional to and , respectively, so they are
essentially the same coordinates.
6. Three particles are arranged along a line as shown in Fig. 4.8. The two end
particles are identical, each with mass M, and the middle particle has mass m.
The particles are connected by identical springs each having a spring constant k,
and they are constrained to move along the x-axis so that there is no bending
motion. This system can be thought of as representing the linear vibrations
a linear triatomic molecule such as CO2 . The coordinates x are displacement
coordinates, and x1 D x2 D x3 D 0 when the system is at rest.
Fig. 4.8 Problem 6
(a) Construct the Lagrangian in terms of the xi and find the Lagrangian equations
of motion.
(b) Assume sinusoidal motion so that xi D Ai sin !t where Ai is the amplitude
of the oscillations in the xi coordinate. Show that substituting this expression
for xi into each of the three Lagrange equations leads to three simultaneous
equations for ! 2 and solve for each value of ! 2 . These are the normal mode
frequencies.
(c) Find the relationships between the xi that describe the normal mode motion
and discuss this motion.
Solution
(a) The kinetic energy for each displacement coordinate is
1 2 1 2 1 2
TD MPx C mPx C MPx (4.102)
2 1 2 2 2 3
and the potential energy is
1 1
UD k .x2 x1 /2 C k .x3 x2 /2 (4.103)
2 2
LDT U
1 2 1 2 1 2 1 1
D MPx1 C mPx2 C MPx3 k .x2 x1 /2 k .x3 x2 /2 (4.104)
2 2 2 2 2
The Lagrangian equations of motion are given by
d @L @L
D0 (4.105)
dt @Pxi @xi
which yield
MRx1 k .x2 x1 / D 0
mRx2 C k .x2 x1 / k .x3 x2 / D 0
MRx3 C k .x3 x2 / D 0 (4.106)
(b) Substituting
xi D Ai sin !t (4.107)
into the Lagrange equations of motion, Eq. (4.106), we obtain

k k
!2 A1 C A2 D 0
M M

k 2 k k
A1 C ! 2 A2 C A3 D 0
m m m

k 2 k
A2 C ! A3 D 0 (4.108)
M M
The only way these homogeneous equations can have a non-trivial

solution is if the determinant of the coefficients vanishes. The secular
equation is
ˇ 2 ˇ
ˇ ! k k
0 ˇ
ˇ M M ˇ
ˇ k
!2 2 m k k ˇD0 (4.109)
ˇ m 2 m k ˇˇ
ˇ 0 k
! M
M
or

2 k 2 k 2 k k2
! ! 2 !
M m M mM

k k k
!2 D0 (4.110)
M m M
which reduces to

2 k 4 k 2 k 2 k2 k2 k2
0D ! ! ! 2 ! C2
M M m mM mM mM

2 k 4 k 2 k 2
0D ! ! ! 2 !
M M m

k k k
0 D !2 !2 !2 C2 (4.111)
M M m
There are three different values of ! 2 that are solutions to this equation.
They are
! 21 D 0
k
! 22 D
M

2 m C 2M
!3 D k (4.112)
mM
(c) To describe the motion of each of the three normal modes we must find
the relations among the mode amplitudes for each frequency listed in
Eq. (4.112). We must therefore replace ! 2 in Eq. (4.108) with ! 1 , ! 2 and
! 3 and examine the motion for each of these normal modes.
• ! 2 D ! 21 D 0. From the first of Eq. (4.108) we see that A1 D A2 . From
the third of these equations we obtain A3 D A2 . There is therefore no
relative motion of the masses. In this normal mode each atom in the
molecule is simply shifted along the axis by the same amount, so that
the entire molecule moves (translates) without compressing or stretching
the springs. This is referred to as “free” motion.
• ! 2 D ! 22 D k=M. From the first (or third) of Eq. (4.108) we find
that A2 D 0. Inserting A2 D 0 into the second equation we obtain
A1 D A3 . For this mode of vibration the center mass m is stationary
while the outside masses vibrate in opposite directions, each with the
same amplitude. This is referred to as the “symmetric stretch” mode.
• ! 2 D ! 23 D Œ.m C 2M/ =mM k. Inserting ! 23 into the first of Eq. (4.108)
we find that
References 117

k k k k 1m
C2 A1 C A2 D 0 ) A1 D A2 (4.113)
M m M M 2M
Now we eliminate A2 in the second of Eq. (4.108)

k .m C 2M/ k M k
A1 k2 2 A1 C A3 D 0
m mM m m m

k k k k M k
A1 C2 2 2 A1 C A3 D 0
m M m m m m
k k k
A1 2 A1 C A3 D 0
m m m
A1 D A3 (4.114)
In this mode we see that the two outer masses vibrate in the same direction
with the same amplitude while the center mass vibrates in the opposite
direction with a different amplitude. This is called the “asymmetric stretch”
mode.
References
1. Burkhardt CE, Leventhal JJ (2006) Topics in atomic physics. Springer, New York
2. Goldstein H (1980) Classical mechanics, 2nd edn. Addison-Wesley, Reading.
3. Thornton ST, Marion JB (2004) Classical dynamics of particles and systems, 5th edn. Harcourt
Brace Jovanovich, New York
Part II
Quantum Mechanics
Chapter 5
Introductory Concepts
The important relationships and notation are in the appendices.
Problems
1. In creating the Bohr theory of the atom Niels Bohr assumed that the electron
travels around the proton in a series of circular orbits. He set forth two
postulates.
I. An atom exists in a series of energy states such that the
accelerating electron does not radiate energy when in these
states. These states are called stationary states.
II. Radiation is absorbed or emitted during a transition
between two stationary states. The frequency of the
absorbed or emitted radiation is given by Planck’s theory.
In many books a third postulate is added that has the form

III. The angular momentum of the orbits of the allowed states
is quantized in units of „.
In fact, Bohr made no such postulate as the last one. Why would he? Did he
have an epiphany? Ask yourself why would he choose „ and not, for example
h or 2h or some other constant. Using Postulate III in a derivation of the
Bohr model of the atom minimizes Bohr’s contribution. Bohr did have a third
postulate, known today as the correspondence principle, but he did not state it
as a postulate. The correspondence principle states that when quantum numbers
become large the system behavior tends toward that of comparable classical
systems. Bohr used his correspondence principle together with the observations
of the hydrogen atom (H-atom) spectrum made by the spectroscopist Johann
Balmer to deduce the correct energy levels. Using “Postulate III” above, the
correct energy levels are obtained without appealing to the correspondence

122 5 Introductory Concepts
principle. This principle is then presented as a consequence of the Bohr model

of the atom, but this was not Bohr’s line of reasoning.
In addition to the postulates, Bohr drew heavily on previous work by others,
particularly Planck, Einstein, and Balmer. Planck and Einstein showed that
there is a relation between the frequency of emitted light and the energy of
the photon emitted.
E D h D „! (5.1)
where , the frequency in hertz, is equal to !=2. Balmer had shown that the
wavelength of the lines in the visible spectrum of hydrogen were given by an
empirical formula

1 1 1
D RH 2
2 (5.2)
n 2 n
where n is an integer and RH is a constant known as the Rydberg constant

because of earlier work on spectroscopy by Johannes Rydberg. In fact, he
generalized Eq. (5.2) by replacing the 2 in the denominator with another integer.
Now, the problem: From the Bohr model [1], using the correspondence
principle, the radius of the nth Bohr orbit rn is given by
rn D n2 a0 (5.3)
where a0 is the radius of the first Bohr orbit
.4 0 / „2 e2 „2
a0 D ) D
me e2 4 0 me a0
„
D (5.4)
me c˛
The constant ˛ in Eq. (5.4) is called the fine structure constant and is given by

e2
˛D
.4 0 / „c
1
w (5.5)
137
This constant ˛ is a fundamental and unitless quantity.
Beginning with Eq. (5.3) derive Postulate III given above. Show also that the
speed of the electron in the nth Bohr orbit is vn D ˛c=n w .1=137/ c=n.
5 Introductory Concepts 123
Solution
In the construction of the Bohr model of the atom the centripetal force is
equated to the Coulomb force giving

me v 2 e2 1
D (5.6)
r 4 0 r2
which is a relationship between the orbital radii r and the orbital velocity v.
Assuming quantized orbits (Postulate I above), attaching subscripts to the
parameters of these orbits and solving Eq. (5.6) for vn we obtain
s
1 e2 1
vn D (5.7)
me 4 0 n2 a0
Using Eq. (5.4), Eq. (5.7) may be rewritten as

s
1 „2 1
vn D
me me a0 n2 a0
1 „
D (5.8)
me na0
The angular momentum in the nth orbit is

1 „
Ln D me vn rn D me n2 a0
me na0
D n„ (5.9)
so Eq. (5.9) shows that the angular momentum is quantized in units of „. This
result was obtained using Postulates I and II and the correspondence principle.
To find v in terms of the speed of light and the fine structure constant we
solve Eq. (5.9) for vn and use Eqs. (5.4) and (5.5) to obtain
n„
vn D
me rn
n„ 1
D
me n2 a0

„ me c˛
D
nme „
˛ 1 c

D c' (5.10)
n 137 n
This is an important relation because it shows that the orbital speed vn of the
electron is much smaller than c. Moreover, the electronic speed decreases as n
increases, thus justifying a nonrelativistic treatment of the quantum mechanical
H-atom (ignoring spin which is an inherently relativistic phenomenon). One
way of describing the fine structure constant is that it is the ratio of the electron
velocity in the first Bohr orbit to the speed of light.
Some of the concepts that Bohr employed to construct his model are
incorrect. For example, the idea of a well-defined orbit for the electron
is incorrect. Additionally, the result derived in this problem that the
angular momentum is equal to n„, Eq. (5.9), is not entirely correct because
n D 1; 2; 3 : : : and we now know that the angular momentum of an electron in
an atom can be zero. Nonetheless, Bohr’s application of empirical facts and his
assertion of the correspondence principle were the stepping stones to modern
quantum physics. Although physicists are well aware of the wave nature of
matter and the consequent position and momentum distributions for an electron
bound to a nucleus, most occasionally find themselves thinking in terms of the
Bohr model if they are not careful.
An important property of the Bohr model is that it gives the correct order
of magnitude for atomic parameters. The Bohr model, in most cases, gives the
correct dependence of the atomic parameters on the principal quantum number.
Perhaps the most dramatic consequence of the Bohr model is that it gives the
correct quantized energies of the H-atom.
2. Find the expression for the radius of the first Bohr orbit ignoring the electro-
magnetic interaction between the electron and the proton and assuming that the
only force that binds them is the gravitational force.
Solution
The Coulomb attractive force between the proton and the electron in the Bohr
atom is
2
e 1
F .r D a0 / D (5.11)
4 0 a20
Denoting the radius of the first Bohr orbit that is due to only the gravitational
force by aG
0 the gravitational force between them at a0 is
G
1
0 D Gme mp G 2
F r D aG (5.12)
a0
where G is the gravitational constant; me and mp are the masses of the electron
and proton, respectively. Comparing the two expressions for the forces in
Eqs. (5.11) and (5.12) we see that the quantities in parentheses are analogous.
The expression for a0 , the “real” Bohr radius [see Eq. (5.4)] is

4 0 „2
a0 D
e2 me
Ð 5:3 1011 m (5.13)
so we simply replace 4 0 =e2 by 1=Gme mp to obtain aG

0 . This gives
„2
0 D
aG Ð 1:1 1029 m (5.14)
Gmp m2e
A currently accepted value for the diameter of the observable universe is the
order of 1027 m, roughly one hundred times smaller than aG 0 which suggests that
gravity plays little role in determining interatomic forces.
3. Show that the de Broglie wavelength of an electron in the nth Bohr orbit is
constant and is 1=n times the circumference of the nth Bohr orbit. This shows
that de Broglie waves “fit” into the Bohr orbits thus justifying the designation
of the Bohr states as “stationary states.”
Solution
Equating the centripetal force to the Coulomb force that binds the electron to
the proton in the nth Bohr orbit we have

me vn2 e2 1 e2 1
D ) p2n D me (5.15)
rn 4 0 rn2 4 0 rn
Bohr’s quantized radii are given by [1]

4 0 „2 e2 „2
rn D n2 a0 where a0 D ) D (5.16)
e2 me 4 0 a0 me
so the linear momentum in the nth Bohr orbit is

2
2 e 1
pn D me 2
4 0 n a0
2
„ 1
D me
a0 me n2 a0
h2 h
D H) pn D (5.17)
4 2 n2 a20 2na0
The de Broglie wavelength in the nth orbit is [see Eq. (L.6)]
h
n D D 2na0
pn
1
2
D 2 n a0 (5.18)
n
or
nn D .2rn / D circumference of the nth Bohr orbit (5.19)
It is interesting to note that the de Broglie wavelength for a Bohr orbit

(a stationary state) is consistent with the correspondence principle. The angular
momentum in the nth Bohr orbit is
Ln D pn rn (5.20)
so that

h 2
Ln D n a0 n„ (5.21)
2na0
4. Use the uncertainty principle to estimate the ground state energy of the H-atom.
Solution
The uncertainty principle is
xp „=2 (5.22)
Now, one often sees these “back of the envelope” calculations with the
uncertainty principle performed in such a way that the answer conforms with
the exact (known) answer. We take the approach that we do not know the answer
until the end and then compare the computed and exact results. To that end we
take the uncertainty in x for the ground state of the H-atom to be some fraction
ˇ of the orbital radius r of the ground state orbit. Thus, x Ð ˇr. Because we
know nothing about the momentum we take the uncertainty in the momentum
p to be the momentum itself so p Ð p. Inserting these estimates into the
uncertainty principle we have
ˇrp Ð „=2 (5.23)

Using Eq. (5.23) total energy is
p2 e2
ED
2me 4 0 r
„2 e2
D 2
(5.24)
8ˇ me r2 4 0 r
where me is the rest mass of the electron.

To estimate the ground state energy we must minimize E with respect to r to
obtain the corresponding value, rmin :
dE „2 e2
D 2 3
C 2
D0 (5.25)
dr 4ˇ me rmin 4 0 rmin
3
Multiplying through by rmin we have

„2 e2 rmin 1 4 0 „2
D H) rmin D (5.26)
4ˇ 2 me 4 0 4ˇ 2 me e2
Now, the Bohr radius of the atom, which is the radius of the first Bohr
orbit, is
.4 0 / „2
a0 D (5.27)
me e2
If we had chosen ˇ D 12 , then we would have obtained the exact value of

a0 that leads to the exact energy of the ground state of the H-atom. The exact
value of ˇ and the uncertainties x and p that are chosen are not important
because this is an estimate. For example, if the reasonable values
ˇ D 1 ; x Ð r ; p Ð p ; xp D „ (5.28)
are chosen, then the exact answer (rmin D a0 ) is obtained, but this is a designed
coincidence. What is important is that the order of magnitude is correct.
In terms of the fine structure constant ˛ [see Eq. (C.1)] the energy of the
H-atom ground state is given by
1
E D ˛ 2 me c2 (5.29)
2
where c is the speed of light and the fine structure constant is

e2
˛D
.4 0 / „c
1
w (5.30)
137
It is easy to calculate the ground state ionization energy of the H-atom

using Eq. (5.29). Given the electron’s rest mass (which should be known by all
physicists) is me c2 D 0:51 MeV and ˛ D 1=137, one obtains E D 13:6 eV.
5. Equate the electron’s rest energy to the electrostatic energy of the electron
regarded as a charge e that is uniformly distributed throughout a sphere of radius
re . Use this relation to estimate re . This radius is known as the classical radius
of the electron. Put your answer in terms of the fine structure constant ˛ and
the Bohr radius a0 (see Problem 4 of this chapter). The electrostatic energy of
a uniform distribution of charge e throughout the sphere is given by [3]

3 e2
Eelectrostatic D (5.31)
5 4 0 re
but, because this approach gives a very rough approximation to re , the 3=5 is
usually omitted. Comment on the magnitude of the answer obtained.
Solution
Following the directions given in the statement of the problem we have
e2
me c2 D (5.32)
4 0 re
so that
e2
re D (5.33)
4 0 me c2
Using Eqs. (5.27) and (5.30) as guides we write

e2 .4 0 / „2 1 me e2
re D 2 (5.34)
4 0 me c2 me e2 .4 0 / „
where the expression in curly brackets is unity. Regrouping we have

2
e2 .4 0 / „2
re D
4 0 „c2 me e2
2
1
D ˛ 2 a0 a0 (5.35)
137
Thus, from this point of view, the electron’s “size” is the order of 105 the
orbital radius of the electron.
6. The magnetic moment of a uniformly charged sphere of radius R is spinning

with an angular frequency ! and carrying a total charge Q is given by
Q!R2
D (5.36)
5
Use this result to find the velocity vs of a point on the surface of a uniform
spherical charge distribution, the radius of which is the classical radius of the
electron re as given by Eq. (5.35). Show that this point on the surface would
be moving at several hundred times the speed of light if the magnetic dipole
moment due to the spinning electron is taken to be the measured value, one
Bohr magneton B .
e„
B D (5.37)
2me
Solution
In this problem we may insert into Eq. (5.36) R ! re , Q ! e, ! ! vs =re

where vs is the speed of a point on the surface of the assumed spherical electron.
Equating this magnetic moment to B in Eq. (5.37) we have
evs re e„
B D D (5.38)
5 2me
so that

5 „
vs D (5.39)
re 2me
At this point we must put in values for the quantities in Eq. (5.39) to
determine the order of magnitude of vs relative to c the speed of light. It is
much easier if we use atomic units (a.u.) (see Table C.1 of Appendix C) for
these quantities because all quantities in Eq. (5.39) except re are unity in that
system. From Eq. (5.35) we can write re in terms of a0 (which is unity in a.u.),
i.e. re D ˛ 2 a0 where ˛ is the fine structure constant [see Eq. (C.1)]. We have

5 5
vs D D 1372 340c (5.40)
2˛ 2 2
because the speed of light in a.u. is 137. Therefore, in this model a point on
the surface of the sphere would be moving about 340 times the speed of light.
The crux of this exercise is to show that the term “spin,” which was invented to
satisfy the physicists concept of reality, is not to be taken literally. Remember
that we are working with nonrelativistic quantum mechanics in which particles
are point particles. The idea that a point is spinning is, of course, nonsense.
Nonetheless, most physicists think of a spinning electron even if they will not
admit it.
7. Use the uncertainty principle to estimate the ground state energy of an

SHO. Use the information in Appendix O that the ground state position and
momentum eigenfunctions are Gaussians. Moreover, Gaussians produce the
minimum uncertainty wave packet so that
xp D „=2 (5.41)
Solution
We take the uncertainties to be the quantities themselves, a reasonable assump-

tion (because we know we will obtain the correct answer). That is
x D x and p D p (5.42)
so
„
pD (5.43)
2x
The total energy is then
p2 1
ED C m! 2 x2
2m 2
„2 1
D 2
C m! 2 x2 (5.44)
8mx 2
Minimizing E with respect to x to obtain xmin we have
ˇ
dE ˇˇ „2
ˇ D 3
C m! 2 xmin D 0
dx xDxmin 4mxmin
r
„
H) xmin D (5.45)
2m!
Substituting this value for xmin back into the expression for the total energy,
Eq. (5.44), we obtain the estimated ground state energy E0
„2 1
E0 D 2
C m! 2 xmin
2
8mxmin 2

„2 2m! 1 „
D C m! 2
8m „ 2 2m!
1 1
D „! C „!
4 4
1
D „! (5.46)
2
That we obtained the exact value in this case is not surprising in view of the
fact that we knew the correct answer a priori. The most suspicious move that
we employed was that of taking the uncertainties to be equal to the quantities
themselves.
8. The normalized wave function at t D 0 for a particle of mass m is given by

"p p #
2 1 10
‰ .x; 0/ D .x/ C .x/ C 3 .x/ (5.47)
4 1 2 2 4
O
where the n .x/ are the orthonormal eigenfunctions of the Hamiltonian H.
Each n .x/ has energy eigenvalue En .1=n/ E0 where E0 is a positive
number.
(a) If an energy measurement is made, what are the possible results of the
measurement?
(b) What is the probability of measuring each of these energies?
(c) What is the expectation value of the energy?
(d) Suppose there is another physical quantity that may be measured, a quantity
that is represented mathematically by Q. Assume that the n .x/ are
O with
also eigenfunctions of the operator that represents this quantity, Q,
eigenvalues nQ0 . That is
QO n .x/ D nQ0 n .x/ (5.48)
If the energy is measured first and found to be E3 D .1=3/ E0 and then

a measurement of Q is made, what will be the value of Q that is measured?
Solution
(a) The only constituents of ‰ .x; 0/ are those representing n D 1, 2, and 3.

Therefore the only values of the energy that can be measured are
E0 E0 E0
I I (5.49)
1 2 3
(b) Because ‰ .x; 0/ is normalized, the sum of the squared coefficients of the
n .x/ is unity and the probability of measuring each of the energies is
(in order)
2 4 10
I I (5.50)
16 16 16
(c) Remembering that the n .x/ are orthonormal we have

Z 1
hEi D O .x; 0/ dx
‰ .x; 0/ H‰
1

2 E0 4 E0 10 E0
D C C
16 1 16 2 16 3
1
D .2 C 2 C 10=3/ E0 D 0:46E0 (5.51)
16
(d) This part of the problem is truly fundamental quantum mechanics. Any
measurement of a quantum mechanical system forces the system into an
eigenstate of the operator corresponding to the quantity that was measured.
Therefore, because the energy was measured to be E3 the measurement
“collapses the wave function” into the n D 3 state. This means that the
wave function after the measurement is no longer given by Eq. (5.47). It is
now simply the eigenfunction 3 .x/. Thus, after the measurement at t D 0
the collapsed wave function ‰collapsed .x; 0/ is simply
‰collapsed .x; 0/ D 3 .x/ (5.52)
From Eq. (5.48) we know that 3 .x/ is an eigenfunction of QO such that
QO 3 .x/ D 3Q0 3 .x/ (5.53)
so the measurement of Q yields 3Q0 .
9. The wave function at t D 0 for a one-dimensional harmonic oscillator (SHO) is

given by
1 2
.x/ D p 1 .x/ C p 2 .x/ (5.54)
5 5
where the n .x/ are eigenfunctions of the time independent harmonic oscillator
Hamiltonian.
r
˛ 2 2
1 .x/ D p 2 Œ.˛x/ e˛ x =2
2
r
˛ h i 2 2
2 .x/ D p 2 .˛x/2 1 e˛ x =2 (5.55)
2
(a) Find ‰ .x; t/, the wave function for t > 0.

(b) Find the expectation value of the energy for t > 0.
(c) Find hx .t/i, the expectation value of the position for t > 0.
Solution
(a) Simply multiply each term in the expansion by the time dependent part of
the eigenfunction. Therefore

1 3 2 5
‰ .x; t/ D p 1 .x/ exp i!t C p 2 .x/ exp i!t (5.56)
5 2 5 2
(b)
Z 1
hEi D O .x; t/ dx
‰ .x; t/ H‰
1
Z 1
1 3
D ‰ .x; t/ p E1 1 .x/ exp i!t
1 5 2

2 5
C p E2 2 .x/ exp i!t dx
5 2
1 4
D E1 C E2
5 5
1 3 4 5 23
D „! C „! D „! (5.57)
5 2 5 2 10
where we have omitted writing the components of ‰ .x; t/ in the integrand.
Note that the cross terms in Eq. (5.57) vanish because the n .x/ are
orthonormal. As a result hEi is independent of time. Also, because .x/
is heavily weighted toward 2 .x/ (by a factor of 2) the average energy is
skewed toward E2 D .25=10/ „! as is clear from Eq. (5.57).
(c)
hx .t/i D h‰ .t/j xO j‰ .t/i
Z 1
D ‰ .x; t/ x‰ .x; t/ dx (5.58)
1
Comparing this integral to that in part (b) we see that the cross terms do
not vanish here because the components of ‰ .x; t/ are not eigenfunctions
O Therefore, there will be a time dependence in the
of x as they were of H.
expectation value.
Z 1
hx .t/i D ‰ .x; t/ x‰ .x; t/ dx
1
Z 1 Z 1
1 2 4
D j 1 .x/j xdx C j 2 .x/j2 xdx
5 1 5 1
Z 1 Z 1
2 2 i!t
C ei!t 1 .x/ x 2 .x/ dx C e 2 .x/ x 1 .x/ dx
5 1 5 1
(5.59)
The first two integrals vanish because they have odd integrands. The
second two must, however, be evaluated. The eigenfunctions n .x/ are
real so these two integrals are identical. We have
Z
2 i!t 1
hOx .t/i D e C ei!t 1 .x/ x 2 .x/ dx
5 1
Z 1
4
D cos !t 1 x 2 .x/ dx (5.60)
5 1
The integral is
Z 1

ID 1x 2 .x/ dx
1
Z 1n h i o
˛ 2 2
D p Œ2 .˛x/ x 2 .˛x/2 1 e˛ x dx
2 1
Z 1
˛ 2 2 2 2
D p 2˛ 3 x4 e˛ x ˛x2 e˛ x dx
1
Z Z 1
2˛ 4 1 4 ˛2 x2 ˛2 2 2
D p x e dx p x2 e˛ x dx (5.61)
1 1
To evaluate the integral in Eq. (5.61) we use Eq. (G.4) which is

Z 1
2 Œ.m C 1/ =2
xm eˇx dx D (5.62)
0 2ˇ .mC1/=2
where Œ.m C 1/ =2 is a -function (see Appendix I). Changing the limits
of integration in Eq. (5.61) to match those in Eq. (5.62) we have

2˛ 4 Œ5=2 ˛ 2 Œ3=2
I=2 D p 5
p (5.63)
2˛ 2˛ 3
From Appendix I we have

p p
3
Œ3=2 D and Œ5=2 D (5.64)
2 4
so, using Eq. (G.4)
p p
2˛ 4 1 3 ˛2 1
I=2 D p p
2˛ 5 4 2˛ 3 2
3 1 1 1
D D )ID (5.65)
4˛ 4˛ 2˛ ˛
and
4
hOx .t/i D cos !t
5˛
10. For an L-box with potential energy
U .x/ D 0 0xL
D1 1<x<0;L<x<1 (5.66)
the eigenfunctions are given in Eq. (M.2). They are

nx
n .x/ D A sin 0 x L n D 1; 2; 3 : : :
L
D1 1<x<0;L<x<1 (5.67)
(a) Calculate the classical probability density of finding the particle in an

increment of space x within the box. ˝ ˛
(b) Find the classical average values hxiclassical and x2 classical and show that the
˝ ˛
quantum mechanical values hxi and x2 approach the classical values as the
quantum number n ! 1.
Solution
(a) The speed of the particle in the box v is constant because the potential
energy within the box is constant (zero). The probability of finding the
particle in a given interval x must then be constant. Therefore, the
classical probability of finding the particle in an increment x is simply
x
Pcl .x/ D (5.68)
L
(b) The average value of any function f .x/ over an interval .x1 ; x2 / is
Z
1 x2
hf .x/i D f .x/ dx (5.69)
.x2 x1 / x1
so
Z
1 L
hxiclassical D xdx
L 0
L
D (5.70)
2
and
Z
˝ 2˛ 1 L
x classical D x2 dx
L 0
L2
D (5.71)
3
Quantum mechanically
Z nx nx
2 L
hxin D sin x sin dx
L 0 L L
Z nx
2 L
D x sin2 dx (5.72)
L 0 L
where we have used the eigenfunctions for an L-box as given in Eq. (5.67).
Let
nx
yD (5.73)
L
so that
2 Z
2 L n
hxin D y sin2 ydy
L n 0
n
2L y2 y sin .2y/ cos .2y/
D
n2 2 4 4 8 0
L
D (5.74)
2
where we have used the integration formula in Eq. (G.9). Clearly the
classical and quantum mechanical values of hxi are identical. Note that the
quantum mechanical value is independent of the quantum number n.
The quantum mechanical expectation value of x2 in the nth eigenstate is
Z nx
˝ 2˛ 2 L
x nD x2 sin2 dx (5.75)
L 0 L
Using the same substitution as above, Eq. (5.73), we have

3 Z
2 L n
hxin D y2 sin2 ydy
L n 0
2 n
2L y3 y2 1 y cos .2y/
D sin .2y/
n3 3 6 4 8 4 0
3 3
2L2 n3 3 n L2 n n
D 3 3 D 3 3 (5.76)
n 6 4 n 3 2
2
L 3
D 1 2 2 (5.77)
3 2n
where we have used the integration formula ˝in Eq.
˛ (G.10).
While the quantum mechanical value of x2 n differs from the classical
value we see that the quantum mechanical value approaches the classical
value as n ! 1. This is a manifestation of Bohr’s correspondence
principle.
11. A rectangular potential barrier of length L and height U0 has a beam of particles
each of mass m and kinetic energy E < U0 incident on it from the left as shown
in Fig. 5.1.The potential energy is
U .x/ D 0 ; x < 0
D U0 ; 0 < x < L
D0;x>L (5.78)
Find the transmission coefficient T and the reflection coefficient R for this
barrier.
Fig. 5.1 Problem 11. The U0

rectangular potential barrier
Mass of particle = m
Kinetic energy = E
I II III
0 L x
Solution
Classically, because E < U0 the particles would simply bounce off the
barrier. Quantum mechanically, however, this total reflection is not the case.
The occurrence of transmission can be understood if we first write the wave
functions in each of the three regions of space. Because the potential energies
in each of these regions is constant the wave functions are

‰I .x; t/ D Aeikx C Beikx ei!t
‰II .x; t/ D .Cex C Dex / ei!t

‰III .x; t/ D Feikx C Geikx ei!t (5.79)
where
k2 D 2m .E U0 / =„
2 D 2m .U0 E/ =„ (5.80)
and
! D E=„ (5.81)
Quantum mechanics permits the incoming beam to penetrate the classically

forbidden region (inside the barrier). If the wave function is non-zero at x D L,
then particles can emerge from the barrier into region III. These particles have
“tunneled” through the barrier.
To solve the problem we must find the coefficients in Eq. (5.79) using the
usual requirements of continuity of the wave function and its first derivative at
the boundaries. We need not consider the time component of the wave functions
in Eq. (5.79) because !, the energy corresponding to the spatial part of the wave
function is the same in all three regions of space.
We immediately set the constant G D 0 in Eq. (5.79) because eikx
represents a plane wave traveling in the x direction. While it is possible that
such a wave will be present in region I (reflection), there cannot be any particles
travelling in the x direction in region III (see Problem 13 of this chapter).
Because a plane wave traveling in the x direction is possible in region I we
must assume that B ¤ 0. There is also a wave function in region II, but, E < U0
so the exponents are real. The wave functions without the time dependence are
I .x/ D Aeikx C Beikx x<0
II .x/ D Cex C Dex 0<x<L
III .x/ D Fe ikx

L<x (5.82)
To better understand the boundary conditions and to solve for the constants
in each region, we employ the equation for the probability current density j .x/,
which must be conserved and continuous. The probability current density is
defined as [see Eq. (L.12)]

„ @ .x/ @ .x/
j .x/ D .x/ .x/ (5.83)
2im @x @x
Applying Eq. (5.83) to regions I and III the probability currents are
„k 2
jI .x/ D jAj jBj2 (5.84)
m
and
„k
jIII .x/ D jFj2 (5.85)
m
Because the velocities, „k=m, are the same in regions I and III

jAj2 jBj2 D jFj2 (5.86)
The two probability terms in jI .x/ represent the incident and reflected
probability currents. It is then clear that the transmission coefficient T must
be the ratio of the transmitted probability current to the incident current so that
jFj2
TD
jAj2
Similarly, the reflection coefficient is
jBj2
RD (5.87)
jAj2
Because the wave must be either reflected or transmitted, we have
RCT D1 (5.88)
In this problem the potential, and therefore the wave number k, is the same in
regions I and III. If the potential in region III were different from that in region
I, the definitions of R and T would be the same, but their relative values would
change.
Before evaluating the constants we can sketch the wave functions in each
region. The nature of these wave functions is shown in Fig. 5.2.
Fig. 5.2 U0
Problem 11—solution.
Sketches of the wave
functions that apply to
penetration of the rectangular
potential barrier
I II III
0 L x
The de Broglie wavelengths in regions I and III are identical. The amplitude
in region III is smaller because there will be some reflection at the barrier. The
curvature of the wave function inside the classically forbidden barrier is away
from the x-axis.
To evaluate the constants A, B, and F, we require the wave functions and
their derivatives to be continuous at the boundaries, x D 0 and x D L. We have
ACB D CCD
ik .A B/ D .C D/
L
Ce C DeL D FeikL

CeL DeL D ikFeikL (5.89)
These are four simultaneous equations with five unknown coefficients. We are
free to choose any one of them so we set A D 1 for convenience. Adding and
subtracting the third and fourth relations in Eq. (5.89) allows us to express C and
D in terms of F. Using these expressions in the first and second relations, with
A D 1, allows evaluation of A and B, and thus the reflection and transmission
amplitudes, R and T. After some algebra we arrive at
" 2 #1
1 k2 C 2 2
TE<U0 D 1C sinh .L/ (5.90)
4 k
After substituting the and k into Eq. (5.90) we obtain
1
TE<U0 D (5.91)
U02 Lp
1C sinh2 2m .U0 E/
4E .U0 E/ „
and, using Eq. (5.88), we have

2 31
6 4E .U0 E/ 7
RE<U0 D 6
41 C 7
5 (5.92)
Lp
U02 sinh2 2m .U0 E/
„
Equation (5.91) shows that, as expected, there is transmission through the

barrier even when E < U0 . It is instructive to take this system to the classical
limit, „ ! 0. In this limit the hyperbolic sines become infinite and TE<U0 ! 0
while RE<U0 ! 1.
It is relatively easy to obtain TE>U0 and RE>U0 using the same equations
as above with appropriate modifications. In essence, the hyperbolic functions
become conventional circular trigonometric functions and variations in these
coefficients occur as functions of the de Broglie wavelengths relation to the
barrier width.
4
hOx .t/i D cos !t
5˛
12. A beam of particles each of mass m and kinetic energy E > 0 is incident from
the left on a potential well, the parameters of which are
U .x/ D 0 x<0
D U0 0<x<L
D0 x>L (5.93)
where U0 is a real positive number.

(a) Sketch the de Broglie wavelength of the particles in each region of the
potential well.
(b) Find the transmission and reflection coefficients, TE>0 and RE>0 . [Hint: Use
the results of Problem 11.]
(c) Find the values of E for which TE>0 1 and show that for these kinetic
energies RE>0 0.
(d) Show that the values of E for which TE>0 1 are the same as the quantized
energies of an L-box (infinitely deep).
Solution
(a) The de Broglie wavelength will be shorter inside the well, region II, than
in region I because the kinetic energy of the particles in region II is greater
than E. The wavelength in region III will be the same as that in region I,
but the amplitude will be smaller because of reflection (see Fig. 5.3).
Fig. 5.3
Problem 12—solution U(x)
Y(x)
E
I I I I I I
0 L
x
U0
(b) The potential well of this problem can be constructed by inverting the
barrier in Problem 11, i.e. U0 ! U0 . Thus, the transmission and reflection
coefficients of Problem 11 are identical with those of this problem with the
substitution U0 ! U0 in Eqs. (5.91) and (5.92). Making this substitution
we have
4E .E C U0 /
TE>0 D (5.94)
2 2 L
p
4E .E C U0 / C U0 sin 2m .E C U0 /
„
and
2 31
6 4E .E C U0 / 7
RE>0 D 6
41 C 7
5 (5.95)
Lp
U02 sin2 2m .E C U0 /
„
(c) For TE>0 to be a maximum the sine in the denominator of Eq. (5.94) must
vanish. To meet this condition
Lp
2m .E C U0 / D n (5.96)
„
where n is an integer. Notice that when TE>0 1 the reflection coefficient
RE>0 0, Eq. (5.95), as it must.
(d) The kinetic energy inside the box is E C U0 D p2 =2m where p is the
momentum of the particles when over the box. Solving Eq. (5.96) for the
kinetic energies that maximize the transmission we have
n2 2 „2
p2 =2m D (5.97)
2mL2
The de Broglie wavelength of the particles when over the well is
2„ 2L
D D (5.98)
p n
Thus, the de Broglie wavelengths associated with TE>0 D 1 are 2=n
times L, the width of the well. This is precisely the condition required to
quantize the energies of an L-box (see Appendix M), which can be deduced
by fitting half de Broglie waves in the box of width L (see Problem 1 of
Chap. 6). This quantum mechanical transmission behavior is in contrast to
that of an analogous classical system. Classically, the transmission across
the potential well is unity for any E > 0, although the particle does speed
up when it is over the well.
13. A monoenergetic beam of particles moving from x ! Cx, each of mass m, is

incident on a potential barrier at x D 0. This potential energy U .x/ is
U .x/ D Vı .x/ where U0 > 0 (5.99)
The constant V has units . J m/ because the units of ı .x/ are m1 (see
Appendix J). Find the transmission and reflection coefficients, Tı and Rı . [Hint:
Integrate the TISE across the discontinuity in U .x/ at x D 0 from ! C

and then take the limit as ! 0.]
Solution
In general, the wave function is
I .x/ D eikx C Beikx x < 0
II .x/ D Ceikx C Deikx x > 0 (5.100)
where
k2 D 2mE=„2 (5.101)
The terms eikx represent plane waves travelling in the Cx direction and the
terms eikx are plane waves travelling in the x direction. While it is possible
that B ¤ 0 in region I (reflection), we must have D D 0 in region II because
only transmitted waves, necessarily travelling in the Cx direction, are present
there. Also, in Eq. (5.100) we have set the coefficient of eikx in region I to unity
so the transmission coefficient T and the reflection coefficient R are
Tı D jCj2
Rı D jBj2 (5.102)
Because the wave function must be continuous even if the derivative of the
wave function is discontinuous (as it is in the case of the ı-function potential),
we must have
I .0/ D II .0/ H) 1 C B D C (5.103)
The TISE for this potential is
d2 .x/
ı .x/ .x/ D k2 .x/ for all x (5.104)
dx2
where
D 2mV=„2 (5.105)
We can find the magnitude of the discontinuity in the derivative of .x/ at

x D 0 by integrating the TISE from to C (across the ı-function) and taking
the limits as ! 0 from the left and from the right. We have
Z Z Z
d2 .x/ 2
dx ı .x/ .x/ dx D k .x/ dx (5.106)
dx2
or
ˇ ˇ
d .x/ ˇˇ d .x/ ˇˇ
lim lim .0/ D 0 (5.107)
!0 dx ˇxD !0 dx ˇxD
where the right-hand side vanishes because lim .x/ D lim .x/. Substi-
x!0 x!0C
tuting the wave function and taking the limit, we have
ik .1 B C/ C D 0 (5.108)
Together with Eq. (5.103) we have two simultaneous equations and two
unknowns. After eliminating C using Eq. (5.103) and replacing k and with
the system parameters that constitute them we have

2E„2
Rı D jBj2 D 1= 1 C
mV 2
mV 2
D (5.109)
2„2 E C mV 2
Eliminating B and solving for C we obtain

mV 2
Tı D jCj2 D 1= 1 C 2
2„ E
2„2 E
D (5.110)
2„2 E C mV 2
It is comforting to note that the sum R C T 1 as it should. Our comfort
should not lead to complacence because the unusual units of V suggest that
we check the units of Tı and Rı , both of which should be unitless. Since
Eqs. (5.109) and (5.110) each contains the same factors we need only check
the units of mV 2 =„2 E. We have
mV 2 kg . J m/2 kg m2 kg m2
2
D 2
D 2 D 2
D1 (5.111)
„ E . J s/ J s J s2 kgs2m
This is not conclusive proof that our answer is correct, but at least we know
that the units are correct, which is encouraging.
14. Show that the transmission coefficient TE<U0 for the rectangular barrier of
Problem 11 of this chapter, Eq. (5.91), reduces to the transmission coefficient Tı
for the ı-function barrier of Problem 13 of this chapter in the limits U0 ! 1,
L ! 0 with U0 L D constant.
Solution
The quantity U0 L has the same units as V (Problem 13) so these quantities
are comparable in this limit. We apply the limits given in the statement of the
problem to the terms in Eq. (5.91) to effect the conversion. In the limit U0 ! 1
the argument of the hyperbolic sine is
s
L E Lp
2mU0 1 ! 2mU0 (5.112)
„ U0 „
The leading term in the sinh x is x (see Appendix H.1) so

p
2 L L2
sinh 2mU0 ' 2 2mU0 (5.113)
„ „
In these limits Eq. (5.91) becomes

1
TE<U0 D
U0 L2
1C 2mU0
4E„2
2„2 E
D (5.114)
2„2 E C m .U0 L/2
Tı is given in Eq. (5.110). It is
2„2 E
Tı D (5.115)
2„2 E C mV 2
which is identical to Eq. (5.114).
15. A semi-infinite potential barrier has a beam of particles each of mass m and
kinetic energy E incident on it from the left as shown in Fig. 5.4 for E < U0 .
Fig. 5.4 Problem 15 I II

U0
0 x
(a) Use the results of Problem 11 of this chapter to find the reflection and
transmission coefficients, R and T for E < U0 .
(b) Find R and T for E > U0 by matching the wave function and its derivatives
at the step as was done using Eq. (5.82).
Solution
(a) This is a special case of the barrier of width L with L ! 1. For this
part of the problem, E < U0 , it is a simple matter to take the limit in
Eqs. (5.91) and (5.92) as L ! 1 because lim sinh2 x D 1. Thus, from
x!1
Eqs. (5.91) and (5.92) we see that the reflection coefficient is unity while
the transmission coefficient vanishes. This is sensible because, no matter
how far the particle penetrates the semi-infinite barrier it must eventually
be squirted back into region I as indicated by the sketched wave function in
Fig. 5.4.
(b) The case for E > U0 presents a bit of a mathematical dilemma. We might
consider converting Eqs. (5.91) and (5.92) which pertain to Ep < U0 to
the case for E > U0 . This would, however, make the factor .U0 E/
imaginary thus converting the hyperbolic sines to circular sines because
[see Eq. (K.14)]
sinh .ix/ D i sin x (5.116)
This makes it difficult to take the limit as L ! 1. In addition, we must

consider the fact that the particle velocities in the two regions of space are
different. Recall that in regions I and III in the case of the finite barrier
the particle velocities are the same. Using the same notation as that in
Problem 11 we write the wave function for this step as
I .x/ D Aeikx C Beikx x < 0

0
II .x/ D Feik x 0<x (5.117)
Because the wave

p numbers are realpin both regions of space we designate
them by k D 2mE=„2 and k0 D 2m .E U0 / =„2 for regions I and II,
respectively. The probability currents are then
„k 2
jI .x/ D jAj jBj2
m
„k0
jII .x/ D jFj2 (5.118)
m
and, as before,
jBj2
RD (5.119)
jAj2
In this case, however, T must be the ratio of the intensity of the transmitted
probability current to the incident probability current, which is given by
„k0
jFj2
TD m (5.120)
„k 2
jAj
m
Solving for the coefficients A, B, and F yields R and T and, including the
results for E < U0 , we have
p 2
1 1 U0 =E
RD p 2 E > U0
1 C 1 U0 =E
D1 E < U0 (5.121)
and
p
4 1 U0 =E
TD p 2 E > U0
1 C 1 U0 =E
D0 E < U0 (5.122)
From Eqs. (5.121) and (5.122) it can be seen that there is complete reflec-
tion and zero transmission when E D U0 , but that the reflection decreases
monotonically for higher values of the incident energy. As expected, the
transmission approaches unity for high values of the incident energy.
Figure 5.5 shows graphs of the reflection and transmission coefficients. The
feature of this problem that is not present in Problem 11 is the necessity to
account for the different particle momenta in the two regions of space.
T
T or R
0
0 1 2 E/U0

Chapter 6
Bound States in One Dimension
6.1 Degeneracy
Before beginning with the problems it is worthwhile to point out some features
of one-dimensional bound states. First, there is no complication arising from
degenerate states. Degeneracy exits when two or more states have the same energy.
Although a very important consideration in two- and three-dimensional problems,
there is no degeneracy in one-dimensional problems. We can prove this by assuming
that the assertion is not true. We assume that 1 .x/ and 2 .x/ are linearly
independent eigenfuntions that have the same eigenvalue E and write the TISE (time
independent Schrödinger equation) in the form
1 d2 .x/ 2m
D 2 ŒE U .x/ (6.1)
.x/ dx2 „
Now, the right-hand side of Eq. (6.1) will be the same for both 1 .x/ and 2 .x/.
Therefore
1 d2 1 .x/ 1 d2 2 .x/
D
1 .x/ dx2 2 .x/ dx2
or

d d 1 .x/ d d 2 .x/
2 .x/ D 1 .x/
dx dx dx dx
which, upon integration, gives

d 1 .x/ d 2 .x/
2 .x/ 1 .x/ DC
dx dx

150 6 Bound States in One Dimension
where C is a constant. C may be evaluated by taking the limit of both sides as

x ! 1 because for bound states .1/ D 0. This indicates that C D 0 leaving
1 d 1 .x/ 1 d 2 .x/
D
1 .x/ dx 2 .x/ dx
Another integration gives
ln 1 .x/ D ln 2 .x/ C ln K
where we have written the constant of integration as ln K for convenience. Taking

antilogs we have
1 .x/ D K 2 .x/
which shows that 1 .x/ and 2 .x/ are not linearly independent thus contradicting
the original assumption of two different degenerate states.
6.2 Parity
If the potential has even parity so that U .x/ D U .x/, then the eigenfunctions must
have either even or odd parity. The number of nodes in the eigenfunctions increases
with the energy of the state; the ground state has no nodes and is therefore of even
parity. To see this we write the TISE with the assumption that U .x/ is an even
function. If we now let x ! x we obtain

„2 d 2
C U .x/ n .x/ D E n .x/ (6.2)
2m dx2
where the subscripts emphasize that the s are eigenfunctions. It is obvious that
n .x/ and n .x/ are solutions of the same TISE, and that they have the same
eigenvalue En . Because n .x/ and n .x/ have the same eigenvalue, they can
differ only by a constant. That is,
n .x/ D ˇ n .x/ (6.3)
If we change the sign of x again, we have
n .x/ D ˇ n .x/
D ˇ Œˇ n .x/
2
Dˇ n .x/ (6.4)
Thus ˇ D ˙1 so that if the potential energy is an even function, the eigenfunctions

of the TISE have definite parity, i.e. .x/ D ˙ .x/.
6 Bound States in One Dimension 151
6.3 Characteristics of the Eigenfunctions
A graph of the potential energy function allows one to qualitatively sketch its
eigenfunctions without detailed calculations. To do this it is necessary to have
knowledge of the characteristics of bound state eigenfunctions. These characteristics
include curvature of the eigenfunction, number of nodes, symmetry and limiting
behavior. While we will not derive any of these characteristics we present a
summary in Table 6.1. They are developed in most textbooks on the subject.
6.4 Superposition Principle
A quantum mechanical system need not be in a single eigenstate. In general it is

not! Inasmuch as the eigenfunctions constitute a complete set, any allowable wave
function may be expanded in terms of this complete set of eigenstates which we
will designate by i .x/. From this point of view systems are in a superposition of
(eigen)states. The state of the system, .x/, is thus represented as
X
.x/ D ci i .x/ (6.5)
i
where the sum is over all eigenstates. For .x/ to be normalized we must have
Z

.x/ .x/ dx 1 (6.6)
all space
Table 6.1 Properties of Asymptotic behavior of .x/ lim .x/ D 0

n n
bound state one-dimensional x!˙1
eigenfunctions Continuity of n .x/ All x
d n .x/
Continuity of dx
All x except at infinite
discontinuity in U .x/
Curvature of n .x/ Toward x-axis
(classically allowed region)
Curvature of n .x/ Away from x-axis
(classically forbidden region)
Number of nodes Increases with
energy .n D 0; 1; 2 : : :/
Symmetry Even/odd only for
even potentials
As a result of orthogonality of the i .x/ we have

Z X

.x/ .x/ dx D jci j2 D 1 (6.7)
all space i
Moreover, the probability of measuring a particular eigenvalue is given by
Pi D jci j2 (6.8)
Problems
1. A particle is in one of the bound stationary states of a one-dimensional potential

well. An eigenfunction for this potential well is shown in the graph below
(Fig. 6.1).
Fig. 6.1 Problem 1 yn(x)
(a) If a measurement of the location of the particle in space were made, would
it most likely be at a positive value of x or negative value of x? Explain.
(b) Sketch a possible potential on the same axes as that on which the eigenfunc-
tion is displayed. Point out salient features?
(c) To which of the allowed energies, i.e. ground state, first excited state, etc.,
does this wave function correspond? Explain.
Solution
(a) Positive x because

Z 0 Z 1

.x/ .x/ dx < .x/ .x/ dx (6.9)
1 0
(b) U .x/ is skewed because the wave function .x/ is skewed. This is only a
sketch (Fig. 6.2).
Fig. 6.2
yn(x)
U(x)
(c) The first excited state because there is one node.
2. This problem illustrates some of the characteristics of one-dimensional bound

states and their eigenfunctions.
(a) Obtain the expression for the energy eigenvalues of a particle of mass m
trapped in an infinitely deep box of width L, an L-box (Appendix M), by
fitting de Broglie waves in the box. Do not use Schrödinger’s wave equation.
(b) The graph shown in Fig. 6.3 is that of an energy eigenfunction n .x/ which
is not an eigenfunction for the particle in-a-box in part (a) of this problem.
The solution of part (a) should, however, help you solve this part of the
problem. Note that n .x/ 0 for x 0. Which state of the (unknown)
potential energy function does n .x/ represent? That is, ground state, first
excited state, etc.?
Fig. 6.3 Problem 2

y(x)
-2 2 10 x
(c) On the same graph as the energy eigenfunction of part (b) sketch a potential
energy function U .x/ for which n .x/ could be an eigenfunction. Point
out salient features of n .x/ that lead you to represent the potential energy
function as you sketch it. Pay particular attention to the classically allowed
and forbidden regions and the curvature of n .x/.
Solution
(a) The potential energy for an L-box is
U .x/ D 0 0xL
D1 1<x<0;L<x<1 (6.10)
Because of the infinite walls at x D 0 and x D L that stretch to x D ˙1 the

eigenfunctions must vanish in the regions 1 < x < 0 and L < x < 1.
Thus, the eigenfunctions n .x/ vanish outside the box and the probability of
finding the particle in these regions is zero. (This would not be the case if the
walls were of finite height.) These eigenfunctions and the de Broglie waves
associated with them must vanish at both x D 0 and x D L. The only way
the de Broglie waves can vanish at x D 0 and x D L is if the wavelengths
are integral or half-integral multiples of L so they “fit” into the box. This is
illustrated in Fig. 6.4 which shows the L-box and the first two eigenfunctions
1 .x/ and 2 .x/.
Fig. 6.4 y2(x)

Problem 2a—solution
2L=2l
y1(x)
2L=1l
0 L/2 L x
The situation is identical with standing waves or “resonances” set up by a

vibrating string of length L with both ends fixed. To comply with the criterion
that the wavelength must be integral or half-integral multiples of L we must
have
L h hn
n D 2 D ; n D 1; 2; 3 H) pn D (6.11)
n pn 2L
Because the potential energy inside the box is zero the total energy is

p2n 1 hn 2
En D D
2m 2m 2L
„2 2 n2
D (6.12)
2mL2
which is indeed the L-box energy eigenvalue for state n. Notice the similarity
between this problem and Problem 3 of Chap. 5 in which it was shown that
de Broglie waves of fixed wavelength fit into the quantized Bohr orbits.
(b) Because n .x/ D 0 for x 0 we must have U .x/ D 1 for x 0.
There are five other nodes in the classically allowed region. Therefore, this
eigenfunction represents the fifth excited state.
(c) The graph below points out the salient features of a possible potential energy
curve based on the given wave function (Fig. 6.5).
U(x)
•
classically allowed region for y5(x)
inflection curvature
point away from
curvature toward the axis axis
2 10 x
classically
forbidden
eigenfunction region for y (x)
vanishes for 5
x<0 so
U(x) •
at x=0
Fig. 6.5 Problem 2c—solution
This problem illustrates some important characteristics of one-dimensional

bound state eigenfunctions. We may generalize from the simple standing
wave picture of the particle-in-box of part (a) where the eigenfunctions
have a fixed de Broglie wavelength throughout the box. If we imagine
deforming the side walls of the box to some arbitrary shape, it is clear
that the finite potential energy of the particle in the vicinity of these non-
vertical walls requires the kinetic energy T of the particle to vary as the
potential varies. A thorough understanding of the concepts introduced in this
problem will greatly facilitate the student’s understanding of eigenfunctions
and the significance of fitting de Broglie wave into potential energy curves.
If T D T .x/, then the momentum p D p .x/ and we may regard the de
Broglie wavelength as also being a function of position when the potential
energy is not constant. Thus, fitting de Broglie waves in a potential well of
arbitrary shape is obviously not as simple a task as fitting them in an infinite
square well, but the principle is the same. Additional features appear because
“soft” walls make it possible for the particle to penetrate into the classically
forbidden region. As shown in most elementary textbooks, this penetration

is marked by a change in the second derivative of the eigenfunction at the
boundary between the classically allowed and forbidden regions.
3. A one-dimensional attractive ı-function potential is given by
U .x/ D U0 ı .x/ (6.13)
where U0 is a positive constant.

(a) Find the bound state energy E. Show that there is only one eigenvalue no
matter what the strength of the well U0 . Find the unnormalized eigenfunc-
tion.
(b) Normalize the eigenfunction.
Solution
(a) To find the bound state energy (or energies) we first examine the TISE.
„2 d2 .x/
U0 ı .x/ .x/ D E .x/ for all x (6.14)
2m dx2
Integrating Eq. (6.14) across x D 0 (from to ) we have
Z Z Z
d2 .x/
dx C ı .x/ .x/ dx D 2 .x/ dx (6.15)
dx2
where
2mU0
D (6.16)
„2
Taking the limit of all factors in Eq. (6.15) as ! 0 we have
ˇ ˇ
d .x/ ˇˇ d .x/ ˇˇ
lim lim C .0/ D 0 (6.17)
!0 dx ˇxD !0 dx ˇxD
where the right-hand side vanishes because the wave function must be a
continuous function. That is, lim .x/ D lim .x/.
x!0 x!0C
From Eq. (6.14) it is seen that in the region for which x ¤ 0 the term
containing the delta function vanishes. The solution of the resulting equation
for negative energies is
.x/ D Aex C Bex (6.18)

where
2m jEj
2 D (6.19)
„2
Because the eigenfunction must be exponentially decaying on both sides of
x D 0, the wave function of the bound state must be of the form
.x/ D Aex x > 0

D Aex x < 0 (6.20)
where A D B because of continuity of the wave function.

Inserting the eigenfunction given in Eq. (6.20) into Eq. (6.17), we have
lim .Ae / lim .Ae / C A D 0 (6.21)

!0 !0
or

D (6.22)
2
so that
2m jEj m2 U02
2 D D (6.23)
„2 „4
from which we obtain the eigenvalue
mU02
ED (6.24)
2„2
We see that there are no other eigenvalues because there are no other
solutions for E. That is, there are no quantum numbers that arise naturally
from the solution as appeared in the cases of, for example, the square well or
the harmonic oscillator (SHO). Thus, there is one and only one bound state
for a ı-function potential well no matter what the strength of the well U0 .
The key to working this problem was understanding and using the
characteristics of the one-dimensional bound state eigenfunctions as sum-
marized in Table 6.1. In particular, knowledge that the eigenfunction must
be continuous, even at an infinite discontinuity in the potential energy (the
ı-function at x D 0), was crucial [see Eq. (6.17)].
(b) To normalize the eigenfunction, we find A as follows.
Z 0 Z 1
jAj2 e2x dx C jAj2 e2x dx D 1 (6.25)
1 0

2 1 2x ˇˇ0 1 ˇ
2x ˇ1
jAj e 1 C e 0
D1
2 2

2 1 1
jAj C D1
2 2
Therefore
jAj2 D (6.26)
s
2m mU02
D
„2 2„2
mU0
D
„2
and the normalized eigenfunction is
r
mU0 x
.x/ D e x>0 (6.27)
„2
r
mU0 x
D e x<0
„2
4. Shown in Fig. 6.6 are two potential wells U1 .x/ ŒD U1 .x/ and U2 .x/ that are
drawn on the same horizontal and vertical scales. U2 .x/ is the same as U1 .x/
for x > 0. The left-hand wall of U2 .x/ is, however, impenetrable. Sketch the
ground state and first excited state eigenfunctions on the well on the left at their
approximate energies. Sketch the ground state eigenfunction at the appropriate
energy on the well on the right. Keep the same energy scale for both wells.
Comment on the following features of all three eigenfunctions: parity, number
of nodes, curvature in different regions of x.
Fig. 6.6 Problem 4
Solution
For U1 .x/: The ground state is an even function because U1 .x/ is evidently an
even function. The eigenfunction has no nodes in the classically allowed region
and it has inflection points at the classical turning points as shown in Fig. 6.7.
The first excited state is an odd function with a single node at x D 0. It too has
inflection points at the classical turning points. All eigenfunctions curve toward
the x-axis in the classically allowed region and away from it in the classically
forbidden region.
Fig. 6.7
For U2 .x/: For x 0 the ground state eigenfunction is the right-hand side
of the first excited state eigenfunction of the well on the left. For x < 0 the
eigenfunction is zero because the potential energy is infinite. The energy is that
of the first excited state of the well on the left. This ground state has no definite
parity, no nodes in the classically allowed region, one inflection point at the right
hand classical turning point and a zero at x D 0 at which the derivative of the
eigenfunction is discontinuous. The eigenfunctions are the odd eigenfunctions of
U1 .x/ for x > 0. For x < 0 they vanish. The curvature of the eigenfunctions is
the same as described in the previous paragraph.
5. A particle of mass m is in the ground state of an a-box, i.e. a potential well of

width a for which the potential energy is given by
U .x/ D 0 a=2 x a=2

D1 x > a=2 ; x < a=2 (6.28)
The width of the box is suddenly doubled symmetrically so that the walls of the
box are located at x D a and x D Ca as illustrated in Fig. 6.8.
Fig. 6.8 Problem 5
(a) Calculate the probability that the particle will be found in the ground state of
the new box.
(b) Calculate the probability that the particle will be found in the first excited
state of the new box.
Solution
This is a classic problem the solution to which should be understood thoroughly.

It is fundamental to an understanding of eigenstates (stationary states), non-
stationary states, and the superposition theorem.
(a) First we must interpret the word “suddenly” in this context. It means that
the transition from the original box to the new box occurs in such a short
time that the wave function cannot adjust to this change. Thus, the wave
function in the new box is the same as that in the original box, the “sudden
approximation.” We are told, however, that the wave function in the original
a-box represents the ground state in that box. While the wave function was
an eigenfunction of the old box, it is not an eigenfunction of the new box.
The transition between boxes is illustrated in Fig. 6.8. In keeping with the
superposition theorem we can expand this wave function in terms of the
complete set of eigenfunctions of the new box. The probability of finding
the particle in any state of the new box will be the square of the expansion
coefficient of the corresponding eigenfunction.
The eigenfunctions for the original a-box with the potential energy is
given by Eq. (6.28) are
r nx
2 a a
n .x/ D cos x n odd (even parity)
a a 2 2
r nx
2 a a
n .x/ D sin x n even (odd parity)
a a 2 2
a a
D0 x < ; x > for all n (6.29)
2 2
so the wave function (not eigenfunction) of the new box (which is the ground
state eigenfunction of the old box) is
r x
2 a a
new
.x/ D cos x
a a 2 2
a a
new
.x/ D 0 x and x (6.30)
2 2
Noting that this wave function, Eq. (6.30), is an even function, its expansion
in terms of the eigenfunctions of the new box cannot contain the sine
eigenfunctions. Using Eq. (6.30), this expansion is
r x r 1 X 1
2 jx a a
cos D cj cos ; x (6.31)
a a a j odd 2a 2 2
The probability Pj of finding the particle in the jth eigenstate of the new box
is given by the square of the jth expansion coefficient in Eq. (6.31),
ˇ ˇ2
Pj D ˇcj ˇ (6.32)
We seek
P1 D jc1 j2 (6.33)
To find c1 we multiply both sides of Eq. (6.31) for new

.x/ by the ground
state eigenfunction of the new box, which is
r x
1
new
1 .x/ D cos (6.34)
a 2a
Integrating over all space leads to c1 .

Z a=2
p Z a=2 x x Z 1
2
0dxC cos cos dxC 0dx D c1 (6.35)
1 a a=2 a 2a a=2
To evaluate the integral on the lhs we let

x
yD (6.36)
2a
so the non-zero integral becomes
Z a=2 x x
ID cos cos dx
a=2 a 2a
Z =4
4a
D cos .2y/ cos ydy (6.37)
0
where we have used the fact that the integrand is even and integrated over
symmetric limits.
To perform the last integral we use the trigonometric identity in Eq. (E.6)
1
cos A cos B D Œcos .A B/ C cos .A C B/ (6.38)
2
The integral is
Z =4
4a 1
ID Œcos y C cos 3y dy

0 2
=4
2a 1
D sin y C sin 3y
3 0

2a 1 1 1 8a
D p C p D p (6.39)
2 3 2 3 2
Returning to Eq. (6.35) we have

p
2 8a 8
c1 D p D (6.40)
a 3 2 3
so the probability of finding m in the ground state is
2
2 8
P1 D jc1 j D
3
64
D Ð 0:72 (6.41)
9 2
(b) As noted above, the eigenfunction for the old box is even while that for
the first excited state of the new box is odd. Therefore, the overlap integral
vanishes and the probability of finding the particle in any of the odd states of
the new box is zero.
6. Let us now modify Problem 5. We start with the particle in the ground state of an
infinitely deep L-box (see Appendix M) for which the potential energy function is
U .x/ D 0 0xL
D1 1 < x < 0I L < x < 1 (6.42)
The width of the box is doubled by suddenly moving only the right-hand
wall. Figure 6.9 shows the wave functions before and after the box is suddenly
expanded. Use the sudden approximation to find the probability of observing the
particle in the ground state of the new box.
Fig. 6.9 Problem 6
Solution
In contrast to the situation in Problem 5, the spatial symmetry about x D L=2

that existed originally was destroyed by the asymmetric way in which the box
was expanded. Nonetheless, the calculation is relatively straightforward. Again
we expand the old wave function, which is not an eigenfunction of the new box,
in a series of eigenfunctions of the new box. Thus, omitting the regions for which
the functions are zero
r x
2
new
.x/ D sin
L L
r x r 1
2 1X jx
sin D cj ; 0 x 2L (6.43)
L L L j 2L
We seek P1 D jc1 j2 so we multiply both sides of Eq. (6.43) by the ground state
eigenfunction of the new box and obtain
Z "r # "r #
L
2 x 1 x
c1 D sin sin dx (6.44)
0 L L L 2L
Note that the upper limit of integration is L because old .x/ D 0 for x > L. To
perform the last integral integral we use the trigonometric identity in Eq. (E.5)
1
sin A sin B D Œcos .A B/ cos .A C B/ (6.45)
2
so that
Z x
1 L
3x
c1 D p cos cos dx
2L 0 2L 2L
x 2L L
1 2L 3x
D p sin sin
2L 2L 3 2L 0
p
1 2L 2L 4 2
D p .1/ D (6.46)
2L 3 3
and
p !2
2 4 2
P1 D jc1 j D Ð 0:36 (6.47)
3
Notice that this result is different from that of part (a) of Problem 5. This is
because in this problem there are both even and odd terms (with respect to x D 0)
in the wave function expansion while the symmetry of Problem 5 demands that
only the even terms are present.
7. A particle is trapped in an a-box potential
U .x/ D 0 a=2 x a=2

D1 1 < x < a=2 ; a=2 < x < 1 (6.48)
The state of the particle is described by the wave function

( "r # "r #)
2 x 1 2 3x
.x/ D A cos cos ; a=2 x a=2
a a m a a
D0 1 < x < a=2 ; a=2 < x < 1 (6.49)
where m is a positive constant. A graph of Eq. (6.49) with m D 3:5 is shown in

Fig. 6.10 together with graphs of each of the constituent cosine terms.
Fig. 6.10 Problem 7
cos(px/a)
y(x)
x
–a/2 a/2
cos(3px/a)
(a) Find the value of A that normalizes .x/ in Eq. (6.49) for arbitrary m.
(b) Find the probabilities that a measurement of the energy will yield E1 and E3 ,
the energy eigenvalues of the ground and first excited states of the a-box.
Note that there are no states with even quantum numbers because .x/ is an
even function of x (see Fig. 6.10).
Solution
(a) To normalize .x/ we must force the integral of .x/ .x/ over all space
to equal unity.
Z a=2

.x/ .x/ dx D 1 (6.50)
a=2
We note that the two terms in square brackets in Eq. (6.49) are normal-
ized eigenfunctions of the a-box potential, 1 .x/ and 3 .x/. Therefore,

.x/ .x/ is the sum of three terms two of which are the squared
eigenfunctions. The integral of the cross term is manifestly zero because of
orthogonality so we have
Z a=2
1
.x/ .x/ dx D A2 1 C 2 (6.51)
a=2 m
so

m2
A2 D (6.52)
1 C m2
Thus, the wave function on the interval a=2 x a=2 is

r
m 2 x 1 3x
.x/ D p cos cos (6.53)
1 C m2 a a m a
(b) Because .x/ is a normalized linear combination of the two even eigenstates
of the a-box the probabilities of measuring their eigenvalues, E1 and E3 ,
are simply the squares of their expansion coefficients in Eq. (6.53). The
individual probabilities Pi are therefore

m2 1
P1 D and P3 D (6.54)
1 C m2 1 C m2
It is clear that as m increases, the probability of measuring E1 increases and

that of measuring E3 decreases as illustrated in Fig. 6.11.
Fig. 6.11
Problem 7—solution 1
P1
Pi
P3
0
m
0 2 4 6
For this particular wave function the only two eigenstates represented
in .x/ are those with eigenvalues E1 and E3 so that these are the only
measurable values of a measurement of energy.
8. A particle is subjected to an a-box potential given by
U .x/ D 0 a=2 x a=2

a a
D1 x< ;x> (6.55)
2 2
The particle is in a state such that the wave function is
x
.x/ D A cos2 a=2 x a=2
a
a a
D0 x< ;x> (6.56)
2 2
A graph of Eq. (6.56) is shown in Fig. 6.12.
Fig. 6.12 Problem 8

(a) Find the value of A that normalizes .x/ in Eq. (6.56).

(b) Find the probability that a measurement of the energy will yield E1 , the
eigenvalue of the first eigenstate.
Solution
(a) To normalize .x/ we must force the integral of .x/ .x/ over all space
to equal unity.
Z a=2

.x/ .x/ dx D 1 (6.57)
a=2
so that, using Eq. (G.12) we have

Z a=2 x
A2 cos4 dx D 1
a=2 a
3a
A2 D1 (6.58)
8
so
8
A2 D (6.59)
3a
and therefore
r x
8
.x/ D cos2 a=2 x a=2
3a a
a a
D0 x< ;x> (6.60)
2 2
(b) Now for the physics. To exploit of the superposition theorem we must expand
the wave function given by Eq. (6.60) in terms of the complete set of a-box
eigenfunctions n .x/.
1
X
.x/ D Cn n .x/ a=2 x a=2
nD1
a a
D0 x< ;x> (6.61)
2 2
The square of the nth expansion coefficient Cn is then the probability of

measuring the energy eigenvalue of the nth state. The eigenfunctions are
r nx
2
n .x/ D cos ; a=2 x a=2 ; n odd (even)
a a
r nx
2
n .x/ D sin ; a=2 x a=2 ; n even (odd)
a a
D0 x < a=2 and x > a=2 all n (6.62)
The wave function, Eq. (6.60), is an even function so there can be no sines
in the expansion. Thus, the solution will contain only odd values of n so the
series representation of .x/ is
r "r #
8 x X 1
2 nx
cos2 D Cn cos ; a=2 x a=2
3a a odd n
a a
q (6.63)
2
x
Multiplying both sides of Eq. (6.63) by .x/ D
1 cos a and integrat-
a
ing over the interval a=2 x a=2 we have, using Eq. (G.11)
r Z
16 a=2 3 x
C1 D cos dx
3a2 a=2 a
r " #a=2 r
16 a sin .x=a/ a sin3 .x=a/ 2 16 2
D2 D
3a2 3 3a2 3
0
16
D p (6.64)
3 3
The probability of finding the system in the ground state and therefore
measuring the energy to be E1 is
256
jC1 j2 D
27 2
0:961 (6.65)
The major component of the given wave function .x/ is the ground state of
the a-box. This is illustrated in Fig. 6.13 which shows that the original wave
function and the ground state eigenfunction 1 .x/ are very similar.
Fig. 6.13
Problem 8—solution. The y(x)
heavy line is the given wave
function and the lighter line is
the ground state
eigenfunction y1(x)
–a/2 0 a/2
9. A particle of mass m is initially in the ground state of a parabolic potential

given by U1 .x/ D .2k/x2 . The force constant is suddenly halved so that the new
potential becomes U2 .x/ D kx2 .
(a) Use the sudden approximation to find the probability that the particle is in
the ground state of the new potential U2 .x/.
(b) What is the probability that the particle will be in the first excited state of the
new potential?
Solution
(a) The normalized ground state eigenfunction for the parabolic potential (SHO,
see Appendix O) has the form
p
˛ 2 x2 =2
0 .x/ D e˛ (6.66)
1=4
where for U1 .x/

v " #
r u r
m! u m .2k/1=2 1
˛D D t 1=4
D m .2k/ 1=4
(6.67)
„ „ m 1=2 „
The ground state eigenfunction for U2 .x/ is

p
0 ˇ ˇ2 x2 =2
0 .x/ D e (6.68)
1=4
where
r
1
ˇ D m1=4 .k/1=4 D 21=4 ˛ (6.69)
„
After the potential suddenly changes from U1 .x/ to U2 .x/ the wave function
remains that in Eq. (6.66), but it is not an eigenfunction for the new potential.
As in Problem 5 we can expand the original wave function in Eq. (6.66) in
terms of the eigenfunctions for the new potential. We use the bra/ket notation
commonly used for the harmonic oscillator to make our equations more
concise. In terms of kets we may write
X ˇ ˛
j0i D ci ˇn0 (6.70)
i
where j0i corresponds to the eigenfunction in Eq. (6.66) and the jn0 i are the
eigenkets of the harmonic oscillator in the potential U2 .x/.
We require only c0 because jc0 j2 is the desired probability so we take the
inner product of both sides with h00 j. Because the jn0 i are orthonormal we
have
˝ 0
0 j0i D c0
Z 1 p p
ˇ ˇ2 x2 =2 ˛ ˛2 x2 =2
D 1=4
e e dx
1 1=4
r Z
˛ˇ 1 .˛2 Cˇ2 /x2 =2
D e dx (6.71)
1
This integral is given in Eq. (G.3).

Z 1 r
2 y2
e dy D (6.72)
1 2

In this case 2 D ˛ 2 C ˇ 2 =2 so
r p s
˛ˇ 2˛ˇ
c0 D v D ˛2 C ˇ2
uu ˛2 C ˇ2
t
2
s 1=4 s
2˛ 2 ˛ 23=4
D p D p (6.73)
2 2
˛ C˛ = 2 1 C 1= 2
Squaring, we have the probability P that the system will be found in the
ground state on the new oscillator.
23=4
P D jc0 j2 D p Ð 0:98 (6.74)
1 C 1= 2
Thus, the overlap between the ground-state eigenfunctions of the two

parabolic potential wells is nearly unity as may be seen in Fig. 6.14.
y*(x)y(x)
(b) As in Problem 7 parity considerations make the answer to this part very
simple. It is impossible for there to be any odd contribution to an even wave
function so the answer is zero.
Chapter 7
Ladder Operators for the Harmonic Oscillator
The ladder operator method of solving the harmonic oscillator problem is not only
elegant, but extremely useful. In fact, the general method transcends the harmonic
oscillator inasmuch as there are other systems for which ladder operators exist, most
notably angular momentum. We devote an entire section to these operators because
of their importance for solving problems and for understanding how to use ladder
operators (also known as raising and lowering operators).
Because these operators simplify many problems it is usually advisable to
employ them whenever possible. We state this as a general rule:
• Always attempt to solve harmonic oscillator problems using ladder opera-
tors before embarking on arduous calculations.
Although the definition of harmonic oscillator ladder operators is not universal
the most commonly used definition is
r
m! i 1 1
aO D xO C pO D p ˛Ox C i pO (7.1)
2„ m! 2 ˛„
and its complex conjugate

r
m! i 1 1
aO D xO pO D p ˛Ox i pO (7.2)
2„ m! 2 ˛„
where
r
m!
˛D (7.3)
„

174 7 Ladder Operators for the Harmonic Oscillator
As will be shown in Problem 2
aO D lowering operator (7.4)

aO D raising operator (7.5)
A number of relations involving the harmonic oscillator oscillator ladder operators

are summarized in Table O.2.
Problems
1. In preparation for Problem 2 of this chapter you are asked to verify a fundamental
commutation relation, that between the position operator xO and the momentum
operator pO . Commutation relations are discussed in Appendix N. In one dimen-
sion the xO , pO commutation relation is
ŒOx; pO D xO pO pO xO D i„ (7.6)
To verify this relation operate on an arbitrary wave function .x/ with this
commutator. It will be helpful to recall that the momentum operator equivalent
of pO is
„ d
pO D (7.7)
i dx
Also, verify Eq. (7.6) for an arbitrary wave function .p/ using operator
equivalent of xO in momentum space
„ d
xO D (7.8)
i dp
Solution
Using the wave function .x/ demands that we replace the momentum operator
pO with it position space differential operator so we write
ŒOx; pO .x/ D .OxpO pO xO / .x/

„ d .x/ „ d Œx .x/
Dx
i dx i dx
„ d .x/ „ d .x/ „
Dx x .x/
i dx i dx i
D i„ .x/ (7.9)
7 Ladder Operators for the Harmonic Oscillator 175
Analogously we write
ŒOx; pO .p/ D .OxpO pO xO / .p/

„ d Œp .p/ „ d .p/
D Cp
i dp i dp
„ d .p/ „ „ d .p/
D p .p/ C p
i dp i i dp
D i„ .p/ (7.10)
2. This problem demonstrates some of the essential features of the ladder operators.

(a) Show that aO ; aO D 1.

(b) Show that H O D „! aO aO C 1 .
2 h i
(c) Define NO D aO aO and show that N; O H
O D 0 and therefore NO and H O have
simultaneous h eigenkets.
i h i
(d) Show that N; O aO D Oa and N; O aO D aO .
(e) Show that, if jni is an eigenstate of NO with eigenvalue n, i.e. aOp
aO jni D n jni
then O
a
jni and O
a jni are also eigenstates of O with eigenvalues .n C 1/ and
N
p
.n 1/, respectively. Moreover,
p
aO jni D n jn 1i (7.11)
and
p
aO jni D n C 1 jn C 1i (7.12)
justifying the designations of aO and aO as raising and lowering operators.

(f) Show that

O jni D „! n C 1 jni
H (7.13)
2

so that the energy eigenvalues are given by En D n C 12 „!.
(g) Construct the ground state eigenfunction in coordinate space.
Solution
(a) Using the definitions of the ladder operators, Eqs. (7.1) and (7.2)

1 1 1
aO ; aO D ˛Ox C ipO ; ˛Ox i pO
2 ˛„ ˛„

1 i i
D ŒOx; pO C ŒOp; xO
2 „ „
i
D f2 ŒOx; pO g
2„
D1 (7.14)
(b) Solving Eqs. (7.1) and (7.2) for xO and pO we have

r
„ 1
xO D aO C aO D p aO C aO (7.15)
2m! 2˛
and
r
m!„ ˛„
pO D i aO aO D i p aO aO (7.16)
2 2
Inserting these in the Hamiltonian we have
2
O D pO C 1 m! 2 xO 2
H
2m 2
„! 2 „! 2
D aO aO C aO C aO
4 4
„! ˚
D aO aO aO aO aO aO C aO aO C aO aO C aO aO C aO aO C aO aO
4
„!
D aO aO C aO aO (7.17)
2
From Eq. (7.14) we have
aO aO D 1 C aO aO (7.18)
so

O D „! aO aO C 1 C aO aO
H
2

1
D „! aO aO C (7.19)
2
(c) Using part (b) we have

h i h i
O H
N; O D „! aO aO ; aO aO C 1
O D aO aO ; H 0 (7.20)
2
(d) Using Eq. (N.3) we have

h i

O aO D aO aO ; aO D aO ; aO aO C aO ŒOa; aO D Oa
N; (7.21)
and
h i

O aO D aO aO ; aO D aO ; aO aO C aO aO ; aO D aO
N; (7.22)
(e) Operate on the vector aO jni with NO and use the commutation relation
Eq. (7.21) to obtain

NO aO jni D Oa C aO NO jni
NO fOa jnig D .Oa C aO n/ jni

D .n 1/ fOa jnig (7.23)
Similarly, operate on aO jni with NO and use the commutation relation

Eq. (7.22) to obtain

NO aO jni D aO C aO NO jni
˚
NO aO jni D aO C aO n jni
˚
D .n C 1/ aO jni (7.24)
Equations (7.23) and (7.24) show that
aO jni D c1 jn 1i (7.25)
and
aO jni D c2 jn C 1i (7.26)
Now we must find c1 and c2 . Because the eigenvectors must be normalized

we must have
1 D hn 1 jn 1i
1
D hnj aO aO jni
jc1 j2
1
D hnj NO jni
jc1 j2
n
D (7.27)
jc1 j2
p
Choose c1 to be real so c1 D n and
p
aO jni D n jn 1i (7.28)
Similarly
p
(f) From parts 2(b) and 2(e) of this problem we know that

1
O
H D „! aO aO C
and NO D aO aO (7.30)
2
so

O O 1
H D „! N C (7.31)
2
Also
aO aO jni D NO jni D n jni (7.32)
Therefore

O O 1
H jni D „! N C jni
2

1
„! n C jni (7.33)
2
so, clearly

1
En D n C „! (7.34)
2
(g) Use the fact that the lowering operator lowers the ground state eigenket to
zero. Therefore
aO j 0 .x/i D 0 (7.35)
In coordinate space this becomes

r
m! „ d
xC 0 .x/ D 0 (7.36)
2„ m! dx
The last equation is a separable first-order differential equation which is

easily solvable. The solution is
2 x2 =2
0 .x/ D Ke˛ (7.37)
where K is a constant of integration, which we obtain by normalization.

Assuming K is real
Z 1
j 0 .x/j2 dx D 1
1
Z 1
2 2 x2
K e˛ D1 (7.38)
1
Using the definite integral given in Eq. (G.3) we have

p 2 1=4
˛
K2 D 1 H) K D (7.39)
˛
so
1=4
˛2 2 x2 =2
0 .x/ D e˛ (7.40)

3. Aspects of this problem, particularly part (c), are identical with Problem 9 of
Chap. 5. The method of solution is, however, different.
The state vector at t D 0 for a one-dimensional harmonic oscillator is given by
1 2
j‰ .0/i D p j1i C p j2i (7.41)
5 5
where jni is an eigenket of the time independent harmonic oscillator Hamilto-
nian. Use the ladder operators

1 1
aO D p ˛Ox C i pO
2 ˛„

1 1
aO D p ˛Ox i pO (7.42)
2 ˛„
and their properties as given in Problem 2 of this chapter for the computations in
part (c).
(a) Find j‰ .t/i, the state vector for t > 0.
(b) Find the expectation value of the energy for t > 0 using Dirac notation.
(c) Find hOx .t/i, the expectation value of the position for t > 0 using ladder
operators.
Solution
(a) Simply multiply each term in the expansion by the time dependent part of
the eigenfunction. Therefore
1 2
j‰ .t/i D p j1i e3i!t=2 C p j2i e5i!t=2 (7.43)
5 5
(b)
O j‰ .t/i
hEi D h‰ .t/j H
1 4
D E1 h1 j1i C E2 h2 j2i
5 5
13 45 23
D „! C „! D „! (7.44)
52 52 10
(c)
hOx .t/i D h‰ .t/j xO j‰ .t/i (7.45)
Solving Eq. (7.42) for xO we have
1
xO D p aO C aO (7.46)
2˛
Inserting xO from Eq. (7.46) into Eq. (7.45) we have

p 1 2
2˛ hOx .t/i D h1j aO C aO j1i C h1j aO C aO j2i ei!t
5 5
2 4
C h2j aO C aO j1i ei!t C h2j aO C aO j2i (7.47)
5 5
The first and last terms of Eq. (7.47) clearly vanish so, using Eqs. (7.28)
and (7.29) on the remaining terms we have
p 2 2
2˛ hOx .t/i D h1j aO C aO j2i ei!t C h2j aO C aO j1i ei!t
5 5
2

D h1j aO j2i C h1j aO j2i ei!t
5
2

C h2j aO j1i C h2j aO j1i ei!t
5
2 p 2p
D 2 h1 j1i ei!t C 2 h2 j2i ei!t
5 5
2 p i!t ˚ i!t
D 2e e C ei!t
5
4p
D 2 cos !t (7.48)
5
So
r
4 4 m!
hOx .t/i D cos !t D cos !t (7.49)
5˛ 5 „
which is the result that was obtained in Problem 9 of Chap. 5, but here
the matrix element h1j xO j2i was evaluated using ladder operators. This was
not only more expeditious, but it also reduced the possibility of error. The
comparison of the solution to this problem with that to Problem 9 of Chap. 5
illustrates the utility of the “rule” stated in the introduction to this chapter:
Always attempt to solve harmonic oscillator problems using ladder
operators before embarking on arduous calculations.
4. The normalized state vector at t D 0 for a particle subject to a one-dimensional

harmonic oscillator potential is
1 2
j‰ .x; 0/i D p j2i C p j4i (7.50)
5 5
where thepjni are eigenvectors of the Hamiltonian and the number operator N.
Let ! D k=m.
(a) Find the state vector as a function of time j‰ .x; t/i.
(b) Find the expectation value of the energy as a function of time. Does the
answer agree with that predicted by the Ehrenfest theorem [1]?
(c) Find the expectation value of the position as a function of time using ladder
operators.
Solution
(a) The time dependent state vector is obtained by multiplying each “compo-
nent” of j‰ .x; 0/i by

1
exp .iEn t=„/ D exp i n C t=„ (7.51)
2
with the appropriate value of n. This gives
1 2
j‰ .x; t/i D p j2i e.5=2/i!t C p j4i e.9=2/i!t (7.52)
5 5
(b)
O j‰ .x; t/i
hEi D h‰ .x; t/j H

1 5 4 9
D „! C „!
5 2 5 2

1 5 C 36 41
D „! D „! (7.53)
5 2 10
The important point here is that hEi is independent of time. This is expected
based on the Ehrenfest theorem which, for an operator AO that does not contain
the time explicitly, is [1]
D E
d AO i h i
D h‰j H; O AO j‰i (7.54)
dt „
D E
Inasmuch as H O commutes with itself, hEi D H O is independent of time.
(c) From Eq. (7.15)
r
„
xO D aO C aO (7.55)
2m!
the expectation value hOx .t/i is
r
„
hOx .t/i D h‰ .x; t/j aO C aO j‰ .x; t/i (7.56)
2m!
We need not, however, write out these inner products. First we note that
the exponential time factors cancel. Second, the only inner products that
will occur are of the form h2j aO j4i, h2j aO j2i and analogous inner products
with the operator aO . All these inner products vanish because the actions of
both aO and aO only change the bra or ket by one [see Eq. (7.55)]. Therefore,
hOx .t/i D 0.
5. The state vector at t D 0 for a particle in a harmonic oscillator potential is

given by
r
1 2
j‰ .x; 0/i D p j1i C j2i (7.57)
3 3
where the jni are eigenvectors of the Hamiltonian and the number operator.
(a) Find the state vector as a function of time j‰ .x; t/i.
(b) Find the expectation value of the energy as a function of time. Does the
answer agree with that predicted by the Ehrenfest theorem?
(c) Find the expectation value of the position as a function of time using ladder
operators. Calculate this expectation value using the Ehrenfest theorem and
compare the results.
Solution
p
(a) With ! D k=m the time dependent state vector is obtained by multiplying
each “component” of the state vector by the exponential factor containing
the energy, that is

1
exp .iEn t=„/ D exp i n C t=„ (7.58)
2
so
r
1 2
j‰ .x; t/i D p j1i e.3=2/i!t C j2i e.5=2/i!t (7.59)
3 3
(b)
O j‰ .x; t/i
hEi D h‰ .x; t/j H

1 3 2 5
D „! C „!
3 2 3 2

13
D „! (7.60)
6
which is independent of time and agrees with the Ehrenfest theorem as in

Problem 4 of this chapter.
(c) In this problem we are not as fortunate as we were in Problem 4 of this
chapter where the basis kets that constituted j‰ .x; t/i were both even so
that all cross (inner) products vanished. Here we must actually do the
computations. We can, however, ignore those cross products with a ladder
operator sandwiched between a bra and ket of equal number because they
vanish. Using Eq. (7.15)
r
„
xO D aO C aO (7.61)
2m!
we have
r
„
hOxi D h‰ .x; t/j aO C aO j‰ .x; t/i
2m!
r p n
„ 2
D h2j e.5=2/i!t aO C aO j1i e.3=2/i!t
2m! 3
o
C h1j e.3=2/i!t aO C aO j2i e.5=2/i!t (7.62)
Again leaving out the terms that vanish due to orthogonality and using
p p
aO jni D n jn 1i and aO jni D n C 1 jn C 1i (7.63)
we have
r p
„ 2
i!t
hOxi D e h2j aO j1i C ei!t h1j aO j2i
2m! 3
r p
„ 2 i!t p p
D e 2 C ei!t 2
2m! 3
r p
„ 2
D 2 cos !t (7.64)
m! 3
Now, let us do the computation
h i using Ehrenfest’s theorem, Eq. (7.54). We

require the commutator H; O xO D 1 pO 2 ; xO pO 2 ; xO which is easily calculated
2m

using the identity xO ; pO nx D i„nOpn1
x (see, e.g., [1]).
h i

O xO D 1 xO ; pO 2 D i„ pO
H; (7.65)
2m m
Therefore, according to the Ehrenfest theorem
d hOxi i h i
D h‰j H;O xO j‰i
dt „

i i„
D h‰j pO j‰i
„ m
1 hOpi
D h‰j pO j‰i D (7.66)
m m
The relationship between d hOxi =dt and hOpi is identical to that for a classical
oscillator, i.e. dx=dt D p=m [Eq. (2.33)]. This classical-quantum correspon-
dence is unique to systems quadratic in p and x.
To evaluate this inner product we must use Eq. (7.16)
r
m!„
pO D i aO aO (7.67)
2
The resulting
r equationrwill be identical with Eq. (7.64) except that the
„ m!„
constant ! i and aO ! Oa . We have
2m! 2
r p
d hOxi m!„ 2 1
i!t
D i e h2j aO j1i C ei!t h1j aO j2i
dt 2 3 m
r p
„! 2 i!t
D i e C ei!t
m 3
r p
„! 2 2 ei!t ei!t
D
m 3 2i
r p
„! 2
D 2 sin !t (7.68)
m 3
Integrating, we have
r p
„! 2 1
hOxi D 2 cos !t
m 3 !
r p
„ 2
D 2 cos !t (7.69)
m! 3
where we have set the constant of integration equal to zero. This result is in
agreement with that obtained using ladder operators, Eq. (7.64).
6. Computation of harmonic oscillator matrix elements using the eigenfunctions

involving Hermite polynomials can be carried out, but it is tedious. Find the
following matrix elements using any method you choose. Ladder operators are
recommended.
(a) Use the h1j xO j2i matrix element evaluated in Problem 3 of this chapter to find
the general expression for hmj xO jni.
(b) Find the general expression for hmj pO jni.
Solution
(a) As shown in Eq. (7.15), the position operator may be written as

r
„ 1
xO D aO C aO D p aO C aO (7.70)
2m! 2˛
so the required matrix element is
1
hmj xO jni D p hmj aO jni C hmj aO jni (7.71)
2˛
We already know the actions of aO and aO on the eigenkets jni of the harmonic
oscillator. They are given by Eqs. (7.28) and (7.29) which are
p
aO jni D n jn 1i
p
Therefore, the first term on the right-hand side of Eq. (7.71) is non-vanishing
only when m D n 1 while the second terms vanishes unless m D n C 1.
This is written compactly in terms of Kronecker deltas as
1 p p
hmj xO jni D p nı m;n1 C n C 1ı m;nC1 (7.73)
2˛
Note the if m D n then Eq. (7.73) gives the expectation value of x for
any oscillator state jni, hxin 0. This is the obvious answer because the
oscillator potential is symmetric about the ordinate (even). Mathematically
this is equivalent to saying that hnj xO jni represents an odd integral over
symmetric limits.
(b) Evaluation of hmj pO jni proceeds in exactly the same manner as that for
hmj xO jni. Using Eqs. (7.16) and (7.72) we have
and
˛„
hmj pO jni D i p hmj aO jni hmj aO jni
2
˛„ p p
D i p nı m;n1 n C 1ı m;nC1 (7.74)
2
If m D n, then Eq. (7.74) gives the expectation value of the momentum,
which is zero for the same reason hmj xO jni is zero.
7. Find the expectation value of pO 2 for the nth eigenstate of DtheE harmonic oscillator.
Find the average value of the kinetic energy of a particle TO in the nth eigenstate
and relate it to the total energy of the nth state. Show that the result is consistent
with the virial theorem, which, for any one-dimensional potential, is
D E dU
2 TO D x (7.75)
dx
Solution
Again using Eqs. (7.16) and (7.72) we have

˛ 2 „2
hnj pO 2 jni D hnj aO aO aO aO jni
2
m! „2
D hnj aO aO C aO aO jni
„ 2
m!„ p p
D hnj aO n C 1 jn C 1i C hnj aO n jn 1i
2
m!„
D .n C 1 C n/
2

1
D m nC „! (7.76)
2
The average value of the kinetic energy is

pO 2
hnj TO jni D hnj jni
2m
1
D En (7.77)
2
The one-dimensional virial theorem states
D E dU
2 TO D x (7.78)
dx
so, for the harmonic oscillator
D E
OT D 1 x d 1 m! 2 x2
2 dx 2
1 ˝ ˛
D m! 2 x2 (7.79)
2
Therefore

˝ 2˛ 1 „
x D nC (7.80)
2 m!
Moreover, the expectation value of the potential energy is the other half of the
total energy, that is,
1 ˝ ˛ 1 ˝ ˛ D E
hUi D k x2 D m! 2 x2 D TO (7.81)
2 2
which is consistent with the virial theorem.
8. Use the matrix element for hmj xO 2 jni, Eq. (7.82) below, to find the expectation
value of xO 4 for arbitrary state jni of the harmonic oscillator.
1 hp
hmj xO 2 jni D n .n 1/ı m;n2
2˛ 2
p i
C .2n C 1/ ı m;n C .n C 1/ .n C 2/ı m;nC2 (7.82)
1
P
[Hint: Use the identity operator jki hkj D IO (See [1]).]
kD0
Solution
1
X
hnj xO 4 jni D hnj x2 jki hkj x2 jni
kD0
1 Xh p
1
D . n .n 1/ı k;n2 C .2n C 1/ ı k;n
4˛ 4 kD0
hp i
C .n C 1/ .n C 2/ı k;nC2 /
p
. n .n 1/ı k;n2 C .2n C 1/ ı k;n
p i
C .n C 1/ .n C 2/ı k;nC2 / (7.83)
When these six terms are multiplied the only survivors will be the squares of
each of the three terms that make up hkj x2 jni. Therefore
1 h i
hnj xO 4 jni D n .n 1/ C .2n C 1/2 C .n C 1/ .n C 2/
4˛ 4
1

D n2 n C 4n2 C 4n C 1 C n2 C 3n C 2
4˛ 4
3
D 2n2 C 2n C 1 (7.84)
4˛ 4
Notice that the units of hnj xO 4 jni in Eq. (7.84) are m4 , as they must be, because
˛ has units m1 (see the eigenfunction in Appendix O).
Chapter 8
Angular Momentum
In quantum mechanics angular momentum includes the usual angular momentum

that we learn about in classical mechanics. This angular momentum is usually
designated by L and defined as
LDrp (8.1)
In quantum mechanics the term “angular momentum” has a much more general
meaning. It is a “generalized angular momentum.” A vector operator JO is defined to
be an angular momentum if its components obey the commutation rules
h i
JOi ; JOj D i„JOk ijk (8.2)
where any of the i, j, and k represent Cartesian coordinates x, y, and z. The quantity
ijk is known as the Levi-Cevita symbol. If the indexes i, j, and k are in cyclic order
(e.g. jki), ijk D C1. If they are out of order (such as kji), then ijk D 1. If any two
indexes are the same, ijk D 0.
It is shown in introductory quantum mechanics textbooks [1] that
h i
JO2 ; JOi D 0 where i D x; y or z (8.3)
ˇ ˇ
ˇ ˇ
Thus, the magnitude of the angular momentum ˇJOˇ can be specified along with any
one of the components of J,O customarily chosen to be JOz . Additionally, JO2 and JOz
have simultaneous eigenkets which we designate by jjmi. The eigenvalue equations
are
JO2 jjmi D j .j C 1/ „2 jjmi and JOz jjmi D m„ jjmi (8.4)
where j and m are the quantum numbers associated with the operators JO2 and JOz .

190 8 Angular Momentum
It is useful to define angular momentum ladder operators JO˙ analogous to the

harmonic oscillator ladder operators
JO˙ D JOx ˙ iJOy (8.5)
These definitions together with the commutation relations given in Eq. (8.2) lead to
the raising and lowering properties of the ladder operators :
p
JOC jjmi D „ j .j C 1/ m .m C 1/ jj .m C 1/i
p
D „ .j m/ .j C m C 1/ jj .m C 1/i (8.6)
and
p
JO jjmi D „ j .j C 1/ m .m 1/ jj .m 1/i
p
D „ .j C m/ .j m C 1/ jj .m 1/i (8.7)
so that JOC and JO are “raising” and “lowering” operators. They raise and lower the
m-values of the eigenkets by unity.
For orbital angular momentum, the angular momentum that we learned about
in elementary classical mechanics, JO2 ! LO2 and JOz ! LOz and the orbital
angular momentum operators can be written in coordinate representation (see
Appendix Q). The eigenfunctions of these operators are the spherical harmonics
Y`m .; / (see Appendix R) where ` .` C 1/ „2 and m„ are the eigenvalues of LO2
and LOz , respectively [see Eq. (8.4)].
Another important angular momentum is “spin.” Although spin is an intrinsically
relativistic concept, it is necessary to include it in nonrelativistic quantum mechan-
ics. The spin angular momentum operator is designated by the symbol SO so that for
spin angular momentum JO2 ! SO2 and JOz ! SOz and the components of SO obey the
commutation rule
h i
SOi ; SOj D i„SOk ijk (8.8)
Unlike orbital angular momentum L there are no spatial coordinates associated with
spin. As will be seen in Problem 2, angular momentum eigenvalues can be only
integral or half-integral multiples of „. The most common spin encountered is that
of a spin- 12 particle, the spin of an electron. For spin- 12 there are only two possible
values of the eigenvalues of SOz , ˙ 12 „, so the eigenvalues of SO2 are 34 „2 [see Eq. (8.4)].
It is convenient to use matrix algebra when dealing with spin- 12 calculations. The
matrices are

3 10 1 1 0
SO2 D „2 ; SOz D „
4 01 2 0 1

1 01 1 0 i
SOx D „ ; SOy D „ (8.9)
2 10 2 i 0
8 Angular Momentum 191
where, as is customary, SOz is chosen to have the simultaneous eigenket with SO2 . Note
that the matrix representations of the
two operators SO2 and SOz are both diagonal. For
simplicity, the Pauli spin matrices
O x ;
O y ;
O z are often used for spin- 12 calculations.
These are the same as the SOi matrices above with the „=2 omitted:

01 0 i 1 0

O x D ;
O y D ;
O z D (8.10)
10 i 0 0 1
The simultaneous eigenkets of SO2 and SOz are designated j˛i and jˇi and have
eigenvalues .3=4/ „2 and ˙„=2, respectively. In matrix notation the eigenkets are

1 0
j˛i D and jˇi D (8.11)
0 1
and are referred to as spinors. The eigenkets of SOx are designated j˛ix and jˇix with
analogous notation for the eigenkets of SOy .
It often happens that there is more than one angular momentum present in
a system. The most common case is when orbital angular momentum and spin
angular momentum exist simultaneously, e.g. an atomic electron. For generality
we designate the two angular momenta by JO1 and JO2 , but there will also be a
total angular momentum JO D JO1 C JO2 . The quantum numbers associated with
these operators are .j; mj /, j1 ; mj1 and j2 ; mj2 , respectively, where we have used
numerical subscripts to denote the particular angular momentum with which a
z-component is associated.

The operators JO21 ; JO1z ; JO22 ; JO2z are mutually commuting [1]. Therefore, the
eigenvalues associated with these operators may be simultaneously obtained. The
simultaneous eigenkets of these four operators are designated
ˇ ˛
ˇj1 ; mj1 I j2 ; mj2 (8.12)
This set of eigenkets is referred to as the “uncoupled set” and the set of commuting
operators that produces them
constitutes the “uncoupled representation.”
O2 O2 O2 O
The set of operators J 1 ; J 2 I J ; J z are also mutually commuting so the set of
quantum numbers associated with them are good (simultaneously obtainable) and
the eigenkets of these four operators are designated
ˇ ˛
ˇj1 ; j2 I j; mj (8.13)
These are the “coupled set” and the operators that produced them constitute the
“coupled representation.” Because each set of eigenkets is a complete set, any
arbitrary ket may be written as a linear combination of eigenkets from one of
the sets. If the arbitrary ket is expanded in terms of the coupled kets, we obtain
information about the probability of measuring the total angular momentum, its
z-component, and the individual angular momenta. No information can be obtained

about the individual z-components though. On the other hand, we could expand the
arbitrary ket in terms of the complete set of uncoupled kets and obtain information
on the individualh z-components,
i hbut none
i on the total angular momentum. This
follows because JO1z ; JO2 ¤ 0 ¤ JO2z ; JO2 (see Problem 9).
Additionally, we can expand an eigenket of one of the sets in terms of the
eigenkets
ˇ of the˛ other complete set. For example, let us expand one of the coupled
kets ˇj1 ; j2 I j; mj on the uncoupled set. Starting with the identity
ˇ ˛ ˇ ˛
ˇj1 ; j2 I j; mj ˇj1 ; j2 I j; mj (8.14)
we can operate on the right-hand side of Eq. (8.14) with the identity operator [1, 2],
rearrange and obtain
0 1
ˇ ˛ X
j1
X
j2
ˇ ˛˝ ˇ
ˇj1 ; j2 I j; mj D @ ˇj1 ; mj1 I j2 ; mj2 j1 ; mj1 I j2 ; mj2 ˇA
mj1 Dj1 mj2 Dj2
ˇ ˛
ˇj1 ; j2 I j; mj
0 1
X j1
Xj2
˝ ˇ ˛
D@ j1 ; mj1 I j2 ; mj2 ˇ j1 ; j2 I j; mj A
mj1 Dj1 mj2 Dj2
ˇ ˛
ˇj1 ; mj1 I j2 ; mj2 (8.15)
The expansion coefficients are the summations in the parentheses in Eq. (8.15). Of
course, we could have expanded an uncoupled ket on the coupled set and obtained
analogous results. These expansion coefficients are known as the Clebsch–Gordan
or Vector Coupling coefficients, some of which are listed in Table S.
Problems
1. Use the ladder operators JO˙ to show the following for a system in an eigenstate
jjmi.
D E D E
(a) JOx D JOy D 0
D E D E 2

(b) JO2x D JO2y D „2 j .j C 1/ m2
Solution
(a) From the equations for JO˙ , Eq. (8.5)
1 O 1 O
JOx D J C C JO and JOy D J C JO (8.16)
2 2i
so that
D E 1
JOx D hjmj JOC C JO jjmi
2
D0 (8.17)
and
D E 1
JOy D hjmj JOC JO jjmi
2i
D0 (8.18)
because of the raising and lowering action of JO˙ . In short, they raise and
lower the bras and kets into orthogonality.
D E D E
The method of showing that JOx D JOy D 0 using ladder operators is
very simple, but there is yet another way to do this by employing the
commutation relations, Eq. (8.2), that define an angular momentum. The
two methods are entwined because the existence of the ladder operators
depends upon the commutation relations,
D E but it is worthwhile to go through
the algebra. We begin by writing JOx in terms of the commutator
D E 1 h i
JOx D hjmj JOy ; JOz jjmi
i„
1 n o
D hjmj JOy JOz jjmi hjmj JOz JOy jjmi
i„
mn o
D hjmj JOy jjmi hjmj JOy jjmi
i„
0 (8.19)
D E
It is clear that the same method will produce JOy D 0.
Let us consider this problem from a physical viewpoint and pretend that
the angular momentum under consideration is orbital angular momentum
(which we can visualize). We have chosen LOz to be the component for which
we can simultaneously know p the eigenvalue m„ and the eigenvalue of the
total angular momentum ` .` C 1/„. This choice precludes knowledge of
the eigenvalues of LOx and LOy : these components can have any value between
j and j with equal probability so their expectation values must vanish.

As we will see in part (b) of this problem though, this is not true of the
expectation values of the squares of these operators.
(b) We write the square of the total angular momentum as the sum of the
squares of its individual components.
JO2 D JO2x C JO2y C JO2z (8.20)
Rearranging and taking the average values we have

D E D E D E D E
JO2x C JO2y D JO2 JO2z (8.21)
D E D E
But, by symmetry JO2x D JO2y because the choice of these axes is arbitrary.
We will, however, prove this assertion below.
Remembering that the system is in an eigenstate jjmi we have
D E D E 1h i
JO2x D JO2y D hjmj JO2 jjmi hjmj JO2z jjmi
2
2

„
D j .j C 1/ m2 (8.22)
2
D E D E
Now prove that JO2x D JO2y . Whenever possible we appeal to the ladder
D E D E
operators and invoke Eq. (8.16). We write JO2x and JO2y as
D E 1 2 2
JO2x D hjmj JOC C JO jjmi
2
D E 1 2 2
JO2y D hjmj JOC JO jjmi (8.23)
2i
The first of these equations becomes

D E 1
JO2x D hjmj JO2C C JOC JO C JO JOC C JO2 jjmi
4

1
D hjmj JOC JO C JO JOC jjmi (8.24)
4
The second is
D E 1
O2
Jy D hjmj JO2C JOC JO JO JOC C JO2 jjmi
4

1
D hjmj JOC JO C JO JOC jjmi (8.25)
4
Equations (8.24) and (8.25) are identical, proving that the expectation
values of the squares of the x- and y- coordinates of JO must be equal.
As long as we have Eqs. (8.24) and (8.25) we may as well evaluate them
(one of them) because it must yield the same answer as the easily obtained
result in Eq. (8.22). Applying Eqs. (8.6) and (8.7) to each of the terms in
Eq. (8.24) we have
p
JO JOC jjmi D „ .j m/ .j C m C 1/JO jj .m C 1/i
p p
D „2 .j m/ .j C m C 1/ .j C m C 1/ .j m/ jjmi
D „2 .j m/ .j C m C 1/ jjmi

D „2 j2 C j m2 m jjmi (8.26)
and
p
JOC JO jjmi D „ .j C m/ .j m C 1/JOC jj .m 1/i
p p
D „2 .j C m/ .j m C 1/ .j m C 1/ .j C m/ jjmi
D „2 .j C m/ .j m C 1/ jjmi

„2 j2 C j m2 C m jjmi (8.27)
Inserting these terms into Eq. (8.24) [or Eq. (8.25)] we have
D E 1
JO2x D hjmj JOC JO C JO JOC jjmi
4
2
„
D 2 j2 C j m2
4
2
„

D j .j C 1/ m2 (8.28)
2
Equation (8.28) is identical with the more easily obtained Eq. (8.22), but
it is comforting to obtain the same result in a seemingly different way,
albeit using more effort. Note that the ease with which Eq. (8.22) was
obtained depended upon using the symmetry of the problem to simplify
our calculation.
2. This problem once again emphasizes the utility of the ladder operators and
shows that angular momentum can take on only integral or half-integral values
of „.
(a) Show the following relations.
h i h i
JOz ; JO˙ D ˙„JO˙ ; JO2 ; JO˙ D 0 (8.29)
and

JO JO˙ D JOx
iJOy JOx ˙ iJOy (8.30)
(b) Use the ladder operators and the above results to show that, in general, the
eigenvalues of JO and JOz have eigenvalues j and m which can only be integers
or half-integers. Moreover, the quantum number j must be positive while m
can range from Cj to j.
Solution
(a) Writing JO˙ in terms of its components and using the fundamental commu-
tation relations that define angular momentum, Eq. (8.5), we have
h i h i h i
JOz ; JO˙ D JOz ; JOx ˙ i JOz ; JOy

D i„JOy ˙ i i„JOx D .„/ ˙iJOx C JOy

D ˙„ iJOx ˙ JOy D ˙„JO˙ (8.31)
and

JO JO˙ D JOx
iJOy JOx ˙ iJOy
D JO2x ˙ iJOx JOy

iJOy JOx C JO2y
h i
D JO2 JO2z
i JOx ; JOy
D JO2 JO2z
„JOz (8.32)
where we have added and subtracted JO2z in the third line of the last equation.
(b) In view of Eqs. (8.6) and (8.7) it is clear that the actions of JOC and JO
are to raise and lower the values of the quantum number m by unity when
operating on the eigenkets jjmi of JO2 and JOz . Let us assume that we have
operated on an eigenket with JO2 . We have

JO2 jjmi D JO2x C JO2y C JO2z jjmi (8.33)
or, from Eq. (8.4), we have

JO2x C JO2y jjmi D j .j C 1/ m2 „2 jjmi (8.34)

Because JO2x C JO2y is manifestly positive (or zero) we must have
j .j C 1/ m2 (8.35)
and there must be a maximum value of m, call it mmax , beyond which the
eigenstate jjmmax i cannot be raised by JOC . That is
JOC jjmmax i D 0 (8.36)
If we now operate on the left of the last equation with JO and apply
Eq. (8.32), we obtain

JO2 JO2z „JOz jjmmax i D 0

j .j C 1/ „2 m2max „2 „mmax jjmmax i D 0 (8.37)
Solving this equation for j .j C 1/ „2 we have
j .j C 1/ „2 D .mmax „/ .mmax „ C 1/ (8.38)
There must also be a minimum value of the quantum number m so
JO jjmmin i D 0 (8.39)
We may now apply JOC to this last equation again using Eq. (8.32) we obtain
j .j C 1/ „2 D .mmin „/ .mmin „ 1/ (8.40)
Comparing Eqs. (8.38) and (8.40) we must have
.mmin „/ .mmin „ 1/ D .mmax „/ .mmax „ C 1/

m2min „ mmin D m2max „ C mmax (8.41)
Thus,
mmax D mmin (8.42)
Equation (8.42) shows that successive values of the quantum number m
must differ by unity. Therefore, m must be either an integer or a half-
integer (try it). Moreover, because j .j C 1/ m2max [Eq. (8.35)] we must
have j 0.
3. The normalized angular part of a certain wave function for a particle is given in
spherical coordinates by
" r #
1 1 p 5 2

.; / D p p 1 C 2 3 cos C 3 3 cos 1 (8.43)
14 4 4
(a) If a measurement is made of the total angular momentum, what are the
possible values that could be measured and with what probabilities for
each?
(b) If a measurement is made of the z-component of the angular momentum,
what are the possible values that could be measured and with what
probabilities for each?
(c) What are theD expectation
E values of the z-component of the orbital angular
O
momentum Lz and of the square of the total orbital angular momentum
D E
LO2 ? The spherical harmonics given in Appendix R should be helpful.
Solution
(a) We must first write . ; / in terms of the spherical harmonics. Inspection

of Table R.1 shows that the first term is proportional to Y00 . ; /, the
p to Y10 . ; /, and the third to Y20 . ; /. Multiplying the factor
second
1= 4 into the square brackets we have
" r r #
1 1 3 5 2

.; / D p p C2 cos C 3 3 cos 1
14 4 4 16
1
D p ŒY00 . ; / C 2Y10 . ; / C 3Y20 . ; / (8.44)
14
Therefore, the only possible values of ` are 0, 1, and 2. The total angular
momentum possibilities are
p p p
` .` C 1/„ D 0; 2„; 6„ (8.45)
The probabilities are 1=14, 4=14, and 9=14, respectively, reflective of the
admixture of the spherical harmonics.
(b) Because m D 0 in each of the spherical harmonics that constitute . ; /
the only possible z-component of the angular momentum for the system
described
D by
E this wave function is zero.
O
(c) Clearly Lz D 0 because of the result of part (b).
D E
To calculate LO2 we must compute the indicated expectation value.
Remembering that the spherical harmonics are orthonormal we have
D E
LO2 D h . ; /j LO2 j . ; /i
„2 2
D 1 0 C 22 2 C 32 6
14
4:6„2 (8.46)
2
D E ` D 2 for which ` .` C 1/ „ D
The admixture is heavily weighted toward
6„2 [see Eq. (8.44)]. Therefore, while LO2 is expected to be less than 6„2 it
is considerably greater than 2„2 , the value for ` D 1.
4. This problem is in the same spirit as Problem 6 of Chap. 5. It is known that the
intrinsic angular momentum, the spin, of an electron is „=2. Assume that the
electron is a uniform solid sphere spinning at a frequency !, with total mass
me and radius re , the classical radius of the electron (see Problem 5 of Chap. 5).
Find the speed of a point on the surface of the sphere vs D !re and show that it
exceeds the speed of light.
Solution
Equating the spin angular momentum, „=2, to the mechanical angular momen-
tum of the spinning sphere we have
„
D I! (8.47)
2
where I D .2=5/ me re2 , the moment of inertia of the solid sphere. In Eq. (8.47)
it is assumed that the spin angular momentum and the mechanical angular
momentum are parallel so we do not concern ourselves with vectors for this
order of magnitude calculation. Substituting for I and ! in Eq. (8.47) we have

„ 2 vs
D me re2 (8.48)
2 5 re
Solving for vs we obtain
5 „
vs D
4 me re

5 „ 4 0 2
D me c
4 me e2

5 4 0 2
D .me / .„/ c (8.49)
4 e2
where we have replaced re with its value deduced in Eq. (5.33) of Chap. 5. In
atomic units (a.u.) all quantities in parentheses in Eq. (8.49) are unity except the
speed of light c which is 137 (see Table C.1). This result is reminiscent of that
obtained in Problem 6 of Chap. 5. We see that in a.u. vs is
5
vs D 1372
4
Ð 170c (8.50)
about one-half the estimate obtained Problem 6 of Chap. 5. It, however,

confirms that the notion of the electron being a spinning sphere of finite
dimension is unrealistic.
5. An electron has total angular momentum quantum number s D 12 .

(a) A measurement is made of the x-component of the spin. What are the
possible results of this measurement?
(b) If the system is in the normalized state represented by the spinor

1 1
ji D p (8.51)
5 2 7
on the basis set of SOz find the probability of measuring the z-component of
the spin to be 12 „?
Solution
(a) Electrons are spin 12 particles so the only value of any component that can
be measured is ˙ 12 „.
(b) The normalized spinor ji represents the spin wave function of the electron
which in Eq. (8.51) is written in terms of j˛i and jˇi the eigenkets of SOz
[see Eq. (8.11)].

1 1 0
ji D p C7
5 2 0 1
1
D p Œj˛i C 7 jˇi (8.52)
5 2
where j˛i and jˇi are the spin-up and spin-down eigenkets of SOz . Thus, the
probability of measuring the spin angular momentum to be 12 „ is given
by the absolute square of the coefficient of jˇi which is
72 49
p 2 D
50
5 2
Because the admixture of ji is so heavily weighted toward jˇi the

eigenvalue 12 „ will be measured most often, in fact 98 times out of 100.
6. Use the Pauli spin matrices to find j˛ix and jˇix , the spin up and spin down
eigenkets of SOx , in terms of j˛i and jˇi, the spin up and spin down eigenkets of
SOz . Do the same for j˛iy and jˇiy , the spin up and spin down eigenkets of SOy .
Solution
The eigenvalue equation for the operator

O x is

01 a a
D (8.53)
10 b b
where represents the eigenvalues of
O x . We know, however, that, D ˙1
because all components of spin angular momenta are on an equal footing. By
convention SOz is chosen to be the one that commutes with SO2 .
Multiplying the matrices on the left side of this equation and equating matrix
elements we have
b D ˙a and a D ˙b (8.54)
Thus, the matrix elements of the eigenkets either have the same sign or the
opposite sign. The eigenkets are to be normalized so the one for which the
signs of the elements are the same is

1 1
j˛ix D p
2 1
1 1
D p j˛i C p jˇi (8.55)
2 2
The one for which the elements have different signs is

1 1
jˇix D p
2 1
1 1
D p j˛i p jˇi (8.56)
2 2
The subscripts on j˛ix and jˇix in Eqs. (8.55) and (8.56) indicate that they
correspond to spin up and spin down with respect to SOx . The eigenkets of SOy
can be obtained in an analogous manner.

0 i a a
D (8.57)
i 0 b b
so
ib D ˙a and ia D ˙b (8.58)
and we have

1 1
j˛iy D p
2 i
1 i
D p j˛i C p jˇi (8.59)
2 2
and

1 1
jˇiy D p
2 i
1 i
D p j˛i p jˇi (8.60)
2 2
7. Use the spin ladder operators to show that

1
SOx j˛i D „ jˇi
2
1
SOx jˇi D „ j˛i
2
i
SOy j˛i D „ jˇi
2
i
SOy jˇi D „ j˛i (8.61)
2
Solution
The x- and y-components of the spin operators are given in terms of the ladder
operators as
1 O 1 O
SOx D SC C SO and SOy D SC SO (8.62)
2 2i
Thus
p
SO j˛i D „ .s C ms / .s ms C 1/ jˇi
s
1 1
D„ C jˇi
2 2
D „ jˇi (8.63)
Similarly
p
SOC jˇi D „ .s ms / .s C ms C 1/ j˛i
s
1 1
D„ j˛i
2 2
D „ j˛i (8.64)
so
1 O
SOx j˛i D SC j˛i C SO j˛i
2
1
D .0 C „ jˇi/
2
1
D „ jˇi (8.65)
2
1 O
SOx jˇi D SC jˇi C SO jˇi
2
1
D .„ j˛i C 0/
2
1
D „ j˛i (8.66)
2
1 O
SOy j˛i D SC j˛i SO j˛i
2i
1
D .0 „ jˇi/
2i
i
D „ jˇi (8.67)
2
1 O
SOy jˇi D SC jˇi SO jˇi
2i
1
D .„ j˛i 0/
2i
i
D „ j˛i (8.68)
2
8. At time t D 0, an electron is in the spin up state in the x-direction, which we
designate by the ket j˛ix . The electron is now placed in a constant magnetic
field in the positive z-direction. The Hamiltonian is given by
HO D B (8.69)
where O D gSO is the magnetic moment of the electron, which is proportional

O g is a constant of proportionality.
to the spin S;
(a) Find j .t/i, the time dependent state vector that describes this electron for
t > 0 in terms of j˛i and jˇi, the spin up and spin down eigenkets of SOz .
(b) Find the time dependent probability that the electron is in the j˛ix state if
a measurement is made of the x-component of the electrons’s spin for any
time after t D 0.
Solution
(a) From Eq. (8.55) we know j˛ix in terms of j˛i and jˇi, so that
1 1
j .t D 0/i D j˛ix D p j˛i C p jˇi (8.70)
2 2
To find the time dependence of this state vector we must multiply each term
by the appropriate exponential time factor which means that we must have
the eigenenergies. We must thus solve the TISE with the Hamiltonian given
in Eq. (8.69), which may be written
HO D B D gSOz B (8.71)
because B is in the z-direction. Thus, the eigenvalue equation is
HO j .t D 0/i D gBSOz j .t D 0/i

gB„
D
j .t D 0/i (8.72)
2
Note that the minus sign corresponds to spin up and the plus to spin down.
With ! D gB=2 we may now write j .t/i as
1 1
j .t/i D p j˛i ei!t C p jˇi ei!t (8.73)
2 2
(b) To find the probability that the electron spin will be measured to be in the
j˛ix state we must re-cast Eq. (8.73) in terms of the kets j˛ix and jˇix . This
is easily done by solving Eqs. (8.59) and (8.60) of Problem 6 of this chapter
for j˛ix and jˇix in terms of j˛ix and jˇix . The result is
1 1
j˛i D p j˛ix C p jˇix
2 2
1 1
jˇi D p j˛ix p jˇix (8.74)
2 2
Inserting these into Eq. (8.73) we have
1 1
j .t/i D j˛ix ei!t C ei!t jˇix ei!t ei!t
2 2
D j˛ix cos !t jˇix sin !t (8.75)
The probability that the system is in the state j˛ix is given by the square of
the expansion coefficient, in particular
Probability in j˛ix D cos2 !t where ! D gB=2 (8.76)
Note that j .t/i is normalized since cos2 !t C sin2 !t D 1.

h i h i
9. Show that JO1z ; JO2 ¤ 0 ¤ JO2z ; JO2 , which show that neither JO1z nor JO2z can be
O2
specified simultaneously with J.
Solution
h i
We will work it for JO1z ; JO2 .
h i h i
JO1z ; JO2 D JO1z ; JO2x C JO2y C JO2z
h i h i h i
D JO1z ; JO2x C JO1z ; JO2y C JO1z ; JO2z
2 2
D JO1z ; JO1x C JO2x C JO1z ; JO1y C JO2y C0 (8.77)
h i h i
Because JO1u ; JO2u 0 where u D x; y; z and JO1z ; JO1z 0 we have
h i h i h i
JO1z ; JO2 D JO1z ; JO21x C 2JO1x JO2x C JO1z ; JO21y C 2JO1y JO2y
h i h i h i
D JO1z ; JO21x C JO21y C 2 JO1z ; JO1x JO2x C 2 JO1z ; JO1y JO2y
h i
D JO1z ; JO21 JO21z C 2i„ JO1y JO2x JO1x JO2y
¤0 (8.78)
10. In an effort to demystify “Clebsch-Gordanry” you are asked in this problem

to construct the table of Clebsch–Gordan coefficients for coupling two angular
momenta j1 D 1 and j1 D 1=2. This is not as difficult as it might at first appear.
Begin by writing the coupled kets as linear combinations of the uncoupled kets
and then cleverly apply the ladder operators to the top and bottom of the ladder
states. Note that you are being asked to construct Table S.3.
[Hint: First count and enumerate the different states in each representation.]
Solution
Following the hint we first enumerate the states using the quantum numbers
in either representation. Of course, we had better get the same number of
states no matter which representation we choose so we will use both as a
check. Let us use the coupled quantum numbers first, j D 3=2 and j D 1=2.
For j D 3=2 there are four values of mj : 3=2; 1=2; 1=2; 3=2. For j D 1=2
there are two values of mj : 1=2; 1=2. Thus, there are a total of six states. In
the uncoupled representation the quantum numbers are j1 D 1 and j2 D 1=2
with mj1 D 1; 0; 1. Because each mj1 state can have mj2 D ˙ 1=2 there are
(again) six states.
Now we correlate the coupled and uncoupled states. First the easy ones, the
top and bottom of the ladder states. It is clear that the only way to make up
mj D ˙3=2 is when the two individual z-components, mj1 and mj2 , have the
same sign. For convenience we recall the notation.
ˇ ˛ ˇ ˛
Coupled: ˇ1; 12 I j; mj ; Uncoupled: ˇ1; m` I 12 ; ms (8.79)
Now, the top and bottom of the ladder states are

ˇ 1 3 ˛ ˇ ˛
ˇ1; I ; ˙ 3 D ˇ1; ˙1I 1 ; ˙ 1 (8.80)
2 2 2 c 2 2 u
The other four states must be made up of linear combinations of two of the
states of the other representation. The correlations are listed in Table 8.1.
From this enumeration of states we can see that our table of Clebsch–Gordan
coefficients will have only ten non-zero entries, two each for four of the states
and one for each of the end of the ladder states. ˇ ˛
Our task is to evaluate the constants, the Ci . Inasmuch as ˇ1; 12 I 32 ; 12 c can be a
ˇ ˛ ˇ ˛
linear combination of only ˇ1; 1I 12 ; 12 c and ˇ1; 1I 12 ; 12 u we write
ˇ 1 3 1˛ ˇ ˛ ˇ ˛
ˇ1; I ; D C1 ˇ1; 1I 12 ; 12 u C C2 ˇ1; 0I 12 ; 12 u (8.81)
2 2 2 c
ˇ ˛ ˇ ˛
Table 8.1 Coupled and Coupled ˇ1; 12 I j; mj Uncoupled ˇ1; m` I 12 ; ms
uncoupled states for an ˇ 1 3 3˛ ˇ ˛
ˇ1; I ; ˇ1; 1I 1 ; 1
electron with orbital angular ˇ 21 23 21 ˛ ˇ 2 2 ˛
momentum ` ˇ1; I ; ˇ1; 1I 1 ; 1
2 2 2 2
ˇ ˛2
ˇ1; 0I 1 ; 1
ˇ 1 3 ˛ ˇ 2 2 ˛
ˇ1; I ; 1 ˇ1; 1I 1 ; 1
2 2 2 ˇ 2 2
˛
ˇ1; 0I 1 ; 1
ˇ 1 3 ˛ ˇ 2 2 ˛
ˇ1; I ; 3 ˇ1; 1I 1 ; 1
ˇ 21 21 1 ˛ 2 ˇ 2 ˛ 2
ˇ1; I ; ˇ1; 0I 1 ; 1
2 2 2 2 2
ˇ ˛
ˇ1; 1I 1 ; 1
ˇ 1 1 ˛ ˇ 2 2˛
ˇ1; I ; 1 ˇ1; 1I 1 ; 1
2 2 2 ˇ 2 2
˛
ˇ1; 0I 1 ; 1
2 2
The left-hand column is a listing of the possible

coupled states. The right-hand column contains the
uncoupled states that correlate with the adjacent
coupled states. Therefore, in the right-hand col-
umn, some uncoupled states are listed more than
once
Table 8.2 Clebsch–Gordan coefficients for j1 D ` D 1 and j2 D 1=2

j1 D 1I j2 D 1=2 j D 3=2 j D 1=2
m` ms 3=2 1=2 1=2 3=2 1=2 1=2
1 1/2 1
1 1=2 C1 C2
0 1/2 C3 C4
0 1=2 C5 C6
1 1/2 C7 C8
1 1=2 1
The state described in Eq. (8.81) can, however, be easily obtained by applying
the lowering operator, Eq. (8.7), to the top of the ladder ket in each representa-
tion in Eq. (8.80). This technique gives the Clebsch–Gordan coefficients, the Ci
listed in Table 8.2.
We have
ˇ ˛ ˇ ˛
JO ˇ1; 12 I 32 ; 32 c D LO C SO ˇ1; 1I 12 ; 12 u
p ˇ 1 3 1˛ ˇ ˛ ˇ ˛
3 ˇ1; 2 I 2 ; 2 c D LO ˇ1; 1I 12 ; 12 u C SO ˇ1; 1I 12 ; 12 u
p ˇ ˛ ˇ ˛
D 2 ˇ1; 0I 12 ; 12 u C ˇ1; 1I 12 ; 12 u (8.82)
p p
Comparing Eq. (8.82) with Table 8.2 we see that C1 D 1= 3 and C2 D 2=3
so we have the first non-trivial entry in our Clebsch–Gordan table. To obtain the
next one down we continue the procedure. We can, however, save some effort
by applying the raising operator to the bottom of the ladder state thus requiring
only minimal work for the computation. The final result is Table S.3.
ˇ ˛
11. A system is known to be in a particular coupled eigenstate ˇj1 ; j2 I j; mj c
that
ˇ is known˛ to be a linear combination of the uncoupled eigenstates
ˇj1 ; mj1 I j2 ; mj2 as
ˇ ˛ ˇ ˛ ˇ ˛
ˇj1 ; j2 I j; mj D C1 ˇ1; 0I 1 ; 1 C C2 ˇ1; 1I 1 ; 1 (8.83)
c 2 2 u 2 2 u
where C1 and C2 are real constants of opposite sign.

(a) Find the specific coupled ket in Eq. (8.83). That is, find C1 and C2 in
Eq. (8.83).
(b) If a measurement is made of the z-component of the individual angular
momenta, the quantum numbers mj1 and mj2 , what are the possible values
of the measurement and with what probabilities will these values be
measured?
Solution
(a) This problem is an exercise in combining the various quantum numbers

in the two different representations, coupled and uncoupled. After deduc-
ing the correct relationship between the coupled and uncoupled kets in
Eq. (8.83), determination of the Clebsch–Gordan coefficients C1 and C2
is an exercise in selecting the correct table in Appendix S.
First we note that because the kets on the rhs are uncoupled kets the values
1
of the individual angular momenta ˇ 1 are j˛1 D 1 and j2 D 2 . Thus, the
coupled ket must be of the form ˇ1; 2 I j; mj c so for convenience we re-write
Eq. (8.83) as
ˇ 1 ˛ ˇ ˛ ˇ ˛
ˇ1; I j; mj D C1 ˇ1; 0I 1 ; 1 C C2 ˇ1; 1I 1 ; 1 (8.84)
2 c 2 2 u 2 2 u
Now, to determine j and mj we see that for the first ket on the rhs mj1 Cm
j2 D
0 C 12 D 12 . For the second ket on the rhs mj1 C mj2 D 1 C 12 D 12 .
This tells us that the quantum number mj in the coupled ket on the lhs is
mj D mj1 C mj2 D 12 . The total angular momentum quantum number can,
however, be either j D j1 C j2 D 32 or j D j1 j2 D 12 . Therefore, the two
possible coupled kets are
ˇ ˛ ˇ ˛
ˇj1 ; j2 I j; mj D ˇ1; 1 I 1 ; 1
c1 2 2 2 c1
or
ˇ ˛ ˇ 1 3 1˛
ˇj1 ; j2 I j; mj D ˇ1; I ; (8.85)
c2 2 2 2 c2
Because j1 D 1 and j2 D 12 we require Table S.3 to decide between the two

coupled kets in Eq. (8.85). We see that for j D 12 one of the Clebsch–Gordan
constants is negative, but for j D 32 both are positive and it was stated in
the problem that C1 and C2 are of oppositeˇ sign. ˛We conclude, therefore,
that j D 12 and the ket that we seek is ˇ1; 12 I 12 ; 12 . Moreover, we see that
p
for mj1 D 0, mj2 D 12 the CG coefficient is 1=3 while for mj1 D 1,
p
mj2 D 12 the CG coefficient is 2=3. Therefore, the coupled eigenket is
r
ˇ 1 3 1˛ 1 ˇ ˛ 2 ˇˇ ˛
ˇ1; I ; D p ˇ1; 0I 2 ; 2 u C
1 1
1; 1I 12 ; 12 u (8.86)
2 2 2 c
3 3
(b) From the uncoupled kets on the rhs of Eq. (8.83) we see that the only
possible
valuesof the quantum numbers mj1 and mj2 that could be measured
are 0; 12 and 1; 12 . Of course, this was known at the outset because it
is essentially given in Eq. (8.83). What is not given are the probabilities
of measuring these couplets. These probabilities are simply the squares of
h are 1=3
the CG coefficients, so the probabilities i and 2=3, respectively. It
O O
is worthwhile to recall that because J 1z ; J 2z D 0, mj1 and mj2 can be
measured simultaneously.
12. The Hamiltonian for the interaction between the spins of the electron and the
proton in a H-atom, the hyperfine interaction is
2
HO HF D 2 SO1 SO2 (8.87)
„
where SO1 and SO2 represent the spins of the electron and proton and is a positive
constant. SO1 and SO2 can represent either the electron or the proton since the
Hamiltonian is symmetric in these operators. Using the coupled representation
and the notation jSMi for the coupled kets find the energy eigenvalues and the
eigenkets for the hyperfine interaction.
Solution
Noting that

SO2 D SO1 C SO2 SO1 C SO2
D SO21 C 2SO1 SO2 CSO22 (8.88)
we can solve for SO1 SO2 so that

HO HF D 2 SO2 SO21 SO22 (8.89)
„
Because Eq. (8.89) contains only the squares of the total and individual angular
momenta, it is clear that the coupled kets, the jSMi are the eigenkets of HO HF .
Additionally, there is no operator in HO HF that involves M so this quantum
number remains constant. Therefore, the states j1Mi are degenerate, in fact
threefold degenerate (M D 0; ˙1), and are known as triplet states. The
remaining coupled state, the j00i state, is non-degenerate and is known as a
singlet state. To find the energy eigenvalues we apply HO HF to the coupled kets.
For the triplet we have

HO HF j1Mi D 2 SO2 SO21 SO22 j1Mi
„

1 1 1 1
D 1 .1 C 1/ C1 C 1 j1Mi
2 2 2 2

D j1Mi (8.90)
2
For the singlet it is

HO HF j00i D 2 SO2 SO21 SO22 j00i
„

1 1 1 1
D 0 C1 C 1 j1Mi
2 2 2 2
3
D j1Mi (8.91)
2
13. As noted in Problem 12 of this chapter the Hamiltonian for the interaction
between the spins of the electron and the proton in a H-atom, the hyperfine
interaction is
2
HO HF D 2 SO1 SO2 (8.92)
„
where SO1 and SO2 represent the spins of the electron and proton and is a positive
constant.
(a) Show that this Hamiltonian can be written in terms of the uncoupled spin
ladder operators as
2
HO HF D 2 SO1C SO2 C SO1 SO2C C 2 SO1z SO2z (8.93)
„ „
(b) Write the singlet and triplet states, j1Mi and j00i, in terms of the uncoupled
states jms1 ; ms2 i.
(c) Using this result show that the eigenvalue of the triplet and singlet coupled
kets are 3=2 and =2, respectively.
Solution
(a)
2
HO HF D 2 SO1 SO2
„
2
D 2 SO1x SO2x C SO1y SO2y C SO1z SO2z
„
h i
D 2 SO1C C SO1 SO2C C SO2 SO1C SO1 SO2C SO2
2„
2
C 2 SO1z SO2z
„
O O 2
D 2 S1C S2 C SO1 SO2C C 2 SO1z SO2z (8.94)
„ „
(b) To write the coupled states as linear combinations of the uncoupled states
we, as usual, appeal to the tables of CG coefficients. In this relatively simple
case we require Table S.2 which, for convenience, we reproduce here.
S pms2 D 1=2 p
ms2 D 1=2
1 .1 C pM/ =2 .1 pM/ =2
0 1= 2 1= 2
Using the table we find

ˇ ˛
S D 1; M D 1; ms2 D 1=2 ) ms1 D 1=2 ) j11i D ˇ 12 ; 12
S D 1; M D 0; ms2 D ˙1=2 ) ms1 D
1=2 ) j10i
1 ˇ ˛ ˇ ˛
D p ˇ 12 ; 12 C ˇ 12 ; 12
2
ˇ ˛
S D 1; M D 1; ms2 D 1=2 ) ms1 D 1=2 ) j1 1i D ˇ 12 ; 12
S D 0; M D 0; ms2 D
1=2 ) ms1 D ˙1=2 ) j00i
1 ˇ ˛ ˇ ˛
D p ˇ 12 ; 12 ˇ 12 ; 12 (8.95)
2
(c) We first evaluate HO HF j00i using the Hamiltonian as evaluated in Part (a) of
this problem.
„2 O h i
H HF j00i D SO1C SO2 C SO1 SO2C C 2SO1z SO2z

1 ˇ ˛ ˇ ˛
p ˇ 12 ; 12 ˇ 12 ; 12
2
1 ˇ ˛ ˇ ˛
D p „2 ˇ 12 ; 12 ˇ 12 ; 12
2
2
1 „ ˇˇ 1 1 ˛ ˇˇ 1 1 ˛
C2 p 2; 2 2; 2
2 4
1 ˇ ˛ ˇ ˛
D p „2 Œ ˇ 12 ; 12 ˇ 12 ; 12
2
1 ˇ ˛ ˇ ˛
C ˇ 12 ; 12 ˇ 12 ; 12
2

3„2 1 ˇˇ 1 1 ˛ ˇˇ 1 1 ˛
D p 2; 2 C 2; 2 (8.96)
2 2
But the quantity in brackets is j00i so we have
„2 O 3„2 3
H HF j00i D j00i H) HO HF j00i D j00i (8.97)
2 2
Now evaluate HO HF j10i:
„2 O h i
H HF j10i D SO1C SO2 C SO1 SO2C C 2SO1z SO2z

1 ˇ ˛ ˇ ˛
p ˇ 12 ; 12 C ˇ 12 ; 12
2
1 ˇ ˛ ˇ ˛
D p „2 ˇ 12 ; 12 C ˇ 12 ; 12
2
2
1 „ ˇˇ 1 1 ˛ ˇˇ 1 1 ˛
C 2p 2; 2 C 2; 2
2 4
1 ˇ ˛ ˇ ˛
D p „2 Œ ˇ 12 ; 12 C ˇ 12 ; 12
2
1 ˇˇ 1 1 ˛ ˇˇ 1 1 ˛
; C 2; 2
2 2 2

„2 1 ˇˇ 1 1 ˛ ˇˇ 1 1 ˛
D p 2; 2 C 2; 2 (8.98)
2 2
But the quantity in brackets is j10i so we have
„2 O „2
H HF j10i D j10i H) HO HF j10i D j10i (8.99)
2 2
We could have made it easy on ourselves and worked it out for the bottom
or top of the ladder triplet state. For example,
ˇ ˛
HO HF j11i D HO HF ˇ 12 ; 12
2
O O ˇ1 1˛
D S 1C S 2 C O
S 1
O
S 2C C O
S 1z
O
S 2z
ˇ ;
2 2
„2 „2

2 „2 ˇˇ 1 1 ˛
D .0 C 0/ C ;
2 2
„2 „2 4
ˇ ˛
D ˇ 12 ; 12 D j11i (8.100)
2 2
These Problems 12 and 13 illustrate how a judicious choice of basis set can
greatly simplify a problem. While the use of the uncoupled set for obtaining
the energy eigenvalues and the eigenkets was not too laborious, use of the
coupled kets led almost immediately to these quantities.
14. This problem treats elastic scattering of quantum particles from a hard sphere
of radius R; it is the quantum mechanical analogue of Problem 12 of Chap. 1,
where the particle mass m is much less than the target sphere mass. The solution
shows that incorporating the wave nature of the scattered particles, taken to
be electrons or light atoms (not BBs!), leads to a substantially different result
for the low-energy cross-section than that obtained in the classical treatment.
Figure 8.1 is a schematic depiction of the problem.
Fig. 8.1 Problem 14
A plane wave with wave vector k and kinetic energy E directed along the z-
axis represents the particle flux incident on a spherical target. The magnitudes
of the energy and the wave vector are related by
p
2mE
kD (8.101)
„
The wave vector after the collision is designated k0 . In the general central
potential case, the scattered flux intensity varies with angle . No energy is
transferred, so k0 D k and both the incident and scattering flux intensities are
symmetric around the z-axis. The incoming plane wave and outgoing scattered
wave can be expanded in terms of spherical waves, exp .ikr/ =r, and Legendre
polynomials, P` .cos /, where ` is the angular momentum quantum number of
the particular expansion term. The problem is simplified if only the ` D 0 terms
are significant; this occurs in the low collision energy limit. In this ` D 0 limit
(known as s-wave scattering) the wave function that represents the outgoing
particles depends only on the radial coordinate r, and outside the range of the
scattering potential .r/ can be written [2]
1 ikr eikr
`D0 .r/ D e eikr f`D0 .k/ (8.102)
2ikr r
where f`D0 .k/ is the scattering amplitude. It is the coefficient of the term that
represents the outgoing spherical wave, eikr =r, which is the solution of the
scattering problem. This coefficient is related to the differential scattering cross
section d
=d by
D jf`D0 .k/j2 (8.103)

d
Equation (8.102) is valid for any central potential that falls off faster than 1=r.
(a) Obtain the scattering amplitude f`D0 .k/ for a hard sphere potential of
radius R. Use Eq. (8.103) to find the total cross section
in the limit k ! 0.
Compare this cross section result with the classical result from Problem 12.
(b) Use a classical argument to define the term “low energy collision,” such that
only ` D 0 scattering is important. Remember that the angular momentum
is quantized in units of „.
Solution
(a) For the hard-sphere the wave function .r/ must vanish at r D R. Setting
.R/ D 0 in Eq. (8.102) and rearranging, one obtains
1 ikR
f`D0 .k/ D e eikR eikR
2ik
1
D eikR sin kR (8.104)
k
so
sin2 kR
jf`D0 .k/j2 D (8.105)
k2
Taking the limit as k ! 0 leads to
lim jf`D0 .k/j2 D R2 (8.106)

k!0
To find the total cross section

we integrate Eq. (8.106) over all and
obtain

D 4R2 (8.107)
This value of
is four times larger than the geometric classical cross
section obtained in Problem 12 of Chap. 1. This result illustrates the inter-
ference phenomena associated with the wave nature of quantum particles.
The fact that diffraction of the incoming particles occurs at the surface of
the hard sphere causes more particles (four times more) to be scattered than
in the classical case for which no such diffraction occurs.
(b) The approximation to the total cross section in Eq. (8.107) is accurate to
within roughly 10 % for kR . 0:6 via Eq. (8.105). For a given R this defines
a limiting k. The validity of Eq. (8.106) by itself does not guarantee that
higher angular momenta (e.g., ` D 1) are not important. The classical
angular momentum for a collision that just grazes the hard sphere (see
Fig. 1.13) is given by
p
Lclassical D R 2mE (8.108)
p the quantum angular momentum for ` D 1 is „. Now we argue that if

and
R 2mE < „, a classical collision corresponding to ` D 1 with collision
energy E is not possible, so only ` D 0 collisions are important. This sets
an energy Emax for ` D 0 collisions to dominate:
„2
Emax < (8.109)
2mR2
To particularize, we can assume a hard sphere target of radius R D 0:5 nm,
or R ' 10 a.u. (see Appendix C), being impacted by a beam of electrons
(m D 1 a.u.). In atomic units „ D 1, so Eq. (8.109) yields Emax < 0:005 a.u.
(or 0.14 eV). If the irradiating particles were neutrons (m D 1839 a.u.) or
the sphere radius 0:5 m, s-wave scattering would lose dominance at very
low collision energy.
Chapter 9
Indistinguishable Particles
The defining characteristic of indistinguishable particles is that their interchange

cannot be detected. In this chapter we will concentrate on two-particle systems
because they are relatively simple and they contain the essential physics. For a two-
particle system we designate a state by ja bi where a and b represent the states (both
spatial and spin states) of each of the particles. The positions of a and b in the ket
ja bi designate particle 1 or particle 2, respectively, under the assumption that they
can be labelled. Thus, the ket ja bi represents a state in which particle 1 is in state a
and particle 2 is in state b. The two particles are, however, indistinguishable so that a
ket such as ja bi is unacceptable because the probability distribution is not invariant
under particle exchange. To accomplish this invariance the kets representing a given
state must be a linear combination of ja bi and jbai that is either symmetric or
antisymmetric under particle interchange to insure that the probability distribution
ˇ ˛
is invariant. The acceptable symmetric and antisymmetric kets are designated ˇ C
and j i, respectively, and are
ˇ ˛ 1
ˇ D p .ja bi C jb ai/ (9.1)
C
2
and
ˇ ˛ 1
ˇ D p .ja bi jb ai/ (9.2)

2
For these eigenfunctions exchange of all particle coordinates leads to the original
eigenfunction or the negative of the original eigenfunction.

218 9 Indistinguishable Particles
There are two types of indistinguishable particles, fermions (half-integral spin)

and bosons (integral spin).
Fermions are characterized by antisymmetric state vectors and bosons by
symmetric state vectors.
Examples of fermions are electrons, protons, and neutrons; examples of
bosons are photons and ˛-particles (bare helium nuclei).
Problems
1. For two identical bosons

(a) find the average positions of each of the particles. Do the same for a pair of
fermions.
(b) Find average of the squares of the separations between the bosons. Do
the same for the fermion pair. Contrast the results with those obtained for
distinguishable particles.
Solution
(a) According to Eqs. (9.1) and (9.2) the allowable eigenfunctions for identical
particles may be written
1
j ˙i D p .ja bi ˙ jb ai/ (9.3)
2
The average value of the position of, for example, particle 1 is
hx1 i D h ˙ j x1 j ˙i
1
D .ha bj ˙ hb aj/ x1 .ja bi ˙ jb ai/
2
1
D Œha bj x1 ja bi ˙ h abj x1 j bai ˙ h baj x1 j abi C h baj x1 j bai
2
1
D Œhx1 ia C hx1 ib (9.4)
2
where hx1 ia is the average of x1 when the particle is in state a; similarly for
hx1 ib . The terms with ˙ vanish because of orthogonality of the states a and b.
Clearly hx1 i hx2 i.
Let us understand the meaning of each term in Eq. (9.4). Because the
particles are identical, each can be in either state a or state b so hx1 ia is
9 Indistinguishable Particles 219
simply the average value of the position of particle 1 when it is in state a.

The term hx1 ib represents the average value of the position of the particle
when in state b. It is not unexpected that hx1 i, the average value of x for
particle 1 when it is in either state, a or b is the average of hxa ia and hxa ib .
Note that the values of hx1 i and hx2 i are the same for bosons and fermions.
(b) The average of the separation between the two particles is
D E
.x1 x2 /2 Dh ˙ j .x1 x2 /2 j ˙i
˙
2
D h ˙ j x12 2x1 x2 C x22 j ˙ i
˝ ˛ ˝ ˛
D x12 C x22 2 h ˙ j .x1 x2 / j ˙ i (9.5)
Now, inserting the kets j ˙i from Eqs. (9.1) and (9.2) let us examine the last
term in Eq. (9.5).
h ˙ j .x1 x2 / j ˙i D h.ha bj ˙ hb aj/j .x1 x2 / j.ja bi ˙ jb ai/i

D ha bj .x1 x2 / ja bi ˙ ha bj .x1 x2 / jb ai
˙ hb aj .x1 x2 / ja bi C hb aj .x1 x2 / j bai
D hx1 ia hx2 ib ˙ haj x1 jb i h bj x2 jai
˙ hbj x1 jai h aj x2 jbi C hx1 ib hx2 ia (9.6)
The matrix elements are the same regardless of whether we use x1 or x2 so

we simply use x. Equation (9.6) becomes
D E ˝ ˛ ˝ ˛
.x1 x2 /2 D x2 a C x2 b 2 hxib hxia
2 h aj x jbi2 (9.7)
˙
where we have taken advantage of the fact that hbj x jai D haj x jbi in the
last term.
Equation (9.7) contains some very important physics. If the particles were
distinguishable, then the last term would not be present. The
sign shows
that the average separation between the two indistinguishable particles is
smaller for bosons and larger for fermions. We see then that bosons like
to congregate while fermions are antisocial. This effect results from the
so-called exchange force, which is not a force in the usual sense. The effect
is a consequence of the symmetry requirements on the eigenfunctions. The
exchange force is responsible for the ordering of the periodic table as well as
many other effects such as boson condensation.
2. Two identical non-interacting spin- 12 particles (fermions) of mass m are subject

to a one-dimensional harmonic oscillator potential. Determine the ground state
and first excited state kets and their corresponding energies
220 9 Indistinguishable Particles
Solution
The Hamiltonian is given by

2
p21 1 p2 1
HO .x1 ; x2 / D 2 2
C m! x1 C 2 2
C m! x2
2m 2 2m 2
D HO .x1 / C HO .x2 / (9.8)
Because the Hamiltonian is separable as indicated in Eq. (9.8) the eigenfunctions

will be the products of the eigenfunction of the individual Hamiltonians and
the energy eigenvalues will be the sums of the individual eigenvalues. Thus, the
spatial part of the eigenfunctions may be written
jn1 n2 ix D jn1 i jn2 i (9.9)
The energy eigenvalues are
E .n1; n2 / D .n1 C n2 C 1/ „! (9.10)
To construct the total eigenfunctions we must include the spin eigenfunctions.

The coupled kets are designated jSMispin where S D 0; 1. For two spin- 21 particles
there are four states, the singlet state j00ispin and threej1Mispin states (the “triplet”
states). In terms of the uncoupled kets jms1 ms2 i these four states are
1 ˇ ˛ 1 ˇ ˛
j00ispin D p ˇ 12 21 C p ˇ 12 12
2 2
ˇ1 1˛
j11i Dˇ ;
spin 2 2
1 ˇ ˛ 1 ˇ ˛
j10ispin D p ˇ 12 21 C p ˇ 12 12
2 2
ˇ 1 1˛
j1 1i ˇ
D (9.11)
spin 2 2
Note that under exchange of particles the singlet spin state is antisymmetric and
the triplet spin states are all symmetric.
Ground state: The spatial part of the ground state is the symmetric state j00i with
energy
E00 D „! (9.12)
Because these two fermions are in a symmetric spatial state the spins must be in
the antisymmetric (singlet) state, clearly j00ispin (see Eq. (9.11)). Thus, the ket
representing the ground state of the system j i00 is
j i00 D j00ix j00ispin (9.13)

9 Indistinguishable Particles 221
First excited state:

The first excited state is more complicated because the states are fourfold
degenerate. The spatial parts of the total eigenkets are j01ix and j10ix each of
which has the same energy eigenvalue
E01 D E10 D 2„! (9.14)
These eigenkets are, however, neither symmetric nor antisymmetric so they must
be symmetrized. Inasmuch as they are degenerate, linear combinations of them
are also eigenfunctions so that using Eqs. (9.1) and (9.2) we have
1
jis D p .j10ix C j01ix /
2
1
jia D p .j10ix j01ix / (9.15)
2
where is used to designate the spatial part of the eigenkets. The total kets for
the first excited states are
1
jis j00ispin D p .j10ix C j01ix / j00ispin
2
1
jia j11ispin D p .j10ix j01ix / j11ispin
2
1
jia j10ispin D p .j10ix j01ix / j10ispin
2
1
jia j1 1ispin D p .j10ix j01ix / j1 1ispin (9.16)
2
Chapter 10
Bound States in Three Dimensions
Three-dimensional problems in quantum mechanics almost always involve central

potentials of the type that were dealt with in classical mechanics in Chap. 3. These
central potential problems are usually solved using spherical coordinates. When the
TISE can be separated in an additional coordinate system an additional symmetry
exists (beyond the obvious isotropic symmetry associated with a central potential).
This additional symmetry signals an “accidental degeneracy” [1], a degeneracy
beyond the .2` C 1/-fold degeneracy associated with the isotropy of the central
potential [1]. In this chapter we concentrate on central potentials with a small dose
of Cartesian potential problems just to show that there are other coordinate systems
used for three-dimensional problems in quantum mechanics.
For quantum mechanical central potential problems the symmetry of the potential
dictates that the eigenfunctions are given by
n`m .r; ; / D Rn` .r/ Y`m . ; / (10.1)
where the Y`m .; / are the spherical harmonics (see Appendix R and [1]) and
Rn` .r/ is the solution to the radial part of the separated TISE which is

„2 d 2 d ` .` C 1/ „2
r C C U .r/ Rn` .r/ D En` Rn` .r/ (10.2)
2mr2 dr dr 2mr2
Making the substitution un` .r/ D rRn` .r/ Eq. (10.2) is sometimes re-written in the
form

„2 d2 un` .r/ ` .` C 1/ „2
C C U .r/ un` .r/ D En` un` .r/ (10.3)
2m dr2 2mr2

224 10 Bound States in Three Dimensions
Using Eq. (10.3) and the trial solution u .r/ D rs the behavior of u .r/ and of R .r/
as r ! 0 can be deduced. The result is
lim u .r/ Ï r`C1 ) lim R .r/ Ï r` (10.4)

r!0 r!0
Additionally, again using Eq. (10.3), the behavior as r ! 1 for bound states
ŒE < U.1/ is determined to be
er
lim u .r/ / er ) lim R .r/ / (10.5)
r!1 r!1 r
p
where D 2mE=„2 .
Problems
1. A particle of mass m is trapped inside a cube, each side of which has length L.
The walls of the cube are impenetrable. It is a three-dimensional L-box.
(a) Find the energy eigenfunctions in Cartesian coordinates.
(b) Find the energy eigenvalues and compare the ground state energy with the
ground state energy of the one-dimensional L-box?
(c) What are the degeneracies of the first three energy levels?
Solution
(a) Inside the cube the Hamiltonian is
p2 p2y p2
HO .x; y; z/ D x C C z
2m 2m 2m
D HO x .x/ C HO y .y/ C HO z .z/ (10.6)
where each Hamiltonian HO xi .xi / is a one-dimensional L-box Hamiltonian

in each of the three coordinates. Because the Hamiltonian is the sum
of three Hamiltonians, each containing only a single coordinate, and the
boundary conditions are the same as those for an L-box, the solution to
the three-dimensional TISE is a product of the wave functions for the
individual Hamiltonians. Using quantum numbers nx ; ny ; nz D 1; 2; 3; : : :
we have
r !3 n n n
2 x y z
.x; y; z/ D sin x sin y sin z (10.7)
L L L L
10 Bound States in Three Dimensions 225
(b) Again because the Hamiltonian is the sum of three Hamiltonians, each
containing only a single coordinate, the eigenvalues are the sums of the
individual eigenvalues. Therefore
2 „2 2
Enx ny nz D 2
nx C n2y C n2z (10.8)
2mL
Inasmuch as there are three quantum numbers, each with a minimum value
ni D 1; it is clear from Eq. (10.8) that the ground state energy of the three-
dimensional L-box is three times that of a comparable one-dimensional
L-box.
(c) The sets of quantum numbers for the ground state and the first excited
state are
ground state: .1; 1; 1/

1st excited state: .1; 1; 2/ I .1; 2; 1/ I .2; 1; 1/
so the degeneracies are g1 D 1 and g2 D 3. For the second excited state we

have
2nd excited state: .1; 2; 2/ I .2; 1; 2/ I .2; 2; 1/
so that g3 D 3 as well.
Notice that g2 D g3 because in each triplet of quantum numbers for a given
energy level two of the three quantum numbers are identical, for example
.a; a; b/. If the quantum numbers are all different .a; b; c/, the degeneracy
is 6 because the number of permutations of 3 distinct objects is 3Š, and this
is the highest degeneracy that can occur for any level of a cubic box.
2. Consider a 3-dimensional L-box as in Problem 1 of this chapter. Assume the

box is large enough so that the energy levels are closely spaced and states with
high values of nx , ny , and nz are occupied. In this case it is useful to think about
the number ofstates present
in a narrow energy band E rather than a single
specific state nx ; ny ; nz . When the energy levels are closely spaced, we can
treat the ni ’s as continuous variables, and we define a quantity
q
nD n2x C n2y C n2z (10.9)
The number of states N with n-values from zero to a specific value n0 is

1 4 3
ND n0 (10.10)
8 3
which is simply
the volume of a sphere in n-space divided by 8 to account for
the fact that nx ; ny ; nz are all positive. The density of states g .E/, which is the
number of states in an energy interval dE, is defined as
dN
g .E/ D (10.11)
dE
Find an explicit expression for the density of states for a particle in an L-box
[use Eq. (10.8)].
Solution
From Eq. (10.8),
1 p
nD p E (10.12)
2 ˇ
where
2 „2
ˇD (10.13)
2mL2
Rewrite Eq. (10.11) as
dN dn
g .E/ D (10.14)
dn dE
and use Eqs. (10.10) and (10.12) to obtain the density of states

g .E/ D E1=2 (10.15)
4ˇ 3=2
The density of states is essentially the energy degeneracy in the continuous limit
where the energy states are very closely spaced. The density of states is used in
treating radiation (photon modes in an enclosure) and in statistical physics.
3. Consider a diatomic molecule as a pair of atoms with masses m1 and m2

connected by a rigid massless rod of length r. The moment of inertia is I D r2
where D m1 m2 = .m1 C m2 /. This problem is referred to as the “rigid rotor”
and may be considered the simplest of all central potential problems. Find the
energy eigenfunctions, the energy eigenvalues and the degree of degeneracy for
this rigid rotor.
Solution
The rod is rigid so the distance between the masses is fixed. There is therefore
no potential energy and the Hamiltonian consists of only kinetic energy T. O
This kinetic energy is purely rotational and can be written in terms of the total
angular momentum operator LO as
LO2
TO D HO D (10.16)
2I
We need go no further in solving the TISE because we have already solved it.
We know that the Hamiltonian operator is proportional to LO2 so the eigenfunc-
tions must be the spherical harmonics, the Y`m . ; /. Moreover we know that
O `m . ; / D 1 LO2 Y`m . ; /
HY
2I

1 2
D ` .` C 1/ „ Y`m . ; /
2r2
D E` Y`m . ; / (10.17)
so the eigenvalues must be

` .` C 1/ „2
E` D (10.18)
2r2
Because the eigenvalues do not depend upon the quantum number m this system
has the usual degeneracy associated with a central potential. Physically, the
central nature of the potential means that there is no preferred direction in space.
4. Another simple central potential is the three-dimensional analog of the one-

dimensional infinitely deep square well. The potential energy function is thus
U .r/ D 0 r < a
D1 r>a (10.19)
Find the eigenenergies and the energy eigenfunctions for the ` D 0 states. Note
the difficulty of finding the same quantities for ` ¤ 0.
Solution
Clearly, Rn` .r/ D 0 for r > a. For r < a the radial TISE for un` D rRn` .r/ is

„2 d2 un` .r/ ` .` C 1/ „2
C un` .r/ D En` un` .r/ (10.20)
2m dr2 2mr2
For ` D 0, however, this equation is
d2 un0 .r/ 2mEn0

C k2 un0 .r/ D 0 where k2 D (10.21)
dr2 „2
We see that we have already solved this equation, which is the same as that
encountered in the solution of the L-box (see Appendix M). The well-known
solutions are
un0 .r/ D A0 sin .kr/ C B0 cos .kr/ (10.22)
where the subscript zero on the constants of integration reminds us that this
solution is peculiar to the ` D 0 case.
Inasmuch as cos .kr/ ! 1 as r ! 0 we must set B0 D 0. In addition, the
boundary condition that Rn0 .a/ D 0 demands that ka D n so that
n2 2 „2
En0 D (10.23)
2ma2
which is the same result at that for an L-box or an a-box (see Appendix M).
Moreover, we already know the normalization constant An0 so we may write
the complete ` D 0 eigenfunctions. They are
r
2 sin .kn0 r/
n00 .r; ; / D Y00 . ; /
a r
r
1 sin .kn0 r/
D (10.24)
2a r
These radial eigenfunctions are the spherical Bessel functions of order zero [1].
Note that in Eq. (10.24) for the eigenfunctions, k now has a subscript n0 because
it corresponds to the energy eigenvalues En0 .
This problem was relatively simple only because we chose to solve it for the
` D 0 case. If ` ¤ 0, then the radial TISE is considerably more complicated
and the solutions are higher order spherical Bessel functions.
5. While Problem 4 of this chapter is likely too elementary to be found on a PhD

qualifying exam and variations of it for ` > 0 are too complicated because they
require knowledge of spherical Bessel functions, it is possible to construct a
problem using this theme which might be acceptable on such an exam. Consider
the shell potential
U .r/ D 1 r < a
D0 a<r<b
D1 r>b (10.25)
We ask the same questions as those asked for Problem 4, namely, find the
eigenenergies and the energy eigenfunctions for the ` D 0 states. Do not bother
to normalize the eigenfunction.
Solution
For r < a and r > b, R .r/ D 0. For the region in the shell a < r < b the radial
TISE for u D rR .r/ and ` D 0 is
d2 u .r/ 2mE
C k2 u .r/ D 0 where k2 D 2 (10.26)
dr2 „
the solution to which is again
u .r/ D A0 sin .kr/ C B0 cos .kr/ (10.27)
The boundary conditions are R .a/ D 0 D R .b/ so that u .a/ D 0 D u .b/.

Inserting these boundary condition into Eq. (10.26) leads to two equations and
two unknowns for A0 and B0 .
sin .ka/ A0 C cos .ka/ B0 D 0

sin .kb/ A0 C cos .kb/ B0 D 0 (10.28)
Inasmuch as these are homogeneous equations the only non-trivial solution is

the one obtained by setting the determinant of the coefficients equal to zero.
Thus
sin .ka/ cos .kb/ cos .ka/ sin .kb/ D 0 (10.29)
which, using Eq. (E.1), is
sin k .b a/ D 0 (10.30)
Reminiscent of square well problems already worked, Eq. (10.30) demands that
k .b a/ D n so that
2mE
.b a/2 D n2 2 (10.31)
„2
Solving for En0 as in Problem 4 of this chapter we have
n2 2 „2
En0 D (10.32)
2m .b a/2
Additionally, from Eq. (10.28) we see that A0 and B0 are related by
B0 D tan .ka/ A0 (10.33)
so that, from Eq. (10.27) the radial energy eigenfunction is

u .r/ A0
Rn0 .r/ D D Œsin .kn0 r/ tan .kn0 a/ cos .kn0 r/ (10.34)
r r
Because we are dealing with only ` D 0 states the angular part of the
eigenfunctions is Y00 . ; /.
6. Find the probability that the electron in the ground state of a H-atom will be
found in the classically forbidden region? Use atomic units (see Appendix C)
in which the Coulomb potential is 1=r, the ground state eigenfunction is
1
.r; ; / D p er
100 (10.35)

2
and the energy eigenvalues are En D 1= 2n for all values of n.
Solution
For the ground state n D 1, ` D 0 so we must first find the value of rc the
classical turning point for the effective potential with ` D 0, i.e. the pure
Coulomb potential as shown in Fig. 10.1.
Fig. 10.1
The classical turning point occurs when the kinetic energy is zero. Therefore,
the total energy, which is the energy eigenvalue, must equal the potential energy
and
1 1
2
D H) rc D 2n2 (10.36)
2n rc
Therefore, the probability Pforbidden of finding the electron in the classical

forbidden region for the ground state is
Z Z Z 2
1
1 r 2 2
Pforbidden D dr p e r sin d d
rc D2 0 0
Z
4 1 2 2r
D r e dr (10.37)
rc D2
Using the integral in Eq. (G.1) we have

ˇ2
e2r 2 2r 2 ˇˇ
Pforbidden D4 r C C 2 ˇ
2 2 2 1

22 1
D 2e4 22 C C
2 2
D 13e4
Ð 0:24 (10.38)
It is fairly surprising that the electron spends nearly one quarter of its time in
the classically forbidden region.
˝ ˛
7. Find the expectation value r2 for the ground state of a H-atom. The
normalized radial eigenfunction for a H-atom in the ground state is (see
Table T.2)
3=2
1
R10 .r/ D 2 er=a0 (10.39)
a0
where a0 is the Bohr radius.
Solution
The expectation value of r2 for the ground state of hydrogen is (see Table R.1
for Y00 .; /)
Z
1 1
D Y00 . ; / R10 .r/ Y00 . ; / R10 .r/ dV
r2 100 all space r2
Z Z 1
ŒR10 .r/2 2
D ŒY00 . ; /2 d r dr
all 0 r2
( " # Z )
1 3 1
2r=a0
D f1g 4 e dr
a0 0
3
1 a0
2r=a0 0
D4 e 1
a0 2
2
1
D2 (10.40)
a0
˝ ˛
It is satisfying that our computed value of r2 has units 1= m2 .
8. Kramer’s relation permits calculation of hrs i for arbitrary values of n and `, and
integral values of s (see the last entry in Table T.3). This relation is
.s C 1/ s ˝ s1 ˛ s 2 h 2 2
i˝ ˛
hr i .2s C 1/ a0 r C a0 .2` C 1/ s rs2 D 0 (10.41)
n2 4
˝ ˛
(a) Using Kramer’s relation evaluate ˝ r1˛ and hri in that order.
3
(b) Using Kramer’s relation
˝ 2 ˛ evaluate r for the ground state. [Hint: To avoid
having to evaluate r use Table T.3.]
Solution
˝ ˛
(a) Choosing s D 0 in Eq. (10.41) immediately leads to r1 because,
conveniently, the last term vanishes for s D 0. We have
1 ˝ ˛ ˝ ˛ 1
2
a0 r1 D 0 ) r1 D 2 (10.42)
n n a0
To evaluate hri we choose s D 1 and obtain
2 a20 h 2
i˝ ˛
hri 3a0 C .2` C 1/ 1 r1 D 0 (10.43)
n2 4
˝ ˛
It is clear from Eq. (10.43) that r1 is required to evaluate hri. This
illustrates the bootstrapping necessary to use Kramer’s relation (except for
s D 0). Solving Eq. (10.43) for hri and using the result in Eq. (10.42) we
have
n2

a20 h i 1
2
hri D 3a0 .2` C 1/ 1
2 4 n2 a0
2
3n 1 2
D 4` C 4` a0
2 4

2 a0
D 3n ` .` C 1/ (10.44)
2
˝ ˛
(b) To evaluate r3 we use s D 1 in Kramer’s relation to obtain
˝ ˛ 1 h i˝ ˛
a0 r2 a20 .2` C 1/2 1 r3 D 0 (10.45)
4
˝ ˛
Although we have computed r1˝ , this ˛ result is˝ not
˛ helpful for other
negative values of s. We require r2 to obtain r3 . Using the hint in
the statement of the problem we have from Table T.3
( )
˝ 2 ˛ 1 1
r D 2 (10.46)
a0 n3 ` C 12
˝ ˛
Solving Eq. (10.45) for r3 we have
˝ ˛ 4a0 ˝ ˛
r3 D h i r2
a20 .2` C 1/2 1
( )
4a0 1 1
D 2
a0 Œ4`2 C 4` a20 n3 ` C 12
1
D 3
(10.47)
.na0 / Œ` .` C 1/ ` C 12
This problem illustrates the usefulness of the Kramer relation as well as its˛
˝ 2
limitations. Fortunately there
˝ ˛ are several ways to obtain the value of r
that is required to obtain r3 (see, for example, [2]).
9. A particle of mass m is subjected to an unknown central potential U .r/ that

vanishes as r ! 1. A particular (unnormalized) eigenfunction in spherical
coordinates is .r; ; / D eˇr where ˇ is a real positive constant. Use atomic
units (see Appendix C).
(a) Does this state have definite angular momentum? If so, why and what is the
total angular momentum and the z-component of the angular momentum of
this state?
(b) Is the state bound or free? What is the energy eigenvalue of this state?
Hint: Because lim U .r/ D 0 all the energy at r D 1 is kinetic energy. In
r!1
a.u. the kinetic energy operator is given by
1
TO D r 2
2

1 1 @ 2 @
D f 2 r C
2 r @r @r

1 1 @ @ 1 @2
sin C g (10.48)
r2 sin @ @ sin2 @ 2
(c) What is the potential energy?
Solution
(a) Because the given function is known to be an eigenfunction for a cen-

tral potential it must have definite angular momentum and z-component
of angular momentum.
p The angular momentum eigenfunction is clearly
Y00 .; / D 1=4 because it is the only spherical harmonic with no
p that we must have ` D 0
angular dependence [see Eq. (10.1)]. This means
and m D 0 and the total angular momentum is ` .` C 1/ D 0.
(b) If we operated on .r; ; / with HO D TO C U .r/, we would get the
eigenvalue. We do not, however, know U .r/ other than that it vanishes
as r ! 1. Therefore, we operate on .r; ; / with HO and take the limit
as r ! 1. Because lim ŒU .r/ .r; ; / D 0, the eigenvalue will be
r!1
the eigenvalue of TO .r; ; / as r ! 1 which is the total energy. Using
1
TO D r 2 from the hint
2
1
lim TO .r; ; / D lim r 2 .r; ; /
r!1 2 r!1
D E .r; ; / (10.49)
We have then

2 1 @ 2 @
r .r; ; / D Œ 2 r eˇr
r @r @r
1 @
2 ˇr
D ˇ r e
r2 @r
1
D ˇ 2 2r ˇr2 eˇr
r

2ˇ
D C ˇ2 .r; ; / (10.50)
r
Taking the limit as r ! 1 and multiplying by 12 we see that the

eigenvalue of TO at 1 is .1=2/ ˇ 2 which is a manifestly negative number
meaning that at r D 1 the particle is in the classically forbidden region
and thus this energy represents a bound state.
Actually, the bound nature of the state was evident at the outset because
lim .r; ; / D 0 (10.51)

r!1
The derivation above was worthwhile, however, because it provided both

the nature of the state and its eigenvalue.
(c) Because HO D E D TO .r/ C U .r/
HO .r; ; / D E .r; ; / D TO .r/ .r; ; / C U .r/ .r; ; /

(10.52)
Therefore
U .r/ .r; ; / D E .r; ; / TO .r/ .r; ; / (10.53)

h i
so we simply subtract TO .r; ; / not lim T .r; ; / from the total
r!1
energy found above. Remembering to include the 12 that makes the
eigenvalue of r 2 the kinetic energy we have

1 1 2ˇ
U .r/ D ˇ 2 C ˇ2
2 2 r
" #
1 2 ˇ ˇ2
D ˇ
2 r 2
1 ˇ ˇ2
D ˇ2 C
2 r 2
ˇ
D (10.54)
r
which is the Coulomb potential. We should have anticipated the Coulomb
potential because .r; ; / D eˇr is the (unnormalized) ground state
eigenfunction of the H-atom.
10. The effective potential for the H-atom in atomic units for which e D „ D me D
1 (see Appendix C) is
1 ` .` C 1/
Ueff .r/ D C (10.55)
r 2r2
where ` is the usual angular momentum quantum number. In the same units,
the quantized energy is
1
En D (10.56)
2n2
(a) Find the position of the minimum of Ueff .r/, call it r0 , and also find
Ueff .r0 /. Sketch Ueff .r/ vs. r for several values of ` (say three), including
` D 0.
(b) By comparing Ueff .r0 / with nearby values of the H-atom eigenvalues En
determine the relationship between the magnitudes of ` and n. Namely, for
a given n, what is the restriction on `. This problem provides a graphical
way of determining the relationship between principal quantum number n
and the angular momentum quantum number ` for the H-atom.
Solution
(a) Taking the derivative and setting equal to zero leads to

ˇ
dUeff .r/ ˇˇ 1 ` .` C 1/
D 2 D0
dr ˇrDr0 r0 r03
H) r0 D ` .` C 1/ (10.57)
The second derivative is

ˇ
d2 Ueff .r/ ˇˇ 2 3` .` C 1/
2 ˇ D 3 C
dr rD`.`C1/ r r4
2 3` .` C 1/
D 3
C >0 (10.58)
Œ` .` C 1/ Œ` .` C 1/4
Therefore r0 D ` .` C 1/ is a minimum. Moreover, as ` increases r0
increases. Also
1 ` .` C 1/ 1
Ueff .r0 ; `/ D C D (10.59)
` .` C 1/ 2 Œ` .` C 1/2 2` .` C 1/
Thus, as ` increases, both the position of the minimum and the minimum
value of Ueff .r0 ; `/ increase as shown in Fig. 10.2.
Ueff(r)
=2
0.1
=3
2 4 6 8 10 12 14 1
r
0 6
n=4
n=3
–0.1
n=2
–0.2
=1
–0.3
–0.4
=0
n=1
–0.5
Fig. 10.2 Problem 10a—solution. The dashed lines represent the hydrogen atom energy eigen-
values
(b) Using Fig. 10.2 we focus on two adjacent energies and their relation to
Ueff .r; `/. Although any two adjacent energies could be chosen, we elect
to examine the relationship of EnD2 and EnD3 to the effective potential that
falls between them, Ueff .r; ` D 2/. It is seen that EnD2 is not compatible
with Ueff .r; ` D 2/ because the energy EnD2 does not intersect this effective
potential. Notice that EnD2 is compatible with both ` D 0 and ` D 1 and
EnD3 is compatible with ` D 0; 1; 2.
This behavior is characteristic of all adjacent energy levels En1 and En .
The energy corresponding to the smaller of the two principal quantum
numbers is always lower than the minimum value of the effective potential,
Ueff .r0 ; ` D n/. Therefore, for compatibility between n and ` we must have
En Ueff .r0 ; `/ (10.60)
or
1 1
2
(10.61)
2n 2` .` C 1/
Solving this inequality leads to
n2 ` .` C 1/ (10.62)
Because n and ` are integers we must have n > `. The relationship between
n and ` in Eq. (10.62) is usually stated as:
`n1 (10.63)
Equation (10.63) can also be obtained directly from the solution of the TISE
for the H-atom [1].
11. In Problem 10 of this chapter it was determined that the relation between the
principal quantum number n and the azimuthal (angular momentum) quantum
number ` is
`n1 (10.64)
The energy levels of the H-atom in atomic units are
1
En D (10.65)
2n2
Because En depends only upon n it is clear from Table T.1 that the states of
the H-atom are degenerate because there can be several values of ` and m for a
given n. Find the degree of degeneracy of the H-atom gH (do not consider spin).
That is, determine how many states that have a given value of n have the same
energy.
Solution
To determine gH we must sum all `-states for a given n over all their possible
sub-states. Using the relationships given in Table T.1 we can calculate the total
number of states for a given principal quantum number. For a given ` there are
2` C 1 possible values of m. Therefore, we must sum these values for every
possible value of ` for a given n, that is, from ` D 0 to ` D n 1. We have then
X
n1
gH D .2` C 1/ (10.66)
`D0
First Method
We use Gauss’s trick, which provides a formula for summing the first j integers
[1]. This formula is
XM
M .M C 1/
jD (10.67)
jD0
2
Thus,
X
n1 Xn1
gH D 2 `C 1
`D0 `D0
.n 1/ n
D 2 Cn
2
D n2 (10.68)
Second Method (This is Actually a Derivation of Gauss’ Trick)
We first write the terms of Eq. (10.66) as follows:
X
n1
.2` C 1/ D 1 C 3 C 5 C .2n 3/ C .2n 1/ (10.69)
`D0
Now re-write Eq. (10.69) with the terms in reverse order.
X
n1
.2` C 1/ D .2n 1/ C .2n 3/ C 5 C 3 C 1 (10.70)
`D0
We now add Eqs. (10.69) and (10.70) term by term starting from the equal signs
to obtain
X
n1
2 .2` C 1/ D 2n C 2n C 2n C 2n C 2n C 2n (10.71)
`D0
Inasmuch as there are n terms on the right-hand side (we started counting at
` D 0) we may write
X
n1
2 .2` C 1/ D n .2n/ (10.72)
`D0
or
X
n1
gH D .2` C 1/ D n2 (10.73)
`D0
This second method is really no different from the first method because it is the
same trick of adding the re-written sums that are used to establish Gauss’ trick.
The origin of the trick and thus its appellation are interesting themselves [1].
12. While the H-atom energy eigenvalues depend only upon the principal quantum
number n, other atoms that resemble hydrogen atoms have nearly hydrogenic
energy level formulas. This is especially true of the alkali atoms, which have
a single valence electron outside a closed shell rare gas configuration. At the
end of the nineteenth century Rydberg found that the energies of many atoms,
especially the alkali atoms, could be described by a hydrogen-like formula that
was `-dependent. In atomic units (see Appendix C)
1
En` D (10.74)
2 .n ı ` /2
where ı ` is called the quantum defect and is dependent upon the angular
momentum of the valence electron. Notice that this formula reduces to the H-
atom energy if ı ` D 0. In this model the atom is viewed as having a single
valence electron under the influence of a spherical ball of charge consisting of
the point nucleus of charge CZe surrounded by .Z 1/ electrons. If the angular
momentum of the valence electron is high, it stays away from the ionic core
and the system behaves in a very hydrogenic manner. If the angular momentum
is low, the valence electron encounters the electron cloud and deviates from
hydrogenic behavior.
The effects of core penetration by the valence electron may be approximated by
assuming that the potential as seen by the valence electron is, in atomic units,
1 b
U .r/ D 2 (10.75)
r r
Of course, the first term is simply the Coulomb potential while the second term
accounts for the increasing attraction of the valence electron by the nucleus
as the valence electron penetrates the electron cloud that shields the nuclear
charge. Multi-electron atoms for which the energy is given by Eq. (10.74) are
called Rydberg atoms.
Now for the problem: Find an approximate expression for the quantum defect
ı ` in terms of b and `.
Solution
The key point here is that the non-Coulombic term in the potential U .r/ has the
same dependence on r as does the centrifugal term in the effective potential.
We may thus write the effective potential in the form
1 b ` .` C 1/
Ueff .r/ D 2 C
r r 2r2
1 Œ` .` C 1/ 2b
D C (10.76)
r 2r2
We already know the solution to the eigenvalue problem with an effective
potential in the form of Eq. (10.76). We merely identify the numerator in the
second term with ` .` C 1/ in the solution of the pure hydrogen problem. We
therefore let

`0 `0 C 1 D ` .` C 1/ 2b (10.77)
At this point it is necessary to recall the relation between n and ` that led to the
solution of the radial Schrödinger equation for the H-atom [1]. It was necessary
to terminate an assumed infinite series to avoid non-physical conditions. We
had
n D nr C .` C 1/ (10.78)
where nr is the index of the last term in the series. Incidentally, this relationship
provides an analytical method of arriving at the conclusion arrived in Prob-
lem 10 of this chapter, Eq. (10.63).
The energy of the H-atom in terms of nr is
1
EH D (10.79)
2 Œnr C .` C 1/2
By analogy with the H-atom the Rydberg atom energy is
1
ER D
2 Œnr C .`0 C 1/2
1
D
2 Œnr C .` C 1/ C `0 `2
1
D (10.80)
2 Œn .` `0 /2
Comparing Eq. (10.80) with Eq. (10.74) we see that

ı ` D ` `0 (10.81)
Now we must eliminate `0 from Eq. (10.81) using Eq. (10.77)

2

`0 C `0 `2 C ` D 2b (10.82)
which is a quadratic equation for `0 .

2
`0 C `0 C Œ2b ` .` C 1/ D 0 (10.83)
Then
1 1p
`0 D ˙ 1 4 Œ2b ` .` C 1/
2 2
1 1p
D ˙ 1 8b C 4` .` C 1/
2 2
r
1 1
D ˙ `2 C ` C 2b
2 4
s
2
1 1
D ˙ `C 2b (10.84)
2 2
and
2 s 3
2
1 1
ı ` D ` 4 ˙ `C 2b5
2 2
s
2
1 1
D `C
`C 2b (10.85)
2 2
Now we must consider the appearance of the ˙ sign in Eq. (10.85). There
cannot be two values of the quantum defect. We can use a limiting case to decide
which root is physically meaningful. Suppose b D 0. If this were the case we
would have a pure H-atom potential [Eq. (10.76)] and ı ` 0 which can only
occur with the minus sign in Eq. (10.85). The plus sign gives an extraneous root
so we can ignore it and move on.
Now, let us consider the properties of ı ` . Assuming that ` > b (because we
are interested in high angular momentum states) we expand the expression for
ı ` using the binomial theorem.
s
2
1 1
ı` D ` C `C 2b
2 2
8 " #1=2 9

1 < 2b =
D `C 1 1
2 : ` C 12
2 ;
" #
1 b
`C 2
2 `C 1 2
b
D (10.86)
` C 12
Two important conclusions can be drawn from Eq. (10.86).

1. This relation between ı ` and ` is consistent with our premise that higher
angular momentum states are more nearly hydrogenic than states having
lower angular momentum inasmuch as the valence electron “stays away”
from the electron core that causes the non-hydrogenic effects.
2. The accidental degeneracy of the H-atom is broken by the presence of
the electronic core because the energy depends upon the quantum defect
which depends upon the angular momentum, a parameter of which the true
hydrogen energy is independent.
Chapter 11
Approximation Methods
Most practical problems in quantum mechanics do not lead to Schrödinger equations

that allow exact solution. For this reason approximation methods are necessary. In
this chapter we present some problems that illustrate the most frequently used of
these methods.
11.1 The WKB Approximation
This approximation technique is developed by expanding the TISE for a single

independent variable in powers of „. Retaining the first two terms in the expansion
(the first term is the classical limit, „ D 0; the second term is linear in „) one
eventually obtains expressions for the wave function and energy. The details are
discussed at length in most textbooks [1].
For a potential function U .x/ with one minimum and classical turning points at
x D a and x D b, the wave function for U .x/ < E (the classically allowed region
between a and b) is given by
Z x Z x
B C
.x/ D p exp i k .x/ dx C p exp i k .x/ dx (11.1)
k .x/ a k .x/ a
where B and C are constants; k is the classical momentum divided by „.
1p
k .x/ D 2m ŒE U .x/ (11.2)
„

244 11 Approximation Methods
In the classically forbidden regions x < a and x > b, where U .x/ > E, the wave
functions are
Z a
A
.x/ D p exp .x/ dx x < a
k .x/ x
Z x
D
.x/ D p exp .x/ dx x < a (11.3)
k .x/ b
with
1p
.x/ D 2m ŒU .x/ E (11.4)
„
Finally, if the potential permits bound states, the discrete energies are found from
Z

b
1
p .x/ dx D n C „ (11.5)
a 2
where p .x/ is the classical momentum

p
p .x/ D „k .x/ D 2m ŒE U .x/ (11.6)
It is also possible to use the WKB approximation to estimate the transmission

coefficient through a barrier. The approximation to the transmission coefficient is
Z
2 bp
TWKB D exp 2m ŒU .x/ Edx (11.7)
„ a
Problems
1. Use the WKB approximation to find the quantized energies for an L-box.
Compare the result with the exact solution. What happens as n ! 1?
Solution
Applying Eq. (11.5) to the L-box we have

Z b Z L p p
p .x/ dx D 2mEdx D L 2mE
a 0

1
D nC „ (11.8)
2
11 Approximation Methods 245
which leads to

1 2 2 „2
En D n C (11.9)
2 2mL2
This value of the quantized energy is very nearly the same as the exact solution
that is obtained by solving the TISE (see Appendix M). This exact energy is
n2 2 „2
En D (11.10)
2mL2
so it is evident that as n ! 1, which is the region of validity of the WKB
approximation, the approximate energy expression approaches the exact one.
2. Use the WKB approximation to find the energy levels for a harmonic oscillator.
How do the results compare with the exact solution?
Solution
For the harmonic oscillator the momentum is given by

p 1
p .x/ D 2m .E Kx2 / with K D m! 2
2
At a given energy E the classical turning points of the harmonic oscillator
occur when the classical momentum vanishes. Thus, the classical turning points
for x < 0 and x > 0 are
r r
E E
xa D and xb D (11.11)
K K
From Eq. (11.5) we have

Z
xb
1
p .x/ dx D n C „
xa 2
p Z xb p
D 2 2mK .E=K x2 /dx
0
p Z q
xb
D 2 2mK xb2 x2 dx (11.12)
0
To evaluate the integral we employ the integration formula of Eq. (G.5).

2 q 3xb
Z xb p x x 2
x 2 2
6 b x x7
p .x/ dx D 2 2mK 4 C b sin1 5
0 2 2 xb
0

p xb2
D 2 2mK
2 2
r
p E m
D 2mK D E
2 K 2K
1
D E (11.13)
!
Inserting this result into Eq. (11.5) we have

1 1 1
En D n C „ ) En D n C „! (11.14)
! 2 2
which is the same as the exact result.
3. A particle of mass m and kinetic energy E < U0 is incident from the left on a
one-dimensional parabolic potential barrier given by

x2
U .x/ D U0 1 2 (11.15)
x0
as shown in Fig. 11.1. Use the WKB approximation to find the approximate
transmission rate through the barrier as a function of E < U0 .
Fig. 11.1 Problem 3 U(x)
m
E
-x0 -a 0 a x0 x
Solution
The WKB approximation to the transmission coefficient T is given in Eq. (11.7).

Inserting the parameters of this problem we have
Z
2 ap
TWKB D exp 2m ŒU .x/ Edx (11.16)
„ a
where x D a when U .x/ D E (see Fig. 11.1) so that

a2 E
E D U0 1 2 ) a2 D x02 1 (11.17)
x0 U0
Then
" p Z 1=2 #
2 2m a x2 a2
TWKB D exp U0 U0 2 U0 C U0 2 dx
„ a x0 x0
2 s 3
p Z a 2
2a 2mU 0 x
D exp 4 1 2 dx5 (11.18)
„x0 a a
We let x=a D u to put Eq. (11.18) in a form consistent with the integral given
in Eq. (G.6), which is
Z 1 p
1 u2 du D (11.19)
1 2
Using Eq. (11.17) along with the new variable u converts Eq. (11.18) to
p Z
E 2 2mU0 1 p
TWKB D exp x02 1 2
1 u du (11.20)
U0 „x0 1
which, using Eq. (11.19), becomes

" s #
.U0 E/ 2m
TWKB D exp x0 (11.21)
„ U0
The first thing we notice about Eq. (11.21) is that when E D U0 the
transmission coefficient is unity. Just to be sure that our answer makes sense let
us check the units. Inasmuch as we are dealing with an exponential the exponent
must be unitless.
s p
.U0 E/ 2m m kg J
x0 D
„ U0 . J s/
p
m kg
D q (11.22)
2
s kgs2m
which is unitless. This does not prove that the answer is correct, but it gives us
some confidence.
11.2 The Variational Method
The principle behind the variational method is simple to understand. We start with
the assumption that the Hamiltonian has a set of eigenfunctions and associated
eigenvalues so that we can write
HO j n i D En j n i (11.23)
where the j n i are assumed to be the exact orthonormal eigenkets and the En the
corresponding energy eigenvalues. Neither are known; the only thing known in
Eq. (11.23) is the Hamiltonian.
We begin by guessing a normalized trial ground-state eigenfunction, which we
designate by j i. We should choose a trial wave function that matches the boundary
conditions of the unknown exact function and that has any known characteristics,
e.g. no nodes for a ground state, and appropriate symmetry. We can expand j i in
terms of the complete, but unknown, set of orthonormal eigenkets j n i.
1
X
j iD cn j n i (11.24)
nD0
The expectation value of the energy in the state represented by j i is
hEi D h j HO j i
1
X
D jcn j2 h n j HO j n i
nD0
1
X
D jcn j2 En (11.25)
nD0
The value of the ground state energy, E0 , is lower than any of the other eigenvalues.
Therefore, if in Eq. (11.25) we replace all of the En by E0 we obtain the inequality
1
X
hEi E0 jcn j2
nD0
E0 (11.26)
where, because j i is normalized, the summation in Eq. (11.26) is unity. In

coordinate space this inequality is
Z
hEi D
.r/ HO .r/ dV
E0 (11.27)
so, given any trial wave function, we are assured that the integral in Eq. (11.27) will
be greater than the true ground state energy.
In typical applications, parameters are incorporated into the trial wave function.
After computing the integral in Eq. (11.27), j i is then minimized with respect
to these parameters to obtain the lowest energy possible for the chosen functional
form. No matter how many parameters are included in the trial wave function hEi
will be greater than the ground state energy, although with modern computers and
sophisticated functional forms one can come very close to the actual value.
It is sometimes possible to obtain useful variational approximations to the first
excited state provided symmetry precludes the ground state from consideration. For
example, if a one-dimensional potential well is an even function and has more than
one bound state, the ground state eigenfunction will be an even function and the
first excited state eigenfunction will be an odd function (see Sect. 6.2). If the trial
function is chosen to be odd, the variational method will converge toward the first
excited state energy. In a sense the first excited state may be thought of as the ground
state for odd symmetry.
Problems
1. Estimate the ground state energy of a harmonic oscillator using the variational
2 2
method with the trial wave function .x/ D Aeˇ x . Because this trial wave
function is of the correct form of the actual ground state eigenfunction we expect
to obtain the exact ground state energy after minimizing hEi with respect to the
p ˇ. Naturally we
parameter p expect that the value of ˇ that minimizes hEi will be
ˇ D ˛= 2 where ˛ D m!=„. Show all of this.
Solution
First normalize using Eq. (G.3).

Z 1
2 2
h j i D 1 D A2 e2ˇ x
dx
1
s
2
DA (11.28)
2ˇ 2
so
s
2ˇ 2
A2 D (11.29)

Then
hEi D h j HO j i
s
Z
2ˇ 2 1 ˇ2 x2 „2 d2 1 2 2
D e C m! eˇ x dx
2
1 2m dx2 2
s !Z Z
2ˇ 2 1 2 2„2 ˇ 4 1
2 2 ˇ 2 „2 1 2ˇ2 x2
D m! x2 e2ˇ x dx C e dx
2 m 1 m 1
s ! s s
2ˇ 2 1 2 2„2 ˇ 4 1 ˇ 2 „2
D m! 2 2
C
2 m 4ˇ 2ˇ m 2ˇ 2
m! 2 ˇ 2 „2
D C (11.30)
8ˇ 2 2m
Because ˇ occurs only as ˇ 2 we may as well minimize with respect to ˇ 2 to

find ˇ 0 , the value of ˇ that minimizes hEi.
d hEi m! 2 „2
D C D0 (11.31)
dˇ 2 8ˇ 40 2m
so
m! 1
ˇ 20 D H) ˇ 0 D p ˛ (11.32)
2„ 2
Therefore, we know that the energy “estimate” will yield the exact value, 12 „!.
Moreover, the trial wave function does indeed reduce to the exact ground state
eigenfunction
!1=4 2 1=4
2ˇ 20 ˇ 2 x2 ˛ 2 2
.x/ D e ! 0 .x/ D e˛ x =2 (11.33)

2. Estimate the energy of the first excited state of a harmonic oscillator using the
2 2
variational method and the trial wave function .x/ D Af .x/ eˇ x where A D
constant and ˇ is an adjustable parameter. It is up to you to choose f .x/. You
may choose any function you wish, but it is strongly advised that you choose the
simplest possible function thatp
will meet the necessary conditions, in this case an
odd function. Recall that ˛ D m!=„.
Solution
The first excited state energy can be calculated using the variational method
because the eigenfunctions of the harmonic oscillator potential have definite
parity. Therefore, the ground state is even and the first excited state is odd. The
first excited state is thus the “ground state” of the odd eigenfunctions. We choose
f .x/ D x because the first excited state eigenfunction must have only one node
and f .x/ D x is the simplest function that makes the given trial wave function
odd.
We require the integral given in Eq. (G.4) which may be re-written as
Z 1
2 Œ.m C 1/ =2
xm eax dx D (11.34)
1 a.mC1/=2
First we normalize by finding A.
Z 1
2 2
jAj2 x2 e2ˇ x dx D 1
1
.3=2/
jAj2 3=2 D 1
2ˇ 2
p
1 25=2 ˇ 3
jAj2 3
D 1 H) jAj2 D p (11.35)
2 23=2 ˇ
We must first evaluate hEi.
hEi D h j HO j i
D hTi C hUi (11.36)
where j i represents the trial wave function.

We evaluate hTi and hUi separately.

Z 1
2 ˇ 2 x2 „2 d2 2 2
D
hTi jAj xe 2
xeˇ x dx (11.37)
1 2m dx
Part of the integrand is
2
d 2 2 d 2 2
2
xeˇ x D 1 2ˇ 2 x2 eˇ x
dx dx
d 2 ˇ2 x2
2 2
D 2ˇ 2 xeˇ x
2ˇ 2
x e
dx
2 2
2 2
D 2ˇ 2 xeˇ x 2ˇ 2 2x 2ˇ 2 x3 eˇ x
2 2
D 6ˇ 2 x C 4ˇ 4 x3 eˇ x (11.38)
so
Z
2
1
„2
ˇ 2 x2 2 4 3 2 2
hTi D jAj xe 6ˇ x C 4ˇ x eˇ x dx
1 2m
Z 1 2
„ „2 2 2
D jAj2 3 ˇ 2 x 2 ˇ 4 x3 xe2ˇ x dx
1 m m
2 2 Z 1 Z 1
2 „ ˇ 2 2ˇ 2 x2 2 4 2ˇ 2 x2
D jAj 3 x e dx 2ˇ x e dx
m 1 1
2 3
!
2 2
„ ˇ 6 1 2 .5=2/ 7
D jAj2 43 2 2ˇ 5=2 5
m jAj 2ˇ 2
! p
2 „2 ˇ 2 1 3 1
D jAj 3 2 2ˇ 2
m jAj 4 25=2 ˇ 5
! ! !
„2 ˇ 2 3 „2 ˇ 2 3 „2 ˇ 2
D3 D (11.39)
m 2 m 2 m
Now, hUi is given by

Z 1
1 2 2
hUi D m! 2 jAj2 x4 e2ˇ x
dx
2 1
p !
1 m! 2 2 1 m! 2 3 25=2 ˇ 3
D 5=2 .5=2/ jAj D p
2 2 25=2 ˇ 5 4
2ˇ 2
3 m! 2
D (11.40)
8 ˇ2
Then
3 „2 ˇ 2 3 m! 2
hEi D C (11.41)
2 m 8 ˇ2
To minimize hEi with respect to the parameter ˇ we differentiate Eq. (11.41)
and set this derivative to zero. Because ˇ occurs only as ˇ 2 in Eq. (11.41) we
may take the derivative with respect to ˇ 2 .
d hEi 3 „2 3 m! 2
2
D D0 (11.42)
dˇ 2m 8 ˇ 40
Solving for ˇ 0 , the value of ˇ that minimizes hEi, we have

1 m! 1
ˇ 20 D D ˛2 (11.43)
4 „ 2
Thus, the energy is given by
3 „2 1 2 3 2
hEiˇDˇ0 D ˛ C m! 2 2
2m2 8 ˛
3 „2 1 m! 3 2¯
D C m! 2
2m2 ¯ 8 m!
3 3 3
D ¯! C ¯! D ¯! (11.44)
4 4 2
which, as expected, is the exact value. With ˇ D ˇ 0 the trial wave function is of
course the exact eigenfunction of the first excited state, i.e.
s
25=2 ˇ 3 ˇ2 x2
1 .x/ D p xe

r
˛ 2 2
D p 2 Œ.˛x/ e˛ x =2 (11.45)
2
3. A particle of mass m is trapped in a potential given by
U .x/ D Bx x > 0
D1 x0 (11.46)
(a) Using the variational method estimate the ground state energy using the trial
wave function
.x/ D Axeˇx (11.47)
where A is a constant and ˇ is an adjustable parameter.

(b) The exact energy of the ground state is (see, for example, [1])
1=3
B2 „2
E1 D 2:338 (11.48)
2m
Show that the energy calculated using the variational method is greater
than the exact energy.
Solution
Note that the trial wave function was chosen so it meets the boundary condition
.0/ D lim .x/ D 0.
x!1
(a) First find the value of A that normalizes .x/. Using the integral given in
Eq. (G.2) we have
Z 1
2
jAj j .x/j2 dx D 1
0
Z 1
jAj2 x2 e2ˇx dx D 1
0
.3/
jAj2 D1
.2ˇ/3
jAj2 D 4ˇ 3 (11.49)
Next we compute hEi D h j HO j i.

Z 1
„2 d2
hEi D 4ˇ 3 xeˇx 2 xeˇx dx
2m 0 dx
Z 1
C 4ˇ 3 B x3 e2ˇx dx
0
!Z
2 3 1
2„ ˇ d2 ˇx
D xeˇx xe dx
m 0 dx2
.4/
C 4ˇ 3 B
.2ˇ/5
!
3 Z 1
2„2 ˇ d
D xeˇx Œx .ˇ/ C 1 eˇx dx
m 0 dx
3Š
C 4ˇ 3 B
.2ˇ/4
!
3 Z 1 h i
2„2 ˇ 3B
D xeˇx .ˇ/2 x ˇ ˇ eˇx dx C
m 0 2ˇ
! Z
2„2 ˇ 3 1
3B
D ˇ ˇx2 2x e2ˇx dx C
m 0 2ˇ
!
2„2 ˇ 3 2 .3/ .2/ 3B
D ˇ 3
2ˇ 2
C
m .2ˇ/ .2ˇ/ 2ˇ
!
2„2 ˇ 3 1 1 1 3B
D C
m ˇ 4 2 2ˇ
„2 ˇ 2 3B
D C (11.50)
2m 2ˇ
Now minimize hEi with respect to the parameter ˇ and set to zero.
d hEi „2 ˇ 3B
D 2 D0 (11.51)
dˇ m 2ˇ
so
3Bm
ˇ3 D (11.52)
2„2
Therefore, the estimate of the energy using this trial wave function and
the variational method is
1=3
„2 3Bm 2=3 3B 2„2
hEi D C
2m 2„2 2 3Bm
1=3
35=3 B2 „2
D 4=3 (11.53)
2 2m
(b) Comparing the variational answer hEi with the exact answer E1 as given in
Eq. (11.48) above, we have
1=3 1=3
hEi 35=3 B2 „2 1 2m
D 4=3
E1 2 2m 2:338 B2 „2
35=3 1
D Ð 1:06 (11.54)
24=3 2:338
so, indeed, the energy obtained using the variational method lies above the
exact value.
4. Estimate the ground state energy and the first excited state energy of a particle of
mass m trapped in a quartic potential given by
U .x/ D Cx4 (11.55)
Use the variational method with appropriate

p harmonic oscillator eigenfunc-
tions as the trial wave functions and use ˛ D m!=„ as the variable parameter.
Solution
The Hamiltonian for this problem is
pO 2
HO D C Cx4 (11.56)
2m
We must first compute hEi D hnj HO jni where the bras and kets are harmonic
oscillator eigenfunctions. We then minimize hEi using ˛ as the variational
parameter.
Before diving into the calculations let us map out the strategy. First, we use the
matrix elements calculated in Chap. 7 to calculate hEi. In fact, we have already
calculated these expectation values in Problems 7 and 8 of Chap. 7, which we
reproduce here.

1 3 2
hnj pO 2 jni D n C ˛ 2 „2 and hnj x4 jni D 4
2n C 2n C 1 (11.57)
2 4˛
Because the quartic potential is an even potential we can use the variational
method to calculate upper bounds to both the ground state energy and the first
excited state energy, n D 0 and n D 1. From parity considerations and the
required number of nodes we see that j0i and j1i are the appropriate trial wave
functions for the ground and first excited state, respectively.
In terms of ˛ the n D 0 and n D 1 values of hEi are
˝ q ˛ 1
En .˛/ D hnj HO jni D hnj pO 2 jni C C hnj x4 jni
2m

˛ 2 „2 1 3C
D nC C 4 2n2 C 2n C 1 (11.58)
2m 2 4˛
Checking units as we proceed we note that in Eq. (11.58) both terms on the
right-hand side have units of energy because ˛ has units m1 and C has units
J= m4 .
Minimizing Eq. (11.58) with respect to ˛ we have

dEnq .˛/ 1 ˛„2 3
D nC C 5 2n2 C 2n C 1 (11.59)
d˛ 2 m ˛
Setting the derivative equal to zero we obtain

2n2 C 2n C 1 3mC
˛ 6e D 2 (11.60)
n C 12 „
˝ ˛
where ˛ e signifies the value of ˛ that minimizes Enq .˛/ . Thus
6mC 10mC
˛ 6e .n D 0/ D 2
and ˛ 6e .n D 1/ D (11.61)
„ „2
Note that the values of ˛ e are different for the two states under consideration.
q q
We can calculate E0 .˛ e / and E1 .˛ e / by inserting ˛ e from Eq. (11.61) into
Eq. (11.58).
For n D 0 we have
" # " 2 2=3 #
q 1 „2 6mC 1=3 3C „
E0 D C
4m „2 4 6mC

1 „4=3 3C1=3 „4=3
D 2 2=3 21=3 31=3 C1=3 C 2 2=3 2=3 2=3
2 m 22 3 m

C1=3 „4=3 31=3 31=3 C1=3 „4=3 31=3 31=3
D C D C
m2=3 25=3 28=3 m2=3 25=3 25=3 2
4 1=3
C1=3 „4=3 31=3 1 C„
D 2=3 5=3
1C D
m 2 2 m2
4 1=3 4 1=3
34=3 C„ C„
D ' 0:681 (11.62)
2 25=3 m2 m2
For n D 1 we have
2 2=3
q 3 „2 10mC 1=3 15C „
E1 .˛/ D 2
C
4m „ 4 10mC
1=3 2=3
3 „2 1=3 mC 15C 1 2=3 „2
D .10/ C
4m „2 4 10 mC

3 „2 1=3 1=3 mC 1=3
D Œ2 5
4m „2
2=3 2=3 1=3

1 1 1
C5 C1=3 m1=3
5 2 „2

3 „2 mC 1=3
1=3 1=3 1=3 2=3

D 2 5 C 5 2
4 m „2
2
1=3 „ mC 1=3 21=3 C 22=3
D 35
m „2 22
1=3 4 1=3
2 C 22=3 C„
D 3 51=3 2
2 m2
4 1=3
C„
' 1:282 (11.63)
m2
1=3
Both E0 .˛/ and E1 .˛/ are proportional to C„4 =m2
q q
which has units of
energy.
1=3 " #1=3

C„4 J . J s/4
D
m2 m4 kg2
" 2 #1=3
5 kg m2
D J
s2
D J (11.64)
q q
This makes us happy. It is also comforting that E0 .˛/ < E1 .˛/.
Note that the use of the matrix elements obtained using the harmonic oscillator
ladder operators obviated the need to evaluate tedious integrals. This observation
emphasizes the general rule:
• For problems involving the harmonic oscillator first see if a solution using
the ladder operators can be effected.
11.3 Non-degenerate Perturbation Theory
Time independent perturbation theory is an often used approximation method

(especially on examinations). It is most useful when the Hamiltonian for the system
may be written as the sum of an “unperturbed” Hamiltonian for which the solution
is known and a term that is much smaller, the perturbation. We begin with notation:
HO D true Hamiltonian
HO 0 D unperturbed Hamiltonian
HO 1 D perturbing Hamiltonian.
A subscript on an energy or a wave function (state vector) signifies the state. The
superscript denotes the degree of approximation. No superscript refers to the exact
energy or wave function. Thus, for example
En D exact energy of the nth state

En.0/ D zeroth order approximation to En
En.1/ D first-order correction to En.0/
n .x/ D exact eigenfunction for the nth state (x-space)

.0/
n .x/ D first-order approximation to n .x/
.1/ .0/
n .x/ D first-order correction n .x/
Synopsis of results:
First order:
D ˇ ˇ E
.0/ ˇ O ˇ
En.1/ D n ˇ H1 ˇ .0/
n (11.65)
The first-order correction to the energy is the expectation value of the perturbing
Hamiltonian on the unperturbed state.
D ˇ ˇ E
.0/ ˇ O ˇ .0/
ˇ E X k ˇ H1 ˇ ˇ E
ˇ .1/
n ˇ .0/
ˇ n D ˇ k (11.66)
.0/ .0/
k¤n En Ek
Second order:
ˇD ˇ ˇ Eˇ2
ˇ .0/ ˇ O ˇ .0/ ˇ
Xˇ k ˇ H1 ˇ n ˇ
En.2/ D (11.67)
.0/ .0/
k¤n En Ek
Problems
1. Find the first-order correction to the energy of a hydrogen atom due to the
gravitational attraction between the proton and the electron. Assume the electron
is in a circular orbit the radius of which is the Bohr radius a0 . Find the ratio of
.0/
this correction to the unperturbed energy EnD1 D 13:6 eV.
Solution
The perturbing Hamiltonian is the gravitational potential between the electron

and the proton. Because we assume a circular orbit it is a constant and is
Gme mp
H1 .r D a0 / D (11.68)
a0
where G is the gravitational constant; me and mp are the masses of the electron
.1/
and proton, respectively. The first-order correction to the energy EnD1 is hH1 i.
Because H1 is a constant the computation is trivial.
.1/ Gme mp
EnD1 D hH1 i D
a0
Ð 1:24 1038 eV (11.69)
.0/
The ratio of this quantity to EnD1 is
.1/
EnD1 1:24 1038 eV
.0/
Ð Ð 1039 (11.70)
EnD1 13:6 eV
This very small ratio shows that gravitational attraction between the electron
and the proton is insignificant and has virtually no effect on electronic structure.
This result is compatible with that obtained in Problem 2 of Chap. 5 where it was
deduced that gravity had little effect on the binding of electrons to a nucleus.
2. A particle of mass m is trapped in an L-box potential (see Appendix M) given by
U .x/ D 0 0xL
D1 1 < x < 0I L < x < 1 (11.71)
The L-box potential is perturbed by a ı-function
HO 1 .x/ D ˛ı .x L=2/ (11.72)
where ˛ is a positive real constant. Find En.1/ , the first-order correction to the
energy for all eigenstates of the L-box (Fig. 11.2). For what values of n does En.1/
vanish and why?
Fig. 11.2 Problem 2
Solution
The eigenfunctions for an L-box are

r nx
2
n .x/ D sin 0xL
L L
D 0 1 < x < 0I L < x < 1 (11.73)
Therefore
En.1/ D h O
n j H 1 .x/ j n i
Z L nx
2
D˛ ı .x L=2/ sin2 dx
L 0 L
n
2˛
D sin2
L 2
8
<D 0 n even
2˛ (11.74)
:D n odd
L
The reason that there is no first-order correction to the energy for even values
of n is that these eigenfunctions have nodes at x D L=2. Therefore, the even
states never “feel” the perturbation because it exists only at x D L=2. En.1/ must
vanish for even n.
3. Find the first-order correction to the energy of an L-box
U .x/ D 0 0xL
D1 1 < x < 0I L < x < 1 (11.75)
for a linear perturbation of the form

U0
HO 1 .x/ D x (11.76)
L
and show the quantum state dependence of this correction. Note that this problem
is equivalent to a charged particle in the L-box with an applied electric field
(Fig. 11.3).
Fig. 11.3 Problem 3
Solution
The first order perturbation to the energy En.1/ for the nth state is
D ˇ ˇ E
.0/ ˇ O ˇ
En.1/ D n ˇ H1 ˇ .0/
n (11.77)
For an L-box the normalized eigenfunctions are

r nx
2
n .x/ D sin 0xL
L L
D0 1 < x < 0I L < x < 1 (11.78)
so
Z nx
2U0 L
x
En.1/ D sin2 dx (11.79)
L 0 L L
nx
Let y D H) x D 0 ! y D 0 and x D L ! y D n (11.80)
L
Then
Z y Z n
2U0 n
L 2U0
En.1/ D 2
sin y dy D 2 2 y sin2 ydy (11.81)
L 0 n n n 0
To perform this integral we use that given in Eq. (G.9) so

n
2U0 y2 y sin .2y/ cos .2y/
En.1/ D 2 2 (11.82)
n 4 4 8 0
The second term vanishes at both limits and the third term is the same at both
limits so we have
2U0 .n/2 1
En.1/ D D U0 (11.83)
n2 2 4 2
and the answer is independent of the quantum state. We could have saved
ourselves the work of computing this integral with a bit of forethought. Let us
examine the integral in Eq. (11.77) and re-write it as follows.
D ˇ ˇ E
.0/ ˇ O ˇ
En.1/ D n ˇ H1 ˇ .0/
n
D ˇ ˇ E
U0 .0/ ˇ ˇ .0/
D n ˇxˇ n
L

U0
D hxi
L

U0 L
D
L 2
U0
D (11.84)
2
This calculation was facilitated because it is obvious that hxi, the average
value of x for an L-box, is L/2. Both approaches were presented to emphasize
that forethought about the problem can often save a lot of work.
4. A particle of mass m is confined to an L-box. The width of the box L is such that
the rest mass energy mc2 >> En.0/ for low values of n.
(a) Find En.1/ the first-order correction to the energy due to the relativistic kinetic
energy of the particle. Write En.1/ in terms of En.0/ , the nonrelativistic energy
of the nth level of the L-box. (The total relativistic energy of a particle of
mass m having momentum pO is given by E2 D pO 2 c2 C m2 c4 .)
(b) Show that for a particle-in-a-box the nonrelativistic energy eigenfunctions
are also eigenfunctions of the relativistic Hamiltonian.
(c) Find the exact relativistic energy and show that for mc2 >> En.0/ the solution
reduces to the perturbation theory result.
Solution
(a) Inside the box the Hamiltonian is the kinetic energy. The relativistic kinetic
energy is the total energy minus the rest energy mc2 . Therefore the relativistic
Hamiltonian for a particle-in-a-box is
p
HO D pO 2 c2 C m2 c4 mc2
1=2
pO 2
D mc2 1 C 2 2 mc2
mc
pO 2 1 pO 4
Ð C (11.85)
2m 8 m3 c2
The first term in Eq. (11.85) is the nonrelativistic kinetic energy, which
we may take to be the unperturbed Hamiltonian HO 0 . We therefore take the
second term to be the perturbation Hamiltonian HO 1 . Before jumping in and
performing an integration that might be avoided we stop and examine HO 1 and
note that it can be written in terms of HO 0 as follows.
1 pO 4
HO 1 D
8 m3 c2
1 2
D 3 2 2mHO 0 (11.86)
8m c
We have thus managed to write HO 1 in terms of the Hamiltonian HO 0 which
is a great simplification because the unperturbed eigenfunctions are now
eigenfunctions of HO 1 . This will save time in the calculation and simplify
the computation, thus minimizing the possibility of error.
Designating the unperturbed kets by jni the first-order correction to the
energy is
En.1/ D hnj HO 1 jni

1
D hnj HO 20 jni
2mc2
1 .0/ 2
D E (11.87)
2mc2 n
where
n2 2 „2
En.0/ D (11.88)
2mL2
Alternative Method (Brute Force)
We use direct integration with the spatial eigenfunctions to obtain En.1/ , in

contrast to the simple method used above.
En.1/ D hnj HO 1 jni

Z nx nx
1 1 2 L
D sin pO 4 sin dx
8 m3 c2 L 0 L L
1 1 nx „ d 4
Z L nx
D sin sin dx
4 m3 c2 L
0 L i dx L
Z
1 „4 n 4 L 2 nx
D sin dx (11.89)
4 m3 c2 L L 0 L
To evaluate the integral we make the substitutions
y D .n=L/ x H) dx D .L=n/ dy
x D 0 ! y D 0I x D L ! y D n (11.90)
which lead to
Z n
1 „4 n 4 L
En.1/ D sin2 ydy
4 m3 c2 L L n 0

1 „4 n 4 L n
D
4 m3 c2 L L n 2
2 2 2 2
1 1 n„
D 2

2 mc 2mL2
1 .0/ 2
D E as above (11.91)
2mc2 n
(b) Rewriting the radical in the relativistic Hamiltonian, Eq. (11.85), provides a
polynomial in the operator pO 2 as shown below.
( 1=2 )
pO 2
HO D mc2 1C 2 2 mc2
mc
( 1=2 )
2 2 pO 2 2
D mc 1C mc (11.92)
mc2 2m
When the radical is expanded using the binomial theorem an infinite

series in powers of pO 2 will be obtained. Inasmuch as the eigenfunctions of
the unperturbed L-box, which we designate jni, are eigenfunctions of any

operator that is proportional to pO 2 they will also be eigenfunctions of the
relativistic Hamiltonian.
(c) To find the values of the relativistic Hamiltonian we simply operate on the
eigenfunctions jni with the relativistic Hamiltonian. This procedure amounts
to expanding the radical, operating on jni and then contracting the infinite
series back to an expression within a radical. This is tantamount to replacing
pO 2 =2m in Eq. (11.85) by En.0/ .
( 1=2 )
pO 2
HO jni D mc 2
1C 2 2 mc 2
jni
mc
( 1=2 )
2 2 pO 2 2
D mc 1C mc jni
mc2 2m
( 1=2 )
2
D mc2 1 C En.0/ mc2 jni (11.93)
mc2
Therefore the exact relativistic energy eigenvalues are

1=2
2
En.rel/ D mc 2
1C En.0/ mc2 (11.94)
mc2
which, when expanded becomes

( 2 )
1 2 1 2
En.rel/ 2
mc 1 C .0/
En .0/
En mc2
2 mc2 8 mc2

.0/ 1 22 .0/ 2
Ð En En C
8 mc2

1 1 .0/ 2
D En.0/ En C (11.95)
2 mc2
Not surprisingly, this produces the same result for En.1/ as that obtained in
Eqs. (11.87) and (11.91).
5. An electron is subjected to a harmonic oscillator potential, U .x/ given by
1 2 1
U .x/ D kx D m! 2 x2 (11.96)
2 2
A constant electric field F is applied in the x direction.
(a) Determine the first nonvanishing correction to the harmonic oscillator energy
using perturbation theory. Find the polarizability [1].
(b) Solve the problem exactly and show that the exact solution yields the result
from perturbation theory.
Solution
(a) Treating the electric field as a perturbation we have
HO 1 D eFx (11.97)
Note that for this potential energy the negative of the derivative yields a
field in the x direction.
The first-order correction to the energy is given by
En.1/ D hnj HO 1 jni D eF hnj x jni (11.98)
Clearly En.1/ 0 because the integrand in Eq. (11.98) is odd over

symmetric limits. We must go to the second-order correction which is given
by
ˇ ˇ2
ˇ ˇ
X ˇhmj HO 1 jniˇ
En.2/ D
.0/ .0/
m¤n En Em
X jeF hmj x jnij2

D (11.99)
.0/ .0/
m¤n En Em
Next we evaluate the matrix elements eF hmj x jni using the ladder operators.
The action of the ladder operators is given by
p
aO jni D n jn 1i

p
aO jni D n C 1 jn C 1i (11.100)
Also, see Eq. (7.46),

r
1 m!
xO D p aO C aO ; ˛ D (11.101)
2˛ „
Therefore
eF p p
eF hmj x jni D p nı m;n1 C n C 1ı m;nC1 (11.102)
2˛
so that only two terms in Eq. (11.99) survive the summation. We have
X jeF hmj x jnij2
En.2/ D
.0/ .0/
m¤n En Em

eF 2 n nC1
D p C
2˛ .En En1 / .En EnC1 /
2 2
e F „ n nC1
D C
2 m! „! „!
e2 F 2
D (11.103)
2m! 2
Note that this correction to the energy is independent of n so the same
correction applies to all levels. The polarizability is
d2 E e2
˛D D (11.104)
dF 2 m! 2
(b) To solve the problem exactly we note that the total potential energy
may be rewritten by completing the square. It is still a parabola and the
eigenfunctions and eigenvalues remain those of a harmonic oscillator. The
total potential energy is
1
UT .x/ D m! 2 x2 C eFx (11.105)
2
which, after completing the square, is

1 eF 2 e2 F 2
UT .x/ D m! 2 x C 2
(11.106)
2 m! 2m! 2
The minimum in the potential energy with the field on occurs at

dUT .x/ eF
D m! 2 xmin C D0
dx m! 2
eF
H) xmin D (11.107)
m! 2
so the minimum value of UT .x/ is

1 eF eF 2 e2 F 2
UT .x D xmin / D m! 2 2
C 2

2 m! m! 2m! 2
e2 F 2
D (11.108)
2m! 2
Figure 11.4 shows the potential energy with and without the electric field on.
Fig. 11.4
F=0
F=0 -n=0
xmin
-n=0 x
e2F2
2mw2
The shapes of both parabolas are the same as is easily seen by taking
the second derivative of Eq. (11.105). Therefore the energy level spacing is
the
2 same
with the
field on or off. The total energy is merely displaced by
e F 2 = 2m! 2 . The energy eigenvalues are therefore

1 e2 F 2
En D n C „! (11.109)
2 2m! 2
We can show this directly from the TISE using the same substitutions as
above. The TISE is

„2 d2
C UT .x/ .x/ D E .x/ (11.110)
2m dx2

Letting y D x C eF= m! 2 we have

„2 d2 1 2 2 e2 F 2
C m! y .y/ D E .y/ (11.111)
2m dy2 2 2m! 2
or

„2 d2 1 e2 F 2
C m! 2 y2 .y/ D E C .y/ (11.112)
2m dy2 2 2m! 2
which is the TISE with energy eigenvalues given by Eq. (11.109).

e2 F 2 1
En C D n C „!
2m! 2 2

1 e2 F 2
H) En D n C „! (11.113)
2 2m! 2
Note that the second-order correction to the energy derived in part (a) of
this problem is the exact energy added to the system when the electric field
is turned on. In this special case the perturbation treatment of the problem
provides the exact answer so the magnitude of the “perturbation” F is not
restricted to small values.
6. A H-atom is subjected to a repulsive ı-function potential at the origin of

.1/
coordinates so that HO 1 D U0 ı .r/. Find En00 , the first-order correction to the
energy of any zero-angular momentum state. [Hint: Rn0 .r D 0/ D 2 .1=na0 /3=2 .]
What is the first-order correction to the energy if ` ¤ 0?
Solution
The first-order correction to the energy of any state jn00i is
.1/
En00 D hn00j HO 1 jn00i (11.114)
.1/
Because the perturbation is a ı-function it is easy to evaluate En00 using the
corresponding integral. We require n00 .r/, which is
n00 .r/ D Y00 . ; / Rn0 .r/ (11.115)

p
Using the hint and the fact that Y00 . ; / D 1= 4 (see Appendix R) we
have
r
1 2
n00 .r D 0/ D (11.116)
4 .na0 /3=2
so that
Z
.1/
En00 D j n00 .r/j2 ŒU0 ı .r/ d3 r
all space
D U0 j n00 .r D 0/j2
U0
D (11.117)
n3 a30
We note that the integration in Eq. (11.117) can be made more transparent by
.1/
evaluating En00 in Cartesian coordinates and using ı .r/ D ı .x/ ı .y/ ı .z/.
Z
.1/
En00 D j n00 .x; y; z/j2 ŒU0 ı .x/ ı .y/ ı .z/ dxdydz
all space
D U0 j n00 .0; 0; 0/j2

U0
D (11.118)
n3 a30
The first-order correction to the energy for ` ¤ 0 vanishes because the

eigenfunctions for ` ¤ 0 are all zero at the origin. Although there is a quantum
mechanical derivation of this fact [1] it is easy to remember by appealing to the
classical definition of angular momentum L D r p. It is clear that if L ¤ 0 then
r cannot vanish.
7. Two H-atoms, each of mass mp (ignore the electronic masses), are joined by
a covalent bond to form the H2 molecule. Small amplitude vibration of this
molecule may be treated as harmonic oscillation with a force constant k D ! 2
where is the reduced mass of the nuclei and ! is the frequency of the vibration.
The force constant k is obtained from the coefficient of the quadratic term in an
expansion of the intermolecular potential.
.1/ .0/
(a) Find E0 the first-order correction to the ground vibrational state energy E0
due to the relativistic motion. ˇ ˇ
.1/ .0/ ˇ .1/ .0/ ˇ
(b) Write E0 in terms of E0 to obtain the approximate ratio ˇE0 =E0 ˇ for the
ground electronic state of the H2 molecule in order to evaluate the validity
of the use of perturbation theory. The energy separation between adjacent
vibrational levels of the electronic ground state is „! 0:54 eV and mp c2 is
Ï 940 MeV.
Solution
(a) From Problem 4 of this chapter we know that the relativistic expansion of TO
the kinetic energy operator is [see Eq. (11.85)].
pO 2 1 pO 4
TO Ð C (11.119)
2 8 3 c2
p
Letting ! D k=m the complete Hamiltonian is therefore
HO D TO C VO
pO 2 1 pO 4 1
D 3 2
C C ! 2
2m 8 c 2
1 pO 4
HO 0 (11.120)
8 3 c2
where HO 0 is the unperturbed harmonic oscillator Hamiltonian. The correction

to the unperturbed vibrational ground state energy is simply h0j HO 1 j0i where
1 pO 4
HO 1 D (11.121)
8 3 c2
Now the question is what is the best (read: easiest) way to calculate
h0j HO 1 j0i? Inasmuch as the commonly used eigenfunctions are those in
coordinate space we can either convert them to momentum space or convert
the momentum operator to its equivalent coordinate space operator. This con-
version would then require four derivatives of the ground state eigenfunction.
While neither method is difficult, they are both tedious enough that an error
is likely. The most efficient approach is through the use of ladder operator as
has been emphasized in Chap. 7.
Now, using Eq. (7.16) HO 1 may be written in terms of the ladder operators
4
1 1 ˛„
HO 1 D p aO aO (11.122)
8 3 c2 2
.1/
and E0 D h0j HO 1 j0i may be written [see Eq. (7.16)] as
.1/
E0 D h0j HO 1 j0i
1 ˛ 4 „4
4
D h0j O
a O
a j0i
83 c2 4
˛ 4 „4
2

D 3 2
h0j aO aO aO aO aO aO j0i (11.123)
32 c
where the terms in square brackets have been written in a way to facilitate
evaluation. We first evaluate the terms in square brackets using Eqs. (7.28)
and (7.29) which, for convenience, we reproduce here.
p p
aO jni D n jn 1i and aO jni D n C 1 jn C 1i (11.124)
Using these raising and lowering operations we have

h0j aO aO D h1j and aO aO j0i D j1i (11.125)
where we have used the fact that aO and aO are Hermitian adjoints so that
operating on a ket with aO is equivalent to operating with aO on a bra. To
arrive at Eq. (11.125) we have also used the fact that applying the lowering
operator to j0i annihilates it, that is aO j0i 0.
Continuing, we have
.1/ ˛ 4 „4

E0 D 3 2
h1j aO aO aO aO j1i
32 c
and

p
h1j aO aO D 2 h2j h0j

p
aO aO j1i D j0i 2 j2i (11.126)
and
.1/ ˛ 4 „4 p p
E0 D 2 h2j h0j j0i 2 j2i
323 c2
˛ 4 „4
D .2 1/
323 c2
r 4
3˛ 4 „4 3„4 m!
D 3 2
D
32 c 323 c2 „
3„2 ! 2
D (11.127)
32c2
.1/
Note that E0 has units of energy as it must.
(b) The ground state vibrational energy of the hydrogen
ˇ molecule,
ˇ the zero-point
.0/ 1 ˇ .1/ .0/ ˇ
energy, is E0 D 2 „! 0:27 eV so the ratio ˇE0 =E0 ˇ is
ˇ ˇ
ˇ E.1/ ˇ 12 12 „!
ˇ 0 ˇ
ˇ .0/ ˇ D
ˇE ˇ 32 .mc2 /
0
3 .0:27/ eV
D
8 .4:7 108 / eV
D 2:16 1010 (11.128)
This is a very small number so the use of perturbation theory is justified.
8. Find the first-order correction to the energy levels of the H-atom due to
relativistic motion of the electron.
Solution
The relativistic energy is
E2 D pO 2 c2 C m2 c4
pO 2 1 pO 4
EÐ (11.129)
2m 8 m3 c2
so the relativistic hydrogen atom Hamiltonian to first order is

pO 2 1 pO 4 1 e2
HO D
2m 8 m3 c2 4 0 r
D HO 0 C HO 1 (11.130)
where

pO 2 1 e2 1 pO 4
HO 0 D (11.131)
2m 4 0 r 8 m3 c2
and
2
1 pO 4 1 pO 2
HO 1 D 3 2
D (11.132)
8m c 2mc2 2m
Because pO 2 =2m is the nonrelativistic kinetic energy we may write

2
1 pO 2
HO 1 D
2mc2 2m
2 2
1 O0 C 1 e
D 2
H
2mc 4 0 r
( 2 2 2 2 2 ) 2
1 O 2 O 1 e e 1 e
D 2
H 0 C 2H 0 C
2mc 4 0 r r 4 0 r
(11.133)
The first-order correction to the nonrelativistic energy is the expectation value

of HO 1 , Eq. (11.65) so
8 ! !2 9
<
1 .0/
2 e2 1 e2 1 =
hn`mj HO 1 jn`mi D E C 2En.0/ C
2mc2 : n
4 0 r 4 0 r2 ;
(11.134)
Putting Eq. (11.134) in terms of the fine structure constant ˛ and using Table T.3
we have
D E 2
.1/ O 1 .0/ .0/ 1 2 1
En D H 1 D En C 2En „c˛ C .„c˛/
2mc2 r r2
8 2 2 !2 2 4 4 !
1 < mc ˛ mc ˛ 1
D 2
2mc 2 : 2n 2 2n2 n2
" #)
2 4 4
1
C mc˛
n ` C 12
3
( " #)
m2 c4 ˛ 4 1 2 1 1
D 2 C 3
2mc2 2n n4 n ` C 12
( " #)
mc2 ˛ 4 1 2 n
D 1C
2n4 2 ` C 12
2 2 ( " #)
mc ˛ =2n2 ˛ 2 1 n
D 1C
2n2 4 ` C 12
" !#
2
˛ 3 n
D En.0/ 2 (11.135)
n 4 ` C 12
The ratio of the energy correction to the unperturbed energy is

" !#
En.1/ ˛2 3 n
.0/
D 2 1
En n 4 `C 2
" !#
5 105 3 n
Ð 1
(11.136)
n2 4 `C 2
This relativistic correction to the hydrogen atom energies, termed “fine structure”
correction, is proportional to ˛ 2 and is small. There are additional corrections
such as that resulting from the coupling between the orbital magnetic moment
and the intrinsic magnetic moment of the electron (the spin). These various
corrections account for the splitting of spectral lines observed in the spectrum
of atomic hydrogen.
9. A harmonic oscillator is perturbed by a potential H1 D Cx4 where C is a positive

constant.
(a) Find the first-order correction to the ground state energy of the harmonic
oscillator.
(b) If the quartic perturbation is replaced by a cubic perturbation, H1 D Bx3 (B is
a constant) find the ground state energy correction to first order. [Hint: Think
about symmetry and you should be able to find the answer without writing
anything down.]
Solution
(a) We write the Hamiltonian as
pO 2 1
HO D C m! 2 x2 C Cx4 (11.137)
2m 2
The first-order correction to the ground state (n D 0) is
.1/
EnD0 D h0j HO 1 j0i
D C h0j x4 j0i (11.138)
The integral in Eq. (11.138) has already been computed using the har-
monic oscillator ladder operators (see Problem 8 of Chap. 7, Eq. (7.84)). For
an arbitrary state jni the expectation value of x4 is
3 2
hnj xO 4 jni D 4
2n C 2n C 1 (11.139)
4˛
so the first-order correction to the energy for a quartic perturbation to a
harmonic oscillator is
.1/ 3C
EnD0 D (11.140)
4˛ 4
(b) The cubic perturbation is an odd function, so hnj xO 3 jni vanishes for any n, and
the first-order correction is zero. This is important because a real diatomic
potential function can be expanded around its equilibrium position in a
Taylor series, with the first term being a quadratic, the second a cubic, and so
forth. The vanishing of the first-order cubic correction enhances the utility of
the quadratic (harmonic oscillator) approximation for low vibrational levels.
This utility decreases for higher vibrational levels where the second order
cubic and first-order quadratic corrections become important.
10. A matrix that has been constructed on an orthonormal basis set and that
represents the Hamiltonian of a particular system is given by
0 1
1 0
O
HD @ 2 0 A (11.141)
0 0 3
(a) Write this matrix as the sum of two matrices HO D HO 0 C HO 1 for the purpose
of applying perturbation theory to approximate the eigenvalues of H. O
(b) Find the eigenvalues to second order.
(c) Solve the problem exactly and compare with the perturbation theory result.
Solution
(a) Splitting the matrix as suggested we have

0 1 0 1
100 0 0
HO D @ 0 2 0 A C @ 0 0 A (11.142)
003 0 0
where
0 1 0 1
100 0 0
HO 0 D @ 0 2 0 A and HO 1 D @ 0 0 A (11.143)
003 0 0
Notice that we could have written the HO 0 matrix with 3 as the entry
in the lower right-hand corner in which case the HO 1 matrix would have a
zero in the lower right-hand corner and we could have reduced the problem
to working with 2 2 matrices. We will, however, continue this solution the
hard way.
(b) The unperturbed eigenkets are the eigenkets of HO 0 , namely
0 1 0 1 0 1
ˇ E 1 ˇ E 0 ˇ E 0
ˇ .0/ ˇ ˇ
ˇ1 D @ 0 A ; ˇ2.0/ D @ 1 A ; ˇ3.0/ D @ 0 A (11.144)
0 0 1
and the unperturbed eigenvalues are
.0/ .0/ .0/
E1 D 1 ; E2 D 2 ; E3 D 3 (11.145)
First order:
D ˇ ˇ E
.1/ ˇ ˇ
Ei D i.0/ ˇ HO 1 ˇi.0/ i D 1; 2; 3 (11.146)
so
0 10 1
0 0 1
.1/ @ A @
E1 D 100 0 0 0A D 0
0 0 0
0 10 1
0 0 0
.1/
E2 D 010 @ 0 0 A@1A D 0
0 0 0
0 10 1
0 0 0
.1/ @ A @
E3 D 001 0 0 0 A D (11.147)
0 0 1
Second order:
ˇ˝ ˇ ˇ ˛ˇ2
ˇ ˇ
.2/
X ˇ k.0/ ˇ HO 1 ˇi.0/ ˇ
Ei D (11.148)
.0/ .0/
k¤i Ei Ek
Then
ˇ˝ ˇ ˇ ˇ˝ ˇ ˇ
ˇ .0/ ˇ O ˇˇ .0/ ˛ˇ2 ˇ .0/ ˇ O ˇˇ .0/ ˛ˇ2
.2/
ˇ 2 H1 1 ˇ ˇ 3 H1 1 ˇ
E1 D C (11.149)
.0/ .0/ .0/ .0/
E1 E2 E1 E3
Digress to evaluate all three off-diagonal matrix elements. (Only three are
needed because HO is Hermitian.)
0 10 1
D ˇ ˇ E 0 0 1
ˇ ˇ
2.0/ ˇ HO 1 ˇ1.0/ D 0 1 0 @ 0 0 A@0A D
0 0 0
0 10 1
D ˇ ˇ E 0 0 1
ˇ ˇ
3.0/ ˇ HO 1 ˇ1.0/ D 0 0 1 @ 0 0 A@0A D 0
0 0 0
0 10 1
D ˇ ˇ E 0 0 0
ˇ ˇ
3.0/ ˇ HO 1 ˇ2.0/ D 0 0 1 @ 0 0 A@1A D 0 (11.150)
0 0 0
Now back to second order.
.2/ 2 0
E1 D C D 2 (11.151)
.1 2/ .1 3/
ˇ˝ ˇ ˇ ˛ˇ2 ˇ˝ ˇ ˇ
ˇ .0/ ˇ O ˇ .0/ ˇ ˇ .0/ ˇ O ˇˇ .0/ ˛ˇ2
.2/
ˇ 1 H1 2 ˇ ˇ 3 H1 2 ˇ
E2 D C
.0/ .0/ .0/ .0/
E2 E1 E2 E3
2 0
D C D 2 (11.152)
.2 1/ .2 3/
ˇ˝ ˇ ˇ ˛ˇ2 ˇ˝ ˇ ˇ
ˇ .0/ ˇ O ˇ .0/ ˇ ˇ .0/ ˇ O ˇˇ .0/ ˛ˇ2
.2/
ˇ 1 H1 3 ˇ ˇ 2 H1 3 ˇ
E3 D C
.0/ .0/ .0/ .0/
E3 E1 E3 E2
0 0
D C D0 (11.153)
.3 1/ .3 2/
The energies correct to second order are therefore

.0/ .1/ .2/
Ei D Ei C Ei C Ei
E1 D 1 C 0 2 D 1 2
E2 D 2 C 0 C 2 D 2 C 2
E3 D 3 C 0 D 3 (11.154)
(c) To solve exactly we solve the secular equation for the whole Hamiltonian.
ˇ ˇ
ˇ 1E 0 ˇ
ˇ ˇ
ˇ 2E 0 ˇD0 (11.155)
ˇ ˇ
ˇ 0 0 3 E ˇ
Expanding along the bottom row we have

.3 E/ .1 E/ .2 E/ 2 D 0

or .3 E/ 2 3E C E2 2 D 0 (11.156)
Clearly one root is
E D3 (11.157)
The other two are given by solving the quadratic equation

E2 3E C 2 2 D 0
and are
1 p
ED 3 ˙ 9 4 .2 2 /
2
1 p
D 3 ˙ 1 C 4 2 (11.158)
2
Now the root E D 3 is equal to E3 from above. This is expected
because it is the only entry in the last row and column of H. O To compare
the other two roots of the quadratic equation we expand the radical using the
binomial expansion.
1 p
ED 3 ˙ 1 C 4 2
2

3 1 1
Ð ˙ 1 C 4 2 C
2 2 2

D 2 C 2 ; 1 2 (11.159)
so the exact result reduces to the second-order perturbation theory result for
small .
11.4 The Helium Atom
Although it looks deceptively simple, the Schrödinger equation for this two-electron
system cannot be solved analytically. The helium atom is, however, amenable to
treatment using one or more of the approximation methods discussed in this chapter.
It is also a system for which indistinguishability of the two electrons plays an
important role. The Hamiltonian in atomic units (see Appendix C) for a two electron
system having Z protons in the nucleus is
p21 p2 Z Z 1
HD C 2 C (11.160)
2me 2me r1 r2 r12
where r1 and r2 are the distances of electrons 1 and 2 from the nucleus; r12 D
jr1 r2 j is the distance between electrons 1 and 2. If the r12 term were absent,
the Hamiltonian discussed in chapter would represent two independent hydrogenic
atoms with nuclear charge Z (Z D 2 for He). The eigenfunctions would be products
of H-atom eigenfunctions scaled by Z, and, using the notation of Sect. 11.3 the
zeroth-order energies would be
Z2 Z2
En.0/ D (11.161)
1 n2
2n1 2n2
where n1 , n2 are the principal quantum numbers of electrons 1 and 2 (we have
ignored indistinguishability here, but including it would not change the allowed
energies). For Z D 2, the ground state energy neglecting the r12 term would be
.0/
E11 D 4 a.u. or about 109 eV. The observed energy for the He ground state is
Eex D 78:8 eV. Introducing the r12 electron repulsion term obviously produces a
significant positive correction to the energy. Utilization of the approximation meth-
ods discussed above allows considerable improvement to the calculated energies
and wave functions for He, or any other two-electron atom, such as H or LiC .
Problems
1. Evaluate the ground state energy for He using first-order perturbation theory.
Solution
First we must construct the proper zero-order ground state wave function. This
function is simply the product of two ground state hydrogenic wave functions as
mentioned above. The spatial part of this wave function in ket notation is
jn1 `1 m1; n2 `2 m2 i D jn1 `1 m1 i jn2 `2 m2 i

j100; 100i D j100i j100i (11.162)
where n D 1 and ` D m D 0 for electrons 1 and 2. This ket is symmetric when

the coordinates for electrons 1 and 2 are interchanged, so the proper spin function
must be antisymmetric. As in Problem 2 of Chap. 9, we designate this spin state
j00i, and from Problem 13 of Chap. 8 the spin zero coupled state j00i in terms of
the uncoupled states is
1 ˇ ˛ ˇ ˛
j00i D p ˇ 12 ; 12 ˇ 12 ; 12 (11.163)
2
The complete ground state ket, symmetric in space and antisymmetric in spin,
is simply j11 01 01 12 02 02 i j00i and the first-order correction to the zeroth-order
ground state energy is
.1/ 1
E11 D h11 01 01 12 02 02 j j11 01 01 12 02 02 i (11.164)
r12
The nonrelativistic Hamiltonian, Eq. (11.160), does not contain spin variables,
so the inner product of the spin bra and ket is unity, and thus does not affect the
.1/ .1/
calculated E11 value. The actual evaluation of E11 requires evaluation of the
integral that is the inner product in Eq. (11.164). This becomes
Z
.1/ 22 Z 6 1
E11 D exp Œ2Z .r1 C r2 /
.4/2 all space r12

.4/ r12 sin 1 d 1 dr1 r22 sin 2 d 2 dr2 (11.165)
This integral can be evaluated by expanding 1=r12 in terms of Legendre

polynomials as shown in Appendix P. The result after some algebra is
.1/ 5
E11 D Z
8
5
D a.u. for He (11.166)
4
so the first-order correction to the total He energy is
.0/ .1/
E11 C E11 D 4 a.u. C 1:25 a.u.
D 2:75 a.u. D 74:8 eV (11.167)
This is an approximation to the total amount of energy required to liberate

.1/
both electrons from the Coulomb attraction of the nucleus. While E11 due to the
.0/
electron–electron repulsion is a significant fraction of E11 , in excess of 30 %, the
answer is remarkably close to the actual energy, 78:95, when one considers that
it was evaluated using perturbation theory.
2. Use the perturbation result from Problem 1 along with the variational method to
improve the perturbation result for the ground state energy of a He atom. Use
Z as the variational parameter; that is, assume that the He wave function can be
improved by using a nuclear charge less than or greater than Z D 2.
Solution
We begin by replacing Z by & in the two-particle unperturbed hydrogenic radial

wave function (Table T.2), so that (in a.u.)
&3
j100i j100i ! exp Œ& .r1 r2 / (11.168)

We must find the value of & that minimizes the calculated first-order pertur-
bation theory energy as obtained in Problem 1. Although we have introduced a
variable & in the wave function, the Hamiltonian still contains the actual Z, so
we simplify the problem by rewriting the Hamiltonian, Eq. (11.160), as

p21 p2 & & 1 1 1
hHi D C 2 C C .& Z/ C (11.169)
2me 2me r1 r2 r12 r1 r2
Neglecting the term in brackets, this Hamiltonian is identical to that in

Problem 1 with Z ! & so, using Eq. (11.161), we can immediately write

5 1
hHi D & 2 & C 2 .& Z/
8 r
5
D & 2 & C 2 .& Z/ & (11.170)
8
where we have used the expectation value of 1=r (see Table T.3). Differentiating
with respect to the adjustable parameter & and setting the result equal to zero, we
obtain & D & 0 the value of & that minimizes hHi.
5
&0 D Z (11.171)
16
Inserting this into Eq. (11.170) for hHi, the variational energy E.v/ .Z/ for
ground state of He becomes
2
.v/ 5
E .Z/ D Z in a.u. (11.172)
16
which for Z D 2 is
E.v/ .Z D 2/ D 2:85 in a.u.

D 77:5 eV (11.173)
This result is fairly close to the experimental value of 78:8 eV, but remains
above it as guaranteed by the variational method. Indeed, it is considerably closer
to this value than the 74:8 eV obtained using only perturbation theory as in
Problem 1. The fact that the optimum value of Z (& 0 ) is less than Z D 2 reflects
the screening each electron provides the other, reducing the effective nuclear
charge.
As might be evident from the discussion of the variational method in
Sect. 11.2 the total energy may be calculated with progressively increasing
accuracy as we add more variational parameters.
11.5 Degenerate Perturbation Theory
Examination of Eqs. (11.66) and (11.67) reveals a serious problem if one or more of
the unperturbed levels
ˇ Eare degenerate.
ˇ E Using the same notation as that in Sect. 11.3
ˇ .0/ ˇ .0/ .0/
it is seen that if ˇ n and ˇ k are degenerate then En.0/ D Ek and the first-
order corrections to the unperturbed wave functions blows up. This problem can be
solved by constructing linear combinations of the degenerate eigenfuntions that are
not only eigenfunctions of HO 0 , but also eigenfunctions of the perturbing Hamiltonian
ˇ E
ˇ .0/
HO 1 . We call this new set of eigenfunctions the select set and designate them ˇ i .
Thus
ˇ E X
q ˇ E
ˇ .0/ ˇ .0/
ˇ i D cij ˇ j (11.174)
jD1
is a select ket such that
ˇ E ˇ E
ˇ .0/ .1/ ˇ .0/
HO 1 ˇ i D Ei ˇ i (11.175)
.1/
where Ei is the first-order correction to the energies of each of the degenerate
states. If there are q degenerate eigenstates of HO 0 , there will be q select states. The
task then is to find the expansion coefficients cij in Eq. (11.174).
After applying first-order time independent perturbation theory [1] we arrive at
the following relationship
X
q D ˇ ˇ E
.0/ ˇ O ˇ .0/ .1/
cij k ˇ H1 ˇ j D Ei cik kq (11.176)
jD1
D ˇ ˇ E
.0/ ˇ ˇ .0/
where k ˇ and ˇ represent degenerate (unperturbed) states of HO 0 . Equa-
j
tion (11.176) provides the matrix elements for O

ˇ theE matrix representation of H 1 .
ˇ .0/
Because the degenerate unperturbed eigenkets ˇ i are not eigenkets of H 1 the HO 1O
ˇ E
ˇ .0/
matrix is not diagonal on this basis set. It is, however, diagonal on the ˇ j basis set,
which is the reason for constructing this select set. Note that the matrix representing
the unperturbed Hamiltonian, HO 0 , is diagonal on both basis sets. Problem 1 below
illustrates this procedure in a step-by-step fashion.
Problems
1. A particle of mass m is trapped in a two-dimensional a-box and a perturbation

HO 1 .x; y/ D Kxy is applied.
(a) Find the eigenfunctions and eigenvalues for the unperturbed ground state and
for the doubly degenerate first excited state.
(b) Find the matrix that represents the complete Hamiltonian HO for the excited
states and find the energies of the ground state and the first excited states to
first order.
Solution
(a) The unperturbed Hamiltonian is separable in x and y so
HO 0 .x; y/ D HO x .x/ C HO y .y/ (11.177)
When the Hamiltonian can be separated as in Eq. (11.177) [1] the eigen-
functions are products of the eigenfunctions of the individual Hamiltonians
and the energies are the sums of the individual energy eigenvalues. The
eigenfunctions for a one-dimensional a-box are (see Appendix M):
r nx
2 a a
n .x/ D cos ; x n D 1; 3; 5 : : :
a a 2 2
r nx
2 a a
n .x/ D sin ; x n D 2; 4; 6 : : :
a a 2 2
D 0 ; 1 < x < a=2 ; a=2 < x < 1 for all n (11.178)
Therefore
.0/ .0/ .0/
11 .x; y/ D .x/ 1 .y/
1
2 x y
D cos cos (11.179)
a a a
One of the degenerate first excited states is
.0/
12 .x; y/ D .x/ 2 .y/
1
2 x 2y
D cos sin (11.180)
a a a
while the other is

.0/
21 .x; y/ D .x/ 1 .y/
2
y
2 2x
D cos sin (11.181)
a a a
The unperturbed energies of these states are (Appendix M)
.0/ 12 2 „2
E11 D 2
2mL2
.0/ 2 „2 22 2 „2 5 2 „2 .0/
E12 D C D D E21 (11.182)
2mL2 2mL2 2mL2
.0/ .0/
Therefore, 12 .x; y/ D j12i and 21 .x; y/ D j21i are degenerate states
because their (unperturbed) energies are the same.
(b) The first-order correction to the ground state energy is clearly zero because
the perturbation is odd in both x and y.
For the excited states the Hamiltonian in matrix form is
HO D HO 0 C HO 1

h12j HO 0 j12i h21j HO 0 j12i
D
h12j HO 0 j21i h21j HO 0 j21i

h12j xy j12i h21j xy j12i
K (11.183)
h12j xy j21i h21j xy j21i
.0/
The diagonal matrix elements of HO 0 are simply E12 while the off-diagonal
matrix elements of HO 0 vanish. Therefore
!
.0/
E 0 h12j xy j12i h21j xy j12i
HO D 12
.0/ K (11.184)
0 E12 h12j xy j21i h21j xy j21i
Because HO 0 is represented by a diagonal matrix the unperturbed kets must

be the column matrices

1 0
j12i D and j21i D (11.185)
0 1
Symmetry considerations dictate that the diagonal matrix elements in the
second matrix of Eq. (11.184) vanish because the perturbation Kxy is an
odd function. We must, however, evaluate the off-diagonal elements.
h12j xy j21i D h21j xy j12i

Z Z a=2 x 2y
4 a=2
D K 2 dx dy cos sin
a a=2 a=2 a a
y
2x
.xy/ cos sin (11.186)
a a
The x-integral is
Z x
a=2
2x
Ix D x cos sin dx
a=2 a a
Z a=2 x x
D2 x cos2 sin dx
a=2 a a
8a2
D (11.187)
9 2
where we have integrated by parts and used Eq. (E.8). Because the y-integral
is identical to the x-integral we have
2
4 8a2
h12j xy j21i D K 2
a 9 2
256a2
D K
81 4
(11.188)
The Hamiltonian in matrix form is then

!
.0/
O E12 0 01
HD .0/ C (11.189)
0 E12 10
If the second matrix can be diagonalized it will be a simple matter

to obtain the perturbed eigenvalues. To do this we solve the eigenvalue
problem for that matrix to obtain the eigenkets (in terms of the unperturbed
eigenfunctions) that make the matrix diagonal. These eigenkets are the select
set.
We have

01 a .1/ 1 0 a
D E1 (11.190)
10 b 01 b
where the column matrix represents the select kets. We have inserted the
identity matrix on the rhs for clarity. Multiplying the matrices in Eq. (11.190)
gives two equations.
.1/
b E1 a D 0
.1/
a E1 b D 0 (11.191)
These are homogeneous equations in a and b so the only non-zero solution

is obtained by solving the secular equation which is
!
.1/
E1
det .1/ D0 (11.192)
E1
which yields
h i2 256a2
.1/ .1/
E1 2 D 0 ) E1 D ˙ D
K (11.193)
81 4
Thus, the degenerate levels are split, one higher than the degenerate unper-
.0/
turbed energy E12 and the other lower. That is,
.0/ .1/ 5 2 „2 256a2

E12 C E12 D 2

K (11.194)
2mL 81 4
.1/
Additionally, inasmuch as E1 D ˙ we see from Eq. (11.191) that the
select set of eigenkets is

1 1 1 1
p and p (11.195)
2 1 2 1
2. A particle of mass is confined to move on a circle of radius R in the xy-plane

( D =2 in spherical coordinates). Along the circular path there is a ı-function
perturbation applied in the form HO 1 ./ D U0 ı ./ where is the azimuthal
angle in spherical coordinates.
(a) Find the energy eigenfunctions and eigenvalues of the unperturbed Hamilto-
nian HO 0 ./ for all states of this system.
(b) Find the matrix that represents the complete Hamiltonian HO ./ and find the
energies of all states to first order.
(c) Find the select set of kets and show that they are eigenkets of both HO 0 and
HO 1 .
Solution
(a) There is no potential energy in HO 0 ./ so the total unperturbed energy is the
kinetic energy of a rigid rotor in a plane. Using Eq. (Q.3), the unperturbed
Hamiltonian is

O
Lz2 1 „ d 2
H 0 ./ D D (11.196)
2I 2R2 i d
where I is the moment of inertia of the mass about the origin. This problem
is a one-dimensional problem in . The unperturbed Hamiltonian HO 0 ./
is simply the square of the angular momentum operator so the eigenvalue
equation is

O .0/ 1 „ d 2 .0/
H 0 ./ ./ D ./ (11.197)
2R2 i d
.0/
Solving this differential equation for m ./ and normalizing we have
r
.0/ 1 im
./ D e m D ˙1; ˙2; ˙3; : : : (11.198)
m
2
where m must be an integer to satisfy the criterion that the wave function
must be single valued. The ˙ m-values correspond to clockwise and coun-
terclockwise motion on the circle. Inserting Eq. (11.198) into Eq. (11.197)
leads to Em.0/ the unperturbed energy levels
m2 „2
Em.0/ D m D ˙1; ˙2; ˙3; : : : (11.199)
2R2
Because m is squared in Eq. (11.199) the mth energy is twofold degener-
ate. This is the same as saying that the energy is independent of the direction
of rotation.
.0/
(b) Designating the eigenkets corresponding to ˙m ./ by j˙mi and using
the pair of kets j˙mi as the basis set to split the twofold degeneracy, the
complete Hamiltonian in matrix form is
HO D HO 0 C HO 1
.0/
Em 0 hmj ı jmi hmj ı jmi
D C U0 (11.200)
0 Em.0/ hmj ı jmi hmj ı jmi
where ı D ı ./. Using the sifting property of the ı-function (see Table J.1),
the diagonal elements in HO 1 are
hmj ı jmi D hmj ı jmi

Z
1
D eim ı ./ eim d
2
1
D (11.201)
2
The off-diagonal integrals are
hmj ı jmi D hmj ı jmi

Z
1
D eim ı ./ eim d
2
1
D (11.202)
2
Then

hmj ı jmi hmj ı jmi U0 11
HO 1 D U0 D (11.203)
hmj ı jmi hmj ı jmi 2 11
Now we solve the eigenvalue problem for HO 1 .

U0 11 a .1/ a
D Em (11.204)
2 11 b b
so
U0 a C U0 b D 2aEm.1/
U0 a C U0 b D 2bEm.1/ (11.205)
or
h i
U0 2Em.1/ a C U0 b D 0
h i
U0 a C U0 2Em.1/ b D 0 (11.206)
The only non-zero solutions to these simultaneous equations are found by

solving the secular equation which is
h i2
U0 2Em.1/ U02 D 0
4Em.1/ U0 C 4 2 Em.1/ D 0
h i
Em.1/ Em.1/ U0 D 0 (11.207)
Thus the corrections to the unperturbed energy Em.0/ are
U0
Em.1/ D 0 and Em.1/ D (11.208)

The perturbation separates the m-level energies by U0 = to first order.

The total energies to first order are
m2 „2 m2 „2 U0
Em.0/ C Em.1/ D 2
and Em.0/ C Em.1/ D 2
C (11.209)
2R 2R
(c) Solving Eq. (11.206) for a and b we have
U0
aD h ib
.1/
U0 2Em
U0
b D h ia (11.210)
.1/
U0 2Em
First consider the case for which Em.1/ D 0. From Eq. (11.210) we have
a D ˙b so the eigenkets corresponding to Em.1/ D 0, the select set, are

1 1 1 1
p and p (11.211)
2 1 2 1
We would get the same result if we had chosen Em.1/ D U0 = because

there is only one select set. Now we wish to solve the eigenvalue problem
with the Hamiltonian HO D HO 0 C HO 1 to verify that the eigenkets in
Eq. (11.211) are eigenkets of both HO 0 and HO 1 (the necessary criterion for
the select set). The eigenvalues that are obtained should be the same as those
in Eq. (11.209).
Using the first of the select kets in Eq. (11.211), call it .1; 1/, we have

O .0/ 1 10 1 U0 1 11 1
H j.1; 1/i D Em p C p
2 0 1 1 2 2 1 1 1

E.0/ 1 U0 1
D pm Cp
2 1 2 1

U0 1 1
D Em.0/ C p (11.212)
2 1
Using the second of the select kets in Eq. (11.211), .1; 1/, we have

O .0/ 1 10 1 U0 1 11 1
H j.1; 1/i D Em p C p
2 01 ˙1 2 2 1 1 ˙1

E.0/ 1 U0 1 0
D pm C p
2 1 2 0

1 1
D Em.0/ p (11.213)
2 1
The respective energy eigenvalues, Em.0/ C0 and Em.0/ CU0 =, are the same
as those in Eq. (11.209) as they should be.
3. A perturbation HO 1 .x; y/ D Cxy is applied to an isotropic two-dimensional

harmonic oscillator (! x D ! y ).
(a) Find the eigenfunctions and eigenvalues for the unperturbed ground state and
for the doubly degenerate first excited state.
(b) Find the matrix that represents the complete Hamiltonian of the excited states
and find the energies of the perturbed first excited states to first order.
Solution
(a) The unperturbed Hamiltonian is separable in x and y so
HO 0 .x; y/ D HO x .x/ C HO y .y/ (11.214)
The Hamiltonian HO 0 .x; y/ may be written in terms of the ladder operators

(see Chap. 7).

O 1 1
H 0 .x; y/ D „! aO x aO x C C „! aO y aO y C
2 2

D „! aO x aO x C aO y aO y C 1 (11.215)
where the x and y subscripts on the ladder operators indicate that these
operators affect only that coordinate. As in Problem 1 of this chapter the
eigenfunctions of HO 0 .x; y/ are a product of the eigenfunctions of HO x .x/ and
HO y .y/. In Dirac notation this means that an arbitrary unperturbed eigenket
with quantum numbers nx and ny is written
ˇ ˛ ˇ ˛
jnx i ˇny D ˇnx ny (11.216)
and the unperturbed eigenvalues are

En.0/
x ny
D nx C ny C 1 „! (11.217)
The eigenvalues of the ground are
h00j HO 0 j00i D „! (11.218)
The first excited states j01i and j10i are clearly degenerate with energy
h10j HO 0 j10i D 2„! (11.219)
(b) The first-order correction to the ground state vanishes because the perturba-
tion is odd in both x and y.
Analogous to the Hamiltonian in Problem 1 of this chapter the matrix that

represents the total Hamiltonian for the degenerate excited states j01i and
j10i is

10 0 h10j xy j01i
HO D 2„! CC (11.220)
01 h01j xy j10i 0
where the diagonal elements of HO 1 .x; y/ vanish because of parity. We must

diagonalize the HO 1 .x; y/ matrix. Rather than use integrals of the algebraic
eigenfunctions with their Hermite polynomials to evaluate h10j xy j01i and
h01j xy j10i. We will use the ladder operators. Replace x and y with
1 1
xO D p aO x C aO x and yO D p aO y C aO y (11.221)
2˛ 2˛
so that
1
xO yO D 2
aO x aO y C aO x aO y C aO x aO y C aO x aO y (11.222)
2˛
and employ the raising and lowering properties (see Appendix O) to evaluate
the four terms in h10j xy j01i.
1 1 p
h10j aO x aO y j01i D 2 h20j 00i D 0
2˛ 2 2˛ 2
1 1
h10j aO x aO y j01i D 2 h20j 02i D 0
2˛ 2 ˛
1 1 1
h10j aO x aO y j01i D h00j 00i D
2˛ 2 2˛ 2 2˛ 2
1 1
h10j aO x aO y j01i D p h00j 02i D 0 (11.223)
2˛ 2 2˛ 2
so that the matrix elements of HO 1 are
C h10j xy j01i D C h01j xy j10i

C „2
D D C (11.224)
2˛ 2 2m2 ! 2
We must now diagonalize the HO 1 matrix. Because the unperturbed
Hamiltonian is symmetric in x and y and the perturbation is the same as
that in Problem 1 we know that the HO 1 matrix will be of the same form.

C 01
(11.225)
2˛ 2 10
Using the result of Problem 1 we have
.1/ C„
E1 D ˙ (11.226)
2m!
Therefore, the energies of the degenerate levels of HO 0 .x; y/ are shifted by

.1/
E1 ,one lower and the other higher. That is,
.0/ .1/ C„
E12 C E12 D 2„! ˙ (11.227)
2m!
The select set is

1 1 1 1
p and p (11.228)
2 1 2 1
4. Consider a Hamiltonian that is represented by the matrix

0 1
1 0
HO D @ 1 0 A (11.229)
006
O
(a) Find the select basis set and show that these eigenvectors diagonalize H.
O
(b) Verify that the off-diagonal matrix elements of H vanish with the select set
and show that the eigenvalues are .1 ˙ / and 6.
(c) Solve the problem exactly and compare with the perturbation theory result.
Solution
(a) The entry in the lower right is a non-degenerate eigenvalue because there are
no other entries in the same column or row. The rest of the Hamiltonian is a
2 2 matrix which may be written
HO D HO 0 C HO 1

10 0
D C (11.230)
01 0
Thus, the unperturbed eigenkets for this 2 2 system are

ˇ E ˇ E
ˇ .0/ 1 ˇ 0
ˇ1 D and ˇ2.0/ = (11.231)
0 1
But these are not the select set. To find the select set we must diagonalize
HO 1 which we do by solving the secular equation

E
det D 0 H) E D ˙ (11.232)
E
To find the select set of eigenvectors we must find the eigenkets corre-
sponding to these eigenvalues.

01 a a
D ˙ (11.233)
10 b b
where the column vectors represent the as yet unknown select kets. We have
b D ˙ a H) b D ˙a (11.234)
ˇ ˛
Therefore, the (orthonormal) select set which we designate ˇi.s/ is
ˇ E
ˇ .s/ 1 ˇˇ E ˇˇ E 1 1
ˇ1 D p ˇ1.0/ C ˇ2.0/ D p
2 2 1
ˇ E
ˇ .s/ 1 ˇˇ .0/ E ˇˇ .0/ E 1 1
ˇ2 D p ˇ1 ˇ2 D p (11.235)
2 2 1
ˇ ˛
The unperturbed state ˇ3.0/ with eigenvalue 6 is non-degenerate so no
manipulation
ˇ is˛ required.
ˇ ˛
(b) Although ˇ1.s/ and ˇ2.s/ were constructed so that the off-diagonal matrix
elements of HO on this select set vanish, let us verify this.
D ˇ ˇ E
O .s/ ˇ
ˇ .s/ 1 10 0 1
1 ˇ H ˇ2 D 11 C
2 01 0 1

1 1 1
D 11 C 11
2 1 2

1 1
D 11
2 Œ1
1
D Œ.1 / .1 / D 0 (11.236)
2
˝ ˇ ˇ ˛
Clearly 2.s/ ˇ HO ˇ1.s/ also vanishes. Now for the diagonal elements.
D ˇ ˇ E
ˇ ˇ 1 10 0 1
1.s/ ˇ HO ˇ1.s/ D 11 C
2 01 0 1

1 1 1
D 11 C 11
2 1 2

1 1C
D 11
2 1C
1
D .2 C 2 / D 1 C (11.237)
2
and
D ˇ ˇ E
.s/ ˇ
O ˇ .s/ 1 10 0 1
2 ˇ H ˇ2 D 1 1 C
2 01 0 1

1 1
D 1 1 C 1 1
2 1

1 1
D 1 1
2 Œ1
1
D .2 2 / D 1 (11.238)
2
O the part of the matrix

(c) Ignoring the already diagonalized portion of H,
representing HO is

1
(11.239)
1
The eigenvalue equation is

1 A A
DE (11.240)
1 B B
which becomes

1E A
D0 (11.241)
1E B
Expanding this equation into two simultaneous equations we notice that

these will be homogeneous equations so the only non-trivial solution can
occur only if the determinant of the coefficients vanishes. We have
.1 E/2 2 D 0
or

E2 2E C 1 2 D 0
Using the quadratic formula we have

p
E D 2 ˙ 4 4 .1 2 / =2
D 1˙ (11.242)
Although this problem involved a 3 3 matrix the fact that one of the
states in this problem was non-degenerate reduced it to a 2 2 problem.
11.6 Time Dependent Perturbation Theory
In time dependent perturbation theory the Hamiltonian HO .r; t/ is written as the sum
of two terms, the unperturbed Hamiltonian HO 0 , which is time independent and the
time dependent perturbation WO .r; t/.
HO .r; t/ D HO 0 .r/ C WO .r; t/ (11.243)
It is assumed that WO .r; t/ is turned on at t D t0 and that the eigenkets and energy
eigenvalues of HO 0 are known. Additionally, each eigenket j n i has an unperturbed
energy eigenvalue En .
Time dependent perturbation theory is used to determine the probability of
finding a system that was initially in state i to be in some final state f after
having turned on the perturbation WO .r; t/ at t D t0 . The details of this subject are
covered in most quantum mechanics textbooks [1]. The fundamental result is that
.1/
the probability of the transition i ! f designated by Pi!f is given by
ˇZ ˇ
.1/ 1 ˇˇ t O 0 i! fi t0 0 ˇˇ2
Pi!f .t/ D 2 ˇ W fi r; t e dt ˇ (11.244)
„ t0
where
Ef Ei
! fi D (11.245)
„
and
˝ ˇ
WO fi D ˇ WO .r; t/ j
f ii (11.246)
It is important to note that Eq. (11.244) is valid only when the sum of the
.1/
probabilities Pi!f (f ¤ i) is much less than unity.
If the limits of integration in Eq. (11.244) are 1 to 1 the integral is simply the
Fourier transform of the perturbation evaluated at the transition frequency.
Problems
1. A particle of mass m is in the ground state of an L-box (see Appendix M). At

t D 0 a perturbation WO .x; t/ is applied;
2
WO .x; t/ D Axe.t= / (11.247)
where A and are constants. What are the units of A? Use first-order time-
dependent perturbation theory to find the probability that the system will undergo
a transition to the first excited state after a long time, t D C1. Do the calculated
probabilities have the correct units?
Solution
The perturbation must have units of energy so the units of A are J= m (SI).
Use the fundamental result of time dependent perturbation theory,
Eq. (11.244), to obtain
ˇZ ˇ
1 ˇ 1 O 0 i! 21 t0 0 ˇˇ2
P1!2 D 2 ˇˇ
.1/
W 12 x; t e dt ˇ (11.248)
„ 0
where „! 21 is the difference between the L-box ground state and first excited
state energies (see Appendix M). That is
22 2 „2 12 2 „2
„! 21 D
2mL2 2mL2
3 2 „2
D (11.249)
2mL2
The matrix element WO 12 .x; t0 / is

2
WO 12 x; t0 D Ae.t= / h1j x j2i (11.250)
where the bra and ket represent the ground and first excited states, respectively.
The matrix element h1j x j2i is, using the eigenfunctions given in Appendix M
Z x
2A L
2x
A h1j x j2i D sin x sin dx
L 0 L L
Z x x
4A L
D x sin2 cos dx
L 0 L L
16AL
D (11.251)
9 2
This result is obtained by integrating Eq. (11.251) by parts and using the
integral given in Eq. (G.13).
We have now
2 ˇZ ˇ2
16AL ˇ 1
/ dt0 ˇˇ
ˇ e.i! 21 t t
.1/ 0 02 = 2
P1!2 D ˇ ˇ (11.252)
9 2 „ 0
We can solve this integral by completing the square in the exponent. Dropping
the primes and subscripts for convenience

t i dt
qD 2
! ) dq D 2
2
t2 i!t ! 2
q2 D (11.253)
4 2 4
so that

t2 !2
i!t 2 D 2 C q2 (11.254)
4
Inserting this into Eq. (11.252) we have
2 ˇZ ˇ
16AL ˇ 1
2 .! 2 =4Cq2 /
ˇ2
.1/
P1!2 D ˇ e dq ˇˇ
2
9 2 „ ˇ
0
2 ˇZ ˇ2
16AL ˇ 1 ˇ
D 4 e!
2 2 =4
ˇ e
2 q2
dqˇˇ (11.255)
9 2 „ ˇ
0
Before evaluating this integral, which is finally in a comfortably recognizable

form, it is worthwhile to check the units of the constants in this problem. Because
exponents must be unitless it is clear from the equations above that and q have
units s and s1 , respectively.
Now, we simply use the integral given in Eq. (G.3) and find that
2 ˇ r ˇ2
16AL 2 ! 221 2 =4
ˇ1 ˇ
.1/
P1!2 D e ˇ ˇ
9 2 „ ˇ2 2 ˇ
2
8AL 2 2 =4
D e! 21 (11.256)
9 2 „
This probability must be unitless. To check we use our deduction above that
A has units J= m. We have already shown that the exponential is unitless, so

2
J Œ m2 Œ s2
P D
Œ m2 Œ J s2
! unitless (11.257)
2. An electron is in the n D 1 state of an L-box (see Appendix M) for which the

potential energy is
U .x/ D 0 0xL
D1 1 < x < 0I L < x < 1 (11.258)
At t D 0 a uniform electric field of magnitude F is applied in the x-direction

and turned off after a short time . Obtain an expression for the probability of
finding the electron in eigenstate n of the L-box after the field is turned off.
Evaluate the probability expression for n D 2 and n D 3.
Solution
Using the fundamental result of time dependent perturbation theory, Eq. (11.248),
we have, in the notation of this problem
ˇZ ˇ
.1/ 1 ˇˇ O 0 i! fi t0 0 ˇˇ2
P1!f ./ D 2 ˇ Wf 1 t e dt ˇ (11.259)
„ 0
where the final state f will be either the first or second excited state of the L-box,
nf D 2 or 3. The difference in energies is
n2f 2 „2 12 2 „2
„! f 1 D
2mL2 2mL2
2 „2
D n2f 1 (11.260)
2mL2
The perturbing Hamiltonian WO .r; t/ is simply the potential energy of a particle

of charge e in an electric field during the time interval 0 < t < . Because
the field F is uniform and in the x-direction the spatial part of the perturbing
Hamiltonian is
U .x/ D eFx (11.261)
where e is the charge (negative) on the electron. The matrix element WO f 1 .r; t0 / is
then given by

WO f 1 r; t0 D eF h1j x jni 0 < t < (11.262)
where the bra and ket represent L-box eigenfunctions and we have let n D nf .
Note that the time dependence of the perturbation is given by the time interval
over which the spatial part of the electric field acts.
Using the L-box eigenfunctions (see Appendix M), the trigonometric identity
given in Eq. (E.5) and the integral given in Eq. (G.8) we have
Z x nx
2 L
h1j x jni D sin x sin dx (11.263)
L 0 L L
(Z Z )
1 L h x i L h x i
D cos .1 n/ xdx cos .1 C n/ xdx
L 0 L 0 L
h
L 1 x i 1 h x i L
D 2 cos .1 n/ cos .1 C n/
.1 n/2 L .1 C n/2 L 0

L 1 1
D 2 cos Œ.1 n/ cos Œ.1 C n/
.1 n/2 .1 C n/2
h i h i
L 1n L
D 2
./ 1 2
./1Cn 1 (11.264)
2 .1 n/ 2 .1 C n/
where
h i
cos Œ.1 n/ D ./1n 1
h i
cos Œ.1 C n/ D ./1Cn 1 (11.265)
and
h i h i
./1n 1 D ./1Cn 1 (11.266)
so we have
h i
L L
h1j x jni D ./1Cn 1
2 .1 n/2 2 .1 C n/2
h i
L 4n 1Cn
D ./ 1 (11.267)
2 .1 n2 /2
We see immediately that excitation to the n D 3 state is forbidden; it cannot

occur because the matrix element vanishes for n D 3 and for all final states with
n odd. The probability of a transition to n D 2 is given by Eq. (11.259). Using
Eq. (11.260),
3 2 „
! 21 D (11.268)
2mL2
and the time independent matrix element WO 21 .r; t0 / D WO 21 is
WO 21 D eF h1j x jni

L 8
D .eF/ .2/
2 9
16 eLF
D (11.269)
9 2
so we obtain
ˇZ ˇ
.1/ 256 .eLF/2 ˇˇ i! 21 t0 0 ˇˇ2
P1!2 ./ D e dt
81 „2 4 ˇ 0 ˇ

256 2 1 1 ˇˇ ˇ
i! 21 ˇ2
D .eLF/ 1e
81 „2 4 ! 221

256 2 1 1
D .eLF/ Œ2 2 cos .! 21 /
81 „2 4 ! 221

44 1 1

D 4 .eLF/2 2 4 2
4 sin2 .! 21 =2/ (11.270)
3 „ ! 21
where we have used Eq. (E.9).

For short time intervals ! 21 << 1 we may use sin .! 21 =2/ .! 21 =2/ so
that

.1/ 45 2 1 1 ! 21 2
P1!2 ./ 4 .eLF/
3 „2 4 ! 221 2
" #
44 .eLF/2 2
D 4 (11.271)
3 „2 4
It is interesting that under the condition that sin .! 21 =2/ .! 21 =2/ the
probability of a transition is proportional to 2 and is independent of the energy
separation between states, „! fi .
3. A particle in a one-dimensional harmonic oscillator potential is in the ground

state. At t D 0 a perturbation is applied
WO .x; t/ D Axet= (11.272)
where A and are constants. What are the units of A? Use first-order time-
dependent perturbation theory to find the probability that the system will undergo
a transition to any excited state after a long time, t D C1. Do the calculated
probabilities have the correct units? It will be helpful to use the matrix element
for the harmonic oscillator hmj xO jni that was calculated in Problem 6 of Chap. 7,
Eq. (7.73). This matrix element is
1 p p
hmj xO jni D p nı m;n1 C n C 1ı m;nC1 (11.273)
2˛
Solution
The perturbation must have units of energy so, as in Problem 1 of this chapter,
the units of A are J= m.
Again as in Problem 1 of this chapter we apply the fundamental result of
time-dependent perturbation theory, Eq. (11.244).
ˇZ ˇ
.1/ 1 ˇˇ 1 O 0 i! f 0 t0 0 ˇˇ2
P0!f D 2ˇ W f 0 x; t e dt ˇ (11.274)
„ 0
where ! f 0 is the difference between the ground state and final state energies. In
the present case the perturbation is given by Eq. (11.272) so the matrix element is

WO f 0 x; t0 D Aet= h0j x jf i (11.275)
where the ket jf i represents any of the final states.

The matrix element we seek is h0j xO jf i which from Eq. (11.273) is
1 p p
h0j xO jf i D p f ı 0;f 1 C f C 1ı 0;f C1 (11.276)
2˛
The only way the first term will not vanish is if the final state is the first excited
state f D 1. The second term will vanish for any final state inasmuch as there
is no state for which f D 1. Therefore, the only transition possible under this
perturbation is the 0 ! 1 transition so
.1/
P0!f 0 f > 1 (11.277)
For f D 1 we have
1
h0j xO j1i D p (11.278)
2˛
To obtain the remaining part of the transition probability we require the integral
Z 1 Z 1
0 0 0
WO 10 x; t0 ei! f 0 t dt0 D A h0j x j1i et = ei! 10 t dt0
0 0
Z 1
A 0
D p e.i! 10 1= /t dt0
2˛ 0
A 1
D p (11.279)
2˛ .i! 10 1= /
where ! 10 D ! the oscillator frequency. The transition probability is then
.1/ jAj2 1 1
P0!1 D (11.280)
„ 2˛ .! C 1= 2 /
2 2 2
Using ˛ 2 D m!=„ we have
.1/ 1 jAj2 „ 1
P0!1 D (11.281)
2 „2 m! .! 2 C 1= 2 /
This probability must be unitless. To check we use SI units and our deduction
above that A has units J= m.

2
1 J 1 Œ J s 1
PD
2 Œ m2 Œ J s2 1 1
Œ kg
s s2
Œ J Œ J
D D (unitless) (11.282)
Œ. kg m2 = s2 / Œ J
We can calculate the average oscillator energy change E0!1 corresponding

to this probability 0 ! 1. This is simply the probability multiplied by „!.
.1/
E0!1 D „! P0!1
jAj2 1
D (11.283)
2m .! 2 C 1= 2 /
In Problem 5 of Sect. 11.3 we calculated the energy shift of a harmonic

oscillator due to the application of a constant electric field F. The perturbation
was given by
HO 1 D eFx (11.284)
In this problem we have an electric field that is decaying with time constant . If,
however, ! 1 the “decaying” field would be constant in time so we should
arrive at the same result for the average energy change of the oscillator as that
obtained in Problem 5 of Sect. 11.3 which was
e2 F 2
E D (11.285)
2m! 2
Comparing the perturbation in this problem, Eq. (11.272) with that of the
previous problem, Eq. (11.284), we see that A in this problem corresponds to
eF in the previous problem. Taking the limit as ! 1 of E0!1 Eq. (11.283)
we have
jAj2 1
lim E0!1 D lim
!1 !1 2m .! 2 C 1= 2 /
jAj2
D (11.286)
2m! 2
so that replacing jAj2 with e2 F 2 does indeed produce the same energy change as
that obtained for a constant field.
4. A particle of mass m is in the ground state of a one-dimensional harmonic

oscillator potential. The oscillator frequency is !. At t D 0 a weak constant
force F is applied and acts until time t D so the perturbing potential is
WO .x; t/ D Fx 0<t< (11.287)
Use first-order time dependent perturbation theory to find the value (or values)
of , call them max , that maximize the probability of a transition to n D 1.
Solution
Using Eq. (11.244), the probability of the n D 0 ! 1 transition is

ˇ Z ˇ2
.1/ F 2 ˇˇ ˇ
i!t0 0 ˇ
P0!1 ./ D 2 ˇ h0j O
x j1i e dt ˇ
„ 0
ˇ 0 ˇ2
F2 ˇ 1 i!t0 t D ˇˇ
ˇ
D 2 ˇh0j xO j1i e ˇ
„ ˇ i! t0 D0 ˇ
F2 ˇ
2 ˇ i! ˇ2
D jh0j O
x j1ij e 1 ˇ (11.288)
„2 ! 2
The matrix element h0j x j1i, using Eq. (11.273), is
1 p p
h0j xO jf i D p f ı 0;f 1 C f C 1ı 0;f C1 (11.289)
2˛
so
1 p
h0j xO j1i D p ı 0;0 C 2ı 0;2
2˛
r
1 „
D p D (11.290)
2˛ 2m!
The probability that the harmonic oscillator will be found in the first excited
state after a time is
F 2 „ ˇˇ i! ˇ2

e 1 ˇ
.1/
P0!1 D 2 2
„ ! 2m!
F 2 ˇˇ i! =2 i! =2 ˇ2
D 3
e e ei! =2 ˇ
2m„!
ˇ ˇ2
2F 2 ˇˇ ei! =2 ei! =2 ˇˇ
D ˇ ˇ
m„! 3 ˇ 2i ˇ
2F 2 !
D 3
sin2 (11.291)
m„! 2
The sin2 .!=2/ term is the only one that depends upon . It has a maximum
of 1 when ! D ; 3; : : :. In general the values of max are

max D .2j C 1/ where j D 0; 1; : : : (11.292)
!
While Eq. (11.292) is the answer we sought, it is worthwhile to make sure the
units are correct. The max result has units of time as it should. How about the
coefficient of the sin2 .!=2/ term in Eq. (11.291)? It must be unitless because
probabilities do not have units. We have
2F 2 N2
! 2
m„! 3 . kg/ . J s/ . s3 /
N2
!2
. kg/ . N m/ s2
N
!2 (11.293)
. kg m s2 /
Inasmuch as N D kg m s2 the coefficient is unitless giving us some

confidence that the answer is correct.
Because we have used first order perturbation theory we must assure ourselves
that F is weak enough so that the probabilities are much less than unity at max .
.1/
A reasonable limit is P0!1 . 0:2 so that, from Eq. (11.291), F 2 . 0:1m„! 3 .
5. A H-atom in the ground state is immersed in an electric field F that is constant

in the z-direction, but varies in time as
FD0 for t < 0

F D F0 et= for t > 0 (11.294)
Use time dependent perturbation theory to determine the probability that after
a long time, i.e. t D 1, the atom is in the n D 2; ` D 1; m D 0 state.
Solution
Again we require the fundamental result of time dependent perturbation theory,

Eq. (11.244).
ˇZ ˇ
.1/ 1 ˇˇ 1 O 0 i! ki t0 0 ˇˇ2
Pi!f D 2ˇ W fi z; t e dt ˇ (11.295)
„ 0
where ! is the difference in the Bohr energies of the n D 2 and n D 1 levels of

atomic hydrogen.
.E2 E1 /
!D (11.296)
„
The time dependent perturbation term in the Hamiltonian is the potential
energy associated with the electric field

WO .z; t/ D e F0 zet=
D eF0 r cos et= (11.297)
In the present case the matrix element WO fi is
WO fi D h210j WO .z; t/ j100i (11.298)
To perform the integral in Eq. (11.298) we require the eigenfunctions given

in Appendix T. The bra and ket in Eq. (11.298) correspond to the H-atom
eigenfunctions
3=2
1 1 r
210 .r; ; / D p er=2a0 cos
2 2 a0 2a0
3=2
1 1
100 .r; ; / D p er=a0 (11.299)
a0
Noting that these functions contain no -dependence we have
WO fi D h210j WO .z; t/ j100i

Z Z
eF0 et= 1 4 1 4 3r=2a0
D p r e dr cos2 sin d
2 2 a0 0 0
" # 3 0 Z 1
eF0 et= 1 4 cos
D p r4 e3r=2a0 dr
2 2 a0 3 0
" #Z
eF0 et= 1 4 1
D p r4 e3r=2a0 dr (11.300)
3 2 a 0 0
Let
3r
uD (11.301)
2a0
so that
2a0 2a0
rD u and dr D du (11.302)
3 3
The last line of Eq. (11.300) becomes
" 4 # 5 Z 1
eF0 et= 1 2a0
WO fi D p u4 eu du (11.303)
3 2 a0 3 0
Using Eq. (G.2) we have

" 4 # 5
eF0 et= 1 2a0
WO fi D p 4Š
3 2 a0 3
28 a0
D p eF0 et=
5
3 2
D Cet= (11.304)
where C is
28 a0
CD p .eF0 / (11.305)
35 2
Inserting this into Eq. (11.295) we have

2 ˇZ 1 ˇ2
.1/ C ˇˇ t0 = i! ki t0 0 ˇ
ˇ
Pi!f D ˇ e e dt ˇ
„ 0
2 ˇZ 1 ˇ2
C ˇˇ ˇ
t0 = Ci! ki t0 0 ˇ
D ˇ e dt ˇ
„ 0
ˇ ˇ
2 ˇ et= Ci! ki t 1 ˇ2
C ˇ ˇ
D lim ˇ ˇ
„ t!1 ˇ i! 1= ˇ
2 ˇ ˇ2 2
C ˇˇ 1 ˇ
ˇ D C 2
D
„ ˇ i! 1= ˇ „ !2 2 C 1

215 .eF0 a0 /2 1
D (11.306)
310 .E2 E1 /2 C .„= /2
Suppose the final state had been chosen to be j200irather than j210i, in which
case WO fi would be
WO fi D h200j WO .z; t/ j100i (11.307)
This integral vanishes because WO .z; t/ / z so the integrand is necessarily

odd and is taken over all space. Recall that the spherical harmonics have definite
parity determined by the value of ` (see Appendix R). The fact that an ` D 0 !
` D 0 transition is forbidden (to first order) is an example of the selection rules
that govern electric dipole (first order) electromagnetic transitions because the
electric dipole operator is proportional to the vector r [1, 1, 2].
6. Consider a system prepared in initial state jii and perturbed by a periodic

potential
U .x; t/ D Axei!t (11.308)
where A is a constant. The potential is switched on at time t D 0. What is the

probability that, at some later time t, the system is in state jf i? Assume that first-
order perturbation theory is valid.
Solution
Use Eqs. (11.244)–(11.246), and write for the probability of an i ! f transition

2 ˇZ t ˇ
.1/ A ˇ
2ˇ 0

ˇ2
0 ˇ
Pi!f .t/ D hf j x jii ˇ dt exp i .! fi !/ t ˇ (11.309)
„ 0
Integration of the exponential and its complex conjugate are straightforward,

.1/
and Pi!f .t/ becomes
2 2
.1/ A 2 sin Œ.! fi !/ t=2
Pi!f .t/ D hf j x jii (11.310)
„ .! fi !/ =2
When .! fi !/ t=2 << =2, the transition probability is

2
.1/ A
Pi!f .t/ Ð hf j x jii2 t2 (11.311)
„
and is independent of the both the transition frequency ! fi and the applied
frequency !. The transition probability reaches its first maximum when
.! fi !/ t=2 D =2 so that
2 2
.1/ A 1
Pi!f .t/ Ð hf j x jii2 (11.312)
„ .! fi !/ =2
.1/
The subsequent time variation is such that Pi!f .t/ oscillates between 0
.1/ .1/
and Pi!f .max/. The value of Pi!f .max/ must be much less than unity for
perturbation theory to be valid; this requirement sets a limit on the magnitude
of the potential parameter A.
The “system” in this problem could be an atom or a molecule irradiated by
an electric field oscillating at visible or infrared frequencies. Rather than the
transition probability at a particular time t, spectroscopists are more interested
in the rate of transition Ri!f .t/ defined by
.1/
Ri!f .t/ lim Pi!f .t/ =t (11.313)
t!1
If we set a D .! fi !/ =2, the bracketed term in Eq. (11.310) becomes

Œsin.at/=a2 , and we have for large t (Appendix J and references therein).
lim Œsin.at/=a2 =t D ı .a/ D 2ı .2a/ (11.314)

t!1
Using Eqs. (11.310) and (11.314) in Eq. (11.313), we obtain

2
A
Ri!f .t/ D 2 hf j x jii2 ı .! fi !/ (11.315)
„
Equation (11.315) is called “Fermi’s Golden Rule” and it emphasizes the

importance of resonance in radiative transitions.
A final remark: The units of the delta function are always the inverse of the
argument units. In this case, the argument has units of s1 , so the delta function
has units of seconds. Using this fact, one sees that Ri!f .t/ has units of s1
(probability per unit time) as it should.
References
1. Burkhardt CE, Leventhal JJ (2008) Foundations of quantum physics. Springer, New York
2. Townsend JS (2000) A modern approach to quantum mechanics. University Science Books,
Sausalite
3. Wangsness RK (1986) Electromagnetic fields, 2nd edn. Wiley, New York
Appendix A
Greek Alphabet
Table A.1 The letters of the Greek alphabet

LC UC Name Usage in this book
˛ A Alpha Fine structure constant (LC), harmonic oscillator (LC), spin up (LC)
ˇ B Beta General parameter, spin down (LC)
Gamma function (UC), square root transmission of coefficient T (UC)
ı Delta Dirac ı-function (LC), small increment
, " E Epsilon Unitless energy parameter (LC), small quantity (LC)
Z Zeta General parameter
H Eta General parameter
, # ‚ Theta Polar angle (LC), function (UC)
I Iota –
K Kappa Real exponent (LC), hyperfine energy (LC)
ƒ Lambda Wavelength (LC)
M Mu General parameter, reduced mass (LC)
N Nu Frequency (radians/ s ) (LC)
„ Xi Unitless length harmonic oscillator (LC)
o O Omicron –
… Pi 3:14159 : : :(LC)
P Rho Parameter, unitless length hydrogen

† Sigma Pauli matrices (LC), summation (U. C.)
T Tau Increment of time
‡ Upsilon –
ˆ Phi Azimuthal angle (LC), function (UC)
X Chi Spin state (LC)
‰ Psi Wave function (LC) and (UC)
! Omega Frequency (radians/ s) (LC), Bohr frequency
Where appropriate, their primary usage in this book is indicated. LC and UC refer to lower and
uppercase

and Quantum Mechanics, DOI 10.1007/978-3-319-46664-4
Appendix B
Acronyms, Descriptors and Coordinates
B.1 Acronyms and Descriptors

Table B.1 Acronyms and Descriptors
Term/Abbreviation Meaning
a-box Infinite square well with a=2 x a=2
L-box Infinite square well with 0 x L
TDSE Time dependent Schrödinger equation
TISE Time independent Schrö dinger equation
TME Total mechanical energy (the “energy”)
WKB Wentzel, Kramers, Brillouin approximation
SHO Simple harmonic oscillator
H-atom Hydrogen atom
B.2 Coordinate Systems
Table B.2 Coordinate systems

System Coordinates Unit vectors
Cartesian x; y; z {O; |;O kO
Spherical r (radial); (polar); (azimuthal)
aO r ; aO ; aO
Cylindrical (radial); (azimuthal); z aO ; aO ; kO
r
Plane polar r;
a
aO r ; aO or ; aO
r
a
Plane polar D cylindrical coordinates with z D 0, ! r, and
!

Appendix C
Units
In keeping with modern usage SI units are used throughout except in problems
where it is inconvenient and other units are appropriate. Useful units for calculations
involving atomic dimensions are
“atomic units”
abbreviated
a.u. In a.u. the unit of
length is the Bohr radius a0 D .4 0 / „2 = me e2 , the electronic charge e, the
mass of the electron me , and „ are all set equal to unity. The conversion between
a.u. and SI units is given in Table C.1.
A few points about a.u. are worth noting:
1. The unit of velocity is the velocity of the electron in the first Bohr orbit.
2. The unit of time is the period of the electron in the first Bohr orbit divided by 2.
3. The unit of energy is twice the ground-state Bohr energy (27:2 eV).
Table C.1 Atomic units (a.u.)

Quantity a.u. SI
Mass me D 1 9:10 1031 kg
Charge eD1 1:60 1019 C
Angular momentum „D1 1:06 1034 J s
Length a0 D 1 5:29 1011 m
Velocity v0 D ˛c 2:20 106 m= s
Time a0 =v0 D 1=˛c 2:42 1017 s
Energy e2 = .4 0 a0 / D 1 4:36 1018 J

Electric field e= 4 0 a20 D 1 5:14 1011 V= m
Bohr magneton e„= .2me / D 1=2 9:274 1024 J= T
Fine structure constant ˛ D e2 = Œ.4 0 / „c D 1=137 1=137

318 C Units
4. The speed of light c D 137 a.u.D Œ(length in a.u.)/(time in a.u.).

5. The fine structure constant ˛ is a dimensionless number (the same in any system
of units). It is easily evaluated using a.u.

1 1 1
˛D D (C.1)
.4 0 / „c 137
Appendix D
Conic Sections in Polar Coordinates
Most students are familiar with the equations of conic sections , ellipses, hyperbolas,
and parabolas in Cartesian coordinates. For the Kepler problem in which the
gravitational force law is an attractive inverse square law the orbits of the bodies
are conic sections which are most conveniently written in polar coordinates .r; /.
Although a conic section may be defined as the curve that results from the
intersection of a plane and a right circular cone, we require an equivalent, but more
quantitative definition.
Discussion of the details is facilitated using Fig. D.1 in which the origin of polar
coordinates O is the focus and the vertical line a distance p away from O is the
directrix. A conic section is the locus of all points having a constant ratio , the
eccentricity.
jOPj
D (D.1)
jMPj
The eccentricity is the key parameter in ascertaining the path of the particle. The
point P lies on the curve if and only if
Fig. D.1 Parameters for the
derivation of the equation of a
conic section

320 D Conic Sections in Polar Coordinates
jOPj D jMPj (D.2)
Equation (D.2) may be cast in terms of the polar coordinates by noting that
jOPj D r (D.3)
and
jMPj D p r cos (D.4)
Therefore,
r D .p r cos / (D.5)
Solving for r we have

p
rD (D.6)
1 C cos
This is the equation of a conic section in polar coordinates with the origin at one
focus. Different values of the ratio yield different shapes.
D 1 parabola
< 1 ellipse
> 1 hyperbola (D.7)
It is known that a circle is an ellipse with D 0 which, from Eq. (D.1) means that
the distance jMPj ! 1.
Equation (D.6) is often written in the form
˛
D 1 C cos (D.8)
r
For Keplerian orbits, where the Keplerian potential is given by
k
U .r/ D (D.9)
r
the constants in Eq. (D.8) are given by
s
`2 2E`2
˛D and D 1C (D.10)
k k2
D Conic Sections in Polar Coordinates 321
where E is the TME of the particle in its orbit and 2˛ is known as the latus rectum.
Notice that if the orbiting object is bound to the force center, an elliptical orbit, then
E is negative. Further, if
k2
ED (D.11)
2`2
then, from Eq. (D.10), D 0 and the orbit is a circle (see Eq. (D.8)).
Appendix E
Useful Trigonometric Identities
sin .A ˙ B/ D sin A cos B ˙ cos A sin B (E.1)

cos .A ˙ B/ D cos A cos B
sin A sin B (E.2)
tan A ˙ tan B
tan .A ˙ B/ D (E.3)
1
tan A tan B
cot A cot B
1
cot .A ˙ B/ D (E.4)
cot B ˙ cot A
1
sin A sin B D Œcos .A B/ cos .A C B/ (E.5)
2
1
cos A cos B D Œcos .A B/ C cos .A C B/ (E.6)
2
1
sin A cos B D Œsin .A B/ C sin .A C B/ (E.7)
2
sin 2A D 2 sin A cos A (E.8)
2 2
cos 2A D 1 2sin A D 2cos A 1 (E.9)
sin 3A D 3 sin A 4sin3 A (E.10)

Appendix F
Useful Vector Relations
A .B C/ D C .A B/ D B .C A/ (F.1)
A .B C/ D B .A C/ C .A B/ (F.2)
d dB dA
.A B/ D A C B (F.3)
dt dt dt
d dB dA
.A B/ D A C B (F.4)
dt dt dt
Z I
.r A/ dV D A da Gauss’ Divergence theorem (F.5)
Z I
.r A/ da D A d` Stokes’ theorem (F.6)

Appendix G
Useful Integrals
Z
eax 2x 2
x2 eax dx D x2 C C 2 (G.1)
a a a
Z 1
Œ.m C 1/ mŠ
xm eax dx D mC1
D mC1 (G.2)
0 a a
Z 1 r
2
eax dx D (G.3)
1 a
Z 1
2 Œ.m C 1/ =2
xm eax dx D (G.4)
0 2a.mC1/=2
Z p p
x a2 x2 a2 x
a2 x2 dx D C sin1 (G.5)
2 2 a
Z 1 p

1 u2 du D (G.6)
1 2
Z
dx x
p D sin1 (G.7)
a2 x2 a
Z
cos .˛x/ sin .˛x/
x cos .˛x/ dx D 2
Cx (G.8)
˛ ˛
Z 2
x x sin .2˛x/ cos .2˛x/
xsin2 .˛x/ dx D (G.9)
4 4˛ 8˛ 2
Z
x3 x2 1 x cos .2x/
x2 sin2 xdx D sin .2x/ (G.10)
6 4 8 4
Z
sin .ax/ sin3 .ax/
cos3 .ax/ dx D (G.11)
a 3a

328 G Useful Integrals
Z
3x sin .2ax/ sin .4ax/
cos4 .ax/ dx D C C (G.12)
8 4a 32a
Z
cos ax cos3 ax
sin3 .ax/ dx D C (G.13)
a 3a
Z
sin ax x cos ax
x sin .˛x/ dx D (G.14)
a2 a
Z
x2 x sin .2˛x/ cos .2˛x/
xsin2 .˛x/ dx D (G.15)
4 4˛ 8˛ 2
Appendix H
Useful Series
H.1 Taylor Series
The Taylor series expansion of a function f .x/ about a point x D a is
.x a/ 0 .x a/2 00
f .x/ D f .a/ C f .a/ C f .a/ C
1Š 2Š
1
X .x a/n
D f .n/ .a/ (H.1)
nD0
nŠ
where the primes signify differentiation with respect to x. For example, f 00 .a/ is the
second derivative of the function f .x/ with respect to x evaluated at x D a.
There are at least four Taylor series that every physics student should have at
their command. These series, Eqs. (H.2) through (H.5) below, are easy to remember
and are frequently used in problems.
x x2 x3
ex D 1 C C C C (H.2)
1Š 2Š 3Š
x x3 x5 x7
sin x D C C (H.3)
1Š 3Š 5Š 7Š
x2 x4 x6
cos x D 1 C C (H.4)
2Š 4Š 6Š
x2 x3 x4
ln .1 C x/ D x C C (H.5)
2 3 4
Two other useful series are those for the hyperbolic sine and cosine, which are
identical to those for the circular sine and cosine.

330 H Useful Series
H.2 Binomial Expansion
Binomial series are special cases of Taylor series for f .x/ D .1 C x/m and a D 0.
The exponent m may be positive or negative and is not restricted to integer values.
The binomial expansion is, in three equivalent forms,
m .m 1/ 2 m .m 1/ .m 2/ 3
.1 C x/m D 1 C mx C x C x C
2Š 3Š
1
X mŠ
D xn
nD0
nŠ .m n/Š
1
!
X m n
D x (H.6)
nD0
n
where
!
m mŠ
(H.7)
n nŠ .m n/Š
is called the binomial coefficient. A few of the most common binomial expansions
are listed below:
.1 C x/1 D 1 x C x2 x3 C (H.8)
.1 C x/2 D 1 2x C 3x2 4x3 C (H.9)
1 1 2 1
.1 C x/1=2 D 1 C x x C x3 C (H.10)
2 24 246
1 13 2 135 3
.1 C x/1=2 D 1 x C x x C (H.11)
2 24 246
Appendix I
-Functions
I.1 Integral -Functions

.n/ D .n 1/Š n D 1; 2; 3; : : : (I.1)
.n C 1/ D n .n/ D nŠ n D 1; 2; 3; : : : (I.2)
.1/ D 1 (I.3)
.2/ D 1 (I.4)
.3/ D 2 (I.5)
.4/ D 3Š D 6 (I.6)
.5/ D 4Š D 24 (I.7)
I.2 Half-Integral -Functions

1 1 3 5 .2m 1/ p
mC D m D 1; 2; 3; : : : (I.8)
2 2m
p
.1=2/ D (I.9)
p

.3=2/ D (I.10)
2
p
3
.5=2/ D (I.11)
4
p
15
.7=2/ D (I.12)
8

Appendix J
The Dirac Delta-Function
The Dirac ı-function is defined as follows:

Z 1
1 0
ı x x0 eik.xx / dk (J.1)
2 1
The dimension of the ı-function is the reciprocal of the variable of integration.

For example, the units of ı .x/ are 1/length.
See Ref. [3], page 665 for more information on the Dirac ı-function.
Table J.1 Some properties Mathematical operation Property

of the Dirac delta-functions R1
f .x0 / D 1 ı .x x0 / f .x/ dx Sifting property
ı .x/ D ı .x/ Parity: even
R1
1 ı .x x0 / dx D 1 Normalization
ı .ax/ D .1= jaj/ ı .x/ Scaling

Appendix K
Hyperbolic Functions
K.1 Manipulations of Hyperbolic Functions
Some proofs of relations involving hyperbolic functions.

Show that:
p
cosh1 x D ln x C x2 1 (K.1)
Proof. Let
x D cosh y y>0
y
e Ce
y
D
2
1 2y
D y e C1 (K.2)
2e
or
e2y 2xey C 1 D 0 (K.3)
Using the quadratic formula

p
ey D x ˙ x2 1 (K.4)
or
p
y D cosh1 x D ln x C x2 1 (K.5)
where the minus sign has been dropped because the principal values of the cosh1 x
are positive.

336 K Hyperbolic Functions
Show that:

1 1 1Cx
tanh x D ln (K.6)
2 1x
Proof. Let
x D tanh y
ey ey
D
ey C ey
e2y 1
D (K.7)
e2y C 1
Then
xe2y C x D e2y 1
xe2y e2y D .1 C x/
e2y .x 1/ D .1 C x/ (K.8)
so

2y 1Cx
e D (K.9)
1x
or

1 1Cx
y D tanh1 x D ln (K.10)
2 1x
Show that:
1
tanh1 z D cosh1 p (K.11)
1 z2
p
Proof. Let x D 1= 1 z2 in Eq. (K.5).
r !
1 1 1
cosh1 p D ln p C 1
1z2 1z 2 1 z2
r
1Cz 1Cz
D ln p D ln
1z2 1z

1 1Cz
D ln
2 1z
D tanh1 z (K.12)
where the last step utilized Equation (K.10).
K Hyperbolic Functions 337
K.2 Relationships Between Hyperbolic and Circular

Functions
sin .ix/ D i sinh x (K.13)
sinh .ix/ D i sin x (K.14)
cos .ix/ D cosh x (K.15)
cosh .ix/ D cos x (K.16)
Appendix L
Useful Formulas
L.1 Classical Mechanics
The Lagrangian
LDT U (L.1)
Lagrange’s Equation

@L d @L
D0 (L.2)
@qi dt @Pqi
Lagrange’s Equation with undetermined multiplier
@L d @L @f .q1 ; q2 /
C D0 (L.3)
@qi dt @Pqi @qi
The Hamiltonian
X
H .qi ; pi ; t/ D qP j pj L .qi ; pi ; t/ (L.4)
j
The Hamiltonian is the TME under the following conditions:

(a) The transformation equations from Cartesian to generalized coordinates are time
independent.
(b) The potential energy contains only the coordinates and not the velocity.

340 L Classical Mechanics
Under these conditions:
pO 2x
HD C U .x/ (the total energy) (L.5)
2m
L.2 Quantum Mechanics
The de Broglie wavelength
h
de Broglie D
p
h
D p ; E D kinetic energy (L.6)
2mE
L.2.1 One Dimension
Time dependent Schrödinger Equation (TDSE)

„2 @2 „ @‰ .x; t/
2
C U .x; t/ ‰ .x; t/ D (L.7)
2m @x i @t
Time independent Schrödinger Equation (TISE)

„2 d2
C U .x/ .x/ D E .x/ (L.8)
2m dx2
TISE in terms of the Hamiltonian
HO .x/ D E .x/
2
pO x
D C U .x/ .x/ (L.9)
2m
Definition of the probability current

„ @‰ .x; t/ @‰ .x; t/
j .x; t/ D ‰ .x; t/ ‰ .x; t/ (L.10)
2im @x @x
L Classical Mechanics 341
The equation of continuity for probability current j .x; t/
@ j‰ .x; t/j2 @
C j .x; t/ D 0 (L.11)
@t @x
Probability current for an energy eigenfunction (time independent)

„ @ .x/ @ .x/
j .x/ D .x/ .x/ (L.12)
2im @x @x
The equation of continuity for time independent probability current j .x/
@
j .x/ D 0 (L.13)
@x
L.2.2 Three Dimensions (Central Potentials)
The radial TISE where Rn` is the radial part of the eigenfunctions and the
angular parts are the spherical harmonics Y`m . ; /

„2 d 2 d ` .` C 1/ „2
r C C U .r/ Rn` .r/ D En` Rn` .r/ (L.14)
2mr2 dr dr 2mr2
The radial TISE with u .r/ D rR .r/

„2 d2 u .r/ ` .` C 1/ „2
C C U .r/ u .r/ D Eu .r/ (L.15)
2m dr2 2mr2
The effective potential
` .` C 1/ „2
Ueff .r; `/ D C U .r/ (L.16)
2mr2
Interparticle distance r12 in terms of spherical harmonics

1 `
1 1X r<
D P` .cos /
r12 r> r>
`D0
X
1 X̀ `
4 1 r<
D Y`m . 1 ; 1 / Y`m . 2 ; 2 / (L.17)
2` C 1 r> r>
`D0 mD`
Appendix M
The Infinite Square Well
The time honored one-dimensional problem of a particle trapped in an infinite

square well potential (also known as a particle-in-a-box) is frequently formulated
in one of two different configurations which we refer to as an L-box and an a-box.
Often one of these configurations is more advantageous for problem solving than
the other.
M.1 The L-Box
In this configuration the potential energy is given by
U .x/ D 0 0xL
D1 1<x<0;L<x<1 (M.1)
which is shown in Fig. M.1
Fig. M.1 The L-box

configuration

344 M The Infinite Square Well
The eigenfunctions and energy eigenvalues are

nx
n .x/ D A sin 0 x L n D 1; 2; 3 : : :
L
D0 1<x<0;L<x<1 (M.2)
and
n2 2 „2
En D (M.3)
2mL2
Advantage: The eigenfunctions are the same function for each level, sines.
Disadvantage: The symmetry of the potential well about x D L=2 is not helpful
in evaluating integrals in problems involving this configuration.
M.2 The a-Box
In this configuration the potential energy is given by
U .x/ D 0 axa
D1 1 < x < a=2 ; a=2 < x < 1 (M.4)
which is shown in Fig. M.2.
Fig. M.2 The a-box

configuration
The eigenfunctions are

r nx
2 a a
n .x/ D cos x n D 1; 3; 5 : : : (even parity)
a a 2 2
M The Infinite Square Well 345
r nx
2 a a
n .x/ D sin x n D 2; 4; 6 : : : even (odd parity)
a a 2 2
D0 1 < x < a=2I a=2 < x < 1 for all n (M.5)
and the energy eigenvalues are the same as they are for the L-box with L ! a.
Advantage: The symmetry of the well and the accompanying definite parity of
the eigenfunctions can be quite useful in problems involving this configuration.
Disadvantage: The eigenfunctions alternate between cosine (odd n) and sine
(even n) making evaluation of integrals slightly more challenging than it is for the
L-box.
Appendix N
Operators, Eigenfunctions, and Commutators
N.1 Eigenfunctions and Eigenvalues of Operators
An operator is a mathematical object that when applied to a function gives a new

function. If we have a function f .x/ and an operator A, O then Af
O .x/ is a some new
O .x/ D
function .x/. In some cases .x/ is proportional to f .x/; this means that Af
af .x/ where a is a constant. In this case f .x/ is called an eigenfunction of AO and a
is the corresponding eigenvalue.
An operator AO is linear if AO Œf .x/ C g .x/ DAfO .x/ C AgO .x/ where f .x/ and
g .x/ are any two appropriate functions. The operators of importance in quantum
mechanics, such as position xO , linear momentum pO , and the Hamiltonian H, O are all
linear operators. In quantum mechanics observable quantities such as xO ; pO , and HO
are represented by Hermitian operators . An operator AO is Hermitian if
Z h i Z h i
A O .x/ dx
O .x/ .x/ dx D .x/ A (N.1)
Hermitian operators have real eigenvalues, and conversely real eigenvalues result
from Hermitian operators . The operators xO ; pO , and HO are Hermitian.
N.2 Operator Algebra; Commutators
The product of two operators is defined by operating with them on a function. Let
the operators be AO and BO operate on f .x/. The expression AOBf O .x/ is a new function,
and AOBO is an operator which we call C,O that is defined as the product of AO and B.
O The
meaning of AOBf O .x/ is that BO operates first on f .x/, giving a new function, and then

348 N Operators, Eigenfunctions, and Commutators
AO operates on that new function. Combinations of operators of the form AOBO BOAO
frequently arise in quantum mechanical calculations. An abbreviated notation that
is customarily used is
h i
AOBO BOAO A;
O BO (N.2)
h i h i
The symbol A; O BO is referred to as the commutator of AO and B.
O If A;O BO ¤ 0, then AO
h i
and BO do not commute; if A; O BO D 0 , then AO and BO commute. If two operators AO and
BO commute, then they have common eigenfunctions. This means that the physical
quantities associated with AO and BO can (in principle) be measured simultaneously
and exactly. If AO and BO do not commute, the associated physical quantities cannot
be accurately measured together. The classic examples of non-commuting operators
are position and momentum; the operators, xO D i„d=dp and p D i„d=dx do
not commute, and therefore the position and momentum of a particle cannot be
simultaneously obtained—the uncertainty principle.
N.3 Commutator Identities

h i
A;O BO AOBO BOAO (N.3)
h i
O AO D 0
A; (N.4)
h i h i h i
O BOCO A;
A; O BO CO C BO A;
O CO (N.5)
h i h i h i
AOB;
O CO A;O CO BO C AO B;O CO (N.6)
h i h i h i
AOB;
O CO DO A;
O CO BODO C AO B;O CO DO
h i h i
CCO A; O DO BO C CO AO B;
O DO (N.7)
N.4 Some Quantum Mechanical Commutators

Position and momentum: Œx; pO D i„ (N.8)

Position and powers of momentum: x; pO nx D i„nOpn1
x (N.9)
h i
Components of angular momentum: JOi ; JOj D i„JOk ijk (N.10)
Appendix O
The Quantum Mechanical Harmonic Oscillator
Together with the particle-in-a-box the most often studied problem in one-
dimensional bound state quantum mechanics is that of a particle of mass m under
the influence of a quadratic potential, that is
1 2
U .x/ D kx
2
1
D m! 2 (O.1)
2
p
where k is the force constant and ! D k=m, the harmonic frequency. One of the
most powerful methods for attacking this problem is the use of ladder operators.
This subject is summarized in Sect. 7 of this book. In this appendix we summarize
the eigenfunctions of the harmonic oscillator which, because the potential is an
even function of x, alternate between even and odd functions. The oscillator
wave functions are always the product of a Hermite polynomial and a Gaussian
exponential. The first four normalized eigenfunctions are listed in Table O.1.
where
r s r
m! m k1=2 1=4 1=4 1
˛D D 1=2
Dm k (O.2)
„ „ m „
Note that the units of ˛ are m1 .

Table O.2 summarizes the properties of the harmonic oscillator ladder operators
discussed in Sect. 7.

350 O The Quantum Mechanical Harmonic Oscillator
r
Table O.1 The first four ˛ 2 2
normalized harmonic 0 .x/ D p e˛ x =2
r
oscillator eigenfunctions ˛ 2 2
1 .x/ D p Œ2 .˛x/ e˛ x =2
r 2
˛ h i
p 2 .˛x/2 1 e˛ x =2
2 2
2 .x/ D
r 2
˛ h i
p 2 .˛x/3 3 .˛x/ e˛ x =2
2 2
3 .x/ D
3

Table O.2 Relations 1 1
involving the ladder operators aO D p ˛Ox C i pO
2 ˛„
1 1
aO D p ˛Ox i pO
2 ˛„
1
xO D p aO C aO
2˛
NO D aO aO
1 1
HO D „! aO aO C D „! NO C

2 2
haO ; aO i D 1 h i
O aO D Oa H) H;
N; O aO D „! aO
h i h i
O aO D aO H) H;
N; O aO D „! aO
p
aO jni D n jn 1i (lowering)
p
aO jni D n C 1 jn C 1i (raising)
Appendix P
Legendre Polynomials
P.1 Properties
The Legendre polynomials P` .cos / D P` .x/ are used in a variety of applications

in which there is azimuthal symmetry, that is, no dependence in spherical
coordinates. They are the solutions of Legendre’s differential equation
d2 y dy
1 x2 2x C ` .` C 1/ y D 0 (P.1)
dx2 dx
We present here a few properties of Legendre polynomials as well as the first few
of these polynomials that are useful for the problems in this book. The substitution
x D cos is conventional and made for convenience. The variable x is in no way
related to the Cartesian coordinate. More detailed information can be found in [1, 3]
and [4].
P.2 Legendre Series
A function f .x/ that is defined on the interval .1; 1/ may be expanded in a series of
Legendre polynomials because they form a complete set. Thus, using the properties
in Table P.1
1
X
f .x/ D a` P` .x/
`D0
Z 1
2` C 1
where a` D f .x/ P` .x/ dx .` D 0; 1; 2; / (P.2)
2 1

352 P Legendre Polynomials
Table P.1 Some properties of Legendre polynomials

1 P1
Generating function p D P` .x/ t`
1 2tx C t2 `D0 `
1 d 2 `
Rodrigues’ formula P` .x/ D ` x 1
2 `Š dx
R1 2
Orthogonality 1 P ` .x/ P`0 .x/ dx D ı ``0
2` C 1
`
Parity P` .x/ D ./ P` .x/
Value at x D 0 P` .0/ D 0 ` odd
1 3 5 .` 1/
D ./`=2 ` even
2 4 4`
Value at x D ˙1 P` .1/ D 1
P` .1/ D .1/`
P.3 The Function 1= jr1 r2 j
There is a useful relationship between the quantity 1= jr1 r2 j and the Legendre
polynomials. This quantity arises in many physical applications. The geometry is
shown in Fig. P.1.
Fig. P.1 Coordinates for

deriving the expression for
1= jr1 r2 j in terms of
Legendre polynomials. The
angle between the two
directions r1 and r2 is
Applying the law of cosines we write 1= jr1 r2 j in terms of the angle .
1 1=2
D r12 C r22 2r1 r2 cos (P.3)
jr1 r2 j
If r1 is placed along the z-axis, then is the spherical coordinate . Now assume
that r2 > r1 . We may write Eq. (P.3) as
" 2 #1=2
1 1 r1 r1
D 1C 2 cos (P.4)
jr1 r2 j r2 r2 r2
P Legendre Polynomials 353
Comparing Eq. (P.4) with the first entry in Table P.1 we recognize the generating
function for P` ./. Letting cos D and t D r1 =r2 we write
1 `
1 1X r1
D P` ./ (P.5)
jr1 r2 j r2 r2
`D0
If we had r1 > r2 rather than r2 > r1 , then the subscripts in Eq. (P.5) would be
reversed. The combined equation for both cases is usually written as
1 `
1 1X r<
D P` ./ (P.6)
jr1 r2 j r> r>
`D0
where r> and r< designate the larger and smaller of r1 and r2 .
P.4 Polynomials
Table P.2 Some Legendre P0 .x/ D 1

polynomials
P1 .x/ D x

P2 .x/ D .1=2/ 3x2 1
3
P3 .x/ D .1=2/ 5x 3x
4
P4 .x/ D .1=8/ 35x 30x2 C 3
5
P5 .x/ D .1=8/ 63x 70x3 C 15x
Appendix Q
Orbital Angular Momentum Operators in
Spherical Coordinates

„ @ @
LOx D sin C cot cos (Q.1)
i @ @

O „ @ @
Ly D cos cot sin (Q.2)
i @ @
„ @
LOz D (Q.3)
i @

@ @
LOC D „ei C i cot (Q.4)
@ @

O i @ @
L D „e i cot (Q.5)
@ @

Appendix R
Spherical Harmonics
The spherical harmonics Y`m . ; / are the eigenfunctions of the angular momentum
2
operator LO . They are given by [1]
s
.2` C 1/ .` m/Š m
Y`m .; / D ./ m
P .cos / eim (R.1)
4 .` C m/Š `
where the Pm ` .cos / are associated Legendre functions. Spherical harmonics have
definite parity which is determined by the value of `. Simply
parity of Y`m . ; / D .1/` (R.2)
Below is a list of some spherical harmonics (Table R.1).
Table R.1 The first few Y`m . ; / Spherical harmonic

spherical harmonics p
Y00 . ; / 1=4
p
Y10 . ; / 3=4 cos
p
Y1 ˙1 .; / 3=8 sin e˙i
p
Y20 . ; / 5=16 3 cos2 1
p
Y2 ˙1 .; / 15=8 cos sin e˙i
p
Y2 ˙2 .; / 15=32 sin2 e˙2i

358 R Spherical Harmonics
The Addition Theorem for Spherical Harmonics
The addition theorem for spherical harmonics is a useful formula when there are
two specified directions as shown in Fig. P.1. The theorem expresses the Legendre
polynomial of the angle in terms of the spherical harmonics for each of the
directions r1 and r2 . The formula is
4 X̀
P` .cos / D Y`m . 1 ; 1 / Y`m . 2 ; 2 / (R.3)
2` C 1
mD`
Appendix S
Clebsch–Gordan Tables
Table S.1 Clebsch–Gordan coefficients for any value of j1 and j2 D 1=2

j ms D 1=2 ms D 1=2
q q
j1 C 1=2 j1 C 1=2 C mj = .2j1 C 1/ j1 C 1=2 mj = .2j1 C 1/
q q
j1 1=2 j1 C 1=2 mj = .2j1 C 1/ j1 C 1=2 C mj = .2j1 C 1/
Table S.2 Clebsch–Gordan S ms2 D 1=2 ms2 D 1=2

coefficients for two spin-1/2 p p
particles 1 .1 C M/ =2 .1 M/ =2
p p
0 1= 2 1= 2
Table S.3 Clebsch–Gordan coefficients for j1 D ` D 1 and j2 D 1=2

j1 D 1I j2 D 1=2 j D 3=2 j D 1=2
m` ms 3=2 1=2 1=2 3=2 1=2 1=2
1 1=2 1
p p
1 1=2 1=3 2=3
p p
0 1=2 2=3 1=3
p p
0 1=2 2=3 1=3
p p
1 1=2 1=3 2=3
1 1=2 1

Appendix T
The Hydrogen Atom
This appendix summarizes the properties of the quantum mechanical H-atom (one-
electron atom, Z D 1). The Coulomb potential

1 e2 1
U .r/ D D in a.u. (T.1)
4 0 r r
is a central potential so the angular parts of the eigenfunctions are the spherical
harmonics Y`m .; /. The H-atom quantum numbers are therefore the usual angular
quantum numbers ` and m which represent the total and z-components of the
angular momentum. The energy quantum number is n. The energy eigenvalues,
En D 1=2n2 (a.u.), depend upon the single quantum number n rather than n and `
(spherical symmetry precludes m). This “accidental degeneracy ” is the result of an
“extra” symmetry of the Coulomb potential that eliminates the dependence of the
energy on ` [2].
When discussing the H-atom the Greek letter ˛ is almost universally used to
designate one of the most important constants in physics, the unitless fine structure
constant ˛. This constant is

e2
˛D
.4 0 / „c
1
' (T.2)
137
In Tables T.1, T.2, and T.3 are some important quantities and properties of the
one-electron atom with nuclear charge Z.

362 T The Hydrogen Atom
Table T.1 Relationships Name of quantum no. Notation Range

between the one-electron
atom quantum numbers Principal n n1
(integers) Azimuthal (ang. momentum) ` 0 ` .n 1/
Magnetic m (or m` ) ` m` `
Spin quantum numbers not included
Table T.2 Properties of the quantum mechanical one-electron atom

Ze2 1 1
Potential U .r/ D D Z .˛„c/
4
2 2 0 r r
1 2
2
2 1
Energy eigenvalues En D 4 0 2n2 a0 D Z me c ˛ 2
Z e
2n
Eigenfunctions n ` m n`m .r; ; /
3=2
1 Z
1 0 .s/ 0 p eZr=a0
a0
3=2
1 Z Zr
2 0 .s/ 0 p 1 eZr=2a0
2 2 a0 2a0
3=2
1 Z Zr
2 1 .p/ 1 p eZr=2a0 sin ei
4 a0 2a0
3=2
1 Z Zr
2 1 .p/ 0 p eZr=2a0 cos
2 2 a0 2a0
3=2
1 Z Zr
2 1 .p/ 1 p eZr=2a0 sin ei
4 a0 2a0
Table T.3 Expectation values of rs for the one electron atom

˝ 2 ˛ a0 2 n2
2
r D 5n C 1 3` .` C 1/
2
a 1

Z
0
hri D 3n2 ` .` C 1/
Z 2
˝ 1 ˛ 1 1
r D
.a0 =Z/ (n2 )
˝ 2 ˛ 1 1
r D
.a0 =Z/2 ( n3 ` C 12 )
˝ 3 ˛ 1 1
r D
.a0 =Z/3 n3 ` ` C 12 .` C 1/
.s C 1/ s ˝ ˛ s h i˝ ˛
2
hr i .2s C 1/ a0 rs1 C a20 .2` C 1/2 s2 rs2 D 0
n 4
The last entry is Kramer’s relation for rs
References
1. Arfken GB, Weber HJ (2005) Mathematical methods for physicists, 6th edn. Elsevier, New York
2. Burkhardt CE, Leventhal JJ (2008) Foundations of quantum physics. Springer, New York
3. Boas ML (1983) Mathematical methods in the physical sciences. Wiley, New York
4. Spiegel MR, Lipschutz S, Liu J (1961) Mathematical handbook of formulas and tables.
Macmillan, New York
Index
A model, 121, 122

acceleration, 18 orbit, 122, 125, 127, 317
angular, 5, 112 radius, 127, 128, 231, 317
gravitational, 14, 15, 18, 40–42, 50, 52, 54, brachistochrone, 27
58–62, 64
angular momentum
classical, 26, 67, 68, 71, 73, 76, 80, 81, 83, C
89 calculus of variations, 25, 27, 30
quantum, 121, 124, 126, 189–196, 198– canonically conjugate, 26, 27
200, 208, 213–215, 227, 233–235, center of mass, 112
237, 239, 242, 270, 271, 289, 317, central forces
357, 361 classical, 67, 69–72, 79, 81, 83, 89
ladder operators, 190, 192–196, 202, quantum, 223
205, 210 central potential
apocenter, 88, 92, 93 classical, 67–73, 81, 83, 90
approximation methods, 243, 280, 281 quantum, 223, 226, 227, 233, 234, 341
degenerate perturbation theory, 284–286, centrifugal potential
288, 292, 294 classical, 67
time dependent perturbation theory, 297, quantum, 240
298, 300, 303, 305, 307, 310 Clebsch-Gordan
time independent perturbation theory, 258, coefficients, 192, 205–208
261–263, 267, 270, 271, 273, 277, table, 206–208, 359
280, 281 commutator, 174, 184, 193, 205, 348
variational method, 248, 249, 251, conic sections, 85, 89, 319
253–256, 282, 283 constraint
WKB, 243–247 equation of, 58, 61–63
Atwood’s machine, 40, 58 undetermined multiplier, 59–61
cyclic coordinates, 27, 44, 69, 70
B
barrier, 137–140, 142–146
Bertrand’s theorem, 72, 89 D
Bohr, 121 de Broglie wavelength, 139, 340
correspondence principle, 121, 126, 137 deBroglie wavelength, 126

364 Index
degeneracy Hermitian operator, 347

classical, 94 hydrogen molecule, 271
quantum mechanical, 149, 223, 225–227,
237, 242, 289, 361
density I
mass, 13, 15 inclined plane, 18, 52, 60
of states, 226 indistinguishable particles, 217
probability, 135, 138 boson, 218, 219
determinant of coefficients, 100, 115, 229, 296 fermion, 218–220
dipole
electric, 309
magnetic, 129 K
dipole moment Kepler
magnetic, 129 laws of planetary motion, 69, 71, 72
Dirac orbit, 92–94, 320
ı-function, 142–144, 156, 157, 170, 260, potential, 83, 89, 320
270, 281, 288, 289, 333 problem, 84, 85, 89–92, 319
notation, 179, 182, 281, 292, 298, 301, 308
L
E Lagrangian, 3, 19, 25–27, 30–36, 38–45, 47,
effective potential 49–56, 59, 60, 62, 65, 68–70, 73, 97,
classical, 67, 81, 82, 84, 87 104, 105, 109, 113–115
quantum, 230, 235, 237, 240 Levi–Cevita symbol, 189
Ehrenfest theorem, 181, 182
elastic collision, 8, 11, 22, 23, 213
O
orbit, 67, 71–74, 81, 82, 89, 92, 94, 126
G Bohr, 121–123, 125, 126, 155
-functions circular, 74, 75, 81, 84, 85, 92, 94
half-integral, 331 elliptical, 69, 71, 83, 85, 87–90, 92, 94
integral, 134, 331 equation of, 78–80, 85, 89
Gauss hyperbolic, 88
law, 13, 14, 16 parabolic, 88
trick, 238 spiral, 76–80
H P
Hamiltonian pendulum, 3, 33, 35, 54, 99, 108
classical mechanics, 26, 27, 32, 33, 43–48, double, 103
56, 58 pericenter, 90, 92, 93
quantum mechanics, 131, 132, 176, 179, polar coordinates, 67, 68, 71, 72, 81, 82, 319,
181, 182, 203, 204, 209–211, 220, 320
224, 225, 227, 248, 256, 258, 276, Principal quantum number, 124
279–281, 283–289, 291–294, 297, projectile, 47
300, 307
relativistic, 263–266, 271, 272
harmonic oscillator Q
classical, 31 quantum defect, 239, 240, 242
isotropic, 81–83
quantum, 132, 157, 170
isotropic, 292 R
ladder operators, 173–175, 179, reflection coefficient, 137, 139, 143, 146
181–183, 185–187, 258 Runge-Lenz vector, 89, 92
Index 365
S T
secular equation, 100, 103, 106, 110, 115, 279, transmission coefficient, 137, 139, 143, 144,
288, 290, 295 146, 244, 247
spring, 10, 11
superposition
classical, 16 U
quantum, 151, 160, 167 uncertainty principle, 126, 348

Kelley, J. Daniel - Leventhal, Jacob Joseph-Problems in Classical and Quantum Mechanics - Extracting The Underlying Concepts-Springer (2017)

Uploaded by

Copyright:

Available Formats

Kelley, J. Daniel - Leventhal, Jacob Joseph-Problems in Classical and Quantum Mechanics - Extracting The Underlying Concepts-Springer (2017)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Kelley, J. Daniel - Leventhal, Jacob Joseph-Problems in Classical and Quantum Mechanics - Extracting The Underlying Concepts-Springer (2017)

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What topics are covered in the book?

What topics are covered in the book?

J. Daniel Kelley · Jacob J.

ISBN 978-3-319-46662-0 ISBN 978-3-319-46664-4 (eBook)

Library of Congress Control Number: 2016951927

© Springer International Publishing AG 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

This book is not a textbook. It is a collection of problems intended to aid students in

Advice for Students

St. Louis, MO, USA J. Daniel Kelley

Part I Classical Mechanics

16. Rotating massless rod - Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Part II Quantum Mechanics

12. Transmission and reflection over a square well . . . . . . . . . . . . . . . . . 141

9 Indistinguishable Particles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

3. 2-D SHO with xy perturbation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292

A Greek Alphabet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

B Acronyms, Descriptors and Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

D Conic Sections in Polar Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319

E Useful Trigonometric Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

F Useful Vector Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

G Useful Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327

H Useful Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329

J The Dirac Delta-Function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333

K Hyperbolic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

L Useful Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339

M The Infinite Square Well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343

N Operators, Eigenfunctions, and Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . 347

O The Quantum Mechanical Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . . . . . 349

P Legendre Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351

Q Orbital Angular Momentum Operators in Spherical Coordinates . . . 355

R Spherical Harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357

S Clebsch–Gordan Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359

T The Hydrogen Atom. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

The force along the arc length s D ` is simply mg sin so according to

mRs D mg sin (1.1)

© Springer International Publishing AG 2017 3

Fig. 1.1 Problem 1

But, because s and ` are related according to s D ` we have

In the case of small oscillations sin Ð so that

.t/ D max sin .!t / (1.4)

For small oscillations

D r F D `mg sin (1.8)

which produces the same differential equation, Eq. (1.2).

2. A particle is dropped into a viscous fluid from rest at y D 0 and t D 0. Take

Fig. 1.2 Problem 2

Using Fig. 1.2 and F D ma we have

Separating this differential equation we have

Before proceeding we note that the units of k must be s1 . As we proceed

Using the initial conditions to evaluate the constant of integration we find

Now we must eliminate t to obtain y D y .v/. To do this we solve Eq. (1.15)

If there were no fluid resistance the result is simple to calculate using

3. A mass m2 is at rest a distance d from a brick wall. It is struck by a mass

Fig. 1.3 Problem 3

Fig. 1.4 Problem 3—

First, in terms of v1i , the initial velocity of m1 find the post-first-collision

m1 xR 1 D k Œx1 .x2 `/

Notice that if D 0 or if D =2, XR vanishes as it should. Moreover, as

The motion is restricted to the range =2 < x < 3=2.