A. Z. Capri-Problems & Solutions in Nonrelativistic Quantum Mechanics-World Scientific (2002)

(pro 6 Cents
ions 1
i n /
NONRELATIVISTIC
QUANTUM
MECHANICS
Anton Z. Capri
*
1 i
•
World Scientific
cpfo 6kwis
Soíupons
NONRELATIVISTIC
QUANTUM
MECHANICS
Anton Z. Capri
Department of Physics
University of Alberta, Canada
V
\
<y 2 M
Published by
World Scientific Publishing Co. Pte. Ltd.

P O Box 128, Farrer Road, Singapore 912805
USA office: Suite 202, 1060 Main Street, River Edge, NJ 07661
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
Library of Congress Cataloging-in-Publication Data

Capri, Anton Z.
Problems & solutions in nonrelativistic quantum mechanics / Anton Z. Capri.
p. cm.
Includes bibliographical references.
ISBN 9810246331 (alk. paper) — ISBN 9810246501 (pbk.: alk. paper)
1. Nonrelativistic quantum mechanics - Problems, exercises, etc. I. Title: Problems and
solutions in nonrelativistic quantum mechanics. II. Title: Nonrelativistic quantum
mechanics. III. Title.
QC174.24.N64 C374 2002

530.12'076--dc21 2002029614
British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.
Copyright © 2002 by World Scientific Publishing Co. Pte. Ltd.

All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the Publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright
Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to
photocopy is not required from the publisher.
This book is printed on acid-free paper.

To bkaidrite, who knows that physics is simple because "everything equals
zero".
Preface
Soon after the first edition of Nonrelativistic Quantum Mechanics appeared,

I received numerous requests for solutions to the problems in that book. To
remedy this situation I started by writing out solutions to the more difficult
problems, but as I proceeded with the third edition of Nonrelativistic Quantum
Mechanics I also revised some of the problems and added quite a few others.
Since in constructing these new problems I had to solve them, in the first place,
to be sure that they were indeed problems that students could solve, I finally
went on to write out solutions to all the problems. However, I did not simply
want a compendium of solutions of the Schrodinger equation since with pro-
grams such as Maple or Mathematica these solutions are accessible to every
student. Instead I wanted to concentrate on problems that teach quantum me-
chanics. It is with this in mind that I began to collect and solve problems. My
idea was to provide a means for students to learn quantum mechanics by "doing
it". This is why the book begins with extremely simple problems and progresses
to more difficult ones.
Some of the problems extend results that are usually taught in a course on
quantum mechanics. But, by having the students obtain the results themselves
they are more likely to retain the ideas and at the same time gain confidence in
their own abilities.
As usual, I tested most of these problems on my students. Sometimes they
came up with very original ways of looking at old problems. I have learned a lot
from my students. It is this learning process that led me to occasionally intro-
duce more than one way of solving a problem since the solutions are intended
to help students to obtain a better understanding of the techniques involved in
tackling problems in quantum mechanics.
The notation and methods used are those explained in Nonrelativistic Quan-
tum Mechanics and I frequently refer to chapters from that book. The chapter
headings are also the same as in Nonrelativistic Quantum Mechanics. Never-
theless, the present book is independent and should serve as a companion to
any of the numerous excellent books on quantum mechanics. Throughout the
book I have used Gaussian units since these are the units most commonly used
in atomic physics. I also tried to arrange the problems according to increasing
degree of difficulty. This, was not always possible since it would have meant
losing the possibility of arranging them according to topic.
It is a pleasure to thank Professor M. Razavy for his generous help in, not
only providing me with some wonderful problems and supplying me with nu-
merous references, but also for his constant moral support.
Of course the students who suffered through the courses in which I subjected
them to all sorts of quantum problems also deserve my heartfelt thanks. To their
credit, the undergraduates seldom complained. On the other hand, there was
many an evening, after I had assigned some more than usually difficult problems
in the graduate course on quantum mechanics, that walking down the hall of
the fourth floor of the physics building I heard my name muttered with less than
flattering epithets. Nevertheless, the graduate students survived and many, after
they completed their degree, even thanked me for what they had learned.
It is my hope that these problems and solutions will be of use to future
generations of physics students. At any rate they should provide more enter-
tainment than solving cross word puzzles.
A.Z.Capri
Edmonton, Alberta
July, 2002.
Contents
1 The B r e a k d o w n of Classical Mechanics 1

1.1 Quantum Number of the Earth 1
1.2 Thermal Wavelength 2
1.3 Photons in a Beam 2
1.4 Hydrogen Atom and de Broglie 3
1.5 Vibrations in NaCl 4
1.6 Crystal Powder 4
1.7 Einstein Coefficients 5
2 R e v i e w of Classical Mechanics 8
2.1 Lagrangian and Hamiltonian for SHO 8
2.2 Lagrangian and Hamiltonian: Simple Pendulum 9
2.3 Bohr-Sommerfeld Quantization: SHO 9
2.4 Bohr-Sommerfeld: Particle in a Box 10
2.5 Larmor Frequency 11
2.6 Applicability of Bohr-Sommerfeld Quantization 12
2.7 Schrodinger and Hamilton-Jacobi 12
2.8 WKB Approximation 13
2.9 Dumbbell Molecule: Bohr-Sommerfeld 14
3 Elementary Systems 15
3.1 Commutator Identities 15
3.2 Complex Potential 16
3.3 Group and Phase Velocity 17
3.4 Linear Operators 18
3.5 Probability Density 19
3.6 Angular Momentum Operators 20
3.7 Beam of Particles 21
3.8 Time Evolution of Wave Function 22
3.9 Operator Hamiltonian 23
3.10 Zero of Energy 24
3.11 Some Commutators 25
3.12 Eigenfunction for a Simple Hamiltonian 26
4 One-Dimensional Problems 29
4.1 Potential Step 29
4.2 Deep Square Well 30
4.3 Hydrogenic Wavefunction 32
4.4 Bound State Wavefunction, Current, Momentum 33
4.5 Time Evolution for Particle in a Box 34
4.6 Particle in Box: Energy and Eigenfunctions 35
4.7 Particle in a Box 36
4.8 Particles Incident on a Potential 36
4.9 Two Beams Incident on a Potential 38
4.10 Ramsauer-Townsend Effect 39
4.11 Wronskian and Non-degeneracy in 1 Dimension 40
4.12 Symmetry of Reflection 41
4.13 Parity and Electric Dipole Moment 43
4.14 Bound State Degeneracy and Current 44
4.15 Car Reflected from a Cliff 45
5 More O n e - D i m e n s i o n a l P r o b l e m s 48
5.1 Motion of a Wavepacket 48
5.2 Lowest Energy States 50
5.3 Particle at Rest 51
5.4 Scattering from Two Delta Functions 51
5.5 Reflection and Transmission Amplitudes: Phase Shifts 53
5.6 Oscillator Against a Solid Wall 55
5.7 Periodic Potential 56
5.8 Reflection and Transmission Through a Barrier 57
5.9 Hermite Polynomials: Integral Representation 58
5.10 Matrix Element Between Degenerate States 60
5.11 Hellmann-Feynman Theorem 61
6 M a t h e m a t i c a l Foundations 63
6.1 Cauchy Sequence in a Finite Vector Space 63
6.2 Nonuniqueness of Schrodinger Representation 64
6.3 Degeneracy and Commutator 64
6.4 von Neumann's Example 65
6.5 Projection Operator 67
6.6 Spectral Resolution 68
6.7 Resolvent Operator 68
6.8 Deficiency Indices 69
6.9 Adjoint Operator 71
6.10 Projection Operator 72
6.11 Commutator of Lz and <p 74
6.12 Uncertainty Relation: Lz and cos tp, siny? 74
6.13 Domain of Kinetic Energy: Polar Coordinates 75
6.14 Self-Adjoint Extensions of p 4 76
7 Physical I n t e r p r e t a t i o n 79
7.1 Tetrahedral Die 79
7.2 Probabilities, Expectation Values, Evolution 79
7.3 (Lx) and (Ly) in an Eigenstate of Lz 80
7.4 Free Particle Propagator 81
7.5 Minimum Uncertainty Wavefunction 82
7.6 Spreading of a Wave Packet 84
7.7 Time-dependent Expectation Values 85
7.8 Ehrenfest Theorem 87
7.9 Compatibility Theorem 89
7.10 Constant of the Motion 90
7.11 Spreading of a Gaussian Wavepacket 90
7.12 Incorrect Time Operator 92
7.13 Probability to Find a Particle 92
7.14 Sphere Bouncing on Sphere 93
7.15 Cloud Chamber Tracks 95
7.16 Spin 1 Measurement in Two Directions 95
7.17 Particle in a Box: Probabilities and Evolution 97
7.18 Free Wave Equation: Translation Invariance 99
7.19 Free Wave Equation: Accelerated Frame 101
7.20 The Wigner Function 103
7.21 Properties of the Wigner Function 105
7.22 Uncertainty Relation and Wigner Function 106
8 D i s t r i b u t i o n s a n d Fourier T r a n s f o r m s 108
8.1 Properties of the Delta Function 108
8.2 Representation of Delta Function 110
8.3 Normalization of Scattering Solution 112
8.4 Tempered Distribution 114
8.5 Fourier Transform of V 114
8.6 Tempered Distribution of Fast Decrease 115
8.7 A Useful Identity 116
8.8 A Representation of ¿(a:) 118
8.9 Fourier Transform of ¿ ' " ' ( x ) 118
8.10 Value of x m S < - n \x) 119
8.11 Distribution Occurring in Fermi's Golden Rule 119
9 Algebraic M e t h o d s 122
9.1 An Operator Identity 122
9.2 Expectation Values: Simple Harmonic Oscillator 123
9.3 Angular Momentum Matrices 124
9.4 Displaced Oscillator 127
9.5 Dipole Matrix Elements 128
9.6 Scalar Operator 129
9.7 Probability to Obtain I, m 129
9.8 Probability to Obtain I, m Along Different Axis 130
9.9 Commutators of x and p with L 132
9.10 Some Eigenfunctions of Angular Momentum 132
9.11 Expectation Value of Lx 134
9.12 Rotation Invariance of the Hamiltonian 136
9.13 Uncertainty Relation for SHO 137
9.14 Baker-Campbell-Hausdorff Formula 138
9.15 A Useful Commutator 139
9.16 Uncertainty in Lz 140
9.17 Expectation Values of Angular Momentum 140
9.18 Validity of Ehrenfest's Theorem 142
9.19 Wigner Problem: Annihilation and Creation 143
9.20 Wigner Problem: SHO 145
9.21 Identity for Pauli Matrices 147
9.22 Operator Identity - Spin Rotation 148
9.23 An Operator Identity 149
9.24 Commutator with Inverse Operator 150
9.25 Schwinger Method for Angular Momenta 150
9.26 Minimum Uncertainty in Jx 152
9.27 Unsold's Theorem and its Application 153
9.28 Rotation Matrix for j= 1 154
9.29 Algebra and Constants of the Motion 155
9.30 Coherent State and Normal Ordering 158
9.31 Normal Ordering of xn 158
10 Central Force P r o b l e m s 162

10.1 Isotropic SHO in Two Dimensions 162
10.2 Attractive Exponential Potential 164
10.3 Reduction of the Two-body Problem 165
10.4 Particle in a Spherical Potential Well 166
10.5 Particle on Surface of a Cylinder 168
10.6 Expectation Values: Electron in a H-atom 169
10.7 Parity in Spherical Coordinates 170
10.8 Magnetic Moment due to Orbital Motion 170
10.9 Spherical Square Well 171
10.10 Binding Energy and Potential 173
10.11 Generating Function: Laguerre Polynomials 174
10.12 Normalization of Hydrogen Wavefunction 175
10.13 Kramers' Relation 177
10.14 Quantum Mechanical Virial Theorem 180
10.15 Ehrenfest Theorem for Angular Momentum 181
10.16 Angular Momentum of a Two-Particle System 182
10.17 Hulthén Potential: Ground State 183
10.18 Hydrogenic Atom in Two Dimensions 184
10.19 Runge-Lenz Vector: Constant of the Motion 186
10.20 Runge-Lenz Vector: Hydrogen Spectrum 188
11 Transformation Theory
11.1 Fourier Transform of Hermite Functions 196
11.2 Schrodinger Equation in Momentum Space 197
11.3 Heisenberg Equation for a Free Particle 198
11.4 Dirac Picture for Displaced SHO 199
11.5 Heisenberg Picture for Displaced SHO 202
11.6 Heisenberg Picture: SHO and Constant Force 203
11.7 Heisenberg Picture: Constant Force 204
11.8 Schrodinger Picture: Constant Force 205
11.9 Dirac Picture: Constant Magnetic Field 206
11.10 Coherent State 208
11.11 Coherent State: Overlap of Two States 208
11.12 Coherent State: Wavefunction 209
11.13 Squeezing Operator 210
11.14 No Eigenstates for Creation Operator 211
11.15 Spin Coherent State: Euler Angles 212
11.16 Minimum Uncertainty Spin Coherent States 214
11.17 Spin Coherent States: Complex Variables 216
11.18 Useful Commutator 219
11.19 Forced SHO 219
11.20 3-D Simple Harmonic Oscillator 221
11.21 Quadrupole Tensor 221
11.22 Eigenfunction of J 2 , L2, and Jz 223
11.23 SHO: A Time-independent Operator 225
11.24 SHO with Time-Dependent Spring 226
12 N o n - d e g e n e r a t e P e r t u r b a t i o n T h e o r y 232
12.1 Expansion of l / | r \ — r^\ 232
12.2 Second Order Correction to State 234
12.3 1/2 Ax2 Perturbation of SHO 235
12.4 1/4 Ax4 Perturbation of SHO 237
12.5 1/4 \x4 - Brillouin-Wigner Perturbation 238
12.6 Two-level System 239
12.7 Approximate SHO 242
12.8 Two-dimensional SHO 243
12.9 Kuhn-Thomas-Reiche Sum Rule 246
12.10 Electron in Box Perturbed by Electric Field 248
12.11 Positronium 249
12.12 Rigid Rotator in Electric Field 250
12.13 Electric Dipole Moment Sum Rule 251
12.14 Another Sum Rule 253
12.15 Gaussian Perturbation of SHO Bosons 254
12.16 Gaussian Perturbation of SHO Fermions 255
12.17 Polarizability: Particle in a Box 256
12.18 Atomic Isotope Effect 258
12.19 Relativistic Correction to H atom 260
12.20 van der Waals' Interaction 260
13 D e g e n e r a t e P e r t u r b a t i o n T h e o r y 263
13.1 Stark Effect for n = 2 Level in H 263
13.2 Perturbation of Particle in a Box 264
13.3 Perturbation of Isotropic Two-dimensional SHO 266
13.4 Two-dimensional SHO with Off-diagonal Term 267
13.5 Non-diagonal Two-dimensional SHO 270
13.6 Particle in a Box Perturbed by Electric Field 272
13.7 Unusual Particle on Interval 273
13.8 Rigid Rotator in Magnetic Field 274
13.9 Axp y Perturbation of SHO 276
13.10 Paschen-Back Effect 277
13.11 H Atom: Weak Field Stark Effect 281
14 Further A p p r o x i m a t i o n M e t h o d s 284
14.1 Variational Ground State of SHO 284
14.2 Variational Ground State of H2 Molecule 285
14.3 Square Barrier: W K B Approximation 287
14.4 Variational Ground State in Gaussian Potential 289
14.5 Variational Ground State: Quartic Potential 292
14.6 WKB: Ball Bouncing on a Floor 294
14.7 Ground State of H * 295
14.8 Variational Solution: Particle in a Box 298
14.9 Hydrogen Atom: Variational Technique 299
14.10 Bound State in One Dimension 301
14.11 Field Emission: W K B Approximation 302
14.12 Deuteron: Variational Principle 303
14.13 Bouncing Ball: Variational Calculation 305
14.14 Beta Decay of Tritium 306
14.15 Anharmonic Oscillator 307
14.16 Nonlinear SHO: Variational Calculation 308
14.17 WKB Solution and Parity 311
15 T i m e - D e p e n d e n t P e r t u r b a t i o n T h e o r y 314
15.1 Transition Probability: Bound State to Free 314
15.2 Photo-disintegration of Deuteron 316
15.3 Excitation of SHO 319
15.4 Excitation of SHO by Stiffer Spring 321
15.5 Periodic Perturbation 322
15.6 Excitation of H-atom 323
15.7 Expanding Box 325
15.8 Sudden Displacement of SHO 326
15.9 Sudden Perturbation of Two-level Atom 327
15.10 Berry's Phase 329
15.11 Neutron in Rotating Magnetic Field 333
- -1 • JO Jí ' 11NI I
15.12 Excitation of Electron by Electric Field 335

15.13 Neutron Magnetic Moment 336
15.14 Electron Passing Through Magnetic Field 337
15.15 SHO: Sudden Transition 337
15.16 Coulomb Excitation 338
16 P a r t i c l e i n a U n i f o r m M a g n e t i c F i e l d 342
16.1 Estimate of Magnetic Energies 342
16.2 Radii of Landau Levels 343
16.3 Equation of Continuity 344
16.4 Gauge Invariance 345
16.5 Gauge Transformations and Observables 346
16.6 Spin 1/2 Particle in Magnetic Field 348
16.7 Spin 1/2 in Magnetic Field: Heisenberg Equations 349
16.8 Separation of Spin and Space for Spin 1/2 352
16.9 Spin 1/2 in Time-dependent Magnetic Field 352
16.10 Spin 1/2 in Rotating Magnetic Field 354
17 A n g u l a r M o m e n t u m , E t c . 358
17.1 Operator to Lower Total J 358
17.2 Energy Shift Due to a Magnetic Field 359
17.3 Coupling of Spin 1 to Spin 1/2 361
17.4 Example of Wigner-Eckart 362
17.5 Rotations for Spin 1/2 and Spin 1 364
17.6 Spin 1/2 Coupled to Spin 3 / 2 366
17.7 Coupling of Spin 1 or 0 and Spin 1/2 367
17.8 Identity for Constant Magnetic Field 368
17.9 The State \n,j, m,l) 369
17.10 Landé ^-factor 369
17.11 Spin and Space Coordinates 372
17.12 Clebsch-Gordon for 3/2 373
17.13 Rigid Rotator in a Step Potential 373
17.14 Spin Dependent Operators for Two Particles 375
18 S c a t t e r i n g - T i m e D e p e n d e n t 377
18.1 Cross-section from Experiment 377
18.2 Green's Functions for Free Particle States 378
18.3 Dispersion Relations 379
18.4 Kalien-Yang-Feldman Equations 380
18.5 Born Approximation 381
18.6 Scattering in CM and Laboratory Frame 382
18.7 Propagator for a Free Particle 385
18.8 Propagator for Simple Harmonic Oscillator 386
19 Scattering - T i m e I n d e p e n d e n t 389
19.1 Equations for Spherical Bessel Functions 389
19.2 Rodrigues Formula: Spherical Bessel Functions 392
19.3 Wronskian for Spherical Bessel Functions 393
19.4 Superposition of Yukawa Potentials 393
19.5 B o r n Approximation for Gaussian Potential 394
19.6 Born Approximation for Square Well 395
19.7 Phase Shifts for Delta-function Potential 397
19.8 Phase Shifts for Yukawa Potential 399
19.9 Low Energy s-Wave Amplitude 400
19.10 Spherical Potential Shell 402
19.11 Expressions for j0{x) and n0(x) 403
19.12 Effective Range, Scattering Length 404
19.13 Effective Range, Scattering Length: Yukawa Potential 405
19.14 Shape-independent Parameters 406
19.15 Phase Shifts for Hard Sphere 407
19.16 Resonance for Square Well 411
19.17 Double Slit 412
19.18 Born Approximation: Spherically Symmetric Potential 413
19.19 Scattering from a Separable Potential 414
19.20 Generalized Optical Potential 415
19.21 Free Particle Eigenfunctions 416
19.22 Scattering from an Inverse Square Potential 418
19.23 Neutron-Proton Scattering: Spin Flip 419
19.24 Reflectionless Potential in One Dimension 420
19.25 n-p Scattering: Singlet and Triplet States 422
19.26 Phase Shift, Scattering Length, Etc 423
19.27 Scattering off a Diatomic Molecule 425
19.28 WKB s-Wave Phase Shift: Attractive Potential 426
19.29 WKB s-Wave Phase Shift: Hulthén Potential 427
19.30 WKB Approximation for Phase Shifts 428
19.31 Zero-Range Potential 431
19.32 Calogero Equation 435
20 S y s t e m s of Identical Particles 439

20.1 Periodic Table 439
20.2 Identical s = 1/2 Particles in I = 0,1 States 439
20.3 Two Identical s = 0 Particles 440
20.4 Identical s = 1/2 Particles in Centre of Mass 441
20.5 Heisenberg Field Operator for Bosons 442
20.6 Heisenberg Field Operator for Fermions 444
20.7 Two-body Interaction 444
20.8 Diagonalization of Boson Hamiltonian 445
20.9 Formula for exp(—^4)5 exp(>l) 447
20.10 Bogoliubov Transformation: Fermions 448
20.11 Diagonalization of Bose Hamiltonian 450
20.12 Diagonalization of Quadratic Hamiltonian 452
20.13 Bogoliubov Transformation: Bose Operators 453
20.14 Density Matrix for a Subsystem 456
20.15 Density Matrix and S-Matrix 457
20.16 Zero Energy Bound States of Two Fermions 458
20.17 Bose Number Operator: Constant of Motion 460
20.18 Bose Operator: More Constants of Motion 461
20.19 The Pauli Problem 462
20.20 Atomic Isotope Effect 464
21 Q u a n t u m S t a t i s t i c a l M e c h a n i c s 467
21.1 Average Energy of Assembly of SHO's 467
21.2 Properties of the Density Matrix 469
21.3 Expectation Values for Spin 470
21.4 Expectation Value for Number of Particles 471
21.5 Spin 1/2 Polarization 472
21.6 Density Matrix for Spin s = l 473
21.7 Polarization Vector for Spin j 475
21.8 Composite Density Matrix 476
21.9 Arbitrariness of Composite Density Matrix 477
21.10 Two Energy Levels Bose Gas 479
21.11 Density Matrix: Particles Coupled by Spring 481
21.12 Particles: Dissimilar, Bose, and Fermi 483
21.13 A Three-Level Laser 484
21.14 Integrals from Q u a n t u m Statistical Mechanics 486
Index 489
Chapter 1
T h e B r e a k d o w n of Classical
Mechanics
1.1 Quantum N u m b e r of the Earth

Calculate the principle quantum number for the earth in its orbit about the sun.
What is the energy difference between two neighbouring energy levels?
Hint: For large n, E„ ss Eciassicai .
Solution
The classical energy of the earth in its orbit about the sun is
where
m is the mass of the earth = 6 x 10 27 g ,
M is the mass of the sun = 2 x 10 33 g ,
G is the gravitational constant = 6.67 x 1 0 - 8 dyn cm 2 /g 2 and
R is the sun-earth distance = 1.5 x 10 13 cm .
Proceeding as for the hydrogen atom we find that we need only replace e 2 by
GmM. Thus, we find
GmM (1.1.2)
Solving for n we get
GMm2 = 18 x 10 (1.1.3)
Therefore
n « 4 x 10 73 . (1.1.4)
1
Also we see t h a t since
1 /2tr \2 An An ,
AE — - I —-GmM ) m —3 = E\ (1.1.5)
2 \ h J n n
we get AE = 6.4 x 1 0 - 3 4 erg .
1.2 Thermal Wavelength

W h a t is the wavelength associated with gas molecules at a temperature T?
Estimate this wavelength for a typical gas at room temperature and compare it
to visible light.
Solution
The energy U of a molecule at temperature T is given by the equipartition
principle
U = \ k B T = y . (1.2.6)
Therefore, the wavelength is

A = hc/U . (1.2.7)
At room temperature T « 300 K. So,
U = 1.5 x 1.38 x 10" 2 3 x 300 J = 6.21 x 1 0 - 2 1 J . (1.2.8)
6 63 X 10 3
A = - 6 2~ ^ o 3 _ Q 0 x 1 Q 8 = 3.20 x 1 0 - m . (1.2.9)
This is considerably longer than visible light which has a wavelength of about
5 x 10~ 7 m .
1.3 P h o t o n s in a B e a m
For a monochromatic beam of electromagnetic radiation of wavelength (A «
5000 Á) , intensity 7 = 1 w a t t / m " , calculate the number of photons passing an
area of A = 1 cm 2 normal to the beam in one second.
Solution
The energy of a photon is given by
-^photon —- —y • (1.3.10)
1 he total energy of the beam is E — IA = N E p h o t o n where N is the number of

photons. Therefore,
.. I AX 1 x 10" 4 x 5000 x 10 - 1 0
N = —— = = 9 k v i n14
he 6.63 x 10 - 34 x 3.00 x 10» ^ X 10" phot°nS " ( L 3 1 1 )
i .4. H r vnulTÜl\ Aiurn ó
1.4 Hydrogen A t o m and de Broglie

Show that if one assumes that the circumference of a stationary state orbit of an
electron in a hydrogen atom is an integral multiple of the de Broglie wavelength,
one also obtains the correct energy levels.
Solution
The de Broglie wavelength is given by
A = h/p .
The circumference is 2irr. Thus, we write
2nr= n\ .
The energy is given by
it P2 e2
E =
to " 7 - <L412>
We substitute for p and r in terms of A and introduce the fine structure constant
2ne2
« = -¡¡r (1.4.13)
to get
h2 hca
E = (1.4.14)
2mA2 n\
We now equate the Coulomb force with the mass times the centripetal acceler-
ation
mv2 2xe2
— = ^x- ( L415 )
This yields
h2 27re2
S P = 7 T ' C-4-16'
We can now solve for A.
1 mac 1
A=—~n- (1.4.17)
Combining this with our previous result we find
ma2c2 ma2c2 1 ma2c2 , ,

E =
^ 5 S5- = - j — • C-418)
1.5 Vibrations in N a C l
The shortest possible wavelength of sound in sodium chloride is twice the lattice
spacing, about 5.8 x 1 0 - 8 cm. The sound velocity is approximately 1.5 x 105
cm/sec.
a) Compute a rough value for the highest sound frequency in the solid.
b) Compute the energy of the corresponding phonons, or quanta of vibrational
energy.
c) Roughly what temperature is required to excite these oscillations appreciably?
Solution
The shortest wavelength is given by
(1.5.19)
A m , n = 5.8 x 1 0 - 8 cm = 2 x lattice spacing
»o = l , 5 x 105 cm/s .
(1.5.20)
a)
«o
A '
(1.5.21)
Therefore,
vmax = = 2.6 x 10 12 Hz .
Amin
(1.5.22)
b)
Emax = hvmax = 1.7 x 1 0 - 2 1 J = 0.011 eV

c (1.5.23)
)
u = \kBT .
(1.5.24)
Assume u « Emax. Therefore,
T ~ Êmax on
(1.5.25)
3k B - 8 3 K •
1.6 Crystal Powder

Estimate the effect on the specific heat of reducing a crystal to a fine powder of
dimensions of about 1 0 - 5 cm.
Hint: Study the Debye model [1.1] of specific heat and realize that the size of
the crystal now also imposes an upper limit on the wavelength of the sound
waves in the crystal.
Solution
The effect of grinding up the crystal into a powder is to limit the maximum
wavelength of a standing wave in the crystal to roughly the size of the crystal
particles. This changes the integral in the Debye expression [1.1] for the internal
energy from
1 2 t t ( k B T ) 4 r — x3dx
u
= - i ^ ~ J 0 ~ i (i.e.26)
to
l2n(kBT)4 iXmax x3 dx
^powder — h3y
0 Jxmtn ex - 1
rr 12Tv(k B T) [*""" x3dx
4
" — V S T J. — r f'-"7»
Here, we have introduced
h =
h Do
•Emm — ^ ^min ~j ™ T (1.6.28)
k B A m a x
where Á m a x = 1 0 - 5 cm « size of the powder particles. If we now estimate x m i n
at room temperature by using that vo ps 500 m / s , we get x m ¡„ « 8 x 1 0 - 4 < < 1.
Therefore, we can approximate the exponential in the last integral by ex ss 1 + a:
and get
rxmtn x3dx rrm,„ j
Jo e ' - l ™J0 X dX
~ ~Z X
min- (1.6.29)
This means that the internal energy U is reduced by
A u
12n-(*Br)4 x /I hv2 o \ 3 4tt kBT
= itur- (£rrr --) = ^ - (1-6.30)
h3V¡¡ X
1^3 kBXmax ) ~ K
Therefore, the specific heat per unit volume at constant volume
\_dU_
V dT
is decreased by a constant amount, namely
47rkB _ 4tt x 1.38 x 1 Q - 2 3 J / K

0.173 J / ( K m 3 ) . (1.6.31)
^max (10- 7 )3m3
1.7 Einstein Coefficients

For a collection of atoms with energies En , n — 1, 2, 3, . . . submerged in
a background of radiation at a temperature T, the following transitions may
occur:
1) spontaneous from n —>• m En > Em
2) induced from n - » m
3) induced from m —» n .
At equilibrium, at a temperature T, the emission and absorption probabilities
are given by
Pmn = KNn[Bmnp(v)+Amn] emission (1.7.32)
Pnm = KNm[Bnmp(v)] absorption. (1.7.33)
Here, hu = En — Em, K is a proportionality constant, and Nn, Nm are respec-

tively the number of atoms in the states n and m . The coefficients Amn, Bmn
are known respectively as the "Einstein Coefficients" of spontaneous and in-
duced emission, whereas the coefficient Bnm is known as the "Einstein Coeffi-
cient" of induced absorption. [1.2] Use these equations together with Planck's
radiation law for the radiation density p(v) at equilibrium to show that
1) the Einstein coefficients of induced absorption and emission are the same,
that is that Byim — Bum and t h a t
2) the Einstein coefficients of spontaneous and induced emission are related by
Anm ~ Bnm .
có
Solution
In equilibrium, at a temperature T, if the number of atoms in the state n and
m is given by Nn and Nrn respectively, we have t h a t
Nn=Ne~E"/kBT , Nm = Ne~Bm/kBT (1.7.34)
where N is the total number of atoms. Therefore,
Nm = Nn e(En-Em)/kBT _ N n ehv/kBT (1.7.35)
Also at equilibrium the number of transitions from n —»• m equals the number
of transitions from m —> n. Thus, we have
NmPnm = NnPmn (1.7.36)
or
Nnehv'kBT Brnnp{v)=Nn[Bnmp{v) + Anm} . (1.7.37)
Thus, solving for the radiation density p we get
<•<">=^rJBBm\B
c
• <17.38)
J^nm/
But, at equilibrium the radiation density is given by Planck's Law
Therefore, comparing these two equations we see that we have
Bnm — Bmn (1.7 .40)
and
871" , 3 n
Aim — ' (1.7.41)
Bibliography
[1.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 1.12, (2002) .
[1.2] F.K. Richtmyer, E.H. Kennard, and J.N. Cooper, Introduction to Modern
Physics, 6th edition, sec 13.12, McGraw-Hill, New York, (1969).
Chapter 2
R e v i e w of Classical
Mechanics
2.1 Lagrangian and Hamiltonian for SHO

Find the Lagrangian for a harmonic oscillator. Use the definition of conjugate
momentum p to find it and also the Hamiltonian H .
Solution
For a simple harmonic oscillator we have
1
2
T -mv
(2.1.1)
Therefore,
L = T — V = \-rnv2 - -kx2 . (2.1.2)

2 2
Next we compute the canonical momentum
dL
p = -x- = mv . (2.1.3)
ov
Then, we can substitute p/m for v and get
(2.1.4)
8
2.2 Lagrangian and Hamiltonian: Simple
Pendulum
Repeat problem 2.1 for the simple pendulum. Interpret the momentum p con-
jugate to the angle variable 9 .
Solution
Here, the kinetic energy is given by
T = ^mi;2 = ^r2 + r292^j (2.2.5)

But, r = / = constant. Therefore,
T=^ml292 . (2.2.6)
The potential energy is

V = mgl(l - cos 9) . (2.2.7)
So,
L = T — V = ^ml292 — m<¡r/(l — cos#) . (2.2.8)
The conjugate momentum is given by

dL
pe = —- = rnl29 . (2.2.9)
09
This provides an equation for 9 in terms of pg. Then,
H = d-±9-L
09
= ml2 9 — -ml2 92 + mgl(l — cos 9)
= ^ml292 + mgl(l — cos9) . (2.2.10)
But the Hamiltonian is a function of the generalized coordinate and the conju-
gate momentum. So we get
Pe
H = + mgl( 1 — cos9) . (2.2.11)
¿ml¿
Furthermore, the expression for the conjugate momentum (2.2.9) shows that it
is just the angular momentum.
2.3 Bohr-Sommerfeld Quantization: SHO

Use Bohr-Sommerfeld quantization to calculate the energy levels of a one-
dimensional simple harmonic oscillator.
Solution
We use Bohr-Sommerfeld quantization
pdq = nh — 2nnh . (2.3.12)
Also, the total energy is

2
E = ^ + hx . (2.3.13)
2m 2
Therefore,
p= ±V2Ê]Jl-'^ • (2.3.14)
The motion is bounded by ± x 0 = ±\/2E/k. The two signs in front of p indicate

that we have momentum to the right + and then momentum to the left —.
Therefore,
f p d , = l-k-^-d,
2E
— 2\/2 mE J (2.3.15)
We now call w2 = k/m and, after evaluating the integral, set it equal to nh to
get
2nnh = 2 — \ . (2.3.16)
UJ 2
So, the quantized energy levels are
En = nhu> . (2.3.17)
2.4 Bohr-Sommerfeld: Particle in a B o x

Use Bohr-Sommerfeld quantization to calculate the energy levels of a particle
confined to a box of length L. For simplicity assume this is a "one-dimensional
box".
Solution
As in the problem above we use Bohr-Sommerfeld quantization
pdq = nh = 2nnh . (2.4.18)
Again we have that the momentum is related to the energy by

p = ±V2mE (2.4.19)
so that we have the particle moving to the right with constant momentum V¿mE
and then returning to the left wit) constant momentum -y/2mE. Hence, we
get
(jipdq = nh
= V2 mE 11: dx — J dxj
= \ZlmE x 2L . (2.4.20)
Solving for E we find
h2
£ = !
s ^ " - I2-4'21)
2.5 Larmor Frequency

Suppose a gyroscope has a magnetic moment ¡2 proportional to its angular
momentum j. according to
¡1 = ML .
The potential energy due to placing the gyroscope in a magnetic field B is
V = -¡I B .
Assume that B is constant and derive the equation of motion for L. Show that
the gyroscope precesses with the angular Larmor frequency
u)L — MB .
Solution
For a gyroscope with angular momentum L we have t h a t if the angular velocity
is Q and the moment of inertia is 7 then
L = Id . (2.5.22)
Thus, the kinetic energy is
2
T=ÎJ = ^lé2. (2.5.23)
Also,
V =-p. • B = -ML • B = -M1Q •B . (2.5.24)
Then, the Lagrangian is
L = \ie2 - I' = l-lé2 + MI9B. (2.5.25)

^ z
The equation of motion is
— lié + MIB1 = 0. (2.5.26)

dt L -I
This yields immediately a first integral
9 + MB = constant . (2.5.27)
But, for no magnetic field (B = 0) we have the gyroscope at rest and we

immediately get 0 = 0 for 5 = 0. So we see that the integration constant
vanishes. Hence,
9 = -MB = -lol . (2.5.28)
2.6 Applicability of Bohr-Sommerfeld

Quantization
The system of quantization proposed by Bohr in 1913 and late gneralized by
Sommerfeld is not applicable to all systems. To what general kinds of physical
systems is Bohr's procedure applicable? For what kinds of systems is it not
applicable.
Solution
The system of quantization proposed by Bohr and Sommerfeld is applicable to
systems with repetitive (periodic) motion. It is not applicable to systems with
unbounded motion.
2.7 Schrodinger and Hamilton-Jacobi

Consider the time-dependent Schrodinger equation and put
xjr = A e ' s / h
where A = constant. Show t h a t in the limit as h —» 0 the equation
h2 a 2 * 3*
2m dx2 + VV = ik- dt
reduces to the Hamilton-Jacobi equation [2.1]
1
fdS\2 , 8S
+ V { q )
2m f e j -~dT-
Solution
If we put
V = AeiS'h
in the Schrodinger equation we get
Now let H - » 0 then we get
" + -lV (2.7.30)

dt 2m
This is just the Hamilton-Jacobi equation.
2.8 W K B Approximation
In the problem above set S = W — Et and let W = Wo + hW\ + h¿W^ + • • • for
the case of a one-dimensional Schrodinger equation. Find the equations for Wo
and Wi and solve the equation for Wq . This is the so-called Wentzel-Kramers-
Brillouin or W K B approximation.
Solution
Clearly if in the problem above we set S = W — Et (before letting h —>• 0) we
get
(2.8.31)
Letting W = Wo + hW\ + . . . and equating the coefficients of powers of h we get
(2.8.32)
So,
(2.8.33)
and
(2.8.34)
So we have an equation for W\

2.9 Dumbbell Molecule: Bohr-Sommerfeld
A dumbbell molecule of moment of inertia I is rotating about its centre of mass.
a) Write the langrangian for this system and find the canonical momentum.
b) Use Bohr-Sommerfeld quantization to find the energy levels of this molecule.
Solution
a) If we take the plane in which the molecule is rotating as the plane z = 0 then
the Lagrangian is just the kinetic energy
L = iI<p 2 . (2.9.35)
The canonical momentum conjugate to ip is

dL T.
Pv = iT = I<P • (2.9.36)
Cv
b) Bohr-Sommerfeld quantization states that
j ) p^dtp = J p^d<p = nh . (2.9.37)
But,
Pv = Itp = V2IE (2.9.38)
where E is the total energy which is conserved and therefore a constant. Thus,
/* 27T
V2IE / dip • 27t\/2IE = nh . (2.9.39)
Jo
So,
2
1 , / h \ n2h2
E = — (2.9.40)
21 \2n 21
Bibliography
Chapter 3
Elementary Systems
3.1 Com mut ator Identities

a) Verify the identity
[AB,C] = A[B,C] + [A,C]B.
b) Using the result above and [x, p] = ih prove that
[x~,p] = 2 ihx
and
[xn ,p] = nihx""1.
Hence prove that for any function g(x) analytic at the origin
[g{x),p] = •
Solution
a) To verify the identity we simply write it out
[AB,C] = ABC-CAB
= ABC - ACB + ACB - CAB
= A[B,C] + [A,C]B. (3.1.1)
b) By applying this identity we immediately get
[a;2,p] = x[x,p] + [x,p]a; = xih + ihx = 2ihx . (3.1.2)
If we replace x2 above by xn we can provide a proof by induction since
[xn,p] = xn-1[x,p]+[xn-1,p}x (3.1.3)

and assuming the result holds true for n — 1 we immediately get the result for
n. Also, we have already shown the result to be true for n = 2.
The second part may be proved in two different ways.
i) Since g(x) is analytic at x — 0 we can write a Maclaurin series
, 1 dng
g(x) = ^cnx , Cn = _ — (3.1.4)
07 = 0
Hence,
[g{x),P\ = l ^ c n [ z n , P ] (3.1.5)
and using the result above
[ff(z)>p] = ~Y^c n nihx n ~ l = i H ^ ~ Y 2 C n X n =

(3.1.6)
ii) A second method is to use the representation of p as the differential operator

h d
P= (3.1.7)
~idi
and act with the commutator on some function f(x). Then,
fx dj h, d dg
[g{x),p]f{x) = g(x 7 — (g(x)f(x)) = [ i h - ) f ( x ). (3.1.8)
dx dx
So, we again find
[ffW.p] = • (3.1.9)
3.2 Complex Potential

Assume that the potential V is complex of the form V = U + iW. Show that
W corresponds to a sink or source of probability.
Hint: Show that
l + v.j = \w„.
This proves that unless W = 0 probability is not conserved.
Solution
We begin with the Schrodinger equation and its complex conjugate.
d$! h2
dV*
—ih + (U - iW)y . (3.2.10)
dt 2m
Next, we multiply the first equation by and the second by $ and subtract
the two equations to get
(*¥*) = (**V 2 tf - l - V 2 * * ) + 2 i W W . (3.2.11)
After rearranging and using the definitions

p =
=
> (3-212)
we get the desired result

dp — 2
- + V.J = -Wp. (3.2.13)
3.3 Group and P h a s e Velocity

a) In deep water the phase velocity of water waves of wavelength A is
v= I9±
2tt
What is the group velocity?

b) The phase velocity of a typical electromagnetic wave in a wave guide has the
form
v =
\J 1 - (w0/w)J
where uio is a certain characteristic frequency. What is the group velocity of

such waves?
Solution
a) The phase velocity is quite generally given by
^ /9 ^ /—rr
= =
=I vs ^ • <3314»
So we see that w = \fgk. Therefore,
du) 1 i—— 1 u> 1 I gX

°' = ü ' 2 ^ = ? ' = ü = 2 ] l k ' (3 315)
'
b) This time we have
( 3 3 1 6 ,
- - '
So, c2k2 = ui2 — uIq. Differentiating this equation with respect to k we get
= . (3.3.17)
This implies that
£ = cjt = £ (3.3.18)
* dk uj vp
3.4 Linear Operators

Which of the following operators are linear?
a) Kifi(x) = f K{x,y)ip(y)dy .
b) K3 where K is defined above.
c) AB if A and B are linear.
d) B_1 if B is linear and B~l is defined by B~1B = BB"1 = 1 .
if A is linear
e) exp A = -
f) Ail> = exp(AV>) •
Solution
In all cases, to see if the operator is linear, we have to check whether
=
A(X\ipi + \2ip2) ^1 Aifii + \2A1fi2 (3.4.19)
where Aj and A2 are constants. The results are:
a) Linear since
J K(x,y) [XiAifii(y) + \2Aifi2(y)] dy
= Ai J K(x,y)ifii(y)dy + A2 J I<{x,y)ifi2(y) dy . (3.4.20)
b) Linear since
K3ifi{x) = J K(x,y)dy J K{y,z)dz J K(z,w)ifi{w) dw . (3.4.21)
c) Linear since
AB(\iifii + A2i/,2) = A(XiBifii + \2Bifi-¿)

= XiABifii + \2ABifi2 • (3.4.22)
d) Linear because the identity operator is linear and we have BB~l = 1. Thus,
BB ^"(Aiî -(- A2^2) — AiBB îfii X2BB 'ifi2

1 1
= B(XiB- 1 fii + X 2 B- ifi 2 ) (3.4.23)
where we have used the fact that B is linear. We now operate on this equation
with A " 1 from the left to get the desired result
B~1(Xiipi + *2^2) = V1 + A2B V2 • (3.4.24)

e) We first use the definition of exp A
A"
(exp A)i> = ^2 • (3.4.25)
n=0
But by an argument the same as for part b) above and induction we see that
An is linear. Furthermore, a sum of linear operators is again linear. Therefore,
exp A is linear. In fact, by the same sort of argument, if A is linear then any
function of A say f(A) is also linear.
f) Not linear because
exp[A(AiV>i + \2ip2)] j 1 Ai exp[Aî] + A2 exp[A^ 2 ] (3.4.26)
3.5 Probability Density

a) Compute, in closed form, the probability density p(t, x) for the wave function.
^(í,a;) = J dk A(k) exp—i - kx
where
_ „-i 2 fc 2 /2
A(k) =
b) W h a t is the "width" of the probability density at time t = 0 and at time t ?
Solution
We first compute the wavefunction $!(t,x) explicitly.
OO
/
dk exp [~k2(L2/2 + iht/2m) + ikx] . (3.5.27)
•OO
Now, we complete the square in the exponent
—k2(L2/2 + iht/2m) + ikx 2ikx

= ~2^ +JWm)
L2-\-ihl/m (L2 + iht/m)2
k2
(3.5.28)
2(L2 + iht/m)
Also, we define
IX
q= k — (3.5.29)
L2 + iht / m
Then,
r°°~ (í.5+.*«<
V(t,x) = exp dq
2 (L2 + iht/m) J — OO — , , n
(I^ + I nt/r,
x exp [ — q 2 ( L 2 / 2 + iht/2m)] (3.5.30)
The contour of integration may now be shifted to the real axis since the integrand
has no singularities (poles or cuts) in the region between the real axis and a line
parallel to the real axis. Also, for real q oo, the integrand vanishes. Thus,
tf(i,x) exp
2 (L2 + iht/m) / OO
dq exp [—q2(L2/2 + iht/2m)]
2tr -OO
• exp (3.5.31)
(L2 + iht/m) 2 (L2 + iht/m)
The probability density is p = Therefore,
2n LJ2X^ 2
P= exp (3.5.32)
\JLA + (ht/m)'- (L + (ht/m)2
4
To see the spreading of the wave packet we find the values of x such that p has
dropped to roughly 1/e of its maximum value. These points are obtained by
setting the argument of the exponential equal to —1. Then we find
L22x™ 2
1 . (3.5.33)
(L4 + (ht/m)2)
So the width is given at any time t by
h2t2
w = L\ 1+ (3.5.34)
m2L4
This shows how the wave packet gets wider with time.
3.6 Angular M o m e n t u m Operators

Show that if we write
L = r xp = -fx V
i
then
LxLy LyLx lhLZ
LyLz LZIjy — lhLx
LzLx LXLZ — ihLy

Solution
The three components of L are explicitly given by
Lx = VPz ~ zpy
Ly ¿Px %Pz
Lz = xpy-ypx • (3.6.35)
Now, use the results of problem 3.1. Then,
[Lx, Ly] = [yPz - zpy,zpx - xpz]
= [yPz,ZPx] + [zpy,xpz]
— —ihypx + ihxpy
= ihL z . (3.6.36)
The other pairs of commutators are obtained in exactly the same manner and
yield
[Ly, Lz] = [zpx - xpz,xpy - ypx]
= [zPx,xpy] + [xpz,ypx]
— —ihzpy + ihypz
= ihLx (3.6.37)
as well as
[Lz, Lx] = [xpy - ypx, ypz - zpy]
= [xpy,ypz] + [ypx,zpy]
= —ihxpz + ihzpx
= ihLy . (3.6.38)
3.7 B e a m of Particles
A beam of free particles is moving along the x-axis with velocity v such that
there is one particle in a volume V.
a) W h a t is the corresponding normalized, time-dependent wave function for
such a particle?
b) W h a t is the number of particles crossing a unit area, normal to the x-axis,
per unit time?
Solution
a) For a free particle moving with velocity v along the x-axis we have
f(i,x) = Ae^kx~^ (3.7.39)
where
mv mv2
(3.7.40)
22 CHAPTER 3. ELEMENTARY SYSTEMS
The particle is confined to the volume V. This means t h a t the normalization

integral is over the volume V. So,
= f |tf|2(íl/= [ \A\2dV = \A[2V . (3.7.41)

Jv Jv
Therefore,
A = (3.7.42)
W
b) The current density is given by
, x) dy*(t, x)
~ V(t,x)
2 im dx dx
h 1 —Hkx—ujt) ^ i(kx-ujt) i(kx-cot) ^ -i(kx-uJt)
c a ^ a
2im V OX OX
hk mv v
= (3.7.43)
mV ~ mV V
The number of particles passing through an area S normal to the x-axis is
therefore
jS = v
V
3.8 Time Evolution of Wave Function

A free particle has the wave packet at time t = 0 given by
„ikx
¥(0,*)= [ dk
J —( k + a2
2
Determine an expression for the wavefunction for a later time t . Do not a t t e m p t

to evaluate the resulting integral.
Hint: Use the equation
-«(*)«] M
/ OO i[kx w
f{k)e -
-OOthe energy for a free particle is given by the equation
and remember that
(hk)2
hw =
2m
Solution
For a free particle the wavefunction at any time t is of the form
OO
(3.8.44)
/ dk f(k)e^kx-^k^
3.9. OPERATOR HAMILTONIAN 23
where
"<*> = TT = iS ' <"-45>

So we have to find an appropriate f(k) and carry out the integral. To do this
we use the initial condition that
oo
/ J k
k T ^ 2 - (3.8.46)
This means that
/(*) = • (3-8.47)
Therefore,
oo gi[kx — hk2t/2m]
/ J k
k2 + a2 • (3.8.48)
Fortunately we have been told not to attempt the integral. Therefore, we are
done.
3.9 Operator Hamiltonian

You are given the classical Hamiltonian for the motion of a particle in the form
H = £-e~ax
2m
where a is a constant. Find an acceptable Hamiltonian operator for this system.
Notice that the answer is by no means unique.
Solution
The classical Hamiltonian that we are given is
^classical = e~ax . (3.9.49)

2m
To make an operator out of this we must do two things:
a) We must replace p by the operator
Pop=-~ (3-9-50)
i ax
and
b) rearrange the order of the operators P o p and xop so that the resultant Hamil-
tonian is hermitian. The easiest, and by no means unique, way of doing this is
to symmetrize the operators. Therefore we propose
3 9 5 1
• i - - »
24 CHAPTER ó. ELEMENTARY SYSTEMS
Other possibilities are
Hop= ¿Pope-axPop (3.9.52)
or any convex linear combination of (3.9.51) and (3.9.52).
3.10 Zero of Energy

In classical mechanics, the reference level for the potential energy is arbitrary.
What are the effects on the wave function and energy of adding a constant
potential in the time-dependent Schrodinger equation?
Solution
Suppose we have the Hamiltonian H and we add a constant Vo to it to get the
Hamiltonian H'. Then we have that
ih— = HV (3.10.53)
as well as
d^'
ih—= H'V = {H+ V0)V' . (3.10.54)
If we try adding a constant phase —a/h to $ we have
V = e-ia/n^ . (3.10.55)
Then,
-ia/h i t . t w (3.10.56)
th——=:e ' + ih——
dt [ dt.
So, if we choose
<* = Vo
we get that
ilr' = e- , v °/ A *J . (3.10.57)
satisfies equation (3.10.54) .

Now, suppose and are stationary states so that we have
V(t,x) = e-iEt'nxP(x) (3.10.58)
and
9'(t,x) = . (3.10.59)
Then, we see that if we set
E' = E+V0 (3.10.60)
3.11. SOME COMMUTATORS 25
we find that
•hdV _ [ e-iv0t,nih<W
ih— = Ko + e i>i m
= e~iV°t/n[H + V0]V
_ e -iVot/h{ E + V [ ^ _ (3.10.61)
So, we have found that
tf'(i,x) = e~iV't/h9(t,x) (3.10.62)
and
E' = E+V0 . (3.10.63)
So, we only get the same constant added to the energy and a constant phase
shift to the wavefunction.
3.11 Some C o m m u t a t o r s
Given the Hamiltonian operator
" = £ + vW
where the operators x and p satisfy the commutation relation
[x,p] = ih
find the following commutators

a) [x,H]
b) [p, H] .
Hint: Use the representation
h d
P
i dx
and let the expressions [x, H] and [x, H] operate on an arbitrary function /(ab-
solution
a) Using the hint we have
[»,«]/<.) = ¿
1 2
2 h ^ = - h ^ — . (3.11.64)
2m dx m i dx
So,
[x,H] = ih— . (3.11.65)

m
26 ( J H A f 1 bin if. ZLrjivirjiv i j \ h y s r j i o j w a
b) Proceeding in the same way
\p,H]f{x) = -—1V(x)f(x)}-V(x)*$L
(3.11.66)
Thus,
dV
(3.11.67)
lP H] =
' ~ i H
^ •
3.12 Eigenfunction for a Simple Hamiltonian

A system (in a particular set of units) is described by the Hamiltonian operator
a) Show t h a t both
ipi = Ai e-*2/ 2
and
x
rp2 = A2xe ¡2
are eigenfunctions of this Hamiltonian and find the corresponding eigenvalues.

b) Find the constants A\ and A2 that correctly normalize these eigenfunctions.
c) Find the expectation value of x for these eigenfunctions.
Solution
a) To verify that the given functions are indeed eigenfunctions we simply act on
them with the given Hamiltonian operator. Thus,
H Aie-x2/2
[1 - x2 + x2\A\ e l2/2
2 2
1 .A i e ~* ' . (3.12.68)
Thus, Ai e r 2 / 2 is indeed an eigenfunction of the given hamiltonain and the

corresponding eigenvalue is 1. Similarly,
r2/2
H A2xe = +x2 A2xe x2/2
= [2 - x2 + x2\A2x e~x*,2
= 2 . A2X e-*2/ 2
. (3.12.69)
3 1? EIGENFUNCTION run A SIMPLE HAMILTONIAN 27
Thus A2xe~ x *l 2 i s also an e

'Senfuncti°n of
the given hamiltonain and the
c o r r e s p o n d i n g eigenvalue is 2
b) Normalization requires t h a t we calculate the integrals
CO
/
\ip¡\2 dx i =1,2. (3.12.70)
• OO
One way to proceed is to first compute the integral

roo 2
1(a) = / e ax dx . (3.12.71)
J — OO
To do this we use a trick
f OO /«OO 2
/ OO 0 2 /*( e""" dy
e" *
2n /* oo dx /
•OO e~ar J—rdr
- í J
= "a Jo
/ z
OO
Jo JO e~ dz
7T (3.12.72)
a
Therefore,
'(a) = ^ • (3.12.73)
This gives us immediately the normalization for ifii by setting a = 1.

Al
= " (3-12.74)
For A-¿ we need the integral
But, f e~ax x2dx
e
.
-¿r = ~L 2^-
Jd—
/ (Ca ) f00 2 „ 1 . Af
xdx=
~ (3-12-75)
So, again setting a = 1 we get
A
2=(^) • (3.12.76)
c) The expectation values of x in the two states are given by

f°° 2
W l , xipi) = / e _ r xcte = 0 (3.12.77)
J — OO
since the integrand is an odd function. Similarly,
OO
/
= 0 e~x2x3dx (3.12.78)
-OO
since in this case also the integrand is an odd function.
Bibliography
f
Chapter 4
One-Dimensional Problems
4.1 Potential Step

A beam of particles of mass m moving from left to right encounters a sharp
potential drop of amount VO. Let E be the kinetic energy of the incoming
particles and show that the fraction of particles reflected at the edge of the
potential (located at x = 0) is given by [4.1]
/ VE + VO-VE\2
V VE + VO + \ÍE J
In view of this result, what will happen to a car moving at 10 km/hr if it meets
the edge of a 200 m cliff? Is this answer reasonable? If not, why not?
Solution
The Hamiltonian for this problem may be taken to be
where
x < 0
x > 0
The relevant differential equations now are
(4.1.1)
nn
Since the incoming beam is from the left we have only a transmitted beam
on the right. Thus, the appropriate solutions are
V>i(x) = A(eik*' + Re~ikim) x<0

ik3X
iP2(x) = Be *>0 (4.1.2)
where
2m{E Vo)
k¡ = ^ f , k\ = ~ . (4.1.3)
Using the continuity of the wavefunction and its derivative at x = 0 we get
A(1 + R)=B , ikiA{l - R) = ik2B . (4.1.4)
Therefore, solving for R we get
k1
R = ~ *2 . (4.1.5)
K
ki+k2 '
Cancelling a factor of 2 m / h 2 we get
2
\R\2 = VE + Vo-VE (4.1.6)

VE + V0 + VE
Regarding the car, the question is a red herring. The car going over a cliff
has nothing to do with the problem we have just solved. For the car going
over the cliff the potential is not a step function but varies smoothly from a
potential that is 0 for x < 0 to a potential which is mgy for x > 0. This is a
step function in the configuration of matter, not in the potential energy of the
car. The car problem involves a potential that acts perpendicular to the car's
original direction of motion.
This explains why a probability of |ii| 2 = 0.84% is unreasonable, apart from
the fact that we know that cars don't as a rule bounce away from unobstructed
cliff edges.
The problem of a car going over a cliff is considerably more complicated and
is handled at the end of this chapter (problem 4.15).
4.2 D e e p Square Well

A particle moving in one dimension interacts with a potential of the form
V(x) = 0 |ar| > a
V = < 0
^ ~â "
Find the equation determining the energy eigenvalues of this system. Solve it
approximately assuming a is very small. What happens in the limit a —> 0?
Solution
Here,
x / 0 for \x\ >a
_ for
^ I -f£ l®l < a '
We define U = vo/2a. The problem now is exactly the same as the problem dis-
cussed in [4.2]. The solutions divide into positive and negative parity solutions
and read
f AeKx x < —a
1
B cos kx |x| < a.
1
iP+(x) =:
{ A e~Kx x >a
AeKx
ip-(x) =
=
|r
^ B sin
sinkxkx
x < —a
|x| < a (4.2.8)
A
o
I{ -Ae-Kx x >a
where
,4.2.9)
The equations that yield the energy eigenvalues are:

Positive parity
k tan ka = K (4.2.10)
Negative parity
kcotka = —K . (4.2.11)
Next, we solve for small a and use the explicit form of U

2 m ( E
k* = + 'U°/2°) . (4.2.12)
h"
Thus, as a gets smaller the term vo/2a dominates so we can replace k'¿ by
ImU/h2. This means that ka « 1 so we can set tan ¿a — ka (and cotA:a =
1/ka). The even parity solution now yields k2a = K. This gives that
E
= -J¡T- (4.2.13)
The odd parity solution requires that k/ka = —K. This leads to an energy that
diverges as I/a2. Thus, we get only one bound state in the limit as a —> 0. The
*+&={!?-*• itr
corresponding wavefunction is clearly
=^-"w.
Furthermore in this limit we find that U —>• oo. The result is a J-function
potential.
4.3 Hydrogenic Wavefunction
The wavefunction of an electron in the ground state of an hydrogen-like atom
is
Zr a
ip(r) = Ae ! , a = —-z
where Z is the charge on the nucleus.

a) Determine the constant A, so that the wavefunction is normalized to unity.
Remember that you are in three dimensions and r represents the radial variable
in spherical coordinates.
b) At what distance from the origin is the probability of finding the electron a
maximum?
c) Determine the average value of: the kinetic energy, the potential energy and
the total energy.
This verifies the virial theorem for the Coulomb potential.
Solution
a) To normalize we simply integrate and impose
1 = J\ip\2dv. ( 4 . 3 . 1 5 )
Since the wavefunction is spherically symmetric we have dv = 4irr2dr and get

oo
I = 4n Í \ i p ( r ) \ 2u rJJdr
,
Jo rOO
= 47r|^4|2 / e~2Zr/a r2dr
a'
Jo 3
= 4tt|¿|-*4^3 ' (4-3-16)
So, after choosing the phase of A
"=v5-
b) Let the probability of finding the electron between r and r + dr be P(r) dr.
Let the probability of finding it in the volume between-!) and v + dv be pdv.
Due to the spherical symmetry we again have dv = 4irr2 dr and so the two
probabilities describe the same thing.
P(r) dr = 4np(r) r2 dr . ( 4 . 3 . 1 8 )
Hence,
P = |V>|24ttR 2
= —-J- e~2Zr'a 4ttr2 . ( 4 . 3 . 1 9 )
4.4. t f U U i y u s I'A'IHJ RAVCRUIV^KLILV, O U Í T I T W I , MUMCJVI UM 33
We now maximize P.
dP n
^ = 0. (4.3.20)
This yields
2 r e - 2 Z r / a [1 - Zr/a\ = 0 . (4.3.21)
So, the answer is r = a/Z.

c) We know that the average energy in this state equals the energy in this state
since we have an eigenstate of the Hamiltonian. Therefore,
Z2e2
M = --2^- (13.22)
The average potential energy is given by
(V) = J \ip\2V(r)dv . (4.3.23)
So,
2
- Z e
(V) = \ e -Zr!a r2dr
Jo r
Ze2
(4.3.24)
a
Then, the average kinetic energy (T) is given by
<T> = (H) - (V) = -7^- • (4-3.25)
4.4 Bound State Wavefunction, Current,

Momentum
a) Show that the wavefunction for a particle in a bound state may always be
chosen to be real.
b) By computing the current density give an explanation of the physical meaning
of this result.
c) Use the results obtained to show that for a bound state the expectation value
(p) of the momentum vanishes.
Solution
a) To obtain the wavefunction for a bound state we have to solve a differential
equation with all the coefficients real and real boundary conditions. As a conse-
quence, we may always choose the normalization constant so that the solution
is real. This is never the case for scattering problems where the boundary con-
ditions describe travelling waves and are of necessity complex.
34
*-XJ ± Jl
b) Now, suppose that we have a real wavefunction V- The probability current

in this case is, as always,
7 = — = 3{V-*VV>} • (4.4.26)
J
2im K
-r > l m
Since V is real, this current vanishes. This means, using the equation of conti-
nuity, that the probability density remains constant in time. Thus, a particle
in a certain region of space, determined by this wavefunction, remains in that
region of space. We have a bound state.
c) Writing out the expression for the current we have
7=o = ^ - ( r ^ - ^ r ) = ^ - [ r ^ + ( p r ) i ] . (4.4.27)
2im lm
Therefore,
= x
(p) \ J ^ Wi»l> + (pV'*)^] = 0 . (4.4.28)
Thus, the average momentum of a particle in a bound state vanishes. On the

average the particle remains where it is.
4.5 T i m e Evolution for Particle in a B o x

For a particle in a box with sides at x — ± a the eigenfunctions and eigenvalues
are
TTX ti2 7T2
ip+}n = 4 cos[(n + 1/2).—] , E+ }n = (n + 1/2) 2
a ¿ma¿
„ . , TTX. , h2 7T2
ip->n = Bn sin n — J , E- n = n t .
a 2ma¿
If we have a particle in such a box and its wavefunction at time t = 0 is given
by
7TT
\P(0,£) = -7=sin[5—]
y/a a
find \P(t, x) .
Solution
We are given the wavefunction at time t = 0 as
^(0,2;) = -^= sin(57ra;/a) . (4.5.29)

y/a
This is an energy eigenfunction with energy
£ 5 1
= £ • <4-5-30'
4.6. PARTICLE IN vu*- '1VJ OO
Th we have a stationary state and the corresponding time-dependent wave-

function is
^ sin(57ra;/a) e x p ( - i E t / h ) (4.5.31)
or
1 . . , ,25hn2t
V(t,x) = -^sin(57rx/a)exp(-z-^^-) . (4.5.32)
4.6 Particle in Box: Energy and Eigenfunctions

Find the
the energy levels and wavefunctions
wavefum for a particle in the potential
rr/ \ Í oo
OO ifif a;X << 00 ,, xX >; a
V
(*)=( -V0 ifO<x<a
Solution
The Hamiltonian for this problem is
P~
H = ^ - V0 0 < x < a . (4.6.33)
The fact that the potential is oo outside the interval 0 < x < a means that the
wavefunction must vanish at both x = 0 and x = a. If we call
2 2m(E + Vo)
¿
* = — ^ (4-6.34)
the Schródinger equation becomes
d2ip „
— + k V = 0 0 < a; < a . (4.6.35)
The physically acceptable solution is therefore of the form
ip(x) = ^ s i n ( ^ x ) 0< x < a (4.6.36)
and we require that
Ha) = 0 (4.6.37)
so that
ka = nn n = 1,2,3,... . (4.6.38)
Thus, the (unnormalized) wavefunctions are
« * ) = ( *.»"<»«/«) for 0<«<a
v
^ U tor z < 0, z > a
The corresponding energies are
„ h2Tr7n2
EN = ~ 2 Vq n = 1 , 2 , 3 , . . . . (4.6.40)
The normalization is given by
An = y j l . (4.6.41)
4.7 Particle in a B o x
If a particle, in a box with sides at x = ±a, is in a state described by the
wavefunction
f(x) = Acos(nx/2a) + B sm(7Tx/a)
a) Choose A and B so that the particle is in the lowest possible energy state.
b) Choose A and B so that the state has parity = + 1 .
Solution
a) For a particle in a box with sides at x = ± a the (unnormalized) eigenfunctions
are:
parity = + 1
h2k2
ip = coskx |x| < a , ka = (n + 1/2)7t , En = —— (4.7.42)
2m
or for parity = — 1
h2k2
ip = sin kx |a:| < a , ka = nir , En = —— . (4.7.43)
2m
Thus, for the wavefunction
f(x) = A cos(7rx/2a) + B sin(nx/a) (4.7.44)
the lowest energy is obtained with 5 = 0.
b) The solution with positive parity requires that f(—x) = f(x). This also
requires B = 0.
4.8 Particles Incident on a Potential

A beam of free particles with intensity N particles per second and energy E > Vo
is incident on a potential
( 0 x < 0
Vo 0 < x < a
0 x > a
Find the number of particles that are reflected in one second. How many par-
ticles per second are transmitted?
Solution
For a beam of particles with energy E > Vo incident from the left we have only
a transmitted beam on the right and the wavefunction is given by
iA(eikx + Re~ikx) x<0
ip=\ A {CeiKx + D e~iKx) 0 <x<a (4.8.45)
I ATeikx x > a
The constant A is determined by the condition that the incident beam A e'kx
has a current density of N particles per second. This means that
Hk . - o
N = —\A\ . v( 4 . 8 . 4 6 )
m '
The wavevector k is defined by the energy of the incident beam
h2k2
£ = — (4.8.47)
and K is defined by
h2I\2 „
= +
~2m~ ' <"'48>
Matching the wavefunction and derivatives at x = 0 and i = « w e get
1 + fl = C +D
ik{l - R) = iK(C — D)
iKa iKa
Ce + De~ = Teika
iK (CeiKa - De~iKa) = ikTeika. (4.8.49)
Solving for R and T wefirstfind
C = ~[I< + k + R(I< ~ k)}
D = ~ [ K - k + R(I< + k)]. (4.8.50)
Then, we get two expressions for T e ' k a
Teika = ^-[K + k + R{K-k)]eiKa
+ ~[I<-k + R(K + k)]e-iKa
Teika = ~[K + k+R{I<-k)]eiKa

ZK
- ^[I< - k + R(I< + k)]e~iKa . (4.8.51)
Thus, by subtracting these two equations, we finally get that
R (k2 — K2) sin Ka

(k2 + K2) sin Ka + 2 i k K cos Ka
2ikK e-i(k-2K)a
rp JV C ^ g
(k2 + K2) sin Ka + 2ikK cos Ka
So, the number of particles per second that are reflected are given by
N\R\* = N (k2-I<2)2sin2K a
(4.8.53)
(k + K ) sin 2 Ka + 4k2K2 cos 2 Ka
2 2 2
3$
J
V_v_i J / 1 I -1. JUI1/ ~r • vxj * ^ i jj x" X l/V-***—* -
and the number of particles per second t h a t are transmitted are given by
dí»^ Íí'2
2
TV i n = N r——— . (4.8.54)
2 2 2
(k + K ) sin Ka + 4k2K2 cos 2 Ka
Notice t h a t
N\R\2 + N\T\2 = N (4.8.55)
as conservation of number of particles requires.
4.9 T w o B e a m s Incident on a Potential

A. beam of free particles with intensity NL particles per second and energy
EL > Vo is incident from the left on a potential
Í 0 x < 0
V(x) = i v ó 0<x<a
\ 0 x > a
At the same time another beam of free particles with intensity NR particles
per second and energy ER > Vo is incident from the right. Calculate the total
particle current travelling to the right. [4.1]
Hint: The waves from the left and right are completely independent and their
scattering from the potential can be handled independently.
Solution
To solve this problem we simply use the results of problem 4.8. Then the total
particle current to the right is given by the transmitted part of the current
incident from the left plus the reflected part of the current incident from the
right. Corresponding to these two currents we have a transmission amplitude
TL for the current from the left given by
T 2ikLKLe^k--2K<•)"
L
(k2L + A'2) sin K i a + ' H k i K i cos K ^ a
where
h2k2
EL = ^ (4.9.57)
and also
h2 KL2
E l = ^ -
+ V0 . (4.9.58)
2m
Similarly, we have a reflection amplitude RR for the current incident from the
right given by
Rr = — (k2R + K2R)smKRa g
K K
( R + R ) s i n KR(I + 'LIKJI KR C O S KRQ
wianw xw j i u u i i~ x » 3y
where
h?kf>
f * = - j — (4.9.60)
and also
VKl Tr
ER = — — — h Vó • (4.9.61)
Then, the total particle current JR to the right is given by
Ír = NL\TL\2 + NR\RR\2 . (4.9.62)
4.10 Ramsauer-Townsend Effect

Consider scattering of a particle with energy E > V0 > 0 from a potential
Í 0 if x < 0
V(x) = ^ V0 if 0 < x < a .
[ 0 if x > a
Show t h a t if the wavelength of the particle in the region 0 < x < a is such t h a t
n\ = a then no reflection occurs. Give a physical explanation for this result.
The actual Ramsauer-Townsend Effect was observed in the scattering of
electrons off atoms of the noble gases since due to their closed shell configurations
these atoms have very sharp outer boundaries.
Solution
This is exactly the case of resonance transmission discussed in [4.3]. The
Schródinger equation reads
d ip k ib x < 0
—K¿xp 0 < x < a (4.10.63)
dx2
—k~ib x > a
where
k2 ^ K> = 2
""E - F
°> . (4.10.64)
h¿ ñ¿
Assuming that we have an incoming wave from the left so that for x > a we
have only a transmitted wave, the solution may be written
i eikx + Re~ikx x < 0

V>=< A eiKx + B e~iKx 0 <x<a (4.10.65)
I Te+ikx x > a
Matching the wavefunction and first derivative at x = 0 and x = a we find
1+ R = A + B
ik(l — R) = iK(A — B)
Ae iKa
+ Be -iKa Te ika
%K (A JKa Be -¡Ka
ikT eika (4.10.66)
Solving these equations for the transmission amplitude T and computing the
transmission probability T = |T| 2 we get
r= i +
4E{E - V0)
i2 Ka (4.10.67)
Clearly for sin K a = 0 we have T = 1. Thus, if
Ka — nn (4.10.68)
which is the same as
nX = 2 a (4.10.69)
we get total transmission.

A physical picture of what happens is that the wave inside the barrier
matches exactly onto the wave outside the barrier. We have nodes at the edges
of the barrier.
4.11 Wronskian and N o n - d e g e n e r a c y

in 1 Dimension
a) For the one-dimensional Schrodinger equation with potential V(x) and any
two independent solutions u(x) and v(x) , corresponding to the same energy,
show that the Wronskain
W(x) = u'(x)v(x) — u(x)v'(x)
is a constant.
Hint: Write out the Schrodinger equation for the two solutions and multiply
each of the equations by the other solution.
b) Use the result of part a) to show t h a t in one dimension the energy eigenvalues
of a Hamiltonian
H = £ . + V(x)
defined on €^{—00, 00) are non-degenerate.

Hint: Assume that one of the eigenvalues is degenerate and show that the
Wronskian of the eigenfunctions corresponding to this eigenvalue vanishes. Use
this fact to obtain a contradiction.
Solution
a) Since the eigenvalue E is degenerate the two solutions corresponding to this
eigenvalue satisfy
p iz + (k2 + U(x)) u = 0
dx
2
^ l + ( k + U(x))v = 0, (4.11.70)
where, as always,
*> - , u{z) =
n2 ' w
h2 '
If we multiply the first of (4.11.70) by v and the second of these by u and
subtract, we get
<4n-7i>
This may be rewritten as
dW _ d du dv
V 0 . (4.11.72)
dx dx di ~ Udi
Therefore, W is a constant and we have shown the desired result.
b) Since, the Wronskian is a constant, we can evaluate it at any point such as
x = ±oo where we have that
u(±oo) = v(±oo) = 0 .
Thus, we have that W = 0. From this it follows that we can write
1 du 1 dv
(4.11.73)
u dx v dx
Integrating this equation we obtain that
l n u = lnv + ln^4 (4.11.74)
where A is an integration constant. Thus,
u = Av (4.11.75)
so that u and v are linearly dependent, contrary to our assumption that there
are two linearly independent eigenfunctions corresponding to the same energy
E.
4.12 S y m m e t r y of Reflection
Consider a repulsive potential V(x) > 0 such that it vanishes outside a finite
region a < x < b . Show t h a t the reflection coefficient is the same for a particle
incident from either the left or the right. Assume that the solutions in the
interval a < x < b have been normalized such t h a t their Wronskian (see problem
4.11) is 1.
Solution
Let u(x) and v(ar) be two linearly independent solutions of the Schrodinger
equation in the interval a < x 6
Imposing as usual the condition of continuity of xfi and ip' we find t h a t
eika + Re~
ika
ika
ika
= Au(a) + Bv(a)
ik (e - Re~ ) = Au'(a) + Bv'{a)
ikb
Te = Au(b) + Bv(b)
ikTeikb = Au'(b) + Bv'(b) . (4.12.77)
Eliminating A and B we find from the last two equations
A = Teikb [v'(b)-ikv(b)]
ikb
B = -Te [u'{b) -iku(b)] . (4.12.78)
Here, we have used the fact t h a t the Wronskian u(x)v'(x) — v(x)u'(x) = 1. After
substituting these into the first pair of equations (4.12.77) we find
eika + Re~ika
_ rp £ikb _ ikv(b)u(a) — u'(b)v(a) — ikv(b)u(a)]
ika ika
ik (e -Re~ )
_ ji eikb [v'(fyu1 _ ikv(b)u'(a) — u'(b)v'(a) + iku(b)v'(a)]. (4.12.79)
Now, eliminating T we get
_ 2ika E -f k2P + ik(F — C)
R e
- ~ E — k2D + ik(F + C) ( 4 1 2 8 0 )
where
C — u(a)v'(b) — v(a)u'(b)
D = v(a)u(b) — u(a)v(b)
E = u'(a)v'(b) — u'(b)v'(a)
F = u(b)v'(a) — u'(a)v(b) (4.12.81)
are real numbers. Therefore,
,„l2 (E + k2D)2 + k2(F — C)2
' ' ~ (-Ek2D)2 + k2(F + C)2 • (4.12.82)
Similarly, for waves incident from the right we have

C eikx + R e~ikx x>b
ip = < Au(x) + Bv(x) a < x < b (4.12.83)
I T eikx x < a
v vyi/j-/ l u v y j r i r / i * i 43
Again imposing the boundary conditions at x = a and x = b and solving for

\R\2 we get
(E + k2D)2 + k2(F - C)2

' ' ~~ (E-k2D)2 + k2(F + C)2 ' (4.12.84)
T h e numbers C, D , É , F are obtained from the numbers C, D ,E ,F respec-

tively by replacing a by b and changing k to — k. Under this transformation we
find that
E ^ - E (4.12.85)
so t h a t
C=F
F = C
D = —D
E = - E . (4.12.86)
T h e transformation (4.12.85) clearly leaves | f i | 2 unchanged and thus proves

our result.
4.13 Parity and Electric Dipole M o m e n t

Prove Laporte's rule, t h a t is, show t h a t a system of N particles with charges
<?i, 92 • • •, (¡N in a state of definite parity cannot have an electric dipole moment.
Solution
If the wavefunction for the state of the system is
V » ( « i , x2, x N ) (4.13.87)
and if P is the parity operator then, by assumption
PV = ± 1 V • (4.13.88)
The expression for the electric dipole moment is
We also have t h a t
PxP = - x P
^2 2
= - x .
q Xi
'
(4.13.89)
(4.13.90)
If we write out the expectation value on the right hand side of (4.13.89) and
apply the operator P2 — 1 we have
N
( N \
pp
pip,p^2^> ^j
( N \
± v , 5 > ( p i i p )(±)v>
\ i=i /
= - Y^qiXi^j
= -d . (4.13.91)
Thus, the electric dipole moment must vanish.

4.14 Bound State D e g e n e r a c y and Current
a) Show that a single-particle system, having a non-zero current and evolving
according to a Hamiltonian H that commutes with the time-reversal operator
T, must be in a degenerate state of this Hamiltonian.
b) Furthermore show that if all the states of the Hamiltonian H are non-
degenerate then the system cannot have a non-zero current and may conse-
quently be described by a real wavefunction. (See also problem 4.5.)
Solution
a) The current is given by
/= — — . (4.14.92)
Now suppose that x¡) is an eigenstate of the Hamiltonian H. Thus,
Hi¡> = Eip . (4.14.93)
Next we use the fact t h a t the time-reversal operator commutes with H to act on
(4.14.93) from the left with T and write (remember that the energy eigenvalue
E is real)
THT~1Tip = ETip . (4.14.94)
Using the properties of the time reversal operator we can rewrite this equation
as
= HiP* = ExP* (4.14.95)

Thus, ip* is again an eigenfunction of H . This leaves only two possibilities
1) T h e state is non-degenerate so t h a t a p a r t f r o m a phase \¡) and tp* are the
same
ip*=eiai/> . (4.14.96)
In this case the current (4.14.92) clearly vanishes. Thus, we are left with the
second possibility.
2) T h e state is degenerate so t h a t
ip*¿eiaip . (4.14.97)
In this case we have a non-zero current.

b) As already seen, if the state is non-degenerate we have
v>* = e i a I p . (4.14.98)
a2
Now, since the phase of ip is arbitrary we can multiply ip by e~' I . In t h a t
case we get
ia/2 ia ia 2 ia 2
e = e e ~ ' ip = e ' ip (4.14.99)
so t h a t , after cancelling the factors of e I a / ' 2 on b o t h sides we see t h a t
= ip . (4.14.100)
4.15 Car R e f l e c t e d from a Cliff

Suppose a car with a mass of 1000 kg travelling very slowly at 1.0 c m / s ap-
proaches a very sharp drop. C o m p u t e the probability t h a t the car is reflected
from this cliff.
Solution
To solve this problem we need the potential energy V^(x) of the car t h a t is
travelling in the x-direction. We assume t h a t the edge of the cliff is located at
x = 0 , y = 0. T h e n for x < 0 the potential energy of the car is a constant
which we set equal to zero. For x > 0 the potential energy is m g y . But, the
classical trajectory of the car as it goes over the cliff is given by
2
X = vt , y = - ^ g t (4.15.101)
or
2
3 / = - ¿ z - (4.15.102)
Thus, the potential energy for x > 0 is given by
V { x ) = ~ T ^ ¿- x 2 x>0 . (4.15.103)
¿v
So, the H a m i l t o n i a n is
H = { % 2 * - ° (4.15.104)
2
1 *>°
If we now set
k 2 = ^ ¿f , <y2 = ^ ¿ 4¿ (4.15.105)
ñ hv
t h e n t h e Schrodinger equation for t h e effective H a m i l t o n i a n becomes
d2ib ,
¿
+ k ip = 0 x < 0
ax
^ 2 + (k2 + a2x2) $ = 0 x > 0 . (4.15.106)

dx
For x < 0 we have an incoming wave plus the reflected wave. So, we have
ip(x) = eikx + fie-'*1 X< 0 . (4.15.107)
For x > 0 we have a purely t r a n s m i t t e d wave. So, we try a solution of the f o r m
ip = e i k x f ( x ) x > 0 . (4.15.108)
T h e n , f ( x ) satisfies
/" + 2ik f + a2x2f = 0 . (4.15.109)
T h e solutions of this equation can be expressed in t e r m s of t h e confluent hy-

pergeometric f u n c t i o n . However, since we are only interested in o b t a i n i n g the
reflection a m p l i t u d e R a n d not t h e entire solution, we can simplify t h e problem
even f u r t h e r at this stage. T h u s , we only need enough of the solution for f ( x )
t o be able to m a t c h the wavefunction and its drivative at x = 0. So, we only
need t h e solution near x = 0 and we can try
f ( x ) = a o + a.iX + a2X 2 + a . 3 X 3 + - - - . (4.15.110)
A f t e r s u b s t i t u t i n g this expansion into (4.15.109) we get t o order x2
2ü2 + &Ü3X + 2ik(ai + 2a2x + 3a3X2) + a2aox2 —0 . (4.15.111)
Therefore,
2a 2 + 2ika\ — 0
603 + 4ika-i — 0
2
6ika3 + a a o = 0 . (4.15.112)
T h u s , we find
• 2
ia
«i = - ^ « 0
a2 = ~ a 0 • (4.15.113)
So, the wavefunction near x = 0 is given by
r eikx + Re~ikx x < 0

=
[ a0 eikx ( l - + • • •) x > 0 ' (4.15.114)
We next match the wavefunction and derivative at x = 0 and get
1 -(- R — a0
ik(l-R) = ika0
r a2 1
I-TT4 (4.15.115)
So, we immediately get
ñ
a2
«o«l , = ¿ 4 - (4.15.116)
The reflection probability is therefore given by
= | j R | 2 =
^ 1 ^ ' (4.15.117)
Next we replace a and k in terms of
o m2q2 l9 2m(mv2/2)
to get
K 1 ( gh
=ió(^J • < 415U8 >
An order of magnitude calculation now shows t h a t
11 < 1CT 128 . (4.15.119)
This shows why classical mechanics works so well when applied to macroscopic
objects.
Bibliography
entific Publishing Co. Pte. Ltd., section 4.3 and 4.4, (2002) .
[4.2] ibid, section 4.5.
[4.3] ibid, section 4.6.1 .

Chapter 5
More One-Dimensional
Problems
5.1 Motion of a Wavepacket

A wavefunction
oo
/
f(k)e^kx~u^ dk
-oo
is normalized such t h a t
r oo
/ mt,- , x)|2 dx = 1 .
J — oo
Assume /(A:) is a smooth function vanishing rapidly at infinity. Show that the
velocity of the centre of mass x defined by
oo
/
x)|2 dx
-oo
is given by
Hint: Use the fact t h a t
oo
/
«.»'<*-«)* dx = 2irS(k -q) .
-oo
Show furthermore t h a t if ui(k) is adequately approximated by
w(fc) « w(0) + (k - k0) ~

0.1. yJr /i \vrí v _l 4y
then
dx duj
dt ~ cf/:
the group velocity.
Solution
We first write out j*?!2 in detail.
i[kx—u>(k)t]
dk
/ oo /»00
r ( f f ) c — d q f ( k ) e'
-oo J — OO
OO pOO (5.1.1)
/
Now, / diwg / * ( g ) / ( a ) e '(fe-9)^ e -'^(fe)-w(g)]i
-oo J — OO
/ oo /«oo /«oo
•oo J — OO J — OO
/ OO roo poo
/ dkdqr(q)f(k)e-i^-u^t xe^-^dx. (5.1.2)
But, -oo J — oo J — OO
Ri(k-q)x _ ei{k-q)x
(5.1.3)
dq
Therefore,
/ oo /«oo i /»oo
/ dkdqf*(q)f(k)e~i^k'>-w^ti-j- / jV-ti'dx
-oo J—oo dq J—QQ
OO roo I
/
/
If we differentiate underd kthe
d q integral
f * ( q ) f {we - i ^ - uget
k ) enow ( 9 ) ] « f í - 2 j r í ( * - «) . (5.1.4)
-oo «/ — oo dq
dx
2n
dt / OO poo
/ dfcfg/*(?)/(*)[«(«)
•oo J — OO
(5.1.5)
Now integrate by parts to get the desired result

dx - r°°
í°° ... ^dw(k)
— = 2TT dk\f(k)f (5.1.6)
J — oo dk
If
io(k) « u0 + (k — k0) — (5.1.7)

dk
then,
du> „ du>
— sa 0 + — (5.1.8)
dk dk
ko
so t h a t
dx = [°° dk\f(k)\>.
dk (5.1.9)
dt \ ko J-CO
Now we use the fact that
J \ip\2dx = 27r J \f(k)\2dk = 1
to get the final result,

dx doj
(5.1.10)
dt dk k
0
5.2 Lowest Energy States

A particle is in the potential
V(z) = Vo exp[ax 2 /2] .
Estimate the energy of the 2 lowest eigenstates. W h a t are the parities of these
states?
Solution
The potential is
V = VQ exp (5.2.11)
The lowest lying states will be centred near x = 0. In the vicinity of this point
we have
1
l + -az2 (5.2.12)
This approximate potential is just the potential for a simple harmonic oscillator
with its zero shifted and with a spring constant k = Vo a. Therefore,
m = x/Voa/ra . (5.2.13)
Then,
Eo*Vo + ^hui (5.2.14)
corresponding to the ground state of a simple hamonic oscillator with parity

= + 1 , and
E i « V0 + - hw (5.2.15)
corresponding to the first excited state of a simple hamonic oscillator with parity
= -1 .
5.3 Particle at Rest
Find the wavefunction for a particle at rest at the origin of a coordinate system
fixed in space.
Solution
We have to solve the eigenvalue problem
(5.3.16)
The only function with this property is
(5.3.17)
5.4 Scattering from T w o D e l t a Functions

Calculate the transmission probability for a particle of mass m incident on a
potential
V(x) = A[i(x + a) + S(x — a)] .
Compute also the phase shifts.
Solution
The potential is an even function of x so parity considerations m a y be used for
the phase shifts. But first we need the correct boundary conditions. These are:
1) xp(x)is continuous at both x = ± a .
2) The S-functions indicate t h a t there is a discontinuity in dip/dx at x = ± a .
As usual, we find this discontinuity by integrating the Schródinger equation
over a tiny interval t h a t includes the discontinuity. T h e Schródinger equation
in terms of the reduced variables
2 _ 2mE 2mAo
(5.4.18)
h2 ' K
~ h2
reads
^2" + = K[S(x + a) + S(x — a)]ip . (5.4.19)
The discontinuity in the derivative at x = ± a is therefore given by
(5.4.20)
The solution may be written
x < —a
—a < x < a (5.4.21)
x > a
Here we have already imposed the physical boundary conditions t h a t for x < -a
we have an incoming and a reflected wave and for x > a we have a purely
transmitted wave. For — a < x < a we have a combination of waves travelling
to the left and to the right.
Applying the boundary conditions at x = —a we get
e~ika + R e'ka = A e~ika + B eika (5.4.22)
and
ik (e~ika - Reika)-ik (.Ae~ika - B eika) = -K (e~ika + Reika) .(5.4.23)
Solving for A and B we find
K + 2ik „ K ^ „
A = + R
^ r 2 T k ' <5-4-24>
and
r, K ->ikn „—I\+2ik
= + (5.4.25)
Similarly, applying the boundary conditions at x = a we get
A eika + B e~ika = T eika (5.4.26)
and
ik (A eika + B e~ika) + ikT eika = KT eika . (5.4.27)
Again, solving for A and B we find
A B =
= ' ~ékTe2lka • (5-4-28)
Equating the two expressions for A and B we get two equations for R and T.
(K - 2ik)T + KRe2ika = -(K + 2ik) (5.4.29)
and
KTe2ika + + 2 i k jR _ Ke-2ika (5.4.30)
After solving these for T and R we get
[ * ' + ( * • + 2,•*)»].-»»•
2iK2 sm(2ka) — 4k2 e~2,ka
and
K \(K - 2ik) e~4ika + (K + 2ik)]

R = — — (5 4 '421
2iK2 sin(2A:a) — 4k2 e~2ika '
To compute the phase shifts we divide the solutions into positive parity solutions
ip+ and negative parity solutions rp-
{ cos(kx —
ylcos(fcx)
x < —a
-a < x < a (5.4.33)

c o s ( k x + Í+) x > a
and
{ isin(kx
Bsin(kx)
— S-)
—a<x<a
x < —a
(5.4.34)
isin(kx + ¿-) x > a
Writing out the boundary conditions at either x = a or else x = — a we get
A cos(ka) = cos(ka + 6+) (5.4.35)
—ksin(ka + ¿+) + ksin(ka) = K cos(ka + ¿ + ) . (5.4.36)
Therefore, we get
tan (ka) = tan(&a + 6+) + — . (5.4.37)
K
So,
£+. = arctan[tan(fca) — K/k] — ka . (5.4.38)
Proceeding in the same way for the negative parity solution we find
B sin(ka) = isin(ka + S-) (5.4.39)
ik cos (ka + 6-) — kB cos (ka) = iK sin(A;a + ¿_) . (5.4.40)
Therefore,
cot(fca) = cot (ka + J _ ) — — . (5.4.41)
k
So,
= arccot[cot(/?a) + K/k] — ka . (5.4.42)
5.5 Reflection and Transmission Amplitudes:

P h a s e Shifts
Complete all the steps in going from
1 z,0+
R = - [e2i<5+ 1
2
T= i [e2iS+ +e2iS~]
= arctan tan KA ka
k
and
1
S- = cot — cot KA — ka
k
to
r
(k2 — K< 22\-
)e — 2ika
T = 2 2
(k + A' ) + ikK(cot Ka — tan Ka)
and
ikK (cot Ka + tan I\a)e' -2 ika
R =
( k + A' 2 ) + ikK (cot Ka — tan Ka)
2
Solution
Our starting point is
T = I(e2« + + e 2.í-) ) R = 1 (c«¿+ _ e2iS-
Also,
á + = a r c t a n f — tan(A'a)] — ka
k
and
JS
S_ = arccot[—cot(A'a)] — ka .
From (5.5.44) we get
t a n ( J + + ka) = — tan Ka .
K
But,
ei(i++ka) _ e-i(S++ka)
t a n ( í + + ka) = êi(i++ka) + g_i(i++/ca)j
e2i(S++ka) _ i
j[e2i(S++ka) _|_
Therefore,
e2i(S++ka) _ l _ ê2i(S++ka) + 1], K tan A'a

k
So,
k
e2is+ _ + lK tan Ka ^_2ika
k — iK tan Ka
Similarly, starting with (5.5.45) we get
i[e2'(s-+ka^ + 1]
c o t ( o _ + ka) — e2i{S_+ka) _ i]
K
— — cot K a . (5.5.50)
k
Therefore,
¿lib k — iK cot Ka -2 ika

(5.5.51)
k + iK cot Ka
Combining these results we have
T = - (e2iS+ + e2i6~)
k + iK tan Ka k — iK cot Ka -2ika

k — iK tan Ka k + iK cot Ka
k2 - K2 — 2ika
(5.5.52)
k2 + K2 + ikK[cot Ka — tan A'a]
and
R = e
)
1 k + iK tan Ka k — iK cot Ka — 2 ika
k — iK tan Ka k + iK cot Ka
ikK [cot Ka + tan Ka\ 2ika
(5.5.53)
k2 + K2 + ikK [cot Ka — tan A'a]
5.6 Oscillator Against a Solid Wall

Find the energy levels and normalized wave functions of the stationary states
of a particle moving in the potential
i
oo if x < 0
V(x)
kx2 if x > 0
~{
Compare the zero point energy in this potential with that for the simple har-
monic oscillator with the same force constant k. Explain any differences you
find.
Solution
For x > 0 we again have just the simple harmonic oscillator solutions
iM*) = ffn(y)e-1/2»a , y = J ~ x , En=(n+ l/2)hu . (5.6.54)

However, we now have to impose the additional boundary condition t h a t the
wavefunction should vanish at x = 0. T h e easiest way to do this is to take the
solutions of the simple harmonic oscillator above and keep only the odd parity
solutions since they must vanish at x = 0. This requires t h a t n should be an
odd integer
n = 2p+l , p = 0,1,2,3,... . (5.6.55)
This means t h a t the lowest energy is now the same as that of the first excited
state of the simple harmonic oscillator. Since the particle is more confined,
the ground state energy is greater. Furthermore all the even parity states are
missing.
5.7 Periodic P o t e n t i a l
Find the equation t h a t determines the allowed energy bands for a particle mov-
ing in the periodic potential (see figure 5.1)
V(x)
V0
a 2a + b 3a + 26 4a + 36 x
(a -(- 6) 2(a -|- 6) 3(a -(- 6)
Figure 5.1: Periodic potential with period (a + 6).
0 0 < x < a
V =
Vo a < x < a+ 6
where V(x) = V(x + (a + 6)) = V(x + 2(a + 6)) = . . . .
Solution
We first solve the problem in the interval —b<x<a and then use Bloch's
Theorem [5.1] to extend the solution to a < x < a + 6.
For 0 < x < a we have with
k\ = 2mE
V
that
1p = c1eiklX + c2e-'fc'r 0 < a: < a (5.7.56)

For — 6 < x < 0 we have with
2 _ 2m(E - V0)
2
h2
that
= c 3 e'752* + C4 e~ik2X - 6< x < 0 (5.7.57)
Now using Bloch's Theorem we get that
= e i i f ( a + 6 ) c3c ik-2(x — a —+ b)C 4 e — ik-2(x — a — b)a < x < a + b . (5.7.58)
The wavefunction and its derivative must be continuous. Imposing these con-
ditions at x = 0 and x = a we get
C j + C2 = C3 + C4 .
mci-c2) = mc3-c4) •
ikia ikia
Cle +c2e~ = eiK{a+b)[c3.e-ik*b + c4eik*b] .
kx [Cleikia -c2e-ikia] = k2 e
iK a+
( V [c 3 e~ik>b - c 4 eik*b] . (5.7.59)
To obtain a nontrivial solution for the parameters c¿ we require that the deter-
minant of their coefficients vanish
1 1 -1 -1
k1 —k 1 —¿2 k2
Jk ia _eiK(a+b)e~ik2b _eiK(a+b) eik2b = 0
e-ikia
kxeikia -kxe~ikia —jfc2 eiK(a+b)e-ik3b ¿ 2 GiK(a+b) e ik 2 b
(5.7.60)
Expanding this out and after some tedious simplifications we get that
¿fe? + k 2
cos I\(a + b) = cos k\a cos k2b—-—-— sin k\a sin k2b . (5.7.61)
ZK1K2
If E > Vo this equation determines the energy bands since it requires that
k\ + k2
— 1 < cos k\a cos k2b — • sin ¿ j a sin ^26 < 1 for E > Vo • (5.7.62)
2k\k2
For E < Vó we have to replace k2 by Í K 2 . In this case the energy bands are
determined by
k\
— 1 < cos k\a cosh k2b • sin k\a sinh K2b < 1 for E< Vq .(5.7.63)
1k\K2
5.8 R e f l e c t i o n and Transmission T h r o u g h

a Barrier
A particle is incident on a smooth, one-dimensional potential barrier whose
maximum height is greater than the energy of the particle. Furthermore, the
potential goes to zero rapidly at x —> ±00. Verify unitarity. T h a t is, show that
the sum of the reflection and transmission probabilities is unity.
Hint: Use the equation of continuity for the probability current.
Solution
Since the potential vanishes rapidly for large values of |ar| we can look in the
asymptotic regions. If we assume t h a t the particle is incident from the left, then
for x - » —oo we have the incident wave plus a reflected wave, say
ip = eikl + Re~'kx . (5.8.64)
For x —> oo we have only a transmitted wave
ip = Teikx . (5.8.65)
T h e probability current in the two regions is

m ^ 2 ^ _qo
J
m
fib
j = —\T\2 x —> + o o (5.8.66)
The equation of continuity states t h a t

= 0 . (5.8.67)
ax
If we integrate this equation from — L to L for very large L we get
J l ^ ~ ^ d x = 0=j{L)-j(-L). (5.8.68)
Writing this out using the asymptotic forms for j we get
— [ 1 - |fi|2] = — |T|2 . (5.8.69)

m m
Therefore,
|fl|2 + |T|2 = 1 . (5.8.70)
5.9 Hermite Polynomials: Integral

Representation
Show t h a t if we define the Hermite polynomials by the integral representation
on poo
Hn(x) =—= / (x + is)n e~s2 ds (5.9.71)
V 71 " J — oo
then the following results hold.

a)
H'n(x) = 2n Hn_1(x)
b)
2x Hn(x) = Hn+i(x) + 2n Hn-i(x)
and finally
c)
H'¿(x) - 2xH'n[x) + 2n Hn(x) = 0 .
f h i s last equation shows that the functions Hn(x) defined by the integral
(5 9 71) satisfy the differential equation for the Hermite polynomials.
Solution
a) If we differentiate the integral representation under the integral sign we im-
mediately get
on roo
H'n{x) = = / n(x + i s ) " - 1 e~s ds = 2n Hn-i(x) . (5.9.72)
v71" j—oo
b) Also, we immediately see that
2 n +i r 0 0 2
2xHn(x) = —7=- / x ( x - f ¿ s ) " e _ s ds
v*" —00
on + l /-oo
= tfn+i(x) 7=r- / ¿s(x + ¿s) n e _ s ds . (5.9.73)
71
v " ./-oo
But,
oo /«oo
/
(x -f- zs) n de~s = / —2s(x + is)" e _ s ds . (5.9.74)
-oo J — OO
After an integration by parts this yields
oo ,00
/
is(x + is)" e~ s ds = — — / (x + i s ) n _ 1 e _ s ds . (5.9.75)
•oo ¿ J —OO
So, combining this with (5.9.73) we have t h a t
2 xHn(x) = Hn+1(x) + 2nHn-i{x) (5.9.76)
as desired.
c) Now, using (5.9.72) together with (5.9.76) we have that
H'n(x) = 2nHn_1(x) = 2xHn(x)-Hn+1(x) . (5.9.77)
From this it follows, after another application of (5.9.72) that
K(x) = 2Hn(x) + 2xH'n(x)-H'n+1(x)

= 2Hn(x) + 2x H'n(x) — 2(n + l)Hn(x)
= 2x H'n(x) — 2n Hn(x) . (5.9.78)
uu
5.10 Matrix Element B e t w e e n D e g e n e r a t e States

Let <f>i and <f>2 be two degenerate eigenfunctions of the Hamiltonian
T h a t is they correspond to the same eigenvalue E. Show t h a t the matrix element
(xp+px)<t>2) = J <t>l(xp + px)<f>2dx = 0 .
Hint: Use the fact t h a t xp + px is self-adjoint.
Solution
Using the hint we have t h a t
(<f>1,(xp + px)<t>2) = ((xp + px)<j>\, (j>2) • (5.10.79)
Writing this out we get
[ x +
7/ ^ ^ l{xh)]dx =
-U [ x
^ +
i{xri)]hdx- (5 10 80)
- -
Using this we can rewrite the m a t r i x element as
(xp + px)<j>2)
1h d<f>2 d<j>\ „ d d
<PiX— x-j—<p2 + <p1—{x(p2) - -- yx<pi)(p2 dx
27 ax ax ax dx
h 2 d
Í l * ^ ^1 dx . (5.10.81)
~ i J
[ dx x
dx < 01
Furthermore, using the fact t h a t the two eigenfunctions are degenerate we have
h2 d2<f>\
+ V<¡>\ = Eft
2 to dx2
h2 d2(j>2
+ ^02 = E(p 2 . (5.10.82)
2m dx2
So, multiplying the first of these equations by ^>2 and the second by <j>\ and
subtracting we get
d2<
- *1 ( A d
^ l*
&1
t>2
= 0 (5.10.83)
2 m V 2
c t e 2
rfx2
This may be rewritten to read
(5.10.84)
c?x dx dx
From this it immediately follows t h a t
( 5 10 8 5 )
•
If we now consider the expression

2 I J d<f>* ,, d(f>2
i ± dx (5.10.86)
J dx
we see that it vanishes since both <j>* and <j>2 vanish rapidly at ±00. If we
furthermore carry out the indicated differentiation and use (5.10.85) we get the
desired result that
/ 2X
~ ^ dx =
° ' (5.10.87)
5.11 Hellmann-Feynman T h e o r e m
Consider a Hamiltonian H(A) t h a t depends on a parameter A and let A be some
operator such that [A, H] is well defined. Prove that for any stationary state
^(A) of this Hamiltonian we have that
dX " ' \ 5a / v 5a
This is known as the generalized Hellmann-Feynman theorem. For A = 1 it is
the Hellmann-Feynman theorem.
Solution
The equation for the stationary states reads
H(X)iP{X) = E(X)iP(X) . (5.11.88)

If we multiply this equation from the left by A, take the expectation value of
this equation, and differentiate the resultant equation with respect to A we get
E
^' + + A H
= + + A
lfx^ ' (5.11.89)
But,
e
»-4i> =
= (ip, HA--) . (5.11.90)
Hence, we get the required result

^ ( A ) = {A^-) + W,[A,H]^-) . (5.11.91)
62 i J t l A f l Ml O. IVIVSJII^ \J1
Bibliography
Chapter 6
Mathematical Foundations
6.1 Cauchy Sequence in a Finite Vector Space

Show t h a t every Cauchy sequence in a finite dimensional vector space converges
strongly (converges using the norm).
Solution
We are given a sequence of vectors {vj j = 1 , 2 , 3 , . . . } where each of these
vectors is an element of an r-dimensional vector space and may therefore be
written for a fixed basis in terms of its components as
Vj = {vji,vj2,...,vjr) . (6.1.1)
Furthermore, we are told t h a t these vectors form a Cauchy sequence. This means
that given an e > 0 we can always find a number N such t h a t for n,m > N we
have
(l^nl I "i- | ^ n 2 ^m21 "I" ' ' ' ~t~ | ^ n r Vmr \ ) ^ ^ * (6.1.2)
As a consequence, we can bound each term |v n s — vms \ where 1 < s < r by e.

This means t h a t each component of the sequence converges to some finite limit
u s - Hence, it follows t h a t the limit vector is
U = (ui,li2, ...,ur) (6.1.3)
and the norm of this limit vector is finite.

Notice t h a t if our space were infinite dimensional, and t h a t even if every
component us were finite, we could not conclude t h a t
OO
is finite since we would now have to sum over an infinite number of components.
This is w h a t makes completeness in an infinite dimensional space non-trivial.
64 ( J H A f l ' E t í O. MAL nrAVl/l t iua L tXJUSXUAí IV1\S
6.2 Nonuniqueness of Schrodinger

Representation
Show t h a t if
[x, p] = i
then there exists a unitary operator S such t h a t for a given a real function f ( x )
we have
SpS~l=p + f{x) .
Solution
T h e proof is by displaying the operator S. In the Schrodinger representation
we have t h a t
P = - 4 - , x = x. (6.2.5)
i ax
Thus, we try with a real d i f f e r e n t i a t e function h(x) ,
5 = , 5_1 = S* = eih(-x) . (6.2.6)
Then, for a function g in the domain of p we have
' s = Is
= eih^\p+h'(x)}g(x) (6.2.7)
where
v w - | w .
Thus,
SpS-1 =P+h'(x) . (6.2.8)
So, we simply have to choose

h'(x) = f(x) . (6.2.9)
T h e unitary operator S serves to give the same additional phase to

every state in the hilbert space and thus has no observable effect. This means,
we are free to use the Schrodinger representation and don't have to worry about
this nonuniqueness.
6.3 D e g e n e r a c y and C o m m u t a t o r
Show t h a t if all eigenvalues of the self-adjoint operator A are non-degenerate
and t h a t if the self-adjoint operator B commutes with A, i.e.
[A, B] = 0
then B is a function of A.
74. VWr* I*iûJnni*IU O ü A / l i r u
'
Solution
We have
¿|„> = a„|n) - (6.3.10)
The kets |n) form a complete set. Also, using the commutator we have
B A \ n ) = a„B\n) = AB\n) . (6.3.11)
Thus,
A(B\n)) = an(B\n)) (6.3.12)
and because the eigenvalues an are non-degenerate we have that
B\n) = c„\n) . (6.3.13)
Now, corresponding to any ket |n) there is a unique eigenvalue an. Con-
versely, corresponding to any eigenvalue an there is a unique ket |n) . This
means that there is a one-one correspondence between n and an. So, we can
write
c„ = c ( a n ) . (6.3.14)
Using (6.3.13) it therefore follows t h a t
B\n) = c(a„)|n) = c(i4)|n) (6.3.15)
and since, as stated, the |n) form a complete set we have that
B = c(A) . (6.3.16)
6.4 von N e u m a n n ' s Example

Consider the operator
An = px2n+l + x2n+1p , n = 1, 2, 3, . . .
where
h d
P =
~idi '
Find the eigenvalues and eigenfunctions of An. W h a t are the deficiency indices
°f An? The Hilbert space in this case is C2(—oo, oo). For n = 1 this example is
due to von Neumann.
uu wiini J. JJJJ.i v. lVAsxx njuiris*.± ik^Í\LJ U¿M- lUISS
Solution
We write the eigenvalue as Ah. Then, the eigenvalue equation for An reads
2n-f 1 d/ _=
? a ( vx 2 « + 1 / ) + ?. Ah f . (6.4.17)
dx i dx
After rearranging this becomes
^_^ x -(2n + l) / + ?!l±i / = 0. (6.4.18)

dx 2 2x
Separating we get
1 = !ax-(2n + l)dj, _ . (6.4.19)
/ 2 2x
For a; 0 this can be integrated to yield

—¿A
/ = , ( n + l / 2 ) exp 4n - 2 n (6.4.20)
Since this function must be square integrable we consider
l^l 2 9(A), -2n

l/l 2 ,(2n + l)
exp (6.4.21)
2n
This requires t h a t 9(A) < 0 so t h a t the singularity at x = 0 does not cause the
function to blow up too strongly at the origin. Therefore, all eigenvalues are of
the form Ah with
A h = a — i(3 , /? > 0 . (6.4.22)
So we see t h a t we get exactly one solution of the equation
A\f = -if (6.4.23)
and no solutions of the equation

— +»/ •
(6.4.24)
This means that the deficiency indices are (0,1) and there are no self-adjoint
extensions of this operator. This example shows that, given two observables A
and B, with corresponding self-adjoint operators A and B, it is not generally
true that the operator AB + BA is self-adjoint and corresponds to an obervable.
6.5. PROJECTION OPERATOR 67
6.5 Projection Operator

A projection operator is a self-adjoint, non-negative operator P satisfying
2
P = P .
Let <j>n be a normalized eigenfunction of a self-adjoint operator A with only

discrete eigenvalues \ n .
a) Show that the operator P T¡J = <J>N{<L>n, *P) is a projection operator.
N
b) Show that A can be written
X
~ H / N<F>N(X)<T>N{Y)LP(Y) DY = ^ ANPN1¡J .
n n
This is called the spectral resolution of the operator A.
Hint: Assume completeness of the eigenfunctions.
Solution
We are given
Pn1> = <t>n(<t>nA). (6.5.25)
Hence, using that ( ) = 1 we get

P& = Pn(Pn^)
=
0n(<^n) v1))
=
0n)(<^n, vo
= KiKA)
= P n ( 6 . 5 . 2 6 )
b) Assume (using completeness) that
^ = Yân(j)n (6.5.27)
n
so that
= (</>„, V1) • (6.5.28)
Then,
V = (6.5.29)
n
and
= y^tA<í)n(<t)n,ip)
n
^ ^ (fan (<ftn ? *0)
n
X n P n
= J 2 ^ - (6.5.30)
68 CHAPTER 6. MATHEMATICAL FOUNDATIONS
6.6 Spectral R e s o l u t i o n
Find the spectral resolution (see problem 6.5) of the operator
A _
— (, 0- O
3 a
i + ia2
~~ ai~ — ia io
2 -a3
Solution
The operator is
as a
¿ = ( . i + m2 ) (6.6.31)
a\ — la 2 — a3
The eigenvalues are given by
- ( a 3 - x ) ( a 3 + x) - (a 2 + a 2 ) = 0 (6.6.32)
or
x2 = a 2 + a 2 + a | = a 2 . (6.6.33)
So that
x = ±a . (6.6.34)
The corresponding normalized eigenvectors are

, _ 1 f a + a3
+ a ia2
~ \j2a(a + a3) v i -
1
/_ = ( ,a~ a 3
, ) . (6.6.35)
y/2a(a - a3) V -( a i ~ l a 2 ) J
Thus, the projection operators are
a
P+ = f+f+f t =
_ J_ ( + a3 «i + ia 2
a — ia
2a I i 2 a - a3
P- = = ,a as
v («1+^2) \ (6.6.36)
2a V ~ ( a i - ía
2) a + a3 J
So, it is easy to see t h a t
aP+ + (—a)P_ = A . (6.6.37)
6.7 Resolvent Operator

For any operator A the corresponding operator
1
R(z) = ( A - z i r ,
where z is a complex number and 1 stands for the unit operator, is called the
resolvent operator. Show that for any square matrix A, R(z) is analytic in z
with poles at the eigenvalues of A.
6.8. DEFICIENCY INDICES 69
Solution
If A is a square n x n matrix then
R(z) = (A- z)'1 (6.7.38)
may be written as
matrix of cofactors o f M — z)
1 ( 6 9 )
= d^m-z) ' "
The cofactors are polynomials of degree < n — 1. The determinant is a poly-

nomial of degree n. Thus, R(z) is the ratio of such polynomials and is analytic
except at the zeroes of the denominator which occur when
det ( 4 - *) = 0 . (6.7.40)
All of these singularities are poles.
6.8 Deficiency Indices

Find the deficiency indices and hence all self-adjoint extensions of the Hamilto-
nian
H = ~*
2m dx2
defined on the interval (a,b).
Hint: It may be useful to express the boundary conditions on a function / £ DH
in terms of 2-component quantities
_ ( /(«) ^
and
_ ( m
F(a) = { W J = { fib)
and assume t h a t F(b) = UF(a) where U is a non-singular 2 x 2 matrix.
Solution
The Hamiltonian is
h2 d?
H (6 8 41
= • - - »
To find the deficiency indices when the Hilbert space is L2(a,b,dx) we have to
find solutions of
Hf± = ±if ± (6.8.42)
belonging to L2(a,b;dx). All solutions of these differential equations are given

by
f± — A exp(k±x) (6.8.43)
where
t 22 1.2
H k± = ±i . (6.8.44)
2m
Therefore, we find
_ v w e ,- 3 i r / 4 ^ _a/2^^/4 (6.8.45)
h h
k _y/}mei,r/4 ^ gi7r/4 (6.8.46)

h h
All four solutions belong to L2(a,b,dx) and thus, the deficiency indices are
(2,2).
To find the self-adjoint extensions we consider
*2
f
2mja dx2'
bn
( r * f _ drt + (Hf,g). (6.8.47)
2m y dx |o dx ^
So, we need
/ * ( % ' ( * ) - / * (a)5'(a) - /*'(W) + r'(a)j(o) = o . (6.8.48)
The same boundary conditions must apply to both f ( x ) and 5(2:). We therefore
define
F
«=((i-'$'(«)) ' GW = U-$-<«)) ,6 8 49)
--
(the factors of 6 — a are simply to have both components with the same dimen-
sion) and look for boundary conditions of the form
F(6) = A F ( a ) , G{b) = AG(a) (6.8.50)
where A is a 2 x 2 matrix. The condition for self-adjointness when now written

out reads
G(b) Ft(a)
FHb) ( J V ) = ( l "o1 ) g ( a )
' (6 8 51)
' '
In terms of the matrix A this requires:
A A =
' ( 1 ~ o ) ( i ~o ) • <6 8'52)
A general 2 x 2 matrix involves 8 real parameters and we have 4 conditions.

Thus, we will find 4 real parameters. The most general 2 x 2 matrix may be
written
A = eia ( a° + 0,3 ai
~ ta2 ^ (6.8.53)
\ aj + ia2 a0 — a3 J
6.9. ADJOINT OPERATOR 71
Since we need only 4 parameters we begin by choosing all 5 parameters in A to

be real. Now, writing out the condition on A we find
0 -1 \ , f 0 - a 2 + a\ + al + al
1 0 ) a = 0
0 -1
(6.8.54)
1 0
Therefore, we need only choose
ao = \Jl + af + g¡2 + a | . (6.8.55)
This fixes the matrix A and since it has exactly 4 parameters as required we
are finished. Thus, the domain of H consists of all twice differentiable functions
f(x) on the interval (a, 6) such that
ih i { % ( k \ ) = e
'" ( a ° l a 3 a i
~ t a 2 ) (ih \ v6-8-56)
(b-a)f(b) J \ai + ia2 a0 - a3 J \ (b - a)f (a) )y '
where a , a i , a2, a3 are real and ao = y/l + a\ + a^ + a§.
6.9 Adjoint Operator

Given an orthonormal basis set {u„ : n = 0,1, 2 , . . . } and an operator a which
has the following action on this basis:
aun = \fñ un-1 n > 0 .
Find the adjoint operator at by explicitly giving its action on this basis set.
Also find the commutator [a, a*] .
Solution
We have that the set {«„} is complete and orthonormal. Also,
aun = ^/ñunî . (6.9.57)
Therefore, we get
(Mm i ® ^n) —- — l) — Sm,n — 1 • (6.9.58)
This last term on the right may be rewritten as y/m+ 1 i m + i , n Using the defi-
nition of the adjoint operator we therefore have t h a t
(a 1 um,un) = \/m+ . (6.9.59)
Therefore,
« t " m = \ / m + l Um+1 - (6.9.60)

The commutator [a, at] is now computed using
aat un = Vn + 1 a u n + i = (n + 1) u„ (6.9.61)
and
a t a u „ = y/na) w„_i = nu„ . (6.9.62)
Therefore,
[a, at]u„ = (aat — a t a ) u n = (n + 1 — n) wn = u n (6.9.63)
Thus,
[a, at] = 1 . (6.9.64)
6.10 Projection Operator

Let A be an operator on a Hilbert space 7i and let {a n } be the set of all
eigenvalues of A. Show that if A has a complete set of eigenstates {|n)} such
that
A|n) = a„\n)
then
A
II ~üm (6.10.65)
dn 0-m
m^n
is a projection operator onto the eigenspace of a n , that is it projects onto the
space spanned by the states belonging to the eigenvalue an.
Hint: Prove that
Pn\l) = Snl\l) . (6.10.66)
Solution
To use the hint we consider a eigenstate |i) , I ^ n. In that case
m
Pn |/) = J J ' |/) = 0 (6.10.67)
an ctm
m^n
since in the product we arrive at a term with m = I. Now consider the eigenstate
I ft). Then,
Pn In) = J ] In) = \n) . (6.10.68)

mz£n
Thus, applied to an eigenstate we have
Pn\l) = Snl\l) . (6.10.69)

6.10. PROJECTION OPERATOR 73
It now follows that
Pi |/> = SnlPn\l)
= snl\l)
= Pn\l) • (6.10.70)
Since the eigenstates of A form a complete set we may expand any state in
terms of these
|¿) = 5 > I 0 • (6.10.71)

i
Then,
pn\4>) = X>P„|Z>
i
=
i
= bn\n) . (6.10.72)
Applying Pn to this equation once more, we get t h a t
P2|<¿>) = bnPn |n)

= bn |n)
= Pn\<t>) • (6.10.73)
Thus, we have shown that
Pn=Pn • (6.10.74)
This shows that Pn is idempotent as required for a projection operator. If we

now consider any state \<j>) as in (6.10.71) and apply Pn we get
n n I
= J2J2 ' "'\l)

bs
n I
6 Z
= E ' I > = I<¿> • (6.10.75)
I
Thus, we have shown that
P = 1
E « (6.10.76)
n
and PN is indeed the projection operator t h a t was claimed.

v y i i / l i JL JJJI t u . ivis\ ± jnn/ivis\ ± £ (J{JjyiJA± 1UNS
6.11 C o m m u t a t o r of Lz and (p
Formally one can derive t h e relation
[Lz, <p] = ih
and deduce f r o m it t h a t
A L z Aip > fi/2 .
Now, A h/2 .
Explain this a p p a r e n t p a r a d o x .
Hint: E x a m i n e the d o m a i n of Lz on which it is self-adjoint. A similar a r g u m e n t
also holds for [p, ar] = ih and a particle confined to a finite interval on the line.
See also [6.1] and [6.2].
Solution
Lz is defined as (h/i) (d/dtp) for f u n c t i o n s f(ip) differentiable and periodic in ¡p.
T h i s defines the d o m a i n of Lz and m e a n s
f((p + 2tt) = f(ip) (6.11.77)
So if / is such a function t h e n ipLzf(ip) is well defined, b u t g(tp) = is not

periodic and therefore does not belong to t h e d o m a i n of Lz. T h u s , we c a n n o t
even write Lz[tpf(ip)]. T h i s m e a n s t h a t t h e c o m m u t a t o r
[Lz,ip] = -ih (6.11.78)
is incorrect .
6.12 Uncertainty Relation: Lz and cos sin<¿>

In the previous problem, if instead of ip we choose periodic f u n c t i o n s of (p such
as cos ip or sin ip we have
[Lz, sin p] = —ih cos <p (6.12.79)
[L z , cos ip\ = +ih sin ip . (6.12.80)
These equations involve well-defined quantities. Use these c o m m u t a t o r s t o com-

p u t e t h e uncertainty relations between Lz a n d cos p as well as sin <p.
13. DOMAIN OF KINETIC ENERGY: POLAR COORDINATES 75
Solution
We know from general considerations [6.3] that
h,
A L z A sin ip > —(cos ip) (6.12.81)
ALZ A cos tp > ^(simp) . (6.12.82)
This does not lead to any contradictions. For example, if we are in an eigenstate
of Lz s 0 that the wavefunction is
m
^ * (6-12.83)
and AL z = 0 we see that we also have
(cos (p) = (sin <p) = 0 . (6.12.84)
This time, however, no contradiction occurs. The results simply mean that the
uncertainty relations read 0 > 0, a statement t h a t is clearly true.
6.13 D o m a i n of Kinetic Energy: Polar

Coordinates
a) Show that if ip(r, 9, ip) is to be in the domain of the kinetic energy operator
T = -(/i2/2m)V2
then, near r = 0 the function ip(r, 9, ip) must be bounded by A ra , a > —1/2
where A is independent of r.
Hint: Examine the condition that
Ti¡>(r,.,.) G £ 2 (0,OO)
near r = 0 by writing
V > ( r , 9, ip) = ra [a 0 + a i r + a 2 r 2 H ] (6.13.85)
where the a,¡ are independent of r, i.e. they depend only on 9 and ip.
Solution
Using the hint we have t h a t near r = 0
Tip = -*1 d2ip 2 dip L2ip

2m gr2 r qr r2
ra~2 [b0 + bir + b2r2 + • • •] (6.13.86)

where the bj are new functions of 6 and <p and are independent of r just like the
a j . Since the integration over r has the measure r2 dr it now follows t h a t near
r = 0 we have
(ip,Tip) = i drr2a[aobo + (a*1bo + a*0b1)r+•••] . (6.13.87)

Jo
For this integral to converge at the lower limit we therefore need that
a > ~ \ • (6.13.88)
This is the required condition for ip(r,.,.) E Dr-

Notice that if in a Hamiltonian
H =T + V
we have a potential V(r) such that V(r) is less singular than r~2 near r = 0
then the condition (6.13.88) obtained also sufRcies for ip to be in the domain of
this Hamiltonian, at least as far as the behaviour near r = 0 is concerned.
6.14 Self-Adjoint Extensions of p4

For the operator p4 defined on £2(0,00) find the deficiency indices and condi-
tions to make the operator self-adjoint.
Solution
For these computations it is convenient to choose units such t h a t h = 1. Then,
2 d2
P
dx!
í>4 = • (6.14.89)
To find the deficiency indices we need to find the solutions belonging to £2(0, 00)
of
^ = (6.14.90)
We try
f± = ea±r (6.14.91)
and find that
a 4_ei„/2 _ Q 4_e-,>/2 (6.14.92)
Thus,
i ± n 8
Q± — e / ^ ei±*/a+iv/2 ^ e«±jr/8+»ir ei±7r/8+»3jr/2 (6.14.93)
^ 6.14. SELF-ADJOINT EXTENSIONS OF p4 77
In order t h a t these solutions belong to £ 2 ( 0 , o o ) we need that 9i(a±) < 0 . So

we find that acceptable values of a ± are
i'7r/8+i7r/2 !7r/8+i7r
(6.14.94)
and
—in/8+in —in/8+i3ir/2
Ct — — 6 J c (6.14.95)
So, the deficiency indices are (2, 2) and we have a 2 x 2 = 4 parameter family
of self-adjoint extensions.
To find the self-adjoint extensions we consider the matrix element of p4 and
integrate by parts several times. Thus,
oo d4ip ,, d3ip loo
d
d<p* dd3rprp ,
[°° d<f>*
—-4 ax =
T dx * dx3 J0 3
oo
0 •°° d2r d2i/>
00
xpw ipW + dx
0 0 dx2 dx2
[°° d3<j>* dip
v>(3) v<2> 00 +
0 0
, V
dx
JQ dx3 dx
0 0
v>(3) 00 v-<2> 00 + .*<»>• v r
0 0
I J
d
+ 'v 1 (6.14.96)
Here we have used the notation

d
= l± .
dxn
Since both ip and <f> as well as their derivatives have to vanish at x = oo we are
left with the following condition for self-adjointness of p4.
<l>*(0)ip(3\0)—<f>^*(0)ip(2\0)+<p^*(O)ip^(O)—^3^*(0)ip(0) = 0 .(6.14.97)
To satisfy this condition we insist that all ip G £2(0,00) should be of the form
ip{3)(0) = aV>(2)(0) + bip{1){0) + cip(0) . (6.14.98)
Then, we also have t h a t
^3),(0) = oV(2),(0) + r f ) * ( 0 ) + c^(0) . (6.14.99)
Now writing out (6.14.97) we find
<£(2)*(0) [ ^ ' ( O j - a X O ) ] -</> (1) *(0) [V>(2)(0) + 6*V(0)]
+ [ a ^ ( 2 ) ( 0 ) + 6V' (1) (0) + ( c - c * ) V ' ( 0 ) ] = 0 . (6.14.100)

1 v,n o. íviAi jtimiviAi 1 (JAL FOUNDATIONS
If we now cause all three of the square brackets to vanish separately we have
that
^x)(0) = a v ( 0)
V)(2)(0) = -6*V( 0)
-ab* +a*b + c-c* = 0. (6.14.101)
This last equation is easily satified by choosing
c = ab* . (6.14.102)
Writing,
a = \a\eip , b=\b\eil (6.14.103)
we have that
c=\ab\ei^~l) . (6.14.104)
Thus, finally the conditions for self adjointness are
v>(1)(0) = |a| e l / 3 v(0)

^(2)(0) = —16| e _ , " i v(0)
^(0) = -\ab\e-^-^(0) . (6.14.105)
It is also clear t h a t the four real parameters |a|, |6|, /?, and 7 label the self-adjoint
extensions.
Bibliography
[6.1] P. Carruthers and M.M. Nieto - Phase and Angle Variables in Quantum
Mechanics, Rev. Mod. Phys. 40, 411, (1968) .
entific Publishing Co. Pte. Ltd., chapter 6, (2002) .

Chapter 7
Physical Interpretation
7.1 Tetrahedral Die

An unbiased tetrahedral die is thrown and shows the number n. If this die is
thrown very many times calculate (n) as well as A n .
Solution
Since each side of the die is equally probable and the sides are numbered from
1 to 4 we get
( n ) — 4 ( 1 + 2 + 3 + 4) — - (7.1.1)
Similarly,
An = ^ [(l-5/2)2 + ( 2 - 5 / 2 ) 2 + (3-5/2)2 +(4-5/2)5

5
(7.1.2)
4 '
Notice t h a t we could also calculate A n from
An = (n2)-((n»2
\ [ l 2 + 22 + 32 + 4 2 ] -
(7.1.3)
7.2 Probabilities, E x p e c t a t i o n Values, Evolution

A particle is in a state given at t = 0 by:
, 1 . i\J2 y/E
tp = -U 0 (x) + — Ui(x) - — U 2 (x)
80 CHAPTER 7. PHYSICAL INTERPRETATION
where uo , « i , «2 are simple harmonic oscillator eigenfunctions corresponding

to the energies 1/2 hu , 3 / 2 ñu; and 5 / 2 fko , respectively.
a) W h a t is the most likely value of the energy that will be found in a single
observation on this system? W h a t is the probability of finding this value?
b) W h a t is the average of the energy that would be obtained if the experiment
in part a) could be repeated many times? W h a t is the probability of getting
this value?
c) A measurement of the energy yields a value 3 / 2 hui . The measurement is
immediately repeated. W h a t is the resultant value of the energy? W h a t is the
wavefunction immediately after the second measurement?
d) What is the wavefunction of the undisturbed system after a time t has
elapsed?
Solution
Since uo, Mi, «2 are normalized and the moduli of their coefficients add up to
unity, the whole wavefunction is normalized.
a) The most likely value of the energy corresponds to the eigenfunction with
the largest coefficient ie. «2- This value is 5 / 2 huj with a probability of
I — a/6/31 2 = 2 / 3 .
b) The average energy is
1 lt 2 3. 6 5, Q7
(E) = x x
= ' <7'2-4)
9 * 2 "*"9 2 9 2
The probability of observing this value is zero since it is not one of the eigen-
values.
c) The resultant energy value obtained is 3/2 hui. The wavefunction is
«i(x) exp(—3¿/2 oj¿) .
d) The wavefunction is
* = iuoOr) e~iut'2 + (*) e ' ^ ' 2

- ^ u 2 ( z ) e~i5^2 . (7.2.5)
o o o
7.3 ( L x ) and (Ly) in an Eigenstate of Lz

Show that in an eigenstate of Lz the expectation values ( L x ) and (Ly) vanish.
Hint: Use the commutation relations for the angular m o m e n t u m operators.
Solution
The commutation relations for the angular momentum operators are
[Ly, Lz] = ihLx . (7.3.6)
[Lz , Lx] = ihLy . (7.3.7)
FREE PARTICLE PROPAGATOR 81
7-4-
If we assume that the system is in an eigenstate |ra) of Lz such t h a t
lz\m) = mh\m) (7.3.8)
then taking the expectation value of (7.3.7) we get
ih(m\Lx\m) = (m\[Ly, Lz]\m)

= mh(m\Ly\m) — mh(m\Ly\m)
= 0 . (7.3.9)
In exactly the same manner, taking the expectation value of (7.3.7) we find that
(m\Ly\m) = 0 . (7.3.10)
7.4 Free Particle Propagator

A free particle is located at x = a at t = 0 ; i.e. its wavefunction at t = 0 is
given by
$ (0, x) — S(x — a) .
Find the wavefunction for t > 0. This solution is called the free particle propa-
gator.
Hint: To evaluate an integral of the form
f
J —C
exp{j'Ax 2 + i/3x} dx
pretend that A is A + ie with e > 0 so that the integral is convergent. Then

complete the square in the exponent of the exponential and change variables.
The ie in your answer will allow you to decide whether to take the positive or
negative square root. Finally let e —> 0.
Solution
The eigenfunctions for a free particle are
k x
uk(x) = — e ' (7.4.11)
v27t
with energy
h?k 2
• p-4-12)
The initial condition is
«(O.ar) = S(x-a) = J A(k)uk{x) dk = J A(k) eikx dk . (7.4.13)

^fiAriun / . rnrsKJAL i i \ 1 ERFKtilA TI ON
Therefore,
1 i°° — ika
dx = (7.4.14)
A { k ) = 8 [ x
^ L ~a) e
~,kx ~
Then, we find that
/ °°
Wfe(x) e-*E>'tlh dk (7.4.15)

•OO
1 /*°° 2
= _ / e>'fc(i-a) e-i7ifc í/2m ^
2?r 7 _ o o
This integral is not obviously convergent. To remedy this we replace the co-
efficient —iht/2m of A:2 in the exponential by —iht/2m — f ,e > 0 , with the
intention of letting e go to zero after the computation is done. This process is
called, "regularizaron". Thus,
i r°°
w(t,x) = — eik(x-a) e-i(htl2m-U)k* dk
2?R J — OQ exp
2ix y e + iht/2m — a)2
,m(x 2ht
1—i ,m(x — a) 2
exp 1 (7.4.16)
~2^ 2ht
It is in the last step t h a t we used the fact t h a t t —> 0 + to obtain the correct
square root .
7.5 Minimum Uncertainty Wavefunction

We have seen that
Ax Ap > h/2 .
Assume
(x) = {p) = 0 .
Now use the Schwarz inequality
II/II2II<HI2>I(/,<7)I 2
and put
f = xip , g = pip .
Show that the equality in the uncertainty will hold only if

p = A xip
with A a constant and
(ip, (xp + px)ip) = 0 .
Hence derive an equation for rl> and solve it explicitly.
7.5. MINIMUM UNCERTAINTY WAVEFUNUTIUN
Solution
We s t a r t with f ( x ) = xip(x) , g(x) = pip(x). For t h e Schwarz inequality to
become an equality requires t h a t
g = Xf (7.5.17)
where A is a c o n s t a n t . T h i s m e a n s t h a t
piP{x) = XxiP(x) . (7.5.18)
Furthermore,
I(ip, [x,p]i/>)\ = \î^(xip,pi>)l < 2\(xip,pip)\ . (7.5.19)
Therefore in order to get equality we f u r t h e r require t h a t
$t(xip,pip) = 0 . (7.5.20)
This m e a n s
(xip,pip) + {pip, xip) = 0 (7.5.21)
or
(ip, xpip) + (ip, pxip) = (ip, (xp + px)ip) = 0 . (7.5.22)
Now,
(ip, xpip) = X(xip, xip) , (ip,pxtp) — X*(xip, xip) . (7.5.23)
Adding we get
(ip, xpip) + (ip,pxip) — (A + A * ) ( x i p , xip) . (7.5.24)
This m e a n s t h a t A h a s t o be p u r e imaginary so t h a t A = ia with a real. T h u s ,

we find
pip(x) = iaxip(x) . (7.5.25)
Therefore,
h dip
— — = iaxip(x) . (7.5.26)
idx
So finally,
2
ip(x) = Aexp ( )l (7.5.27)
2h
and for this t o be a square-integrable f u n c t i o n f u r t h e r requires t h a t a > 0.

J.XJIXV #. x 111 o 11^sajú iiv i ft t i t * 1 A'l'lO.N
7.6 Spreading of a Wave Packet

A free particle of mass m is at t = 0 in a state described by
$ ( 0 , x) — [27tL 2 ] 1/i4
exp — (x/2L)2
W h a t is the wavefunction for an arbitrary time t > 0? C o m p u t e the uncertain-

ties Ax and Ap as functions of time. This illustrates the "spreading" of a wave
packet.
Solution
We first express ^ ( 0 , x) as a superposition of free particle states. Since the most
general solution for a free particle is
(7.6.28)
/ OO
, fe:E fifc2f 2m
A(k) e '( - / ) dk
•OO
we write
(7.6.29)
/ OO
A(k) eikx dk .
•OO
Then, the coefficient of A(k) is determined from the explicit expression for
tf(0,x).
i r°°
A(k) = — <I<(0 ,x)e~ikxdx. (7.6.30)
2n
J-OO
Inserting this in the integral and completing the square in the exponent, we can
evaluate A(k).
A(k) = ~(2nL2)1/4e~k'L2 . (7.6.31)
Hence, we find (upon again completing the square in the exponent) t h a t
(87ri2)1/,4exp
4(L2+iht/2m)
*(*,*) =
2n
ix
e x p -(L2 + iht/2m)
C { - 2(L2 + M/2r
(87TZ,2)1/4 exp [ - 4 ( L ! + x ¡ t / 2 r n )
(7.6.32)
27r L2 + iht/2m
This is a Gaussian of the form
$ ( T , X ) = (2KW2)~1^4 AA
exp (7.6.33)
\w2 )
7.7. TIME-DEPENDENT EXPECTATION VALUES 85
where
2
h2t
= ¿ 2 + £ í 27 22 (7-6-34)
Am L
and
"1 f ht \ htx2
A = — arctan 4- (7.6.35)
2 \ 4 mL2 J 8mL2w2
So, we can immediately conclude t h a t
(Ax) 2 = (x2) = 2w 2 . (7.6.36)
Furthermore since this is a Gaussian it is a minimum uncertainty wavefunction.

Therefore,
< a <» 2 = i j s j i - b ? • <7-6 371
7.7 Time-dependent E x p e c t a t i o n Values

A free particle is, at t = 0, in a state described by the wavefunction
. . f Asin 2 —• |x| < a

=| o " ¡J| > 0 •
Find for t > 0 the following expectation values
{p)t > (x)t , {P2)t , (xp + px)t , and (x2)t .
Jt(A)t = ^(V{t,x),[H,A]y{t,x)) .
If you try to evaluate these results using the time-dependent wave-function

ty(t, x) you will get some impossible integrals. [7.1]
Solution
In this case we have a free particle so
(7-7-38>
Also we are using the equation
~(A) = ^ , [ H , A } ^ ) . (7.7.39)
Since
[H,p] = 0 (7.7.40)
we therefore have that
jt(p) = 0 (7.7.41)
so that
(p)t = (P)o = W » / s i n 2 ( 7 r a ; / a ) y ^ (sin 2 (7rx/a)) dx = 0 . (7.7.42)
Next, we find that
= (7.7.43)
so that
0
= <£> = • <7 7 44
»
Therefore,
(x)t = (x)o = \A\2 Í sm4(Tvx/a)x dx = 0 . (7.7.45)

J—a
Similarly, we see that
(P 2 )t = (p 2 )o - h 2 \ A \ 2 J s i n 2 ( 7 T x / a ) - j ^ (sin 2 (7rx/a)) dx
a
2n2
/
sin 2 (7rx/a)—Y {cos2 (nx/a) — sin 2 (7rx/a)) dx . (7.7.46)
-a ^
2 2
Now, we rewrite cos as 1 — sin and use the fact that
\A\2 f sin4(nx/a) dx = 1 . (7.7.47)

J —a
So, we find that
. 2\ 4ft27T2 2h7TT2 2
<A = — 2 IT-«HI 2 - (7-7.48)
Next, we need to evaluate |j4| 2 . This turns out to be 4/(3a). Therefore,
<A = ^ ( 2 - 4 / 3 ) = ^ . (7.7.49)
We now find
4-px]
[H,xp + rurl==-2ih~
—9¿/„ . V2
(7.7.50)
2m
Therefore,
d
\ 2
/ 2x 2(2nh) 12
-{ l p + p , ) = - - & ) = - L _ f . (7.7.51)
7.8. EHRENFEST THEOREM
So, we get
. . 2(27T h)2
(xp + px)t = (xp + px)o + , t
3ma,2
We now compute
(xp + px) o
= \A\2^ J sin2(rex/a) (^¿x-jj-+ 1*J sm2(irx/a) dx =
Therefore,
2(2tt/í) 2
(XP + Px)c — -3^2" ' •
Finally we find
iH
[H, x2] = x[H, x] + [H, x]x = — — (xp + px)
Hence,
d , 2, *' - i h , v 1 2(2tt/í) 2 4
- x ) = (xp + px) t = ^ f-i .
dt nm m 3mcr
Therefore,
/ 2\ _ / _ 2 \ 1 (27rfi)~
(x ) t _ (x >0 + 3 m 2 f l 2 j 2 .
Now,
^>o = |A|2/ x2sin4(?zx/a)dx = ^(tt-15/8).
Thus, we finally have
/ 2v 4 / a\3 ir, , _,„x (2nh)2 ,

<® >t = - 5 - ( - ) t t - 1 5 / 8 h - w r 4 •
3a \7T/ 4 ?>m¿a¿
7.8 Ehrenfest T h e o r e m
Consider a particle under the influence of a Hamiltonian
H = ^ + V(x)
so that
Show that if (x) is the "centre of mass" of the wave packet and (p) the average
momentum of the particle then
d(x) _ (p)
dt m
and
m = <*•(«)> = _ ) .
These are known as the Ehrenfest equations. To be equivalent to Newton's

equations requires that
(F(x)) = F((x)) .
Discuss under what circumstances this condition is approximately valid.
Solution
We have
n2
H = £ - + V{x). (7.8.60)
¿m
Therefore,
[H,x] = -ih— (7.8.61)

m
and
[H,p] = i h ~ . (7.8.62)
From this we find that
= !([«,«]) = * H » , ( £ ) = M (7 . 8 . 63)
and
Jt(p) = = - ( % ) = <FM> • (7.8.64)
For these results to be equivalent to the classical Newton's equations requires

that
(F{x)) = F((x)) . (7.8.65)
To see what this means, we Taylor expand F(x) about (x).
Fw = f (m) + £ - (7.8.66)
Clearly we require that

< ( z - 0 c » " } « 0 n > I . (7.8.67)
This is exactly true if the potential is a quadratic function of x. It is also

approximately true if the dispersion in x, that is Ax is not too large. This
means that the wavefunction , P(t,x) must be fairly smooth and narrow.
7.9. COMPATIBILITY THEOREM 89
7.9 Compatibility T h e o r e m
Prove the compatibility theorem which states:
The following statements are equivalent for any pair of observables A, B.
i) A and B are compatible.
ii) The operators A and B, representing the observables A and B, possess a
common eigenbasis.
iii) A and B commute, that is
[A, B] — AB - BA = 0 .
Solution
The compatibility theorem states that, given two observables A , B represented
respectively by the self-adjoint operators A , B the following statements are
equivalent.
i) A and B are compatible observables.
ii) A and B possess a common eigenbasis.
iii) [A, B]= 0 .
We begin by recalling that two operators A , B are compatible if a measurement
of A followed immediately by a measurement of B and another measurement of
A yields the same result for the value of A in both cases. To prove the theorem
we now show t h a t
i) ->• ii) - » iii) - y i).
i) - » ii)
Suppose the measurement of A yields a . Then, immediately after the measure-
ment the state of the system has to be an eigenstate of A say <j)a. To guarantee
that after the measurement of B when we again measure A we again get the
value q requires that after the measurement of B the state be 4>a¡p where
B<t>a,p = (7.9.68)
and
â,/3 = • (7.9.69)
This shows that A and B have a common eigenbasis.

ii) —» iii)
Since, A and B have a common eigenbasis we have t h a t operating on any com-
mon eigenstate
AB<t>a,p = PA^p = pa^p = BA<¡>a i)
Since [A, B] = 0, the two operators may be diagonalized simultaneously. Then,
a measurement of A followed by a measurement of B will yield a state <j>a>p as

above. Thus, if a measurement of A is now performed, we are guaranteed to
get the result a . This proves the desired result.
7.10 Constant of t h e Motion

Show that if [H, A] = 0 then AA does not change in time.
Solution
Since,
[H,A] = 0 (7.10.72)
it follows that
[H, A2] = A[H, A] + [H, A]A = 0 (7.10.73)
Hence, we can conclude t h a t
5<A> = ° (7.10.74)
as well as t h a t
(7.10.75)
This immediately implies t h a t
±(AA)>=±{{A-{A))2) = 0 (7.10.76)
7.11 Spreading of a Gaussian Wavepacket

Consider a free electron described at time t = 0 by the Gaussian wavepacket
^ ( O , ^ ) = [27tL2] 1/i4
exp — (x/2L)2 .
Using the experience from problem 7.6 answer the following questions:
a) W h a t is (x) at any time?
b) W h a t is Ax as a function of time?
This is known as the spreading of a wave packet. If this packet corresponds to
an electron with 20 eV energy and a width of 1.0 Á at t = 0, what is its width
after travelling 100 m?
7.11. SPREADING OF A GAUSSIAN WAVEPACKET 91
Solution
We are given as initial condition
\¡r(0,a;) = (2xL2)~1^4 exp + ikx (7.11.77)

4 L2
Now following the steps in problem 7.6 we first Fourier transform Vf(0,x).
F(p)= [ V(0,x)e~ipx dx = ^(2nL2)1/4 (7.11.78)

J — OO
and
hp2t/2m)
dp (7.11.79)
/ OO
i{px
F{p) e -
•OO
Thus, carrying out the integral, we see that the wavefunction after a time t is
again a shifted Gaussian of the form
-(k-p)2L2 -x2 + 4ikL2x + 4k2 L2
2\-i/4 2 exp . (7.11.80)
«(*,«) (27tL , [i + iht/2mL ] 2
4 (L + iht/2m)
From this we compute that
.x . hkt
( ) = t;— (7.11.81)
2m
and
h2t2 \
(Ax) 2 = (x2) - (x)2 = L2 I 1 + (7.11.82)
4 in2L4 •
10
For the problem under consideration we have L = 1.0 Á = 10 m . The
speed of the electron is given by
E = -mv~ (7.11.83)
2
Substituting 20 eV for the energy we find that v = 8.4 x 10® m / s . Therefore
the time to travel 100 m is
t = 1.2 x 10~ 5 s . (7.11.84)

Hence, we get t h a t
ht
Ax 69 m (7.11.85)
2mL
This means t h a t in going down a 100 m beam tube the wavepacket has spread to
macroscopic dimensions. The reason for this is that originally it was extremely
narrow and therefore had a very large spread in momentum. If instead the
wavepacket originally had a width of 1.0 m m the width after travelling down
the tube would have increased by a negligible 6.9 x 1 0 - 3 m m .
7.12 Incorrect T i m e Operator

Consider the space of functions of E belonging to £2(0,00). The relationship
[E,t] = ih
can be represented on this space by
Ef(E) = Ef(E)
and
Show that the operator t so defined cannot be an observable, i.e. that it has
no self-adjoint extensions. This proves that if the energy has a lower bound, a
relationship such as
[E, t] = ih .
cannot hold if time is to be an observable. [7.2]
Solution
We first find the deficiency indices for the putative time operator —ihd/dE.
Thus, we look at
- i h ^ f ( E ) = ±iTf(E) . (7.12.86)
The solutions are
f+(E) = e~ET'h , /_(E) = eET'h . (7.12.87)
Only the first of these is square integrable on the interval (0,oo). Thus, the
deficiency indices are (0,1) and this operator has no self adjoint extensions.
Since for any physical system the energy must have a lower bound we can
always shift the minimum of energy to E — 0. This allows us to consider the
space £ 2 ( 0 , 0 0 ) as the appropriate space for this problem. Our result means
that the relationship
[E,t] = ih
is false.
7.13 Probability to Find a Particle

A particle is in a state described by the unnormalized wavefunction:
f{x) = Ae-'W a> 0.
Find the length of an interval around the origin such t h a t the probability of
finding the particle in this interval is 40 % .
7.14. SPHERE BOUNCING ON SPHERE 93
Solution
We have the unnormalized wavefunction
f ( x ) = Ae~aW a > 0. (7.13.88)
To normalize it we require that
oo poo
/ 2 2
\f(x)\ dx = 2\A\ e~2axdx= 1 (7.13.89)
or -oo J0
A = V¿ . (7.13.90)
Now, the probability that the corresponding particle is found between — L <
x < L is given by
f \f(x)\2dx = 2\A\2 [ e~2axdx = 1 - e~2aL . (7.13.91)

J-L JO
We want this probability to be 40%. Thus, we get
1 _ e - 2 « ¿ _ o.4 . (7.13.92)
Hence, we find that
1 = - ® ! ! . (7.13.93)
2a
7.14 Sphere Bouncing on Sphere

A 1.0 g sphere of radius r = 1 cm is dropped from a height of I = 1.2 m
(separation of the two centres) onto a similar ball rigidly attached to the ground.
If the collisions are perfectly elastic, find the number of times the ball could be
expected to bounce if the only limitation in the precision of releasing the ball
to fall directly on top of the lower ball is due to the uncertainty principle.
Solution
Assume the ball falls in the y-direction and is deflected in the x-direction. The
ball actually falls a distance of 1 m. Its terminal momentum just before contact
is
3
Py = m\/2g(l — 2r) « 10 x \/2 x 10 x 1 kg m / s
-3
= 4.47 x 1 0 kg m / s . (7.14.94)
Suppose t h a t on contact the vertical makes an angle Atp with the line connecting
the two centres (see figure 7.1). Then, on bouncing, the angle increases to 2A<p.
After n bounces the angle will be nAip.
CHAPTER 7. PHYSICAL INTERPRETATION
Figure 7.1: Sphere bouncing on a sphere. The angle <p is the same as the angle
A f in the text.
Now, if the deflection is Aip, then the momentum in the x-direction after
the collision is
Apx « PyAlf (7.14.95)
since Aip is a very small angle. The angular momentum of the upper ball with
respect to the point from which it was dropped is
A L =px(l- 2 r ) « (/ - 2r)pyAip . (7.14.96)
On the other hand, for very small angles we have the uncertainty relation
ALAip > h/2 . (7.14.97)
So, if this uncertainty relation is the limiting factor we get
ALAip ra h/2 w (/ — 2r)m\j2g(l — 2r)(A<^) 2 • (7.14.98)
Therefore,
h
(A r ) 2 « 1.1 x 1(T 3 2 . (7.14.99)
2(1 — 2r)m\/2g(l — 2 r)
So,
Aip « l x 10~ 16 (7.14.100)

2
The last bounce will occur when the angle is approximatly <p ra r/l = 10
radians. Thus, the number of bounces is given by
<p = nAip (7.14.101)

or
n = JL « « 1014 (7.14.102)
aip 10-16
7.15. CLOUD CHAMBER TRACKS 95
7.15 Cloud Chamber Tracks

The tracks made by an electron in a cloud chamber consist of small droplets of
about 1 fJ. diameter. Use this fact to show t h a t the track made by a 1.0 keV
electron is essentially indistinguishable from the classical trajectory.
Solution
The momentum of the electron in the direction of the track (say x) is given by
px = V2mE
= y/2 x 9.1 X 10-31 X 1.6 X 10-19 X 1.0 x 10 3
-23
= 1.71 x 1 0 kg m / s . (7.15.103)
This is also the momentum for the classical trajectory. Quantum mechanically
there is also a transverse component (say py) of the momentum given by the
Heisenberg uncertainty principle.
ApyAy « h . (7.15.104)
Here, Ay is given by the size of the droplets. Thus,
=
Apy ~ -7— = ^ ^ ^ ^ 1-05 x 1 0 - 2 8 kg m / s . (7.15.105)
Ay 10-6
This means t h a t the fractional change in the momentum is of the order of
^ « 6 x 10" 6 (7.15.106)
Px
and hence, the trajectory is essentially the same as the classical.
7.16 Spin 1 Measurement in T w o Directions

A particle has spin h. A measurement of the spin of this particle along a given
direction yields the largest possible value. Next, a measurement of the spin is
made along a new direction making an angle 6 with the original direction.
a) W h a t are the possible results of this measurement?
b) W h a t are the probabilities for the results obtained in part a)?
Solution
a) Since the particle has spin h the outcome for the measurement of spin along
any direction must be one of h, 0, or — h.
b) Since we obtain the largest possible value, in the initial measurement, we
obtain h. To compute the probabilities for the new measurement we call the di-
rection of the first measurement the z-direction. In that case the corresponding
eigenfunction is
ip = | 0 (7.16.107)
0
We let the new direction lie in the x-z plane. In that case the appropriate
angular momentum operator for this measurement is
( cos 0 0
72sin
Ln = Lz cos 9 + Lx sin 9 = h 0 (7.16.108)
72sin( 72sinf
0 ^ sin 9 — cos 9
We now compute the eigenfunctions of Ln for the eigenvalues h, 0, and —h

respectively. In every case we assume the eigenfunction is of the form
ip (7.16.109)
The equations for the coefficients a, 6, c then read

h:
a cos 9 -\—~ sin 9 b = a

V2
—= sin 9b — cos 9 c = c . (7.16.110)
v/2
Thus, after normalization, we get
| ( 1 + cos 9)
= ^ sin 6 (7.16.111)
| ( 1 — cos 9)
0 h:
a cos 9 + —= sin 9 b = 0
v2
—= sin 9 b — cos 9 c = 0 . (7.16.112)
y/2
Hence, after normalization, we get

1
v2sml
ipo = cos 9 (7.16.113)
sin0
72
7.17. PARTICLE IN A BOX: PROBABILITIES AND EVOLUTION 97
-h:
a cos 9 H—\= sin 9 b = -a

v2
— s i n 9 b — cos9 c — —c (7.16.114)
V2
and again, after normalization, we get
—
5(1 ~ c o s ^ ) \
/
V-i = "^sin0 . (7.16.115)
\ - | ( 1 + cos9) /
The desired probabilities are now given by
P . = IW, r»,)\ 2 (7.16.116)
and are
pi = i ( l + cos0) 2
4
P0 = ^ sin 2 9
P—\ = ^(l-cos0)2 . (7.16.117)
7.17 Particle in a Box: Probabilities

and Evolution
The wavefunction for a particle in a box 0 < x < L is given at t = 0 by
*(x,0)=(\ #0 ~7L °otherwisi

0
P"
- x
- L
0 otherwise
a) Find the probability P„(0) that at t = 0 a measurement of the energy yields
the value
h2n2 7T2
= 2
" 2m i '
b) Find the corresponding probability Pn(t) for a time t.
Solution
The eigenvalues and eigenfunctions for a particle in a box of length L are
h2n7 7T2
and
un = \ j — sm(nnx/L) . (7.17.119)
v jl/
The probability t h a t at t = 0 a measurement of the energy yields the value

•,2—2
K2'
En =
2mL2
is given by
r
Pn( 0 ) ik,$(x,o))|2
_2_
sm(mrx/L) dx
L2
Jo
r [l — cos(n7r)]¿
n
T&r — 1,3,5,...
(7.17.120)
0 n = 2,4,6,...
b) To find the probability at time t we can proceed in two different ways.

1) We can find the wavefunctions at time t and use this. Thus, each eigenstate
evolves according to
«„( x)e~iB*t'h
and the total state evolves according to
V{x,t) = J2C' w/(z)e~ii?'i/ft (7.17.121)
where, as just calculated,
2V2
CL = ($(x,0),M¡(x)) 1 = 1,3,5,.. (7.17.122)
In
Then,
Pn{t) = |(w( e~iE,t!h, $(x, <))|2

2
y t ci(un,ui)
1=1,3,5,...
2
Y] c;¿ní
/ = 1,3,5,...
n = 1,3,5,...
(7.17.123)
0 n = 2,4,6,...
as before.
2) A simper method is to realize that the evolution is unitary. This means that
both the eigenstates un and the total state $ evolve with the unitary operator
U(t) = exp {—iHt/h} (7.17.124)

7.18. FREE WAVE EQUATION: TRANSLATION INVARIANCE 99
where H is the total Hamiltonian for this system. Then,
Pn(t) = |(l/(í)^(x,0),í/(í)«„(x))|2
= |K,^(x,0))|2 (7.17.125)
which is the same result as for t = 0 .
7.18 Free Wave Equation: Translation Invari-

ance
a) Show t h a t the Schrodinger equation for a free particle
= (7.18.126)
at 2m
is form invariant under a translation
x —»• x' = x — vt
y -»• y' = y
z z'= 2 . (7.18.127)
That is, show that if the wavefunction in the new coordinate system is Í '(x1, y', z', t)
then it satisfies the Schrodinger equation
DV H2
i h ^ = ~ — W . (7.18.128)
dt 2m
Hint: The wavefunction transforms as a scalar with an additional phase trans-
formation.
b) Apply this result to a plane wave and discuss the results so obtained.
Solution
a) Using the hint we write the wavefunction in the new coordinate system as
^ ' ( x ' , y'> z ' , t ) and relate it to the wavefunction in the old coordinate system by
an additional phase factor. Thus,
* ( x , t ) = e-,'/(*'-t>tf'(x'It) . (7.18.129)
This may be rewritten to read
9 ( x ' + vt,y,z,t) = e~if<'x'-y-z-t^'(x',t) . (7.18.130)
Since the transformation is only on the x coordinate, we need not consider the
y and z coordinates and so we choose / independent of y and z. Then,
dV dx' d^_.fdj_d¿_ df\ ¡
e~if
dt dx' dt dt \c?x' dt ^ dt )
dv dv .( df df\ /
e~i} (7.18.131)
-vd^ +
- d T - l
[ - v
d ^ +
m r
Also,
DV_ d*' . df ;
= e -if j——w (7.18.132)
dx dx' dx'
and
D2V D2V •D2F , , D
F .2 df d*'
= e~i} 2 2 + ( ^ 7 ) 2
p ' " ^ (7.18.133)
2
dx dx' dx'
Substituting all this into the Schródinger equation we get

DV DV
ih + l 1
- V
D ^ - m ~ dx' dt
•d2f . ,df, . DF DV
2l (7.18.134)
2m dx' 2
dx'2 + (»..»)
dx' )* dx'dx'
Therefore, to recover the Schródinger equation in the new coordinate system we

need
.ihv = 2 (7.18.135)
2m dx'
DF D F
J J. \ y (7.18.136)
2m dx'> dx'2
N V +
[ - D * M ) =
From (7.18.135) we get t h a t

df mv (7.18.137)
dx' h
so that
,, , , mv , , ,
f(x ,<) = ——x -(- a(t) (7.18.138)
where a(t) is an arbitrary function of t. After substituting this result into

(7.18.136) we get
mv da mv2
(7.18.139)
dt 2h
Thus,
,,, mir
a(t) = 1 (7.18.140)
V;
2h
where we have set the arbitrary constant of integration equal to zero since it
would only yield an irrelevant constant phase factor. So, finally
/(*',!) = - ^ ( * ' + trt/2) (7.18.141)
The wavefunction in the primed coordinate system is therefore given by

y(x',y',z',t) = + vt,y,z,t) (7.18.142)
7.19. FREE WAVE EQUATION: ACCELERATED FRAME 101
and satisfies the Schrodinger equation
in w(*'<*) = _|!_v'2$'(£',<) . (7.18.143)

dt 2m
b) For a plane wave we have
«(aií) = e"<[ü't"í^ • (7.18.144)
According to (7.18.143) the wavefunction in the primed coordinate system is
$ ' ( í ' , < ) = E-'ÑR&'+VT/2)
mv2\ /, mv\ . ,
: exp —I w+ I \k* +
~~fr~)x - kyy- kzz (7.18.145)
This means that in the moving frame (primed system) the frequency and wavenum-
bers are given by
, mv 2
UJ = UI
2n
, < i
i +
x
~ x
=
BY
k'z = kz . (7.18.146)
If we express these results in terms of the energy E = fiuj and momentum p = hk

we get
„ 1
E' = E + 2mz;
Px = px + mv
Py = Py
P'z = Pz • (7.18.147)
These are exactly the same results that would be obtained classically.
7.19 Free Wave Equation: Accelerated Frame

Show that if in the Schrodinger equation for a free particle
d V ft2
7 1 9 I 4 8
< >
we go to a uniformly accelerated frame
x -» x' = x + -at2 (7.19.149)
then the wavefunction ^'(a:', t) in the new coordinate system satisfies a Schrodingei
equation of the form
d^'(x' t) h2 d2y/'(x' t)
where
V(x') (7.19.151)
is the potential for a constant force t h a t produces a uniform acceleration a.
Hint: In transforming the wavefunction to the accelerated frame allow for a
phase factor so t h a t wavefunction in the stationary frame 9(x,t) and the wave-
function in the accelerated frame are related by
9(x,t) = e-iflx''t)*{x',t) (7.19.152)
See also problem 7.18.
Solution
We begin with equation (7.19.152) and rewrite it using (7.19.149) as
< £ ( x ' - 1/2 at , t) —

= pe-¡J(x',t) yf.ll
V (x/ ,t) . (7.19.153)
Then,
09 d9' d9' dx'
P-'F + ]$/
dt dt dx' dt dx' dt dt
p-ij d9' d9'
+ at (7.19.154)
~dT dV
Also,
09_
= e~iJ 99' .df
I—— (7.19.155)
dx dx' dx'
and
d29 = e~iJ d29' df d9' f.d2f .df
- 2 i (7.19.156)
dx2 dx'2 dx' dx' dx'2 dx
Substituting this into the Schrodinger equation we get
d9' d9'
ih + at — i at 9'
dt dx' dx' dt
d29' . df 09' ,d2f
'' vvdx' / (7.19.157)
2m dx'2 dx' dx' dx'2
Therefore, to recover a Schrodinger equation, with the possibility of a potential,
in the new coordinate system we need
ihat = 2 i—^L (7.19.158)

2m dx'
as well as
. d2 f
2m L
D ^
+ (ÉL)'
\dx' J
—h A T
df
D * +
Of
M
= V(x'). (7.19.159)
20. THE WIGNER FUNCTION 103
From (7.19.158) we get
= ?£-*' +Hi) (7.19.160)
where b(t) is an arbitrary function of t. Substituting this result into (7.19.159)

we find
ma2t2 max' db
_ ^ = V(x')/h . (7.19.161)
2h h dt
Therefore we must take
V(x') = —max' (7.19.162)
and
db ma2t2
(7.19.163)
dt ~ 2/t~
So,
b{t) = ma2t3
- ^ T (7.19.164)
where we have dropped an irrelevant constant of integration. Hence,

mat f , 1 2
f ( x ' , t ) = ^ ( x ' - ^ a t j . (7.19.165)
Furthermore the wavefunction \ í ' ( x ' , t ) satisfies the Schrodinger equation
+ ( 7 ,9,66)
where, as stated,
V(x') = —max' (7.19.167)
is the potential due to a constant force t h a t produces a uniform acceleration a.
7.20 T h e Wigner Function

Show that if for a given wavefunction ^(x, t) one defines the "Wigner function"
[7.3]
W(x,p,t)=-j-- J d3y*(x + y/2,t)**(x-y/2,t)e-i™"i
then it satisfies the following properties.

a)
QX / H
W{X, P, t ) = ~ J d3q * ( p + q/2, t ) * * ( p - 9/2, t)
b)
J d3pW(x,p,t) = \V{x,t)\2 .
c
)
J d3x W(x, p, t) = |V(p, t)|2 .
Hint: These properties depend only on the properties of Fourier integrals.
Solution
We begin by writing out the Fourier transforms.
+ y/2, t) = -L J d3g e'i-P+f/a)/» 9{q, t)
9*(x-y/2,t) = L Jd3ke-ik (t-y/2)/*y*(k; t) . (7.20.168)
Therefore,
W(x, p, t)
= ~ J d3yd3q d3k e-'P-y/" ei^+y/2)/K e~ik\x-y ¡2) / h t)
= J d3qd3kS{p-q/2 -k/2)eiU-*Wl*{q, <)**(ik, t)
= ~Jd3ke2i{P-h'>£/n9(2P-k,t)9*(k,t). (7.20.169)
We now let
p— k = q/2
so t h a t
k=p — q/2 and -d3q = d3k .

8
Then,
W(x, p,t)= -L J d3q9(p + q/2, t)9*(p-q/2, t)e^s/n . (7.20.170)
b) In this case we simply integrate W over p and get
J d3p W(x, p, t) = ~ J d3yd3p9{x + y/2,t)9*(x-y/2,t)e-iP s/n
(2nh)3 [ ,
= - J ^ - J d3y9(x + y/2, t)9*(x - y/2, t)S(y)
2
= 9(x, t)9*(x, t) = | $ ( x , i)| . (7.20.171)
7.21. PROPERTIES OF THE WIGNER FUNCTION 105
This result shows that
J d3p W(x, p, t)
gives the probability density for finding a particle at the point x at the time t.
c) Here, just as in part b) above we simply integrate over x. Thus,
Jd3xW(x,p,t) = J d3q d3x^(p + q/2, t)^* (p-q/2, t) e'^ x/n
= ¿ / d3qy(p + ? / 2 , W*(p- 9/2,
= ^ « ( p , *)**(£ t) = ¿ ^ ( p , <)|2 • (7.20.172)
now gives the probability density for finding a particle with momentum p at the
time t.
7.21 Properties of t h e Wigner Function

Show that although the Wigner function (see the previous problem) seems to
behave like a classical probability density for x and p it cannot be a true prob-
ability distribution since it is not positive definite.
Hint: Try the wavefunction
* ( « ) = Axe~r2'2a2
where A is a normalization constant and r 2 = x • x.
Solution
Using the suggested wavefunction we have
W{x,p,t)
= J d3pe,P ^fi(a; + í/2)e-(I+í''2)2/2a2

2
x e-(y+l/2)
2
/2a2 e -( 2 + C /2) /2a 2
x (¡B _ £ / 2 ) e - ( * - € / 2 ) a / 2 a > e - ( y - , / 2 ) 2 / 2 a 2 e -(z-(/2f/2a\ (7.21.173)
Here,
P= (£> V, 0 •
After some simplification we find
W{x, p, t)
2 2
= M ! J dtdT)dCeiP P/h (x2 - í 2 /4) e " 2 ^+i /4)/2a
2 2
e-2(yw/4)/2a e _2(, +c 2 /4)/2 a 2
= M! J die-*'*"1 (x*-e/4)e-**'a3e-?'4a3
2 2 2
x J dr, e-ip^'h e -»
a
/« a e -i a /4« a J d( e"- < p « c / f i e - ^ / « e - c / 4 a
= m ! e - r 2 / a 2 e - p 2 a 2 / a 2 ^3/2^5 + 2a^£ _ ^ (7.21.174)
So, for
2z 2 2a 2 p 2 ,
1 — < 1
a2 2
/i
we get negative values. This means that, in this case, W(x, p, <) cannot be a
true probability density. Notice the region for which we get negative values is
the interior of an ellipse in the x — px plane with semi-major and semi-minor
axes a/\/2 and fi/(\/2a). Inside this region we would have
xpx < a/\J2 x h/(V2a) = h/2 .
7.22 Uncertainty Relation and Wigner Function

Show that if 9 ( x , t) is a properly normalized wavefunction then
\W(x,p,t)\2
< 2nJ 6d3y\9(x + y/2, t)\2 J d3z |tf(f - z/2, t)\2
(7.22.175)
This means that
m
u 2N 3
and states that a particle can not be localized in a cell of phase space smaller
than (A/2) 3 .
Hint: Use the Schwarz inequality.
BIBLIOGRAPHY 107
Solution
The Schwarz inequality states that for any two square integrable functions /
and g we have
|(/, 9)I2 < ( / , f)(9, 9) (7.22.176)
where
( / , 9) = J f'(x)g(x)d3x .
We now choose
h
FW = ^{s-Y/2,T)E^yl
9(y) = + $/*>*) • (7.22.177)
Then,
(/,<?) = +
= W(x,p,t) . (7.22.178)
So,
( / . g)(g, f ) = \w(x, p, *)|2
< J d3y\v(x + y/2, t)\2 j d3z\V*(x - z/2, t)\2
26
J d3(y/2)\V(x + y/2,t)f J d3(z/2)\V* (x - z/2, <)|2
he _
6
|J . (7.22.179)
Bibliography
[7.2] ibid section 7.10 .
[7.3]E.P. Wigner, Phys. Rev., 40, 749, (1932).

Chapter 8
Distributions and Fourier

Transforms
8.1 Properties of the D e l t a Function

Use the representation
¿(z) = lim hc(x)
e—v0+
where
hn (x) ~ f & 'f < 6

^x>-\ 0 if |x| > e
to verify the following equations
a)
/ OO
f(x)S(x-a)dx — f(a) (8.1.1)
b) -OO
S(ax) dx =-S(x) (8.1.2)
c)
f(x)S(x — a) = f(a)S(x — a) (8.1.3)
d)
OO
/
S(x - y)S(y - z) dy = S(x - z) (8.1.4)
-OO
S(x2 - a 2 ) = 2j~j[¿(* ~ a ) + S ( x + a ) ] • (8.1.5)
Hint: Integrate both sides of the equation with a well-behaved function f(x) .
81. PROPERTIES OF THE DELTA FUNCTION 109
Solution
a) We begin with
(8 L6
m . ) = { * ¡:¡>: »
Therefore,
lim f f(x)h((x-a)dx = limi f ^ "J" ' dx . (8.1.7)

f->0+7-oo 2e
Now, for f(x) sufficiently smooth near x = a we can use a mean value theorem
to write (with - 1 < 9 < 1) the above expression as
f(
lim f >X + a)
dx = lim f(a + 9e) = / ( a ) . (8.1.8)
e-t-0+ J-e 2e f >0+
-
b) In this case we consider
f°° > fe,a f{z/a) dz

lim / f(x)hf:(ax) dx = lim /
2e a
<-*0 + J _ 0 0 eÔ + J _ e / a
= lim 2<-?{it/a)
•>•0-1- a,
= -/(0) . (8.1.9)
a
Here we have changed variables from x to z = ax in going to the second line
and again invoked a mean value theorem to get to the third line where 9 is a
number bewtween 0 and 1. So, as required,
¿(ax) = 1 % ) . (8.1.10)
a
c) We evaluated S(x — a) in part a) so now we simply compute
lim f l^ldx = Jf(a)

y . (8.1.11)
e-+o +J_c 2e '
This equals the result from part a) and therefore we have verified that
/ ( x ) á ( x — a) = f(a)6(x — a) . (8.1.12)
d) In this case we first write out the two functions whose limit we need to take.
he(x~y) = { h x-e<y<x + e
\ 0 otherwise
< z +< 13
*•(»-') = {t — • <«>
110 CHAPTER 8. DISTRIBUTIONS AND FOURIER TRANSFORMS
Therefore,
OO
/
f(x)hei (x - y)he2 (y - z) dx dy
•OO
1 1 /*z+e2 r%+e i
= lim f lim - — - — / dx fix) / dy
ei-*o+ 2->o+ 2ei 2e2 j z _ £ 2 z+e2 jx-ti
1 1 f
= lim lim - — - — 2 e i / dx fix)
E1-+0+ £2-V0+ 2tl 2f2 J2 —c2
= f(z) . (8.1.14)
r
Therefore, we have again verified that
S(x - y)S(y - z) dy = S(x - z) . (8.1.15)

)
J—(
e) In this case we have to write out the limiting function

Va2 — e < x < Va2 + c
¿ -%/a 2 + e < x < —%/a2 — . (8.1.16)
2 2
he(x — a ) = 0 otherwise
Therefore,
/ OO
/ ( x ) / i f ( x 2 — a 2 ) / ( x ) dx
•OO
1 r r-y/a2-e /•V<i2+£
lim — / / ( x ) dx+ fix) d
e-yo+ 2e j-/^l
1 r /~(o-«/2|o|) /-(a+e/2|a|;
= x dx
liW, ^7 / /( ) + f(x) d x
£->0+ |_./_(o+£/2|a|) J(a-e/2|a|)
| [ /-f/2|a| /-i/2|a|
lim — / fix + a)dx + / / ( x - a) dx
e-f0+ L-/_e/2|al J- e /2|al
= ói-7 [ / ( « ) + /(—«)] • (8.1.17)

2|a|
Thus, we have also verified that
<5(x 2 - a2) = r^-r[<S(x - a) + S(x + a)] . (8.1.18)

z\a\
8.2 Representation of D e l t a Function

Repeat the previous problem (problem 8.1) using the representation
1 /"0< ikx ck2

¿(x) - lim — / e ~ dk .
e - + 0 + 27T 7 _ c
8.2. REPRESENTATION OF DELTA FUNCTION 111
Solution
a) In this case we again consider
1 yOO pOO
1= lim ^ - / dx f ( x ) eikx-ek dk . (8.2.19)
c-to ¿ir j_00 J-oo
First, we evaluate the integral over k (by completing the square in the argument
of the exponential and changing variables).
f
-Í7
ikx — ek
dk e x p l - - (8.2.20)
J — C
Then we get that
í= (8.2.21)
»mo¿/„vêxp("í;) n
')jx•
Now, as e —> 0 the exponential in the integrand is non-zero only near x = 0 i.e.
within a distance 9c where 9 is some finite number. So we can use a mean value
theorem to write
(8.2.22)
b) For this part we proceed exactly as in part a) but we change the integration
variable from x to y = ax. The desired result then follows immediately from
part a).
c) This result is also obvious from the calculation in part a).
d) In this case we first consider the expression
W
2
I = lim lim _ dy dk dq e ik(x-y)-e1k Jq(x-y)-e2q2 (8.2.23)
¿i—^0+ é2—•o-j- (2
and integrate over k and q to get
w
(z - y)2 (y - z) 21
I = lim lim , dy\ exp .(8.2.24)
d->0+e 2 -+0+ (2 f 1^2 4ci 4í 2
The argument of the exponential simplifies as follows
e 2 (x - y)~ + ex(y - z) 2 _ (x - z) 2 + gt + e2 e2x + t\Z

(8.2.25)
4ti€2 4(ei+c2) 4eic2 f l + f2
If we now substitute this result back into (8.2.24) and integrate over y we get
I = —
{x
exp (8.2.26)
2tt y e i + £2 4(ei + e 2 )
^Jiauuiiujvj ai\U PUUHIER TRANSFORMS
Now taking the limit e 2 —» 0 + we immediately obtain an expression that we

recognize from part a) to be the desired result.
e) In this case we consider the expression
= lim -7- í f(x)dx [ ¿H*'—')-*' dk

e->o+ 2tt J
/(i),fa expí l
J
11
= ,!ísv¿/ \/f * (8.2.27)
In the limit as e —> 0 + the integrant vanishes everywhere except in the vicinity
of x2 = a2 which is the same as in the vicinity of x = ±a. So we can use a mean
value theorem and write
(x — a) 2 4a 2
I = lim 7 7 - ^ / (xa + 0e) / \/*f(x)dxexp
>0+ 2?r 1/ ' ' J V £"V ' 4e
J fa [f{a+0e)+f{ a+e (x + a ) 2 4 a 2 1
= ^+h - ^\f^]/
+ f(-a + 0c) x) dx exp
= ¿ ] [ ^ a ) + ^(-a)] • (8-2.28)
8.3 Normalization of Scattering Solution

Show that the appropriate normalization for the positive parity solution for
scattering from a square well
{ cos (kx — 6+)

j4cos KX
c o s ( k x + J+)
if x <—a
if |x| < a
if x > a
to yield ¿-function normalization is l/y/ir. You will have to use the continuity
of the wavefunction at x = ± a as well as the equation
kta,n(ka + S+) = K tan Ka .
Solution
We begin with the given solution.
{ cos(kx - S+(k))
A(k) cos(A'x)
c o s ( k x + S+(k))
x < a
—a<x<a
x > a
where throughout we have k > 0 and, from continuity of the wavefunction, that
(8.3.29)
cos ka
A(k) = ( +S+(k))
cos (Ka) (8.3.30)
8.3. NORMALIZATION OF SCATTERING SOLUTION 113
as well as
ktan(ka + S+(k)) = A'tan(A'a)

q tan(qa + S+{q)) = Qtan(Qa) (8.3.31)
with
+ ^ (8.3.32)
The normalization is given by the integral
•^2I=^2 J ip+(k,x)il>+{q,x)dx (8.3.33)
or
[~a
I = / dx cos[^x — <?+(&)] cos[ga; — i+(g)]
J—oo
r
+ >l(fc)^4(g) / dx cos(Kx) cos(Qx)
J—a
+ poo
/ dx cos[fca: + Ó+ (k)] cos[gx + <5+ (g)]
Ja
poo
= 2 dx cos[&£ + i+(&)] cosf^rx + ¿+(9)]
Jo
+ 1A(k)A(q) / dx cos(Kx) c o s ( Q x )
J—a
— 2 dx cos[kx + S+(k)]cos[qx + 6+(q)]

Jo
2
= 7T6(k -q)~ fc2 _ q2 [ksm(S+(k)) - qs\n(S+{q))]
cos[&a + ¿ + (k)] cosfga + 6+ (<7)]

cos(A'a) cos(Qa)
(I< + Q) sin((A" - Q)a) + (K - Q) sin((A' + Q)A)
X
k2 — q2
{(k + q)sin[(k-q)a + 6+(k)-S+(q)]

k2 — q2
+ {k - q)sm[(k + q)a +S+(k) + 6+(q)]} . (8.3.34)
We now use equation (8.3.31) to simplify and get
I = v6{k-q). (8.3.35)
Therefore, we find that
C = ~ (8.3.36)
o. I / U j. iiiDu IJL/JVJ /IIVL* R (JUKIKH THANSt (JRMS
8.4 Tempered Distribution

Show t h a t T is a tempered distribution if T is defined by
k=oJ~°° ax
where Fk are continuous functions bounded by
ii*(*)i<c fc (i+m
for some Ck and j depending on k. As a matter of fact every tempered distri-
bution can be written in this form. Symbolically one then writes
* dkFk(x)
r = £(-i)
dxk
k=0
This formula cannot be taken literally however since the Fk (x) need not be
differentiable. It arises from a formal integration by parts of the first equation
above.
Solution
Let / G S. Then,
dkf
^ £ 5 for all k . (8.4.37)
Furthermore, \x\*f(x) <E S for all j. Therefore, if we define

m
f°° dk f
T F
(/) = E / «( )Vkdx
x
(8.4.38)
k=oJ-°°
we find t h a t
m
roo .
dkf
\T(f)\<¿2 / + dxk
dx. (8.4.39)
k=0J-oo
Each term on the right can be bounded by one of the semi-norms || / || r n and
hence, T ( f ) is a continuous functional on S.
8.5 Fourier Transform of V

The test function space V consists of the space of functions of bounded
support. The support of a function / , (supp / ) is the complement of the largest
open set on which the function vanishes. Show that if / € T V then / is an
entire function.
8.6. TEMPERED DISTRIBUTION OF FAST DECHEASE
Solution
Consider
(8.5.40)
/ OO
ip
e- *f(x)dx
-oo
Since the support of / , s u p p / , is bounded, the integral on the right converges
uniformly for all complex values of p. We may therefore differentiate under the
the integral sign with respect to the complex variable p to get
(8.5.41)
Thus, / is an entire function.
8.6 Tempered Distribution of Fast Decrease

Prove the Theorem: The Fourier transform of a tempered distribution of fast
decrease is a C^°°^ function bounded by a polynomial. A tempered distribution
of fast decrease F is of the form
F = fT
where / £ S and T is also a tempered distribution.

Hint: To prove that the Fourier transform of F is bounded by a polynomial use
the result of problem 8.2.
Solution
We have t h a t a tempered distribution of fast decrease is of the form / T where
/ G S and T is a tempered distribution. Now, f ( x ) elkx is itself a test function
belonging to S and therefore
(8.6.42)
is uniformly convergent for all real values of k. Using the representation of T

obtained in problem 8.2, namely
RU^N-tr
(8.6.43)
we can perform the N=0differentiation

J-CO and obtain t h a t the highest possible power
of k that occurs is km. Furthermore, since the integral is uniformly convergent
we may differentiate under the integral sign any number of times to obtain that
the Fourier transform is a C°° function.
8.7 A Useful Identity

Let f ( z ) be an entire function vanishing rapidly at large |9?(z)|. Show that
Unr i r 1 1 x
f i )
+ dx
[x — a + te x — a — ic x—a
£-+o+ 2 J_,
where the principle value integral is defined by
•f f /(«) r f(x).
x
f ( )
dx = lim J — C
x—a eô+ Ja x —a
J —I
Furthermore show t h a t
e/ir
lim = ¿(z) .
e^+0+ x + e 2
2
Hence conclude that, considered as distributions,

1 „ 1
lim =F iir8(x — a)
£-•0+ x — a ± ic x —a
that is,
f°° f(x)dx f(x)dx . .. .
hm / — , . = P / qp m f ( a )
*->°+ J-oo x - a±ie J_00 x —a
Solution
To study
P
1 1
/ = lim — + f ( x ) dx (8.7.44)
e-»0 + ^x — a — i c x — a + ic
we consider the 1contour integral
I f .J-I 1 1
a
2 R \x — a — ic + x — a + te
f ( x ) dx
1 1
+ f ( z ) dz
a — ic z — a + ic
- -27t? f(a + ¿e) (8.7.45)
around the contour shown in figure 8.1. We could equally well have closed the
contour in the lower half plane. Next we consider the integral
-L rr R ñ ± i 1 m f /(*)
x + d i + dz
R x -a JA+e x a
jci 2 a
jcr z—a
Prmdx+r n±dz+i
R
dz (8.7.46)
where the contour isX shown in figure 8.2. To evaluate the integral around the
J-R ~ A JCR z -a JCI z - a
8.7. A USEFUL IDENTITY 117
S(X)
CR
a
. R »(*)
v
x a — te '
Figure 8.1: The contour used to evaluate the integral I.
CR
rrs CI
R »(z)
Figure 8.2: The contours C\ and CR
semicircle C\ of radius e we let
z = a + e e,e . (8.7.47)
We then obtain
I Í
f(a + eeie
i e e'e dd = —inf(a) . (8.7.48)
Here we have already taken the limit of e —)• 0 under the integral sign since the
j cconverges
integral ^ - A jn
uniformly. This means we can conclude that
P ¡ R ñ * L J X = _ ¡ lü*+,•„/(„). (8.7.49)
J-R x - a JCr z-a
Next taking the limit e 0 + of equation (8.7.46) and comparing with (8.7.49)
we conclude that
1
ii,„ I f l — L + f(x)dx . (8.7.50)
J-R x — a c-fo+ 2 J_R \x — a it x — a + ic
After taking the lim/j-ôo we have the desired result.
o n / i r i M . o. JJ1S1 tilt!U11UNS AND FOURIER TRANSFORMS
For the next part we simply do a contour integration by closing the path of
integration with a large semicircle (at infinity) in either the upper or lower half
plane. In either case we get
dx =
lim
€—•o-f"/ oo 21
TTTTi fW
+ ' " '
~ 2 7 r i 42ie?
£-+o+ 7T
= /(°) • (8.7.51)
•oo
Now a simple partial fraction decomposition yields t h a t
e/n _ 1 1 1
2 2 (8.7.52)
x + e 2rr x — a — te x — a + ie
Combining this with our previous computation we again have the desired result.
8.8 A Representation of S(x)

Using the result of problem 8.4 and defining
1 roo rU
-2 wf_ c eikx dk = lim / e ' k ( x + ' e ) DK + / ) dk
c«fe(*-«e
>o+ 2tt
.J 0 J—oo
Prove that
f OO
dk = <J(x) .
/ OO
•oo eikr
Solution
The desired result follows immediately by writing out the integrals. Thus,
ik x ie
fHm+ -L Qf e ( + ) dk + J ««(-'O dkj
,. 1 / -1 1 e/n
= lim -—: I 1 lim (8.8.53)
«-»o+2ni \x + ie x — ie e - + 0 + X¿u +, 2e '
8.9 Fourier Transform of 5^ n \x)

Calculate the Fourier transform of ¿( n )(x).
Solution
In this case we can proceed quite formally
j 8^\x)eikxdx = ( - 1 ) n J ó(x)£¿eikx dx
= (-ik)n J S(x) eikx dx = (-ik)n . (8.9.54)

8.10. VALUE OF 119
8.10 Value of xrnS^n\x)
Show t h a t
{ 0
(—l) m m!i(a:)
if
if
n < m
n = m
(-l)mt^á(n- m )
(*) if
" > m -
Solution
Let f ( x ) be a test function. Then,
J xm «(»)(«) f ( x ) dx = (-1)"¿T (xmf(x))l=o - (8.10.55)
Now we consider the three cases separately and remember that / ( 0 ) is finite.
Thus, we find:
i) for n < m the righthand side vanishes. So,
zm¿(n>(x) = 0 T K m . (8.10.56)
ii) for n = m the righthand side yields (—l)"n!/(0). So,
xm S^(x) = (—l) n n!á(x) n = m. (8.10.57)
iii) for n > m the righthand side yields (—1)" ( ^ r ¿ ^ y í / ' " _ m ' ( 0 ) . So,
xm<S<">(x) = (-1)" "! X|(-l)("-")f("-m)(«)

(n — raj!
= (—l) m " ! X ,¿("~ m )(x) n > m . (8.10.58)
(n — raj!
8.11 Distribution Occurring in Fermi's

Golden Rule
Show t h a t the function
*(«,*) = - 1 - |e í - É - 1| 2 (8.11.59)
1
U¿T '
that occurs in the derivation of Fermi's Golden Rule satisfies the following prop-
erties.
a)
OO
/
<t>(u),t) dui = 2n . (8.11.60)
-OO
b)
( lim — 2TTS(UI) . • (8.11.61)

Solution
a) We begin by rewriting
^ (e*" - l ) (e-<¿U>t
1 — iujt j
( 1 — COS U)t) . (8.11.62)

JH
So,
• COS Ljt
du
/ OO 2 f°° 1 ujz
4>(u>,t) du = - — cos z
—
-oo J — OO dx . (8.11.63)
.ifi
After an integration by parts«/ — oo
this integral reduces to
2 [ ^ d x = 29 f —dx (8.11.64)
J — oo X J— oo X
and may be evaluated by contour integration. To do this we realize that this

integral is to be understood as a prinipal value integral so t h a t what we want is
r-t- € eo%x
,x fOO
roo e ,xx
x
lim / — dx+ — dx
e—f0+ .J-oo X Jt X
Thus, to evaluate this integral we close the contour with a large semicircle C in
the upper half plane and a small semicircle Co of radius e around the origin as
shown in figure 8.3. Since there are no poles inside the contour the total contour
9(z)
Co
R 8(*)
Figure 8.3: Contour used to evaluate the integral ~—dx .
integral vanishes and we get

f f~£ pix poo ix
0 = lim / — dx-j- — dx
£-*•0+ [7_ OO X J£ X
í eix f eix 1
+ / dx + / —dx\ . (8.11.65)
Jc x Jc 0 x J
BIBLIOGRAPHY 121
The contribution from the large semicircle vanishes and after taking the limit
for the first two terms we see that
f°° — dx = - lim [ — dx
x
J-oo x <-*°+Jco
f° exp(ze e !
— lim I "—-iee,ed9
£->0 +
Jw eei6
r°
= — i / dO = ix . (8.11.66)
J IT
Therefore,
OO
/
</>(u,t)duj = 2tt . (8.11.67)
-OO
b) To prove that
( |im 0(«M) = 2tt(5(w) (8.11.68)
we integrate the left hand side with a test function f(uj) before taking the limit.
Thus, we consider
oo »oo 2
/
<t>(w,t)f(u)du> = lim / ——(1 — cosuit)f(oj) du
• OO '-•"j-oo w '
f°° 2
= lim / — (1 — cos x)f(x/t) dx
'-"loo z2
f oo 2
=
/(0) I —r(l—cos x) dx
d-oo
= 2tt/(0) . (8.11.69)
Bibliography
Chapter 9
Algebraic M e t h o d s
9.1 A n Operator Identity

Show that for any two operators A, B such that
[A, B] = c a c-number
it is true that
[A,eB] — ceB .
Hint: Expand eB and use the fact that
[A,Bn] =ncBn~1 .
Solution
We are told that
[A,B] = c a c-number (9.1.1)

and thus it follows that
[A,Bn] = ncB"-1 . (9.1.2)

But,
oo
n=0
oo
= Y-^ncb"-1
„ nn!I
n=0
oo
= c y - B"-1
v 7
n=1
= ceB . (9.1.3)
9.2. EXPECTATION VALUES: SIMPLE HARMONIC OSCILLATOR 123
9.2 E x p e c t a t i o n Values: Simple Harmonic

Oscillator
Use algebraic techniques to evaluate the following expectation values as a func-
tion of time for a simple harmonic oscillator state which at t = 0 is given by
ip(0, x) = Au0(x) + Bui(x) +Cu3(x)
where UQ(X) is the ground state, U\(x) is the first excited state and u3(x) is the
third excited state of a S.H.O.:
(H) , (p2/2m) , (1/2 kx2> , (p) , (x) , ( A p ) 2 , (Ax)2 .
Solution
The wavefunction at t = 0 is given by
$ — Auo Bu\ Cu3 (9.2.4)

2 2 2
where \A\ + \B\ + \C\ = 1 and the Hamiltonian is
H = ¿ + \k„' , (9.2.5)
so that
Hun = (n + l/2)fojun w = \fkjrn . (9.2.6)
We want ( H ) , (p2/2m), (1/2 fea:2), (p), and (x) as functions of time. Therefore,
(H) = hjj \\A\2 + l\B\2 +7-\C\2 (9.2.7)
Now, using p = (mhui)/2(a) — a) we have
= —hu>(a — a*) 2 = —hu>(a2 + a* 2 — aa* — aâ) . (9.2.8)

2m 4 4
Also the state at time t is
9(t) = Au0e~iut'2 + Buie-3iwt'2 + Cu3e~7iut'2 . (9.2.9)
Next, using that
aun = i/ñun_i , a)un = \ / n + l « n + i (9.2.10)

we get
af 29(t) = y/2Au2e-iut'2 + y/6u3 e~3iut/2 + ^20Cu5 e~

7iwt/2
a2 V(t) = V&Cu-ie-7^1'2
aW(i) = B u \ e ~ 3 i w t / 2 _|_ 3Ci¿3 e ~ 7 i w t ¡ 2 . (9.2.11)
<jtiAfl'ER 9. ALGEBRAIC METHODS
Furthermore,
a a ' = aâ + 1 . (9.2.
So, we find
= [ s ' C V 6 e - 4 i w t / 2 + C*B\fáe4i"t/2 - 2 [ B * B + 3 C * C ) - l]
= ^ [ 2 ( | 5 | 2 + 3|C| 2 ) + 1 - 2V6\B*C\ cos(2ut + 4)
where
B*C= \B*C\ e'* .
Since
it follows that
<i*^>=<*>-<£>.
Next, we again use that
p = i —-—(at — a) as well as x = + a)
and find
(p) = —V2mfiw|j4*5| sin(w< + a )
/ ih
(i) = a / |A*B| cos(ait + a )
V mu
where
= |^*s|eia .
9.3 Angular M o m e n t u m Matrices

a) Compute the three 3 x 3 matrices
— (Yjm, LjY\mi)
where j = ar, y, z.
b) Show that they satisfy the cyclic commutation relation
, Ly\ — xfi L j etc.

9.3. ANGULAR MOMENTUM MATRICES 125
c) Furthermore show that each matrix Lj satisfies the characteristic equation
{L) - h21) Lj= 0.
d) Evaluate in closed form the expression for the matrix e x p ( i L z 9 ) .

Hint: Write out the series and resum it after using the characteristic equation to
simplify. Compare this result to a rotation matrix corresponding to a rotation
through an angle 9 about the z-axis for the quantities corresponding to
x + iy x — iy
2
' '
Solution
a) We begin with
Lz Yhm = mhYim . (9.3.21)
Therefore,
(¿ 2 )m,m' = (Yi,m,LxYiimi) - mh8m,m' • (9.3.22)
As a matrix this has the following form
1 ü
0 0 0 (9.3.23)
0 0
Also
= {Y\,m,L± y 1 > m .) = \Jl — m'(m' ± . (9.3.24)
In matrix form this becomes

/ 0 1 0 \ / o o o
L+ = 0 0 1 j , = >/2fi [ 1 0 0 ) . (9.3.25)
\ 0 0 0 / \ o 1 0
So we finally get
Lx - i ( l + + 1 _ ) = | 1 0 1 ) (9.3.26)
V2 0 1 0
0 -i 0
L ^ y ( I + - q = A | ¿ o -i i (9.3.27)
b) By straightforward matrix multiplication we now find that
[Lx,Ly] = ihLz , [Ly, Lz] = ihLx , [Lz, Lx] = ihLy . (9.3.28)

126 CHAPTER 9. ALGEBRAIC METHODS
c) In a similar manner we find that these matrices satisfy the characteristic

equation
(Ll - h2)Lx = (Ly - h2)Ly = (L] - h2)Lz = 0 . (9.3.29)
d) Writing simply L for Lx or Ly or Lz we now have that
L3 = h2L , L4 = h2L2 , L5 = h4L etc. (9.3.30)
Thus,
iL9 , iLO i f u e 1 ÍL6

exp 1 + + + ...
h x + 2! ( T 3! ~h~
(iO)3 (id)5
4- + V r - + ••
3! 5!
2
(¿fl) | (i0)< | (fg)« |
+ t
2! 4! 6!
, L2 iL s. m L2
1 — TT + ~T~ " + T7T cos 9 . (9.3.31)
h2 h?
H
Now,
i o 1 0 0
u
0 0 0 0 0 0 (9.3.32)
h
0 0 h2
0 0 1
and
0 0 0
l 0 1 0 (9.3.33)
h2
0 0 0
Therefore writing (9.3.31) out for L — Lz and 9 = 7 we find
e't 0
0 (9.3.34)
6 X P
( ^ 0 0 e 7
To see what this means, consider a rotation about the z-axis
x' = x cos 9 + y sin 9

y' = —xsin9 + ycos9
z' = z . (9.3.35)
Instead of x, y, z we now introduce
x + iy
Zl = 7=-
v2
x0 = z
x — iy
x_i (9.3.36)
v2
9.4. DISPLACED OSCILLATOR 127
since then
xm = ^ y l i m . (9.3.37)
For these variables the transformation properties are
®i = nei7
Xg — X()
x'_! = (9.3.38)
This allows us to write
= 5Zexp ("T") Xm
'' (9.3.39)
9.4 Displaced Oscillator

Diagonalize the following Hamiltonian, that is, find all of its energy eigenvalues.
H = E a) a + V(a + a*)
where E and V are constants and

[a, at] = /?2 ,
a positive constant.
Hint: Try to transform to operators
6 = ua + v , = u* a} + v*
where u and v are complex numbers, and recall the simple harmonic oscillator.
Solution
We are given
H = E a)a + V(a* + a) (9.4.40)
where
[a,a*] = 02 (9.4.41)
So we try as suggested
6 = ua + v , = u*a* + v* . (9.4.42)
Then,
[6, 6*] = u2[a, a*] = /? 2 |u| 2 . (9.4.43)

This suggests t h a t we should choose u = 1//3. Then we get
[6,6*] = 1 (9.4.44)
v_,iuxr-1 an y. ALGEBRAIC METHODS
and we find
a = (3(b-v) , a1 = - V*) . (9.4.45)
So,
H = /32E(bH — vtf — v*b + |u| 2 ) + /3V(b^ + b - v — v*) . (9.4.46)
Therefore, if we now choose v real and such t h a t (32Ev = ,6V then the terms
linear in b and 6* go away and we have just a simple harmonic oscillator Hamil-
tonian.
H = P2Eb^b-^. (9.4.47)
Jb
The energy eigenvalues are therefore given by
En = (n + 1 / 2 ) 0 2 E - — . (9.4.48)
9.5 Dipole Matrix Elements

Evaluate matrix elements of the form (Y¡tm, zY¡iim>). These occur in the evalu-
ation of dipole radiation rates.
Hint: Use the following recurrence relation:
(21 + 1 )xPF(z) = (l + m)P[21 («) + ( / - m + l)P^(x) .
Solution
In spherical coordinates we have 2 = r cos 9. So we have to evaluate the integral
1 = J sin 9 d9 dip Y[*m (6, ip) cos 9 Yr ¡m> (0, tp) . (9.5.49)
Using the explicit form for the spherical harmonics
1/2
Y¡¡m(0,ip) = '21 + 1 ( / - m ) P Pi"1 (cos0) etmip (9.5.50)
4 7T (I + m)\
we see that the ^-integration yields ó m ¡ m >. After we make the substitution
u = cos 9 the integral becomes:
J = y/(2l+l)(2l'+ l)(l-m)\(l'-m)\
4tr (I + m)\(l' + m)!
But, i: du P™(u) u P™(u) . (9.5.51)
u p¡Jl{u) =
2ÜTÍ[(/'" m + 1)p[r+1 + ( /
'+ • (9 5 52)
- -
9.6. SCALAR OPERATOR 1'29
Therefore we find t h a t
y/(21 + 1)(2V + !)(/ — m)\(l' - m)!

I =
4tt (I + + m)\ m ' m '
2
[(/ — m)6iji+i + (/ + 1 + fn)8iy
( 2 / + l ) ( 2 / ' + l)
1 (/ — m)!(/' — m)!
27r>/(2/ + 1)(2/' + 1) (I + "»)!(/' + m)! '
x [ ( / - m)¿i 7 c + i + (/-(-1 + . (9.5.53)
9.6 Scalar Operator

Let A be an operator such that
[A, Lx] = [i4, Ly] = 0 .
Calculate [A,L2].
Hint: First compute [A,Lz\-
Solution
We have
[^37) Ly\ Lz
Therefore,
ih[A, Lz] — [A, [Lx, Ly]] = [A, LxLy] — [A, LyLx]

= [v4, Z/r]Ly + LX[A, Ly] — [y4, Ly]LX] — LyA, Lx
= 0. (9.6.54)
Thus, [A, Lz] = 0 and hence it follows that
[A,L2} = 0. (9.6.55)
9.7 Probability to Obtain l,m

Let a wavefunction be given by
% — y + 2 ixy 3z
V> = A + 2 2 2
+ 5 exp(—oí\J x + y + z )
«.2
-j.2 +I y 2 I y2
+ z 2
y J x 2_s r y2j s _
Find the probability of obtaining any (l,m) value.

Solution
The wavefunction is
x2 - y2 + 2ixy 3z -ay/x^+yi+z' 2
r¡) = A 2 2 + 2 2 2
+ 5 (9.7.56)
X2
+ y + z ' y/x + j/ + z
In spherical coordinates this can be rewritten
ip = A [sin 2 9 e2itp + 3 cos 9 + 5] e _ a r . (9.7.57)
In terms of spherical harmonics (see for example table (9.1) of [9.1]) we find
'XO'TT I t+jr.
tp = A - ¡ t - Y2 2 + 3 y — Yi 0 + bViivYo 0 (9.7.58)
15 V o
Thus, a measurement of L2, Lz will only yield the values (2,2), (1,0), (0,0). The
corresponding probabilities are
32tt 1 _ 32
^2,2 = 15 327r/15+9(4?r/3)+25(47r) —
1712
p _ 12ir(15) _ _180
1712ir ~ 171'
1712
0,0
1007t(15) 1500
(9.7.59)
1712ir 1712
9.8 Probability to Obtain /, m A l o n g Different

Axis
a) A measurement of the z-component of angular momentum is made on a
particle in a state of total angular momentum 1 and z-component 1. W h a t are
the probabilities for obtaining the values 1, 0, —1?
b) For a particle in a state of total angular momentum 1 and z-component 1
a measurement is made of the component of angular momentum along an axis
lying in the x — z plane and making an angle 9 with the z-axis. W h a t is the
probability of getting the values 1 , 0 , - 1 ?
Solution
a) We have
Lxxp = hip , L2ip = 2 h2%l> (9.8.60)
So, the total angular m o m e n t u m is h i. We can therefore represent the angular

momentum operators by the following matrices.
0 1 0
1 0 1
0 1 0
9.8. PROBABILITY TO OBTAIN I, m ALONG DIFFERENT AXIS 131
n °.
= —=\ I 0 -I
i o
1 0 0
= h | o o o (9.8.61)
0 0 - 1
The eigenstate of Lx corresponding to Ih is
*=5 ^ (9.8.62)
The eigenstates of Lz are
1 0
vi = i 0 , v'o = tf-i = I 0 (9.8.63)
0
Thus, the required probabilities are: Pm = |(V>, ipm)\2• This yields,
P+i = - , Po = - , = - • (9.8.64)
b) In this case the state corresponding to Lz = Ih is
ip = | 0 (9.8.65)
0
We want the eigenstates of the operator
/ cos 9 4= sin f 0 \
>/2
L$ = Lz cos 9 + Lx sin 9 — h 0 cos 9 (9.8.66)
72sin(
0
72 sin V — cos ( /
The eigenvalues are, of course, still úzh, 0. The eigenstates are found in the
standard way to be (up to an arbitrary phase factor)
j / ± ( 1 ± cos 9) \/2sin 9
\/2sin 9 , ip o —2 cos 9 (9.8.67)
± ( 1 cos 9) —\/2sin 9
The probabilities are as always given by Pm = \(ip, ipm)\2. So we get
P+i = - ( 1 + cos#) 2 , P o = ^ s i n 2 0 , F _ i = i ( l - cos0) 2 . (9.8.68)

9.9 Commutators of x and p with L

Compute the commutators [x, L] and [p, L] and compare the results with [L, L],
W h a t does this suggest about the commutator [A, L] where A is an arbitrary
vector operator? Note that each of the commutator brackets above constitutes
nine commutators.
Solution
By definition
L —r x p . (9.9.69)
This means
Lx = VPz - zpy , Ly = zpx - xpz , Lz = xpy - ypx . (9.9.70)
Then we find
x, Lx] = 0
X, Ly] = z[x,px] = ihz
x,Lz] = -y[x,Pr] = —ihy
V) Lx] = -z[y,Py} = — ihz
V,Ly] = 0
y, LZ] = ~x[y,Py] = ihx
Zj Lx] = y[z,pA = ihy
Z
i Ly] = -x[z,pz] = —ihx
z,Lz] = 0. (9.9.71)
This means that r and L satisfy the same sort of cyclic commutation relations
as L and L. Similarly we find t h a t p and L satisfy
[Px, Ly] = ihpz etc. in cyclic fashion . (9.9.72)
It is therefore quite reasonable to guess t h a t for a vector operator A we will

have
[Ax, Ly] = ihAz etc. in cyclic fashion . (9.9.73)
9.10 Some Eigenfunctions of Angular

Momentum
Let
ip*=xf{r) , xpy = y f(r) , tp{z) = z f{r) .

Show, by explicit calculation, t h a t they are respectively eigenfunctions of Lx , Ly ,
and Lz as well as of L2 and find the corresponding eigenvalues.
9.10. SOME EIGENFUNCTIONS OF ANGULAR MOMENTUM 133
Solution
By straightforward computation we have
Lxipx = (ypz - zpy)xf(r)

z y dj_
xy xz—
i dr
0 . (9.10.74)
Similarly,
Lylpy (.zpx - xpz)yf(r)

h r x z~\ df
- yz yx- \ —
i I r ri dr
0 (9.10.75)
and
Lzipz = [xpy - ypx)zf{r)

y x d¿
zx zy—
r r dr
0 (9.10.76)
Thus, all three are eigenfunctions of the corresponding component of angular

momentum with eigenvalue 0.
Continuing we find
Ly 1pX (zpx - xpz)xf(r)

2] df h
= - \xz x - — + -zf(r
i dr i
h
-ipi (9.10.77)
i
and
Lzrpx = (xpy - ypx)xf[r)

h r 2y x df h
— ix xy—
i L r r Tv - l »/(r)
h
I (9.10.78)
—7ll>y •
In exactly the same way we find
l x 1py 71pz
I
k
L z Ipy I
~Wx
I
(9.10.79)
and
fi ,
L x rp z = yV'y
L y 1 fi z = . (9-10.80)
Therefore,
=
—LyXpz .L2rl>y
I 4
= -[-2^]ipx
I I
2h2x¡>x , (9.10.81)
, fi, h ^ r ,
LJrpy
(
= -7 J¿ 1 ^ 2 +
" iI I
2 h2 Ipy
and (9.10.82)
9 fir / fir/
Z/ ipz — T LxYy ~7 Lylpx
1 1
2h 2 rp z . (9.10.83)
9.11 E x p e c t a t i o n Value of Lx
A system is in a state of angular momentum given by
$ = a y j i + fcyô + c y i , - !
where
|a|2+|fe|2+|c|2= 1 .
a) Compute the expectation value of Lx

Hint: The equation
L±Yl¡m = y/1 (I + 1) - m(m ± l)fiYi, m ±i
may be useful.
b) Compute the expectation value of L2.
c) W h a t are the possible values of the coefficients a, b, c in order t h a t
Lx9 = h$t ?
It may again be useful to recall that
Lx = 2^+ + L-) .
9.11. EXPECTATION VAL UE OF Lx
Solution
The wavefunction is
i¡j = a v i j + 6 ^ , 0 + cYi,-i
where
|a|2+|6|2+|c|2=l
a) we use
(Lx) = \(L+) + \{L_)
and
L+ip = V^hbYi^ + >/2hcYlfi
l_v> = \ í 2 h a y \ f i + \ í 2 h b y \ ^ i .
So,
(L+) = (ip, L+tp) = V2h(a*b + 6*c)
(£_) = {il>,L+rl>) = V2h(b*a + c*b) .
Therefore,
(Lx) = ^ ( 2 3 i ( a * 6 ) + 23fi(6*c)) = x/2fi(3?(a*6) + H(6*c)) .
b) Since ip is an eigenstate of L 2 we have that
(L2) = 1(1 + l)h2 = 2h2 .
c) We want an eigenstate of Lx with eigenvalue h. This means
Lx tp = h ip .
But, writing this out we get
Lxip = i ( L + + L-)ip = ~=h(bYiA + (c + a)Yi, 0 + bY-í¡
= h(aYiti + 6 y l i 0 + cYx-i) .
This requires that
b=V2a , c + a = \/26 , b = \/2c .
Therefore,
= + y i +
^ r 7 5 ' ° •
9.12 Rotation Invariance of the Hamiltonian

Starting from the equation,
URHU^ = H
and using
, it „
U = exp{— n • L)
n
with |e| < < 1 obtain to first order in e that
[H, Ñ • L] = 0 .
Hence conclude that
[H,L] = 0 .
Solution
We are given
U r ( 0 ) = exp (9.12.97)
1
UR(9) = exp -9N • L (9.12.98)
For 9 e we get
UR(0) w 1 + r f ñ - L (9.12.99)
n
Uñl(0)t* 1-êñ L .
h (9.12.100)
Therefore, the equation
UFÍHUÑ1 = H
(9.12.101)
becomes to this approximation
1 + —EÑ • L I H ( 1 — - e ñ • L
J \ N J " ' (9.12.102)
2
or to order e
[ñ-L,H] = 0 . (9.12.103)
Since h is an arbitrary direction, this means t h a t
[L,H} = 0 . (9.12.104)
9.13. UNCERTAINTY RELATION FOR SHO 137
9.13 Uncertainty Relation for SHO

Show that for any eigenstate of the simple harmonic oscillator
(Ax) 2 = <x2>
(Ap) 2 = <p2)
and also show for the state with quantum number n t h a t
(Ax)(Ap) = (n + 1 / 2 ) h .
Solution
We have
X = ( a t + a )
\ / S j (9.13.105)
p=iJ^-(a^ - a) . (9.13.106)
Thus,
(x) = (n|x|n> = y + a)ln)
[\/n + l ( n | n + 1) + \/n(n\n — 1)] = 0 . (9.13.107)

2mu>
Here we have used the fact t h a t (n|m) = SN M. For the very same reason we
find that (p) = (n\p\n) = 0. Therefore,
(Ax) 2 = (x 2 ) - (x) 2 = (x 2 ) (9.13.108)

and
(Ap) 2 = (p 2 ) - (p) 2 = (p2> . (9.13.109)
Now, computing explicitly we have
(Ax) 2 = (x 2 ) = (n\a^ 2
+ a 2 + a*a + aa* |ri)
2 raw
= (nlaâ + aa^ln)
2 mui
( 2 + 1
= db " >' (9.13.110)
Similarly,
(Ap)" = (p 2 ) = 2
+ a 2 — aâ — aa^|n.)
TTlhuJ
= — — (2n + l) . (9.13.111)
So, we get that

h2
(Axf(Ap)2 = — (2n + l ) 2 . (9.13.112)
Therefore,
Ax Ap = h(n + 1/2) . (9.13.113)
9.14 Baker-Campbell-Hausdorff Formula

Show that if you have two operators A, B such t h a t
[A,[A,B]] = [B,[A,B}] = 0
then
eA+B = eA eB e-l/2[A,B]
Hint: Consider the operator
f ( x ) = exA exB
and show that
^ = (A+B + [A,B]x) f(x) .
Integrate this equation and obtain the desired result. This result is a special
case of the Baker-Campbell-Hausdorff formula.
Solution
The following derivation is due to Glauber. We have
Hi [A, 5]] = [B, [A, B]] = 0 . (9.14.114)
Now, consider the operator
f ( x ) = exA exB . (9.14.115)

Then,
~ = AexA exB +exA BexB

ax
= (A + exA Be~xA)f(x) . (9.14.116)
We also have t h a t
[B,An] = nAn~1[B,A] (9.14.117)

so that
[B,e~xA] — —xe~xA[B,A] (9.14.118)

9.15. A USEFUL COMMUTATOR 139
and
exA Be~xA = B -[B,A]x . (9.14.119)
Therefore,
P = (A + B + [A,B]x)f(x) (9.14.120)
ax
with / ( 0 ) = 1. Furthermore, since
[A + B,[A,B]] = 0 (9.14.121)
we can integrate this like an equation for a c-number function to get

x{A+B)
f(x) = e . (9.14.122)
If we now set x = 1 and multiply from the right by e - 1 / 2 ' ' 4 " 0 ' we get the desired
result
eA+B _ eA eB e-i/2[A,B] (9.14.123)
9.15 A Useful C ommutator

Show that if
[A, B] = \A
then
AeB = exeB A .
This formula can be used, for example, if the operator B is proportional to the
Hamiltonian / k j ( a t a + 1/2) and the operator A is proportional to either a or a*.
Solution
Given
[A, B] = XA (9.15.124)
we find, using a proof by induction, that
ABn = (a + B)nA . (9.15.125)
Thus,
AeB = V l ( A + S ) " A
* J ni
nI
n
= ex+BA
= exeBA. (9.15.126)
CHAPTER 9. ALGEBRAIC METHODS
9.16 Uncertainty in Lz
A particle is in a state described by the wavefunction
- -7=1Yi,i + 2iYi,i-i - Yij-2] .

v6
a) W h a t are the most probable values that would be obtained in a single mea-
surement of L2 and Lz? W h a t are the possible values of Lz and what are their
corresponding probabilities?
b) Compute ( A L z ) 2 the uncertainty in Lz for this state.
c) If the state is now simply Yi)Tn find the uncertainty ( A L z ) 2 in Lz.
Solution
The particle is in the state given by
tP = 4 = {Yi.i + 2íY¡i,/_i - YIJ-2) • (9.16.127)

v6
a) A measurement is then sure to yield that the value of
L2 = 1(1 + l)fi 2 (9.16.128)
since the state is an eigenstate of L2 with this eigenvalue. The possible values
of Lz are:
lh , (l-\)h , (l-2)h
with the corresponding probabilities:
1/6 , 4 / 6 , 1/6 .
Thus, the most likely value to be obtained is (/ — 1 )ft.
b) ( A L z ) 2 is given by
(A Lz)2 = (L2Z)-({LZ))2
= j [ l 2 + 4(I - l ) 2 + (/ - 2) 2 - {l[l + 4(1 - 1) + (/ - 2)]} 2
= J^2 . (9.16.129)
c) In this case the state is an eigenstate of Lz and thus we are certain to get
the eigenvalue namely mh. So, A L z = 0.
9.17 E x p e c t a t i o n Values of Angular M o m e n t u m

Given the state |Z, m)
a) compute the expectation values ( £ 2 ) , (L2),and ( L 2 ) as well as the uncertain-
ties A L x , A L y , a n d A L z .
b) Discuss how the uncertainties computed in part a) conform with the com-
mutator
[LX , Ly\ — ihLZ

g.17. EXPECTATION VALUES OF ANGULAR MOMENTUM 141
Solution
a) We begin with the step-up and step-down operators
L± = Lx±iLy . (9.17.130)
So,
Lx=l-(L+ + L_) , Ly = i ( L + -L_) . (9.17.131)
Therefore,
1
Ll = -(Ll + Ll + L+L- + L.L+)
L2 = -1-{L\+L2_-L+L--L-L+) . (9.17.132)
Now,
(/, m\L± |/, ra) = (/, m\L± |Z, m) = 0 . (9.17.133)
Therefore,
(/, m | L 2 | / , m ) = (/, m\L2y\l, ra) = ^(l,m\L+L-+ L-L+\l,m) . (9.17.134)
But,
L± |/, m) = \/l(l + 1) — ra(ra ± l)h\l, m) . (9.17.135)
Therefore,
L^L±\l,m) = \/l(l + 1) — ra(ra ± l)HL^\l, m ± 1)

= \/l(l + 1) - m ( m ± 1 + 1) — (ra ± l)mft 2 |/, m)
= [1(1+ l ) - r a ( r a ± l)]ft 2 |/,ra) . (9.17.136)
So,
ft2,
(£*) = (¿y) = — [ 2 / ( / + 1) - ra(ra+1)-ra(ra - 1)]
4
ft2,
= y['('+l)-"»2] . (9.17.137)
Also, since the state |Z, m) is an eigenstate of Lz we immediately have that
(Lz) = mft
(Ll) = m2h2 . (9.17.138)
Also, as already seen,

(Lx) = ( L y ) = 0 . (9.17.139)
Therefore
ALx — a ty =
= AL — - 7 = y / l { l + 1) - m 2 (9.17.140)
v2
and
ALZ = 0 . (9.17.141)
b) From the commutator
[Lx, Ly\ — ihLz
we obtain on general grounds the uncertainty relation
ALxALy > — |([¿zi ¿y])|
= 1(LZ) = y m . (9.17.142)
So, consistency requires that
y [ / ( / + 1) - m 2 ] > y m . (9.17.143)
But, quite generally

I> m . (9.17.144)
Therefore the inequality is satisfied and the results are consistent. Notice that
for a state with / = 0 the right hand side of (9.17.142) vanishes so that the
product of the uncertainties
ALxALy = 0 .
This is also consistent since the individual uncertainties A L x and A L y vanish

separately.
9.18 Validity of Ehrenfest's T h e o r e m

Ehrenfest's theorem states t h a t
!«=-<£>=<"<*»•
Suppose that
F(x) = Ax2 , A = constant .
Given that the system of interest is a simple harmonic oscillator and the wave
function at t = 0 is
^(0,ar) = i[\/3w„(a:) + ¿u„+i(z)]
compute the difference between F((x)) and ( F ( x ) ) as a function of time.

9.19. WIGNER PROBLEM: ANNIHILATION AND CREATION 143
Solution
As a first step we need to compute the wavefunction for all times. This is given
by
rj;(t,x) = i [>/&„(*) + ¿ « n + 1 ( x ) e - ' ü ' t ] e - ( " + i / 2 M . (9.18.145)
Now,
F((x)) = A(xy \ 2 . (9.18.146)
Thus, we need
W = + (9.18.147)
But,
(i¡)(t,x), (a f + a)ip(t,x)) - ^ ^{\/3u„(x) + ¿u n + i(a;) e ,UJt

},
{V^[>/"un-i + Vn + lun+i] + i[Vn + lwn + Vn + 2un+2]e~lut })
sinuit . (9.18.148)
Thus,
FUx))
u
= AÛ+ ^ sin2 cut . (9.18.149)
v
" 4 2 mu '
Similarly,
(F(x)) = A(x2) = x), (af + a)2ip{t, z))
+
= 8 = (<^°»
{ V 3 [ \ / n ( n - l)w„_2 + (2n + l)u„ + \/(n + l ) ( n + 2)w„ + 2 ]

+i[\/n(n + l ) u „ _ i + (2 n + 3 ) u n + i + \ / ( n + 2)(n 4- 3)u n +3]e
Ah
= (2n + 3/2) — . (9.18.150)
2 mui
9.19 Wigner Problem: Annihilation

and Creation
This problem is based on a paper by Wigner [9.2]. For more discussion of that
paper see also the following problem.
Consider the Hamiltonian
H = hui(ba + 1/2) (9.19.151)
Show t h a t the Heisenberg equations
á = i[H, a] = —ihjja (9.19.152)

b = i[H,b] = ihujb (9.19.153)
imply that
[o,6]=l (9.19.154)
so that a is an annihilation operator and b is a creation operator. Assume that

all the energy eigenvalues are positive and non-degenerate.
Hint: Work with matrix elements in the energy representation.
Solution
If we take matrix elements of (9.19.152) and (9.19.153) in the energy repre-
sentation where H is diagonal and the energies are labelled EQ, E\, j? 2 , • •. we
get
ihojanm — i{En Em)anm

ifux}bnm — i(En Em)bnm .
These equations imply
En Em = -hu (9.19.155)
En-Em = hui . (9.19.156)
Thus, we have that
En = Em±fuj . (9.19.157)
Since the energies are all positive we have
Ei — EQ + hui
E2 — Ei -(- fiu - EQ -|- 2 hu)
En = EQ + nhui . (9.19.158)
Furthermore, we see that the only non-zero matrix elements are: an,n+1 a l l ( l
. We now impose that the diagonal Hamiltonian should have the eigen-
values that we obtained. Thus, we get
H = ñ w ( 6 n i „ _ i a n _ i ) n + 1/2)
= hui (n + • (9.19.159)
Therefore,
bn,n-ian-i,n = (n-l/2) + p- . (9.19.160)

ftjjj
9.20. WIGNER PROBLEM: SHO 145
We can now compute the matrix elements of the commutator.
— ^ n , n + l ^ n + l,n ^n,n — l@n — l,n
= 1 . (9.19.161)
Thus,
[a,b] = l (9.19.162)
and a and b satisfy the commutation relations for an annihilation and creation
operator, respectively.
9.20 Wigner Problem: SHO

The following problem was first considered by Wigner [9.2], Consider the Hamil-
tonian
H=l-(p2 + x2) (9.20.163)
with units such t h a t h — 1. Show t h a t the Heisenberg equations
p= x = i[H,x] (9.20.164)
—x=p = i[H,p] (9.20.165)
imply that
(\p,x] + i)2 = -{2E0-l) (9.20.166)
where EQ is the ground state energy. You may assume that all the energy eigen-
values are positive and non-degenerate as well as that all the matrix elements
of x in the energy representation are real.
Hint: Work with matrix elements in the energy representation.
Solution
Since we are in the energy representation, H is diagonal and has eigenvalues
Eo, Ex, E 2 , . . . where EQ is the lowest eigenvalue and the eigenvalues are ordered
by increasing size. Taking matrix elements of (9.20.164) and (9.20.165) we get
Pnm — i(En Em)xnm

~xnm - i(En — Em)pnm . (9.20.167)
Combining these equations we find
Xnm = (En - Em)2xnm . (9.20.168)

Thus, to have a solution we need

( E n - E m f = 1 (9.20.169)
or
(En-Em) = ±1 - (9.20.170)
So,
En = Eo + n . (9.20.171)
It also follows that the only non-zero matrix elements of x and p are of the form
I n n + 1 , x n +i,n and Pn,n+1 and p„+i,„. By hermiticity we also have t h a t
Xn,n + 1 — 2-n+l,n
Pn,n +1 = Pn+l,n • (9.20.172)
Also, from the Heisenberg equations we have
Pn,n + 1 = ÍXn,n + l > Pn + l,n ^ n + l,n — ixn}n + l - (9.20.173)

At this stage we have satisfied the Heisenberg equations. We still have to impose
the condition that the Hamiltonian (9.20.163) has only the diagonal matrix
elements En = EQ + n. So, for n ^ 0 we get
E0 ~(~ TL — ^ ^ (XNMXMN PnmPmn )
2 VXn>n + l X r l + l J n + x
n}n — lXn — l , n ~i~ Pn,n+lPn + l,n Pn,n— lPn — l , n )
X x
2 (^n.n + l n,n — 1 + •''n.n + l + n,n — l)
= <„+i + <„-i • (9.20.174)

For n = 0 we get
E0 = z 2 i • (9.20.175)
For the next few values of n we find
EQ + 1 —- Xj -1 2 + * 1 , 0 = > « 1 , 2
2
= 1
Eo + 2 -
—
x 2
2,3 + XL,2
„2
2,3
= EQ + 1
2 2
EQ + 3 - T x + X
2,3 X3 , 4
x
= 2
— 3,4
2
Eo + 4 —- T4 , 5 x + X
3,4 => X24 , 5 = i?o + 2
(9.20.176)
These results mean that
" = odd <n+i = ^("+l)

n
n = even = E0 + - . (9.20.177)
9.21- IDENTITY FOR PAULI MATRICES 147
As a consequence we find t h a t the commutator [p, x] is also diagonal due to the

relationships (9.20.173). For n ^ 0 the matrix elements of this commutator are
(n\[p, x]|n)
— Pn,n + l*En + l,n %n}n + lPn + l,n 4" Pn,n — l^n — l,n ^nyn — lPn — l,n
= 2i(xn¡n+i - xn-it„) . (9.20.178)
For the first few values of n these become
-2iE0, —2¿(1 - E0), -2iE0, -2i(l - E0), etc.
This can be summarized in the single equation
(\p,x] + i)2 = -{2E0- l)2 (9.20.179)

as required.
For further discussions of how the [p, x] commutation relations are deter-
mined by the Heisenberg equations and the Hamiltonian the reader should read
the paper by Wigner [9.2].
9.21 Identity for Pauli Matrices

Show that if A and B are two vector operators t h a t commute with the Pauli
matrices a then the following equation holds
(<x • A)(a • B) = A • É + ia • (Á x B) .
Solution
We want to show t h a t if A and B are two vector operators that commute with
the Pauli matrices 5 then the following equation holds
(a A)(a • B) = A • B + ia • (A x B) .
To see this we write out the left hand side
• (9.21.180)
After explicitly multiplying out these two matrices we get
f A - B + i{AxBy - AyBx) AZBX — AXBZ + i(AyBz — AzBy)\

\AXBZ — AZBX -f i(AyBz — AzBy) A • B — i(AxBy — AyBx) J
which establishes the required identity
(a-A)(a • B) = A-B+ ia • (Áx B) . (9.21.181)

9.22 Operator Identity - Spin Rotation

a) Verify to order A3 that for any two operators G, A
eiXG A e-iXG = A + ¿ A [ G i + M - [ G , [G, A]]+... .
b) Use this result to show that if Sx, Sy, Sz are spin operators then
e»s, v/fi e-tS,<p/h _ ^ c o s y, _ sin ¡p
eiS,v/h gê-iSt<p/h _ sycosip + Sxsinip .
Solution
a) The proof follows simply by expanding the exponentials. Thus,
ÍXG
,AG r, (¿AG)2 (iXGf
eiXG A e " = [1 + i\G + v nl' + - — +
2! 3!
„ri _ (-¿AG) 2 (-¿AG)3 - .
x A[\ H ^ 1 t " ]
2! 3!
(¿A)
= A + iX[G,A]+^-[G'A-2GAG+AG¿¡
(¿al! 3
+ [G A + 3G¿4G2 - 3 G 2 A G - AG3} + • (9.22.182)
3!
But,
[G, [G, A]] = [G, GA - AG]
G2A-2GAG+ AG2 . (9.22.183)
Similarly,
[G, [G, [G, A]]] = [G, G2A - 2GAG + AG2]
= G3A + 3GAG2-ZG2AG-AG3 . (9.22.184)
Substituting these results back into 9.22.182 we have the desired verification,
b) Since Sx, Sy, Sz are spin operators we have that
[S2 , S r ] = ih Sy
[S„Sy] = -ihSx . (9.22.185)
Using this we have
eiS,<p/h Sx e-is,<p/n = S x + ihSy + i .1 ( /2E)

iip tf .
S
2!
1 (ixp
+ f J ih J
S y + ---
3!
ip3 {p5
= Sx — Sy — — + h
2! 4! V
3! 5!
= Sx cos ip — Sy sin ip . (9.22.186)
9.23. AN OPERATOR IDENTITY 149
Similarly, we get
ismVm S y e-iS.ip/h = S y + + I ( ^ ) ft2Sy
+ ( ik)3Sx +
3! ( f t ) '
5
y2 'fi4 < p +\ ? +
+ Sx 5! ~
— <7y i — — + 1- •
~~ 2! 4!
= Sy cos tp + Sx sin tp . (9.22.187)
9.23 A n Operator Identity

Let A and B be two arbitrary operators and define
„CA BE-AA
Z(a) = e~
Show that
dZ_
= [A,Z] .
da
Use this result to derive the operator identity
2 3
c«a B e-aA = B + B] + B] j + ^[A[A, [A, B]} ] +
Hint: The right hand side is a Taylor expansion.
Solution
To prove the first part we simply differentiate Z(a), keeping in mind t h a t the
operator order must be maintained. Then,
dZ r]eaA\ Í dp~aA
de
=
^ B e~oA ++ eaA B ' d£
6
di W V~da
= AeaABe~aA - eaA Be~aAA
= [A,Z] . (9.23.188)
We now use induction. Thus, we assume that
(9.23.189)
n
Differentiating this equation it follows that
dn+lz
n+1 = [A,[A,--[A,^]]]..:i
da
= [A, [A,...[A, Z ]]]..] (9.23.190)

n+ l
Thus, we have that
dnZ
= [A,[A,...[A,B)YIJ^ (9.23.191)
dan
and therefore the Taylor expansion of Z becomes

2
eaA B e~aA = B+a[A, B]+^[A, [A, B) }+^[A[A, [A, 5 ] ] ] + - • -(9.23.192)
as desired.
9.24 Comm utator with Inverse Operator

Show that
[A,B~l] = -B~1[A,B}B-1 .
Hint: Use the fact that
[A,BB~1] = 0.
Solution
Using the hint we have that
[A,BB~1] = 0= B[A,B'1] + [A,B]B~1 . (9.24.193)
Therefore,
B[A,B~l] = -[A,B]B~l . (9.24.194)
So, acting with B~l on the left of this equation we obtain the desired result
[A, B~l] =—B~l[A, B]B~l . (9.24.195)
9.25 Schwinger M e t h o d for Angular M o m e n t a

Let (ii, a\ and , a\ be simple harmonic oscillator annihilation and creation
operators. Define
= ^(a2al+ala2)
J
2 = y(aj,ai-a{a2)
= âjai-aía2)
j = ^ (a\ai + a\a2^ • (9.25.196)

9.25. SCHWINGER METHOD FOR ANGULAR MOMENTA 151
Show that Ji, J2 , J3 obey the angular momentum commutation relations
[Ji, J2] = ih'h and cyclic permutations
and that
[J, A = 0
as well as that
J2 = j ( j + l)^2 •
Solution
The transformations given are due to Schwinger [9.3]. We first verify the angular
momentum commutation relations using t h a t
[a,-, a]] = Si:i . (9.25.197)
Then,
ih2
[•/i,J 2 ] = — [ a í a x + a j a 2 , aâ! - a{a 2 ]
= ([aía2,a2ai] _
[a2ai,aía2])
= ( [ a í , a 2 a i ] a 2 + a}[a 2 , a|,ai])
ih2 ( t t \
= — ^-aâ2 + alaij
= ihJ3 - (9.25.198)
Similarly we find
[«/2, t/3] — ^ [ci2ai aâ 2 i a 2 o. 2 ]
i^ 2 / t t t t \
= — [a'2ai + aia^ + a2a[ + a[a2 J
= ihJi . (9.25.199)
And finally
h2
[«^3, </i] = — [a{ai - a l a 2 , a \ a i + a j a 2 ]
= — a \ a i + a\a2 — a!,ai + a 2 a { )
= ihJ2 . (9.25.200)
Next we compute J 2
n2
J2 = — [(<4)2a2 + (a{)2a2 + aâia2al + a\a2a\a i
— ( a 2 ) 2 a 2 — (a}) 2 a 2 + a\a\a2a\ + a\a2a\ai
+ (a\ai + a2a2)2 — 4a\aia2a2
h2
= — (a\a\ + a\a2)2 — 4a\aia\a2
-+- 2a2aj(aja2 + 1) -f a\ai(a\a2 + 1)
h2 1
= — (a\ai + a\a2)2 + 2(a\ai + a\a2)
= JÜ + 1) • (9.25.201)
This last equation also shows that
[ J 2
J ] = 0 • (9.25.202)
9.26 M i n i m u m Uncertainty in Jx
Given the eigenstates |j, m) of J 2 , Jz find the value of m t h a t minimizes the
uncertainty in Jx.
Solution
We first write
1
Jx — 2 (J+ + J-) (9.26.203)
where
m m
J±\j, ™) = ViU + !) - ( ± !)lh m i l ) . (9.26.204)
It thus follows that
( j , m\Jx\j, m) - 0 . (9.26.205)
Also,
Jx — — (./£ + Jl + J+J- 4- J-J+) . (9.26.206)

But,
J+J— — J2 + J y — i[JX , Jy] — J~ — J2 + /lj2

J-J+ = J2 + Jy + i[JX, Jy] = J2 — J2 — HJZ (9.26.207)
Combining all these results we find that
( A J x ) 2 = (J2) - (Jx)2 = \ [ j ( j + 1) - m 2 ] . (9.26.208)
This result is clearly minimized by \m\ = j.

9.27. UNSOLD'S THEOREM AND ITS APPLICATION 153
9.27 Unsold's T h e o r e m and its Application

a) Prove that
i
m=—l
is a constant. This is Unsold's Theorem. [9.4] Also evaluate the constant.

Hint: Use the addition theorem for the spherical harmonics.
b) Use this result to show that the probability of finding a particle at the point
r, if it is acted on by a potential V(r) and is in a state of definite angular
momentum I, is only a function of r = | f | .
Hint: Assume that there is no special alignment so t h a t all the z-components
(m-values of the angular momentum) are equally probable.
Solution
a) The addition theorem for the sperical harmonics states t h a t
i
£ * Y
v.* (0 (f)' w'l =
l*m{0,ip)Yim(0',<p')
= -^ ± l p ; ( C o s 7 ) (9.27.209)
47T
m1—
=——Ll
where
cos 7 = sin 0 sin 0' cos(</? — ip') + cos 0 cos 0' . (9.27.210)
But, setting 9 = 9' and ip = ip' in (9.27.209) we get
cos 7 = sin 2 9 + cos 2 9 = 1 . (9.27.211)
This means that
Pi(cos7) = Pi(l) = 1 . (9.27.212)
Therefore,
¿ Yrm(9,v)Ylm(9,<p) = . (9.27.213)
m= — l
b) Since we have a state of definite I the wavefunction must be of the form

i
V w ( » 0 = Y . CmRni(r)Y l m (9,v) . (9.27.214)
m=—l
But, all values of m are equally probable. This means t h a t all the states
Rni{r)Y¡m(9, <p) are equally probable. So, we can take
so t h a t all states
1
1pn,l,m (0 — :Rni(r)Yim(9,<p) (9.27.216)
V2T+T
are equally probable. Therefore, the probability density for finding the particle
at r is the sum over all m values of these states
i ' i
hMr)|2 = 2TTT £ 2rn"|/?"'(r)l*Y¿n(0,<P)Yim(0,<p)
m— — l
|Bn,(r)[22¿+l
2/ + 1 4tt
2
i^wi (9.27.217)
47T
and is a function in terms of r only.
9.28 R o t a t i o n Matrix for j— 1

we showed
/ 0 1 0
Jx — 1 0 1 (9.28.218)
V 0 1 0
( 0 —i
3y - — i 0 (9.28.219)
~ V2
\ 0 i
i 0 0
0 0 0 (9.28.220)
0 0 -1
it follows that for a unit vector ñ
R = exp [¿(J/ft) • ñ0}j = 1 H -J^- sin 9 + ^ J ( c o s 0 + 1 ) . (9.28.221)
Use this result to show that the trace of this expression determines the rota-
tion angle 9 and t h a t the antisymmetric part of this expression determines the
rotation axis ñ.
Solution
Again using the matrix representations (9.28.218) - (9.28.220) we find that
( nz
V2 ^ 0 \
J •ñ nx+in y nx — inv
0 (9.28.222)
\
VT
0
nx+iny "yr
~T7o~ —n.
9.29. ALGEBRA AND CONSTANTS OF THE MOTION 155
Also, recalling t h a t
(ñ) 2 = n 2 + n 2 + n 2
= l (9.28.223)
we find t h a t we can write

2
/ - \
I J ñ\
h J
( £(1 + nz) -^(nx iny) i(nr - iny)2 \
= h¿ ~^{nx + zn
y) 1 — n? (9.28.224)
^(n^+my)2 ~^{nx + iny) §(l + n j )
Taking the trace of the right hand side of (9.28.221) we now obtain the result
that
Trace = 3 + 2(cos 0 + 1) = 4 + 2 cos 9 (9.28.225)
Thus, the trace determines the rotation angle 0. Also, taking the three different
antisymmetric parts of the right hand side of (9.28.221) we get
R21 — R12 — %/2 i[ny sin 9 + nzny (cos 9 + 1)]

P31 — R\3 — 2 inxny (cos 9 + 1)
R23-R32 = V2i[ny sin# - nzny(cos9 + 1)] . (9.28.226)
The three components nx,ny,nz are now easily solved from these expressions.
Thus, the antisymmetric part determines the rotation axis ñ.
9.29 Algebra and Constants of the M o t i o n

Consider the Hamiltonian for a symmetric two-dimensional harmonic oscillator
H = +
\Mu>2(x2 + y2)
as well as the two operators
Q = ¿ ( p x - py) + \MU2(X2 - y2)
and
Lz = xpy - ypx
a) Show that
[L„H] = 0
b) Show that if
[A, H] = 0 and [B, H] = 0
then
[[A,B],H] = 0
and hence deduce that
[i?, H] = 0 where R = [Lz, Q] .
c) Finally show that the algebra of the operators H , L2 , Q , R closes. In other

words show that no other new operators can be obtained by continuing in this
way. This shows that, by proceeding in this way, one can construct at most a
finite number of independent constants of the motion of the Hamiltonian H .
Solution
a) Since the given Hamiltonian H is invariant under rotations about the z-axis
it follows that
[LZ,H] = 0 . (9.29.227)
This result also follows, of course, from a brute force computation which yields
[LZ,H] = [xpy - ypx, ~(p2x+p2y) + ~MLO2(X2 + y2)

xfi ifi
= - 2P*Py] + -^Mui2[-1xy + 2xy]
= 0 . (9.29.228)
To evaluate [Q, H] we first rewrite the Hamiltonian as
- = +
k J^{Px
^f(Px-Py)2+
+ Py? + | m u 2 ( i + y)2
^Muj2(x-y)2 (9.29.229)
and Q as
Q = ^(Px-Py){Pr+Py)+ ^Mu)2(x - y)(x + y) (9.29.230)
Now, we easily find that
\px -py,x + y] = 0
\Px+Py,x-y} = 0 . (9.29.231)
So, we get
[(Px-Py)(Px+Py),(x + y)2] ~ (Px - Py)[(Px +Py),(x + y)2]

- -4ih(px - py)(x + y) • (9.29.232)
9.29. ALGEBRA AND CONSTANTS OF THE MOTION 157
Similarly,
2
[(Pr-Py)(Px + Py),(x-y) ] = [(Pr - P y ) , {x ~ t / ) 2 ] ( P r + Py)
= —4ih(x — y)(px + Py) • (9.29.233)
In an exactly similar manner we find
[ ( z - ? / ) ( x + 2/),(Px+Py)2] = ( x - J/)[(x + 2/),(Pr + P y ) 2 ]

= 4ih(x — y)(px + Py) (9.29.234)
and
[{x-y)(x + y ) , ( p x - p y ) 2
] = [ ( x - y ) , { p z - p y ) 2
] { x + y)
= 4ih(px - Py)(x + y) . (9.29.235)
Therefore, combining these results we find that
[Q,H] = 0 . (9.29.236)
b) Now, consider operators A, B , H such that
[A,H} = 0 and [B,H] = 0 . (9.29.237)
Then,
[[A,B],H] = [AB,H]-[BA,H]
= A[B,H] + [A,H]B-B[A,H]-[B,H]A
= 0 (9.29.238)
since each of the four commutators vanishes. Therefore, we also have that
[R,H] = [[L2,Q],H]
= 0 . (9.29.239)
c) To show that the algebra closes we explicitly calculate the operator R.
R = [LZ,Q]
2 2
= [xp y - y p x ,^(p x -p ) + ^Mu2(x2 - y2)]
= 4ih (^~-+ ^Mui2xy^j . (9.29.240)
Proceeding in t h e s a m e spirit we c o m p u t e
[Lz, R/(4ih)] = [xpy - ypx, + ~Mui2xy]
= ~ih ( ¿ ( p 2
~ py) + \mu2(X2 ~ 2/ 2 )^
= —ihQ . (9.29.241)
Also,
[Q,R/(4ih)] = [ ~ ( p I - p D + ^Mu)2{x2-y2),^- + ^Mu;2xy]
= ihuj2(xpy - ypx)
= ihLz . (9.29.242)
This shows that the algebra closes and no further constants of the motion can
be produced in this manner.
9.30 Coherent State and N o r m a l Ordering

In many cases it is convenient to "normal order" operators constructed from
annihilation and creation operators. An operator F(a, a*) is said to be normal
ordered if all the annihilation operators a appear to the right of all the creation
operators a*. In that case the operator is written with colons on each side
: F(a, a*)
Show that for any normal ordered operator : F(a, a*) : and any two coherent
states | z) |ui) we have that
( z | : F ( a , A*) : | W ) = F ( z * , U;)(Z|UJ) .
Solution
For the coherent states we have
a|u;) = w|tu) , (z|a* = (z\z* . (9.30.243)
Since : F(a, a*) : is, as indicated by the colon, normal order we have that if F
is written out as a series that
: F(a, a*) := ^ e„m(at)"am . (9.30.244)

nm
But, using (9.30.243)
( z | c „ m ( a t ) " a m | i ü ) = cnm(z*)nwm(z\w) . (9.30.245)
Hence,
(z| : F(a, a*) : |u;) = F(z*, u>)(z|ui) . (9.30.246)
9.31 N o r m a l Ordering of xn
Suppose we have units such that
« = 4(«+«») -
9.31- NORMAL ORDERING OF x" 159
Show t h a t the n o r m a l ordered o p e r a t o r (see t h e previous problem) : xn : is given

by
where H n ( x ) are the Hermite polynomials.

Hint: It may be convenient to write
'•
x
" • = /»(*)
and show that fn{x) satisfies a certain recursion relation.

Hint: The Hermite polynomials satisfy
Hn+1(x) = 2xHn(x) - 2nHn.l(x) .
Solution
We start with the algebra of the annihilation and creation operators
[a, a*] = 1 . (9.31.247)
It then follows immediately that
a(at)n = («*)"«+ n(at)"-1

a"at = a) an + nan_1 . (9.31.248)
Now, following the hint we assume that
(9.31.249)
It is also clear that
: xn : = 2 - " / 2
£í " 1 (at)"-rar . (9.31.250)
r=0 ^ '
Therefore, using (9.31.248)
: xn : x = 2~("+1>/2¿ f \ [(at)n"rar+1 + (at)"-rarat]

r=0
_ 2 -( n +l ) / 2 ^ / n \ [(at)"-rar+l + (at)« + l - ' V
r=0
+r(a^)n-rar_1] . (9.31.251)
But,
£ ( " ) [(atr-ra'"+1+(at)"+1-'-ar]
r=0 ^ '
= (at)"+1+ar,+1
r (n + 1)! ^ n\
+ £ ( a t^n
) +l-rar
n + l ( n + l - r)!r! (n — r)!r!
r=l
n!(r+n+l-r) t +1 .rqr
= (at)n+1 + a n + 1
¿
(n + 1 — r)!r!
r=1
n+ 1
= E " + 1
^(at)n+1-rar (9.31.252)
r=0
Also,
£ ( ; J r(«t)-'«-'
r=0
n—1
= S> (a tV
) »-l-(r-l)ar-l
r —1
r=l
n—1
n — 1 (a Tt )^ n - l - r a r
= » l (9.31.253)
r=0
Therefore,
:x : x =: x ^ : -\-n : x : . (9.31.254)
So, we have obtained that
xfn(x) = fn+i(x) + |/„_i(z) . (9.31.255)
However, according to the hint, the Hermite polynomials satisfy
xHn(x} — — Hn+i(x) -f- nHn-i(x) (9.31.256)
and we see immediately that
gn{x) = 2n f n ( x )
satisfy the same recursion relation as Hn(x). Therefore,
: xn : = 2~nHn(x) . (9.31.257)
Bibliography
[9.2] E.P. Wigner, Phys. Rev. 77, 711, (1950) .
[9.3] J. Schwinger, "On Angular Momentum". US Atomic Energy Commission

NYO - 3071, (1952)
BIBLIOGRAPHY 161
[9.4] A. Unsold, Ann. Phys., 82, 355, (1927).
[9.5]P. Carruthers and M.M. Nieto - Phase and Angle Variables in Quantum
Mechanics, Rev. Mod. Phys. 4 0 , 411, (1968).
C h a p t e r 10
Central Force P r o b l e m s
10.1 Isotropic SHO in T w o D i m e n s i o n s

Solve the isotropic simple harmonic oscillator problem in two dimensions in both
Cartesian and cylindrical coordinates.
Hint: Lz commutes with the Hamiltonian.
Solution
We are given
h + k f q ifi = x 2 +
~ 2 m \ ' y2 > f=pl+p 2
y (10.1.1)
a) Cartesian coordinates
H = HX + Hy [Hx,Hy] = 0 . (10.1.2)
Therefore, we simply have two one-dimensional simple harmonic oscillators and

the energy levels are
Eni,n2 — (î 4" 1 / 2 ) h u -(- (712 1/2)hw — (iii ~1" ^2 ~l" l)ft¿^ (10.1.3)
where, as always, ui2 = k / M . Also the eigenstates are (here we are setting
a2 = Muj/h)
(1014)
•
b) Cylindrical coordinates: The Schródinger equation reads
h2 ( d2 I d I d 1
+
2 m \dr* +^ + f *r'J ^ = Exp
- (10 L5)
Since [ L z ,H] = 0 we can now set
Tp = R(r) eim,fi (10.1.6)

10.1. ISOTROPIC SHO IN TWO DIMENSIONS 163
and w 2 = k/M . Then,
Now let
2 = ^ r 2 ) a = 2£ a n d fi(r) _ g(j/) ( 1 q a 8)
ñ /kj
Then,
r a g 0
t t + - t - - t - - » ' + ) = - (w.1.9)
d?/2 y dy y¿ J
We first examine the asymptotic behaviour of the solutions of this equation.
G ~ e±y /2
for i/—» oo . (10.1.10)
± m
G~y for y->0. (10.1.11)
Therefore we set
G=yme-y*'2F(y) (10.1.12)
and find the differential equation satisfied by F. After some simple algebra we
find
0 + ( ^ _ 2 s ) ^ + ( A _ 2 m _ 2 ) f = 0. (10.1.13)
Next, we look for a solution as a power series in y.
F = . (10.1.14)
n
Substituting in the equation for F and e q u a t i n g t h e coefficients of equal powers

of y we find
2(n + m + 1) — A / m i ic\
0,1+2
~ (n + 2)(2m + 2 + n) ' ( • • )
If this series does not terminate it leads to behaviour like exp(j/ 2 ) for large y and
is therefore unacceptable. Therefore the series must terminate and we require
that
2E
A = — = 2{n + m+ 1) . (10.1.16)
T h u s , the allowed energies are

E = (n + TTI + 1 )HUI . (10.1.17)
164 CHAPTER 10. CENTRAL FORCE PROBLEMS
10.2 Attractive Exponential Potential

Consider the attractive potential
V(r) = -V0e~ar
for / = 0 . This is one of the few solvable problems.
Hint: Change variables to u = e~ar. The resultant equation is Bessel's equation.
Discuss carefully the boundary conditions to be obeyed by
<t>{u) = R(r)
and find the equation that determines the energy eigenvalues.
Solution
The Schrodinger equation, for the s-wave (/ = 0), reads
(10.2.18)
We now set
2 mE 2mVo = /32
rip = R , • —K (10.2.19)
Then, the equation reads

d?R
+ P'2 e~ar R = K2R (10.2.20)
dr2
Let
aT 2
— e ! and F(u) = R(r) . (10.2.21)
Then, we find that the equation for F becomes
a2u2~^r + a2u-j— + a2u2F = k2F . (10.2.22)

du¿ du
This can be rewritten
(10.2.23)
This is Bessels equation of order K/a. The boundary conditions derive from the
boundary conditions on R. Thus,
R(0) = 0 and R(r) —> 0 for r —>• oo . (10.2.24)
This means that
F(2f3/a) = 0 and F ( 0 ) = 0. (10.2.25)

Therefore the solutions are
F(u) = AJK/a(u) (10.2.26)
with the quantization of energy given by
JK/a(2/3/a) = 0 . (10.2.27)
j0.3. REDUCTION OF THE TWO-BODY PROBLEM 165
jO.3 Reduction of t h e T w o - b o d y Problem

Show that the Hamiltonian
H = P - + + V(fi - r 2 )
2mi 2m2
reduces to the Hamiltonian
P2 p2
H = + — + K(r)
W
2M 2m
under the transformations
w + w i ( 1 0 3 28)
mi + m2
r = fi — f , (10.3.29)
M = mx + m 2 , (10.3.30)
and
mim2
m = . (10.3.31)
mi + m2
Solution
The Hamiltonian is
" = ék+é;+v<f>-r~>} • (103

-32»
We define
?
o "iñ + m2r2
it = — , M = mi + m 2 (10.3.33)
m
17111712
r = f i — r2 , m— (10.3.34)
mi + m 2
Now consider any differentiable function
^ ( n , ^ ) = F(xi,j/i,21;x2,y2,z2) . (10.3.35)
Then,
dF__dF_dx_ w d ^ _ m i d F OF
dxi dX dx\ dx dxx M dX dx
and
d2F __ /mi\2 d2F 2mi dd2FF d2F

1 = f ü ü r f ü2 L , . £ Z2 n( o 3 37)}
dx?? \ M / ax M SxdX dx '
, iu. uujyiTHAL FORCE PROBLEMS
Thus we find,
2 v f + v x v f + v f (10
= © * i ¿ r ' * * • -3-38)
Similarly we find
dF_ _ 5F 5X 5F ax m? 5F _ 0F not*»
ax2 ax dx2 dx dx2 M 8X dx
and
a2F (m2\2d2F 2m2 a2F a2F /in o â\
ax2 _
( M ) ax2 M dxdX +
dx2 ( • • )
So,
V 2 F =
-V-F + V
'F- (10-3-41)
Therefore,
—V? + —1—Vr^j F = — V y F + - ( — + — )VzFr

2mj 2m2 / 2m 2 \mj m2
< 10 " 2 '

So we see that
H + + v
= m ^ ^ - < , 0 3 - 43 >
10.4 Particle in a Spherical Potential Well

A particle is in a spherical potential well
— Vo for r < a
V(r) = |
0 for r > a
Find the transcendental equation which yields the energy eigenvalue for the
state with angular m o m e n t u m I.
What is the minimum degeneracy of this state?
If a proton and neutron are bound in an I = 0 state with an energy of 2.2
MeV, determine Vo given that o « 2 x 1 0 - 1 3 cm.
Solution
The radial wave equation becomes
h2 d2u 1(1 + 1 )h2

—ñ p¿ : H 5 — u — Volt = —Eu for r<a
zra ar ¿mrz
2 2
h du 1(1 + 1 )h2 , .
— -——— H -—u = —Eu for r > a . (10.4.44)
¿m dr¿ ¿mr¿
10.4. PARTICLE IN A SPHERICAL POTENTIAL WELL 167
Here we have w r i t t e n —E for t h e energy since we are looking for b o u n d states.

Defining as usual
£
, fc- = > (10.4.45)
the equations become
d?u 1(1 +1) 2 n c

—2- H 1
—u + A u = 0 for r < a
dr r
-Sj—-u — = 0 for r > a . (10.4.46)
dr J r2
The solutions are
u = A^/rZi+\/2(Kr) for r < a

u = B\/rZi+i/2(ikr) for r>a (10.4.47)
where Z¡+i/2 = Ji+1/2 or ^1+1/2- F° r r < a we must choose J¡+i/2 whereas

for r > a we need H^.îkr) = J¡+i/2(ikr) + iNi+i/2(ikr) = Ki+i/2(kr) since
this damps exponentially. Also at r = a both u and its first derivative are
continuous. Therefore, it suffices to match the logarithmic derivatives
dr [•y/rJ,+1/2{Kr)] = [yÂ; + 1 / 2 (A-r)]

(10.4.48)
y/rJi+i/2 (A'r) VrKi+i/2(kr)
This is the transcendental equation for the eigenvalue E.

Since this equation does not involve the magnetic quantum number m, the
minimum degeneracy of each level En¡¡ is 21 + 1 corresponding to the possible
m values m = 0, ± 1 , ± 2 , . . . , ±1. For / = 0 we do not need Bessel functions
since the differential equations are simply
d?u o
- 7 - 7 + A u = 0 for r < a
dr¿
— k2u = 0 for r > a . (10.4.49)
Thus, we get
u — A sin A'r r > a
u = B e~kr r > a . (10.4.50)
Matching the logarithmic derivatives at r = a we get
tan Ka = . (10.4.51)
Let
A z
T = - (10.4.52)
CHAPTER 10. CENTRAL FORCE PROBLEMS
Then we have
tan kaz = —z .
Substituting the values appropriate for a deuteron and remembering that we

have to use the reduced mass which is about 1 / 2 the proton mass we find that
ka = 0.46. Now a glance at a table of ± z versus tan kaz shows that we want z
somewhat greater than TT/2. This corresponds to z = ir/(2ka) = 3.41. Choosing
a series of slightly larger z values we get the following table. Therefore, to an
z tan kaz
3.6 -11.71
3.8 -5.58
3.9 -4.40
3.95 -3.97
accuracy of about 1% we have z = 3.95. Then, we get
V0 = E + Z2E = 1 6 . 6 £ = 36.5 MeV . (10.4.53)
10.5 Particle on Surface of a Cylinder

A particle is free to move on the surface of a circular cylinder of radius R. The
Laplacian in cylindrical coordinates is given by:
2
V - —2 — — — 2
dr r dr r2 dip2 <9z
Find the energy eigenvalues and eigenfunctions for this motion.
Solution
Since the particle has to move on the surface of the cylinder we have that r = R,
the radius of the cylinder. Thus, the Schrodinger equation reads
1 d2ip d2ip ,
+ -?r~ = -k2ip . (10.5.54)
R? dip2 dz 2
Here we have set

,, 2 ME
(10.5.55)
h2 '
Since Lz commutes with the Hamiltonian for a free particle we can write,
V> = e i m *>F(z) (10.5.56)
where F satisfies
d2F / , m2\
2
+ U 2 - —2 L P = 0 . (10.5.57)
dz ' V RJ
10.6. EXPECTATION VALUES: ELECTRON IN A H-ATOM
por a free particle the t o t a l energy
Li
E> (10.5.58)
- 2MR2
Therefore,
m 2
2 (10.5.59)
k2 > —
k
- R2
and the eigenfunctions are given by
2 2 2 2 2
v> = eimip A e x p ( i z \ / k — m / R ) + B e x p ( — i z y / k — m / R ? ) . (10.5.60)
10.6 E x p e c t a t i o n Values: Electron in a

H-at o m
An electron in the Coulomb field of a proton is in a state described by the ket:
J [ 4 | 1 , 0 , 0 ) + 3 | 2 , 1 , 1 ) - |2,1,0> + \ / l 0 | 2 , 1 , - 1 ) ]
where the labelling is \n,l,m). Find

a) The expectation value of the energy.
b) The expectation value of L2.
c) The expectation value of Lz.
Solution
We are given
| v > = l í4|l,0,0) + 3 | 2 , 1 , 1 ) - | 2 , 1 , 0 ) + vT0|2,1,-1) (10.6.61)

6 l
a) The energy expectation value of the electron, neglecting the motion of the
proton-electron center of mass, is given by
(H) = (¿¡,\H\1>)
1 1
2 2
1 6 9 + 1 1 0
y2+ ¿ ¿ + ¿ -mc a
36 2
1_ *) 2
mc" a (10.6.62)
24'
b)
(l2) = ^ [16 • 0 + 9 • 2 + 1 • 2 + 10 • 2] fr2 = y h2 . (10.6.63)
¿ [16 - 0 + 9 • i + 1 • o + 1 0 • ( - i ) ] h = - ~ h . (10.6.64)
10.7 Parity in Spherical Coordinates

Show that under a parity transformation
9 -> 9' = tt — 9
tp tp' = x' = — x = rsin 9' cos ip'

y — r s i n 9sin ip —> j/ = — y = r s i n # ' sin93'
z = rcos6 —t z' — —z — rcos6' . (10.7.65)
Therefore,
cos 9' = — cos 9 = cos(7r — 9) . (10.7.66)
Since 0 < 9 < n and 0 < 9' < 7r we therefore get that
(10.7.67)
Using this result we have that
cos ip' = — cos ip

sin<¿>' = — sin y . (10.7.68)
Therefore,
<p' = tp + x . (10.7.69)
10.8 Magnetic M o m e n t due to Orbital Motion

Consider an electron in the state \n,l,m) of a hydrogen atom.
a) Compute the current in the direction tp.
b) Compute the current flowing through an element of area normal to e v . Use
this result to compute the magnetic moment of the electron due to its orbital
motion.
Hint: The magnetic moment due to a current loop is the magnitude of the
current times the area enclosed by the current loop and points in a direction
given by the right hand rule in tracing the loop around the area.
10.9. SPHERICAL SQUARE WELL 171
Solution
a) The electric current in the direction e v due to an electron of charge —e and
mass M is given by
*, = — • • (10.8.70)
* 2iM r s i n # dip dip
For the state \n,l,m) the wavefunction is of the form
V'n,/,m(r,0^) = ñ n , i ( r ) P ¡ m ( c o s 0 ) e ' ^ . (10.8.71)
Therefore carrying out the required differentiation in (10.8.70) we see that
( 1 0 8 72)
•
b) Consider an element of area dA = dz dp normal to ev. Here we have written

x = pcos<p , y = ps\nip . (10.8.73)
Then we have
p=rsin0 , z = rcos9 (10.8.74)
so that
dzdp = rdrd0 . (10.8.75)
The element of current through this area is
dl = j<pdA = —tt-iti——t \^n,i,m(ñ\2 rdrdO . (10.8.76)

M rsint'
The area enclosed by the current loop is
7rp2 = 7rr2 sin 2 9
and the normal to this area points in the e z direction. Thus, the magnetic
moment due to the electron's orbital motion is
—eh
=
^ I T -méz J \rpn,i,m(r)\2 nr2 dr sin 9 d9
eh
~ 2 M ':me2 . (10.8.77)
Here we have used the fact t h a t the wavefunction is normalized to unity.
10.9 Spherical Square Well

For a particle in an s-state (I = 0) acted on by a spherically symmetric square
well potential
-Vo 0 < r < a
V •{ 0 r > a
a) Find the e q u a t i o n for t h e b o u n d s t a t e energies.
b) W h a t is the m i n i m u m s t r e n g t h for Vo in order t h a t at least one b o u n d s t a t e
exists?
Solution
a) The Hamiltonian is
H = ^ - + V(r) . (10.9.78)
2m
The eigenvalue equation for the I = 0 bound states now becomes
d2u ,2 n
—-2 + k u = 0 r < a
dr
^1-K2U = 0 r > a (10.9.79)

dr-
where
= = (10.9.80)
fi-1 ft*1
and
u = rip (10.9.81)
so that w(0) = 0. The solutions are
u = A sin kr r < a
u = B e~Kr r > a . (10.9.82)
Imposing the condition of continuity of u and its derivative at r = a by equating

the logarithmic derivatives we get the equation for the bound state eigenvalues
k
k cot ka — —K or tan&a = — . (10.9.83)
K
This completes part a).
b) For a solution to exist requires that the equation
tan ka • - - = -</ °, (10.9.84)

k y \E\
2j
should have a solution. This requires that
'•4»- py>~ c»9-85)

So, for l^l > 0 we need
n2 h2
- t s s ? • < 10 - 9 - 86 >
This result is also clearly the minimum condition since using the equal sign, by
taking the limit from above, yields the eigenvalue E = 0.
10.10. BINDING ENERGY AND POTENTIAL 173
10.10 Binding Energy and Potential

A spinless particle (mass m) moves in a short range central potential V(r). The
wavefunction describing the state of the particle is
r
where A is some constant and A < (3 are both positive constants.
a) What is the angular momentum of this particle?
b) What is the binding energy of this state?
c) What is the potential that produced this state?
Solution
a) Since the wavefunction has no angular dependence it follows that
L2V> = 0 (10.10.87)
so that 1 = 0.
b) To find the binding energy we look at the asymptotic behaviour of this
wavefunction for large r. This must be of the form
— exp j^—V— 2mE/h^ .
Since ¡3 > a we find that
for r oo . (10.10.88)
Therefore,
h2a2
E = (10.10.89)
2m
c) To find the potential that produced this state we simply stick the wavefunction
and the given energy into the Schrodinger equation and solve for the potential.
Thus, we find
h2 1 d2(r<p) h2 ,a2e ar
(32 e -Pr
= —z—A
2m r dr2 2m
= (E — V)(J> = A-
2 _2
/rev -ar _ p-pr
- V A- (10.10.90)
2m
Therefore,
h2 „ e~&r
v w = - _ ( ^ - ! ) _ £ _ (10.10.91)
17 4 CHAPTER 10. CENTRAL FORCE PROBLEMS
10.11 Generating Function: Laguerre

Polynomials
a) Use the integral representation for the Laguerre polynomials [10.1]
l { )=
" ' é¡l'~''í±ftt-dt
to derive the generating function
/(*.s) = £ e x p [
r?75 (10.11.92)
n=0
for these polynomials. The contour in the integral is a circle about the origin.
b) By differentiating (10.11.92) / times and using the definition
dl
Lln.l{x) = (-\)l—[Ln{x) (10.11.93)
obtain the generating formula
^Jn(x) n 1 r —sx
£ «p1tt71 • (10.11.94)
Solution
a) By changing the integration variable from t to — xt in (10.11.92) we get
/(«,.) = ¿ / e - ¿ ( . l l í ) " £ . (10.11.95)
We now take |s| < 1 to insure that for sufficiently large t the series converges.
Thus,
1
tt \ I t 1 dt
f { x s )
' ~ 2;rife 1-s(t-x)/t t
dt
= JL/e«
2iri J t — s(t — x)
et
—T-—— / : jr, r • (10.11.96)
2ni
ml 1 —
— ss jJ tt +
+ sx/(l
sx/(l —
— s)
s)
The integrand has a simple pole at
SX
t = —
1—s
so we immediately get the desired result
n
f ( x , s) = J2 ^ r ~ s = j 1 - exP["r~~] • (10.11.97)
' ni 1—s 1—s
10.12. NORMALIZATION OF HYDROGEN WAVEFUNCTION 175
t,) If we differentiate formula (10.11.97) I times we get
< i o i L 9 8 )
n—l
Now using the definition of Lln_¡(x), namely,
£»-<(*) = ( - l ) ' % r * (10.11.99)
we see that we have
( i o n i o o )
n=l
10.12 Normalization of Hydrogen Wavefunction

a) Use the generating function for the associated Laguerre polynomials
n=0
(see problem 10.11) to evaluate the integral

ROO
/ e - ' x P L ^ W L W ^ W d x
Jo
fcr p = 21 + 1 as well as p = 21 + 2.
The value p = 21 + 2 yields the normalization for a hydrogenic wavefunction.
Solution
a) We first rewrite the generating formula in the form
1
£ - t i w ' " - ' - = « " f c l <1012101>
n=( +1
and then use this given generating formula twice. This gives
cn-l-lin -l-l r°°

roo
£i ,V (»(+i /)\!(«' +
n,n'—l+1
+2 /)! Í
y „ r 2,+2 00
p
e
~'^ln-'-llx)ln#-l(*)
i - .sí
d x
, i - s ; v 1^7
i - < /J X
io - - P [
l - ( T T
(i-»)(i-f).
poo
= (1 - s ) p _ 2 i _ 1 ( l - i ) p - 2 i - 1 ( l - st)~ip+1) / zp e~z dz
Jo
= (1 — s ) p _ 2 ' _ 1 ( l — í ) p - 2 i - 1 ( l — s i ) ~ ( p + 1 ' p ! . (10.12.102)
r rífjtí I .hMS
If p — 21 + 1 then the right hand side reduces to

(21 + 1)! _ [n + 21 +
2i+2
(1 - s i ) TV.
n=0
= £ (n-^-l)!**0""1"1 (10.12.103)
n=l+1
where we have used the binomial expansion
= t • (10.12.104)
For this to equal the left hand side for all values of s and t we clearly need
n = n' so that the left hand side reduces to
00
I i°°
E [(re -f Z)!]2 J0 e
~X x2l+1 L
n-i-1 ( * ) £ * - / - ! ( * ) d* • (10.12.105)
Hence, we find
jo e-xx 2 l + l
l2j+^1(x)lyjl_1(x)dx=-^±^s n t n . . (10.12.106)
If we now put p = 21 + 2 and n = n' we find that the right hand side yields
JiffS = (2/ + 2)!(1 — s ) ( l — i ) ( l — st)~2l+3
+ + 2 ) !
= ( i - « - . + » 1 ) e ' " ! ' , c r
n\
n=0
( l _ s _ í + s/) • (10.12.107)
n=i+1
On the other hand, the left hand side reads

n
v f (str-'-v s(st) -'-1)
LHS =
2 +
¿í
n . n ' z r +i +1l
l [(" + 0 ' ] (n + I + l)[(n + /)!] 2 Ó "'"' +1
¿(st)"'-'-1)
+ Jn
(n' + i + l ) [ ( n ' + /)!]2 '
poo
/X • (10.12.108)
Jo
Now, the right hand side may be further rewritten as
RHS =
£ {f^TTT + («")-')V-
n,n'=l-1-1
l'zzl + 1 ^ ^ '
(n + l + 1)! n 1
1 )T!s^( s t Í„
° n1'„" ' + 1
(«' + / + ! ) !
t(stn ^ • ' v '0.12.109)
1)!
lU.iú. i\RU\IVIRJTLD~ ¡XCJL,J\11y
Choosing the diagonal terms n = n' we get
OO / \n-l-1 r°°
/»C
y - l£í) í e -x 21+2 L2l + l ,x)L2í+X í x ) d x
^ [ ( n + lVVJo n-í-lV ; n-l-lV ;
n=/+1
oo
(n + l + 1)! (n +i /)!
7\t 1
= E (n — l — 1)! (n — / — 2)!
(sí)*n —/ — 1
n=/+1
£
n=l +1
(n + 0 ¡ o„C„j\n-/-l
2n(st)
(n — / — 1)!
(10.12.110)
This immediately yields that
jf «»+' ¿ l + . ' - . m g , ' - . ( » ) j» = i), • (10.12.111)
Hence, we obtain the desired normalization for the hydrogenic wavefunctions.
2Z \ (n — I — 1)
^ = \ ^
nao/ 2n[(n + /)!] 3
(10.12.112)
10.13 Kramers' Relation

a) Use the radial equation for the hydrogenic atom
1(1+1) j_ 1 1
Ki + -Ki O h~ ñn; = 0
to derive the recursion relation
1
- (2p + l ) ^ " ) + £ [(2/ + l ) 2 - p 2 ] a 2 ^ " 2 ) = 0 p > 1
nz 4
This result is also known as the Pasternach relation [10.2]. Here,

roo
(rp) = / \Rni\2rp+2dr .
Jo
b) Use this result and the results of problem 10.12 to calculate
(r) , ( r _ 1 ) , and (r 2 ) .
Solution
a) If we write the radial equation in the form
'/(/ + 1) 2 1 1
K , = -~Ki + Rnl (10.13.113)
1Y8 CHAPTER 10. CENTRAL FORCE PROBLEMS
and multiply it by rp+2Rn¡ and integrate over all r we get (after dropping the
subscripts on Rn¡)
roo roo
/ R"rp+2Rdr = -2 / R'rp+1Rdr + /(/ + l)(rp~2)
Jo Jo
+ (i
- ¿ 5 " • °' 13114 )
If we integrate the left hand side of this equation by parts we obtain
roo r°° r0 0
/ R"rp+2Rdr=-(p + 2)/ R'rp+1Rdr- R'rp+2R'dr. (10.13.115)
Jo Jo Jo
After another integration by parts we further find
pOO r OO
/ R"rp+2Rdr = (p + 2) R[(p + l)rpR + rp+1 R'] dr
Jo Jo
r OO
+ / R[(p + 2)rp+1R'+ rp+2R"] dr . (10.13.116)
Jo
Hence, it follows that
r^ 1 1
J^ R'rp+1Rdr = - ^ - ( r p
~ 2
} . (10.13.117)
We now multiply the radial equation (10.13.113) by rp+sR' and again integrate
over all r to get
[ R"rp+3R'dr = I [°° r»+3d(R')2 = f ° ° R'r**2 R'dr

Jo ¿Jo ¿ Jo
roo poo
= - 2 / R'rp+2R'dr + l(l + 1)/ R'rp+1Rdr
Jo Jo
2 roo 1 roo
- - R'rp+2Rdr+^— R'rp+3Rdr
a Jo « a Jo
= -2 í R'rp+2R'dr-l(l + l)^l(r p
- 2
)
Jo 2
+
• <1013-118>
Therefore,
rOO
/
p+2
R r R' dr
'
Jo
= l(l+l)P±l(r»-2)--P±l( r
p
- l ) + - 2^2 P ± ^ ( r p
) .(10.13.119)
p- 1 ap —1 n a p- 1
Combining the results of (10.13.114), (10.13.115), and (10.13.119) we get
~ ^ ( p + 2)(r»-2) - 1(1 + l)ÍP±Íl(rP-2)

¿ P—1
10.13. KRAMERS' RELATION 179
+ 2p+2 ! l_P±A/rP\
ap —1 n2a2 p — 1
= (P+i)(r"-2) + l(l+l)(rp-2)
- - ( r " - 1 ) + ~ ¿^¿( r p ) • (10.13.120)
a na
Simplifying this result yields the desired expression
^ ¿V ) - (2p + l M r - P " 1 ) + V- [(2/ + l ) 2 - p2] a2(r»-2) . (10.13.121)

n 4
Throughout the calculation we have dropped the terms resulting from the inte-
gration by parts. This required that p > 1. Hence, that condition.
For an alternate derivation that relies on the generalized Hellmann-Feynman
theorem see the article by Balasubramanian [10.3].
b) The radial wavefunctions are
(!0.13.122)
Also, from problem 10.11 we have that

r°° /Or\ 21+2
+
•>
e 2r/n
Jo ~ " (Ln-l-i(2r/na)) d(2r/na)
2n[(n + /)!]
(10.13.123)
(n — I — 1)!
as well as
f
2 / + 1
,'9r\ ?
e 2r/n
" — ) {Ln-i-i(2rlna)) d(2r/na)
na J
[(" + *)']
(n — I — 1)! (10.13.124)
J0
Equation (10.13.123) is just the normalization of the wavefunction and states
that
(r°)=l (10.13.125)
while equation (10.13.124) states that
1
( r ~ ) =z ~ . (10.13.126)
na
Now, using these results and (10.13.121) with p = 1 we get
2 2
^¿( r ) = 3 a - i [ ( 2 / + l ) - lJ ] a ^ . (10.13.127)
n 4 n2a
After simplification this reads
(r) = [3n2 -1(1+1)] . (10.13.128)

Similarly, putting p = 2 in (10.13.121) we get
A ( r 2 ) = | a 2 [3n 2 - /(/ + 1)] - ^ [(2/ + l ) 2 - 4 ] a 2 . (10.13.129)

TI Z Z
Simplification now yields

2
(r 2 ) = y f i 2 [ 5 n 2 + 1 - 3 / ( / + 1)] . (10.13.130)
10.14 Quantum Mechanical Virial T h e o r e m

a) Prove the quantum mechanical Virial Theorem for stationary states, namely
2(T) = ( f - V V ) . (10.14.131)
ihjt(A) = ([A,H}) (10.14.132)
where A is an operator, corresponding to the radial momentum, and H is the

Hamiltonian.
b) Use the result of part a) to show that for a potential of the form
V(r) = Vo rn
this yields
2(T) = n(V) .
Solution
a) The proof of the first part is similar to the proof used in classical mechanics.
For a stationary state, the expectation value of r • p, the radial component of
momentum, is independent of t. Therefore, using (10.14.132) we have that
^(r p) = 0 = . (10.14.133)
But,
Px + + Vz
[f p,H] = [xpx + ypy + zpz, ^ + v(f)]
ih . 2 „ 2 / dV 8V dV
= - ( p r + p y + p 1 ) - t f t ^ _ + y _ + z--
= ih2T-ihr- VV . (10.14.134)
Therefore,
2(T) = (r • V V ) (10.14.135)
10.15. EHRENFEST THEOREM FOR ANGULAR MOMENTUM 181
b) If
V(r) = V0rn (10.14.136)
then,
f• V V ( f ) = nV(r) . (10.14.137)
Therefore the desired result i m m e d i a t e l y follows and
2 ( T ) = n(V) . (10.14.138)
10.15 Ehrenfest T h e o r e m for Angular

Momentum
a) Show that for a particle evolving according to the Hamiltonian
1 + v(f)
the rate of change of the expectation value of the angular momentum operator
L = r x p
is given by
¿(í> = - ( - v v > .
b) Under what conditions is this equation exactly the classical equation of mo-
tion for L ?
Solution
a) The result follows immediately if we use the equation for the time evolution
of the expectation value of an operator A.
^{A) = ±([A,H]). (10.15.139)
Thus,
~(L) = jr([vxp1H}) . (10.15.140)
But,
[rxp,H] = fx[p,H] + [r, H] x p

= —ihr x VV . (10.15.141)
Therefore,
~(L) = —(fx VV) . (10.15.142)

i mi iu. i_y&i\ L KAL FORCE PROBLEMS
b) For this result to be the same as the classical equation requires that, if
F(r) = - V V
then,
( f x F ) = (r)xF((f)) . (10.15.143)
This is possible only if F(r) is a constant vector, or parallel to r.
10.16 Angular M o m e n t u m of a Two-Particle

System
For a two-particle system with coordinates £ 1 , x 2 introduce the centre of mass
and relative coordinates
miXi+m2X2 _
A = — , X = X i - X 2
M
where
m\m2
M = mi + m2 and m -
M
are the total and reduced mass respectively.
Show that the total angular momentum of the two-particle system is
L = X x P + x x p
where
P = - V x , p=-Vx
I I
are the centre of mass momentum and relative momentum respectively. Give
an interpretation of this result.
Solution
Comment: This calculation is the same as the corresponding classical one if we
maintain the order of the operators.
The total angular momentum is given by
L = xi x pi + x2 x p2 . (10.16.144)
But,
,7 mo _
xi = X +—x
M
x2 = X - ^ - x (10.16.145)
M
10.17. HULTHÉN POTENTIAL: GROUND STATE 183
and
P - P1+P2
P = j f P i - ^P2 • (10.16.146)
Inverting these two equations we find

m i
5 , -
« = ~MP +P
P'i = Jj-P'P- (10.16.147)
Combining these results we now get
L = Xi x pi + X2 x pi
m
i ,7 ^ ,7 - mim2 _ - m2 _
= — X x P + X x p + — — i x P l - x x p
m M M
m2 - - - toim2 _ - mi ^
+ x P - I x p ——x x P -\——x x p
M M M
= X x P + x x p . (10.16.148)
This result states that the total angular m o m e n t u m of a two-particle system

consists of the angular momentum of the centre of mass of the system plus the
angular m o m e n t u m of the relative motion.
10.17 H u l t h é n Potential: Ground State

Find the ground state energy and wavefunction for the Hulthén potential
Vne~ar
V(r) = - 1 _ e _ t t r • (10.17.149)
Hint: Try a solution of the form
$ = -e~iT (1 -e~aT\ . (10.17.150)

r
Give a reason why this is indeed the ground state.
Solution
The ground state is an s-state (I = 0) and so setting the radial wavefunction
R(r) equal to 1 / r u(r) we get the Schródinger equation
h2 d2u Vo e~ar
2
u=Eu . (10.17.151)
2m dr 1 — e~ar
If we set
2mV 2 2 m E
ffi = 2
° , K = - 2
(10.17.152)
h h
IV. ^niVTKAL FORCE PROBLEMS
the equation reduces to
d2u 2 /32 e~ar 1
K u = 0 . (10.17.153)
d ^ ~ 1 — e~ ar
We now insert the suggested solution
u = Ae"yr( l-e~ar) (10.17.154)
into (10.17.153). This equation reduces to
[7 2 - k 2 ] e ,r
+ [ - ( 7 + q ) 2 + K2 + /3 2 ] e ~ ' 7 + a ' r = 0 . (10.17.155)
Thus,
7 2 = K2 or 7 = K (10.17.156)
and
(7 + a ) 2 = K2 + f32 . (10.17.157)
Solving this equation for K we get
P2 a
K (10.17.158)
2a 2
Therefore,
mVo a
77 == kK ==
i ^ - 2 <1017159>
and
E = h2K2 h2 (mVn a
- ^ - =-2^{T^-2) ' I101"60»
This is an exact solution of the 1 = 0 Schródinger equation and does not have
any nodes. Therefore, it is indeed the ground state solution.
10.18 Hydrogenic A t o m in T w o Dimensions

Solve the hydrogenic atom for its eigenvalues and eigenfunctions in two dimen-
sions. [10.4]
Solution
The Hamiltonian for the hydrogenic atom is
10.18. HYDROGENIC ATOM IN TWO DIMENSIONS 185
In cylindrical coordinates, and in two dimensions, this leads to the Schródinger

equation
h2 d2ip 1 dip 1 d2ip Ze2

ip = Exp (10.18.162)
2m dr2 r dr r2 dip2 r
The equation separates into
1 tmtp
ip{r,(p) = R{r) m = 0, ± 1 , ± 2 , (10.18.163)
y/2ñ
where the radial function satisfies
h2 d2 R 1 dR m2 Ze2
R= ER (10.18.164)
2m [ dr 2
r dr r 2
J r
It is convenient to introduce new variables

h2
2me2
h2 E
e =
2me 2 e 2
r
(r) = ~i=y{ x ) • (10.18.165)
v*
The radial equation now becomes
d2y ( m 2 — 1/4) Z
-7-^ 2 2 / + - j / + ey = 0 . (10.18.166)
¿ ¿
ax x x
The asymptotic behaviour near x = 0 yields the solutions
y^x±(m±l/2) as £ 0 . (10.18.167)
The only acceptable (square integrable) solution is therefore the one with the
asymptotic behaviour £ l m l + 1 / 2 .
For i - + o o w e get that y satisfies
d2
y
-7—5- + ey « 0 as x oo . (10.18.168)
dxz
In this case the acceptable solution is of the form
y —> e~y^x as x —>• oo . (10.18.169)
So, we look for a solution of the form
y= x f(x) . (10.18.170)
fl
Substituting this into (10.18.166) we get the equation for /
dx2 +
2|m| + 1
— 2\/—e £ + i [Z - (2|m| + 1 ) > / = ! ] / = 0. (10.18.171)
As usual we now look for a series solution
y = Ya>=xk (10.18.172)
k
and substitute this into (10.18.171). The resulting equation is
yâfc {[{k + 1)(& + 2|m| + l)ar f e _ 2

k
- [(2(|m| + *) + l ) - / ^ e - Z ] a : ' ! - 1 } = 0 . (10.18.173)
Therefore we obtain the recursion relation

[2(|m| + k) + l]y/ê - Z
a
*+1 " (*+l)(i + 2 M + l "l • (10.18.174)
Unless this series terminates it leads to exponential growth yielding a non square
integrable wavefunction. Therefore we need that the series terminates. This is
acomplished by setting the numerator equal to zero. The result is
^=5(wfi)TT I10'18175)
or
*= - f Í (10.18.176)
where
N = 2(\m\ + k) + l = 1, 3, 5 (10.18.177)
Thus the energy eigenvalues are
2 me 4 Z2e2 1
E=-—e = 10.18.178
ao (n + 1/2) 2
where
h2
do = J is the Bohr radius (10.18.179)
me
and
n = |m| + * = 1, 2, 3, . . . • (10.18.180)
10.19 Runge-Lenz Vector: Constant of the

Motion
The classical Kepler (attractive 1 /r potential) problem has closed orbits. Classi-
cally, this means that the vector pointing from the origin along the semi-major
10.19. RUNGE-LENZ VECTOR: CONSTANT OF THE MOTION 187
axis of the orbit is a constant of the motion. This is called the Runge-Lenz
vector. In quantum mechanics the Runge-Lenz vector is given by
2
R= — (px L-L xp)-e - . (10.19.181)
2m r
Given the hydrogen Hamiltonian
-*2 2
H = ( 1 0 . 1 9 . 1 8 2 )
2m r
1) Verify that R is formally hermitian.

2) Show that R is a constant of the motion. That is, show that
[R, H) = 0 .
Solution
1) It is convenient to work with individual components of the Runge-Lenz vector.
This is made even simpler by using the Levi-Civita antisymmetric tensor symbol
{ 1 if ijk is an even permutation of 123

— 1 if ijk is an odd permutation of 123
0 if any two of ijk are equal

. (10.19.183)
Now,
fit _ J _ [(pr x £ ) t _ x p)t"| _ e 2Ü (10.19.184)

2m L J r
Also,
(p x L)i = CijkPjLk (10.19.185)
and
(p x L)\ = eijkLkPj = ~(L x p)i . (10.19.186)
Similarly,
(L x p)i = eijkLjPk (10.19.187)
and
(Lxp)] = eijkpkLj = -(pxL)i . (10.19.188)

Therefore,
t
ñ = -í- \-(L x p) + (p x L)1 - e 2 - = R . (10.19.189)
2m L J
2) To show that R is a constant of the motion we consider the commutator
"l nXi p2 e2
[Ri, H] = ijkiPjLk - LjPk) - e2y , — - -
= (b>j> - Aib*, - [ y , p2] J • (10.19.190)
Here we have used the fact that
bj.p2] = 0
[Lk,p2] = 0
[Lk,r] = 0 . (10.19.191)
Continuing, we find
6^ f 2 2 x¿
[Ri,H] = ih— {Ujk^XiXjPj - -Pi - 2ih^
2 €
2 x x
.2 txih 1
+ - P < - H ^ j P j ' + ^ j
= 0 . (10.19.192)
10.20 Runge-Leiiz Vector: Hydrogen Spectrum

In the previous problem (problem 10.19) we introduced the Runge-Lenz vector
and found that it is a constant of the motion. The "accidental" degeneracy
of the hydrogen atom is due to the fact that this vector along with angular
momentum is conserved. Use the Runge-Lenz vector
-* 1 —* —* V
R = -— (pxL — L x p) — e 2 - (10.20.193)
2m r
together with the hydrogen Hamiltonian
—2 e
2
H = ^ - - (10.20.194)
2m r
to prove the following steps.
1)
R • L = L • R —0 (10.20.195)
2)
— ih í i j k Lk (10.20.196)
3)
[Ri i 7 ] — ihtjjk Rk (10.20.197)

10.20. RUNGE-LENZ VECTOR: HYDROGEN SPECTRUM 189
4)
2H
R2 _ e4 + (L2 + , (10.20.198)
By restricting yourself to the space of bound states (energy E < 0) define
= (10.20.199)
and
5) show that the components of K satisfy the commutation rules
[I<i,Lj] = iheijkLk (10.20.200)
as well as that
[Ki, Lj] = iheijkKk . (10.20.201)
Next define the two operators
M=l-(L + K) , N=^[L-K) (10.20.202)
and verify that they both satisfy angular momentum commutation relations as
well as that they commute with each other. Finally,
6) use all these results to deduce the hydrogen atom energy levels
E
" = - 5 ? T - zi (10.20.203)
Ih- n
as well as the degeneracy of these levels.
Solution
1) In preparation for subsequent computations we list two properties of the
Levi-Civita symbol e¡jk as well as some commutators. Throughout we assume
the Einstein convention that repeated indices are to be summed over.
tijkCiim = 8j¡S k m — Sj m Si k (10.20.204)
e i j k e i j m = 2S k m . (10.20.205)
[p¿, Lj] = ih(ijkpk . (10.20.206)
Also,
2 2
\p¡, xjr] = -ihr'~3Xi . (10.20.207)
We are now ready to prove the first result. To do this we first compute
r • L = xi€ijkxjpk = 0 (10.20.208)
as well as
L " V — ( iJ k Xj Pk J'i
— (¡j k ( Xj X1 l*k ihXjSik)
= 0 . (10.20.209)
Similarly,
L • p = CijkXjPkPi = 0 (10.20.210)
and, with a little more algebra
=
p •L £ijkPi%jPk
— ¿ij k ( Pi Pk ® hpk &ij )
= 0 . (10.20.211)
Therefore,
R L = (p x L — L x p) • L
2m
— 2tt2 ^ ^ ^ ^ ^
— 2 m ^*3 k ^1
— 7: €ijkLj{LiPk ih^kilPl)
Zm
ih
= —2sjiljpi = 0 (10.20.212)
where we have used that
etjkPjLkLi = p j L j = 0 . (10.20.213)
Similarly,
L •R — -^—£jjkL¡(pjLk LjPk)
= ~UjkLiPjLk = 0 . (10.20.214)
2) To compute the commutators between the components of R we use our results

from above. Then,
[ i Rj] — ^[îkl(PkLl LkPl) , tjmniPmLn LrnpJl )]
~ ^ {eilk^llk ~ l l p k
' 7"]
- W — , PiLm - L,pm]\ • (10.20.215)
To evaluate this we need several commutators. These are:
[pkL¡ , PmLn] — ifo{tknqPmPqLl — ClmqPkPqLn + QnijPfcPmLq)

[LkPl , Lmpn] = ifr(tlmq LkPqPn H" CkmqLqPnPl CknqLmPqPi)
[LkPl j PmLri\ — Ífo(tlnqLkPmPq f-kmqPqLnPl ~f* CknqPmLqPl)
\pkLl , LmPn\ ~ Í^{tlmqPkLqPn "1" tkmqPnPqLl f-lnqPkLmPq)
(10.20.216)
as well as
ibí.,2-] =
[2 + . ,10.20.217)
Putting all this together we get that
. f P2 2e2 |
[-fíj, Rjj — ih < Cijk^^Lk îjkLk |
9 tj
= ih eijk
J
Lk . (10.20.218)
m
3) The next result is
1 e2
[Ri, Lj] — - Cikl\Pk Ll LkPl i Lj] [Xj , Lj\
Zm r
— 2mîkl{tljmPkLjn ~1~ CkjmPmLl
CkjmLmPl CljmLkPm}
_ ife..
í ít
' z?mc
r
= itx
2m îm^jk)(PkLm LkPm)
(îj&lm &im&jl) (pm LmPl) }
_ i fa.. e2^m
r
1 — f
X
— 2 m [ Pj Li + PiLj — LiPj + ¿jP¡] ijme ^ }
= iheijmRm . (10.20.219)
4) Now we compute
•R 2 = e
¡ f c í ( P f c ¿ / - ¿fcPi) - e2 — }
{ 2m£'k'^Pk^' ~ ^kp'^
1
{^km^ln 3kn^ml){PkLl LkPl){P m Ln LmPn)
4m 2
2
c X' X'
-—eiki[{pkLi - LkPi)^ + y (PkLi - Lkpi)]
+ e4
— 4m2 (pmln pn^m "1" -^'npm)(pm^'n ^mpn)
C r/ ^ X¿ Xi
- 7r~€iki[(PkLi - Lkpi) h — (;VkU ~ LkPi)]
2m r r
+ A
e . (10.20.220)
But,
PmLnPmLn — Pm[Pm Ln i^mnqPq) Ln

2 r2
= p*L
Pm Ln L-m p„ = (Lnpm + i He
mn qPq) LmPn
— ihtmnqPqiPn
2
Lm nmrPr)
= - 2 típ
Pn LmPm Ln — 0
pnLmLmpn = (Lmpn + ihen
m qPq)L m Pn
— l m [ l m p n i h € n r n q P q s ) p n -j- Íhcnrnq(LmPq Ífl€qmrPr)Pr
= L2p2 + 2 h2p2 = P 2 L 2 + 2ft 2 p 2
LmPnPm Ln — LmPn{LnPm ~t" i^mnqPq) —0
LmPnLmPn — Lfn^L-mPn ih€nm qPq)Pn
= L2p2 = p2L2
LnPmPm Ln — L"p" ~ p L"
Also, n p m l m p n
l = 0 . (10.20.221)
1 1 Xk
£iklPk~L[Xi — £ikl{~Pk ~l~ ifo ó i h t i l m X m )
y» jiO
/I r *fe i fc2 xk%n
— £ikl\ ^%ipk ^n€ii m pkx m -f- a¿ í¿/m ^3
1 ñ2
= —L2 -f 2 i hp • r — 2 —
r r
1 r 2 o-fc- - ,
= - I +2ifir-p + 4—
r r
X' 1
tikiLkPi— = tikiLk ~{xipi — ihSu)
r r
— ~îklLkXiPl
--L2 =
r
%i j 1 r2
tikl—PkU - -L
r r
X' i1
€% kl — LkPl — ~~tiklxi{PlLk ihtlkmPm)
r v
- -~L2 + 2 i h r p . (10.20.222)
r
R2 = e
4
+ f ^ - 2 — ) (l2+ft2)
\mz mr J
= e4 +—(L2 + h2) . (10.20.223)

m
We now restrict ourselves to the subspace of bound states of the Hamiltonian
and define
K = J - ^ R . (10.20.224)
5) It now follows immediately from the commutation relations satisfied by the

components of R that
[I<i, Kj] = ihujkLk (10.20.225)
so that K satisfies angular momentum commutation relations. Furthermore

from the commutators between i?, and Lj it also follows immediately that
[Ki,Lj] = iheijkKk • (10.20.226)
Next, we define the two operators
M=^(L + K) , N = ^ ( L - K ) . (10.20.227)
It then follows that
[Mi, Mj] = — {[Li, Lj] + [Li, Aj] + [A,-, Lj] + [A¿, Aj])
ih
— (Lk + Kk + a k + Lk)
= iheijkMk . (10.20.228)
Similarly,
[Ni, Nj] = i ([Li, Lj] - [Li, Kj] - [I<i, Lj] + [A,-, Kj])
ih
— ~êijk (Lk — Kk — Kk + Lk)
= iheiikNk . (10.20.229)
CHAPTER 10. CENTRAL FORCE PROBLEMS
Also,
[Ni, Mj] = — ([¿i, Lj] + [Li, Kj] — [Kt, Lj] — [A,-, A'j])
= — fijfc (¿fc + Kk — Kk — Lk)

= 0 . (10.20.230)
2
This means we can simultaneously diagonalize M and N 2. Their respective
eigenvalues are
h 2 n ( n + 1) and /i2i/(i/+l) p , u = 0,1/2,1,3/2,... .
6) To get the hydrogen atom spectrum we now use (10.20.223)
R2 = e4+ — (L2+h2) . (10.20.231)

m
In terms of K2 this reads
-—K2 = e4 + —(L2 + h2). (10.20.232)

m m
Solving for H we find
777,6^
( 1 2 0
" = + ' ° -233'
But,
A'2 + L2 = 2 ( M 2 + N2) . (10.20.234)
On the other hand, we also have that
R L = L R=K L = L K = 0 . (10.20.235)
Therefore,
M2-N2 . (10.20.236)
So, we can write the spectrum of the hydrogen Hamiltonian as
H = -
2h2[4p(fi + 1) + 1]
(10.20.237)
2/i 2 (2/z + l ) 2 '
But,
2// + 1 = 1 , 2 , 3 , 4 , . . . . (10.20.238)
So, using more conventional notation we call 2p + l = n and find that the energy
levels for the hydrogen atom are given by the Bohr formula
= " = 1,2,3,4,... . (10.20.239)

2 n¿n¿
BIBLIOGRAPHY 195
The degeneracy is also easy to count since for a given value of n = 2 / i + 1 = 2v+1
we have
(2fi + l)(2u + 1) — n2
different states with the same eigenvalue. Therefore, including the spin, the
degeneracy is 2 n 2 .
It was by a technique of this sort that Pauli [10.5] first succeeded to use
matrix mechanics to find the spectrum of the hydrogen atom before Schródinger
came up with his result.
Bibliography
entific Publishing Co. Pte. Ltd., section 10.6.1, (2002) .
[10.2] S. Pasternach, Proc. Natl. Acad. Sci. USA, 23, 91 (1937). ibid 23, 250,
(1937).
[10.3] S. Balasubramanian, Am. J. Phys. 68, 959, (2000)
[10.4] B. Zaslow and M.E. Zander, Am. J. Phys. 35, 1118, (1967).
[10.5] W. Pauli, Zeitschrift f. Phys. 36, 336, (1926).

Chapter 11
Transformation T h e o r y
11.1 Fourier Transform of Hermite Functions

Verify the equation
/ OO
dx e~'pxu„(x) = in un(k)
-oo
where
w =
* " m
and Hn(x) is a Hermite polynomial. That is, compute the Fourier transform of
the Hermite functions.
Solution
Let {u n (ar)} be the set of normalized Hermite functions with k/(huj) = 1. So
that
( 1 U
"
and
OO
/
e~i'"un(x)dx . (11.1.2)
•OO
Since the Hermite functions damp very rapidly at ± o o we can integrate by parts
and use the following general results.
r OO *00 fip-ip* f00
/ e-'P*-f-dx=- f(x)— dx = ip e—"f(z)dx.( 11.1.3)
J— OO dx J-na dx J-OO
11.2. SCHRÓDINGER EQUATION IN MOMENTUM SPACE 197
Also,
oo J roo
/
e~i'"xf(x)dx = i-=- e-'r*f(x)dx. (11.1.4)
Thus, dp J-co
Vp,n
^/ñl \dp J tt1/4 J-oo
2
- h r ( ± r / í a - p v - l e ^
1 0
- u p v *1/4
= ( - i ) B «„(*>) - (11.1.5)
11.2 Schrodinger Equation in M o m e n t u m Space

Transform the Schrodinger equation
fi2
V2ip(r) + V(f)ip(r) = Etp(r) (11.2.6)
2m
to momentum space.
Solution
It is convenient to first rewrite the Schrodinger equation in the form
( V 2 + k2)iP(r) = U(r)ip(f) (11.2.7)
where
, u( ? , = 2Jnm . (11.2.8)
We then define the momentum space wavefunction <¡>(p) by
ip r
<¡>{p) = J e " "i>(r) d3r . (11.2.9)
The inverse of this relation is
m = J d s m ^ ' • (n.2.10)
To proceed we take the Fouier transform of (11.2.7). Thus,
J d3r(V2 + k2)iP(P)eif ?
= J d3rU(r}rp(f)eiP f
. (11.2.11)
The left hand side may be integrated by parts twice to become
J d3r ( - p 2 + k2)ip(r) e'P-r~ = ( - p 2 + k2)<f>(p) . (11.2.12)

iy» CHAPTER 11. TRANSFORMATION THEORY
To simplify the right hand side we replace ip(r) by its expression in terms of
<t>(p). Then,
J d3rU{r)xl>(r)eif f
= J d3rU{r)eip f
J d3q 4>(q)
=
(¿F / d3qHql / U{rleÍ{P'~ql ' r
=
( ¿ ) 3 / A ^ Í P - í X í ) (11.2.13)
where
Ü(p-q) = j d3rU(f)ei^-^j' (11.2.14)
is the Fourier transform oiU(r) . Thus, the Schrodinger equation in momentum

space reads
{-p2 + k2)<j>(p) = J d3q Ü(p— q)4>(q) • (11.2.15)
11.3 H e i s e n b e r g E q u a t i o n for a Free Particle

Find a: and solve the Heisenberg equation of motion for a particle with a Hamil-
tonian
T n 1 r* Yi
»=¥-•
2m
If at t = 0 the particle is in the state |0, 0, 0) of a harmonic oscillator basis, find
(r2) as a function of t .
Solution
We are given the Hamiltonian
*=£• <il3i6>
The corresponding Heisenberg equations of motion are:
¡»f = k*] = o
ih^ = [r,H] = - p . (11.3.17)

at m
Therefore p = p0 where po is the momentum operator in the Schrodinger picture
p = -ihV (11.3.18)
11.4. DIRAC PICTURE FOR DISPLACED SHO 199
and
- P° 4 , - (11.3.19)
r —— t + x
m
where x is the position operator in the Schródinger picture. The wavefunction
in the Heisenberg picture is the same as the wavefunction in the Schródinger
picture at time t = 0. Thus, we have
tp(x) = (k/ir)3/í4 e (11.3.20)
where
k = (11.3.21)
Also,
(r2) = {Po)~j ¿ + (po-x + x-p0)— + (x2) . (11.3.22)

m m
So, we get
•2 roo
(f2) = (Ar/ir)3^2—J47T / — h2 (k2r2 — k) — kr r" dr
m
Jo
2
+(k/n)3/2—47t f —
2
(—2kr2 + 3) , - f c r 1
dr
m jo
f c 2 kr
+(k/ir)3t24ir
/ r e- - * r 2 dr (11.3.23)
Jo
Evaluating these integrals we get
/o /o , i a\ f 6k 3
( f 2 ) = 47t— tv + . (11.3.24)
"m2" / ^ +
2 ^ (,2(2fc) 2 ~ 4k 8fc2
Simplifying, we find
,-2\ 3 /i 2 fci 2
(11.3.25)
^ ^ ~ 2 k ~ 2m2 '
Substituting the value of fc we finally obtain
,-2, 3A hot2
(r ) = . (11.3.26)
2 mw 2m
11.4 Dirac Picture for Displaced SHO

Transform the displaced one-dimensional SHO
H = ^ b l-kx2 + /3x H0 + f3x

Zm 2
to the Dirac picture. Solve for both ipo(t), the wavefunction of the system and
xd(t), the position operator of the system. Assume ipo{0) describes the ground
state of H0 in the Schródinger picture.
Hint: Rewrite everything in terms of the Schródinger picture annihilation and
creation operators and use the results of problem 9.15 . The expression for
ipD{t) cannot be completely evaluated.
200 CHAPTER 11. TRANSFORMATION THEORY
Solution
We have the Hamiltonian in the Schrodinger picture
p2 1
H = — +-kx2 + (3x = H0 +fix . (11.4.27)
Then we find the unitary

.H0t operator Uo(t) that transforms to the Dirac picture
U0(t) = exp 1 (11.4.28)
h
If we write Ho in the form
H0 = hu(a} a + 1/2) (11.4.29)
and
/& =
' V 2 ^ ( 0 , + a) (11.4.30)
then
U0(t) = exp [-iu/tâ + 1/2)] . (11.4.31)
Next we compute
xD(t) = U0](t) x U0(t) (11.4.32)
So we need to compute
x¡o(t) 2mw CXP

Cl'wâta 1/2)] (a^ + a) exp [—iujt(aâ + 1/2)] . (11.4.33)
Using the results from problem 9.15 that if
[A, B] = XA (11.4.34)
then
AeB = ex eB A (11.4.35)
we find that
ae a g — iujt g — aa
(11.4.36)
and
a t g—iwtaâ __ eiwt g —¿wtaâ^t

(11.4.37)
so that
x
o(t) = \j —[e~'ut a + eiwt a*]
V ¿muí
Ps
= xs cos ojt-\ sin cot . (11.4.38)
771LÜ
11.4. DIRAC PICTURE FOR DISPLACED SHO 201
An alternate way to solve this part is to use the equations of motion for
XD(t) and pD(t).
^ dXD r TT i I F I D X D
_ P D
ik =
~dT ix^Ho\= -PD or
~df ~ m
<
i t = \PD, Ho] = -ihkxo or ^-=-kxD. (11.4.39)
Therefore,
= }_d_PD = = _ J x d ( n 4 40)
dt2 m dt m
and
= x s cosujt + sinwi (11.4.41)

mu
just as before.
b) To find the state in the Dirac picture we first solve for the state in the
Schrodinger picture. The Hamiltonian is
H = hu(aâ+l/2) + P\l —^—(a^ + a) = fiw[at a + c ( a t + a ) + 1/2](11.4.42)

2mu)
where
<n 3)
«-Vsb "
Therefore the Hamiltonian may be rewritten as
H = hu(bH+ 1 / 2 - c2) (11.4.44)
where
6= a+ c , 6^ — a t _|_ c (11.4.45)
So, the ground state has an energy
E0 = tu(l/2- c2) (11.4.46)
and satisfies
6|0) = 0 . (11.4.47)
The Schrodinger ground state is then
|tfs(t)> = e - i " t < 1 / 2 - c 2 ) | 0 ) . (11.4.48)
So, the equation for the state in the Dirac picture is

ih~9D(t) = PxD(t)9D(t) (11.4.49)
,iun ii. i HANSFORMATION THEORY
where ^ D (0) = ^ 5 ( 0 ) . To proceed we convert this to an integral equation
^D(í) = ^s(<) — -jr J xo(t')^D(t') dt' . (11.4.50)
This integral equation is of a type known as a Volterra equation and can always
be solved by simply iterating. Thus, we find
ip r
y¡d(<) = ^ s ( < ) - y J dtiXD(ti)yS{tl)
l
dt2XD(t2)iSf5(^2)
= + ±{^pj
+ j dtiXo(tl) J
*s(t) j\tlXD(h)
[tl rtn-1
x dt2xD(t2)...
dtnxD(tn)ips(tn) . (11.4.51)
Jo Jo
This is as far as one can go since this last expression cannot be simplified.
11.5 Heisenberg Picture for Displaced SHO

Repeat the previous problem for the Heisenberg picture.
Solution
As in the previous problem, 11.4 above, we have
p2 1
H = - + -kx2 + px. (11.5.52)
The Heisenberg equations of motion are:

ih = P_
lE ~ (11.5.53)
m
or
dx p
(11.5.54)
dt m
and
dv
= [p, H] = —ihmui2x — ih/3 (11.5.55)
or
dp n
— = -mu x - p . (11.5.56)
11.6. HEISENBERG PICTURE: SHO AND CONSTANT FORCE 2U3
The solutions are:

Po
x = So cos uit-\ sinwt
m rwjj2
p = po cos uit — muxo sin uit . (11.5.57)
Also the Heisenberg state at any time t is the same as the Schródinger state at
time t — 0.
|* H (<)> = | * s ( 0 ) > = |0) • (11.5.58)
11.6 Heisenberg Picture: SHO and

Constant Force
Use the Heisenberg picture to find the expectation values for p and x for a
particle in a harmonic oscillator potential and also acted on by a constant force
F. Assume that the state of the system is the ground state of the simple
harmonic oscillator without the constant force F.
Solution
The Hamiltonian is
H = ^ - + \kx2 - Fx (11.6.59)
2m 2
The equations of motion are
dx p
dt m
(11.6.60)
k +F
t =~' •
The state in the Heisenberg picture is |0), the ground state of the simple har-
monic oscillator at t = 0 in the Schródinger picture. This state is annihilated by
the Schródinger picture annihilation operator which coincides with the Heisen-
berg picture annihilation operator at time t = 0. Also, in terms of the annihi-
lation and creation operators, the position and momentum are given by
\l (a^ + a)
V 2 mui
p = i (a _ a
) (11.6.61)
or
mujh ip
x +
muj
,t - mujh ip
(11.6.62)
moj
un/irj UK 11. TRANSFORMATION THEORY
The equation of motion for the annihilation operator, in the Heisenberg picture
is, by direct substitution
S= (u 6 63)
--
Therefore,
a = «(0) e-<" + i F ( l - e'"") (11,6.64)
and
at = » t ( 0 ) « ' » ' - ¡ F ^ / ( 1 - e""«) . (11.6.65)
Hence using that
(0|a(0)|0) = <0|a t (0)|0> = 0 (11.6.66)
we find
<0|x|0) = 0 (11.6.67)
and
(0|p|0) = — ( 1 — c o s w i ) . (11.6.68)
id
11.7 Heisenberg Picture: Constant Force

A free particle is acted on by a constant force F . Use the Heisenberg picture to
find the expectation values of p and x if the state of the particle is described by
a Gaussian wavefunction.
Solution
Here, the Hamiltonian is given by
7)2
H = £ - - F z . (11.7.69)
¿m
The state is
ô(s) = . (11.7.70)
The Heisenberg equations of motion yield

dx p
dt m
^ = F. (11.7.71)
11.8. SCHRÓDINGER PICTURE: CONSTANT FORCE 205
Therefore, we find
P = PO + Ft (11.7.72)
and
^ . - E l + L t (11.7.73)
dt m m
so that
2
x = x0 + ^ t + ^ t . (11.7.74)
m 2m
Since the Gaussian wavefunction is an even parity state we immediately find
that
(ipo,Paipo) = (ij>o,xoil>o) = 0 . (11.7.75)
So, we get
(4>o,prl>o) = Ft (11.7.76)
and
2
(V>o,*Vo ) = ^ - (11.7.77)
11.8 Schrodinger Picture: Constant Force

Repeat problem 11.6 in the Schrodinger picture.
Hint: Do not try to actually find the time-dependent wavefunction, but try to
find an alternate method. It may help to review some of the results of [11.1].
The point that this problem illustrates is that in this case (a constant force) it
is much simpler to do the calculation in the Heisenberg picture.
Solution
In this case we could try to first solve the time-dependent Schrodinger equation
99 h2 d29
ih
m = ~F'" • <1L8'78)
We would do this by first separating out the time
9{t,x) = ip(x)e~iEt'R (11.8.79)
so that
h2 d2ip
Fx%¡) = Exp . (11.8.80)
2m dx2
This equation leads to Bessel functions of order ± 1 / 3 . We would then have to
expand the initial state, namely
2 2
m*) = ê-* ' (11.8.81)
\ a
.ii. i n/\ IMS FORMATION THEORY
in terms of these eigenfunctions to get the time-dependent solution. This is

rather complicated. The expectation values of p and x could then be evaluated
using this result.
A better way to do this is to use the results discussed in [11.1]
^ = ~(ii[pi/t]*) = («if®) = f (11.8.82)
This simply says that
(p) = Ft . (11.8.83)
Therefore,
,—tV) = —t (11.8.84)
m m
and hence,
2
(x) = (x)\t=0 + ^ t . (11.8.85)
But at t — 0 we have that ( x ) = 0. Thus, we recover the results obtained in the

Heisenberg picture.
11.9 Dirac Picture: Constant Magnetic Field

An electron in the potential V(x) is also acted on by a weak constant magnetic
field so that the resultant Hamiltonian is
h ea
~ 2m (x^) +
with
A = — -x x B .
2
Introduce the interaction picture with
H
°=L+V(*>
and find the equations of motion for p and x in this picture. Write also the
Schrodinger equation for the wavefunction in the interaction picture.
Hint: Use the fact that the magnetic field is weak as well as, the result that you
will prove later (problem 17.8), that if B is a constant vector then
p•A + A •p = B • L
11.9. DIRAC PICTURE: CONSTANT MAGNETIC FIELD 207
Solution
In the Interaction picture the equation of motion for any dynamical variable is
given by
ih
~ 4 r - ~ I a D,HO] • (11.9.86)
at
Therefore using
Ho = ^ + V(x) (11.9.87)
2m
we get
~~rr~ ~~ j ^ = -VV(xD). (11.9.88)

at m at
Here we have used the fact that for any function of momentum
Fd(P) = U¡F(p)Uo = F(U¡pU0) = F(PD) (11.9.89)
where
U0=e-iHot/n. (11.9.90)
The Schródinger equation is
M^\9D) = H'D\9D) (11.9.91)
where
H'd = UqH'UQ . (11.9.92)
Now for a weak magnetic field we can drop the term ( e 2 / 2 m ) A • A. Thus,
H'd = eiHot/h \-4-{P-A + A-P)\ e~

iHotlh
. (11.9.93)
2m j
But, using the hint that
p • Á+ A • p= B •L (11.9.94)
we see (since B • L commutes with IIQ) that
H'd = —B • L . (11.9.95)
2m
So, the Schródinger equation now reads
J I T , e
ih!L\9
l D) = -J-B.L\*
_ D) . (11.9.96)
di' ~~ 2m
± riAivawRMAl'lUN THEORY
11.10 Coherent State

Show that the normalized coherent state
\z)=e~^'2eza'\0)
may also be written
\z) = e[2at-2'al|0> .
Hint: Use the results of problem 9.14 (the Baker-Campbell-Hausdorff formula).
Solution
We have that
|z) = e -*2/Vat|0) (11.10.97)
as well as the result of problem 9.14 that if there are two operators A and B
such that
[A,[A,B]] = [B,[A,B)) = 0 (11.10.98)
then
eA+B =eAeBe~1'2^BK (11.10.99)
The operators za) and z*a satisfy this condition. Therefore we have that
_ gza* e - 2 ' a e - l / 2 | ¿ | 2 (11.10.100)
And since
a|0) = 0 (11.10.101)
so that
e _ 2 ' a | 0 ) = |0) . (11.10.102)
It follows that
\z) = e * a t - i , a | 0 ) . (11.10.103)
11.11 Coherent State: Overlap of T w o States

Verify the equation
|(u>|¿)| 2 = e x p - [|w| 2 + \z\2 - w*z - wz*] =
for any two coherent states \z) , |u>) .

11.12. COHERENT STATE: WAVEFUNCTION 209
Solution
We want to show that
|Hz)|2 = - e-l»-*!2 (11.11.104)
where
°° 2"n
\*) = e_k|2/2 E 4=rln> • (11.11.105)
n = 0t\ vn!
YL 11 *
So, we have
2
(w\z) = e-(M +M2)/2 y * ™ (m|n)
(ma+ma)/a ^
El( w* z)
n=0
w z n
a S
e -(|z| +M -2««**)/2 (11.11.106)
Similarly,
2
(z|u>) = -2^')/2 . (11.11.107)
Thus,
|h¿)|2 = = e-l"'-zl2 . (11.11.108)
11.12 Coherent State: Wavefunction

Verify the equation
| 2 ) = ( l - | 2 | 2 ) 1 / 4 e x p ( | a t 2 ) |0)
where
i*) = ( i - N 2 ) 1 / 4 E ' w > / ( 2 / ^ ) ! ! N
Solution
We have that
2 1 / 4 1)M
|z) = ( 1 - | z | ) y V ^ " |2n) . (11.12.109)
But, it is easy to see that
(2n)!! = (2n)(2n - 2)(2n - 4) . . . (2) = 2" n! (11.12.110)

¿w CHAPTER 11. TRANSFORMATION THEORY
and
(2n)! = (2n)ü (2n - 1)!! . (11.12.111)
Also,
|2n)=-7=L=(at)2"|0). (11.12.112)
V(2«)!
Thus,
00
1 1
|2\ 1/4 ^ n
oo 1
= — (a»)'"|0)
n=0
= (1 - | z | 2 ) 1 / 4 e x p [ z / 2 ( a t ) 2 ] |0) (11.12.113)
as required.
11.13 Squeezing Operator

Use the result of problem 9.15 to derive the equation
5* (r)xS(r) = \ 5 f ( r ) ( a + a^S^r) = erx

V ¿TTILÜ
as well as the result
(r)pS(r) - —¿y —-—S*(r)(a — a^)5(r) - e~rp
where
S(r) = exp ^ ( a t 2 - a 2 ) ] .
Solution
From problem 9.15 we know that if
[A,B] = XA (11.13.114)
then
AeB=exeBA. (11.13.115)
We apply this result to A = (a ± a*) and
C 8 = S ( r ) = exp ^-(a* 2 — a 2 ) ] . (11.13.116)

11.14. NO EIGENSTATES FOR CREATION OPERATOR 211
As a first step we find
(11.13.117)
[fl. ^ *- q 2
M = 22<jt = rat
and
[at,^(at2 -fl2)] = g2a = ra

(11.13.118)
Therefore,
(a ± a^)5(r) = (aiaêxp -(at2-a2)
exp (a^ 2 —a 2 ) e ± r (i na ±-4-a/ Ti t )\
= S(r) e±r (a ± a*) (11.13.119)
This is the desired result since
SS* = SfS = 1 . (11.13.120)
11.14 N o Eigenstates for Creation Operator

Show that the creation operator has no normalizable eigenstates.
Hint: Assume that eigenstates exist and show that they are not normalizable.
Solution
We assume that the creation operator has a normalizable eigenstate |z) so
that
at|z) = z\z) , (z|z) < oo . (11.14.121)
Next, we expand |z) in terms of the harmonic oscillator eigenkets |n) so that
c
l*> = E » - (11.14.122)
Then, using (11.14.121) we get
Y . cny/n + l|n + 1) = z ^ c„|n) (11.14.123)

n n
Therefore,
y/ñ c \frú
c
n — n —1 =
Co . (11.14.124)
Z z"
Hence,
(11.14.125)
and
2
(z\z) = \ c o \ Y^- (11.14.126)
n
This sum clearly diverges for all values of z. Thus, there are no normalizable
eigenstates of the creation operator a f .
11.15 Spin Coherent State: Euler Angles

The Euler angle representation of a spin coherent state is defined by
Ie, rp) = e~i¡fiJ''h e-iBJ^n \j,m) (11.15.127)
where | j, m) is an eigenstate of J 2 and Jz corresponding to the eigenvalues

j{j + l)'®2 a n d respectively. Since the action oí e~'^Jz^n on \j,m) is trivial
it is usual to take ip — 0 • So we shall do the same and consider the spin coherent
states
\9, ip) = e-iipJ'/h e~iejy/n \j,m) . (11.15.128)
For the state |j, m) with j = 1 evaluate the spin coherent states corresponding
to
a) m = 1 .
b) m = 0 .
c) In both cases evaluate the overlap (6', ip'\9, ip) .
Hint: Use the result obtained in problem 9.3 that for any component Jk k =
x, y, z of the angular momentum operator corresponding to j = 1 you have
-ieJk/n = i - ^z + l
-^ s i n 9 c ¿o s d . (11.15.129)
n ft fi
Solution
For j = 1 we have t h a t
1 0
¿1 0 0 (11.15.130)
h
0 0
and
(11.15.131)
yfi
Therefore
1 ü 0
0 0 0 (11.15.132)
n?
o o i
11.15. SPIN COHERENT STATE: EULER ANGLES 213
and
jJ22 __ i 1 0 -1
Ji- 0 2 0 (11.15.133)
ff
-1 0 1
Now, using the hint we have that
0 0
— iipjz/h _ 0 0 0 (11.15.134)
0 0 é*
and
/ ^(l+cos0) —^sin# ^(1 — cos 9) ^
-iBJy/h sin 9 cos 9 —^ sin 9 (11.15.135)
^ | ( 1 — cos9) —^sin9 | ( 1 + cos#) j
a) If m = 1 then
IU> = (11.15.136)
and
—\ e (1 + cos 9)
|0, f ) = (11.15.137)
,72sind
h ,(p
e (1 — cos 9)
b) If m = 0 then
|1,0) = (11.15.138)
and
~72e ivsin0
|0> <fi) = cos 9 (11.15.139)
S 9
72^ ™
c) For m — 1 the overlap is
(0\ <p'\0, <p)

-i( ip-tp') (1 + cos 9) (1 + cos 9') + - sin 9 sin 9'
6
4
+ — e1îp~'p ' (1 — cos 0)(1 — cos 9')
= — (1 + cos 9 cos 9') cos()

— ig «(v V ) s¡ n Q s ¡ n Q' _|_ c o s 0 c o g Q'
2
_j_ Ig'tv v) sin0 sin#'
2
= cos 9 cos 0' + sin 0 sin 9' cos(ip - y>') . (11.15.141)
11.16 Minimum Uncertainty Spin

Coherent States
Show that for any choice of the "fiducial vector" | j, m) not all the spin coherent
states
\ e , <p) =e-i,fij'/ne-iej»/n Ij,m) (11.16.142)
can be simultaneously minimum uncertainty states for any pair of non-commuting

operators. This is to be contrasted with the coherent states based on the cre-
ation and annihilation operators of the simple harmonic oscillator.
Hint: For the required proof first find an equation to be satisfied by any min-
imum uncertainty state. Then show t h a t this equation does not have enough
solutions to encompass all the spin coherent states.
Solution
Supppose the operators of interest are A, and B. To derive the Heisenberg
uncertainty relation we first define
A' = A - (A) , B' = B-(B) , (11.16.143)
where the expectation values are taken with the state i/>. Then,
( A A ) 2 = (A'1p,A'1p) , (AB)2 = (£?V,BV) . (11.16.144)
We now consider
\([A,B])\* = \([A',B'])\2
= ( < m ' b v ) - w*. s ' a ' v ' )
= (¿v,5'v>)-(b'v,¿v)
= (A'rp,B'rp)-(A'ip,B'i¡;y
= 2i^(A'ip,B'ip) . (11.16.145)
Therefore,
\{[A,B])\=\2Q(A'i,B'1p)\<2\(A'4>,B't)\ • (11.16.146)
11.16. MINIMUM UNCERTAINTY SPIN COHERENT STATES 215
By the Schwarz inequality we get for the last term

2\(A'rp,B'iP)\ < (v4V, ^ ' V ) 1 / 2 ( ^ V , -e'V) 1 / 2 = 2AAAB . (11.16.147)
Thus, we obtain the Heisenberg uncertainty relation
AAAB>l-\^,[A,B]n • (11.16.148)
For a minimum uncertainty state this must be an equality and therefore both
of the inequalities (in (11.16.146) and (11.16.147)) must also be equalities. The
first of these requires that
1 B ' i 0 1 = B'ip)| (11.16.149)
and the second requires that

A'ip = \B'ip (11.16.150)
where A is some constant. Substituting this into (11.16.149) we get that
|9A||(B'rP, B V ) I - |A||(B'V>, B'tP)| (11.16.151)

so that A is pure imaginary
A — ia , a = real . (11.16.152)
We use these results to write out in detail (11.16.150) (the equation for the
minimum uncertainty state ip).
{A - (A))i = ia(B - (B))iP . (11.16.153)
This equation can be cast in the form of an eigenvalue equation
[A - iaB)ip = ((A) - ia{B))xp = pip . (11.16.154)
Here,
P = (A) — ia(B) (11.16.155)
is the eigenvalue.
Now, for spin coherent states, any operators A and B are simply ( 2 j + 1) x
(2j + 1) dimensional matrices. Therefore, to obtain solutions to the eigenvalue
equation (11.16.154) we have to solve the characteristic equation
det(A-iaB-P) = 0 . (11.16.156)
However, this equation yields at most 2 j + l distinct complex eigenvalues depend-
ing on the real parameter a . The corresponding eigenvectors may be labelled
|k(a)),p(a)) k = 1,2, • • •, (2j + 1) .
These states can never represent all the states 10, f ) which are labelled by two
continuous real variables. So, not all the spin coherent states \0, <p) can be
minimum uncertainty states.
Incidentally, those spin coherent states t h a t are minimum uncertainty states
are called "intelligent spin states". [11.2]
ü AU CHAPTER 11. TRANSFORMATION THEORY
11.17 Spin Coherent States: Complex Variables

Consider the spin coherent state
|z) = N e{zJ-)/ihV2)\j,j) , (11.17.157)
where z is a complex number, N is a normalization constant, and
J— — Jx iJy *
This is a complex variable representation of a spin coherent state.

a) Find the normalization constant N.
Hint: Use the fact that
m
J-1h j) = VU ~ + 1)0* + ™)h \j, j) (11.17.158)
as well as
J - \ j , - j )= 0 . (11.17.159)
b) Show that the overlap of two such states is given by
/ i n (l + z*z') 2 j
(1l17
' (i + i . - m i í | z t h • '160)
c) Verify
2j+lthe completeness
ft w , d2zrelation
= 1
> < (11.17.161)
7T J ( 1 + Z*Z')2
where
d z = dx dy , z = x + iy . (11.17.162)
Hint: To evaluate the integral it may be useful to use polar coordinates. Also
the following integral may prove useful.
r xn dx n\(m — n — 2)!
/ m • \rn= < n > —I , m > n + l . (11.17.163
m v
Jo (1 + x) (m - 1)!
Solution
a) To start we evaluate the action of J" on | j , j ) by making repeated use of
(11.17.158).
= y/lMlJ.i ~ 1)
Jl\j,j) = VH2j)V2(2j-l)\jJ-2)
Jt\j,j) = v/iMv/2(2j - l ) V m ~ 2 ) \ j j - 3)
J-\hi) = > / l ( 2 j ) > / 2 ( 2 j — ! ) • • • ^ 2 ( 2 j - 1 ) ^ 1 ( 2 j ) \ j , - j ) (11.17.164)

11.17. SPIN COHERENT STATES: COMPLEX VARIABLES 217
and finally
Jlj+1\jj) = 0 • (11.17.165)
Expanding the exponential in the definition of |z) we get
i|2> = + í í m ü j í
2
= \j, j) + zVWU J ~ !) + \ZJW~1)1 li. i - )
2i
+ ••• + z \j, ~j) • (11.17.166)
It now follows that

1
(z\z) = 1 + |z| a + (|z| 2 ) 2 j ( 2 j " 1}
+ Vl( \ 1z y\ ^ j ( 2 j ~ l ) ( 2 j
~ 2)
|TV"J2 x 1 ' 2! 3!
+ ••• + ( M 2 ) 2 j
= E(W2r
2J
xn ( 2j
n=0
= (l + | 2 | 2 ) 2 j . (11.17.167)
Therefore,
AT= (l + | r | 2 ) _ i (11.17.168)
b) We begin with the definition of \z). Then,
(z\z') = (i + \z\y3(i + \-'^-

z r
3
x
(zV+)/(fiV2) (z J )/(fiN/2)
0',j|e e ' - |i,j) . (11.17.169)
But, as we saw above,
e(zJ.)/(hV2) _ Ij, j ) + z^W)\j,j - 1) + ^ y \ / j ( 2 j - 1)|i, J - 2)
+ ••• + 2 2 j | i , - j ) • (11.17.170)
Therefore,
/,' } -| e (^j + )/(a n /2) Jz'j_)/(hV2)
+ ••• + (z*z')2j
2
• h*-*r( !)
= (l + z V ) 2 i
n=0 X
•
7
(11.17.171)
So,
(.z\z') = (1 + |,|2)-J (l + |z'|2)

(1 + z*z') 2j
/|2\-j
(11.17.172)
c) To verify the completeness relation we simply write it out.

2j + l d2z
;\z)(A
(1 + m 2 ) 2 '
z
= — ~ J (1 + | 2 |2)2j + 2 ( i j ' j ) + v j \ j j ~ !) + i" z ' 2 j |j, - í ) )
X ((j,j\ + z*\/j(j,j-M + --- + {z*)2j(j,-j\) • (11.17.173)
The typical term involves an integral of the form
I
d2zzn{z*)m
(1 + |z|2)2¿+2 '
If, in this integral, we let
x = r cos 9 , y = r sin 0 , dx dy = reír d9 (11.17.174)
so that
z = re'e _* -»0 (11.17.175)
then,
d2zz"(z*2)2m 2
/ (l + |z|)J+
,00 j . n + m + 1 ^ y2,r
-m)9 M
2 2 2 6
Jo (l + r ) ¿ + j >
yoo r „+m + l d r
2ttS, (11.17.176)
I (l + r 2)2j+2 "
n,m
So, only the diagonal terms survive. Jo
But these are now all of the form
«±12, ')!. °o
7T \
2j
n 2n + l
(l r + r 2)2j+2
d r I j,j ~ n)(j,j ~ 1
)j;
2j
= ( 2 j + 1) | j, j — n)(j, j — n\
( l + x ) 2 j + 2
2j \ nl(2j — n)!
= (2j + l ) \ j , j - n ) ( j , j - n \
n J (2.7 + 1)!
= \j,j-n)(j,j-n\. (11.17.177)
Here we have used the integral given in the "hint". Thus, finally we have
^/|í>WÜT??F =
2j + 1
E t í J - n W . } - »
n=0
j
m
= E Ii'm——j
l
= i . (11.17.178)
11.18. USEFUL COMMUTATOR 219
11.18 Useful Commutator

prove that
[fl , a 1 (fl t ) # ] = n ( « , ) B •
Hint: Use induction.
Solution
We want to prove, using induction, that
[ata, (o^)"] = n (at)» . (11.18.179)
p o r n = 1 we have
[ota,at]=at[«,ot] = ot . (11.18.180)
Now we assume that (11.18.179) holds for n and we consider n + 1. Then we

have
[aâ,(a t )" + 1 ] = [ a ^ , (at) B ] a t + (at)»[0t0|0t]
= n(at)"at + (at)"at
= (n + l ) ( a t ) " + 1 . (11.18.181)
11.19 Forced SHO

Given the Hamiltonian for a "forced" simple harmonic oscillator
v2 1
H(t) = —- + -muj2 x2 — xF(t) — pG(t)
2m 2
where both F(t) and G(t) vanish outside the interval 0 < t < T, calculate the
probability P(n) that the system is in the state of n quanta of the Hamiltonian
H(t) (t > T) if it was originally in the ground state of the Hamiltonian
H(t) (t < 0) .
Hint: Use the Heisenberg picture.
Solution
The Hamiltonian may be rewritten to read
H = ^hw(aâ + 1/2) — hf(t)* a} — h f(t) a (11.19.182)
where we have defined
<'<> = ;75kn,)-VWG(,)- (11.19.183)

itiAi'lER 11. TRANSFORMATION THEORY
The Heisenberg equations are

di
da ,
1- IUI a — if
1dt
dat
iuja'
i.w = - i f . (11.19.184)
dt
These equations are solved by using integrating factors e laJt and e~iwt respec-
tively to give
a(t) = ie-iwt f eiut'r(t')dt' + ae~iut
Jo
a^(t) = ieiwt Í e~iut'f(t') dt' + a* eiwt (11.19.185)

Jo
where
a = a(0) , = af(0) . (11.19.186)
We now define
b(t) = a e~iut + i e~iu,t [ eiwt' f* (<') dt'

Jo
= e 'wt (a + c)
tf(t) = eiwt(a]+c*). (11.19.187)
In this case the state |n; 6) of n quanta of the Hamiltonian H(t > T) is given by
(6t\n
|n;6) = i - 7 % | 0 ; 6 ) (11.19.188)
Vn!
where
6|0;6) = e~iut (a + c)|0;6> = 0 . (11.19.189)
This shows that the state |0; b) is a coherent state, namely an eigenstate of the
annihilation operator a. Thus, we can write
10; b) = e~iut e - W 2 / 2 e ~ c a t |0; a) (11.19.190)
where |0, a) is the state annihilated by a and gives the ground state of the
Hamiltonian H(t < 0) .
We want the amplitude
(n; 6|0; a) = e~inut e~^'2 (0; a\e~c' ° |0; a)

Vn!
= e~inut
e-W2/2 ~ . (11.19.191)
Vn!
Therefore, the desired probability is given by
P(n) = |(n; 6|0; a)| 2
|r|2n
= -w
ll_ .
e (11.19.192)
n\
This is a Poisson distribution.
11.20. 3-D SIMPLE HARMONIC OSCILLATOR 221
11.20 3-D Simple Harmonic Oscillator

A particle moves as a three-dimensional simple harmonic oscillator so that its
Hamiltonian may be written
+
" = ¿ •
At time t — 0 the position of the particle is measured to be at r = rb. Show

that a measurement of the position of the particle at time t = T / 2 , where T is
the period of the oscillator, is sure to yield the value — ro.
Hint: Use the Heisenberg equations of motion.
Solution
The Heisenberg equations of motion are formally identical with the classical
equations. Furthermore, in the Heisenberg picture the state vector does not
change with time. Thus, the Heisenberg equations are
df p
dt m
-jj- = —mw 2 r . (11.20.193)
The solutions of these equations are formally the same as the classical solutions.
r = ro cos ujt + sinut . (11.20.194)

mw
Here we have already imposed the initial condition that at t = 0 we have r=r0
where ro is t h a t operator whose eigenvalue at time t = 0 is r"o-
Now, the period T of the oscillator is given by
u
Therefore, at t = T / 2 we have t h a t
r = r0 cos 7T + sin n = — ro . (11.20.195)

muj
But, as stated, the state of the system is an eigenstate of ro. Therefore, we are
certain to measure the value — ro at time t = T / 2 .
11.21 Quadrupole Tensor

The quadrupole tensor of a system is defined by
Qik — X{Xk ^7c|r|-) j

r uhmatION THEORY
where f is the position vector and are the components of r. For a system in
the state of definite angular momentum
V>(r, 9, <p) = R{r)Yim(9, ip)
evaluate the components of Qik •

Hint: Use the relation
/ (I + l ) 2 — m2 / I2 — m 2
<+1,m + ,-1,m
cos 9 Y l m = y (2/ + l)(2/ + 3) \ j (2/ - 1)(2Z + 1) "
Solution
We begin by writing the components xt in spherical coordinates and use the
fact that
J \R{r)\2r2dr= 1 (11.21.196)
to write
Qxx = (r2) J[3sm2 9cos2 <p - \)Y*mYim d£l
Qxy = Qyx = (r2) J 3 sin 2 9 cos 2 ip Y*mY¡m dd
Qyy = (r 2 ) J ( 3 sin 2 9 sin 2 <p - 1 )Yl*mY,m dtt
Qxz — Qzx = (r2) J 3 sin 9 cos <p cos 9 Y*mY¡m dCl
Qyz = Qzy = {r2) J 3 sin 9 sin tp cos 9 Y*mY¡m dQ
Qzz = {r2) J(Zcos2 9 — \)Y*mYim dQ, . (11.21.197)
Next we recall that
Yim(0, ip) oc Pr (cos 9) eimv . (11.21.198)
In this case it follows that
sin
/ V y*m (6, <p)Ylm (0, <p) dip
J0
/>2ir
= cos v? Y*m(9, <p)Y¡m(9, <p) dip
Jo
= 0 (11.21.199)
and hence t h a t
Qxy = Qyx = Qxz = Qzx = Qyz =Qzy = 0 . (11.21.200)

11.22. EIGENFUNCTION OF J2, L2, AND Jz 223
By exactly the same argument and using t h a t
sin 2 if = - [ 1 — cos2yp]
cos 2 ip = - [ 1 + cos 2<p]
we see that
Qxx = Qyy = (r2) J(^sin2 6 - l)Y^Y¡mdü
l
= -(r2) J(1-Z cos2 9)YrmYlr¡ dd
(11.21.201)
Q z z
2 '
Now using the hint twice we have t h a t
COS2 6 Y¡m - AY,+2,m + BY¡-2,n

(1+ l ) 2 - : /2 - m 2
+ + Yi, (11.21.202)
. ( 2 / + l)(2/ + 3) ( 2 / + 1 ) ( 2 / - 1).
where A and B are irrelevant constants. It now follows that
(l + l)2 — m2 i2 m
Qz + <r 2 ) . (11.21.203)
(2/ + l)(2/ + 3) (21 + 1)(2Z — 1)
After some simplification this reads
21(21 + 1) - 6 m 2 22
Qz (r ) . (11.21.204)
(21 — 1)(2/ + 3)
As a direct check we also see immediately from the definitions for Qxx, Qyy and
Qzz that
Qxx "1" Qyy Qzz — 0 (11.21.205)
11.22 Eigenfunction of J 2 , L2, and Jz

A spin 1/2 particle is in a central potential. Find the explicit form of the
wavefunction which is simultaneously an eigenfunction of
J2, L2 and Jz — Lz + Sz .
Solution
In terms of the Pauli spinors we have that
J = Ll + ^ha. (11.22.206)
So, in matrix form we have
/ + 2 0 \
• (11.22.207)
V o -i-fc ~ I /
The eigenfunctions of Jz corresponding to the eigenvalue mh are
/ ñi(r)fi(0)e1'(m-1/2)^ \
V = . (11.22.208)
\ R2{r) F2{6) J
Since we also want these to be eigenfunctions of L 2 we require that
L 2 V = /(/ + l ) / i 2 ^ (11.22.209)
or, written out, that
I2fli(r)F1(0)ei(m-1/2>'' = /(/ + l)/rfl1(r)JF1(<?)ei(m-1'2>¥'

L2R2(r)F2(e)e^m+1'2^ = 1(1 + l)h2 R2(r) F2(9) e '( m +i/2)v> (n.22.210)
This means that
Fl{0)ei(m-1/2)V = Y l i m _ 1 / 3 (e,p)
F2(e)e^m+l'2^ = Yl)m+1/2(6,<p) . (11.22.211)
Finally, we require that $ also be an eigenfunction of J2. To achieve this we

write out J2 explicitly as a matrix using the fact that, acting on Y¡m
J2 = 1(1 + l)h2 + ^h2 + ha • L . (11.22.212)
Then
1(1 + 1) + f + \LZ
J< = h (11.22.213)
\L+ 1(1 + 1) + 1 —
where
L±Y¡im — (Lx ± iLy)Yltm = \Jl(l + 1) — m(m ± l)hYiim±i • (11.22.214)
So, writing out the pair of equations resulting from the eigenvalue equation
J2q> = j ( j + \)h2 (11.22.215)

11.23. SHO: A TIME-INDEPENDENT OPERATOR 225
we get
[(/ + 1/2) 2 - j{j + 1) + m]Ri + s/(l + 1/2) 2 - m?R2 = 0
y/(l + 1/2)2 - m? R\ + [(/ + 1/2)2 _ + J) _ = o .(11.22.216)
f o r a nontrivial solution we require that the determinant of the coefficients of
Rx and 7?2 must vanish. This yields
[(/ + 1/2) 2 — j ( j + l)] 2 —TO2— (/ + 1/2) 2 + m 2 = 0 . (11.22.217)
The two roots are

j = 1+1/2 and j = 1-1/2 . (11.22.218)
The corresponding (normalized) eigenfunctions are
j = I + 1/2, Z = 0 , 1 , 2 , •••
R , in(r) { V l + m+ 1/2y/,m-i/2
*,j=i+i/a.m = ~ ^ = f - I (11-22.219)
V ¿ 1 + 1
\ y/l-m + 1/2y,,m+i/2
j = l - 1/2, Z = l , 2 , 3,
R• Ir) ( — m + 1/2 i | , m - l / 2 \
. (11.22.220)
\ — -*// + m + 1 / 2 y ¡ j m + i / 2 /
11.23 SHO: A Time-independent Operator

Show that for a simple harmonic oscillator, the operator
A(t) = raui(í) cos wí — p(i) sin uit
is independent of the time t.
Can this operator be simultaneously diagonalized with the Hamiltonian?
Solution
The equation of motion (Heisenberg equation) for a time-dependent operator is
f = I [ A . * ] + £ . (11.23.221)
Here, the Hamiltonian is

d2 1
H=£- + -nvJ-x2 . (11.23.222)
2m 2
Thus, we get
dA muí 1 ,
= ——2»coswi-|—mui2x sinui
dt 2m 2
— mu>2x sin uit — uipcosujt
= 0. (11.23.223)
This shows that the operator A is independent of t. We can go even further and
obtain an explicit form for A by using the solutions for x(t) and p(t).
4^ P° • 4
x = xocoswr-l sinwr
mui
p = po cosut — mwxo sinwi . (11.23.224)
Here, XQ and po are the initial conditions at t = 0 for x and p respectively. They
are nothing other than the Schródinger picture operators for x and p. Then, we
see that
. . Po . .
A — mw Xo cos uit H sin u>t cos uit
mui
— [po cosut — rriLJXa sin uit] sinwi
= mujxo . (11.23.225)
Thus, it is obvious that A is indeed a constant operator, independent of t.

It is also clear that
[A,H]¿ 0 .
Therefore, it is not possible to simultaneously diagonalize A and H .
11.24 SHO with Time-Dependent Spring

Consider a simple harmonic oscillator whose Hamiltonian is
F(í) =
l^+^mw2(<)x2 "
a) Define the time-dependent annihilation and creation operators
a ¿
mult) x
. 1
() = \ nt +1 / =r^P
V 2h \Jlmhw(t)
ar(í) = 1
p . (11.24.226)
K
V 2h yj2mhui(t)
Use the Heisenberg equation, either directly or by first finding the equations of
motion for x and p, to obtain the equations of motion for a(t) and al(t).
b) Now rewrite H(t) in terms of a(t) and (t). Next use the fact t h a t the oper-
ators a(t) and a^(t) are related to the operators a(0) and a^(0) by a Bogoliubov
transformation to show that the expectation value of H(t) for a state |n), that
is an eigenstate of H(0),
tf(0)|n) = (n + l / 2 ) / k j ( 0 ) | n ) (11.24.227)
is of the form
(ra|tf(i)|n) = (n + 1/2 )hw(t)f{t) (11.24.228)

11.24. SHO WITH TIME-DEPENDENT SPRING 227
where f ( t ) is some function of t.

c) Find expressions, in terms of the parameters of the Bogoliubov transforma-
tion, for the expectation values at time t of x, p, as well as x 2 and p2 and hence
also for A x and Ap at time t. Use the results obtained to conclude t h a t it is
possible to alternately squeeze x and p.
Hint: Use the Heisenberg picture. Also do not a t t e m p t to obtain explicit solu-
tions for the time dependence. [11.3]
Solution
The Heisenberg equations are obtained in the standard manner by using
ih^ = [A, H] + ih~ . (11.24.229)

at Ot
As a consequence we get
= i • f = -"'«>*• ('I24 230>

Next, we introduce the explicitly time-dependent annihilation and creation op-
erators
. . mui(t) . 1
" ( 0 = \ r ^ 2th~ x + l -
«'(<) = • (11.24.231)
A straightforward computation now shows t h a t these satisfy the commutation

relations for creation and annihilation operators.
[a{t), a^(t)] =-^[x,p]+~\p,x] = 1 . (11.24.232)
b) Also, by straightforward computation we see that
lhw(t)[a(t)a^(t) + a\t)a(t)] = 7^ + \Tmj2(t)x2
= H(t) . (11.24.233)
Thus, using the commutation relation (11.24.232) we find that
H{t) = ftw(i)[at(<)a(i) + 1/2] . (11.24.234)
The Heisenberg equations for a(t) are
Íh
t=
=
huj(i)a{t) +
+
• (11.24.235)
But, using the explicit form for a(t) and denoting differentiation with respect
to t by a dot the last term on the right hand side of (11.24.235) yields
da w r rnuiit) . 1 , 0 ; +
— = —f\/——x — i—. =p] = —a' . (11.24.236)
dt 2u> V 2h y^2mhuj(t) 2u
Thus,
— = -iu{t)a + • (11.24.237)
dt
In a similar manner or by taking the dagger of (11.24.237) we find that
rl n t , (jJ
í í f - = iiü(t)al + —a . (11.24.238)
dt Zio
One can check that these equations are correct by writing out the Heisenberg
equations for x and p
1 7 ~ ~ ' IT = - m u j 2 { t ) x • (11.24.239)
dt m dt
Now inverting equations (11.24.231) and substituting a(t) and a^(t) for x and p
we again get (11.24.237) and (11.24.238).
The operators a(t) and a^(t) must be related to the operators a(0) and a f (0)
by a Bogoliubov transformation. We parametrize the Bogoliubov transforma-
tion with the three real functions a(t), f3(t), and 7 ( t ) to get:
a(t) = e ~ , a ' f ' a ( 0 ) cosh f3(t) + e ' ^ ^ a ^ O ) sinh/?(¿)

a^(t) = e _ , 7 ^ a ( 0 ) sinh/?(<) + e ' ^ ' a ^ O ) cosh/?(<) . (11.24.240)
Now consider a state |n) such that
/f(0)|n) = (n + l/2)fiw(0)|n) . (11.24.241)
In other words we also have that
a(0)|n) = \/ñ\n — 1)
at(0)|n) = Vn + l|n + 1) . (11.24.242)
Using (11.24.234) we see that
(n\H(t)\n) = (n\hu(t)[a^(t)a(t) + l/2]|n)

= (n|/jw(i)[sinh /3a(0)at(0) -I- cosh 2 ^0^0)0(0) + 1/2]|n)
2
= /kj(i)[(n + 1) sinh 2 ¡3 + n cosh 2 /? + 1/2]

= (n+l/2)fiw(f)cosh(2/?) . (11.24.243)
c) The expectation values of x and p vanish since both of these operators are
linear in a(0) and at(0). Thus,
<7i|a?(í)|7z> = <n|p(í)|n) = 0 . (11.24.244)

11.24. SHO WITH TIME-DEPENDENT SPRING 229
On the other hand,
(n\x2(t)\n) - M[at(°) + a(°)]2|")
h
( n | c o s h / ? s i n h ^ ( e - ' ' ( a - 7 ) + e i ( a - ^ ) [ a ( 0 ) a t ( 0 ) + a t (0)a(0)]
2 mu¡(t)
-f (cosh 2 (3 + sinh 2 /?)[a(0)a t (0) + a t (0)a(0)]|n)
= 2 — + l)[cosh(2/3) + cos(ar — 7 ) sinh(2/?)] . (11.24.245)
Similarly,
(n|p 2 (í)|n> = -m^(¿)(n|[at(0)-a(0)]2|n)

mfríü(t)
= —y-1(2n+l)x
x [cosh(2/?) — COS(Q — 7 ) sinh(2/3)] . (11.24.246)
Since,
(n\x(t)\n) = (n\p(t)\n) = 0
it follows that
(Ax) 2 = (n|a: 2 (i)|n)
(2N + l)[cosh(2/3) 4- C O S ( Q — 7) sinh(2/?)] (11.24.247)
2 mu)(t)
and
(Aja) 2 = {n\p2(t)\n)
mhu>(t)
(2n-I- l)[cosh(2/?) — cos(a — 7 ) sinh(2/?)] . (11.24.248)
2
Notice that if
COS(Q — 7 ) = 1 (11.24.249)
then
(n|p 2 (i)|n) = mhu>(t)(n + 1/2) e~213
Hx2(i)|rc) a _A_( n + 1/2) e 2 " . (11.24.250)
Thus, if ¡3 > 0 the variable p is squeezed. Similarly, by taking

cos(a — 7 ) = — 1 (11.24.251)
we get
(n|p 2 (i)|n) = mhw(t)(n + 1/2) e 2/J
(n\x2(t)\n) = — ^ - r ( n + 1/2)e"2" (11.24.252)

7TILÜ (t)
and for positive /? the variable x is squeezed. Of course, in both cases, if (3 is

negative the situtation is reversed.
For completeness we write out the equations that determine the functions
a(t), /?(<), and 7(t). Substituting (11.24.240) into the equations of motion for
a(t) or else a*(i) we get
^ cosh/? + /? sinh/?j e _ , a a ( 0 )
+ ^¿7 sinh /? + ¡3 cosh /?j e%1a) (0)
= —iui ( e ~ ' a cosh /?a(0) + en sinh fia) (0))
+ — (e~n sinh /?a(0) + e'a cosh /3a^(0)) . (11.24.253)

UJ
Now using the algebra
[a(0), a^(0)] = 1 (11.24.254)
we get
iá cosh /? + /? sinh (3^ e~,a = —iu>e~'a cosh /? + @
(i~) sinh/? + /?cosh ¡3^ en = - ¿ w e ' 7 sinh/? + ^ - e , a cosh/? . (11.24.255)
This simplifies to
(-iá + iui) cosh /3 + (/? - -^-e , ( a- " 1 '^) sinh/? = 0

2uj
(¿7 + iui) sinh / ? + ( / ? — ^—e%â~1'>) cosh ¡3 = 0 . (11.24.256)
2uj
We next separate real and imaginary parts and find
(• UJ .
(a — a;) cosh ¡3 + 2cj
— sin(a — 7) sinh ¡3 = 0
/ • ú .
( 7 + w) sinh /? — — sin(a — 7 ) cosh ¡3=0
lu>
P-^cos(a- 7 ) = 0 . (11.24.257)
The initial conditions are
Q
(0) = /?(0) = 7(0) = 0 .
We would have to solve these three equations, subject to these initial conditions,
if we wanted the explicit time dependence.
BIBLIOGRAPHY 231
Bibliography
[11.2] C. Aragone, E. Chalbaud, and S. Salamó, J . Math. Phys. 17, 1963,

(1976).
[11.3] H.R.Lewis Jr. and W.B.Riesenfeld, J. Math. Phys. 10 ,1458, (1969).

see also
H.R.Lewis Jr., J. Math. Phys. 9,1976, (1968).
C h a p t e r 12
Non-degenerate
Perturbation Theory
12.1 Expansion of 1 / 1 — r^l

Verify the formula
00
i ' A *
=
In - r21 ^ ^ 2/tt ¿ ,
' /=0 ^ m— — l
where
ñ = (0, <£>) , r 2 = (0', v?') .
Hint: Solve the problem
V2V»(r) =6{r- r)
by
a) expanding in spherical harmonics,
b) realizing that ip(r) is the potential for a unit charge located at r", and com-
paring the two solutions.
Solution
We begin with Poisson's equation for the potential ip due to a unit point charge
V 2 ^ = i ( r — r1) . (12.1.1)
The solution of this equation is just
= • (12-1.2)
4tt r — PI
12.1. EXPANSION OF l / | n - r 2 | 233
Now, we also solve this differential equation in spherical coordinates. For this
purpose we write
V>(f) = £ a l t m ( r ) Y , , m ( 0 , p ) • (12.1.3)
1,171
In spherical coordinates we can use the completeness of the spherical harmonics

to write
6(r- \_S(r - rO£Y,;mmm(r) (12.1.4)

l,m
where
f=(9,(p) , r> - (6',ip') . (12.1.5)
Substituting all this in the Poisson equation we get
~[rBillB(r)] - l-^-~^-ai,m(r) = ^S{r - r')Y*m{r') . (12.1.6)
Integrating this equation with rdr about r' from r' — e to r' + e we get in the
limit as e —> 0
Ida i ¡
¿2 [ra,¡m{r)]dr=
da¡,
^Yím{?)
1
.
This shows t h a t there is a discontinuity in the first derivative of a ; m ( r ) .

(12.1.7)
= • (12.1.8)
dr r'-f e
dr
For r ^ r' the radial equation has the solutions
ai:m(r) = Ai¡m rn with n = I or n = —(I + 1) (12.1.9)
so we have that
Ai<m r' r <r'

(12.1.10)
B¡ m r_'i+1' r > r'
Here we have already imposed the boundary conditions that the solution has
to be finite at r = 0 and has to vanish for r —>• oo. The solution also has to
be continuous at r = r' and to satisfy the condition of discontinuity in the first
derivative. Imposing these two conditions we get
Ai,m r'1 = Bi n (12.1.11)
or
r/ ( 2 Í + 1 )
bl^rn — a l ) t n (12.1.12)
^riAf i an 12. J V U N - D E G E N E R A T E PERTURBATION THEORY
as well as
Al¡m\-(l + iy2l+1r'-{W) ^,('-1)] = • (12.1.13)
Thus,
Al m (12.1.14)
' ~ 21 + 1 r'V+1)Yl,m^ '
So finally we get:
1 í Ei¡m2^^YiUr')Yltm(r) r < r'

(12.1.15)
l^-^l l E í i m 2 í U ^ % m m ( r ) r > r'
But, this is just the desired result.
12.2 Second Order Correction to State

In the equation for the r t h order correction to the state |m)
(0)
<m|">(r) =
„«>) p (0) [ ( 0 )
h # ' - ^l">(r_1) - 4 2 ) (0)
(m|n)(r-2)
En Em
(0)
- ^ (m|n)(r_3) - . ( ° ) { m | n ) ( 1 )
m^n (12.2.16)
set r = 2 and derive the equation
i n \(2) _ ^(m\H'\r)(°) (°)(r|#'|7i)(°)
(stf* - E{°>) - E ^ j
y--. |m)(°) '°)(m|//'|n)(°) (°)(n|.ff'|n)( 0 )

mjén ( E Í 0 ) - E£])
1 |m) ( 0 ) | ( ° ' ( m | / i ' | n ) ( 0 ) | 2

(12.2.17)
m¿n I En - Efn \
for the second order correction to the wave function.
Solution
We start with the equation (12.2.16) for the r t h order correction to a state,
namely
(°><m|n)( r ) = 1
[<°)(m|jy' - Eln\n)^ - E<2> ^ h n ) ^ 2
)
En Em
- •E'^3) (0)
(m|n)^r_3) - . . . - (°)(rn|n) ( 1 'j m^n (12.2.18)
12.3. 1/2 Ax2 PERTURBATION OF SHO 235
r
and we set r — 2. Thus, we get
(0)(m|
(°)( m |„>(2) = ^' ~ g%'n)(1) - (12.2.19)
E(n0) -
Inserting, from equation (12.2.16) the expression for ("'(rain)' 1 ), namely
m
w»>(" = "vlff'-fll-)"»
_(.) ' „(.) " " h " > * * (12.2.20)
we find, after writing everything out, the desired result, namely equation (12.2.17).
12.3 1/2 Aa;2 Perturbation of SHO

H = —H \kx2 + 7-Ax2 A: > 0

2m 2 2
a) Find the exact energy of the nth state of this Hamiltonian and expand it to
order A2 assuming |A| < k.
b) Use perturbation theory, treating (l/2)Ax 2 as a perturbation, and find the
energy of the nth state to order A2 .
c) Find a bound for the r t h order correction and hence show that the pertur-
bation series converges for |A| < k.
Hint: For part c) find a simple diagonal bound for the perturbation Hamiltonian.
Solution
a) Here
H = ~ + h x 2 + l - Ax2 *>|A|. (12.3.21)

zm z z
We now define
ft2 = ^ , lj 2 — — , q 2 — — . (12.3.22)
m m m
The exact energy eigenvalues are given by
En = {n+ l/2)hQ = (n + l / 2 ) / i \ / w 2 + a2 . (12.3.23)

If we now expand this in a binomial series in A to get the corrections due to the
perturbation we find
En — (n + \/2)hui\/\ + a2/u2
q2 a4
—• (n -(- l/2)/k¿ + +
2 ^ ~ &7 '"
<„ + i)«u, + („ + i),j l - („ + 1 ) ^ + ... . (12.3.24)

236 CHAPTER 12. NON-DEGENERATE PERTURBATION THEORY
b) Perturbation Theory
We have that
E{n0) = (n+l/2)hu . (12.3.25)
Then the first order correction is given by
£(!) = A <0><n|ff'|n)<0>
= ^(0)(n|x2|n)(0)
A fi (°'(n|a 2 + at" + 2ata + l|n)'°)

4 mw
Ah
(n+1/2) . (12.3.26)
2 mu>
So this result agrees with the exact result to this order. Next we have
|(°)(n|tf'|s)(°)| 2
E ]
n X
~ 'Y1 (12.3.27)
E(0) E(0)
s¿n ~ rj >
For s n we have
Ah
\W(n\H'\s)W '°)(n[a 2 + at" + 2ata + l|s)' 0 ^
2 mw
Ah
\/s(s — 1)¿„J_2 + \ / ( s + l ) ( s + 2 ) ¿ n s + 2 (12.3.28)
2mu
So,
p ( 2 ) _ fe"(A2/4) V ^ s ( s - 1)¿„|S_2 + ( s + l)(a + 2 ) S n s + 2 , 1 9 „ 0 Q ,
4m.2ui2 ^ hui(n + 1 / 2 — s — 1 / 2 ) '
Then,
h\2 (ro + 2 )(n + 1) n(n — 1)

m2)
16 m 2 w 3
h\2 1
(12.3.30)
8 m2u>3 " +
2
This again agrees with the exact result to this order,
c) The total Hamiltonian may be rewritten to read
H = H0 + xir = £ + !*.>+ £[!*,»] . (12.3.31)
This Hamiltonian is clearly bounded by
(12.3.32)
2m 2 k 2m 2
It is now clear t h a t
|A|H" < H for IAI < k (12.3.33)

12.4. 1/4 Ax4 PERTURBATION OF SHO 237
But, as stated, we also have
H' < II" . (12.3.34)
Therefore, the perturbation series is bounded by the perturbation series for H"
as long as |A| < k. On the other hand, the r t h order term for the perturbation
series for H" is
ATE^ = (A/Ar)r ( 0 )
H ¿ + ^x2|n)(r-1)
= (X/k)r(n+ \/2)huj '°'(n|n)(r_1)

= (\/k)r(n+ l/2)hhjST:i . (12.3.35)
From this we see t h a t we can bound the perturbation series for H' by a conver-
gent result if |A| < k. In fact,
\En-E^\ < ^^(A/fc)r(n + l/2)huiSrdr.l

r=l
= (\/k)(n + l/2)huj . (12.3.36)
12.4 l / i X x 4 Perturbation of SHO

a) Find the approximate ground state energy to second order for the Hamilto-
nian
H = + hx2 + Âx4 k > 0

2m 2 4
using the Rayleigh-Schrodinger perturbation theory.
b) Find the ground state correct to order A.
Solution
a) The Hamiltonian is
2
H = H0 + \H'= f - + h x +-Xx4 . (12.4.37)
2m 2 4
The unperturbed ground state energy is
E^ 0 ) = l-huj . (12.4.38)
The first order perturbation gives
a^1' = (o)
(0|íax4|0)w
= +«>>><«>
M.M.* M.M. -L X J I t. URATE PERTURBATION THEORY
=h
=
\ 2 mu J
4
-A
4
\ 2 raw
f—V
) (°'{0|a 2 a^" + aataatlO)^ 0 ^
(12.4.39)
The second order perturbation correction is given by

2
X2 P (2) _ A V l(°)(0|x4[n)W|2
x E
o - i«
16 2 ^ m l / 2 1 / 2 )
A2 / h | (°)<0|(qt + a) 4 |n)(°>| 2
16 \2muí^ £
n?£0
nhui
1 / \4 (\z4!¿ n 4 + 6 \ / 2 j n 2) 2
16/iu; Ix 2raw y n
' n
_ 2 i j _ / fta y
(12.4.40)
8 huí \ 2mw y
b) The wavefunction correct to order A is given by

i)(°) W{m\H'|0>(°>
10) = |0><°) + A £
0) (12.4.41)
m^n 4 - ^m 0)
But, for m ^ 0 we have
<°>Hff'|0> ( O ) = 1 (0)
<m|(at + a) 4 |0)W
1
' ° ) ( m | a t 4 + a t 2 a a t + a t a a t 2 + aa t 3 |0)^°^
4 V 2 raw
1
>/24 ¿ m ,4 6 a/2 ¿im.2 (12.4.42)
4 V 2 mw
Therefore,
|0> = |0)<°> - - ( h
4 \2mui
= |0>(O) A ft
"l6 4 ^ [l2V2 !2><°) + 2V6 |4>(°)] . (12.4.43)
12.5 1/4 Ax4 - Brillouin-Wigner Perturbation

Repeat part a of the problem above (problem 12.4) using Brillouin-Wigner per-
turbation theory.
12.6. TWO-LEVEL SYSTEM 239
Solution
To first order the calculation is the same as for Rayleigh-Schródinger perturba-
tion theory. In second order the appropriate formula is
2
a ^ w<01^|0)(°)
A
(12.5.44)
- 4 L (0) •
h
0 0 ~ en
Substituting the results from problem 12.4 this becomes
2 (o)
\ 2 p(2) (0l(at + a)4|0)(o)
A ¿
° - 4 \2muj) ¿- Eo_E(o)
A2 v4í¿n,4 + 2 + v2sn¡ 2
£0 - ^ 0 )
4 V 2raw
A2 ^4! 3>/2
(12.5.45)
4 K2mu) E0 - (4 + l/2)huj E0-{2 + l/2)hw
To this order we have that
1 3 . / fi
EQ — —huí + —A I + A2 E ( 0 2) (12.5.46)
2 4 V 2mu>
After substituting this in the equation above and rearranging we get the follow-
ing equation for EQ.
E o - ^ f k j - â
2 4 2 raw
Vi! 3v/2
= A2 + (12.5.47)
2muj Eo - (4 + l/2)hu> Eo-(2 + l/2)hv
Solving this equation to order A2 we get the same result as in 12.4 .
12.6 Two-level System

H = H0 + A H'
where
*•=(*„' l ) * ' = ( l í )
a) solve for the exact eigenvalues and eigenfunctions.
b) Solve for both eigenvalues and eigenfunctions to second order using Rayleigh-
Schródinger perturbation theory.
Solution
The Hamiltonian is as always
H = H0 + XH' (12.6.48)
where
I) • = t) <12-6-49>
Exact solutions
The eigenvalues are given by
«*{E-áa E Í - E ) = ° ' < 1 2

- 6
- 5 0
)
We assume that E2 > E\ so that \E2 — Ex \ = E2 — Ex- Then we have to solve

the quadratic equation
E2 - (Ex + E2)E 4- ExE2 - A 2 a 2 = 0 . (12.6.51)
The solutions are
E± = ¿(Ex + E2)±^s/(E2-Ex)2 + 4\2a2 . (12.6.52)
Expanding in powers of A we get
E+ = ±(E1+E2)+1-(E2-E1) + ^ 2
—
X2a2
— Ey "4" .
- E2-EX
E_ = \ { E x + E 2 ) - \ { E 2 - E x ) - - £ ^
X2a?
( 1 2 6 5 3 )
= E x - ~ ^ E [ " '

o\2~2
E+-E- = E2 -Ex
+ - — (12.6.54)
E2 — Ex
so that the two energy levels are "repelled" by the perturbation. This result is
quite general.
The eigenvectors are of the form
* = ( 0 ) (12.6.55)
where
[Ex — E±)a± + iXa0± = 0 (12.6.56)

12.6. TWO-LEVEL SYSTEM 241
so that
a± = — ^ f 3 ± . (12.6.57)
hi — h±
We can now write
—i\a
(12 6 58)
e 7 - E ± ) - -
where, with an arbitrary choice of phase
1
M i l - — — — (12.6.59)
1 ±!
OT
Now, to later compare with the perturbation theory we let A —>• 0 and find t h a t
2 2
Ei — E+ —(E 2 — Ei) — A a
E2 — Ei
(12 6 60)
E i - E - ^ ^~É[ ' '
A2 a 2
A+ —y (E2 — Ei) +
2(E2-EI)
A. -> Aa . (12.6.61)
So, with an arbitrary choice of phase we have
i\a
1
^ ^ j ^ j . (12.6.62)
Perturbation Theory
The Oth order eigenvalues and eigenvectors are
Ei(0) = Ei , E2(0) = E2 (12.6.63)
and
ô) = (I) ^ = (5) • (i2-6-64>

The first order correction to the eigenvalues is zero since
0) 0)
(</>! , H'^) = (v4 , H'^) = 0 . (12.6.65)
The first order correction to the wavefunctions is
^(1) _ v , 2° ) (v'2 0 ) .^ / v , i 0 ) )
1
Ei{0) - E2{0)
— ia ( 0
Ei — E2 \ 1
^(i) _ ip[0) (Tfi{0), H'IP{20))
2
^ E2w _ Ei(o)
ia ( 1
(12.6.66)
£i-£2 v 0
So, to first order we have
ipi -
1
iXa
E-2 — EI
o i\a
1p2 :
+
1 / E2 — Ei
iXa
E? — Ei (12.6.67)
1
To order A we now have agreement with the exact solutions.
The energy to second order is
(2) _
E{ ' p(0) (0)
1 "^1 E. £1 — E2
(2) _
= £ (12.6.68)
J?(Q) (0)
£"2 — Ei
n¿2 2
Therefore, the two energies to order A2 are
A2 a 2
£•_ = £"! -
E2 — Ei
2 2
E+ = E2 + A a (12.6.69)
£2 — Ei
These results again coincide with those obtained by expanding the exact solu-
tions in powers of A.
12.7 Approximate SHO

A particle of mass m moves in a potential
V=~k\x\2+Í |e| < 1 .
Estimate the energy of the ground state.

Hint:
-k\x\2+e = i k x 2 - ^ k ( x 2 - | x | 2 + f ) sa ^kx2 + ^kx2 In \x
Also,
where
/
Jo
x2 In x dx —
4
a 3¡2 1 — - ( c + ln4a)
c = 0.577216 . . . = Euler's constant .

12.8. TWO-DIMENSIONAL SHO 243
Solution
Using the first hint and writing, £is suggested,
^k\x\2+e = ^kx2^kex2 ln\x\ e« 1 (12.7.70)
we immediately find t h a t the ground state energy may be approximated by
E0 = + lfce(0|a; 2 In |x|0) (12.7.71)
where
u¡2 = k/m (12.7.72)
and |0) is the corresponding simple harmonic oscillator ground state. Thus, we
need to evaluate
(0|x 2 In |z|0> = f x2 In \ x \ d x . (12.7.73)
Now using the second hint
/
/ = 4 -3/2
a~ax2x2 In |z| dx = —— [1 - l / 2 ( c + ln(4a))] (12.7.74)
Jo a
where c = 0.577 216 . . . = Euler's constant, we get
€
E0 - ^hui 1+ (12.7.75)
2v/2
12.8 Two-dimensional SHO

For a particle of mass m moving in the potential
V = ^kix2 + ^k2x2 + Axy |&i - k2\ > 2A
a) find the exact energy levels.

b) Use perturbation theory to find to order A2 the energy of all the levels and
compare with the exact solution to this order.
Solution
The Hamiltonian H is
H
= 7T- + v(x, y) = + \kxx2 + + \k2y2 + A xy . (12.8.76)
2m 2m 2 2m 2
This can be rewritten as
H — ~— -f l~muji2x2 + ^— + l-mu22y2 + ma2xy , (12.8.77)

2 ra 2 2 ra 2
•,-unKj.mWKATE PERTURBATION THEORY
where
wi2 = ki/m , u22 = k2/m , a 2
-\/m. (12.8.78)
a) Exact Solution
We now rotate the axes to remove the cross term.
x — X cos 9 + Y sin 9
y = - X s i n 0 + Ycos0 (12.8.79)
Then,
— = u\ [X2 cos 2 9 + Y 2 sin 2 9 + XY sin 20]

m
+ u>$ [X2 sin 2 9 + Y 2 cos 2 9 - XY sin 20]
+ a2 [—X2 sin 29 + Y 2 sin 29 + 1XY cos 20] . (12.8.80)
So, to remove the cross term, we choose the angle 9 to be given by
(w2 — w 2 ) sin 29 + 2 a 2 cos 29 = 0 . (12.8.81)
Also, since this is a rotation we have
f l + ü l - Ex. + (12 8 821

2m 2m 2m 2m
" é, é
So.
= + +
lmí>'x'+12ms>'y2 <12'8-83»
where
Í22 = w 2 cos 2 9 + w 2 sin 2 9 — a2 sin 29
= I (wi + | \J{"1 - )2 + 4 a 4
a4
—^ Wi -(- — ^ ~ as Of — 0 .
w2 -
fi2 = w 2 cos 2 9 + sin 2 9 + a2 sin 20
=
2 ( w i + wi>) + 2 \ / ( w 2 - ^ i ) 2 + 4 q 4
, a4
^ ^2 — —2 2 ^ ® ^0 • (12.8.84)
u>1 — w 2
The (exact) energy eigenvalues are
E
nitm = (m + l/2)/iQi + (n 2 + 1/2)htt 2 • (12.8.85)
So, to lowest order in A, for later comparison with perturbation theory, we have
=
£«i,ií2 ini + l/2)hui\ + (n2 + l/2)hu>2
h\- (n2 4- 1/2)î - (ni + \¡2)u2 (|9 8 8fii
+
w ™(w2-w2) ' (12 8 86)
- '
12.8. TWO-DIMENSIONAL SHO 245
b) Perturbation Theory
The Hamiltonian H0 represents two uncoupled SHO's. The energy eigenvalues
are
£ ( ° ) n 2 = ( n i + 1 / 2 ) f k j 1 + (n2 + l / 2 ) f t u 2 . (12.8.87)
Also, the eigenkets may be written
|„ „ 2 )(0) = ( ° í r ( 4 ) " a |0|0)(0) (12.8.88)

V n i ! n2!
The first order correction to the energy is zero since
(°)(ni,n2\xy\ni,n2)'0) = 0 . (12.8.89)
In second order we get
E (12.8.90)
£j
ml9in1,m2?£n2 n1,n2 ^mum 2
Now,
n2\xy\mu m2)(0)
(0)
= ("i,n2| 7 T — 7 = = ( a I +
a
i)(a2 + a2)|mi,m2)(0)
Zm y/u)\u)2
= [ ( 0 ) ( n 1 , n 2 | \ / ( m i + l ) ( m 2 + l ) | m i + l , m 2 + 1) (0)
(0)
+ ( n i , n 2 | \ / ( m i + l ) m 2 | m i + l , m 2 - 1) (0)
(0)
+ ( " i , n2\\/mi(m2 + l ) | m i - 1, m 2 + 1) (0)
+ ^ 0 ) (ni, n 2 | 1 / m i m 2 | m 1 - l , m 2 - l)^°'j
+ 1 )(m 2 + 1) ^1,1111+1 ^n2,ra2 + l

2m ui2
+ \/(mi + 1 )m2Sni
+ > / m i ( m 2 + 1) Sni
,mi —1 ^ri2,"i2 + l
-j- ^ni,nii-l ^n 2 ,m 2 -l] • (12.8.91)
So, after collecting terms we get
z?(2) _ ^ ( n 2 + l / 2 ) w j — (ni 4- l/2)w 2

Ml,Mo r> 9 / 9 9\ (12.8.92)
cjicj2(^2 ~~ )
This is the same result that we obtained by expanding the exact solution.
12.9 Kuhn-Thomas-Reiche Sum Rule

Classically the polarizability a of an atom is defined as the induced electric
dipole moment e\r\ divided by the strength of the inducing electric field É. So
a = —
\E\
and for harmonically bound electrons takes the form
e2 ^ fj
a =
An2m Ei/? Jj u -
2
'
3
j
Here, f j are dimensionless constants called the "oscillator strengths". In quan-
tum mechanics these are defined by
47rm „ ,
fj = 3^2ô|mi0|
47rm Ei — EQ , _
1 1 9
= sV "»
For N uncoupled electrons one then has the Kuhn-Thomas-Reiche sum rule
j
This polarizability can be used to describe the absorption of light which
carries an electron from its ground state |0) to an excited state |n) in an atom.
If, ^no = (n|r|0) and the Hamiltonian for the bound electron is
a) Show t h a t
(En - E o ^ n o = - — (n|plO) .
m
Hint: Use the commutator [H, r] and work component by component.
b) Use the commutators
[x,Px] = [y,Py] = [z, Pz] = ih

together with the results of part a) to prove that for a single electron
2 m e
£ ' ^ » ' l k „ 0 p + |to0|>+,2„„n _ i
n
and hence deduce the Kuhn-Thomas-Reiche sum rule.

For this problem and all subsequent problems dealing with sum rules it may
be useful to consult [12.11.
¡2.9. KUHN-THOMAS-REICHE SUM RULE 247
Solution
We start with the Hamiltonian
H = f - + V(r) . (12.9.93)
2m
Then,
2h
(12.9.94)
and similarly for [y, H] and [z, H\. Therefore,
— (n\px\k) - (n|[x,//]|fc)
m
= (Ek - En)(n\x\k) . (12.9.95)
Hence, we have
ih
- < 0 | P x | * > = (E„ - £o)<0|x|*> (12.9.96)
771
and by complex conjugation
--(k\p,\0) = (Ek-Eo){k\x\0) . (12.9.97)

m
- £o)|x f c o | 2 = £ ( £ * - £o)<0|x|fc><*|s|0>
k k
= L (0|p r |fc)(fc|x|0)
—' m
k
= —(0|p x x|0) . (12.9.98)
m
Similarly,
Y,(Ek - E0)\xk0\2 = £ ( £ * ~ £o)(0|x|fc)<*|x|0>

k k
_ j2—(0\ \k)(k\px\0)
z
—' mx
k
= -^(0|xpr|0) . (12.9.99)
Therefore,
^T(Ek - E0)\xk0\2 = ^ ( 0 | p x X - xp x |0)

k
ti2
This means that for a single electron

2 rr,
- E0)\xk0\2 = 1 . (12.9.101)
fc
Hence, since all three directions x, y z are equivalent we have for a single electron
that
J 2 ^ { E k - £ o ) [ m 2 + ko|2 + ko|2] = 1 • (12.9.102)

k
For N independent (uncoupled) electrons we therefore find the Kuhn-Thomas-
Reiche sum rule
£ ~(Ek - E0)[\xk0\2 + I»o|2 + \zko\2} = Y , h = N

• (12.9.103)
k k
12.10 Electron in Box Perturbed by Electric Field

An electron confined to a one-dimensional box 0 < x < L is acted on by an
electric field £ acting in the x-direction. Assuming that eSL « energy of the
lowest unperturbed level, find the lowest energy in the presence of the electric
field £ .
Solution
The unperturbed energy for the ground state is
i 2 1 0 1 0 4
• ( - )
The corresponding ground state wavefunction is

l~2
ipo = y — sin(7rx/¿) (12.10.105)
The perturbation Hamiltonian is

V = e£x . (12.10.106)
Therefore, the energy correction is
E^ = (rp0,vip0)
2 fL
= —e£ / sin2(irx/L) xdx
Jo
= ê£L . (12.10.107)
Thus, to this accuracy, the lowest energy is
=
(12.10.108)
12.11. POSITRONIUM 249
12.11 Positronium
Positronium is a hydrogen-like system consisting of a bound state of an electron
and a positron (positive electron). The ground state consists of a singlet and
three triplet substates. The singlet state is the most stable lying about 8.2x 1 0 - 4
eV below the triplet levels which are degenerate. Field theoretic calculations
show that this is due to a spin-spin interaction of the form
A
h 0 = - t x s i • «2 •
a) Determine the value of the constant A.

b) Using the fact t h a t the positron has a charge and magnetic moment opposite
to that of the electron calculate the effect of a magnetic field on these levels.
Solution
a) The Hamiltonian for these four levels, in the absence of a magnetic field, is
as given
tfo = - ^ s i - s 2 • (12.11.109)
If we choose the representation of good total spin | s , m ) where
(s,m) = (1,1); (1, 0); (1, —1) and (0,0)
then, in terms of the spin states for the electron and positron, respectively, we
have
(12.11.110)
Also, since
si • s 2 = 1 [(si + s 2 ) 2 - « i - ® ! ] (12.11.111)
we see that Ho is diagonal in this representation.
/ 1 0 0 0 \
_ a 0 1 0 0
0
(12.11.112)
~ 2 0 0 1 0
\ 0 0 0 -3 /
Therefore, the splitting of these levels is
A E = j[l - ( - 3 ) ] =2A = 8.2 x 10" 4 eV . (12.11.113)

So, we have
A = 4.1 x 1 0 - 4 eV (12.11.114)
b) If we introduce a magnetic field B pointing in the z-direction then the Hamil-

tonian becomes
A _ eh
H = H0-fi-B = s i s 2 +
B • (<?i - ¿ r 2 ) (12.11.115)
^ " 2 ^
where <?,• are the Pauli matrices and the index 1 refers to the electron and the
index 2 refers to the positron. Calling
eB
= u>
2mc
we find that in the representation already used we get that
(00 0
0
0
2
0 \
0
—¡I • B = hw (12.11.116)
0 0 0 0
1 0 2 0 0 )
The eigenvalues of the total Hamiltonian are therefore (listed in the order in
which the states are listed in (12.11.110))
A/2 , A/2 + 2hu> , A/2 , A/2 - 2hui.
12.12 Rigid Rotator in Electric Field

Consider a three-dimensional rigid rotator with moment of inertia I and electric
dipole moment P parallel to the axis of the rotator. This rotator is placed in
an uniform electric field E. Compute, to lowest non-vanishing order in E, the
strength of the electric field, the ground state energy of the rotator.
Solution
The unperturbed Hamiltonian is
L2
Ho = — . (12.12.117)
The energy eigenvalues are

E¡ = l(-l+^h (12.12.118)
The interaction Hamiltonian, if we take the axis of the rotator parallel to the
z-axis, is
H' = -P • É = -PEcosd . (12.12.119)

12.13. ELECTRIC DIPOLE MOMENT SUM RULE 251
The effect of this perturbation is to produce (in lowest order) a change in energy
£;(1) = -PE{lm\cos 9\lm) . (12.12.120)
This yields no change. So we have to go to second order. In this case, the

ground state is shifted by
= . (i2.i2.i2i)
lm
The only non-zero matrix elements in this sum are
(00|cos#|10) = (10|cos#|00) . (12.12.122)
This illustrates the selection rules
A/ = 1 , Am = 0
that apply to electric dipole transitions. Using the fact that
I 4tt 1
cos# = \ — Yio and YQO
3
we get
(001 cos #| 10) = (10| cos#|00) = 4 = . (12.12.123)

V3
So, finally
e(2) _ _(pE) (12.12.124)
12.13 Electric Dipole Moment Sum Rule

Show that for a system of N particles with charges qj and masses Mj, j =
1 . . . N confined to a finite region of space we have the following sum rule [12.1]
for the electric dipole moment.
00 n
2 - a?
~ En)\dnm\2 = ^2 .
J
m j =1
The sum here extends over a complete set of energy eigenstates.

Solution
For any operator A the matrix element of the time derivative of the operator is
given by
(n\Á\m) = - j:<np, H]\m) (12.13.125)
where H is the total Hamiltonian. Applying this to the operator d we find
d„m = -TH[d,H]\m)
(n\dH\m) — (n\Hd\m)
:Êm - En)d (12.13.126)

nm -
From this it follows that
OO CO
y ](Em — En)\dnm | = y ^\Em — En)dnm •d

m m
ih
m ' dmn dnm • dmn
ih
(12.13.127)
— 2 Id d d-d |nn .
Now, using (12.13.125) again on the operator d we get
ih(n\d\m) = (n|[c£ 7/]|ra)
(12.13.128)
J
j
So,
ih
~ En)\dnr 2 53 (d)mn (d)nm ' (p))mnj
m j
=
J
3
<ñ (12.13.129)
Mi
as required.
12.14. ANOTHER SUM RULE 253
12.14 Another Sum Rule

Use the double commutator
1ri
Ee_i -iqrk
to derive the sum rule (see [12.1]
£ ( £ m - En) <n|£e«^| m) = N ML
2m
for a system of ./V interacting particles. Here |n) represents an energy eigenket
of the Hamiltonian H .
Solution
If we first write out the double commutator explicitly we find
a, £ ^
3 k
e 'q-rk
H - £ ^ ' f jH • £ e
~''4
j j k
= H ^2 lq r
e ' '> e ig r
~ ' '" - £ eq r J H
'~~ £e~ S-'irk
j k j k
e _ i f r% H e< e_,í r
~ £ £ * + £ " £ e'V' r"J H • (12.14.130)
k j k j
Next, we take the diagonal matrix elements (in the energy representation) of
this expression and insert a complete set of intermediate states.
(n | |n>
e
= En(n\^ J2 "q'fk + £ e iq fk
~ ~ £ ^ \ n
)
j k k j
- h e |m)(m| Y . z ~ i q f k i")
m
\ j k
+ <«l £ e~i9'r~k l"»><m|(n| £ |n> 1

= 2yi(en~e m ) {n\ Y , Im) (12.14.131)
On the other hand, if we evaluate the double commutator we find
iqrk
IIY.'-"" • £ « -
+e<i
= ^-E jk
áhq)2
- - 2 - Y e"q'fk ¿
2M
jk
= -2#" - (12.14.132)
2m
This proves the desired result.
It is worth noting t h a t if we make the dipole approximation by putting
e'^'rj fa 1 + iq • f j (12.14.133)
we get the result derived in the previous problem.
12.15 Gaussian Perturbation of SHO Bosons

Two identical bosons move in the one-dimensional simple harmonic oscillator
potential
V= imw2 (x\ + x\)
and also interact with each other via the potential
Vint = V0e-a(x'-Z2)2 Q>0.
Find the ground state energy correct to first order in Vo-
Solution
The unperturbed ground state of the two bosons is
rP(Xl,x2) = M*i)Mx2) (12.15.134)
with energy
£"o = 2 x — hu> = huí

12.16. GA USSIAN PERTURBATION OF SHO FERMIONS 255
Here
«•>=(&r-»(-£•') • <i2i5i35>
Also, the wavefunction ip(xi,x2) is already properly symmetrized. The energy
shift A E due to the perturbation is given to lowest order by
AE = (tp,vintip)
V (mu)Y/2 oo
/
e_i-(xj+x»)e_a(xi_a;3)2 d x i d x 2 (i2.i5.i36)
We now change variables to •oo
d *1+2:2
R =
~~2
r = x!-x2 • (12.15.137)
The Jacobian of this transformation is 1. Therefore,

/mu\ 1/2 [°°
A E = Vb
\2nhJ / oo ar
e~ exp ^j^-(2R2 + r 2 / 2 ) j dRdr
1 2
/mwN / / 2irh \ 1/2 / nh \1/i2
\2nh) \Kmuj + 2ah) 1^2 muj)
2
= % ( '"o y • (12.15.138)
2 \mio + 2ahJ
12.16 Gaussian Perturbation of SHO Fermions

Two identical spin 1 ¡2 fermions move in the one-dimensional simple harmonic
oscillator potential
V = ^mw2 (x\ + x\)
and also interact with each other via the potential
vint = V0e~a^-x^ a > 0 .
a) Find the ground state energy correct to first order in Vo for the case of the
singlet spin state.
b) Find the ground state energy correct to first order in Vo for the case of the
triplet spin state.
Solution
a) In the singlet case, the spin wavefunction is antisymmetric and the space
wavefunction is therefore symmetric in the interchange of the two coordinates.
Thus, this case is identical to the case of two bosons discussed in problem 12.15.
—- i vyj y xtlüUHY
b) In the triplet case the spin wavefunction is symmetric so the space wave-
function must be antisymmetric. This forces one of the particles to be in the
first excited state. Therefore, the unperturbed spatial wavefunction for the
ground state of the two fermions is
ip(xi,x 2 ) = -^= [<po{xi)<j)i(x2) - <l>i{xi)(j)o(x2)] (12.16.139)
with energy
1 3
eQ — ~ huí -(- ~ hijj - 2 huí
Here
, „ /mui\1/4 ( mu ,\
e,p r
*><*> = (s») ra )
<l>i{x) = ^fap-x<t>o(x) . (12.16.140)
The energy shift A E due to the perturbation is now given to lowest order by
a x x
AE = Í e- ( i- 2) [tj)l(xi)<f>21(x2) + <j>\(xi)(j)l(x-i)
¿ J — OO
- 2<l>o(xi)<f>i(x2)<j>i(xi)<j)o{x2)] dx\dxi . (12.16.141)
We now again change variables to
D XX+X2
r =
- y -
r = Xl -x2 . (12.16.142)
The Jacobian of this transformation is again 1. Therefore,
. „ vo / m w \ 1 / 2 2 m w f°° ,
ae exp
= t(s») —J_x' (—ir(2B + r / 2 ) ) r dRd
'
\2ivh) \2mu J 27t \ m w + 2ah J
12.17 Polarizability: Particle in a Box

For a particle of mass m and charge e in its ground state when confined to a
box —a <x<a,—b<y<b,—c<z<c find the electric polarizability.
Hint: The polarizability a is obtained from the shift in energy
A £ = ~âS2
when the particle is placed in an electric field S.

12.17. POL ARIZ ABILITY: PARTICLE 11M A m y A
Solution
If we take the electric field pointing in the x-direction then the perturbation is
V = -eSx. (12.17.144)
The problem is now, for all practical purposes, one-dimensional. The ground
state wavefunction is
^j+ 0 Q = (afcc) -1 ^ 2 cos(7rx/2a) cos(ny/2b) cos(nx/2c) . (12.17.145)
Here the superscript -f refers to "positive parity". The corresponding ground

state energy is
= + • (12.17.146)
To lowest order the perturbation has no effect since
« 0 , 0 . ^V>o+o,o) = ° . (12.17.147)
Therefore we require second order perturbation theory. This gives
( ^ 0 , 0 ' ^ v'n,0,Cl)(v'n,0,0 ' ^V'o^Q.o) ^

E _
n*0 0,0,0 ^n.O.O
where we have included only the non-vanishing matrix elements and the super-
script " refers to parity = — 1.
^)~ 0 0 = (o6c) - 1 ^ 2 s'm(mrx/2a) cos(ny/2b) cos(nx/2c) (12.17.149)
as well as
h2ir2 (An2 1 + 1 1 2 I 7 I 5 0 )
= • <
Thus, we need the matrix element

1 f
M = — I x cos(7rx/2a) s'm(nirx/a) dx
a
Ja
1 f 1 f
= — / x sin[(n + l/2)7ra:/a] dx + •— I x sin[(n — l/2)7rx/a] dx
2a J a 2a J a
+ 1
= (~1)" ^ " , • (12.17.151)
7x¿ (4n ¿ — \) ¿
Also,
4 o , o - ^ o , o ^ 2( l - V ) • (12.17.152)
8 ma
Combining these results we obtain
After equating this to the expression for the energy shift in terms of the polar-
izability we find that the polarizability is given by
4096 e 2 ma 4 n2
a =
~ ? ^ (4n — l) 5 •
2
(12.17.154)
n= 1 '
12.18 Atomic Isotope Effect

Every nucleus has a finite radius R = r0Al¡3 where
13
r o = 1.2 x 1 0 "
and A is the atomic number of the nucleus. Thus, the potential energy experi-
enced by an electron near a nucleus is not simply
V(r) = - ~ •
r
If we assume that the charge density in the nucleus is constant then we have
instead the potential energy
( Z¿_ [ r2 _ 3
/ { 22
v ( ) = {J
r R
\ l[22R 22j r
~ r • (12.18.155)
I r > R
a) Use perturbation theory to calculate the isotope shift, that is the depen-
dence on A of the K-electron (Is state) for an atom with Z protons and atomic
number A.
b) Use this result to compute the energy splitting for the K-electron between
the heaviest lead (Z — 82) isotope A — 214 and the lightest A — 195.
Neglect the presence of the other electrons.
Solution
a) The unperturbed Hamiltonian is
p2 Ze2
— - (12-18-156)
The perturbation is
f \j±. _ 3 , «] < D
= < r w 2 + r j r
s « (12.18.157)
i 0 r> R
The unperturbed ground state energy of the K-electron is
E 0)
° = - ¿ $ (12-18158)
12.18. ATOMIC ISOTOPE EFFECT 259
where a = 5.292 x 1CT9 cm is the Bohr radius. The corresponding wavefunction

is
(12.18.159)
The first order correction to E ^ ' is given by
E(01] = (40), H ' f f l ). (12.18.160)

Thus,
(1) i ( 2 z \ ' z ¿ r' -2Zr/a f _ _ 3 R
0 +
r2 dr (12.18.161)
2 V a J R Jo 2R? ~ 2 r
We now let
2 ZR 2 Zr
a = (12.18.162)
Then,
(i) 1 Ze2 r _x x2 3 a
r2 dx
K 6
2 R Jo 2a2 _
2 +
x
1 Ze 2 2 f 12 „ 12 12 „
3 + q —x 4 1-3 (12.18.163)
2 ^ z a r ~
If we now make the dependence on the atomic number A explicit by writing
a = _ 7j4i/3 (12.18.164)
a
we have the desired dependence on A.
(i) _ 1 Ze2 2 .-1/3 12 ^ - 2 / 3 _ 3 + 7j4i/3
E
° ~ 2 a/Z t l72
_ e - v i " 3 ( 1 | ¿ - 2 / 3 + h a - i / 3 + 3n (12.18.165)
\7 7 >
b) If we take Z = 82 and >1 = 195 we get that a = 0.238. Substituting these
values we find that
E^ 1 ] (A = 195) = x 9.91 x 10" 3 . (12.18.166)

¿ CL ¡ ZJ
Similarly, for Z = 82 and ^4 = 214 we get that a = 0.245 . Thus, repeating the
calculation we find that in this case
E ^ ( A = 214) = x 1.08 x 10~ 2 . (12.18.167)

Í a/Z
Thus, recalling t h a t
1 e2
- — = 13.6 eV
¿ a
the energy difference in energy of the K-electrons between the two isotopes is
A E = 1 Ze2 x 9.4 x 1 0 - 4 = 86 eV (12.18.168)
2 aj Z
12.19 Relativistic Correction to H atom

The kinetic energy for a relativistic particle is
2
2 \
T = VW2 + «V " ^ (12.19.169)
Use the last term as a perturbation to calculate the first order correction to the
energy levels of a hydrogenic atom.
Hint: The following expectation values for a hydrogenic atom may be useful
_1 1_
(1/r) =
n2 a/Z
2
(l/r
K
) 1
= — (12.19.170)
' n3(l + 1/2) ( a / Z ) 2 '
Here, a is the Bohr radius.
Solution
To save labour we use the fact that
(12.19.171)
2m 2 (a/Z) n 2
Thus, the first order correction is given by
i ,2n 2 >
(P)
E
n \
n ,1 2 mc 2 2m
1 1 Ze2 1 Ze2Y
2
2 mc2 2 (a/Z) n r J
1 1 Z2e4 1 Z2e4 1
(l/r) + Z2e2(l/r2)
2 mc2 2 4
.4 (a/Z) ^ ~ (a/Z)
Z2eA 1 /zn 2
1 Z\2 1 1
2 mc2 4 v7 a ,/ n^ 4 +' V a J n3(l+ 1/2)
z 22e„ 4 1
(12.19.172)
mc' \ a 8 n4 n3(2l + 1)
12.20 van der Waals' Interaction

Two widely separated hydrogen atoms interact via a dipole-dipole interaction
whose potential, known as the van der Waals potential, is given by
12.20. VAN DER WAALS' INTERACTION 261
where R is the separation of the centres of the two hydrogen atoms and f\ =
( * 1 , 2 / 1 , 2 1 ) and r*2 = ( # 2 , 2 / 2 , 22) are respectively the coordinates of the electrons
associated with atom 1 and 2. Using perturbation theory calculate the interac-
tion energy of two widely separated hydrogen atoms. As an approximation for
the unperturbed energies use
Ei - En « Ei .
Solution
The total Hamiltonian for the two hydrogen atoms is
H = H01 + H02 + V (12.20.173)
where
n-2 7e2
Hoi = — ¿ = 1,2 . (12.20.174)
2m r,-
The unperturbed energies for HQÍ are
e2 1
En = =• (12.20.175)
ao n¿
with the corresponding eigenstates |n¿,/¿,m¡). We are interested in the energy
shift of the ground state
| V ( 1 , 2 ) ) = | 1 i , 0 1 , 0 1 ) | 1 2 , 0 2 , 0 2 > = |1)|2) . (12.20.176)
To lowest order this energy shift is
ew = (^(1,2)1^(1,2))
= ^ [ ( l | « 1 | l ) ( 2 | « a | 2 ) + ( l | W | l ) ( 2 | i f t | 2 ) + (1|zi|1)(2| 2 2 |2)]
= 0 . (12.20.177)
So we have to go to second order. In this case we find
e(2) _ y- 1(^(1,2)|V>1,/i,mi)|n2,Z2,m2)|2
ni,n«il 2Ei- Eni- Eni
~ £ <V , (l,2)|V / |n 1 ,/i,mi)|fi2,/2,m 2 )

«1 .«2^1
x (n1,li,m1\(n2,l2,m2\V\i(l,2))
= ¿-<V(1,2)|K2|V(1,2)> . (12.20.178)
Here we have used the indicated approximation to go to the second line and the
completeness relation to obtain the last line. These matrix elements are now
easy to evaluate using the symmetry of the ground state wavefunction. In fact
the only non-zero matrix elements are
(i|*?|t) = (»|j/i I») = (i\zf\i) = . (12.20.179)
But,
4 r°°
<í|r?|¿> = "3
a
/ e~*r'a°r4dr
o Jo
= 3a20 . (12.20.180)
Thus, we finally have the desired result
,,1 3 e4<Zo
£<" = • (12.20.181)
Bibliography
[12.1] R.Jackiw, Phys.Rev. 157, 1220, (1967) .
C h a p t e r 13
Degenerate Perturbation
Theory
13.1 Stark Effect for n = 2 Level in H

Find the shift in the energy of the n = 2 levels of a hydrogen atom, to first
order due to a constant electric field (linear Stark effect). The potential is
V' = —eE • r = — eSz .
Solution
We first write out the n = 2 levels for the hydrogen atom in units such that
length is scaled by the Bohr radius.
^2,0,0 = -^L(l-r/2)e"r/ 2
1 r 2
^2,1,0 = — re / cos#
VÍÜñ
1 r/ 2
^2,1,1 — re ' sin 9 e
\/647r
1
^2, i , - i — r e r / 2 sin 9 e~ (13.1.1)
\/647r
The matrix elements of the perturbation in this degenerate subspace are
(ip2,i,m\eE r cos9\rp2,i',m') •
The only non-zero elements are between ^2,0,0 and ip2,i,o- Thus, the matrix to
be diagonalized is
/ 0 -3 0 0 \
-3 0 0 0
eSa
0 0 0 0
\ 0 0 0 0 /
264 CHAPTER 13. DEGENERATE PERTURBATION THEORY
Here we have restored the usual units so t h a t the Bohr radius a is no longer of
unit length. The eigenvalues of this matrix are +e£a, —e£a, 0, 0. Therefore
the degeneracy of two of the states is lifted in first order and the new energies
are + e£a, E^ - e£a, E2°\ and E^ where
2
E{20) = ~ • (13.1.2)
13.2 Perturbation of Particle in a Box

A particle is in a two-dimensional box of sides a. If a perturbation
V' = Xxy
is applied, find the change in the energy of the first excited state to first non-
trivial order.
Solution
The eigenfunctions for the particle in the box are given by
2 . nnx . knu
4>n,k = - s i n sin 0 <x,y<a. (13.2.3)
a a a
The corresponding energies are:
e
h2 7T2 ,
+ t )2
" ' * = w < " - i132-4'
This shows that the ground state with n = k = 1 is the only nondegenerate state.
All other states are clearly degenerate. The first excited state is two-fold de-
generate. The two degenerate states are |1,2)(°> and |2,1)( 0 '. The perturbation
Hamiltonian in the degenerate subspace has matrix elements ^(i,j\xy\i,j)^°\
When this is written out we get
/ a 16a \
9
H ' = ( xL f ) • (13.2.5)
\ 97t2 2 /
The corresponding eigenvalues and eigenvectors are respectively
1 _16_ 1 _
(13.2.6)
2 + 9tt 2 2 9tt 2
and
v§( 1 ) ' v2 ( - 1 ) ' (13.2.7)
Thus, instead of |1,2)^ 0 ' and |2,1)'°' we begin with
|+) = ~ = [¡1,2)(°) + |2,1)(°)] (13.2.8)

13.2. PERTURBATION OF PARTICLE IN A BOX 265
and
|_) = - l [ | 1 , 2 ) ( ° ) - | 2 , 1 > w ] (13.2.9)
in our perturbation calculation. The first order correction is given by A times

the eigenvalues above. The second order terms are given by
|(±|xy|n,Ar)(°)| 2
p(2) _ v \2 (13.2.10)
±
+k¿3 (h2n2)/(2ma2)(b — n2 — k2)
Now,
. irx . 2ny
(±\xy\n,k)(0) = [ dx dy sin — sin
a
Jo a a
. 2nx . iry nnx . kivy
± sin sin — xy sin sin (13.2.11)
a a a a
But,
2 fa . nnx . mn.
—2 / xdx sin sin
a J0 a a
J_ f xdx .cos (n — m)nx cos
(n + m)nx.
a2 Jo
1
f x r a . (N — m) TVX
d, sin
= —
a bn m r~
Jo ( ~ )*"
a , . (n + rn)irx.
- d sin
(n + m)n a
1 f . (n — m)7tx
7 i— / Sln dx
[
(n - m ) 7 r a J0
a
1 . (n + m)nx ,
sin dx
(n + m)na J0
1 1
[(-i)»-™ - 1 ] [(-1) n + m 1]
(n — m)2n2 (n + m) 2 7T2
1 Anm
= [(-i)"+m - 1 ] 2 2 (13.2.12)
7T (n — m 2 ) 2
With this result we find
[(-l)"+2-l] [(-l)*+i-1]
(13.2.13)
(n 2 — 4) 2 (k 2 — l ) 2
Substituting this in the sum for the perturbation series we have the desired
result.
266 CHAPTER 13. DEGENERATE PERTURBATION THEORy
13.3 Perturbation of Isotropic Two-dimensional

SHO
For the two-dimensional simple harmonic oscillator with
Ho = ftw(a|ai + a\a2)
calculate the effect, to second order, of the perturbation
H' = AíajaíaiOi + a\a\a2a2)
on the second excited states and to first order on the third excited states. What
are the effects on the ground state and first excited states?
Solution
In this case we have
H0 = hu> + ala2) (13.3.14)
H' = A â{a{a 2 a 2 + a ^ a j O ! j (13.3.15)
and
Ho\n!,n2)W = (ni + n2)hoj\ni, n 2 )' 0 ) (13.3.16)
This shows that all states except the ground state are degenerate. For the
second excited state we have three degenerate states:
|2,0)w |1,1>(°>, |0,2)(°) .
All have the same unperturbed energy 2hw. In this degenerate subspace, the
perturbation Hamiltonian is
/ 0 0 2A \
H' = 0 0 0 . (13.3.17)
V 2A 0 0 /
The eigenvalues are
2A , 0 , —2A .
The corresponding eigenvectors are respectively
—= 0 (13.3.18)
V2 v2 \ j
l3 4. TWO-DIMENSIONAL SHO WITH OFF-DIAGONAL TERM 267
Thus, we have as new basis states in this degenerate subspace
IV-2.+) = - U | 2 , 0 ) < o > + |0,2)(°>]

V2
= |i,i> ( 0 )
= 71 | 2 , 0 ) (o)
-|0,2)(o) (13.3.19)
Therefore, writing r = +1, 0, — 1 we find
ni+n3ji2 \ l~r
— Ihoj + r2A + 0 to second order in A . (13.3.20)
The third excited states are

|3,0)(°> , |2,1)(°> , |1,2)(°) , |0,3>(°) .
The perturbation Hamiltonian in this degenerate subspace is
/ o o o \
H' o o o Vñ\ (13.3.21)
v/l2A 0 0 0
V 0 VÍ2X 0 0 /
The eigenvalues are:
0, ,0, \Zl2X, and, —Vl2\ .
Not all the degeneracies are lifted. For this reason we only compute to first
order in A. To this order the eigenvalues are
E 3iT = 3 huj + r\/l2A (13.3.22)

where r = 0, 0, ±1.
13.4 Two-dimensional SHO with

Off-diagonal Term
a) Repeat the previous problem with
H' = A[(aja 2 + <4 a i)] •
b) This problem can also be solved exactly by introducing operators
Ai = a j cos 9 + a 2 sin 9
A2 = — ai sin O + a.2 cos 9
and choosing 9 appropriately. Do this and compare with the perturbation result.
Solution
a) Again we have
H0\ni,n2)W = (ni + n2)Hui\ni,n2)('0) . (13.4.23)
We first solve for the eigenvalues to second order in A, then we solve the problem
exactly and compare. For n\ + n2 = 2 (second excited state) we have the three
degenerate states:
|2,0>(°) , |1,1>(°) , and |0,2><°).
0 V2X 0
H' =\ a/2A 0 v/2A (13.4.24)
0 y/2X 0
The eigenvalues are: 2A , 0 , —2A. The corresponding eigenvectors are respec-

tively
1/2 \ 1/V2 1/2 \

l/y/2 —1/v2 . (13.4.25)
1/2 J -l/y/2 i/2 /
Thus, our new basis states are
iv>2,+> i|2,0)(°) + -1|1,1)(o)+i|0,2)(o)
1^2,0) _l|2,0)<o>- ^|0,2>(°)
1^2,-) i|2,0>(°> - - ^ | 1 , 1 ) < ° > + ||0,2>c°> . (13.4.26)
This yields:
E2,r = 2fru> + r2X + £

' 2hu> — (ni + n2)huj
y
ni+n27<2
= 2hw + r2X . (13.4.27)
Since H' does not change the sum over n\ + n2 there is no second order contri-
bution. The third excited states are:
|3,0>(°\ |2,1>(°), |1, 2)<°>, and|0,3)<°> .
0/ s/3X 0 0 \
H' = \/3A 0 2A 0
(13.4.28)
0 2 X 0 y/3X
\ 0 0 y/3X 0 j
13.4. TWO-DIMENSIONAL SHO WITH OFF-DIAGONAL TERM 269
The eigenvalues for this matrix are: 3A , -3A -A. The corresponding
eigenvectors are
( 1/v2 \ ( 1/V5 \
v/3/2 -%/3/2
V3/2 V3/2
V 1/V5 ) V -1/V2 /
( v/3/2 \ ( \/3/2 \
l/v/2 -1/v2
(13.4.29)
-1/V2 -1/72
\ -v3/2 V V3/2 /
We could now again compute the new basis vectors and go on to second order,
but because [ i í o , # ' ] = 0 there is no second order contribution. Thus, correct
to second order the energies are
E2 = 3 / k j ± 3A , 3TKJ ± A . (13.4.30)
b) We next examine the exact solution. For this purpose we make the principal
axis transformation
A = ai cos 9 + a2 sin 9
B — — ai sin 9 4- 02 cos 9 . (13.4.31)
Then, A , A^ , B , B^ satisfy the same commutation relations as a\ , a j , a2 , a!,.

Rewriting the Hamiltonian we find
Ho = huj(a\ai + a\a2)
= HU(AÂ + B^B) (13.4.32)
and
H' = A [(^t cos 0 — fit sin 0) (^4 sin 0 + B cos 9)

+ (At sin 9 + B^ cos 0) (¿4 cos 9 — B sin 0)]
= A [ A t ^ s i n 2 0 - B ^ B sin 29+ (A*B + B1 A) cos 29] (13.4.33)
Therefore, choosing cos 29 = 0 so that sin 29 = 1 we get
H = (hw + \)Á[A+{huj-\)B^B . (13.4.34)
The eigenstates may be labelled \N\,N2) and the eigenvalues are
H — (ñ¿j + A)Aî -t- (hui — X)N2 — (N\ + N2)hui + — 7V2)A . (13.4.35)
For N\ + N2 — 2 we get 2hu> + 2r\ where r — 1 , 0 , - 1 .

For Ni + N 2 = 3 we get 3huí ± 3A , 3Su' ± A.
These results are exactly the same as those obtained from perturbation theory.
13.5 Non-diagonal Two-dimensional SHO

For a particle of mass m moving in the potential
V = ^kix2 + ^k2y2 + Xxy |&i — k2\ < 2A .
Find to order A2 the energy of the ground state and first excited state. Compare
your answer with the exact solution obtained in problem 12.8c.
Solution
In this case ki « k2 and all levels except the ground state are almost degenerate.
Thus, for the first two excited states we need to use degenerate perturbation
theory.
Ground state: n\ = n2 = 0
£o?o = | A ( w i + w a ) (13.5.36)
and
4!o = 0 (13.5.37)
while
f( 2)_A2 y- l(0M™i)(%K)|2
E
°'°~X ^ —h(oJimi + ui2m2) (13 5 38)
-'
V 1 1
m1+m2¿0
where we have
<0|x|mi)(0|y|m2) = L = [áo,mi-i¿o,ms-i] (13.5.39)
Thus, we get
(2) A 2 /i 2 1 1
K,o = Z—2 -r, r • 13.5.40
UI\U>2 H[UJ\-\-UI2)
Notice, that this expression does not involve
2 9 ^1 — k 2
UJ2 - U J 2 = y 13.5.41
M
First two excited states: ni = 0 , n2 = 1 ; ni = 1 , n2 = 0

The Hamiltonian is
2 2
H
+ ^k2y2 + Xxy . (13.5.42)
Our perturbation is
H' = A x y = (aj + a ¡ ) ( a 2 + a\) . (13.5.43)
13.5. NON-DIAGONAL TWO-DIMENSIONAL SHO 271
To do degenerate perturbation theory we diagonalize this perturbation in the

degenerate subspace. T h a t is, we diagonalize the matrix
/ I rr'l / /\ ^ ( 0 1 (13.5.44)
The eigenvalues are
h\
Ei1' = ±- (13.5.45)
and the eigenvectors are
fa =
^(±i (13.5.46)
The energies are therefore
HX
e ± — 2 huí i (13.5.47)
2RRIY/U¡\U>2
To compare with the exact solution obtained in problem 12.8a we must use
the fact that \kx — k2\ < 2A and put uii — ui 2 . It is then easy to see that the
results agree to this order.
We can now do higher order perturbation theory using that linear combi-
nation of 10,1) and | 1 , 0 ) which diagonalizes H'. In other words, to find the
perturbation to second order of the levels | 0 , 1 ) and |1, 0) we use as a basis the
set:
|0, 0) , | ^ + ) , 1^-) , | « i , n2) where nx + n2 > 2.
Thus, we get
E{2) - V (13.5.48)
" i , n ' 2 ~ n n^l
/ , E
n'lin'2 ~ E„1}n2
1+
But,
A ft 1
(ip±\H'\ni,n2) = (0, l | ( a | + ai)(<4 + o,2)\nx,n2)
2rriy/ujiu 2
± (l,0|(a| + ai)(aí + a2)|ni,n2) (13.5.49)
Here, we have already introduced the explicit form
v± = ^ [ | 0 , 1 > ± | 1 , 0 > ] . (13.5.50)
So, for n j + n 2 ^ 1 we then find that
Aft
(ip±\H'\n1,n2) (13.5.51)
2M^/LÜ\ÍÜ2
A< ^ gruirían i J. DEGENERATE PERTURBATION THEORY
Combining these results we get
f( 2 ) _ \2h 1
^0,1 — (13.5.52)
2
4M ÜJIU)2 UI I + U>2 2wi
and
2
£•(2) _ Ah 1
+ (13.5.53)
1uir
—
w
1,0
4M2U>I<JJ2 1 + w2
Again, if we set uj\ = w 2 , we find that the results agree to order A2 with the
exact results obtained in problem 12.8a.
13.6 Particle in a B o x Perturbed by

Electric Field
A particle of mass m and a charge q is placed in a box of sides (a, a, b) where
b < a. A weak electric field
É = £(y/a,x/a,0)
is applied to this particle. Find the energy of the ground state and first excited
states correct to order \£\.
Solution
For a box with sides (a, a, b) the energy eigenstates for a particle in this box are
,FOL ¿\J¿
2^2 . ¡nTTX\
, , . ,tmny\ . /r7rz\
,m r
' ~ ^ a s m sin
(~ít)sin (~r) (13.6.54)
where n, m, r = 1, 2, 3 , — The corresponding energies are:

,22
E(O) _
(n + m 2 W 2 r2^2"1
T~L— + (13.6.55)
n
' m
' r
2 m b2
Since 6 < < a the ground state energy is given by E[°¡ : and the first excited
states are degenerate and their energy is given by E[°] : = E^\ Thus, the
ground state perturbation is given by taking the expectation value of the per-
turbing potential which is given by
, e£
—e<p = —xy . (13.6.56)
a
The result is
£•(0)
^-(1,1,11^11,1,1)
xdx
1,1,1
= e£a
~4~
sin2
( t ) l s1 2
" ( ? ) y d y
(13.6.57)
13.7. UNUSUAL PARTICLE ON INTERVAL 273
For the first excited states we need to diagonalize the perturbation in the de-
generate subspace. Thus, we have to evaluate
—<1,2,11*011,2,1)= —(2,1,11^12,1,1)
A A
si 2
= —4 f "> (—) ¡ ' ^ ( — ) »<ly
a a2 J0 Va ) J0 \ a J
eSa
(13.6.58)
~T~
as well as
—(1,2,11^12,1,1)= —<2,1,1m1,2,1)
a a
4 [a . /nx\ . (2nx\ fa . (iry\
7ry\ . flny
eS —2 I sin — sin ] x ax sin — I s i n ydy
a Jo \ a / \ a J J0 \ a / v a
a
256
— e£a (13.6.59)
81 7T4
Next we have to diagonalize the matrix
( 2Í6
I
817T 4
8
256
V
4
4
(13.6.60)
The corresponding eigenvalues are:

1 256
e£a (13.6.61)
4 817T4
Thus, the degeneracy is lifted.
13.7 Unusual Particle on Interval

P2
Ho = on — a < x < a
2m
with the domain for p given by
Dp = { f ( x ) e C 2 | f[-a) = -f(a)} .
Find the shift in energy of all the energy levels to first nontrivial order due to
a perturbation
AH' = A* .
Hint: The eigensolutions for HQ are given in [13.1] and are
fn(x) = ~ei" i n + 1 / 2 ) x , a
n = 0,±1,±2,... (13.7.62)
v2 a
with the corresponding energy eigenvalues
En = E _ i n + 1 ) = £ ^ ( n + 1/2)2 . (13.7.63)
^nAf iUH 13. DEGENERATE PERTURBATION THEORY
Solution
T h e eigenfunctions of the unperturbed Hamiltonian are, as in [13.1], given by
f„(x) = —L= e'w("+1/2)x/a n = 0, ± 1 , ± 2 , . . . (13.7.64)

y/2a
T h e energy eigenvalues are
2 »-2
En = E_{n+1) = ^ (n + 1 / 2 ) 2 . (13.7.65)
This shows that every eigenvalue is two-fold degenerate. We thus need to diag-
onalize the perturbation Hamiltonian in the degenerate subspaces. This means
we have to evaluate the matrix elements
i f n , x f n ) = ( / _ ( „ + ! ) , ¡r/_(„+!)) = 0 (13.7.66)
as well as
(/ni x / - ( n + l)) =
(f-{n + l)t x f n ) ) — 7r(2n + 1) ' (13.7.67)
T h e energy corrections are now given by the eigenvalues of the matrix
Aa ( 0 1 1
, ,
' (13.7.68)
(2n + 1) 7t \ 1 0
T h e eigenvalues are
Aa
(13.7.69)
(2 n + 1) 7r
Thus, all the degeneracies are lifted and the energies corresponding to En and
E-(n+i) are
TT2h2 9 Aa
13.8 Rigid R o t a t o r in Magnetic Field

A system with m o m e n t of inertia I has the Hamiltonian
h0 = — .
21
a) W h a t are the energies of the lowest and first excited states?

b) A perturbation
H' - yg — Lxx
Mc
is applied. Find the splitting of the first excited states.
13.8. RIGID ROTATOR IN MAGNETIC FIELD 275
Solution
We have
= ' » ' = OWC1-• ( 1 3 8

-71)
The eigenfunctions of L2 are Y¡¡m with eigenvalue /(/ + l)h2. Therefore, we have
l{l+
H0Yl¡m = 2 j)h2Y,,m. (13.8.72)
Every eigenvalue is (2/ + l)-fold degenerate. Now, we can also write
Lx = i(L+ + L.) . (13.8.73)
Thus, the first order corrections to the energy are given by
9
Wc \{L+ +
= 0 (13 8J4)
-
since
L±Yi¡rn = \/l(l + 1) - m(m ± l)ftYj, m ±i . (13.8.75)
For the first excited states we have to diagonalize the matrix with matrix
elements
(Y.,m'} LxYitm)
h
^ [1/2 - m(m + l)¿m',m+i + \J2 - m(m - l)<5 m ', m _ij . (13.8.76)
2
Written out as a matrix this looks like
0 1 0
~ I 1 0011 I . (13.8.77)
0 1 0
The eigenvalues are: 0 , ± 1 . Thus, the energies correct to second order are
given by
% + y
, £ , 5 l _ y, f £ . v
(13.8.78)
/ Mc ' I ' I Mc '
Incidentally, since Lx c o m m u t e s with L2 and is known to have the eigenvalues
mh with m = —I, —(I— 1), . . . , (I— 1), I we can write down the exact eigenvalues
of the total Hamiltonian. These are
El}m = ^-l(l + l)+g^-mh (13.8.79)

1 MC
in complete agreement with the perturbation result.

13.9 A x p'Vy Perturbation of SHO

Find the energy correct to order A2 for the second excited state of the Hamil-
tonian
H = H0 + A H'
where
Ho = -1—{pI+ P2y) + ^rnu2(x2 + y2)
and
H' = xpy .
Can this problem be solved exactly?
Solution
T h e energy levels of the second excited state are |2, 0) , | 1 , 1 ) , |0, 2). They all
have the same energy 3hu>. We now diagonalize the perturbation part of the
Hamiltonian in the degenerate subspace. To do this we use
h
(aí+ai) , Py = » y ~ ~ í r ~ ( a 2 ~~ a 2 ) • (13.9.80)
2 mw
Then,
AH' = ¿ A - f a i a í + a\a\ — a{a2 — aia2] . (13.9.81)
In the degenerate subspace the perturbation matrix to be diagonalized is now

easily computed to be
1 0
Ah
I— i —1 0 1 (13.9.82)
0
T h e eigenvalues and corresponding eigenvectors are \h/2 , 0 , —\h/2 and
-t= 2i 0 2i (13.9.83)
V6 ' V2 ' 76
Thus, the new basis states we use for starting the perturbation computation are
|v+> = -j=(|2,0> + 2 ¿ | 1 , 1 > - | 0 , 2 »
|Vo> = ^=((|2,0> + |0,2»)
l^-> = ^=(|2,0>-2í|l,l>-|0,2» .
(13.9.84)
13.10. PASCHEN-BACK EFFECT 277
We now have to compute the terms H'\ip+) , H'\ipo) , and which do not
contain terms |n, m) with n + m — 2. Then we have (after omitting the terms
with n + m = 2)
A(n, m\H'\±)
—Z —
— '~~/= 3^71,1 — ¿n,l¿m,3) ± 2í(2¿„ i 2¿m,2 ¿n,0<5m,0) (13.9.85)
and
A(n, m\H |0) — ¿Aft "I" ^n,l^m,3] (13.9.86)
so that
p( 2 ) _ (Aft) 2 _2 2 16 4
+ + +
2,±
16ftw . ( 3 - 5 ) (3-5 3-5 3- 1
2
4A ft (13.9.87)
3w
Similarly we find
(2) A2 ft
E.2 , 0 (13.9.88)
Yes, this problem can be solved exactly since the total Hamiltonian is quadratic
in the annihilation and creation operators.
13.10 Paschen-Back Effect

Calculate (to first order) the splitting of the n,l levels for / = 0 , 1 of the energy
levels of the valence electron of an alkali atom in a strong magnetic field. For
weak fields this is the Zeeman effect. For magnetic fields so strong that their
effect is greater than the spin-orbit splitting the result is known as the Paschen-
Back effect.
Hint: For the Paschen-Back effect it may be convenient to use some of the
results of problem 17.7 .
Solution
We take the magnetic field to point in the z-direction. In that case the Hamil-
tonian for the valence electron is
" = £ + ^> + 2¿;Í £ s"-|S(í'+2S-> • <1310'89»

For a given solution of the unperturbed Hamiltonian with quantum numbers
n, I and energy eigenvalue Eni and
+ (13.10.90)
ÍJ I (J
^tiAf iKK 13. DEGENERATE PERTURBATION THEORY
we have the weak field or Zeeman effect. In that case it is convenient to choose
states of good j,rrij and this situation is solved in general in problem 17.10.
On the other hand if
( n , ¿ S > , í ) <
f e ¥ ' —(n,l\Lz + 2Sz\n,l) (13.10.91)
we have the Paschen-Back effect. In this case we want states of good m¡ , ms.
For this we use the results of problem 17.7 where we find
|3/2,3/2> = |l,l)|l/2,l/2)
|3/2,l/2> = y | | l , 0)11/2,1/2)+ - ^ | l , l ) | l / 2 , - l / 2 )
13/2,-1/2) = -L|l,-l)|l/2,1/2) +^||1,0)|l/2,-l/2)
13/2,-3/2) = |l,-l)|l/2,-l/2)
|l/2,l/2> = -^=|l,0)|l/2,l/2)-y^ll, l)|l/2,-l/2)
11/2,-1/2) = y | | l , - 1 ) 1 1 / 2 , 1 / 2 ) - - ^ | l , 0 ) | l / 2 , - l / 2 ) . (13.10.92)
The inverse of these equations is
|l,l)|l/2,l/2) = |3/2,3/2)
|l,0)|l/2,l/2) = y||3/2,1/2) +-L|l/2,l/2)
|l,l)|l/2,-l/2) = -^|3/2,l/2)-y||l/2,l/2)
|l,-l)|l/2,l/2) = y||3/2,-1/2) + -L|l/2,-l/2)
|l,0)|l/2,-l/2) = -L|3/2,-l/2)-y||l/2,-1/2)
|1,-l)|l/2,-l/2) = |3/2, —3/2) . (13.10.93)
For I = 0 the only possibilities are
|0,0)|l/2,ms) = 11/2, rrij)

|0,0)|l/2,l/2) = |l/2,l/2)
|0,0)|l/2,-l/2) = 11/2, —1/2) . (13.10.94)
Thus, we can evaluate the magnetic part of the energy. If we call
fc2 / ,, 1 1 dvt
2mc 2 r dr
eB
= w (13.10.95)
2 mc
10. PASCHEN-BACK EFFECT 279
the result is
|Z,m/)|l/2,m5) E
|1, l ) | l / 2 , l / 2 ) £„,i+2hu
|1,0)|l/2,l/2) e n ,i + hu
|l,l)|l/2,-l/2) 0
|l,-l)|l/2,l/2) 0
11,0)|l/2,-l/2) i-fiw
|1,-1)11/2,-1/2) e „ , i - 2 hu
|0,0)|l/2,l/2) c„,i + ftw
|0,0)|l/2,-l/2) e B ,i + /Sw . (13.10.96)
The spin-orbit term is now simply a perturbation. Since,
2
1 - 5 = I P - P - S ] . (13.10.97)
We immediately get that for the states | j = 3 / 2 , ± 3 / 2 )
(n, / , j = 3 / 2 , m.j = ±3/2|jL • 5 | n , l,j = 3 / 2 , mj = ± 3 / 2 )
= ^[3/2(3/2 + 1)-1(1 + 1)-3/4]
h2
= — . (13.10.98)
Similarly, for the states |j = 1/2, ± 1 / 2 ) corresponding to I = 0 we immediately

get that
(n,l, j = 1/2, mj = ± 1 / 2 | L • 5 | n , /, j = 1/2, = ±1/2)
= y[l/2(l/2+l)-0-3/4] = 0 . (13.10.99)
This leaves only four states to consider. These split into two pairs: those with
rrij = 1 / 2 and those with mj = —1/2. Thus, we can consider them separately.
However, L • S is not diagonal in this representation. In fact, we find that
£ • § = - ( Ltz l
7 ^ . (13.10.100)
2 y L.j. —L
For mj = 1 / 2 , when evaluated with the corresponding states, namely
|1,0)|l/2,l/2) and |1, l ) | l / 2 , - 1 / 2 )
we get
tnl Í
^
0 yfi. • (13.10.101)
2 v
CHAPTER 13. DEGENERATE PERTURBATION THEORY
T h e energy corrections are j u s t the eigenvalues of this matrix. T h e s e are easily

found to be
—Cni and cni .
T h u s , the energy shifts are
ê„¡ and - e„¡ . (13.10.102)
Similarly for rrij = — 1 / 2 we have the s tates
|l,0)|l/2,-l/2) and |1,-l)|l/2,1/2) .
T h i s leads t o the energy m a t r i x
£n¡_ Í 0 \/2
(13.10.103)
2 v yfi- 1
with the eigenvalues
2fn¡ and tni .
T h u s , in this case, the energy shifts are
-ê„; and e„i . (13.10.104)
T o identify which energy level belongs to which s t a t e it is convenient to include

the m a g n e t i c energy in these considerations. Since the m a g n e t i c energy is di-
agonal this causes only a slight c o m p l i c a t i o n . T h u s , for rrij = 1 / 2 w e get the
energy m a t r i x
(I310105)
WV5 wf) "
T h e eigenvalues are
€
2 ~ m / 2 ± \ J h 2 u 2 + c n ihjj + 9 / 4 e2nl
If we now let
e„i/(ñw) -> 0
we recover the strong field (Paschen-Back) case. T h e energy shifts then approach
2 — en¡/2 + sjh-u2 + £nifíu + 9/4 cní) —

^ ^
¿ - c»i/2 - + e n / fiw + 9 / 4 e O -> (13.10.106)

13.11. H ATOM: WEAK FIELD STARK EFFECT 281
This shows that the term with the + sign corresponds to the case m¡ = 0 , ms —
1/2 and the term with the — sign corresponds to ra¡ = 1 , m, = —1/2 .
Similarly, for m j = - 1 / 2 we find the eigenvalues
- + Cni/2 ± h 2 ui 2 + Ztn\hu) + 9 / 4 f 2 ¡^ .
Again, letting
fni/(^) 0
we recover the strong field (Paschen-Back) case. We then get
— + eni/2 + \J h2ui2 + Zenitux> + 9/4 —> tiui + e„;
I ^ + fni/2 - + 3e n ,fkü + 9 / 4 ^ ) -+ . (13.10.107)
So, the + sign corresponds to m; = 0 , ms — —1/2 while the — sign corresponds

to m¡ = — 1, m , = 1/2.
13.11 H A t o m : Weak Field Stark Effect

Using the results of problem 10.11 and 11.22 calculate the energy shift in the
hydrogen atom energy levels (fine structure) due to a weak electric field.
Solution
We have the weak perturbation
V { f ) = e£z (13.11.108)
where we have chosen the electric field to point in the z-direction. Also, from
problem 11.22 we have that the unperturbed normalized hydrogen a t o m wave-
functions are
L M 1 2Y
R,B j - l / j ( r ) ( V + + / ',M-L/2
**n,l,j=l-\-l/2,m
VW+I
\ f l — m + 1 / 2 y,, m + i/2
(13.11.109)
m 1
Rn,j+ !/2(r) / V ' - + / 2 V i / 2
® n , I, j=l —1/2, m
^ * \ — \/l + m + 1/2 Y¡,171+1/2
(13.11.110)
To simplify the writing we call
^n,l,j=l+l/2,m = 1p 1
^n,l,i=l-l/2,m — V>2 • (13.11.111)
The energy levels in the unperturbed hydrogen atom are labelled by n , j. Under
the perturbation the orbital angular m o m e n t u m no longer commutes with the
Hamiltonian and thus, I is no longer a good quantum number. However, the z-
component of the total angular momentum J 2 , namely m, continues to be a good
quantum number. Thus, the matrix elements of the perturbation Hamiltonian
V between states with different m vanish. The diagonal matrix elements of V
also vanish
(9n,l,j=l+l/2,m, V9niltj=l+i/2,m) = {ip 1, Vlpi) = 0

va>n,iij=l-i,2tm) = (rl>2,Vri>2)=0 . (13.11.112)
This means we must diagonalize V in the degenerate subspace corresponding to

the states *J / n ,i=i+i/2,m and We call these matrix elements
V\2 = V21 = {9nJ-i/2,m, V9nj + i/2,m)

eS f°°
= r 3 R
o /•!• , / n,j-l/2(r)Rn,j+l/2(r)dr
2 \ / j ( j + 1) Jo
m c o s
x v u + ) ( j - m + 1) / ^j-l/2,m-l/2^J + l/2,m-l/2 # dd
- V Ü - m ) ( j + m + 1) / — l / 2 , m + l / 2 ^ J — l / 2 , m + l / 2 COS 6 dQ j .
(13.11.113)
To perform the angular integration we use that
axr ,Q s. (I + m+ 1)(/ — m+ 1)
cos = J { 2 l + 1 ) { 2 l + i)
Our matrix elements now reduce to

poo
e£
V12 = V21 =
7,3 dr
2\/j(j + 1) Jo
171
X . . 1
13.11.115
\/7(7tt)
The remaining radial integral is
roo
/ r3fínj_1/2(r)ñnj+1/2(r)dr (13.11.116)
JO
and may be evaluated using the generating function as in problem 10.11 to yield
--ny/n2 - (j + 1/2) 2 .
BIBLIOGRAPHY 283
(13.11.117)
The perurbation matrix to be diagonalized is now
V = ( ¿ "¿' ) . (13.11.118)
The resulting energy shifts are
3 " \ / n 2 — ( j + 1/2)2m
= ±Ki2 = ± t — ., , —e£ . 13.11.119
4 ju + 1
Finally we see that for a fixed value of n all terms of the fine structure, except
the term with j = n — 1/2, are split into 2 j + 1 equidistant levels corresponding
to m = — j , . . . , j- The term with j — n — 1 / 2 is not split at all since / has a
fixed value I = j — 1 / 2 = n — 1 and is not degenerate as regards the quantum
number /.
Bibliography
C h a p t e r 14
Further Approximation
Methods
14.1 Variational Ground State of SHO

Use the trial wavefunction
iP(x) = Ae~aW
to estimate the ground state energy of a simple harmonic oscillator.
Solution
T h e trial wavefunction is
v> = AE~AW . (14.1.1)
Normalization yields,
A = -Y/A . (14.1.2)
Since this wavefunction has a discontinuity in its first derivative, something that
is rather unphysical, we expect that it will yield a rather poor estimate of the
energy.
T h e Hamiltonian is
+ —muix (14.1.3)
Applying the kinetic energy operator to ip we get
= [a2 - 2a¿(x)] e - a | E | . (14.1.4)

14.2. VARIATIONAL GROUND STATE OF H2 MOLECULE 285
Thus,
h2a2
(T) = • (14.1.5)
2m
Similarly,
I1416)
j2
W = Sr- 4q2
So, the expectation value of the Hamiltonian is
+ ( 1 4 L 7
= ^ >
The extremum of this function is obtained by differentiating to get
_0 (14.1.8)
m laA
The solution of this equation yields
a2 = ± mw (14.1.9)
y/2 h
Only the + sign is acceptable (since ip has to be square integrable) and it yields
a minimum for the energy expectation value given by
E0 « (H) = - i = hw. (14.1.10)

v2
As expected this is larger than the exact ground state energy of 1 / 2 hut by about
41 %.
14.2 Variational Ground State of H2 Molecule

73
V-(ri,r 2 ) = — 3 e - z ( r i + r ^ ' a °
7T a0
to evaluate the expectation value of the Hamiltonian
B = f + f - 2 e * ( ± + l ) + ¿ .
2m 2m \r i r2 J ri2
286 CHAPTER 14. FURTHER APPROXIMATION METHODS
Solution
T h e Hamiltonian is
rr pi2 2e 2 p2 2e 2 e2 „ rr e2
H = h 1 — Hi + H2 H (14.2.11)
2m ri 2m r2 r 12 rj2
where
2
2e 2
= ¿=1,2 (14.2.12)
2m r,-
are hydrogenic Hamiltonians.

As an approximate wavefunction we choose a product of "screened" hydro-
genic wavefunctions since we expect the electrons to screen the nucleus from
each other.
z r
i>(ri,?2) ~ — E ( i+r=)/a = (j)(ri)<j)(r2) (14.2.13)
7t a
where
h2
(14.2.14)
is the Bohr radius. Now
H2 1 D2
Z2 2Z
(14.2.15)
2m cr ar
The same result holds for T2<t>(r2). Next using that h2/m = e2a we find that
(Ti) = (t2)
„2 7 3 /-oo
Z2e2
= ^ " 1 l^-2ZRA]^"^R (14.2.16)
2a
Also,
2e 2 = 2e 2
= — 1e2^— 47T I
2Ze2
7t a°
Jo
f e~2Zvlardr
(14.2.17)
14.3. SQUARE BARRIER: WKB APPROXIMATION 287
Thus,
(Hi) + (Hx) = 2 - 2 = (Z2 - 4 Z ) ~ . (14.2.18)

2a a a
Next we evaluate (e2/ri2) by expanding in spherical harmonics, j us t as we
did
in problem 12.1. Thus, we get
5Ze
/il) = ~ (14.2.19)
V127 8a "
Collecting all these terms we get
9 27 (14.2.20)
(H) = - Z2 - —Z
To complete the computation we now minimize the energy witli respect to Z.

Thus, we compute
4-m=o = 2 z - % . ("-2-21)
oz 8
Then,
z = 27 (14.2.22)
16
So, clearly Z < 2 and we see that the electrons do indeed scr?611 e a c h other.
The best estimate for the ground state or "ionization" energy is n o w
£=_ rny f i =_2.85-. <i4-2-23>
\16y a a
This is clearly lower than the perturbation result
_llf! = _2.00- . (14-224)
2 a a
14.3 Square Barrier: W K B Approximation

a) Use the W K B approximation to solve the tunneling p r o b l e m f ° r a s c l u a r e
barrier using the fact that in this case you have no need to us1? the connection
formulae. T h i s result agrees with the exact result in [14.1].
b) Repeat part a) by using the connection formulae.
Solution
a) In this case we have the potential
Í 0 x <0
V(x) = I Vo 0 < x < a . (14.3.25)
Also, in this case we have special boundaries so we do not have to use the
connection formulae. Calling
2 mE _ 2m(V 0 - E)
k2 = K
2
= (14.3.26)
2
h ' " h2
we can write our solutions as follows for a wave incoming from the left
'rPj(x) = + R' e-(fc®-T/4)j x < o

Kx Kx
xp(x) ipu(x) = [A'e + B'e~ ] 0 <x<a . (14.3.27)
x>a
If we define new constants
A' = A ^ r , B' = B^¡r , R' = Re~in'2 (14.3.28)
and use the fact that the solutions are valid right up to the turning points, we
can impose the boundary conditions that both ip and dip/dx are continuous at
x = 0 and x = a. Thus, we get
£-i,r/4 [1 + fl] = A + B
in/4
ike~ [1 - R] = K[A-B]
E-IN/4T _ AEKA + BE-KA
Ka
ike-iit/4t _ K[Ae - Be~Ka] . (14.3.29)
From the first two of these equations we get
Í „-i>/4
[{1 + ik/K)+ R{1-ik/K)]
i v
B = le- ' 4 [ ( l - i k / K ) + R(l + ik/K)] . (14.3.30)
Substituting this result in the second pair of equations we get
T = - [(1 + ik/K) eKa + (1 - ik/K) e'Ka]
+ ~R [(1 - ik/K) eKa + (1 + ik/K) e~Ka]

iK
T = —^[(l + ik/K)eKa-{l-ik/K)e~Ka]
- ~ R [(1 - ik/K) eKa — (1 + ik/K) e~Ka] (14.3.31)
Solving for R we find
(k2 + K2) sinh Ka

R = (14.3.32)
(k2 — K2) sinh Ka + 2ikK cosh Ka
Next, we use that
T = e' 71 ^ 4 \AeKa + Be~Ka 1 (14.3.331

14.4. VARIATIONAL GROUND STATE IN GAUSSIAN POTENTIAL 289
and substitute for A and B to get, after a little algebra, the result that
T = . (14.3.34)
(k2 — A' 2 ) sinh Ka + 2ikI\ cosh Ka
Except for an irrelevant phase factor these results agree with the exact results
obtained in [14.1].
b) T h e solutions are exactly those given by (14.3.27). However, this time we
have to use the connection formulae. To do so we first rewrite ipj and ipm as
follows.
ipj = {(1 + R) cos(kx — 7r/4) + ¿(1 — R) sm(kx — 7r/4)}
ipui = —j=T {cos[fc(ar — a)7r/4] + ¿sin[fc(x — a) — 7r/4]} . (14.3.35)

Vk
Now using the connection formulae at the turning points x = 0 and x = a and
defining
S = e~Ka (14.3.36)
we get
^ = ^(1 + fl)
B = i(l-R) (14.3.37)
as well as
T = 2 B S ' 1 = 2 i ( l - R)S~1
T = iAS= ^{1 + R)S . (14.3.38)
We solve these equations for T and R and get

T ^ ^ • — Ka
~ l + S2/4
1 _ S2 /4 e-2Ka
R (14 3 39
• - - »
Clearly, the results from part a) and part b) are very different. On the other
hand, to the approximation used, we still have conservation of probability since
p — 4Ka
\T\2 + | i l | 2 = e~2Ka + 1 - e~2Ka + — — w 1 . (14.3.40)
14.4 Variational Ground State in Gaussian

Potential
a) Use a variational approach to find the ground state energy for a particle in
the potential
V(r) = -V0e~ar2
zau CHAPTER 14. FURTHER APPROXIMATION METHODS
if the particle involved is an electron and a = 5.29 x 10 1 3 c m - 2 , Vo = 20 eV.

A numerical answer is required.
b) For comparison approximate the Gaussian potential by its first two terms of
a Taylor expansion and estimate the energy that way.
Solution
a) The potential
V(r) = -V0e (14.4.41)
looks like a simple harmonic oscillator near the bottom of the well. So a rea
rea-
sonable choice of wavefunction for the ground state seems to be
rp(r) = A e~^r ^2 (14.4.42)
with the normalization
a = 2(p3l*)l,a . (14.4.43)
Since this is an / = 0 state we have

h2 1 d2
2m r dr¿
= A -/3rs/2
- Í L ^ R ' - W - V O E - (14.4.44)
Therefore,
rc
f
M W,H^) = \A\'< dr(/32r2 — 3 (3)e -PR*
. 2 m J0
Vr J 0 r2dre (a+^)r (14.4.45)
So,
3/3 2m V0 /? 3 / 2
M = (14.4.46)
2m . 2 h? (a + /?)3/2
We now minimize £"(/?). To do this we take the derivative with respect to ¡3

and set the result equal to zero. Furthermore, we call
K =
2mVÓ
(14.4.47)
h2a
and get
3 j A'o/? 1 / 2 ^ A'a/? 3 / 2
:0 . (14.4.48)
2 ~ (/? + a ) 3 / 2 T
(/? + a ) 5 / 2
After setting /? = a x , to make everything dimensionless, we find
(1 + x ) 5 / 2 = Kx1/2 . (14.4.49)
14.4. VARIATIONAL GROUND STATE IN GAUSSIAN POTENTIAL 291
Now, substituting the numerical values we get K — 10 3 . Thus, we see that x

must be large compared to 1. In fact we see that
x fa \f~K rs 30 .
To find x we therefore make a table of values for the left side and the right side.
Thus, we conclude that, to the accuracy of the data given, x — 30.4. Putting
•A (1 + z ) 5 / 2 Kx1'2
30.0 5.35 x 10 3 5.48 x 10:
30.2 5.44 x 10 3 5.50 x 10:
30.4 5.52 x 10 3 5.51 x 10:
30.5 5.57 x 10 3 5.52 x 10:
this back into the expression for the energy we find E ft¡ —15.4 eV. Thus, the
ground state energy is about E « —15.4 eV .
b) If we Taylor expand the potential we get
V(r) «-V0 + V0ar2 . (14.4.50)
This is just a simple harmonic oscillator potential and the ground state energy
is given by
E K - V o + ^Hlj (14.4.51)
where
(14.4.52)
so that
E = -V0 + ~ V 0 . (14.4.53)
VK
Substituting the numerical values we get
i? ss —18.1 eV . (14.4.54)
This shows that the two calculations give similar results and agree to within
about 20% . Although the variational calculation gives a larger result it should
be considered the more reliable since the approximate potential is wider in the
region where the wavefunction is appreciable and thus gives a lower kinetic
energy.
14.5 Variational Ground State:

Quartic Potential
An electron is in the spherically symmetric potential
t / ( r ) = A r 2 ( r 2 — a2)
where
a - 2.00 x 1 0 - 1 nm
a4A = 1.90 eV.
Use Rayleigh Ritz to estimate the ground state energy. A numerical answer is
required.
Solution
We have
H
= k+vl-r) (14.5.55)
where
V(r) = A r 2 ( r 2 — a2) . (14.5.56)
Since we are looking for the ground state we have I = 0 and the kinetic energy
is
1
T ! ^ ^ I L\ (14.5.57)
=
~2MRDR^
We now scale in terms of dimensionless variables so that x = r/a. T h e n we find

that
1 d2 a4 A 2 2 '
HX!> = € ---T-r(ar^) H x2(x2 - \)xp (14.5.58)
x dx¿ e
with
h2
e = ^ ~¿ = 0.95 eV (14.5.59)
¿ma
and
a4 A
= 2.00 . (14.5.60)
As possible trial wavefunctions we choose
r¡> = xne~bx2/2 (14.5.61)

14.5. VARIATIONAL GROUND STATE: QUARTIC POTENTIAL 293
since near the b o t t o m of the potential well we have something that looks like
a simple harmonic oscillator and we want as few nodes as possible. To proceed
we need the following integrals.
(«,,« = j f f - . (14.5.62)
Also,
— — -^—r(xrp) = — \n(n + l)xn~2 — b(2n + 3)x" + b2xn+2} e~bx / ' 2 . (14.5.63)

x dxz
Combining these results we find that the expectation value of the Hamiltonian
is given by
,-7/ M (v'.ffv')
£(
"<6> = "fimt
2 n 2 + 6n + 3 4n2 + 16ra+15 2n + 3
= 6
+ 2 (14.5.64)
4n + 2 2b b
Differentiating with respect to 6 to find the m i n i m u m energy we get after some

rearranging that 6 is given by
2n2 + q n :]
+ h3 + ^2n + 3)6 _ ^ 4n 2 + l g n + 15) _ o. (14.5.65)
4n + 2
We now look at this for a couple of different values of n.

n = 0
^ 6 3 + 36 - 15 = 0 . (14.5.66)
A few numerical attemp ts show that 6 = 1.847 to the accuracy warranted by

the data. Substituting this value back into the expression for the energy we find
E = 3.345 e = 3.18 eV . (14.5.67)
n=l
3
y6 + 5 6 - 35 = 0 . (14.5.68)
Again a few simple computations show that b = 2.335. This yields
E = 4.921 e = 4.667 eV . (14.5.69)
Since this result is greater than the previous one we conclude that the ground
state enerev for this electron is E < .*?_18 pV
14.6 W K B : Ball Bouncing on a Floor

A particle is in a potential
v _ í rngz z >0
{| oo z =0
This corresponds to a perfectly elastic ball bouncing on a floor.
a) Find the W K B solution for all the energy levels.
b) Find the W K B solution for the ground state wavefunction of this particle.
Solution
a) We have the potential
(14.6.70)
This means that the point z = 0 is a special boundary and the wavefunction
must vanish at z = 0. The other turning point is given by
E = mgzo or zo - (14.6.71)
mg
Thus, the regions are as follows
I 0 < z < ZQ is classically allowed.
II z > zo is classically forbidden.
In the classically forbidden region we need an exponentially damped solution.

Writing,
K(z) = -y/2m2g(z - z0) z > z0 (14.6.72)
(14.6.73)
we have in the classically forbidden region
(14.6.74)
This connects onto the solution xp¡ in the classically allowed region if ipi is given
by
(14.6.75)
But, rpi(0) = 0. This means that
(14.6.76)
14.7. GROUND STATE OF f i j 295
Therefore,
rz o
[ ° k(z') dz' - j = (n + 1/2)7T n = 0,1, 2,... (14.6.77)
4
Jo
So we find,
[¿O
m / \/zo — z dz = (n + 3/4)irh . (14.6.78)
Jo
Integrating this expression and substituting for z0 in terms of the energy En we
get
1/3
(3 n + 9)27r2_1
En [MG2H2)^ (14.6.79)
128
In particular the ground state energy, which corresponds to the classical situa-
tion of the ball at rest on the floor, is given by
1/3
81tt
E0 = (:mg'h2)^3 « 1.84(my 2 /i 2 ) 1 / ' 3 . (14.6.80)
128
b) T h e ground state wavefunction is given (up to normalization) by n = 0 and

is
hl/2
ip = 2B 0 < z < zo
[2m2g(z0 - z)]1/4
ft1/2 2m
rp = 2B 1/4
exp z > z 0 . (14.6.81)
2
[2m g(z - z 0 )]
Here,
1/3 1/3
•Eo "81tt 2 ' r h2 i
20 (14.6.82)
mg m?g
oo
ÍO
r •1
I—
This solution corresponds to the classical case of a particle sitting on an impen-

etrable floor.
This solution should also be compared with the variational calculation (prob-
lem 14.16) which yields the value
E0 = 1.36 (mg2h2)1/3 . (14.6.83)
Clearly the variational calculation gives a better result.
14.7 Ground State of

Estimate the ground state energy of H J , an ionized hydrogen molecule.
zyo CHAPTER 14. FURTHER APPROXIMATION METHODS
TA
R
B
A B
Figure 14.1: An ionized hydrogen molecule.
Solution
In this case the Hamiltonian is (see figure 14.1)
p2 e2 e2 e2
H = f + - 14.7.84
2m ta rg R
where ta and rg are the distances between the electron and the protons at A
and B respectively and R is the distance between the two protons. As a trial
wavefunction we choose a superposition of a wavefunction centered at A and at
B. Thus, we use
r A a r B a
Â = - ^ = e - ' , rl>B = - L = e - ' (14.7.85)
where a is the Bohr radius, R is a variational parameter, and N is a normaliza-

tion constant. We now find
JV = ( ^ + V > B , ^ + l M = 2 [ l + 5 ] (14.7.86)
where S is the overlap integral
S = ¿ (14.7.87)
and is evaluated by using elliptical coordinates.

rA + rB rA-rB
< = — ^ — . n = — ^ — > V (14.7.88)
where tp> is the angle of rotation about the line joining the two protons. The
volume element in these coordinates is [14.2]
d3
d3rA = -¿-(£2 - r]2)d£dr)d<p (14.7.89)
o
with the range of integration
1<£<OO, — 1 < T) < 1 , 0 < , Hip) _ H A A + HAB + HBA + H BB

E{R) = (14.7.92)
(•ip,ip) 2[1 + S]
where
HAA — HBB - (IPA, Hip A) (14.7.93)
HAB — HBA = (4>A, Hips) • (14.7.94)
Thus,
HAA + HAB
E(R) = (14.7.95)
1+ 5
and we are left with three integrals to evaluate. Now,
p2 e2 „2
HAA = [IPA, IPA) ~ (*PA, + ( v m , -FTLPA
2m ta
t„2
2
e-22 ( e2
= + I IpA, IpA (14.7.96)
2a R V rs
The last term is again evaluated by using elliptical coordinates and yields
2
— d3r
y;
IpA, IpA
R
B J\*. A\
RB
e2R2
t e - r v a d £ , J X e - r " / a dr¡
2a3
L
POO
-Rt/ -Rri/a
+ Tj e df]
e"
~R
e
1 — e -2 R/a i+5 a
(14.7.97)
The term HAB breaks up into

r
p2 „2
HAB - (ipA,
2m rA
ipB " ( IPA, ) +
(*"• Ti*")
E E'
S - [ip A (14.7.98)
~2A~^ fl
The final integral is also evaluated using elliptical coordinates and yields
L
p.2 — v2 r2lt
di> DIP
(R/2a)(p — v)
(*•£*») = &Í J!
1
1 + - I e~R'a . (14.7.99)
Combining all these results we find with
R
x = — (14.7.100)
a
that
2x x
1 1 ( l + x)e — x ( l + a;)e -
E(x) = e2 2 x (14.7.101)
2 ^ x ar(l + x + x /3)e~ +x
A graph of this function shows that E(x) has a minimum at x w 2.16. This
yields a value of about 1.76 eV below the energy for x = oo, i.e. when one of
the protons is removed to infinity.
14.8 Variational Solution: Particle in a B o x

Use a variational technique to find the ground state energy of a particle of mass
m in the potential
{ oo
vo
oo
x < —a
—a < x < a
x >a
Hint: Pay close attention to the boundary conditions that the trial wavefunction
has to satisfy.
Solution
Since the wavefunction must vanish at x = ± a and must be a positive parity
solution with no nodes we choose
ip = A[\a\a-\x\a] . (14.8.102)
Here, a is the variational parameter. Normalization yields
'-41' = " - <14-8-103>
Differentiation of this wavefunction does not produce any singularities at x = ± a

since Tp(±a) = 0. The kinetic energy operator T acting on V yields
h2
Tip = —-—A[—a{a — l)xa~2} . (14.8.104)
2m
14.9. HYDROGEN ATOM: VARIATIONAL TECHNIQUE 299
Hence we find
roo
(T) = 2\A\2J ( | a r - 1 * 1 ° ) — a r ( a - i)xa~2
h
= ' 0»"+!)("+!) . (14.8.105)
Ama2 (2a — 1)
Also
(V) = V0 . (14.8.106)
Therefore,
w = ^' te £-°) + " + *- <14-8107»

To minimize this we differentiate with respect to a to get
—(H) = 0 = ( 2 a - l ) ( 4 a + 3) - 2 ( 2 a 2 + 3 a + 1) . (14.8.108)
da
The solution yields
1 +
a = 2 ^ « 1.7247 . (14.8.109)
Substituting this value back into (H) we obtain
(H) = E0 = (H) « 1.003 x + K0 . (14.8.110)
This agrees with the exact result to within better than 0.3 % since we had a
very good choice of trial function.
14.9 H y d r o g e n A t o m : Variational Technique

Use the ground state wavefunction for a three-dimensional simple harmonic
oscillator, namely
3 / 4
, (2 A\ *
= e
* { t )
as a trial wavefunction to calculate an approximate value for the hydrogen a t o m

ground state energy.
-i u n 11. r umtiuti APPROXIMATION METHODS
Solution
The Hamiltonian for the hydrogen atom is
p' e'
H = (14.9.111)
2m ~ 7 '
Using the trial wavefunction we get
(14.9.112)
This expectation value breaks up into two parts, the kinetic energy T plus the
potential energy V. For the kinetic energy we get
h2 A f2a \ 3 ' 2 f°° _ a r , l d2

r~dr
<t> = - 2 i ? 4
* b ) I
fi2 /9 /i\3/2 /"°°
. ,'2a\3/23 /it
2M
2M ) 8 V 2a
h2a
(14.9.113)
2M
For the potential energy we find
3/2
<k> = -e 2 47r
vfi' - e
r
r2dr
-e 2 47r
ífjf
2ar
e reír
3/2
e247r ( — ) —
7r / 4a
—2e^ (14.9.114)
Therefore,
£<„) = 3 ^ - 2 e = (14.9.115)
The variational principle states that the best approximation is obtained by

minimizing this expectation value with respect to the variational parameter a.
Thus we compute dE/da and set it equal to zero. Therefore,
dE(a)
3— e' — = 0 . (14.9.116)
da 2M 7T a
Solving for a we get
8 M2e4
a= (14.9.117)
97T h4
14.10. BOUND STATE IN ONE DIMENSION 301
After substituting this into the expression for E(a) we get the best estimate for
the ground state energy
£ = . (14.9.118)
37t 2 n-
This is to be compared with the exact ground state energy
1 Me4
Eo = - ^ • (14.9.119)
It is also to be noticed that the exact energy is lower than the estimate. This
is in keeping with the variational principle.
14.10 Bound State in One D i m e n s i o n

Use the variational principle to prove that, in one dimension, a potential that
is everywhere attractive always has at least one bound state.
Solution
What we need to show is that for a potential V(x) that is everywhere attractive
V(x) < 0 for all x

we can always find a solution such that the energy E is less than zero. To do
this we assume a normalized trial function
Í2a\1'4
rP=[-1 -ax (14.10.120)
7t J e
together with the Hamiltonian
+ (141(U2I
" = >
Using this wavefunction the expectation value of the energy is, by the variational
principle such that
E{a) = (i¡>, Hip) > E for all a (14.10.122)
where E is the energy of the possible bound state. Evaluating E(a) we get
E(a) = — + J — f e-2ax2V(x)dx . (14.10.123)

2m V 7T J ^
To obtain the best estimate for the energy we now minimize E(a) with respect
to the variational parameter a.
dE(a) h2
da
12a
2 x 2 e-2"*2 V(x)dx . (14.10.124)
Solving for ^ by setting
dE(a) _ Q
da
we find
h2a Í2a f°° 2 — 2ar 2 \rt \ j

= \ — / 2ax e V(x)dx
2m v 7T 7 - o o
1 /2a
e -lax V{x)dx . (14.10.125)
2 V 7T
Substituting this value back into (14.10.123) we get that
[2a
E < \/ — J {\/2+2ax¿)e~¿ax V(x)dx< 0 . (14.10.126)
Thus, we have proven the desired result.
14.11 Field Emission: W K B Approximation

In field emission the electron is literally torn from inside the surface of a metal.
A model for this process is as follows. T h e surface of the metal, located at
x = 0, is subject to an electric field S outside the metal, i.e. for x > 0 so that
the potential experienced by the electron is
\/i \- Í ~u x
< 0
— | _e£x x > 0
T h e electron is bound by an amount <f>, the so-called "work function". Thus,

the energy of the electron is E = —<j>.
Use the W K B approximation to show that the transmission probability for
an electron incident on the surface of the metal is well approximated by
T(E) = exp
4}
where
c 4V2m\E\3'2
S c
~ 3 ~h '
Solution
The transmission probability is given by | S | 2 where, in W K B approximation,
2
S = exp j^- J K(X') dx' (14.11.127)
14.12. DEUTERON: VARIATIONAL PRINCIPLE 303
and the integral runs over the classically forbidden region. Also,
k(X) = — E) . (14.11.128)
In our case we have

E
xi = 0 and x2 =
et
so that
fij rx 2 o /IFI\3^2
J K(X) dx = J \/xT^-x dx = -Ve£ J . (14.11.129)
Therefore,
2
T(E) \ =
- |IQI
Sf =- epyn
x p {í - ^ \ (14.11.130)
as desired.
14.12 Deuteron: Variational Principle

Assume that a deuteron is an I = 0 bound state of a neutron and a proton in
the potential
V(r) = -V0e~r'a
where
V0 = 32 MeV and a = 2.2 fm.
i(a) = Ae~ar'a
in a variational calculation to determine the binding energy of the deuteron.
Solution
T h e Hamiltonian of this system in the center of mass is
H = ^ - + V(r)
w
(14.12.131)
2m
where m is the reduced mass of the proton-neutron system. T h e energy is
estimated from
_ (14.12.132)
1
> WA)M<*)) '
i r,K 14. FURTHER APPROXIMATION METHODS
Evaluating the denominator (normalization) we get

rc
(rp(a),ip(a)) = \A\2 / e -2ar/ar2dr
J0
„3
= • (14.12.133)
T h e kinetic energy yields
T = (V'(a).^V'(a))
[°° e-°r/a i i ^ r e - a r / a \ ^ d j ,
2m Jo \r dr2 J
AW i°° e~2ar/a ( * ± - ^ dr
2m J0 \ a2 a J
/¿2|j4|2 a
(14.12.134)
2m 4a
The potential energy contributes
= M<*), V(r)ip(<*))
roo
2
= -\A\ V0 / e-2arlae-r'ar2dr
Jo
= ~w2v'J^TW • ( 1 4 1 2 1 3 5 )
Therefore,
h2
= ( 1 4 1 2 1 3 6 )
•
To use the variational principle we must minimize the energy with respect to
the parameter a. So we compute a from
0 a V a2
~~T^ - = — J ~ °~7 /0\4 • (14.12.137)
da ma2 2(q+1/2)4
After inserting the appropriate numbers we get the following equation for a.
(a + 1 / 2 ) ' = 3 w = 2 80 (14.12.138)
2
a 2h
A few straightforward numerical attempts yield the value
a = 0.67. (14.12.139)
After substituting this value back into the equation for the energy we find
2
E = 8.52 a -32 f ) MeV = - 2 . 2 MeV . (14.12.140)
va+1/2 j
14.13. BOUNCING BALL: VARIATIONAL CALCULATION 305
14.13 Bouncing Ball: Variational Calculation

A perfectly elastic ball, mass m is bouncing on a recoilless surface. Use a
variational calculation to find the energy of its lowest eigenstate. Choose your
own trial function.
Solution
We choose the potential energy to be zero when the ball is in contact with the
surface at say x = 0. Furthermore, since the ball cannot penetrate into the
region below the surface we must have that the wavefunction must vanish at
x = 0. Thus, we choose as a trial function
rl> = Axe~ax . (14.13.141)
The normalization constant A, with an arbitrary choice of phase, is
A = 2a3'2 . (14.13.142)
P
H = - 1- MGX . (14.13.143)
2m
Then we have, using this trial function, that the expectation value for the energy
is
rOO h2 d2
e + mgx x e dx
2m dx2
E(a) = = 4 a 3 H2A2x 2
{IP, HTP)hra — 2 ax
4a 3 x + mgx3 dx
M 2m
h2a2 3mg
Jo+ Lm (14.13.144)
2m 2a
The best value is obtained by minimizing E(a) with respect to the variational
parameter a. Thus, we compute a from
dE h2 a 3mg
0 (14.13.145)
m 2 a2
This yields the value
2 1/3
3m g
2 (14.13.146)
2h
Substituting this back into the expression for E we get
£
= (I) (-v»2)"3
1/3
= 1.36 (mg2h2)1^3 . (14.13.147)

i vi/MiuiS METHODS
This result should be compared with the result obtained from the W K B ap-
proximation (problem 14.6) which yielded
E= 1.84 (mg2h2)1/3 . (14.13.148)

It is clear that the variational calculation gives a better result since the energy
obtained is lower than the value obtained from the W K B approximation.
14.14 B e t a D e c a y of Tritium
A tritium atom (H 3 ) in its ground state /^-decays to form a singly ionized helium
a t o m (He 3 ). W h a t is the probability that this new a t o m will be found in its
ground state? Assume that both nuclei have infinite mass and that there is no
interaction between the beta-decay electron and the rest of the system.
Solution
W i t h the stated assumptions we can neglect recoil and need only compute the
overlap between the initial and final states. For the ground state of a hydrogenic
at om with nuclear charge Z we have
1 /2z\3^2
(1414149)
where do is the Bohr radius

K2
"o = ñ •
me¿
For the initial state we have Z = 1 so that
1 / 2 \ 3 '' 2
, 1 4 1 4 1 5 0 )
* •
For the final state we have Z = 2 so that

1 ( 4 \
e (1414151)
""" •
T h e probability amplitude for finding the atom after the decay in its ground
(*'•*>) 3
state is therefore given by
_ g3/2 í
0,
>oo
e~xx2dx
= j'- "-
JO
8 3/2
- g4 ~ 0.559 . (14.14.152)
Thus, the desired probability is

p
=\{i>J, V>f)| 2 « 0.312 . (14.14.153)
14.15. ANHARMONIC OSCILLATOR 307
14.15 Anharmonic Oscillator

For a potential of the form
V = Ax 4
use the variational m e t h o d to find the energies of
a) the lowest energy level and
b) the first excited state.
The corresponding energies as found by numerical integration are respectively:
2 3
T2 \ 2/3 / T2 \ /
1.060 -?-) A1/3 and 3.800 ( Í - ) A1/3
2m j 2m j
Solution
a) For the ground state we choose a trial wavefunction similar to the ground
state wavefunction for the harmonic oscillator.
ipo(x) (14.15.154)
Then,
r OO
2 2 ft2 (i2
(H) = E0(a) = -= / e-"2* / + a
4 -a2x2/2
J —CO 2^5^ *
• ^£h£ ( ° v -' , ' ) + A i

h2 a2 3 A
2 m ~2~ +
4o4 ' (14.15.155)
The best value for a is obtained from

DEO h2 3A „
= 0 -—A = 0. (14.15.156)
da 2m a5
So we find
1/,S
( 6mA\
a 6 = 6mA or a = (14.15.157)
2
h
Substituting this value for a in the expression for Eo(a) we get
_^_/6mA\1/3 3A f h2 \ 2 / 3
En 2
4m \ h J "^4 \6mA J
= ^ y ' v 3 . , 0 8 2 x ( i i y , a a - (14.15.158)
2m J 2m /
This is within about 2 % of the value obtained by numerical integration,
b) In this case we need a trial wavefunction orthogonal t o the wavefunction
for the ground state. For this reason we choose a wavefunction similar to the
wavefunction for the first excited state of the simple harmonic oscillator.
4cr,6 \ 1/4 -a2x2/2

IPI(X) = xe (14.15.159)
Then,
2a3 t" 2 2
E
' w
= x e -a x /2 . * L ^ L +
2
Xx*
2 mdx
2a3 f Í ft / 4 4 2 2o
= —¡= / - — (q x - 3 a x ) + Ax 6 e~a x dx
V* J-oo L 2 m
ft2 3 a 2 15 A
+ 4
(14.15.160)
2 ^ ¡ ~ y a '
Again the best value for a is obtained from
DEI 15A
= 0 (14.15.161)
da 2m 2 2a5
Thus, we find
1/6
10mA / 10mA\
2 or (14.15.162)
ft Q =
Hence, our best estimate for the energy of the first excited state is
. 2/3 / t2 \ 2/3
-i/d
EX 847 x í - ) a1/3 . (14.15.163)
2m J 2m y
This is within better than 1.5% of the value obtained by numerical integration.
14.16 Nonlinear SHO: Variational Calculation

\t[>) = cos 0|O) + sin 9\2) ,

where |0) and |2) are the ground state and second excited state of the unper-
turbed simple harmonic oscillator respectively corresponding to the Hamiltonian
HO = ^ — i - - m u ; 2 * 2
2m 2
and 9 is a variational p a r a m e t e r , to calculate t h e ground s t a t e energy of the
Hamiltonian
H = H0 + \\x4
4
Compare the result with that obtained by first order perturbation theory. (Prob-
lem 12.4 a).
14.16. NONLINEAR SHO: VARIATIONAL CALCULATION 309
Solution
The trial wavefunction is already normalized. So we need only compute
E(0)= (rp\H\4>) (14.16.164)
and minimize this result. To this end we need the matrix elements (0|Jio|0) ,
(2|tfo|2) , and ( 0 | t f o | 2 ) = (2\Ho\0), as well as ( 0 | x 4 | 0 > , ( 2 | x 4 | 2 ) , and <0|x 4 |2) =
4
( 2 | x | 0 ) . Now, we have
<0|tfo|0> =
<2|tf 0 |2) = |/»u>

<2|// o |0> = (0|tf 0 |2> = 0 . (14.16.165)
Also, in terms of creation and annihilation operators,
h 4
x^ -
= I V ( a t + a )
2 mui
= | — | (a 4 + a 3 a f + a 2 aâ + a 2 a t _ + a a f a 2
+aaâa^ + aa* a + aa* + aâ 3 + a}a2a} + a}aa)a

+ a ^ a ' + a ^ a a * + a^ 3 a + a* 4 ) . (14.16.166)
Next, using that
a|n) = \ / ñ \ n — 1)
a^n) = V n + l|n + l) (14.16.167)
we get
( ^ (0|x4|0) = ( 0 | a 2 a f " + aataa^lO)

Y 2TTLLD J
= 3 (14.16.168)
as well as
(d=)" <2|l4|0> - (¿)~ (0|I<|2 >

-2 / t \ - 2
(2|aa^ + aâa) + aâa^\0)

6\/2 . (14.16.169)
Finally,
F H \ O
4
( 2mu ) ^2|x |2) = (2\cra}~ + aa)aa} + aa}~a + a)aa)a + a^"a 2 |2)
fjtiAfl'KH 14. FURTHER APPROXIMATION METHODS
Combining all these results we get
L A f h
E(6) = cos 2 — hu) -|- 3— [
2 4 I 2 rwu
5. ( h
+ sin -hu + 39— (
2 4 \ 2 mu
A ( h \4
+ 2 sin 9 cos 9 6\/2— ( ] (14.16.171)
4 \ 2 raw J
Using that
cos 0 = -[l+cos(20)]
sin 2 9 = - [ 1 —cos(2#)] (14.16.172)
we find
E(0) = - cos (20) hoi + —A + sin(20) ^ - ^ A

2mu> 2 mw /
21A / h
+ -hu + (14.16.173)
4 V 2 mu
To minimize this energy we differentiate with respect to 9 and set the result
equal to 0 to get
sin(20) hu -f- —A . = —cos(20) — A (14.16.174)

2 V 2 mu 2mu
Therefore,
3\/2
tan(20) = — (14.16.175)
9 + 8(hu>/X)(mu/h)2 '
if A is sufficiently small so that we can compare this result with perturbation

theory we get
3a/2A
9 rj - (14.16.176)
16fiw
If we now substitute this result into the expression (14.16.171) for the energy we
get the approximate ground state energy. To simplify the writing we introduce
the parameter
A h2
(14.16.177)
4 m2u2
14.17. WKB SOLUTION AND PARITY 311
Then,
1 3 9 72 , 45 f _ 9 72
0
~ 2 4^ 16 huj 16 hu 2 hu>
2
1 3 9 7
= —ñu H—t . (14.16.178)
2 4 4fiw
A comparison with the perturbation result (Problem 12.4 a) shows that these
results agree to order A and differ by about 14 % in the A2 term.
14.17 W K B Solution and Parity

Show that for a parity invariant potential
V(x) = V(-x)
and an energy E such that there are four turning points, say at
x = ±a , x = ±6 b>a> 0
it is not possible to write W K B solutions that are simultaneously eigenstates of

the parity operator.
Solution
We begin by writing the solutions in the four "external" regions
x < — b , —b<x< —a and a < x b .
We do not need to consider the "central" region —a < x < a . Thus, with
2 777,
«(*) = —5-[V(z) — E] E<V, x <—b or x>b
¿
N
k(x) = 2m
[E — K(x)] E > V , —b<x<—a or a < x b
Here, we have already imposed the conditions that the solutions must be ex-
ponentially damped for |x| —> oo as well as the matching conditions at x = — 6
and x — b. Letting x —T — x and remembering that both K(X) and k(x) are even
functions of x the solutions become
^ y e x p [-¡:*K(x')dx> x > b
cos [ / _ * k(x') dx' - | a < x < b
ip(-x) = (14.17.181)
2B
c o s -b < x < —a
y f e
exp •FB X
K(X')DX' x < —b
Vk(x)
We now impose the conditions that these solutions should be eigenstates of the
parity operator. Thus, we want
ip(—x) = ±xp(x) . (14.17.182)
Hence, we get the conditions
A exp J K[x')dx' ±B exp •J K(X') dx' (14.17.183)
as well as
A cos f k(x')dx'—— = ±B cos Í k(x')dx' —— (14.17.184)

4
.J-b J i Jb 4
From (14.17.183) we get by changing the integration variable on the right hand
side from x to — x that
B -- ±A exp —2I J K(X') dx' (14.17.185)
Similarly by changing the integration variable on the right hand side of (14.17.184)
we get
Acos^J k(x')dx'—— =±5cos J k(x') dx' + ^ (14.17.186)
But, both A and B have to be constants independent of x. So, the only possi-
bility is that
A = B = 0 . (14.17.187)
In this case we have no solution at all. Thus, these solutions can never be parity
eigenstates.
BIBLIOGRAPHY 313
Bibliography
14.1 A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
ibid, section 14.7.
ibid, section 14.8.
14.2 P.M. Morse and H. Feshbach, Methods of Theoretical Physics , Vol.1 -

McGraw-Hill Book Co., Inc., New York (1953).
Chapter 15
Time-Dependent
Perturbation Theory
15.1 Transition Probability: B o u n d State to Free

A particle is in the ground state of the Hamiltonian
where
y _ í 0 x < -a , x > a
~ {1 — Vo —a<x<a
Find the transition probability per unit time to a state of energy Ek > 0, due
to a perturbation
x
H'(t) = ve Ia sinujt
where i; is a constant and a « a.

Hint: The normalized bound state solution is given in [15.1] and the normalized
continuum solution is given in problem 8.3.
Solution
We need both the normalized ground state solution and the normalized contin-
uum solution. The normalized ground state solution is the even parity solution
given in section 4.5 of [15.1].
If we define
2m\E\
~hT~
2m(\E\ + V0
K.2 (15.1.1)
K2
15.1. TRANSITION PROBABILITY: BOUND STATE TO FREE 315
then the ground state solution is
{ A eK°x
B cos
A e~K°x
The equation for the binding energy is
KX
x < —a
|x| < a
x >a
(15.1.2)
NO
tan na = — . (15.1.3)
K
The normalization follows from

(15.1.4)
A — B eK°a cos Ka
'i
and
-2K0a
1 = \A\2 e-2KoXdx + ' dx

Ja /'
1 tan JKa
a
\A\2e-2Koa h +
KQ
\A?e~ a+ (15.1.5)
«o
Therefore, we get,
K0 K0
A= B = (15.1.6)
1+ KOA 1 KQÜ
The normalized continuum solution is obtained from the solution given in prob-
lem 8.1, as well as
cos (ka + ¿+)

A= (15.1.7)
cos K a
The overall normalization is found in problem 8.1 to be 1/%/^-
We only need the positive parity solution since the perturbation potential is
an even function of x and the ground state is even parity. Now we are ready
to use Fermi's golden rule. To do this we need the density of final states.
This follows from putting the continuum solution in a box. Fortunately the
perturbation is extremely short range and we only need the solution inside the
well. So if the box is of length L » 2a we get that
2nn t 2 p fc2 js2

K = £ + ÍLÍL = . (15.1.8)
2m 2M
So,
dE dK and dK = — dn . (15.1.9)
m L
^ •
Therefore, the density of final states is
i?/
(15.1.10)
316 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORY
We next need the matrix element. T h e perturbation is
H' = ^ e x p [ - x 2 / a 2 ] 2 s i n w i . (15.1.11)
Since a « a we need only consider the wavefunctions near x = 0, but we can

extend the integrals t o all of x. Therefore the relevant matrix element is
v I «o 1 f° „ , ,/ 2, j
cos
M = 7= / A x cos KX exp — x A \dx
2 v 1 + k0a j-oo
^ ~[exp[_<,<A + K) /2]
v I «ó 1 a-vA' r r / T.r \2/(VI
= 2V
+ exp[—q(A' — k)2/2]] . (15.1.12)
Putting all this together we finally get w, the transition probability per unit
time.
w = —\M\2p(E) (15.1.13)
or more explicitly
w = mv2a2 «o a
3
32fi K 1 + kqü
x [exp[—q(A' + k ) 2 / 2 ] + exp[—a(A' — k ) 2 / 2 ] ] . (15.1.14)
15.2 Photo-disintegration of D e u t e r o n
The deuteron is an s-wave (I = 0) bound state of a proton and neutron with
a binding energy of 2.226 MeV. It is well approximated as a bound state in a
square well of depth Vo = 36.2 MeV and width a = 2.02 x 1 0 - 1 3 cm . Using
these data, c o m p u t e the probability for photo-disintegration of the deuteron.
A s s u m e the incident photon can be approximated by a perturbation
TT __ f e^4 • r s i n w í t > 0
0 t < 0
where A is a constant vector of m a g n i t u d e about 1 x 10 3 V / c m . Use whatever
other approximations seem reasonable.
Solution
Since, Vó is the depth of the well, we define
2
i» = - ^ > 0 , K ' -- -^lf S > 0 . (15.2.15)
Then the radial equation for / = 0 reads
id2 „
-~r-^(rR)
r dr2 + K R = 0 for r < a
1^2
— -j-^(rR) — k2 R = 0 for r>a. (15.2.16)
• dr2
15.2. PHOTO-DISINTEGRATION OF DEUTERON 317
T h e solutions are
„ sin K r _
R — A for r <a
r
R = B - for r > a . (15.2.17)
At r = a both R and dR/dr are continuous. Therefore,

ka
B = Ae~ sinA'a . (15.2.18)
Equating the logarithmic derivatives now yields
Ka cot Ka = —ka . (15.2.19)
The solution of this transcendental equation yields the bound state energy which
for the values given is stated to be 2.226 MeV. So the wavefunction is (up to
normalization)
f ¿si"Kr r < a
R(r) = \ .r . (15.2.20)
1 A sin A a r>a
V R
Throughout, m is the reduced mass « 1 / 2 m pro ton- Thus,
K « 9.66 x 10 1 2 c m - 1 . (15.2.21)
This means that we can neglect the part of the wavefunction that is inside the
well and simply write
„-k(r-a) -kr
R(r) « B = B' . (15.2.22)
r r
The normalization of this wavefunction yields
\B'\ 2 = ¿ (15.2.23)
so that
= í : (15 2 24
*» v Í t - - ' »
This specifies the initial state with
2 m
] f ^ w 2.32 x 10 1 2 c m " 1 . (15.2.25)
h¿
For the outgoing state we take a plane wave
«w= (1^-26)
Next we c o m p u t e the density of final states. To do this we take the outgoing

plane wave and discretize the energy levels (to be able to count them) by placing
the particle in a cube of sides L. In this case the permitted wavenumbers are
2 7T v
q = -¡rinr, ny, nz) (15.2.27)
¿IB CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORY
where nx, ny, nz are integers. Now we can count the levels. The number of
modes lying between qx and qx + dqx, qy and qy + dqy, as well as qz and qz + dqz
is
dN - AnxAnyAnz - ~dqx dqy dqz

z7t z7t z7t
= q2 d q s i n 6 d < f i
= ^ (j^r) ' (15.2.28)
We use spherical coordinates in momentum space since energy conservation fixes

the magnitude \q\ = q according to
Eq = Eb + hj . (15.2.29)
Now using (15.2.28) we get the density of final states

m
p{q) = (152
J 2 ^ V Q D Ü
- -30)
We now have to evaluate the matrix element
(ipout, ^ A • ripin) = A • J re'^r ^——r2 dr sin 9 dd dip
e y/k - -
= (15.2.31)
where
r oo /»7t
1= rdre~kr I f eiqr cosB
sin 9 d9 . (15.2.32)
Jo Jo
By symmetry, this integral points in a direction parallel to q so
I = ql or I = ~ q l . (15.2.33)
Thus,
/ = i r°° f1
J rdre~kr J qrue'qru du
4 ik2
q2(q2 + k2)2' (15.2.34)
So,
4z k 2
1 q
^ q*{q2 + k2)2 (15.2.35)
and the matrix element squared is
e2
. ,2 & (q A)216k4
' ' ~ T (2tt) 2 qs(k2 + q2y ' (
^
15.3. EXCITATION OF SHO 319
We now put all these results into Fermi's golden rule

27t
dw = — x square of the matrix element x density of final states
H
e2 q2Á2 cos2 9 m
= s 2 2 4
87xh q (k + q ) (2tt) 3 /í 2<?
4e2 1 mk5 A2 cos2 9
3 5 2 2 4
~ ( 2 t t ) 4 h q {k + q ) ( •• • )
15.3 Excitation of SHO

An ato m is initially in the ground state of a simple harmonic oscillator
H = hui aâ .
At t = 0 a perturbation
V' = h£l(a^ + a)
is turned on. Find the transition probability to any excited state of the system
for t > 0. W h a t is the probability that the a t o m remains in its ground state for
t >0?
Solution
Here we have an example of the sudden approximation. So for í < 0 we have
= |0,a)e-iwt/2 <<0 (15.3.38)
where
a|0, a) = 0 . (15.3.39)
For t > 0 we have to solve for the eigenstates of the Hamiltonian
H = huj oâ + Kl(a) + a) . (15.3.40)
Writing b = a + c where c is just a c-number we find that
H = fiw(6t-c*)(6-c) + ftO(fct+6-c*-c)
q2
= Hb = hwbH - h— . (15.3.41)
UJ
Here we have chosen c = f2/w. Thus, in terms of the operators 6, 6* the

Hamiltonian Hb is diagonal. T h e energy eigenvalues are
q2
En = (nb + l/2)hcj - h— . (15.3.42)
UJ
We have to re-express |\&(i)) in terms of the eigenstates of this Hamiltonian.

For this purpose we find the unitary operator U such that
b=UaUl . (15.3.43)
320 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORy
This is easily done if we realize that for any two operators A, B such that
[A, B] = c-number we have
AeB = eB A + [A,B]eB (15.3.44)

So we try
U = ec (15.3.45)
Then,
i/t = e-(ca-c*at)
(15.3.46)
So,
U aU^ = UU^ a + U[a, —(ca - c*a})]U^ = a + c* (15.3.47)

Therefore, choosing
(15.3.48)
u
yields the desired transformation. It then follows that
|0,a) = C/t|0i6) (15.3.49)

where also
|n,6) = 4 r ( f r t r i 0 1 i ' > (15.3.50)
is the eigenstate of H¡, corresponding to the eigenvalue
n2
En = (nb + l/2)hu> - h— . (15.3.51)
U!
The probability to find the system at time t in the state
|n, 6, t) = | n, b) e~lEnt!h (15.3.52)
is
Pn(t) = |(n,6,*|*(i))|2
= ¿|(0,6|6"C/t|0,6)|2
= l|(0,6|6"|0,a>|2. (15.3.53)
Therefore,
Pn(t) (0,6|6ne-n6t/w e«f>/ê-(n3/2^)|oj6)|2

n\
= _ e - ( n 2 / " 2 ) (0, b\e~^lu
j1_ enb/u (b - Q/uj)n |0, 6) | 2
n\
2n
j_ e-(«2/^2) I (15.3.54)
o. iimu-UUPUNDENT PERTURBATION THEORY
15.5 Periodic Perturbation

An atom has two energy levels ±hQ. A weak disturbance V(t) connecting these
two levels and varying periodically in time such that
(l|y(i)|2) = hq.i sinw¿
is turned on at t = 0.
a) Find a model Hamiltonian for this system.
b) If the atom was originally in its ground state, estimate the probability P(T)
that it is in its excited state at time t.
Solution
a) The model Hamiltonian is H = HQ + H' where
(15.5.63)
for t > 0 . (15.5.64)
b) Let
(15.5.65)
The Schrodinger equation now reads for t > 0
ih—- = Ml a + hCli sinu>t b

dt
ih-— = — h£l b + h£li sin uit a . (15.5.66)
DI
After simplifying these read
— = — id a — iQi sinwi 6
~77 = iClb — iCli sin uit a . (15.5.67)
At t _ 0 we have a = 0, b = 1. So, we put

int
a = e~ A(t)
mt
b = e B(t) . (15.5.68)
Then,
dA 2 nt
—iQi sinwí e ' B
dt
dB
—iCli sin uit e 2'nt A (15.5.69)
dt
15.6. EXCITATION OF H-ATOM 323
where at t = 0 we have ^4 = 0 and B = 1 . Thus, to lowest order in Qi we get
A(t) » —iCli f sin uit' e 2 , í i í ' dt'

Jo
¿ííi 1r e , ' ( w + 2 n ) t - 1 + e «(«-2n)t _ I
+ 2ÍÍ - 2 Q
2¿12i
2 e 2 , n ' [w(coso;< — e 2,f2t
) — 2¿Q s i n w í ] (15.5.70)
UL -fi2
and
B(t) « 1 . (15.5.71)
The probability to find the particle in the excited state is given by
P(t) = |¿(*)|2 - (15.5.72)
When evaluated this yields

2
fi2 • 2 ft •
2 + 3 sin uit — 2 cos uit cos 2S7i + 4 — sin uit sin 2Qt .(15.5.73)
(w - fi2)2
2
Ul
15.6 Excitation of H - a t o m
A hydrogen atom in an excited state \n,l,m) is perturbed by a uniform electric
field. If the interaction can be written
eE-f2s'muit 0 <t<T
V(t)
0 t < 0 , t > T
Find an expression for the transition probabilities to a lower level \n', m'). Do
not attempt to evaluate the radial integrals. T h i s is how intensities of spectral
lines can be computed. You may use the results of problem 9.54. See also [15.2],
Solution
T h e state | n , l , m ) is described in configuration space by the wavefunction
= Rn i(r) (fi) (15.6.74)
T h e perturbation is
eS r(sin 9 cos ip, sin 9 sin <p, cos 9) sin uit 0 < t < to
V(t) = | (15.6.75)
0 t < 0 , t > to
The transition probability is given by
P = N',V
gl(w+w„„,)to _ x
(15.6.76)
2(ui -f uinni)
Here we have En < Eni and
En — E„i 1 1
^nn' — 2
(15.6.77)
2Hao \n n'2
N o w , we have
J DNY;m(0,<p) sin 0e^Y,'m,(9, F)
I (I — m + \ ) ( l + 2 — m )
= <5rn,m'
, +1 S, ,¡+i
(2/ + 3 ) ( 2 / + 1)
I (I + m)(l + m — 1)
+ (21 - 1 ) ( 2 / + 1) I-1 (15.6.78)
/ dQY*m(9, ip) sin 9e ,lfi

Yi'm'(9,<p)
I(/ 771 2)(/ ~f~ fTl + 1)

VM.M' — 1 SI' ,/+l
(2/ + 3)(2/ + l )
I (/ — m)(lm — 1)
+ SI\ / - i (15.6.79)
(21 — 1)(2/ + 1)
/ dtt Y*m (9, tp) cos 9 YVm,
2(1 — m — 1)(/ + m ) !
(9, <p)
= <5, SI',I+I
(2/ + 3)(2/ + 1)(/ — m)\
2(1 + M + 2)(L + m ) !
+ °L',I-1 (15.6.80)
(21 — 1) (2/ + 1)(/ — m ) !
T h i s explicitly displays the selection rules A / = ± 1 , A m = 0 , ± 1 . We are now

left with only the radial integrals.
n oo
I„y = / Rn,i(r)r Rnj±i(r)dr . (15.6.81)
Jo
T h e s e integrals can be expressed in t e r m s of hypergeometric functions. T h e

results, however, are n o m o r e i l l u m i n a t i n g than the integrals themselves so we
leave t h e m in this form. T h u s , we finally get
4 e 2 £ 2 sin2[¿(w„,„' + w)i0] (I + m)i

P =
K2 (un,n' + w)2 . ( 2 / + 1)(/ - m ) !
2
{(4^) w+(- + m — 1x 2
K I (15.6.82)
21 - 1
15.7. EXPANDING BOX 325
15.7 Expanding B o x
A particle is in the ground state in a one-dimensional box of length L. Sud-
denly the box expands (symmetrically) to a length 2 L leaving the wavefunction
undisturbed. Calculate the probability that the particle will be found in the
ground state of the expanded box.
Solution
For a box extending from -L/2 to L/2 the ground state wavefunction is given
by
, / \ ^ í 7T«E \ •«-> J-/ /i r w nn\

cos
M*) = y ¿ ( T ) ~2 2' (15.7.83)
Also for a box extending from — L to L the ground state wavefunction is given
by
= \J~-j- cos —L < x < L . (15.7.84)
At t = 0, where this is the time at which the box expands, the state is the same
as that of the unexpanded box. Thereafter the state evolves according to the
Hamiltonian described by the box of length 2L. Thus, for t > 0 the state is
described by
005
*«,*) = \/5 (fir)cos (x) ^ { i \2LJ
( 7VX \
—J (15.7.85)
where
„ = . (15.7.86)
This is the sudden approximation. The probability for finding the particle in the
gound state is now given by the overlap of this wavefunction with the ground
state wavefunction, namely
e,W c o s x
$0(<,x)=i V L (Y) 2 ^ ^ 2 . (15.7.87)
0 |z| > f
Thus, the result is

2
~2~ rL/2 cos 2
P =
l2 j-L/2COS (S) (iir ] d x
' 37t
(15.7.88)
1
+.J. 11iviívutiPENDENT PERTURBATION THEORY
15.8 Sudden Displacement of SHO

A particle is bound by a simple harmonic oscillator potential and is in the first
excited state. If a perturbation
V' = Ax
is turned on at time t = 0, find the probability that the particle will be in the
new ground state for times t > 0.
Solution
For t < 0 we have a simple harmonic oscillator with
Ha = hu>(aâ + 1/2) . (15.8.89)
For < > 0 w e have a displaced simple harmonic oscillator with
H = a + 1/2) + X\ (a* + a) . (15.8.90)

V ¿TT10J
If we now introduce new operators
b= a+ c , tf = ¿+c (15.8.91)
where
C2 =
2 ^ <15-8'92)
then, the Hamiltonian for t > 0 is again diagonalized and becomes
Hb = hu(bH + 1/2) - ^ 2 • (15.8.93)
If we call the eigenstates of H b , |n, 6) with corresponding energies
A2
En,6 = (n + l/2)ftw - J • (15.8.94)
Then the ground state |0, a) annihilated by a is a coherent state |c) and is given
by
| C ) := e
~ C
' 2
Y , - n |n'6>' (15.8.95)
n=»v«!
At t — 0 the state is given by the first excited state of 11a and is
at|c) = (&t + c ) | c ) . (15.8.96)
But,
OO
C
+ c)|c) = e ~ ° 2 t 2 y;_=[c|n,&)+v/ir+T|n+l,6)]. (15.8.97)
v"!
15.9. SUDDEN PERTURBATION OF TWO-LEVEL ATOM 327
This means that for t > 0 the state is given by
w
rn
9
X Y -^=[c\n, b) e~inwt + V^Tl\n + 1,6) e — ^ + 1 ^ t ] . (15.8.98)
nr r=f0, VN\
Therefore the probability amplitude for finding the particle in the new ground
state is (0,6|\P) and the probability is given by
| ( 0 , 6 | ^ ) | 2 = c 2 e~°2 . (15.8.99)
15.9 Sudden Perturbation of Two-level A t o m

An atom has two energy levels of energy ±E. So the Hamiltonian may be
written
H = £V3
where
1 0
0"3 = 0 -1
If this atom is in the ground state, and at time t = 0 a perturbation
0 1
H' = V .
1 0
is turned on, find the probability t h a t this atom is still in its ground state at
some later time t.
Solution
If we make the replacements
E = nno , v = hSh (15.9.100)
the Hamiltonian is
H = /t«0 ( ¿ _ ° j ^ = H0 = HN0<T3 T< 0 (15.9.101)
and
H = Mlacrz + hSlicri = H\ t > 0 (15.9.102)
or
1 0
H = hQo ( J ^ ) + «I! ( J J ) = > • (15.9.103)
The eigenstates of Ho are, as before,
!+>=( J ) '->=(l) (15.9.104)
with
E± = ±KLo . (15.9.105)
If we call the eigenfunctions and eigenvalues of H , and respectively, we

have for
' a
M- [ b ) (15.9.106)
the eigenvalue equation
(15 9 107
* ( nt ) ( » ) = <' ( I ) ' -- '
The eigenvalues are:
£± = ± f t ^ í 2 g + í2j . (15.9.108)
Calling
(15.9.109)
we have
e± = ± h Q . (15.9.110)
We then get:
V± ± 0
~ v ^ Q ( 0 - fi0) ( - ) (15.9.111)
If at t = 0 the system is in the state u+ = |+), then for t > 0 the system will
be in the state
tf(f) = b+v+e~int + fe_w_eint (15.9.112)
where b± are determined by the initial condition
u+ = b+v+ + 6_t;_ (15.9.113)
or
b±=(v±,u+). (15.9.114)
Thus,
6± = . Ql (15.9.115)
v
^ ( q í f l o )
Combining these results we get
(>«»•)
This wavefunction is an exact solution of the Hamiltonian (15.9.101) and
(15.9.102) with the initial condition *£(0) = u+. The approximation was made
in writing the Hamiltonian (15.9.101) and (15.9.102) in the first place.
X t/ • J- vy • j-rj-jx. u i i iJ ±
15.10 Berry's P h a s e
where
5 ( i ) = Bo [sin 0cos éx + sin 0sinwí ey + cos 9 éz]
so t h a t
B(t + 2n/u)) = B(t) .
If U « HBQ show that
(9(t = In/uj)\y>(t = 0)) = exp pBo^j exp[—¿7r(l — cos#)]
-L
where ^J-/J,BO is just the dynamical phase
27t/uj
E(t) dt
and
—7r(l — cos 9) - — A i l / 2
is the geometric phase (Berry's phase).
a) Use the formula
7„(C) = £(n,B\VB\n,B)dB
for the geometrical or Berry's phase.

b) Use the formula
d-li .( dv± \
= L V ±
-DF { ' ^ R )
where v± are the "instantaneous" eigenvectors of H(t).

c) Finally solve the exact eigenvalue equation for H(t) and compute
(9(t = 2 j r / w ) | t f ( * = 0)> .
— v j- j i r i J -iyr,riiwBJVT PERTURBATION THEORY
Solution
We are given
H f f
(t) = - ^ B ( t ) (15.10.117)
where
B(t) = So (sin 0 cos ut ex + sin 9 sin ut éy + cos 9ez) (15.10.118)
so that
B(t + 2x) = B(t). (15.10.119)
Using hw « i-iBo we want to show t h a t
(9(t = 2 j t / w ) | ¥ ( t = 0)) = exp e-"(1±coswt) (15.10.120)
where the dynamical phase is

F2IR/U)
± 7 r^ ° = _ i / ' jB(i')d<' (15.10.121)
nu) n Jo
and
- 7 r ( l ±COS0) = ( 1 5 . 1 0 . 1 2 2 )
is the geometric phase (Berry's phase).

We begin by writing
• r2x/a>
1^(0) = ex
P - j( E(t')dt' p!'Tn(í)| ,B(t)). (15.10.123)
Now we apply the theory developed for Berry's phase (see [15.3]). To illustrate
the theory we solve the problem by three different methods.
a) Thus,
1n(C) = j ) (n, 5|Vs|n, B) • dB (15.10.124)
where the contour C is due to the fact t h a t
S ( 0 ) = B(2TT/U>) . (15.10.125)
To compute the dynamical phase we need the "instantaneous" energies which

are the eigenvalues of
H(t) = - ^ f f - B ( t ) . (15.10.126)
Writing out the eigenvalue equation

det \E — H(t)\ = 0 (15.10.127)
15.10. BERRY'S PHASE
we get the two eigenvalues
£ ± = ± ^ | i . (15.10.128)
The corresponding eigenvectors are

sin(<?/2) \ „ _ ( cos(0/2)
(15 10 129
- i « ( $ ) * & j ' = { ri»(?/2) ) ' - - )
It now follows that the dynamical phase is given by
if - íjf * TSIDV
2
= . (15.10.130)
hid
The geometrical phase (Berry's) phase is given by
7„(C) = i j ( n , §\VB\n, B) • dB (15.10.131)
where
B(t) = 5o(sin 9 cos cot ex + sin 9 sin cot ey + cos 9 ez) . (15.10.132)
Thus, using "spherical coordinates" in parameter space we have
VBv±= ( ¿ - . - ¿ - 4 . - 5 - ( 1 5 . 1 0 . 1 3 3 )
\ a B o #o aw B 0 sin y awi /
Carrying out these differentiations and taking inner products we get
K , Vbw+) = (0, 0, — cot(0/2)) = — cot(0/2)e„ (15.10.134)
and
(v_,VBv_) = (0, 0, tan(0/2)) = - ¿ - t a n { 9 / 2 ) é v . (15.10.135)

¿n o o
We still need to compute dB/dt. This is straightforward and yields

D
B
= Bow sin o(— s m w i e x + cosuitey)
= Bqco s'm9ev . (15.10.136)
Now, we are finally ready to compute the geometrical phase using (15.10.131)
7+ ( C ) = i f —l— cot(9/2)BQU) sin 9 dt

Jo 2Bq
= — 7T c o t ( 0 / 2 ) s i n 0
= -7r(l+cos0) (15.10.137)
7-(C) = i ——- tan(0/2)Bow sin 9 dt

JO2b0
= — 7T tan(0/2) sin 9
= TT(1 COS 9) . (15.10.138)
b) Another more direct way to compute the geometrical phases is to use the
equations
í a = , ' ( „ ± , ^ ) . (15.10.139)
In this case we immediately find
^ = -wcos2(0/2) . (15.10.140)
Then,
/j 2n/u)
7+(C) = uit cos 2 ( —)
¿
0
2
= —27t cos (0/2)
= — 7r(l + c o s 0 ) . (15.10.141)
Similarly,
DF.
= —to sin 2 (0/2) (15.10.142)
dt
7-(C) = - w i sin 2 ( —)
= —27r sin 2 (0/2)

= —7r(l — cos 9) . (15.10.143)
c) Finally, we can also simply just solve the time-dependent Schrodinger equa-
tion exactly. To do this we look for a solution of the form
/ p-iwt/2 \
*(*) = ,„.W2
beiwt/2
K * (15.10.144)
with
M2+|6|2 = l .
The time-dependent Schrodinger equation
¿/I— = #(*)*(*)
15.11- NEUTRON IN ROTATING MAGNETIC FIELD 333
now reads
ft(w/2 - A ) a e - W 2 - A ) '
-h(u/2 + A)6 e ' ( w / 2 + A ) i
hb0 -iwt ae-'(W2-A)t
cos Í7 sin c/ e (15.10.145)
Oicot f)ei(w/2+\)t
2 V sintfe""' —cos
Thus, after some simplification, we get the eigenvalue equation

//is0 w\ [¿BO .
—A a = —— cos 0 — — a + —— sin 0 o
\ 2ft 2J 2ft
i Lcos#
Hence,
\ a
—Xb= • 6/a
—— sin d
2N (^ -?" (15.10.146)
/ /xbo \ w
— cos 0 +
2/5 V 2h ) 2
II is (15.10.147)
v 2h 2
Here we have used the approximation that
u « fib0 •
With the initial condition
¥(0) = (15.10.148)
we again obtain the same result as before.
\2tt
($(27r/w), >í(0)) = exp i—A — in
U>
= EXP ITT n _l a\
( 1 ± c o s i )
(15.10.149)
±2^7-
15.11 N e u t r o n in R o t a t i n g Magnetic Field

Consider a neutron (charge = 0), magnetic moment
eft _
=
» -G2M*
in a magnetic field consisting of a uniform component B in the z-direction and

a component b(t) rotating in the x — y plane. Assume that at t = 0 the neutron
has its spin pointing in the negative z-direction. Find P+(t), the probability
that at time t the neutron has spin h/2 in the positive z-direction. Find the
condition between B and b such that the amplitude of oscillation of the spin
between +h/2 and —ft/2 is a maximum.
334 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEOliy
Solution
In this case the Hamiltonian consists of
H = 4- hij[a-x cos at + cry sin at] (15-11-150)
where a is the angular frequency of rotation of the magnetic field b(t) and
eB eb
= U = i
^2mc ' 2 ^ ' (15.11.151)
The state at t = 0 is given by
(15.11.152)
The state for a time t > 0 is taken to be
* ( « ) = ( l+_ e > i T ) • (15.11.153)
To first order we then have
a_ = 1 (15.11.154)
and
a+(t) = — ioj J (1, 0)(<rx cos at + ay sin at) ^ ^ ^ e _ 2 , n t dt . (15.11.155)
Therefore,
«+(<) = ^ — [ e ~ ' ( 2 n + a ) t ~ ll • (15.11.156)
2í2 -|- GL j
Then the probability

P+ = |a+(i)| 2 (15.11.157)
is given by
A, .2
p + sin2
~ 10o _l \2¿ ( 2 q + a )T • (15.11.158)
+ a)
For the oscillations to be as large as possible requires t h a t
4w 2
(2fi + a) 2
be a maximum. This means t h a t we want w or b as large as possible.

12. EXCITATION OF ELECTRON BY ELECTRIC FIELD 335
15.12 Excitation of Electron by Electric Field

Compute to lowest non-trivial order the probability t h a t an electron in the
ground state of a simple harmonic oscillator at t — —oo will be found in an
e x c i t e d state at t = oo if perturbed by an electric field
£ T 2
e -(*/ )
y/ÑR
The electron continues to be bound by the simple harmonic oscillator potential.
Solution
The perturbation potential is due a dipole interaction and is
£ 2
V(t) = -ex-ê~<-t/T)2 . (15.12.159)
Y/TTT
The probability amplitude for a transition from the state |0) to the state |m) is
1 f°°
an,o = ¿n,o + — Yt / dt (n\V(t)\m)amfie,u'nmt (15.12.160)
tfl
M J~°°
where
Unm = {En — Em)/fr= (n - m)uj . (15.12.161)
The matrix element ( n | \ / ( i ) | m ) is given by

at
(n\V(t)\m) = + )lm> e~(t/T)2
e_(t/T)2
= + Vm + l i „ , m + i ) . (15.12.162)
Therefore,
a
1 CS I TX
n,0 = ¿«,0 - {VmSn,m-l + Vm + l¿n,m+l)
/ oo
dt e_(t/,T'2 e
iu n m t
(~ )
-oo
= ¿n,0 - (V™&n,m-1 + V™ + l¿n,m+l)
v
in J-KT V 2 ma; '
m
xam,0
%
=
<5n,0 + y/2mtvJe ( v / n T T o n + 1,0 + l / " a
n-l,o) • (15.12.163)
This equation is still exact. To use perturbation theory we simply iterate this
equation. Then,
= w T
«n,0 ¿n,0 + ^ ^ e t4 (y/n + 1 + i o + y/ñ CIrí-l,o)
•
« Sn,o + ^ Í - ^ - - e - " 2 T 2 / 4 ¿ n _ 1 | 0 . (15.12.164)
This equation also shows that, to lowest order, a transition can only occur to
the state with n = 1. The resultant probability is
|ai,o|2= e_w2r2/2
• (15.12.165)
Zmriuj
15.13 N e u t r o n Magnetic M o m e n t
A neutron has a magnetic moment fi n . A free neutron is placed in a magnetic
field B pointing in the z-direction and at time t = 0 the x-component of the
spin of the neutron is measured to be h/2. If this measurement is repeated at
some later time i, calculate the probability of obtaining the same result.
Solution
The Hamiltonian describing the evolution of the spin of the neutron is given by
H = —pn • B (15.13.166)
and may be written
H = tkj<T3 (15.13.167)
where, as always,
<r 3 = ( J ^ ) . (15.13.16K)
At time t = 0 the state is described by
V(< = 0) = - ^ ( J ) . (15.13.169)
The equation for the evolution of this stal e reads

lb
ih^=hw<T3iP . (15.13.170)
The general solution of this equation is

m = ( H - L ) (15.13.171)
15.14. ELECTRON PASSING THROUGH MAGNETIC FIELD 337
where a, b are integration constants. Imposing the initial condition we find

( tiwt
1 \
m = ^ { e~iwt J • (15.13.172)
Therefore, the probability of obtaining the value h/2 for a measurement of spin
along the x direction is given by
l(v>(0),^))|2 = i l e - ' + e--""! 2
= cos 2 uit . (15.13.173)
15.14 Electron Passing Through Magnetic Field

An electron (magnetic moment ^) enters a uniform magnetic field B which acts
at right angles to its velocity v. If the path length through the field is L and the
electron's spin is initally perpendicular to both B and v find the probability that
the electron emerges from the field with its spin pointing in the same direction
as when it entered.
Solution
If we take the magnetic field pointing in the z-direction and the velocity in the
¡/-direction then we can take the spin pointing in the ^-direction. In this case
the problem is like problem 15.13 above and the probability of obtaining the
spin to point in the same direction after a time t as it was pointing at time t = 0
is given by
cos 2 uit
where
¡XB
w =
x '
In this case the time t is just the time that the electron spends in the magnetic
field and is given by
t = L/v .
Therefore, the probability that the electron's spin points in the same direction
after emerging from the magnetic field is
P = cos2(uiL/v) . (15.14.174)
15.15 SHO: Sudden Transition

A one-dimensional simple harmonic oscillator with spring constant k is in its
ground state. At t = 0 the spring constant is suddenly changed to 4k. Calculate
the probability that the system will end up in the new ground state.
Solution
Before the spring constant changes the state of the system is described by the
wavefunction
(15.15.175)
where
U) = Y/K/R (15.15.176)
At t = 0 the sping constant increases to 4k. Therefore, u> changes to 2u and

the new ground state wavefunction is
M X ) eXP
={-^) H
x
'
(15.15.177)
For t > 0 the wavefunction evolves according to the new Hamiltonian and is
given by
ü n
$(x,t) = Y (15.15.178)
where (f>n is the wavefunction for the nth excited state corresponding to the
energy (2n + 1 )hu¡ of the new Hamiltonian. At t = 0 this wavefunction must
coincide with ip. So,
(15.15.179)
The probability that the system winds up in the new ground state is now given
by
r
2
m =
a/2 mu> 3 raw
exp dx
7T h 2h '
J —c
Y/2MUI 2IVH
7t h 3 muí
^ = 0.9428 (15.15.180)
15.16 C o u l o m b Excitation
A particle with charge Q and initial momentum
P -- Mvéz
is incident from infinity on an atom with Z electrons and energy levels En such
that the impact parameter is b.
15.16. COULOMB EXCITATION 339
a) Use first order perturbation theory to show t h a t the probability amplitude

for an excitation from the level
|¿) = |n, I, m) to the level |/) = |n', m')
is given by
iOe [e^ •
A}i =
~~hvbJ (xij sin9 + zíj cos 6) d9 (15.16.181)
where
hiofi — En< En
and the following approximations have to be made.

1) The incident particle follows a classical trajectory R(t) t h a t lies entirely in
the x — z plane and is determined by the classical equation
d9 bv
(15161821
s = m -
Here we have used the fact t h a t
R(t) = R(t)(sin 0(t), 0, cos 0(i)) . (15.16.183)
2) The dimensions of the target atom are much smaller than R(t) = |R(t)\ •
3) A dipole approximation is used.
For details on the validity of these approximations and further discussions see
[15.4],
Solution
Let Ho be the unperturbed Hamiltonian so t h a t
H0\n, I, m) = En\n, I, m) . (15.16.184)
The perturbation is given by

z
~n
V e
V = Y „ — (15.16.185)
where r¡ are the coordinates of the electrons in the target atom. Since the
dimensions of the target atom are much smaller than R(t) we can write
1 oo I
r'
f 0 w ) ' (15.16.186)
| R(t)
Since we are using a dipole approximation we may truncate this series after
1 = 1. However, for 1 = 0 the perturbation is just a constant and since the
initial and final states are different and therefore orthogonal we see that the
I = 0 term does not contribute. Thus, the perturbation reduces to
" -
Z _Q
= sin
DU\-> + Z> c o s
(01 • (15.16.187)
Here we have used the fact that the trajectory of the incoming particle lies in the
x — z plane so that $,-(<) is the instantaneous angle between the z-axis and R(t),
the vector from electron i to the instantaneous position of the incoming particle.
Again, since |r¿| < < R(t) we see that all angles #¿ as well as coordinates x¡, Zi
are the same and we may as well call them 9 and x, z. Therefore,
—OZe
V= 9 [xsin0(¿) + zcosfl(<)] . (15.16.188)
R(ty
Thus, using first order perturbation theory we have t h a t the transition ampli-
tude Aif from |¿) to | / ) is given by
1 r°° •
Afi « jr e-^(i|V| f)dt
in
J-OO
_ Q f etwfit _|_ z C o s 0 ( i ) | i ) dt .(15.16.189)

^ J— oo
But, the classical trajectory is given by (15.16.182) so t h a t

R2(t)d9
dt = — — . (15.16.190)
Therefore, we can write
,0(6)
. iQZe 1 °\°)
A =
" — V V /
etu>fit (f\xsin9
+ zcos0\i) d9 . (15.16.191)
•7r
Here, 9(b) is the asymptotic value of 9(t) for t —>• oo .
Bibliography
entific Publishing Co. Pte. Ltd. (2002), section 15.9.1.
[15.2] The computation of transition probabilities for atomic systems is carried

out in
E.U. Condon and G.H. Shortley, The Theory of Atomic Spectra - Cam-
bridge University Press (1963).
entific Publishing Co. Pte. Ltd. (2002), section 15.11 .
341
BIBLIOGRAPHY
[15.4] N. Bohr, Kgl. Danske Videnskab Selskab, Mat.-fys. Medd, b f l 8 , No.8,

(1948).
C h a p t e r 16
P a r t i c l e in a U n i f o r m
M a g n e t i c Field
16.1 E s t i m a t e of Magnetic Energies

The strongest static magnetic fields currently achieved in laboratories are of the
order of 3 x 10 5 gauss. For fields of this strength estimate the magnitude of the
terms
2
6 -* -» 6 — -»
-—B • L and -—-A •A .
2mc 2mc
Solution
We want to estimate the two terms
e
* = Ú-csz ' =
For this purpose we take
B-L = hB and \A\ = ^\r\\B\ (16.1.2)
with r = |r| = typical atomic distance = 1 0 - 9 m.
Then we find
eh 4.8 x 1Q- 1 0 x 1.05 x 10~ 2 7

n_ 5
1
2mc 8 x 9.1 x 1 0 - 2 8 x 3 x 10 10
= 6.92 x 1 0 - 1 6 erg = 4.33 x 1 0 - 4 eV . (16.1.3)
eV 2 (4.8 x 1Q~ 10 ) 2 x (10~ 7 ) 2 5 2
2mc 2 8 x 9.1 x 1 0 - 2 8 x (3 x 10 1 0 ) 2
17
= 3.16 x 10~ erg = 2 x 10~ 5 eV . (16.1.4)
16.2. RADII OF LANDAU LEVELS 343
16.2 Radii of Landau Levels

Solve the eigenvalue problem for the Hamiltonian
»o = + 2/o
where x0 and y0 are given by
nx + Muy = Muyo (16.2.5)
7ry — Mux = —Mux o (16.2.6)
and 7rx and 7ry are given by
n = p+-A. (16.2.7)
c
Here, A is the vector potential for a uniform magnetic field and u = eB/Mc is
the Larmor frequency. Interpret the meaning of this result.
Hint: Use the symmetric gauge
A = y(-2/,x,0) •
Solution
We want to solve the eigenvalue equation
r
o f i — al fi (16.2.8)
where
r l = x l + yl . (16.2.9)
Now consider the "transverse" Hamiltonian
Ht = ¿ t é + «*) = ~MW[(x - xo)2 + (y-2/o ) 2 ] • (16.2.10)
When written out this reads
Ht = + Py) +-M(u/2)~(x2 + y2) +-uLz . (16.2.11)
On the other hand
xpy - ypx
= xny - yirx - ^Mu(x2 + y2)
= Mu[x(x - x0) + y(y - y0) - ^{x2 + y2)]

1
= *Mw[(x - x0)2 + ( y - y0)2 - x20 - y20]
2
1 „ 1
= -Ht--Mu(x2 + yl) . (16.2.12)
344 CHAPTER 16. PARTICLE IN A UNIFORM MAGNETIC FIELD
Therefore,
' (16.2.13)
Notice that, by definition, r% is a positive operator so t h a t the eigenvalues of

Ht are greater than the eigenvalues oiu>Lz. Now,
[l„ht] = 0 .
So we can diagonalize them simultaneously to get

ÜJ 1
Ht = (2n + l)h—+-hu (16.2.14)
since Lz = mh. Therefore,
2h h
r = + 1/>2 =
° ~ ~mZ^H ~ m +
^ " (16.2.15)
If we now introduce the "magnetic length"
.o HE FI
A 1 6 2 1 6
= 73 = MZ I ' )
and let 2n — m = I, we get
r2 = A 2 ( / + l ) . (16.2.17)
Furthermore, the eigenfunctions of r 2 are products of the eigenfunctions of Ht

and Lz. This may be interpreted to mean t h a t the various energy levels with
fixed z-component of angular momentum correspond to circular orbits with
definite radii. However, the orbits are not observable since the "centres of the
orbits" satisfy the uncertainty relation
A2
ax0ayo > y . (16.2.18)
16.3 Equation of Continuity

Show that in the presence of a time-dependent electromagnetic field (<f>, A) the
equation of continuity holds for a particle of charge —e if the charge density is
given by
p = -e '*$
and the current density is given by
J = - e ^ - [vfiV + ie/^c)^")^- *(V-ie/(&0^)**] .
Hint: Start with the time-dependent Schrodinger equation.

GAUGE INVARIANCE 345
golution
The proof is a straightforward use of the Schrodinger equation and its complex
conjugate
ih^ ~ [ V + ie/(hc)Á}2*-e<W
ot 2m
-ihd-^~ = -^-[V-ie/{hc)Á]2V*-e<t>V . (16.3.19)
ot ¿m
If we multiply the first of these equations by and the second by \f> and
subtract the two equations from each other we get
F) f 1£ —*
IH—(V*<A) = 1 + — [«'a-
dtx ' 2m 1 hcL

+ $ * V • (yfo) + VÁ • V<T + * V ( j f r * ) ] }
= — - V • {*P*[V + ie / (he) Á]*!/ — \t[V — ie/(/ic)¿4]^* 1 . (16.3.20)
2m I. J
After dividing both sides by ih we obtain the desired equation of continuity for
the probability current density.
^ = V-J. (16.3.21)
After multiplying by the electric charge — e we have the electric current and
charge densities given by
2im i J
p = (16.3.22)
16.4 Gauge Invariance

Show that the Schrodinger equation is form invariant under the time-dependent
gauge transformations.
Hint: The wavefunction acquires a phase under the gauge transformation.
Solution
We again start with the time-dependent Schrodinger equation in the presence
of an electromagnetic field (<¡>, A)
ifi
^ =¿(f-F)2,1,+«^ (16 4 23)
'-
Now define the gauge transformations
= exp T^-A (16.4.24)

he
and
A' A + VA
- <T>--D4- (16.4.25)
c dt
Then, inverting these definitions and substituting in the Schrodinger equation
ih
we find
i
1
exp | - ^ A | » ' P~ ~{A' — VA)j' exp ill A
2m he
i , 1<9A\
exp ( - j p - A ] \P' (16.4.26)
The left hand side now becomes

f—iq . \ qdA f —iq \ .
e x p ( _ a j , f i ^ - + - - e x p ^ a j * . (16.4.27)
Also, by straightforward differentiation, we have
p — — (A' — VA) exp ( —r—A ) = exp ill 1 p- -A'j . (16.4.28)

c' 'J ' \ he he
So that the right hand side reduces to
e x p , _!ía) j_ p— - A ' j + <7</>exp ( - Í 1 A \ V (16.4.29)

he J 2m c J v he
Combining all this, and cancelling the phase factor, we get
1 / . a -a 2
ih—-— -— {p - -Áj + q(j>ty' . (16.4.30)
dt 2m v c
Thus, the Schrodinger equation is form invariant under a gauge transformation.
16.5 Gauge Transformations and Observables

Under a gauge transformation we have t h a t for static electromagnetic fields
Á -+ÁA = Á+ VA
<t> —> <f>A = <j>

A i
• $ = e (i/hc)A
so t h a t the Schrodinger equation remains form invariant. We also require, how-

ever, that observables be gauge invariant in the sense that their matrix elements
remain invariant under gauge transformations. This means that if under a gauge
transformation an observable
O(v.A.A) -j- 0(v.AA.éA)

r. l 6 5. GAUGE TRANSFORMATIONS AND OBSERVABLES 347
\ve require that
= (*A,o(p,ÁA,A)*A).
The left side may be rewritten as
(tf,0(p,Í,<¿)<5) = (*A,oA(p,Á,<t>)<¡>A),
where
0 a ( p , a <T>) = e , '(9/' ic ) a 0 ( p , a , FT e - ' ( 9 / f t c ) a
is the unitary transform of the operator 0(p,A,<j)) t h a t maps O —> 0A. Show
that the resulting necessary and sufficient condition, namely
0A{p,Á,<p) = 0(p,ÁA,<j>A)
is satisfied if and only if
0{p, A, <j>) = 0(p + q/cA, <f>) .
See also [16.1] for further discussion.
Solution
Clearly the condition
0(p,A,<t>) = 0(p+Â,<j>) (16.5.31)
is sufficient since under the gauge transformation

A
Á^Á = Á + VA
A
<T> —> <T> = <F>
-» $ a = e*(9/ftc)a ^ (16.5.32)
we have that
( p + ^ A ) 1 > A = e t '(<'/' ic > A (p+ | y í ) 4 ' . (16.5.33)
Therefore,
0 ( p , Í A , <¡!>A)tf A = e i ^ ' h c 1 A 0 ( p + - A , <t>)V (16.5.34)
c
and hence, with
Ox(p,AA) = 0(p, Ax, 4>x) (16.5.35)
we have
(tf.OÍp, A 4 > ) $ a ) = ( ^ A , 0 ( p . A a . < 6 a ^ a ^ . (16.5.36)

ohamm 16. PARTICLE IN A UNIFORM MAGNETIC FIELD
To prove necessity we try to consider the canonical momentum as an ob-

servable. In this case we have
0(p,A,<f>)=p (16.5.37)
and we need
p$A = e ^ / ^ V l r . (16.5.38)
This condition cannot be satisfied. Thus, we need to replace p by p + q/cA .

This means that, in the presence of a magnetic field, the canonical momentum
is not an observable, but the mechanical m o m e n t u m
mv = p+ (q/c)A (16.5.39)
is.
16.6 Spin 1 / 2 Particle in Magnetic Field

A neutral spin 1/2 particle with a magnetic moment ¡I is placed in a constant
uniform magnetic field B. Compute how the average values of the three spin
components vary with time if
a) the initial spin is in the direction of the magnetic field.
b) The initial spin is normal to the magnetic field.
Solution
H = —¡1 • B = -/Z<7 • B . (16.6.40)
a) In this case we take
B~(0,0,B) , s=-a . (16.6.41)
The initial state is
i(0) = ( I ) • (16.6.42)
^0 ,
The Schrodinger equation may then be written
- —hua z i¡) (16.6.43)
where
w= ^ . (16.6.44)
A
16.7- SPIN 1/2 IN MAGNETIC tlULu: ^
T h e solution is
7p(t) = exp(iutaz)ip(0) = e,wt ( q ^ • (16.6.45)
T h e expectation values of the spin operators in this time-dependent s t a t e are
(sx) = 0
(sy) = 0
(s2) = J. (16.6.46)
b) In this case we t a k e
B = (B, 0 , 0 ) . (16.6.47)
T h e inital s t a t e is as in case a) above, b u t the Schródinger equation m a y now

be written
ih^- = —tiuicrxip • (16.6.48)

at
T h e solution is
1>(t) = exp(iu>tcrx)il>(0) = ( ^ . (16.6.49)
In this case the expectation values for the spin o p e r a t o r s are
(sr) = 0
. . H .
(sy) = -sm2ui
(sz) = — cos2wi . (16.6.50)
T h e average values of t h e spin clearly precess a b o u t the m a g n e t i c field.
16.7 Spin 1 / 2 in Magnetic Field:

Heisenberg Equations
A n e u t r a l spin 1 / 2 particle with a m a g n e t i c m o m e n t p is placed in a c o n s t a n t
u n i f o r m m a g n e t i c field B . Using t h e Heisenberg equations, c o m p u t e how the
average values of t h e three spin c o m p o n e n t s vary with t i m e if y
a) the initial spin is in the direction of the m a g n e t i c field.
b) T h e initial spin is n o r m a l t o t h e m a g n e t i c field. T h i s is the s a m e problem as
16.6, except t h a t we now c o m p u t e in t h e Heisenberg picture.
OüU i^riAr r&lí 1 o. fAHl'KJLE IN A UNIFORM MAGNETIC FIELD
Solution
The Hamiltonian for this problem is as before
H = —¡1 • B = —[/.a • B . (16.7.51)
a) In this case we again take
5 = (0,0,b) , s = - a . (16.7.52)
The state, for all times, is
M = (J (16.7.53)
The Heisenberg equations for the spin operators read
Íh< = (16.7.54)
^t ^
and when written out in detail become
dsx
dt - 2 UlSy
DSY
^ = -2 usx
dt
ds2
= 0 (16.7.55)
dt
where, as before,
fiB
w = (16.7.56)
x '
The solutions are
sx(t) = s r (0) cos 2w< + sjn 2ojt

ZUJ
s
y (') sy (0) cos 2 + 'y W sin 2u>t
1U>
Sz(t) = S*(0) (16.7.57)
Here, s x (0), s y (0), and s 2 (0) are operators such that
<M0)> = o
(s y (0)> = 0
h
< M 0 ) > (16.7.58)
2
The two operators ¿ r ( 0 ) and sy(0) are determined by the initial conditions that
may be read off the Heisenberg equations.
sx(0) = 2ws y (0)
s =
y(0) —2ws j; (0) . (16.7.59)
16.7. spin 1/2 in magnetic field: hmijuivunnu zu¿uj\±iviyj ool
Thus,
(¿,(0)) = 2w(s y (0)) = 0
<¿y(0)> = —2w(s r (0)) = 0 . (16.7.60)
Therefore, we obtain for the time-dependent expectation values
(s x (t)) = 0
0
<«»(<)> =
(Sz(t)) = ^ • (16.7.61)
b) In this case we take

B = {B, 0,0) . (16.7.62)
The Heisenberg equations now read
dsx
dt = 0
DSY
— =
2 2 OJS
dt
dsz
- —2 u s „ (16.7.63)
dt
The solutions are
sx(t) = sx( 0)
sy(t) = SY (0) cos 2uit + -7^—^- sin 2ut

2cJ
s z (t) = sz (0) cos 2ut + sin 2u>t (16.7.64)
ZUJ
where, as above, s x ( 0 ) , Sy(0), and sz (0) are operators such t h a t
<«.(0)> = 0
( S y (0)) = 0
(s2(0)) = \ . (16.7.65)
Also, this time we find, by the same method as in part a), t h a t

(¿y(0)) = 2w(s z (0)) = hu
(¿,(0)) = —2w(sy(0)) = 0 . (16.7.66)
Therefore, the expectation values for the time-dependent spin operators are
(sx) = 0
(sy) = — sin2wi
/ \ H
\sz) - — cos 2uit . (16.7.67)
These are, of course, the same results as obtained by using the Schrodinger
equation.
16.8 Separation of Spin and Space for

Spin 1 / 2
Show that for a spin 1/2 particle in a uniform, but time-dependent, magnetic
field the wavefunction separates into the product of a spin and space function.
Solution
The Hamiltonian is
H = H0 - na • B = ( p ~ c / r - 4 ) +ev . (16.8.68)
2 777
We attempt to find a solution of the form
= ip(t,r) ° l f j ^ = ip(t,r)\(t) . (16.8.69)
If we substitute this into the time-dependent Schródinger equation (which in

this instance is also called the Pauli equation) we find that
íh
~^x + íh
^^t =
(h°^ x
~ • (16.8.70)
Thus, if we have that
dib
i h = H0ip (16.8.71)
and
ih^- = ~na • B\ (16.8.72)

at
we have a solution. So, we have shown that the equation separates.
16.9 Spin 1 / 2 in Time-dependent Magnetic Field

A spin 1/2 particle (magnetic moment fi) is placed in a uniform, but time-
dependent, magnetic field B(t) that points along the ¿-axis. The wavefunction
at time t = 0 is given by
, ( cosae-'^ \
^ u n a e ' - » ) •
a) Calculate the expectation values of the x and y-components of the spin.
b) Find the direction in space such that at any time t the 2-component of the
spin has a well-defined value along this direction.
16.9. SPIN 1 /2 IN TIM K-DEPENDENT MAGNETIC FIELD 353
Solution
If we write
= ( b(t) (16.9.73)
the Pauli equation becomes

da
ih—
dt
db
ih— = +fj.B(t)b . (16.9.74)
dt
If we write
hio(t) = pB(t)
the solutions may be written
a{t) cio exp [ u{t')dt'

Jo
b(t) = 6oexp iJ u(t')dt' (16.9.75)
After imposing the initial condition we find that

-•73
ao =
¿0 (16.9.76)
bo -
This shows that except for acquiring a phase, the two z-components of the spin
remain the same. In other words, the probability of measuring a z-component
h/2 remains at cos 2 a and the probability of measuring a z-component —h/2
remains at sin a .
a) The expectation values of the x and y-components are
(«x) ag&o exp -2i f ui(t')dt' + ab" exp 2i í ui(t')dt'

Jo J Jo
— cos a sin a 2 cos -2i Lj(t')dt' + P
h .
sin(2a) cos HJ uj(t') dt' + ¡3 (16.9.77)
2
Similarly,
(,sy) = —i— ^ a^bo exp - 2 i f u>(t')dt' — ab*exp 12ii £ uj(t')dt' j

Jo
— cos a sin q2 sin 12i f Oj(t')dt' -P
Jo
3Ín(2a) sin \î j ui(t') dt' - P (16.9.78)
b) In order t h a t the ñ-component of the spin have a well-defined value h/2

we need that ip(t) should be the eigenfunction of ñ • s corresponding to the
eigenvalue h/2. Writing this out for
ñ = (sin 9 cos ip, sin 9 sin <p, cos tp)
we get
cos 0 a(t) + sin 6 e~ttf b(t) = a(t) . (16.9.79)
Hence,
— = tan(#/2) e = tan a e 2l/3 exp -2i [ u(t') dt' (16.9.80)

a Jo
Therefore, a solution is
9 = 2a
<P P- [ Lj(t')dt' (16.9.81)

Jo
16.10 Spin 1 / 2 in Rotating Magnetic Field

A spin 1/2 particle (magnetic moment fi) is placed in a magnetic field
B(t) = B(sin 0 cos wf, s i n ^ s i n w i , cos0)

rotating with frequency w about an axis making an angle 9 with the z-axis. If
at t = 0 the component of the spin parallel to B(t) has the value h/2 find the
probability that at some later time t the component of the spin parallel to B(t)
has the value —h/2 .
Solution
The Schrodinger equation for the wavefunction
(16.10.82)
is
-ITOT
ih- d_ f a(t) \ =_ —^B cos a sin o e a(t)
(16.10.83)
dt b(t) sin f e icut COS ( b(t)
If we write
hq — [iB cos 9 , hp = nB sin 9

the Schrodinger equation becomes
da(t)
= iqa(t) + ipe b(t)
dt
db(t) iut
IPE a(t) - iqb{t) (16.10.84)
dt
16.10. SPIN 1/2 IN ROTATING MAGNETIC FIELD 355
If we differentiate the first equation we obtain

d2a . da db
iojt —iuib + — (16.10.85)
DP = , ,
I I +
" " dt
After substituting for ip(db/dt) and ipb from (16.10.84) we get
d2a . da / i i \ r\
+ lüJ (16.10.86)
dp ~dt + ^P~ q
~ + Uq a
^ =
'
If we now call
fii = + \ / w 2 / 4 + p2 + q2 + u>q
fi2 = ~ \Ju2¡\ + p2 + q2 + uq (16.10.87)
then the solution for a(t) may be written
a(t) = aiemit +a2eiU3t . (16.10.88)

From the first of (16.10.84) we get
da
6(f) = — — eiut (16.10.89)
P dt~lqa
Hence, we find that
bit) = ~ qa1 e'(ni+w)t + —LZ! a2 e i(Cl2+w)t (16.10.90)

P P
The initial conditions require that
a(0)
M =
6(0)
a
l + a
2
(16.10.91)
should be the eigenfunction corresponding to the eigenvalue h/2 of the operator
, • a „ - h f cos 6 sin 0
v(sin 9, 0, cos 0) • s = - . . „ (16.10.92)
' ' 2 V sin 0 — cos 9
After a short computation we get that, up to normalization
0,2 - q~
ai = A tan 0/2 +
P
Sli-q
a2 = A tan 9/2 — (16.10.93)
p
The normalization yields t h a t
|2 p2 cos 2 0 / 2
km 2 = (16.10.94)
(n2-«i)2
ooo itlAflUH 16. f ARTICLE IN A UNIFORM MAGNETIC FIELD
We also need the state </>(t) t h a t at time t is the eigenstate of
(sinOcosu!t,sm0sinu>t,cos9)s= — i . cos
. t) ,, sin 9 e~-iuit \
v ;
2 V sin 6 e J (16.10.95)
— cos t
corresponding to the eigenvalue -h/2. Again, a short computation yields
sin( 2
V (16.10.96)
W) - y -cos9/2eiut
Then, the probability t h a t at time t the particle will be found to have the spin
value -h/2 aligned along the magnetic field is given by
p(t) = !(</>(<),v(*))|2 .
Thus,
p cos 9/2 ft2 - q
P = sin 9/2 — tan 9/2 +
2
(ft2-fti)
CLI-Q
+ sin 9/2 tan 9/2 —
fti - Q ft2 - q
— cos 9/2 p — tan 9/2 + -iflit
P
^2 -q fti - q jn2t
— cos 9/2 tan 9/2 —
p
ill - q
2
sin 9/2 — cos 9/2
(ft2-«i)
\
sin 9/2 — ——-cos9/2J [1 — cos(ili — Q2)<] (16.10.97)
Next, using the definitions of fij and f22 and writing

c o s ( 1 - Q 2 )f = 1 — 2cos 2 (fii - fi2)i/2
we can simplify the result to read
2p 2
P = sin 0/2 i cos 9/2) x
w 2 + 4(p 2 + q2 + uq) P
^2 - q
sin 0/2 — - c o s 9 / 2 j cos 2 (^JUJ2/A + p2 + q2 + uiq .
(16.10.98)
Here,
q= cost/ , p = w¿sin( (16.10.99)

and
pB
w¿ = (16.10.100)
is the Larmor frequency.

BIBLIOGRAPHY
Bibliography
[16.1] D.H. Kobe, Am. J. Phys. 54, 77, (1986).
C h a p t e r 17
Angular M o m e n t u m , Etc.
17.1 Operator to Lower Total J

Show that the operator
T = A\j2J\- — j\ J2-]
where A is a normalization constant, has the property that
T\j,j,ji,h) = \j ~ 1, j ~ l j u h )
for j = ji + j 2 -
Solution
We consider the "unnormalized" operator
T={hJi-hh)/h. (17.1.1)
Then, we find that
T\ji + j2, h + Í2, j\,h) = T\ñ,h)\h,h)

= Í 2 y f i E \ j \ , j \ - 1) - 1> • (17.1.2)
Now we have t h a t
•Z2 = J\ + J2 + 1JizJ2z + J1+J2- + J1-J2+ • (17.1.3)
Also we see that
( J 2 + J2 + 1J\Z J2z)T\j\ + J 2 , ji + 32,j\ih)

2
= [ji ( j i + 1) + J2U2 + l)]h T\ji + J2, ji + 32) ji, Í2>
2
+ j'¿ \ / 2 i i 2 ( j i — l ) í 2 ^ | j i , í i — 1) ií2, j2)
- J i \ / 2 j l 2 j i ( j 2 - l ) f t 2 | i i , i i ) | í 2 , j2 - 1) • (17.1.4)
17.2. ENERGY SHIFT DUE TO A MAGNETIC FIELD 359
Furthermore,
(JI+J2- + JI-J2+)T\ji + 32, ji + Í 2 , h ^ h )

2
= 2JD2V*HT> \H,H)\J2J2-L)
2
-jij2\fíj~ih \h,ji - 1>|Í2,Í2> . (17.1.5)
So adding all these results we find that
J2T\j\ + Í 2 , j l + J 2 ) JI1J2)
= [ji (¿1 + 1) + h (¿2 + 1) + 2JU2
— 2(ji + +Í2, jl + J2, Jl J J2)
= (¿1 + 32 - l ) ( i i + j 2 ) / j 2 r | j ' i + 32, ji + 32,ji,32) • (17.1.6)
This shows t h a t T\ji + 32, j i + j 2, ji,j 2) is proportional to a linear combination

of |ji + 32 — 1, Til, ji, 32) for various m. The last step is to show t h a t m is
restricted to only one value, namely j 1 + j2 — 1. But this follows immediately
from applying J\z + J2z to
T\h + 32, ji + 32, j i , 32) •
17.2 Energy Shift D u e to a Magnetic Field

Evaluate the expectation value ( S z ) for the states | / ± 1/2, m, I) . These expecta-
tion values are used to obtain the shift in energy due to a uniform magnetostatic
field.
Solution
The state \l + 1/2, m, I) is obtained by applying
/ (/ + 1/2 + m)! t—ff+l/2 —m) Al+l/2-m)

V (2/+l)!(/+l/2 -m)!
to the state |/, i ) | l / 2 , 1 / 2 ) . Here
J_=L_+S_. (17.2.7)
So we find
/C \ _ ( / + 1/2 + m)! (2l+l-2m)
[ z)
(2/ + 1 ) ! ( / + 1 / 2 -m)\
(I + 1/2 + m)! t-(2l+l-2m)

(2/ + l)!(Z + l / 2 - m)\
± i < ¿ + + S + Y - ^ S A L - + S - Y - ^ I ) \ \ , |>
360 CHAPTER 17. ANGULAR MOMENTUM, ETC
(/ + l / 2 + m ) ! fc —(2/+1 —2ro1
(2/ + 1)!(/ + 1/2 — m)!
-(I - m + l / 2 ) y ( / , Z | l ' - ( m + 1 / 2 ^ ' _ " ( m + 1 / 2 ) | / , / > (17.2.8)
where we have used the fact t h a t more than one application of S_ to the state
j 1 / 2 , 1 / 2 ) annihilates it. Thus, what we have to evaluate are quantities like
(/, / | ¿ ' + - ( m - 1 / 2 ) L ( _ - ( m - l / 2 ) |Z) ^ a n d ^ Z |i/-(m+l/2)¿l-(m+l/2)
11,1).
Now, using that
(17.2.9)
we imediately get that

/-(m-l/2) r / - ( m - l / 2 ) |? A _ (2/)!(/ in -+ 1 / 2 ) ! ^ 2 i - 2 m + l
( U \ L i'» 0 = (17.2.10)
(I + m — 1/2)!
and
(m+l/2)Ll-(m+l/2)^ ^ _ (2/)!(/ - m - 1/2)! h2l-2m-1
(z,/|i'+- (Z + m + 1/2)!
(17.2.11)
Hence we get t h a t
(/ + 1/2) 2 -m2 + 1
( 5 2 ) = /i (17.2.12)
( 2 / + 1 ) ( 2 / + 1 - m) '
For the states |/ — 1 / 2 , m, I) we proceed in exactly the same manner starting
with the equation
\l- 1/2,1- 1/2)
1
= \l,l - l ) | l / 2 , 1 / 2 ) - y Z | T | Z , / ) | l / 2 , - 1 / 2 ) . (17.2.13)
V2T+
Then,
(/ + m + 1/2)! 1
(5.) = (1,1- 1|(1/2,1/2|
(2/ + l ) ! ( / - m + l / 2 ) ! 2Z + 1
21
J jl-m+1/2
'1(1/2, — 1/21 +
21+ 1
X 5,j'_m+1/2 v^m-1)11/2,1/2)
2 ^ - l í , 011/2,-1/2) (17.2.14)
COUPLING OF SPIN 1 TO SPIN 1/2 361
This simplifies to
h [ l m + : ) + m ( m _ 2 ) + ? / 4 ] ( 1 7
(Sz> = ~ ' '2' 1 5 )
17.3 Coupling of Spin 1 to Spin 1/2

A particle of angular momentum 1/2 is coupled to a particle of angular momen-
tum 1. List the states t h a t are eigenstates of
J" — (JI + JI)~
and
Jz = J lz + J2z
and express them in terms of the eigenstates of
(J^Ju) and (^2' -hz) •
Solution
The possible eigenstates of J2 and JZ are
|3/2,3/2) , |3/2,l/2) , ¡3/2,-1/2) , |3/2,-3/2)
and
|l/2,l/2> , |l/2,-l/2).
These are obtained as follows. We begin with
¡3/2,3/2) = |1, l ) | l / 2 , l / 2 ) . (17.3.16)
Now apply J_ to both sides to get
y/Z\3/2,1/2) = V 2 | l , 0)11/2,1/2) + |1, l ) | l / 2 , - l / 2 ) (17.3.17)

so that
13/2,1/2) = ^ j | l , 0 ) | l / 2 , 1 / 2 ) + - 1 | 1 , l ) | l / 2 , - l / 2 ) . (17.3.18)
Now apply J_ to both sides again to get
2|3/2, —1/2) = -||1,-1)11/2,1/2)+ ^ | l , 0 ) | l / 2 , - l / 2 )
+ ^|l,0)|l/2,-l/2) (17.3.19)
so that
v2,
| 3 / 2 , - l / 2 ) = ^=11, - l ) | l / 2 , l / 2 ) + ^ | l , 0 ) | l / 2 , - l / 2 ) . (17.3.20)
362 CHAPTER 17. ANGULAR MOMENTUM, ETC.
Finally,
|3/2,-3/2)= | l , - l ) | l / 2 , - l / 2 ) . (17.3.21)
The remaining two states must be linear combinations of
|l,0)|l/2,l/2) , |l,l)|l/2,-l/2)
and
|l,-l)|l/2,l/2) , |l,0)|l/2,-l/2)
and must be orthogonal to all the states above. Thus we find (up to an arbitrary
phase factor)
11/2,1/2) = -^|l,0)|l/2,l/2)~ ^|l,l)|l/2,-l/2) (17.3.22)
11/2,-1/2) = -^11,-1)11/2,1/2)+ -L|l,0)|l/2,-l/2). (17.3.23)
Another way to proceed for these two states would be to use the operator T of
problem 17.1. Then we find that, up to normalization,
11/2,1/2) = A[S-- 1/2 L _ ] | l , l ) | l / 2 , l / 2 )
= A[\l, l > | l / 2 , - l / 2 ) - ^ | l , 0 ) | l / 2 , 1 / 2 ) ] . (17.3.24)
Normalization now yields, up to an arbitrary phase factor,

A = - 0 7 3 - (17.3.25)
Thus,
11/2,1/2) = - ^ | l , 0 ) | l / 2 , 1 / 2 ) - ^ E | l , 1)11/2, - 1 / 2 ) (17.3.26)
which agrees with our previous result. The state 11/2,—1/2) is obtained by
applying J_ to this result.
17.4 Example of Wigner-Eckart

Consider a set of three operators Tm m — 1, 0, —1 such that
T*
X
—T
M —
[J±, Tm] = >/2 - m ( m ± 1 )hTm
[Jz, Tm] =- mhTm

where J are the total angular momentum operators. Evaluate the total m' , m"
dependence of the matrix elements ( j , m'\Tm |j, m").
Hint: Express Tm in terms of 3 x 3 matrices. This is an example of the Wigner-
Eckart Theorem.
17.4. EXAMPLE OF WIGNER-ECKART 363
Solution
We begin by taking
( a d e \
f b g . (17.4.27)
h k c )
Then using
[J*,T 0 ] = 0 (17.4.28)
we find
( a 0 0
To = 0 6 0 | . (17.4.29)
\ 0 0 c
Next we use
[J+,T0} = y/2hTx (17.4.30)
and find
/ 0 6—a 0 \
Ti = 0 b c-b . (17.4.31)
\ 0 0 0 /
Similarly we find using
[J-, To] = V^hT-i (17.4.32)
that
/ 0 0 0 \
T_i = j a — b b 0 J (17.4.33)
\ 0 b— c 0 J
The condition
T l = (—l)mT_m (17.4.34)
simply shows that the three numbers a,b,c are all real. If we now use
[J_,T_i] = [J+,Ti] = 0 (17.4.35)
we find the condition that
a — 26 + c = 0 . (17.4.36)
Thus, we get
6= . (17.4.37)
Therefore,
„ c-ol ,
T\ - —-ZZT—J+
2Y/2 H
m c—a 1 T c+ a
T o = J Z
2 ~ H ~ ~ 2 ~
711 =
* i17"4'38)
17.5 Rotations for Spin 1 / 2 and Spin 1

Consider the unitary operator
RNITP) =
where J is the angular momentum operator.

a) If j = 1/2 expand Rn(<p) in a Taylor series to obtain a simpler expression
and apply it to the states
and
o) (l
b) If j = 1 repeat part a) but consider the states
1\ / o x / 0
0 , 1 and 0
0 / V 0 / V 1
Solution
a) We start with
ñ = (sin a cos /?, sin a sin/3, cos a ) . (17.5.39)
Then for j = 1/2
^ or2) (17.5.40)
and
ñ •J _ 1 / cosa sinae-1^
(17.5.41)
h 2 \ sin a e8'3 cos a
2
/ -\
I ñ • J\ l / l 0 \ 1
L (17 5 42)
n o i r ? ' '
Thus,
2n
¿ 1 (17.5.43)
17.5. rotations f o r spin 1/2 and spin 1 365
f ^ —A 2n + l
ñ •J \ 1 fi • J
= (17.514)
y~~h~J 22"+1 ~~h~ •
Hence we find by expanding the exponential that
Rn(f) = cos(<¿>/2)l + zsin(<p/2)^-^— . (17.5.45)
Applying this to the states indicated we find

1 \ ( cos(<p/2) + ¿sin(95/2) c o s a
ün(v) (17.5.46)
isin(^>/2) sin ae 1 / ?
i sin(<p/2) p-'/ 3
Rn(f) (17.5.47)
cos(<p/2) — ¿sin(y>/2) c o s a
b) Similarly we find, by explicitly writing everything out, that for j = 1 with

0 1 0
JX_
i 0 1
h V2 0 1 0
0 -i 0
J,y _ i U —i
V2 ü i Ü
1 0 0
¿L
0 0 0 (17 5.48)
h
0 0 - 1
we have
^ sin a e 0
ñ • J
4 ^ sin a e'1* 0 cos o ; 17.5.49)
0 _l ... j b
\ — C O S (V
and by multiplication that

3
h • J ñ •J
17.5.50)
Thus,
2n + 1 2n
n •J Tl • J ñ •J n • J
(17.5.51)
where
H•J
i ( l + cos a ) 2 i sin q cos a e -ip -2i(3

SÑ.
= I -4= sin a cos a e,/3 sin 2 a (17.5.52)
-~7r.
i sin a cos a e1'5 n/2 sin a cos a e
i sin 2 a e 2 , / J |(1 + cosa)2
71
Expanding the exponential in a power series we get

-\ 2
+1
R U ) = I - ^ 1 ) i | | f M Ü
n W + +
{ * ) * h ^ + i v - { « J ¿ w
V j V f ñ - f V
= i i
I + ~7j _ sin<?!,
+ I J COS(
^- (17.5.53)
Applying this to the three states we get
( 1 \ / | sin 2 a + ¿cos a s i n <j> + | ( 1 + c o s a ) cos <j>

0 I = I ^ | s i n a c o s a e , / 3 ( l - cos <j>) +
0 / \
sin a sin cfre'P
— | sin 2 a e 2 l | í (1 — cos^)
o> ^ sin a cos a e ,/3

(l — cos <f>) + sin a s i n (j) e
2 2
RnW I 1 cos a + sin a cos <
0; y ^ s i n a c o s a e I / 3 (l — cos <¡>) + sin a s i n (pe'^ j
/ 0 \ / — | sin 2 a e - 2 " ' ' (1 — cos <j>)

RN{<J>) i 0 j = i ^ sin a c o s a e~ , / 3 (l — cos <F>) + ^ sin a sin 0 e~ ,/3
\ 1 / \ | sin 2 a — i cos a sin <j> + 5 (1 + cos a ) cos <f>
(17.5.54)
17.6 Spin 1 / 2 Coupled to Spin 3 / 2

A particle of total angular momentum j i = 1/2 is coupled to another particle
with total angular momentum j'2 = 3/2. W h a t are the states of possible total j ?
Express all the states with the lowest possible j in terms of the states |ji, mi),
\h, m2).
Solution
The states of total possible angular momentum j are given by
j = 3/2 + 1/2 = 2 , 3/2-1/2=1.
Thus, we only have two possibilities j = 2 and j = 1. The possible states of

total j = 1 are
11,1) , |1,0) , |1,-1> .
To get the state |1,1) we use the result of problem 17.1 that, u p to a normal-
ization factor A we have
|1,1) = AT\2,2) = J \ JI/*-) |3/2,3/2)|l/2,l/2)

17.7. COUPLING OF SPIN 1 OR 0 AND SPIN 1/2 367
A
| ( | h-1) - | ( | - 1)13/2,1/2)11/2, l/2>
2
+ l) 1)|3/2
- Y V ~ ~ ' 3/2)|l/2, -1/2)
= | [\/3|3/2, l / 2 ) | l / 2 , 1 / 2 ) - 3 | 3 / 2 , 3 / 2 ) 1 1 / 2 , - 1 / 2 ) ] . (17.6.55)
Normalizing we find
|1,1) = i | 3 / 2 , 1 / 2 ) 1 1 / 2 , 1 / 2 ) - ^ | 3 / 2 , 3 / 2 ) 1 1 / 2 , - 1 / 2 ) . (17.6.56)
Then, again up to a normalization factor A,
¡1,0) = a7_|l,l)
= A[\3/2,-1/2)11/2,1/2) -|3/2,l/2)|l/2,-l/2)] (17.6.57)
so that
¡1,0) = ^ = 0 3 / 2 , - l / 2 ) | l / 2 , 1 / 2 ) - |3/2,1/2)11/2, - 1 / 2 ) ] . (17.6.58)
Finally,
|1,-1) = AJ.\L,0)
= ^ [ > / 3 | 3 / 2 , - 3 / 2 ) | l / 2 , 1 / 2 ) - |3/2, - l / 2 ) | l / 2 , - 1 / 2 ) ] .
(17.6.59)
So
|1- - 1 ) = ^ 1 3 / 2 , - 3 / 2 ) | l / 2 , 1 / 2 ) - | | 3 / 2 , - l / 2 ) [ l / 2 , - 1 / 2 ) . (17.6.60)
17.7 Coupling of Spin 1 or 0 and Spin 1 / 2

An electron (spin = 1/2) is in a state of either I = 0 or I = 1. Express all states
of total angular momentum |j, mj) in terms of the states |/, m ; ) | l / 2 , ms) where
I = 0 or I = 1.
Solution
Starting with the state of highest weight we have
|3/2,3/2)= |l,l)|l/2,l/2) . (17.7.61)
Applying
J- — L- + S- (17.7.62)
to this equation we get

V 3 | 3 / 2 , 1 / 2 ) = V 2 | l , 0)11/2,1/2) + |1, l ) | l / 2 , - 1 / 2 ) . (17.7.63)
Thus,
| 3 / 2 , 1 / 2 ) = y/2/S\l, 0)11/2,1/2) + 1/V3|1, l ) | l / 2 , - 1 / 2 ) . (17.7.64)
Applying J_ once more to this equation we find
|3/2,-l/2> = 1/^311,-1)11/2,1/2)4-^27311,0)11/2,-1/2) . (17.7.65)
The state of lowest weight is simply
¡3/2, - 3 / 2 ) = |1, - 1 ) 1 1 / 2 , - 1 / 2 ) . (17.7.66)
If / = 0 we only have two states of total angular m o m e n t u m 1/2 and these are
11/2,1/2) = |0,0)|l/2,l/2)
11/2,-1/2) = |0,0)|l/2,-l/2) . (17.7.67)
17.8 Identity for Constant Magnetic Field

Show that if
A = —- r x B
2
where B is a constant vector, then
p-Á+Á-p=B-L .
Solution
We are given that
A = ~ r x B (17.8.68)
where B is a constant vector. Then, using the Lorentz condition

v-i"=0 (17.8.69)
we see that
p Á = Á p . (17.8.70)
Now, using this we can write
pÁ+Á-p = — - p - ( f x B ) — -(rxB)-p
= — (r x B) • p
= (B x r) • p . (17.8.71)
Next, using the fact t h a t the dot and cross product may be interchanged as long
as the order of the vectors is maintained we get
p Á+Á-p= B (rxp) = B L . (17.8.72)
j7.9. THE STATE \n,j,m,l) 369
17.9 T h e State | n , j , m,/)
Use induction to show that
Solution
I n,j,m,l)
We want = J w +tom show
to use induction
)!
that (17-9-73)
( 2 j ) \ ( j - m)!
We start with m = j. In this case the (17.9.73) is clearly correct. For m = j - 1
we have
J-\n,j,j,l) = \ / j { j + 1) - j ( j - l)H\n,j,j - 1,1) . (17.9.74)
So solving for \n,j,j — 1,1) we again see t h a t (17.9.73) holds. Now assume t h a t
(17.9.73) holds for m ~ j — M. Then, using
J-\n,j,j - M,l)
= VIU + 1) - U - M)(j - M - l)h\n,j,j — M — 1,1) (17.9.75)
so that
In,j,j- M -1,1)
1
h lJ^\n,j,j — M,l)
VjU + 1) - a - M)(j - M - 1)
l(2j - M)\ 1
H-W+VJ"+1\N,J,J,L)
(2j)!M! sj(2j — M)(M + 1)
(17 97<i
= - »
as required.
17.10 Landé g-factor

Show that if A is a vector operator such that
[</;ri-Ax] — 0 , x i Ay\ — ihAz , — ihAy

and cyclic permutations. Then,
a)
[J2,[J2,A]] = 2h2(J2A + ÁJ2) - 4fr2(Á J)f.

b) Use this result to show t h a t
and
c) hence evaluate the m a t r i x element for the weak-field Zeeman effect
AE = {Mh + (nLSJM\SZ \nLSJM)} = nBBMg

h
where
, J(J + 1) + S(S + 1) - L(L + 1)
g +
~ 2 J ( J + 1)
is the Landé g-factor.
Solution
T h e first p a r t is simply a tedious c o m p u t a t i o n . We therefore first c o m p u t e
[J2, [J2,AX\[ . T h e n , by s y m m e t r y , t h e result also holds for Ay and Az. As a
first step we c o m p u t e
[ J 2 , Ax] = J • [J, Ax] + [Ax, J ] • J

= ih[-JyAz+JzAy-AzJy + AyJz]. (17.10.77)
Similarly we find
\J2, Ay\ = ih[JxAz — JZAX AZJX — AXJZ]
= ih [— JXAy + JyAX — Ay Jx + Ax Jy] . (17.10.78)
Then,
[j 2 , [j 2 , AX]] = Íh[-jy[j2, + JZ [J2, Ay]

2 2
~ [J ,AZ]JY + [J ,AY]JZ]
= —h"[jyJxAy — JyAX + Jy Ay JX ~ Jy AX Jy
+ JyAy
Jy Jy Ax Jy + Ay JxJy AXJ~
+ JZJXAZ — JZAX + JZAZJX — JZAXJZ
+ JXAZJZ — JZAXJZ + AXJXJZ — Ax J~]
2 2 2 2 2
= h [(J + J )AX + AX(J + J )]
h [Jy(2AyJx + ihAz) — (AxJy — ihAz)J, y
Jy(Jy ""t~ ihAZ) -f- (AyJX + ihAZ)Jy + Ay(jyjx + ÍHJZ)
+ JZ(2AZJX — ihAy) — (AxJ)z + ihAy)Jz — JZ(JZAX — ihAy)
+ [AZJX - ihAy)Jz + AZ{JZJx - ihjy)] . (17.10.79)
Here we have used
[Jr,,4x] = 0 . (17.10.80)
17.10. LANDÉ g-FACTOR 371
Now collecting terms and using (17.10.80) again we find t h a t

[J\[J\AX]\ = h2 [2 ( J 2 + J?)AX + Ax{J2y + J2)
— 4(y4y Jy + AZ Jz)Jx)\
= fc2[2(J2 + J2 + J2z)Ax + Ax(J2x + J2 + J2)
— 4(AXJX + Ayjy + AZJZ)JX)]
= 2 / i ( J A x - M * . / 2 ) -4h2(Á-
2 2
J)JX . (17.10.81)
Thus, applying identical computations for Ay and Az we find the desired result
[J2,[J2,Á]] = 2h2(J2Á + ÁJ2)-4h2(ÁJ)J . (17.10.82)
b) We now choose A = 5 and then we immediately get
(JM\[J2,[J2,§\]\JM)
= 2h2(JM$J2S + SJ2)\JM) - 4h2(JM\(S • J)J\JM) . (17.10.83)
But the left-hand side of this equation vanishes. Thus, evaluating this for the
z-component we obtain
0 = 2h2(JM\(J2SZ + S Z J 2
) \ J M ) - 4h2(JM\(S •J ) J Z \JM) (17.10.84)
or
0 = 4h4J(J + ${JM\SZ\JM) — 4h3M(JM\S • J\JM) . (17.10.85)
Thus, we have the desired result
{JM\SZ \JM) = . (17.10.86)
c) Here we have to evaluate
(nLSJM\SZ |nLSJM) = Mh^nLSJ1^}f ' , (17.10.87)

J (J + 1 )n¿
But,
L2 = (J — S)2 = J2 + S 2
- 2 S J . (17.10.88)
Therefore,
\nLSJM\S • J\nLSJM)
Mh±
J{J+l)h2
= Mh3J(J + l) + S(S + l)-L(L + l) (17.10.89)
v
2J(J+l)h°-
and the desired result follows t h a t
AE = fiBBMg (17.10.90)
where the Landé g-factor is given by
• | • / ( j + l) + s ( s + l ) - l ( t + i) . 0 9 1 )
17.11 Spin and Space Coordinates

A spin 1/2 particle has its spin components aligned with a vector
ñ = (sin a cos /?, sin a sin /?, cos a) .
This means that the eigenvalue of s ñ is h/2. The angular momentum wavefunc-
tion of this particle is given by the wavefunction for total angular momentum j
derived in section 17.7 of [17.1], namely
yjl ± m + 1/2 Y,, m _ 1 / 2 (0, <p)
±y/l + m + \/2 Y,,m+i/2(0, <p)
Find the relationship between the angles (a, /?) and (6, tp).
Solution
The eigenfunction of s • ñ corresponding to the eigenvalue h/2 is given by
2 (
h cos a sin a e ,IS H
(17.11.92)
sin a e l P — cos a 2
From this we get that
cos aa + sin a e~1/3b = a . (17.11.93)
The second equation is an identity. Therefore, solving for a / 6 we get
a( 1 — 1 + 2sin 2 a/2) = 2 s i n a / 2 c o s a / 2 e 6 (17.11.94)
and
1/3
Y = cot a/2 e . (17.11.95)
B
For the given wavefunction we have
(17.11.96)
Comparing this with (17.11.95) we get
P = <P
cot a/2 (17.11.97)

r.
17.12.
17.12
CLEBSCH-GORDON
Clebsch-Gordon for j= 3/2

FOR j= 3 / 2
A spin 1 / 2 particle in a P - s t a t e (/ = 1) has a t o t a l angular m o m e n t u m j = 3 / 2

a n d z - c o m p o n e n t of the t o t a l angular m o m e n t u m rrij = 1 / 2 . W h a t is the
probability of finding the z - c o m p o n e n t of the spin of the particle to have t h e
373
value fflj = 1 / 2 ?
Solution
By s t a r t i n g f r o m
| 3 / 2 , 3 / 2 ) = |1, 1) 11/2, 1/2) (17.12.98)
and applying
J_=L_+S_ (17.12.99)
to b o t h sides of this equation, we find t h e angular m o m e n t u m wavefunction for

the particle
13/2, 1/2) = [ V 2 | l , 0)11/2, 1/2) + |1, 1)j 1/2, - 1 / 2 ) ] . (17.12.100)
Therefore, the probability t h a t t h e z - c o m p o n e n t of the spin is + 1 / 2 is

2
V2 2
v3
17.13 Rigid Rotator in a Step Potential

A rigid plane r o t a t o r , m o m e n t of inertia / , r o t a t e s subject to a p o t e n t i a l field
0 0 < < 7T
V(v) =
-VO IT < tp, 2n
where tp is the a z i m u t h a l angle of t h e r o t a t o r relative to s o m e fixed axis and Vo

is a positive constant. Find the transcendental equation for t h e positive energy
eigenvalues.
Solution
T h e H a m i l t o n i a n for this problem m a y be written
H = £ + V(<p) (17.13.101)
where we impose periodic b o u n d a r y conditions on the eigenfunctions, n a m e l y
V>(2rr) = <¿>(0) . (17.13.102)

If, using the fact that we restrict ourselves to positive energies, we write
2 IE 21(E + V0)
K2 =
2
(17.13.103)
h2 ' h2
the eigenvalue equation for the energy becomes
2
:d ip
J-~2 + k2ip = 0 0 < ip < TT
dip
d2i>
= 0 7t < if < 27t . (17.13.104)
dipr¡2¿
The solutions are
= Aeihv + Be-ikv 0 < <p < tt

v» = CE'^ + DE-'** 7t<<p<27t. (17.13.105)
Imposing the condition t h a t ip and its first derivative are continuous at tp = n
we get
Aeihn + Be~ikn = C eiKn + D e~iKn

k [A B d -tfetl] =— K [CeiK* - De~iKn] (17.13.106)
Finally, imposing the periodic boundary condition
IP( 27r) = ^(0) (17.13.107)
we get
A eik2n + B e~ik2w = A + B
id is2
Ce * + fle" * = C+D (17.13.108)
Writing all this out we find a set of four homogeneous equations in four un-
knowns. For a nontrivial solution to exist requires t h a t the determinant of
the coefficients must vanish. This is the desired transcendental equation. The
required determinant is
0iirk e-ink
ilrk
kr e - k :e e~ i n k —K E
D = i2nk (17.13.109)
-1] [e -i2nk - i ] 0 0
i2 -Í2TTK
0 o [e
0 Í2HK™ - 1] -i]
This determinant when evaluated yields the value
D = —4in sin 7t(k + k) . (17.13.110)
Therefore, the energy eigenvalues are given by the equation
K + k —n n = l , 2 , 3, . . . . (17.13.111)
Negative values of n are excluded since
K = n — k > 0 and k > 0 . (17.13.112)

T
17.14.
17.14
SPIN DEPENDENT OPERATORS
Spin D e p e n d e n t Operators for T w o

FOR TWO PARTICLES 375
Particles
If we have two particles with spin operators
h h
e - Q ~
O\ - 20"1 ' '"'2 =
20"2
separated by a distance r = fir show t h a t any positive integral power of either
of the o p e r a t o r s
A = 5\ • <72
or
Si2 = 3(<?i • ñ)(<?2 • n) - <?i • <?2

as well as any p r o d u c t of such powers can again be w r i t t e n as a linear combi-
nation of A, Si2 and the unit m a t r i x .
Solution
We show t h e result s t e p by step. Recall t h a t for the Pauli m a t r i c e s we have
<r2 = <r2 = <r2 = 1 (17.14.113)
and
crxay = i<rz , crycrz = iax , az(Tx = icry . (17.14.114)

We are now ready t o c o m p u t e
A2 = ai • <72<71 • <?2
= (a\x<^2x + ""1 y&2y + V\z&2z) (&IX&2X + &\y&2y + &1 z&2z)
=
3 — 2 ( a i x a 2 x + Vly&2y + (Tlz^z)
= 3 - 2 A. (17.14.115)
Next we c o m p u t e
S2U = 9(<?i • ñ) 2 (<? 2 • ñ ) 2 + A2 - 6(<?i • ¿ f 2)(ó : i • Ñ){<?2 • Ñ) . (17.14.116)
B u t , by using (17.14.113), we see t h a t
(<7i • ñ ) 2 = (<72 • ñ ) 2 = 1 . (17.14.117)
Also, by simply m u l t i p l y i n g out the m a t r i c e s we see t h a t
a(a • ñ) = ñ + i(a x ñ) . (17.14.118)
Therefore,
(<?i • <?2)(<?i • ñ)(<?2 • ñ) = [ ñ + i(a 1 x ñ)][ñ + i(a2 x ñ)]
= 1 — (<?i x ñ)(<x2 x ñ)
= l-ai + -ñ)(a2 ñ) . (17.14.119)
So, we find
Sf2 = 9 + A2 — 6 + 6A — 6(<?i • ñ)(a2 • ñ)

= 6 + 4A - 2(5i2 + A)
= 6 + 2 A - 25i2 • (17.14.120)
Finally,
AS\2 = —A2 + 3(<?i • <T2)(<Ti • ñ)(a2 • ñ)

— —3 + 2 a + 3 — 3 A + 3(íti • 72) ((72 • ñ)
= 5i2 • (17.14.121)
In exactly the same manner we find that
si2^4 = 5 i 2 . (17.14.122)
This proves the desired result.
Bibliography
C h a p t e r 18
Scattering - Time
Dependent
18.1 Cross-section from Experiment

A p r o t o n b e a m producing a current of 5 x 1 0 - 9 a m p s is incident on a target
of copper. Assume the target thickness is such t h a t the areal density is 0.2
m g / c m 2 . T h e detector has an area of 0.5 c m 2 , n o r m a l to the scattered b e a m ,
and is 20 cm f r o m the t a r g e t . If 10 p r o t o n s are counted by the detector every
second at a particular angle, calculate the differential cross-section for protons
scattering off copper at t h a t angle.
Solution
We use the equation
AN = J N 0 A O . (18.1.1)
ail
Now we have
9
I = 5 x 10 A = incident p r o t o n current
3.12 x 10 1 0 p a r t i c l e s / s (18.1.2)
1.6 x 10~ 1 9 C / p a r t i c l e
A = cross-sectional area of the b e a m (18.1.3)
J = I / A (18.1.4)
p = 0.2 m g / c m 2 = areal density

0.2mg/cm2 _
= 1.84 x 10 1 8 a t o m s / c m 2 (18.1.5)
1.09 x 10~ 1 9 m g / a t o m
378 CHAPTER 18. SCATTERING - TIME DEPENDENT
AN = 10protons/s (18.1.6)
AS = 0.5 cm 2 = area of the detector (18.1.7)
R = 20 cm = distance of the target from the detector (18.1.8)
AQ = AS/R2 = 1.25 x 1 0 - 3 . (18.1.9)
From our equation above we find
da 1 AN A I AN I AN
(18.1.10)
dfi J No AÍ2 / Ap Aft Ip Aft '
Therefore we have
= 1.34 x 10 - 25
cm2/atom/steradian . (18.1.11)
aft
18.2 Green's Functions for Free Particle States

Use the expressions
g t m = - l e - i h o t / n e,ts (18.2.12)
H
GQ (t) = +ê-iHot/h0(-t) (18.2.13)

FX
and evaluate the matrix elements {p\G^ (<)|£) where |p), |£) are free particle
states of momentum p and k respectively.
Solution
We begin with
_I ( MOT
G U t ) = T ^ e x p ( - i - J - ) 8{±t) . (18.2.14)
Now let |P) be an eigenstate of HQ.
H0\p)=^\P). (18.2.15)
Then,
(p\G±\k) = T faexp(-i?jf\\k)0(±t)
= exp ( - i ^ L ) S(p— k) 0(±t) . (18.2.16)

18.3. DISPERSION RELATIONS 379
18.3 Dispersion Relations

a) Let / ( z ) be defined by
f ( z ) = J dt eitz F(t)
where F(t) = 0 for t < 0. Show that / ( z ) is an analytic function for Si(z) > 0.
b) Consider a function f ( z ) which is analytic in the upper half of the complex
z-plane as well as on the real axis and vanishes rapidly for |z| —> oo. In that
case we may write
(18 317
= é ¡ f - '
where the contour runs along the real axis from — R to R and is closed by a
semicircle in the upper half-plane. Let R —>• oo and let z —> the real axis from
above and use this result to express 5f(/) in terms of 9 ( / ) and vice versa.
Hint: Recall that
lim —^— = P— ¿7rá(a;) . (18.3.18)
e-*0 + X ± It X
c) Verify the result obtained by applying it to the function
f ( z ) — —-— a > 0 . (18.3.19)

z + la
Solution
a) The proof that / ( z ) is an analytic function for 9 ( z ) > 0 is straightforward
since by assumption the integral
f ( z ) = J dteitz F(t) (18.3.20)
is assumed convergent. This convergence is only improved by letting

z = a; + iy , y > 0 .
In that case we may differentiate with respect to z under the integral sign and
get t h a t
^-(z) = i J t dt e'tz F(t) (18.3.21)
is finite for all 9 ( z ) > 0. Thus, / ( z ) is an analytic function for 9 ( z ) > 0.

b) Following the instructions we see t h a t for R oo the contribution to the
integral from the semicircle in (18.3.17) vanishes and by taking the limit as z —>
the real axis from above we get (writing z = x + ie)
f(x + ie) = -L r M , dz'

27VI J_00 z x ie
D Z L D Z
= y - P R 4 ^ ~ ' + ^ - R F ( * ' W - Z ) ' • (18.3.22)
2th J_00 z x 1-KI J_00
38U CHAPTER 18. SCATTERING - TIME DEPENDENT
p 1 d
= h jZ¥ * *' •
So carrying out the delta function integration and simplifying we find
Next, we take the real and imaginary dz'

parts of this equation and find
sfi/w] = i p r «
*" J-OO z
~ x
P
c¡[f(x)] = ~ l J • (18-3-24)
This is the required result. The pair of equations (18.3.24) are known as dis-
persion relations. These relations may be combined with a statement of causal-
ity, such as is assumed in part a), to yield important conditions for scattering
amplitudes. In optics the resulting equations are known as "Kramers-Kronig
relations".
c) If the given function is
/ (K« ) = — r - a > 0 K(18.3.25)

z + ia '
then we have that
x
B[/(s)]
x + a2
2
• (18.3.26)
To verify the dispersion relations for this example we simply write out the
integrals
i P R M 9I[ / I( * 'D) ] Z J .. = z- «I nP Rr dz'

n
J-oo z
' -
x
t J-c (z' + a 2 ) ( z ' — x)
= 3t[/(x)] (18.3.27)
x2 + a
as required. Similarly,
_ i P R M G I D Z , = i R R
z x
*" J-oo ' - * J-oo iz'2 + a 2 ))(z'-x)
(z' -
a
=9[/(i)] . (18.3.28)
x¿ + a
This is again the expected result.
18.4 Kállén-Yang-Feldman Equations

Assume that V is independent of time and use the equation
OO
/
G%(t-t')VG±{t')dtl (18.4.29)
•OO
± ±
to obtain an equation for the Fourier transform G (Ü;) of G ( I ) in terms of the
Fourier transforms of G^(<). Write a formal solution for G ± (a>).
18.5. BORN APPROXIMATION 3»i
Solution
We begin with the integral equation for G ± ( ¿ ) .
/ O O
G j f t i - f ' ) V <?*(<') <&'• (18.4.30)

-OO
Now define
/ O O
eiuiG^{t)dt (18.4.31)
-OO
/ O O
eiutG±(t)dt . (18.4.32)
-OO
Then,
/ O O r OO
dt / e i w ( t - t ' ) e i w f ' G Í ( í - í ' ) l / G ± ( < ' ) di'.(18.4.33)
- OO J — OO
So, changing t h e variable of integration f r o m t t o t—t' and since V is independent

of t, we have:
G ± (w) = G f ( w ) + G Í ( w ) K G ± ( w ) . (18.4.34)
Therefore solving formally we get
[ l - G j ( w ) K ] G ± ( a ; ) = Gf(w) (18.4.35)
and
G ± ( w ) = [1 - G f (w) K ] _ 1 G f (w) • (18.4.36)
18.5 Born Approximation

In t h e equation
Tfi = {*0(Ef)\V\*+[Ei)) for Ef = Et (18.5.37)
approximate by a free particle s t a t e . If V is a screened C o u l o m b

potential
e~^r
V = -V0
calculate the scattering a m p l i t u d e . T h e a p p r o x i m a t i o n used is known as the

first Born a p p r o x i m a t i o n .
Solution
The T - m a t r i x on shell is defined by
Tfi = (*°(E})\V\y+(Ei)) for E}=Ei. (18.5.38)
We are using the approximation that
(r\V+{Ei)) = (2tt)" 3 / 2 e i S r " (18.5.39)
where
TFK2 o /i 2 9 2
EJ —
Thus,
_i R E-V-R .r _
T i = , q r V o (18
' j2ñyiJ ~ ~ -541)
So,
Tfi = r rdre-»r f s[ndde

(2tt) 7O JO
VO{
v
°<m [ " r d r e -^-iiüm (18.5.42)
(2 tt) 3 y 0 pr
where
P = |*-«1 (18.5.43)
is the magnitude of the change in momentum that the particle experiences. So

finally,
Tfi = . (18.5.44)
271"" \k — q]2 + p.2
18.6 Scattering in C M and Laboratory Frame

Establish the connection between the centre of mass frame scattering amplitude
f(8,<p) and the laboratory frame scattering amplitude / l ( 0 l , ¥ ? l ) for particles
of mass mi and m 2 scattering off each other.
Solution
Comment: The first thing to realize is that the calculations involved here are
completely classical.
In the laboratory frame a particle of mass mi, velocity vx is incident along the
¿-axis on a particle of mass rn-¿ at rest so that its velocity is v? = 0. This gives
rise to a differential cross section dcr^^L, ipi,)/dS2L for the incident particle.
18.6. SCATTERING IN CM AND LABORATORY t namn,
The number of scattered particles t h a t are observed in a unit time in a cone

A f t i centered at 9¡ , ipL is then given by
ND°L(H, V L)ASJi
DÍ¿£
where we have assumed a single scattering centre and an incident flux of N

particles per unit area per unit time.
p - mv
p = mi(vi — V) = mvi
—p = 7772(0 — V) = —mv1
p = — mv
Center of Mass Frame
P1 TTL\ VY
Pi = mxvi 9L,<PL
PI = Pi Pi
Laboratory Frame
Figure 18.1: Scattering in the center of mass and laboratory frames.
The situation in the two cases is as depicted in figure 18.1. In the centre of
mass frame a "fictitious" particle with the reduced mass
m\ + m2 M
is incident on a fixed centre of force. The centre of mass moves with a constant
speed
m i
1/
V (18.6.46)
~ T7V1 •
M
We still need the relation between 61, tpL and 9, (p. This is obtained from the
relation between the velocity v[ of the scattered particle in the laboratory frame
where it has components
7^ = i>i(sin 9L cos ipL, sin 9 i s i n ^ L , cos 9 i ) (18.6.47)
and the centre of mass velocity V and the velocity v1 of the scattered particle
in the centre of mass frame. The speed v' of the scattered particle in the centre
of mass frame is given by
v' -- jj-Vl . (18.6.48)

The components of the velocity of the scattered particle in the centre of mass
frame are therefore
FTL 2
if = (sinocos <p, sin 9 sin <p, cos 9) . (18.6.49)
The relationship between the two sets of angles follows from

v[ = if + V . (18.6.50)
Writing out the components, we get
m
• A
VI s i n 0L c o s IPL =
2 •
s m A COS
F
m2
•a • • a s i•n
Sln
ui sm 0 l sin ifiL — V
m
2 „ mi
üicos9L = —1>1 cos9 +—VI . (18.6.51)
M M '
Thus, we have the desired relation
sin9
tan 9l =
MI/M,2 + cos 9
<PL = <P • (18.6.52)
Now, as stated at the beginning of this section, the number of scattered

particles that are observed in a unit time in a direction AÍ2/, is given by
NDAL(9L, <PL) (18.6.53)

DI ¿L
The same number of scattered particles must be observed in the centre of mass
frame. Therefore, we have
d a i d<7
^ ^ sin 9Ld9Ld<pL = ^' ^ sin9d9dp> (18.6.54)
aí]¿ DIÍ,
Hence,
dcriifii, ¡pi) da(9, ip) sin# d9 dip
(18.6.55)
dQi dQ sin 6i d9i dpi
Now, from (18.6.52) we get
cos eL = \rni/m2+cos 9\ ( l g 6 56)

, / ( m i / m 2 ) 2 + 2 ( m i / m 2 ) cos# + 1
So, after differentiating and some rearranging we find
<1(tl{9l,<Pl) _ [ ( m L / m 2 ) 2 + 2(mi/m 2 )cos6> + l]3/2(¿<T(6>, p) g ^
c/fiz, | m i / m 2 + cos(?| dfi
Therefore, in terms of the scattering amplitudes we obtain
|Tii/m2 + cos 9\
18.7. PROPAGATOR FOR A FREE PARTICLE 385
18.7 Propagator for a Free Particle

The retarded propagator is defined as an operator by the differential equation
ih^--H]G+(t) = S(t) 1 (18.7.59)

at J
with the initial condition that
G+(t) = 0 for t < 0 .
a) Show t h a t if the complete set of eigenfunctions of H is
= {(*!">}
with corresponding eigenvalues En, then the configuration space representation

of G + ( I ) , n a m e l y ( S | G + ( ¿ ) | Y ) m a y b e w r i t t e n
(x\G+(t)\y) = e~iB"t,h M*)K(v) • (18.7.60)

n
Hint: Insert a complete set of intermediate states at the necessary places.

b) Use this result to obtain the propagator for a free particle.
Solution
a) Formally the differential equation (18.7.59) with its accompanying initial
condition is solved by
G+(t) = -l-e{t)e-iHt'h (18.7.61)

N
where 0(t) is defined by

9 (i8 7 62)
w={; ',<i • -
Taking matrix elements of this equation we get
G+(x,y;t, 0) = (x|G + (í)|í/)
n,n'
n,n'
= (18.7.63)
n
b) For a free particle the normalized complete set of eigenfunctions are

1
with corresponding energy

h2k2
= (18.7.65)
Thus, the free particle propagator is given by
G+{x,y-t,0) = -je{t)-^— [ d3ke-i^t/2meik (r-p) (18.7.66)

n {¿ny J
We next evaluate the integral by going to spherical coordinates and aligning the
fc2-axis with x - y to get
J d3k e~ihk^tl2m e>k (2-<!)
i-OO r1
= 2tt y k2 dk e-ihk t/2m

J dueikû (18.7.67)
exp im!f exp dq
i fit
2n m
{ i¡r} /„, {" ( ^ ) "1
(18.7.68)
2
2nm f . láf— u| 1 I 2nm
= ——exp i/— • (18.7.69)
iht \ 2ht J V M + e
In the second last line we have added a term — (e/2m)q 2 . This was simply to
make sure that the integral converges. We can now let e —» 0 + and we see that
this serves for us to choose the proper square root so that the integral becomes
(2ttm\3/2 J\ |£ - y\2 \
exp m
{-m) r ^r|
Substituting this back into the expression for the propagator we get the desired
result
0) = »p{i'» ! £
2sr} (18.7.70)
18.8 Propagator for Simple Harmonic Oscillator

Find the retarded propagator for a one-dimensional particle bound by a har-
monic oscillator.
Hint: Use the result of part a) of problem 18.7, as well as the integral represen-
tation (5.9.71) of problem 5.11.
Solution
The result of problem 18.7 shows t h a t the retarded propagator is given by
G+(x,y,t,0) = -U(t)J2e~iE"t/n<t>n(x)fc(y) • (18.8.71)

18.8. PROPAGATOR FOR SIMPLE HARMONIC OSCILLATOR 387
For a simple harmonic oscillator we have t h a t

2 \ 1/4 1 —a2r2/2
<F>N{X) = — Hn(ax)e (18.8.72)
7r J 2un\
with the corresponding energy
E„ = {n+ l/2)hu (18.8.73)
where we have also written

, mui
a
= J T •
Now, replacing the Hermite polynomials by their integral representation (5.9.71)
2n roo 2
H„(x) = —f=
71
/ (x + ip)n e p
dp (18.8.74)
V " j-oo
we get
G+(x,y,t, 0) = - ! , ( , ) z
H — ' V 7T J 7T71!
oo
/
(ax + ip)na 2e~p 2 dp(ay
3 + iq)n e~q dq e~a ^ +y
' e~'(n+i/2)wt
-OO
e - (® +y ) E-IWT/2
h 7T^/2
dpdq
/ oo 1
— [2(aa; + ip)(ay + ¿9) e~'wt]n e~( p +q
n
-00
1 „ a - 2
— 0(f) r-<x (*2+y2) --iut/2
~ fa *•' jj-3/2 e e
/ OO
exp {2(ax + ip){ay + iq) e~'ut — p2 — q2} dpdq . (18.8.75)
•OO integration variables from p, q to r, a where
If we now change
p = [ T + ] = [ T 1
^ " » ^ - '
(18.8.76)
the argument of the exponential in the integral becomes, after a little algebra,
-icot
i a(x + y) e
- {l + e~iwt) iu
V2 1 + e~
i a(x — y) e~
- (1 — e~iwt) O• +
V^2 1 - e~iut
g — ildt
I _ e-2Út [ q 2 ( ® 2 + y 2 ) e _ , w t - 2a2xy] . (18.8.77)
This now leads to a pair of standard Gaussian integrals. Thus, we can combine
these results to get
G+(x,y]t, 0)
—z x OL A' / 2 ,,1+e-^" 4xye
= -T 2 iuit
-iiot/2
2 g —2iujt
(18.8.78)
After replacing a by s/rmJJh and some simplification we have
G*(x,y,t.0) = i S L u t
expi ?mü; $x2 + y1)smujt - 2xy] J 1 . (18.8. 79)

L 2h sin u)t L J
Bibliography
[17.1] For a totally different way of computing propagators see
E.P. Feynman and A.R.Hibbs, Quantum Mechanics and Path Integrals,
New York: McGraw-Hill, (1965).
r
C h a p t e r 19
Scattering - Time
Independent
19.1 Equations for Spherical Bessel Functions

Use the integral representations for the spherical Bessel functions [19.1]
ji(x) = J_exu{u2 + $ldu
X U 2
^ ( X ) = J ^ £ ^ E ( U + L ) ' D U
a
l roo e'
h?\x) = j^T+T/T J . (u a + 1)'d«i (19.1.1)
to show that
1
a) Zi-i(x) + Zi+i(x) = Zi(x) for / > 1
x
b) -^f-Zi(x) = Z¡-i (x) - ^ * Zi(x) for / > 1

dx x
C) ^ [x_iZ;(i;)] = -x~'zt+1(x) .
Here Zi(x) may be any one of the four spherical Bessel functions.
Solution
We demonstrate these formulae explicitly for ji(x). The calculations for the
other two Bessel functions is exactly the same.
390 CHAPTER 19. SCATTERING - TIME INDEPENDENT
a)
jl = (
"2 + 1 }
' d u
• (19 L2
)
Now,
u d(u2 +1)' = 2lu2(u2 +1)'_1 du = [2Z(u2 + 1)' - 2/(u 2 + 1 ) ' _ 1 ] du .(19.1.3)
So,
(u 2 + 1)' du = <Z(u2 + 1)' + (u 2 + l ) ' - 1 du . (19.1.4)
Therefore, writing this out and integrating by parts we get

„¡-i 21
=
. c t 1/ .2 . 1\/
'• ¿ g « ( l + 1)! /_, * '*"<l("2 + ¿ W
=
1 . . .
Ly xi+1
x '+l
+
1
1 /•'
(1
+ ")<'" +
i 2 t + 1 (/ + 1 ) ! 2f / , < " +
1) «"»•'«+a á - . w
1 a;'+ 2 1 í' r
= -2iJ' M +
. 2 H . ( / + l ) ! 2 i /_.(«I + l ) ' + ' « ' « < ¡ » + 2 í A - , W
= + + • (19.1.5)
Thus finally,
2/4-1 T
— y - i i i x ) = 2y [ Á + i ( * ) + ( 1 9 . 1 . 6 )
and
21 + 1
— - — j ¡ ( z ) = j¡+i(z)+.?'/-i(x) • (19.1.7)
b) Again we begin with

3l{x) = t ^ r y y J \ e™ (« 2 + 1)' du . (19.1.8)
Then,
2/ + 1 . . . (2/ + 1 W - 1 r Tn , , w ,
d
—*<«> = i2'W[! J ' <-+'> "
(2Z+ l ) z —2 /•• 2
J (u + l)ldexu . (19.1.9)
¿ 2 ,+1 Z!
19.1. EQUATIONS FOR SPHERICAL BESSEL FUNCTIONS ayi
So,
ji(x) = — [' <u2 + l)1 dexu

j { í+1
' i2 (/ - 1)!
y J' exu l(u2 + l ) ' - 1 udu . (19.1.10)

i 2'+!/!
Thus,
( + 1),_ UdU 1191

AM = -¡y^'y'-l)! L r "' ' ">
and
dji(x) _ Ix1-1 r
'
l+1 - f e ™ (u 2 + 1)' du
dx i2 l\
J
+
T¥*1i/.«"V+ !)'«<<»
= (19.1.12)
X
Now use the result of part a) in the form
i l + 1
jl+i(x) = ji(x) - j , _ i ( x ) . (19.1.13)
X
Then,
dji(x) . , . , f l 2/+l\
= j , + 1 ( * ) + ( ~ ^ j j , ( * )
= (19.1.14)
c) Here we also begin with (19.1.1)
Ji(*) = 72¡TT¡I J\exu(n2 + l)'du (19.1.15)
and simply differentiate to get
d + i )
^'+i
= T ¥ ^ . r h L ' " ^ '
—a;
- Í exu (u2 + 1)' + 1 ¿ií
i 2<+2 ( / + i )) !• J — i
= -x~l ji+1{x) . (19.1.16)
As stated at the beginning, the calculations for hj 1 ^ and hj ' follow exactly
the same lines.
19.2 Rodrigues Formula: Spherical Bessel

Functions
A useful formula for generating any of the spherical Bessel functions is the
Rodrigues formula
«(*>=*' i r l ú ) ' z
"
Use the results of problem 19.1 to get
dx \ x '
and hence derive the generating formula above.
Solution
To get the desired expression we start with the results of problem 19.1 a)
Zi-i(x) = HLÜ Zi(x) - Zi+i(x) (19.2.17)

X
and combine it with 19.1 b) to get
dZi(x) 21+1 I+ 1
- f a - = —¿-Zi(x)-Z,+1(x)-—Z,(x)
= -Zl(x)-Zl+1(x) . (19.2.18)
X
Now change I to / — 1 and rearrange terms to get:
d Z
Z,(x) = — Zi-i(g)- '-l(x) . (19.2.19)
x dx
Next, we use the identity
d {Zi-1\ I- 1 + _J_ dZ¡-1 (19.2.20)

dx y or' - 1 J x x'_1 x ' - 1 dx
Therefore,
Iterating this expression yields the desired result.

r
19.3.
19.3
WRONSKIAN FOR SPHERICAL
Wronskian for Spherical Bessel Functions

BESSEL FUNCTIONS 393
Show t h a t the Wronskian
W = ji{x)n[(x) - 7i;(x)j,'(:c)
satisfies the differential equation
™ = -2-W.
dx x
Solve this equation and use the behaviour of ji(x), ni(x) for small x to fix the
constant of integration to get
w = i .
Hint: Start with the differential equations for ji(x) and n¡(x) .
Solution
We start with the differential equation for both and n/.
d~ 2 d_ /(/+!)
j,(x) = 0
dx2 x dx x¿
1(1 + 1)
d~ 2 d_ : _
ni(x) = 0 . (19.3.22)
dx2 x dx
Now we multiply the first by n¡(x) and the second by ji(x) and subtract to get
d
-^ = -2-w.
(19.3.23)
dx xx
Therefore,
(19.3.24)
x- '
We can evaluate the constant C for any value of x. So we choose x near 0
and use the asymptotic form of the spherical Bessel functions as well as the
definition of W. Then we find C = 1.
19.4 Superposition of Yukawa Potentials

Find an expression, in Born approximation, for the scattering amplitude due to
scattering off a potential which is a superposition of Yukawa potentials
V(r) = f
JA
M
DN<R(N)
e~"r
R
(19.4.25)
Solution
In Born approximation the solution of the Schrodinger equation has the asymp-
totic form
9m 1 p'kr r
— J e^V{r')d\' (19.4.26)
where
q = it-It' (19.4.27)
is the momentum transfer. So the scattering amplitude is given by
e í (
/(m) = -f?¿/ "~£' ) r
"v(r')dv
wL •
2m f°° a(n)
Here, 1 9 4 2 8
< ' >
2 2 2
q = 4k sin (0/2) . (19.4.29)
19.5 Born Approximation for Gaussian

Potential
a) Calculate the differential cross section in first Born approximation for the
potential
v(r).= v 0 e - ^ 2 .
b) To the same approximation compute the s-wave (I = 0) phase shift.
Solution
a) We have
V
V{r) = V 0 e . (19.5.30)
Then in first Born approximation we have the scattering amplitude
/(*, 0) = ~ J e " " ' ' d3r , (19.5.31)
where
g = 2fcsin(0/2) . (19.5.32)
19.6. BORN APPROXIMATION FOR SQUARE WELL 395
'
Therefore, we get
,,, m 2tt 2mV0 f°° 2 j C i qru

'<M> = 2
T
* d r
L ' du
4n h
2mV0 r°°
r
_ . F
/ re sm qrdr
JO
2mV0 yjit
(19.5.33)
fi2 4 p 3 / 2
Hence we find that
(19.5.34)
b) We have in general that
00
1
/ ( M ) = i X](2/ + 1)e'á' siná
' p '( c o s é l ) • (19.5.35)
;=o
Therefore, we can extract the I = 0 phase shift from
je,<5°sin¿o= f f (k, 9) sin 9 d9 . (19.5.36)

k Jo
Applying this to our result above, we find
1 mV
HO0 • X °
—e sm óo = — 2 exp ( sin 2 (0/2) ) sin 9 d9 . (19.5.37)
2fi p
Therefore we get
1
iS0 • r 2mV 0 [ ¥ / fe2/mu>
(19.5.38)
Solving for the phase shift we get
2mVn [ k
ío = ¿ l n 1- 2 2 - (i_e-
fc2 mu
/ ) (19.5.39)
HK V ^
19.6 Born Approximation for Square Well

Repeat the previous problem for the potential
—Vq for r < a

V(r) = [
0 for r > a
Solution
This time we have
in \ f ~V° r < a
V
^ = 1 0 r > a (19.6.40)
Therefore, as before, with

q = 2ksin(9/2) (19.6.41)
we get
f(k,6) = — J r2 dr J e l q r s i n ddd dip
f
2mVo
r sin qrdr
qh2
2mVo<z sin qa
cos qa (19.6.42)
Jo qa
a) Hence,
da 2 mV 0 2 a 2
|JV cos qa — sin qa
n (19.6.43)
dQ ' qa q4h4
b) Proceeding as for part b) in question 19.6 above we find
e's° sin Jo = [ f{k, 9) sin 9 dd , (19.6.44)

* Jo
so t h a t
— e's° sin ¿o
k
Í
2mVo a cos(2ak sin 9/2) sin(2a&sin 9/2)
sin I (19.6.45)
h2 (2k sin 9/2Y- (2A: sin 9/2)3
Thus, letting
x = 2ak Jo
sin 9/2 (19.6.46)
we can rewrite the integral as
.3 /-2 ka
piSo0 sin jq = — 2mVoa
i e' cos x sin x dx (19.6.47)
u
k h2 Jo x x-
The integrand is a perfect differential
cos x sin x d (sin x
2 (19.6.48)
x x dx
So, we immediately get
1 3 /
id0 • c 2mV0a ^ sin 2ka
r ° =
W~ (19.6.49)
2 ka
Then finally, using the fact t h a t the phase shift is a small angle, we find
nd
. 2mVoa2(ka) ( sin2&a
19.7. PHASE SHIFTS FOR DELTA-FUNCTION POTENTIAL 397
19.7 Phase Shifts for Delta-function Potential

Compute the phase shifts for scattering by a potential
V(r) = Voa 6(i— a) .
Solution
With the potential
V(r) = Vba ¿(r — a) (19.7.51)
we have basically free solutions so t h a t we can write the radial solutions as
Bi ji(kr) r<a
Ri(r) = (19.7.52)
Ai[cosS¡ji(kr) — sinn¡(A:r)] r > a
We have already imposed the physical boundary condition t h a t the wavefunction

is free of singularities at the origin. We now impose that the wavefunction is
continuous at r = a and that the first derivative is discontinuous at 7* = a. The
discontinuity is computed by integrating the radial differential equation about
r = a. Thus, we have
Bi ji(ka) — Ai[cosS¡ji(ka) — sin<5;(Ara)] (19.7.53)
and
h2 dRi dRi
= — Vqü R¡(a) . (19.7.54)
2m dr dr
a+0 a —0_
Writing out this last equation and substituting for B¡ from the first equation
we find with
2mVoa~
Q (19.7.55)
h2(ka)
that
[cosSij'Aka) — sin¿/n¡] — [cosS¡j[(ka) — sinS¡ ,'j j¡]

Ji(ka)
= Q[cosS¡ji(ka) — siní;íi¡] . (19.7.56)
Now, using the Wronskian (see problem 9.3)
ji(x)n[(x) - j',(x)ni(x) = (19.7.57)
to simplify, we can solve for tan to get
(2mV0a2)/h2 jf(ka) (19.7.58)

tan S¡ =
(2mVoa )/h2 ni(ka)ji(ka) — l/(ka)
2
OTO CHAPTER 19. SCATTERING - TIME INDEPENDENT
Another way to solve this problem is to use the partial wave Lippmann-
Schwinger equations
rp(+\r) = + J G\+)(r,r')U(r')^+\r') r'2dr' (19.7.59)
where
G¡+\r,r') =-ikji(kr<)h\l\kr>) . (19.7.60)
Since we are interested in the asymptotic region we have r > a and with
U(r) = 2"^°a S(r - a) (19.7.61)
we find that
rp¡+)(r) = tp{°\r) - ika2^^-ji(ka)h'¡1)(kr)i>¡+)(a) . (19.7.62)
Putting r = a we get
ip\+\a) = — ika—^—-j¡(ka)h\1\ka)ip[+^ (a) (19.7.63)
that
V>| + '(a) = 1 + ika^m^°a ji(ka)h[l\ka} ^0)(a) . (19.7.64)
Thus,
í-M ini ika2my$a ji(ka)xb<l0\a) /i\,

W ( r ) = W (r) 7—— "7TT h\ (kr) . (19.7.65)
1 + ika ^."aji(ka)h\ \ka)
Now, in general we have
^ ( 0 ) ( r ) = i'(2l + 1 )j,{kr) (19.7.66)
and
ip¡+\r) i/>[°\r) + i'(2l + l)ie' S l sin¿¡ /¡¡^(kr) . (19.7.67)
Combining these results we find
, ika2mY§ha ji(ka)i l
(2lV + l)ji(ka)
i'(2l + 1 )ie's' sin J, = V
^ — - . 19.7.68)
1 + ika2mâ ji(ka)h¡ (ka)
From this it follows t h a t
e 2 it,_1 ka2mVn°ljl(ka) 2
(19.7.69)
2i
1 + ika2m^r2 ji(ka)h{l1\ka)
19.8. PHASE SHIFTS FOR YUKAWA POTENTIAL ¿y»
Therefore, after some simplification, it follows t h a t
2iS¡ _ 1 ~ ka2m$a2 ji(ka)n¡{ka)

e—• = ' ' m (19.7.70)
1 + ika^slji(ka)h\l\ka)
If we now write
e 2IS, _ j 2
t.ná, = x (19.7.71)
we see that the result coincides with our previous result (19.7.58).
19.8 Phase Shifts for Yukawa Potential

Compute approximate 1 = 0 and I = 1 phase shifts for scattering a high energy
particle of mass m by a short range potential
r-ar
V(r) = V0-
r
Use whatever seems to be an appropriate approximation.
Solution
For a high energy particle, the Born approximation should be valid. Further-
more, for the Yukawa potential
e-ar
V{r) = V0- (19.8.72)
r
we have the scattering amplitude (in Born approximation) as given, for instance,
in [19.2],
pri /I\ 2mV 0 1 2m.vQ 1 MQ c 7^^

/ (
' v (jfc_jfe/)2 + q2 " 4&2 s in 2 (0/2) + a 2 (
Also the partial wave expansion is given by

OO
f(k,9) = ^Y/(2l+l)[l-Sl(k)]P,(cos6) (19.8.74)

1=0
where
Si(k) = e 2íá '( fc ) . (19.8.75)
However, the Born approximation assumes that the scattering amplitude is

small. Thus, we must have small phase shifts. In this case we replace 1 — Si(k)
by
(19.8.76)
Now projecting out the I = 0 partial wave (by integrating with Po(cos0) = 1)
we get
2mVo i n si d0
—2 iS0(k) =
2 2 2
h J 0 4k sm (0/2) + a2
2mV 0 2sin(fl/2)cos(fl/2)d0
JQ 4k 2 _:„2
A 1.1
sin (0/2) + a 2
2 mVo 4k2'
In 1 + (19.8.77)
' h2k2
Hence we find
2mVb 4/fc2
S0(k) = In 1 + (19.8.78)
~H^K~
Similarly for the 1 = 1 partial wave we get by integrating with Pi (cos 0) = cos f
I sin 0 cos 0 d0
2« , ( * ) =
2k o 4k2 sin2(0/2) + a2
2mV 0 f
I
2MV0 [* 2sin(0/2)cos(0/2)[l - 2sin 2 (0/2)]i
a 2
4k2 sin" (0/2) 4- a 2
, 4/fc2
+ ( i ) , n 2~ (19.8.79)
w r - f q-
Therefore,
, 4arl
}• (19.8.80)
a¿
19.9 Low Energy s-Wave Amplitude

Show that the scattering amplitude for low energy s-waves may be written as
fo(k) = - Co (19.9.81)
r
+ icok — 1/2 cokr"o
as well as in the form
1
fo(k) = (19.9.82)
k cot ó — ik
where c 0 is the scattering length and ro is the effective range [19.3]. Also verify
that both versions of the amplitude satisfy the optical theorem.
Solution
In terms of the phase shift the scattering amplitude is
AW = ¿(1 (19.9.83)
19.9. LOW ENERGY S-WAVE AMPLITUDE 401
This may be rewritten as
fo(k) = ¿ ( l-e2iá)
e's
= —— sin S
k
sin <5 sin <5
'lS k cos S — ik sin
sinSó
1
(19.9.84)
k cot S — ik
On the other hand for low energies the phase shift is given in terms of the
scattering length c 0 and effective range r 0 by
A:cot¿ = — — + Irkr* . (19.9.85)

co 2
Substituting this into (19.9.84) we get
fo(k) 1
-1/co — ik +1/2 kr^
Co
(19.9.86)
1+ ICOK — l/2coa;rg
The optical theorem states that the total cross-section a is given by
<r = y$>[/„(*)] • (19.9.87)
Using the second line of (19.9.84) we see t h a t
(19 9 88)
ywo(*)] = x n r • - -
a
=J
On the other hand the total cross-section is given by
da
dQ
dQ
\f(k)\2dQ = 4n\f(k)\2
- I
4tt . 2
sin" S . (19.9.89)
P
So, in this case, the optical theorem is verified.

If we start with (19.9.81) we have
47T
a = 'MFO(K)}
t
47T —Cp(—Cpk)
k (1 — 1/2 cokrl)2 + c%k2
Aire, o (19.9.90)
(1 — 1/2 c0kr%)2 + cqAt2
4U2 CHAPTER 19. SCATTERING - TIME INDEPENDENT
On the other hand by direct computation we see t h a t

da
=1
a = d n
/ DN
\F(K)\2DN = 4*\F(K)F
AttcI
(19.9.91)
(1 - l/2c 0 A:rg) 2 + cgfc2
Thus, in this case we have also verified the optical theorem.
19.10 Spherical Potential Shell

Consider a potential "shell" of value Vo between r = a and r = b and zero
otherwise. Calculate the s-wave phase shift and show that for large Vo (with
respect to what?) resonances occur approximately at energies which would be
bound states if Vo were infinite.
Solution
The general scattering solution for the Ith partial wave can be written
Í
ji(kr) r < a
Aiji(Kr) + Bini(Kr) a<rb
where
fí2 K2 H2 K
2
E=~— , E-V0 = ^ — . (19.10.93)
2m 2m
Imposing continuity of the wavefunction and the first derivative at r = a we get
ji(ka) = Aiji(Ka) + Bini(Ka)
k j¡(ka) = K[A¡j¡(Ka) + Bim(Ka)'] . (19.10.94)
Then, after some algebra and using the Wronskian for the spherical Bessel func-
tions we find
A¡ = Ka2[kji(ka)n'i(Ka) — Kj'i(ka)ni(Ka)]
B¡ = —Ka2[kj¡(ka)ji(Ka) — Kji(ka)j[(Ka)]. (19.10.95)
Equating logarithmic derivatives at r — b we get
k[cos ólj¡(kb) - sin óirí^kb)] K[A¡j¡{I<b) + Bm'^Kb)] (19 10 96)
cosSiji(kb) — sinSini(kb) A¡j¡(Kb) + B¡n¡(Kb)
Dividing the numerator and denominator of the left side by cos S¡ and solving
for tan J; we get
tan S¡
_ K[A¡j¡(Kb) + Bin't(Kb)]ji(kb) - k[A¡ji{Kb) + B¡ni(Kb)]j¡(kb)
k[A¡ji(Kb) + Bini(Kb)]n¡(kb) — K[A¡j¡(Kb) + Bin[(Kb)\ni(kb)
19.11. EXPRESSIONS FOR jo(x) AND n0(x) 403
For the special case of I = 0 we have
¿„(*) = ^ , „„(*) = - — • (19.10.98)

x x
Thus, we can either simplify the general case using these expressions or start
from scratch, as we now do.
r
f V <a
s
Ro{r)=\ A -^-B^0^ a<r b
Matching the functions and derivatives at r = a we find
feos ka sin ka , . eos Ka sin Ka \ „ / sin Ka eos Ka
= K A[—r, -nn7Í+B\—z— +
ka (ka)2 Ka (Ka)2) \ Ka (Ka
sin ka , sin Ka „ eos Ka
. = A— B——. (19.10.100)
ka Ka Ka
Solving this pair of equations we get
A = — sin Ka sin ka + cos Ka cos ka

k
B = —— cos Ka sin ka + sin A'a cos ka . (19.10.101)
K
Next we match logarithmic derivatives at r = b and get after a little algebra
c o t ( t t + « , ) ' + + ' t m - - s i . (19 ,„,„2)

kb k[A tan Kb-B]
At energies corresponding to a box extending from r = a to r = b we have that
R0(a) = R0(b) = 0 (19.10.103)
so that we get
B = Ata.nKa and B = AtanKb. (19.10.104)
Equation (19.10.102) now clearly shows that at such energies
cot(Ar6 + ¿o) = ±oo . (19.10.105)
Therefore,
kb + ¿o = nn + tt/2 . (19.10.106)
For small values of kb this is the condition for a resonance. So, this approxima-
tion means that Vq is large compared to the energy.
19.11 Expressions for jo(x) and uq(x)

Verify directly by using the differential equation and their behaviour near x = 0
that
• ( \ s i n x an i , , cosx
Mx) - d n0(x) =
Solution
We have t h a t near x = 0
sin x , , cos x 1 , „ „ ,
= 1 and = . 19.11.107
X X X
Thus, they have the correct asymptotic behaviour to be jo(x) and no(x) respec-
tively. Furthermore we find
d ( sin x \ cos x sin x

(19.11.108)
dx V x
and
D2 s m i
s.nx_2cos£ »m£ (19.11.109)
+ 2
2
dx V x
Therefore,
0
•dx¿£ + —
x ax+ > )J (V —x 1 = ' (1911110>
Similarly
d /_c«£\ slnx+cos. (19.11.111)

dx V x / x x2
and
= (19.11.112)
dx¿ V x ' x x1 xó
Therefore,
( £z + l ± + l ) ( ™ ) = 0. (19.11.113)
\dx x dx J \ x /
19.12 Effective Range, Scattering Length

Given the potential
w r ) = / -ô r<a
k
' \ 0 r> a
find the effective range and the scattering length for the s-wave (1 = 0).
19.13. EFFECTIVE RANGE, ETC.: YUKAWA POTENTIAL 405
Solution
As a first step we need to find the s-wave phase shift. To a good approximation
this is given by equation [19.2]
f
2mVo\ sin2(A:r)
sin ¿o r'dr
h2 J (kr)2
mvo 1
ka sin 2ka (19.12.114)
'WK*2 2
Jo
For k —» 0 this yields
l k "T^ 3
2 mvoa 2
sin ¿o
H2
1 1
= —A- ka — -(ka)3 (19.12.115)
5
Here we have introduced the dimensionless parameter
2mVoa2
= (19.12.116)
° h2 '
The scattering length and effective range are defined for k —• 0 by
1 1 (19.12.117)
k cot do ' + ^2RO
Co 2
Using our result for sin ¿o we find
cos¿o = \j 1 — sin 2 So « 1 — — a2(ka)2 . (19.12.118)

v
lo
Therefore, after a little algebra we get
k cot íq — + \k2 ( a / 3 - 6/(5a)) a . (19.12.119)
aa /
So we can read off the answer.
1 (19.12.120)
c 0 = -aa , t*o = [a/3 - 6/(5a)]a .
19.13 Effective Range, Scattering Length:

Yukawa Potential
Repeat the previous problem for the Yukawa potential
V =
Solution
Again we start with the approximate equation for sin Sq
3ín¿o «
r
I ¡ dr -
2mV\
2
e /ir
sin2(kr)
(19.13.121)
Jo h J fir (Arr)2
In this case the parameter a is determined by l//z. However, it is easier, and

more exact, to take the upper limit in the integral as oo. We then get
2mV\ e **r sin2(Arr)

sin ¿o
Í Jo
ImV
dr -
h2 J
2
4k ]
fir (Arr)2
2
In 1 +
H FIK
2 mV k_ _ k3
h2 fi2
k2
1 - ""a
K
-P - (19.13.122)
¡i 2 j
a*
where again we have introduced a dimensionless parameter
2 mV
(19.13.123)
So, after a little algebra we get
k cot ¿o » + ;U 2 (/? - 1 / / ? ) - . (19.13.124)

P ¿ ¡I
So,
P
c0 - r 0 = (/? - 1//?) (19.13.125)
H FT
19.14 Shape-independent Parameters

Use the results of problems 19.12 and 19.13 to fix the parameters of the Yukawa
potential in terms of those of the square well so that both yield the same s-wave
scattering length and effective range. The fact that this is possible is what is
meant by calling this a "shape-independent" approximation.
Solution
From the previous two problems we get:
For the Yukawa potential
P
Co = - r 0 = (/? - 1//3) (19.14.126)
r
» 19.15. PHASE SHIFTS FOR HARD SPHERE 407
where
For the square well we get
c0 = i a a , r 0 = ( a / 3 - 6/(5a)) a (19.14.128)
o
where
2 mvoa 2
a - (19.14.129)
h2
Equating the two scattering lengths we find
2 mV 1 2 mVoa2
a . (19.14.130)
ftV 3 h2
So that
V =^(fia)3V0 . (19.14.131)
Then, we also find that
P=\a(na). (19.14.132)
So, equating the two effective ranges, we get
( 1 9 1 4 1 3 3 )
< < " • » •
This immediately yields that
fta = y j | . (19.14.134)
So, finally
1 /ci\3/'2
V =
Z \ 2 j ^ ' (19.14.135)
19.15 Phase Shifts for Hard Sphere

a) Find the phase shifts for scattering by a hard sphere
r < a
V(r) = { °° .
0 r >a
b) Find the total cross-section for an incoming energy
E = ™
in the two limits:

k-> 0
k oo .
Give a physical explanation for the factors of 4 and 2.

Hint: For k —• oo use the asymptotic forms of j; and n¡ to obtain a simple form
for sin 2 ¿¡. Furthermore, replace the sum over / by an integral so that
l=ka
RKA
u = V ] cr¡ sí ^ f (21 + 1) sin 2 ¿i dl
i
l—n
1=0 * JO
Solution
a) We first have to find the phase shifts for the potential
If we now define
^2=2m_e (19.15.137)
n¿
d2 equation
the radial 2 d becomes
1(1 + 1) 2
2 Ri(r) = 0 r>a. (19.15.138)
d r ^ r dr r2
The solutions of this equations satisfying the condition that Ri(r) corresponds
to a fixed incoming flux for large values of r is
Ri(r) = A[ e'Sl [cos 5; ji(kr) — s i n í ; n¡(kr)] r>a. (19.15.139)
The appropriate boundary condition is that R¡(a) = 0. Thus, we get
tan <5; = 4 ^ 1 . (19.15.140)

ni(ka)
b) To get the total cross-section we use the equation
00
A
<7 = — Y^(2/ + 1) sin 2 Si . (19.15.141)
1=0
Now, for k —> 0
tan S¡ ss sin S¡ = (19.15.142)

ni(ka)
and we can use the asymptotic form of the spherical Bessel functions. Thus,
sin¿i = Á M « (A'fl)2'+1 (19.15.143)

J' 19.15. PHASE SHIFTS FOR HARD SPHERE 409
Clearly in the limit only the I = 0 term survives and we get
sin 2 ¿o = (Ara)2 . (19.15.144)
Thus,
it = 47ra2 for k —> 0 . (19.15.145)
Classically, the differential cross-section is related to the impact parameter

b (see figure 19.1) by
4/ A
9/2
Figure 19.1: Classical scattering off a hard sphere.
or
da^ = _b_db (19.15.147)

dCl sin 9 d9
For a hard sphere we have (figure 19.1)
b = a sin 9/2 . (19.15.148)
So,
da | a 2 sin 9/2 cos 9/2 1 2

= —A (19.15.149)
dQ sin 9 4
Thus,
a = 47T ^ a 2 = 7ra2 , (19.15.150)
a result that is quite obvious since that is the cross-sectional area that intercepts
the incoming beam.
The result for low energy a = 47ra2 (19.15.145) is clearly 4 times the classical
cross-section of na2. The reason for this is that for k —• 0 the wavelength A of
the particle gets infinitely large and the particle scatters off the whole surface
of area 47ra2 of the sphere rather than just off the cross-sectional area na2 per-
pendicular to the incident beam.
In the limit as k —>• oo we have t h a t
tan Si = ^ \^. a \ —y — tan(A:a — lir/2) . (19.15.151)

ni(ka)
Therefore,
sin 2 SI —> sin 2 (fca — LIR/2) . (19.15.152)
Now using the approximation suggested we have t h a t
(T « ^ Í (21 + 1) sin 2 S¡ dl
* Jo
-J J0
_8_ r
k2 J-ka
(21 -(- 1) sin 2 (&a — lx/2) dl
ka(it/2— 1)
Ka{
— (x + ka) + 1 sin 2 x dx
— •
- 27ra (19.15.153)
In this case, since we are approaching a classical limit, we would expect the
cross-section to be just the classical cross-section of na2 and not 27ra2 that we
found. The factor of 2 is due to the way the quantum mechanical cross-section is
defined in terms of the scattered wave. Thus, one defines the total wavefunction
as
= ^ incident + IP.scattered (19.15.154)
ip = 0
Figure 19.2: Shadow scattering off a hard sphere by high energy particles: The
vertical lines depict wavefronts.
19.16. RESONANCE FOR SQ UARE WELL 411
In the case of the hard sphere we have for Ar —> o o that in the shadow of the
sphere the wavefunction vanishes identically (See figure 19.2). This means that,
in this region, one defines
^scattered — V 7 incident (19.15.155)
and so one obtains a cross-section that is twice the classical cross-section. This
effect is known as shadow scattering.
19.16 Resonance for Square Well

Use the exact solution for the square well to find the condition on the potential
for the s-wave (I = 0) to produce a resonance at an energy Eq = (h2kl)/2m.
Comment on the result t h a t you obtain.
Solution
The s-wave phase shift is given by
l t u 6 1 5 6
< >
k jo(ha)n'0(ka) — A j'0(Ka)no(ha)
Using the explicit form in terms of trigonometric functions for j0 and no we find
that
. k tan(A'a) - Atan(fca)
tan¿0 = „ , ,f ,, w /„ ' • (19.16.157)
A + Artan(A:a) t a n ( A a )
This can be rewritten as
1 (19.16.158)
tan So - tan tan •( j7 tan(A'a) ) — ka
For a resonance we require that ¿o = 7r/2. Therefore we need that
Ka = — . (19.16.159)
2
Writing out the definition of K we get
(Ka)2 = — = 2 m
^ 2
+ V
°^ • (19.16.160)
So, for a resonance to occur we need
e o = £ ¿ - v 0 . (19.16.161)
This is the energy for a particle in a box whose zero of energy is at — Vo-
19.17 Double Slit

a) Find the differential cross-section for scattering of electrons from a double
slit. Assume the slits are cut into very thin material and that the incident beam
is normal to the plane of the slits. Also assume that the scattering is weak.
Solution
Using the assumptions stated in the problem we may take the scattering poten-
tial to be
Vo¿(z) for \(d - a) < |z| < ¡¿(d + a) , \y\ <
V (19.17.162)
0 otherwise
Notice that Vo has the dimensions of energy xlength. Also, here we have chosen
the slits to lie in the x — y plane and thus we choose the incident beam along
the z-axis. Since the scattering is weak we may use the Born approximation.
Defining
2 mV(r)
U(Ñ = (19.17.163)
h2
we find that
Ü(q) = J eiqr~U(r) d3
(d+a)/2 rh/2
dz ¿(z)e iqr
UO -(d-a)/2 Ma+a)/¿ t-n/¿ /-c
/ (d+a)/2 (d-a)/2 -h/2
dx+ dx dy
sin•(d+a)/2
(qyh/2) J(d—a)/2 J—h/2 J—
WO [sin(g x (d + a)/2) - s m ( q x ( d - a)/2)]
QXLY
sm(qyh/2)
8t/ 0 s i n ( g r a / 2 ) cos(qxd/2) (19.17.164)
QxQy
Here
q = k — k' . (19.17.165)
We also have that
£ = (0, 0, k) , k' = (k sin 6, 0, k cos 9) (19.17.166)
so that
q = (k sin 0, 0, fc(l — cos 9)) = (¿sin 9, 0, 2k sin 2 (9/2)) . (19.17.167)
In this case we find that the differential cross-section is given by
2
da 1 t2sin (qyh/2) sin2(qxa/2)
r64 U ; \qxd/2)
DQ 167T- Q^
íHy Qx
4iti2VQ A2 sin2(ka/2 sin 9)

cos2(kd/2 sin 9) . (19.17.168)
h4 A-k2 (ka/2)2 sin 2 i
Here, A = ah is the area of the slits.
19.18. SPHERICALLY SYMMETRIC POTENTIAL 413
19.18 Born Approximation: Spherically

Symmetric Potential
Show that for a spherically symmetric potential V(r) the total cross-section in
first Born approximation is given by
r2fc
M
cr / \V(q)\2qdq
2nh 4 k 2 Jo
where V is the Fourier transform of V. Use this result together with the proper-
ties of the Fourier transform to conclude that for high energies and a potential
of finite range a the scattering is appreciable only in the forward direction where
the scattering angle 9 satisfies
sin(«/2) < ¿ •
Solution
The scattering amplitude is given in Born approximation by
(19.18.169)
where
q = 2k\ sin 9/2 \ .
The differential cross-section is then
• (19.18.170)
The total cross-section is now given by
J dCl
n
rrr f ~
= 2n-—/ 2 4 \V(q)\2 s\i\ 9 d9
47T /i Jo
-.2 /-t
[ \V(2ksm9/2)\72sm9/2cos9/2d9 . (19.18.171)
2tt h 4 Jo
Therefore, letting q = 2ksm9/2 we find that
2
m '[•¿K
2k
a = ( 1 9 1 8 I 7 2 )
Now, we are told t h a t the potential V has a finite range a. This means that
the Fourier transform V(q) is appreciably different from zero only for q < 1 /a.
But, from the definition of q this means that
2A*sin 9/2 < l/a . (19.18.173)
Rewriting this we have t h a t for appreciable scattering to occur we need
sin 9/2 < — . (19.18.174)
2 ka
v u / i r i jC/A I y. SUA'1 "1ERING - TIME INDEPENDENT
19.19 Scattering from a Separable Potential

Nonlocal potentials occur in situations where reactions are possible as in the
case of scattering off a bound state t h a t may break up. Nonlocal potentials are
such that in configuration space they are of the form
(r|V|r") = V { f y ) .
If
V ( f y ) = V(r)0(f-f)
then the potential is local. Otherwise it is nonlocal. For a nonlocal potential we

have t h a t
(f\V\ip) = J d3r'{r\V\ff){r'\%l>)
= J d3r'V(r, r')rl>(r') . (19.19.175)
A particularly simple form of nonlocal potential is a separable potential. This

is of the form
V{fy) = /(r)p(f) . (19.19.176)
Solve the scattering problem for a separable potential for an incoming plane
wave
^(0) =
(2^ e ¿ f c
" " (19.19.177)
Solution
The Schrodinger equation for this case is
h2 „ r
— ^^V2V' + /(r) d3r' g{r')'ip(r') = Eip(r) . (19.19.178)
Introducing
a 2 = 2 ^ = (19.19.179)
we get
(V 2 + k2)il> = F(r) J d3r' g(r,)ip{f>) =CF{r) (19.19.180)
where C is a constant defined by the integral. To solve this equation for incoming
boundary conditions we require the incoming Green's function given by
(V 2 + k2)G^+\f, ?) — S(r — r1) . (19.19.181)

19.20. GENERALIZED OPTICAL POTENTIAL
/ g f . aJ
. • / ^ P v^ -9/
The solution is / ^ o s; <J? '
/ v UJ
... 1 e»fc|r-r | " N ^ /
G(+)(r-r-) = _ _ _ _ . (19.19.182)
The Schrodinger equation is now solved by

*'*' = ' f , / ' • <I919183>
To determine the constant C we multiply this equation by g(r) and integrate

over d3r to get
c = { k )
W f T ^ ' - i / • (19.19.184)
This is now an algebraic equation for C. Thus, calling

1 Í Q e'M F _ f / l
A = d r d r F
4 ' y = p \ ^ W ' (19.19.185)
we finally get
(2-K)3/2g(k)
C =
I+ A (19.19.186)
where
g{k) = J e'i' r g(r) d3r (19.19.187)
is the Fourier transform of £r(r). So, we have
= ( 5 ^ ' " ]¡1t+a J • <19i9188>
19.20 Generalized Optical Potential

Suppose two identical spinless particles, mass mo, can scatter into two identical
spinless particles, mass m¿, if their energy is greater than the threshold energy
Ei - 2(m¿ — m 0 )c 2 .
The outgoing particles, of type i, interact via a potential V¿¿. The coupling
between different types of particles (channels) is due to potentials
Vij = Vj i .
If there are N coupled channels the resultant Schrodinger equation reads

N-i
J2 (Tihj + Vtj) IVJ) = (E- E,)\j>i) i = 0,1,2,..N —1 (19.20.189)
j=°
V-/. i i xa. l íiiiuviT - ii¡vi£j írsutJfESSUEN T
where
T, = | ¿ (19.20.190)
is the kinetic energy of particle i .

Assume t h a t there are only two channels (N = 2) and find an equation of
the form
(T0 + V)\rl>0) = (E-E0)\il>0)
for the wavefunction lô) by eliminating the wavefunction \-tpi) - The resultant
"potential" V in this equation is known as a generalized optical potential. Note
that this potential is complex and allows for absorption or emission of particles
of type i = 0.
Hint: Use a Green's function to express \ipi) in terms of |^ 0 )-
Solution
The Schrodinger equation for this case is
—
(To + Voo — E + Eo)\ipo) = Voilî) (19.20.191)
{Ti + Vn - E + E i W Ú = -Violto) • (19.20.192)
From the second of these equations we get
\ipi) =—(Ti + E —E i i e V \ i ) 1
Violto) • (19.20.193)
Here we have imposed outgoing wave boundary conditions as indicated by re-

placing E by E + ie. Substituting this result into (19.20.191) we get
l
[To + Vóo — Vói(Ti + E — Ei + ie + Vii) Vio]\ipo) = (E — Eo\*l>o) (19.20.194)
which is the desired equation.

The generalized optical potential is
V = Voo-Voi(T1+E-Ei+ie + Vn)-1Vio - (19.20.195)
A detailed discussion of optical potentials can be found in the article by

Feshbach [19.4],
19.21 Free Particle Eigenfunctions

Show t h a t the free particle wavefunctions for states of definite angular momen-
t u m (not normalized) may be written
V f c l m M . p ) = (Lx - iLyy-m(Px + i P y ) l
^ j ^ .
19.21. FREE PARTICLE EIGENFUNCI l(JI\b lit
Solution
For a free particle the states of definite angular momentum satisfy
h2k2
2M
L2Í>klm = 1(1 + 1 )ti21pklm
Lztpkim = rnhipkim . (19.21.196)
Here,
H
=%M • (19.21.197)
The solution for I = 0 is

sin(fcr)
ipkoo = kr • (19.21.198)
We now introduce
P+ = Px + ipy
L± = Lx±iLy. (19.21.199)
Then we find that
\p+,H] = [L±,H) = 0 (19.21.200)

so that these two operators do not change the energy eigenvalue. On the other
hand we also find that
[L*,p+] = hp+
[L2,P+] = 2h2p+ + 2h(p+Lz-PzL+) . (19.21.201)
If now consider the equation
L2rpkii = 1(1 + l)h2ipkll (19.21.202)
and act on it from the left with p+ and use the results of (19.21.201) we get
[L2p+ - 2 h 2 p + - 2 h ( p + L z - pzL+)]ipku = 1(1 + 1 )h2p+ipku . (19.21.203)
But,
L+1>ku = 0 . (19.21.204)
Thus, we find
L2p+^kU = (l+ 1)(/ + 2)h 2 P + 1 p k u . (19.21.205)
Now, starting with

Lzipkim = Tnhi>kim (19.21.206)
i i i JL JüIliiMLr liME INDEPENDENT
and acting on it from the left with p+ we find
p+Lzxl>kim = (Lz - h)p+ipklm = mhp+ ipkim . (19.21.207)
Therefore,
LzP+ipkim = ( m + l^p+^fcim • (19.21.208)
Thus, the effect of p+ is to raise both I and m in ipkim by one unit. On the
other hand we know t h a t L_ lowers the value of m in ipkim by one unit, but
leaves the value oil unchanged. So,
L L-Ipkim — "i" 1 ) ^ L-Xpklm
LzL-ipkim = {m - l)hL-ipkim • (19.21.209)
Finally, we use the fact that, up to normalization,
^ = fco. (19.21.210)
is an energy eigenstate corresponding to
i = m = 0 .
Then, we have that
L2p'+ipkoo = 1(1 + l)h2pl+ipkoo

LzP+ipk oo = lhpl+ipk oo (19.21.211)
so that, up to normalization,
P+ipkoo = Ipkii • (19.21.212)
Now, acting with L_ on ipkii for I — m times we get, again up to normalization,

that
L l I m ip l ku — ipkim • (19.21.213)
Therefore, we have proven the desired result.
19.22 Scattering from an Inverse Square

Potential
Find the phase shifts for scattering from a potential
V(r) = V , V o > 0 .
19.23. NEUTRON-PROTON SCATTERING: SPIN FLIP 419
Solution
The Schródinger equation for the radial function reads
d2 2d 1(1+1) b2 2
R i ( r ) = 0 . (19.22.214)
dr2 r dr r2 r2
Here we have introduced the parameters
,, 2mE , ,9 2mVoa2
K A N D B
= - J ¿ T ~ = — • (19.22.215)
The solution finite at the origin is j\(kr) where A is the positive root of
A(A + 1) = l(l+l) + b2 (19.22.216)
so that,
2
A = + y/(l + 1/2) + b2 . (19.22.217)
The asymptotic form of the solution for large kr is

„ , . sin (kr — A W 2 )
R
RI( ) -+ — " — ; — • (19.22.218
kr
This asymptotic form must coincide with
^ s m ( t r - / , / 2 + ¿,) (19.22.219)
kr
Therefore,
SI = /tt/2 - Att/2 = [(/ + 1/2) - y/(l + 1/2)2 + 62] | . (19.22.220)
19.23 Neutron-Proton Scattering: Spin Flip

Assume that the scattering amplitudes for the scattering of a slow neutron by
a proton are f s and f\ for the triplet and singlet states respectively. Find the
probability for the reversal (flip) of the neutron spin if before the collision the
neutron spin was up and the proton spin was down.
Solution
The spin wavefunction of the neutron proton system before the collision is
I t , " ) l i,P) •
This may be rewritten in terms of the triplet and singlet spin wavefunctions as
1
|t,")U,P> = ~ A \ t , " ) l I.P> + I -I-.")I t , p ) )
+ ^ ( 1 t , n ) U , P ) - I 4-.")I t , P ) ) (19.23.221)
Using the given data we have t h a t the scattering amplitude is
1
^ = ( 1 t,rc)l i,p) + 14-,«)l t , p ) )
V2
fi
= ^ ^ I t,")U,p)+ ^ 2 ^ I l,")lt,p) • (19.23.222)
The probability for finding the spins flipped is then given by
1(/3-/I)/212 = I/3-/1I2
2 2 (19.23.223)
|(/ 3 + /i)/2| + | ( / 3 - / I ) / 2 | 2(|/ 3 | 2 + |/ 1 | 2 ) •
19.24 Reflectionless Potential in One Dimension

A particle (mass m) moves in the potential
h2 A2 Ae~Xx
V(x) = -
m (Ae"Al + i f
where A and A > 1 are real constants.

a) Verify that <p(k, x) are solutions of the Schrodinger equation
h2 d2<j)
2 + [V{x) - E]<t> = 0
2m dx
where
-ikx a ae_^ 1 _ 2.k

x)
ae~xx + 1
and
h2k2
E
2m
b) Use these solutions to show t h a t there is no reflection from this potential.
c) Find the energy of the bound states, if any.
Solution
a) By direct differentiation we find t h a t
h2 d2cf> ft2 A2 A, -Xx

) J^D) 2-
2m dx2 2m i (Ae~ Xx
+ l)2 Ae~x + 1
= (E-V)4> • (19.24.224)
19.24. REFLECTIONLESS POTENTIAL IN ONE DIMENSION 421
Thus, as required, <j)(k,x) satisfies the Schrodinger equation for all k.

b) To solve the scattering problem we assume an incoming wave from the left.
In this case the appropriate solution is
Ae~Xx-I 0.f
<t>(k,x) = e'ikx A— x~—7—-—b
x 2z/?
ae~ + 1
-ikx A e~Xx — 1
+ Rc A— x7x — 2 ik x < 0
ae~ + 1
ikx ~xx - 1
a ee"
<j>(k,x) = Té + 2 ik x > 0 (19.24.225)
ae~xx +1
Matching the solutions and their derivative at x = 0 we get
A — I .4-1 \ A - 1
A— —K 2ik + R 2 ik = T A~r — + lik
A+ 1 A+ l A + 1
. A - 1 2A2 A- 1 2A2
ik A — r + 2 ik — ikR — 2 ik - R
A + l A ^ + 1 A
<A- 1 2A2
= ikT A— + 2 ik -T- (19.24.226)
^ + 1
This pair of equations may be rewritten
A —1 2A2 2A2"
R ik [ A ; 7 — 2ik + (T- 1) ik(X——— + 2ik ) — 2A °
A+l A A+l
\ A - 1 A = 0. (19.24.227)
R A— 2ik (T-l) + 2ik
A+ 1 A+l
The determinant of the coefficients of R and T — 1 does not vanish for any real
k. Therefore, the only allowed solution is
T - l = 0
R = 0 . (19.24.228)
This shows that there is no reflected wave.
c) For the bound state the parameter k must be imaginary to give us square
integrable solutions. Thus, we take
Ae~Xx - 1
= Ce" A— - Xx
T7——7 ~í~ 2/Í x < 0
+ 1
A e - Xx i
= Ct xx 2/c x > 0 (19.24.229)
ae~ + 1
Matching the function and derivatives at x = 0 yields
,A-1 A - 1
C A— —|- 2K - D 2K = 0
4+1 A+l
C
K ^ A + l
+ 2/c ] —
AA
(A+l)2
AA
ka
( ?ÍT-2'<) +
+ D = 0 . (19.24.230)
(A+l)2.
For a nontrivial solution the determinant of coefficients must vanish. This yields
1 2A,4
-4 k + A + = 0 . (19.24.231)
Â + LJ ' ( a + 1) 2
The energy is now given by
HHc2 K2 [ A 2 Q 4 - 1 ) 2 + 2\Á\
E= — (19.24.232)
2m 8 m(A + l)2
19.25 n-p Scattering: Singlet and Triplet States

A neutron and proton scatter off each other in a zero orbital angular momentum
state in their centre of mass system. For the triplet spin state the scattering
amplitude is at, whereas for the singlet spin state the scattering amplitude is as.
a) Write the scattering amplitude for this scattering as a matrix in spin space.
b) If we designate the spin up state by (+) and the spin down state by (—) write
the differential cross sections for the following reactions.
1) n(+) + p ( + ) n(+) +p(+)

2) "(+) + p H "(+) +P(-)
3) " ( + ) + P(~) " H + P(+)
4) n
(~)+P{~) "(-) +P(~)
5) " h +P(+) "(-) +P(+)
6) » H + P(+) "(+) +P(~)
c) W h a t is the differential cross section for the scattering of unpolarized neutrons
by unpolarized protons?
Solution
a) There are two different expressions for the scattering amplitude as a matrix.
1) If we write it for states of total spin namely the states
l¿>, M) = | 1 , 1 ) , | 1 , 0 ) , |1, —1), |0,0) .
we have a diagonal matrix

/ at 0 0 0
0 at 0 0
-F(&, 0)s,M-,S'M' = (19.25.233)
0 0 at 0
0 0 0 a,
On the other hand, if we write the scattering amplitude in terms of the neutron
and proton spin states we need the Clebsch-Gordon coefficients
( i 0 0 0 \
0 1 0 1
(l/2,mi;l/2,m2|5M) = 72 72 (19.25.234)
0 i 0 1
"v2
K 0 0 1 0
19.26. PHASE SHIFT, SCATTERING LENGTH, ETC. 423
Here the rows are labelled by m i , m i — H—|-,H—,—I-,—- and the columns

are labelled by (5, M ) = (1,1), (1,0), ( 1 , - 1 ) , (0,0). The scattering amplitude
matrix labelled by m'1, rn'2 is given by
.RRIJ
(1/2, mi; 1/2, m 2 | 5 M ) ( l / 2 , m ' i ; 1/2, m' 2 |S'M') (19.25.235)
and the differential cross-section labelled in the same manner is
Cm ,m'2 = \F(k,0)mitma.m/,mi|2 . (19.25.236)
b) The differential cross sections for the reactions listed may now be written
down by inspection from equations (19.25.234) and (19.25.235).
1) n(+) +p{+) -> n ( + ) + p ( + ) : ^ = |at|2
2) n ( + ) + p ( - ) - » n ( + ) + p ( - ) : — = -|at-t-as|2
3) n ( + ) + p ( — ) - » • n ( — ) + p ( + ) : — = -|at-as|2
4) n ( - ) + p ( + ) ->• n ( - ) + p ( + ) : ^ = i|at + a5|2
5) " ( - ) + ? ( + ) - > • " ( + ) + P ( ~ ) : ^ = | | a t - aa|2
6)n(-)+p(-)->nH + p H : ~ = |at|2. (19.25.237)
c) For unpolarized neutrons and protons the differential cross section is obtained
by averaging (19.25.233) over the initial spin states and summing over the final
spin states. So, using the orthogonality of the Clebsch-Gordon coefficients we
get
%=\ £
= ^ [3|a(| 2 + |a s | 2 ] • (19.25.238)
19.26 Phase Shift, Scattering Length, Etc.

The regular s-wave (I = 0) function for scattering by some central potential is
. . . , . b2 — a 2 &sinh(6r) cos(kr) — 6cosh(6r) sin(fcr)

«(')=«»(*') 6cosh(f>r) -+ asinh(6r) '
a) Calculate the phase shift.

b) Calculate the scattering length and the effective range.
CHAFTER 19. SCATTERING - TIME INDEPENDENT
Solution
a) In the asymptotic region r —> oo the wavefunction must be of the form
u(r) Asin(kr + J) . (19.26.239)
Thus, we take the asymptotic form of the given wavefunction. To do this we
use that
cosh(ftr) —> — ebr and sinh(fcr) —> — e (19.26.240)
Then,
b2 — a2 kcos(kr) — bsin(kr)
u(r) Si {kr) +
" b2 + k2 b+ a
sin(kr) + ^ ^ [kcos(kr) — 6sin(&r)]
• ? ^ , o V ( k 2 + ab) 2 + fc2(¿ - a) 2 x
6*= +
k2 + ab
s'm(kr)
v / ( F + a6) 2 + P ( 6 - a ) 2
k(b — a)
+ 2
cos(Arr) (19.26.241)
\J(k + a6) 2 + &2(6 — a) 2
This is of the desired form (19.26.239) if we define the phase shift J by
k2 + ab
cos S :
2
V(k + afc)2 + k2(b — a) 2
&(6 — a)
sin <5 = (19.26.242)
i / (A; + ab)2 + k2(b — a) 2
2
b) To get the scattering length cq and the effective range r*o we use the result
that for k —>• 0
1 1,9
k cot o —y 1—k ro . (19.26.243)
c0 2
But,
(k2 + ab)
k cot S = (19.26.244)
b— a
Thus, we find that
1 1 9 k2 ab
-~ + -k2r 0 = (19.26.245)
cq 2 b— a
So,
6 —a _ 1 1
co (19.26.246)
ab b a
and
i-o (19.26.247)
19.27. SCATTERING OFF A DIATOMIC MOLECULE 425
19.27 Scattering off a Diatomic Molecule

A diatomic molecule, such as H2 or O2, consisting of two identical atoms is
modelled by two identical spherically symmetric scattering centres separated by
a distance R. (See figure 19.3.) Suppose the scattering amplitude for scattering
an electron off one of these centres is known to be f(0). Find the scattering
cross-section for scattering off the molecule.
Hint: Neglect the effect of multiple scattering.
Solution
With reference to figure 19.3 we see t h a t the electron arrives at atom A earlier
\
f
> r\
( «>íY
)
\,A_y
Figure 19.3: Scattering off a diatomic molecule.
than at atom B. So, its phase at atom A relative to the centre of the molecule
is
eikR/2
Similarly, its phase at B relative to the centre of the molecule is
e-¿fc fi/2
Thus, asymptotically as r —> 00 the scattered wave must be of the form

. . pikrA _ „ pikrB
0 ( r ) = eikR'2 f(0) + e~ik R2
' f(9) . (19.27.248)
r
B
But,
kr¿ = kr — -k' • R and krg = kr + —k' • R (19.27.249)

¿ A
where hk' is the momentum of the scattered electron. We define the "momentum
transfer"
q = k-k' (19.27.250)
and in the denominator of (19.27.248) we replace TA and RG by r. Then,
/(#)-— Lit-fifi-iV-*/* _|_ £-ik R/2+ik' R/2

IP(R)
r
Jkr
eiqR/2 e-iq ig fi/2j
= M — +
(19.27.251)
So the scattering amplitude for the molecule in this particular orientation (fixed
value of R) is
£iqR/2 _|_ e-iq R/2
F(M) = M
= 2f(6)cos(q R/2) . (19.27.252)

The corresponding differential cross-section is
= 4 | / ( f ) | 2 c o s 2 ( g ^ R/2) . (19.27.253)
However, in any given experiment, the direction of the vector R is random. So,
writing
q-RzzqRcosa (19.27.254)
we must average over all values of a . The result is
j i r
— = 4|/(0)| 2 — / cos2(-qRcos a) da . (19.27.255)
as I it Jq 1
If we now change integration variables from a to
x = -qRcos a (19.27.256)
we get
L
da 4|/(0)|2 fqR/2 cos 2 x dx
DCL
qR/2 2
2
= 2\f(9)\ [1 + J0(qR)] . (19.27.257)
19.28 W K B s-Wave Phase Shift: Attractive

Potential
Use the W K B approximation to calculate the s-wave phase shift for scattering
off an attractive potential.
Hint: The most general radial solution is asymptotically of the from
R¡(r) A¡ [cosSi ji(kr) — s\nS¡ni(kr)] . (19.28.258)
19.29. WKB S-WAVE PHASE SHIFT: HULTHÉN POTENTIAL 427
Solution
If we replace the radial function R¡ (r) by
u¡(r) = rR¡(r) (19.28.259)
the Schrodinger equation for 1 = 0 becomes
dru 2. .
— + g (r)u = 0 (19.28.260)
where
,V) = ?!Mr)! = p _„ ( r ) . (19.28.261)
Here we also have the additional boundary condition t h a t

«(0) = 0 . (19.28.262)
The appropriate W K B solution of this equation is therefore
A
u(r) f q(r') dr' (19.28.263)
Jo
To extract the phase shift we need the asymptotic behaviour of
R0(r)
= -u(r)
r
to compare it with the general solution given by (19.28.258). This asymptotic
behaviour is
A
Ro(r) — sin [ q(r')dr'
kr Jo
A .
— sin [ {q(r') — k} dr'+ kr (19.28.264)
kr .Jo
Comparing this with (19.28.258) and taking the limit r —> oo we get
r OO
¿o =
= I [<7(r) — k] dr
Jo
poo
/ [y/k2 — U(r) — k] dr (19.28.265)
Jo
19.29 W K B s-Wave Phase Shift: Hulthén

Potential
Use the results of problem 19.28 to calculate the s-wave phase shift for the
Hulthén potential
v0
V(r) (19.29.266)
1
Hint: Change the integration variable from r to
2 mV(r)
<k2 ft5
Solution
From problem 19.28 we get that the s-wave phase shift is given by
k2_ 2mVAr)_k
-/
Jo
h2
dr (19.29.267)
We now change the integration variable from r to
2 mV(r)
z — \lk2 — (19.29.268)
ft2 '
Therefore,
2mV 0 1
22 = k 2 + (19.29.269)
h2 ear - 1
If we call
p l22 = 2 mvo
(19.29.270)
/i 2
we get that
/? 2
ear = 1 + (19.29.271)
z -fc2
2
'
So, we find that
2 /?2 r dz
dr = —- (19.29.272)
a (z — k )(z — k2 + /?2)
2 2 2
Hence,
= Ml [ a
2 /*oo
JK
(z + fc)(2 2
: dz
— k2 + (32)
(19.29.273)
This integral is easily evaluated and yields
^ In ^ ^ arctan \ / P 2 — k2 f3 > k
8 = (19.29.274)
2k , aA 2 -/? 2 In k-y/k*-P 0< k
f lnf + K + ^/K*-0\
19.30 W K B Approximation for Phase Shifts

The radial part of the Schrodinger equation for a spherically symmetric potential
with
rR(r) = u(r)
becomes
d2u Lo . 1(1 + 1)"
+ kz - U(r) - u = 0 .
dr2
19.30. WKB APPROXIMATION FOR PHASE SHIFTS 429
Here,
U r
( )= and =
'
By changing the independent variable from r to x where
kr = ex
and the dependent variable from u(r) to v(x) where
u(r) = ex¡2 v(x)
one obtains an equation of the form
^ + Q2{x) v = 0
where
Q2{X) = e2x (1 - U/k2) - (I + 1/2) 2 .
Here, the variable x runs over the region —oo < x < oo. Assume that V(r) goes
to zero rapidly at r —> oo and is no more singular than 1 / r near r = 0. Also
assume t h a t there is only one point xo such t h a t
Q J (xo) = 0 .
Clearly, for x —> large positive, Q 2 ( x ) > 0 and for x —> large negative, Q2(x) < 0.
Thus, such a turning point always exists.
Use the WKB approximation to solve the equation for v(x) and then convert
back to the variables r and u(r) and look at the asymptotic region r —^ oo to
obtain the phase shifts.
Solution
For the equation for v(x) the W K B solution in the asymptotically forbidden
region x < XQ must be exponentially damped (since u(r) vanishes at r = 0) and
is
A
i)(a;) : exp - \Q(x')\dx' (19.30.275)
V\Q(x)\
In the classically allowed region x > XQ the solution is
v(x) = l=[cexp-i{£<3(*v*'-i
Vol
¡{/
D exp (19.30.276)
The appropriate matching condition is the one that joins the exponentially
damped solution onto the corresponding solution in the classically allowed re-
gion. The solution in the classically allowed region corresponding to the solution
(19.30.275) is
2A í r 7t i
°{') =
7 m C ° S
\ L Q {
' ' ) d
" ' ~ * l • (19.30.2T7)
Comparing this with the solution (19.30.276) we see t h a t this is accomplished

by choosing
C=D = 2A. (19.30.278)
To find the phase shifts we now revert to the original variables. Thus, we have
u(r) = Vkrv(x) (19.30.279)
and
Q2(x) = r2 ík2 — U(r) — ^ +

V2) ) • (19.30.280)
We also define
kr0 = ex° . (19.30.281)
Then,
I z
u(r) - 2A\
k — U (r) — (/ + l / 2 ) 2 / r 2
2
xcosjj^ yjk2 - U[r') - (I + 1/2 ) 2 / r ' 2 dr' - . (19.30.282)
Now, in general, the asymptotic form of the /th partial wave is

u(r) —• Constant x sin (kr — /7t/2 + S¡) for r —> oo . (19.30.283)
Comparing this with the asymptotic behaviour of our W K B solution we have
to rewrite it as
u (r) = 2A\
k2 - U(r) - (/ + l / 2 ) 2 / r 2
x sin I i : v k2 - U(r') - (/ + l / 2 ) 2 / r ' 2 - k

+k(r-r0)- 1+
4 J2 } . (19.30.284)
The phase shift can now be read off
<5, = (/ + 1 / 2 ) | - kr0
dr1 . (19.30.285)
+
L : V k2 - U(r') - (/ + 1 / 2 ) 2 / r ' 2 - k
Although this equation is quite different from the one obtained earlier (equation
(19.28.265)) they yield similar results.
19.31. ZERO-RANGE POTENTIAL 431
19.31 Zero-Range Potential

Consider the 1 = 0 free Schrodinger equation and define a solution ip by the
boundary condition t h a t the r —> 0 limit of the logarithmic derivative of rip is
a constant i.e.
lim
1 d
, n = a constant independent of energy . (19.31.286)
j—>-0
The resulting problem mimics a potential and is a solution for a zero-range

potential. For such a "potential" both the scattering amplitude and the bound
state energy (if a bound state exists) are determined by a single parameter
namely the constant in (19.31.286) .
Now, consider a zero-range potential which is known to have a single 1=0
bound state with energy
h2a2
e (19 3i 287
' = - 1 í t • ' - >
a) Find the bound state wavefunction ipo(r) as well as the scattering solutions
IK(R) •
b) Show t h a t the solutions are orthogonal and can be normalized to satisfy
d3r \ipo(r)\2 = 1
/'
/' D3RIP0[R)IPJ;{R) = 0
/ ' d rip~{f)^{f)
3
= S(k — q) . (19.31.288)
c) Show t h a t the solutions form a complete set. T h a t is, show t h a t
V,o(f?)V,o(':') + J d3k — S(r — r1) . (19.31.289)
Hint: It may be useful to recall that an integral of the form
/•d k e
3 ikr
has to be regularized and is defined by

r OO rOO
/ dkeikr = lim / d k e ^ ^
Jo £-V0+ Jq
1
= - lim
•o+ ¿(r + ie)
= i[P- - Í7rJ(r)] . (19.31.290)

± vn. ív. jo/t 11 uhimu - TIME INDEPENDENT
Solution
a) If we set
u = rip (19.31.291)
then the function u, for the bound state, satisfies
d2u 2mEo i
d^ = —vr- u = a u
• (19.31.292)
So, the bound state solution is of the form
u = Ae-ar . (19.31.293)
After imposing the boundary condition (19.31.286) we get

,. 1 du
k m —— = —a (19.31.294)
i—•() u dr
which is indeed a constant. Normalization yields
(19.31.295)
A =
The scattering solution has to satisfy the free Schrodinger equation and must
have only an / = 0 outgoing spherical wave. Therefore, it must be of the form
pikr
1
v» s (f) 3 2
e ,fc f + 7f - (19.31.296)
(2tt) / r
We now impose the boundary condition (19.31.286) and get
lim
r-+ 0 [4¿(r,As) (27T) 3 / 2
+ ikf (19.31.297)
Therefore,
1
/ = - (a + ik) (19.31.298)
(2tt)3/2
and
Akr
eik f _ 1
^(f) = 3 2 (19.31.299)
(2tr) / a + ik r
b) The normalization of the bound state solutions has already been demon-
strated so we next consider the orthogonality between the bound state and
scattering solutions.
/ d3rrp0{f]ip^(f)
Q
I J3
°~
:
aT
-e -
e»*-r _ 1 0 — ar-\-ikr
(19.31.300)
(2tt) /2 y -—
3
2TV / d r a + ik
1 y.ó 1. zUIÍU-RANGE POTENI ¡AL
To carry out the integrals we use spherical coordinates and line up the z-axis
with the vector k. In that case the integration over ip yields simply a factor 2n
and the integration over 9, after setting cos 9 = t, becomes
[ eikrt dt = -77— (eikr - e~'kr) (19.31.301)

J—I IRV
Therefore,
J d3r rp0(r)ip¿(r)
VA
27T ik a — ik
VE/I a — ik a + ik A2 + K2 }
27T \ ik
= 0 (19.31.302)
as required. Next, we consider the (delta-function) normalization of the scat-
tering solutions
(r)
J d3rrpt(f)ip?i
d 3 r -t'fcr gtq f
(27 r
_ j _ [fleê-r k r + _i !1 _ [Í d—e*q
a — ik J r a + iq a —— ikik J,J r2r2 J
= S(K-Q) + - ^ { H + I2 + I3} (19.31.303)
where
o- r°°
h = r—————-7- I dr e~ikr (e,qr — e~lqr)
iq(a - ik) J0
2?r
- f dr (e~'(k~q',r — e~i<-k+q)r\ (19.31.304)
q(k + ia) Jo ^ >
H = (19.31.305)
• U R M L
and
H =
47T
(k + ia)(q — ia-f
¡O) JO
dre —i(k — q)r (19.31.306)
Hence we get
^1+^2 + I3
27r (fc-g)r _ i(k+q)r
(k + ia)(q — ia) )
-sr2r*('- 2 ^ 0 ° d r e - i ( f c - 9 ) r j . (19.31.307)
(k-q)r _ —i(fc+9): •j +
We now evaluate the terms in the braces. Thus, we have
= ia(rl)CJre~"k~')r
+ Í5L R DRE^Y - R DRE-^+I)r
kr O OJo rOO
q Jo
+ / DRE^K+^R + / DRE-^+I^
Jo Jo
1 1\ —i ia i ia —i
q kJ k — q — ie k k + q + ie q k + q — ie
i —i
+ +
k + q + ie k + q — ie
a a
« 0,1
___+2,i(i+?)
0 (19.31.308)
since both k > 0 and q > 0. Thus, we have shown t h a t
J d3rtp*-{r)ip$(r) = 6(k - q) . (19.31.309)
c) To show completeness we consider
J d3k
e >£ (f-f) d3k eikr e-ik-r'

3
(2*) { / r J a + ik
3
d3k —ikr' ik.r J_ r dd3kk c<fc(r-r'A
-r' ÍJ - a ik e etk r + rr' J a2 + k2 J
6 ( ? - ? ) + 7 ^ a{ H + H + H} • (19.31.310)
(2V)
Here we have
H + H
"J
d3k ikr „ — ik FF _ 1 / D3K „ — ikr' ikr
e e
a + ik r' J a — ik e e
oo g»fc(r-r') _ g i k ( r + r ' ) eik(r-r') _ g-ik(r+r')
k dk
fc - zq k + ia
2tt r ik(r+r')
k2 2tt R e
J_c
rr' Jo Jk(r-r')
dk k dk-
k2 + a 2 +
rr' k — ia
4tr f c 2
K Jk(r-r')
rr' Jo dk 2
K + A2
I
4tt
rr' Jo
r CO ik(r+r')
ik( r+r
+ dk e - "> - ia dk-
k — ia
43(5 CHAPTER 19. SCATTERING - TIME INDEPENDENT
Solution
a) To obtain the desired form we simply write sinfc(r — r') = sin kr cos kr' —
cos kr sin kr' in the integral equation to get
i R
= sinfcr f — / dr' cos kr' U (r')ip(r')
k Jo
1 fr 1
+ cos kr — / dr' sin kr' U(r')ip(r')
k Jo
= C(r) sin kr + S(r) cos kr . (19.32.319)
So,
1 F
C(r) = 1+ — dr cos kr' U(r')xj)(r')
* Jo
1 R
S(r) — — / dr sin kr' U(r')ip(r') . (19.32.320)
k Jo
b) For r —y oo we have the general result that
ip(r) A sm(kr + i 0 ) (19.32.321)
where A is a constant. Also for U(r) —> 0 sufficiently rapidly for r —> oo we
have that
C(r) —> C(oo) , a finite constant

S(r) —y S(oo) , a finite constant . (19.32.322)
This means that
rp(r) —>• \ / C ( o o ) 2 + 5(oo) 2 x

C(oo) . , 5(oo)
2
: sinin kr +r cos kr
yC(oo) + S(oo) °" ' " y/C(oo)2 + S(oo)2
2
= A [cosio sin kr + sin So cos kr] . (19.32.323)
Therefore,
f SYoo) .
tan óo = — — - = lim t(r) . (19.32.324)
G(oo) r~y oo
c) From equation (19.32.320) for C(r) and S(r) we obtain

dC U
u it \
—— = —cos krw(r)
dr
dS kU
• t It \ (19.32.325)
— = - smfcrV(r) .
But,
dt_ _ CS' - C'S _ N
(19.32.326)
dr ~~ C2 _
C2
BIBLIOGRAPHY 437
where
N = — — [Csmkrip + 5 cos krip]

k
U ,2
=
- J *
u
— [C sin kr + S cos kr]¿ (19.32.327)
k
Therefore,
dt U_
sin kr + — cos kr
dr ' k o
U ;
= — —- [sinfer + t(r) cosfcr]' (19.32.328)
k
d) If |¿(r)| < < 1 we may may expand the right hand side of (19.32.328) and
drop the term involving t 2 to get
u
— fs — — [sin 2 kr + 2<(r) sin kr cos krl
d r k
= — y [sin 2 kr + t(r) sin 2kr\ . (19.32.329)
K
So,
dt
U • , U . 2,
——I—— sin ¿kr i = —— sin kr . (19.32.330)
dr k k
This is a linear equation with integration factor
U(R')
exp / dr sin 2kr' .
Jo
Therefore, we have t h a t
1 i [r U(r') ,1
tlr) = —— exp — / dr'—-—sin 2 kr
k I Jo *
/ dr' U(r') sin 2 r' exp sin 2kr" (19.32.331)

Jo Jo
Thus, using (19.32.324) we get
1 í°° Í f°° U(r')

(19.32.332)
tanáo = — — J dr U(r) sin 2 r exp I — J dr'—-—sin 2kr'
Bibliography
1Í8 <JH AFTER 19. SCATTERING - TIME INDEPENDENT
[19.4] H. Feshbach, Ann. Phys. (N.Y.), 19, 287, (1962).
[19.5] F. Calogero, Nuovo Cimento 27, 261, (1963).

r
C h a p t e r 20
S y s t e m s of I d e n t i c a l
Particles
20.1 Periodic Table

Using the Pauli exclusion principle, determine the maximum number of electrons
in any energy level n of an atom. Neglect the interactions between the electrons.
Solution
If we neglect the interaction between the electrons then each electron experiences
simply a central Coulomb potential and all of them have the same energy levels
given by
7 p1 1
En = J. (20.1.1)
a n¿
If we neglect the spin, all of these energy levels except n = 1 are degenerate.
Including spin, the degeneracy of the rath energy level is given by
n —1
(20.1.2)
1=0
This, according to the Pauli Exclusion Principle, is the maximum number of

electrons that may be accommodated in the energy level En.
20.2 Identical s = 1/2 Particles in I = 0,1 States

Two identical spin 1/2 particles collide.
a) Assume that the wavefunction for the two particles after the collision cor-
responds to a relative orbital angular momentum with 1 = 0. Write out the
possible spinor wavefunctions.
b) Repeat part a) if Z = 1.
i l / l l £t\J. j i j i UJIVIO l t jL/ü/jv i iOAL fAitl'KJLEb
Solution
a) If I = 0 the spatial wavefunction is even under an interchange of spatial
coordinates. Since spin 1/2 particles are fermions, this requires that the spinor
wave function be odd under an interchange of particles. The only possibility is
|«,m„ 1/2,1/2) = | 0 , 0 , l / 2 , l / 2 )
= ± = [|i/2, l / 2 ) | l / 2 , - 1 / 2 ) - | l / 2 , - l / 2 ) | l / 2 , 1 / 2 ) ] . (20.2.3)
b) If I = 1 the spatial wavefunction is odd under an interchange of spatial coor-

dinates. Therefore the spinor wavefunction must be even under an interchange
of particles. So, we have three possibilities
\a, m„ 1/2,1/2) = |1, m, 1 / 2 , 1 / 2 ) . (20.2.4)
When written out these three cases are
11,1,1/2,1/2) = |l/2,l/2)|l/2,l/2)
11,0,1/2,1/2) = -^[|l/2, l/2)|l/2,-1/2)+|l/2,-l/2)|l/2,l/2)]
11,-1,1/2,1/2) = |l/2,-l/2)|l/2,-l/2). (20.2.5)
20.3 Two Identical s — 0 Particles

Consider the collision of two identical spin 0 particles. Separate the wave func-
tion into centre of mass and relative coordinates. Discuss the symmetry required
of the relative wave function. Use this to obtain the modifications required
for the scattering amplitude and hence an expression for the differential cross-
section.
Solution
In the center of mass system the Hamiltonian is
P2 P2
H = — + f— + V(r) (20.3.6)
4m m
where
ñ = 2 (î + ^2) , r = ri - r 2 (20.3.7)
The wavefunction separates into
= eiÑ ñtP(f) . (20.3.8)
The first factor (describing the center of mass motion) is clearly symmetric under
the interchange of particle labels. Thus, since the spin is 0 and we are dealing
20.4. IDENTICAL 5 = 1 / 2 FAKl'lolí» íív oísív / n c IVIADS 111
with bosons we require t h a t the wavefunction i/)(r) must also be symmetric

under the interchange of particle labels. In other words, we need
V>(f) = tf-(-r) • (20.3.9)
In general we would write the asymptotic form of the solution as

Akr
Akrr~ + f(o,<p) — • (20.3.10)
r
After symmetrization, this becomes
- L + V(-r)]
1 eilr + e-ik-f ^ _
7i
+ [f{dj + f { n 9 i p +
(20.3.11)
Now, the detector cannot distinguish between the two particles. Therefore, the
differential cross-section is the sum of the cross-sections for particles 1 and 2.
However, they have identical scattering amplitudes. Thus,
cr{d,<p) = 2 [ / ( 0 , <FI) +F(*-0,<P+ *-)]
I + /(*" - 0,<P+ *")|2 (20.3.12)
20.4 Identical s = 1/2 Particles in Centre of Mass

Repeat the previous problem for two identical spin 1/2 particles. Again, sepa-
rate the wavefunction into centre of mass and relative coordinates as well as spin
coordinates. Assume the particles scatter from a spin-independent potential.
a) Discuss the symmetry required of the relative wavefunction if the scattering
occurs in a singlet (s = 0) state. Use this to obtain the modifications required
for the scattering amplitude and hence an expression for the differential cross-
section.
b) Repeat part a) if the scattering occurs in a triplet (s = 1) state.
Solution
We proceed as in the problem above. However, this time the total wavefunction,
in the center of mass system, is given by
ñ
»í(R, r, S, m s ) = e'^ tp(r)x(S,m,) , (20.4.13)
where x{S, rns) is the wavefunction for the total spin of the two particles,
a) If the scattering occurs in a singlet (s = 0) state then x{S> m ' ) antisym-
metric under the interchange of the two particles. In this case we need that
the total wavefunction must be antisymmetric under the interchange of the two
2U. SYSTEMS OF IDENTICAL PARTICLES
particles and hence, that the spatial part of the wavefunction should be sym-
metric just like in the spin 0 case considered above. Thus, for the singlet case
we get that the appropriate scattering amplitude is
/(.=o)(0, <p) = f(0, <p) +f(n-9,<p+ ?r) (20.4.14)
and the differential cross-section is given by
"•(fl.V7) = \f(,=o)(6,v)\" = 1/(0, V?) + / ( í t - 0,ip+ 7r)|2 . (20.4.15)
Here, we have deliberately dropped the factors of 2 and 1 / \ / 2 that cancel each
other.
b) If the total spin is S = 1 the spin wave function is symmetric under the
interchange of particles. Thus, we need that the scattering amplitude should be
antisymmetric. The result is
/(»=i)(0> <P) - /(0, <P) ~ / ( T - 0, ip + tt) (20.4.16)
and the differential cross-section is given by
a
(0,<p) = |/(»=i)(0,¥>)| 2 = \f(0,<p) - f ( n ~ 0,<P+ t ) | 2 • (20.4.17)
Again we have deliberately dropped the factors of 2 and l/y/2 that cancel each
other.
20.5 Heisenberg Field Operator for B o s o n s

Consider a system of non-interacting bosons and write the Hamiltonian in the
form
OO
h = ^ 2 fw k a\ak •
k=o
Find an explicit expression for
exp[¿Hí/fi] ip(x, t) exp[—iHt/h]
where
OO
^»(®) = •
k=0
Hint: Expand the second exponential and commute ak through showing that
akeXala" = eA(aIa'=+1) afc .

20.5. HEISENBERG FIELD OPERATOR FOR BOSONS 443
Solution
The field Hamiltonian is
H=^hLjka\ak. (20.5.18)
k=0
So, the field operator in the Heisenberg picture is related to the field operator
in the Schrodinger picture by
ip(t, x) = exp{iHt/h} ips(x) exp{-¿Hí/£} (20.5.19)

where
OO
V's(i) = ^ { x \ k ) a k . (20.5.20)
k=o
To evaluate exp{¿Hí/h} ak exp{—iWt/h) we first consider
v a —1
^k (^n^n) — (^n^n)
5—1
— (^/cn + ^k (^n^n)
5-2
— (&kn "i" ^k^n^n (®n®n)
= {Skn + aÍan)2 ak (aân)3 . (20.5.21)
So, after n steps, we have
ak ( 4 a t ) ) ' = ( 4 n + alan)' ak . (20.5.22)
Hence,
ak exp (casein) = ak ^ (aâ„)

5=0
oo g
— ^ ^" y (s k n (z]jizn) a k
S
3=0 '
= exp[c(4„ + aj,a„)] ak . (20.5.23)

So, we get
exp(iWt/K) ak exp(—¿Hi/ft)
= exp (it ak exp (-it u;nqtqn)
= e~ii0kt ak . (20.5.24)
Therefore,
ex
ip(t,x) = exp{¿Hf/ft} ips{x) p{ — ¿Hi/ft}
OO
= ^(x\k) e~iwkt ak . (20.5.25)

444 CHAPTER 20. SYSTEMS OF IDENTICAL PARTICLES
20.6 Heisenberg Field Operator for Fermions

Repeat the previous problem for fermions.
Solution
The solution for fermions starts as in the problem above. However, now it is
even simpler since ak commutes with aj,a n for k ^ n and for k = n we find
analan = -a\anan + an = a„ . (20.6.26)
This means that
exp(¿Hí/ñ) ak exp(—iHt/h)
= exp it ak exp -it an)
= e~iWkt ak . (20.6.27)
So that, as in the problem above, we get
ip(t,x) = exp{¿H<//i} rps(x) exp{—iW}t/h

OO
= j 2 ( s \ k ) e~iwkt ^ • (20.6.28)
k=0
20.7 T w o - b o d y Interaction
Obtain the equation
V=^ J d3xd3y^(x)ip^y)V(x,y)rp(y)ip(x)
for the occupation number space representation of the interaction specified by
1 n
v(xi,x2,...,xN) = ~Y^y(xj,xk).
Solution
For any configuration space operator F its representation in occupation number
space is given by [20.1]
(n0,ni,... |F|m 0 , m i , . . . ) = J d3x i . . .d3xN(n0,ni, ...\

] N
ip (xN). ..ipl(xi)F (xi,f2, • ..,xN)ip(xi) . ,.i/>(xN)
|m0,mi,...). (20.7.29)
20.8. DIAG ONA L IZA TION OF BOSON HAMILTONIAN 445
Now we replace F by
1 JV N
V{xux2,...,xff) = = ^2v(xj,xk) . (20.7.30)
j<k
In this case we easily see that
(no,ni,...|V|mo,mi,...)
=
N(N — 1) I d3Xjd3Xkip^ (£j)
j<k
^ {xk)V(xj, xk)ip(xk)i¡j(xj)\m0, mi,...)
= (n0,ni,...\^ J d3xd3yi^x)ip](y)V{x,y)rp(y)xp{x)\rn0,Tril,...) .
(20.7.31)
Therefore,
V=\J d3xd3y ^(x)tp](y)V(x, y)ip(y)ip(x) (20.7.32)
as required.
20.8 Diagonalization of B o s o n Hamiltonian

Show that the Hamiltonian
H = ^2 Êka\ak + Afcaj.aj. + \kakakj , Ek > 2\\k\

k
can be diagonalized if a t , aj. are bose operators.

Hint: Introduce operators
bk — ukak -(- vkak
and choose the constants uk and vk appropriately. What happens if ak, a[ are
fermi operators? These are simple examples of Bogoliubov transformations.
Solution
We want the transformations
bk = uka{ + vkak
b[ = vâl + umkak (20.8.33)
to maintain the bose commutation relations

[&*,*!] = [«*,«£] = i (20-8-34)
and that all other commutators vanish. This implies that

2
M - \u k \ 2 = 1 • (20.8.35)
So, we can parametrize these coefficients as follows
Vk = e"*k cosh 6k
uk = e1/3t sinh 6k (20.8.36)
Solving for ak , in terms of bk , b\ we get
ak — e~,ak cosh 6k bk — e'^k sinh#fc6£
al = e'ak cosh 6k bl - e~,/3k sinh 6kbk • (20.8.37)

The Hamiltonian now is a sum of terms of the form
Hk = Ek a\ak + Afc<4<4 + A*kakak
2 c t+/3
Ek cosh 6kb\bk — e^ " *' s i n h ^ cosh
+ sinh 2 6kbkb\ — e sinh0¿ cosh

2,ak
+afc e cosh 2 6kblbl — e,^c"k &k' sinh 6k cosh 6kblbk
2 k
+e '^ sinh2 6kbkbk — e,l-c'k l3k
^ sinh 9k cosh 6kbk
+Xk e~2,ak cosh 2 6kbkbk — e~,(ak~Pk^ sinh 6k cosh 6kbkb\.
+e2i0k sinh 2 0*6*6* - sinh 9k cosh 0*6¿&J . (20.8.38)
To get the (off-diagonal) terms involving ój.6], as well as those involving bkbk to
cancel we require that
1 at
— Ek e ( +'3t) sinh 6k cosh 6k
+ Xk e2iak cosh 2 6k + X*k e2i/3k sinh 2 9k = 0 . (20.8.39)
Therefore, writing
A/c = | Ajt |eitTfc (20.8.40)
we see that if we choose
afe = - y , = y (20-8.41)
then all we need is
£'fc-sinh20/ ! = |Afc|cosh20fc . (20.8.42)
So, we take
tanh 26k = . (20.8.43)

Ek
20.9. FORMULA FOR EXP (-A) B EXP (A) 447
The Hamiltonian now reads

Hk = Ek ^cosh 29k b\bk + sinh"
- (Afc + Afc) sinh20/c b\bk + ^ sinh20 fc . (20.8.44)
After substituting for cosh 9k and sinh 9k we get
Hk =
~JWk — 4|Afc|2 ^ ~ 4|Afc|2cos<r] [b\bk + 0 - ^ . (20.8.45)
If we are dealing with fermion operators the Hamiltonian is already diagonal

since
« M = ai*ak = 0 • (20.8.46)
20.9 Formula for e~A B eA

Let A and B be arbitrary operators. Derive the following formula
1
e~A BeA =Y,~AB,A]
z
—• n\ n
n=0
where,
[B,A]0 = B , [B,A]n+1 = [[B,A]ntA].
Solution
We prove the result in two steps. First we prove by induction that
BA
" = ¿ ( " ) Ar[B,A]„-r (20.9.47)
where
[B,A]0 = B , [B,A]n+1 = [[B,A]n,A]. (20.9.48)
Clearly the result is true for n = 0. We now assume the result holds for n — 1.
Thus,
n-1 •
BA"'1 = ( n\ — 1 J Ar[B, A] n _ x _ r . (20.9.49)
r=0
Then,
n—1
BA n 1
" = E ( r )Ar[B,A]n^rA
r=0 ^ '
n 1
X) ( ~ ) X ] „ - i - r + [S, A]„_ r } . (20.9.50)
The coefficients of the terms [B, A]s (1 < s < n — 1) are

71—1 í n —1
An~' + An'
n — 1— ¡ \ n-s
("-!)!
= A* [(n - s) + s]
(n — s)!s!
(20.9.51)
For s = n there is only one term, namely
71—1
(20.9.52)
0
Also for s = 0 there is only one term, namely
= »Z\ )=W 0 (20.9.53)
Thus, in all cases we have the desired result
BAn
=Í2( " )Ar[B,A]n-r. (20.9.54)
r=0
Now consider
n
°° An °° i
= = b.A].. (20.9.55)
v
n=0 n=05=0 '
We can interchange the order of summation. Then this equation reads
D A _ An~' [B,A],
2^Z^(n-s\!(n — s)! «!s!
5—0n=s
oo oo
Ar [B,A],
= E E (r)\ s!
a=0 r=0
[B,A],
= e
E
5= 0
(20.9.56)
A
Multiplying by e on the left we obtain the desired result.
20.10 Bogoliubov Transformation: Fermions

Let ak, aI, (Ar = 1,2) be fermion operators and define
b\ — ua\ — va,2 , &i — u*ai — v* a\
62 = ua{ + ua.2 , b\ = v*a\ + u* a\

20.10. BOGOLIUBOV TRANSFORMATION: FERMIONS 449
with
|U|2+|t,|2=l , UV*~U*V = 0.
a) Verify that b k , b\ are fermion operators.

b) Show that for an appropriate choice of the constant c the unitary operator
U = exp | c a j a 2 — c* a 2 a i j
has the action
b\ = Ua\U] , bi = UaiU]
b\ = Ua\Ut , b2 = Ua2Uf .
This is known as a Bogoliubov transformation.

Hint: Use the results of the previous problem.
Solution
a) We first verify the anticommutation relations.
bib\ + = [(u*ai — v*a\), (wa{ — vc¡2)]+
= M 2 [a}, ai]+ + m 2 [g4, a 2 ]+ = 1 . (20.10.57)
Similarly
6262 + 6362 = [(vai + ua\), (v*a\ + «*02)] +
= H 2 [ a { , a i ] + + |w| 2 [aí,a 2 ]+ = 1 . (20.10.58)
It also follows quite trivially that
M i = M2 - 0 (20.10.59)
and hence that
b\b\ = b\bl = 0. (20.10.60)
This shows that the bj , 6] j = 1,2 are indeed Fermi operators.

b) To verify this part we use the result of problem 20.9 and take
A = c a\al — c* 0.20.1 , B = aj , j = 1,2. (20.10.61)
Now,
a
[a.j, A] 0 = j • (20.10.62)
utiaftisk -ju. systems of identical particles
Then,
[ai,A]1 = [al,A] = ca\ , [a2,A]l = [a2,A] = ca\
[ai,A]2 = [ca\,A] = - | c | 2 a i , [a2, A}2 = [ca\, A] =-\C\2CL2
K,^]3 = -[|c| 2 «i,-4] = - c | c | 2 a í , [a.2,A)7i=[\c\'2a2,A] = c\c\'2a\
[ai,^] 4 = -c[\c\2a\,A] = \c\Aax , [a 2 , A]4 = c[|c| 2 aj, A] = |c| 4 a 2
etc.
Thus,
[«i>^] 4 „ = k l 4 " a i 1 [a2,A]4n = \c\4na2

[ a l > ^] 4n -|-l — c | c | a 2 ' [a2>AI4n + l = - c | c | 4 n a j
fai,^j4„+2 = - I c l 4 n + 2 « 1 . [ 2-^]4„+2 = - | c | 4 n + 2 a 2
a
[ai,^]4„+3 = -c|c|4"+24 , [a 2 ,A] 4 n + 3 = c|c| 4 n + 2 aí .

Therefore,
oo
e~A a\ eA = V ] — [ai,,4] n
^ ' 71!
n=0
Je¡4n _ ~ |c|4„+2 ^
1
1w o ( 4 " ) ! n2 ^^0 ( 4 " + 2 ) 7
£ t / f j c t j e t
|c| 2 4n + 1 ! 4n + 3 !
V¿( ) ¿( ),
c
I I , t • I I (20.10.63)
= a\ cos c + — aó sin c .
let
Similarly
4 , i ü t • i i (20.10.64)
e a2 e = ü2 cos c — — a\ sin c .
\c\
So if we choose
* ii * ^ • i i * (20.10.65)
i/ = cos c , v — sin |c| , c —c
\c\
then we have exactly the transformation required.
20.11 Diagonalization of B o s e Hamiltonian

Diagonalize the Hamiltonian
oo 1
H v k
= H + 9 ( )[akdk + «14].
k=0 k
where oj(k) and V(k) are given functions of k and
= Sk,k' , [afc,a A ] = 0 .
20.11. DIAGONALIZATION OF BOSE HAMILTONIAN 451
Hint: Use the Bogoliubov transformation

bk = ukak - = u*kal - v*kak
with
kl2 - m 2
= i.
Solution
We have
H= + (AKAK + 4 4 ) (20.11.66)
k k
where ui(k) and V{k) are given functions of k and
["fc-4'1 = Sky ' [«£'«£'] = 0

• (20.11.67)
Following the hint we use the Bogoliubov transformation
b
k = ukak ~ vka\ > 4 = "H ~ v a
lk (20.11.68)
with
|2-k-|2=l. (20.11.69)
This means we can parametrize
u¿ = coshx(£) , = sinhx(fc) • (20.11.70)
Inverting the transformation we find
4 = u b
k l + v'kbk ' a
k = u*kbk + vkbl • (20.11.71)
Inserting this into the Hamiltonian we get
H = YJAk [i";;!2bRb\ + |wg|26g6jf + "fc«fc44 + u

*kv*kbkbk]
k
V
+ \ £ W [(«s^je + vkbl)(u*kh + vkbl)
k
+(ukbl + vlbk)(ukbl + vlbk)] • (20.11.72)
After collecting terms, and inserting the parametrization this becomes
H= ^ j^Lj(fc) cosh 2 x ( k ) + V(k) sinh 2x(£)J 4 b k
cos
j(k) sinh2x(fc) + ^ h ^x(k) (»í»í+'i'O
k
J
(k) g (cosh 2 x ( k ) - 1) + sinh %x{k) (20.11.73)
wJinriün ¿\). O 1 S I MS Ut IDEN TICAL PARTICLES
For the Hamiltonian to be in diagonal form requires that the terms involving
fclfcl
k k
+ Kbrbr
K vanish. This means that
So we find
u(k)
cosh2x(fc) =
y/u[k)2 - V(k)2/4
sinh2x ( k ) = _ (20.11.75)
^fu;{k)2 -V{k)2/4
Then we have that
.(k)2-VCk)2/2 « U{k)>/2 + vM/4 „(*) ( 2 o n 7 6 )
yjuj(k)2 -V(k)2/4 k
\/u(k)2 - V(k)2/4 2
20.12 Diagonalization of Quadratic Hamiltonian

Use a Bogoliubov transformation to diagonalize the Hamiltonian
H +
= óX ^ * lfl)( a I a fc + alfcO-ik) + + afca_fc)] .
fc> 0
Here, a k , are Bose creation and annihilation operators respectively.
Solution
The Bogoliubov transformation may be written
ak = 6/. cosh Xk ~ &Lfc sinh Xfc
a_ fc = 6_fc cosh x-fc - ¿¡sinh x~k
a
l = ¿fcCoshx/t - 6-fcsinhx/t
a}_k = 6Lfc cosh x-fc - sinh x-fe (20.12.77)
where bj., bk etc. are again Bose creation and annihilation operators. The
Hamiltonian, in terms of these new operators, becomes
H 6
= \ X IM(-Ea + If I) (cosh2 Xk + sinh 2 X-fc)
k> 0
- 2|¿f| cosh Xk sinh x~k
+ &L k b- k [(E k + |^|) cosh2 x-fc +sinh 2 xfc) - 2\g\cosh X-fc sinh Xfc]
+ \ X ^ + |5|)[sinh2Xfc + sinh 2 x-fc

fc> o
20.13. BOGOLIUBOV TRANSFORMATION: BOSE OPERATORS 453
— cosh Xk sinh X-k ~ cosh X-k sinh Xfc]

+
+ l5l)( cosh Xfc sinh Xk + cosh x-fc sinh x-fc)
fc> o
+ |¿r|(cosh Xk cosh x-k + sinh Xk sinh x~k]
+
\ ^ M - f c [ - ( £ f c + | <71) (cosh Xk sinh Xk + cosh x~k sinh x~k)
k> 0
+ |^| (cosh Xk cosh x~k + sinh Xfc sinh x-fc] (20.12.78)
We now choose
(20.12.79)
X-k = Xk
and
(EK + |5|)sinh(2xk) = |ff|cosh(2x*) . (20.12.80)

So,
l e k + | <71
cosh(2xfc) =
Ek
sinh(2xic) = (20.12.81)
The Hamiltonian is now diagonal and reads

1
H = ^ ^ 2 ( b l b k + b l k b - k ) [ ( E k + |íf|)cosh(2xfc) - M sinh(2xfc)]
fc>0
+ ttEk +
M)[cosh(2X*) - 1] - \g\sinh(2xfc)] . (20.12.82)
k> 0
After we replace cosh(2x/c) and sinh(2x/c) by their values we get
(Et + | 9 l ) 3 ' ! - l « l 3 "

H = 5X>I 6 *+ 6 -* 6 -*)
VE~K
fc> o
(ek + m ) 3 / 2 - \g\ 3 ' 2
+ (Ek + \g\) (20.12.83)
y/E~k
k> 0
The energy eigenvalues may now be read off.
20.13 Bogoliubov Transformation: B o s e

Operators
Consider a finite set N of bose operators AK such that
\p>k 5 ®p] — ^k,p } [Â: 5 ^ — p] — 0 j [ük, ^p] — 0 .

454 <JHAFTEK 2U. SYSTEMS Ut IDENTICAL PARTICLES
Define
bk = cosh A « t + sinh A a\
b\ = cosh A a [ + sinh A ak •
a) Find the commutation rules for bk, b\.

b) Find a unitary operator
Vjv = exp(iTjv)
in terms of a/¡, al such that
vjv <*k VN = bk .
c) Show that
lim (\C|W|$) = 0
N-too
where I1!'). |$) are any states of the form

a
II * l°> "
k
What does this last result mean?
Solution
a) Straightforward computation shows that
[6fc, b\] = 1 and [bk,bk] = [6^, b{] = 0 . (20.13.84)

b) We now want to find an operator 7/v such that
bk = cosh A afc + sinh A

= e'TrJ ak e~iT"
~ £ 777 [ÍTN, 1k]n (20.13.85)

n n\
where (see problem 20.9)
[¿Xjv5Cl/c]n — \}Tn j [^JV i ^k\n — l] j , GE/e]o — &k • (20.13.86)

A short computation shows that we require
[iTN,ak]2n+1 = A2n+1at
[iTN,ak]h2n n = a 2 n ak • (20.13.87)
Thus, we try
ÍTn a "
= \ £ (a* - aI2) (20.13.88)
a= 1
¿u.u. BOGOLIUBOV THAiysrunmAiiuiy: aus>& ur&nAi uns 4óó
and we see that this works.

We can also write this expression as
A N
iT N = - S 2 A k . (20.13.89)
k=1
We also have that

[Ak, Aq] = 0 for k ± q .
c) We begin with states of the form
*|> = («l)"l°> . I*> = («ln°> • (20.13.90)

Then, for fixed n and m and a sufficiently large N, the expression (<E>j |^)
always contains values of j such that there are terms exp Aj with no a j to the
right or aj to the left. This leads us to consider terms of the form (0|e A j |0)
To evaluate such a term we temporarily drop the subscript j and expand the
exponential. This means we need to evaluate
¿ í ^ ( 0 | («2-at T | 0 ) . (20.13.91)
n=0
For the typical mth order term, in this sum, all contributions come from terms
having an even value of m or m = 2n and of the form (—l)"a 2 .. .a*" with an
equal number of a 2 and a*** and the at 2 to the right of the a 2 . Using this we
find that
(0| (a 2 - a* 2 ) 2 |0) = (0|a 4 - 2at2<z2 - 1 + a t4 |0> = ( - 1 ) (20.13.92)
and
(0| (a 2 - a* 2 ) 4 |0) = (0| (a 2 - a* 2 )" (a 2 - |0) = ( - 1 ) 2 2 ! (20.13.93)
and generally
(0| (a 2 - a* 2 ) 2 " |0) = (—l)"n! . (20.13.94)
Thus,
£ ^ < 0 | ( a ' - « ' y i 0 > =

n=0 n=0
< ¿ (~A2/4)n = e -A'/4 , (20.13.95)

n=0
Now, for limjv^-oo the number of such terms tends to oo and we get a term
bounded by
lim [e /4]-'v r
x irrelevant finite factors .
N—too
J I R / J I I X.L/. ÍJ ± KJ JL JLJIVIID V_/R 1JL/JC/Í\ 1 I O A L I V L I Í L IGLJFC/S
Also r is some finite number. Thus, the limit is 0.

This means that norms are not preserved and the operator T = lim
cannot be unitary. The fancy way of saying this is to say that in this case, "the
Bogoliubov transformation is not unitarily implementable".
20.14 Density Matrix for a S u b s y s t e m

Consider the normalized wavefunction \P(ar¿, qj) i — i.. .n , j = 1 . . . N of a
system where the coordinates x¡ refer to a subsystem. Let A^ be any operator
that acts only on the subsystem. Thus, H M represents the total Hamiltonian
for the subsystem. Let
(z|p|2/) = J V*(yi,qj)W(zi,qj)dqi. ..dqN
be the reduced density matrix for the subsystem.

a) Express the expectation value of any operator in terms of A^ and p.
b) Find the normalization condition satisfied by p.
c) Find the equation of motion satisfied by p.
Solution
a) The expectation value of A i s given by
= J y * ( x i , q j ) A ( r ) y ( x i , q j ) d x i .. ,dxndqi .. ,dqN
= J(x\A^\x)dx\.. .dxn . (20.14.96)
But,
(x|a< r >|x)
= J<®|4(r)|y}¿yi • . dyndx 1 . ,.dxndpi .. .dpNV* (yi,pj)V(xi,pj)
= J(x\A(r)\y)dyi .. ,dyn(y\p\x) . (20.14.97)
Therefore,
(^(r)) = J dxi .. .dxn(x\A^r)\y)dy1 .. ,dyn{y\p\x)
= Tr[A( r V] • (20.14.98)
b) The normalization condition follows from the normalization of ^ by simply

putting AW = 1 to get
1 = J dx\ .. ,dxn(x\p\x) = Tr[p] . (20.14.99)

I 2U.15. DENSITY MATRIX AND S-MA1 MX 457
c) To get the equation of motion for the reduced density matrix we expand it
in terms of the eigenfunctions of the Hamiltonian H ^ for the reduced
system. Thus, the time-dependent density matrix is
(x\p\y) = J2anm(f'n^^y)e~,{En~Em)t/h • (20.14.100)

nm
The coefficients anm give p in the energy representation. If we differentiate this
equation with respect to time we find
ih(x\^\y) = ânm[En<fin(x)<f>*m(y) - <j>n(x)Em<f>*m(y)]e-,(-E"-E'»)t/h

nm
a m
= " j\.(t>m{y)H<t>n(x) - n(x)H<t>*m(y)}e~l(Bn~E'n)tlh
nm
= J2anm f[(x\H\¡/)<¡)n(¡/)<j>*m(y)-<t>n{x)<t>*m(y')(y'\H\y)}
nm
x dy[...dy'ne-i(En-E^t'h
= J [(x\H\y')p(y' ,y) - p(x, y')(y'\H\y)dy[ ...dy'n
= (x\Hp - pH\y) . (20.14.101)
Therefore the equation of motion for the reduced density matrix is
ih~ = [H, p] . (20.14.102)
20.15 Density Matrix and S-Matrix

Suppose that the 2 x 2 S-matrix for scattering of a spin 1/2 particle off a spin
zero target is given by
S = f l + g(ñ • (?)
where ñ is a unit vector and
l/l 2 + m 2 = i •
Find the polarization of a beam of particles if the incident beam is totally

unpolarized. Notice that this S-matrix is not unitary.
Solution
Since the incident beam is unpolarized, it is described by the density matrix
Pi = l-l . (20.15.103)
tüO <^ru\r inn. zu. sxSTEMS Ut IDENTICAL PARTICLES
After the scattering the density matrix is
pj = SpiS^
= iss'
= ¿1 + »(/'s)(S • s) . (20.15.104)
Therefore, the polarization after the scattering is given by
P = TT(¿Pj) = ${f*g)n . (20.15.105)
20.16 Zero Energy B o u n d States of T w o Fermions

Two identical spin 1 / 2 fermions of mass m interact via the potential
V(r) = Vi + V¡¡ (?i • <r2
where <?i and <72 are the Pauli matrices associated with the spins of particles 1
and 2 respectively and where
vi = \i
r < a
V2 = ( —U r <a
0 r > a " \ 0 r >a
Here, V and U are positive constants. If this system is known to have a single
zero energy 1 = 0 bound state and a single zero energy 1 = 1 bound state, find
V and U.
Solution
For the 1 = 0 state we have a symmetric spatial wavefunction in the relative
coordinates and therefore the spin state must be antisymmetric or s = 0. For
the / = 1 state the spatial wavefunction is antisymmetric and therefore the spin
wavefunction corresponds to s = 1. Now,
s2 = (si + s 2 ) 2 (20.16.106)
so that
<?i • <72 = J r [ ^ 2 ~ ^l2 - s | ] • (20.16.107)
Thus, for the singlet state we get
• ^2 = ^-[0 — 3/4 — 3/4]ñ 2 = —3 singlet state . (20.16.108)
For the triplet state we get

2
• <?2 = -pr[2 — 3/4 — 3/4]h 2 = 1 triplet state . (20.16.109)
n-
20.16. ZERO ENERGY BOUND STATES OF TWO FERMIONS 459
We first consider the singlet state. In this case we have that
V W
V(r) = V i - 3 V 2 = { - + H I . (20.16.110)
The Schrodinger equation for an I = 0 bound state with energy E < 0 becomes
H2 1 d2(rip)
—» t~5 (-(—V + 3 U — E)ip = 0 r<a
zm r dr¿
h2 1 d2(rip)
-2¿>rd¿r~E* = 0 r > 0
- (2016U1)
The solution is
J j4sin(Arr) r < a ¿ 2 = 2m(v-3t/+£)
r,> = (2 16 112
{ Be~" r>a «' = # 1 " ° ' »
Matching the logarithmic derivative at r = a we get
kcot(ka) = -a . (20.16.113)
So, if we let |i?| —> 0 we find that either
k =0 => V — 3[/ = 0 (20.16.114)
or
h2
ka = (n + l/2)7r =>• V — 3C/ = j ( n + l/2) 2 7r 2 . (20.16.115)
The first case V — 3Í7 = 0 is not possible since this would mean that for this
situation there is no potential at all and hence there is no bound state. So,
.2
V-3U = . (20.16.116)
8 ma¿
We now repeat the solution for the I = 1 case. Here the Schrodinger equation
reads
h2 Id2(rip) 2h2
~2mr~dri ^ 2mr^ + (—V — U — E)i> = 0 r < a
2»!W ) _ ^ = 0 r > 0 . (20.16.117)

2m r dr¿ 2mr¿
The solution to this pair of equations is
( A sin{Kr)-Kr cos(Kr) K2 _ 2M(V+C/+£)
r r <
v> = . \ ar
° - 2mfh . (20.16.118)
B + a) e r >a a2 = 2
ft
Again matching the logarithmic derivatives at r = a we get
(K2a2 + 1) sin(A'a) — Ka cos (Ka) a2 a2 + aa -f- 1
(20.16.119)
Ka cos(Ka) — sin(A'a) aa + 1
» _i- rS\I\± 1 UL£/i5
Now, letting l^l -4 0 we find that the right hand side goes to —1 so that the
equation for the energy becomes
(A' 2 a 2 + 1) sin(Aa) — A'acos(Aa) = —A'acos(A'a)+sin(A'a) .(20.16.120)
Hence, we find
A' 2 a 2 sin(A'a) = 0 (20.16.121)
so that Ka = 0, which is impossible since both U and V are positive, or else
h2
Ka = pn =>• U + V = 2ma2p2x2 * (20.16.122)
So,
U + V = § ¿ . (20.16.123)
Combining this with the result for the singlet case (20.16.115) we get
u r 120 16 1241
= =IS •
20.17 B o s e N u m b e r Operator: Constant
of M o t i o n
Show that for a system of bosons with the Hamiltonian
E
H = (k)alak + A
£ a}alaíafi^+kSr+ñV0~ «)
k jklñ
the number operator
a
n = p h
is a constant of the motion.
Solution
We first compute
[atdp-, ata g ] = [atap-, at]a¿ + at[atap-, a £ ]
= — alap^f,k
= 0 . (20.17.125)
Next we compute
[alap, ata^.aâ^] = aí[ap, ata^-jajja.ís] + ata^-[at, aña^ñ]af
a
= UaU6P¿+jp,-fcada"a-"
- a\aKk{a_ñS?¡ii+añ8Pi-ñ)ap . (20.17.126)
20.18. BOSE OPERATOR: MORE CONSTANTS OF MOTION 461
Now carrying out the sum over p, k, ñ we get

aaaa S
\ l i " j+kSr+ñV0~ ")]
jklñ
= ]C [ah-f(S?,i +6p,-k)a*a-fi - alalk(sf,ñ + < V , - r i W a - p - 1 V(-k-n)
pkñ
— ^ ^ 1 2 A Í A ^ J . A ^ A _ ^ V(—k — TÍ)
kñ
= 0. (20.17.127)
Combining these results we have that
[N,H] = 0 (20.17.128)
which was to be shown.
20.18 B o s e Operator: More Constants of

Motion
Given the Hamiltonian
H = atan + aKa_-k + g ((a 0 ) 2 ata|_- + (aj) 2 aâ_¿)
show that
a) the operator
N = a¡¡ao + at + a^-a_¡¡
is a constant of the motion as well as that
b) the operator
D = a*a-k - a[.a_'k
is a constant of the motion.

Solution
a) If we remember that k ^ 0 we see that
[a 0 ,a^-] = 0 . (20.18.129)
Also, for Bose operators we have that

[aUj,aj] = —a,j
[aUj,al] = a\ (20.18.130)
~Z\J ¿J \^£inr ± n,n zv. siDimviS Uf 1D£¿I\T1(JAL FAtil ICLES
as well as
[ a o a o; ( a o) 2 ] = -2(ao) 2
[aoao> ( a o) 2 ] = 2(aJ) 2 . (20.18.131)
[aja 0 , H] = - 2 g | ( a j ) 2 a ¿ a _ ¿ - a l a ^ . ( a 0 ) 2 } (20.18.132)
as well as that
[ 4 a Á ' H\ = 9 {(ao)2«fca_fc - 4 a _ f c ( a ° ) 2 } (20.18.133)

and, by symmetry, that also
[aUa-k> H
l = 9 {(«í) 2 «fc a -fc - 4 a - f c ( a ° ) 2 } • (20.18.134)
So, after combining these results we immediately see that
[N,H] = 0 . (20.18.135)
b) Furthermore, it is also immediately clear that
[D, H] = 0 . (20.18.136)
20.19 T h e Pauli P r o b l e m
a) Using the equation
E
m = —
cr
calculate the total mass of one gram of electrons confined to a cube 1.0 cm on
a side.
b) Repeat the calculation assuming that instead of electrons the particles are
bosons with the same mass as the electrons.
Solution
a) This problem shows the dramatic effect of the Pauli exclusion principle. In
one gram of electrons there are
=
N
i ^ fn n = 1.1 x 1027 electrons .
-28
9.1 x 1 0
The energy levels for a particle in a cube of side a are
E„ = £Lk> (20.10.13T)
where
k7 = kl + k2y + k] (20.19.138)
and each of the kx, ky, kz is an integer.
As a first step we need to evaluate the degeneracy of each level k. This means
we need the number of different combinations of kx, ky, k2 such that (20.19.138)
is satisfied for a given k. In general, with a few exceptions, this degeneracy is
3. Since the number of electrons is very large, we can take the degeneracy to
be 3. This means that, according to the Pauli exclusion principle, we can place
in each energy level
2x3 = 6 electrons .
The factor of 2 is due to the two possible spin values. To get the total energy
of the N electrons we have to sum all the energies from k = 1 up to k = Ar/6.
Thus, the total energy is
h2*2 ^
e
= 5 = ? e * "
k=1
h27T2 (N/6)[{N/6) + 1]
2
2 ma 2
2 2 2
h 7T N
* h w ' < 20 1 9 139
»
Substituting in the values we get
E = f-°S x 1 0 - " ) ' ( 1 . 1 x 1 0 - " ) ' 26 (20.19.140)

(144) x (9.1 x 10- 2 8 ) x 1
Now using the equation
E
m = —
c-
we convert this to a mass and find
10 26
= 105 gm = 100 kg! . (20.19.141)
9 x 1020
b) If the particles were bosons they would all be in the ground state and the
total energy would be simply
E = ^ - N ^ 14 e r g s . (20.19.142)
2 ma-
in this case the additional mass would be a negligible
14
= 1.6 x 10~ 20 gm . (20.19.143)
9 x 1020
20.20 Atomic Isotope Effect

Every nucleus has a finite radius R = roA1^3 where
r 0 = 1.2 x 10" 1 3 cm
and A is the atomic number of the nucleus. Thus, the potential energy experi-
enced by an electron near a nucleus is not simply
V(r) = •
r
If we assume that the charge density in the nucleus is constant then we have
instead the potential energy
Ze*
R
H r <R
V{r) = <! \2R2 2
J ' ~ " . (20.20.144)
I r >R
a) Use perturbation theory to calculate the isotope shift (that is the dependence
on A of the K-electron (Is state) for an atom with Z protons and atomic number
A.
b) Use this result to compute the energy splitting for the K-electron between
the heaviest lead (Z = 82) isotope A = 214 and the lightest A = 195.
Neglect the presence of the other electrons.
Solution
a) The unperturbed Hamiltonian is
v2 Ze2
Ho=£ • (20.20.145)
v
2m r
The perturbation is
f 111 _ 3 , fil < D

= ; R [2R2 2 r — (20.20.146)
[0 r >R
The unperturbed ground state energy of the K-electron is
E
*) =
~ \ % (20.20.147)
z a, I ZJ
where a = 5.292 x 10" 9 cm is the Bohr radius. The corresponding wavefunction

is
4 0 ) ( r ) = -TL ( — ) ' e~Zrla . (20.20.148)

v87t v a /
20.20. ATOMIC ISOTOPE EFFECT 465
The first order correction to E{00) is given by
E^ = (40), #'40)) • (20.20.149)

Thus,
(1)= i -2Zr/a r2 _ 3
f '
Jo r~ dr (20.20.150)
2ft 2 ~ 2 + 7
0
We now let 2 Va J R
2ZR
a - x - (20.20.151)
2Zr
Then,
1 Ze 2 x2 3 a
I?(!)
c/0
_
— 2 fi Í ' - + x 2 i/x
Jo 2a2 ~ 2 x
1 z e 2 2 r 12 ,12 12 „
3+ a —e [ —— -| 1- 3 (20.20.152)
2 a / 2 a [a2 Ot
If we now make the dependence on the atomic number A explicit by writing
a = a i/3 = 7j4i/3 (20.20.153)
we have the desired dependence on A.

2
(i) _ 1 4-1/3 1 | ^ - 2 / 3 _ 3 + 7ai/3
~ 2 a/Z 7
1/3
_ e-^ (l2A-2'3 + -A'1'3+ 3 (20.20.154)
7 7
b) If we take Z = 82 and A = 195 we get that a = 0.238. Substituting these
values we find that
17
E ¿ ' ( A = 195) = x 9.91 x lO" 3 (20.20.155)
2 a/Z
Similarly, for Z = 82 and A = 214 we get that a = 0.245. Thus, repeating the
calculation we find that in this case
)(A =
- 9214)
1 A\ -= I £ l _ x 1.08 X 10-2 (20.20.156)
2 a/Z
Thus, recalling that
1 e2
= 13.6 eV
a2
the energy difference between the two isotopes is
1 7 f>2
AE = x 9.4 x 10- 4 = 86 eV . (20.20.157)
2 a/Z
Bibliography
C h a p t e r 21
Q u a n t u m Statistical
Mechanics
21.1 Average Energy of Assembly of SHO's

Compute the average energy of an assembly of identical simple harmonic oscil-
lators using:
a) the microcanonical ensemble
b) the canonical ensemble
c) the grand canonical ensemble .
Solution
a) The microcanonical ensemble is given by
(21.1.1)
E<En<E+A
Let
E = (N! + ^)hu , E + A=(N2 + ^)ho: . (21.1.2)
Then we have
N7
(21.1.3)
n=N,
So
i N,+I-N,
huj rjV2(Ar2 + 1) _ (Ni - l)iVi N2 + 1 - Ni
N2 + 1 - Nx 2 2 2
468 CHAPTER 21. QUANTUM STATISTICAL MECHANICS
ñu {N2 + N!){N3 - Ni + 1)
1+
~2 N2 + 1 - JVi
— [1 + JV2 4- Ni] = E + — . (21.1.4)
b) For a single harmonic oscillator the partition function is given by T ^ e - ^ )

where
H = hiL>(a^ a + 1/2) (21.1.5)
This gives
zi = E 1 ,-/3fi(n +1/2)
n=0
e~ptUo/2
~ j _ e-phu
1
(21.1.6)
2sinh(/?/iu;/2)
Next we have that
= [ZX)N = [2 sinh(/3ñw/2)]_'ÍV (21.1.7)
Then,
U = ((E)) = -~\nZN
p\
= N— ln[2sinh(/?/k<;/2)]
OP
= N^-coth(/3hw/2) . (21.1.8)
c) In this case we simply compute

;v
z N
X]
N=0 N=0 \^2sinh(/?fiw/2)
v
vr / y
2 sinh(/?/iui/2)
(21.1.9)
2 sinh(/?fiw/2) — z
Then,
U =
=
fruí z coth(/3 hw / 2)
(21.1.10)
2 2 sinh(/?fiw/2) — z
But we also have that
N
~ zjrlnZG(P>z) = o - utnl /<)\ (21.1.11)
dz 2 sinh(/?/kj/2) — z
21.2. pROpERTIES OF THE DENSITY MATRIX 469
Therefore,
U = N^-coth(phoj/2) (21.1.12)
which is the same as the result for the canonical ensemble.
21.2 Properties of the Density Matrix

Prove the following properties of a density matrix.
a)
P2 0.
b)
Tr([p,A]) = 0 .
To see that this is not trivial, consider Tr([x,p]) and discuss a necessary condi-
tion on the operator A for this to hold.
Solution
a) To prove
p2<p (21.2.13)
we begin with p in diagonal form. This is always possible since p is self-adjoint.

In this case we have
P =^\n)Pn(n\ (21.2.14)
n
with
Y2Pn = 1 Pn
- 0
' (21.2.15)
n
Now, using completeness we immediately get
P2 ~ \n)pl(n\ < Y2 l">Pn<n| = P • (21.2.16)

n n
b) For this part we simply write out the commutator for a general matrix element
and sum over the diagonal elements
^ ] ( m | [ p , A]\m) = ^ {(m|n)p„(n|yl|m) - ( m \ A \ n ) p n ( n \ m ) } = 0 .(21.2.17)

rn m,n
Thus, as required,
t r b , ¿ ] = 0. (21.2.18)
4YU CHAPTER 21. QUANTUM STATISTICAL MECHANICS
This result relied heavily on the fact that the density matrix is a bounded op-
erator, that is, its eigenvalues have a finite bound. Clearly the commutator of
the two unbounded operators x and p is ih and its trace blows up. This is the
case for all unbounded operators since their eigenvalues are not bounded and
therefore neither is the sum of their eigenvalues. So, a necessary condition is
that one of the operators be bounded.
21.3 Expectation Values for Spin

In a gas of electrons, a fraction p is known to have their ¿-component of spin in
the up direction. Assume the remainder are random with equal probability for
up and down.
a) What is the average value of sx, sy, and sz ?
b) If nothing is known about the spins of the remaining fraction 1—p of electrons
what are the maximum possible values of ((s x )), ((sy)) and ((sx)) ?
Solution
a) Using the given information we have that
p = p | + ) ( + | + (l ~P) 5I+X+I + 5 I - X - I 2 u + p°3] • (21.3.19)
Therefore, we find
«»*» = Tr(ps x ) = —Tr(crj + P<J\<T3) = 0
((«y» = Tr(ps y ) = ^Tr(o- 2 + p(r2a3) = 0
((«*)) = T r ( p s z ) = ^Tr(o3 + p(T3a3) = . (21.3.20)
b) In this case we we can only write
P = P|+X+| + (1 ~p)p' • (21.3.21)
Then we have that
{(s x )) = Tr(ps r ) = ^(1 - p ) T r ( p V i ) (21.3.22)
is maximized when
Tr(Ai) = 1 • (21.3.23)
Then the maximum value is (h/2)(1 — p). This requires that
p' = a ¡3
(21.3.24)
¡3* 1 — a
21.4. EXPECTATION VALUE FOR NUMBER OF PARTICLES 471
where
/? + /?* = 1 , 0 < a < 1 . (21.3.25)
Similarly we get that
((s y )) = T r ( p s y ) = ^ ( 1 - p)Tr(pV2) (21.3.26)
is maximized when
Tr(pV 2 ) = 1 . (21.3.27)
Then the maximum value is (h/2)(1 — p). This requires that
>>' = ( r l - a ) P13'28)
where
/?-/T=-¿ , 0< a < 1. (21.3.29)
Finally we have that
Tr
((«*» = (Ps¿) = P^ + -p)Tr(p'<r 3 ) (21.3.30)
is maximized when
Tr(pV 3 ) = 1 . (21.3.31)
Then the maximum value is h/2. This requires that
£ ) (21.3.32)
and
21.4
P = P
I (1
7" ) • I21'3 33»
E x p e c t a t i o n Value for N u m b e r of Particles
Verify the equation
JV = « N » = z ^ l n Z G
where ZQ is the grand canonical partition function and z is the fugacity.

Tt I A u u n r i rjr\ ¿1. 1 U1V1 3 j a j j 3 u o a l ME(.JtlANl(J!}
Solution
To verify
JV = ((N)) = z — \nZG, (21.4.34)
we begin with
N = ^ (21.4.35)
where
p = zNe~pH (21.4.36)
and
ZG=rTrp. (21.4.37)
Then,
Z lnZa = 2 ln TT zN e P
lL ^ [ ( ~")}
TT(UZn e~0H)
t ^ j
TV(Np) (21.4.38)
Tr(p)
as required.
21.5 Spin 1 / 2 Polarization

a) Show that any pure state of spin 1/2 is completely polarized.
b) Find the direction of polarization.
Solution
a) The most general pure spin 1/2 state is of the form
|X) = a|l/2) + 6 | - l / 2 ) , |a| 2 + | 6 | 2 = l . (21.5.39)
The polarization P for such a state is defined by
P = (X\A\X) • (21.5.40)
In this case we find
P={a*b + ab')ex + i(ab'-a'b)ey + (\a\2-\b\2)éz . (21.5.41)

21
-6- DENSITY MATRIX FOR SPIN S=1 473
It now follows that

P •P •=. (a*b + ab*)2 + (ab* — a*b)2 + (|ce|2 — |6| 2 ) 2
= (|a|2+|6|2)2= 1 • (21.5.42)
Thus, the state is completely polarized.

b) If we write the polarization vector in terms of polar angles so that
P = (sin6cos<p, sin0sin¡£>, cos#) (21.5.43)
we can read off the angles immediately.
sin# cos = 25i(a6*)

sin# sin = 2i>(a6*)
cos# = |ct|2 — |6|2 . (21.5.44)
This gives the direction of polarization.
21.6 Density Matrix for Spin s = l

Show that for spin s = 1, the density matrix can be completely specified by the
unit matrix, the polarization vector p, and the quadrupole polarization tensor
Qik defined for spin j by
- (J)
P -
_ (JiJk + JkJi) 2
~ j ( j + l)^ 2 ~ 3 <fe
'
Hint: Show that the unit matrix 1, the polarization vector p, and the quadrupole
polarization tensor Qik form a basis for expanding any 3 x 3 matrix.
Solution
We first need to show that the unit matrix 1, the polarization vector p, and
the quadrupole polarization tensor Qik form a basis for the expansion of the
density matrix. To do this we need 9 linearly independent matrices. We start
by considering the 9 matrices ( l / h 2 ) J i J j i,j = 1, 2, 3 where
0 -i
í - i í i or ii ül i v/2
J3
0 0 (21.6.45)
h
0 0
VÛÍT.1» X UÍW j l / l l i O l I U A L M£/OiÍAiViOi!)
We break these nine matrices up as follows.
— JiJj = + JjJi) + ~ JjJi)
1 1 4 2 i
^2 (JiJj + JjJi) — + J +
2 3 '¿h2 e'jk k
=
3^ ,J 2K¿ e i i k ^ k ' (21.6.46)
Notice that all of these matrices, except the unit matrix, are traceless. Also,
only five of the Q¡j can be independent since
Qli + Q22 + Q33 = 0 and Qij = Qji . (21.6.47)
To show that these 8 traceless matrices plus the unit matrix form a basis for
3 x 3 matrices we need to show that if
S — Gol -I~ o. • J -j- ^ ' aijQij — 0 (21.6.48)
then,
ao = 0, a = 0 , a,j = 0 .
To complete this proof we simply evaluate (by multiplying out the various ma-
trices) the following traces.
Tr (J?) = 2h2
Tr (Ql) = 2/3
Tr (Qfj) = 1/2
Tr ( J i J j ) = 0 i
Tr(JfcQij) = 0
Tr(QkiQij) = 0 i (21.6.49)
We now take the trace of (!
:
3ao 0 .
Next we multiply (21.6.48) from the left by Jk and take the trace to get
2ak = 0 .
Finally we mutliply (21.6.48) from the left by Qk¡ and take the trace to get
a,j = 0 .
This shows that these 9 matrices are linearly independent. Thus, we can write
1 1 ^ 1 j"2 , *^3 ,
p — ao 1 + ai ——b a>2 ——h 03 ——h
n n n
+ «11q11 + CI12Q12 + 113Q13 + 0.22Q22 + 0.23Q23 • (21.6.50)
I 21.7. POLARIZATION VECTOR FOR SPIN j 475
Normalization requires that a o = 1/3. We now find that if we define
- (j)
P=\k (21.6.51)
then
Px-ai , py = a2 , pz = a3. (21.6.52)
Also the remaining five constants a n , , a 2 3 are expressed in terms of the five
linearly independent quantities
(Qij) • (21.6.53)
21.7 Polarization Vector for Spin j

Show that for the case of general spin j, if we again define the polarization
vector by
( / )
*
P—
jh
and are given a "hamiltonian"
H = -7/• B
where B is a magnetic field, then we have

dp --
~ = 1PxB
Í Í2 = 0 .
dt
Solution
This is a straightforward computation.
i k f = L t h l i , .-{•><•}
at jn dt
- • ^ T r { / ( / • B)p - Jp(J • / ) }
-j-hTx{[J{J-B)-J{J-J)]p}
jjîhTr | ( J x 5 ) p |
ihyp x B . (21.7.54)
476 CHAPTER 21. QUANTUM STATISTICAL MECHANICS
Therefore,
^ = 7 PxB. (21.7.55)
It now follows that

d2p d /_
(pxS)
dt2 ^ dt
= f2 (px É) x B = 0 . (21.7.56)
21.8 C o m p o s i t e Density Matrix

Show that if and p1-2'1 are two density matrices and p is defined by
=
Pm,n\m',n' Pm]m' Pnji' (21.8.57)
then p satisfies the general properties of a density matrix
a)
b)
Tr p = 1
c
)
P2 (2)*
r m' ,mr n',n

= (p (1)t )m,,n<(p (2)t kn<
= p^m'p^n' • (21.8.61)
21.9. ARBITRARINESS OF COMPOSITE DENSITY MATRIX 477
b)
Tr p — ^ ^ pm,n]m,n
mn
= = l • (21.8.62)
mn
c
)
=
(P)m,n;k,l ^ ^ Pm,n\m' }n' Pm' ,n'\k }l
m'n'
= v 0 ( 1 ) p (2) 19(1) 0 (2)
/ v rm,m' rn ,n' Pm' tkPn' ,1
m'n'
= (p(1))m,fc(/9(2))2,i • (21.8.63)
Therefore,
(P)2 = (P(1))V2))2 < P(1V(2) = P• (21.8.64)
d)
, (i) r . p )
,n — Pm,m' / ,Pn.n
= Prn,m' • (21-8.65)
Similarly,
Y^P^m.n' = Pn}n-
m \ m /
= p™n. • (21-8.66)
21.9 Arbitrariness of C o m p o s i t e Density Matrix

Show that if we have
pW = ai<7(1) + / ? i r ( 1 ) , qi+/?i = 1 , Qi , /?i > 0
= q20-(2) + / ? 2 r ( 2 ) , C*2 + 02 = 1 1 «2 , 02 > 0
and
Tr<r(i> = Trr ( i ) = 1 .
Then any combination

p = off-l1) (gi cr' 2 ' + 0a^ <g) r ' 2 ' + 77"'1' ® <r^2' + ® r'"'
with
q-)-/3 = a i a + 1 = 012
4YO CHAflEli 21. LJUAN1 UM al Al loll CA L MECHANICS
1 + 6 = fa p + S = p 2
satisfies
=
Pm]m' P m
, n
, m
' , n
(21.9.67)
n
and
P n j i ' ~ ^ ^ Pm,n;m,n' (21.9.68)

m
and is a possible density matrix for the composite system. This establishes the
necessity of the condition that p' 1 ' and p^> correspond to pure states in order
that
Pm,n,m' ,n' = p\n,m'p\i,n' (21.9.69)
give a unique solution for a density matrix for the composite system.
Solution
We have that
Pm,n,m ,n m,m' n,n' ' P m,m' n,n'

r
+ n2n' • (21.9.70)
It now follows that

X Pm,n-,m',n = aar^m, + p(7^m, + 7 ^ , + Sr^m,
n
= (" + + (7 +
= «1 Vm,m- + PlTSm>
= PÍ]m" (21-9.71)
In exactly the same fashion we find that
a a T a Ó T
X / V » ; - . » ' = n ^ > + P n , l + 1 n ! n > + n %
m
= ( a + 7 ) ^ ' + (^ + í ) t " s »
(2) , a (2)
= <X2*n,n' + foTn,i,'
= til • (21-9.72)
Z1.10. TWO ENERGY LUVULS BUS je, Ü/IO <i/y
21.10 T w o Energy Levels B o s e Gas

Consider a "gas" of Bose particles with energy either +E or —E. The Hamil-
tonian for this system is
H = £"(0202 — ajai)
where a¿ , a] (i = 1,2) are the usual annihilation and creation operators for
bosons.
a) Show that the canonical partition function is given by
_ sinh[/3E(N + 1)]
N
~ sinh P E
and that the grand canonical partition function is given by
ZG = [l - 2z cosh pE + 2 2 ] " 1
b) Compute the internal energy U, and the average number of particles ((N))
and express U as a function of /? and ((N)) rather than as a function of ¡3 and
z.
Solution
The Hamiltonian is
H = E(a\a2 - a j a j ) = E(N2 - Ni) . (21.10.73)
a) To compute the canonical partition function we use the fact that the total
number of particles is fixed. So,
N = N1 + N2 or NX=N-N2 (21.10.74)
and
H = E(2N2 — N) . (21.10.75)
So,
N
/3(2N2-N)
ZN = Tre-"" = Y , e~
N2=0
1 _ e-/32£(iV+l)
E0EN
1 — e~p2e
smh[(3E(N + 1)]
(21.10.76)
sinh /3E
For bosons we have
= no-»-'*)-1
k=1
1 1
= (1 - z e - ^ y {l-ze^y
= (l — 2z cosh (3E + 2 r 2 ) 1
. (21.10.77)
tou ON/URI^RI ¿±. U¿UA¡\ 1 UM S1A11SUUAL MECHANICS
An alternate way to compute ZQ is to use

oo
ZG = Yl Z"ZN
N=0
1, OO
OO
V Z N
( ,P(n+I) _ ,-p(N+iy
ih((3E) ^¿7-1
2 sinh(/?£) V
1 / e?e e~pe
2sinh(/?¿?) — zePE 1 — ze~PE
1 / 1 1
2sinh(/?Z?) \e~PE — z eÊ — z
E
1 eP _ e-0E
2sinh(/3£') 1 — z(e@ + e~PE) + z2 E
= (1 - 2 z c o s h ( / ? £ ) + z 2 ) - 1 . (21.10.78)
b) The internal energy is given by

f) f)
U = Zc(f3,z) =—\n[l - 2zcosh/3E + z2]
—2z£"sinh (3E
(21.10.79)
1 — 2z cosh ¡3E + z 2
The average number of particles is given by
N = m ) = z~ In ZG((3,z)
2 z cosh (3 E — 2 z 2
(21.10.80)
1 — 2z cosh (3E + z 2
Therefore,
u = -AT - i E ^ J E . (21.10.81)
2 cosh j3E — z
We can further solve for z in terms of N using the equation for N. This is
simply a quadratic and yields
2 =
fiT+2 { ( ^ + ! ) c o s h p E ± y / { N + 1)2 cosh 2 (3E-N{N + 2) | . (21.10.82)
Now
z = eí3N > 1 . (21.10.83)
Therefore we see that for E —y 0 we have to choose the + sign in the above
expression.
21
-U. DENSITY MATRIX: PARTICLES COUPLED BY SPRING 481
21.11 Density Matrix: Particles Coupled by

Spring
Consider two point masses connected by a spring so that the Hamiltonian is
í, = + + í ( i i ! ) !
¿ ¿ 5 '- • <2111-84>
Find the density matrix for the nth excited state
Pn (*£l, X2} X2) — ,n (^1» %2)lpK,n i j ^2) (21.11.85)

where
h2i<2
HlpK,n{xi,X2) = + (n 1/2)hu> 4 > k a x 1 . 4 ) (21.11.86)
2(m\ + m2)
and
u = y/k/m , m = —— . (21.11.87)
mi+m2
Finally compute the reduced density matrix for particle 1 and show that it is
not idempotent; i.e. that correlations between the two particles persist.
Solution
In the centre of mass system defined by
_ m1xi+m2x2
M
x = x\ — x2 (21.11.88)
with
M•= mi + m2
17111712
m = (21.11.89)
mi + m 2
the Hamiltonian becomes
H< h2 d2<f> h2 d2<t> 1 2 2 roí 1 1 am

t> = - 7 ^ 7 - 7 ^2 1 ~ 7T- 1 T2 T + Ó m u ; r * (21.11.90)
2M dX 2m dx 2
where
u> = \fk/m . (21.11.91)
The eigenfunctions for the nth excited state written in the centre of mass system
are
<t>(x, X) = — elKX un(x) (21.11.92)

Z7T
Kynnr jen ¿í. 1 UM SlAlISllUAL MECHANICS
where
d
~§m cfo 2 ^ +
\ r m j j 2 x 2 u ^ i x ) = (" + 1/2 )hu un(x) . (21.11.93)
In the laboratory frame the wavefunction is
i¡>(xi,x2) = ^ e x p \iK {^~xi + ^ - * 2 ) ] u„(a:i - x2) . (21.11.94)
The density matrix is therefore given by
p(x\ , j #2> 2^2) • exp i K { x i x i ) + { x x 2 )

(27t)2 ( w ~ ' w ' ~ ' )
X«n(Xl ~ X2)u„(xí — x'2) . (21.11.95)
The reduced density matrix is
pi{xi,x'x) = Jp(Xl,x[;x2,x2)dx2 = jêxpîK ( ^ ( x x - x ' ^
x J un(xi - x 2 ), un(x\ - x 2 ) dx2 . (21.11.96)
If we set x2 = x[ — z this can be rewritten as
Pi(x i,«i)
iK
1
2
(2*)
exp (~jfi~(xi ~ xii) J dz un(z)un(z + (x i - a r i ) ) . (21.11.97)
Since this is not just of the form
exp [z'/i ( ^ - ( x i - * i ) ) Un(«i)«n(®i) (21.11.98)

(2tr)
the density matrix p\ is not idempotent and we see that correlations persist.
To convince ourself of the fact that correlations persist we check that p\ is not
idempotent.
e x iK _ x
J PÁxux'^p^x'^x'Ddx'x = ^ ? P "))]
x J dz un(z)un(z + (xi - x[))dz'un(z')un(z'+ (x[ - x")) dx[
x J dzun(z)un(z + (®1 - x"))
= Pi(xi,x") . (21.11.99)
21.12. PARTICLES: DISSIMILAR, BOSE, AND FERMI 483
21.12 Particles: Dissimilar, Bose, and Fermi

Two particles of mass m, confined to a cubical box of sides L, interact with each
other via an interaction
v= v0 S{n - f 2 )
where the particle positions are given by f i and r 2 . Use first order perturbation
theory to discuss the effect of this interaction on the energy levels of this system
for the following cases.
a) The particles are distinguishable.
b) The particles are identical bosons with spin 0.
c) The particles are identical fermions with spin 1/2.
Solution
For the non-interacting particles the energies are just the sums of the individual
energies for a particle in a box and so are given by
(21.12.100)
where
Tlx — (jl{x, Tljy , Tliz) i — 1,2 .
The eigenfunctions for the two states are

/ 2 \ 3/2
un^ft) = ( — J sin(riixnx/L) sm(niyny/L) sm(nizxz/L) .(21.12.101)
a) Distinguishable Particles
In this case the unperturbed wavefunctions are
^ñuñ2{ri,r2) = uñl(fy)uñi(r2) • (21.12.102)
To simplify the computation we work with one coordinate at a time. The change
in energy in first order perturbation theory is given by
4%, = V^,*a) • (21.12.103)

Thus, we first compute
dx\dx2 sm2(nixTTXi/L) sin 2 (n 2a; 7r2: 2 /L)8(x\ — a;2)
dx sm2(n\xnx/L) sin2(n2l:7ra;/L)
l
2 1 f
dx [1 — cos(2nij;7ra;/L)][l — cos(2ri2Xnx/L)]
^(2 + ¿n li n 2 x) • (21.12.104)
484 UHAr i UK 21. QUANTUM STATISTICAL MECHANICS
So, for distinguishable particles the energy correction is given by
~ 0^(2 + ¿ni x n 2 l )(2 + ¿n ly n 2 j,)(2 + • (21.12.105)

b) If we have Bose particles of zero spin, the wavefunction is given by
ipñ l t ñ 3 (ri,r2) = ^ [ « ü j n j u f l j í f b ) + «fll(f2)tiaa(fi)] (21.12.106)
The change in energy in first order perturbation theory is again given by
4 S i l , = (ifo..*». • (21.12.107)
After carrying out the integrations over the delta functions we get exactly double
the integral in the previous case for each direction. Thus,
=
"g"(2 + ^ni x n 3 l )(2 + íni v n 3¡) )(2 + ¿nij,n3B) • (21.12.108)
c) In the case of fermions of spin 1/2 we have two possibilities.
1) If the particles are in a singlet state (s = 0) then the spatial wavefunction is
as for the bose case above and we get
=
^1^2 ~g"(2 + ¿r»i I n 3l )(2 + ¿n lB n 2s )(2 + in ly n 3 y) • (21.12.109)
2) If the particles are in a triplet state (s = 1) the spatial wavefunction is
antisymmetric. In this case carrying out the delta function integrations causes
each integrand to vanish and we get
4 ' u = 0 • (21.12.110)
21.13 A Three-Level Laser

Consider a three-level system with energies E 3 > E 2 > E\ with populations
J'v3, yv2, Ni respectively. If this system is irradiated with monochromatic radi-
ation of frequency
_ E3-E1
" h
it is possible, at equilibrium, to populate level 3 higher than level 2 so that
N 3 > N 2 . This is called a population inversion and permits the possibility of a
laser with the laser frequency
_ £ 3 — E 2
^laser — t •
h
Assume that the only radiation present is the pumping frequency
_ E3 — Ei
h
and use the Einstein Coefficients A¡j, D t j (see problem 1.8) to determine the
ratio of N3/N2 as well as the ratio N-2/N\.
Hint: Write out the equation for the rate of change of the three populations and
impose the equilibrium conditions.
21.13. A THREE-LEVEL LASER 485
Solution
The rate of change of the population in level 3 is entirely due to emission of
radiation. Thus,
dN3
— = -N3Bl3p(u) - N3(A23 + A13) . (21.13.111)
The first term on the right describes the induced transitions from level 3 to level
1, while the second term describes the spontaneous transitions from level 3 to
level 2 as well as from level 3 to level 1. For level 2 we have only spontaneous
transitions from level 3 to level 2 and from level 2 to level 1. Thus,
dN2
— = N3A32 - N2A21 . (21.13.112)
Level 1 is only able to absorb radiation. Therefore,
^ - = N l B 3 1 p(v) . (21.13.113)
At equilibrium, we have that

**¡1 = 0 = N3A32 - N2A21 (21.13.114)
dt
so that
N3 _ A2\
(21.13.115)
N2 A32
Thus, if A2\ > A32 we have population inversion. Also, at equilibrium, the rate
of change of population 3 equals the rate of change of population 1. Therefore,
N3B\3p(v) + N3(A23 + Ai3) = NiB3\p(y) . (21.13.116)
From this we get that
N3 _ B3ip(v)
(21.13.117)
Ni B\3p(v) + (^23 + ^13)
Now, writing
N2 _ N3 N2 (21.13.118)
~Ñ[ ~ ~Ñ~X~Ñ~3
we find that
N2 B3ip(v)A32 (21.13.119)
N\ A2\[B\3p(v) + (A23 + A\3)\
lOU \jnj\r run zi. ljuawium STATISTICAL MECHANICS
21.14 Integrals from Quantum Statistical

Mechanics
In quantum statistical mechanics one encounters integrals of the form
r r00 xkdx
,±lh)
= L ? ± t - <2114120>
a) Show that I+(k) can be expressed in terms of I~(k).
b) Evaluate I±(k) in terms of the Riemann zeta function [21.1]
00 1
= • (21.14.121)
n = 1
Solution
a) To express I+(k) in terms of I-(k) we use the identity
1 1 2_
(21.14.122)
ex + 1~ ex - 1 e2x -
Then,
, r xkdx n r°° xkdx
'*{t) =
L — i - 2
1 (2i 14123)
'
The first integral is simply I~(k) and in the second integral we replace the
integration variable x by t — 2x. Then this integral becomes 2 - f c /_(fe). Thus,
we have that
/+(*) = ( l - 2 - * ) / _ ( / f c ) . (21.14.124)
b) To express I-(k) in terms of the Riemann zeta function we expand the
/_(*) in
denominator = the/f integrand
x* dx in a binomial series.
Jor°° e T.k- , 1
^L
n=00 .©o
* e—' dx
1
=
n=1 J o n = l
OO
= r l'tHEj •
1
r(A+l)C(Jfc+l) (21.14.125)
Bibliography
[21.1] E.T.Whittaker and G.N.Watson, A Course of Modern Analysis - ^th
Edition, Chapter 13, Cambridge, (1963).
[21.2] Kerson Huang, Statistical Mechanics, John Wiley and Sons, Inc., (1963).
References
There are several books that deal with problems and solutions in quantum
mechanics. The following is a fairly exhaustive list of such books presently
available in English.
1. F. Constantinescu and E. Magyari, Problems in Quantum Mechanics, Perg-

amon Press (1971).
2. S. Fliigge, Practical Quantum Mechanics I, Springer-Verlag, (1971).

S. Fliigge, Practical Quantum Mechanics II, Springer-Verlag, (1971).
3. I.I. Gol'dman and V.D. Krivchenkov, Problems in Quantum Mechanics,

Addison-Wesley Publishing Co., (1961).
4. D. ter Haar, Selected Problems in Quantum Mechanics, Academic Press,

(1964).
5. C.S. Johnson and L.G. Peedersen, Problems and Solutions in Quantum

Chemistry and Physics, Addison-Wesley Publishing Co., (1974).
6. Lim, Yung-kuo, Problems and Solutions on Quantum Mechanics: Major

American Universities Ph.D. Qualifying Questions and Solutions, World
Scientific, (1998).
7. L.P. Kok and J. Visser, Quantum Mechanics: Problems and their Solutions,
Coulomb Press Leyden, (1987).
8. H.A. Mavromatis, Exercises in Quantum Mechanics, 2nd reveised edition,

Kluwer Academic Publishers, (1992).
9. G.L. Squires, Problems in Quantum Mechanics with Solutions, Cambridge

University Press, (1995).
On quantum mechanics itself there are very many excellent texts on the
market. A small selection of representative texts is listed below.
10. I. Bialnycki-Birula, M. Cieplak, J. Kaminsky, Theory of Quanta, Oxford

University Press, (1992).
11. D. Bohm, Quantum Theory, Prentice-Hall, (1951).

400 nn,r tvitíijv ü j o
12. A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd Edition, World Scien-
tific, (2002).
13. C. Cohen-Tannoudji, B. Diu & F. Laloe, Quantum Mechanics, John-Wiley

and Hermann, (1977).
14. A.S. Davydov, Quantum Mechanics, Pergamon Press, (1965).
15. E.H. Dicke and J.P. Wittke, Introduction to Quantum Mechanics, Addison-
Wesley Publishing Co., (1960).
16. P.A.M. Dirac, The Principles of Quantum Mechanics, Oxford University

Press, (1958).
17. H. Friedrich, Theoretical Atomic Physics - Second Edition, Springer-Verlag,

(1998).
18. S. Gasiorowicz, Quantum Physics, John Wiley and Sons, Inc., (1974).
19. L.D. Landau and E.M. Lifshitz, Quantum Mechanics, Pergamon Press,
(1985).
20. P.M. Mathews and K. Venkatesan, A Textbook of Quantum Mechanics,

Tata McGraw-Hill Publishing Co. Ltd., (1977).
21. E. Merzbacher, Quantum Mechanics, John Wiley and Sons, Inc., (1970).
22. A. Messiah, Quantum Mechanics Vol.1 & Vol.2, North-Holland Publishing

Co., (1961).
23. M.M. Morrison, Understanding Quantum Physics, Prentice-Hall, (1990).
24. J.J. Sakurai, Modern Quantum Mechanics, Benjamin/Cummings, (1985).
25. L.I. Schiff, Quantum Mechanics, McGraw-Hill Book Co. Inc., (1968).
26. R. Shankar, Principles of Quantum Mechanics, Plenum Press, (1981).

Index
absorption compatible observables, 89
of particles, 416 completeness, 63
adjoint operator, 71 configuration space
anticommutation relations, 449 representation, 444
approximation connection formulae, 288
shape independent, 406 cross-section
asymptotic behaviour, 163 differential, 377
total, 408
Baker-Campbell-Hausdorff, 208 current
basis, 71 particle, 39
Berry's phase, 329-331
Bessel functions de Broglie
spherical wavelength, 3
asymptotic form of, 393, 408 Debye model, 4
Bloch's theorem, 56, 57 deficiency indices, 65, 66, 69, 76, 92
Bogoliubov transformation, 226, 445, degenerate subspace, 263, 271, 273,
449, 451 274
Bohr-Sommerfeld quantization, 9, 10, density matrix, 469, 470, 473, 476,
12, 14 478
Born approximation, 381, 394, 399 density of final states, 315
Bose particles, 479 determinant, 69
bosons, 479 deuteron, 316
boundary diatomic molecule, 425
special, 288, 294 dipole moment
boundary conditions, 51 electric, 43
physical, 52 dipole radiation, 128
Brillouin-Wigner perturbation, 238 Dirac picture, 199, 201
dispersion relations, 380
Calogero equation, 435 distribution
Cauchy sequence, 63 tempered, 114
channels, 415 domain, 71
characteristic equation, 125 kinetic energy of, 75
classically allowed, 294 dynamical phase, 329, 331
classically forbidden, 294
coherent state, 208, 209, 326 effective range, 400, 405
commutator, 71, 72 Ehrenfest equations, 88
compatibility theorem, 89 Ehrenfest's theorem, 142
eigenstate, 33 isotope shift, 464
Einstein coefficients, 6
elliptical coordinates, 296, 298 kinetic energy
emission domain of, 75
of particles, 416 Kramers' Relation, 177
energy Kramers-Kronig relations, 380
zero point, 55 Kuhn-Thomas-Reiche sum rule,
ensemble 248
canonical, 467
grand canonical, 467 Lagrangian, 8
microcanonical, 467 Laporte's rule, 43
equation of continuity, 344 Larmor frequency, 11, 356
equipartition principle, 2 Levi-Civita symbol, 187
Euler angle, 212 Lippman-Schwinger
extensions equation, 435
self-adjoint, 66, 69, 70 Lorentz condition, 368
Fermi's golden rule, 315, 319 magnetic fields, 342

fermion operators, 449 magnetic length, 344
fermions, 444 mean value theorem, 109
fiducial vector, 214 molecule
field operator, 443 diatomic, 425
fugacity, 471 momentum
canonical, 348
gauge transformation, 345-347 mechanical, 348
geometric phase, 329, 330
geometrical phase, 331 neutron, 333
Glauber, 138 normal ordered, 158
group velocity, 17
observable, 92
Hamilton-Jacobi, 13 occupation number space, 444
Hamiltonian, 23 representation, 444
hard sphere, 407 operator
Heisenberg equation, 198 adjoint, 71
Heisenberg picture, 199 linear, 18
Hellmann-Feynman theorem, 61, 179 bounded, 470
Hermite functions, 196 projection, 67
Hermite polynomials, 58, 159, 387 self-adjoint, 67
integral representation, 58 operators
Hult.hén potential, 427 annihilation, 199
creation, 199
integral representation optical potential
spherical Bessel functions for, generalized, 416
389 optical theorem, 400
internal energy, 480 orthonormal, 71
ionization energy, 287 oscillator strengths, 246
in ljka 491
parity, 31, 53 Schrodinger picture, 198, 199

positive, 36 Schwarz inequality, 82, 83, 106
particle current, 39 self-adjoint, 66
partition function self-adjoint extension
canonical, 479 p 4 of, 76
partition function, 468 self-adjoint extensions, 92
Pasternach relation, 177 semi-norms, 114
Pauli equation, 352, 353 separable
Pauli Exclusion Principle, 439 see potential, 414
phase shift, 53, 395 shadow scattering, 411
phase velocity, 17 shape independent approximation,
photo-disintegration, 316 406
Poisson distribution, 220 special boundaries, 288, 294
Poisson's equation, 232 spectral resolution, 67, 68
polarization vector, 473 spin coherent state, 212
population inversion, 484 complex variable representation,
positron, 249 216
positronium, 249 spin states
potential intelligent, 215
local, 414 square well, 407
nonlocal, 414 squeezing operator, 210
separable, 414 sudden approximation, 319, 321, 325
zero-range, 431 sum rule
probability density, 19 electric dipole moment for, 251
projection operator, 67, 68, 72 Kuhn-Thomas-Reiche, 246
propagator, 81 support, 115
support of a function, 114
quadrupole polarization tensor, 473
T-matrix
Ramsauer-Townsend Effect, 39 on shell, 382
Rayleigh Ritz, 292 tempered distribution
Rayleigh-Schródinger perturbation, of fast decrease, 115
239 test function, 114
reduced mass, 317 threshold energy, 415
reflection amplitude, 38 trace, 470
regularization, 82 transition probability
resolvent operator, 68 per unit time, 316
resonance, 403 transmission amplitude, 38
transmission, 39 transmission probability, 40
Riemann turning point, 294
zeta function, 486
Rodrigues formula, 392 unitarily implementable, 456
Runge-Lenz vector, 187 unitarity, 57
scattering amplitude, 399 van der Waals potential, 260

scattering length, 400, 405
"TCÍI
variational technique, 298

vector
fiducial, 214
virial theorem, 32
quantum, 180
Volterra equation, 202
von Neumann, 65
Wigner function, 103

Wigner-Eckart Theorem, 362
WKB approximation, 287
Wronskian, 40, 393, 397
Yukawa potential, 399, 405, 406

superposition of, 393
zero point energy, 55

zero-range
potential, 431
zeta function, 486

A. Z. Capri-Problems &amp; Solutions in Nonrelativistic Quantum Mechanics-World Scientific (2002)

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

A. Z. Capri-Problems &amp; Solutions in Nonrelativistic Quantum Mechanics-World Scientific (2002)

Hochgeladen von

Copyright:

Verfügbare Formate

(pro 6 Cents

World Scientific Publishing Co. Pte. Ltd.

Library of Congress Cataloging-in-Publication Data

QC174.24.N64 C374 2002

British Library Cataloguing-in-Publication Data

Copyright © 2002 by World Scientific Publishing Co. Pte. Ltd.

This book is printed on acid-free paper.

Soon after the first edition of Nonrelativistic Quantum Mechanics appeared,

1 The B r e a k d o w n of Classical Mechanics 1

10 Central Force P r o b l e m s 162

15.12 Excitation of Electron by Electric Field 335

20 S y s t e m s of Identical Particles 439

1.1 Quantum N u m b e r of the Earth

Solving for n we get

1.2 Thermal Wavelength

Therefore, the wavelength is

1 he total energy of the beam is E — IA = N E p h o t o n where N is the number of

1.4 Hydrogen A t o m and de Broglie

The circumference is 2irr. Thus, we write

The energy is given by

Combining this with our previous result we find

ma2c2 ma2c2 1 ma2c2 , ,

c) Roughly what temperature is required to excite these oscillations appreciably?

Emax = hvmax = 1.7 x 1 0 - 2 1 J = 0.011 eV

1.6 Crystal Powder

This means that the internal energy U is reduced by

47rkB _ 4tt x 1.38 x 1 Q - 2 3 J / K

1.7 Einstein Coefficients

Here, hu = En — Em, K is a proportionality constant, and Nn, Nm are respec-

Nn=Ne~E"/kBT , Nm = Ne~Bm/kBT (1.7.34)

where N is the total number of atoms. Therefore,

Nm = Nn e(En-Em)/kBT _ N n ehv/kBT (1.7.35)

NmPnm = NnPmn (1.7.36)

Nnehv'kBT Brnnp{v)=Nn[Bnmp{v) + Anm} . (1.7.37)

Thus, solving for the radiation density p we get

Bnm — Bmn (1.7 .40)

2.1 Lagrangian and Hamiltonian for SHO

L = T — V = \-rnv2 - -kx2 . (2.1.2)

Next we compute the canonical momentum

Then, we can substitute p/m for v and get

T = ^mi;2 = ^r2 + r292^j (2.2.5)

The potential energy is

L = T — V = ^ml292 — m<¡r/(l — cos#) . (2.2.8)

The conjugate momentum is given by

= ^ml292 + mgl(l — cos9) . (2.2.10)

2.3 Bohr-Sommerfeld Quantization: SHO

pdq = nh — 2nnh . (2.3.12)

Also, the total energy is

The motion is bounded by ± x 0 = ±\/2E/k. The two signs in front of p indicate

2.4 Bohr-Sommerfeld: Particle in a B o x

pdq = nh = 2nnh . (2.4.18)

Again we have that the momentum is related to the energy by

Solving for E we find

2.5 Larmor Frequency

The potential energy due to placing the gyroscope in a magnetic field B is

Thus, the kinetic energy is

V =-p. • B = -ML • B = -M1Q •B . (2.5.24)

Then, the Lagrangian is

L = \ie2 - I' = l-lé2 + MI9B. (2.5.25)

— lié + MIB1 = 0. (2.5.26)

But, for no magnetic field (B = 0) we have the gyroscope at rest and we

9 = -MB = -lol . (2.5.28)

2.6 Applicability of Bohr-Sommerfeld

2.7 Schrodinger and Hamilton-Jacobi

A. Z. Capri-Problems & Solutions in Nonrelativistic Quantum Mechanics-World Scientific (2002)

A. Z. Capri-Problems & Solutions in Nonrelativistic Quantum Mechanics-World Scientific (2002)

(¥) = (**V 2 tf - l - V 2 * * ) + 2 i W W . (3.2.11)